Neurobiological mechanisms of addiction: Focus on corticotropin-releasing factor

Abstract

Drug addiction is a chronically relapsing disorder that is characterized by a compulsion to take drugs and loss of control in limiting intake. Medications that are on the market for the treatment of drug addiction target either the direct reinforcing effects of abuse (eg, naltrexone) or the consequent protracted abstinence syndrome (eg, acamprosate). Both conceptual and neurobiological advances in research have suggested that brain stress systems contribute to the withdrawal/negative affect and preoccupation/anticipation stages of the addiction cycle that promote the compulsivity of drug-taking in addiction. Validated animal models of the stress component of addiction and improved understanding of the neurocircuitry and neuropharmacological mechanisms involved in perturbations of this component suggest that corticotropin-releasing factor systems are a viable target for the development of future medications for drug addiction.

Introduction

Both conceptual and neurobiological advances in research have suggested that brain stress systems can contribute to the compulsivity of drug-taking and therefore participate in the development and persistence of addiction. Corticotropin-releasing factor (CRF) is a 41-amino acid polypeptide that mediates hormonal, autonomic and behavioral responses to stressors. This review outlines significant progress that has been achieved in understanding how perturbations of neuropharmacological mechanisms involving CRF can drive the addiction process. Based on this research, CRF systems are hypothesized to represent a particularly viable target for the development of medications for treating drug addiction in the withdrawal/negative affect and preoccupation/anticipation stages of the addiction cycle.

Mechanisms of addiction

Three neurobiological circuits have been identified to have heuristic value for the study of neurobiological changes associated with the development and persistence of drug dependence (Figure 1): a binge/intoxication-related circuit, a withdrawal/negative affect-related circuit, and a preoccupation/anticipation (ie, craving)-related circuit. Such mechanisms may be relevant to the use of CRF receptor antagonists in the treatment of drug addiction. The acute reinforcing effects of drugs of abuse that comprise the binge/intoxication stage most likely involve actions with an emphasis on the ventral striatum and extended amygdala reward system, as well as dopaminergic and opioid inputs from the ventral tegmental area (VTA) and arcuate nucleus of the hypothalamus, respectively [1,2]. In contrast, the symptoms of acute withdrawal that are important for addiction, such as dysphoria and increased anxiety associated with the withdrawal/negative affect stage, most likely involve decreases in the reward function of the ventral striatum, but also the recruitment of brain stress neurocircuitry, including CRF and norepinephrine in the extended amygdala. The preoccupation/anticipation stage also involves core elements of the brain stress systems in the extended amygdala (for stress-induced reinstatement), as well as key afferent glutamatergic projections to the extended amygdala and nucleus accumbens, specifically from the prefrontal cortex (for drug-induced reinstatement) and the basolateral amygdala (for cue-induced reinstatement) [1,2]. Compulsive drug-seeking behavior is also hypothesized to activate ventral striatal-ventral pallidal-thalamic-cortical loops that may subsequently engage dorsal striatal-pallidal-thalamic-cortical loops [3], with both effects influenced by concomitant decreases in reward function and the activation of brain stress systems in the extended amygdala [4]. The activation of the brain stress system by CRF that drives dependence and compulsivity in addiction is the conceptual framework for this review article.

Figure 1. The three stages of the addiction cycle and the criteria for substance dependence.

There are three stages in the addiction cycle: binge/intoxication, withdrawal/negative affect and preoccupation/anticipation; these stages are defined from a psychiatric perspective, with different criteria for substance dependence incorporated from the American Psychiatric Association's Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) [92].

(Adapted with permission from The American Association for the Advancement of Science and Koob GF, Le Moal M: Drug abuse: Hedonic homeostatic dysregulation. Science (1997) 278(5335):52-58. © 1997 The American Association for the Advancement of Science)

Corticotropin-releasing factor

The identification of CRF was followed by the discovery of genes encoding three paralogs of CRF (urocortin [Ucn]1, Ucn2 and Ucn3), as well as two GPCRs (CRF₁ and CRF₂) to which the CRF/Ucn peptides bind and activate with varying affinities [5,6]. Pharmacological and transgenic studies have demonstrated that brain and pituitary CRF₁ receptors mediate many of the functional stress-like effects of the CRF systems [7]. Consequently, numerous selective CRF₁ receptor antagonists have been designed that are capable of penetrating the blood-brain barrier [8]. Previous reviews have surveyed the biology of CRF systems [5,7].

CRF is widely distributed throughout the brain, with particularly high concentrations of cell bodies occurring in the paraventricular nucleus of the hypothalamus, the basal forebrain (notably the extended amygdala) and the brainstem [9]. The central administration of CRF mimics the behavioral response to activation and stress in rodents, and the administration of competitive CRF receptor antagonists generally has anti-stress effects [10–12] (for reviews, see references [13–16]). CRF₁ receptor activation is associated with increased stress responsiveness [17], and CRF₂ receptor activation is associated with decreases in feeding [18,19]. CRF₂ activation also may decrease stress responsiveness although there is some controversy in this area. The activation of CRF₂ receptors has also been implicated in increased stress responsiveness or no change in stress-related behavior depending on the neuropharmacological probe, the dose and the brain site used [6,20–22].

The role of CRF in animal models of addiction

The chronic administration of drugs with potential for dependence dysregulates the stress responses mediated by CRF, including the hypothalamic-pituitary-adrenal axis and the brain extrahypothalamic stress system. During acute withdrawal, certain responses are common for all drugs of abuse and alcohol, including an activated hypothalamic-pituitary-adrenal stress response, reflected by elevated ACTH and corticosteroids; and an activated brain stress response, with increased CRF release from the amygdala. However, with repeated cycles of addiction, a blunted hypothalamic-pituitary-adrenal response occurs, but with a sensitized extrahypothalamic CRF stress system response [2,23].

The neuroanatomical entity referred to as the extended amygdala may represent a common anatomical substrate for acute drug reward, as well as a common neuroanatomical substrate for the negative effects on reward function produced by stress that helps to drive compulsive drug administration. The extended amygdala includes not only the central nucleus of the amygdala, but also the bed nucleus of the stria terminalis and a transition zone in the medial subregion of the nucleus accumbens (shell of the nucleus accumbens) [24]; these regions share certain cytoarchitectural and circuitry similarities. A key neurochemical component of this circuitry that delineates the central division of the extended amygdala is a high density of cell bodies and terminals for CRF. The extended amygdala receives numerous afferent neurons from limbic (emotional) structures, such as the basolateral amygdala and hippocampus, and sends efferent neurons to the medial part of the ventral pallidum and a large projection of neurons to the lateral hypothalamus, thus further defining the specific brain areas that interface classic limbic structures with the extrapyramidal motor system and motivational systems [24].

In rats, in vivo microdialysis during acute withdrawal following the chronic administration or self-administration of drugs of abuse increased extracellular CRF in the extended amygdala. During alcohol withdrawal, extrahypothalamic CRF systems have been observed to become hyperactive, with an increase in extracellular CRF levels within the central nucleus of the amygdala and bed nucleus of the stria terminalis in alcohol-dependent rats [25,26]. Extracellular CRF in rats also increased in the central nucleus of the amygdala during precipitated withdrawal from chronic nicotine [27], withdrawal from binge cocaine self-administration [28], and precipitated withdrawal from opioids [29] and cannabinoids [30]. Amygdala CRF tissue content was reduced during acute withdrawal from ethanol exposure [31,32] and from binge cocaine self-administration [31]. CRF mRNA expression in the central nucleus of the amygdala also increased in rats during withdrawal from cannabinoids [33] or ethanol [34].

Another common response to acute withdrawal and protracted abstinence from all major drugs of abuse is the manifestation of a negative emotional state, including anxiety-like responses. Animal models in which the dependent variable is often a passive response to a novel and/or aversive stimulus, such as the open field, elevated plus maze, defensive withdrawal and social interaction tests, or an active response to an aversive stimulus, such as defensive burying of an electrified metal probe, have demonstrated anxiety-like responses to acute withdrawal from all major drugs of abuse [2]. Withdrawal from the repeated administration of cocaine, alcohol, nicotine, cannabinoids and benzodiazepines produces an anxiogenic-like response in the elevated plus maze, defensive withdrawal test and defensive burying test, and such effects are reversed by the administration of CRF antagonists [27,30,35–40]. Moreover, the aversive state of opiate withdrawal and the decreased brain reward function associated with nicotine withdrawal have both been demonstrated to be CRF₁ receptor-dependent [41–44].

The motivational effects of CRF in dependence can be observed in animal models of increased alcohol self-administration in alcohol-dependent animals. A competitive CRF peptide antagonist, which had no effect on alcohol self-administration in non-dependent rats, effectively eliminated excessive drinking in dependent or post-dependent rats [45]. When administered directly into the central nucleus of the amygdala, a CRF₁/CRF₂ receptor antagonist blocked alcohol self-administration in alcohol-dependent rats [32]. The direct administration of a CRF₂ agonist revealed an identical pattern of results, suggesting that, in the amygdala, the activation of CRF₂ systems may act in opposition to the activation of CRF₁ systems [46]. These data suggest an important role for CRF, primarily within the central nucleus of the amygdala, in mediating the increased self-administration that is associated with drug dependence.

Systemic injections of small-molecule CRF₁ antagonists also reduced the increased alcohol intake observed in dependent or post-dependent animals [47–53] (Table 1). The ability of CRF antagonists to block the anxiogenic-like and aversive-like motivational effects of drug withdrawal also predicts the motivational effects of such antagonists in animal models of extended access to drugs. Accordingly, CRF₁ small-molecule antagonists selectively reduced the increased self-administration of drugs during extended access to intravenous cocaine [54], nicotine [27] and heroin [55].

Table 1.

Comparison of the effects of CRF antagonists with naltrexone and acamprosate in alcohol addiction.

	Naltrexone	Acamprosate	CRF antagonist
Baseline drinking	↓ [93]	−[94]	– (antalarmin, MJL-1-109-2 and R-121919 [48]; LWH-63, MTIP and MPZP [47,49–53])
Binge drinking	↓ [95,96]	↓ [95]	– (MPZP [96])
Dependence-induced drinking	↓ [97]	↓ [98]	↓ (antalarmin, MJL-1-109-2 and R-121919 [48]; MTIP and MPZP [47,49–53])
Cue-induced reinstatement	↓ [99]	↓ [100]	– (d- Phe CRF12–41 [99])
Stress-induced reinstatement	− [99]	Not determined	↓ (d- Phe CRF12–41 [99])

Naltrexone and acamprosate, medications that are used to treat alcoholism, and CRF (corticotropin-releasing factor) antagonists were evaluated in various animal models, including rats, mice and primates, reflecting different components of the addiction cycle: binge drinking (binge/intoxication), dependence-induced drinking (withdrawal/negative affect), and cue- and stress-induced reinstatement (preoccupation/anticipation). Dashes (–) indicate no effect.

MPZP N,N-bis(2-methoxyethyl)-3-(4-methoxy-2-methylphenyl)-2,5-dimethyl-pyrazolo [1,5a]pyrimidin-7-amine, MTIP 3-(4-chloro-2-morpholin-4-yl-thiazol-5-yl)-8-(1-ethylpropyl)-2,6-dimethyl-imidazo[1,2-b]pyridazine

(Adapted with permission from George F Koob. © 2010 George F Koob)

Stress and stressors also have an established association with relapse and vulnerability to relapse [23,56]. In human alcoholics, numerous symptoms that can be characterized by negative affect persist after acute withdrawal from alcohol [57]. These symptoms, referred to as protracted abstinence, tend to be affective in nature, may last months to years, and often precede relapse [58,59]. A factor analysis of Marlatt's relapse taxonomy revealed that negative emotion, including elements of anger, frustration, sadness, anxiety and guilt, was a key factor in relapse [60], and was the leading precipitant of relapse in a large-scale replication of the taxonomy [61]. Exposure to negative affect, stress or withdrawal-related distress also increases drug craving [62–64].

The neuropharmacology of the stress-induced reinstatement of drug-seeking follows a pattern of results that is somewhat parallel to the neuropharmacology of the anxiety-like effects of acute withdrawal and dependence-induced increases in drug intake [65,66]. Mixed CRF₁/CRF₂ antagonists injected intracerebroventricularly and CRF₁ small-molecule antagonists administered systemically have been demonstrated to block the stress-induced reinstatement of cocaine-, opiate-, alcohol- and nicotine-seeking behavior [44,53,65–69]. These effects have been replicated with intracerebral injections of a mixed CRF₁/CRF₂ antagonist or a small-molecule CRF₁ antagonist into the bed nucleus of the stria terminalis, median raphe and VTA, but not the amygdala or nucleus accumbens [65,66]. The foot-shock-induced reinstatement of cocaine-seeking was blocked by the administration of a CRF₂ receptor antagonist into the VTA [70,71]. These observations suggest that sites such as the bed nucleus of the stria terminalis, median raphe and VTA may be important for stress-induced relapse, in contrast to the role of CRF in dependence-induced increases in drug self-administration that have been localized to the central nucleus of the amygdala [32]. CRF systems have also been identified in the VTA, and foot-shock stress can release CRF into the VTA [72,73]. Further supporting a role for VTA CRF systems in stress-induced relapse was the observation that the intra-VTA infusion of CRF elicited the reinstatement of cocaine-seeking behavior in rats [73]. The mechanism of reinstatement may involve the activation of the mesolimbic dopamine system; the application of CRF to a VTA slice preparation facilitated excitatory glutamatergic responses of dopamine neurons [70], a potentiation that was greater in cocaine-experienced animals compared with drug-naïve animals [74]. VTA CRF-mediated facilitation of dopamine transmission and reinstatement behavior may also involve the CRF binding protein, a protein that binds free CRF with high affinity [75]. CRF_6–33, a fragment without known receptor agonist activity that competitively displaces CRF from the CRF binding protein, prevented CRF from stimulating VTA dopamine transmission or from eliciting cocaine-seeking in rats [70,71]. These results complement the role of the CRF₁ system in the extended amygdala in stress-induced reinstatement [68].

Stress-induced reinstatement occurs independently from stress-induced activation of the hypothalamic-pituitary-adrenal axis [67,76,77]. Other brain stress systems that are implicated in stress-induced reinstatement and are possibly linked to brain CRF systems include norepinephrine, orexin, vasopressin and nociceptin [65,66] (Figure 2). CRF₁ antagonists are ineffective in blocking cue-induced reinstatement, suggesting the selectivity of these actions on the stress components of the addiction cycle [2]. CRF antagonists failed to block yohimbine-induced reinstatement of cocaine seeking [78], but were capable of reversing yohimbine-induced reinstatement of alcohol and food seeking [69,79], suggesting a specific interaction with cocaine (which also blocks the reuptake of norepinephrine) that is not observed with other drugs of abuse. The hypothesis is not that CRF₁ antagonists will reverse all craving but, rather, that CRF₁ antagonists may be effective in reversing both state- (protracted abstinence) and stressor-induced reinstatement of drug seeking. Substantial evidence supports the hypothesis that some form of a stress-like (negative emotional) state or stressor exposure contributes to the majority of relapses in humans [60,61].

Figure 2. Neurocircuitry associated with the acute positive reinforcing effects of drugs of abuse and the negative reinforcement of dependence, and changes in the transition from non-dependent to dependent drug-taking.

(A) Key elements of the reward circuit include dopamine and opioid peptide neurons that intersect at both the VTA (ventral tegmental area) and the nucleus accumbens that are activated during initial drug use and the early binge/intoxication stage of the addiction cycle. Drugs approved for the treatment of this component of the addiction cycle include naltrexone, buprenorphine, varenicline and disulfiram. (B) Key elements of the stress circuit include CRF (corticotropin-releasing factor) and noradrenergic neurons that converge on GABA interneurons in the central nucleus of the amygdala that are activated during the development of dependence. Drugs approved for the treatment of this component of the addiction cycle include methadone, bupropion, acamprosate, varenicline and buprenorphine.

DA dopamine, NE norepinephrine

(Adapted with permission from Nature Publishing Group and Nestler EJ: Is there a common molecular pathway for addiction? Nat Neurosci (2005) 8(11):1445-1449. © 2005 Nature Publishing Group)

Thus, brain stress systems may have an impact on both the withdrawal/negative affect and preoccupation/anticipation stages of the addiction cycle, albeit by engaging different components of the extended amygdala emotional system (ie, the central nucleus of the amygdala versus the bed nucleus of the stria terminalis). The dysregulations that comprise the negative emotional state of drug dependence persist during protracted abstinence to promote vulnerability to craving by activation of the drug-, cue-and stress-induced reinstatement neurocircuits, driven by a hypofunctioning and possibly reorganized prefrontal system [80].

No laboratory studies or clinical trials have been initiated to explore the effects of CRF antagonists on drug dependence in humans. However, several clinical trials that explored the effects of CRF₁ antagonists, including ONO-2333Ms, CP-316311 and R-121919 (NBI-30775), on anxiety and depression measures have been completed. Double-blind, placebo-controlled trials of ONO-2333Ms (ClinicalTrials.gov identifier: NCT00514865) and CP-316311 (NCT00143091) for major depression have yielded negative results [81], in contrast to the results of an earlier open-label trial for R-121919 [82,83]. However, in all trials, the CRF₁ antagonists were well tolerated, and no common adverse effects were observed [81,84–86].

Given the negative results that have been observed for CRF₁ antagonists in the treatment of depression [84–86], prospects for the efficacy of CRF₁ antagonists in addiction remain unknown. However, addiction is not necessarily equated with depression. Antidepressants generally demonstrate poor efficacy in the treatment of addiction. Indeed, the emotional allostatic state that has been argued to occur in addiction appears to have more similarity with the anxiety-dysphoria continuum. In addition, given that CRF has a prominent role in 'fight or flight' responses in animals, and that CRF antagonists have a mixed efficacy in classic animal models of depression, better therapeutic indications for CRF₁ antagonists may be stress, anxiety and the stress-like components associated with addiction, rather than depression.

Pharmacogenetic interactions with the CRF₁ receptor

Genetic markers may ultimately predict the individuals who will respond to CRF₁ antagonists, and association studies have identified a potential genetic interaction. An association has been observed between SNPs of the CRF₁ receptor gene and excessive drinking in humans and rats [87–90]. One SNP observed in humans, R1876831, is located in an intron, and may influence the transcription of the CRF₁ receptor gene. Adolescents who were homozygous for the C-allele of R1876831 drank more alcohol per occasion, and demonstrated higher lifetime rates of heavy drinking related to negative life events, compared with individuals carrying the T-allele [87].

In preclinical studies, a genetic polymorphism in the corticotropin-releasing hormone receptor 1 (Crhr1) promoter was identified in Marchigian Sardinian alcohol-preferring (msP) rats, which express a phenotype of elevated CRF₁ mRNA expression in the extended amygdala, increased anxiety-like behavior, and CRF₁ antagonist-reversible elevations in alcohol intake and alcohol-seeking behavior [88]. Expression of the Crhr1 transcript was downregulated following ad libitum access to alcohol in alcohol-preferring msP rats [89]; however, the expression of CRF and the Crhr1 transcript within the amygdala were upregulated in post-dependent animals [52]. Similarly, a Crh haplotype predicted alcohol consumption in rhesus macaques [90]. Altogether, these results suggest that certain SNPs in the human population may predict vulnerability to particular subtypes of excessive-drinking syndromes, and may also predict responsiveness to the use of CRF receptor antagonists in the treatment of alcoholism.

Conclusion

The extrahypothalamic CRF systems appear to be involved in mediating the anxiety-like effects of acute withdrawal from drugs of abuse, the increase in drug-taking associated with dependence, and stress-induced reinstatement for all major drugs of abuse, including psychostimulants, opioids, alcohol, nicotine and (in limited studies) cannabinoids. Many of these effects have been localized to the extended amygdala, and acute withdrawal from major drugs of abuse was demonstrated to increase CRF release in the central nucleus of the amygdala, as measured by in vivo microdialysis.

The majority of studies have implicated CRF and the CRF₁ receptor in stress responsivity [84], although CRF₂ receptors and Ucn/CRF₂ ligands have also been demonstrated to have both stress and anti-stress-like effects [6]. The results suggest a major role for CRF systems in mediating the negative emotional states that have motivational significance in maintaining the drug-dependent state [4,91]. Pharmacogenetic studies suggest that certain CRF system-related SNPs may predict vulnerability to the excessive drinking that is associated with stress, and thus may also predict the efficacy of CRF₁ antagonists in addiction. Overall, these results suggest that the CRF system is a significant factor in the withdrawal/negative affect and preoccupation/anticipation stages of the addiction cycle, and pharmacotherapeutic agents directed at the CRF system may be effective in treating some key aspects of addiction.