Stress and addiction: contribution of the corticotropin releasing factor (CRF) system in neuroplasticity (2012)

Corticotropin releasing factor (CRF) system

CRF, also known as corticotropin releasing hormone (CRH), has been shown to induce various behavioral changes related to adaptation to stress. Dysregulation of the CRF system at any point can lead to a variety of psychiatric disorders such as depression, obsessive compulsive disorder, post-traumatic stress disorder and SUDs (Cole et al., 1990; Sarnyai et al., 1992, 2001; Cador et al., 1993; Koob and Kreek, 2007; Koob and Le Moal, 2008a). Footshock-induced stress has been shown to be effective in inducing reinstatement of alcohol (Le et al., 1998, 2000; Gass and Olive, 2007; Richards et al., 2008), nicotine (Buczek et al., 1999), cocaine (Erb et al., 1996), opiate and psycostimulants (Lu et al., 2003) and heroin (Shaham et al., 1997) seeking. Specifically CRF has been associated with drug reinstatement (Shaham et al., 1997; Le et al., 2002; Liu and Weiss, 2002; Funk et al., 2006). CRF has also been shown to produce anxiety-like behaviors during withdrawal from chronic ethanol (Baldwin et al., 1991; Overstreet et al., 2004) and may be responsible for persistent vulnerability and eventual relapse.

The CRF system consists of four ligands: CRF, urocortin (UCN) (Vaughan et al., 1995) 1, 2, and 3, two G-protein-coupled receptors (GPCR), CRF-receptor 1 (CRF-R1) and CRF-receptor 2 (CRF-R2), as well as a secreted CRF binding protein (CRF-BP); see Table Table11 and (Bale and Vale, 2004) for CRF system review.

It was originally identified as a hypothalamic factor responsible for stimulating adrenocorticotropic hormone (ACTH) secretion from the anterior pituitary (Guillemin and Rosenberg, 1955; Saffran et al., 1955) where it stimulates glucocorticoid synthesis and secretion form the adrenal cortex (Turnbull and Rivier, 1997). Its name was established thirty years before its biochemical identification in the 1980’s (Vale et al., 1981) while its gene identifier in the National Center for Biotechnology Information (NCBI) is CRH. It is a 4.7-kilo-Dalton (kDa) peptide and consists of 41-amino acid residues. Neurosecretory neurons of the paraventricular nucleus (PVN) of the hypothalamus synthesize CRF (Meloni et al., 2005). CRF is then released into the afferent portal blood vessels to the anterior pituitary gland where it induces ACTH release in the systemic circulation. The hypothalamic-pituitary-adrenal (HPA) axis is regulated by negative feedback from glucocorticoids that activate glucocorticoid receptors specifically in the PVN and hippocampus. CRF is also expressed outside the HPA axis to control autonomic and behavioral responses to stressors (Palkovits et al., 1983; Swanson et al., 1983) including stress-induced reinstatement of drug seeking.

CRF mediates physiological stress responses by activating CRF-R1 and CRF-R2, which are distributed throughout the periphery and the brain (De Souza, 1995; Bale and Vale, 2004). It is believed that the binding of CRF to CRF-Rs is a two-step mechanism. The N-terminus of the receptor initially binds to the C-terminus of CRF, which initiates a rearrangement of the receptor (Grace et al., 2007). The CRF N-terminus contacts the other sites on the receptor to initiate cellular signaling (Vale et al., 1981; Rivier et al., 1984) and consequently activate the G-protein (Nielsen et al., 2000; Grace et al., 2004; Rijkers et al., 2004; Yamada et al., 2004; Hoare, 2005). The CRF system comprises other peptides with structural homology to CRF. UCN 1 shows 45% sequence identity with CRF and binds with high affinity to both CRF receptor subtypes (Perrin et al., 1995), whereas CRF binds with highest affinity to CRF-R1 (Vaughan et al., 1995; Burnett, 2005). UCN 2, also known as stresscopin related peptide, and UCN 3, also known as stresscopin bind specifically to CRF-R2 (Hsu and Hsueh, 2001; Lewis et al., 2001; Reyes et al., 2001).

CRF-R1 has 415 amino acid residues and it is expressed in the periphery and in the CNS (Chang et al., 1993; Chen et al., 1993; Vita et al., 1993; Potter et al., 1994; Tsai-Morris et al., 1996; Sanchez et al., 1999; Van Pett et al., 2000). Chronic stress mediated by activation of CRF-R1 by CRF has been associated with the development of anxiety disorders (Arborelius et al., 1999); CRF-R1 antagonists have been shown to reduce anxiety-like behaviors (Funk et al., 2007). Transgenic mice with deletion of CRF-R1 (CRF-R1 knock out (KO) mice) have reduced reaction to both stress and anxiety, for comprehensive review see (Bale and Vale, 2004). This anxiolytic effect, however may be attributed to the reduction in circulating glucocorticoids in preclinical models (Tronche et al., 1999). A conditional KO mouse line was generated to differentiate the behavioral from the neuroendocrine CRF-mediated CRF-R1 signaling pathways. The selective inactivation of the limbic structures, but not of the HPA system has shown that CRF-R1 modulates anxiety-like behaviors and it is independent of the HPA (Muller et al., 2003). Furthermore, CRF-R1 is thought to increase susceptibility to alcohol relapse behaviors (Hansson et al., 2006; Heilig and Koob, 2007). A recent study evaluated the role of CRF both within and outside the HPA has shown that CRF via CRF-R1 signaling may have opposite effects on stress-related alcohol consumption (Molander et al., 2012).

CRF-R2 has three variants: α,β, and γ. The α is comprised of 411 amino acid residues and the β is comprised of 413–418 amino acid residues. Both are found in the brain and periphery; however, CRF-R2β is predominantly found in the heart and vasculature (Lovenberg et al., 1995a,b; Kimura et al., 2002; Burnett, 2005). The γ variant is a smaller peptide containing only 397 amino acid residues, and is found only in the human brain (Kostich et al., 1998). The precise role of CRF-R2 in the regulation of the stress response is a subject of intense investigation. Genetic mouse model studies with deletion of CRF-R2 (CRF-R2 KO mice) have demonstrated that CRF activation of CRF-R2 can lead to either an increased or decreased response to stressors (Bale et al., 2000, 2002; Coste et al., 2000; Kishimoto et al., 2000).

The lack of specific antisera that support immunohistochemical experiments and the low resolution of ligand binding approaches have limited the studies to elucidate the CRF-Rs distribution and limit the analysis at the mRNA level. To overcome this impediment, a transgenic mouse that reports expression of CRF-R1 with green fluorescent protein (GFP) has been successfully generated providing a novel tool to investigate the role of CRF-R1 signaling in stress adaptation (Justice et al., 2008).

CRF-BP is a water-soluble, 37 kD protein and consists of 322 amino acid residues (Bale and Vale, 2004). It is a secreted glycoprotein, efficiently stored into secretory granules and released into the extracellular space through exocytosis (Blanco et al., 2011). It contains aspargine N-linked-type oligosaccharides that are critical for CRF-BP binding to CRF (Suda et al., 1989). Previous attempts to identify small molecule inhibitors of CRF-BP have produced limited success due in part to the high affinity (picomolar) of CRF binding to CRF-BP (Behan et al., 1995a) and also because CRF-BP full length (FL) is susceptible to autocatalytic proteolysis (Woods et al., 1999). The spontaneous proteolytic cleavage yields a larger N-terminal fragment of 27 kD, CRF-BP (27 kD), which retains the binding site for CRF and a smaller, 9.6 kD C-terminal fragment, CRF-BP (10 kD) (Woods et al., 1999) with no apparent physiological or pathological role. The unique cleavage site in CRF-BP (FL) has been identified between amino acid residues serine 234 and alanine 235. The generation of two fragments has made it extremely difficult to successfully purify sufficient quantities of CRF-BP (FL) to study the physiological properties of the native protein. CRF-BP is distributed in plasma, amniotic and synovial fluid, the placenta, the pituitary gland, the liver, and in several distinct brain regions, including the cerebral cortex, the hippocampus (Behan et al., 1995a), the amygdala (Herringa et al., 2004) and the VTA (Wang and Morales, 2008). In the periphery, circulating CRF-BP neutralizes the physiological actions of CRF (Kemp et al., 1998). Because of the high affinity with CRF, it is believed that CRF-BP plays a buffer role by reducing the amount of free CRF. In the brain, however, CRF-BP is mostly membrane-bound and expressed in different amounts in neurons and neuroglial cells (Behan et al., 1995b). Within neuronal cells, recent findings demonstrated that discrete subpopulations of VTA dopaminergic and γ-aminobutyric acid (GABAergic) neurons express CRF-BP (Wang and Morales, 2008). The physiological role of CRF-BP in the central nervous system (CNS) is still unclear. Additionally, theories suggest the possibility that CRF-BP may assist the clearance of CRF from the body and may also protect CRF from degradation (Seasholtz et al., 2002). Genetic mouse model studies with deletion of CRF-BP (CRF-BP KO mice) have shown there is an increase in anxiety-like behavior (Karolyi et al., 1999). Electrophysiology studies have shown that CRF signals through CRF-R2 to potentiate N-Methyl-D-aspartate (NMDA)-mediated excitatory postsynaptic currents (EPSCs) in the VTA (Ungless et al., 2003). Furthermore, using CRF (6–33), a peptide that competes with CRF at the CRF-BP binding site, but does not bind to CRF-R2, it was shown that it blocked CRF-induced potentiation of NMDAR-mediated EPSCs (Ungless et al., 2003). Taken together, these results suggest that CRF-BP possesses a diverse role in modulating the CRF-system. As described by in vitro and in vivo studies, purifying human CRF-BP (FL) in sufficient quantities for investigation has not been successful to date (Woods et al., 1997). There have not been any research tools available to characterize the role of CRF-BP in the CNS by expressing CRF-BP on the cell surface. Therefore, it has not been possible to determine whether CRF-BP participates specifically in the CRF-R2 signaling. A summary of the involvement of the CRF binding in addictive behavior is described in Table Table22.

Cellular involvement of CRF in the amygdala

The amygdala is believed to be a pivotal brain region for emotional response and it is critical for providing affective salience to sensory information (Adolphs et al., 1994; LeDoux, 2003; Phelps and LeDoux, 2005). Negative affective responses have been studied in specific nuclei of the amygdala by studying the conditioned fear response (Davis, 1992a,b). The amygdala is widely connected to other limbic regions where it participates in integrating sensory and cognitive information (LeDoux, 1992, 1993). Experimental evidence strongly suggests drugs of abuse act on this system and can modify synaptic events especially during withdrawal. While the VTA has been associated with the reinforcing effects of ethanol (Gatto et al., 1994), the activation of the GABAergic system has been associated with alcohol’s anxiolytic effect (Frye and Breese, 1982). In addition to the rewarding circuits of the shell of the NAcc, and brain regions activated by pharmacological stressor, such as yohimbine and footshock were found to be specific in the basolateral and central amygdalar nuclei, and the bed nucleus of the stria terminalis (BNST) (Funk et al., 2006). Preclinical studies demonstrated that exposure and withdrawal from ethanol induces functional and biochemical changes in the amygdala of rats, demonstrating that this circuit is involved in long-term increases in anxiety-like behavior following chronic ethanol exposure (Christian et al., 2012).

The amygdala mediates conditioned and unconditioned responses to aversive stimuli (Davis and Whalen, 2001) and it has been investigated using Pavlovian fear conditioning by pairing a conditioned stimulus with an aversive unconditioned stimulus. The re-exposure of the unconditioned stimulus elicits a conditioned fear response derived by the conditioned-unconditioned association (Pitts et al., 2009). The association signal takes place in the basolateral amygdala (BLA) and is then transmitted to the central nucleus of the amygdala (CeA) (McDonald, 1998; Maren, 1999; Davis and Shi, 2000; Pitkanen et al., 2000; Pare et al., 2004). This transmission process involves both positive and negative associations.

All components of the CRF system, CRF, CRF-Rs and CRF-BP are expressed in the amygdala (Potter et al., 1994). Furthermore, the amygdala is a major extrahypothalamic source of CRF-containing neurons (Palkovits et al., 1983; Van Pett et al., 2000). Both BLA and CeA nuclei play a role in the stress response (Richter et al., 1995; Merali et al., 1998; Koob and Heinrichs, 1999). Extensive studies have shown that the CRF system participates in memory consolidation that involves the BLA-CeA circuit (Roozendaal et al., 2002; Hubbard et al., 2007). It has been observed that CRF release in the amygdala is increased during acute withdrawal (Richter and Weiss, 1999); therefore, it has been hypothesized that CRF may modulate drug-evoked synaptic plasticity (Ungless et al., 2001, 2003) and for a recent review, see (Luscher and Malenka, 2011). The neuronal basis for negative reinforcement is less well-understood; however, more recent behavioral studies have shown that CRF is capable of potentiating excitatory synaptic currents via CRF-R1 in the CeA two weeks following withdrawal from cocaine (Pollandt et al., 2006).

A recent study has shown that CRF-R1 specifically possess a bidirectional role in anxiety (Refojo et al., 2011). While deletion of CRF-R1 in the mid brain dopaminergic neurons increases anxiety-like behaviors and reduces dopamine release in the prefrontal cortex, deletion of CRF-R1 in the forebrain glutamanergic neuronal network reduces anxiety and disrupts transmission in the amygdala and hippocampus (Refojo et al., 2011).

The role of CRF was also evaluated extensively in voluntary ethanol consumption using gene expression and genetic variation in preclinical models see (Bjork et al., 2010) for extensive review. In ethanol-exposed animals, ethanol intake was reduced by administration of CRF-R1 antagonist, and tested using pharmacological interventions that reduce anxiety-like behaviors (Logrip et al., 2011; Zorrilla and Koob, 2012). The reduction of ethanol intake was also observed in transgenic mice with deletion of CRF-R1 (CRF-R1 KO) (Chu et al., 2007). CRF-R1 antagonists reduce drug withdrawal-associated anxiety and attenuate the negative reinforcing effects of ethanol associated with prolonged ethanol exposure (Ghitza et al., 2006; Marinelli et al., 2007; Li et al., 2007; Koob and Le Moal, 2008b; Richards et al., 2008). CRF-R1 inhibitors have shown to attenuate stress-induced relapse to cocaine and heroin in trained animals (Shaham et al., 1998) and to reduce stress-induced reinstatement and stress-induced reactivation of conditioned place preference in many addictive drugs (Koob and Zorrilla, 2010).

Conflict of interest statement

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.