The endogenous opioid system in cocaine addiction: what lessons have opioid peptide and receptor knockout mice taught us?

Abstract

Cocaine addiction has become a major concern in the UK as Britain tops the European ‘league table’ for cocaine abuse. Despite its devastating health and socio-economic consequences, no effective pharmacotherapy for treating cocaine addiction is available. Identifying neurochemical changes induced by repeated drug exposure is critical not only for understanding the transition from recreational drug use towards compulsive drug abuse but also for the development of novel targets for the treatment of the disease and especially for relapse prevention. This article focuses on the effects of chronic cocaine exposure and withdrawal on each of the endogenous opioid peptides and receptors in rodent models. In addition, we review the studies that utilized opioid peptide or receptor knockout mice in order to identify and/or clarify the role of different components of the opioid system in cocaine-addictive behaviours and in cocaine-induced alterations of brain neurochemistry. The review of these studies indicates a region-specific activation of the µ-opioid receptor system following chronic cocaine exposure, which may contribute towards the rewarding effect of the drug and possibly towards cocaine craving during withdrawal followed by relapse. Cocaine also causes a region-specific activation of the κ-opioid receptor/dynorphin system, which may antagonize the rewarding effect of the drug, and at the same time, contribute to the stress-inducing properties of the drug and the triggering of relapse. These conclusions have important implications for the development of effective pharmacotherapy for the treatment of cocaine addiction and the prevention of relapse.

Introduction

Cocaine is a psychostimulant whose use in the European Union has dramatically increased over the last 10 years with Britain topping the European ‘league table’ for cocaine abuse (European Monitoring Center for Drugs and Drug Addiction, 2011). It is estimated that nearly 1 million people used cocaine last year in England and Wales alone, and the number of cocaine addicts aged 18–24 in treatment has doubled in the last 3 years in the UK (UNODC, 2011). When considering the physical danger, addictive liability and social harm of all drugs of abuse, cocaine is especially destructive, costing the UK tax payer billions of pounds a year on productivity loss, criminal activity and on social and medical care. Despite this, and although a number of potential targets have been identified (Heidbreder and Hagan, 2005), currently, there is no specific pharmacological therapy with established efficacy for the treatment of cocaine addiction and for the prevention of relapse (Kreek et al., 2002; Somaini et al., 2011).

Cocaine is well known to increase extracellular levels of monoamines by blocking monoamine transporters (Heikkila et al., 1975). Elevation of extracellular dopamine (DA) levels in the mesolimbic dopaminergic system, which is composed primarily of dopaminergic neurons projecting from the ventral tegmental area (VTA) to the nucleus accumbens (NAc) has been suggested to play a central role in the reinforcing effect of the psychostimulant, although other brain structures such as the hypothalamus (Hy), bed nucleus of the stria terminalis (BNST), septum (Sept), hippocampus (Hi), caudate putamen (CPu), thalamus (Th), amygdala (Amy) and frontal cortex (FCx) have also been implicated (Koob and Le Moal, 2001; Koob and Kreek, 2007; Le Moal and Koob, 2007). The acute positive reinforcement (i.e. rewarding, euphoric) effect of the drug is thought to involve the activation of the VTA-NAc (reward), BNST and Amy (emotional learning) circuitries (Le Moal and Koob, 2007). The negative withdrawal symptoms that drive drug administration (negative reinforcement) following repeated drug use (dependence state) involve suppression of the aforementioned circuitries and the recruitment of brain stress pathways, namely the Hy and Amy. In terms of relapse, the FCx, involved in impulse control and decision-making, and the basolateral Amy (BLA) have been implicated in drug priming and cue priming reinstatement respectively (Le Moal and Koob, 2007). In addition, other regions such as the Hi (contextual memory) and CPu (habit learning) have been shown to be involved in the transition from recreational drug use to compulsive cocaine abuse, which is characteristic of addiction (Everitt and Robbins, 2005; Le Moal and Koob, 2007). Opioid peptides and receptors are expressed in high density in areas of the reinforcement circuitries and play a key role in their regulation (for extensive review, see Le Merrer et al., 2009).

Three classical opioid receptors, the µ, δ and κ have been identified (see Goldstein and Naidu (1989) and the genes encoding them have been cloned (δ by Evans et al., 1992; Kieffer et al., 1992; Yasuda et al., 1993; µ by Chen et al., 1993; Thompson et al., 1993; Wang et al., 1993; and κ by Meng et al., 1993; Minami et al., 1993; Yasuda et al., 1993; for review, see Corbett et al., 2006; Kieffer 1995) and are referred as µ-opioid (MOP), δ-opioid (DOP) and κ-opioid (KOP) receptors respectively (Borsodi et al., 2011). Endogenous peptides that target opioid receptors have been identified and characterized. Notably, the enkephalins, dynorphins and β-endorphins (ends) are produced by proteolytic cleavage of large protein precursors known as preproenkephalin (PENK), preprodynorphin (PDYN) and proopiomelanocortins (POMC) respectively. Genes encoding these peptide precursors have been cloned (PENK by Gubler et al., 1982; Noda et al., 1982; POMC by Nakanishi et al., 1979; PDYN by Kakidani et al., 1982; Rossier, 1982; see review by Kieffer, 1995) and mice lacking MOP receptor (MOPr) (Matthes et al., 1996; Sora et al., 1997; Tian et al., 1997; Loh et al., 1998; Schuller et al., 1999), DOP receptor (DOPr) (Zhu et al., 1999; Filliol et al., 2000), KOP receptor (KOPr) (Simonin et al., 1998; Ansonoff et al., 2006) or a combination of opioid receptors (MOPr/DOPr, MOPr/KOPr, KOPr/DOPr, MOPr/KOPr/DOPr) (Simonin et al., 2001) as well as mice lacking PDYN (Sharifi et al., 2001; Zimmer et al., 2001), PENK (Konig et al., 1996; Ragnauth et al., 2001) and β-end (Rubinstein et al., 1996) have been generated by homologous recombination by several groups (for review, see Gaveriaux-Ruff and Kieffer, 2002; Kieffer and Gaveriaux-Ruff, 2002).

There is a large body of evidence demonstrating that the endogenous opioid system plays a key role in regulating mood and reward and is central in modulating addictive behaviours. In this context, it is generally accepted that systemic and local region-specific administration of MOPr, and to a lesser extent DOPr, agonists stimulate positive reinforcement, whereas KOPr agonists inhibits positive reinforcement and induces aversion and dysphoria. While there is compelling evidence that the aversive effects of KOPr are mediated via the suppression of DA release in the NAc (Shippenberg and Elmer, 1998; Van Ree et al., 2000), the role of DA in mediating the reinforcing effects of MOPr agonists is less clear. While MOPr and DOPr stimulation induce DA release in the striatum (Di-Chiara and Imperato, 1988; Fusa et al., 2005), at least in animal models, evidence (Robinson and Berridge, 1993; Daglish et al., 2008) suggests that DA may not be critical in mediating the ‘high’ from opioids in opioid-dependent subjects but the ‘drug wanting’ instead (Daglish et al., 2008). MOPr and DOPr antagonists have direct aversive-anxiogenic effects and can also suppress the positive reinforcing properties of natural rewards (see review by Colasanti et al., 2011), whereas KOPr antagonists has been shown to facilitate these effects (Van Ree et al., 2000).

Here, we review the evidence demonstrating a dysregulation of different components of the opioid system, which are likely to contribute towards the transition from recreational drug use to compulsive cocaine abuse, which is characteristic of addiction. We focus on the profound effects of chronic cocaine exposure and withdrawal on the opioid receptor (at the mRNA, protein and activity level) and peptide system, primarily in rodent models. This review does not include findings obtained by microarray experiments as this is reviewed elsewhere (Yuferov et al., 2005). Furthermore, we review the studies utilizing opioid peptide and receptor knockout (KO) mice in order to identify and/or clarify the role of different components of the opioid system in cocaine-addictive behaviours and in cocaine-induced alterations of brain neurochemistry. This review will merely focus on the receptor and peptide components of the classical opioid system and does not include endomorphins, nociceptin, nocistatin and opioid receptor-like 1.

Regulation of opioid receptor and peptide system by chronic cocaine exposure and withdrawal

Opioid receptor regulation by cocaine

Regulation of MOPr

There is large body of evidence indicating a region-specific activation of the MOPr system following chronic cocaine exposure, which persists for a long time after withdrawal from the drug (Table 1). Several brain regions have been reported to be sites of MOPr regulation by cocaine. The most consistent finding is the increase of MOPr mRNA (Azaryan et al., 1996a,b; 1998; Walters et al., 2005; Leri et al., 2006), binding (Hammer, 1989; Unterwald et al., 1992; 1994; Izenwasser et al., 1996; Azaryan et al., 1996a,b; 1998; Unterwald, 2001) and MOPr activation as measured both by (D-Ala², N-MePhe⁴, Gly-ol)-enkephalin (DAMGO), stimulated 5′-(γ-[³⁵S]thio)-triphosphate ([³⁵S]GTPγS) autoradiography (Schroeder et al., 2003; Bailey et al., 2007b) and by DAMGO inhibition of AC activity (Izenwasser et al., 1996; Schroeder et al., 2003; Bailey et al., 2007b) in the NAc following chronic cocaine exposure. This up-regulation of MOPr did not depend on the administration paradigm and is consistent in animals that have been exposed to continuous cocaine administration (mini-pumps) or repeatedly injected in a ‘binge’ paradigm or following conditioned place preference (CPP) paradigm (see Table 1). Interestingly, an increase in MOPr mRNA levels (Yuferov et al., 1999) but a decrease in MOPr occupancy (Soderman and Unterwald, 2009) was observed in the NAc following acute cocaine administration in rats, demonstrating dissociation between mRNA and protein levels. However, acute or chronic withdrawal from 14 day ‘binge’ cocaine administration did not show alterations in MOPr mRNA (Bailey et al., 2005b) or MOPr binding (Bailey et al., 2005a), respectively, in the NAc, demonstrating that these alterations do not persist after the discontinuation of the drug at least in the NAc.

Table 1.

Effect of chronic cocaine exposure and withdrawal on opioid peptide and receptor expression, protein and activity

Opioid system	Administration paradigm	Brain region	mRNA	Protein (binding)	Functional activity	References
MOPr	Acute	FCx, NAc, Amy	↑			(Yuferov et al. 1999)
		Hi, Th, Hy	↔
		NAc		↓		(Soderman and Unterwald, 2009)
		CPu, Tu, CgCx, MCx, VTA		↔		(Soderman and Unterwald, 2009)
	Chronic (repeated)	Amy	↔			(Zhou et al., 2005)
		CgCx, FCx, NAc, CPu, BLA		↑	↑*	(Unterwald et al., 1992; 1994; 2001; Schroeder et al., 2003; Bailey et al., 2005a; 2007b)
		NAc, CPu			↔#	(Unterwald et al., 1993)
		OT, Den, MS, VP, VDB, Hy, Hb, Hi, SN, VTA		↔		(Unterwald et al., 1992; Bailey et al., 2005a)
	Chronic (continuous)	NAc	↑	↑	↑#	(Hammer, 1989; Izenwasser et al., 1996; Azaryan et al., 1996a,b; 1998)
		OB, CPu	↔	↔	↔#	(Izenwasser et al., 1996; Azaryan et al., 1996b)
		VP, LH, MG, IC		↑		(Hammer, 1989)
		SN, VTA, DR, BLA		↓		(Hammer, 1989)
	Acute withdrawal	FCx	↑			(Bailey et al., 2005b)
		CPu, NAc, Amy, Hy	↔			(Bailey et al., 2005b; Zhou et al., 2005)
	Chronic withdrawal	FCx, CgCx, CPu		↑		Bailey et al., 2005a)
		NAc, OT, Den, Sept, VDB, Th, Amy, Hy, Hi,		↔		(Bailey et al., 2005a)
		Limbic Cx, NAc, VDB		↓		(Sharpe et al., 2000)
	CPP	NAc, FCx	↑			(Walters et al., 2005; Leri et al., 2006)
		CPu, Amy, Hy	↔			(Leri et al., 2006)
	Human addicts	CPu, Th, CgCx, FCx, TCx		↑$		(Zubieta et al., 1996; Gorelick et al., 2005; 2008)
		CPu		↓		(Hurd and Herkenham, 1993)
DOPr	Chronic (repeated)	CgCx, OT, NAc, CPu, BLA, SN, VTA		↔	↔*	(Unterwald et al., 1994; Schroeder et al., 2003)
		NAc, CPu			↓#	(Unterwald et al., 1993)
	Chronic (continuous)	NAc, CPu, OB	↔		↔#	(Izenwasser et al., 1996; Azaryan et al., 1996b)
	Acute withdrawal	FCx, Nacb, CPu			↓#	(Perrine et al., 2008)
KOPr	Acute	SN	↔			(Spangler et al., 1997)
	Chronic (repeated)	CPu	↔			(Spangler et al., 1996a; Unterwald et al., 2001)
		NAc, VTA, SN	↓			(Spangler et al., 1996a; 1997; Rosin et al., 1999)
		CgCx, OT, CPu, VTA, Sept		↑		(Unterwald et al., 1994; 2001; Bailey et al., 2007a)
		NAc, BLA, SN		↔		(Unterwald et al., 1994)
		NAc, CPu, CgCx			↔*	(Schroeder et al., 2003)
	Chronic (continuous)	NAc, BLA, End, Sept		↑		(Collins et al., 2002)
	Chronic withdrawal	SN, CPu	↔			(Spangler et al., 1996a)
		NAc, CPu, Septum, BLA		↓		(Turchan et al., 1998; Bailey et al., 2007a)
	Human addicts	NAc, CPu, Amy		↑		(Hurd and Herkenham, 1993; Staley et al., 1997; Mash and Staley, 1999)
PENK	Acute	CPu, Cx	↑			(Daunais and McGinty, 1995; Przewlocka and Lason, 1995)
		Amy	↔			(Turchan et al., 2002)
	Chronic (repeated)	NAc, CPu	↑			(Hurd et al., 1992; Steiner and Gerfen, 1993; Przewlocka and Lason, 1995; Spangler et al., 1996b; 1997; Mathieu-Kia and Besson, 1998; Svensson and Hurd, 1998; Crespo et al., 2001)
		NAc, CPu	↔			(Branch et al., 1992; Hurd et al., 1992; Daunais and McGinty, 1995; Spangler et al., 1996a; Mathieu-Kia and Besson, 1998; Alvarez Fischer et al., 2001; Bailey et al., 2005b; Ziolkowska et al., 2006)
		Pir, Tu	↑			(Crespo et al., 2001)
		FCx, Cx, Amy, Hy, Pit, CG, Cer	↔			(Branch et al., 1992; Daunais and McGinty, 1994; Turchan et al., 2002; Bailey et al., 2005b; Ziolkowska et al., 2006)
	Acute withdrawal	CPu, NAc,	↓			(Przewlocka and Lason, 1995; Crespo et al., 2001)
		Amy	↓			(Crespo et al., 2001)
		CPu, Pir, Tu, NAc	↑			(Crespo et al., 2001)
		FCx, NAc, CPu	↔			(Spangler et al., 1996a,b; Svensson and Hurd, 1998; Arroyo et al., 2000; Bailey et al., 2005b)
	Chronic withdrawal	CPu, Tu, Pir, NAc	↑			(Crespo et al., 2001)
		Amy, VMN	↓			(Crespo et al., 2001)
		CPu, NAc	↔			(Svensson and Hurd, 1998)
	Human addicts	CPu	↓			(Hurd and Herkenham, 1993)
β-end	Acute	NAc		↑&		(Olive et al., 2001)
		Pit		↑T		(Moldow and Fischman, 1987)
	Chronic (repeated)	NAc		↑&		(Roth-Deri et al., 2003; 2006)
		Hy	↔			(Zhou et al., 2005)
		Pit, plasma		↑T		(Forman and Estilow, 1988)
	Acute Withdrawal	Hy	↔			(Forman and Estilow, 1988; Zhou et al., 2005)
	CPP	Hy	↔			(Leri et al., 2006)
PDYN	Human addicts	plasma		↑T		(Vescovi et al., 1992)
	Acute	CPu	↑			(Spangler et al., 1997; Zhou et al., 2002)
	Chronic (repeated)	CPu	↑	↑T		(Sivam, 1989; Smiley et al., 1990; Hurd et al., 1992; Spangler et al., 1993; 1996a; 1997; Steiner and Gerfen, 1993; Daunais and McGinty, 1995, 1996; Romualdi et al., 1996; Mathieu-Kia and Besson, 1998; Svensson and Hurd, 1998; Werme et al., 2000; Yuferov et al., 2001; Bailey et al., 2005b; Schlussman et al., 2005; Ziolkowska et al., 2006)
		NAc, DG, SN	↑	↑T		(Smiley et al., 1990; Hurd et al., 1992; Mathieu-Kia and Besson, 1998; Turchan et al., 2002)
		FCx, Cx, NAc, Hy, Amy, SN	↔			(Spangler et al., 1993; Alvarez Fischer et al., 2001; Yuferov et al., 2001; Turchan et al., 2002; Bailey et al., 2005b; Schlussman et al., 2005; Ziolkowska et al., 2006; Zhou et al., 2008)
	Chronic (continuous)	CPu	↑			(Romualdi et al., 1996)
		Hy	↓			(Romualdi et al., 1996)
		Hi, NAc	↔			(Romualdi et al., 1996)
	Acute withdrawal	CPu, LH, SN	↑	↑T		(Sivam, 1989; Smiley et al., 1990; Bailey et al., 2005b; Zhou et al., 2008)
	Chronic withdrawal	CPu, NAc	↓			(Svensson and Hurd, 1998)
		CPu, LH	↔	↔T		(Sivam, 1989; Spangler et al., 1996a; Zhou et al., 2008)
	Human addicts	CPu, VP	↑			(Hurd and Herkenham, 1993; Frankel et al., 2008)

^*As measured by DAMGO stimulated [³⁵S]GTPγS autoradiography.

^#As measured by DAMGO inhibition of AC activity.

^&As measured by microdialysis for β-endorphin release.

^$As measured by positron emission tomography.

^TPrecursor peptide fragment immunoreactivity.

MCx, motor cortex.

The profile observed for MOPr in the NAc was not reproduced in the CPu. Although no changes in MOPr mRNA and density was observed in the CPu of animals treated with a continuous cocaine administration paradigm (Izenwasser et al., 1996; Azaryan et al., 1996b), or in animals undergoing cocaine CPP (Leri et al., 2006) or after acute withdrawal (Bailey et al., 2005b), an increase in MOPr binding (Unterwald et al., 1992; 1994; 2001; Bailey et al., 2005a) and activity (Schroeder et al., 2003; Bailey et al., 2007b) was observed following chronic ‘binge’ cocaine administration, which persisted for a long time after the withdrawal of the drug (Bailey et al., 2005a). Surprisingly, no alterations in DAMGO inhibition of adenylate cyclase activity was observed following continuous (Izenwasser et al., 1996) or chronic ‘binge’ (Unterwald et al., 1993) cocaine administration in the CPu. The lack of association between mRNA and protein receptor level in the CPu suggests that mechanisms other than increased gene expression are responsible for the regulation of MOPr by cocaine in the CPu. Indeed, mechanisms such as decreased degradation of the receptor, increased recycling or/and changes in the levels of the endogenous ligand for the receptors could account for those alterations of MOPr protein and activity. In this regard, although cocaine has been shown to increase β-end in the NAc in rodents (Olive et al., 2001; Roth-Deri et al., 2003; 2006; 2008), β-end deficiency in regulating hypothalamic–pituitary–adrenal (HPA) axis activity has also been suggested based on a study demonstrating hyper-responsivity to removal of glucocorticoid negative feedback in cocaine-abstinent human individuals following the administration of metyrapone (Schluger et al., 1998; 2001). In addition, there is evidence based on rodent studies suggesting a D₂ receptor (Soderman and Unterwald, 2009) and dynorphin (Bailey et al., 2007b)-dependent mechanism in the regulation of MOPr by cocaine.

The FCx and cingulate cortex (CgCx) were also shown to be regions where MOPr expression, binding and activity were profoundly affected by cocaine exposure and withdrawal. MOPr mRNA levels were shown to be increased in the FCx and CgCx following acute cocaine administration (Yuferov et al., 1999), cocaine CPP (Leri et al., 2006) and acute withdrawal from chronic cocaine administration (Bailey et al., 2005b). In addition, although MOPr binding was unchanged in the CgCx following acute cocaine administration (Soderman and Unterwald, 2009), MOPr binding was elevated in the FCx/CgCx following chronic ‘binge’ (both steady and escalating dose) cocaine administration (Unterwald et al., 1992; 1994; Unterwald, 2001; Bailey et al., 2005a), which persisted for a long time after the withdrawal of the drug (Bailey et al., 2005a). Fourteen-day ‘binge’ cocaine also increased MOPr activity in the CgCx (Schroeder et al., 2003; Bailey et al., 2007b). Together, these studies clearly indicate an activation of the MOPr system in the FCx, which persists for a long time after the discontinuation of the drug.

Some inconsistencies between mRNA and protein levels and between administration paradigm have been observed in the regulation of MOPr in other brain regions. For instance, while at the mRNA level, there was no modification of MOPr in the Amy following chronic cocaine administration (Zhou et al., 2005), acute withdrawal (Zhou et al., 2005) and CPP (Leri et al., 2006); increased levels of MOPr binding was observed in the BLA following chronic ‘binge’ administration paradigm (Unterwald et al., 1992; 1994; Unterwald, 2001), which did not persist following 14 day withdrawal period (Bailey et al., 2005a). In contrast, decreased levels of MOPr binding were found following continuous cocaine administration (Hammer, 1989) in the BLA. Again, these inconsistencies probably suggest that mechanisms other than alterations of gene expression contribute towards the regulation of MOPr by cocaine in the Amy and that this depends on the drug administration paradigm.

No alterations of MOPr mRNA levels were observed in the olfactory bulbs (Azaryan et al., 1996b) and no modification of MOPr binding was observed in the olfactory tubercle (OT), ventral pallidus (VP), substantia nigra (SN), VTA, dorsal endopiriform (Den), medial Sept, Hy, Th, vertical limb of the diagonal band, habenula and Hi following chronic cocaine administration (Unterwald et al., 1992; Bailey et al., 2005a) and following long-term withdrawal from chronic cocaine administration (Bailey et al., 2005a). No modification of MOPr mRNA regulation in Hy following CPP (Leri et al., 2006) and withdrawal (Zhou et al., 2005) was observed. MOPr binding was, however, increased in the VP, lateral Hy (LH), medial geniculate, inferior colliculus but decreased in the SN, VTA, dorsal raphe following chronic continuous cocaine administration (Hammer, 1989). These findings indicate a regulatory role for cocaine on the MOPr system in a large number of brain regions, at least at the protein level.

Similar alterations in MOPr were also reported in brains of human cocaine addicts. Increased levels of MOPr availability measured by positron emission tomography were also observed in cocaine abstinent human addicts in the CPu, Th, CgCx, FCx, temporal cortex (Zubieta et al., 1996; Gorelick et al., 2005), demonstrating the translational relevance of this cocaine regulation of the MOPr system in humans as well as the translational reliability of the animal model of chronic ‘binge’ cocaine administration paradigm not only in terms of mimicking the human pattern of cocaine administration but also in mimicking the neurochemical consequences of chronic cocaine abuse (see also Bailey et al., 2008). It is not clear if the increased availability of MOPr in humans represents a decrease in β-end release in addition to an increase of MOPr in abstinent human individuals. In rodents, although the increase of β-end release in response to acute and chronic cocaine administration has been demonstrated at least in the striatum (Olive et al., 2001; Roth-Deri et al., 2003; 2008), the release of β-end in withdrawn individuals remains to be explored. Increased levels of MOPr availability in frontal and temporal cortical regions was shown to be significantly positively correlated to cocaine craving (Gorelick et al., 2005) and time of relapse to cocaine use (Gorelick et al., 2008) in human cocaine addicts and high MOPr concentration in the PFx, CgCx, NAc and BLA has been shown to be correlated with high impulsivity in humans (Love et al., 2009). Together, these findings clearly demonstrate a region-specific activation of the MOPr system by chronic cocaine exposure at least in the NAc, CPu, FCx, CgCx and Amy, which persists for a long time after the discontinuation of the drug at least in the CPu, FCx and CgCx. Given the role of these regions in decision-making (FCx), impulsiveness (FCx, CgCx, NAc, BLA), emotional learning (Amy) and reinforcement (NAc, CPu, Amy), the previous findings indicate an important role for MOPr in mediating the ‘drug wanting’ effect of cocaine during cocaine administration and craving of the drug during withdrawal followed by relapse.

Regulation of DOPr

No modification of DOPr mRNA (Azaryan et al., 1996b), receptor binding levels (Unterwald et al., 1994) and DOPr activity as measured by D-penicillamine(2,5)-enkephalin (DPDPE) stimulated [³⁵S]GTPγS autoradiography (Schroeder et al., 2003) was reported in any brain regions of rats treated chronically with cocaine irrespective of treatment paradigm. Surprisingly, chronic ‘binge’ (Unterwald et al., 1993) but not continuous (Izenwasser et al., 1996) cocaine administration decreased DOPr activity in the NAc and CPu as measured by DPDPE inhibition of adenylate cyclase activity, suggesting that cocaine may result in an uncoupling of the DOPr from the G protein under that administration paradigm. This uncoupling seems to persist during acute cocaine withdrawal, as inhibition of AC by DOPr agonists was attenuated in the FCx, NAc and CPu (Perrine et al., 2008), clearly indicating desensitization of DOPr during acute cocaine withdrawal, which has been shown to be associated with the emergence of emotional withdrawal symptoms (Perrine et al., 2008). In support of this, internalization of DOPr in the NAc has been demonstrated 48 h following a chronic ‘binge’ cocaine administration paradigm (Ambrose-Lanci et al., 2008). Though limited, these data do not support a regulatory role of cocaine on the DOPr system, at least at the mRNA and protein level. Nonetheless, there is strong evidence to suggest that the regulatory role of cocaine on the DOPr is more likely to be at a signal transduction level.

Regulation of KOPr

As far as the regulation of KOPr by chronic cocaine use is concerned, there are inconsistent findings between KOPr mRNA, receptor binding and KOPr-stimulated [³⁵S]GTPγS studies in the literature. At the mRNA level, chronic ‘binge’ cocaine decreased KOPr levels in the VTA, NAc, SN (Spangler et al., 1996a; Rosin et al., 1999) but not in the CPu (Spangler et al., 1996a; 1997). No modification of KOPr mRNA was observed following withdrawal (Spangler et al., 1996a). Nonetheless, at the protein level, while no modification of KOPr binding was observed in the NAc, BLA, SN (Unterwald et al., 1994), there was an up-regulation of KOPr binding in the CgCx, OT, CPu, VTA (Unterwald et al., 1994; 2001) and in the Sept (Bailey et al., 2007a) following a chronic ‘binge’ cocaine administration paradigm. This up-regulation was also observed following chronic continuous cocaine administration in the NAc, BLA, Den and Sept (Collins et al., 2002). Despite this up-regulation in KOPr binding, no increase of KOPr activity was observed in studies of dynorphin 1- to 17-stimulated [³⁵S]GTPγS autoradiography (Schroeder et al., 2003). As in the case of MOPr, this inconsistency between mRNA, protein and [³⁵S]GTPγS data suggests that mechanisms other than alterations of gene expression contribute towards the regulation of KOPr by cocaine. Withdrawal from cocaine administration decreased KOPr binding in the NAc, CPu, BLA and Sept in rodents (Turchan et al., 1998; Bailey et al., 2007a).

These results clearly demonstrate an activation of the KOPr system following chronic cocaine exposure, but in contrast with the MOPr system, it does not persist following withdrawal from the drug. In agreement with these results, an increase in KOPr binding was also observed in the CPu, NAc, Amy of post-mortem brain of people with a history of cocaine abuse (Hurd and Herkenham, 1993; Staley et al., 1997; Mash et al., 2002), demonstrating the translational relevance of this cocaine-induced regulation of KOPr as well as the translational reliability of the animal model of chronic ‘binge’ cocaine administration paradigm. As KOPr stimulation in the brain produces aversive effects in animals and humans (Zimmer et al., 2001; McLaughlin et al., 2003; Shippenberg et al., 2007), it is likely that this cocaine-induced up-regulation of KOPr in regions associated with reward (CPu, CgCx, VTA) might be part of a protective compensatory neuroadaptive mechanism to counteract the rewarding effect of cocaine and might contribute to the emergence of persistent dysphoria, which is very often reported in humans after the withdrawal of the drug (Gawin, 1991). The functional implications of the alterations of KOPr in brain regions associated with emotional regulation and stress (Amy, Sept) remain to be determined (see KOPr KO section).

Opioid peptide regulation by cocaine

Regulation of PENK

There are several studies that have investigated the effect of chronic cocaine treatment on PENK mRNA, but the results of these studies are inconsistent. For instance, while an increase of PENK mRNA has been reported in the NAc (Branch et al., 1992; Hurd et al., 1992; Przewlocka and Lason, 1995; Mathieu-Kia and Besson, 1998; Crespo et al., 2001) and CPu (Steiner and Gerfen, 1993; Przewlocka and Lason, 1995; Spangler et al., 1996b; 1997; Svensson and Hurd, 1998; Crespo et al., 2001) of rodents following chronic cocaine exposure, a number of other studies have not reported any changes in PENK mRNA in these regions (Hurd et al., 1992; Daunais and McGinty, 1995; Spangler et al., 1996a; Mathieu-Kia and Besson, 1998; Alvarez Fischer et al., 2001; Bailey et al., 2005b; Ziolkowska et al., 2006). In human post-mortem brains, a decrease in PENK mRNA levels with a history of cocaine abuse has been reported in the CPu (Hurd and Herkenham, 1993). It is possible that enkephalins may also affect MOPr regulation, but this is unlikely as a reduction of MOPr levels was observed in the CPu of the same human post-mortem brains where a reduction of enkephalin was observed (Hurd and Herkenham, 1993). In other regions, no changes in PENK mRNA levels were observed in the cortex (Daunais and McGinty, 1994), FCx (Branch et al., 1992; Bailey et al., 2005b), Amy (Turchan et al., 2002; Ziolkowska et al., 2006), Hy, pituitary (Pit), central grey and cerebellum (Branch et al., 1992) following chronic cocaine treatment in rodents. However, cocaine increased PENK levels in the piriform (Pir), tubercle (Tu) in a self-administration paradigm, which persisted following acute and long-term withdrawal (Crespo et al., 2001). Divergent results have also been observed following acute withdrawal from chronic cocaine exposure. While one study reported a decrease of PENK mRNA levels in the CPu and NAc (Przewlocka and Lason, 1995), several others did not demonstrate any changes following acute withdrawal in these regions (Arroyo et al., 2000; Bailey et al., 2005b). A decrease in PENK expression was observed in the Amy following spontaneous withdrawal (Crespo et al., 2001), but no change was observed in the FCx (Bailey et al., 2005b) and in the CPu and NAc (Svensson and Hurd, 1998). Increased levels of PENK mRNA was reported in the CPu, NAc, Tu and Pir and decreased levels were observed in the ventromedial nucleus of the hypothalamus and Amy in animals chronically abstinent from cocaine self-administration (Crespo et al., 2001). These differences could reflect differences in drug administration paradigm or strain/species, and as a result, it is difficult to make clear suggestions about the regulatory role of cocaine on PENK expression other than to say it is commonly susceptible to alterations by cocaine treatment and withdrawal.

Regulation of POMC

More consistent results have been reported for POMC gene expression where no change in mRNA levels were reported following chronic cocaine exposure (Zhou et al., 2005), withdrawal (Zhou et al., 2005) or CPP (Leri et al., 2006) in the Hy, suggesting a lack of POMC regulation by cocaine at least at the level of gene expression in the Hy. However, after acute cocaine administration (Olive et al., 2001) and self-administration (Roth-Deri et al., 2003; 2006) of cocaine, studies have shown increases in β-end release in the NAc, which suggests that β-end might be involved in the initial rewarding properties of cocaine. A cocaine-induced increase in β-end release in the NAc has also been suggested from a study showing a time-dependent decrease in MOPr binding in the NAc following acute cocaine administration in rats (Soderman and Unterwald, 2009). Plasma β-end levels were also shown to be increased following acute cocaine administration (Moldow and Fischman, 1987). Moreover, chronic cocaine administration increased β-end immunoreactivity in plasma, Pit but not in the Hy (Forman and Estilow, 1988), indicating a regulatory role of cocaine on β-end, specifically in the Pit. High levels of plasma β-end have been observed in abstinent human cocaine addicts (Vescovi et al., 1992), supporting the concept of a stimulatory effect of cocaine on β-end release. It is not clear what the β-end release status is in the brain during cocaine withdrawal. β-end deficiency in regulating HPA axis activity has also been suggested based on a study demonstrating hyper-responsivity to removal of glucocorticoid negative feedback in cocaine abstinent human individuals following the administration of metyrapone (Schluger et al., 1998; 2001). More studies need to be carried out in order to clarify the β-end status during withdrawal.

Regulation of PDYN

The most consistent and reliable finding in terms of cocaine-induced regulation of opioid peptide gene expression is undoubtedly the increase of PDYN mRNA levels and immunoreactivity in the CPu following acute, subacute, chronic ‘binge’ and continuous cocaine treatment (Sivam, 1989; Smiley et al., 1990; Hurd et al., 1992; Spangler et al., 1993; 1996a; 1997; Steiner and Gerfen, 1993; Daunais and McGinty, 1995, 1996; Romualdi et al., 1996; Mathieu-Kia and Besson, 1998; Svensson and Hurd, 1998; Werme et al., 2000; Yuferov et al., 2001; Zhou et al., 2002; Bailey et al., 2005b; Schlussman et al., 2005) as well as following cocaine self-administration (Ziolkowska et al., 2006). This increase has been shown to occur irrespective of treatment protocol. In agreement with these findings, an increase in PDYN was also observed in the CPu and VP of post-mortem brain of people with a history of cocaine abuse (Hurd and Herkenham, 1993; Frankel et al., 2008), demonstrating the translational relevance of this cocaine-induced regulation of PDYN as well as the translational reliability of the animal model of chronic ‘binge’ cocaine administration paradigm. However, less consistent results have been reported following withdrawal, where a decrease (Svensson and Hurd, 1998), no change (Spangler et al., 1996a) or increase (Sivam, 1989; Smiley et al., 1990; Bailey et al., 2005b) of PDYN mRNA or immunoreactivity has been reported in the CPu. In the NAc, an increase (Smiley et al., 1990; Hurd et al., 1992; Mathieu-Kia and Besson, 1998; Turchan et al., 2002) or no change (Alvarez Fischer et al., 2001; Schlussman et al., 2005; Bailey et al., 2005b; Ziolkowska et al., 2006) in PDYN mRNA or immunoreactivity levels has been reported following chronic ‘binge’ cocaine administration. This administration paradigm was also shown to increase PDYN mRNA in the dentate gyrus (DG) (Turchan et al., 2002) but not in the Amy (Turchan et al., 2002; Zhou et al., 2008), FCx, Cx (Yuferov et al., 2001) and SN (Spangler et al., 1993). On the contrary, an increase of dynorphin immunoreactivity was observed following chronic ‘binge’ cocaine treatment in the SN, which persisted in acute withdrawal in rats (Smiley et al., 1990). No change was observed in the Hi (Smiley et al., 1990). In addition, a decrease in PDYN mRNA was observed in the Hy following continuous cocaine administration (Romualdi et al., 1996) but not following chronic ‘binge’ administration (Yuferov et al., 2001; Zhou et al., 2008), implying that the regulation of PDYN by cocaine in the Hy depends on the paradigm of administration. However, while an increase of PDYN mRNA was observed in the LH following acute withdrawal from chronic ‘binge’ cocaine, it did not persist into long-term withdrawal (Zhou et al., 2008).

By considering the KOPr and PDYN data together, we can conclude that the PDYN/KOPr system is under profound regulatory control by cocaine and is activated following chronic cocaine exposure. As discussed earlier, it is likely that this cocaine-induced activation of the KOPr/PDYN system in regions associated with reward (CPu, NAc, VTA, CgCx) might be part of a protective compensatory neuroadaptive mechanism to counteract the positive reinforcement effect of cocaine. In addition and considering the role of PDYN/KOPr in inducing dysphoria, it is likely that the alterations observed in brain regions associated with stress (Sept, Amy, Hy) may contribute to the emergence of persistent dysphoria and in triggering relapse in response to stress, which is very common in human cocaine addicts after the withdrawal of the drug (Gawin, 1991). Recent studies carried out with the use of selective pharmacological tools and in KO mice (discussed in the following section) have shed further light into the role of PDYN/KOPr system in cocaine addiction.

Cocaine responses in opioid receptors and peptide KO mice

Opioid receptor KO mice

MOPr KO mice

A great body of literature over the last 50 years has provided clear evidence with the use of pharmacological tools that the MOPr plays an important role in mediating the positive reinforcing effects of natural rewards as well as of opiate and non-opiate drugs of abuse including cocaine (for extensive review on this topic, see Shippenberg and Elmer, 1998; van Ree et al., 2000; Le Merrer et al., 2009). Briefly, administration of selective MOPr antagonists attenuates the development of cocaine CPP (Schroeder et al., 2007; Soderman and Unterwald, 2008) and reduces cocaine self-administration (Ward et al., 2003) and reinstatement (Tang et al., 2005) of cocaine-seeking behaviour in rats, clearly suggesting that the MOPr system is involved in the rewarding as well as the relapse potential of the psychostimulant.

The use of MOPr KO mice clearly demonstrated that the MOPr is the primary molecular target for morphine as morphine CPP (Matthes et al., 1996; Sora et al., 2001; Mizoguchi et al., 2003) and self-administration (Becker et al., 2000) were completely abolished in KO animals. Although MOPr KO studies demonstrated a clear role of MOPr in mediating the positive reinforcing effects of some non-opioid drugs of abuse such as nicotine (Berrendero et al., 2002), delta9-tetrahydrocannabinol (Ghozland et al., 2002) and alcohol (Roberts et al., 2000; Becker et al., 2002), the same was not true for cocaine reinforcement. Cocaine CPP was shown to be unchanged (Contarino et al., 2002), increased (Becker et al., 2002) or decreased (Hall et al., 2004) in MOPr KO (Table 2), which gives a unclear picture of the role of MOPr in cocaine reinforcement and is certainly not in agreement with the pharmacological manipulations described earlier. The discrepancy of results between studies could be due to differences in genetic background of mice strain used, differences in gender of animals and/or differences in experimental protocol used (dose of cocaine, number of conditional sessions, etc.). While the study conducted by Becker et al. (2002) used mixed 129/Ola × C57BL male mice (F2 generation), Hall et al. (2004) used congenic C57 mice (F10 generation) of mixed sexes. Indeed, strains of mice differ considerably in their behavioural and neurochemical effects of drugs of abuse (e.g. Cunningham et al., 1992; Orsini et al., 2005; Glatt et al., 2009; Bailey et al., 2010), and genetic background of mouse strain has been shown repeatedly to influence the phenotypic and neurochemical consequences of gene KO (Hummel et al., 2004; Yoo et al., 2010). This really points towards the importance of using KO mice of homogeneous genetic background by backcrossing over several generations in order to dilute the effect of background strain. In addition, differences in cocaine rewarding effects have been shown between male and female mice (Anker and Carroll, 2011), which might also explain the discrepancies observed. This points towards the importance of using same-sex KO mice in behavioural experiments to minimize the gender effect on phenotypic changes in KO mice. Finally, in most studies, only one or two doses of cocaine was tested for cocaine CPP experiments, which could mask dose–response shift in sensitivity of these mice to the rewarding effect of cocaine. This demonstrates the need to conduct CPP experiments in KO mice with a range of different doses of cocaine. In contrast to cocaine CPP and in agreement with pharmacological studies, cocaine self-administration was reduced in male congenic C57BL MOPr KO in a dose–response manner (Mathon et al., 2005), suggesting a role for MOPr in the operant and reinforcement effect of cocaine.

Table 2.

Behavioural and neurochemical effects of cocaine in opioid receptor and peptide knockout mice

Gene	Gender and background	Dose	Behavioural effects	Biochemical effects	Reference
MOPr	Male	5 or 10 mg·kg−1, i.p.	↓ CPP (5 mg·kg−1)(5 mg·kg−1)		(Becker et al., 2002)
	129/Ola × C57BL/6 Crossed to CB6F1	Two or four parings	↑ CPP (10 mg·kg−1)
	F2 used for the study
		20 or 40 mg·kg−1, i.p.	↔ Locomotion
	Male, female	10 mg·kg−1, i.p.	↔ CPP		(Contarino et al., 2002)
	129Sv × C57BL/6 F2 used for the study	Three parings
		30 mg·kg−1, i.p.	↔ Locomotion
	Male, female	2 × 5 or 10 mg·kg−1, s.c.	↓ CPP (10 mg·kg−1)		(Hall et al., 2004)
	129SvEv × C57BL/6J	Two parings
	>F10 used for the study
		5, 10,20 mg·kg−1, s.c.	↔ Locomotion
		20 mg·kg−1, s.c., 5 days	↔ Sensitization
	Male, Female	15 mg·kg−1, i.p. (injection for 6 days, 6 days withdrawal then challenge with initial dose)	↓ Locomotion	↓ nNOS-ir in DG	(Yoo et al., 2003; 2006)
	129/Ola × C57BL/6		↓ Sensitization	↔ nNOS-ir in Cx, CPu
	Male	10 or 20 mg·kg−1, i.p.	↓ Locomotion (20 mg·kg−1)	↔ DA in NAc	(Chefer et al., 2004)
	129/Sv × C57BL/6
	>N10 backcrossed to C57BL/6
	F15 used for the study
	Male	0.4, 0.8, 1.6 µg·kg−1·1.6 µL−1, i.v.	↓ Self-administration (1.6 µg·kg−1·1.6 µL−1)		(Mathon et al., 2005)
	129Sv × C57BL/6
	N6 or N7 backcrossed to C57BL6/Jico
	Male	3, 10, 20 or 30 mg·kg−1, i.p.	↔ Locomotion		(Lesscher et al., 2005)
	129Sv × C57BL/6
	N6 or N7 backcrossed to C57BL6/Jico
		10 mg·kg−1, i.p., (Injection for 11 days, 72 h withdrawal and then challenge with 20 mg·kg−1)	↔ Sensitization
	Male	15 mg·kg−1, i.p. (injection for 10 days, 7 days withdrawal then challenge with initial dose)	↔ Locomotion		(Hummel et al., 2004)
	129S6 × C57BL/6J		↓ Sensitization
	Male	Same as previous	↑ Locomotion
	129S6 × C57BL/6J		↑ Sensitization
	≥ N10 backcrossed to C57BL/6J
	Male	Same as previous	↓ Locomotion
	129S6 MOPr KO × F10 C57BL/6J		↓ Sensitization
	F1 used for the study
	Male	Same as previous	↔ Locomotion
	129S6 × C57BL/6J crossed to 129S6		↔ Sensitization
DOPr	Male	10 or 20 mg·kg−1, i.p.	↑ Locomotion (10 mg·kg−1)	↓ DA in Nac (20 mg·kg−1)	(Chefer et al., 2004)
	129/Sv × C57BL/6
	>N10 backcrossed to C57BL/6
	F15 used for the study
KOPr	Male	5, 10 or 15 mg·kg−1, i.p.	↑ Locomotion	↑ DA in Nac (10 mg·kg−1)	(Chefer et al., 2005)
	129S6 × C57BL/6J
				↓ c-Fos, jun B mRNA
	Male	15 mg·kg−1, i.p., 8 days (injection for 5 days, 3 days withdrawal then challenge with initial dose)	↓ Sensitization
	129S6 × C57BL/6J		↑ Sensitized state
	N10 backcrossed to C57BL/6J
	Gender not specified	15 mg·kg−1, s.c.	↔ CPP (unstressed)		(McLaughlin et al., 2006a)
	129S6 × C57BL/6J	2 days pairing	↓ CPP (FST stress)
	>N10 backcrossed to C57BL/6
	Male	15 mg·kg−1, s.c.	↓ CPP reinstatement (Foot shock & FTS)		(Redila and Chavkin, 2008)
	129S6 × C57BL/6J backcrossed to C57BL/6	four pairings
PENK	–	–	–	–	–
β-end	Male	15, 30,60 mg·kg−1 i.p.	↓ Locomotion		(Marquez et al., 2008)
	C57BL/6J × 129/SV
	>N10 backcrossed to C57BL/6J
		15, 30, 60 mg·kg−1 i.p.	↓ CPP
		2 pairings
PDYN	Male	10 and 20 mg·kg−1, i.p.	↓ Locomotion (20 mg·kg−1)	↓ DA levels in NAc	(Chefer and Shippenberg, 2006)
	129/SV × C57BL/6,
	N10 backcrossed to C57BL/6
	F15 used for the study
	Male	3 × 15 mg·kg−1/day, i.p.	↔ Locomotion	↓ D2 binding in CPu	(Bailey et al., 2007b)
	129/SV × C57BL/6J background	14 days	↑ Sensitization	↔ D1, DAT binding
	≥N10 backcrossed to C57BL/6J		↔ Stereotypy	↓ MOP activity increase*
				↓ Corticosterone levels
	Gender not specified	15 mg·kg−1, s.c.	↔ CPP (unstressed)		(McLaughlin et al., 2006b)
	129/SvEvTac × C57BL/6	two pairings	↓ CPP (FST stress)
	>N10 backcrossed to C57BL/6J
	Male	15 mg·kg−1, s.c.	↓ CPP reinstatement (foot shock and FTS)		(Redila and Chavkin, 2008)
	129/SvEvTac × C57BL/6	four pairings
	backcrossed to C57BL/6J

^*As measured by DAMGO stimulated [³⁵S]GTPγS autoradiography.

In addition to CPP and self-administration, other behavioural tests are very commonly used in order to address other aspects of cocaine-addictive behaviours such as behavioural sensitization. Behavioural sensitization is a phenomenon whereby repeated exposure to psychostimulant drug elicits progressive enhancement of behavioural responses, which persists after withdrawal from the drug (Robinson and Becker, 1986) and is thought to reflect neuroadaptive/neuroplastic alterations that mediate drug-seeking behaviour (Kalivas and Stewart, 1991; Kalivas et al., 1993; Pierce and Kalivas, 1997). As with CPP, there are discrepancies among studies investigating the locomotor stimulating and sensitizing effects of cocaine in mice lacking MOPr. A decrease (Yoo et al., 2003; Chefer et al., 2004) or no change (Becker et al., 2002; Contarino et al., 2002; Hall et al., 2004; Lesscher et al., 2005) of cocaine-induced locomotor activity and/or locomotor sensitization was observed in MOPr KO mice. As for CPP, the discrepancy in locomotor and sensitization effects of cocaine in MOPr KO mice are likely to be due to gender differences as well as to genetic background. Mixed sexes of mice were used in the Contarino et al. (2002), Hall et al. (2004), Yoo et al. (2003; 2006) studies, whereas only male mice were used in the Chefer et al. (2004), Lesscher et al. (2005) and Hummel et al. (2004) studies. Hummel et al. (2004) specifically carried out cocaine-induced locomotor and sensitization studies in MOPr KO mice of different genetic backgrounds under the same experimental conditions in order to address the influence of genetic background in phenotypic changes in mice lacking MOPr. An increase in the acute locomotor and sensitization effect of cocaine was observed in congenic C57BL/6J MOPr KO mice (>N10 backcrossing to C57BL/6J) versus wild type (WT); a decrease was observed in the mixed 129S6 × C57BL/6J MOPr KO, whereas no changes were observed in isogenic 129S6 MOPr KO (Hummel et al., 2004), clearly displaying that genetic background does influence phenotypic changes in mice lacking MOPr.

Studies using pharmacological tools have demonstrated that MOPr contributes towards the elevation of DA release in the NAc response to cocaine (Shippenberg and Elmer, 1998; Van Ree et al., 2000). In contrast with these studies, no significant genotype effect was detected in either basal or acute cocaine-induced NAc DA levels in MOPr KO (Chefer et al., 2004). This discrepancy may account for compensatory changes taking place in other CNS systems in KO mice, which could influence DA release in the NAc.

The effect of chronic cocaine administration on neuronal nitric oxide synthase (nNOS) expression in the Hi of WT and MOPr KO was investigated in order to assess if cocaine induction of nNOS is mediated by MOPr. Chronic cocaine administration induced up-regulation of the expression of nNOS in the DG of the Hi in WT mice but that was attenuated in MOPr KO clearly suggesting an involvement of nNOS pathway in the mechanism of action of cocaine (Yoo et al., 2006).

DOPr KO mice

The use of pharmacological manipulation has consistently demonstrated that the DOPr is involved, although to a lesser extent than the MOPr, in mediating the positive reinforcing effects of opioid and non-opioid drugs of abuse including cocaine (for review, see Le Merrer et al., 2009). Selective DOPr antagonists decrease systemic cocaine self-administration when microinjected in the NAc but increase cocaine self-administration when injected in the VTA and had no effect when injected into the Amy, indicating that the DOPr can modulate cocaine reinforcement depending on the brain site (Ward et al., 2003).

Studies on DOPr KO mice demonstrated an anxiogenic and depressive phenotype in these animals (Filliol et al., 2000), and DOPr KO exhibited an increase in ethanol self-administration (Roberts et al., 2001), clearly suggesting a role for DOPr on emotional regulation and drug reinforcement. The locomotor activity of DOPr KO mice maintained on a pure C57BL/6J background in response to acute cocaine administration was investigated. Acute cocaine-induced locomotor activity was significantly increased in DOPr KO, clearly suggesting a modulatory role for DOPr in the acute psychomotor effect of cocaine (Chefer et al., 2004). Further behavioural tests such as cocaine CPP, sensitization and self-administration need to be carried out to demonstrate the role of DOPr in cocaine addiction behaviours.

Studies conducted with pharmacological tools have suggested that cocaine-induced release of DA in the NAc is partly mediated by a DOPr dependent mechanism (Shippenberg and Chefer, 2003), which is likely to contribute towards the positive reinforcing effect of cocaine. In agreement with these studies, Chefer et al. (2004) showed a diminished release of DA in the NAc of DOPr KO in response to acute cocaine, clearly suggesting a role for the DOPr in mediating the release of DA in response to cocaine in the NAc. The mechanism underlying the role of DOPr in mediating the cocaine-induced DA release is not clear. However, there is evidence supporting an involvement of glutamatergic and free radical dependent mechanisms (Billet et al., 2004; Fusa et al., 2005). A mechanism also involving MOPr–DOPr heteromeric interactions may also be plausible.

KOPr KO mice

It has been well established that activation of KOPr produces motivational and neurochemical effects that oppose those of MOPr (Shippenberg et al., 1998; Van Ree et al., 2000). This was confirmed in KOPr KO mice, which did not experience CPP aversion in response to a KOPr agonist, in contrast to WT mice, which did (Simonin et al., 1998). In terms of neurochemistry, KOPr agonists have been repeatedly shown to decrease DA release in the NAc, whereas activation of MOPr increases it (Di-Chiara and Imperato, 1988). This was also confirmed in the KOPr KO mice, which were found to have higher basal levels of DA in the NAc compared with WT in a freely moving microdialysis study (Chefer et al., 2005), indicating a tonic inhibitory role of KOPr on DA release. This result was replicated both in KOPr KO mice maintained on a pure C57BL/6J or a mixed 129S6 × C57BL/6J background (Chefer et al., 2005), indicating that that KOPr-mediated regulation of DA release is not influenced by background strain of mouse.

It is also broadly accepted that activation of the KOPr system has anti-addictive properties by antagonizing the acute reinforcing/rewarding effect of cocaine (see Wee and Koob, 2010). Selective KOPr agonists were shown to consistently block cocaine-induced CPP in rodents (Crawford et al., 1995; Mori et al., 2002; Zhang et al., 2004), to decrease cocaine self-administration in rats (Glick et al., 1995; Schenk et al., 1999) and monkeys (Mello and Negus, 1998), to attenuate cocaine-induced behavioural sensitization and to decrease cocaine-induced DA release in the NAc (Heidbreder et al., 1993; 1995; 1998; Heidbreder and Shippenberg, 1994; Shippenberg et al., 1996; Zhang et al., 2004; 2005), clearly suggesting an important role for the KOPr system in modulating the rewarding effect of cocaine by opposing cocaine-induced DA release.

In agreement with the aforementioned studies, the acute effect of cocaine on locomotor behaviour and striatal DA release was enhanced in KOPr KO mice relative to WT maintained on a mixed 129S6 × C57BL/6J background (Chefer et al., 2005). In addition, in contrast to WT mice, repeated administration of cocaine did not induce locomotor sensitization in KOPr KO mice (Chefer et al., 2005). This might signify the development of a chronically sensitized state in the absence of the KOPr. The reduction of cocaine-induced immediate early gene expression observed in KOPr KO mice (Chefer et al., 2005) provides additional evidence of a ‘cocaine-sensitized’ phenotype for KOPr KO mice.

Although it was initially accepted that the activation of the KOPr system has anti-addictive properties, there has been recent emerging evidence showing that it can also trigger cocaine reinstatement via a stress-dependent mechanism (see review by Wee and Koob, 2010). While KOPr antagonists inhibited stress-induced, but not cocaine-primed, reinstatement of cocaine self-administration in rats (Beardsley et al., 2005) and stress-induced cocaine CPP in mice (Carey et al., 2007; Redila and Chavkin, 2008), KOPr agonists induced reinstatement of extinguished cocaine-associated place preference (Redila and Chavkin, 2008), clearly suggesting that stress can induce reinforcing effects of cocaine via a KOPr-mediated mechanism. This suggestion was further reinforced by recent studies carried out in KOPr KO mice. While no differences in cocaine CPP was observed between WT and KOPr KO mice in the absence of stress, repeated swim stress potentiated the rewarding effect of cocaine in WT but not in KOPr KO mice, further demonstrating that KOPr activation induced by stress may be both necessary and sufficient for potentiating the reinforcing actions of cocaine (McLaughlin et al., 2006a). In support of this, foot shock stress and forced swim-induced reinstatement of cocaine CPP was absent in mice lacking the KOPr gene (Redila and Chavkin, 2008).

PENK KO mice

Endogenous PENK has been postulated to mediate the positive reinforcing effects of several drugs of abuse (Gianoulakis, 1993; Berrendero et al., 2005; Marinelli et al., 2005; Shoblock and Maidment, 2007). No alterations of basal levels of DA was observed in the NAc of PENK KO mice versus WT (Berrendero et al., 2005), indicating a lack of tonic regulation of DA release by PENK, at least at a basal state. There are no studies to our knowledge investigating the effect of cocaine on PENK KO mice.

β-end KO mice

Endogenous end has been suggested to modulate positive reinforcement and pleasurable effects of a range of drugs of abuse including cocaine (Olive et al., 2001), alcohol (Gianoulakis, 1993; Jarjour et al., 2009) and nicotine (Roth-Deri et al., 2008). Cocaine has been shown to increase β-end in the NAc (Olive et al., 2001), raising the possibility that this neuropeptide may be important in rewarding effects of the drug. Locomotor activity and CPP effects of cocaine were both reduced in mice lacking β-end in a dose-dependent manner (Marquez et al., 2008), further demonstrating that β-end plays a modulatory role in the motor stimulating and rewarding effects of acute cocaine. Further studies are warranted for the investigation of the operant and chronic effect of cocaine in these animals.

PDYN KO mice

Activation of the endogenous dynorphin system has been proposed to oppose positive reinforcement/reward by means of KOPr activation, resulting in tonic inhibition of DA release and signalling (Zhang et al., 2004; Chefer et al., 2005; Shippenberg et al., 2007). Dynorphin A (1–17) administered directly in the CPu of mice was shown to decrease basal DA release in a dose-dependent manner by more than 60% (Zhang et al., 2004). Dynorphins have also been implicated in behavioural responses to stress. i.c.v. administration of dynorphin A (1–17) was shown to potentiate the immobility response to stress, which was blocked by selective KOPr antagonists (McLaughlin et al., 2003). In addition, when exposed to an inescapable physical or psychological stressor, rodents demonstrate stress-induced analgesia that is blocked by KOPr selective antagonists (Takahashi et al., 1990; McLaughlin et al., 2003; 2006a; Aldrich and McLaughlin, 2009). This effect was also blocked in PDYN KO mice, suggesting that KOPr activation by dynorphin contributes to behavioural responses to stress (McLaughlin et al., 2003).

In terms of responses to cocaine effects, administration of dynorphin 1–17 directly in the CPu was shown to decrease cocaine-induced CPP to attenuate the locomotor effect of the drug and to block the elevation of striatal DA levels in response to cocaine in mice (Zhang et al., 2004). These data, together with the fact that cocaine induces PDYN expression in the striatum (as discussed earlier), demonstrates a critical role of dynorphin in suppressing DA release and, as a result, in opposing the rewarding and reinforcing effect of cocaine. As a consequence, dynorphin could be characterized as an ‘anti-addictive’ peptide. In agreement with this hypothesis, we showed an enhancement of locomotor sensitization following chronic ‘binge’ cocaine administration in mice lacking PDYN gene (Bailey et al., 2007b). In terms of DA release, Chefer and Shippenberg (2006) showed that the deletion of dynorphin in mice, decreased basal DA levels and attenuated the cocaine-induced elevation of DA in the NAc. This paradoxical phenomenon might be due to compensatory up-regulation of KOPr observed in PDYN KO mice (Clarke et al., 2003), which could be activated by endogenous met-enkephalin and/or β-end. We also showed that the deletion of dynorphin induced a decrease of striatal D2 receptors in chronic ‘binge’ cocaine-treated mice but not in WT animals (Bailey et al., 2007b), further demonstrating the protective role of dynorphin in opposing D2 receptor down-regulation, which is thought to be associated with increased vulnerability to the reinforcing effect of the drug (Nader and Czoty, 2005).

In complete agreement with data obtained from KOPr KO mice, use of Ppdyn KO confirmed the role of endogenous dynorphin in stress-inducing effects of cocaine and in triggering cocaine reinstatement via a stress-like mechanism. Genetic deletion of PDYN abolished stress-induced reinstatement of cocaine CPP in mice (Redila and Chavkin, 2008). Moreover, while no differences in cocaine CPP was observed between WT and PDYN KO mice in the absence of stress, repeated swim or foot shock stress potentiated the rewarding effect of cocaine in WT but not in PDYN KO (McLaughlin et al., 2003), further demonstrating that activation of the dynorphin system induced by stress may be both necessary, and sufficient, for the potentiation of the reinforcing actions of cocaine. The mechanism underpinning this effect is not clear, but it might possibly involve the HPA axis and/or extra-hypothalamic mechanisms (e.g. Amy). We demonstrated that while chronic ‘binge’ cocaine elevated corticosterone levels in WT mice, this effect was blunted in PDYN KO (Bailey et al., 2007b), clearly suggesting a PDYN mechanism in the HPA axis stress-activating effect of cocaine. In support of this finding, Wittmann et al. (2009) showed an attenuation of stress-induced corticosterone levels in PDYN KO mice.

Conclusions/therapeutic implications

Together, the opioid peptide/receptor regulation and gene KO studies have clearly demonstrated the induction of profound neuroadaptive changes in the opioid system, which may contribute towards the progressive transition from impulsive drug taking to compulsive cocaine abuse (Figure 1).

Figure 1.

Endogenous opioid neurocircuitries hypothesized to be recruited during the transition from positive to negative reinforcement in cocaine addiction (founded on figures from Le Moal and Koob, 2007; Wee and Koob, 2010). Top left: illustrates an increase in MOPr activity in regions of the brain involved in reward (NAc, CPu, Amy) and impusivity/decision-making (FCx, CgCx), which drives the positive reinforcement effect of cocaine. It also illustrates the involvement of β-endorphin (β end) and enkephalin (enk) in the reinforcement effect of cocaine via their interaction with the mesolimbic dopaminergic system. Top right: illustrates an activation of the dyn/KOPr system in regions of brain regions involved in reward. This activation opposes the positive reinforcement effect of cocaine via its interaction with the mesolimbic dopaminergic system. Bottom left: illustrates an increase in MOPr activity in regions of the brain involved in drug priming (FCx, CgCx) and cue priming reinstatement (Amy). It is not clear what effect this may have in terms of driving negative reinforcement and compulsive drug taking. Bottom right: illustrates an activation of the dyn/KOPr system in brain regions involved in stress responses (Hy, Amy), which will drive negative reinforcement and compulsive cocaine taking and trigger relapse. The images are computer-enhanced autoradiograms of coronal mouse brain sections labelled with MOPr and KOPr selective ligand [³H]DAMGO and [³H]CI977 respectively (for detailed autoradiography methods, see Kitchen et al., 1997). The sections shown are from the level of the caudate (Bregma 0.86 mm) and from the thalamus (Bregma –1.94 mm). The arrows indicate an increase of positive or negative reinforcement and the size of the arrow represents the magnitude of the effect. The mesolimbic dopaminergic system (green line), which projects from the VTA to various parts of the brain, is illustrated. dyn, dynorphin; enk, enkephalin; PAG, periaquetactal grey.

• Cocaine causes a region-specific activation of the MOPr system, which we postulate to contribute towards the positive reinforcement/hedonic effect of cocaine (partly via increasing DA release in the NAc) and towards cocaine craving during withdrawal followed by relapse. Cocaine causes a region-specific activation of the MOPr system, which we postulate to contribute towards the positive reinforcement/hedonic/drug wanting effect of cocaine (partly via increasing DA release in the NAc) and towards cocaine craving during withdrawal followed by relapse. More specifically, we postulate that cocaine would activate MOPr, leading to opioid-induced DA release. There is evidence to suggest that dopaminergic (Azaryan et al., 1996a; Ambrose et al., 2004) and non-dopaminergic (Bailey et al., 2007b) mechanisms may mediate this MOPr activation.

• Cocaine also causes a region-specific activation of the KOPr/PDYN system, which we postulate to antagonize the positive reinforcement effect of cocaine, via opposing DA release in the NAc, but also to contribute to the stress-inducing properties of the drug and the triggering of relapse.

These conclusions have important implications for the development of effective pharmacotherapy for the treatment of cocaine addiction and the prevention of relapse. Indeed, one could suggest that MOPr antagonists and KOPr agonists would be more effective as an acute intervention to suppress cocaine reward and craving than for a long-term treatment (Aldrich and McLaughlin, 2009). Indeed, the non-selective opioid receptor antagonist naltrexone has been tested in clinical trials as a possible treatment of cocaine cravings with mixed results (Heidbreder and Hagan, 2005; Modesto-Lowe et al., 1997), possibly due to the anxiety provoking effect of naltrexone (see review of Colasanti et al., 2011). In terms of KOPr agonism, although centrally acting KOPr agonists are known to induce dysphoria and sedation in humans (Pfeiffer et al., 1986; Mello and Negus, 2000), and for that reason the therapeutic development has been limited (Barber and Gottschlich, 1997; DeHaven-Hudkins and Dolle, 2004), they were shown to decrease symptoms of mania in bipolar disorder patients with no adverse effects (Cohen and Murphy, 2008). As a result, KOPr agonists might be of benefit for the treatment of acute cocaine cravings. On the other hand, the KO data suggest that KOPr antagonists would be better suited as a therapeutic to prevent stress-induced relapse in cocaine-dependent individuals and as consequence would help cocaine-abstinent individuals to maintain their abstinence state. The mixed opioid receptor ligand buprenorphine, which has mixed MOPr agonist/KOPr antagonist activity, was used in combination with naltrexone in heroin and cocaine addicts in a clinical study (Gerra et al., 2006). The outcomes of the study demonstrated enhanced compliance and drug abstinence from heroin and cocaine compared with naltrexone alone in dependent individuals (Gerra et al., 2006), further demonstrating the therapeutic potential of selective KOPr antagonists or a combination of MOPr/KOPr antagonists in cocaine addiction.

Problems, limitations and future directions

Despite its significant contribution towards understanding the mechanisms underlying the effects of cocaine, KO mouse technology does not come without its problems and limitations. Firstly, classic opioid receptor/peptide KO have the opioid receptor or peptide globally deleted from the animal, that is, in all brain regions, spinal cord, peripheral and immune sites. As a result, it is not possible to pinpoint the role of the precise anatomical brain sites or circuitry where the specific receptor or peptide is localized in modulating certain behavioural effects associated with cocaine use such as sensitization, stereotypy, impusivity, habit formation, emotional memory formation, anxiety, positive and negative reinforcement, reinstatement, CPP, etc. For instance, the role of MOPr in the FCx is postulated to be involved in impulsivity (Love et al., 2009), whereas the role of MOPr in the NAc is critical in reward processing (Van Ree et al., 2000). MOPrs in both these regions are up-regulated following chronic cocaine administration; however, the use of conventional KO mice is unable to pinpoint the functional/behavioural consequence of each of these region's specific up-regulations. The development of conditional KO mice that will enable spatial control over the gene deletion will surely allow more fine-tuning in our understanding of the role of receptors and peptides in specific regions involved in addictive biology. Transgenic Cre technology will permit the deletion of opioid peptide or receptor in specific neuronal populations (e.g. GABAergic or aminergic neuron) and, in combination with neuroanatomical imaging approaches, will provide further knowledge of the molecular and cellular mechanisms by which different components of the opioid system would influence addictive properties of cocaine in vivo (Le Merrer et al., 2009).

Furthermore, the genetic background of mouse strain has also been repeatedly shown to influence the phenotypic consequences of gene KO mice as evidenced, for example, by the discrepancy of results between cocaine sensitization, locomotor and CPP studies in MOPr KO mice discussed in this review. This really points towards the importance of using KO mice of homogeneous genetic background by backcrossing over >10 generations.

Furthermore, there is a need to take advantage of opioid transgenic technology to investigate the role of specific components of this system in later stages of cocaine addiction (reinstatement, extinction, habit formation, etc) as well as in the adverse effect of the drug (seizures, psychosis, neurotoxicity, cardiotoxicity, cognitive impairment). Indeed, opioid receptor/peptide KO technology was used predominantly to investigate the role of the endogenous opioid system in modulating the locomotor, behaviour sensitization and drug conditioning effects of cocaine, and a limited number of studies have investigated the influence of those genes in the development of cocaine dependence, extinction, reinstatement and consolidation, which are characteristics of later stages of drug addiction. In addition, although cocaine consumption is well known to be associated with serious cardiovascular complications that may lead to death (Vasica and Tennant, 2002; Afonso et al., 2007), the mechanism underlying its cardiotoxicity is not well understood (Fan et al., 2009). Despite the unique benefit that KO technology can provide to investigate these mechanisms, there has been a very limited use of these animals in this specific field.

Although opioid KO mice have been used for the investigation of genetic components of addiction and there is a large body of literature demonstrating the contribution of ‘environment’ at different stages of the drug addiction cycle, there is very limited literature on the interaction of these two factors, which are generally accepted to play a crucial role in drug addiction. Transgenic mice provide a unique model to study the influence of environment on genes and vice versa in an addiction setting. The further understanding of those interactions is crucial not only to provide new knowledge on the neuropathology of the disease and on risk factors that may influence drug addiction vulnerability but also for the developing of more effective pharmacotherapy for the treatment of this disease.

A final point that has to be made here is that based on recent genome-wide association studies (GWAS) of addiction in humans and multitude of genetic studies carried out in mice, the concordance between human and animal studies have been disappointing. However, this is not the case for MOPr and PDYN where human polymorphisms of these genes have been repeatedly reported to affect the vulnerability of cocaine addiction (for review, see Kreek et al., 2005; Yuferov et al., 2010). Nonetheless, there is a need for closer interaction between human and animal geneticists so that information about candidate genes underlying drug addiction detected by GWAS in human populations is used to inform development of novel transgenic animal models that will better model the genetics of human addiction.

Glossary

Amy

amygdala

BLA

basolateral amygdala

BNST

bed nucleus of the stria terminalis

Cer

cerebellum

CG

central grey

CgCx

cingulate cortex

CPP

conditioned place preference

CPu

caudate putamen

DA

dopamine

DAT

dopamine transporter

Den

dorsal endopiriform

DG

dentate gyrus

DOPr

δ-opioid receptor

DR

dorsal raphe

End

endorphin

FCx

frontal cortex

FST

forced swim test

Hb

habenula

Hi

hippocampus

HPA

hypothalamic–pituitary–adrenal

Hy

hypothalamus

IC

inferior colliculus

KO

knockout

KOPr

κ-opioid receptor

LH

lateral hypothalamus

MG

medial geniculate

MOPr

µ-opioid receptor

MS

medial septum

NAc

nucleus accumbens

nNOS

ir, neuronal nitric oxide synthase immunoreactivity

OB

olfactory bulb

OT

olfactory tubercle

PDYN

preprodynorphin

PENK

preproenkephalin

Pir

piriform

Pit

pituitary

POMC

proopiomelanocortin

Sept

septum

SN

substantia nigra

TCx

temporal cortex

Th

thalamus

Tu

tubercle

VDB

vertical limb of the diagonal band

VMN

ventromedial nucleus of the hypothalamus

VP

ventral pallidus

VTA

ventral tegmental area

WT

wild type

Conflict of interest

Authors declare they have no conflict of interest.