The genetics of the opioid system and specific drug addictions

Abstract

Addiction to drugs is a chronic, relapsing brain disease that has major medical, social, and economic complications. It has been established that genetic factors contribute to the vulnerability to develop drug addiction and to the effectiveness of its treatment. Identification of these factors may increase our understanding of the disorders, help in the development of new treatments and advance personalized medicine. In this review we will describe the genetics of the major genes of the opioid system (opioid receptors and their endogenous ligands) in connection to addiction to opioids, cocaine, alcohol and methamphetamines. Particular emphasis is given to association and functional studies of specific variants. We will provide information on the sample populations and the size of each study, as well as a list of the variants implicated in association with addiction-related phenotypes, and with the effectiveness of pharmacotherapy for addiction.

Introduction

Addiction to drugs is a chronic complex relapsing brain disease that causes major medical, social and economic problems and is caused by genetic, epigenetic, environmental, and drug-induced factors. The endogenous opioid system plays a key role in drug addiction, and mediates the analgesic and rewarding properties of drugs. The endogenous opioid family is a network of genes coding for neuropeptide ligands and their cell surface receptors. This system consists of four major subtypes of 7-transmembrane, G protein-coupled opioid receptors: mu, kappa, delta and receptor-like, encoded by distinct genes (OPRM1, OPRK1, OPRD1, and OPRL1), which are stimulated by endogenous opioid peptides: beta-endorphin, prodynorphin, enkephalin, and orphanin/nociceptin, encoded by POMC, PDYN, PENK, and PNOC, respectively, as well as exogenous opiates. The receptors' genes are highly conserved in their 7-transmembrane domain, but not in their amino and carboxyl termini, reflecting on their different ligand binding ability and signal transduction pathways (for figures see LaForge et al. 2000). The ligands' genes all share overall similar structure with a single intron in the coding region. The opioid system is presumed to have been formed by genome duplications early in vertebrate evolution (Cruz-Gordillo et al. 2010; Li et al. 1996). Each receptor gene also produces multiple mRNA isoforms through the use of alternative splicing, alternative promoters (OPRM1, OPRK1), alternative polyadenylation sites (OPRK1), or inclusion of non-coding exons.

Several comprehensive reviews describe the role of the endogenous opioid family of genes in addiction and treatment responses (e.g. Kreek et al. 2005; Kreek et al. 2009; LaForge et al. 2000; Yuferov et al. 2010). In this review we will focus on the genetics of the opioid system that is relevant to addiction to opioids, cocaine, alcohol and methamphetamines, with particular emphasis given to genetic association studies and functional studies. Cannabis or tobacco dependence will not be discussed. We will also discuss the genetics of the opioid system in relation to specific addiction treatments (methadone, buprenorphine, naltrexone and nalmefene), as well as stress responsivity and the hypothalamic-pituitary-adrenal (HPA) axis.

Association studies

A vast number of studies have been reported the association of variants of the opioid system genes and drug addiction-related phenotypes, but the results are not always consistent. This inconsistency may be explained by several factors, including inconsistency in phenotyping, severity of diagnosis, small sample size, inadequate statistics, ethnic heterogeneity and population stratification, large phenotype range, and different diagnostic criteria. The majority of studies have used individual single nucleotide polymorphism (SNP) analysis, and several studies have used hypothesis-based multi-SNP arrays that are based on linkage disequilibrium (LD)-tagging SNPs and capture a substantial proportion of common genetic variation (e.g. Hodgkinson et al. 2008; Levran et al. 2008; Levran et al. 2009; Maher et al. 2011). In this article, we will define SNP as a variation in a single nucleotide of any allele frequency (see dbSNP). Linkage studies and genome-wide association studies are beyond the scope of this review. The studied populations include Asian (Japanese, Chinese, Taiwanese, and Indians) (As); European (E); European American (EA); African (including African Americans, AA); Native American (NA); and Hispanic (His). Some studies have mixed populations and only a few studies have applied methods to control for population stratification.

The studies included a range of phenotypes, including heroin addiction, opioid dependence (OD); heroin-induced subjective response; cocaine addiction/dependence (CD); alcohol dependence (AD), methamphetamine (MAP) dependence/psychosis (MD); amphetamine-induced euphoria; response to alcohol; adolescent alcohol misuse; antisocial drug dependence in adolescents; response to naltrexone (NTX) treatment; and personality traits. For review of association studies of MAP use disorders see (Bousman et al. 2009).

Different studies have used different criteria for defining specific addictions or different criteria for the definition of controls. For example, in our studies we have analyzed heroin addicts that have met the stringent criteria of entering into methadone maintenance treatment (MMT) (at least one year of multiple daily use), while other studies rely on DSM-IV criteria that may be less stringent. Several studies use the more general phenotypes that are based on the concept of common drug use disorders (DUD) liability (Vanyukov et al. 2003), including substance dependence (SD), substance use disorder (SUD), and illicit drug dependence.

To allow a comprehensive and critical examination of the data, and to assist with planning future studies, we have listed the various association studies, with information about the specific addiction or related phenotype, the numbers of cases (or families) and controls, the specific genes analyzed, and the populations studied (Table 1). In Table 2 we provide the list of all the SNPs (sorted by rs number) that were reported to be associated with drug addiction in at least one study. For each SNP the specific addiction and the specific population in which the association was found (indicated under the specific addiction) are indicated. The references listed are all the studies that analyzed the specific SNP including the one/s that identified the association. In addition, we have listed all the SNPs that were analyzed and were not reported to be associated with drug addiction, along with the gene location and the relevant references (Supplement Table 1). In Supplement Table 2 we provide the allele frequencies of all the variants reported in this review in three HapMap populations: CEU (Utah residents with ancestry from northern and western Europe), YRI (Yoruba, Nigeria) and CHB (Han Chinese in Beijing, China).

Table 1. Association studies of drug addiction and related phenotype.

Publication1	Population		Addiction						Controls		Genes
			OD	CD	AD	AD (Fam)2	DD	MD			OPRM1	OPRK1	OPRD1	OPRL1	POMC	PDYN	PENK	PNOC
(Bart et al. 2004)	E	139						170
(Bart et al. 2005)	E			389				170
(Bergen et al. 1997)	EA			100
	E			324
	NA			367
(Bousman et al. 2010)	EA						117	76
(Briant et al. 2010)	EA	100						76
	AA	59						44
	His	94						29
(Chen et al. 2002)	EA		18					43
	AA		49					33
	His		13					9
(Clarke et al. 2009)	As	484						374
(Comings et al. 1999)	EA	31				89		132
(Crowley et al. 2003)	EA	124						100
	AA	89						96
(Dahl et al. 2005)	AA		167					88
(Deb et al. 2009)	As	87		53				82
(Dlugos et al. 2011)	EA							1623
(Drakenberg et al. 2006)	EA	78						40
(Edenberg et al. 2008)	EA			215	219
(Ehlers et al. 2008)	NA			251
(Franke et al. 1999)	E	233						173
(Franke et al. 2001)	E	287			111			365
	E			221	75			365
(Geijer et al. 1997)	E			70				55
(Gelernter et al. 1999)	Mix4	79	202	100				116
(Gerra et al. 2007)	E	106						70
(Glatt et al. 2007)	As				473
(Gscheidel et al. 2000)	E			327				340
(Hoehe et al. 2000)	AA	33	125					51
(Huang et al. 2008)	E			189				167
(Ide et al. 2004)	As						138	213
(Ide et al. 2006)	As						128	232
(Kapur et al. 2007)	As	123						156
(Kim et al. 2004)	As			112				140
(Kobayashi et al. 2006)	As						170	260
(Kranzler et al. 1998)	AA	11	70	39				34
	EA	22	84	201				84
(Levran et al. 2008)	EA	412						184
(Levran et al. 2009)	AA	202						167
(Loh el et al. 2004)	As			158				149
(Luo et al. 2003)	EA					318		179
	AA					124		55
(Maher et al. 2011)	EA					359		398
(Mayer et al. 1997)	E	103						115
(Nishizawa et al. 2006)	As			64				74
(Nomura et al. 2006)	As						143	209
(Racz et al. 2008)	E			247				247
(Sander et al. 1998)	E			327				340
(Schinka et al. 2002)	EA			179				297
(Shi et al. 2002)	As	145						48
(Smith et al. 2005)	AA	76						71
	EA	94						57
(Szeto et al. 2001)	As	200						97
(Town et al. 1999)	EA			105				122
(Wei et al. 2011)	As	304						300
(Williams et al. 2007)	EA		26					53
	AA		110					49
	His		25					12
(Xu et al. 2002)	As	450			219			304
(Xuei et al. 2006)	EA				219
(Xuei et al. 2007)	EA	83			219	508		832
(Xuei et al. 2008)	EA	83				508		832
(Yuferov et al. 2004)	EA	65						64
	AA	36						34
	His	35						25
(Yuferov et al. 2009)	EA		82					65
	AA		204
(Zhang et al. 2007)	As	3365
(Zhang et al. 2006)	EA	91	171	318				338
(Zhang et al. 2008)	EA	111	225	557				443
(Zhang et al. 2009)	AA	455			319			199
	EA	336			313			483

¹The list is sorted by alphabetical order of the first authors; DD; Illicit drug dependence (marijuana, cocaine, stimulant, sedative or opioid), substance dependence, or substance use disorder;

²COGA families with AD;

³subjective response to amphetamine;

⁴EA, AA, His;

⁵heroin-induced positive responses on first use. Abbreviation: As; Asian; E; European; EA; European Americans; AA; African Americans; His; Hispanics; NA; Native Americans.

Table 2. Opioid system gene variants reported to be associated with specific drug addictions.

SNP	Location	Addictions						References7
		OD	CD	AD	DD	MD	Other
		OPRM1
rs1799971	exon 1	As, E, EA		As, E, EA				(e.g. Bart et al. 2004; Bart et al. 2005. For full list see text)
rs1799972	exon 1						AA1	(Levran et al. 2008; Levran et al. 2009; Crawley et al. 2003; Crystal et al. 2011; Kapur et al. 2007; Xu et al. 2002; Glatt et al. 2007; Hoehe et al. 2000)
rs1461773	intron 1			AI				(Ehlers et al. 2008)
rs2075572	intron 2			AI		As		(Zhang et al. 2006; Ehlers et al. 2008; Ide et al. 2006; Zhang et al. 2007; Xuei et al. 2007; Hernandez-Avila et al. 2007; Bergen et al. 1997)
rs2281617	3′ region						EA5	(Levran et al. 2008; Levran et al. 2009; Dlugos et al. 2011)
rs3778148	intron 1			AI				(Ehlers et al. 2008)
rs3778151	intron 1	EA, As						(Levran et al. 2008; Levran et al. 2009; Zhang et al. 2007)
rs3823010	intron 1				EA			(Zhang et al. 2006)
rs483481	3′ region				EA			(Maher et al. 2011)
rs495491	intron 1			EA	EA			(Zhang et al. 2006; Xuei et al. 2007; Hernandez-Avila et al. 2007)
rs510769	intron 1						EA2	(Levran et al. 2008; Levran et al. 2009; Dlugos et al. 2011; Zhang et al. 2007; Xuei et al. 2007)
rs524731	intron 1			AI				(Zhang et al. 2006; Ehlers et al. 2008; Xuei et al. 2007)
rs534673	intron 1	As						(Zhang et al. 2007) originally called rs696522
rs548646	intron 3			AI				(Zhang et al. 2006; Ehlers et al. 2008; Xuei et al. 2007)
rs553202	intron 1			AI				(Ehlers et al. 2008)
rs648007	intron 3			AI				(Ehlers et al. 2008; Xuei et al. 2007)
rs648893	intron 3			EA				(Zhang et al. 2006; Ehlers et al. 2008; Xuei et al. 2007; Hernandez-Avila et al. 2007)
rs609148	intron 3			EA				(Zhang et al. 2006; Ide et al. 2006; Xuei et al. 2007; Hernandez-Avila et al. 2007)
rs681243	intron 3			AI				(Ehlers et al. 2008)
rs9479757	intron 2						As3; 4	(Ide et al. 2006; Shi et al. 2002)
OPRK1
rs1051660	exon 2	E, His						(Yuferov et al. 2004; Gerra et al. 2007; Xuei et al. 2006; Zhang et al. 2008; Loh el et al. 2004)
rs12548098	intron 2		EA					(Xuei et al. 2006)
rs16918941	intron 2		EA					(Xuei et al. 2006)
rs35991105	5′ region			EA				(Edenberg et al. 2008)
rs6473797	intron 2	EA	EA					(Levran 2008; 2009; Xuei 2006)
rs6985606	intron 2		EA					(Xuei et al. 2006; Zhang et al. 2008)
rs997917	intron 2		EA					(Xuei et al. 2006; Zhang et al. 2008)
OPRD1
rs1042114	exon 1	EA						(Zhang et al. 2008; Xuei et al. 2007; Xu et al. 2002)
rs204055	intron 1			EA				(Levran et al. 2008; Levran et al. 2009; Xuei 2007)
rs2234918	exon 3	E			EA			(Mayer et al. 1997; Franke et al. 1999; Xu et al. 2002; Loh el et al. 2004; Zhang et al. 2008; Levran et al. 2008; Levran et al. 2009; Kobayashi et al. 2006)
rs2236857	intron 1	EA						(Levran et al. 2008; Levran et al. 2009; Zhang et al. 2008; Xuei et al. 2007)
rs2236861	intron 1	EA						(Levran et al. 2008; Levran et al. 2009)
rs2298896	intron 1	EA						(Levran et al. 2008; Levran et al. 2009; Zhang et al. 2008)
rs3766951	intron 1	EA						(Levran et al. 2008; Levran et al. 2009)
OPRL1
rs6010718	intron 1			E				(Huang et al. 2008)
rs6090041	intron 1	EA,AA						(Briant et al. 2010; Levran et al. 2008; Levran et al. 2009; Xuei et al. 2008)
rs6090043	intron 1	EA,AA						(Xuei et al. 2008; Briant et al. 2010; Levran et al. 2008; Levran et al. 2009)
rs6512305	intron 1	EA						(Xuei et al. 2008; Levran et al. 2008; Levran et al. 2009)
POMC
rs1009388	intron 1	EA						(Xuei et al. 2007)
rs1042571	3′ UTR	EA						(Rutters et al. 2011; Xuei et al. 2007)
rs12473543	intron 2	EA						(Xuei et al. 2007)
rs1866146	3′ region	EA, AA	EA, AA	EA				(Zhang et al. 2009)
rs3769671	Intron 1			E				(Racz et al. 2008; Levran et al. 2008; Levran et al. 2009; Zhang et al. 2009)
rs6713532	Intron 3		EA	EA				(Zhang et al. 2009; Levran 2008; 2009; Xuei 2007)
rs6719226	5′ region	AA						(Zhang et al. 2009; Levran et al. 2008; Levran et al. 2009)
rs934778	Intron 1	EA		EA				(Xuei et al. 2007)
PDYN
rs1022563	3′ region	As						(Clarke et al. 2009; Wei et al. 2011)
rs10485703	3′ UTR		EA	EA				(Levran et al. 2008; Levran et al. 2009; Xuei et al. 2006; Yuferov et al. 2009)
rs10854244	5′ region			EA				(Xuei et al. 2006)
rs1997794	5′ region	As		EA				(Levran et al. 2008; Levran et al. 2009; Xuei et al. 2006; Geijer et al. 1997; Clarke et al. 2009; Yuferov et al. 2009)
rs2235749	3′ UTR	As	EA	EA				(Taqi et al. 2011; Wei et al. 2011; Yuferov et al. 2009; Xuei et al. 2006)
rs35286281	5′ region		AA, His			As		(Wei et al. 2011; Chen et al. 2002; Dahl et al. 2005; Williams et al. 2007; Nomura et al. 2006)
rs6035222	intron 3			EA				(Levran et al. 2008; Levran et al. 2009; Xuei et al. 2006)
rs6045784	3′ region			EA				(Xuei et al. 2006)
rs6045819	exon 4		EA	EA				(Xuei et al. 2006; Yuferov et al. 2009)
rs6045868	intron 2			EA				(Xuei et al. 2006)
rs910079	3′ UTR	As	EA	EA				(Yuferov et al. 2009; Wei et al. 2011)
rs910080	3′ UTR		EA	EA				(Xuei et al. 2006; Yuferov et al. 2009)
PENK
rs12545109	3′ region			E				(Racz et al. 2008)
rs1437277	intron 3	EA						(Xuei et al. 2007; Levran et al. 2008; Levran et al. 2009)
rs1975285	intron 2	EA						(Xuei et al. 2007; Levran et al. 2008; Levran et al. 2009)
rs2576581	5′ region			E				(Racz et al. 2008)
rs3138832	3′ region						E6	(Nikoshkov et al. 2008)
rs3219515	5′ region	EA						(Comings et al. 1999)
PNOC
rs17058952	5′ region			EA				(Xuei et al. 2008)
rs351779	3′ region			EA				(Xuei et al. 2008)
rs4732636	5′ region				EA			(Xuei et al. 2008)

The SNPs are sorted by rs number for each gene.

Abbreviations: OD; opioid dependence; CD; cocaine dependence; AD; alcohol dependence; MD; amphetamine dependence; As; Asian; E; European; EA; European Americans; AA; African Americans; His; Hispanics; NA; Native Americans. Other:

¹quantitative measure for cocaine and alcohol use;

²Response to amphetamine;

³MAP psychosis;

⁴Heroin intake dose;

⁵Euphoria; energy and stimulation after amphetamine use;

⁶heroin abuse.

⁷The references listed are all the studies that analyzed the specific SNP, including the one/s that identified the association in the specific population indicated in the specific addiction column.

The mu opioid receptor gene (OPRM1)

The G protein-coupled mu opioid receptor (encoded by the OPRM1 gene) is the major site of action for endogenous opioids, opiate and opioid analgesic drugs, and exogenous opioid drugs such as methadone, heroin and morphine (Kreek 2005). The receptor mediates the action of non-opioid drugs of abuse (e.g. alcohol, nicotine) and the stress-responsive HPA axis. Receptor activation results in the opening of G protein-gated inwardly-rectifying K+ (GIRK) channels, inhibition of voltage-gated Ca²⁺ channels, and reduction of adenylyl cyclase-mediated cAMP production. Binding of beta-endorphin results in disinhibition of dopaminergic neurons and this dopamine influx has been associated with reward and reinforcement and is believed to contribute to the development of drug dependence.

The OPRM1 gene is located at cytogenetic band 6q25.2. The major subtype contains 4 exons. A number of alternatively-spliced variants have been reported in rodents and humans. Two human receptor variants (encoded by splice variants) (hMOR-1X and hMOR-1O) were identified (Pan 2003). An alternatively-spliced exon with a specific promoter was identified -28 kb upstream of exon 1 (Xu et al. 2009). An alternatively-spliced exon with an alternative promoter was identified in rodents and humans (Shabalina et al. 2009). Notably, the genomic organization of the human OPRM1 locus is highly similar to the mouse locus. However, alternative-splicing events display some substantial differences between human and mouse.

Only two SNPs out of the variants described in dbSNP in the OPRM1 coding sequence (17C>T (Ala6Val) and 118A>G (Asp40Asn) in exon 1) are relatively common at least in one population (see sections below). Numerous SNPs have been described in the regulatory region, some of which are population-specific (Hoehe et al. 2000; Ono et al. 2009b), and several have been functionally characterized. Two promoter polymorphisms, -554G>A and -1320A>G, have been shown to affect transcription (Bayerer et al. 2007). SNPs -554G>A, in a STAT6 binding site, and SNP -995C>A, in a nuclear factor (NF)-kB binding site, decreased the amount of transcription factor binding and the transcriptional activity (Kraus et al. 2001; Kraus et al. 2003). SNP -1793T>A is located in an YY1 transcription binding site; SNP -1699insT is located in an AP-1 binding site (Hoehe et al. 2000). The Poly (ADP-ribose) polymerase-1 (PARP-1) was shown to preferentially bind to the -172T allele and positively regulate OPRM1 gene expression (Ono et al. 2009a). Bioinformatics assessment suggests that C/EBP or CREB binds to the SNP -1748G>A region, OCT-1 binds to the SNP -1565T>C region and the SNP -1045A>G region, and GATA, MZF, and SP1 bind to the SNP -172G>T region (Ono et al. 2009b).

Functional evidence for a few coding variants was reported. The 118A>G SNP was the first variant that was shown to be functional by the combined groups of Kreek and Yu (Bond et al. 1998) (see section below). Three rare SNPs in highly evolutionary conserved amino acids 779G>A (Arg260His), 794G>A (Arg265His), and 802T>C (Ser268Pro) in the third intracellular loop of the receptor were shown to decrease receptor signaling activity (Befort et al. 2001; Wang et al. 2001). SNP 802T>C was also shown to lose Ca²⁺/calmodulin-dependent protein kinase-induced receptor desensitization (Koch et al. 2000).

118A>G (Asn40Asp, rs1799971)

The most studied OPRM1 variant is the non-synonymous (changing an amino acid) variant rs1799971 (118A>G, Asn40Asp) that was shown to remove a potential N-glycosylation site in the extracellular domain, to be more potent in beta-endorphin binding, and to reduce receptor signaling efficacy (Bond et al. 1998). The 118G allele is common in persons of European (15–30%) and Asian ancestry (40–50%), and has lower prevalence in African Americans (AA) (1-3%) (Deb et al. 2010; Gelernter et al. 1999). The variant 118G allele is not found in the African HapMap population (Yoruba, Nigeria), suggesting that it arose after the ‘out-of-Africa’ migration (Cavalli-Sforza and Feldman. 2003).

The functionality of this variant was subject to numerous studies (e.g. Befort et al. 2001; Bond et al. 1998; Deb et al. 2010; Filbey et al. 2008; Margas et al. 2007; Ray et al. 2011). The 118G variant was associated with lower cell-surface receptor binding site availability (Beyer et al. 2004; Kroslak et al. 2007), reduced mRNA levels in vivo and in vitro (Zhang et al. 2005), altered signal transduction (PKA and pERK), in vitro (Deb et al. 2010), and enhanced affinity for beta-endorphin (Bond et al. 1998). For more information and references, see a current review (Mague and Blendy 2010).

Recently, in a positron emission tomography (PET) study of 22 smokers and 20 controls, the 118G variant was associated with reduced receptor binding availability in the bilateral amygdala, left thalamus, and left anterior cingulate cortex (Ray et al. 2011). An fMRI study showed that the 118G allele is associated with increased cue-elicited activation of mesocorticolimbic structures and that this activation is extended by a priming dose of alcohol. This activity in the striatum was correlated with drinking behavior in individuals with the 118G allele (Filbey et al. 2008; Ramchandani et al. 2011). Taken together, some results may be interpreted as a loss-of-function while others may be interpreted as a gain-of-function of the 118G receptor. These results are not mutually exclusive and it is possible that the variant causes reduction in receptor numbers but also enhanced binding. It is also possible that some results reflect the action of another SNP that is in high LD with the 118G allele in some chromosomes (haplotypes) but not in others. The exact nature of the physiological changes has yet to be elucidated.

Association studies of 118A>G

A large number of association studies of this SNP across various phenotypes were reported (for a recent review see Mague and Blendy 2010). The 118G allele has been implicated in drug addiction, stress responsivity, and in treatment responses. Several association studies of 118A>G reported negative results, and meta-analyses revealed mixed results (Arias et al. 2006; Glatt et al. 2007). The118G allele was shown to be associated with opioid dependence (OD) and other substance dependencies in several studies in diverse populations. An association was observed between the 118G allele and opioid addiction in Swedish (Bart et al. 2004), Chinese (Szeto et al. 2001), EA (Drakenberg et al. 2006), and Indian patients (Deb et al. 2010; Kapur et al. 2007; Tan et al. 2003), whereas other studies did not detect such association (Franke et al. 2001; Glatt et al. 2007; Levran et al. 2008; Levran et al. 2009; Shi et al. 2002).

Several studies reported an association of the 118G allele with alcohol dependence (AD) in different populations (e.g. Bart et al. 2005; Deb et al. 2010; Kim et al. 2004; Nishizawa et al. 2006; Rommelspacher et al. 2001; Schinka et al. 2002), and with the development of alcohol use disorder diagnoses during adolescence in EA drinkers (Miranda et al. 2010). Carriers of the 118G allele were more sensitive to the euphoric effects of alcohol, more likely to abuse alcohol and had greater cue-induced alcohol craving during neutral condition, when compared to stressful imagery condition (Ray 2011; Ray and Hutchison 2004). Increased dopaminergic sensitivity (assessed by using the apomorphine induced growth hormone secretion as indicator) was reported in abstinent alcoholic individuals who were 118G carriers (Smolka et al. 1999). Several studies did not find association with this variant (e.g. Bergen et al. 1997; Gscheidel et al. 2000; Loh el et al. 2004; Sander et al. 1998; Town et al. 1999).

The SNP 118A>G was also associated with social hedonic capacity (Troisi et al. 2011) and with MAP psychosis, having latency less than three years from first methamphetamine intake in Japanese subjects (Ide et al. 2004).

As the 118A>G SNP changes an amino acid, it has been generally assumed that it is the causative variant for the phenotypes associated with it. There is a possibility that other variants, which are in LD with 118A>G, contribute to these phenotypes. The mixed results for 118A>G may be explained in part by different and/or population-specific haplotype patterns. The 118G allele was shown to be represented by a specific haplotype that includes several SNPs from the 5′ flanking region of the gene in EA subjects, but appears in two haplotypes in the Asian HapMap population. The major haplotype is similar to the one found in EA subjects and the second one includes other SNP combinations (Levran et al. 2011).

Animal models of 118A>G

Two mouse models of 118A>G with the equivalent amino acid substitution have been developed using two different approaches. In one model, knock-in mice homozygous for the equivalent (112G) allele (Asn38Asp) exhibited reduced oprm1 mRNA and protein levels in multiple brain regions and sex-specific reductions in the rewarding properties of morphine (Mague et al. 2009). In the second model, two humanized mice lines were produced by introduction of human exon 1 with and without the variant 118G into the mouse gene. Mice bearing the 118GG genotype showed a greater peak of striatal dopamine response to an alcohol challenge (Ramchandani et al. 2011).

In rhesus macaques, SNP 77C>G (Pro26Arg) in the N-terminal appears to be functionally equivalent to the human 118A>G. Carriers of the 77G variant showed increased psychomotor stimulation in response to alcohol, increased frequency of alcohol consumption to intoxication, increase of alcohol preference following naltrexone treatment, altered stress reactivity, reduced cortisol response to maternal separation in infancy and to acute alcohol exposure later in life and attenuated cortisol levels during the maternal postpartum period (Barr et al. 2010; Barr et al. 2007; Miller et al. 2004; Schwandt et al. 2011; Vallender et al. 2010).

17C>T (Ala6Val, rs1799972)

The function of the 17C>T polymorphism in the extracellular space of the N-terminal is yet unknown. The frequency of the 17T allele varies significantly between populations. It is very rare in subjects of European descent, Hispanics, Middle Eastern subjects (Gelernter et al. 1999), and Asian subjects (dbSNP) (Tan et al. 2003; Xu et al. 2002) and common (15–20%) in subjects of African descent (Crowley et al. 2003; Gelernter et al. 1999; Luo et al. 2003). A relatively high allele frequency was also reported in Indian males from Delhi, India (Kapur et al. 2007).

No association was found with the17T allele and opioid dependence in an Indian sample (Kapur et al. 2007), nor in an AA sample (Crowley et al. 2003). A recent large study of HIV+ and HIV- women identified association between the 17TT genotype and a quantitative measure (Kellogg et al. 2003)) for cocaine and alcohol use in AA (Crystal et al. 2010).

Other OPRM1 SNPs

Several studies analyzed other SNPs in the OPRM1 gene for association with addiction. An association with substance dependence (SD) (heroin, cocaine and alcohol) was shown with three promoter SNPs (-1793T>A, -1699insT and -2044C>A) (Hoehe et al. 2000; Luo et al. 2003). SNP rs1074287, located in the 5′ upstream region (-11.6 kb from exon 1) was found to be associated with heroin addiction in EA subjects (Nielsen et al. 2008).

Several SNPs in intron 1 were associated with drug addiction. In a study of 9 SNPs in Chinese subjects, heroin-induced positive response on first use was associated with SNPs rs534673 (originally called rs696522), rs1381376, and rs3778151 (Zhang et al. 2007). SNPs rs510769 and rs3778151 were found to be associated with heroin addiction in EA subjects, but not in AA subjects (Levran et al. 2008; Levran et al. 2009). SNP rs510769 was also associated with several effects of amphetamine (euphoria, stimulation and blood pressure) in healthy subjects of EA ancestry (Dlugos et al. 2011).

Haplotype analysis revealed that SNPs rs510769 and rs3778151 are part of a haplotype block that spans intron 1. Several SNPs at the -20 kb 5′ region (including SNP rs1074287) are in high LD with this haplotype block (Levran et al. 2011). Several lines of evidence suggest a special functionality to intron 1 (Choi et al. 2006; Shabalina et al. 2009). A potential functional SNP in intron 1 (rs563649) is shown to be located within a structurally conserved internal ribosome entry site (IRES) in the 5′ UTR of a novel exon 13 in an OPRM1 isoform and to affect both mRNA levels and translation efficiency. This SNP was also associated with pain perception (Shabalina et al. 2009). A (CA) repeat in intron 1 was associated with substance dependence (alcohol, cocaine or opioid) in EAs, but not in AAs (Kranzler et al. 1998).

Two SNPs in the 3′ region near the gene (rs483481, rs2281617) showed association with drug use disorder diagnosis (Maher et al. 2011) and response to amphetamine use (euphoria, energy and blood pressure) in healthy EA subjects (Dlugos et al. 2011), respectively.

Several intronic SNPs were associated with response to alcohol in Native Americans (NA), which is correlated with less susceptibility to dependence (Ehlers et al. 2008). In a study of methamphetamine dependence/psychosis (Ide et al. 2004; Ide et al. 2006), 16 variants in the 5′ regulatory region and intron 1 were studied in Japanese subjects. Association was found with SNP rs2075572 (IVS2+691G>C).

No association with opioid dependence was found in two studies of AA and EA subjects analyzing SNPs -1793T>A, -1699insT, -1320A>G, 17C>T, 118A>G, as well as SNPs 540825 and 562859 (Crowley et al. 2003; Smith et al. 2005).

The kappa opioid receptor gene (OPRK1)

The kappa opioid receptors have widespread distribution in the central nervous system and play a role in a wide variety of physiological systems, including pain regulation, addiction to drugs of abuse, neuroendocrine regulation, cardiovascular function, respiration, temperature regulation, feeding behavior, and stress responsivity (e.g. Bruchas et al. 2010; Knoll and Carlezon 2010; Kreek et al. 2005). Kappa opioid receptors play an important role in modulation of opioid, cocaine and other rewarding stimuli, presumably through modulation of basal and drug-induced dopaminergic tone (Kreek et al. 2002). Oprk1 knockout mice showed elevated basal dopamine release in the nucleus accumbens and enhanced cocaine-induced dopamine levels compared to wild-type mice, suggesting association with greater vulnerability to cocaine abuse (Chefer et al. 2005) (also see PDYN section below).

OPRK1 contains four exons and is located on chromosome 8q11.2 (Yuferov et al. 2004). Twenty-seven SNPs are described in the coding sequence, out of which twelve are rare non-synonymous SNPs that alter an amino acid in the protein. Four synonymous SNPs that do not modify an amino acid but may have other functions (rs1051660 (36G>T), rs702764, rs16918875 and rs7815824) are common in at least one population. An indel of net insertion of 830 bp (rs35566036) was described at the 5′ flanking region (-1986 bp) and was shown to reduce transcription in transient transfection assays, in vitro (Edenberg et al. 2008).

Four studies reported association of OPRK1 SNPs with alcohol dependence (AD); a study of a Taiwanese sample reported no association for SNPs rs1051660 and rs702764 (843A>G) (Loh el et al. 2004); in a study of 13 SNPs in 219 EA COGA families, an association of several SNPs in intron 2 was reported (Xuei et al. 2006). A third study of seven SNPs (not including SNPs from intron 2) in EA subjects reported an association of three SNPs (rs1051660, rs6985606 and rs997917) with AD or cocaine dependence (CD), and a seven-SNP haplotype with AD (Zhang et al. 2008). An association of the 830 bp indel (rs35566036) was reported in the COGA sample (Edenberg et al. 2008).

Four studies reported association of OPRK1 SNPs with OD. A study of eight SNPs in EA, AA and His subjects, reported association with the synonymous SNP rs1051660 (Yuferov et al. 2004). A study of an Italian sample replicated this finding (Gerra et al. 2007). A third study of 11 SNPs (not including SNP rs1051660) reported an association with the intronic SNP rs6473797 in EA subjects (Levran et al. 2008). No association with these 11 SNPs was detected in AA subjects (Levran et al. 2009).

The delta opioid receptor gene (OPRD1)

Delta opioid receptors bind enkephalin as its endogenous ligand. Delta opioid receptors have been implicated in the modulation of reward, addiction, affective state, pain perception, and analgesia. A blunted ability to form and/or retrieve drug-context associations was shown in Oprd1−/− mice (Le Merrer et al. 2011).

OPRD1 contains 3 exons and is located on chromosome 1p36. Nine polymorphisms were identified in the coding region (four non-synonymous, three synonymous and one insertion), out of which only one SNP (921T>C, rs2234918) is common in several ethnicities. SNP rs1042114 (80G>T, Phe27Cys) is common mainly in Europeans, and the synonymous SNP rs118175398 was described in Asians. In addition, two SNPs were described in the 3′ UTR, out of which SNP rs4654327 is common in several ethnicities.

The promoter SNP rs569356 is located ∼2 Kb upstream to the transcription start site and is in high LD with SNP rs1042114 (80G>T, Phe27Cys). Functional characterization of SNP rs569356 (A>G) with luciferase reporter gene assay in HEK293 cells demonstrated that the G-allele can enhance OPRD1 promoter activity under basal conditions. Electrophoretic mobility shift assay (EMSA) with human brain nuclear proteins showed enhanced DNA protein binding of the probe with the G-allele (Zhang et al. 2010).

The non-synonymous SNP rs1042114 (80G>T, Phe27Cys) changes the evolutionary conserved phenylalanine to cysteine in the extracellular N terminus. Interestingly, the reference 80G-allele is the minor allele and is absent in Asians, rare in Africans, and has a minor allele frequency (MAF) of 0.1, in Europeans. One of the possible explanations for this phenomenon may be selective advantage for the 80T-allele, but the biological mechanism underlying this potential advantage has yet to be identified. The minor 80G-allele has been found to be associated with OD in EA subjects (Zhang et al. 2008). These results were not found in a family-based association study of 18 SNPs with AD (219 EA families), a case control association study with “illicit drug dependence”, and in a small subsample with OD (Xuei et al. 2007). This SNP was not included in the array used in our studies of heroin addiction (Levran et al. 2008; Levran et al. 2009). Functional studies revealed that the two variants (27Phe and 27Cys) have identical pharmacological properties, but differ in maturation efficiency, stability at the plasma membrane, and Ca²⁺ signaling regulation (Leskela et al. 2009; Tuusa and Petaja-Repo 2011).

The C-allele of the synonymous SNP rs2234918 (921T>C) in exon 3 was reported to be associated with heroin dependence in Germans (Mayer et al. 1997), but this result was not found in several other studies of this SNP in Germans (Franke et al. 1999), Han Chinese (Shi et al. 2002; Xu et al. 2002), EAs (Levran et al. 2008; Zhang et al. 2008), and AAs (Levran et al. 2009). No association of this SNP was found with AD in Taiwanese Han (Loh el et al. 2004) or with methamphetamine dependence/psychosis in Japanese (Kobayashi et al. 2006). A haplotype, which harbors the 80G-allele and the 921C-allele, was associated with OD, CD, and AD (Zhang et al. 2008).

In a study of 11 OPRD1 SNPs, three common SNPs in intron 1 (rs2236861, rs2236857 and rs3766951) and a haplotype block (SNPs rs204055, rs2236857 and rs2298896) showed associations with heroin addiction in EAs (Levran et al. 2008). No associations of these SNPs and heroin addiction were found in AAs (Levran et al. 2009).

The opiate receptor-like 1 (nociceptin/orphanin FQ receptor) (OPRL1)

The opiate receptor-like 1 receptor for the neuropeptide nociceptin shows high homology to opioid receptors and plays an important role in inhibition of the rewarding effects of addictive drugs. The opiate receptor-like 1 receptor couples to inhibitory G proteins and negatively regulates the function of the mesolimbic dopaminergic system. Studies suggest that the endogenous nociceptin system has a role in mediating responses to alcohol (Murphy 2010). Oprl1 knockout rats are more sensitive to the rewarding effect of morphine than wild-type controls (Rutten et al. 2011). Oprl1 knockout mice displayed increased anxiety-related behavior (Gavioli et al. 2007) and nociceptin blocked the development of cocaine-induced locomotor sensitization in mice (Bebawy et al. 2010). In humans, reduction in PNOC mRNA (1.7-fold) in the hippocampus and OPRL1 mRNA (1.4-fold) in the central amygdala of postmortem brain of alcoholics, compared with controls, was reported (Kuzmin et al. 2009).

OPRL1 is located on chromosome 20q13 and has alternative splicing transcript variants that are controlled by alternate-promoter mechanism and contain 4 or 5 exons (Ito et al. 2000; Wick et al. 1995). The protein is encoded by three exons. Seventeen coding SNPs have been described, of which only the synonymous SNP rs2229205 is common in several populations.

The regulator of G protein signaling 19 protein (RGS19/GAIP) which has been shown to modulate signaling of the mu opioid receptor, is located immediately upstream of OPRL1, and is transcribed from the opposite strand. The two genes (RGS19, OPRL1) share their promoter (Ito et al. 2000; Xie et al. 2007). Two unique repeat polymorphisms were described in the promoter region but no significant effect on promoter activities was found (Ito et al. 2000).

Three association studies of OPRL1 and drug addiction were reported. In the first study, 10 SNPs covering OPRL1 (and RGS19) were analyzed and no association was found with alcohol addiction or illicit drug dependence. Two intronic SNPs in high LD (rs6512305 and rs6090043) were marginally associated with OD in a small subsample of 83 affected subjects (Xuei et al. 2008). In the second study, 15 SNPs were analyzed and one SNP (rs6010718) showed an association with AD (Huang et al. 2008). A haplotype of five tag SNPs TTTGC (rs6090043, rs6010718, rs7271530, rs2295448, and rs6089789) was significantly more common in cases than in controls (Huang et al. 2008). In the third study of five SNPs in three ethnicities (EA, AA and His), association of two SNPs in the 5′ flanking region (rs6090041 and rs6090043) with vulnerability to develop heroin addiction was reported in EA only, and a haplotype was associated with heroin addiction in AA and EA (Briant et al. 2010). No association with heroin addiction was detected for six SNPs (including rs6512305, rs6090041 and rs6090043) in a larger cohort of EA and AA from our laboratory. A trend toward association was detected for SNP rs6090041 in EA (Levran et al. 2008; Levran et al. 2009).

The endogenous opioid neuropeptides

The endogenous opioid peptides acting at their cognate opioid receptors modulate the effects of many drugs of abuse. The endogenous opioid peptides, beta-endorphin, dynorphin, enkephalin, and orphanin/nociceptin, are derived from precursors encoded by proopiomelanocortin (POMC), prodynorphin (PDYN), proenkephalin (PENK), and nociceptin/orphanin FQ (PNOC), respectively (for figures see LaForge et al. 2000).

The proopiomelanocortin gene (POMC)

POMC is a polypeptide precursor protein with 241 amino acid residues. Ten different peptides can be derived from POMC through tissue-specific posttranslational processing, including adrenocorticotropin (ACTH) and beta-endorphin, which are principal components of the HPA axis. Studies in rodents have shown that stressors elevate POMC mRNA levels in the pituitary. POMC-derived peptides actively regulate drug-related behaviors (Kiefer et al. 2002; O'Malley et al. 2002). Beta-endorphin is a 31-amino acid peptide that is the major endogenous ligand of the mu opioid receptor and is found mainly in neurons of the hypothalamus and the pituitary gland (Dores and Baron et al. 2011).

The POMC gene is located on chromosome 2p23 and all the variants in the coding regions are rare. In one reported association study, seven SNPs were genotyped in a sample of alcoholic families (COGA) and association was found for intronic SNP rs934778 and general illicit drug dependence. Two intronic SNPs (rs934778 and rs1009388) were associated with opioid dependence in a small subsample of this cohort (Xuei et al. 2007). In a second association study of AD in European sample, nine SNPs were analyzed. Three SNPs were associated with AD, out of which SNP rs934778 was associated after Bonferroni correction. The T-A haplotype (rs934778 and rs3769671) was associated with AD in women only (Racz et al. 2008). A third association study of five SNPs with AD, CD or OD was reported in AA and EA (Zhang et al. 2009). The main finding is of SNP rs1866146 in the 3′ flanking region that was associated with OD or CD in AAs, and with AD, CD or OD in EAs. In addition, the intronic SNP rs6713532 was marginally associated with AD or CD in EAs, and SNP rs6719226 in the 5′ flanking region was associated with OD in AAs (Zhang et al. 2009). In another study, SNP rs1042571 in the 3′ UTR was associated with higher cortisol exposure and reduced negative feedback of the HPA axis, in a nonchallenged condition, in women (Rutters et al. 2011).

The prodynorphin gene (PDYN)

Prodynorphin is the precursor for the opioid peptides alpha- and beta-neoendorphins, dynorphin A and dynorphin B, which are endogenous ligands for the kappa opioid receptor. Dynorphin peptides decrease basal and drug-induced dopamine levels in several areas of the dopaminergic, nigrostriatal, and mesolimbic–mesocortical systems. Expression of the PDYN gene is increased by cocaine (for a recent review see Yuferov et al. 2010). Pdyn knockout mice showed increased explorative behavior in anxiety tests demonstrating the anxiogenic role of prodynorphin-derived peptides (Wittmann et al. 2009).

The PDYN gene contains four exons and is located at chromosome 20p13. Exons 1 and 2 encode the 5′ UTR, exon 3 encodes a signal peptide, and exon 4 encodes the dynorphin peptides. Nine SNPs were described in the coding sequence of which only one synonymous SNP in exon 4 (rs6045819) is not rare. Seven alternative transcripts with different 5′-ends were detected in brain tissue; some have novel initiation sites and some contain new exons (Nikoshkov et al. 2005; Telkov et al. 1998).

One of the first discovered and most studied PDYN polymorphisms is rs35286281 a variable number nucleotide repeat (VNTR) of 1-5 copies of 68-bp tandem which contains a putative AP-1 transcription complex binding site, located 1250 bp upstream of exon 1. SNP rs61761346 was identified at the ninth nucleotide of each repeat (Rockman et al. 2005; Rouault et al. 2011). In vitro, PDYN expression studies using various constructs, different cell lines, and different species reported opposite effects (Babbitt et al. 2010; Rouault et al. 2011; Zimprich et al. 2000).

Several association studies of PDYN SNPs and different addictions were reported. Nine out of the eighteen PDYN SNPs tested in the COGA cohort (including rs1997794) were associated with AD (Xuei et al. 2006). The results were supported by haplotype analysis. SNP rs1997794 was reported to show no association with alcoholism in a small European sample (Geijer et al. 1997). Three SNPs were analyzed in Chinese subjects for association with OD (Clarke et al. 2009). No association in the overall sample, but an interaction between sex and genotype distribution was detected for rs1997794, in females. Also, SNP rs1022563 was associated with OD in females, suggesting a gender-specific role of PDYN. Association between a haplotype (SNPs rs1022563, rs2235749 and rs910080) and OD was reported in Han Chinese (Wei et al. 2011). Several studies of SNP rs35286281 (the 68-bp VNTR) did not produce consistent results for association with OD, CD or MD in several ethnic groups (Chen et al. 2002; Dahl et al. 2005; Nomura et al. 2006; Ray et al. 2005; Wei et al. 2011; Williams et al. 2007). An association between three SNPs (rs910080, rs910079 and rs2235749) in the 3′ UTR, and the haplotype CCT, with both CD and CD/AD, was reported in EAs, but not in AAs (Yuferov et al. 2009). Allele-specific gene expression of PDYN, using SNP rs910079 as a reporter, in postmortem human brains, showed lower expression for the C-allele in the caudate and nucleus accumbens. Total PDYN expression was also lower in carriers of the CCT haplotype (Yuferov et al. 2009).

Three SNPs (rs1997794, rs6045819 and rs2235749) that showed association with AD, CD or OD form a CpG dinucleotide (methylation associated SNPs). Alterations in methylation of a specific allele through epigenetic mechanisms may affect transcription. An increase in methylation levels of the C- allele of the 3′ UTR SNP rs2235749 was reported in the dorsolateral prefrontal cortex in alcohol-dependent subjects (Taqi et al. 2011). A 63 kDa protein showed differential binding affinity for the risk T-allele, and the methylated and unmethylated C-allele in the brain. The T-allele of SNP rs1997794 resides within a noncanonical AP-1-binding element and was associated with lower PDYN expression in brain cortical areas. The C-allele was shown to abrogate AP-1 binding (Yuferov et al. 2009).

The proenkephalin gene (PENK)

Proenkephalin-derived peptides act on mu and delta opioid receptors to produce rewarding actions of substances of abuse in several brain regions, including the ventral tegmental area and nucleus accumbens. Specific inbred mouse strains (e.g. DBA/2J and SWR) are characterized by relative insensitivity to drug reward and reinforcement, compared to C57BL/6J. These differences were shown to parallel inter-strain differences in basal proenkephalin expression in the nucleus accumbens (Gieryk et al. 2010).

The PENK gene is located on chromosome 8q23. Forty-six coding SNPs have been reported, most of them are rare and some are population-specific. A (CA)_n repeat (rs3219515) in the 5′ flanking region was reported to be associated with OD (Comings et al. 1999). Seven SNPs were genotyped in a sample of alcoholic families (COGA) and no association was detected with AD or general illicit drug dependence. Three SNPs (rs2609997 and rs1975285, in the 5′ flanking region, and the intronic rs1437277) provided evidence of association with opioid dependence in a small subsample (Xuei et al. 2007). In another study of a different (CA)_n repeat (rs3138832) in the 3′ flanking region, a significant association of a specific allele (CA)₇₉ (based on PCR product described in Weber and May 1990), with heroin abuse was reported in a European sample. This allele was also associated with higher striatal PENK mRNA expression in the fetal postmortem human brain, but not adult brain (Nikoshkov et al. 2008). In another study of AD in two European cohorts, four SNPs were analyzed and two SNPs (rs12545109 and rs2576581) showed association in the German cohort but not in the Swedish cohort (Racz et al. 2008).

Nociceptin/orphanin FQ (PNOC)

Nociceptin is a 17-amino acid neuropeptide that binds OPRL1. The PNOC gene is located on chromosome 8p21. Ten rare coding SNPs were described and one non-synonymous SNP (rs76786693, Ala118Gly) is common in Asian populations. Only three association studies have been reported with PNOC SNPs. Fifteen PNOC SNPs were analyzed in the COGA cohort. Two SNPs (rs17058952, in the 5′ flanking region, and rs351779, in the 3′ flanking region) were associated with AD, and one SNP (rs4732636, in the 5′ flanking region) was associated with illicit drug dependence (Xuei et al. 2008). No associations were reported for seven SNPs studied in EAs and AAs (Levran et al. 2008; Levran et al. 2009). The only SNP that was common to these studies is rs351784.

Addiction treatment

Methadone, the long-term major treatment of opioid addiction, is a mu opioid receptor agonist and a weak N-methyl-D-aspartic acid (NMDA) receptor antagonist. Part of the large inter-individual variability in drug response may be accounted for by genetic factors. Successful methadone treatment for opiate dependence relies in part on dosage optimization. Several pharmacogenetics studies of methadone have been performed to date, but only a few have studied variants in the opioid system genes that are relevant to this review. No association was found between OPRM1 118A>G or OPRD1 921T>C and response to MMT and methadone dose in a study of 238 European patients (Crettol et al. 2008).

Buprenorphine has been an alternative treatment approved relative recently for opioid addiction. No association studies of the opioid system genes have been reported on the response to buprenorphine treatment (see below a study of the HPA axis reactivity after buprenorphine-maintained patients (Kakko et al. 2008).

The hypothalamic-pituitary-adrenal (HPA) axis

The HPA axis is a neuroendocrine system involved in stress response by regulating ACTH and cortisol secretion (Hernandez-Avila et al. 2003). Disruption of stress pathways is considered to be one of the causes of addictive disorders (Koob and Kreek 2007; Kreek and Koob 1998). The mu opioid system plays an important role in stress response by regulating the HPA axis. Studies have shown that the mu opioid receptor is involved in modulation of the HPA axis by tonic inhibition of corticotropin-releasing hormone (CRH, also called CRF) in the hypothalamus and POMC in the anterior pituitary. The PDYN/OPRK1 system may also modulate the HPA axis by activation. Naloxone, nalmefene or naltrexone challenge in healthy subjects caused transient increase in plasma levels of ACTH and cortisol by disinhibition of the hypothalamic-pituitary part, indicating activation of the HPA axis (King et al. 2002; Schluger et al. 1998). HPA-axis stimulation plays an important role in the neurobiology of AD, CD and OD. Clinical studies and studies of animal models have shown that modest stimulation of the HPA axis is sought or desired by alcohol- and cocaine-dependent individuals. In contrast, opiate dependents find it aversive, since HPA axis activation is a characteristic of opioid withdrawal (e.g. Bond et al. 1998; Culpepper-Morgan and Kreek 1997; Koob and Kreek 2007; O'Malley et al. 2002). Long-term methadone maintenance treatment normalizes HPA axis activity in heroin addicts (Kreek 1973; Kreek et al. 1983).

Several studies report the effect of the OPRM1 118A>G polymorphism on the HPA axis. The 118G allele was associated with a robust cortisol response to naloxone blockade in subjects of predominantly European ancestry (Chong et al. 2006; Hernandez-Avila et al. 2003; Wand et al. 2002), but the effect was not found in Asians (Hernandez-Avila et al. 2007). It was also associated with blunted cortisol response to psychosocial stress (Chong et al. 2006) and with greater baseline concentrations of plasma cortisol in healthy subjects (Bart et al. 2006; Hernandez-Avila et al. 2007).

Metyrapone transiently blocks glucocorticoid production in the adrenal cortex. The 118G allele blunted the ACTH response to metyrapone in healthy subjects (Ducat et al. 2011). 118G carriers in buprenorphine maintenance treatment showed even greater attenuated response to metyrapone compared to 118A carriers, indicating more potent HPA axis suppression by buprenorphine in 118G carriers (Kakko et al. 2008).

Naltrexone and nalmefene

Two antagonists of opioid receptors are used for the treatment for AD: naltrexone (NTX) and nalmefene. Both antagonists reduce drinking and craving in alcoholic individuals in treatment and also in heavy drinkers. It is hypothesized that NTX works in part by occupying opioid receptors, preventing their binding by endogenous opioid peptides released upon alcohol intake, and in part by modest activation of the HPA axis (e.g. O'Brien et al. 2011; O'Malley et al. 2002). Activation of the HPA axis by the mechanism of disinhibition of the tonic mu opioid receptor inhibition is one of the pharmacological effects of NTX (Schluger et al. 1998). A substantial number of patients still do not respond to treatment, and family history of alcoholism was reported to be predictive of NTX response, suggesting the possibility that genetic factors may play a role in its effect (Rubio et al. 2005).

Several studies showed a positive effect of the OPRM1 118G allele on NTX treatment response (for review see Sturgess et al. 2011). Healthy subjects expressing the 118G-allele showed greater cortisol response to opioid receptor blockade by NTX and naloxone (Wand et al. 2002). Since alcoholics seek activation of the HPA axis, we predicted that alcoholics with the 118G allele would respond better to NTX treatment. This prediction has been supported by several studies showing that AD subjects with the 118G allele have better clinical response (Anton et al. 2008; Oroszi et al. 2009; Oslin et al. 2003; Ray and Hutchison 2007). However, these results were not supported by two other studies (Gelernter et al. 2007; Tidey et al. 2008). NTX was also shown to increase the urge for alcohol in 118G allele carriers (McGeary et al. 2006) and to decrease the positive subjective effect of alcohol (Setiawan et al. 2011).

Neurosteroids have been implicated as a factor that increases subjective effects of alcohol (Ray et al. 2010). Carriers of the 118G allele displayed increased neurosteroid levels after NTX treatment, suggesting that GABAergic neurosteroids may be a useful adjunctive therapy for alcoholism in non-carriers of the 118G allele associated with therapeutic efficacy to NTX (Ray et al. 2010).

Similar studies in rhesus macaques showed that carriers of the 77G variant that is functionally equivalent to the human 118G (see animal models of 118A>G, above) were selectively sensitive to suppression of alcohol preference by NTX treatment (Barr et al. 2010), were more sensitive to the effects of NTX, and showed greater reductions in alcohol consumption at lower NTX doses (Vallender et al. 2010).

NTX also modulates amphetamine-induced effects and has a potential in the treatment of amphetamine dependence (Jayaram-Lindstrom et al. 2008a;b). No pharmacogenetics studies were reported for the effects of specific variants on NTX response in amphetamine-dependent subjects. No association was found between the OPRM1 SNP 118A>G, the OPRD1 SNPs rs2234918 (921T>C) and the intronic rs678849, and the OPRK1 SNP rs963549 and outcome of nalmefene treatment for alcoholism (Arias et al. 2008).

Conclusions

In this article we reviewed the genetics of the endogenous opioid system genes, including association studies, which identified variants that may contribute to the vulnerability to develop specific drug addiction or common drug liability. We also discussed the genetics of the HPA axis and the pharmacogenetics of current treatment for addiction. Association studies in complex disorders such as addiction are challenging. Nevertheless, there is growing evidence for association and/or functionality of several SNPs in the opioid system genes that may have implications for understanding the addictive diseases and for improvements in treatment strategies and personalized medicine. More work is clearly needed to verify the role of variants suggested to be associated with addiction in small studies and for better understanding of their role in different populations.