Pharmacogenetics of alcohol, nicotine and drug addiction treatments
Daniel J. Müller, Centre for Addiction and Mental Health, Pharmacogenetics Research Clinic, 250 College Street, R-30, Toronto, ON M5T 1R8, Canada. E-mail: firstname.lastname@example.org
The numerous premature deaths, medical complications and socio-economic repercussions of drug and alcohol addiction suggest that improvements in treatment strategies for addictive disorders are warranted. The use of pharmacogenetics to predict response to medication, side effects and appropriate dosages is relatively new in the field of drug addiction. However, increasing our understanding of the genetic factors influencing these processes may improve the treatment of addiction in the future. We examined the available scientific literature on pharmacogenetic advancements in the field of drug addiction with a focus on alcohol and tobacco to provide a summary of genes implicated in the effectiveness of pharmacotherapy for addiction. In addition, we reviewed pharmacogenetic research on cocaine and heroin dependence. Thus far, the most promising results were obtained for polymorphisms in the OPRM1 and CYP2A6 genes, which have been effective in predicting clinical response to naltrexone in alcoholism and nicotine replacement therapy in smoking, respectively. Opinions differ as to whether pharmacogenetic testing should be implemented in the clinic at this time because clinical utility and cost-effectiveness require further investigation. However, the data summarized in this review demonstrate that pharmacogenetic factors play a role in response to addiction pharmacotherapy and have the potential to aid in the personalization of addiction treatments. Such data may lead to improved cessation rates by allowing physicians to select medications for individuals based, at least in part, on genetic factors that predispose to treatment success or failure rather than on a trial and error basis.
Alcohol and nicotine dependence are among the most prevalent addictive disorders worldwide (Kessler et al. 1997; Grant et al. 2004). Both disorders have numerous negative health impacts with smoking, causing 435 000 premature deaths in the United States and 5 million deaths worldwide per year (Thome et al. 2009). Alcohol use is linked to between 20 and 30% of oesophageal cancer, liver disease, epilepsy and motor vehicle accidents (World Health Organization 2002). Additionally, nicotine and alcohol dependence are associated with large socio-economic impacts; in wealthy nations, 6–15% of health-care costs are estimated to result from smoking alone (Prabat & Chaloupka 2000). Each year, tobacco smoking and excessive alcohol intake cause approximately 440 000 (Fellows et al. 2002) and 75 766 deaths (Mokdad et al. 2004), respectively, in the United States. Cocaine and heroin addiction, although less prevalent, have considerable negative impacts on individuals. These drugs are typically associated with substantial social decline, criminality, a large economic burden to the society and negative health consequences, including increased emergency room admissions (Substance Abuse and Mental Health Services Administration 2006), elevated rates of HIV (McCoy et al. 2004), hepatitis B and hepatitis C (Office of National Drug Control 2004).
Current pharmacotherapy for nicotine dependence, including nicotine replacement therapy (NRT), bupropion and varenicline, significantly increase cessation rates compared with placebo (Etter & Stapleton 2006; Wu et al. 2006; Hughes, Stead & Lancaster 2007; Le Foll & George 2007). However, only approximately one in four smokers is able to maintain abstinence using these treatments (Schnoll & Lerman 2006). Additionally, pharmacotherapy for alcohol dependence shows mixed outcomes; the commonly used medication naltrexone (NTX) benefits some patients but not others (Anton et al. 2008; Kim et al. 2009; Ray & Oslin 2009). Cessation is influenced by many variables; however, 40 to 60% of abstinence from smoking is thought to depend on genetic factors (Xian et al. 2003).
Pharmacogenetics, the study of genetic variation underlying individual differences in drug metabolism (pharmacokinetics) and drug response (pharmacodynamics), in conjunction with other biomarkers to identify disease states and therapeutic response, can aid in personalizing prescription (de Leon 2009). While the concept of using pharmacogenetic testing to tailor drug dependence treatments is relatively new, studies of several genes have yielded significant findings suggesting that pharmacogenetic studies may lead to the selection of individual-appropriate medication for the treatment of drug dependence in the future and to increased cessation rates. This review will focus on genetic factors that influence the effectiveness of therapeutic compounds used to treat addictions. Many additional genetic factors influence the onset or severity of addiction; however, a comprehensive discussion of these latter factors is beyond the scope of this review.
A literature search was conducted in the most common electronic databases (e.g. PubMed) considering papers published between the year 2000 and May 2010 using keywords such as ‘pharmacogenetic’ or ‘genetic’ in combination with ‘addiction’, ‘pharmacotherapy’, ‘nicotine replacement’, ‘bupropion’, ‘naltrexone’, ‘methadone’, etc. All papers evaluating pharmacogenetic effects on the outcome of treatments for addiction were included in the manuscript; papers describing purely genetic factors predisposing to addiction or to the likelihood of abstaining from addictive substances in the absence of treatment were excluded. Studies were included if it was unclear whether effects were of a genetic or pharmacogenetic nature, and limitations in the ability to interpret these findings are described in the review.
PERSONALIZING PHARMACOTHERAPY FOR ALCOHOL DEPENDENCE
Once in the brain, ethanol is thought to act on the dopaminergic system to modulate the brain's reward system, similar to the action of other drugs. At low doses, ethanol facilitates dopaminergic and GABA-A receptor signalling and this is thought to be responsible for its behavioural effects (Heinz et al. 2003). Additional neurotransmitters including opioid peptides and glutamate are also thought to be important for the acute reinforcing effects of ethanol (Koob, Sanna & Bloom 1998). Alcohol is metabolized in the liver by alcohol dehydrogenases, converting it to a toxic metabolite, acetaldehyde. Acetaldehyde is then converted to acetate by aldehyde dehydrogenases (Yin & Agarwal 2001; Higuchi et al. 2004). Pharmacogenetic studies in alcohol dependence are summarized in Table 1.
Table 1. Summary of pharmacogenetic studies in treatment for alcohol dependence.
|OPRM1||A118G (rs561720)||NTX||40||European||G allele was associated with greater feelings of intoxication (P = 0.04); naltrexone (NTX) blunted alcohol-induced high more strongly in these individuals (P = 0.007)||Ray & Hutchison (2007)|
|A118G||NTX||604||European||G carriers had a greater number of days abstinent (P = 0.07) and fewer heavy drinking days (P = 0.04) on NTX while individuals with an AA genotype showed no benefit; this was only true for groups not receiving behavioural treatment||Anton et al. (2008)|
|A118G||NTX||31||Korean||Subjects with at least one copy of the G allele took significantly longer to relapse (P = 0.014)||Kim et al. (2009)|
|A118G||NTX||141||European and African American||Subjects with the G allele receiving NTX had lower rates of relapse (P = 0.04) and took longer to return to heavy drinking than AA individuals (P = 0.044)||Oslin et al. (2003)|
|A118G||NTX, acamprosate||108||Dutch||No significant effect of genotype on NTX or acamprosate response was found (P > 0.05)||Ooteman et al. (2009)|
|A118G||NTX||215||European AND African American, male only||There was no effect of genotype on response to NTX (P > 0.05)||Gelernter et al. (2007)|
|A118G||NTX||93||European||The G allele was associated with increased urge to drink following presentation of a cue in patients taking NTX but not placebo (P < 0.01)||McGeary et al. (2006)|
|A118G||Nalmefene||272||Central European and Finnish||There was no effect of genotype on the outcome of nalmefene treatment (P > 0.05)||Arias et al. (2008)|
|rs6848893||Nalmefene||272||Central European and Finnish||There was no effect of OPRM1 genotype on outcome of nalmefene treatment (P > 0.05)||Arias et al. (2008)|
|rs6848893||NTX||215||European and African American||There was no effect of genotype on response to NTX (P > 0.05)||Gelernter et al. (2007)|
|DRD4||VNTR||NTX||93||European||Study participants with the long variant (seven or more repeats) showed a trend towards increased cue-elicited urge to drink than those homozygous for the short variant (P = 0.09)||McGeary et al. (2006)|
|DRD2||rs6276||Tiapride||110||German||Patients with an A/A genotype required significantly more tiapride than patients carrying a G allele (P = 0.037) and had higher scores on some measures of psychopathology (P < 0.05)||Lucht et al. (2001)|
|ANKK1||Taq1A (rs1800497)||Acamprosate, NTX||108||Dutch||Acamprosate had greater efficacy at reducing cue-induced craving in individuals with an A1A1 genotype while NTX had greater efficacy in A2A2 individuals (P = 0.086)||Ooteman et al. (2009)|
|Taq1A||Bromo-criptine|| || ||Individuals with at least one A1 allele treated with bromocriptine improved the most in terms of anxiety and craving and did worst on placebo||Lawford et al. 1995)|
|DRD1||rs686||Acamprosate, NTX||108||Dutch||No significant effect of genotype on NTX or acamprosate treatment was found (P > 0.05)||Ooteman et al. (2009)|
|OPRD1||rs2234918||Nalmefene||272||Central European and Finnish||There was no effect of genotype on the outcome of nalmefene treatment (P > 0.05)||Arias et al. (2008)|
|rs2234918||NTX||215||European and African American||There was no effect of genotype on response to NTX (P > 0.05)||Gelernter et al. (2007)|
|rs678849||Nalmefene||272||Central European and Finnish||There was no effect of genotype on outcome of nalmefene treatment (P > 0.05)||Arias et al. (2008)|
|rs678849||NTX||215||European and African American||There was no effect of genotype on response to NTX (P > 0.05)||Gelernter et al. (2007)|
|OPRK1||rs963549||Nalmefene||272||Central European and Finnish||There was no effect of genotype on the outcome of nalmefene treatment (P > 0.05)||Arias et al. (2008)|
|rs963549||NTX||215||European and African American||There was no effect of genotype on response to NTX (P > 0.05)||Gelernter et al. (2007)|
|GABRA6||T1519C||NTX, Acamprosate||108||Dutch||The C allele predicted increased ability of acamprosate to block cue-induced cravings while the T allele predicted greater NTX efficacy (P = 0.066)||Ooteman et al. (2009)|
|GABRB2||C1412T||NTX, Acamprosate||108||Dutch||Upon presentation of an alcohol cue, individuals with a TT or CC genotype showed reduced physiological responses when given acamprosate (P > 0.1)||Ooteman et al. (2009)|
|GABRG2||G + 3145A(rs211013)||NTX, Acamprosate||108||Dutch||No significant effect of genotype on NTX or acamprosate response was found (P > 0.05)||Ooteman et al. (2009)|
|GRIN2B||C2664T(rs1806201)||NTX, Acamprosate||108||Dutch||No significant effect of genotype on NTX or acamprosate response was found (P > 0.05)||Ooteman et al. (2009)|
One of the most common treatments for alcohol dependence is NTX. NTX is an antagonist of opioid receptors, with the highest affinity for the µ-opioid receptor (Verebey & Mule 1975). Several previous studies on NTX response have focused on the relationship between the A118G polymorphism (rs561720) in the µ-opioid receptor gene OPRM1, resulting in a change from asparagine to aspartic acid at amino acid position 40. Initially, it was thought that this substitution resulted in an increased affinity of the receptor for β-endorphin (Bond et al. 1998); however, subsequent studies have been unable to replicate this finding (Befort et al. 2001; Beyer et al. 2004). In humans, carriers of the G allele were found to have altered activity of the hypothalamic–pituitary axis under conditions of opioid receptor antagonism (Wand et al. 2002; Hernandez-Avila et al. 2003). This substitution was also associated with a decreased cortisol response following psychosocial stress (Chong et al. 2006) and an enhanced increase in cortisol following treatment with the opioid receptor antagonist naloxone (Wand et al. 2002; Chong et al. 2006). In vivo, alcohol-dependent patients with a G allele showed reduced µ-opioid receptors in the ventral striatum compared with patients with an A allele (Heinz et al. 2005).
Three placebo-controlled trials have found that alcohol-dependent individuals with a G allele have better clinical responses, including lower relapse rates, on NTX than those with the A allele (Oslin et al. 2003; Ray & Hutchison 2007; Anton et al. 2008):
In a study of 141 alcohol dependent patients, Oslin et al. (2003) found that subjects with a G allele showed lower rates of relapse with NTX, while those receiving placebo did not differ from other genotypes in terms of relapse. In agreement with this, a study of 604 alcohol-dependent individuals of European ancestry found that presence of a G allele was associated with reduced drinking and increased abstinence on NTX, while individuals with an AA genotype showed no difference in response when treated with NTX compared with placebo (Anton et al. 2008). Further support came from a study of 40 heavy drinkers showing that the ability of NTX to block alcohol-induced stimulation, positive mood, craving and enjoyment was higher in patients with the G allele (Ray & Hutchison 2007). Alcohol-dependent subjects with a G allele had the best treatment outcome, with a trend towards a higher number of days abstinent and fewer heavy drinking days, when treated with NTX compared with placebo. Also consistent were findings from a study of 31 Korean patients: Individuals with a G allele were shown to take significantly longer to relapse when taking NTX compared with individuals with the A allele. However, because there was no placebo group, interpretation is limited and findings may be due to an increased propensity for G carriers to abstain from drinking even in the absence of treatment (Kim et al. 2009). These four studies suggest that the G allele predicts greater effectiveness of NTX treatment. However, one placebo-controlled study of 215 male individuals failed to replicate these findings: Gelernter et al. (2007) found no effect of genotype on the outcome of treatment with NTX. However, the authors indicate that the larger sample from which the study sample was taken did not show a positive response to NTX treatment in general. With little variation in general treatment response, a pharmacogenetic effect may be more difficult to detect.
Previous work showed surprisingly that individuals with the G allele taking NTX reported an increased urge to drink alcohol following presentation of an alcohol-related cue relative to individuals on placebo (McGeary et al. 2006). However, because this study did not measure alcohol consumption and relapse, interpretation remains limited. Moreover, alcohol craving per se does not predict treatment outcome and can sometimes even act as a warning sign to help patients avoid relapse (Heinz et al. 2005). Two additional polymorphisms in OPRM1, rs6848893 and rs17180691, were not found to have any effect on the outcome of treatment with NTX (Gelernter et al. 2007).
ANKK1 Taq1A, GABAergic, glutamatergic and opiodergic polymorphisms
Additional investigations focused on the ANKK1 Taq1A (C32806T) (rs1800497) polymorphism, previously thought to be located in the dopamine D2 receptor (DRD2) gene, as well as on genes involved in GABAergic, glutamatergic and opioidergic neurotransmission. In a randomized, double-blind placebo-controlled study of 108 Dutch individuals, an A2A2 genotype at the Taq1A polymorphism of ANKK1 was associated with greater NTX efficacy, while an A1A1 genotype was associated with an increased ability of acamprosate, an NMDA receptor modulator, to blunt cue-induced cravings (Ooteman et al. 2009). Additionally, the T allele of the GABA receptor subunit alpha-6 (GABRA6) T1519C polymorphism was associated with greater NTX efficacy while the C allele predicted improvement with acamprosate. A TT genotype at the GABA receptor subunit beta-2 (GABRB2) C1412T polymorphism was associated with reduced physiological responses, measured by change in heart rate, to alcohol-associated cues after the administration of acamprosate. Additionally, this study found no interaction between physiological response and genotype at the DRD1 rs686 polymorphism, glutamate receptor subunit 2B (GRIN2B) C2664T (rs1806201) polymorphism or the GABA receptor subunit gamma-2 (GABRG2) G + 3145A (rs211013) polymorphism after NTX or acamprosate treatment and subsequent presentation of alcohol (Ooteman et al. 2009).
The role of the dopamine D4 receptor (DRD4) variable number of tandem repeats (VNTRs) polymorphism in response to NTX was also investigated (McGeary et al. 2006). Study participants on NTX who had seven or more repeats at the VNTR experienced a trend towards higher cue-elicited urges to drink alcohol than individuals homozygous for alleles with fewer than seven repeats; however, because drinking behaviour was not assessed, the effect of the increased desire for alcohol consumption is unclear. Additionally, the δ-opioid receptor gene (OPRD1) polymorphisms rs2234918 (T921C) and rs1042114 (F27C) and the κ-opioid receptor gene (OPRK1) polymorphism rs963549 were not found to have any effect on the outcome of treatment with NTX (Gelernter et al. 2007).
In summary, the most consistent results were obtained for the OPRM1 gene, where carriers of the G-allele of the A118G (rs561720) polymorphism showed better response patterns to NTX.
Nalmefene is an additional opioid receptor antagonist used to treat alcohol dependence. The effect of polymorphisms of several opioid receptor subtypes on the outcome of treatment with nalmefene was investigated by Arias et al. (2008). In this randomized, placebo-controlled trial of 272 heavy drinkers of Central European and Finnish background, no association was found between the A118G (rs561720) polymorphism of OPRM1, the rs2234918 (T921C) and rs678849 polymorphisms in OPRD1, and the rs963549 polymorphism of OPRK1 and outcome of treatment with nalmefene (Arias et al. 2008).
Additional studies have investigated the influence of genetic factors on treatment with tiapride, a dopamine D2 and D3 receptor antagonist and bromocriptine, a DRD2 agonist. One study assessed the role of the DRD2 A to G substitution downstream of rs6276 (exon 8 polymorphism) and a second investigated the Taq1A polymorphism of ANKK1, on tiapride and bromocriptine therapy, respectively.
Patients with an AA genotype at the DRD2 rs6276 polymorphism required significantly more tiapride than patients carrying a G allele (Lucht et al. 2001) and scored higher on measures of anxiety and depression following detoxification.
In a placebo-controlled trial, patients with an A1 allele at the Taq1A polymorphism of ANKK1 improved the most on measures of anxiety and craving following treatment with bromocriptine and were found to deteriorate on placebo (Lawford et al. 1995).
PERSONALIZING PHARMACOTHERAPY FOR NICOTINE DEPENDENCE
Nicotine acts as an agonist of nicotinic cholinergic receptors (nAChRs). However, after initial stimulation of nAChRs by nicotine, the receptors desensitize and subsequently up-regulate. As a consequence, nicotine behaves functionally as an antagonist at the nAChR (Picciotto et al. 1998). Nicotinic receptors are pentameric ligand-gated ion channels composed of α (2–10) and β (2–4) subunits (McGehee & Role 1995; Role & Berg 1996; Gotti, Fornasari & Clementi 1997; Jones, Sudweeks & Yakel 1999). α4β2-containing nAChRs are believed to be of primary importance for mediating the rewarding effects of nicotine; both α4 and β2 subunit genes have been shown to play a role in sensitivity and response to nicotine in mice (Picciotto et al. 1998; Tapper et al. 2004; Maskos et al. 2005). Nicotine is metabolized primarily by the cytochrome P450 2A6 (CYP2A6) enzyme (Nakajima et al. 1996; Messina, Tyndale & Sellers 1997); however, other enzymes, including CYP2B6, may also play a minor role (Johnstone et al. 2006). Pharmacogenetic studies in nicotine dependence are summarized in Table 2.
Table 2. Summary of pharmacogenetic studies in treatment for nicotine dependence.
|Genome-wide association study||Many||Bupropion and nicotine replacement therapy (NRT)||550||European||A total of 41 SNPs associated with abstinence specifically on NRT were identified and 26 SNPs associated with abstinence specifically on bupropion were identified||Uhl et al. (2008)|
|CYP2A6|| ||Bupropion||414||European||Fast metabolizers had significantly greater cessation on bupropion versus placebo (P > 0.01) while rates were high but did not differ between bupropion and placebo groups for slow metabolizers (P > 0.05)||Patterson et al. (2008)|
| ||NRT||480||European||3-HC/cotinine ratio predicted the success of transdermal NRT; higher plasma nicotine levels were associated with increased probability of quitting (P = 0.005)||Lerman et al. (2006b)|
| ||NRT||568||European||Faster metabolizers were less likely to be abstinent after eight weeks of transdermal NRT (30%) than slow metabolizers (47%) (P = 0.003)||Schnoll et al. (2009)|
|*2, *4, *9, *12||NRT||394||European||Slow metabolizers (one or two copies of *2 or *4; two copies of *9A and/or *12A) had higher plasma nicotine than normal metabolizers (P = 0.02) when using the same number of patches/week, slow metabolizers used fewer doses of nicotine spray/day (P < 0.02)||Malaiyandi et al. (2006)|
|CYP2B6||C1459T (rs3211371)||Bupropion||426||European||Individuals with one or two T alleles were less likely to remain abstinent during the treatment phase and bupropion counteracted this increased relapse in female smokers (P = 0.01)||Lerman et al. (2002)|
|C1459T||Bupropion||291||European||The T variant of CYP2B6 C1459T in combination with A2/A2 predicted the highest abstinence (P = 0.01)||David et al. (2007b)|
|*6||Bupropion||326||European||Cessation in CYP2B6*6 carriers was significantly higher on bupropion versus placebo at the end of treatment (P = 0.01) and after six months (P = 0.008); bupropion did not improve outcomes compared with placebo for wild type individuals (P = 0.93)||Lee et al. (2007b)|
|*4||Bupropion||121||Healthy Europeans||Individuals with the CYP2B6*4 allele had significantly greater clearance of bupropion than wild type (P = 0.001)||Kirchheiner et al. (2003)|
|*1, *4, *6||NRT||369||European||Cessation/abstinence was not influenced by CYP2B6 genotype (P > 0.05)||Lee et al. (2007a)|
|CHRNB2||rs2072661||Bupropion||412||European||The A allele of rs2072661 was associated with significantly reduced abstinence at the end of treatment (P = 0.0004) and at the six-month follow-up (P = 0.000 06), but this was not associated with treatment||Conti et al. (2008)|
|rs2072661||NRT||156||European||Smokers with the GG genotype had significantly greater abstinence on NRT compared with placebo (P < 0.01) and increased probability of quitting on the target quit date (P < 0.05)||Perkins et al. (2009)|
|OPRM1||A118G (rs561720)||NRT||320||European||Individuals with the G allele (Asp40) variant were more likely to be abstinent at the end of treatment (P = 0.03) than those homozygous for Asn40; this was significant for transdermal but not nasal nicotine||Lerman et al. (2004a)|
|A118G||NRT||374||European||The G allele was significantly associated with increased cessation at the end of treatment (P = 0.032)||Ray et al. (2007)|
|A118G||NRT||710||European||AA individuals had improved cessation with NRT, while G carriers did not benefit compared with placebo (P = 0.048)||Munafo et al. (2007)|
|CHRNA4||Rs2236196||NRT||316|| ||TC genotype at rs2236196 was significantly associated with abstinence (P = 0.01)||Hutchison et al. (2007)|
|SLC6A3 (DAT1)||3′ UTR VNTR||Bupropion||418||European||Smokers with DRD2 Taq1 A2 and SLC6A3-9 genotype had the highest rates of abstinence (P = 0.03) but genotype did not influence treatment response||Lerman et al. (2003)|
|3′ UTR VNTR||NRT, Bupropion||583||European||Smokers with the 9-repeat allele were more likely to quit than those with the 10/10 genotype (P = 0.012); smokers with the 2-repeat allele of the intron-8 polymorphism more likely to quit (P = 0.03); however, these findings were not related to whether NRT or bupropion was used||O'Gara et al. (2007)|
|3′ UTR VNTR||Bupropion||416||European||No main effect of SLC6A3 genotypes on smoking cessation; DRD2A1 and SLC6A3-9 showed poorer response to bupropion at the 12-month follow-up (P < 0.05)||Swan et al. (2007)|
|3′ UTR VNTR||Bupropion||291||European||There was no interaction of SLC6A3 genotype with DRD2 Taq1A genotype and the outcome of treatment (P = 0.93)||David et al. (2007b)|
|3′ UTR VNTR||Bupropion||553||European||Genotype was not associated with weight gain at the end of treatment with bupropion (P > 0.05)||Hu et al. (2006)|
|5-HTT||5-HTTLPR||NRT||397||European||There was no association between genotype and response to NRT (P > 0.05)||Munafo et al. (2006a)|
|5-HTTLPR||NRT||741||European||There was no interaction between genotype and response to NRT (P > 0.05)||David et al. (2007a)|
|DRD2||-141C Ins/Del (rs1799732)||NRT||368||European||Smokers with at least one copy of the Del C allele had higher quit rates on NRT compared to individuals who were homozygous for the Ins C allele (P = 0.006)||Lerman et al. (2006a)|
|-141C Ins/Del||NRT||363||European||Smokers with at least one copy of -141 Del allele and two copies of FREQ rs1054879 A were more likely to be abstinent at the end of the NRT trial than smokers with other alleles (P = 0.022)||Dahl et al. (2006)|
|-141C Ins/Del||Bupropion||414||European||Smokers homozygous for the Ins C allele had higher abstinence at the end of treatment with bupropion compared with smokers with the Del C allele (P = 0.01)||Lerman et al. (2006a)|
|-141C Ins/Del||Bupropion||553||European||Genotype was not found to be associated with bupropion-induced weight gain (P > 0.05)||Hu et al. (2006)|
|C957T||NRT||368||European||The C allele was associated with increased abstinence on NRT compared with the TT genotype (P = 0.03)||Lerman et al. (2006a)|
|C957T||Bupropion||414||European||Genotype was not associated with quit rates at the end of bupropion treatment (P = 0.85)||Lerman et al. (2006a)|
|C957T||Bupropion||553||European||TC genotype was associated with increased weight gain at six months follow-up (P = 0.03)||Hu et al. (2006)|
|Exon 8 A/G (rs6276)||NRT||755|| ||There was no association between genotype and the effectiveness of NRT after 12 weeks of treatment (P > 0.05)||Johnstone et al. (2004)|
|ANKK1||Taq1A (rs1800497)||NRT||755||European||Smokers with both an A1 allele and a DBH 1368A allele showed the greatest benefit from NRT after 12 weeks of treatment (P = 0.04)||Johnstone et al. (2004)|
|Taq1A||NRT||752||European||Quit rates on NRT were higher for women with the Taq1 A1 allele at the end of treatment (P = 0.03). There was no effect of genotype on quit rates at the end of treatment with NRT in men (P = 0.46)||Yudkin et al. (2004)|
|Taq1A||Bupropion||32|| ||Craving, anxiety and irritability were reduced by bupropion in Taq1 A2/A2 but not A1/A1 or A1/A2||David et al. (2003)|
|Taq1A||Bupropion||291||European||A2/A2 had greatest reduction in craving and greatest abstinence on bupropion (P = 0.038); the T variant of C1459T was found to increase the effect of the Taq A1 variant on abstinence (P = 0.01)||David et al. (2007b)|
|Taq1A||Bupropion||722||European||There was a significant interaction between genotype and the effectiveness of bupropion (P = 0.009) with A2/A2 carriers more likely to be abstinent compared with placebo at the end of treatment||David et al. (2007c)|
|Taq1A||Bupropion||416||European||A1 carriers were more likely to report discontinuing bupropion because of side effects (P < 0.05)||Swan et al. (2005)|
|Taq1A||Bupropion||71||European||A1 carriers reported significant increases in the rewarding value of food following smoking cessation and this increase was blocked in individuals taking bupropion (P = 0.03).||Lerman et al. (2004b)|
|Taq1A||Venla-faxine||134||European||Smokers with an A2/A2 genotype responded to venlafaxine with a significant reduction in negative affect but A1 carriers did not (P = 0.01)||Cinciripini et al. (2004)|
|DRD4||40 bp VNTR||NRT||720||European||Presence of at least one copy of the long allele (at least seven repeats) was associated with reduced cessation at 12-week follow-up (P = 0.034); however, this effect was independent of treatment||David et al. (2008)|
|C-521T||NRT||720||European||There was no interaction between genotype and the effectiveness of NRT (P > 0.05)||David et al. (2008)|
|C-521T||NRT||706||European||DRD4 status was not associated with smoking status but CC genotype was associated with increased body mass index (BMI) following cessation (P = 0.046)||Munafo et al. (2006b)|
|HINT1||rs3852209||NRT||374|| ||TT genotype at the HINT1 rs3852209 polymorphism was associated with abstinence six months after treatment with NRT (P = 0.013); however, it may not be a pharmacogenetic effect since treatment was eight weeks only||Ray et al (2007)|
|COMT||Val108/158Met||NRT||290||European and African American, women only||Women with the Met/Met genotype showed higher rates of prolonged abstinence with NRT than women with Val/Val (P = 0.03)||Colilla et al. (2005)|
|Val108/158Met||NRT||749||European||There was a significant interaction between genotype and treatment at 12 weeks with a higher percentage of Met/Met individuals abstinent on NRT versus placebo than individuals with a Val allele on NRT versus placebo (P = 0.05)||Johnstone et al. (2007)|
|Val108/158Met|| ||785||European and African American, women only||European women with a Met/Met genotype were more likely to be former smokers (P = 0.03)||Colilla et al. (2005)|
|rs737865||Bupropion||511||European and African American||Individuals with a GG genotype at both rs165599 did not benefit from bupropion (P = 0.05) while individuals with an A allele showed increased abstinence after bupropion treatment relative to placebo (P = 0.04). No significant effects were observed for the rs737865 polymorphism||Berrettini et al. (2007)|
|rs4818||Bupropion||553||European||Val/Met genotype was associated with less weight gain than the Met/Met genotype at the end of treatment (P = 0.0232)||Hu et al. (2006)|
Nicotine Replacement Therapies (NRT)
NRT are currently the most commonly used treatment forms for nicotine dependence. Delivery of nicotine can be achieved many ways, for example, through transdermal patches, nasal spray, gum, lozenges or an inhaler. The rate of nicotine metabolism has been found to influence the effectiveness of NRT. Two open label NRT trials have shown an association between CYP2A6 activity and response to NRT (Malaiyandi et al. 2006; Lerman et al. 2006b; Schnoll et al. 2009). The rate of nicotine metabolism was determined by measuring a nicotine metabolite phenotype measure, the 3′-hydroxycotinine/cotinine metabolite ratio, which is highly correlated with nicotine metabolism and can be used as a marker of CYP2A6 activity (Dempsey et al. 2004). Faster nicotine metabolism when using transdermal NRT was linked to a decreased probability of smoking cessation (Lerman et al. 2006b; Schnoll et al. 2009). Malaiyandi et al. (2006) found an association between slow nicotine metabolism and both higher plasma nicotine levels on transdermal nicotine and the use of fewer doses of nicotine spray per day when using nasal nicotine. An investigation of the effect of CYP2B6 activity on response to NRT found no association between CYP2B6 genotype and smoking cessation (Lee et al. 2007b).
CHRNB2 and CHRNA4 polymorphisms
Two studies have shown that pharmacodynamic factors, in this case, genotype at nAChR subunit polymorphisms, are associated with smoking cessation. The 3' untranslated region polymorphism rs2072661 of the β2 nAChR subunit(CHRNB2) was studied in a sample of 156 Europeans. In this study, a within-subjects cross-over design was chosen where smokers received 21 mg nicotine and placebo patch in counter-balanced order, during two separate 5-day simulated quit attempts. Subjects with the GG genotype had higher abstinence rates during the nicotine versus placebo patch week compared with carriers of the AG or AA genotypes (Perkins et al. 2009). However, this effect was observed regardless of whether a placebo patch or nicotine patch was being used, suggesting individuals with this genotype may be able to quit more readily even in the absence of pharmacotherapy. Additionally, the rs2236196 polymorphism of the α4 nAChR subunit (CHRNA4) was associated with cessation, specifically with nicotine nasal spray without the nicotine patch. The authors acknowledge, though, that further studies including a control group receiving no treatment are necessary to rule out the possibility that this polymorphism is associated with relapse, in general, rather than response to pharmacotherapy (Hutchison et al. 2007).
Studies on genes related to dopaminergic neurotransmission have yielded several positive findings. An open label trial of transdermal nicotine versus nasal spray in 368 European smokers linked the DelC allele of the DRD2 -141C Ins/Del (rs1799732) polymorphism to higher quit rates on NRT (Lerman et al. 2006a). Consistent with this study, a second open label trial also implicated the DelC allele with increased abstinence on NRT; however, it was only when the DelC allele was present in combination with two copies of the neuronal calcium sensor-1 protein (FREQ) rs1054879A allele (Dahl et al. 2006). Genotype at a second DRD2 polymorphism, C957T, was also found to be related to the success of NRT. In an open label study of 368 European smokers, individuals with a C allele were less than two thirds as likely to be abstinent than individuals homozygous for the T allele following eight weeks of NRT (Lerman et al. 2006a). However, the authors indicate this may have been primarily due to an effect at the DRD2 -141C Ins/Del (rs1799732) polymorphism because of the linkage disequilibrium between these two polymorphisms. A randomized, placebo-controlled trial also failed to find an association between the DRD2 rs6276 polymorphism and the effectiveness of NRT (Johnstone et al. 2004).
Previous work has also investigated the effect of two DRD4 polymorphisms on NRT. In a randomized, placebo-controlled trail, presence of at least one copy of the long allele (at least seven repeats) of the DRD4 VNTR was associated with reduced cessation at 12 week follow-up; however, this finding was independent of treatment (David et al. 2008). In this study and in an additional randomized, placebo-controlled trial, no association was found between genotype at the C521T polymorphism of DRD4 and success of NRT (Munafo, Murphy & Johnstone 2006b; David et al. 2008).
Catetchol-O-methyltransferase (COMT) Val108/158Met
Two studies have linked the rs4680 polymorphism (Val108/158Met) of COMT to response to NRT. A randomized, open label study in 290 women found that individuals with the Met/Met genotype (AA) had higher rates of prolonged abstinence with NRT than women with a Val (G) allele (Colilla et al. 2005). This study also showed that European women with a Met/Met genotype were more likely to be former smokers, suggesting that the effect of genotype on abstinence could be independent of treatment. Consistent with this study, a randomized, placebo-controlled trial in 749 European individuals found a significant interaction between genotype and treatment, with a higher number of Met/Met individuals abstinent on NRT versus placebo than carriers of the Val allele (Johnstone et al. 2007). Both independent studies suggest that COMT gene variants influence response to NRT.
Two studies have linked genotype at the ANKK1 Taq1A locus with the efficacy of NRT. In a randomized, placebo-controlled study by Johnstone et al. (2004), individuals with at least one Taq1A A1 allele and one dopamine beta-hydroxylase (DBH) 1368A allele had significantly higher cessation rates after 12 weeks of NRT than other genotypes and the lowest quit rates on placebo. In agreement, a second study found that the A1 allele was associated with higher quit rates on the nicotine patch, while A2A2 (CC genotype) individuals did not benefit; however, this effect was only observed in women (Yudkin et al. 2004).
Interestingly, a randomized, placebo-controlled NRT trial indicated that individuals with an AA genotype at the OPRM1 A118G (rs561720) polymorphism, the same variant associated with greater abstinence from alcohol during NTX treatment (Oslin et al. 2003; Ray & Hutchison 2007; Anton et al. 2008; Kim et al. 2009), benefited from NRT relative to placebo while G allele carriers did not (Munafo et al. 2007). Opposite results were obtained from a clinical trial of transdermal versus nasal spray NRT. In this study, the OPRM1,118G allele was associated with increased smoking cessation at the end of treatment (Lerman et al. 2004b). An extended study of Lerman et al. (2004b), found again an association between the G allele of OPRM1 and abstinence at the end of treatment (Ray et al. 2007). This study also found an interaction that approached significance between the beta arrestin-2 (ARRB2) polymorphisms and OPRM1 A118G (rs561720) genotype after six months of abstinence (Ray et al. 2007). However, because the two latter studies did not include placebo groups, it is difficult to compare the findings to those of Munafo et al. (2007) and to determine whether these are pharmacogenetic effects or independent of treatment.
Two studies on the influence of the serotonin transporter gene-linked polymorphic region (5-HTTLPR) on the outcome of NRT found no association between genotype and abstinence at the end of NRT treatment (Munafo et al. 2006a; David et al. 2007b). A study of histidine triad nucleotide binding protein 1 (HINT-1) found an association between the T allele of rs3852209 and abstinence six months after treatment (Ray et al. 2007). Further studies will be necessary to determine if this is a pharmacogenetic effect or an effect of genetic factors on cessation in general.
At present, CYP2A6 polymorphisms have had the most consistent association with the outcome of treatment with NRT. Polymorphisms leading to decreased activity of CYP2A6, and hence, increased levels of nicotine when using NRT, are associated with a better outcome on NRT.
Bupropion is an atypical antidepressant and in its sustained-release formulation (Zyban), is an FDA-approved treatment for nicotine dependence. It acts as a norepinephrine and dopamine reuptake inhibitor and also as a non-competitive antagonist at high-affinity nAChRs (primarily α3β4) (Fryer & Lukas 1999; Slemmer, Martin & Damaj 2000).
The rate of nicotine metabolism influences the effectiveness of NRT, and this also seems to be true for bupropion; however, findings are in the opposite direction. In a randomized, placebo-controlled trial, CYP2A6 rapid metabolizers, as determined by 3-HC/COT ratio, had higher quit rates on burpropion compared with placebo (34% versus 10%). Slower metabolizers had higher quit rates than rapid metabolizers on placebo (32%); however, this rate was not significantly improved in slower metabolizers receiving bupropion (32%) (Patterson et al. 2008).
Polymorphisms of the CYP2B6 gene also seem to influence bupropion efficacy. In a trial of 121 healthy individuals, Kirchheiner et al. (2003) showed that individuals with the *4 allele (A785G) had increased clearance of bupropion than individuals with the wild type allele. This could potentially lead to an altered therapeutic efficacy in *4 individuals; however, future studies will be necessary to determine the effect of this allele on the outcome of bupropion treatment. In a randomized, placebo-controlled bupropion trial, a link was found between the C1459T (rs3211371) polymorphism of CYP2B6, where the 1459T allele confers reduced enzymatic activity with poorer outcome of bupropion treatment. On placebo, males had lower rates of relapse than females. Within each gender group, smokers with a CC genotype, resulting in normal CYP2B6 activity, reported less abstinence-induced cravings and were more likely to be abstinent at the end of treatment. However, in the bupropion group, female T allele carriers had cravings and abstinence rates that were almost equivalent to those of smokers with a CC genotype (Lerman et al. 2002). In contrast, a second randomized, placebo-controlled trial found that the 1459T variant, in combination with the ANKK1 Taq1A A2A2 genotype, was associated with the highest abstinence rates on bupropion (David et al. 2007a).
The *6 allele has also been found to influence the outcome of buproprion treatment. A placebo-controlled buproprion trial showed that smokers carrying at least one *6 allele (i.e. C516T and A785G polymorphisms) of the CYP2B6 gene had significantly lower abstinence rates on placebo than with buproprion after 10 weeks of treatment an at 6-months follow-up (∼14% vs ∼32% and ∼13% vs. ∼31%, respectively). However, smokers with the *1 alleles had similar rates of abstinence with placebo and with buproprion (∼31% after 10 weeks and ∼22% at 6-months follow-up (Lee et al., 2007a). Based on these findings, it was concluded that smokers carrying the *6 allele (*1/*6 or *6/*6) are more likely to relapse on placebo, however these individuals are also more likely to benefit from smoking cessation treatment with buproprion.
Genes involved in the pharmacodynamic effects of nicotine have also been implicated in bupropion response. A randomized, placebo-controlled trial of 412 European smokers found that individuals with a GG genotype at the rs2072661 polymorphism of CHRNB2 who quit smoking stayed abstinent longer than individuals with other genotypes (Conti et al. 2008). This polymorphism was found to be associated with withdrawal symptoms; individuals with at least one copy of the A allele experienced more severe withdrawal (Conti et al. 2008).
SLC6A3/DAT1 3′ UTR VNTR
Three studies have assessed the effect of polymorphisms of the dopamine transporter gene (SLC6A3, DAT1) on treatment with bupropion. In a placebo-controlled study of 418 smokers by Lerman et al. (2003), an association between the 9-repeat allele and abstinence was shown. Individuals with the 9-repeat allele and the A2 variant at the ANKK1 Taq1A locus had the highest quit rates and showed the longest time intervals to relapse. In agreement with this finding, an open label bupropion and NRT study by O'Gara et al. (2007) found that individuals with the 9-repeat allele were more likely to be abstinent after bupropion or NRT. A study assessing the effectiveness of different combinations of counselling and bupropion was somewhat consistent, showing that individuals with the 9-repeat allele of the SLC6A3 3′ UTR VNTR were more likely to be abstinent 12 months after treatment than other genotypes. However, this was in the absence of the ANKK1 Taq1A A1 allele (rs1800497), opposite to the findings of Lerman et al. (2003) (Swan et al. 2007). Additionally, a randomized, placebo-controlled study of 291 smokers by David et al. (2007a) did not find an interaction between the number of repeats at the SLC6A3 3′UTR (rs27072) polymorphism and ANKK1 Taq1A (rs1800497) genotype or a main effect of SLC6A3 genotype on treatment outcome.
Several studies have investigated the Taq1A polymorphism of ANKK1. In a randomized, placebo-controlled bupropion trial, bupropion was found to reduce craving, anxiety and irritability in individuals with an A2A2 genotype (David et al. 2003). In agreement with this finding, an open label effectiveness trial showed a trend towards increased abstinence in women with an A2A2 genotype, while women with the A1 allele were more likely to discontinue bupropion treatment due to side effects (Swan et al. 2005). Also, in support of this finding, data from two pooled, randomized, placebo-controlled trials showed that individuals with the A2A2 genotype were three times as likely to be abstinent at the end of treatment as individuals of the same genotype taking placebo. This was not true for individuals with the A1 allele where bupropion was not found to confer a benefit relative to placebo (David et al. 2007c). Consistent with these findings, a randomized, placebo-controlled trial found that individuals with an A2A2 genotype reported the greatest reduction in cravings and highest abstinence while on bupropion. Presence of the T variant at the CYP2B6 C1459T polymorphism was found to modify the effects of Taq1A A1 genotype on abstinence; individuals with the A2A2 genotype were more likely to remain abstinent if they also had a T allele at the C1459T polymorphism of CYP2B6 (David et al. 2007a).
The role of genotype at three polymorphisms of DRD2, -141 Ins/Del, the intron 8 VNTR and C957T, in bupropion response has also been investigated. In a randomized, placebo-controlled trial by Lerman et al. (2006a), smokers with the Ins C allele were found to have higher quit rates than smokers with the Del C allele, while smokers with the Del C allele quit more frequently on placebo. O'Gara et al. (2007) found that individuals with the 2-repeat allele of the intron 8 VNTR were 8.5% more likely to quit within the first week than individuals with other genotypes; however, additional studies will be necessary to determine whether this effect is due to pharmacogenetic factors or independent of treatment. In a randomized, placebo controlled trial, the C957T polymorphism of DRD2 was not found to be associated with quit rates at the end of bupropion treatment (Lerman et al. 2006a).
Polymorphisms of COMT have also been associated with the effectiveness of bupropion treatment. A double-blinded, placebo-controlled trial found that the individuals with a GG genotype at rs165599 had higher quit rates on placebo while individuals with an A allele did significantly better on bupropion than placebo (Berrettini et al. 2007).
Polymorphisms associated with weight gain
The genetic basis of weight gain during and after treatment with bupropion has also been investigated. In a randomized, placebo-controlled trial, carriers of the ANKK1 Taq1 A1 allele reported increases in the rewarding value of food following smoking cessation and this was blocked by bupropion treatment (Lerman et al. 2004a). This increase in food reward was associated with increased weight gain in the placebo but not bupropion group at the six-month follow up (Lerman et al. 2004a). Genotype at the rs4818 polymorphism of COMT was also linked to weight gain following smoking cessation. Individuals with the Val/Met genotype were found to have gained less weight at the end of treatment with bupropion, though the interaction with treatment was not significant (Hu et al. 2006). The DRD2 C957T polymorphism has been linked to weight gain; individuals with a CC or TC genotype had gained more weight six months after the end of treatment with bupropion than other genotypes (Hu et al. 2006).
The DRD2 -141 Ins/Del polymorphism was not found to be associated with weight gain at the end of bupropion treatment (Hu et al. 2006).
In summary, several trials have shown an association between the Taq1A A2A2 genotype and effective reduction of the negative effects of withdrawal, in addition to increased abstinence, on bupropion.
The antidepressant venlafaxine, a serotonin and norepinephrine reuptake inhibitor, has also been investigated for use in the treatment of nicotine dependence (Yardley et al. 1990). A double-blinded, placebo-controlled trial found that smokers with a Taq1A A1 allele quit less often than individuals with an A2A2 genotype; however, this effect was independent of treatment (Cinciripini et al. 2004). A2A2 individuals also reported a significant reduction in negative affect on venlafaxine while A1 carriers did not (Cinciripini et al. 2004).
Additionally a genome-wide association study found associations between several clusters of polymorphisms and the ability of individuals to quit smoking using either bupropion or NRT (Uhl et al. 2008). The genes that were identified and coded for several different classes of molecules included molecules involved in cell adhesion, signal transduction, receptor function and enzymatic activity. In addition, several regions of the genome that were identified by the analysis are currently of unknown function. (Uhl et al. 2008). Interestingly, there was only modest overlap between genes that were associated with susceptibility to addiction to nicotine and genes associated with the success of bupropion or NRT treatment. This pattern is observed for other pathologies. For example, in the case of schizophrenia the DRD2 gene is highly implicated in the effectiveness of pharmacotherapies (Malhotra, Murphy & Kennedy 2004) but has not been strongly linked to susceptibility to the illness itself (MacDonald & Schulz 2009).
PERSONALIZING PHARMACOTHERAPY FOR COCAINE AND HEROIN DEPENDENCE
Few pharmacogenetic studies have been conducted on treatments for cocaine and heroin dependence. This discussion will focus on pharmacogenetic factors influencing disulfiram treatment for cocaine addiction and methadone treatment for heroin addiction.
Disulfiram was originally developed for the treatment of alcohol dependence. It acts to block the conversion of acetaldehyde to acetate, leading to an accumulation of the toxin acetaldehyde in the body. Disulfiram also blocks the DBH enzyme, which is involved in the breakdown of dopamine (Musacchio et al. 1966). This leads to increased dopamine, similar to when an individual takes cocaine. Disulfiram may act to increase the dysphoric effects of cocaine, making drug use unpleasant, or to blunt drug reward, euphoria and cravings (Gaval-Cruz & Weinshenker 2009). Previous work has indicated that the DBH C-1021T (rs1611115) polymorphism influences plasma levels of DBH; the T allele results in the lowest circulating levels of DBH and the lowest enzymatic activity (Zabetian et al. 2001; Kohnke et al. 2002; Deinum et al. 2004; Bhaduri & Mukhopadhyay 2008). It is predicted that the TT genotype would be associated with increased disulfiram efficacy because the lower levels of DBH would be inhibited by a smaller quantity of drug. In rats and mice, a genetic deficiency in DBH or treatment with disulfiram, both similar to having a TT genotype, were associated with enhanced anxiety and behavioural effects in response to cocaine (Schank et al. 2006; Schank, Liles & Weinshenker 2008). This work suggests that the DBH C-1021T (rs1611115) polymorphism may influence the effectiveness of disulfiram treatment; further studies in humans will be necessary to clarify the role of this polymorphism in treatment outcome.
Pharmacogenetic studies in levomethadone and methadone are summarized in Table 3.
Table 3. Summary of pharmacogenetic studies in levomethadone/methadone for treatment of heroine dependence.
|OPRM1||A118G (rs561720)||Levo-methadone||51||Healthy volunteers||Carriers of the G allele showed less pupil dilation in response to levomethadone than non-carriers (P < 0.01)||Lotsch et al. (2006)|
|CYP3A4||*1B||Methadone||245||Swiss sample||Patients with low CYP3A activity had higher trough methadone levels (P = 0.0002), but there was no influence on the effectiveness of treatment||Crettol et al. (2006)|
|CYP2B6||*6||Methadone||179||European||Patients with the *6/*6 genotype, slow metabolizers, had an increased risk of prolonged QTc when receiving methadone treatment (P = 0.017)||Eap et al. (2007)|
|*4, *5, *6, *9||Methadone||245||Swiss sample||*6 individuals had higher trough (s)-methadone levels (P = 0.0001), but there was no influence on the effectiveness of treatment||Crettol et al. (2006)|
|CYP2D6||*1, *2, *3, *4, *5, *6, *7, *8, *9, *10, *12, *14, *15, *21, and duplications of alleles *1, *2, *4||Methadone||205||Spanish||Ultrarapid metabolizers had lower satisfaction scores when on methadone maintenance treatment than poor or extensive metabolizers (P < 0.01)||Perez de los Cobos et al. (2007)|
|*1, *3, *4, *5, *6, *xN||Methadone||245||Swiss sample||Ultrarapid metabolizers had lower trough methadone levels (P = 0.04), but in poor metabolizers, there was no difference||Crettol et al. (2006)|
|*1, *2, *4||Methadone||59|| ||CYP2D6 status did not affect clearance (P > 0.39)||Coller et al. (2007)|
|ABCB1||C3435T||Methadone||245||Swiss sample||Lower trough levels were observed in individuals with a 3435 TT genotype (P = 0.01)||Crettol et al. (2006)|
|CYP1A2||*1F||Methadone||245||Swiss sample||Genotype did not influence plasma methadone levels (P > 0.4)||Crettol et al. (2006)|
|CYP2C9||*2, *3||Methadone||245||Swiss sample||Genotype did not influence plasma methadone levels (P > 0.3)||Crettol et al. (2006)|
|CYP2C19||*2, *3||Methadone||245||Swiss sample||Genotype did not influence plasma methadone levels (P > 0.1)||Crettol et al. (2006)|
|CYP3A5||*3||Methadone||245||Swiss sample||Genotype did not influence plasma methadone levels (P > 0.1)||Crettol et al. (2006)|
|UGT2B7||*2a||Methadone||245||Swiss sample||Genotype did not influence plasma methadone levels (P > 0.09)||Crettol et al. (2006)|
|DRD2||Taq1 A1/A2 (rs1800497)||Methadone||95||European||Individuals with the A1 allele tended to drop out of methadone treatment more frequently (P = 0.000 02) but had also used more heroin than other genotypes (P = 0.003)||Lawford et al. (2000)|
Similar to heroin, methadone is an agonist of µ-opioid receptors. Methadone, however, has a longer half-life. A study of 179 European individuals showed that individuals with a CYP2B6 *6 allele (rs3745274) had an increased risk of prolonged QTc interval when administered methadone (Eap et al. 2007). A second study of 245 individuals being treated with methadone found that the CYP2B6 *6 allele (rs3745274) was associated with higher trough levels of the drug; however, there was no effect on the success of treatment (Crettol et al. 2006). Though future studies will need to replicate these results, at present, CYP2B6 activity seems to influence the frequency of methadone-induced side effects but does not appear to alter treatment outcome.
CYP2D6 activity has been linked to the effectiveness of methadone treatment (Perez de los Cobos et al. 2007). In a study of 205 Spanish patients receiving methadone maintenance treatment, CYP2D6 ultrarapid metabolizers reported lower satisfaction with methadone than poor or extensive metabolizers. Additionally, male ultrarapid metabolizers rated significantly lower satisfaction than female ultrarapid metabolizers (Perez de los Cobos et al. 2007). Polymorphisms of the CYP2D6 gene were found to influence blood levels of methadone. (Crettol et al. 2006). Ultrarapid metabolizers had lower trough levels of methadone; however, methadone levels did not differ between poor metabolizers and individuals with normal CYP2D6 activity. Additionally, genotype was not found to influence the effectiveness of methadone treatment. A retrospective study of 59 patients receiving methadone maintenance treatment did not find an association between CYP2D6 genotype and clearance of methadone (Coller et al. 2007).
Other genes involved in drug metabolism
The role of CYP3A4 activity on the effectiveness of methadone treatment has also been studied (Crettol et al. 2006). Low CYP3A4 activity was associated with higher trough methadone levels; however, the serum levels were not associated with response to treatment. The same study assessed the influence of ABCB1, CYP1A2, CYP2C9, CYP2C19, CYP3A5 and UGT2B7 on methadone pharmacokinetics. Genotypes were not found to influence methadone levels or response to treatment.
An observational study of 95 patients receiving methadone maintenance treatment found that individuals with an A1 allele at the Taq1A locus of ANKK1 discontinued methadone treatment significantly more often than carries of other genotypes (Lawford et al. 2000). However, these individuals also reported taking more heroin in the year prior to treatment than other genotypes. This suggests that these individuals may have been more addicted to heroin, potentially leading to reduced treatment efficacy and a poor outcome on methadone.
A methadone trial in 51 healthy volunteers indicated that central nervous effects by means of pupil dilation to levomethadone, one of the two enantiomers that make up the more common racemic mixture of methadone, was lower in carriers of the OPRM1,118G allele than in non-carriers (Lotsch et al. 2006). However, because this study was conducted among healthy volunteers, additional work will be needed to determine if physiological responses to methadone are also reduced in heroin-addicted 118G allele carriers.
While additional research is necessary, the association of the CYP2B6 *6 allele (rs3745274) with methadone-induced prolonged QTc interval has potential clinical relevance because of the serious nature of this side effect.
SUMMARY AND FUTURE DIRECTIONS
The pharmacogenetic studies summarized in this review provide a solid indication that response to pharmacotherapy for drug addiction is influenced by genetic factors. However, at present, many of the positive findings require replication and further evaluation to determine clinical utility. A simulation by Gartner, Barendregt & Hall (2009) indicated that the clinical utility of a predictive genetic test is based on the magnitude of the effect at a given locus. This suggests that for modest effect sizes, a test will have limited predictive ability. Munafo (2009) argued that because most of the alleles discovered to date that affect the outcome of pharmacotherapy have relatively small effects or are relatively rare, clinical applications are currently limited. However, Lerman, Schnoll & Munafo (2007) indicated that some variants associated with poor response to treatments, including Taq1A A1 and variants associated with increased CYP2A6 activity, are quite common (43% and 77% of the population, respectively). Additionally, some alleles, including the InsC allele and the Taq1A A1 allele, respectively, have substantial effects on the outcome of smoking cessation treatments and predict cessation rates that are greater by 15 and 10% (Yudkin et al. 2004; Lerman et al. 2006a). This indicates there is potential use for pharmacogenetic approaches in the development of personalized medicine strategies for the treatment of tobacco and other addictions.
Opinions differ with respect to whether evidence is adequate to warrant implementation of pharmacogenetic testing in the clinic. Some polymorphisms have demonstrated clinical relevance consistently; however, clear quantitative evidence is still necessary to demonstrate clinical utility. The most promising data at present for the treatment of alcohol dependence with NTX are for the OPRM1 A118G (rs561720) polymorphism. Three randomized, placebo-controlled studies (Oslin et al. 2003; Ray & Hutchison 2007; Anton et al. 2008; Kim et al. 2009) have shown that individuals with a G allele display longer abstinence and a greater reduction in the positive effects of alcohol when taking NTX. Further support for the importance of this allele in alcohol consumption and treatment with NTX comes from work social drinkers and mice showing that striatal dopamine responses to alcohol were only observed in carriers of the 118G allele (Ramchandani et al. 2010). Additionally, a study in rhesus monkeys showed that animals with the 118G allele equivalent had stronger alcohol preferences and were more likely to respond to NTX than animals with the allele equivalent to the A variant (Barr et al. 2010). In a previous review of the A118G (rs561720) polymorphism and treatment with NTX, Oslin, Berrettini & O'Brien (2006) indicated that future efforts should be made to test and develop adaptive approaches to treatment based on genotype information.
The most promising pharmacogenetic data for smoking cessation treatment relate to CYP2A6 polymorphisms. Two open label studies of approximately 500 European patients each found a link between nicotine metabolism and the effectiveness of NRT. The ratio of metabolites was found to predict the outcome of treatment with NRT (Schnoll et al. 2009); CYP2A6 rapid metabolizers were less likely to achieve abstinence with transdermal NRT (Lerman et al. 2006b).
Studies investigating the role of the ANKK1 Taq1A (rs1800497) polymorphism in smoking cessation suggest that specific genotypes at this polymorphism may predict whether bupropion or NRT will be more effective for a given patient. Several studies have found that individuals homozygous for the A2 allele had higher quit rates on bupropion (David et al. 2007a; David et al. 2007c) and responded to bupropion with reduced craving, irritability and anxiety, while carriers of the A1 allele did not show positive effects (David et al. 2003). The finding that A2A2 individuals have higher quit rates on bupropion was replicated in an additional study; however, only in women (Swan et al. 2005). Interestingly, two studies have found that individuals with an A1 allele and a DBH A allele respond better to NRT (Johnstone et al. 2004). A second study also found an association between the A1 allele and higher quit rates on NRT in women, while A2A2 individuals did not benefit (Yudkin et al. 2004). Inconsistency in findings is further complicated by the lack of functional significance of this polymorphism.
While pharmacogenetic findings for some genes have been replicated, differences in study design and participant selection make comparing results between different studies challenging. The level of dependence of the subjects enrolled in the study can interact with pharmacogenetic factors that affect the outcome of addiction pharmacotherapy. Subjects who are more dependent on a substance are likely to have greater difficulty quitting, and this could make pharmacogenetic effects on treatment outcome difficult to detect. It is possible that some of the variability in results between studies investigating the same polymorphisms can be explained by the fact that researchers have selected subjects taking a different quantity of cigarettes, alcohol or other substances. In the future, genetic analyses that take into account the level of dependence of research subjects, and possibly genes previously linked to persistent addiction, in addition to genes of interest for treatment efficacy may improve study replicability and allow new genes that influence treatment outcome to be identified.
Several ethical issues related to genetic testing, including privacy, discrimination and the readiness of primary care physicians for these techniques, have been discussed (Shields, Lerman & Sullivan 2004; Munafo et al. 2005). Genetic discrimination, in the form of denied health-care coverage or increases in cost, is less likely to be a concern associated with pharmacogenetic compared to genetic testing for the prediction of disease states. However, the implementation of legislation protecting individuals from genetic discrimination, similar to what has been implemented in the UK (Department of Health 2003) and in some states in the United States (Department of Health and Human Services 2009), is a logical preventative step.
Many physicians are optimistic that the use of genetic information will improve treatment drastically. According to a recent study, 73.5 % of the 1121 physicians surveyed said they would be likely to adopt genetically tailored smoking cessation treatments (Shields et al. 2005a). However, many primary care physicians feel unprepared for counselling patients on genetic matters, indicating that educational services, resources and guidelines will need to be made available (Gurwitz, Weizman & Rehavi 2003; Gurwitz & Rehavi 2005).
Another important issue in pharmacogenetic testing is the difference in allele frequency between individuals of different ethnic backgrounds (Munafo et al. 2005; Lerman et al. 2007). In the case of smoking, there are large differences between ethnic groups in the frequency of alleles that influence the activity of nicotine metabolizing enzymes. These alleles have been found to influence the likelihood of addiction and the outcome of smoking cessation treatments. There is a possibility that ethnicities with higher frequencies of the alleles associated with risk phenotypes could experience discrimination. Additionally, since most studies have been conducted in individuals of European descent, there is a large gap in scientific data for other ethnic backgrounds. These issues have been discussed in a review (Shields et al. 2005b) and will need to be considered during the implementation of pharmacogenetic testing.
Another topic of discussion with respect to the implementation of pharmacogenetic testing is cost-effectiveness (de Leon 2009). Welton et al. (2008) performed a cost-effectiveness analysis evaluating which of four treatment strategies (NRT alone, bupropion alone, bupropion and NRT or standard care) would be the most cost-effective in the presence and absence of ANKK1 Taq1A (rs1800497) genotype information. Results indicated that the most cost-effective practice would be to prescribe both NRT and bupropion regardless of genotype. The authors indicate that these results may not hold true for tests including multiple gene variants. The most costly factor associated with testing is the physician time required for counselling related to genetic findings. Additionally, it is difficult to estimate the reduction in treatment sessions genetic testing may cause by reducing trial-and-error approaches to prescription, which may act to offset the increased time required for conveying genetic information. As mentioned earlier, there are many social and ethical issues that should be weighed with cost variables, for example, the importance of providing individuals with the most effective cessation strategy when determining how to implement genetic testing (Shields et al. 2005b).
Since addiction poses such a large individual and societal burden and response to current addiction treatments is often inadequate, the use of pharmacogenetic information may help to provide substantial improvements in treatment and avoid serious side effects (e.g. QTc prolongation). As described earlier, several gene variants have been shown to influence individual response to pharmacotherapy for drug addiction, notably OPRM1 A118G (rs561720), polymorphisms of CYP2A6 and ANKK1 Taq1A. It remains to be seen how this genetic information will be incorporated into clinical practice given the controversy regarding the implementation of genetic testing for complex phenotypes. Prospective studies evaluating the use of genetic testing in a clinical setting and the effect on treatment outcome are warranted to further evaluate the benefits and risks of this approach.
Canadian Institutes of Health Research (CIHR) operating grant, a National Alliance for Research and Schizophrenia (NARSAD) Young Investigator Award, a CIHR New Investigator Award and an Ontario Mental Health Foundation (OMHF) New Investigator Fellowship to DJM.
All authors critically reviewed content and approved final version for publication.