Assessment of genetic variation as a predictor of smoking cessation success: a review

Abstract Introduction Recent research strongly suggests that a specific genetic background influences smoking behaviour and may also determine the efficacy of pharmacotherapies used for smoking cessation. The aim of this critical review was to provide an overview of the developments in the pharmacogenetics of smoking cessation treatment. Discussion Several (combinations of) genetic variants in smoking-related genes (e.g. genes influencing the response to nicotine (e.g. nicotine metabolism, nicotinic receptors) and genes that may predispose to addictive behaviour due to their effects on key neurotransmitter pathways (e.g. dopamine, serotonin, opioid)) have been found to influence the level of nicotine dependence. Furthermore, the different aspects of nicotine dependence seem to be influenced by genetic variants in different pathways; ‘morning smoking’ by genes that may influence the response to nicotine, but ‘smoking pattern’ by genes that influence key neurotransmitter pathways. Moreover, several variants in smokingand treatment-related genes influence the efficacy of smoking cessation therapies, which are often distinctive for the different forms of pharmacotherapy, especially when they have a different mechanism-ofaction. Conclusion Much progress has been made in unravelling the effects of genetic variants on smoking behaviour and smoking cessation treatment, but much research still remains to be done and prospective trials should be set up to fully confirm the effect of the variants before genetically tailored smoking cessation therapy can be implemented in standard clinical practice.


Introduction
Although the risk of smoking is well documented, tobacco smoking continues to be the largest preventable cause of disease and premature death throughout the world. It is estimated that there are currently still over 1.5 billion smokers world-wide and this is expected to reach about 2 billion by 2025 1 . Smoking results in many of the most common diseases: cancers, cardiovascular diseases, and chronic lung diseases such as chronic obstructive pulmonary disease (COPD) and asthma 2 . Moreover, about half of all the smokers who continue to smoke will eventually die from a smoking-related disease, resulting in over 6 million deaths per year world-wide 3,4 . Smoking cessation can reverse many of the adverse effects of smoking 5 . However, although the majority of smokers are highly motivated to quit and much progress has been made in the (pharmacological) treatment of nicotine dependence (ND), the efficacy of available treatments is limited; only ~15%-30% continue to abstain from smoking (see Table 1) 6 .
Recent research strongly suggests that a specific genetic background influences smoking behaviour 6,7 . Since pharmacological therapies used for smoking cessation are directed at the modulation of the pathways involved in smoking behaviour, genetic variation in candidate genes for smoking behaviour will probably also influence the efficacy of smoking cessation therapies. Furthermore, genetic variants in genes influencing the metabolism and/or secretion of smoking cessation pharmacotherapies, thereby determining the level and duration of the medication in the body, might also influence the efficacy of smoking cessation treatment. Therefore, the overall effectiveness of smoking cessation therapy could be increased if the therapy is matched with the smoker's genetic background. This is expected to result in a more efficient use of anti-smoking therapies, increased cessation rates, and ultimately, in reduced morbidity and mortality deaths from smoking.
In this paper we present an overview of the developments in the assessment of genetic variation as a predictor of smoking cessation success. For this, we first shortly discuss the biological pathways associated with smoking. This is followed by an overview of the influence of genetic variants on smoking behaviour and smoking cessation treatments (for more extensive reviews see 6,7 ).

Discussion
The authors have referenced some of their own studies in this review. The protocols of these studies have been approved by the relevant ethics committees related to the institution in which they were performed.

Biological pathways associated with smoking
Nicotine is the primary reward component which is associated with the addictive effects of smoking 8 . The level of ND has been shown to depend on the amount and the way in which nicotine is delivered 9 .
The amount and the duration of nicotine in the body are determined by the rate the nicotine is metabolised. The major genes responsible for the nicotine metabolism are the hepatic enzymes cytochrome P450 2A6 (CYP2A6) and 2D6 (CYP2D6) 10 . Of these, CYP2A6 is believed to be the most important predictor of the rate of nicotine metabolism, because it is responsible for roughly 90% of the metabolic inactivation of nicotine to cotinine 10 .
Upon entering the brain, nicotine binds to nicotinic acetylcholine receptors (nAChRs), which activates them and results in the release of several types of neurotransmitters and hormones 8 . Release of these neurochemicals induces the behavioural effects associated with smoking (see Figure 1) 8 . Dopamine is the most important neurotransmitter released by nicotine; it is responsible for the pleasurable (rewarding) effects of nicotine and is therefore critical for its reinforcing effects 8,11 . Furthermore, dopamine is responsible for compelling urges such as eating 8 . Nicotine also seems to mediate the release of glutamate, which is believed to play a role in learning and memory, and facilitates the release of dopamine 8,12 . It also stimulates the release of γ-aminobutyric acid (GABA), associated with a reduction in anxiety and tension, and inhibits dopamine release 8,12 . Furthermore, nicotine has been shown to increase the secretion of serotonin in the brain, which is involved in mood regulation, impulse control, appetite and aggression 8  All authors contributed to conception and design, manuscript preparation, read and approved the final manuscript. All authors abide by the Association for Medical Ethics (AME) ethical rules of disclosure.
significant genetic influences on several aspects of the smoking behaviour 6,7,[16][17][18][19] . It has been demonstrated that genetic factors account for approximately 40%-75% of the variation in smoking initiation, 70%-80% of the variation in smoking maintenance, about 50% of the variance in cessation success and 30%-50% of the variance in risk of withdrawal symptoms 6,7,[16][17][18][19] . Variants in two broad classes of candidate genes have been suggested to contribute to the smoking behaviour: (1) genes that may influence the response to nicotine (e.g. nicotine metabolism, nicotinic receptors) and (2) genes that may predispose to addictive behaviour due to their effects on key neurotransmitter pathways (e.g. dopamine, serotonin, opioid) 6,7 .
Overall, it seems that smokers with a reduced nicotine metabolism and increased dopamine levels are less addicted to smoking, smoke less, and have a higher chance of quitting, and quit for longer periods of time 6,7 . Smokers with an increased nicotine metabolism and reduced dopamine levels, on the contrary, seem to be more addicted, have a higher chance of becoming a smoker, start smoking at a lower age, smoke more cigarettes, and undergo fewer and less successful smoking cessation attempts 6,7 . This is probably because the latter group has to smoke more to compensate for the increased nicotine metabolism and the relative deficiency in dopamine (and possibly also other neurotransmitters) that they experience. There also seems to be a relationship with variants in the serotonin pathway, but the nature of the relationship is not yet clear 6,7 . A more recent research also shows an effect of variants in the nicotinic acetylcholine receptors, especially the CHRNA5-A3-B4 cluster 20 .
So far most studies investigated only single genes. However, regarding the number of genes that have been shown to be implicated in smoking, and the large number of variants  and have an antidepressant effect, and lower serotonin re-uptake with several behavioural traits (e.g. neuroticism, novelty seeking and anxiety-related personality traits) that are related to an increased incidence of smoking, increased ND, and difficulty in quitting smoking [13][14][15] .
With chronic exposure to nicotine, tolerance develops to the release of many of these neurochemicals 8 . In the absence of nicotine, this results in a relative deficiency state, which is generally characterised by symptoms that are opposite to the acute effects of nicotine 8 . Furthermore, with long-term exposure to nicotine, some nicotinic cholinergic receptors become desensitised but some do not. As a result, GABA-mediated inhibitory tone diminishes while glutamate-mediated excitation persists, thereby increasing excitation of dopaminergic neurons and enhancing responsiveness to nicotine 12 .

Influence of genetic variation on smoking behaviour
While early reports suggested that the influence of heredity on smoking was modest, more recent twin and adoption studies have found Competing interests: declared in the article. Conflict of interests: none declared.
All authors contributed to conception and design, manuscript preparation, read and approved the final manuscript. All authors abide by the Association for Medical Ethics (AME) ethical rules of disclosure.
serotonin transporter do not seem to influence the response to NRT. It has been suggested that variants in the serotonin pathway, mainly the serotonin transporter (SERT), may be associated with the response to pharmacologic treatments for smoking cessation with an antidepressant action. We therefore investigated the influence of three functional genetic variants in the SERT gene (SLC6A4 5-HTTLPR, STin2 and rs25531) on smoking cessation rates using the antidepressants bupropion and nortriptyline 22 . Both bupropion and nortriptyline seemed to increase, possibly even double, smoking cessation rates, but only among the carriers of serotonin transporter high-activity variants. This is probably because they block the increased serotonin transporter activity, thereby increasing serotonin levels. In addition, several other variants in smoking-and treatment-related genes were found to influence cessation rates using bupropion and nortriptyline 21 . Some of these variants had comparable effects for both bupropion and nortriptyline treatment, but some genetic variants were also identified that had distinct effects depending on the type of antidepressant that was used. Genetic variants associated with a dopamine or serotonin deficiency resulted in increased cessation rates using both antidepressants. Furthermore, both seemed to attenuate the decreased cessation rates among individuals with a higher nicotine metabolism (CYP2D6), although this was not significant for nortriptyline, possibly because this variant also results in an increased nortriptyline metabolism. Moreover, cessation rates using bupropion were found to be decreased among individuals with a lower bupropion metabolism, while a variant in the acetylcholine pathway was found to be associated with nortriptyline only. Finally, variants in nicotinic receptor subunits also seemed to play the metabolism and/or secretion of smoking cessation pharmacotherapy, thereby determining the level and duration of the medication in the body, are also expected to influence the efficacy of smoking cessation therapy.
Most research on the role of genetic variation on smoking cessation pharmacotherapy have been directed to the two most widely accepted and licensed forms of smoking cessation therapy: nicotine replacement therapy (NRT) and the antidepressant bupropion (see Table 2) 6,7 . They include both placebo-controlled studies and studies comparing multiple active medications. The former may be maximally informative to determine the underlying mechanisms of the pharmacogenetic effect, while the latter may help to identify subgroups of patients who will respond optimally to a particular medication given a range of therapeutic options. In most of these studies, the available pharmacotherapies are combined with behavioural interventions. However, since participants in the different treatment groups usually all receive the same behavioural interventions, the effect of the behavioural interventions is comparable in the treatment groups.
Overall it seems that smokers with genotypes associated with reduced dopamine levels achieve better-quit rates with NRT, while genotypes associated with increased dopamine availability predict a better response to bupropion. Furthermore, a decreased metabolism of the used medication seems to increase cessation rates as well. Moreover, smokers who carry genetic polymorphisms associated with reduced nicotinic receptor (and possibly also dopaminergic) activity may experience greater benefit from the greater rewarding effects of nicotine spray (NS), while smokers with increased activity variants in the µ-opioid receptor (MOR) may have better success with the higher levels of nicotine delivered by transdermal nicotine (TN) patches. Variants in the present in these genes, approaches analysing single variants or single genes will probably fail to fully determine the influence of genetic variation on smoking behaviour. Therefore, we investigated the effect of multiple genetic variants and their interactions in a large number of smoking-related genes 21 . Several variants were found to influence ND levels and combinations of genetic variants were found to be able to have a significant effect, even if the variants do not show an effect on their own. Moreover, the different aspects of ND seemed to be influenced by genetic variants in different pathways. While 'morning smoking' (FTND [Fagerström Test of Nicotine Dependence] items 1, 3 and 5) was found to be associated mainly with genes that influence the response to nicotine, 'smoking pattern' (FTND items 2, 4, and 6) was found to be associated mainly with genes that influence key neurotransmitter pathways. These results confirm that ND is multifactorial rather than one common trait.
Therefore, we may conclude that it is important to investigate the effect of multiple genetic variants in smoking-related pathways on both the level and the different aspects of ND. This could help to better understand the underlying biological mechanisms that cause (the different aspects of) ND, which can help direct treatment to the individual needs of smokers who want to quit. Furthermore, it can guide the development of new pharmacotherapies, since these can then be directed at the pathways found to be involved in ND.

Influence of genetic variation on smoking cessation treatments
Since pharmacotherapies used for smoking cessation are directed at the pathways involved in smoking behaviour, variants in candidate genes for smoking behaviour are also expected to influence the efficacy of smoking cessation therapies. Furthermore, genetic variants in genes influencing Licensee  -Increased probability of quitting on TQD with G/G (p < 0.01).
-Increased probability of quitting on any day with G/G (p < 0.05). -Trend for bupropion being more effective than placebo for quitting with G/G (OR vs.
Thus it seems that several genetic variants in smoking-and treatmentrelated genes influence the efficacy of smoking cessation therapy and that these effects are often distinctive for the different forms of pharmacotherapy, especially when they have a different mechanism-of-action. Therefore, genotyping smokers before a cessation attempt may give directions in determining at forehand which treatment would be most effective for an individual smoker. In Figure 2 it is hypothesised how smoking cessation therapy might be genetically tailored based on the present knowledge.
Smokers with variants resulting in a dopamine or serotonin deficiency (e.g. decreased synthesis, increased re-uptake or increased metabolism) seem to achieve better-quit rates with antidepressant therapies, such as bupropion and nortriptyline. On the other hand, smokers with an effective neurotransmitter response and nicotinic receptors (e.g. increased/normal number and activity receptors), and a decreased nicotine metabolism (determined primarily by CYP2A6 genotype) achieve better-quit rates with NRT. Since varenicline is a (partial) nicotinic receptor agonist, like NRT, it is expected to be more effective among smokers with genotypes associated with an effective neurotransmitter response and nicotinic receptors as well. However, differences in the metabolism or elimination of these drugs and pathways involved in their mechanism-of-action could make one drug within these categories (e.g. nicotinic receptor agonists and antidepressants) more effective than the other, or result in less side-effects in certain subgroups of smokers. For instance, bupropion has been found to be more effective in the presence of a high bupropion metabolism and high-expression nicotinic receptors, but nortriptyline among carriers of variants resulting in a low nortriptyline metabolism and differences in the acetylcholine pathway. Furthermore, since NRT has been found to be less effective among smokers with a high nicotine metabolism, varenicline might be indicated for these smokers. In contrast, since varenicline is eliminated Competing interests: declared in the article. Conflict of interests: none declared.
All authors contributed to conception and design, manuscript preparation, read and approved the final manuscript. All authors abide by the Association for Medical Ethics (AME) ethical rules of disclosure.
tine metabolism may benefit from high-dose nicotine patches without experiencing the generally occurring side-effects. Fifthly, a marked racial/ethnic diversity exists in smoking behaviour (e.g. age of initiation, smoking rate, level of dependence) and in the frequency of functional polymorphisms. To date, the vast majority of studies have been conducted with Caucasians only to avoid population stratification. Thus, the effect of genetic variants in other racial/ethnic groups should be investigated as well.
Furthermore, some researches suggest that pharmacotherapies might work through different processes and/or are subject to different genetic influences in men and women. Therefore, the effect of genetic variations should be assessed for men and women separately.
And finally, several practical, policy and ethical considerations have to be addressed. Additional research should be conducted to examine the benefits, risks, and challenges of conveying genetic information about smoking predisposition to the patients, clinicians and the public. Economic analyses of the cost-effectiveness of using genotype information to tailor smoking treatment would also be necessary and appropriate legal and regulatory frameworks should be set up to ensure privacy and to protect against genetic discrimination.
In order to make it possible to implement genetically tailored smoking cessation therapy in general medical practice, future studies should thus investigate the effect of multiple susceptibility genes as well as their mutual interactions on several smoking cessation therapies in large-scale, comparable trials in different ethnic/ racial and gender groups. Furthermore, prospective trials should be set up to fully confirm the effect of the variants. Finally, several practical, policy and ethical considerations will trials. Therefore, findings of the previous studies should be validated across independent trials and prospective studies should be set up to fully confirm the effect of the variants.
Secondly, multiple genes (and environmental factors) are likely to influence the response to smoking cessation therapies. However, so far most studies investigated only single genes. This approach will fail to fully determine the role of genetic variation in the individual susceptibility towards smoking cessation therapies. The commonly occurring alleles have only relatively small effect (ORs in the range of 1.1-2.0) and explain only a minor proportion of the observed phenotypic variance. Combinations of genetic variants have been found to have a significant effect even if the variants do not have an effect on their own. Thus it is important to investigate the combined effect of multiple genetic variants in smoking-and treatment-related genes. However, this will require larger scale genetic trials, of hundreds to thousands of participants, to achieve significant statistical power to evaluate these gene-gene and gene-environment interactions.
Thirdly, until now the pharmacogenetics of only a couple of smoking cessation therapies has been investigated extensively (e.g. NRT and bupropion). Newer compounds (e.g. varenicline), as well as current second-line medications for smoking cessation (e.g. nortriptyline), will also require investigation.
Fourthly, genetic associations with tolerability and side-effects should also be examined. It is likely that some individuals are predisposed to have unusual reactions to drugs due to the presence of certain genetic defects. Certain subgroups of individuals may also exist who respond well to certain medications that are normally not well tolerated. For instance, individuals with a high nico-by the organic cation transporter 2 (OCT2), carriers of variants resulting in a high transporter activity might benefit more from NRT. Moreover, smokers carrying genetic polymorphisms associated with reduced nicotinic receptor (and possibly also dopaminergic) activity may experience greater benefits from the greater rewarding effects of NS, while smokers with increased activity variants in the MOR may have better success with the higher levels of nicotine delivered by TN.

Conclusion
Much progress has been made in unravelling the effect of genetic variants on smoking behaviour and smoking cessation treatment, and promising results have been found. It seems that several genetic variants in smokingrelated (e.g. influencing the response to nicotine and influencing key neurotransmitter pathways activated by nicotine) and treatment-related (e.g. metabolism and/or secretion) genes influence the efficacy of smoking cessation therapy and that these effects are often distinctive for the different forms of pharmacotherapy, especially when they have a different mechanism-of-action. Therefore, genotyping smokers before a cessation attempt may give directions in determining at forehand which treatment would be the most effective for an individual smoker. However, much research still remains to be done before genetically tailored smoking cessation therapy can be implemented in standard clinical practice.
Firstly, and most importantly, although many genetic variants have been reported to influence smoking cessation rates using NRT or bupropion, most have not yet been replicated in other studies or inconsistent results have been found. Only a few genetic variants have been investigated in more than one trial, and only a very few variants have been investigated in more than two have to be addressed. When these steps are taken, genetically tailored smoking cessation therapy will likely become available in general medical practice in the next decades, which will result in a more efficient use of anti-smoking therapies, increased cessation rates, and ultimately, in reduced morbidity and mortality deaths from smoking.