Genetic Analysis of Polymorphisms in Dopamine Receptor and Transporter Genes for Association with Smoking among Cancer PatientsGordiev M.a · Engstrom P.F.b · Khasanov R.a · Moroshek A.a · Sitdikov R.a · Dgavoronkov V.a · Schnoll R.A.c
aTatarstan Regional Clinical Cancer Center, Kazan, Russia; bDivision of Medical Oncology, Extramural Research Program, Fox Chase Cancer Center, and cDepartment of Psychiatry, University of Pennsylvania, Philadelphia, Pa., USA Corresponding Author
Background: Smoking among Russian cancer patients may be related to variations in the DRD2/ANKK1 (Taq1), DRD4 (exon III VNTR), and SLC6A3 genes. Methods: Seven hundred fifty patients provided smoking history and DNA. Results: Current smokers were more likely to be DRD2 A2 allele carriers versus nonsmokers (former/never smokers; 69 vs. 56%; OR = 1.69; 95% CI 1.13–2.53, p = 0.01) and former smokers (69 vs. 59%; OR = 1.54; 95% CI 0.97–2.46, p = 0.07). Ever smokers (current/former smokers) were more likely to be DRD2 A2 allele carriers versus never smokers (65 vs. 55%; OR = 1.50; 95% CI 1.00–2.27, p = 0.05). The risk of current smoking among DRD2 A2 allele carriers was present if the DRD4 short allele was also present (OR = 1.76; 95% CI 1.12–2.78, p = 0.02), and the risk of ever smoking among DRD2 A2 allele carriers was present if the DRD4 short allele was also present (OR = 1.62; 95% CI 1.02–2.55, p = 0.04). DRD2 A2 allele carriers had a shorter period of previous abstinence versus DRD2 A1 carriers (p = 0.02). Effects were not statistically significant when controlling for multiple comparisons. Conclusions: The DRD2 A2 allele may increase the risk of smoking among cancer patients, convergent with studies using non-Western samples. However, additional replication is needed.
© 2012 S. Karger AG, Basel
Continued smoking among cancer patients remains an ongoing public health problem, particularly in developing countries such as Russia. Upwards of about one third of Russian cancer patients continue to smoke following diagnosis, and the intention to quit is relatively low [1,2]. Given the risks of continued smoking in this population, including diminished treatment efficacy, increased risk of death, and worsening of the quality of life , identifying correlates of smoking among cancer patients to guide treatment implementation is a priority.
Research on the neurobiology of nicotine dependence underscores the role of dopamine . Nicotine binds to and stimulates nicotinic acetylcholine receptors (nAChRs) [5,6], increasing dopamine levels [7,8]. As with other drugs of abuse, this dopamine increase is experienced as rewarding and perpetuates drug dependence . Polymorphisms in genes that modulate the effects of nicotine on nAChRs, therefore, have been examined as risk factors for nicotine dependence  and biomarkers of treatment response .
Variants of dopaminergic genes have been evaluated as correlates of smoking behavior. The Taq1A polymorphism situated on the ANKK1 gene, located 10 kb downstream in the 3′-flanking region of the dopamine receptor 2 gene (DRD2), decreases D2 receptor binding . Initial studies linked the A1 allele to a greater risk of smoking , nicotine dependence , and progression from experimenting with tobacco to regular use ; however, this relationship has not been consistently replicated . The dopamine transporter gene SLC6A3, which transports extracellular dopamine out of the synapse, has a 40-bp functional repeat polymorphism, and the 9-repeat allele lowers transporter expression . The SLC6A3 9-repeat polymorphism has been related to longer quitting durations and a greater likelihood of cessation [18,19,20], although not in all studies . Lastly, the functionality of DRD4 exon III variable number tandem repeat (VNTR) polymorphisms has been identified, with the ‘long’ alleles (≥7 repeats) related to a blunted intracellular response to dopamine in vitro versus ‘short’ alleles (≤6 repeats) . Individuals with long alleles are less likely to quit smoking [23,24] and report higher rates of smoking and a younger age of smoking initiation [14,25]. However, this relationship was not replicated in a recent study .
To date, no study has examined these variants in association with smoking among cancer patients. Such a study should be large and include a sample with low ethnic admixture [27,28]. The present study evaluated the relationship between variants in genes functionally associated with dopamine availability with smoking in a large sample of Russian cancer patients. Results from this study may help guide treatments for a subgroup of smokers who are understudied and at particular risk for adverse health consequences from smoking.
The study sample was comprised of 750 patients diagnosed with cancer within 30 days of study enrollment and receiving care at Tatarstan Regional Clinical Cancer Center (TRCC) in Kazan, Russia (table 1). Participants were diagnosed with head and neck, colorectal, or lung cancer and were over age 18. These tumor sites were selected since they have a high frequency of smoking . Eighty-nine patients refused to enroll or provided incomplete data (participation rate = 89%).
|Table 1. Descriptive sample data (n = 750)|
All procedures were approved by the ethics committee of the TRCC. A designated research assistant used daily physician schedules to identify patients and determine eligibility and willingness to enroll into the study. Informed consent was ascertained. The research assistant conducted an assessment in a private clinic area, which included surveys and the collection of a blood sample for genetic analysis. Participants were given USD 3.00 for completing the assessment. Smokers were given a smoking cessation treatment manual .
Demographic and Medical Data
Demographic information (e.g. gender, age) and medical data (e.g. tumor site) were collected.
The Center for Epidemiologic Studies depression scale (CES-D) is a 20-item Likert measure used to assess depressive symptoms . The CES-D was administered to control for variation in smoking behavior attributable to depression symptoms .
Blood samples were collected on FTA bloodspot cards and analyzed at the TRCC using polymerase chain reaction (PCR) methods. Genotyping of the DRD2 TaqIA was performed using a fluorogenic 5′-nuclease assay as described in the NCI SNP500 Cancer Database (http://variantgps.nci.nih.gov/cgfseq/pages/home.do). A 25-ul TaqMan reaction was setup in a 96-well plate using 20 ng of genomic DNA, 2× TaqMan Universal PCR MasterMix, 900 nm primers (forward: 5′-GTGTGCAGCTCACTCCATCCT, reverse: 5′-GCAACACAGCCATCCTCAAA), and 200 nm TaqMan-MGB probes (T probe = FAM 5′-TGCCTTGACCAGCAC, C probe = VIC 5′-TGCCTCGACCAGCA). Patients were identified as A1 (A1/A1 or A1/A2) or A2 (A2/A2) allele carriers. Genotyping of SLC6A3 was performed according to past studies . Genomic DNA was amplified for 35 cycles; each cycle consisted of denaturation for 1 min at 93°C and annealing/elongation for 1 min at 72°C. The following oligonucleotide primers were used: 5′-TGTGGTGTAGGGAACGGCCTGAG-3′ and 5′-CTTCCTGGAGGTCACGGCTCAAGG-3′. PCR products were electrophoretically separated on a 5% polyacrylamide gel and their molecular weights calculated by comparing their rate of migration with that of known molecular weight standards. The number of copies of the 40-base pair repeat present was determined from the size of the product. Participants were categorized as having the 9-repeat allele (i.e. 9/9 or 9/* vs. */*, where * is alleles other than 9). Lastly, the 48-bp variable nucleotide tandem repeat of DRD4 was assessed following procedures used previously . The polymorphic region within exon 3 was amplified by PCR and fragments ranging from 270 to 570 bp, containing 2–8 repeats, were resolved by electrophoresis on a 3% agarose gel and detected with ethidium bromide staining. Patients were coded as short (≤6 repeats) or long (≥7 repeats) allele carriers.
Our primary outcome was smoking status based on self-report and defined as current (i.e. smokes regularly, cut down, or once in a while), former (i.e. used to smoke but no longer does), or never (i.e. never smoked even a puff of one cigarette) smoker as done previously  and recommended with cancer patients . As secondary outcomes, assessed only among current smokers, we examined the age of initiation, years smoked, number of previous 24-hour quitting attempts, longest duration of the previous quitting attempt, current smoking rate, and level of nicotine dependence measured by the Fagerström test for nicotine dependence (FTND) .
Descriptive statistics were computed to delineate the characteristics of the present sample. The χ2 likelihood ratio test was used to assess Hardy-Weinberg equilibrium. Separate multiple logistic regression models were examined for each gene and gene × gene interaction predicting smoking status. Following the methods of Das et al. , models compared current smokers to former smokers, current smokers to nonsmokers (former and never smokers combined), and ever smokers (current and former smokers combined) to never smokers. Models controlled for gender, tumor type, and depression since these variables have been associated previously with smoking behavior among cancer patients [1,2]. OR and 95% CI were computed for predictors. By convention, the probability value of 0.05 or less was considered statistically significant but, to be more conservative, a Bonferroni correction was applied for multiple comparisons as well. As done in a previous study , the correction was applied to each set of comparisons for the primary outcome (e.g. current vs. former smokers for the 3 main effects of each gene and the three interaction effects) and the secondary outcome (e.g. 6 dependent variables across DRD2). Thus, for each comparison, the adjusted p value for significance testing was 0.009. Analysis of variance was used to assess differences between genetic alleles in terms of the age of smoking initiation, years smoked, number of previous 24-hour quitting attempts, longest duration of the previous quitting attempt, current smoking rate, and level of nicotine dependence. Power estimates calculated during the planning of this study assumed effect sizes for genetic associations based on previous studies in the general population [19,24,27]. Analyses were conducted using the Statistical Package for the Social Sciences (version 20).
The frequencies of alleles for each gene are shown in table 2. The present frequencies did not differ from the expected frequencies based on the Hardy-Weinberg test of equilibrium (for DRD2, χ2  = 1.95, p = 0.16; for DRD4, χ2  = 1.75, p = 0.19, and for SLC6A3, χ2  = 1.41, p = 0.23). The DRD2 genotype was related to smoking status in two models based on the conventional p value cut-off. Current smokers were more likely to be DRD2 A2 allele carriers (A2/A2) versus nonsmokers (former and never smokers; 69 vs. 56%; OR = 1.69; 95% CI 1.13–2.53, p = 0.01). Ever smokers (current and former smokers) were more likely to be DRD2 A2 allele carriers (A2/A2) versus never smokers (65 vs. 55%; OR = 1.50; 95% CI 1.00–2.27, p = 0.05). In addition, current smokers were more likely to be DRD2 A2 allele carriers (A2/A2) versus former smokers, although this comparison represented only a trend based on convention (69 vs. 59%; OR = 1.54; 95% CI 0.97–2.46, p = 0.07). Models for DRD4 and SLC6A3 yielded no significant relationships between genotypes and smoking status. Furthermore, no model reached statistical significance based on the Bonferroni correction for multiple testing.
|Table 2. Frequencies of genetic alleles by smoking status|
Two models that tested gene-gene interactions yielded significant relationships with smoking status based on the conventional p value cut-off. Participants were more likely to be current smokers versus nonsmokers (former and never) if they possessed the DRD2 A2 allele (A2/A2) and the short allele for DRD4 (72 vs. 55%), but not the DRD4 long allele (62 vs. 58%; OR = 1.76, 95% CI 1.12–2.78, p = 0.02). In addition, participants were more likely to be ever smokers (current and former) versus never smokers if they possessed the DRD2 A2 allele (A2/A2) and the short allele for DRD4 (67 vs. 54%), but not the DRD4 long allele (60 vs. 57%; OR = 1.62; 95% CI 1.02–2.55, p = 0.04). Further, no model reached statistical significance based on the Bonferroni correction for multiple testing.
No associations were found between DRD2,DRD4, and SLC6A3 alleles and age of initiation, years smoked, number of previous 24-hour quitting attempts, current smoking rate, and level of nicotine dependence. Participants who were DRD2 A2 allele carriers reported a significantly shorter previous quitting attempt (M = 56.8 days; SD = 160.1) versus carriers of the DRD2 A1 allele [M = 152.9 days; SD = 411.7; F(1, 190) = 5.44, p = 0.02], although this was not statistically significant based on the Bonferroni correction.
The results of the present study suggest that the risk of smoking may be associated with the A2 allele of DRD2. Using the conventional p value cut-off, the A2 allele was related to a greater likelihood of continued tobacco use following diagnosis as well as ever having been a smoker. In addition, DRD2 A2 allele carriers reported a shorter average duration of past smoking cessation. As such, this patient subgroup may require targeted smoking cessation interventions, particularly if the short allele for DRD4 is also present. However, when correcting for multiple comparisons, these relationships were no longer considered statistically significant. Such a correction may also increase the probability of a type II error. Thus, we discuss the present findings in relation to past studies, yet with caution given the results when considering the correction for multiple testing.
The present results are divergent from studies that have identified the A1 allele for DRD2 as a correlate of smoking [13,35]. However, studies using non-Western samples have found that the A2 allele for DRD2 was associated with smoking. Two studies reported that the DRD2 A2 allele was associated with a greater risk of tobacco use in Japanese samples [36,37], as did a study with a Polish sample . Taken together, these results are suggestive of the possibility that ethnic differences may influence the relationship between DRD2 alleles and smoking. It is possible that other functional genetic variants, which differ across these ethnic groups, influence the relationship between DRD2 alleles and smoking. Alternatively, cultural differences that covary with geographic and ethnic factors could moderate the relationship between DRD2 alleles and smoking and explain these divergent findings across samples. It is also worth noting that a recent meta-analysis showed that the effects of DRD2 genetic alleles vary across different smoking phenotypes (e.g. smoking persistence vs. smoking rate) and depend on the proportion of men in a particular sample . The gender effect noted in this meta-analysis is particularly relevant given that the present sample was comprised of 65% males. Further, since the DRD2 A1 allele is in linkage disequilibrium (LD) with other DRD2 variants and possibly other variants of the ANKK1 gene, it is plausible that other functional variants in DRD2 influence smoking phenotypes. While it is unclear if the LD patterns for DRD2 are different across the geographic regions, a previous study found variation in the relationship between DRD2 variants and Parkinson’s disease across racial/ethnic groups that may be attributable to differences in LD .
The interaction with DRD4, suggesting a greater risk of smoking among the short allele carriers, is inconsistent with previous studies as well [14,23,36]. However, among depressed smokers, carriers of the DRD4 short allele are more likely to smoke to alleviate depression symptoms, an effect not detected among DRD4 long allele carriers . The high rate of depression in the present sample (64%, based on a cut-off of ≥16 on the CES-D) may explain the present finding, indicating that the relationship between DRD4 alleles and smoking, in part, depends on the presence of depression symptoms .
The present results also suggest that variants of DRD4 and SLC6A3 may have little association with smoking phenotypes. While null findings concerning smoking phenotypes and these genes have been reported [21,26], it may also be the case that differences across studies in the definitions of smoking phenotypes explain contradictory findings. There was some indication in the present analyses that DRD4 variants interact with DRD2 variants to predict smoking phenotypes, so other gene × gene interactions not tested here may be important for other smoking phenotypes for DRD4 and for SLC6A3 variants.
Study limitations should be considered. First, although the present sample was relatively large, it may still have been inadequate to provide sufficient statistical power to consider gene × gene interactions. Second, when considering the results in terms of the correction for multiple testing, there were no comparisons that reached statistical significance. As such, replication is critical to verify that the possible relationships noted in the present study were not found simply by chance. Third, while the rate of misreporting of smoking status is low in this population and unrelated to prediction models , the lack of bioverification of self-reported smoking may have affected study results. Fourth, the present study focused on only a certain subset of candidate genes and very recent data suggests a possible role of variants of genes not tested here such as variation in the neuropeptide Y gene promoter . Fifth, the candidate gene approach assumes a narrow conceptual framework for understanding nicotine dependence. Recent methods in genetic analysis, including genome-wide and deep sequencing analysis, offer more powerful approaches to identifying genetic contributions to disease risk. Finally, the prevalence of never smoking in the present study (52%) was higher than that found in previous studies (37–43%) [1,2], which may have affected the present results.
Nevertheless, the present findings contribute to the growing literature suggesting that the relationship between DRD2 genetic variants and smoking may differ across Western and non-Western samples. Further, the present findings offer some support for the identification of DRD2 A2 allele carriers and targeted treatment with smoking cessation treatments including bupropion. Previous studies have suggested that carriers of the DRD2 A2 allele are more likely to respond to bupropion compared to A1 allele carriers . Given the high rate of tobacco use following diagnosis identified in these patients in the present study and in past studies [1,2], the need for empirical methods for targeted treatments for nicotine dependence is critical and hopefully will yield improved clinical outcomes for this understudied population.
This study was funded by grant R03 TW007164 from the National Cancer Institute. The authors thank the American-Russian Cancer Alliance and Ms. Sophia Michaelson and Mr. Alan Howald for their assistance with coordinating this study. The authors thank Dr. Paul Wileyto as well for assistance with this paper.
Robert A. Schnoll, PhD
Department of Psychiatry, University of Pennsylvania
3535 Market Street, 4th Floor
Philadelphia, PA 19104 (USA)
Received: May 9, 2012
Accepted: July 7, 2012
Published online: November 2, 2012
Number of Print Pages : 7
Number of Figures : 0, Number of Tables : 2, Number of References : 42
European Addiction Research
Vol. 19, No. 2, Year 2013 (Cover Date: February 2013)
Journal Editor: van den Brink W. (Amsterdam), Kiefer F. (Mannheim)
ISSN: 1022-6877 (Print), eISSN: 1421-9891 (Online)
For additional information: http://www.karger.com/EAR