The significant worldwide health burden introduced by tobacco smoking highlights the importance of studying the genetic determinants of smoking behavior and the key factor sustaining compulsive smoking, that is, nicotine dependence (ND). We have here addressed the genetic background of smoking in a special study sample of twins, harmonized for early life events and specifically ascertained for smoking from the nationwide twin cohort of the genetically unique population of Finland. The twins and their families were carefully examined for extensive phenotype profiles and a genome-wide scan was performed to identify loci behind the smoking status, ND and the comorbid phenotype of ND and alcohol use in 505 individuals from 153 families. We replicated previous linkage findings on 10q (max logarithm of the odds (LOD) 3.12) for a smoker phenotype, and on 7q and 11p (max LOD 2.50, and 2.25, respectively) for the ND phenotype. The loci linked for ND also showed evidence for linkage for the comorbid phenotype. Our study provides confirmatory evidence for the involvement of these genome regions in the genetic etiology of smoking behavior and ND and for the first time associates drinking and smoking to a shared locus on 10q.
Tobacco use is a major health hazard worldwide,1, 2 highlighting the importance of studying the genetic determinants of smoking behavior and the key factor sustaining compulsive smoking, that is, nicotine dependence (ND; MIM 188890). Twin and family studies have suggested a high heritability (56–75%) for ND;3, 4, 5, 6, 7 clearly, social, psychological and other environmental factors as well as comorbidities, such as alcohol use and depression, contribute as well.8, 9
Several genome-wide scans have been performed suggesting loci linked to smoking behavior and ND. However, ascertainment criteria as well as study samples have been diverse, and not surprisingly, only a few findings seem consistent across studies. Depending on the study population and the phenotype criteria used, some evidence of linkage (logarithm of the odds (LOD)/Z scores ⩾2) for ND has been reported on chromosomes 2, 6, 7, 8, 10, 11 and 18,10, 11, 12, 13 linkage for smoking rate on chromosomes 3, 4, 5, 10, 11 and 17,7, 12, 13, 14, 15, 16 and linkage for ever smoking, habitual smoking or smoking initiation on chromosomes 6, 9 and 11.7, 16, 17, 18 Thus, suggestive findings for smoking-related loci exist for most chromosomes.
Drinking and smoking often co-occur,19, 20 and alcohol and tobacco are also frequently coabused,20, 21, 22, 23 indicating either that the use of one substance leads to the use of the other, or that there are shared predisposing factors. Such factors may include a common genetic component.3, 24, 25 The majority of alcoholics (80–95%) smoke, compared to 25–30% of non-alcoholics, and 70% of alcoholics are heavy smokers.26 Smoking rate has been shown to correlate with the amount of alcohol consumed, and there seems to be a correlation between the severity of alcohol and nicotine dependencies.26 Four chromosomal regions (on chromosomes 1, 2, 11 and 15) have shown modest evidence for linkage to the comorbid phenotype of habitual smoking and alcohol dependence18 in families selected for alcoholism.
In the present study, we made an effort to minimize both the genetic heterogeneity and the divergence in environment during early life events by collecting the study samples from the Finnish Twin Cohort with extensive phenotypic profiles. We specifically looked for loci linked to smoking status and ND, as well as the comorbid alcohol use and ND. We performed a genome-wide scan on 153 Finnish twin families (a total of 505 individuals) ascertained for cigarette smoking in at least two sibs and identified three major loci, all replicating previous findings and implying the real biological significance for these loci. Of special interest are the ND locus on 7q and the 10q locus exposing linkage to both the smoker status and binge drinking.
We ascertained the twins for this study based on questionnaire data collected from a nationwide cohort of twins. Concordant dizygotic (DZ) twin pairs ascertained for smoking were then targets of more detailed interviews and phenotyping for nine different phenotypes measuring smoking, ND and alcohol use/dependence (Table 1). Strong correlations were detected between several traits (Table 2) as can be expected due to the overlap in phenotypic criteria and the overall association between smoking and alcohol drinking. Our genotyping strategy was to identify encouraging regions (pairwise non-parametric linkage (NPL)/LOD score >2) in the initial genome scan in 153 twin families and then follow those regions with a denser set of markers. Here, the linkage data are provided for critical phenotypes.
Significant evidence of linkage for the smoker phenotype was observed at 10q25 (D10S597, two-point NPL 3.35, max two-point LOD 3.12) (Table 3, Figure 1). Flanking markers did not yield significant two-point NPL scores; thus, the NPL multipoint curve lies flat. Suggestive evidence of linkage for the smoker phenotype also emerged at 5q12 (D5S647, two-point NPL 2.55, max two-point LOD 2.66) and at 2q33 (D2S325, two-point NPL 1.50, max two-point LOD 2.56) (Table 3).
Suggestive evidence of linkage for FTND (Fagerström Test for Nicotine Dependence) was detected, the most interesting signals obtained for markers at 7q21 (D7S524, two-point NPL 2.00, max two-point LOD 1.61) and 7q31 (D7S486, two-point NPL 1.71, max two-point LOD 2.50) (Table 3, Figure 1). Furthermore, in the male-specific analysis, an apparent distinct locus (D7S661, max two-point LOD 2.90) emerged at 7q35, approximately 30 cM telomeric to the 7q21–31 locus (Table 4). Suggestive evidence of linkage for FTND was observed also at 5q34 (D5S400, two-point NPL 2.66, max two-point LOD 2.66) and at 11p15 (D11S4181, two-point NPL 1.76, max two-point LOD 2.25). The 11p15 locus also showed suggestive evidence for linkage with DSM-IV (Diagnostic and Statistical Manual of Mental Disorders) ND (D11S4046, two-point NPL 1.29, max two-point LOD 1.84, data not shown). Finally, DSM-IV defined ND showed some evidence for linkage at 20p13 (D20S117, two-point NPL 1.07, max two-point LOD 2.36) (Table 3).
Interestingly, the loci with suggestive evidence for linkage with FTND: 5q34, 7q21–31 and 11p15, also showed suggestive linkage with the comorbid phenotype of ND (FTND) and alcohol use. The highest LOD score for the comorbid phenotype was obtained at 7q31 (D7S486, two-point NPL 2.07, max two-point LOD 2.82) (Table 3). The linkage signal at 7q was predominantly contributed by males (Table 4). Similar to the analyses of the FTND phenotype, a male-specific locus (D7S661, max two-point LOD 2.74) emerged at 7q35 for the comorbid phenotype (Table 4).
Alcohol use phenotypes
Suggestive linkage was observed for binge drinker at 10q25 (D10S597, two-point NPL 2.09; D10S1741, max two-point LOD 2.49), overlapping the linkage locus for the smoker status. Suggestive evidence for linkage was also seen at 20p13 (D20S117, two-point NPL 1.88, max two-point LOD 1.35), overlapping the linkage locus for DSM-IV ND (Table 3). Clear sex-specific differences were only seen in chromosome 7 for the FTND and the comorbid phenotypes.
In Finland, almost one quarter of the population (26% of men, 18% of women) are daily smokers,27 with similar numbers reported throughout westernized countries. Although the prevalence of smoking has decreased significantly in many countries, particularly among men, the number of highly dependent smokers has not declined correspondingly.28
In previous linkage analyses, smoking behavior as a phenotype has been defined in several different ways: average7, 15 or maximum29, 30 number of cigarettes smoked per day, ever smoking,31 habitual smoking17, 18 and total number of cigarette packs per day for 1 year.14 Similarly, ND can be defined in various ways, the most commonly used definitions being based on the DSM-IV32 and the FTND.33 The Heaviness of Smoking Index (HSI) contains the two most informative questions of the FTND (time to first cigarette in the morning and number of cigarettes smoked per day).34
According to a wide range of family, adoption and twin studies, inter-individual differences in ND and smoking behavior are in part accounted for by inter-individual genetic differences. To begin to identify the relevant genes, several genome-wide scans have been performed. Multiple loci linked to ND and smoking behavior have been suggested; however, very few findings seem consistent across studies and very few replicated loci exist today. Explanations for the lack of consistency and difficulty to replicate include imprecise phenotype assessment, diversity in phenotype definitions, confounding population admixture effects, genotyping errors and limited sample sizes.
The current study utilizes a Finnish twin cohort with extensive phenotypic profiles, and has a number of advantages. The population of Finland represents a well-established isolate with minuscule population admixture; furthermore, Finland is ethnically and culturally highly homogeneous. In isolates, the genetic drift may lead to an overabundance of disease-alleles for particular disorders, and a high proportion of patients share these alleles identical by descent. Although the effect is strongest for rare disease alleles, isolates are also advantageous for genetic studies of common disorders.35
In contrast to many earlier studies, our study sample of twins was ascertained specifically for smoking and ND. We sampled sibpairs (DZ twin pairs) concordant for smoking based on questionnaire surveys of the Finnish twin cohort. We contacted all the heaviest smoking pairs; thus, the index subjects represent exceptionally heavy smokers at the population level. The prevalence of heavy smoking (20 cigarettes or more) among smokers in Finland is about 30%,36 whereas in the current study population (9 out of 10 were smokers), 46% were heavy smokers. Our sampling strategy resulted in an overrepresentation of tobacco use in the families giving us an extra power to detect linkage. The collected phenotypes are also more detailed than in most of the earlier studies. Importantly, the interviews on smoking history were performed by highly experienced and well-trained SSAGA (Semi-Structured Assessment for the Genetics of Alcoholism) interviewers, with quality controls to assess and minimize intra- and inter-observer variability. Since our study sample is biased toward heavy smokers, we did not determine heritability estimates for the traits using the study sample. The use of a telephone interview was the only feasible option we had for data collection. Compared to face-to-face interviews, telephone interviews have disadvantages and strengths.37 In particular, telephone interviews may be better at eliciting responses to sensitive items such as substance abuse or dependence-related symptoms, but a telephone interview does not allow picking up facial or other non-verbal cues during the interview. Biochemical markers of smoking exposure were not assessed; however, we assessed lifetime smoking behavior and only some of our smokers were current smokers. There are no biochemical markers for having ever been a smoker or for the degree of ND. However, earlier studies in Finland have indicated that the current smokers do report their smoking status very accurately.38 Based on previous studies of the Finnish Twin Cohort, heritability estimates for smoking-related traits in Finland are similar to those reported for other Caucasian populations.39, 40
The smoker phenotype yielded the highest linkage scores in this study (10q25, D10S597, NPL 3.35, LOD 3.12), less than 5 Mb apart from a previously reported linkage locus for the Fagerström Tolerance Questionnaire- (an earlier version of FTND)41, 42 based ND in 130 New Zealand families ascertained for ND,10 and 37 Mb apart from a locus for smoking quantity recently identified in an African-American population.13 As the flanking markers yielded unsubstantial two-point NPL scores, the multipoint NPL curve remains flat. The definition for the smoker phenotype, assessed with a single question ‘have you smoked ⩾100 cigarettes during your lifetime?’ based on World Health Organization criteria, can be considered as too encompassing of all kinds of smokers. However, the smoker phenotype significantly correlates (Spearman r=0.94) with a regular smoker phenotype assessed with a question ‘have you smoked regularly for at least 2 months?’ This reflects the high addictive potential of cigarettes. The vast majority of smokers in our study had smoked for years. We confirm the smoker phenotype43 as an easy-to-obtain, reliable measure for smoking behavior.
Interestingly, the 10q locus with significant evidence of linkage for the smoker status also showed suggestive evidence of linkage for the binge drinker phenotype. This may reflect the shared genetic etiology of these traits, and warrants future studies.
We have utilized the two most widely used measures of ND: the DSM-IV32 and the FTND.33 The FTND and the DSM-IV appear to measure different aspects of the tobacco-dependence process: DSM-IV aims to measure loss of control in respect to smoking behavior, whereas FTND focuses on the consumption of cigarettes and the difficulty of tolerating reduced nicotine levels. In a population-based Finnish adult health behavior study, 44% of daily smokers scored 4 points or more on the FTND scale (n=3500, 795 were daily smokers and 641 answered to all six FTND questions) (A Haukkala and S Helakorpi, personal communication;44, 45). Population estimates of DSM-IV defined ND are not available from Finland.
In our data set, the FTND phenotype yielded suggestive evidence for linkage at multiple loci (5q34, 7q21–31, 11p15). In contrast, the DSM-IV phenotype showed suggestive linkage independent from FTND only at 20p13, in agreement with the concept that these measures monitor different aspects of nicotine addiction. DSM-IV and FTND linkage loci coincided, if at all, only in 11p15. As observed earlier,46, 47 the various measures of ND were not always highly correlated among our subjects, indicating that these measure different aspects of dependence; thus, it can also be expected that genetic findings related to FTND and DSM-IV may sometimes be phenotype-specific.
Although the linkage signal at 11p15 was somewhat modest, suggestive evidence for linkage was seen with several phenotypes. The linkage signal appears in the telomeric region raising questions concerning the reliability of the finding; however, the fact that our 11p15 locus replicates a locus for habitual smoking previously identified in 12 US families (142 genotyped individuals) ascertained for panic disorder,17 increases the likelihood of a true finding rather than an artifact. The 11p15 locus indeed seems intriguing, and contains at least two plausible candidate genes, DRD4 and CHRNA10 (MIM 606372). We tested an intragenic minisatellite marker in DRD4 for association with negative results. CHRNA10 has not been shown to be expressed in the brain but should, however, be investigated.
The FTND linked locus at 7q spans some 24 cM and contains a multitude of positional candidate genes. Interestingly, sex-specific analyses showed that the linkage signal is predominantly contributed by males; however, the number of affected females was slightly lower than the number of affected males (90 vs 138). Furthermore, an apparent distinct locus at 7q35, approximately 30 cM telomeric to the 7q21–31 locus, appeared in males. This 7q35 locus overlaps with a ND locus recently reported in 158 US families ascertained for smoking.12 Although none of the linkage signals at 7q reached genome-wide significance, our results, especially for the 7q35–36 region, can be considered as a replication of the earlier linkage finding. It seems evident that in this homogeneous Finnish study sample a wide segment at 7q, potentially representing multiple loci, cosegregates with ND in males.
The comorbidity of ND and psychiatric illnesses, such as depression and schizophrenia, is well recognized but poorly understood. Interestingly, the same marker yielding the highest two-point LOD score for FTND and the comorbid phenotype at 7q (D7S486) also yielded the highest two-point LOD score in a Finnish genome-wide scan for schizophrenia.48 Whether this is a pure coincidence, or a representation of shared genetic etiology, remains to be resolved.
We then stratified the families into those linked to 10q (two-point LOD >0.10; 73 families, α=0.48) and those with no evidence of linkage to 10q (two-point LOD ⩽0.00; 67 families). Interestingly, the 7q linkage for FTND was contributed by those families linked to 10q for the smoker phenotype. This is not explained by a difference in information content as the number of affected sibpairs (ASPs) was higher in families not linked to 10q (130 ASPs for smoker, 46 ASPs for FTND) than in those linked to 10q (111 ASPs for smoker, 27 ASPs for FTND). This initial result may suggest a co-occurrence of genetic predisposition to smoking and ND in these families.
Smoking commonly co-occurs with alcohol dependence,20, 21, 23 suggesting shared predisposing factors, including shared genetic components. Several studies have suggested shared genetic etiology for smoking and alcohol use.18, 25, 49, 50 In our study, the linked loci for FTND and the comorbid phenotype (FTND+binge drinker) consistently coincided, supporting the hypothesis of shared genetic etiology. However, as the study sample has been ascertained for smoking, alcohol use is expected to be more prevalent than in the population at large. The prevalence of the binge drinker phenotype among Finnish young adult twins (n=4667) has been estimated to be 78% (J Kaprio, unpublished data), identical to the prevalence in the study population (79%); however, one would expect a much lower prevalence for binge drinking in older cohorts, such as the one we have been utilizing.
In conclusion, we detected significant linkage for smoking behavior at 10q25, replicating a ND susceptibility locus identified by Straub et al.10 Interestingly, the 10q25 locus also shows evidence for linkage for binge drinking, supporting the hypothesis of a shared genetic etiology between ND and alcohol use. Furthermore, we obtained suggestive linkage for ND at 7q21–31, fully overlapping our linkage locus for the comorbid phenotype of ND and alcohol use. The chromosome 7q linkage appears to further extend some 30 cM toward the telomere, especially in males, thus replicating a ND locus recently identified by Swan et al.12 Finally, we demonstrated suggestive linkage for ND at 11p15, replicating a previously reported susceptibility locus17 around the DRD4 gene locus. As we have tested a total of nine phenotypes and 380 markers, the observed LOD/NPL scores are inflated. Thus, even the NPL score of 3.35 on chromosome 10q25 cannot be regarded as genome-wide significant. However, our findings are supported by previous findings in the literature suggesting that these loci may harbor susceptibility genes with universal impact. By further utilizing this genetically homogeneous Finnish study sample, we have excellent chances of pinpointing the predisposing variant(s).
Materials and methods
Twin pairs concordant for smoking were identified from the Finnish Twin Cohort based on earlier health questionnaire surveys during 1996–1997 (for opposite-sex twins) and in 1975, 1981 and 1990 (for same-sex twins).51 Twins who were born between 1938 and 1957 as well as their siblings and parents were recruited and interviewed. In addition, we recruited a small number of offspring. Index twins smoked on average 20.2 cigarettes per day (s.d. 9.2); in 7% of pairs the index twins smoked less than 15 cigarettes per day. Compared to twin pairs who were invited to participate in the study, but declined to do so, the participating twin pairs were slightly more often lighter smokers and somewhat more often female. The diagnostic interview was based on the SSAGA,52 with the section on nicotine use and dependence based on the Composite International Diagnostic Interview.53 The interview content was modified for the Nicotine Addiction Genetics Study (NAG; an international consortium among Australia, Finland and USA) and was translated into Finnish by professional translators. The interview staff, consisting of Masters of Health Care, Psychology or Education, or registered nurses, had extensive prior experience with SSAGA interviews. The principal investigator of the NAG study (PAF Madden), assisted by other faculty, postdoctoral trainees and senior interviewers from Finland, Australia and US, trained the interview staff in the specifics for the NAG interview instrument. The customized computer-assisted telephone interviews included more than 100 questions on smoking behavior, taking an average of 30–40 min out of the approximately 2 h interview. Two senior interviewers were responsible for monitoring the interview quality and reliability by reviewing each completed interview as well as by blindly reviewing and independently coding a portion of the audiotaped interviews to provide a measure of inter-rater reliability. Each family member was interviewed by a different interviewer, based on assignments created by custom-built in-house software for interview scheduling and management. Subjects returned signed consent forms by mail after the interview. Participants gave blood samples at their local health center or laboratory, and blood samples were sent to the National Public Health Institute in Helsinki, Finland. DNA was extracted from blood samples using standard procedures. For the genome-wide scans, 505 individuals from 153 families with interview data and blood sample were available. In addition, blood samples were available for three individuals who refused to participate in the interview but gave a written informed consent for the use of their DNA sample; we have included these individuals in the analyses to provide phase information. An additional 21 individuals from five families were included at the time of fine-mapping. The study has been approved by the Ethics committee of the hospital district of Helsinki and Uusimaa in 2001.
The study sample consisted of 51% males and 49% females, with a mean age of 58.4 years. Table 5 lists the characteristics for the total of 529 individuals from 158 families. Two-thirds of families had two or three siblings. Because of the mean age of the siblings, very few parents were available for the study. The study subjects were interviewed between February 2002 and March 2004.
All the phenotypes used in analyses are based on the interview data, and their definitions are presented in Table 1. We used two phenotypes for smoking behavior (smoker,43 heavy smoker), three phenotypes for ND (DSM-IV, FTND, HSI+DSM-IV), three phenotypes for alcohol use (regular drinker, binge drinker, DSM-IV alcohol dependent), and a comorbid phenotype of ND and alcohol use (FTND+binge drinker). In the Finnish population, binge drinking has proven to be a good predictor of morbidity from different disorders, including dementia and cardiovascular disease;54, 55 thus, we chose to implement the binge drinker trait into our comorbid phenotype. All phenotypes were assessed as dichotomized (that is, yes/no) variables. As ND develops only after exposure, individuals not fulfilling the smoker phenotype have missing values for the other smoking and ND phenotypes.
Genome-wide scans were performed at the Finnish Genome Center, University of Helsinki, utilizing the ABI PRISM Linkage mapping Set MD10, with 380 microsatellite markers (363 autosomal markers), yielding an approximately 10 cM resolution. Two different platforms, the MegaBACE1000 (Amersham Biosciences, Piscataway, NJ, USA) automated 96-capillary electrophoresis system and the ABI3730 (Applied Biosystems, Foster City, CA, USA) automated 48-capillary electrophoresis system were used. Data generated with the two systems were harmonized using CEPH reference samples present in each plate and 96 samples that have been run on both genotyping platforms. Standard PCR protocols were used for the amplification of fragments using 10 ng of genomic DNA as a template. Each of the sample plates contained one blank control, two CEPH reference samples and one duplicate sample. All of these controls were in shifting positions in the sample plates allowing us to monitor for plate orientation mix-ups. These control samples further helped us in controlling for plate mix-ups, sample mix-ups, genotyping errors, null alleles and marker mutations. In addition, the duplicate samples, as well as any MZ (monozygotic) pairs, allowed us to estimate the overall error rate.
The genotype calls were made with the GeneticProfiler1.5 (MegaBACE1000) and GeneMapper3.7 (ABI3730) software. To minimize the amount of allele calling errors and acceptance of low-quality genotypes, allele calls were verified by two independent reviewers. Genotypes that could not be reliably scored were excluded at this point, as were markers that consistently performed badly. Markers with contamination in the blank wells or with both CEPH reference samples failing were also excluded. Quality checks were performed at two stages: before and after entering the genotype data into the database.
Pre-database quality checks
Custom PERL scripts developed by scientists at the Finnish Genome Center were used to detect suspicious peaks based on signal intensity, peak width, binning accuracy, allele height ratios and to check for identical genotypes (MZ twins), duplicate sample inconsistencies and water controls. Checks were applied for each electrophoresis run separately, and any genotypes failing to meet preset criteria were deleted. PedCheck56 was run to detect Mendelian errors and the electropherograms for genotypes generating non-Mendelian inheritances were manually reviewed to resolve possible genotyping errors. The unresolved errors are likely to represent new marker mutations or allele drop-outs, and the genotypes for the corresponding nuclear family were removed.
Database quality checks
Once each run was checked, the genotype data were uploaded into a relational database developed in-house. The database is implemented in Oracle and is running on a 64-bit Linux server. Several quality checks were performed for the whole data set: (1) CEPH reference sample genotypes were compared with the expected genotypes, (2) a check for identical genotypes (MZ and duplicate sample check) was performed, (3) X-chromosomal check was used to identify males heterozygous for any X-chromosomal marker and females homozygous for all X-chromosomal markers, (4) the program GRR57 was run to detect inconsistencies within family structures, (5) the program PedCheck was then used to detect Mendelian errors and (6) genotyping success rates for samples and markers were calculated. All electropherograms for genotypes generating non-Mendelian inheritances were manually reviewed to resolve possible genotyping errors. The following issues were revealed as likely reasons for non-Mendelian inheritances: sample performing weakly in PCR (low-quality DNA), new mutation (1 repeat unit from the expected allele), null alleles, allele drop-outs, high stutter resulting in wrong peak being labeled, problems in color separation and variable allele amplification. Three markers and six individuals with success rates below 70% were excluded from analyses. Two nuclear families were excluded due to an excess of Mendelian errors most likely indicating errors in the pedigree structure. GRR detected two MZ pairs from which we deleted one individual. Also two sibpairs with too low IBS-sharing were detected and excluded. One sib from a sibship of eight sibs was deleted due to poorly working PCR. The total number of individuals after the exclusions was 529 (158 families). Based on over 2500 genotype comparison from the two MZ pairs and six duplicate pairs, the overall error rate was 0.2%. The average Mendelian error rate was 1.2 errors/1000 genotypes, out of which 0.4/1000 are likely genotyping errors and 0.8/1000 are likely new mutations or null alleles. The average success rate was 94.4% for individuals and 94.8% for markers.
For fine-mapping purposes, additional 19 microsatellite markers selected from the deCODE map58 were genotyped using the above described protocol. In addition, the 48-bp variable number tandem repeat polymorphism in DRD4 (MIM 126452) exon 3 was amplified with a fluorescently labeled forward primer (5′-IndexTermGCG ACT ACG TGG TCT ACT CG-3′) and an unlabeled reverse primer (5′-IndexTermAGG ACC CTC ATG GCC TTG-3′), with modifications of a previously described method.59 PCRs were performed in 10 μl volume, using 20 ng genomic DNA, 1 μ M of each primer, 1 × PCR buffer (Roche, Basel, Switzerland), 1 × MasterAmp PCR Enhancer (Epicentre, Madison, WI, USA), 200 μ M of dATP, dCTP, dTTP, 150 μ M of dGTP, 50 μ M deaza-dGTP (Roche) and 0.9 U of AmpliTaq Gold polymerase (Roche) in 10% dimethyl sulfoxide. PCRs were carried out on MJ Thermocyclers using the following cycling conditions: initial denaturation at 95°C for 15 min, followed by 30 cycles of 95°C for 40 s, 59°C for 20 s, and 72°C for 80 s, and a final extension of 10 min at 72°C. The length of the fragments was determined by running diluted PCR products in ABI3730 automatic DNA sequencer (Applied Biosystems). The genotype calls were made with GeneMapper v3.0 (Applied Biosystems) software and manually verified.
Genetic marker map construction
Since multipoint linkage analyses applied here are highly sensitive to errors in the genetic map, we made every effort to construct a reliable genetic map that was based on the sequence data rather than statistical inference from genotype data. First, we identified the physical location of the markers from the University of California Santa Cruz (UCSC) hg17 assembly (NCBI Build 35) database and ordered the markers based on this sequence information. The genetic location of each marker was defined using the published deCODE genetic map locations,58 which are also stored in the UCSC database. For markers that were not included in the deCODE genetic map, we used linear interpolation for obtaining estimates of the genetic locations of these markers by using the physical and the genetic locations of the immediately flanking deCODE markers. If the physical location from the UCSC database and the genetic location from the deCODE genetic map for a given marker were in disagreement, we obtained an estimate of the genetic location via interpolation using the nearest flanking deCODE markers that were in agreement with the sequence information.
In all our analyses, we defined individuals fulfilling the phenotypic criteria as ‘affected’ and all others ‘unknown’. Marker allele frequencies were obtained from the study sample. Non-parametric two-point and multipoint linkage analyses were performed using Merlin.60 Parametric two-point linkage analyses were performed using Autogscan,61 using dominant and recessive models with conservative parameters (penetrance 90%, disease allele frequency 1%, phenocopy rate (A) 1% and (B) 0.01%). Two different sets of parameters were used due to the difficulty of estimating the model for these complex traits under study. Parametric analyses were performed for the whole study sample as well as separately for males and females.
As DRD4 resides in one of our linked regions, we tested an intragenic DRD4 minisatellite marker for association using Pseudomarker62 with default dominant and recessive models.
Diagnostic and Statistical Manual of Mental Disorders, 4th edition
Fagerström Test for Nicotine Dependence
Heaviness of Smoking Index
University of California Santa Cruz
Doll R, Peto R, Boreham J, Sutherland I . Mortality in relation to smoking: 50 years' observations on male British doctors. BMJ 2004; 328: 1519–1528.
Shafey O, Dolwick S, Guindon GE (eds). Tobacco Control Country Profiles. American Cancer Society: Atlanta, GA, 2003.
True WR, Xian H, Scherrer JF, Madden PAF, Bucholz KK, Heath AC et al. Common genetic vulnerability for nicotine and alcohol dependence in men. Arch Gen Psychiatry 1999; 56: 655–661.
Kendler KS, Neale MC, Sullivan P, Corey LA, Gardner CO, Prescott CA . A population-based twin study in women of smoking initiation and nicotine dependence. Psychol Med 1999; 29: 299–308.
Lessov CN, Martin NG, Statham DJ, Todorov AA, Slutske WS, Bucholz KK et al. Defining nicotine dependence for genetic research: evidence from Australian twins. Psychol Med 2004; 34: 865–879.
Maes HH, Sullivan PF, Bulik CM, Neale MC, Prescott CA, Eaves LJ et al. A twin study of genetic and environmental influences on tobacco initiation, regular tobacco use and nicotine dependence. Psychol Med 2004; 34: 1251–1261.
Vink JM, Beem AL, Posthuma D, Neale MC, Willemsen G, Kendler KS et al. Linkage analysis of smoking initiation and quantity in Dutch sibling pairs. Pharmacogenomics J 2004; 4: 274–282, Erratum in: Pharmacogenomics J 2004; 4: 345–346.
Hämäläinen J, Kaprio J, Isometsä E, Heikkinen M, Poikolainen, Lindeman S et al. Cigarette smoking, alcohol intoxication and major depressive episode in a representative population sample. J Epidemiol Community Health 2001; 55: 573–576.
Johnson EO, Rhee SH, Chase GA, Breslau N . Comorbidity of depression with levels of smoking: an exploration of the shared familial risk hypothesis. Nicotine Tob Res 2004; 6: 1029–1038.
Straub RE, Sullivan PF, Ma Y, Myakishev MV, Harris-Kerr C, Wormley B et al. Susceptibility genes for nicotine dependence: a genome scan and follow-up in an independent sample suggest that regions on chromosomes 2, 4, 10, 16, 17 and 18 merit further study. Mol Psychiatry 1999; 4: 129–144.
Sullivan PF, Neale BM, van den Oord E, Miles MF, Neale MC, Bulik CM et al. Candidate genes for nicotine dependence via linkage, epistasis, and bioinformatics. Am J Med Genet B Neuropsychiatr Genet 2004; 126: 23–36.
Swan GE, Hops H, Wilhelmsen KC, Lessov-Schlaggar CN, Cheng LS-C, Hudmon KS et al. A genome-wide screen for nicotine dependence susceptibility loci. Am J Med Genet B Neuropsychiatr Genet 2006; 141B: 358–360.
Li MD, Payne TJ, Ma JZ, Lou XY, Zhang D, Dupont RT et al. A genomewide search finds major susceptibility loci for nicotine dependence on chromosome 10 in African Americans. Am J Hum Genet 2006; 79: 745–751.
Duggirala R, Almasy L, Blangero J . Smoking behavior is under the influence of a major quantitative trait locus on human chromosome 5q. Genet Epidemiol 1999; 17: S139–S144.
Li MD, Ma JZ, Cheng R, Dupont RD, Williams NJ, Crews KM et al. A genome-wide scan to identify loci for smoking rate in the Framingham Heart Study population. BMC Genet 2003; 4: S103–S107.
Morley KI, Medland SE, Ferreira MA, Lynskey MT, Montgomery GW, Heath AC et al. A possible smoking susceptibility locus on chromosome 11p12: evidence from sex-limitation linkage analyses in a sample of Australian twin families. Behav Genet 2006; 36: 87–99.
Gelernter J, Liu X, Hesselbrock V, Page GP, Goddard A, Zhang H . Results of a genomewide linkage scan: support for chromosomes 9 and 11 loci increasing risk for cigarette smoking. Am J Med Genet B Neuropsychiatr Genet 2004; 128: 94–101.
Bierut LJ, Rice JP, Goate A, Hinrichs AL, Saccone NL, Foroud T et al. A genomic scan for habitual smoking in families of alcoholics: common and specific genetic factors in substance dependence. Am J Med Genet A 2004; 124: 19–27.
Kaprio J, Koskenvuo M, Sarna S . Cigarette smoking, use of alcohol, and leisure-time physical activity among same-sexed adult male twins. Prog Clin Biol Res 1981; 69: 37–46.
Istvan J, Matarazzo JD . Tobacco, alcohol, and caffeine use: a review of their interrelationships. Psychol Bull 1984; 95: 301–326.
Maletzky BM, Klotter J . Smoking and alcoholism. Am J Psychiatry 1974; 131: 445–447.
Kaprio J, Hammar N, Koskenvuo M, Floderus-Myrhed B, Langinvainio H, Sarna S . Cigarette smoking and alcohol use in Finland and Sweden: a cross-national twin study. Int J Epidemiol 1982; 11: 378–386.
Battjes RJ . Smoking as an issue in alcohol and drug abuse treatment. Addict Behav 1988; 13: 225–230.
Swan GE, Carmelli D, Cardon LR . Heavy consumption of cigarettes, alcohol and coffee in male twins. J Stud Alcohol 1997; 58: 182–190.
Madden PA, Bucholz KK, Martin NG, Heath AC . Smoking and the genetic contribution to alcohol-dependence risk. Alcohol Res Health 2000; 24: 209–214.
Batel P, Pessione F, Maitre C, Rueff B . Relationship between alcohol and tobacco dependencies among alcoholics who smoke. Addiction 1995; 90: 977–980.
Helakorpi S, Patja K, Prättälä R, Uutela A . Suomalaisen aikuisväestön terveyskäyttäytyminen ja terveys, kevät 2005. Health Behavior and Health among the Finnish Adult Population, Spring 2005. Kansanterveyslaitoksen julkaisuja B 18/2005 Publications of the National Public Health Institute B 18/2005: Helsinki, 2005.
Warner KE, Burns DM . Hardening and the hard-core smoker: concepts, evidence, and implications. Nicotine Tob Res 2003; 5: 37–48.
Goode EL, Badzioch MD, Kim H, Gagnon F, Rozek LS, Edwards KL et al. Framingham Heart Study Multiple genome-wide analyses of smoking behavior in the Framingham Heart Study. BMC Genet 2003; 4: S102.
Bergen AW, Yang XR, Bai Y, Beerman MB, Goldstein AM, Goldin LR, Framingham Heart Study. Genomic regions linked to alcohol consumption in the Framingham Heart Study. BMC Genet 2003; 4 (Suppl 1): S101.
Bergen AW, Korczak JF, Weissbecker KA, Goldstein AM . A genome-wide search for loci contributing to smoking and alcoholism. Genet Epidemiol 1999; 17: S55–S60.
American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th edn. American Psychiatric Press: Washington, DC, 1994.
Heatherton TF, Kozlowski LT, Frecker RC, Fagerström K-O . The Fagerström Test for Nicotine Dependence: a revision of the Fagerström Tolerance Questionnaire. Br J Addict 1991; 86: 1119–1127.
Heatherton TF, Kozlowski LT, Frecker RC, Rickert W, Robinson J . Measuring the heaviness of smoking: using self-reported time to the first cigarette of the day and number of cigarettes smoked per day. Br J Addict 1989; 84: 791–799.
Peltonen L, Palotie A, Lange K . Use of population isolates for mapping complex traits. Nat Rev Genet 2000; 1: 182–190.
Broms U, Silventoinen K, Madden PAF, Heath AC, Kaprio J . Genetic architecture of smoking behaviour: a study of Finnish adult twins. Twin Res Hum Genet 2006; 9: 64–72.
Armstrong BK, White E, Saracci R . Principles of exposure measurement in Epidemiology. Monographs in Epidemiology and Biostatistics, vol. 21. Oxford University Press: New York, 1994.
Vartiainen E, Seppälä T, Lillsunde P, Puska P . Validation of self reported smoking by serum cotinine measurement in a community-based study. J Epidemiol Community Health 2002; 56: 167–170.
Madden PA, Heath AC, Pedersen NL, Kaprio J, Koskenvuo MJ, Martin NG . The genetics of smoking persistence in men and women: a multicultural study. Behav Genet 1999; 29: 423–431.
Madden PA, Pedersen NL, Kaprio J, Koskenvuo MJ, Martin NG . The epidemiology and genetics of smoking initiation and persistence: crosscultural comparisons of twin study results. Twin Res 2004; 7: 82–97.
Fagerström KO . Measuring degree of physical dependence to tobacco smoking with reference to individualization of treatment. Addict Behav 1978; 3: 235–241.
Fagerström KO, Schneider NG . Measuring nicotine dependence: a review of the Fagerström Tolerance Questionnaire. J Behav Med 1989; 12: 159–182.
CDC. Annual smoking-attributable mortality, years of potential life lost, and productivity losses – United States, 1997–2001. MMWR 2005; 54: 625–628.
Haukkala A, Laaksonen M, Uutela A . Smokers who do not want to quit – is consonant smoking related to lifestyle and socioeconomic factors? Scand J Public Health 2001; 29: 226–232.
Helakorpi S, Berg M, Uutela A, Puska P . Suomalaisen aikuisväestön terveyskäyttäytyminen, kevät 1994. Health behaviour among Finnish adult population, Spring 1994. Kansanterveyslaitoksen julkaisuja B 8/1994 Publications of the National Public Health Institute B 8/1994: Helsinki, 1994.
Moolchan ET, Radzius A, Epstein DH, Uhl G, Gorelick DA, Cadet JL et al. The Fagerström Test for Nicotine Dependence and the Diagnostic Interview Schedule. Addict Behav 2002; 27: 101–113.
Etter JF . A comparison of the content-, construct- and predictive validity of the cigarette dependence scale and the Fagerström test for nicotine dependence. Drug Alcohol Depend 2005; 77: 259–268.
Ekelund J, Lichtermann D, Hovatta I, Ellonen P, Suvisaari J, Terwilliger JD et al. Genome-wide scan for schizophrenia in the Finnish population: evidence for a locus on chromosome 7q22. Hum Mol Genet 2000; 9: 1049–1057.
Madden PA, Heath AC . Shared genetic vulnerability in alcohol and cigarette use and dependence. Alcohol Clin Exp Res 2002; 26: 1919–1921.
Ye Y, Zhong X, Zhang H . A genome-wide tree- and forest-based association analysis of comorbidity of alcoholism and smoking. BMC Genet 2005; 6: S135.
Kaprio J, Koskenvuo M . Genetic and environmental factors in complex diseases: the Older Finnish Twin Cohort. Twin Res 2002; 5: 358–365.
Bucholz KK, Cadoret R, Cloninger CR, Dinwiddie SH, Hesselbrock VM, Nurnberger Jr JI et al. A new, semi-structured psychiatric interview for use in genetic linkage studies: a report on the reliability of the SSAGA. J Stud Alcohol 1994; 55: 149–158.
Cottler LB, Robins LN, Grant BF, Blaine J, Towle LH, Wittchen HU et al. The CIDI-core substance abuse and dependence questions: cross-cultural and nosological issues. The WHO/ADAMHA field trial. Br J Psychiatry 1991; 159: 653–658.
Kauhanen J, Kaplan GA, Goldberg DE, Salonen JT . Beer binging and mortality: results from the Kuopio ischaemic heart disease risk factor study, a prospective population based study. BMJ 1997; 315: 846–851.
Järvenpää T, Rinne JO, Koskenvuo M, Räihä I, Kaprio J . Binge drinking in midlife and dementia risk. Epidemiology 2005; 16: 766–771.
O'Connell JR, Weeks DE . PedCheck: a program for identifying genotype incompatibilities in linkage analysis. Am J Hum Genet 1998; 63: 259–266.
Abecasis GR, Cherny SS, Cookson WO, Cardon LR . GRR: graphical representation of relationship errors. Bioinformatics 2001; 17: 742–743.
Kong A, Gudbjartsson DF, Sainz J, Jonsdottir GM, Gudjonsson SA, Richardsson B et al. A high-resolution recombination map of the human genome. Nat Genet 2002; 31: 241–247.
Lichter JB, Barr CL, Kennedy JL, Van Tol HH, Kidd KK, Livak KJ . A hypervariable segment in the human dopamine receptor D4 (DRD4) gene. Hum Mol Genet 1993; 2: 767–773.
Abecasis GR, Cherny SS, Cookson WO, Cardon LR . Merlin – rapid analysis of dense genetic maps using sparse gene flow trees. Nat Genet 2002; 30: 97–101.
Hiekkalinna T, Terwilliger JD, Sammalisto S, Peltonen L, Perola M . AUTOGSCAN: powerful tools for automated genome-wide linkage and linkage disequilibrium analysis. Twin Res Hum Genet 2005; 8: 16–21.
Göring HHH, Terwilliger JD . Linkage analysis in the presence of errors IV: joint pseudomarker analysis of linkage and/or linkage disequilibrium on a mixture of pedigrees and singletons when the mode of inheritance cannot be accurately specified. Am J Hum Genet 2000; 66: 1310–1327.
We are grateful to the participating families for their patience and contribution. We express our appreciation to the skillful study interviewers: A Häppölä, A-M Iivonen, K Karhu, H-M Kuha, U Kulmala-Gråhn, M Mantere, K Saanakorpi, M Saarinen, R Sipilä, L Viljanen and E Voipio. We are grateful to D Statham for helping to train the interviewers. We thank P Ruokolinna for thorough data entering. We thank the excellent staff at the Finnish Genome Center: S Ekström, A Yliperttula, A Korvenpää and M Baarman for producing the genotype data, A Sarahonka and V Leppä for diligently reviewing the genotype data, A Leinonen and J Hinkula for careful data cleaning procedures and Dr J Muilu for supervising the data management team. The study is funded by the Academy of Finland Post-doctoral Fellowship to A Loukola, by Doctoral Programs of Public Health, University of Helsinki, Finland to support U Broms, by GenomEUtwin (European Union Contract No. QLG2-CT-2002-01254) to J Kaprio and L Peltonen, by NIH grant DA12854 to PAF Madden, and by Center of Excellence in Disease Genetics, Academy of Finland to L Peltonen. Data collection was supported by a NIH grant DA12854 to PAF Madden and by Academy of Finland grants to J Kaprio and M Koskenvuo. ML Pergadia is supported by NIH grant DA019951. This study forms a part of the research program of the Academy of Finland Center of Excellence in Complex Disease Genetics.
Duality of Interest
About this article
Translational Psychiatry (2017)
Reframing video gaming and internet use addiction: empirical cross-national comparison of heavy use over time and addiction scales among young users
Association of the OPRM1 Variant rs1799971 (A118G) with Non-Specific Liability to Substance Dependence in a Collaborative de novo Meta-Analysis of European-Ancestry Cohorts
Behavior Genetics (2016)
Introduction to Deep Sequencing and Its Application to Drug Addiction Research with a Focus on Rare Variants
Molecular Neurobiology (2014)
Genome-wide association study on detailed profiles of smoking behavior and nicotine dependence in a twin sample
Molecular Psychiatry (2014)