UBC Faculty Research and Publications

The genetics of smoking in individuals with chronic obstructive pulmonary disease Obeidat, Ma’en; Zhou, Guohai; Li, Xuan; Hansel, Nadia N; Rafaels, Nicholas; Mathias, Rasika; Ruczinski, Ingo; Beaty, Terri H; Barnes, Kathleen C; Paré, Peter D; Sin, Don D Apr 10, 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


52383-12931_2018_Article_762.pdf [ 2.36MB ]
JSON: 52383-1.0365551.json
JSON-LD: 52383-1.0365551-ld.json
RDF/XML (Pretty): 52383-1.0365551-rdf.xml
RDF/JSON: 52383-1.0365551-rdf.json
Turtle: 52383-1.0365551-turtle.txt
N-Triples: 52383-1.0365551-rdf-ntriples.txt
Original Record: 52383-1.0365551-source.json
Full Text

Full Text

RESEARCH Open AccessThe genetics of smoking in individuals withchronic obstructive pulmonary diseaseMa’en Obeidat1* , Guohai Zhou1, Xuan Li1, Nadia N. Hansel2, Nicholas Rafaels3, Rasika Mathias4, Ingo Ruczinski5,Terri H. Beaty6, Kathleen C. Barnes3, Peter D. Paré1,7 and Don D. Sin1,7AbstractBackground: Smoking is the principal modifiable environmental risk factor for chronic obstructive pulmonarydisease (COPD) which affects 300 million people and is the 3rd leading cause of death worldwide. Most of thegenetic studies of smoking have relied on self-reported smoking status which is vulnerable to reporting andrecall bias. Using data from the Lung Health Study (LHS), we sought to identify genetic variants associated withquantitative smoking and cessation in individuals with mild to moderate COPD.Methods: The LHS is a longitudinal multicenter study of mild-to-moderate COPD subjects who were all smokersat recruitment. We performed genome-wide association studies (GWASs) for salivary cotinine (n = 4024), exhaledcarbon monoxide (eCO) (n = 2854), cigarettes per day (CPD) (n = 2706) and smoking cessation at year 5 follow-up (n =717 quitters and 2175 smokers). The GWAS analyses were adjusted for age, gender, and genetic principal components.Results: For cotinine levels, SNPs near UGT2B10 gene achieved genome-wide significance (i.e. P < 5 × 10− 8) withtop SNP rs10023464, P = 1.27 × 10− 11. For eCO levels, one significant SNP was identified which mapped to theCHRNA3 gene (rs12914385, P = 2.38 × 10− 8). A borderline region mapping to KCNMA1 gene was associated withsmoking cessation (rs207675, P = 5.95 × 10− 8). Of the identified loci, only the CHRNA3/5 locus showed significantassociations with lung function but only in heavy smokers. No regions met genome-wide significance for CPD.Conclusion: The study demonstrates that using objective measures of smoking such as eCO and/or salivarycotinine can more precisely capture the genetic contribution to multiple aspects of smoking behaviour. TheKCNMA1 gene association with smoking cessation may represent a potential therapeutic target and warrantsfurther studies.Trial registration: The Lung Health Study ClinicalTrials.gov Identifier: NCT00000568. Date of registration:October 28, 1999.Keywords: Cessation, Smoking, GWAS, eCO, CotinineBackgroundThe smoking epidemic is one of the biggest public healthchallenges in modern history [1]. Tobacco-attributabledeaths are expected to rise to more than 10 million globallyby 2030 [2–4]. Despite aggressive public health programsaimed at eliminating smoking in the United States (US)and elsewhere, one in 4 adults in the US still use tobaccoproducts and 1 in 5 are daily users. Smoking is the princi-pal modifiable environmental risk factor for chronicobstructive pulmonary disease (COPD), ischemic heart dis-ease, and lung cancer [5]. COPD, for instance, affects 300million people and is the 3rd leading cause of death world-wide [6].Genetic studies of smoking behaviour and smoking-related illnesses such as COPD and lung cancer haveidentified strong associations in the chromosome 15q25region, which contains genes encoding the nicotinic re-ceptor subunits CHRNA3-CHRNA5-CHRNB4 [7–9].However, most of these studies have relied on self-reported smoking status as either the phenotype of inter-est or one of the covariates to “adjust” tobacco exposure.Self-report is vulnerable to reporting and recall bias and* Correspondence: maen.obeidat@hli.ubc.ca1The University of British Columbia Center for Heart Lung Innovation, StPaul’s Hospital, Vancouver, BC, CanadaFull list of author information is available at the end of the article© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.Obeidat et al. Respiratory Research  (2018) 19:59 https://doi.org/10.1186/s12931-018-0762-7has been shown to consistently underestimate totaltobacco exposure [10, 11], which may result in residualconfounding [12]. One way to mitigate this risk is to usean objective biochemical assay to validate self-reports oftobacco use. The most commonly used assays includeexhaled carbon monoxide (eCO) [13] or cotinine inserum, urine or saliva [14]. Because eCO has a shorthalf-life (~ 4 h) it is best suited for short term tobaccoexposure while the longer half-life of cotinine (~ 16 h)makes it a more robust measure to differentiate activefrom non-active smokers with a longer duration of ab-stinence. However, because cotinine is a by-product ofnicotine, enzymatic processes involved in nicotine me-tabolism can affect cotinine levels in smokers. eCOand/or cotinine assays are relatively inexpensive, non-invasive and well-standardized measurements andmost importantly they more accurately quantify to-bacco exposure in smokers compared with self-reportalone [13, 15].To date, very few studies have ascertained the geneticdrivers of smoking (and cessation) using validated biochem-ical assays [16], especially among those with establishedsmoking related diseases such as COPD. Using data fromthe Lung Health Study, we identified the genetic variantsassociated with cigarette smoking using validated objectiveassays in individuals with mild to moderate COPD.MethodsThe lung health study (LHS)The details of the LHS have been published previously[17, 18]. Briefly, LHS was a multicenter clinical studythat evaluated the effects of ipratropium bromide, ashort acting antimuscarinic agent (i.e. ipratropiumbromide), and smoking cessation on lung function de-cline in current smokers with mild to moderate COPD.At the time of recruitment all subjects were activesmokers between the ages of 35 and 60 years (with amean age of 48 years) who had smoked at least 10 ciga-rettes a day within the 30 days prior to initial screeningand who demonstrated mild to moderate COPD on spir-ometry defined by forced expiratory volume in 1 second(FEV1) between 55% and 90% of predicted, in the pres-ence of a FEV1/forced vital capacity (FVC) ratio of < 0.70 after bronchodilation. The mean FEV1 of the cohortwas 75% predicted and the mean FEV1/FVC was 63%post-bronchodilator.After enrolment, these patients were randomlyassigned to one of 3 groups: (1) usual care (UC), whoreceived no intervention, n = 1964; (2) an intense anti-smoking (special) intervention and ipratropium bromide(Atrovent®, Boehringer Ingelheim Pharmaceuticals) n =1961 (SIA); or (3) an intense anti-smoking (special)intervention and an inhaled placebo, n = 1962 (SIP). Tencenters participated in the original study and togetherthey recruited 5887 patients (of whom 37% werefemales). Those who were in the SIP or SIA groups re-ceived a program that consisted of: 1) a strong recom-mendation by attending physician for smoking cessationin an one-on-one encounter; 2) a group program led bya health educator that met 12 times over 10 weeks,which taught behavioural modification techniques; and3) nicotine replacement therapy with nicotine gum(2 mg per piece, Nicorette Gum, Marion Merrell DowInc), which was provided at no cost to patients. Thosewho successfully quit smoking were enrolled in a main-tenance program to prevent relapses.For the first 5 years, the lung function of participantswas measured annually. At each face to face visit thesubjects’ smoking status was determined using a ques-tionnaire, which was validated by salivary cotinine andexhaled carbon monoxide levels as previously described[18]. Participants were classified as smokers if their co-tinine levels were greater than 20 ng/mL or if their ex-haled carbon monoxide concentrations were higher than10 ppm. At year 5 of the study, participants were dividedinto three groups based on smoking history as previ-ously described [19]. Sustained quitters (SQs) weredefined as those who gave a history of total abstinence(no month in which the subject smoked even a singlecigarette per day) and had eCO readings below 10 ppmat each annual follow-up visit over 5 years. Continuingsmokers (CSs) were those who reported smoking at allscheduled follow-up visits. Intermittent quitters (IQs)were current smokers at some but not all of their visits.Given the ambiguity of the IQ group (n = 1210) in termsof cigarette smoking, they were excluded from the smok-ing cessation genome wide association study (GWAS).GenotypingAt year 5 of LHS, venipuncture was carried out on 5413LHS participants who were alive and eligible at this visit.Blood samples were taken when participants were stableand free of exacerbations for at least 4 weeks and were sep-arated into buffy coat and serum [20]. DNA was extractedfrom the buffy coat samples of 4251 European Americansin LHS and SNP genotyping was performed. The details ofgenotyping and quality control have been previously de-scribed [21]. Briefly, samples were genotyped using theIllumina Human660WQuad v.1_A BeadChip. Overall, 98.4% of samples (n = 4181) passed initial quality control stan-dards and genotypes were available for 559,766 SNPs. Anadditional 133 samples were removed because they failedsubsequent quality control, which resulted in a final sampleof n = 4048 for the present analysis. Imputation was under-taken with the Michigan Imputation Server [22] using theHaplotype Reference Consortium (HRC) [23] panel. Vari-ants were excluded if the imputation r2 was < 0.5 and if theminor allele frequency was < 1%.Obeidat et al. Respiratory Research  (2018) 19:59 Page 2 of 10Measurements of expired carbon monoxide (eCO) andcotinineThe details of eCO and cotinine measures have beenpreviously described [24]. To conduct the cotinine assay,LHS participants were asked to deposit at least 1 ml ofsaliva in a plastic vial, which was then frozen and sent ina batch to the American Health Foundation laboratoryin Valhalla, NY. One sample was taken from each par-ticipant and a single cotinine assay measurement wasperformed. The cotinine assessment was conductedusing the radioimmunoassay technique of Langone et al.by personnel who were blinded to the smoking status ofthe participants [24]. In LHS, the sensitivity and specifi-city of using a salivary cotinine cutoff of 20 ng/ml com-pared to self-report was 99% and 92%, respectively.Studies have reported a technical coefficient of variation(CV) value of 5% for salivary cotinine assay [25].Carbon monoxide in expired air was measured usingeither of two instruments: the MiniCO (CatalystResearch) or the EC50 (Vitalograph). The eCO measure-ment procedure involved two attempts. If the two valueswere not within 4 ppm, the measurements were re-peated. The result was the average of the two readingsrounded to the nearest integer [24].Genome-wide association analyses (GWASs)Given that eCO, cotinine and cigarettes per day (CPD)distributions were skewed, the values were transformedinto normally distributed Z scores using the R function‘rntransform’ [26] (Additional file 1: Figure S1). We per-formed GWAS for three transformed phenotypes: eCO,salivary cotinine and CPD using SNPTEST [27] assumingan additive genetic model and adjusting for age, gender,and the first 5 genetic principal components (PCs). Sincecotinine levels were measured in all subjects at baselinevisit, we performed the GWAS for cotinine baseline levelsto make available the largest sample size (n = 4024 andmissing data rate = 1.9%). The eCO and CPD values wereonly measured in smokers at subsequent visits so we usedthe values measured at year 1 follow up to make availablethe largest sample size (n = 2706 and missing data rate = 8.1% for eCO; n = 2854 and missing data rate = 3.12% forCPD). Significant SNPs were defined as the sentinel SNPsmeeting genome-wide significance (P < 5 × 10− 8).Evaluation of previously published variants which relateto smoking cessationWe evaluated previously published hits from two studiesfor smoking cessation. The first study is the Tobaccoand Genetic Consortium (TAG) meta-analysis GWAS ofsmoking behaviour [28]. The TAG study included 41,278former and current smokers and the top 15 associatedSNPs for self-reported smoking cessation were followedup in 64,924 independent individuals from the EuropeanNetwork of Genetic and Genomic Epidemiology (EN-GAGE) and the Oxford-GlaxoSmithKline (Ox-GSK)consortia. This meta-analysis identified only one signifi-cant SNP as being associated with smoking cessation;rs3025343 (P = 1.8 × 10− 8), which was located onchromosome 9, near the dopamine beta-hydroxylase(DBH) gene.The second study by Siedlinski et al. reported GWASresults for self reported phenotypes of smoking includ-ing lifetime average and current CPD, age at smokinginitiation, and smoking cessation in 3441 patients withCOPD [29]. In total, the meta-analysis included 1164current smokers and 1907 former smokers (all using selfreport of yes/no answers); none of the SNPs had showeda statistically significant association with smoking cessa-tion. In the present study, we interrogated the 9 SNPsassociated with smoking cessation reported in Siedlinskiet al. study.Association of smoking related SNPs with extremes oflung functionTo determine whether any of the smoking related SNPsdiscovered in the present study also had an impact onsmoking-related physiological outcomes such as lungfunction, we tested SNPs identified in this study forassociation with lung function in never and, separately,in heavy smokers. The UK Biobank Lung Exome VariantEvaluation (UK BiLEVE) evaluated the genetic determi-nants related to low (mean of 65.6% predicted, average(mean of 90.6% predicted), or high (mean of 118% pre-dicted) forced expiratory volume in 1 second (FEV1) inheavy smokers (mean 35 pack-years) and separately innever smokers. The study included 10,002 individualswith low FEV1, 10,000 with average FEV1, and 5002 withhigh FEV1 from each of the heavy smoker and neversmoker groups. Genome-wide genotyping was performedusing a custom Affymetrix Axiom array (UK BiLEVEarray; Santa Clara, CA, USA). After quality control non-genotyped variants were imputed using a combined1000G Phase 1 and UK10K Project [30] reference panel.Gene drug interactionsTo uncover the potential biological relevance of thesmoking cessation GWAS hits, we used two databasesto search for potential gene drug interactions: the DGIdb[31] http://www.dgidb.org/ and the DRUGBANK data-base [32] https://www.drugbank.ca/ .ResultsDescriptive demographics of LHS participantsThe overall LHS design and the sample size for each ofthe GWASs are shown in Fig. 1, and the demographicsand quantitative smoking values are shown in Table 1.Among smokers at year 1 of LHS, there was a strongObeidat et al. Respiratory Research  (2018) 19:59 Page 3 of 10correlation between eCO, cotinine, and CPD values withthe correlation between eCO and CPD being the stron-gest (r = 0.5, P = 1 × 10− 170, Additional file 1: Figure S2).The levels of quantitative smoking biomarkers such ascotinine were inversely related to cross sectional lung func-tion measures at years 1–5 (Additional file 1: Figure S3).This relationship was significant (P < 0.05) for years 1–4and borderline significant in year 5 (P = 0.08). Furthermore,cotinine levels at year 5 showed negative correlation withFEV1 decline between years 1 and 5 (P = 3 × 10− 11,Additional file 1: Figure S4).Genome-wide association results for cotinine levelsIn the four GWAS analyses, we included 7,807,992 vari-ants with MAF > 1% and imputation quality> 0.5. TheGWAS of cotinine levels at baseline included 4024 indi-viduals. Quantile–quantile (QQ) plots are presented inAdditional file 1: Figure S5, which showed a sharp devi-ation from the expected distribution for low p-values in-dicating strong signals. The genomic inflation factor (λ)was 0.996, suggesting no systematic deviation in the as-sociation statistics due to factors such as populationstructure.A total of 250 SNPs in the 4q13.2 region containingthe UGT2B10 gene achieved genome-wide significance(P < 5 × 10− 8). A Manhattan plot is shown in Fig. 2 andthe region plots are shown in Fig. 3. The mostsignificantly associated SNP for cotinine levels was anintergenic SNP on chromosome 4 (rs10023464) (P = 1.27 × 10-11). Another interesting signal was observed onthe 15q25.1 region with an intronic SNP rs9788721 (P =3.49 × 10-7) in the HYKK gene near CHRNA3/5 genes,though this latter association did not reach genome-wide significance. All SNPs meeting or approachinggenome-wide significance for smoking phenotypes arepresented in Table 2.The association of the two cotinine-associated loci(UGT2B10 and CHRNA3/5) with cotinine levels at sub-sequent years showed that the strength of the associ-ation decays during subsequent follow up visits i.e. the Pvalues increase yet they maintain the same direction ofeffect (Additional file 1: Table S1). This decrease in sig-nificance is directly proportional to the the decrease insample sizes available for the analysis at follow up visitswith a missing rate ranging from 8% in year 1 to 51%missing subjects at year 5 (Additional file 1: Table S2).Genome-wide association results for exhaled carbonmonoxide (eCO) levelsThe GWAS of eCO levels at year 1 included 2706smokers (Fig. 2 and Additional file 1: Figure S5). Thegenomic inflation factor (λ) was 1.00. Only one region;the 15q25.1 region met genome wide significance withone SNP; intronic SNP (rs12914385) mapping to theCHRNA3 gene (P = 2.38 × 10-8) (Fig. 3 region plots).Two regions approached genome wide significance foreCO. The first was the 3q22.3 region with intronic SNPrs546764 (P = 7.76 × 10-8) mapping to the CEP70 gene.Fig. 1 Overall LHS smoking GWAS study design. eCO: exhaled carbon monoxide. CPD: Cigarettes per day. Y1: year 1. eCO was measured in thosereporting current smokingTable 1 Demographics of study subjects. Gender male in n (% of column totals), other variables in mean±SD. CPD: cigarettes perday. eCO exhaled carbon monoxide. * Age at assessmentBaseline Year 1 Year 5Smokers Smokers Quitters Continuous smokers Sustained quitters Intermittent quittersN (%) 4102 2946 1156 2175 717 1210Age* (years) 48.6 ± 6.7 49.4 ± 7 50.0 ± 6.7 53.3 ± 6.7 54.2 ± 6.6 53.7 ± 6.7Gender (male) 2853 (63%) 1837 (62%) 746 (65%) 1366 (63%) 481 (67%) 736 (61%)BMI (kg/m2) 25.5 ± 3.8 25.7 ± 3.9 27.3 ± 4.0 26.0 ± 4.2 28.4 ± 4.3 27.5 ± 4.2FEV1% predicted 78.6 ± 9.0 78.0 ± 9.8 81.8 ± 9.5 72.8 ± 12.0 80.3 ± 10.9 77.2 ± 11.5eCO (ppm) 32.4 ± 16.1 24.8 ± 13.2 4.8 ± 2.5 25.4 ± 13.8 4.0 ± 2.4 9.6 ± 11.1Cotinine (ng/ml) 361.4 ± 199.3 302.7 ± 146.3 92.2 ± 158.7 343.1 ± 195.3 27.8 ± 132.4 117.5 ± 229.6CPD cigarettes/day 21.9 ± 14.5 21.9 ± 14.5 0 ± 0 23.0 ± 12.9 0 ± 0 5.3 ± 10.2Obeidat et al. Respiratory Research  (2018) 19:59 Page 4 of 10The second region was the 14q23.1 region with anintergenic SNP rs140706189 (P = 6.31 × 10-8) near theSIX1/4/6 and MNAT1 genes (Table 2).Genome-wide association results for cigarettes per day(CPD)The GWAS for CPD included a total of 2854 smokers atyear 1 (Fig. 2 and Additional file 1: Figure S5). The gen-omic inflation factor (λ) was 1.00. No regions metgenome-wide significance. Two regions, however,approached this level of significance. These included the13q21.32 region with intronic SNP rs9599114 mappingto PCDH9 gene (P = 7.94 × 10-8), and the 9q34.13 regionwith intergenic SNP rs1412076 (P = 2.24 × 10-8) nearNTNG2 (Table 2).Genome-wide association results for smoking cessationAt year 5 of LHS, there were 717 sustained quitters and2175 continuous smokers (quitters vs. smokers case con-trol GWAS). No loci met genome-wide significance buttwo loci very closely approached this level of significance(Figs. 2 and 3). The strongest association was intronicSNP rs207675, which mapped to the KCNMA1 gene on10q22.3 (P = 5.95 × 10-8). The second loci included theintronic SNP rs212420 (P = 1.50 × 10-8) near theATXN7L1 gene on 7q22.3.Evaluation of previously associated SNPs for smokingbehavioursPrevious reports have identified genetic loci (CHRNB3/A6 region on chromosome 8 and the CYP2A6 region onchromosome 19) that were significantly associated withCPD [33] . In the present study, both of these SNPsshowed no significant association with CPD (P > 0.05).Additionally, we evaluated 15 SNPs that were nominallyrelated to smoking cessation in two previous publica-tions [28] and a number of other SNPs that were nomin-ally associated with smoking cessation among COPDsubjects [29].Additional file 1: Table S3 shows the results of thelook-up. One SNP; rs4362358 near the CHRNA3/5genes that was related to cessation in the TAG consor-tium [28] was associated with eCO (P = 0.03) and cotin-ine (P = 0.002) in our study. Another cessation SNP inthe TAG consortium; rs17178639 in SLC25A21 genewas associated with eCO (P = 0.007) in the presentstudy. The two SNPs were associated with reduced ces-sation in the TAG consortium and were also associatedwith higher eCO and cotinine levels in our study. Of thecessation SNPs in COPD patients from the study ofSiedlinski et al. [29], two near the IPMKP1 gene werenominally associated with cessation in our study:rs9506942, and rs9552733 with P = 0.005 and P = 0.004,respectively and with the same direction of effect.We tested SNPs identified in our study for associationswith smoking phenotypes in the TAG consortium (CPD,cessation, age of onset and ever vs. never phenotypes).Only CHRNA3/5 SNPs (for eCO and cotinine) were sig-nificant in the TAG data for both CPD and for cessation(Additional file 1: Table S4).Associations with lung functionWe tested the SNPs identified in this study for associationswith extremes of lung function: high vs. low FEV1 in heavysmokers and separately for never smokers in a large studyfrom the UK Biobank [34]. Only the CHRNA3/5 regionvariants showed associations with lung function and onlyin the heavy smokers group (not in the never smokers).The results for lung function are shown in Additional file 1:Table S5.Fig. 2 Manhattan plots of smoking GWASs in Lung Health Study. The plots show the P values (−log10 scale) on the Y axes and the SNP positionsacross 22 autosomal chromosomes on the X axes. The horizontal red line represents the genome-wide cut-off of 5 × 10− 08Obeidat et al. Respiratory Research  (2018) 19:59 Page 5 of 10Fig. 3 Region plots of the smoking associated loci. The Y axis represent the P values in the (−log10 scale) and the X axis is the genomic position.Gene names and their corresponding coordinates are shown below. The sentinel SNP is shown as a purple diamond and the color coding ofSNPs reflects the degree of linkage disequilibrium (LD) with the sentinel SNP using 1000G referenceObeidat et al. Respiratory Research  (2018) 19:59 Page 6 of 10DiscussionSmoking places a huge burden on individuals and healthcare systems. Smoking behaviours, and consequently therisk of smoking related illnesses are at least partially gen-etically determined [35, 36]. The majority of previousstudies on the genetics of smoking behaviour have reliedon self-report, which is affected by recall bias and moreimportantly under-reporting bias, which may lead to aninaccurate assessment of smoking exposure. Indeed, sev-eral groups have shown that in approximately 25 to 50%of self-reported quitters, objective assays could not valid-ate the reported smoking status [37] . It is thus crucialto accurately phenotype smoking status to properlyunderstand the molecular mechanisms underlying nico-tine addiction, metabolism, and smoking cessation.In the current study of COPD subjects, we performedGWAS for the phenotypes of self-reported cigarettes perday (CPD) and two biochemical biomarkers of smoking:eCO and salivary cotinine. Additionally, we evaluatedgenetic determinants of biochemically validated smokingcessation over 5 years. Using this approach, we identifiedgenome-wide significant loci associated with salivarycotinine on 15q.25.1 (CHRNA3/5 genes) and 4q13.2(UTG2B10) gene). For eCO, only the 15q.25.1 locusreached genome-wide significance. The smoking cessa-tion GWAS revealed a borderline signal in theKCNMA1 gene on 10q22.3. Finally, of all the loci identi-fied in the current study, only the CHRNA3/5 locusshowed a significant association with lung function inheavy smokers (but not in never smokers) from the gen-eral population.Genetic determinants of smoking can be related tosmoking intensity (addiction), metabolism or both. Themetabolism of nicotine involves several enzymatic path-ways. Approximately 10% of nicotine is excreted un-changed in the urine. The majority of nicotine (~ 80%) isconverted to cotinine in two steps: initial metabolism,which is mediated by the cytochrome P450, family 2, sub-family A, poly-peptide 6 (CYP2A6) enzyme, followed byconjugation by aldehyde oxidase [38]. After these twosteps cotinine is further metabolised by CYP2A6 to 3-hydroxycotinine. Oxidation and glucuronidation processesaccount for the remaining 10% of the metabolic process[38]. We found significant association of salivary cotininewith SNPs in UDP glucuronosyltransferase family 2 mem-ber B10 (UGT2B10), which catalyses both nicotine andcotinine glucuronidation in smokers.Previous GWAS for cotinine in urine, plasma orserum have all identified the UGT2B10 region [16, 39]with stronger associations reported for urinary cotininelevels. In the current study, we replicated the associationsignal for SNPs in the UGT2B10 region for associationwith salivary cotinine. Smokers are thought to self-titratetheir nicotine to meet their physiological need [40] (i.e.high metabolizers are likely to smoke more). If this hy-pothesis were true, then we would expect variations inthe rates of metabolism (and hence consumption) to beassociated with smoking related diseases/phenotypes.However, we failed to find any significant associations be-tween SNPs in the UGT2B10 gene region with impairedlung function in smokers. We may not have sufficientpower to detect a subtle effects of this locus; alternativelythe two mechanisms (metabolism and consumption) maynot be directly linked as previously suggested [34]. Theassociation results of the cotinine and eCO-associatedSNPs with CPD in the same individuals are shown inAdditional file 1: Table S6. The results indeed show thatthe UGT2B10 variant is not associated with CPD (P = 0.28),arguing against the notion that variation in metabolism ofnicotine may affects smoking behaviour.Perhaps the most widely studied and reported region forsmoking is CHRNA3/5 on 15q25. The associations of thisregion in the current study are with biochemical bio-markers of smoking (eCO and cotinine). In the presentTable 2 Genetic loci associated with smoking behaviour in the Lung Health StudyPhenotype SNP Chr Gene(s) Position (hg19) Alleles (REF/ALT) MAF Imputation r2 beta SE P value % VarianceCotinine rs10023464 4 UGT2B10 69,659,738 C/T 9.8% 0.995 0.25 0.04 1.27e-11 1.1%Cotinine rs9788721 15 CHRNA3/5 78,802,869 C/T 39.5% 0.989 −0.12 0.02 3.49e-7 0.6%eCO rs546764 3 CEP70 138,294,336 T/G 2.8% 0.844 0.39 0.07 7.76e-8 0.9%eCO rs140706189 14 SIX1/4/6, MNAT1 61,151,425 T/G 1.7% 0.827 −0.49 0.09 6.31e-8 0.9%eCO rs12914385 15 CHRNA3/5 78,898,723 C/T 43% 1.000 0.12 0.02 2.38e-8 1.2%CPD rs1412076 9 NTNG2/ SETX 135,032,890 A/G 37.6% 0.999 0.10 0.02 2.24e-7 0.8%CPD rs9599114 13 PCDH9 66,987,131 T/C 41.7% 0.957 0.11 0.02 7.94e-8 0.8%Cessation rs212420 7 ATXN7L1/ CDHR3 105,496,412 C/G 6.7% 0.992 −0.60 0.11 1.50e-7 n/aCessation rs207675 10 KCNMA1 79,154,537 T/C 33.6% 0..988 −0.35 0.06 5.95e-8 n/aUGT2B10 UDP glucuronosyltransferase family 2 member B10, CHRNA3 cholinergic receptor nicotinic alpha 3 subunit, CHRNA5 cholinergic receptor nicotinic alpha 5subunit, CEP70 centrosomal protein 70, SIX1 SIX homeobox 1, MNAT1 MNAT1, CDK activating kinase assembly factor, NTNG2 Nitrin G2, SETX senataxin, PCDH9protocadherin 9, ATXN7L1 ataxin 7 like 1, CDHR3 cadherin related family member 3, KCNMA1 potassium calcium-activated channel subfamily M alpha 1, * Refers toodds ratio and 95% confidence intervals (CI)Obeidat et al. Respiratory Research  (2018) 19:59 Page 7 of 10study, the CHRNA3/5 variants were associated with saliv-ary cotinine as well as eCO levels. Importantly, these vari-ants were also significantly related to CPD (P < 0.05),suggesting this genetic region modulates cigarette con-sumption (Additional file 1: Table S6). However, this locuswas not associated with smoking cessation in LHS partici-pants, suggesting that other factors are involved in quit-ting. Other previously reported loci in CYP2A6 andCHRNB3/CHRNA6 genes could not be replicated in ourstudy. However, as previously noted [35, 36], the strengthof the relationship between these loci and smoking is rela-tively modest and may require much larger sample sizesto be detected.We identified a suggestive signal (P = 5.95 × 10-8) forsmoking cessation in the potassium calcium-activatedchannel subfamily M alpha 1 (KCNMA1) gene which isimportant for the control of smooth muscle tone andneuronal excitability [41, 42]. The association betweenthe KCNMA1 variant with cessation in our study couldnot be replicated in the cessation GWAS from the TAGconsortium [28]. This could be due to the fact that ourstudy used a biochemically verified smoking status;whereas the previous studied relied only on self-report.Our data are in keeping with a previously publishedstudy. A GWAS in Australian and Dutch populationsidentified SNPs in the KCNMA1 gene, which weresignificantly associated with nicotine dependence(rs592676, p = 8.91 × 10-6). In our study, the same SNPwas strongly associated with cessation (P = 5.6 × 10-7).Interestingly, a drug repositioning study that integrateddisease and drug expression profiles identified KCNMA1as a potential molecular target for lobeline: a naturalalkaloid that has been used as a smoking cessation aid[43] as well as for amphetamine and cocaine addictions[44] KCNMA1 is a target for the FDA approved drug,chlorzoxazone, which is a centrally acting musclerelaxant. Chlorzoxazone acts as an activator of acalcium-activated potassium channel [45] and iscommonly used as a probe drug to phenotype CYP2E1activity and its metabolism is strongly accelerated bycigarette smoking [46]. Another study proposedchlorzoxazone as a potential treatment for alcoholaddiction [47]. At the gene expression level, NHBE cellsexposed to nicotine-containing e-cigarette vapour demon-strate decreased expression of KCNMA1 [48], while in hu-man lung tissue smokers have significantly higherexpression compared to never smokers (1.5 fold change,P = 3.11 × 10− 08) [49]. Finally, a genome-wide study iden-tified differential hydroxymethylation of potassium chan-nel genes, including KCNMA1, in the nucleus accumbensin methamphetamine addiction and abstinence [50].Taken together, these data suggest the KCNMA1 associ-ation with smoking cessation is biologically plausible withthe potential for drug repurposing.This study has several limitations. The sample sizemay have been too small to detect novel loci for smok-ing biomarkers or cessation. On the other hand, the useof precise biochemical phenotypes on the other handlikely improved the specificity of the smoking cessationsignal. Furthermore, and in line with most publishedGWASs, the proportion of variance explained by theidentified variants is small.ConclusionIn conclusion, we identified genetic loci associated witheCO and cotinine in COPD patients. Our study stronglysupport the need to use objective measures of smokingto capture the genetic contribution to smoking in thesestudies. The KCNMA1 region association with smokingcessation represents a potential target for drug discoveryand repurposing which warrants further studies.Additional fileAdditional file 1: Supplementary Figures and Tables. (DOCX 1183 kb)AcknowledgementsM. Obeidat is a Fellow of the Parker B. Francis Foundation. He is also a recipientof British Columbia Lung Association Research Grant. D. Sin holds a Tier 1Canada Research Chair in COPD.FundingThis study was funded by the Canadian Respiratory Research Network (CRRN)and Genome Canada: Genome British Columbia. CRRN is supported by grantsfrom the Canadian Institutes of Health Research (CIHR) – Institute of Circulatoryand Respiratory Health, Canadian Lung Association/Canadian Thoracic Society,British Columbia Lung Association, and Industry Partners Boehringer IngelheimCanada Ltd., AstraZeneca Canada Inc. and Novartis Canada Ltd. Funding fortraining of post-doctoral students and new investigators within the networkwas supported by the above funding sponsors and as well by GlaxoSmithKlineInc. The funders had no role in the study design, data collection and analysis orpreparation of the manuscript.Availability of data and materialsThe genotype data for the Lung Health Study is available on National Centerfor Biotechnology Information (NCBI) database of genotypes and phenotypes(dbGaP) under phs000335.v3.p2.Authors’ contributionsConceived and designed the study: MO, PDP, DDS. Lung Health Study dataanalysis and genotyping: NNH, NR, RM, IR, THB, KCB. Statistical analyses: GZ,MO, XL. Wrote the manuscript: MO, PDP, DDS. Discussed results and implicationsand commented on the manuscript at all stages: all co-authors. All authors readand approved the final manuscript.Ethics approval and consent to participateWork undertaken in this manuscript was approved by the University ofBritish Columbia Institutional Review Board, certificate numberH16–01201.Consent for publicationNot applicable.Competing interestsDr. Sin has received research funding from AstraZeneca (AZ), Merck, BoehringerIngelheim (BI) and Novartis and has received honorarium for speakingengagements from AZ, Novartis, Regeneron and Sanofi-Aventis.Obeidat et al. Respiratory Research  (2018) 19:59 Page 8 of 10Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.Author details1The University of British Columbia Center for Heart Lung Innovation, StPaul’s Hospital, Vancouver, BC, Canada. 2Pulmonary and Critical CareMedicine, School of Medicine, Johns Hopkins University, Baltimore, MD, USA.3Division of Biomedical Informatics and Personalized Medicine, Departmentof Medicine, University of Colorado School of Medicine, Anschutz MedicalCampus, Aurora, CO, USA. 4Division of Genetic Epidemiology, School ofMedicine, Johns Hopkins University, Baltimore, MD, USA. 5Department ofBiostatistics, Bloomberg School of Public Health, Johns Hopkins University,Baltimore, MD, USA. 6Department of Epidemiology, Bloomberg School ofPublic Health, Johns Hopkins University, Baltimore, MD, USA. 7RespiratoryDivision, Department of Medicine, University of British Columbia, Vancouver,BC, Canada.Received: 16 January 2018 Accepted: 27 March 2018References1. WHO. Tobacco fact sheet: World Health Organization; 2015. http://www.wpro.who.int/mediacentre/factsheets/fs_201203_tobacco/en/.2. Mortality from smoking in developed countries 1950–2020[http://www.ctsu.ox.ac.uk/~tobacco].3. Jha P, Peto R. Global effects of smoking, of quitting, and of taxing tobacco.N Engl J Med. 2014;370:60–8.4. Jha P. Avoidable global cancer deaths and total deaths from smoking. NatRev Cancer. 2009;9:655–64.5. Doll R, Peto R. Mortality in relation to smoking: 20 years’ observations onmale British doctors. Br Med J. 1976;2:1525–36.6. Lozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, Abraham J,Adair T, Aggarwal R, Ahn SY, et al. Global and regional mortality from 235causes of death for 20 age groups in 1990 and 2010: a systematic analysisfor the global burden of disease study. Lancet. 2010;380:2095–128.7. Thorgeirsson TE. A variant associated with nicotine dependence, lungcancer and peripheral arterial disease. Nature. 2008;452:638–42.8. Obeidat ME, Hao K, Bossé Y, Nickle DC, Nie Y, Postma DS, Laviolette M,Sandford AJ, Daley DD, Hogg JC, et al. Molecular mechanisms underlyingvariations in lung function: a systems genetics analysis. Lancet Respir Med.2015;3:782–95.9. Hobbs BD, de Jong K, Shrine NRG, Wyss AB, London SJ, Cho M. A genome-wide association study of chronic obstructive pulmonary disease identifies22 risk loci. Nat Genet. 2017;49:426–32.10. Stelmach R, Fernandes FLA, Carvalho-Pinto RM, Athanazio RA, Rached SZ,Prado GF, Cukier A. Comparison between objective measures of smokingand self-reported smoking status in patients with asthma or COPD: are ourpatients telling us the truth? J Bras Pneumol. 2015;41:124–32.11. Fendrich M, Mackesy-Amiti ME, Johnson TP, Hubbell A, Wislar JS. Tobacco-reporting validity in an epidemiological drug-use survey. Addict Behav.2005;30:175–81.12. Obeidat ME, Ding X, Fishbane N, Hollander Z, Ng RT, McManus B, TebbuttSJ, Miller BE, Rennard S, Paré PD, Sin DD. The effect of different casedefinitions of current smoking on the discovery of smoking-related bloodgene expression signatures in chronic obstructive pulmonary disease.Nicotine Tob Res. 2016;18:1903–9.13. Jarvis MJ, Russell MA, Saloojee Y. Expired air carbon monoxide: a simplebreath test of tobacco smoke intake. Br Med J. 1980;281:484–5.14. Jatlow P, Toll BA, Leary V, Krishnan-Sarin S, O’Malley SS. Comparison ofexpired carbon monoxide and plasma cotinine as markers of cigaretteabstinence. Drug Alcohol Depend. 2008;98:203–9.15. Sato S, Nishimura K, Koyama H, Tsukino M, Oga T, Hajiro T, Mishima M.Optimal cutoff level of breath carbon monoxide for assessing smokingstatus in patients with asthma and COPD. Chest. 2003;124:1749–54.16. Ware JJ, Chen X, Vink J, Loukola A, Minica C, Pool R, Milaneschi Y, ManginoM, Menni C, Chen J, et al. Genome-wide meta-analysis of cotinine levels incigarette smokers identifies locus at 4q13.2. Sci Rep. 2016;6:20092.17. Kanner RE, Connett JE, Williams DE, Buist AS. Effects of randomizedassignment to a smoking cessation intervention and changes in smokinghabits on respiratory symptoms in smokers with early chronic obstructivepulmonary disease: the lung health study. Am J Med. 1999;106:410–6.18. Anthonisen NR, Connett JE, Kiley JP, Altose MD, Bailey WC, Buist AS, ConwayWA Jr, Enright PL, Kanner RE, O'Hara P, et al. Effects of smoking interventionand the use of an inhaled anticholinergic bronchodilator on the rate of declineof FEV1. The lung health study. JAMA. 1994;272:1497–505.19. Nicholas RA, John EC, Robert PM. Smoking and lung function of lung healthstudy participants after 11 years. Am J Respir Crit Care Med. 2002;166:675–9.20. Man SF, Connett JE, Anthonisen NR, Wise RA, Tashkin DP, Sin DD. C-reactiveprotein and mortality in mild to moderate chronic obstructive pulmonarydisease. Thorax. 2006;61:849–53.21. Hansel NN, Ruczinski I, Rafaels N, Sin DD, Daley D, Malinina A, Huang L,Sandford A, Murray T, Kim Y, et al. Genome-wide study identifies two lociassociated with lung function decline in mild to moderate COPD. HumGenet. 2013;132:79–90.22. Das S, Forer L, Schonherr S, Sidore C, Locke AE, Kwong A, Vrieze SI, ChewEY, Levy S, McGue M, et al. Next-generation genotype imputation serviceand methods. Nat Genet. 2016;48:1284–7.23. McCarthy S, Das S, Kretzschmar W, Delaneau O, Wood AR, Teumer A, KangHM, Fuchsberger C, Danecek P, Sharp K, et al. A reference panel of 64,976haplotypes for genotype imputation. Nat Genet. 2016;48:1279–83.24. Murray RP, Connett JE, Lauger GG, Voelker HT. Error in smoking measures:effects of intervention on relations of cotinine and carbon monoxide toself-reported smoking. The lung health study research group. Am J PublicHealth. 1993;83:1251–7.25. Etter JF, Vu Duc T, Perneger TV. Saliva cotinine levels in smokers andnonsmokers. Am J Epidemiol. 2000;151:251–8.26. Aulchenko YS, Ripke S, Isaacs A, van Duijn CM. GenABEL: an R library forgenome-wide association analysis. Bioinformatics. 2007;23:1294–6.27. Marchini J, Howie B, Myers S, McVean G, Donnelly P. A new multipointmethod for genome-wide association studies by imputation of genotypes.Nat Genet. 2007;39:906–13.28. Consortium” TaG. Genome-wide meta-analyses identify multiple lociassociated with smoking behavior. Nat Genet. 2010;42:441–7.29. Siedlinski M, Cho MH, Bakke P, Gulsvik A, Lomas DA, Anderson W, Kong X,Rennard SI, Beaty TH, Hokanson JE, et al. Genome-wide association study ofsmoking behaviours in patients with COPD. Thorax. 2011;66:894–902.30. The UKKC. The UK10K project identifies rare variants in health and disease.Nature. 2015;526:82–90.31. Wagner AH, Coffman AC, Ainscough BJ, Spies NC, Skidmore ZL, Campbell KM,Krysiak K, Pan D, McMichael JF, Eldred JM, et al. DGIdb 2.0: mining clinicallyrelevant drug–gene interactions. Nucleic Acids Res. 2016;44:D1036–44.32. Wishart DS, Knox C, Guo AC, Shrivastava S, Hassanali M, Stothard P, ChangZ, Woolsey J. DrugBank: a comprehensive resource for in silico drugdiscovery and exploration. Nucleic Acids Res. 2006;34:D668–72.33. Thorgeirsson TE, Gudbjartsson DF, Surakka I, Vink JM, Amin N, Geller F.Sequence variants at CHRNB3-CHRNA6 and CYP2A6 affect smokingbehavior. Nat Genet. 2010;42:448–53.34. Wain LV, Shrine N, Miller S, Jackson VE, Ntalla I, Artigas MS, Billington CK,Kheirallah AK, Allen R, Cook JP, et al. Novel insights into the genetics ofsmoking behaviour, lung function, and chronic obstructive pulmonarydisease (UK BiLEVE): a genetic association study in UK biobank. LancetRespir Med. 2015;3:769–81.35. Lutz SM, Hokanson JE. Genetic influences on smoking and clinical disease.Understanding behavioral and biological pathways with mediation analysis.Ann Am Thorac Soc. 2014;11:1082–3.36. Loukola A, Hällfors J, Korhonen T, Kaprio J. Genetics and smoking. CurrAddict Rep. 2014;1:75–82.37. Hilberink SR, Jacobs JE, van Opstal S, van der Weijden T, Keegstra J,Kempers P, Muris JW, Grol RP, de Vries H. Validation of smoking cessationself-reported by patients with chronic obstructive pulmonary disease. Int JGen Med. 2011;4:85–90.38. Benowitz NL, Hukkanen J, Jacob P. Nicotine chemistry, metabolism, kineticsand biomarkers. In: Henningfield J.E., London E.D., Pogun S. (eds) NicotinePsychopharmacology. Handbook of Experimental Pharmacology, vol 192.Berlin, Heidelberg: Springer; 2009.39. Patel YM, Stram DO, Wilkens LR, Park S-SL, Henderson BE, Marchand LL,Haiman CA, Murphy SE. The contribution of common genetic variation tonicotine and cotinine Glucuronidation in multiple ethnic/racial populations.Cancer Epidemiol Biomarkers Prev. 2015;24:119–27.40. Benowitz NL. Nicotine addiction. N Engl J Med. 2010;362:2295–303.Obeidat et al. Respiratory Research  (2018) 19:59 Page 9 of 1041. Zemen BG, Lai MH, Whitt JP, Khan Z, Zhao G, Meredith AL. Generation ofKcnma1(fl)-tdTomato, a conditional deletion of the BK channel α subunit inmouse. Physiol Rep. 2015;3:e12612.42. Bell TJ, Miyashiro KY, Sul J-Y, McCullough R, Buckley PT, Jochems J, MeaneyDF, Haydon P, Cantor C, Parsons TD, Eberwine J. Cytoplasmic BK(ca) channelintron-containing mRNAs contribute to the intrinsic excitability ofhippocampal neurons. Proc Natl Acad Sci U S A. 2008;105:1901–6.43. Buchhalter AR, Fant RV, Henningfield JE. Novel pharmacological approachesfor treating tobacco dependence and withdrawal: current status. Drugs.2008;68:1067–88.44. Polston JE, Cunningham CS, Rodvelt KR, Miller DK. Lobeline augments andinhibits cocaine-induced hyperactivity in rats. Life Sci. 2006;79:981–90.45. Liu YC, Lo YK, Wu SN. Stimulatory effects of chlorzoxazone, a centrallyacting muscle relaxant, on large conductance calcium-activated potassiumchannels in pituitary GH3 cells. Brain Res. 2003;959:86–97.46. Benowitz NL, Peng M, Jacob P 3rd. Effects of cigarette smoking and carbonmonoxide on chlorzoxazone and caffeine metabolism. Clin Pharmacol Ther.2003;74:468–74.47. Addolorato G, Leggio L, Hopf FW, Diana M, Bonci A. Novel therapeuticstrategies for alcohol and drug addiction: focus on GABA, ion channels andtranscranial magnetic stimulation. Neuropsychopharmacology. 2012;37:163–77.48. Garcia-Arcos I, Geraghty P, Baumlin N, Campos M, Dabo AJ, Jundi B,Cummins N, Eden E, Grosche A, Salathe M, Foronjy R. Chronic electroniccigarette exposure in mice induces features of COPD in a nicotine-dependent manner. Thorax. 2016;71:1119–29.49. Bossé Y, Postma DS, Sin DD, Lamontagne M, Couture C, Gaudreault N,Joubert P, Wong V, Elliott M, van den Berge M, et al. Molecular Signature ofSmoking in Human Lung Tissues. Cancer Res. 2012;72:3753–63.50. Cadet JL, Brannock C, Krasnova IN, Jayanthi S, Ladenheim B, McCoy MT,Walther D, Godino A, Pirooznia M, Lee RS. Genome-wide DNAhydroxymethylation identifies potassium channels in the nucleusaccumbens as discriminators of methamphetamine addiction andabstinence. Mol Psychiatry. 2017;22:1196–204.Obeidat et al. Respiratory Research  (2018) 19:59 Page 10 of 10


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items