The CHRNA5–A3–B4 Gene Cluster and Smoking: From Discovery to Therapeutics
2016; Elsevier BV; Volume: 39; Issue: 12 Linguagem: Inglês
10.1016/j.tins.2016.10.005
ISSN1878-108X
AutoresGlenda Lassi, Amy E. Taylor, Nicholas J. Timpson, Paul J. Kenny, Robert J. Mather, Tim Eisen, Marcus R. Munafò,
Tópico(s)Influenza Virus Research Studies
ResumoGenome-wide association studies (GWASs) have highlighted genetic variants in the CHRNA5–CHRNA3–CHRB4 gene cluster associated with smoking heaviness and nicotine dependence as well as known smoking-related diseases such as chronic obstructive pulmonary disease (COPD) and lung cancer. Animal studies have described the anatomy and function of the nicotinic acetylcholine receptor (nAChR) subunits encoded by the CHRNA5–CHRNA3–CHRB4 gene cluster. More in-depth phenotyping in humans is improving our understanding of how these variants contribute to smoking behaviour. Use of this gene cluster in Mendelian randomisation analyses is enabling us to investigate the causal effects of tobacco use. The latest findings suggest that smoking heaviness could be a causal factor for schizophrenia. Validated genetic targets emerging from GWASs that influence smoking behaviour may pave the way towards individually tailored smoking cessation treatments, ultimately reducing the burden of smoking-related diseases. Genome-wide association studies (GWASs) have identified associations between the CHRNA5–CHRNA3–CHRNB4 gene cluster and smoking heaviness and nicotine dependence. Studies in rodents have described the anatomical localisation and function of the nicotinic acetylcholine receptors (nAChRs) formed by the subunits encoded by this gene cluster. Further investigations that complemented these studies highlighted the variability of individuals' smoking behaviours and their ability to adjust nicotine intake. GWASs of smoking-related health outcomes have also identified this signal in the CHRNA5–CHRNA3–CHRNB4 gene cluster. This insight underpins approaches to strengthen causal inference in observational data. Combining genetic and mechanistic studies of nicotine dependence and smoking heaviness may reveal novel targets for medication development. Validated targets can inform genetic therapeutic interventions for smoking cessation and tobacco-related diseases. Genome-wide association studies (GWASs) have identified associations between the CHRNA5–CHRNA3–CHRNB4 gene cluster and smoking heaviness and nicotine dependence. Studies in rodents have described the anatomical localisation and function of the nicotinic acetylcholine receptors (nAChRs) formed by the subunits encoded by this gene cluster. Further investigations that complemented these studies highlighted the variability of individuals' smoking behaviours and their ability to adjust nicotine intake. GWASs of smoking-related health outcomes have also identified this signal in the CHRNA5–CHRNA3–CHRNB4 gene cluster. This insight underpins approaches to strengthen causal inference in observational data. Combining genetic and mechanistic studies of nicotine dependence and smoking heaviness may reveal novel targets for medication development. Validated targets can inform genetic therapeutic interventions for smoking cessation and tobacco-related diseases. GWASs (see Glossary) seek to identify genetic variants associated with specific phenotypes, and cigarette smoking is a complex behavioural phenotype that has successfully been subjected to this approach (Figure 1A , Key Figure). An association between the SNP rs16969968 in CHRNA5 and nicotine dependence was first reported in 2007 in a candidate gene study [1Saccone S.F. et al.Cholinergic nicotinic receptor genes implicated in a nicotine dependence association study targeting 348 candidate genes with 3713 SNPs.Hum. Mol. Genet. 2007; 16: 36-49Crossref PubMed Scopus (659) Google Scholar]. The following year rs1051730 at the same locus (in CHRNA3 but strongly correlated with rs16969968 in samples of European ancestry) was found to be associated with smoking quantity in a GWAS [2Thorgeirsson T.E. et al.A variant associated with nicotine dependence, lung cancer and peripheral arterial disease.Nature. 2008; 452: 638-642Crossref PubMed Scopus (1221) Google Scholar]. This study also reported an association between rs1051730 and two smoking-related diseases: lung cancer and peripheral arterial disease. Importantly, CHRNA5 was not recognised as a strong candidate gene at the time given the then-known neurobiology of tobacco dependence. Animal models had implicated the α4 and β2 nicotinic receptor subunits as critical to nicotine's reinforcing effects [3Picciotto M.R. et al.Acetylcholine receptors containing the β2 subunit are involved in the reinforcing properties of nicotine.Nature. 1998; 391: 173-177Crossref PubMed Scopus (1123) Google Scholar, 4Tapper A.R. et al.Nicotine activation of α4* receptors: sufficient for reward, tolerance, and sensitization.Science. 2004; 306: 1029-1032Crossref PubMed Scopus (589) Google Scholar], and α4β2* partial agonists are one of the most effective treatments available for smoking cessation [5Fowler C.D. Kenny P.J. Nicotine aversion: neurobiological mechanisms and relevance to tobacco dependence vulnerability.Neuropharmacology. 2014; 76: 533-544Crossref PubMed Scopus (118) Google Scholar]. Therefore, while the candidate gene study was published first, the GWAS (which did not require a strong prior hypothesis regarding gene selection) made the greater impact. What followed illustrates the potential for GWASs to advance our understanding of neurobiology in a way that hypothesis-driven investigations such as candidate gene studies (which rely on known or presumed neurobiology) rarely could. These findings have made variants in the CHRNA5–CHRNA3–CHRNB4 region promising targets for the study of nicotine dependence and smoking heaviness, given their association with response to nicotine and its consequent consumption and titration. The CHRNA5–CHRNA3–CHRNB4 gene cluster on chromosome 15 (at 15q25) encodes three (α3, α5, β4) of the eleven (α2–7, α9, α10, β2–4) neuronal nAChR (Figure 2A) subunits. The rs1051730 SNP in CHRNA3 is a coding, synonymous variant that does not result in an amino acid change in the subsequent protein and is therefore unlikely to be of any functional significance; however, rs1051730 is highly correlated with rs16969968 in CHRNA5, a missense mutation that results in the substitution of aspartate (D) to asparagine (N) at the 398th amino acid in the resultant α5 subunit protein (D398N). This variant is functional; in vitro studies have demonstrated that α5 receptor complexes with the aspartic acid variant exhibit a twofold greater maximal response to a nicotine agonist compared with α5 receptor complexes containing the asparagine variant (i.e., the risk variant associated with smoking quantity and nicotine dependence [6Bierut L.J. et al.Variants in nicotinic receptors and risk for nicotine dependence.Am. J. Psychiatry. 2008; 165: 1163-1171Crossref PubMed Scopus (497) Google Scholar]). Interestingly, an analysis of gene expression has shown that the risk allele of the rs1051730 SNP is associated with a lower level of expression of CHRNA5 in the brain and peripheral blood mononuclear cells [7Jackson K.J. et al.Variants in the 15q25 gene cluster are associated with risk for schizophrenia and bipolar disorder.Psychiatr. Genet. 2013; 23: 20-28Crossref PubMed Scopus (35) Google Scholar]. Animal studies have described the role of the α5 nAChR subunit by investigating the α5 knockout-mouse model's phenotype; unfortunately, investigation of the role of the α3 subunit in nicotine intake is challenged by the early postnatal lethality that results from its genetic ablation. Smoke inhaled from a cigarette is a mixture of chemicals that travel through the airways to the lungs. Nicotine is the primary addictive constituent of cigarette smoke; it is absorbed by the alveolar epithelium and then circulates in the bloodstream. Nicotine is an exogenous ligand for nAChRs and binds, after crossing the blood–brain barrier, to nAChRs ubiquitously expressed in the brain (Figure 1C). nAChRs are ligand-gated ion channels that open in response to the binding of acetylcholine and nicotine allowing trafficking of cations (Ca++, Na+, and K+; Figure 2B). Critically, our understanding of the neurobiology of smoking behaviours has been advanced by GWASs, particularly through the identification of how genetic variants that result in functional changes in nAChRs lead in turn to behavioural outcomes. Following the identification of the association of the CHRNA5–CHRNA3–CHRNB4 gene cluster with heaviness of smoking, an elegant series of experiments [8Fowler C.D. et al.Habenular α5 nicotinic receptor subunit signalling controls nicotine intake.Nature. 2011; 471: 597-601Crossref PubMed Scopus (472) Google Scholar] established the underlying mechanism linking this genetic variation to nicotine response. This work utilised an α5 knockout-mouse model, which is analogous to individuals with reduced α5 receptor function (i.e., carriers of the rs16969968 risk allele). Both wild-type and knockout animals showed the expected inverted U-shaped dose–response curve for intravenous nicotine infusions with the difference that knockouts responded more vigorously at high doses. While wild-type mice appeared to titrate the delivery of nicotine dose (through self-administration) to achieve a consistent, desired level, knockout mice did so to a considerably lesser extent, consuming greater amounts as the dose increased. Furthermore, [8Fowler C.D. et al.Habenular α5 nicotinic receptor subunit signalling controls nicotine intake.Nature. 2011; 471: 597-601Crossref PubMed Scopus (472) Google Scholar] also showed that this effect could be 'rescued' in α5 knockout mice through the injection of a lentivirus vector into the medial habenula (MHb), rescuing expression of α5 subunits in this region. The knockout mice did not appear to differ from wild-type mice in their experience of the rewarding effects of nicotine, but the inhibitory effect of high nicotine doses on the activity of reward circuitries observed in wild-type mice appeared to have been largely abolished in knockout mice. This observation aligns with a previous study [9Jackson K.J. et al.Role of α5 nicotinic acetylcholine receptors in pharmacological and behavioral effects of nicotine in mice.J. Pharmacol. Exp. Ther. 2010; 334: 137-146Crossref PubMed Scopus (126) Google Scholar] where the differential effects of nicotine dose on reward between α5 knockouts and wild-types was illustrated using a conditioned place-preference task. In addition, a recent study in humans [10Jensen K.P. et al.A CHRNA5 smoking risk variant decreases the aversive effects of nicotine in humans.Neuropsychopharmacology. 2015; 40: 2813-2821Crossref PubMed Scopus (53) Google Scholar] found an attenuated aversive response to intravenously administrated nicotine in overnight-abstinent smokers who were carriers of the CHRNA5 rs16969968 risk allele genotype. The MHb mainly projects to the interpeduncular nucleus (IPN) and [8Fowler C.D. et al.Habenular α5 nicotinic receptor subunit signalling controls nicotine intake.Nature. 2011; 471: 597-601Crossref PubMed Scopus (472) Google Scholar] observed that diminished IPN activity in response to nicotine was seen in knockout mice. In short, it appears that high doses of nicotine stimulate the MHb–IPN tract through nAChRs containing α5 subunits and elicit aversion, limiting further intake. This does not occur when α5 signalling is deficient and consequently the negative effects of nicotine are attenuated. Similarly, smokers carrying the rs16969968 risk allele are therefore more likely to smoke more heavily than their counterparts without the risk allele. The study reported in [11Frahm S. et al.Aversion to nicotine is regulated by the balanced activity of β4 and α5 nicotinic receptor subunits in the medial habenula.Neuron. 2011; 70: 522-535Abstract Full Text Full Text PDF PubMed Scopus (226) Google Scholar] provides further evidence that the MHb acts as a gatekeeper for nicotine intake. The authors manipulated the concentration of α5 and β4 in vitro while α3 was kept constant, and as a result the cation conductance of the channel was altered. nAChR activity changes according to the local electrostatic charge; therefore, the next step was to test the nicotine-evoked currents in MHb neurons of wild-type mice and transgenic mice (called Tabac mice) characterised by overexpression of β4. The authors [11Frahm S. et al.Aversion to nicotine is regulated by the balanced activity of β4 and α5 nicotinic receptor subunits in the medial habenula.Neuron. 2011; 70: 522-535Abstract Full Text Full Text PDF PubMed Scopus (226) Google Scholar] made patch-clamp recordings after local delivery of nicotine and reported a dramatically higher firing rate in Tabac mouse neurons. The behavioural response observed in vivo also showed significant changes in Tabac mice compared with wild-type mice. Tabac mice exhibited reduced nicotine intake and a strong preference for water rather than low-nicotine-concentration solutions in a two-bottle choice test. By contrast, their littermate controls drank water and nicotine solution equally even at much higher nicotine concentrations. Similarly, wild-type mice showed no preference for a nicotine-conditioned versus saline-conditioned chamber while Tabac mice spent less time in the latter. In addition, expression of the α5 D397N variant (corresponding to the human α5 variant D398N) was elicited in MHb neurons by injecting a lentivirus vector in Tabac mice. The latter restored their nicotine consumption and their two-bottle choice behaviour to a level comparable with wild types. In summary, α5 and β4 compete in regulating nicotine intake: β4 overexpression enhances MHb activity leading to aversion whereas induced α5 expression in Tabac mice normalised nicotine consumption. As we have seen, the first major results of GWASs of smoking behaviours related to nicotine dependence and smoking heaviness. These behaviours are conventionally quantified using self-report measures such as the Fagerström Test for Nicotine Dependence (FTND) and the number of cigarettes smoked daily (Box 1). Both the FTND and the number of cigarettes per day (CPD) are valuable measures, but since they are based on self-reporting they are prone to subjective errors. In addition, individuals differ in how they smoke a cigarette, varying in the number of puffs taken and the volume of smoke inhaled per puff and per cigarette (smoking topography [12Ware J.J. Munafò M.R. Determining the causes and consequences of nicotine dependence: emerging genetic research methods.Curr. Psychiatry Rep. 2014; 16: 477Crossref PubMed Scopus (5) Google Scholar]); therefore, cigarette smokers who consume the same CPD may differ in how much nicotine (and other toxicants) they consume. Individuals homozygous for the major allele at rs16969968 reduce the volume of smoke inhaled per puff when the nicotine content of the cigarette is increased; by contrast, carriers of the risk variant do not [13Macqueen D.A. et al.Variation in the alpha 5 nicotinic acetylcholine receptor subunit gene predicts cigarette smoking intensity as a function of nicotine content.Pharmacogenomics J. 2014; 14: 70-76Crossref PubMed Scopus (27) Google Scholar]. The lack of compensatory behaviour in risk allele carriers when smoking cigarettes with higher levels of nicotine is analogous to the reduced aversive effect of nicotine observed in α5 knockout mice [8Fowler C.D. et al.Habenular α5 nicotinic receptor subunit signalling controls nicotine intake.Nature. 2011; 471: 597-601Crossref PubMed Scopus (472) Google Scholar].Box 1Phenotyping of Smoking BehavioursPhenotyping in humans can rely on subjective methods. Precise phenotypes such as proximate biomarkers better capture genetically disposed interindividual differences in smoking. Here we provide a brief description of common indicators used to quantify smoking dependence and heaviness.CPDCPD is the count of cigarettes smoked in one day from waking up to going to bed. This phenotype has been used in several studies [2Thorgeirsson T.E. et al.A variant associated with nicotine dependence, lung cancer and peripheral arterial disease.Nature. 2008; 452: 638-642Crossref PubMed Scopus (1221) Google Scholar, 50Furberg H. et al.Genome-wide meta-analyses identify multiple loci associated with smoking behavior.Nat. Genet. 2010; 42 (p. 441-U134)Crossref Scopus (839) Google Scholar, 51Berrettini W. et al.α-5/α-3 nicotinic receptor subunit alleles increase risk for heavy smoking.Mol. Psychiatry. 2008; 4: 368-373Crossref Scopus (388) Google Scholar].Questionnaires•The FTND, or Fagerström Test for Cigarette Dependence (FTCD) as Fagerström renamed it in 2012 [52Fagerstrom K. Determinants of tobacco use and renaming the FTND to the Fagerstrom Test for Cigarette Dependence.Nicotine Tob. Res. 2012; 14: 75-78Crossref PubMed Scopus (625) Google Scholar], assesses the intensity of physical addiction to nicotine. The FTND is a six-item questionnaire. The total score range is 0–10 and a higher score indicates higher dependence. Often, smokers with a score greater than 4 are defined as dependent [1Saccone S.F. et al.Cholinergic nicotinic receptor genes implicated in a nicotine dependence association study targeting 348 candidate genes with 3713 SNPs.Hum. Mol. Genet. 2007; 16: 36-49Crossref PubMed Scopus (659) Google Scholar, 6Bierut L.J. et al.Variants in nicotinic receptors and risk for nicotine dependence.Am. J. Psychiatry. 2008; 165: 1163-1171Crossref PubMed Scopus (497) Google Scholar].•The Questionnaire of Smoking Urges (QSU) provides a measure of current craving levels. The QSU is a ten-item questionnaire. Each item requires a 1–5 response and a higher score indicates a greater urge to smoke.Smoking TopographyObjective indicators such as number of puffs, puff volume, and the duration and velocity of the smoking of a cigarette highlight interindividual differences in smoking behaviour. They can lead to different levels of biomarkers for similar levels of CPD.Biomarkers [53Verification S.S.o.B. Biochemical verification of tobacco use and cessation.Nicotine Tob. Res. 2002; 4: 149-159Crossref PubMed Scopus (1541) Google Scholar]Biomarkers are directly correlated with the level of smoking and are measured objectively.•Carbon monoxide (CO): Expired air CO levels in the breath are measured via a portable CO analyser. Smoking status is usually confirmed by a value of 10 parts per million (ppm). CO testing is a rapid, low-cost, and noninvasive method for confirming smoking status [54Emery R.L. Levine M.D. optimal carbon monoxide criteria to confirm smoking status among postpartum women.Nicotine Tob. Res. 2016; 18: 966-970Crossref PubMed Scopus (7) Google Scholar]. Alternatively, carboxyhaemoglobin (COHb) can be measured spectrophotometrically in blood. Environmental sources of CO (e.g., car exhaust fumes) can contribute to the overall level of CO in the organism.•Nicotine: Nicotine levels can be measured in urine, saliva, and blood by means of gas or high-performance liquid chromatography and immunoassays. However, nicotine has a short half-life and is therefore suited to capturing nicotine consumption only in the 8–12 h before testing.•Cotinine: Most (∼90%) nicotine is broken down into cotinine. The half-life of cotinine is relatively long (∼16 h) and it is therefore generally preferred as a biomarker of nicotine intake. Urine tests allow discrimination between smokers and non-smokers while more accurate measures, in saliva and blood, are performed as for nicotine, by means of gas or high-performance liquid chromatography and immunoassays. Phenotyping in humans can rely on subjective methods. Precise phenotypes such as proximate biomarkers better capture genetically disposed interindividual differences in smoking. Here we provide a brief description of common indicators used to quantify smoking dependence and heaviness. CPD CPD is the count of cigarettes smoked in one day from waking up to going to bed. This phenotype has been used in several studies [2Thorgeirsson T.E. et al.A variant associated with nicotine dependence, lung cancer and peripheral arterial disease.Nature. 2008; 452: 638-642Crossref PubMed Scopus (1221) Google Scholar, 50Furberg H. et al.Genome-wide meta-analyses identify multiple loci associated with smoking behavior.Nat. Genet. 2010; 42 (p. 441-U134)Crossref Scopus (839) Google Scholar, 51Berrettini W. et al.α-5/α-3 nicotinic receptor subunit alleles increase risk for heavy smoking.Mol. Psychiatry. 2008; 4: 368-373Crossref Scopus (388) Google Scholar]. Questionnaires•The FTND, or Fagerström Test for Cigarette Dependence (FTCD) as Fagerström renamed it in 2012 [52Fagerstrom K. Determinants of tobacco use and renaming the FTND to the Fagerstrom Test for Cigarette Dependence.Nicotine Tob. Res. 2012; 14: 75-78Crossref PubMed Scopus (625) Google Scholar], assesses the intensity of physical addiction to nicotine. The FTND is a six-item questionnaire. The total score range is 0–10 and a higher score indicates higher dependence. Often, smokers with a score greater than 4 are defined as dependent [1Saccone S.F. et al.Cholinergic nicotinic receptor genes implicated in a nicotine dependence association study targeting 348 candidate genes with 3713 SNPs.Hum. Mol. Genet. 2007; 16: 36-49Crossref PubMed Scopus (659) Google Scholar, 6Bierut L.J. et al.Variants in nicotinic receptors and risk for nicotine dependence.Am. J. Psychiatry. 2008; 165: 1163-1171Crossref PubMed Scopus (497) Google Scholar].•The Questionnaire of Smoking Urges (QSU) provides a measure of current craving levels. The QSU is a ten-item questionnaire. Each item requires a 1–5 response and a higher score indicates a greater urge to smoke. Smoking Topography Objective indicators such as number of puffs, puff volume, and the duration and velocity of the smoking of a cigarette highlight interindividual differences in smoking behaviour. They can lead to different levels of biomarkers for similar levels of CPD. Biomarkers [53Verification S.S.o.B. Biochemical verification of tobacco use and cessation.Nicotine Tob. Res. 2002; 4: 149-159Crossref PubMed Scopus (1541) Google Scholar] Biomarkers are directly correlated with the level of smoking and are measured objectively.•Carbon monoxide (CO): Expired air CO levels in the breath are measured via a portable CO analyser. Smoking status is usually confirmed by a value of 10 parts per million (ppm). CO testing is a rapid, low-cost, and noninvasive method for confirming smoking status [54Emery R.L. Levine M.D. optimal carbon monoxide criteria to confirm smoking status among postpartum women.Nicotine Tob. Res. 2016; 18: 966-970Crossref PubMed Scopus (7) Google Scholar]. Alternatively, carboxyhaemoglobin (COHb) can be measured spectrophotometrically in blood. Environmental sources of CO (e.g., car exhaust fumes) can contribute to the overall level of CO in the organism.•Nicotine: Nicotine levels can be measured in urine, saliva, and blood by means of gas or high-performance liquid chromatography and immunoassays. However, nicotine has a short half-life and is therefore suited to capturing nicotine consumption only in the 8–12 h before testing.•Cotinine: Most (∼90%) nicotine is broken down into cotinine. The half-life of cotinine is relatively long (∼16 h) and it is therefore generally preferred as a biomarker of nicotine intake. Urine tests allow discrimination between smokers and non-smokers while more accurate measures, in saliva and blood, are performed as for nicotine, by means of gas or high-performance liquid chromatography and immunoassays. How and what is measured to capture a phenotype (particularly for complex behaviours such as smoking heaviness) can be critical to the strength of the genetic association observed. Smoking studies that have used cotinine levels as a phenotype illustrate this. Nicotine consumed by a smoker is metabolised principally into cotinine, which is therefore a biomarker for nicotine consumption. It can be reliably detected in the serum, urine, and saliva of smokers (Box 1) [14Ware J.J. et al.Genome-wide meta-analysis of cotinine levels in cigarette smokers identifies locus at 4q13.2.Sci. Rep. 2016; 6: 20092Crossref PubMed Scopus (30) Google Scholar]. In the study reported in [15Keskitalo K. et al.Association of serum cotinine level with a cluster of three nicotinic acetylcholine receptor genes (CHRNA3/CHRNA5/CHRNB4) on chromosome 15.Hum. Mol. Genet. 2009; 18: 4007-4012Crossref PubMed Scopus (93) Google Scholar], the authors showed that the variants in the CHRNA5–A3–B4 gene cluster are more strongly associated with levels of cotinine than CPD. This was subsequently confirmed in an extensive meta-analysis: there is a much stronger signal for the association between the CHRNA5–A3–B4 genotype and heaviness of smoking when this is captured by cotinine levels (4% of phenotypic variation explained) rather than CPD (1% variation explained). Interestingly, the association with cotinine levels remains when CPD is statistically adjusted for in the analysis [15Keskitalo K. et al.Association of serum cotinine level with a cluster of three nicotinic acetylcholine receptor genes (CHRNA3/CHRNA5/CHRNB4) on chromosome 15.Hum. Mol. Genet. 2009; 18: 4007-4012Crossref PubMed Scopus (93) Google Scholar, 16Munafò M.R. et al.Association between genetic variants on chromosome 15q25 locus and objective measures of tobacco exposure.J. Natl. Cancer Inst. 2012; 104: 740-748Crossref PubMed Scopus (135) Google Scholar]. These findings highlight the need for more precise phenotypic measures. However, collecting detailed phenotypic data depletes resources and may not be feasible for large cohorts of individuals. A strategy to select a smaller sample is recall by genotype. For example, in 200 individuals of European ancestry from the general population, we would expect to find only 32 homozygous for the risk variant rs16969968 (AA, asparagine/asparagine) and 78 homozygous for the wild-type variant (GG, aspartate/aspartate) [17Ware J.J. et al.A recall-by-genotype study of CHRNA5-A3–B4 genotype, cotinine and smoking topography: study protocol.BMC Med. Genet. 2014; 15: 13Crossref PubMed Scopus (10) Google Scholar]. If in-depth phenotyping is time consuming and expensive, and collecting data on 200 participants is at the limit of available resources, selectively recruiting 100 minor homozygotes and 100 major homozygotes from a larger genetic screen before conducting in-depth phenotyping is likely to be a far more powerful and efficient use of resources, given the low cost of genotyping. We know that current smokers smoke one more cigarette and show a 138 nmol/l mean increase in cotinine levels for each copy of the rs16969968 minor allele that they may carry [16Munafò M.R. et al.Association between genetic variants on chromosome 15q25 locus and objective measures of tobacco exposure.J. Natl. Cancer Inst. 2012; 104: 740-748Crossref PubMed Scopus (135) Google Scholar]. By recalling participants according to rs16969968 genotype (i.e., the two homozygote groups), we can capture the greatest phenotypic difference in smoking heaviness resulting from genetic difference at the CHRNA5–A3–B4 locus (Figure 1B). As we see in the next section, this approach captures many of the features of a randomised experimental design and therefore any potential confounding factors should be randomly distributed across the sample because it has been selected solely according to genotype. The robust association of variants in the CHRNA5–A3–B4 gene cluster with smoking intensity phenotypes makes variation at this locus a valuable tool for investigating the causal effects of tobacco exposure. Although many of the harmful effects of smoking, such as lung cancer and cardiovascular disease, are well documented (http://www.surgeongeneral.gov/library/reports/50-years-of-progress/full-report.pdf), tobacco use is associated with many other conditions for which causal links remain to be established. For example, in the UK smoking rates are higher in individuals with a mental health condition – estimated to be 33% compared with 19% in the general population (http://www.ash.org.uk/files/documents/ASH_120.pdf). In addition, the mechanisms through which smoking causes some diseases, such as cardiovascular disease, are not yet fully understood. Much of the evidence regarding the health effects of smoking comes from observational data. Causal inference from observational data can be problematic because smoking is associated with a range of lifestyle and demographic factors including socioeconomic status, diet, and other substance use. Therefore, we cannot be certain that associations are due to smoking itself or to these other factors. For example, maternal smoking during pregnancy is linked with numerous offspring outcomes but confounding by familial factors (including genetic factors) is likely to explain some of these associations [18D'Onofrio B.M. et al.Critical need for family-based, quasi-experimental designs in integrating genetic and social science research.Am. J. Public Health. 2013; 103: S46-S55Crossref PubMed Scopus (239) Google Scholar]. Furthermore, individuals may alter their smoking behaviour in response to ill health, so reverse causality may also influence these observational associations. The problems of confounding and reverse causality in the investigation of the causal effects of smoking can be reduced by using variants in the CHRNA5–A3–B4 gene cluster as proxies for smoking heaviness in MR analyses (Box 2). In contrast to direct measurement, germline genotypes reliably associated with risk factors can act as proxy measures of exposure, offering several advantages: genotypes are relatively easy to measure precisely, are stable through time, are largely immutable, and are not correlated with confounding factors, given the mechanisms of Mendelian inheritance [19Smith G.D. Ebrahim S. 'Mendelian randomization': can genetic epidemiology contribute to understanding environmental determinants of disease?.Int. J. Epidemiol. 2003; 32: 1-22Crossref PubMed Scopus (2510) Google Scholar, 20Smith
Referência(s)