Joint Associations of 61 Genetic Variants in the Nicotinic Acetylcholine Receptor Genes with Subclinical Atherosclerosis in American Indians: A Gene-Family Analysis

Abstract

Background

Atherosclerosis is the underlying cause of cardiovascular disease, the leading cause of morbidity and mortality in all American populations including American Indians. Genetic factors play an important role in the etiology of atherosclerosis. While a single SNP may explain only a small portion of variability in disease, the joint effect of multiple variants in a pathway on disease susceptibility could be large.

Methods and Results

Using a gene-family analysis, we investigated the joint associations of 61 tag SNPs in seven nicotinic acetylcholine receptors (nAChRs) genes with subclinical atherosclerosis, as measured by carotid intima-media thickness (IMT) and plaque score, in 3,665 American Indians from 94 families recruited by the Strong Heart Family Study (SHFS). Although multiple SNPs showed marginal association with IMT and/or plaque score individually, only a few survived adjustments for multiple testing. However, simultaneously modeling of the joint effect of all 61 SNPs in seven nAChRs genes revealed significant association of the nAChR gene family with both IMT and plaque score, independent of known coronary risk factors.

Conclusions

Genetic variants in the nicotinic acetylcholine receptors gene family jointly contribute to subclinical atherosclerosis in American Indians participated in the SHFS. These variants may influence the susceptibility of atherosclerosis through pathways other than cigarette smoking per se.

Introduction

Atherosclerosis is the primary cause of cardiovascular disease (CVD), the leading cause of morbidity and mortality in all American populations including American Indians.¹ Although lifestyle and environmental risk factors are believed to be significant contributors to the etiology of atherosclerosis, genetic predisposition also plays a critical role.² Recent genome-wide association studies (GWAS) have identified multiple genetic variants, each of which contributes a small fraction to the interindividual variability in CVD risk.^3-5 It is well accepted that the etiology of atherosclerosis involves many genes, but a single gene does not cause disease individually; instead, multiple genes act synergistically in the context of biological pathways in leading to disease susceptibility.⁶ A gene-based or gene-family approach taking into account the joint contribution of multiple variants with marginal individual effect may capture a large proportion of the associated genetic variants, and thus should have a higher power than single gene analysis in dissecting the complex genetic etiology of atherosclerosis.

Cigarette smoking is a strong risk factor for cardiovascular disease. It disrupts vascular wall morphology, impairs endothelial function,⁷ induces vascular inflammation and oxidative stress⁸ and causes insulin resistance,⁹ whereas smoking cessation could mitigate or reverse the progression of atherosclerotic CVD.¹⁰ American Indians have the highest prevalence of cigarette smoking of all US ethnic groups¹¹ and also suffer from high rates of CVD and diabetes. It is unclear whether and how cigarette smoking contributes to the pathogenesis of atherosclerosis in American Indians. Elucidation of the biological pathways linking smoking to atherosclerosis will provide novel strategies for the prevention and treatment of CVD and its related conditions in this ethnically important but traditionally understudied population, as well as in many others where smoking and diabetes play roles in CVD risk.

Nicotine, the major bioactive component of cigarette smoke, affects the cardiovascular system through stimulating the sympathetic nerve pathways,¹² and causes damages to the vascular wall and endothelial function.¹³ Nicotine acts by binding to nicotinic acetylcholine receptors (nAChRs), a superfamily of ligand-gated ion channels that are widely present within neuronal and non-neuronal cell types.¹⁴ Evidence from human and animal research has documented that genetic polymorphisms in nAChRs are associated with nicotine dependence¹⁵ and lung cancer.¹⁶ However, little research has been done to investigate the potential impact of the nAChRs variants on susceptibility to subclinical atherosclerosis. Moreover, existing studies focused on single gene analysis, which is less powerful in detecting variants with small genetic effect and cannot capture the joint contribution of multiple genes. The goal of this study is to conduct a gene family analysis to examine the joint impact of 61 tag SNPs from seven nAChRs genes on preclinical CVD assessed by carotid intimal medial thickness (IMT) and plaque score in a large, well-characterized American Indian population.

Methods

Study population

This study used data from the Strong Heart Family Study (SHFS), a multicenter, family-based prospective study designed to identify genetic factors for cardiovascular disease, diabetes and their risk factors in American Indians. Detailed descriptions of the SHFS protocols for the collection of phenotype data have been described previously.¹ Briefly, a total of 3,665 tribal members (aged 18 years and older) from 94 families residing in Arizona (AZ), North and South Dakota (DK) and Oklahoma (OK) were recruited and examined between 2001 and 2003. Among the 94 families, 76 are three-generation pedigrees (26 from AZ, 28 from OK, and 22 from DK) and 18 are two-generation pedigrees (5 from AZ, 8 from OK, and 5 from DK). The largest family size is 113 individuals from DK, 61 from OK, and 80 from AZ, with an average family size of 38 (37 in AZ, 34 in OK, and 45 in DK). The largest sibling size is 9 in DK, 9 in OK, and 10 in AZ. The SHFS protocols were approved by the Institutional Review Boards from the Indian Health Service and the participating centers. All participants underwent a personal interview to collect data on demographic characteristics, medical history and lifestyle risk factors including smoking, alcohol consumption, diet and physical activity. A physical examination was given to each participant, including anthropometric and blood pressure measurements and an examination of the heart and lungs. Laboratory methods were reported previously.^{1, 17} All participants have given informed consent for genetic study of cardiovascular disease, diabetes and associated risk factors.

Subclinical atherosclerosis assessment

All study participants underwent carotid ultrasonography using Acuson Sequoia machines equipped with 7 MHz vascular probes on the day of the study visit using a standardized protocol as described previously.^{18, 19} Briefly, the extracranial segments of the left and right carotid arteries were extensively scanned for the presence of discrete atherosclerotic plaque, defined as focal protrusion (intimal-medial thickness, IMT) with a thickness exceeding that of the surrounding wall by 50%. Plaque score, a measure of the extent of atherosclerosis, was calculated by the number of left and right segments (common carotid, bulb, internal carotid, external carotid) containing plaque; thus plaque score ranged from 0 to 8. Intimal-medial thickness (IMT) of the far wall of the distal common carotid artery (CCA) was measured at end diastole on multiple cycles of M-mode images. IMT was never measured at the level of a plaque, and the average of the left and right values was used in statistical analyses. All ultrasound measurements were performed by trained research sonographers and interpreted by a single highly experienced cardiologist who was blinded to the clinical characteristics of the participants.

Measurements of coronary risk factors

A detailed description for the collection and measurement of coronary risk factors was reported previously.¹ Briefly, cigarette smoking was assessed via questionnaire and participants were grouped as smokers (former and current smoker) versus never smokers. Pack-years were calculated by multiplying the number of packs of cigarettes smoked per day by the number of years the person has smoked. Participants were categorized into current drinkers, former drinkers and never drinkers based on their history of alcohol consumption. Physical activity was assessed by the mean number of steps per day calculated by averaging the total number of steps recorded each day during the 7-day period. Body mass index (BMI) was calculated by dividing weight in kilograms by the square of height in meters. Hypertension was defined as blood pressure levels of 140/90 mm Hg or higher or use of antihypertensive medications. According to the 1997 American Diabetes Association (ADA) criteria,²⁰ diabetes was defined as fasting plasma glucose ≥7.0 mmol/L or receiving insulin or oral hyperglycemic treatment. Impaired fasting glucose (IFG) was defined as a fasting glucose of 6.1-7.0 mmol/L. Fasting glucose <6.1 mmol/L was defined as normal.

Tag SNPs selection and genotyping

A total of 61 tag SNPs in seven nAChRs genes (CHRNA3-A6, CHRNB2-B4) were selected and genotyped in 3,665 participants from the Strong Heart Family Study. These genes were consistently reported to be associated with cigarette smoking in previous studies. Information for these SNPs has been described in our previous study.²¹ To choose tag SNPs in each candidate gene, we used the computer program Haploview 4.2 ²² with an r² threshold of 0.80 for linkage disequilibrium (LD). The following criteria were also considered: minor allele frequency, SNP location (i.e., coding region) and Illumina design scores (quantifying how likely a SNP can be genotyped). SNPs that could not be tagged (i.e., singletons) were included as long as their design scores were greater than 0.15. All genotyping was done at the Texas Biomedical Research Institute using the Illumina VeraCode technology (Illumina, Inc., San Diego, CA). The average genotyping call rates were > 98% for the 61 tagSNPs, and sample success rate was 99.5%.

Statistical analysis

Single SNP association analysis

The association of each SNP with IMT or plaque score, separately, was assessed using the computer program FBAT,^{23, 24} adjusting for age, sex, study center (AZ vs. OK vs. DK), smoking status (ever-smokers vs. never), alcohol intake (current vs. former vs. never), waist-to-hip ratio (WHR), diabetes status, high density lipoprotein cholesterol (HDL), low density lipoprotein cholesterol (LDL), systolic blood pressure (SBP), levels of physical activity, plasma fibrinogen, and renal function (assessed by estimated glomerular filtration rate, eGFR). Carotid plaque score, which ranges from 0-8, was treated as a continuous variable in the analysis by FBAT, which was reported to be robust to the distribution of response variables.^{23, 25, 26}

Gene-based and gene-family analysis

We examined the association of a candidate gene (including all SNPs with the gene) with IMT or plaque score by combining p-values from single SNP association analysis. The gene-based association was assessed by the truncated product method (TPM).²⁷ A gene-family analysis was then performed by combining p-value of each gene obtained from gene-based analysis, including all seven genes in the nAChRs family. The TPM method has been evaluated extensively by simulation studies²⁸ and its application to determine gene-based or gene-family associations has been described recently.²¹

Sensitivity Analyses

To evaluate the impact of diabetes on the association between genetic variants and subclinical atherosclerosis, we conducted separate analyses in diabetic patients and normal controls (those with normal fasting glucose). To examine whether the gene-based or gene-family associations are primarily driven by the most significant SNPs, we conducted sensitivity analysis by removing these SNPs from gene- or gene-family analysis. We also performed sensitivity analysis to investigate how different truncation points of p-values in TPM affect the statistical significance of gene-based or gene-family analysis by assuming different truncation cutoffs. In addition, to examine whether cigarette smoking per se affects the relationship between genetic variants and IMT or plaque score, we performed sensitivity analysis by comparing results with or without adjustment for smoking status. Multiple testing was corrected using the adjusted false-discovery rate (FDR),²⁹ and statistical significance was defined as FDR-corrected p < 0.05. To improve normality, continuous variables were logarithmically transformed prior to statistical analyses. Participants with missing information on smoking status or subclinical measures were excluded from statistical analyses (N = 84). Analyses were done using Matlab 7.10.0.499 (The MathWorks, Inc., Natick, MA).

Results

Baseline characteristics of the study participants

Table 1 presents the baseline characteristics of the study participants according to smoking status. Compared to never smokers, smokers were older, more likely to be males, and more likely to be centrally obese. No significant difference in IMT or plaque score was observed between smokers and never smokers.

Table 1.

Characteristics of the study participants according to smoking status

Characteristic	Ever Smoker (n=2,134)* (Mean ± SD or %)	Never smoker (n=1,531) (Mean ± SD or %)	P‡
Age (years)	41.5±15.8	37.9±18.5	0.0034
Male sex (%)	44.6	33.7	<0.0001
Body mass index (kg/m2)	32.2±7.9	32.3±7.9	0.8258
Waist/hip ratio	0.92±0.08	0.90±0.09	0.0003
Systolic blood pressure (mmHg)	123.3±17.0	121.7±17.3	0.2862
Diastolic blood pressure(mmHg)	76.7±10.9	75.5±11.5	0.5729
High-density lipoprotein (mg/dL)	50.5±14.5	51.2±14.7	0.3848
Low-density lipoprotein (mg/dL)	99.3±29.7	96.4±28.9	0.5241
Plasma fibrinogen (mg/dL)	391.6±88.7	388.3±92.3	0.2783
Estimated glomerular filtration rate (eGFR, ml/min/1.73m2)	99.8±27.4	101.8±30.1	0.5994
Total cholesterol (mg/dL)	183.2±38.5	177.2±34.9	0.3726
Fasting glucose (mg/dL)	115.7±52.6	112.1±52.8	0.2176
Insulin (uU/mL)	18.7±20.2	18.8±20.4	0.3017
Type 2 Diabetes (%)	24.2	21.0	0.0251
Insulin resistance (HOMA)	5.58±7.19	5.45±7.07	0.3684
Intima-medial thickness (mm)	0.68±0.16	0.64±0.15	0.8631
Plaque score	0.80±1.46	0.57±1.24	0.4106

^‡P values were obtained by FBAT-GEE24 adjusting for age and sex wherever possible.

^*Former plus current smokers.

Results for single SNP association analysis

Information of the 61 SNPs and their LD patterns were reported elsewhere.²¹ We found evidence for a different genetic architecture among study participants from the three study centers, suggesting a possible population admixture among our study participants. However, this should not be an issue for our analysis because we used the family-based association test (FBAT), which is robust to population substructure.²⁵

Of the 61 SNPs examined, multiple SNPs showed individual association with IMT or plaque score. For example, four SNPs in CHRNA3, two SNPs in CHRNA4, six SNPs in CHRNA5, two SNPs in CHRNB2, one SNP in CHRNB3 and five SNPs in CHRNB4 were significantly associated with IMT, whereas one SNP in CHRNA3, one SNP in CHRNA4, three SNPs in CHRNA5, one SNP in CHRNA6, four SNPs in CHRNB2, one SNP in CHRNB3 and five SNPs in CHRNB4 were associated with plaque score. However, only three low frequency SNPs (rs3811450 in CHRNB2, rs4952 in CHRNB3 and rs12914008 in CHRNB4, minor allele frequencies 1.6%, 1.4% and 0.6%, respectively) survived correction for multiple testing. All three SNPs are in Hardy-Weinberg Equilibrium (all p’s >0.95 based on Chi-square test). Results for single SNP association analysis with IMT and plaque score are shown in Table 2 and Table 3, respectively. The associations of the three low frequency SNPs with IMT and plaque score are presented in Table 4.

Table 2.

Single SNP association of the 61 SNPs in nAChRs genes with IMT by FBAT

SNP	Gene	P*	SNP	Gene	P*
rs1051730	CHRNA3	0.0126	rs905739	CHRNA5	0.0416
rs11637630	CHRNA3	0.0464	rs951266	CHRNA5	0.0176
rs12910984	CHRNA3	0.0558	rs2304297	CHRNA6	0.3074
rs12914385	CHRNA3	0.0133	rs2072658	CHRNB2	0.2557
rs1317286	CHRNA3	0.0206	rs2072659	CHRNB2	0.0048
rs1878399	CHRNA3	0.1993	rs2072660	CHRNB2	0.2341
rs3743074	CHRNA3	0.1187	rs2072661	CHRNB2	0.1482
rs3743078	CHRNA3	0.0857	rs3811450	CHRNB2	2.00×10−13
rs578776	CHRNA3	0.1245	rs10958726	CHRNB3	0.3215
rs6495308	CHRNA3	0.0558	rs13277254	CHRNB3	0.3507
rs660652	CHRNA3	0.1944	rs13280604	CHRNB3	0.2676
rs7177514	CHRNA3	0.0722	rs4950	CHRNB3	0.3229
rs2236196	CHRNA4	0.0210	rs4952	CHRNB3	1.38×10−13
rs2273504	CHRNA4	0.1301	rs4953	CHRNB3	0.6827
rs3787116	CHRNA4	0.3916	rs4954	CHRNB3	0.4007
rs3787137	CHRNA4	0.0259	rs6474413	CHRNB3	0.3690
rs6122429	CHRNA4	0.0744	rs11633223	CHRNB4	0.1259
rs11633585	CHRNA5	0.3405	rs11636605	CHRNB4	0.0085
rs11637635	CHRNA5	0.2355	rs12440014	CHRNB4	0.0104
rs16969968	CHRNA5	0.0176	rs12914008	CHRNB4	5.66×10−14
rs17483686	CHRNA5	0.2905	rs1316971	CHRNB4	0.0086
rs17486278	CHRNA5	0.0176	rs16970006	CHRNB4	0.0995
rs2036527	CHRNA5	0.0158	rs17487223	CHRNB4	0.1496
rs514743	CHRNA5	0.1844	rs1948	CHRNB4	0.2882
rs569207	CHRNA5	0.0955	rs1996371	CHRNB4	0.0199
rs588765	CHRNA5	0.2353	rs3813567	CHRNB4	0.0975
rs615470	CHRNA5	0.0669	rs3971872	CHRNB4	0.1569
rs637137	CHRNA5	0.1030	rs7178270	CHRNB4	0.1259
rs680244	CHRNA5	0.2040	rs8023462	CHRNB4	0.2292
rs684513	CHRNA5	0.5849	rs950776	CHRNB4	0.2784
rs8034191	CHRNA5	0.0259

^*Adjusted for age, sex, study center, smoking status, alcohol intake, WHR, HDL, LDL, SBP, physical activity, plasma fibrinogen, diabetes status, and renal function. P-values in bold indicate statistical significance after correction for multiple testing by FDR (q-value).

Table 3.

Single SNP association of the 61 SNPs in nAChRs genes with plaque score by FBAT

SNP	Gene	P*	SNP	Gene	P*
rs1051730	CHRNA3	0.0450	rs905739	CHRNA5	0.2248
rs11637630	CHRNA3	0.2735	rs951266	CHRNA5	0.0446
rs12910984	CHRNA3	0.2946	rs2304297	CHRNA6	0.0310
rs12914385	CHRNA3	0.0936	rs2072658	CHRNB2	0.0814
rs1317286	CHRNA3	0.1062	rs2072659	CHRNB2	0.0096
rs1878399	CHRNA3	0.1294	rs2072660	CHRNB2	0.0162
rs3743074	CHRNA3	0.1785	rs2072661	CHRNB2	0.0210
rs3743078	CHRNA3	0.3838	rs3811450	CHRNB2	8.77×10−6
rs578776	CHRNA3	0.2243	rs10958726	CHRNB3	0.1105
rs6495308	CHRNA3	0.2989	rs13277254	CHRNB3	0.1204
rs660652	CHRNA3	0.1285	rs13280604	CHRNB3	0.1411
rs7177514	CHRNA3	0.3181	rs4950	CHRNB3	0.1236
rs2236196	CHRNA4	0.1164	rs4952	CHRNB3	0.0006
rs2273504	CHRNA4	0.0498	rs4953	CHRNB3	0.8269
rs3787116	CHRNA4	0.4439	rs4954	CHRNB3	0.2632
rs3787137	CHRNA4	0.4317	rs6474413	CHRNB3	0.1725
rs6122429	CHRNA4	0.2180	rs11633223	CHRNB4	0.0262
rs11633585	CHRNA5	0.4127	rs11636605	CHRNB4	0.4455
rs11637635	CHRNA5	0.1285	rs12440014	CHRNB4	0.4135
rs16969968	CHRNA5	0.0446	rs12914008	CHRNB4	0.0001
rs17483686	CHRNA5	0.3991	rs1316971	CHRNB4	0.4705
rs17486278	CHRNA5	0.0446	rs16970006	CHRNB4	0.0255
rs2036527	CHRNA5	0.0614	rs17487223	CHRNB4	0.2787
rs514743	CHRNA5	0.1422	rs1948	CHRNB4	0.3453
rs569207	CHRNA5	0.2180	rs1996371	CHRNB4	0.2607
rs588765	CHRNA5	0.0928	rs3813567	CHRNB4	0.0047
rs615470	CHRNA5	0.1221	rs3971872	CHRNB4	0.0428
rs637137	CHRNA5	0.2194	rs7178270	CHRNB4	0.1021
rs680244	CHRNA5	0.1745	rs8023462	CHRNB4	0.3242
rs684513	CHRNA5	0.1775	rs950776	CHRNB4	0.2049
rs8034191	CHRNA5	0.2323

^*Adjusted for age, sex, study center, smoking status, alcohol intake, WHR, HDL, LDL, SBP, physical activity, plasma fibrinogen, diabetes status, and renal function. P-values in bold indicate statistical significance after correction for multiple testing by FDR (q-value).

Table 4.

Associations of the three most significant SNPs with subclinical atherosclerosis

SNP	Gene	Risk allele	MAF	OR (95% CI)*
				IMT(highest vs. lowest tertile)	Plaque(plaque >0 vs. plaque=0)
rs3811450	CHRNB2	A	0.016	1.09 (1.04- 1.14)	1.52 ( 1.20- 1.92)
rs4952	CHRNB3	A	0.014	1.16 (1.25- 1.04)	2.02 (1.57- 2.61)
rs12914008	CHRNB4	A	0.006	1.60 (1.32- 1.92)	1.35 (1.03- 1.82)

^*Results were obtained by mixed model regression, adjusting for age, sex, study center, smoking status, alcohol intake, WHR, HDL, LDL, SBP, physical activity, plasma fibrinogen, diabetes status, and renal function.

Results for gene-based and gene-family analysis

Table 5 shows the results of gene-based and gene-family analyses. For gene-based analysis, variants in three genes (CHRNB2, CHRNB3 and CHRNB4) significantly contribute to susceptibility for IMT and plaque score (all p’s ≤ 0.0004). Gene-family analysis comprising all seven genes demonstrates that the nAChRs gene family was significantly associated with both IMT and plaque score (both p’s <10⁻⁵) after correction for multiple comparisons.

Table 5.

Gene-based and gene-family associations of the seven nAChRs genes with IMT and plaque score

	IMT	Plaque score	IMT*	Plaque score*
Gene-based analysis
CHRNA3	0.0432	0.1725	0.0432	0.1725
CHRNA4	0.0488	0.2147	0.0488	0.2147
CHRNA5	0.0396	0.1337	0.0396	0.1337
CHRNA6	0.3076	0.0314	0.3076	0.0314
CHRNB2	6.72×10−8	8.63×10−5	0.0252	9.32×10−5
CHRNB3	3.16×10−8	0.0004	0.3982	0.0682
CHRNB4	5.14×10−9	7.62×10−5	0.0034	0.0061
Gene-family analysis	6.46×10−9	3.82×10−5	0.0086	1.39×10−4

P-values in bold indicate statistical significance after correction for multiple testing by FDR (q-value).

^*P-values after removing the three most significant SNPs (rs3811450, rs4952 and rs12914008).

Results for sensitivity analyses

Diabetes is an important risk factor for CVD in American Indians. To examine whether diabetes modifies the association of nAChR genetic variants with subclinical atherosclerosis, we stratified our analyses by diabetic status. Results for single SNP association analysis stratified by diabetes are shown in Supplementary Tables 1 and 2 for IMT and plaque score, respectively. Results for gene-based and gene-family association analyses, stratified by diabetes status, are shown in Supplementary Tables 3 and 4, respectively. We observed differential effects of diabetes on the gene-based associations. For example, among participants with normal fasting glucose, two additional genes (CHRNA3, CHRNA5) were significantly associated with plaque score but not IMT, whereas another gene (CHRNA4) showed significant association with IMT but not plaque score. However, this pattern was not observed among patients with diabetes. It appears that diabetes also modifies the association of single SNP with subclinical atherosclerosis. For instance, the statistical significance for the associations of three SNPs (rs3811450 in CHRNB2, rs4952 in CHRNB3 and rs12914008 in CHRNB4) with IMT exceed genome-wide significant level (p<10⁻¹³) among subjects with normal fasting glucose but not among those with diabetes (though the associations are also statistically significant). After removing the three most significant SNPs (rs3811450, rs4952 and rs12914008) and correction for multiple testing, the associations of CHRNB2 and CHRNB3 with IMT or plaque score substantially attenuated. The P-values for the associations of CHRNB4 gene with IMT or plaque score were reduced but remained statistically significant. The gene-family association also remained statistically significant after removing the three most significant SNPs from gene-family analysis (Table 5). Our sensitivity analysis also indicates that the statistical significance of gene-based or gene-family analysis is quite robust to the different truncation points of p-values for TPM analysis (Supplementary Table 5). In addition, no significant difference was observed in the relationship between nAChRs variants and subclinical atherosclerosis between ever smokers and never smokers, suggesting that these genetic variants may influence the susceptibility of atherosclerosis through pathways beyond cigarette smoking per se.

Discussion

In a sample of 3,665 American Indians who participated in the Strong Heart Study, we conducted gene-based and gene-family analyses to examine the joint associations of 61 tag SNPs in seven nAChRs genes with preclinical atherosclerosis. We found that, although multiple SNPs showed individual nominal or marginal associations with IMT and/or plaque score, only a few survived corrections for multiple testing. However, a gene-family analysis considering the joint contribution of multiple SNPs revealed a significant association of the gene family with both IMT and plaque score. To our best knowledge, this is the first study examining the joint contribution of multiple candidate genes involved in the nAChRs pathway to the susceptibility of atherosclerosis in any ethnic groups.

Several aspects of our investigation merit comment. First, our study demonstrated that a single SNP may show no or marginal association with IMT or plaque score, but the joint impact of multiple SNPs within a gene or a pathway on disease susceptibility could be substantial. For example, after correction for multiple testing, no SNP in the CHRNA4 gene was individually associated with IMT among participants with normal fasting glucose, but gene-based analysis revealed a significant association of this gene with IMT. Similarly, no SNP in CHRNA3 or CHRNA5 showed individual association with plaque score among subjects with normal fasting glucose, but gene-based analysis demonstrated significant associations of these two genes with plaque score (Supplementary Table 3). This is consistent with previous studies demonstrating that, for complex traits involving multiple variants, the contribution of a single genetic polymorphism could be small, but statistical approaches that take into account the cumulative effect of multiple variants, such as a gene-family analysis employed in this study, could capture the joint contribution of many variants simultaneously, and thus provides a better opportunity in identifying disease genes.³⁰ Second, previous studies repeatedly reported an association of rs16969968 in the CHRNA5 gene with nicotine dependence in European Americans or African Americans.^{16, 31} Our analyses, however, found only a nominal association of this SNP with IMT or carotid plaque score, probably due to difference in genetic background between American Indians and other ethnic populations. It is also possible that this SNP influences subclinical atherosclerosis through pathways beyond cigarette smoking. Third, the gene cluster CHRNA3/CHRNA5/CHRNB4, located on chromosome 15q24, was consistently reported to be associated with nicotine dependence.^{32, 33} Although the CHRNB4 gene showed strong association with both IMT and plaque score, we did not detect a significant association of CHRNA3 or CHRNA5 with subclinical atherosclerosis by either single gene analysis or gene-family analysis. However, our sensitivity analyses revealed that this gene cluster was significantly associated with the extent of atherosclerosis among participants with normal fasting glucose levels, indicating that the effect of this gene cluster on subclinical atherosclerosis might be mediated through blood glucose regulation, a known mechanism involved in cardiovascular disease.³⁴ However, it is unclear why this association was not observed in diabetic patients.

In this study, we detected significant associations of three low frequency variants (rs3811450, rs4952 and rs12914008, minor allele frequencies 1.6%, 1.4% and 0.6%, respectively) with both carotid IMT and plaque score. Although these SNPs (or their nearby SNPs that are in high LD with them) were previously associated with nicotine response or nicotine dependence,^{16, 35, 36} neither was reported to be associated with subclinical atherosclerosis, thus our findings may represent novel rare variants with large effects on subclinical atherosclerosis in American Indians.

Both carotid IMT and plaque score are markers of subclinical vascular disease, but they might reflect different stages or severity of atherosclerosis. Compared to IMT which measures the thickness of carotid artery, plaque score reflects the severity of irregular morphology and lumen narrowing,³⁷ thus is considered as a marker of advanced atherosclerosis. In our analysis, IMT was measured in the common carotid artery with absence of plaque, thus providing a measure that could distinguish carotid IMT from plaque presence. This is different from previous studies that measured combined carotid and femoral artery IMT,³⁸ or measured the wall thickness regardless of the absence or presence of plaque. ³⁹ Utilizing IMT measurement incorporating focal plaque thickness could potentially conflate the two entities and overestimate the effect of IMT.⁴⁰ Interestingly, it appears that genetic associations at several SNP loci are stronger with IMT than with plaque score. This observation is in line with previous studies demonstrating a higher heritability of IMT than plaque score,⁴¹ and may suggest that the studied genes could have a larger effect on earlier atherosclerosis (as measured by IMT) than on advanced atherosclerosis (as assessed by plaque presence). It is also probable that these genes may affect wall thickening through other as yet unknown pathways. This finding does not contradict with previous studies reporting a stronger prognostic value of carotid plaque than IMT in predicting CVD events,⁴² including research from our study population.⁴⁰

Although smokers tend to be thinner than non-smokers,⁴³ they are more likely to have increased abdominal obesity,⁴⁴ a strong risk factor for carotid atherosclerosis.⁴⁵ The mechanism underlying the association between smoking and body weight remains unclear, with a recent study indicating that, by activation of hypothalamic α₃β₄ nAChRs, smoking can stimulate the activity of pro-opiomelanocortin, resulting in a decreased appetite and body weight.⁴⁶ In this study, we did not observe a significant difference in BMI among ever smokers and never smokers, but the WHR was significantly higher in ever smokers than never smokers. However, our results are unlikely to be confounded by central obesity because we adjusted for WHR in all statistical analyses.

The biological mechanisms through which nAChRs gene variants influence atherosclerosis are unclear. According to our results, it seems that cigarette smoking per se may not cause atherosclerosis directly because we controlled for smoking in all statistical analyses. However, given previous evidence for nicotine toxicity in vascular system, ^{7, 10} this hypothesis needs to be confirmed in future research. Given the importance of nicotinic acetylcholine receptors (nAChRs) in neuronal function,⁴⁷ it is possible that cigarette smoking could affect CVD risk through the central nerve system.⁴⁶ Moreover, because atherosclerosis is an inflammatory process,⁴⁸ and smoking increases inflammation,⁴⁹ it is also plausible that these genetic variants may influence carotid atherosclerosis through their impact on inflammatory responses to cigarette smoking. In this study, we did observe a higher level of plasma fibrinogen in smokers than in never smokers (though the difference was statistically insignificant), but no difference in IMT and plaque score was observed between these two groups. Alternatively, the nAChR genetic variations may influence atherogenesis through smoking-induced oxidative stress.⁵⁰ Of course, it is also possible that these nAChRs variants could affect carotid atherosclerosis through other independent yet uncharacterized mechanisms.

Our study has a few limitations. First, though we were able to control many of the potential confounders, we cannot rule out entirely the possibility of residual confounding by other unknown or unmeasured factors. Second, this study used a cross-sectional design, which precluded any causal inference. Third, IMT was measured only in the distal common carotid artery and thus the association of genetic variants with bifurcation or internal carotid IMT cannot be assessed. Finally, our analyses were undertaken among a cohort of a single ethnic group and thus needs to be replicated in other study populations.

In summary, this study provides initial evidence that multiple genetic variants in the nAChRs gene family jointly contribute to the susceptibility of subclinical atherosclerosis in American Indians participated in the Strong Heart Study. The genetic effect of these polymorphisms on atherosclerosis is likely mediated by pathways beyond cigarette smoking per se. Our results may provide valuable information for individualized prevention or intervention on atherosclerosis in American Indians who suffer from disproportionately high prevalence of CVD and diabetes.

Supplementary Material

01

Supplementary Table 1. Single SNP association of the 61 SNPs with IMT according to diabetes status

Supplementary Table 2. Single SNP association of the 61 SNPs with plaque score according to diabetes status

Supplementary Table 3. Gene-based and gene-family associations of the seven nAChRs genes with IMT and plaque score among participants with normal fasting glucose levels

Supplementary Table 4. Gene-based and gene-family associations of the seven nAChRs genes with IMT and plaque score among participants with diabetes

Supplementary Table 5. Gene-based and gene-family associations of the seven nAChRs genes with IMT and plaque score assuming different truncation points of p-values for TPM analysis

Atherosclerotic cardiovascular disease is the leading cause of morbidity and mortality in all American populations including American Indians. The genetic etiology of atherosclerosis is complex, involving many genes, each of which may contribute to a small or modest effect to disease risk, as well as their interactions with environmental factors. Traditional statistical methods usually analyze individual genetic markers, e.g., single nucleotide polymorphism (SNP), independently. However, single SNP association analysis is less powerful in detecting genetic variants with small effect and may not capture the joint contribution of multiple variants on disease risk. Here we used a gene-family approach to investigate the cumulative effect of 61 tagging SNPs in seven nicotinic acetylcholine receptors (nAChRs) on subclinical atherosclerosis, as measured by intima-media thickness (IMT) and plaque score in a well-characterized population of American Indians. Results show that multiple variants in the nAChRs gene-family jointly contribute to the interindividual variability in IMT and plaque score, independent of established coronary risk factors. In addition, our results suggest that the studied nAChRs variants may influence atherosclerosis through pathways beyond cigarette smoking per se. Findings of this research may provide useful information for individualized intervention or prevention on atherosclerotic cardiovascular diseases in American Indians and other ethnic groups as well.