Summary
The rs588765-rs16969968 haplotype modifies lung cancer risk more than effects from individual variations at rs16969968 or rs588765, and therefore may be a marker of genetic susceptibility to lung cancer even among never-smokers. This knowledge may facilitate our understanding of lung cancer etiology.
Abstract
The role of haplotypes and the interaction of haplotypes and smoking in lung cancer risk have not been well characterized. We analyzed data from an Italian population-based, case–control study with 1815 lung cancer patients and 1959 healthy controls in discovery, and performed a validation using a case–control study with 2983 lung cancer patients and 3553 healthy controls of European ancestry for replication. Sliding window haplotype analysis within chromosome 15, evaluating 4722250 haplotypes and pair-wise haplotype analysis identified that CHRNA5 rs588765-rs16969968 was the most significant haplotype associated with lung cancer risk (omnibus P = 8.35×10−15 in discovery and 7.26×10−14 in replication), and improved the prediction of case status over that provided by the individual SNPs rs16969968 or rs588765 (likelihood ratio test P = 0.006 for rs16969968 and 3.83×10−14 for rs588765 in discovery, 0.009 for rs16969968 and 4.62×10−13 for rs588765 in replication, compared with rs588765-rs16969968). Compared with the wild-type homozygous diplotype, CA/CA homozygote exhibited an approximately 2-fold increase risk for lung cancer (OR = 2.12; 95% CI 1.46–3.07 in discovery, and OR = 2.01; 95% CI 1.51–2.67 in replication). Even among never-smokers, CA/CA homozygote showed an increased risk of lung cancer with borderline significance in discovery (adjusted OR = 1.75, 95% CI 0.96–3.19) and statistical significance in replication (adjusted OR = 2.10, 95% CI 1.12–3.96), compared with combined genotypes (CG/CG + CG/TG). Accordingly, rs588765-rs16969968 may be a genetic marker to lung cancer risk, even among never-smokers.
Introduction
Lung cancer is the leading cause of cancer death in the United States and the world, which accounts for 13% of all cases and 23% of all deaths from cancer worldwide and, as such represents a major public health problem (1). Although lung cancer is frequently cited as a malignancy attributable solely to environmental exposures, primarily cigarette smoke (2), numerous studies demonstrated that genetic factors play a significant role in lung cancer susceptibility (3). Several genome-wide association studies (GWAS) have identified tens of common genetic variants associated with lung cancer and provided valuable insight into its genetic architecture (4,5). However, the genetic haplotypes that can influence lung cancer susceptibility have not been explored thoroughly and need to be further investigated.
A region associated with lung cancer risk on chromosome 15q25.1, spanning 203kb, contains several genes of interest in cell growth, signaling and metabolism. It contains the nicotinic acetylcholine receptor gene cluster (CHRNA5-CHRNA3-CHRNB4), which is implicated in nicotine addiction, and mediates the synthesis and release of growth factors and signaling in tumor growth and metastasis (6); the porcine proteasome subunit A4 (PSMA4) gene, which plays a role in cancer cell apoptosis and DNA repair (7); IREB2, which is an iron responsive element-binding protein; and hydroxylysine kinase (HYKK or AGPHD1), which catalyzes the GTP-dependent phosphorylation of 5-hydroxy-l-lysine (8). The 15q25.1 locus was identified by GWAS as a region of susceptibility for lung cancer, smoking behavior (9,10) and nicotine addiction (11) in Caucasians (12) and African-Americans (13). Several single nucleotide polymorphisms (SNPs) within this locus have been specifically associated with smoking phenotypes. For instance, rs16969968, whose SNP changes an amino acid from aspartate to asparagine at position 398 (D398N) in the α5 cholinergic nicotinic receptor (CHRNA5) (3), appears to increase the risk for both lung cancer and nicotine dependence (14). rs1051730 (12,15), rs8034191 (12) and rs12914385 (16), which are located in CHRNA3 and are in tight linkage disequilibrium with rs16969968 (17), reportedly increase lung cancer susceptibility and nicotine dependence. The SNP rs588765, affecting CHRNA5 mRNA expression level in brain tissue and lung tissue (18,19), has been reported to have a strong association with nicotine dependence (14).
The causal haplotypes that affect lung cancer risk have not been identified, even within the chromosome 15q25.1 locus. Two studies determined the haplotype block structure across this locus and found that some haplotypes had significant associations with lung cancer risk, but no specific haplotype was more significantly associated with lung cancer risk than individual SNPs within the haplotype (12,13). Therefore, our objective in this study was to conduct an analysis in which the haplotypes were determined by sliding windows within an extended region of the chromosome 15q25.1 locus, whole region of chromosome 15. We then performed pair-wise haplotype analysis based on the reported SNPs, including functional SNPs or those that presented the most consistent significant associations with lung cancer. In addition, we further explored the role of the selected haplotype with 1815 lung cancer cases and 1959 healthy controls of Italian origin, and replicated the exploration with 2983 lung cases and 3553 controls of European ancestry, aiming to assess and validate whether it influences lung cancer risk and whether it cooperates with smoking in the development of lung cancer. Such work may potentially facilitate our understanding of lung cancer etiology and identify a particularly high risk subset.
Materials and methods
Study participants
In the discovery phase, the case–control comparison was composed of 1815 lung cancer patients and 1959 healthy controls, which were taken from the Environment And Genetics in Lung cancer Etiology (EAGLE) study (20). Between 2002 and 2005, EAGLE study participants were recruited in Italy for a population-based case–control study, which included incident primary lung cancer cases of any histologic type and healthy population-based controls, matched by gender, residence and 5-year age classes. All participants in EAGLE study were Italian nationality and born in Italy. In the replication phase, 2983 lung cancer cases and 3553 healthy controls from two case–control studies, M.D. Anderson Cancer Center (MDACC) study (12) and International Agency for Research on Cancer (IARC) study (21), were used for analysis. The MDACC study participants were recruited at the University of Texas MD Anderson Cancer Center between 1997 and 2007, which included 1154 primary lung cancer patients with adenocarcinoma and squamous cell carcinoma and smokers (former and current), and 1136 healthy smokers, which were matched to patient participants by smoking behavior, ethnicity and 5-year age classes. The IARC study was a multicenter trial from six countries within central Europe, and included participants which recruited newly diagnosed lung cancer patients of any histologic type and healthy control individuals without diagnosed cancers or family history of cancers, matched to cases by sex, age and center or region within European countries. The current case–control comparison included 1829 patients and 2417 controls from available data from the IARC study. All participants used in replication phase were European ancestry. In both the discovery and replication phases, participants who lacked smoking information were excluded. Informed consent forms were signed by all participants in the study. Human participant approval was obtained from by the Institutional Review Board of each participating hospital, center and university in Europe and the United States, and by the National Cancer Institute, Bethesda, MD.
Genotyping
To genotype EAGLE samples at the Center for Inherited Disease Research, we used Illumina HumanHap550v3_B Bead Chips (Illumina, San Diego, CA), which were part of the Gene Environment Association Studies Initiative (GENEVA) funded through the National Human Genome Research Institute (21). Illumina 300K HumanHap v1.1 was used to genotype MDACC samples (21). IARC samples were genotyped using either Illumina 317k or 370Duo arrays (21).
SNP imputation
We imputed EAGLE data with MaCH v1.0 software, imputed MDACC data with Minimac v2 software, and imputed IARC data with IMPUTE v2 software; we imputed all scans over 10 million SNPs using data from the 1000 Genomes Project (phase 1 integrated release 3, March 2012) as reference. Genotypes were aligned to the positive strand in both imputation and genotyping. Thresholds for imputation quality were set to retain both potential common and rare variants for validation. Specifically, poorly imputed SNPs defined by an R-square < 0.80 were excluded from the analyses in the discovery phase.
Haplotype selection
We selected for the most significant haplotypes in the discovery phase. Both imputed and directly genotyped SNPs were studied for haplotype analyses and selected for additional studies. We imputed haplotypes based on multimarker predictors using the standard E-M algorithm and performed association tests based on the distribution of probabilistically inferred sets of haplotypes for each individual with Plink 1.9. Haplotypes were analyzed using the sliding window approach with haplotype windows of 2–26 SNPs in chromosome 15. Omnibus P values were calculated for each haplotype window to evaluate whether the distribution of haplotypes was significantly different between cases and controls with Plink 1.9. Golden Helix 8.1.5 was used to verify the omnibus P values for the ten most significant haplotype. In addition, we selected the five SNPs from chromosome 15q25.1, HYKK rs8034191, CHRNA5 rs588765, CHRNA5 rs16969968, CHRNA3 rs1051730 and CHRNA3 rs12914385, based on published reports (3,12,14–16,18), which could alter the gene function or that were shown to have consistent statistically significant associations with lung cancer risk, to conduct the pair-wise haplotype analysis. Plink 1.9 was used to perform the standard case–control association analysis for the five SNPs. The pair-wise linkage disequilibrium was estimated by Plink 1.9. Haplotypes with frequencies < 0.01 among both patients and controls were omitted in each analysis. Consequently, six healthy controls and one lung cancer case were dropped for the further analysis in discovery set. Comparing the omnibus likelihood ratios, we selected the most statistically significant haplotype as the associated haplotype to perform the further analysis.
Demographic characteristics
Descriptive statistical analyses were used to characterize the discovery and replication study. The differences in the distributions of select variables between the patients and controls were evaluated using a χ2 test, implemented in Statistical Analysis System software (SAS) version 9.1. Principal component analysis was performed based on GWAS data with the EIGENSTRAT program [26] for discovery and replication set, respectively. Two significant principal components (PCs) for discovery and three significant PCs for replication were included in further regression models to adjust for population substructure.
Association analyses
We performed tests for Hardy–Weinberg equilibrium and estimated the association between the risk of lung cancer and genetic variants, including the selected haplotypes and the individual SNPs within the selected haplotype, respectively, by computing the odds ratios (ORs) and 95% confidence intervals (CIs) in univariate and multivariate logistic regression analyses. We further stratified the association of the selected haplotypes and lung cancer risk by smoking status (never, former and current). The interaction effect of the selected haplotype and smoking status on the development of lung cancer was analyzed by univariate and multivariable logistic regression analyses. In multivariable logistic regression analyses, the associations were adjusted for age, gender and significant PCs and/or adjusted for the aforementioned variables and smoking status (never, former, current) and smoking pack-years (0, 1–15, 16–30, 31–50, >50). In addition, we modeled an interaction between smoking status (never = 1, former = 2 and current = 3) by haplotype risk (CG/CG = 1, CG/TG = 2, TG/TG = 3, TG/CA = 4, CG/CA = 5 and CA/CA = 6). We also modeled interaction between smoking status (never = 1 and ever = 2) by number of CA diplotype (CG/CG, CG/TG and TG/TG = 1; TG/CA and CG/CA = 2; and CA/CA = 3). Statistical analyses were performed with SAS 9.1 in both discovery and replication phase; a P value of less than 0.05 was considered to be significant.
Results
In the discovery phase, 188903 imputed SNPs passed quality control steps and were retained for further haplotype analyses. For the discovery analysis, we performed a sliding window haplotype analysis within chromosome 15, and evaluated 4722250 haplotypes. We found that the haplotype, including 10 SNPs, rs76639159-rs116990850-15:78811129-15:78820964-rs72738786-rs78262427-15:78836022-15:78836751-15:78840152-rs113352275 had the most significant association with lung cancer risk (Omnibus P = 3.4×10−14). We also evaluated pair-wise haplotype analysis among previously identified but noncontiguous SNPs rs8034191, rs588765, rs16969968, rs1051730 and rs12914385, and found that rs588765-rs16969968 haplotype was most significantly associated with lung cancer risk (omnibus P = 8.35×10 −15). In addition, to examine whether rs588765-rs16969968 haplotype improved the prediction of case status over that provided by effects from the individual SNPs rs16969968 or rs588765, we compared the joint model and the model when each SNP is analyzed alone, respectively. We identified that the rs588765-rs16969968 haplotype was more significantly associated with lung cancer risk than either individual SNPs in the discovery phase (likelihood ratio test P = 0.006 for rs16969968 and 3.83×10−14 for rs588765, compared with rs588765-rs16969968), and validated the result in the replication phase (likelihood ratio test P = 0.009 for rs16969968 and 4.62×10 −13 for rs588765, compared with rs588765-rs16969968). In addition, rs588765-rs16969968 haplotype was shown to be significantly associated with lung cancer risk in replication (omnibus P = 7.26×10−14). Therefore, we selected rs588765-rs16969968 as the most strongly associated casual haplotype for further analysis.
Table 1 summarizes the demographic characteristics of the current population-based case–control study in the discovery phase, which included 1814 lung cancer patients and 1953 healthy controls, and the additional case–control studies in the replication phase, which included 2983 lung cancer patients and 3553 healthy controls. A difference in gender distribution between cases and controls was not significant for the discovery set (P = 0.09), but was significant for replication set (P = 0.04). In both datasets, the distribution of cigarette smoking was significantly different between patients and controls (P < 0.0001 for discovery and replication) with current smokers being overrepresented among cases (Table 1 In addition, heavy smokers with higher smoking pack-years (31–50 pack-years and > 50 pack-years) were present in 34.5 and 35.8% for lung cancer cases, and 17.5 and 8.6% for controls in discovery (P < 0.0001) and similarly more common in 38.52 and 28.70% for replication cases, and 24.35 and 15.00% for controls in replication (P < 0.0001). Older subjects (≥65 years) were significantly more common in cases than in controls (P = 0.01 and 0.005 for discovery and replication, respectively).
Table 1.
Participant characteristics of lung cancer cases and controls
| Variants | Discovery (n = 3767) | P value | Replication (n = 6536) | P value | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Control (n = 1953) | Case (n = 1814) | Control (n = 3553) | Case (n = 2983) | |||||||
| Age (years) | ||||||||||
| 0–64 | 846 | 43.32 | 711 | 39.2 | 0.01 | 2286 | 64.34 | 1819 | 60.98 | 0.005 |
| ≥65 | 1107 | 56.68 | 1103 | 60.80 | 1267 | 35.66 | 1164 | 39.02 | ||
| Gender | ||||||||||
| Male | 1492 | 76.4 | 1428 | 78.72 | 0.09 | 2397 | 67.46 | 2082 | 69.80 | 0.04 |
| Female | 461 | 23.6 | 386 | 21.28 | 1156 | 32.54 | 901 | 30.20 | ||
| Smoking status | ||||||||||
| Never | 628 | 32.16 | 138 | 7.61 | <0.0001 | 865 | 24.35 | 136 | 4.56 | <0.0001 |
| Former | 840 | 43.01 | 774 | 42.67 | 1279 | 36.00 | 964 | 32.32 | ||
| Current | 485 | 24.83 | 902 | 49.72 | 1409 | 39.66 | 1883 | 63.12 | ||
| Smoking pack-years | ||||||||||
| 0 | 628 | 32.16 | 138 | 7.61 | <0.0001 | 866 | 24.37 | 136 | 4.56 | <0.0001 |
| 1–15 | 462 | 23.66 | 123 | 6.78 | 541 | 15.23 | 241 | 8.08 | ||
| 16–30 | 353 | 18.07 | 278 | 15.33 | 748 | 21.05 | 601 | 20.15 | ||
| 31–50 | 342 | 17.51 | 625 | 34.45 | 865 | 24.35 | 1149 | 38.52 | ||
| >50 | 168 | 8.6 | 650 | 35.83 | 533 | 15.00 | 856 | 28.70 | ||
| Histology | ||||||||||
| Squamous | 466 | 25.69 | 307 | 10.29 | ||||||
| Adenocarcinoma | 753 | 41.51 | 620 | 20.78 | ||||||
| Other | 566 | 31.20 | 226 | 7.58 | ||||||
| Missing | 29 | 1.6 | 1830 | 61.35 | ||||||
Table 2 shows the genotype distribution of rs588765 and rs16969968 among the lung cancer patients and controls in discovery and replication, respectively. In each dataset, the genotype distributions of the two SNPs were in Hardy–Weinberg equilibrium in the control group. rs16969968 showed a significant association with lung cancer risk in discovery (P = 4.92×10−14) and replication (P = 2.91×10−13) phases, and P value for the association between rs588765 and lung cancer risk was 0.006 in discovery and 0.007 in replication. The SNP rs16969968 was significantly associated with lung cancer risk with ORs being 1.45 (95% CI 1.23–1.71) in discovery and 1.22 (95% CI 1.09–1.37) in replication for A/G heterozygote; OR was being 1.87 (95% CI 1.50–2.33) in discovery and 1.72 (95% CI 1.46–2.03) in replication for the A/A homozygote after adjustment for age, gender, smoking status, smoking pack-years and significant PCs. For the rs588765 SNP, compared with individuals with C/C homozygote, those with C/T heterozygote had an decreased risk of lung cancer with borderline significance in discovery (adjusted OR = 0.86, 95% CI, 0.74–1.01) and with statistical significance in replication (adjusted OR = 0.85, 95% CI, 0.76–0.96). We also performed a stratified analyses by smoking status with adjustment for age, gender, smoking pack-years (among smokers) and significant PCs, and found that among never-smokers, individuals with rs16969968 A/A homozygote had an increased risk of lung cancer with borderline significance in the discovery (adjusted OR = 1.70, 95% CI 0.98–3.01, P = 0.06) and non-significance in the replication (adjusted OR = 1.58, 95% CI 0.91–2.73, P = 0.11), compared to those with rs16969968 A/A homozygote. Also, rs588765 showed no significant association with lung cancer risk among never-smokers (Supplementary Table 1, available at Carcinogenesis Online).
Table 2.
Associations between polymorphisms and risk of lung cancer
| SNP | Control | Case | Crude | Adjusteda | ||||
|---|---|---|---|---|---|---|---|---|
| Allele | N | N | % | OR (95% CI) | P | OR (95% CI) | P | |
| Discovery | ||||||||
| rs588765 | C/C | 789 | 827 | 45.59 | 1 | 1 | ||
| C/T | 919 | 771 | 42.5 | 0.80 (0.70–0.92) | 0.001 | 0.86 (0.74–1.01) | 0.06 | |
| T/T | 245 | 216 | 11.91 | 0.84 (0.68–1.04) | 0.10 | 0.91 (0.72–1.16) | 0.47 | |
| Trend test | 0.01 | 0.16 | ||||||
| rs16969968 | G/G | 750 | 514 | 28.34 | 1 | 1 | ||
| A/G | 917 | 904 | 49.83 | 1.44 (1.24–1.66) | <0.0001 | 1.45 (1.23–1.71) | <0.0001 | |
| A/A | 286 | 396 | 21.83 | 2.02 (1.67–2.44) | <0.0001 | 1.87 (1.50–2.33) | <0.0001 | |
| Trend test | <0.0001 | <0.0001 | ||||||
| Replication | ||||||||
| rs588765 | C/C | 1154 | 1070 | 35.87 | 1 | 1 | ||
| C/T | 1753 | 1414 | 47.40 | 0.87 (0.78–0.97) | 0.01 | 0.85 (0.76–0.96) | 0.01 | |
| T/T | 646 | 499 | 16.73 | 0.83 (0.72–0.96) | 0.01 | 0.81 (0.70–0.95) | 0.01 | |
| Trend test | 0.005 | 0.002 | ||||||
| rs16969968 | G/G | 1554 | 1088 | 36.47 | 1 | |||
| A/G | 1599 | 1405 | 47.10 | 1.26 (1.13–1.40) | <0.0001 | 1.22 (1.09–1.37) | 0.0005 | |
| A/A | 400 | 490 | 16.43 | 1.75 (1.50–2.04) | <0.0001 | 1.72 (1.46–2.03) | <0.0001 | |
| Trend test | <0.0001 | <0.0001 | ||||||
aAdjusted for age at diagnosis/interview, sex, smoking status, smoking pack-years and the significant PCs.
The associations of lung cancer risk and genotypes of rs588765-rs16969968 haplotypes are shown in Table 3. The estimated R-square value in the pair-wise linkage disequilibrium measures for rs16969968 and rs588765 was 0.39 (D′ = 0.99) in discovery and 0.41 (D′ = 1.00) in replication, respectively. For the rs588765-rs16969968 haplotype, in both datasets, there are three haplotypes, including CG, TG and CA, and six diplotypes. In discovery phase, we found that the A/A homozygote of rs16969968 were the most likely to develop lung cancer when rs588765 was C/C homozygous, and that G/G homozygote of rs16969968 were the least likely to develop lung cancer when rs588765 was C/C homozygous. In replication, we confirmed this finding. Compared with the CG/CG homozygote, ORs associated with lung cancer for CA/CA homozygote were 2.12 (95% CI 1.46–3.07) in discovery, and 2.01 (95% CI 1.51–2.67) in replication, after adjustment for age, gender, smoking status, smoking pack-years and significant PCs. Because CG/TG had a similar but non-significant, effect on lung cancer risk as CG/CG, we combined the CG/CG genotype with the CG/TG genotype for further stratification analyses.
Table 3.
Association between the rs588765-rs16969968 haplotype and risk of lung cancer
| Haplotype | Control | Case | Crude | Adjusteda | ||||
|---|---|---|---|---|---|---|---|---|
| N | % | N | % | OR (95%CI) | P | OR (95% CI) | P | |
| Discovery | ||||||||
| CG/CG | 122 | 6.25 | 71 | 3.91 | 1 | 1 | ||
| CG/TG | 383 | 19.61 | 227 | 12.51 | 1.02 (0.73–1.42) | 0.91 | 0.96 (0.66–1.40) | 0.83 |
| TG/TG | 245 | 12.54 | 216 | 11.91 | 1.52 (1.07–2.14) | 0.02 | 1.50 (1.01–2.23) | 0.04 |
| TG/CA | 536 | 27.44 | 544 | 29.99 | 1.74 (1.27–2.39) | 0.001 | 1.75 (1.22–2.50) | 0.002 |
| CG/CA | 381 | 19.51 | 360 | 19.85 | 1.62 (1.17–2.25) | 0.004 | 1.49 (1.02–2.17) | 0.04 |
| CA/CA | 286 | 14.64 | 396 | 21.83 | 2.38 (1.71–3.31) | <0.0001 | 2.12 (1.46–3.07) | <0.0001 |
| Replication | ||||||||
| CG/CG | 208 | 5.85 | 115 | 3.86 | 1 | 1 | ||
| CG/TG | 700 | 19.70 | 474 | 15.89 | 1.23 (0.95–1.58) | 0.12 | 1.14 (0.86–0.86) | 0.36 |
| TG/TG | 646 | 18.18 | 499 | 16.73 | 1.40 (1.08–1.81) | 0.01 | 1.26 (0.96–1.66) | 0.09 |
| TG/CA | 1053 | 29.64 | 940 | 31.51 | 1.62 (1.27–2.06) | 0.0001 | 1.43 (1.10–1.85) | 0.01 |
| CG/CA | 546 | 15.37 | 465 | 15.59 | 1.54 (1.19–2.00) | 0.001 | 1.40 (1.07–1.85) | 0.01 |
| CA/CA | 400 | 11.26 | 490 | 16.43 | 2.22 (1.70–2.88) | <0.0001 | 2.01 (1.51–2.67) | <0.0001 |
aAdjusted for age at diagnosis/interview, sex, smoking status and smoking pack-years, the significant PCs.
When we performed the stratified analyses by smoking status adjusting for age, gender, smoking pack-years and significant PCs, we found that even among never-smokers, individuals with CA/CA homozygous diplotype had an increased risk of lung cancer with borderline significance in the discovery (adjusted OR = 1.75, 95% CI 0.96–3.19) and statistical significance in the replication (adjusted OR = 2.10, 95% CI 1.12–3.96), compared with those with combined genotypes (CG/CG + CG/TG). On the other hand, risk associated with rs588765-rs16969968 overlapped among never, former and current smokers for TG/TG homozygote, TG/CA heterozygote, CG/CA heterozygote and CA/CA homozygote, respectively, compared with the combined genotypes (CG/CG + CG/TG) in discovery. This result was also verified in replication (Table 4).
Table 4.
Association between haplotype and risk of lung cancer in never, former and current smokers
| Smoking | Haplotype | Control | Case | Crude | Adjusteda | ||||
|---|---|---|---|---|---|---|---|---|---|
| N | % | N | % | OR (95% CI) | P | OR (95%CI) | P | ||
| Discovery | |||||||||
| No | CG/CG + CG/TG | 150 | 23.89 | 29 | 21.01 | 1 | 1 | ||
| TG/TG | 76 | 12.10 | 15 | 10.87 | 1.02 (0.52–2.02) | 0.95 | 1.01 (0.50–2.05) | 0.97 | |
| TG/CA | 181 | 28.82 | 42 | 30.43 | 1.20 (0.71–2.02) | 0.49 | 1.23 (0.71–2.13) | 0.47 | |
| CG/CA | 124 | 19.75 | 23 | 16.67 | 0.96 (0.53–1.74) | 0.89 | 1.00 (0.53–1.89) | 0.99 | |
| CA/CA | 97 | 15.45 | 29 | 21.01 | 1.55 (0.87–2.75) | 0.14 | 1.75 (0.96–3.19) | 0.07 | |
| Former | CG/CG + CG/TG | 225 | 26.79 | 141 | 18.22 | 1 | 1 | ||
| TG/TG | 107 | 12.74 | 97 | 12.53 | 1.45 (1.02–2.05) | 0.04 | 1.66 (1.13–2.44) | 0.01 | |
| TG/CA | 230 | 27.38 | 233 | 30.10 | 1.62 (1.22–2.14) | 0.001 | 1.76 (1.29–2.39) | 0.0003 | |
| CG/CA | 150 | 17.86 | 142 | 18.35 | 1.51 (1.11–2.06) | 0.01 | 1.41 (0.99–2.00) | 0.05 | |
| CA/CA | 128 | 15.24 | 161 | 20.80 | 2.01 (1.47–2.75) | <0.0001 | 1.78 (1.25–2.53) | 0.001 | |
| Current | CG/CG + CG/TG | 130 | 26.80 | 128 | 14.19 | 1 | 1 | ||
| TG/TG | 62 | 12.78 | 104 | 11.53 | 1.70 (1.14–2.54) | 0.01 | 1.65 (1.06–2.58) | 0.03 | |
| TG/CA | 125 | 25.77 | 269 | 29.82 | 2.19 (1.58–3.02) | <0.0001 | 2.10 (1.48–2.97) | <0.0001 | |
| CG/CA | 107 | 22.06 | 195 | 21.62 | 1.85 (1.32–2.60) | 0.0004 | 1.86 (1.27–2.73) | 0.001 | |
| CA/CA | 61 | 12.58 | 206 | 22.84 | 3.43 (2.36–4.99) | <0.0001 | 3.16 (2.12–4.72) | <0.0001 | |
| Replication | |||||||||
| No | CG/CG + CG/TG | 241 | 27.86 | 29 | 21.32 | 1 | 1 | ||
| TG/TG | 144 | 16.65 | 32 | 23.53 | 1.85 (1.07–3.18) | 0.03 | 1.85 (1.06–3.24) | 0.03 | |
| TG/CA | 248 | 28.67 | 36 | 26.47 | 1.21 (0.72–2.03) | 0.48 | 1.29 (0.76–2.21) | 0.35 | |
| CG/CA | 138 | 15.95 | 15 | 11.03 | 0.90 (0.47–1.74) | 0.76 | 0.96 (0.49–1.87) | 0.89 | |
| CA/CA | 94 | 10.87 | 24 | 17.65 | 2.12 (1.18–3.83) | 0.01 | 2.10 (1.12–3.96) | 0.02 | |
| Former | CG/CG + CG/TG | 327 | 25.57 | 192 | 19.92 | 1 | 1 | ||
| TG/TG | 236 | 18.45 | 165 | 17.12 | 1.19 (0.91–1.56) | 0.20 | 1.16 (0.88–1.54) | 0.29 | |
| TG/CA | 391 | 30.57 | 297 | 30.81 | 1.29 (1.02–1.63) | 0.03 | 1.32 (1.04–1.68) | 0.02 | |
| CG/CA | 179 | 14.00 | 153 | 15.87 | 1.46 (1.10–1.93) | 0.01 | 1.51 (1.13–2.01) | 0.01 | |
| CA/CA | 146 | 11.42 | 157 | 16.29 | 1.83 (1.37–2.44) | <0.0001 | 1.90 (1.40–2.57) | <0.0001 | |
| Current | CG/CG + CG/TG | 340 | 24.13 | 368 | 19.54 | 1 | 1 | ||
| TG/TG | 266 | 18.88 | 302 | 16.04 | 1.05 (0.84–1.31) | 0.67 | 1.03 (0.82–1.28) | 0.82 | |
| TG/CA | 414 | 29.38 | 607 | 32.24 | 1.36 (1.12–1.64) | 0.002 | 1.30 (1.06–1.58) | 0.01 | |
| CG/CA | 229 | 16.25 | 297 | 15.77 | 1.20 (0.96–1.50) | 0.11 | 1.19 (0.94–1.50) | 0.14 | |
| CA/CA | 160 | 11.36 | 309 | 16.41 | 1.78 (1.40–2.27) | <0.0001 | 1.77 (1.38–2.26) | <0.0001 | |
aAdjusted for age at diagnosis/interview, sex, smoking pack-years and the significant PCs.
Furthermore, we evaluated whether rs588765-rs16969968 haplotype and smoking exposure could serve jointly as risk factors for the etiology of lung cancer. Current smokers who carried CA/CA homozygote diplotype had a greater than 3-fold increased risk for lung cancer (adjusted OR = 3.93, 95% CI 1.32–11.71) in discovery; in replication, compared to non-smokers with the combined genotypes (CG/CG + CG/TG), the smokers with the CA/CA diploytpe were verified to have greater than 6-fold increase in lung cancer risk (adjusted OR = 6.88, 95% CI 3.09–15.34) (Table 5). We also found evidence for interaction between the rs588765-rs16969968 haplotype (coded as 1–6) and smoking exposure (coded as 3 levels) on the lung cancer risk in the discovery data (OR = 1.10, 95% CI 1.02–1.17, P = 0.009), but the interaction was not significance in replication data (OR = 1.01, 95% CI 0.96–1.06, P = 0.74). When the haplotype was coded as 3 levels (0, 1 and 2 CA diplotypes) and smoking exposure coded as 2 levels (never and ever), interaction between the haplotype and smoking exposure on the lung cancer risk was non-significant in discovery (OR = 1.24, 95% CI 0.91–1.61, P = 0.18) and in replication (OR = 1.14, 95% CI 0.86–1.49, P = 0.37).
Table 5.
Interaction of haplotypes and smoking status in the risk of lung cancer
| Haplotype | Control | Case | Crude | Adjusteda | |||||
|---|---|---|---|---|---|---|---|---|---|
| Smoking | N | % | N | % | OR (95% CI) | P | OR (95% CI) | P | |
| Discovery | |||||||||
| CG/CG + CG/TG | No | 150 | 7.68 | 29 | 1.60 | 1 | 1 | ||
| Former | 225 | 11.52 | 141 | 7.77 | 3.24 (2.07–5.08) | <0.0001 | 0.65 (0.31–1.37) | 0.25 | |
| Current | 130 | 6.66 | 128 | 7.06 | 5.09 (3.19–8.12) | <0.0001 | 0.36 (0.13–0.99) | 0.05 | |
| TG/TG | No | 76 | 3.89 | 15 | 0.83 | 1.02 (0.52–2.02) | 0.95 | 1.01 (0.50–2.05) | 0.97 |
| Former | 107 | 5.48 | 97 | 5.35 | 4.69 (2.89–7.60) | <0.0001 | 0.77 (0.32–1.84) | 0.56 | |
| Current | 62 | 3.17 | 104 | 5.73 | 8.68 (5.23–14.40) | <0.0001 | 0.50 (0.12–1.98) | 0.32 | |
| TG/CA | No | 181 | 9.27 | 42 | 2.32 | 1.20 (0.71–2.02) | 0.49 | 1.23 (0.71–2.13) | 0.47 |
| Former | 230 | 11.78 | 233 | 12.84 | 5.24 (3.38–8.11) | <0.0001 | 0.88 (0.45–1.73) | 0.70 | |
| Current | 125 | 6.40 | 269 | 14.83 | 11.13 (7.09–17.46) | <0.0001 | 1.48 (0.64–3.40) | 0.36 | |
| CG/CA | No | 124 | 6.35 | 23 | 1.27 | 0.96 (0.53–1.74) | 0.89 | 1.00 (0.53–1.89) | 0.99 |
| Former | 150 | 7.68 | 142 | 7.83 | 4.90 (3.09–7.75) | <0.0001 | 1.08 (0.47–2.48) | 0.86 | |
| Current | 107 | 5.48 | 195 | 10.75 | 9.43 (5.94–14.96) | <0.0001 | 0.30 (0.10–0.88) | 0.03 | |
| CA/CA | No | 97 | 4.97 | 29 | 1.60 | 1.55 (0.87–2.75) | 0.14 | 1.75 (0.96–3.18) | 0.07 |
| Former | 128 | 6.55 | 161 | 8.88 | 6.51 (4.11–10.31) | <0.0001 | 0.85 (0.36–1.99) | 0.71 | |
| Current | 61 | 3.12 | 206 | 11.36 | 17.46 (10.70–28.49) | <0.0001 | 3.93 (1.32–11.71) | 0.01 | |
| Replication | |||||||||
| CG/CG + CG/TG | No | 241 | 6.78 | 29 | 0.97 | 1 | 1 | ||
| Former | 327 | 9.20 | 192 | 6.44 | 4.88 (3.19–7.46) | <0.0001 | 2.23 (1.18–4.22) | 0.01 | |
| Current | 340 | 9.57 | 368 | 12.34 | 8.99 (5.96–13.59) | <0.0001 | 3.61 (1.89–6.89) | <0.0001 | |
| TG/TG | No | 144 | 4.05 | 32 | 1.07 | 1.85 (1.07–3.18) | 0.03 | 1.85 (1.06–3.24) | 0.03 |
| Former | 236 | 6.64 | 165 | 5.53 | 5.81 (3.77–8.96) | <0.0001 | 2.67 (1.31–5.44) | 0.01 | |
| Current | 266 | 7.49 | 302 | 10.12 | 9.44 (6.21–14.35) | <0.0001 | 2.55 (1.22–5.36) | 0.01 | |
| TG/CA | No | 248 | 6.98 | 36 | 1.21 | 1.21 (0.72–2.03) | 0.48 | 1.29 (0.76–2.21) | 0.35 |
| Former | 391 | 11.00 | 297 | 9.96 | 6.31 (4.17–9.55) | <0.0001 | 3.19 (1.81–5.64) | <0.0001 | |
| Current | 414 | 11.65 | 607 | 20.35 | 12.18 (8.18–18.27) | <0.0001 | 3.28 (1.80–5.96) | 0.0001 | |
| CG/CA | No | 138 | 3.88 | 15 | 0.50 | 0.90 (0.47–1.74) | 0.76 | 0.96 (0.49–1.87) | 0.89 |
| Former | 179 | 5.04 | 153 | 5.13 | 7.10 (4.57–11.05) | <0.0001 | 4.82 (2.35–9.88) | <0.0001 | |
| Current | 229 | 6.45 | 297 | 9.96 | 10.78 (7.07–16.44) | <0.0001 | 4.14 (2.02–8.48) | <0.0001 | |
| CA/CA | No | 94 | 2.65 | 24 | 0.80 | 2.12 (1.18–3.83) | 0.01 | 2.10 (1.12–3.96) | 0.02 |
| Former | 146 | 4.11 | 157 | 5.26 | 8.94 (5.72–13.96) | <0.0001 | 3.02 (1.42–6.42) | 0.004 | |
| Current | 160 | 4.50 | 309 | 10.36 | 16.05 (10.44–24.67) | <0.0001 | 6.88 (3.09–15.34) | <0.0001 | |
aAdjusted for age at diagnosis/interview, sex, smoking pack-years and the significant PCs.
We further explored the effect of rs588765-rs16969968 haplotype on the risk for lung adenocarcinoma and lung squamous cell carcinoma (Table 6); we found this haplotype had a significant association with both lung adenocarcinoma and lung squamous carcinoma risk, respectively. Patients with the CA/CA homozygote haplotype showed significant increased risk for lung adenocarcinoma in both the discovery (adjusted OR = 2.30; 95% CI 1.70–3.11) and in replication (adjusted OR = 1.60; 95% CI 1.14–2.25) data, compared with patients exhibiting the combined genotypes (CG/CG + CG/TG). Risk for squamous carcinoma risk was also elevated in CA/CA homozygotes with an approximately 1.7-fold increase risk in discovery (adjusted OR = 1.69; 95% CI 1.16–2.47) and 1.86-fold elevated risk in replication (adjusted OR = 1.86; 95% CI 1.19–2.90) studies.
Table 6.
Association between haplotypes and risk of lung adenocarcinoma and lung squamous cell carcinoma
| Haplotype | Controls | Cases | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| (n = 1953) | Lung adenocarcinoma (n = 753) | Lung squamous cell carcinoma (n = 466) | ||||||||||||
| Crude | Adjusteda | Crude | Adjusteda | |||||||||||
| N | % | N | % | OR (95%CI) | P | OR (95%CI) | P | N | % | OR (95%CI) | P | OR (95%CI) | P | |
| Discovery | ||||||||||||||
| CG/CG + CG/TG | 505 | 25.86 | 114 | 15.14 | 1 | 1 | 84 | 18.03 | 1 | 1 | ||||
| TG/TG | 245 | 12.54 | 88 | 11.69 | 1.59 (1.16–2.19) | 0.004 | 1.64 (1.17–2.30) | 0.004 | 60 | 12.88 | 1.47 (1.02– 2.12) | 0.04 | 1.49 (0.98–2.29) | 0.06 |
| TG/CA | 536 | 27.44 | 235 | 31.21 | 1.94 (1.51–2.51) | <0.0001 | 2.01 (1.53–2.64) | <0.0001 | 131 | 28.11 | 1.47 (1.09– 1.98) | 0.01 | 1.59 (1.13–2.23) | 0.01 |
| CG/CA | 381 | 19.51 | 158 | 20.98 | 1.84 (1.40–2.42) | <0.0001 | 1.85 (1.38–2.49) | <0.0001 | 104 | 22.32 | 1.64 (1.20– 2.25) | 0.002 | 1.57 (1.08–2.28) | 0.02 |
| CA/CA | 286 | 14.64 | 158 | 20.98 | 2.45 (1.85–3.24) | <0.0001 | 2.30 (1.70–3.11) | <0.0001 | 87 | 18.67 | 1.83 (1.31– 2.55) | 0.0004 | 1.69 (1.16–2.47) | 0.01 |
| Replication | ||||||||||||||
| CG/CG + CG/TG | 908 | 25.56 | 134 | 21.61 | 1 | 1 | 61 | 19.87 | 1 | 1 | ||||
| TG/TG | 646 | 18.18 | 100 | 16.13 | 1.05 (0.79–1.39) | 0.74 | 0.95 (0.69–1.30) | 0.74 | 57 | 18.57 | 1.31 (0.90– 1.91) | 0.15 | 1.23 (0.81–1.86) | 0.33 |
| TG/CA | 1053 | 29.64 | 194 | 31.29 | 1.25 (0.99–1.58) | 0.07 | 1.32 (1.01–1.74) | 0.04 | 103 | 33.55 | 1.46 (1.05– 2.02) | 0.03 | 1.50 (1.04–2.17) | 0.03 |
| CG/CA | 546 | 15.37 | 100 | 16.13 | 1.24 (0.94–1.64) | 0.13 | 1.35 (0.98–1.85) | 0.06 | 36 | 11.73 | 0.98 (0.64– 1.50) | 0.93 | 0.99 (0.62–1.57) | 0.95 |
| CA/CA | 400 | 11.26 | 92 | 14.84 | 1.56 (1.17–2.08) | 0.003 | 1.60 (1.14–2.25) | 0.01 | 50 | 16.29 | 1.86 (1.26– 2.75) | 0.002 | 1.86 (1.19–2.90) | 0.01 |
aAdjusted for age at diagnosis/interview, sex, smoking status, smoking pack-years, and the significant PCs.
Discussion
Several GWAS identified a few common genetic variants associated with lung cancer, especially in the chromosome 15q25.1 locus, which provided valuable insight into its genetic architecture. However, the possible casual haplotypes that commonly influence lung cancer susceptibility have not yet been established and needed to be investigated. In the discovery phase, we conducted a large scale, population-based case–control study, performed haplotype analysis with the sliding windows approach within the whole chromosome 15 region, and analyzed pair-wise haplotypes based on the reported SNPs, including rs8034191, rs588765, rs16969968, rs1051730 and rs12914385. We found that rs588765-rs16969968 was the most statistically significant haplotype affecting lung cancer susceptibility. Individuals with CA/CA diplotype of rs588765-rs16969968 had a significant 2-fold elevated lung cancer risk, compared with the individuals with the CG/CG diplotype. In addition, the rs588765-rs16969968 haplotype was more strongly associated with lung cancer risk than each SNP analyzed alone, and the differences reached statistical significance. In replication phase, we validated the role of the rs588765-rs16969968 haplotype in lung cancer risk. A possible explanation for this is that the two distinct biological mechanisms that affect lung cancer risk, including altering the receptor function with changing an amino acid in CHRNA5 by rs16969968 (3,18) and modifying the CHRNA5 mRNA expression level by rs588765 (18,19), could cooperate and jointly influence the carcinogenic potential for lung cancer.
In several independent GWAS analyses, the non-synonymous CHRNA5 SNP rs16969968 has been associated with lung cancer risk (1,13,14,22,23). The CHRNA5 expression-related SNPs rs588765, which is associated with heavy smoking in individuals of European-ancestry (14) and affects the level of bulky polycyclic aromatic hydrocarbon–DNA adducts (24), has a non-significant association with lung cancer risk (14). One study of the risk of lung cancer and nicotine dependence and two studies of the risk for other diseases have examined this joint effect of rs16969968 and rs588765 in population-based analyses. These studies provided the evidence that the joint model affected risk of nicotine addiction, lung cancer and earlier Parkinson’s disease, respectively, more strongly than individual SNPs, which was in agreement with our finding for lung cancer risk. However, the effect of individual SNP rs588765 and rs16969968 in the joint model was diverse. We discovered and validated that, in the joint model, when rs588765 was C/C homozygote, those individuals with the A/A homozygote haplotype of rs16969968 were the most likely to develop lung cancer and those with the G/G homozygote haplotype of rs16969968 were the least likely to develop lung cancer. Wang et al. (18) explored the interaction of rs16969968 and rs588765 in lung cancer risk with 194 patient with familiar lung cancer and 219 cancer-free control subjects of European descent, and found that, when the expression of CHRNA5 mRNA was low or the individuals carried C/C homozygote of rs588765, the A/G heterozygote of rs16969968 would greatly increase the risk for developing nicotine dependence and lung cancer. Saccone et al. (14) performed a meta-analysis of over 38000 unrelated subjects, and observed that rs16969968 could change the direction of rs588765 on nicotine addiction risk. The contradictory results showing the effect of SNPs rs588765 and rs16969968 in the joint model may be due to the differences in disease type, sample sizes and ethnic groups studied. Therefore, in the future, studies should confirm our findings and further explore the mechanism of the rs588765-rs16969968 haplotype.
Another important finding in our study was that individuals—even in never smokers—with the rs588765-rs16969968 haplotype seemed to be at significant greater risk for lung cancer. This indicates that the rs588765-rs16969968 haplotype, a 15q25.1 variation, might be involved in mechanisms unrelated to cigarette smoke. To our knowledge, this is the first report that 15q25 variation is associated with lung cancer risk in never smokers. Most previous studies explored the association between other 15q25 variations and lung cancer risk and found no association in never smokers (12,25,26). A meta-analysis that included 2405 never-smokers lung cancer patients and 7622 controls demonstrated that the strong association between 15q25 locus and lung cancer risk was restricted to ever smokers and strongly suggested no direct effect of 15q25 on lung cancer risk (25). Our study also verified neither individual SNP, rs16969968 or rs588765, was significantly associated independently with lung cancer risk in never-smokers. However, the diplotype-based analysis identified a significant effect on risk for never-smokers for the homozygous high risk diplotype. In addition, a recent small study found that a CHRNA3 polymorphism might influence lung adenocarcinoma susceptibility in Chinese individuals, particularly in non-smoking females (27). This study also supports the association between the chromosome 15q25.1 locus and lung cancer risk both through smoking and independent of smoking. Therefore, more studies of the 15q25 variants should be performed to explore their role in the etiology of lung cancer in never-smokers and to determine whether or not 15q25.1 locus involves non-smoking related mechanisms for lung cancer.
We found that individuals with the rs588765-rs16969968 haplotype had a significant risk for lung adenocarcinoma. A previous study (19), investigating the six genes within chromosome 15q25.1 locus, reported that individuals with lung adenocarcinoma had 30-fold increased CHRNA5 mRNA, 2-fold downregulated CHRNA3 mRNA and absent CHRNB4 mRNA, compared with individuals with normal lung tissue, and that CHRNA5 rs16969968 could modify lung adenocarcinoma risk. This provided evidence for the association between lung adenocarcinoma and CHRNA5-A3-B4, and suggested potential biological mechanisms of chromosome 15q.25.1 locus involved lung adenocarcinoma development. Moreover, a cohort study, examining 302 surgically treated patients with lung adenocarcinoma, showed an association between rs16969968 and survival (28). These studies support the association between the chromosome 15q25.1 locus and lung adenocarcinoma both through smoking and through mechanisms independent of smoking. We know of no previous studies that have explored the role of 16969968-rs588765 haplotype in the etiology of histology-specific lung cancer. Therefore, our results should be confirmed in additional studies.
Our study had some limitations. The exclusion of subjects with missing phenotype data in discovery phase resulted in mild unbalancing of the 5-year age classes between the lung cancer cases and healthy controls. However, we adjusted our analyses for age at diagnosis/interview and gender. In addition, our findings, in both discovery and replication, confirmed the previous GWAS results on the non-significant association of rs588765 and significant association of rs16969968 with lung cancer risk, which indicated that the effects of any residual confounding demographic factors had minimal effect. Another limitation is that a limited number of individuals were included within each category of the stratified analyses by smoking status and the joint analyses for the effect of the rs588765-rs16969968 haplotype and smoking behavior; thus our results should be confirmed in future in different populations.
In conclusion, our study identified and replicated the finding that individuals with the rs588765-rs16969968 haplotype have a significantly greater lung cancer risk and our study improved the prediction of case status over that provided by the individual SNPs rs16969968 or rs588765. We also found that the rs588765-rs16969968 haplotype seems to be a significant risk factor for lung cancer in never-smokers and that individuals with this haplotype had an increased risk of lung adenocarcinoma and lung squamous cell carcinoma. To the best of our knowledge, this is the first study to examine the effects of specific haplotyes on chromosome 15, to explore the joint effect of this haplotype and smoking exposure on lung cancer risk, and to identify and validate that a 15q25.1 variation was associated with an increased risk for lung cancer in never smokers. Such work expands our understanding of the etiology of lung cancer and help to identify a particularly high risk subset.
Supplementary material
Supplementary Table 1 can be found at http://carcin.oxfordjournals.org/
Funding
National Institutes of Health (NIH) for the research of lung cancer (grant P30CA023108, P20GM103534 and R01LM012012); Trandisciplinary Research in Cancer of the Lung (TRICL) (grant U19CA148127); UICC American Cancer Society Beginning Investigators Fellowship funded by the Union for International Cancer Control (UICC) (to X.J.).
Conflict of Interest Statement: None declared.
Supplementary Material
Glossary
Abbreviations
- EAGLE
Environment And Genetics in Lung cancer Etiology
- IARC
International Agency for Research on Cancer
- MDACC
M.D. Anderson Cancer Center
- PCs
principal components
- GWAS
genome-wide association studies
- SNP
single nucleotide polymorphism
References
- 1. Siegel R., et al. (2011) Cancer statistics, 2011: the impact of eliminating socioeconomic and racial disparities on premature cancer deaths. CA. Cancer J. Clin., 61, 212–236. [DOI] [PubMed] [Google Scholar]
- 2. Parkin D.M., et al. (1994) At least one in seven cases of cancer is caused by smoking. Global estimates for 1985. Int. J. Cancer, 59, 494–504. [DOI] [PubMed] [Google Scholar]
- 3. Bierut L.J., et al. (2008) Variants in nicotinic receptors and risk for nicotine dependence. Am. J. Psychiatry, 165, 1163–1171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Spitz M.R., et al. (2012) Multistage analysis of variants in the inflammation pathway and lung cancer risk in smokers. Cancer Epidemiol. Biomarkers Prev., 21, 1213–1221. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Dong J., et al. (2012) Association analyses identify multiple new lung cancer susceptibility loci and their interactions with smoking in the Chinese population. Nat. Genet., 44, 895–899. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Singh S., et al. (2011) Nicotinic acetylcholine receptor signaling in tumor growth and metastasis. J. Oncol., 2011, 456743. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Liu Y., et al. (2009) Haplotype and cell proliferation analyses of candidate lung cancer susceptibility genes on chromosome 15q24-25.1. Cancer Res., 69, 7844–7850. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Veiga-da-Cunha M., et al. (2012) Molecular identification of hydroxylysine kinase and of ammoniophospholyases acting on 5-phosphohydroxy-L-lysine and phosphoethanolamine. J. Biol. Chem., 287, 7246–7255. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Liu J.Z., et al. (2010) Meta-analysis and imputation refines the association of 15q25 with smoking quantity. Nat. Genet., 42, 436–440. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Tobacco & Genetics C. (2010) Genome-wide meta-analyses identify multiple loci associated with smoking behavior. Nat. Genet., 42, 441–447. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Thorgeirsson T.E., et al. (2008) A variant associated with nicotine dependence, lung cancer and peripheral arterial disease. Nature, 452, 638–642. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Amos C.I., et al. (2008) Genome-wide association scan of tag SNPs identifies a susceptibility locus for lung cancer at 15q25.1. Nat. Genet., 40, 616–622. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Walsh K.M., et al. (2013) Fine-mapping of the 5p15.33, 6p22.1-p21.31, and 15q25.1 regions identifies functional and histology-specific lung cancer susceptibility loci in African-Americans. Cancer Epidemiol. Biomarkers Prev., 22, 251–260. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Saccone N.L., et al. (2010) Multiple independent loci at chromosome 15q25.1 affect smoking quantity: a meta-analysis and comparison with lung cancer and COPD. PLoS Genet., 6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Wassenaar C.A., et al. (2011) Relationship between CYP2A6 and CHRNA5-CHRNA3-CHRNB4 variation and smoking behaviors and lung cancer risk. J. Natl. Cancer Inst., 103, 1342–1346. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Broderick P., et al. (2009) Deciphering the impact of common genetic variation on lung cancer risk: a genome-wide association study. Cancer Res., 69, 6633–6641. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Falvella F.S., et al. (2010) Promoter polymorphisms and transcript levels of nicotinic receptor CHRNA5. J. Natl. Cancer Inst., 102, 1366–1370. [DOI] [PubMed] [Google Scholar]
- 18. Wang J.C., et al. ; COGEND collaborators and GELCC collaborators. (2009) Risk for nicotine dependence and lung cancer is conferred by mRNA expression levels and amino acid change in CHRNA5. Hum. Mol. Genet., 18, 3125–3135. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Falvella F.S., et al. (2009) Transcription deregulation at the 15q25 locus in association with lung adenocarcinoma risk. Clin. Cancer Res., 15, 1837–1842. [DOI] [PubMed] [Google Scholar]
- 20. Landi M.T., et al. (2008) Environment And Genetics in Lung cancer Etiology (EAGLE) study: an integrative population-based case-control study of lung cancer. BMC Public Health, 8, 203. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Wang Y., et al. (2014) Rare variants of large effect in BRCA2 and CHEK2 affect risk of lung cancer. Nat. Genet., 46, 736–741. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Wei C., et al. (2011) A case-control study of a sex-specific association between a 15q25 variant and lung cancer risk. Cancer Epidemiol. Biomarkers Prev., 20, 2603–2609. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Le Marchand L., et al. (2008) Smokers with the CHRNA lung cancer-associated variants are exposed to higher levels of nicotine equivalents and a carcinogenic tobacco-specific nitrosamine. Cancer Res., 68, 9137–9140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Tekpli X., et al. (2013) Functional effect of polymorphisms in 15q25 locus on CHRNA5 mRNA, bulky DNA adducts and TP53 mutations. Int. J. Cancer, 132, 1811–1820. [DOI] [PubMed] [Google Scholar]
- 25. Wang Y., et al. (2011) Chromosome 15q25 (CHRNA3-CHRNA5) variation impacts indirectly on lung cancer risk. PLoS One, 6, e19085. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Thorgeirsson T.E., et al. (2008) A variant associated with nicotine dependence, lung cancer and peripheral arterial disease. Nature, 452, 638–642. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. He P., et al. (2014) CHRNA3 polymorphism modifies lung adenocarcinoma risk in the Chinese Han population. Int. J. Mol. Sci., 15, 5446–5457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Sasaki H., et al. (2010) CHRNA5 gene D398N polymorphism in Japanese lung adenocarcinoma. J. Surg. Res., 162, 75–78. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.