Lung DOI 10.1007/s00408-014-9579-4
Association of Lung Function Genes with Chronic Obstructive Pulmonary Disease Woo Jin Kim • Myoung Nam Lim • Yoonki Hong • Edwin K. Silverman Ji-Hyun Lee • Bock Hyun Jung • Seung Won Ra • Hye Sook Choi • Young Ju Jung • Yong Bum Park • Myung Jae Park • Sei Won Lee • Jae Seung Lee • Yeon-Mok Oh • Sang Do Lee
•
Received: 30 November 2013 / Accepted: 21 March 2014 Ó Springer Science+Business Media New York 2014
Abstract Background Spirometric measurements of pulmonary function are important in diagnosing and determining the severity of chronic obstructive pulmonary disease (COPD). We performed this study to determine whether candidate genes identified in genome-wide association studies of spirometric measurements were associated with COPD and if they interacted with smoking intensity. Methods The current analysis included 1,000 COPD subjects and 1,000 controls recruited from 24 hospitalbased pulmonary clinics. Thirteen SNPs, chosen based on
genome-wide association studies of spirometric measurements in the Korean population cohorts, were genotyped. Genetic association tests were performed, adjusting for age, sex, and smoking intensity, using models including a SNP-by-smoking interaction term. Results PID1 and FAM13A were significantly associated with COPD susceptibility. There were also significant interactions between SNPs in ACN9 and FAM13A and smoking pack-years, and an association of ACN9 with COPD in the lowest smoking tertile. The risk allele of FAM13A was associated with increased expression of FAM13A in the lung.
W. J. Kim Y. Hong Department of Internal Medicine and Environmental Health Center, Kangwon National University Hospital, School of Medicine, Kangwon National University, Chuncheon, South Korea
H. S. Choi Department of Internal Medicine, Dongguk University Gyeongju Hospital, Dongguk University College of Medicine, Gyeongju, South Korea
M. N. Lim Department of Statistics, Kangwon National University, Chuncheon, South Korea E. K. Silverman Channing Division of Network Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA J.-H. Lee Department of Internal Medicine, CHA Bundang Medical Center, CHA University, Seongnam, South Korea B. H. Jung Department of Pulmonary and Critical Medicine, Gangneung Asan Hospital, University of Ulsan College of Medicine, Gangneung, South Korea S. W. Ra Department of Internal Medicine, Ulsan University Hospital, University of Ulsan College of Medicine, Ulsan, South Korea
Y. J. Jung Health Promotion Center, Asan Medical Center, University of Ulsan College of Medicine, Seoul, South Korea Y. B. Park Department of Pulmonary and Critical Care Medicine, Kangdong Sacred Heart Hospital, Hallym University, Seoul, South Korea M. J. Park Division of Respiratory and Critical Care Medicine, Kyung Hee University Hospital, Kyung Hee University, Seoul, South Korea S. W. Lee J. S. Lee Y.-M. Oh (&) S. D. Lee Department of Pulmonary and Critical Care Medicine, and Clinical Research Center for Chronic Obstructive Airway Diseases, Asan Medical Center, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 138-736, Republic of Korea e-mail: [email protected]
123
Lung
Conclusions We have validated associations of FAM13A and PID1 with COPD. ACN9 showed significant interaction with smoking and is a potential candidate gene for COPD. Significant associations of genetic variants of FAM13A with gene expression levels suggest that the associated loci may act as genetic regulatory elements for FAM13A gene expression. Keywords ACN9 COPD Genetic association Gene-by-environment interaction Cigarette smoking
Introduction The prevalence and burden of chronic obstructive pulmonary disease (COPD) are increasing, and it is expected to become the third leading cause of death worldwide by 2020 [1]. COPD is characterized by irreversible airflow limitation and results from environmental factors, including cigarette smoking, with contributions from genetic background and gene-by-environment interactions [2]. Spirometric measurements of forced expiratory volume in 1 s (FEV1) and FEV1/forced vital capacity (FVC) are important in diagnosing and determining the severity of COPD. Genome-wide association studies (GWAS) have identified many loci influencing FEV1 and/or the FEV1/ FVC ratio in general population samples [3]. Although many of the genes at these loci have unknown functions and have not previously been considered biologically plausible candidates for respiratory disease pathogenesis, these findings can provide insight into the genetic architecture contributing to lung function and may have important roles in the development of the respiratory system or in airway disease pathogenesis. On examining COPD as an outcome in GWAS meta-analysis, genes associated with pulmonary function in the general population showed significant associations with COPD susceptibility [4, 5]. In a genetic association study of spirometric loci as candidate genes, many were shown to be susceptibility genes for COPD [6, 7]. These studies suggest that genomic loci associated with FEV1 or FEV1/FVC may also be associated with COPD susceptibility. Smoking is the most important risk factor for COPD and has adverse effects on lung function. Genetic factors may have heterogeneous effects on pulmonary function, depending on smoking exposure. In a recent GWAS, novel loci were identified as significantly associated with lung function after investigation of SNP-by-smoking interaction [8]. In the present study, lung function candidate genes identified in GWAS of general population cohorts were evaluated for association and smoking interaction in separate COPD and control samples.
123
Methods Study Participants This study included 1,000 individuals with COPD and 1,000 unaffected controls. Individuals with COPD were recruited from 24 hospital-based pulmonary clinics belonging to the KOLD Study Group or the ASANCOPD Network. Controls were recruited from individuals attending the Asan Medical Center, Seoul, Korea, for a medical checkup. All control subjects were current or exsmokers. Participants performed standardized spirometry in accordance with American Thoracic Society/European Respiratory Society criteria [9]. Individuals with COPD demonstrated a post-bronchodilator FEV1/FVC \ 0.7, while controls had normal spirometry. Pack-years of smoking were calculated as smoking amount per day and duration of years smoked. All participants provided written informed consent, and the study protocol was approved by the respective institutional review boards. Selection of Candidate Genes and Genotyping Thirteen SNPs were chosen based on meta-analysis of the results from GWAS of Korean population cohorts, where spirometric measurements had been taken (KARE3 and HTS) [10]. SNPs were genotyped using TaqMan technology (Applied Biosystems, Foster City, CA) on an Applied Biosystems 7900HT real-time PCR system. Statistical Analyses Hardy–Weinberg equilibrium was assessed using goodness-of-fit tests. We performed genetic association testing using logistic regression with COPD affection status, assuming an additive genetic model and adjusting for age, sex, and smoking pack-years using SAS ver. 9.2 (SAS Inc., Cary, NC, USA). In addition to the overall model, we evaluated models that included a SNP-by-pack-years interaction term and a smoking intensity-stratified model. Linear regression analyses were implemented using FEV1 and FEV1/FVC ratio as quantitative phenotypes after adjusting for age, sex, and smoking pack-years. Gene Expression in Lung Tissue We used gene expression profiling data from a previous RNA-seq experiment using lung tissues from 193 participants with or without COPD [11]. For differential expression according to genotype of candidate genes, fragments per kilobase of exon per million fragments mapped (FPKM) levels were analyzed using a t test.
Lung Table 1 Demographic characteristics of analyzed participants with COPD and controls COPD (n = 1,000)
Control (n = 1,000)
p value
Male (%)
977 (97.7)
983 (98.3)
0.42
Age (years)
69.2 ± 7.8
57.2 ± 9.5
\0.0001
Pack-years of smoking
44.9 ± 23.1
34.7 ± 27.8
\0.0001
FEV1 (L)
1.64 ± 0.61
3.41 ± 0.68
\0.0001
FEV1, % of predicted
55.9 ± 18.6
95.9 ± 32.6
\0.0001
FEV1/FVC
48.9 ± 12.1
78.5 ± 22.8
\0.0001
Data are presented as the mean (±SD) unless otherwise stated
Table 2 List of 13 SNPs genotyped for association with COPD, according to pulmonary function genome-wide association meta-analysis
Chr chromosome, MAF minor allele frequency, Ref reference, Alt alternative
SNP
Chr
Position
Gene
Ref allele
Alt allele
MAF
Beta
rs16825116
2
229406685
PID1
G
T
0.16
?
rs2609264
4
89828080
FAM13A
C
T
0.52
–
rs2609261
4
89835485
FAM13A
A
G
0.53
–
rs2609260
4
89836819
FAM13A
C
T
0.56
–
rs10231916
7
96763456
ACN9
A
T
0.51
–
rs10229181
7
96795041
ACN9
C
G
0.51
–
rs7119465
11
25734037
ANO3
A
G
0.76
?
rs8031759
15
93798810
MCTP2
G
T
0.08
–
rs886282
17
55188150
AKAP1
C
T
0.18
–
rs17178251
17
69176879
SOX9
C
G
0.61
?
rs17765644
17
69179492
SOX9
C
T
0.62
?
rs11870732
17
69195241
SOX9
A
G
0.39
–
rs2269145
21
44124432
PDE9A
C
T
0.53
?
The baseline characteristics of the study participants (1,000 with COPD and 1,000 controls) are presented in Table 1. Individuals with COPD were older and had higher packyears of cigarette smoking than controls. As expected, individuals with COPD had poorer lung function than controls. Therefore, we adjusted for age and smoking intensity in our genetic association analyses.
(p = 0.0015) and rs10229181 (p = 0.002), showed modest deviations from Hardy–Weinberg equilibrium in control subjects but not in the entire population; these SNPs were retained in the analysis. A list of SNPs genotyped and the direction of the association with lung function from prior analysis in a Korean general population cohort are shown in Table 2. One SNP (rs16825116) near PID1 and three in FAM13A (rs2609264, rs2609261, and rs2609260) showed significant associations with COPD, after adjustment for age, sex, and pack-years of smoking (Table 3). The ‘T’ allele of rs16825116, which was associated with increased lung function, was protective against COPD. The three SNPs in FAM13A are in strong linkage disequilibrium (LD) with one another (r2 = 0.8–0.93). The most significant SNP (rs2609264) is intronic and the direction of the odds ratio in FAM13A was consistent with the effect observed on lung function. The rs2609264 ‘T’ allele had a negative beta coefficient for association with FEV1 and an increased risk for COPD (OR = 1.407, Table 3).
Genetic Association Tests
Smoking Interaction and Stratified Analyses
Eleven of the SNPs were in Hardy–Weinberg equilibrium in control subjects. Two of the SNPs, rs10231916
After inclusion of a SNP-by-pack-years interaction term, there were significant interactions between SNPs in ACN9,
ACN9, PID1, and FAM13A were genotyped in 93 participants who had available blood samples, regardless of disease status, to investigate whether variants at these loci influence gene expression. Linear regression was performed to investigate associations between expression level and genotype.
Results Demographic Characteristics
123
Lung Table 3 Results of genetic association analyses with COPD susceptibility and interaction with smoking pack-years
SNP
rs16825116
rs2609264
rs2609261
rs2609260
rs10231916
rs10229181
rs7119465
Chr
Gene
2
PID1
4
4
4
7
7
11
FAM13A
FAM13A
FAM13A
ACN9
ACN9
ANO3
rs8031759
15
MCTP2
rs886282
17
AKAP1
rs17178251
rs17765644
rs11870732
rs2269145
17
17
17
21
SOX9
SOX9
SOX9
PDE9A
pint is the p value of interaction term with smoking intensity
FAM13A, MCTP2, and AKAP1 and smoking pack-years (Table 3). When the study population was stratified into three COPD groups based on pack-years with all controls as the reference group, there was a significant association of ACN9 with COPD in only the lowest smoking tertile subgroup (OR = 1.29, p = 0.028), and SNPs in FAM13A showed significant associations with COPD in all
123
COPD
Control
OR 95 % CI pSNP
pint
GG
727
672
0.780
0.18
GT
254
294
(0.632–0.962)
TT
18
33
0.02
CC
170
239
1.407
CT
512
517
(1.199–1.650)
TT
308
241
2.8 9 10-5
AA AG
160 511
229 508
1.329 (1.133–1.558)
GG
317
261
4.7 9 10-4
CC
139
203
1.327
CT
509
514
(1.130–1.559)
TT
345
282
5.8 9 10-4
AA
271
248
1.043
AT
494
549
(0.893–1.219)
TT
232
203
0.60
CC
265
245
1.064
CG
479
544
(0.910–1.243)
GG
236
204
0.44
AA
58
71
0.932
AG
371
366
(0.779–1.116)
GG
549
561
0.44
GG GT
845 145
842 148
0.967 (0.728–1.283)
TT
5
9
CC
665
711
CT
286
267
TT
29
18
0.51
CC
378
395
1.013
CG
498
471
(0.861–1.192)
3.4 9 10-4
4.0 9 10-5
2.3 9 10-4
1.3 9 10-8
8.1 9 10-9
0.39
3.1 9 10-6
0.82 1.074
1.5 9 10-6
(0.868–1.328)
GG
121
134
0.88
CC
121
133
0.984
CT
497
471
(0.837–1.158)
TT
377
395
0.85
AA
376
392
0.998
AG
492
472
(0.848–1.174)
GG
122
135
0.98
CC
209
208
0.917
CT TT
502 285
469 321
(0.786–1.069) 0.27
0.11
0.12
0.19
0.84
subgroups, whereas MCTP2 and AKAP1 did not demonstrate any associations (Table 4). Association with Lung Function Linear regression analyses revealed that SNPs in ACN9 and FAM13A were significantly associated with FEV1/FVC
Lung Table 4 Results of stratified genetic association analyses according to smoking intensity
Values are odds ratio, (confidential interval), and p values
SNP
Gene
Tertile 1
rs2609264
FAM13A
1.48 (1.18–1.85), 0.0006
1.38 (1.12–1.72), 0.0032
1.61 (1.24–2.10), 0.0003
rs2609261
FAM13A
1.39 (1.11–1.73), 0.004
1.30 (1.05–1.60), 0.018
1.54 (1.19–2.00), 0.0011
rs2609260
FAM13A
1.38 (1.10–1.73), 0.0046
1.36 (1.10–1.70), 0.0055
1.42 (1.09–1.84), 0.0091
rs10231916
ACN9
1.25 (1.00–1.56), 0.051
0.97 (0.78–1.20), 0.77
0.82 (0.64–1.06), 0.13
rs10229181
ACN9
1.29 (1.03–1.61), 0.028
1.00 (0.81–1.24), 0.98
0.82 (0.64–1.06), 0.13
rs8031759
MCTP2
1.15 (0.79–1.65), 0.47
0.71 (0.47–1.08), 0.11
0.89 (0.55–1.44), 0.63
rs886282
AKAP1
1.04 (0.76–1.41), 0.82
1.15 (0.87–1.52), 0.34
1.11 (0.79–1.55), 0.56
across all study participants (Table 5). The ‘T’ allele of rs10231916 in ACN9 was associated with lower FEV1/FVC values, consistent with results from the general population, and the association with SNPs in FAM13A was also in the same direction as that observed in the population study. SNPs in PID1 and FAM13A also showed significant associations with FEV1, with the ‘G’ allele of PID1 associated with lower FEV1 values, consistent with results of GWAS meta-analysis. Significant associations of PID1 with FEV1 and FAM13A with FEV1/FVC were observed only among participants with COPD (Table 5). Expression in RNA-seq Three genes (ACN9, PID1, and FAM13A) were evaluated for mRNA expression in lung tissue samples using data from a previous experiment [11]. RNA-seq analyses demonstrated that lung tissue from COPD subjects had significantly decreased expression of ACN9 (p = 2.6 9 10-8). The ‘T’ allele of rs10231916 (ACN9) showed a tendency toward lower expression in 93 samples (p = 0.07, Fig. 1). PID1 and FAM13A showed increased expression in lung tissue from individuals with COPD (p = 1.5 9 10-4 and p = 1.8 9 10-11, respectively). The risk allele of rs2609261 was associated with increased expression of FAM13A in the lung (p = 0.0009, Fig. 1). SNPs at the PID1 locus did not show significant association with gene expression.
Discussion In this study, we demonstrated that SNPs in FAM13A and PID1 are significantly associated with COPD in the Korean population. Although SNPs in ACN9 were not associated with COPD in the analysis of main genetic effects, a significant interaction with smoking levels among the COPD subgroup who smoked the least was observed. In the current study, rs11435116 near PID1 was associated with COPD and FEV1 in the Korean population. In a previous GWAS, another SNP flanking PID1, rs11435867, was associated with the FEV1/FVC ratio in the European
Tertile 2
Tertile 3
general population [12]. In another report, the association of PID1 with lung function was replicated in individuals with asthma [13], and PID1 has previously been associated with COPD susceptibility, albeit with a nominal p value [6]. The rs16825116 SNP is located at chromosome 2q36 between SPHKAP and PID1, 100 kb from rs11435867, which has previously been reported as associated with lung function in the European population. PID1 showed increased expression in COPD lung tissues in our analyses, while SPHKAP gene expression levels did not differ between the two groups. PID1 is involved in tissue homeostasis, cell growth, and adiposity [14]. Overexpression of PID1 leads to mitochondrial dysfunction in adipocytes [15], and overexpression in rat skeletal muscle results in inhibition of insulin-stimulated glucose transport [16]. Although the PID1 SNP significantly associated with COPD in our study did not appear to influence gene expression, overexpression of PID1 may have a role in COPD pathogenesis through mitochondrial dysfunction and excess reactive oxygen species [17]. In a previous GWAS that used populations of European descent, FAM13A was significantly associated with COPD susceptibility [18] and with lung function in the general population [12]. This region also showed a genome-wide significant association with idiopathic pulmonary fibrosis [19]. In the current study, SNPs in FAM13A were associated with COPD susceptibility and lung function in individuals with COPD and among the entire study population. The directions of association were consistent with our previous analysis in the general population and showed significant interaction with all subgroups of smoking intensity. The three SNPs in FAM13A identified in the current study are intronic. The most significant, rs2609264, is located 10 kb from rs2869967 (r2 = 0.73) and 50 kb from rs1903003 (r2 = 0.53), which were previously reported as associated with lung function or COPD in European populations. These results provide compelling evidence that the FAM13A region influences COPD susceptibility. Differences in LD between the Korean and European populations may assist in localizing functional genetic variants. Interestingly, tight LD between rs2609264 and rs2609255
123
Lung Table 5 Association between SNPs and lung function in all participants and those with COPD
SNP
rs16825116 rs2609264
All participants (n = 2,000)
COPD only (n = 1,000)
FEV1
FEV1
0.087 (0.036)
1.022 (0.942)
0.082 (0.037)
0.02
0.28
0.03
-0.09 (0.027) 0.002
rs2609261
-0.070 (0.027) 0.01
rs2609260
-0.065 (0.028)
rs10231916
-0.028 (0.027)
0.02 0.30
\0.0001 -3.193 (0.709) \0.0001 -3.209 (0.717) \0.0001 -1.460 (0.712) 0.04
-0.043 (0.027) 0.11 -0.036 (0.027) 0.18 -0.022 (0.028) 0.43 -0.018 (0.026) 0.49
1.181 (0.783) 0.13 -1.768 (0.567) 0.002 -1.544 (0.572) 0.0070 -1.490 (0.580) 0.01 -0.496 (0.547) 0.37
-0.028 (0.027) 0.04
-1.524 (0.716) 0.03
-0.007 (0.026) 0.78
-0.427 (0.550) 0.44
rs7119465
-0.018 (0.031)
0.768 (0.802)
-0.028 (0.030)
-0.038 (0.645)
rs8031759 rs886282
-0.007 (0.049)
rs2269145
-0.080 (0.049)
0.95 -1.398 (1.025)
0.67
0.10
0.17
1.283 (0.956)
0.017 (0.034)
1.100 (0.725)
0.012 (0.028) -0.013 (0.028) 0.65
rs11870732
-0.543 (1.269)
0.36
0.038 (0.037)
0.68 rs17765644
0.34
0.89 0.31 rs17178251
0.016 (0.028)
0.18 -0.640 (0.733) 0.38 0.643 (0.734) 0.38 -0.590 (0.734)
0.62 0.011 (0.028) 0.69 -0.011 (0.028) 0.70 0.012 (0.028)
0.13 -0.069 (0.587) 0.91 0.079 (0.587) 0.89 -0.071 (0.588)
0.57
0.42
0.67
0.90
0.032 (0.027)
0.272 (0.692)
0.010 (0.026)
0.124 (0.553)
0.23
0.69
0.71
0.82
was reported in an idiopathic pulmonary fibrosis study (r2 = 0.97), based on the Asian 1000 Genomes database (http://www.broadinstitute.org/mpg/snap/). In the previous European population study, there was no association between FAM13A and smoking intensity and no statistical evidence of interaction between smoking and FAM13A. Using our previous RNA-seq data, we established that expression of FAM13A is significantly higher in the lung tissue of individuals with COPD compared to controls. Expression levels did not differ according to disease status in the idiopathic pulmonary fibrosis study. SNPs in FAM13A showed significant associations with gene expression levels, with increased expression correlated with the risk allele. This suggests that these intronic SNPs, or functional SNPs in tight LD with them, may act as regulatory elements for this gene. Although FAM13A is known to function in intracellular signal transduction [20], its precise role in the respiratory system is currently unknown. ACN9 (chromosome 7q21) is known to have roles in gluconeogenesis [21] and mitochondrial disease [22] and has also been identified as associated with alcohol
123
-3.482 (0.709)
FEV1/FVC
rs10229181
0.56
Values are b (error) and p
FEV1/FVC
dependence in a previous genetic study [23]; however, it has not been previously reported as having a role in respiratory disease. Significant SNPs at ACN9 identified in this study were intronic. Our previous analysis of data from the general population demonstrated association of ACN9 with the FEV1/FVC ratio. This gene showed significant interaction with smoking intensity, measured in packyears, and significant association with COPD susceptibility only in the light-smoking group, i.e., those who smoked less than 35 pack-years. The lack of a significant association in heavy smokers may suggest that the role of this gene in susceptibility has an effect early in the pathogenesis related to smoking exposure, or it may be that only this subgroup has comparable smoking intensity to controls, where the smoking pack-years mean was 35. In our gene expression analysis, we observed a trend suggesting that SNPs in ACN9 may affect its expression level; further studies are warranted to confirm this and elucidate the underlying mechanism. Two additional genes (MCTP2 and AKAP1) showed interaction with smoking intensity; however, they were not associated with COPD in stratified models.
Lung
Fig. 1 Gene expression levels according to genotypes of three genes: a PID1, b FAM13A, and c ACN9. There was a significant association between expression of FAM13A and genotype at rs2609264
(p = 0.0009). Expression levels are in FPKM (fragments per kilobase of exon per million fragments mapped)
There are several limitations of this study. First, controls were younger and had smoked less than the subjects with COPD. We adjusted for age and smoking amount in our analyses. The smoking intensity of the group of individuals with COPD with the lowest smoking intensity was less than 35 pack-years, which is comparable with the control group. When we analyzed only this group compared to all controls; results were consistent with those for the whole COPD group with all controls for all SNPs. The p value for ACN9 was modest and would not be considered significant after correction for multiple testing. The associations with FAM13A and PID1 remain significant using this more conservative threshold. Subsequent functional studies of FAM13A and PID1 in the pathogenesis of COPD would be needed. Genetic association of ACN9 with COPD and lung function may not be generalized to other ethnic groups. Further replication studies are needed for clarifying the association of ACN9 and COPD. In conclusion, we have confirmed the associations of FAM13A and PID1 with COPD susceptibility. We also identified that ACN9 showed a significant interaction with smoking intensity and was associated with COPD in the light-smoking subgroup. Significant associations of genetic variants of FAM13A with gene expression levels of FAM13A suggest that the associated loci may act as genetic regulatory elements for FAM13A gene expression.
Biobank, which is part of the National Biobank of Korea, supported by the Ministry of Health and Welfare, Republic of Korea. For the recruitment of COPD participants and also for the collection of data and samples, we thank all members of the Korean Obstructive Lung Disease (KOLD) Study Group anrsd the ASAN Network: Ji-Hyun Lee, Eun Kyung Kim, Tae-Hyung Kim, Tae Rim Shin, Jin Hwa Lee, Seong Yong Lim, Sang Yeub Lee, Ho Il Yoon, Kwang Ha Yoo, Seung Soo Sheen, Joo Hun Park, Yong Bum Park, Changhwan Kim, Yong Il Hwang, Young Sam Kim, Ji Ye Jung, Yoonki Hong, Seung Won Ra, Joon Beom Seo, Sang Min Lee, In A Jeong, Chang Hoon Lee, Sei Won Lee, Jae Seung Lee, Jin Won Huh, Ji Yong Moon, HyeKyeong Park, Hye Yun Park, Jin Woo Kim, Chin Kook Rhee, Hyoung Kyu Yoon, Woo Jin Kim, Jong Deog Lee, Kang Hyeon Choi, Bock Hyun Jung, Joo Ock Na, Doh Hyung Kim, Hye Sook Choi, Kwang Ha Lee, Myung Jae Park, Sung Soon Lee, Yeon-Mok Oh, and Sang Do Lee.
Acknowledgments This study was supported by the National Project for Personalized Genomic Medicine (A111218-11-GM02) and a grant from the Korean Health 21 R&D Project, Ministry of Health and Welfare, Republic of Korea (A102065). DNA samples were generously provided by the Kangwon National University Hospital
Conflict of interest SDL received payment for lecturing from Nycomed, Takeda, and Norvatis. YMO received payment for lecturing from MSD Korea, AstraZeneca Korea, Boehringer Ingelheim Korea, Norvatis, DongWha, Takeda, and GSK Korea. EKS received a consultant fee from GSK, AstraZeneca, and Merck and payment for lecturing from GSK, AstraZeneca, and Merck. The other authors have no conflicts of interest to disclose.
References 1. Mathers CD, Loncar D (2006) Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med 3:e442 2. Vestbo J, Hurd SS, Agustı´ AG, Jones PW, Vogelmeier C, Anzueto A et al (2013) Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: GOLD executive summary. Am J Respir Crit Care Med 187:347–365 3. Todd JL, Goldstein DB, Ge D, Christie J, Palmer SM (2011) The state of genome-wide association studies in pulmonary disease. Am J Respir Crit Care Med 184:873–880
123
Lung 4. Soler Artigas M, Wain LV, Repapi E, Me Obeidat, Sayers I, Burton PR et al (2012) Effect of five genetic variants associated with lung function on the risk of chronic obstructive lung disease, and their joint effects on lung function. Am J Respir Crit Care Med 184:786–795 5. Wilk JB, Shrine NRG, Loehr LR, Zhao JH, Manichaikul A, Lopez LM et al (2012) Genome-wide association studies identify CHRNA5/3 and HTR4 in the development of airflow obstruction. Am J Respir Crit Care Med 186:622–632 6. Castaldi PJ, Cho MH, Litonjua AA, Bakke P, Gulsvik A, Lomas DA et al (2011) The association of genome-wide significant spirometric loci with COPD susceptibility. Am J Respir Cell Mol Biol 45:1147–1153 7. Hardin M, Zielinski J, Wan ES, Hersh CP, Castaldi PJ, Schwinder E et al (2012) CHRNA3/5, IREB2, and ADCY2 are associated with severe chronic obstructive pulmonary disease in Poland. Am J Respir Cell Mol Biol 47:203–208 8. Hancock DB, Artigas MS, Gharib SA, Henry A, Manichaikul A, Ramasamy A et al (2012) Genome-wide joint meta-analysis of SNP and SNP-by-smoking interaction identifies novel loci for pulmonary function. PLoS Genet 8:e1003098 9. Miller MR, Hankinson J, Brusasco V, Burgos F, Casaburi R, Coates A et al (2005) Standardisation of spirometry. Eur Respir J 26:319–338 10. Kim WJ, Lee MK, Shin C, Cho NH, Lee SD, Oh YM, Sung J (2014) Genome-wide association studies identify locus on 6p21 influencing lung function in the Korean population. Respirology 19(3):360–368 11. Kim WJ, Oh YM, Lim JH, Jeong JG, Kim JH, Lee JS et al (2012) Gene expression profiling of lung tissue from COPD subjects [abstract]. Respirology 17(S2):35 12. Hancock DB, Eijgelsheim M, Wilk JB, Gharib SA, Loehr LR, Marciante KD et al (2010) Meta-analyses of genome-wide association studies identify multiple loci associated with pulmonary function. Nat Genet 42:45–52 13. Li X, Howard TD, Moore WC, Ampleford EJ, Li H, Busse WW et al (2011) Importance of hedgehog interacting protein and other lung function genes in asthma. J Allergy Clin Immunol 127:1457–1465
123
14. Xu H, Xu G, Wang D, Zheng C, Wan L (2013) Molecular cloning and tissue distribution of the phosphotyrosine interaction domain containing 1 (PID1) gene in Tianfu goat. Gene 515:71–77 15. Zhao Y, Zhang C, Chen X, Gao C, Ji C, Chen F et al (2012) Overexpression of NYGGF4 (PID1) induces mitochondrial impairment in 3T3-L1 adipocytes. Mol Cell Biochem 340:41–48 16. Wu WL, Gan WH, Tong ML, Li XL, Dai JZ, Zhang CM et al (2011) Over-expression of NYGGF4 (PID1) inhibits glucose transport in skeletal myotubes by blocking the IRS1/PI3K/AKT insulin pathway. Mol Genet Metab 102:374–377 17. Kirkham PA, Barnes PJ (2013) Oxidative stress in COPD. Chest 144:266–273 18. Cho MH, Boutaoui N, Klanderman BJ, Sylvia JS, Ziniti JP, Hersh CP et al (2010) Variants in FAM13A are associated with chronic obstructive pulmonary disease. Nat Genet 42:200–202 19. Fingerlin TE, Murphy E, Zhang W, Peljto AL, Brown KK, Steele MP et al (2013) Genome-wide association study identifies multiple susceptibility loci for pulmonary fibrosis. Nat Genet 45:613–620 20. Cohen M, Reichenstein M, Everts-van der Wind A, Heon-Lee J, Shani M, Lewin HA et al (2004) Cloning and characterization of FAM13A1—a gene near a milk protein QTL on BTA6: evidence for population-wide linkage disequilibrium in Israeli Holsteins. Genomics 84:374–383 21. Dennis RA, McCammon MT (1999) Acn9 is a novel protein of gluconeogenesis that is located in the mitochondrial intermembrane space. Eur J Biochem 261:236–243 22. de Graaf RM, Ricard G, van Alen TA, Duarte I, Dutilh BE, Burgtorf C et al (2011) The Organellar genome and metabolic potential of the hydrogen-producing mitochondrion of Nyctotherus ovalis. Mol Biol Evol 28:2379–2391 23. Dick DM, Aliev F, Wang JC, Saccone S, Hinrichs A, Bertelsen S et al (2008) A systematic single nucleotide polymorphism screen to fine-map alcohol dependence genes on chromosome 7 identifies association with a novel susceptibility gene ACN9. Biol Psychiatry 63:1047–1053
Association of Lung Function Genes with Chronic Obstructive Pulmonary Disease Woo Jin Kim • Myoung Nam Lim • Yoonki Hong • Edwin K. Silverman Ji-Hyun Lee • Bock Hyun Jung • Seung Won Ra • Hye Sook Choi • Young Ju Jung • Yong Bum Park • Myung Jae Park • Sei Won Lee • Jae Seung Lee • Yeon-Mok Oh • Sang Do Lee
•
Received: 30 November 2013 / Accepted: 21 March 2014 Ó Springer Science+Business Media New York 2014
Abstract Background Spirometric measurements of pulmonary function are important in diagnosing and determining the severity of chronic obstructive pulmonary disease (COPD). We performed this study to determine whether candidate genes identified in genome-wide association studies of spirometric measurements were associated with COPD and if they interacted with smoking intensity. Methods The current analysis included 1,000 COPD subjects and 1,000 controls recruited from 24 hospitalbased pulmonary clinics. Thirteen SNPs, chosen based on
genome-wide association studies of spirometric measurements in the Korean population cohorts, were genotyped. Genetic association tests were performed, adjusting for age, sex, and smoking intensity, using models including a SNP-by-smoking interaction term. Results PID1 and FAM13A were significantly associated with COPD susceptibility. There were also significant interactions between SNPs in ACN9 and FAM13A and smoking pack-years, and an association of ACN9 with COPD in the lowest smoking tertile. The risk allele of FAM13A was associated with increased expression of FAM13A in the lung.
W. J. Kim Y. Hong Department of Internal Medicine and Environmental Health Center, Kangwon National University Hospital, School of Medicine, Kangwon National University, Chuncheon, South Korea
H. S. Choi Department of Internal Medicine, Dongguk University Gyeongju Hospital, Dongguk University College of Medicine, Gyeongju, South Korea
M. N. Lim Department of Statistics, Kangwon National University, Chuncheon, South Korea E. K. Silverman Channing Division of Network Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA J.-H. Lee Department of Internal Medicine, CHA Bundang Medical Center, CHA University, Seongnam, South Korea B. H. Jung Department of Pulmonary and Critical Medicine, Gangneung Asan Hospital, University of Ulsan College of Medicine, Gangneung, South Korea S. W. Ra Department of Internal Medicine, Ulsan University Hospital, University of Ulsan College of Medicine, Ulsan, South Korea
Y. J. Jung Health Promotion Center, Asan Medical Center, University of Ulsan College of Medicine, Seoul, South Korea Y. B. Park Department of Pulmonary and Critical Care Medicine, Kangdong Sacred Heart Hospital, Hallym University, Seoul, South Korea M. J. Park Division of Respiratory and Critical Care Medicine, Kyung Hee University Hospital, Kyung Hee University, Seoul, South Korea S. W. Lee J. S. Lee Y.-M. Oh (&) S. D. Lee Department of Pulmonary and Critical Care Medicine, and Clinical Research Center for Chronic Obstructive Airway Diseases, Asan Medical Center, University of Ulsan College of Medicine, 88 Olympic-ro 43-gil, Songpa-gu, Seoul 138-736, Republic of Korea e-mail: [email protected]
123
Lung
Conclusions We have validated associations of FAM13A and PID1 with COPD. ACN9 showed significant interaction with smoking and is a potential candidate gene for COPD. Significant associations of genetic variants of FAM13A with gene expression levels suggest that the associated loci may act as genetic regulatory elements for FAM13A gene expression. Keywords ACN9 COPD Genetic association Gene-by-environment interaction Cigarette smoking
Introduction The prevalence and burden of chronic obstructive pulmonary disease (COPD) are increasing, and it is expected to become the third leading cause of death worldwide by 2020 [1]. COPD is characterized by irreversible airflow limitation and results from environmental factors, including cigarette smoking, with contributions from genetic background and gene-by-environment interactions [2]. Spirometric measurements of forced expiratory volume in 1 s (FEV1) and FEV1/forced vital capacity (FVC) are important in diagnosing and determining the severity of COPD. Genome-wide association studies (GWAS) have identified many loci influencing FEV1 and/or the FEV1/ FVC ratio in general population samples [3]. Although many of the genes at these loci have unknown functions and have not previously been considered biologically plausible candidates for respiratory disease pathogenesis, these findings can provide insight into the genetic architecture contributing to lung function and may have important roles in the development of the respiratory system or in airway disease pathogenesis. On examining COPD as an outcome in GWAS meta-analysis, genes associated with pulmonary function in the general population showed significant associations with COPD susceptibility [4, 5]. In a genetic association study of spirometric loci as candidate genes, many were shown to be susceptibility genes for COPD [6, 7]. These studies suggest that genomic loci associated with FEV1 or FEV1/FVC may also be associated with COPD susceptibility. Smoking is the most important risk factor for COPD and has adverse effects on lung function. Genetic factors may have heterogeneous effects on pulmonary function, depending on smoking exposure. In a recent GWAS, novel loci were identified as significantly associated with lung function after investigation of SNP-by-smoking interaction [8]. In the present study, lung function candidate genes identified in GWAS of general population cohorts were evaluated for association and smoking interaction in separate COPD and control samples.
123
Methods Study Participants This study included 1,000 individuals with COPD and 1,000 unaffected controls. Individuals with COPD were recruited from 24 hospital-based pulmonary clinics belonging to the KOLD Study Group or the ASANCOPD Network. Controls were recruited from individuals attending the Asan Medical Center, Seoul, Korea, for a medical checkup. All control subjects were current or exsmokers. Participants performed standardized spirometry in accordance with American Thoracic Society/European Respiratory Society criteria [9]. Individuals with COPD demonstrated a post-bronchodilator FEV1/FVC \ 0.7, while controls had normal spirometry. Pack-years of smoking were calculated as smoking amount per day and duration of years smoked. All participants provided written informed consent, and the study protocol was approved by the respective institutional review boards. Selection of Candidate Genes and Genotyping Thirteen SNPs were chosen based on meta-analysis of the results from GWAS of Korean population cohorts, where spirometric measurements had been taken (KARE3 and HTS) [10]. SNPs were genotyped using TaqMan technology (Applied Biosystems, Foster City, CA) on an Applied Biosystems 7900HT real-time PCR system. Statistical Analyses Hardy–Weinberg equilibrium was assessed using goodness-of-fit tests. We performed genetic association testing using logistic regression with COPD affection status, assuming an additive genetic model and adjusting for age, sex, and smoking pack-years using SAS ver. 9.2 (SAS Inc., Cary, NC, USA). In addition to the overall model, we evaluated models that included a SNP-by-pack-years interaction term and a smoking intensity-stratified model. Linear regression analyses were implemented using FEV1 and FEV1/FVC ratio as quantitative phenotypes after adjusting for age, sex, and smoking pack-years. Gene Expression in Lung Tissue We used gene expression profiling data from a previous RNA-seq experiment using lung tissues from 193 participants with or without COPD [11]. For differential expression according to genotype of candidate genes, fragments per kilobase of exon per million fragments mapped (FPKM) levels were analyzed using a t test.
Lung Table 1 Demographic characteristics of analyzed participants with COPD and controls COPD (n = 1,000)
Control (n = 1,000)
p value
Male (%)
977 (97.7)
983 (98.3)
0.42
Age (years)
69.2 ± 7.8
57.2 ± 9.5
\0.0001
Pack-years of smoking
44.9 ± 23.1
34.7 ± 27.8
\0.0001
FEV1 (L)
1.64 ± 0.61
3.41 ± 0.68
\0.0001
FEV1, % of predicted
55.9 ± 18.6
95.9 ± 32.6
\0.0001
FEV1/FVC
48.9 ± 12.1
78.5 ± 22.8
\0.0001
Data are presented as the mean (±SD) unless otherwise stated
Table 2 List of 13 SNPs genotyped for association with COPD, according to pulmonary function genome-wide association meta-analysis
Chr chromosome, MAF minor allele frequency, Ref reference, Alt alternative
SNP
Chr
Position
Gene
Ref allele
Alt allele
MAF
Beta
rs16825116
2
229406685
PID1
G
T
0.16
?
rs2609264
4
89828080
FAM13A
C
T
0.52
–
rs2609261
4
89835485
FAM13A
A
G
0.53
–
rs2609260
4
89836819
FAM13A
C
T
0.56
–
rs10231916
7
96763456
ACN9
A
T
0.51
–
rs10229181
7
96795041
ACN9
C
G
0.51
–
rs7119465
11
25734037
ANO3
A
G
0.76
?
rs8031759
15
93798810
MCTP2
G
T
0.08
–
rs886282
17
55188150
AKAP1
C
T
0.18
–
rs17178251
17
69176879
SOX9
C
G
0.61
?
rs17765644
17
69179492
SOX9
C
T
0.62
?
rs11870732
17
69195241
SOX9
A
G
0.39
–
rs2269145
21
44124432
PDE9A
C
T
0.53
?
The baseline characteristics of the study participants (1,000 with COPD and 1,000 controls) are presented in Table 1. Individuals with COPD were older and had higher packyears of cigarette smoking than controls. As expected, individuals with COPD had poorer lung function than controls. Therefore, we adjusted for age and smoking intensity in our genetic association analyses.
(p = 0.0015) and rs10229181 (p = 0.002), showed modest deviations from Hardy–Weinberg equilibrium in control subjects but not in the entire population; these SNPs were retained in the analysis. A list of SNPs genotyped and the direction of the association with lung function from prior analysis in a Korean general population cohort are shown in Table 2. One SNP (rs16825116) near PID1 and three in FAM13A (rs2609264, rs2609261, and rs2609260) showed significant associations with COPD, after adjustment for age, sex, and pack-years of smoking (Table 3). The ‘T’ allele of rs16825116, which was associated with increased lung function, was protective against COPD. The three SNPs in FAM13A are in strong linkage disequilibrium (LD) with one another (r2 = 0.8–0.93). The most significant SNP (rs2609264) is intronic and the direction of the odds ratio in FAM13A was consistent with the effect observed on lung function. The rs2609264 ‘T’ allele had a negative beta coefficient for association with FEV1 and an increased risk for COPD (OR = 1.407, Table 3).
Genetic Association Tests
Smoking Interaction and Stratified Analyses
Eleven of the SNPs were in Hardy–Weinberg equilibrium in control subjects. Two of the SNPs, rs10231916
After inclusion of a SNP-by-pack-years interaction term, there were significant interactions between SNPs in ACN9,
ACN9, PID1, and FAM13A were genotyped in 93 participants who had available blood samples, regardless of disease status, to investigate whether variants at these loci influence gene expression. Linear regression was performed to investigate associations between expression level and genotype.
Results Demographic Characteristics
123
Lung Table 3 Results of genetic association analyses with COPD susceptibility and interaction with smoking pack-years
SNP
rs16825116
rs2609264
rs2609261
rs2609260
rs10231916
rs10229181
rs7119465
Chr
Gene
2
PID1
4
4
4
7
7
11
FAM13A
FAM13A
FAM13A
ACN9
ACN9
ANO3
rs8031759
15
MCTP2
rs886282
17
AKAP1
rs17178251
rs17765644
rs11870732
rs2269145
17
17
17
21
SOX9
SOX9
SOX9
PDE9A
pint is the p value of interaction term with smoking intensity
FAM13A, MCTP2, and AKAP1 and smoking pack-years (Table 3). When the study population was stratified into three COPD groups based on pack-years with all controls as the reference group, there was a significant association of ACN9 with COPD in only the lowest smoking tertile subgroup (OR = 1.29, p = 0.028), and SNPs in FAM13A showed significant associations with COPD in all
123
COPD
Control
OR 95 % CI pSNP
pint
GG
727
672
0.780
0.18
GT
254
294
(0.632–0.962)
TT
18
33
0.02
CC
170
239
1.407
CT
512
517
(1.199–1.650)
TT
308
241
2.8 9 10-5
AA AG
160 511
229 508
1.329 (1.133–1.558)
GG
317
261
4.7 9 10-4
CC
139
203
1.327
CT
509
514
(1.130–1.559)
TT
345
282
5.8 9 10-4
AA
271
248
1.043
AT
494
549
(0.893–1.219)
TT
232
203
0.60
CC
265
245
1.064
CG
479
544
(0.910–1.243)
GG
236
204
0.44
AA
58
71
0.932
AG
371
366
(0.779–1.116)
GG
549
561
0.44
GG GT
845 145
842 148
0.967 (0.728–1.283)
TT
5
9
CC
665
711
CT
286
267
TT
29
18
0.51
CC
378
395
1.013
CG
498
471
(0.861–1.192)
3.4 9 10-4
4.0 9 10-5
2.3 9 10-4
1.3 9 10-8
8.1 9 10-9
0.39
3.1 9 10-6
0.82 1.074
1.5 9 10-6
(0.868–1.328)
GG
121
134
0.88
CC
121
133
0.984
CT
497
471
(0.837–1.158)
TT
377
395
0.85
AA
376
392
0.998
AG
492
472
(0.848–1.174)
GG
122
135
0.98
CC
209
208
0.917
CT TT
502 285
469 321
(0.786–1.069) 0.27
0.11
0.12
0.19
0.84
subgroups, whereas MCTP2 and AKAP1 did not demonstrate any associations (Table 4). Association with Lung Function Linear regression analyses revealed that SNPs in ACN9 and FAM13A were significantly associated with FEV1/FVC
Lung Table 4 Results of stratified genetic association analyses according to smoking intensity
Values are odds ratio, (confidential interval), and p values
SNP
Gene
Tertile 1
rs2609264
FAM13A
1.48 (1.18–1.85), 0.0006
1.38 (1.12–1.72), 0.0032
1.61 (1.24–2.10), 0.0003
rs2609261
FAM13A
1.39 (1.11–1.73), 0.004
1.30 (1.05–1.60), 0.018
1.54 (1.19–2.00), 0.0011
rs2609260
FAM13A
1.38 (1.10–1.73), 0.0046
1.36 (1.10–1.70), 0.0055
1.42 (1.09–1.84), 0.0091
rs10231916
ACN9
1.25 (1.00–1.56), 0.051
0.97 (0.78–1.20), 0.77
0.82 (0.64–1.06), 0.13
rs10229181
ACN9
1.29 (1.03–1.61), 0.028
1.00 (0.81–1.24), 0.98
0.82 (0.64–1.06), 0.13
rs8031759
MCTP2
1.15 (0.79–1.65), 0.47
0.71 (0.47–1.08), 0.11
0.89 (0.55–1.44), 0.63
rs886282
AKAP1
1.04 (0.76–1.41), 0.82
1.15 (0.87–1.52), 0.34
1.11 (0.79–1.55), 0.56
across all study participants (Table 5). The ‘T’ allele of rs10231916 in ACN9 was associated with lower FEV1/FVC values, consistent with results from the general population, and the association with SNPs in FAM13A was also in the same direction as that observed in the population study. SNPs in PID1 and FAM13A also showed significant associations with FEV1, with the ‘G’ allele of PID1 associated with lower FEV1 values, consistent with results of GWAS meta-analysis. Significant associations of PID1 with FEV1 and FAM13A with FEV1/FVC were observed only among participants with COPD (Table 5). Expression in RNA-seq Three genes (ACN9, PID1, and FAM13A) were evaluated for mRNA expression in lung tissue samples using data from a previous experiment [11]. RNA-seq analyses demonstrated that lung tissue from COPD subjects had significantly decreased expression of ACN9 (p = 2.6 9 10-8). The ‘T’ allele of rs10231916 (ACN9) showed a tendency toward lower expression in 93 samples (p = 0.07, Fig. 1). PID1 and FAM13A showed increased expression in lung tissue from individuals with COPD (p = 1.5 9 10-4 and p = 1.8 9 10-11, respectively). The risk allele of rs2609261 was associated with increased expression of FAM13A in the lung (p = 0.0009, Fig. 1). SNPs at the PID1 locus did not show significant association with gene expression.
Discussion In this study, we demonstrated that SNPs in FAM13A and PID1 are significantly associated with COPD in the Korean population. Although SNPs in ACN9 were not associated with COPD in the analysis of main genetic effects, a significant interaction with smoking levels among the COPD subgroup who smoked the least was observed. In the current study, rs11435116 near PID1 was associated with COPD and FEV1 in the Korean population. In a previous GWAS, another SNP flanking PID1, rs11435867, was associated with the FEV1/FVC ratio in the European
Tertile 2
Tertile 3
general population [12]. In another report, the association of PID1 with lung function was replicated in individuals with asthma [13], and PID1 has previously been associated with COPD susceptibility, albeit with a nominal p value [6]. The rs16825116 SNP is located at chromosome 2q36 between SPHKAP and PID1, 100 kb from rs11435867, which has previously been reported as associated with lung function in the European population. PID1 showed increased expression in COPD lung tissues in our analyses, while SPHKAP gene expression levels did not differ between the two groups. PID1 is involved in tissue homeostasis, cell growth, and adiposity [14]. Overexpression of PID1 leads to mitochondrial dysfunction in adipocytes [15], and overexpression in rat skeletal muscle results in inhibition of insulin-stimulated glucose transport [16]. Although the PID1 SNP significantly associated with COPD in our study did not appear to influence gene expression, overexpression of PID1 may have a role in COPD pathogenesis through mitochondrial dysfunction and excess reactive oxygen species [17]. In a previous GWAS that used populations of European descent, FAM13A was significantly associated with COPD susceptibility [18] and with lung function in the general population [12]. This region also showed a genome-wide significant association with idiopathic pulmonary fibrosis [19]. In the current study, SNPs in FAM13A were associated with COPD susceptibility and lung function in individuals with COPD and among the entire study population. The directions of association were consistent with our previous analysis in the general population and showed significant interaction with all subgroups of smoking intensity. The three SNPs in FAM13A identified in the current study are intronic. The most significant, rs2609264, is located 10 kb from rs2869967 (r2 = 0.73) and 50 kb from rs1903003 (r2 = 0.53), which were previously reported as associated with lung function or COPD in European populations. These results provide compelling evidence that the FAM13A region influences COPD susceptibility. Differences in LD between the Korean and European populations may assist in localizing functional genetic variants. Interestingly, tight LD between rs2609264 and rs2609255
123
Lung Table 5 Association between SNPs and lung function in all participants and those with COPD
SNP
rs16825116 rs2609264
All participants (n = 2,000)
COPD only (n = 1,000)
FEV1
FEV1
0.087 (0.036)
1.022 (0.942)
0.082 (0.037)
0.02
0.28
0.03
-0.09 (0.027) 0.002
rs2609261
-0.070 (0.027) 0.01
rs2609260
-0.065 (0.028)
rs10231916
-0.028 (0.027)
0.02 0.30
\0.0001 -3.193 (0.709) \0.0001 -3.209 (0.717) \0.0001 -1.460 (0.712) 0.04
-0.043 (0.027) 0.11 -0.036 (0.027) 0.18 -0.022 (0.028) 0.43 -0.018 (0.026) 0.49
1.181 (0.783) 0.13 -1.768 (0.567) 0.002 -1.544 (0.572) 0.0070 -1.490 (0.580) 0.01 -0.496 (0.547) 0.37
-0.028 (0.027) 0.04
-1.524 (0.716) 0.03
-0.007 (0.026) 0.78
-0.427 (0.550) 0.44
rs7119465
-0.018 (0.031)
0.768 (0.802)
-0.028 (0.030)
-0.038 (0.645)
rs8031759 rs886282
-0.007 (0.049)
rs2269145
-0.080 (0.049)
0.95 -1.398 (1.025)
0.67
0.10
0.17
1.283 (0.956)
0.017 (0.034)
1.100 (0.725)
0.012 (0.028) -0.013 (0.028) 0.65
rs11870732
-0.543 (1.269)
0.36
0.038 (0.037)
0.68 rs17765644
0.34
0.89 0.31 rs17178251
0.016 (0.028)
0.18 -0.640 (0.733) 0.38 0.643 (0.734) 0.38 -0.590 (0.734)
0.62 0.011 (0.028) 0.69 -0.011 (0.028) 0.70 0.012 (0.028)
0.13 -0.069 (0.587) 0.91 0.079 (0.587) 0.89 -0.071 (0.588)
0.57
0.42
0.67
0.90
0.032 (0.027)
0.272 (0.692)
0.010 (0.026)
0.124 (0.553)
0.23
0.69
0.71
0.82
was reported in an idiopathic pulmonary fibrosis study (r2 = 0.97), based on the Asian 1000 Genomes database (http://www.broadinstitute.org/mpg/snap/). In the previous European population study, there was no association between FAM13A and smoking intensity and no statistical evidence of interaction between smoking and FAM13A. Using our previous RNA-seq data, we established that expression of FAM13A is significantly higher in the lung tissue of individuals with COPD compared to controls. Expression levels did not differ according to disease status in the idiopathic pulmonary fibrosis study. SNPs in FAM13A showed significant associations with gene expression levels, with increased expression correlated with the risk allele. This suggests that these intronic SNPs, or functional SNPs in tight LD with them, may act as regulatory elements for this gene. Although FAM13A is known to function in intracellular signal transduction [20], its precise role in the respiratory system is currently unknown. ACN9 (chromosome 7q21) is known to have roles in gluconeogenesis [21] and mitochondrial disease [22] and has also been identified as associated with alcohol
123
-3.482 (0.709)
FEV1/FVC
rs10229181
0.56
Values are b (error) and p
FEV1/FVC
dependence in a previous genetic study [23]; however, it has not been previously reported as having a role in respiratory disease. Significant SNPs at ACN9 identified in this study were intronic. Our previous analysis of data from the general population demonstrated association of ACN9 with the FEV1/FVC ratio. This gene showed significant interaction with smoking intensity, measured in packyears, and significant association with COPD susceptibility only in the light-smoking group, i.e., those who smoked less than 35 pack-years. The lack of a significant association in heavy smokers may suggest that the role of this gene in susceptibility has an effect early in the pathogenesis related to smoking exposure, or it may be that only this subgroup has comparable smoking intensity to controls, where the smoking pack-years mean was 35. In our gene expression analysis, we observed a trend suggesting that SNPs in ACN9 may affect its expression level; further studies are warranted to confirm this and elucidate the underlying mechanism. Two additional genes (MCTP2 and AKAP1) showed interaction with smoking intensity; however, they were not associated with COPD in stratified models.
Lung
Fig. 1 Gene expression levels according to genotypes of three genes: a PID1, b FAM13A, and c ACN9. There was a significant association between expression of FAM13A and genotype at rs2609264
(p = 0.0009). Expression levels are in FPKM (fragments per kilobase of exon per million fragments mapped)
There are several limitations of this study. First, controls were younger and had smoked less than the subjects with COPD. We adjusted for age and smoking amount in our analyses. The smoking intensity of the group of individuals with COPD with the lowest smoking intensity was less than 35 pack-years, which is comparable with the control group. When we analyzed only this group compared to all controls; results were consistent with those for the whole COPD group with all controls for all SNPs. The p value for ACN9 was modest and would not be considered significant after correction for multiple testing. The associations with FAM13A and PID1 remain significant using this more conservative threshold. Subsequent functional studies of FAM13A and PID1 in the pathogenesis of COPD would be needed. Genetic association of ACN9 with COPD and lung function may not be generalized to other ethnic groups. Further replication studies are needed for clarifying the association of ACN9 and COPD. In conclusion, we have confirmed the associations of FAM13A and PID1 with COPD susceptibility. We also identified that ACN9 showed a significant interaction with smoking intensity and was associated with COPD in the light-smoking subgroup. Significant associations of genetic variants of FAM13A with gene expression levels of FAM13A suggest that the associated loci may act as genetic regulatory elements for FAM13A gene expression.
Biobank, which is part of the National Biobank of Korea, supported by the Ministry of Health and Welfare, Republic of Korea. For the recruitment of COPD participants and also for the collection of data and samples, we thank all members of the Korean Obstructive Lung Disease (KOLD) Study Group anrsd the ASAN Network: Ji-Hyun Lee, Eun Kyung Kim, Tae-Hyung Kim, Tae Rim Shin, Jin Hwa Lee, Seong Yong Lim, Sang Yeub Lee, Ho Il Yoon, Kwang Ha Yoo, Seung Soo Sheen, Joo Hun Park, Yong Bum Park, Changhwan Kim, Yong Il Hwang, Young Sam Kim, Ji Ye Jung, Yoonki Hong, Seung Won Ra, Joon Beom Seo, Sang Min Lee, In A Jeong, Chang Hoon Lee, Sei Won Lee, Jae Seung Lee, Jin Won Huh, Ji Yong Moon, HyeKyeong Park, Hye Yun Park, Jin Woo Kim, Chin Kook Rhee, Hyoung Kyu Yoon, Woo Jin Kim, Jong Deog Lee, Kang Hyeon Choi, Bock Hyun Jung, Joo Ock Na, Doh Hyung Kim, Hye Sook Choi, Kwang Ha Lee, Myung Jae Park, Sung Soon Lee, Yeon-Mok Oh, and Sang Do Lee.
Acknowledgments This study was supported by the National Project for Personalized Genomic Medicine (A111218-11-GM02) and a grant from the Korean Health 21 R&D Project, Ministry of Health and Welfare, Republic of Korea (A102065). DNA samples were generously provided by the Kangwon National University Hospital
Conflict of interest SDL received payment for lecturing from Nycomed, Takeda, and Norvatis. YMO received payment for lecturing from MSD Korea, AstraZeneca Korea, Boehringer Ingelheim Korea, Norvatis, DongWha, Takeda, and GSK Korea. EKS received a consultant fee from GSK, AstraZeneca, and Merck and payment for lecturing from GSK, AstraZeneca, and Merck. The other authors have no conflicts of interest to disclose.
References 1. Mathers CD, Loncar D (2006) Projections of global mortality and burden of disease from 2002 to 2030. PLoS Med 3:e442 2. Vestbo J, Hurd SS, Agustı´ AG, Jones PW, Vogelmeier C, Anzueto A et al (2013) Global strategy for the diagnosis, management, and prevention of chronic obstructive pulmonary disease: GOLD executive summary. Am J Respir Crit Care Med 187:347–365 3. Todd JL, Goldstein DB, Ge D, Christie J, Palmer SM (2011) The state of genome-wide association studies in pulmonary disease. Am J Respir Crit Care Med 184:873–880
123
Lung 4. Soler Artigas M, Wain LV, Repapi E, Me Obeidat, Sayers I, Burton PR et al (2012) Effect of five genetic variants associated with lung function on the risk of chronic obstructive lung disease, and their joint effects on lung function. Am J Respir Crit Care Med 184:786–795 5. Wilk JB, Shrine NRG, Loehr LR, Zhao JH, Manichaikul A, Lopez LM et al (2012) Genome-wide association studies identify CHRNA5/3 and HTR4 in the development of airflow obstruction. Am J Respir Crit Care Med 186:622–632 6. Castaldi PJ, Cho MH, Litonjua AA, Bakke P, Gulsvik A, Lomas DA et al (2011) The association of genome-wide significant spirometric loci with COPD susceptibility. Am J Respir Cell Mol Biol 45:1147–1153 7. Hardin M, Zielinski J, Wan ES, Hersh CP, Castaldi PJ, Schwinder E et al (2012) CHRNA3/5, IREB2, and ADCY2 are associated with severe chronic obstructive pulmonary disease in Poland. Am J Respir Cell Mol Biol 47:203–208 8. Hancock DB, Artigas MS, Gharib SA, Henry A, Manichaikul A, Ramasamy A et al (2012) Genome-wide joint meta-analysis of SNP and SNP-by-smoking interaction identifies novel loci for pulmonary function. PLoS Genet 8:e1003098 9. Miller MR, Hankinson J, Brusasco V, Burgos F, Casaburi R, Coates A et al (2005) Standardisation of spirometry. Eur Respir J 26:319–338 10. Kim WJ, Lee MK, Shin C, Cho NH, Lee SD, Oh YM, Sung J (2014) Genome-wide association studies identify locus on 6p21 influencing lung function in the Korean population. Respirology 19(3):360–368 11. Kim WJ, Oh YM, Lim JH, Jeong JG, Kim JH, Lee JS et al (2012) Gene expression profiling of lung tissue from COPD subjects [abstract]. Respirology 17(S2):35 12. Hancock DB, Eijgelsheim M, Wilk JB, Gharib SA, Loehr LR, Marciante KD et al (2010) Meta-analyses of genome-wide association studies identify multiple loci associated with pulmonary function. Nat Genet 42:45–52 13. Li X, Howard TD, Moore WC, Ampleford EJ, Li H, Busse WW et al (2011) Importance of hedgehog interacting protein and other lung function genes in asthma. J Allergy Clin Immunol 127:1457–1465
123
14. Xu H, Xu G, Wang D, Zheng C, Wan L (2013) Molecular cloning and tissue distribution of the phosphotyrosine interaction domain containing 1 (PID1) gene in Tianfu goat. Gene 515:71–77 15. Zhao Y, Zhang C, Chen X, Gao C, Ji C, Chen F et al (2012) Overexpression of NYGGF4 (PID1) induces mitochondrial impairment in 3T3-L1 adipocytes. Mol Cell Biochem 340:41–48 16. Wu WL, Gan WH, Tong ML, Li XL, Dai JZ, Zhang CM et al (2011) Over-expression of NYGGF4 (PID1) inhibits glucose transport in skeletal myotubes by blocking the IRS1/PI3K/AKT insulin pathway. Mol Genet Metab 102:374–377 17. Kirkham PA, Barnes PJ (2013) Oxidative stress in COPD. Chest 144:266–273 18. Cho MH, Boutaoui N, Klanderman BJ, Sylvia JS, Ziniti JP, Hersh CP et al (2010) Variants in FAM13A are associated with chronic obstructive pulmonary disease. Nat Genet 42:200–202 19. Fingerlin TE, Murphy E, Zhang W, Peljto AL, Brown KK, Steele MP et al (2013) Genome-wide association study identifies multiple susceptibility loci for pulmonary fibrosis. Nat Genet 45:613–620 20. Cohen M, Reichenstein M, Everts-van der Wind A, Heon-Lee J, Shani M, Lewin HA et al (2004) Cloning and characterization of FAM13A1—a gene near a milk protein QTL on BTA6: evidence for population-wide linkage disequilibrium in Israeli Holsteins. Genomics 84:374–383 21. Dennis RA, McCammon MT (1999) Acn9 is a novel protein of gluconeogenesis that is located in the mitochondrial intermembrane space. Eur J Biochem 261:236–243 22. de Graaf RM, Ricard G, van Alen TA, Duarte I, Dutilh BE, Burgtorf C et al (2011) The Organellar genome and metabolic potential of the hydrogen-producing mitochondrion of Nyctotherus ovalis. Mol Biol Evol 28:2379–2391 23. Dick DM, Aliev F, Wang JC, Saccone S, Hinrichs A, Bertelsen S et al (2008) A systematic single nucleotide polymorphism screen to fine-map alcohol dependence genes on chromosome 7 identifies association with a novel susceptibility gene ACN9. Biol Psychiatry 63:1047–1053
