Genetic Linkage and Association Analysis of COPD-Related Traits on Chromosome 8p

Abstract

Genome-wide linkage analysis in the Boston Early-Onset Chronic Obstructive Pulmonary Disease (COPD) Study has demonstrated significant evidence of linkage to chromosome 8p for forced expiratory volume in 1 second, an important COPD-related phenotype. In this study, we sought to fine map the linkage peak and to test variants in two candidate genes for association with COPD and related traits. In a variance component linkage analysis on chromosome 8, including seven additional short tandem repeat markers, the logarithm of the odds of linkage score was reduced from 3.30 to 1.80 (at 1 cM). Five single nucleotide polymorphisms (SNPs) in Defensin Beta-1 (DEFB1) were genotyped in the Boston Early-Onset COPD Study families; none was significantly associated. Four SNPs and an insertion-deletion polymorphism in Macrophage Scavenger Receptor-1 (MSR1) were also genotyped in the family-based study. A coding variant (Pro275Ala) was marginally associated with two qualitative airflow obstruction traits (p ≤ 0.02). This SNP showed a trend toward association in a case-control study comparing participants in the National Emphysema Treatment Trial to smoker controls (p = 0.07). Despite the reduced support for linkage upon further analysis, it remains possible that chromosome 8p contains a gene that influences COPD susceptibility. There is marginal, though not convincing, evidence for association with MSR1.

Key words: :

• Emphysema
• Haplotypes
• Linkage disequilibrium
• Single nucleotide polymorphism
• Spirometry

INTRODUCTION

Although cigarette smoking is the major environmental determinant of chronic obstructive pulmonary disease (COPD), family studies demonstrate a genetic influence on disease susceptibility ([1], [2]). In the Boston Early-Onset COPD Study, extended pedigrees were collected through a proband with severe chronic airflow obstruction diagnosed at a young age (< 53 years old) ([3]). Familial aggregation of airflow obstruction among smoking first-degree relatives in these families has been established ([3]). Genome-wide linkage analysis in the families using a panel of short tandem repeat (STR) markers has identified several regions linked to COPD-related traits ([4], [5]). Fine mapping of linked regions with additional STR markers, followed by candidate gene association studies in both the Boston Early-Onset COPD Study families and in a case-control study has identified transforming growth factor beta-1 (TGFB1) on chromosome 19q and serine proteinase inhibitor, clade E, member 2 (SERPINE2) on chromosome 2q as genes that may influence COPD susceptibility ([6], [7]).

Genome-wide linkage analysis in the Boston Early-Onset COPD Study families has found a region on chromosome 8p to be linked to post-bronchodilator values for forced expiratory volume in 1 second (FEV₁), an important intermediate phenotype for COPD ([5]). The logarithm of the odds of linkage (LOD) score of 3.30 represents borderline genome-wide significant evidence of linkage ([8]). Two candidate genes on chromosome 8p have been associated with COPD in other studies: defensin beta-1 (DEFB1) ([9], [10]) and macrophage scavenger receptor-1 (MSR1) ([11]), though the latter has been reported only in abstract form. In order to identify relevant COPD genes on chromosome 8p, we genotyped additional STR markers to fine map the region in the Boston Early-Onset COPD Study families. This was followed by genotyping of single nucleotide polymorphisms (SNPs) in the two potential COPD candidate genes, DEFB1 and MSR1, in the family study and in a case-control study.

METHODS

Study subjects

Details of recruitment in the Boston Early-Onset COPD Study have been published ([3]). Briefly, extended pedigrees were ascertained through a proband with an FEV₁ less than 40% predicted, at an age less than 53 years, and without alpha1-antitrypsin deficiency. After written informed consent, subjects completed a study questionnaire (modified from ([12])), performed spirometry (pre- and post-bronchodilator), and provided a blood sample for DNA extraction. The genome scan linkage dataset included 585 subjects in 72 pedigrees ([4]); the association analysis dataset was extended to include 949 subjects in 127 pedigrees.

The 304 cases in the case-control study were enrolled in the National Emphysema Treatment Trial (NETT) Genetics Ancillary Study. Participants in NETT had an FEV₁ less than or equal to 45% predicted and bilateral emphysema on chest CT scan ([13]). After written informed consent, they provided a blood sample for genetic analysis. The 441 control subjects were from the Normative Aging Study (NAS), a longitudinal study of aging in men conducted by the Boston VA ([14]). Controls had a smoking history of at least 10 pack-years but had no airflow obstruction on spirometry. Characteristics of participants in the Boston Early-Onset COPD Study and in the NETT-NAS case-control study have been reported ([15]). All studies were approved by the appropriate institutional review boards.

We had previously genotyped a panel of 44 unlinked single nucleotide polymorphisms (SNPs) to test for population stratification between the NETT cases and NAS control subjects; no significant stratification was found ([6]). Only white subjects were included in the case-control study.

Genotyping

A set of 377 short tandem repeat (STR) markers throughout the genome had already been genotyped in the Boston Early-Onset COPD Study by the National Heart, Lung, and Blood Institute Mammalian Genotyping Service ([4]). In the present study, seven additional STR markers were genotyped on chromosome 8p in the same 585 subjects from the genome scan: D8S504, D8S1806, D8S262, D8S518, D8S1706, D8S1742, D8S351. Primer sequences available in the NCBI database <http://www.ncbi.nlm.nih.gov/> were used to design assays. Product sizes were resolved by capillary electrophoresis on an ABI 3100 machine (Applied Biosystems, Foster City, CA). Marker locations were based on the Marshfield map < http://research.marshfieldclinic.org/genetics/ >. Mendelian inconsistencies were assessed using Pedcheck ([16]).

Four haplotype-tagging SNPs ([17]) and one additional coding SNP in DEFB1 were genotyped in the Boston Early-Onset COPD Study families using unlabeled mini-sequencing and mass spectrometry on the Sequenom (San Diego, CA) platform. These 5 SNPs are listed in Table 1A. Assay details are available at <http://www.innateimmunity.net/>. Four SNPs and a 3 bp insertion-deletion (indel) in MSR1 were genotyped in both the Boston Early-Onset COPD Study and in the NETT-NAS case-control study (Table 1B). Assays were designed based on flanking sequences published by Xu and colleagues ([18], [19]). SNPs were genotyped using Sequenom, and the indel was genotyped with the 5′ to 3′ exonuclease TaqMan assay (Applied Biosystems).

Table 1A DEFB1 polymorphisms genotyped in the Boston Early-Onset COPD Study. Assay details are available at http://www.innateimmunity.net/

rs#	IIPGA Identifier	Role	Allele frequency (founders)
rs5743402	IIPGA DEFB1 666	5′ genomic	0.22
rs2741136	IIPGA DEFB1 755	5′ genomic	0.28
rs2738182	IIPGA DEFB1 2181	5′ genomic	0.35
rs2702945	IIPGA DEFB1 3263	Intron	0.22
rs2738047	—	Exon, Val38Ile	0.005

*IIPGA = Innate Immunity Program for Genomic Applications.

Table 1B MSR1 polymorphisms genotyped in the Boston Early-Onset COPD Study. Assays were designed based on flanking sequences as per Xu et al. ([18])

Marker (names per Xu et al.)	rs#	Role	Allele frequency (founders)
PRO3 A > G	rs433235	Promoter	0.083
IVS5-59 C > A	—	Intron	0.049
P275A C > G	rs3747531	Exon, Pro275Ala	0.067
R293X C > T	—	Exon, Arg293X	0.029
INDEL7 3bp insertion-deletion	—	Intron	0.077

*By the extended pedigree family-based association test (PBAT): Moderate-to-severe airflow obstruction (FEV ₁ < 60% predicted, with FEV ₁/FVC < 90% predicted), post-bronchodilator: p = 0.01.

Mild-to-severe airflow obstruction (FEV ₁ < 80% predicted, with FEV ₁/FVC < 90% predicted), post-bronchodilator: p = 0.02.

Statistical analysis

Multipoint linkage analysis was performed using the variance component method in SOLAR ([20]). Post-bronchodilator spirometric traits were analyzed, adjusting for significant covariates ([5]). Stratified linkage analysis in smokers-only was performed by setting phenotypes to missing in lifelong non-smokers. Association testing in the Boston Early-Onset COPD Study was performed using the extended pedigree family-based association test (PBAT) ([21]). Additive genetic models were assumed, in models adjusting for age, sex, height, ever-smoking status, and pack-years of smoking, including quadratic terms for continuous variables. Significance was tested under the null hypothesis of linkage but no association. Haplotype analysis was also performed in PBAT. Linkage disequilibrium in the probands was expressed as pairwise r², calculated using Haploview ([22]).

In the NETT-NAS case-control study, MSR1 allele frequencies were compared in cases and controls using SAS (SAS Institute, Cary, NC). Further investigation for the single SNP (Pro275Ala) showing a trend toward association (p < 0.1) included logistic regression, adjusting for age and pack-years, under both additive and dominant genetic models. Haplotypes were analyzed with the expectation-maximization algorithm and score tests, using haplo.stats ([23]). Power for the case-control study was estimated using Quanto ([24]).

RESULTS

Linkage analysis

In the initial genome scan linkage analysis ([5]), the LOD score for post-bronchodilator FEV₁, was 3.30 (at 2 cM), representing borderline genome-wide significance. With the addition of 7 STR markers, the LOD score decreased to 1.80 (at 1 cM), in a model adjusted for age, sex, height, pack-years of smoking, (pack-years)², and race, as in the initial scan (Table 2). With the additional markers, the LOD scores for linkage of pre-bronchodilator FEV₁ and pre- and post-bronchodilator FEV₁/FVC ratio were all less than one (pre-bronchodilator data not shown). LOD scores were further attenuated in the stratified analysis of smokers only, reflecting reduced power due to the smaller sample size (Table 2).

Table 2 Chromosome 8p linkage analysis of post-bronchodilator FEV₁ and FEV₁/FVC in the Boston Early-Onset COPD Study. Maximum LOD scores are shown

Phenotype (post-bronchodilator)	Initial genome scan (Palmer et al., 2003)	Including additional markers	Smokers only, with additional markers
FEV1	3.30 (2 cM)	1.80 (1 cM)	0.35 (2 cM)
FEV1/FVC^†	0.46 (6 cM)	0.40 (0 cM)	0.03 (0 cM)

*Models adjusted for age, sex, height, pack-years, (pack-years)², race.

^†Models adjusted for age, pack-years, (pack-years)².

Defensin Beta-1 (DEFB1)

Five SNPs in DEFB1 were genotyped in the Boston Early-Onset COPD Study families; allele frequencies in the founders are shown in Table 1A. All SNPs were in Hardy-Weinberg Equilibrium (HWE) in the founders. Using the extended pedigree family-based association test (PBAT), none of the 4 haplotype tagging (ht-) SNPs were significantly associated with quantitative or qualitative spirometric traits. The coding SNP was too rare (minor allele frequency < 1%) to be analyzed in the families.

In the haplotype analysis of the 4 ht-SNPs in the Boston Early-Onset COPD Study families, the haplotype consisting of the common alleles of all 4 SNPs (A-T-T-G, frequency 23%) was associated with a greater FEV₁/FVC ratio, pre-bronchodilator (p = 0.03) (Table 3). None of the other 4 SNP haplotypes were associated with the spirometric traits. In the probands, two non-adjacent SNPs, 755 and 3263 (see Table 1A for SNP numbering), were found to be in linkage disequilibrium (r² = 0.65). Pairwise r² for the 755 and 2181 SNPs was 0.15. None of the other pairs of SNPs were in LD (r² < 0.1 for all other pairwise comparisons). Since the DEFB1 SNPs were not significantly associated in the Boston Early-Onset COPD Study families, they were not genotyped in the NETT-NAS case-control study.

Table 3 DEFB1 haplotype analysis in the Boston Early-Onset COPD Study. Haplotypes with frequency ≥ 5% are shown. SNP identifiers are from Table 1A

DEFB1	DEFB1	DEFB1	DEFB1
666	755	2181	3263	Frequency	Phenotype
A	T	T	G	0.23	FEV1/FVC (pre-BD), p = 0.03
G	T	T	G	0.12	–
A	C	T	G	0.09	–
A	T	A	G	0.28	–
A	C	T	A	0.15	–

Macrophage scavenger receptor-1 (MSR1)

Four SNPs and a 3 bp insertion-deletion (indel) in MSR1 were genotyped in the Boston Early-Onset COPD Study and in the case-control study. All markers were in HWE in the Boston Early-Onset COPD Study founders and in the NAS controls. In the family study, all SNPs had a minor allele frequency less than 10% (Table 1B). Using PBAT, the Pro275Ala coding SNP (rs3747531) was associated with the post-bronchodilator qualitative traits moderate-to-severe airflow obstruction (p = 0.01) and mild-to-severe airflow obstruction (p = 0.02), defined by FEV₁/FVC < 90% predicted and FEV₁ < 60% predicted or FEV ₁ < 80% predicted, respectively. Carriers of the rare allele had a lower risk of airflow obstruction.

Allele frequencies for the MSR1 variants in the NETT-NAS case-control study are shown in Table 4. For the Pro275Ala SNP, the allele frequency difference between cases and controls showed a trend towards significance (p = 0.07). In an additive genetic model, using logistic regression to adjust for age and pack-years of smoking, the trend towards significance persisted (p = 0.07). Since the previous abstract ([11]) had used a dominant model for this SNP, we also preformed logistic regression under a dominant genetic model, finding a significant association with COPD (p = 0.04). In the dominant model, the minor allele was underrepresented in the cases (OR = 0.52; 95% CI 0.27, 0.97), similar to the effect in the Boston Early-Onset COPD Study families, but opposite to the findings in the previous report ([11]) where the minor allele frequency was higher in the cases. However, the same group found the Ala allele to be underrepresented in prostate cancer cases ([18]).

Table 4 MSR1 polymorphisms in the NETT-NAS case-control study. Marker names are from Table 1B

Marker	Allele frequency: cases	Allele frequency: controls
PRO3	0.083	0.082
IVS5-59	0.036	0.043
P275A	0.039	0.060
R293X	0.014	0.016
INDEL7	0.040	0.049

*p = 0.07 vs. controls (allele test).

Haplotype results in MSR1 in the family study are reported in Table 5. One haplotype of the 5 markers (frequency 3%) was associated with a lower risk of moderate-to-severe airflow obstruction, post-bronchodilator (p = 0.04). None of the other 5-marker haplotypes was significant. In the probands, LD was found between P275A and INDEL7 (r² = 0.75) and PRO3 and IVS5-59 (r² = 0.34). None of the other marker pairs were in LD (r² < 0.1). In the NETT-NAS case-control study, the 5-marker haplotype was not associated with COPD. LD patterns were similar to the early-onset COPD probands.

Table 5 MSR1 haplotype analysis in the Boston Early-Onset COPD Study. Haplotypes with frequency ≥ 1% are shown. Marker names are from Table 1B

PRO3	IVS5-59	P275A	R293X	INDEL7	Frequency	Phenotype, post-BD
A	C	C	C	del	0.84	—
G	C	C	C	del	0.05	—
G	A	C	C	del	0.03	Mod-to-severe, p = 0.04
A	C	C	C	ins	0.01	—
A	C	G	C	ins	0.02	—

*Moderate-to-severe airflow obstruction: FEV₁ < 60% predicted, with FEV₁/FVC < 90% predicted.

In the case-control study, we estimated the power to detect an association with COPD for SNPs with minor allele frequencies of 5 and 10%, similar to the actual MSR1 SNPs. For a 5% SNP, there was 90% power to detect a genotype relative risk of 2 (additive model, alpha = 0.05); for a 10% SNP, there was greater than 90% power to detect a relative risk of 1.75.

DISCUSSION

After genotyping additional STR markers on chromosome 8p, the LOD score for linkage to FEV₁ (post-bronchodilator) decreased in the Boston Early-Onset COPD Study families. In the initial genome-wide linkage scan, the LOD score of 3.30 represented borderline significant evidence of linkage, but the LOD score of 1.80 with the additional STR markers did not even reach suggestive evidence ([8]). Two candidate genes on chromosome 8p have been previously associated with COPD ([9], [10], [11]). DEFB1 is located within the 1.5 LOD support interval (analogous to a 95% confidence interval) for the linkage to FEV₁ in the initial genome scan ([5]). MSR1 is located outside of the support interval, but was studied because of an abstract reporting a significant association with COPD ([11]).

Five SNPs in DEFB1 were not associated with quantitative or qualitative traits in the Boston Early-Onset COPD Study families. There were marginal associations in the haplotype analysis, though the association with FEV₁, the phenotype with the highest LOD score in the linkage analysis, was not significant. Based on these results, DEFB1 was not genotyped in the case-control study. A coding variant in MSR1 (Pro275Ala) was associated with qualitative spirometric traits (but not with FEV₁ as a quantitative trait) in the family-based study. The same SNP showed a trend towards association in the case-control study.

The highest LOD score for linkage with FEV₁ in the initial genome scan was located near the p-terminus of chromosome 8. Though the LOD score of 3.30 was based on a multipoint analysis, the linkage signal was largely due to a single marker (ATA27A03 at 2.59 cM, twopoint LOD = 3.40). Since the additional markers led to a decrease in the LOD score, it is possible that the initial linkage was a false positive result. Though not significant or even suggestive evidence of linkage, the LOD score of 1.80 when the additional markers were included does not eliminate the possibility that a COPD susceptibility gene is located on chromosome 8p. The limited sample size may also explain the lack of significant evidence for linkage in the region, though this sample size was sufficient to detect the LOD of 3.30 in the initial genome screen and to identify TGFB1 on chromosome 19q and SERPINE2 on chromosome 2q as potential COPD susceptibility genes ([6], [7]). The power to detect a valid linkage signal should be increased with the inclusion of additional markers on chromosome 8p in the present analysis.

The human Defensin Beta-1 gene product is found in the airway and is involved in the innate immune response ([25]). The gene product is an antimicrobial peptide that is also a chemoattractant for inflammatory cells (26). COPD is characterized by airway inflammation, in which chronic infection may play a role ([27]). In addition to a biologically plausible role, human studies have shown polymorphisms in DEFB1 to be associated with COPD ([9], [10]) as well as with asthma ([17]). However, we were not able to replicate these associations in our study. Our analysis may be limited by the fact that we genotyped only 5 SNPs in the gene. But four of these SNPs were chosen as haplotype-tagging SNPs based on publicly-available sequencing data ([17]), so we should have been able to capture the common variation across the gene. In addition, the fifth SNP genotyped was a coding SNP, which was significantly associated with COPD in one of the previous papers ([9]).

The product of the Macrophage Scavenger Receptor-1 gene is expressed in alveolar macrophages ([28]) and is involved in innate immunity ([29]). MSR1 may be important for the accumulation of macrophages in the lungs in cigarette smokers ([30]). Linkage and association studies have pointed out the role of MSR1 in prostate cancer susceptibility ([18], [19]). A coding SNP in MSR1 (Pro275Ala) has been reported to be associated with COPD ([11]). We found this same SNP to be associated with COPD-related qualitative traits in the Boston Early-Onset COPD Study families. This result must be interpreted with caution, since p-values were marginal and would not remain significant after adjustment for the multiple SNPs and phenotypes studied. Replication in the case-control study could provide a safeguard against false positives due to multiple testing, but there was only a trend towards significance in this analysis, despite excellent power to detect a substantial genetic effect. In addition, in the Boston Early-Onset COPD Study families, the Pro275Ala SNP was not found to be associated with FEV₁, the trait that showed the strongest evidence for linkage. As with DEFB1, we only tested 5 SNPs in MSR1, but these were common variants identified through sequencing by Xu and colleagues ([18], [19]).

In addition to the genes we studied, other potential COPD candidate genes are located on chromosome 8p. For example, several alpha defensin genes are in the region. Human neutrophil peptides, the products of the alpha defensin genes, are important in modulating lung inflammation in alpha 1-antitrypsin deficiency, a known genetic cause of emphysema ([31]). However, chromosome 8p contains a large repeated region that includes several beta-defensin genes (besides DEFB1) ([32]). This copy number polymorphism makes it difficult to accurately genotype variants in these genes with existing genotyping methodologies. Newer methods that allow for accurate assessment of copy number variation may potentially be useful for study of this region ([33]).

Even though we were unable to conclusively identify a COPD susceptibility gene, chromosome 8p remains an interesting region for further study. Other lines of evidence suggest a possible role for a gene or genes on chromosome 8p in COPD. Chromosome 8p has been linked to asthma in a genome-wide linkage scan in the Hutterite population ([34]) and in an additional analysis of African-American families in the Collaborative Study on the Genetics of Asthma ([35]). Loss of heterozygosity on chromosome 8p21-23 has been found in lung cancer, another smoking-related lung disease ([36]).

In summary, in the Boston Early-Onset COPD Study families, we found a reduction in the LOD score for linkage of FEV₁ (post-bronchodilator) to chromosome 8p when additional markers were included in the analysis, yet a weak linkage signal was still present in the region. A coding variant in MSR1, one of the two candidate genes in this region that have been previously associated with COPD, was marginally associated with COPD-related traits in the Boston Early-Onset COPD Study families, and showed a trend for association in a case-control analysis. The use of genome-wide linkage analysis followed by genotyping of positional candidate genes has thus far identified two potential COPD susceptibility genes, TGFB1 on chromosome 19q ([6]) and SERPINE2 on chromosome 2q ([7]), highlighting the potential of this technique to find genes that influence COPD. Further research will be necessary to identify a potential COPD-susceptibility locus on chromosome 8p.

Supported by U.S. National Institutes of Health grants HL080242 (CPH), HL71393 (EKS), and HL075478 (EKS), a grant from the Alpha-1 Foundation (CPH), and an American Lung Association Career Investigator Award (EKS). The National Emphysema Treatment Trial (NETT) was supported by the U.S. National Heart, Lung, and Blood Institute contracts N01HR76101-N01HR76116, N01HR76118, and N01HR76119; the Centers for Medicare and Medicaid Services; and the Agency for Healthcare Research and Quality. The Normative Aging Study is supported by the Cooperative Studies Program/ERIC of the U.S. Department of Veterans Affairs and is a component of the Massachusetts Veterans Epidemiology Research and Information Center (MAVERIC).

The authors thank Alison Brown for her assistance with genotyping and Drs. Scott Weiss, Frank Speizer, Jeffrey Drazen, Hal Chapman, and Leo Ginns for their assistance in developing the Boston Early-Onset COPD Study.