We found a borderline significant association between one tagSNP, rs2071749, and lung function decline in smokers who had mild to moderate airway obstruction. This intronic tagSNP was in perfect LD (r2 = 1) with rs3761439, which is located in the promoter region of the HMOX1 gene. We tested rs3761439 and two other putatively functional polymorphisms (rs2071746 and the (GT)n polymorphism) in reporter gene assays but no effects on gene expression were found.
polymorphism in the HMOX1
gene was investigated in a Japanese study population [13
] and the L class of alleles (≥ 33 repeats) was significantly associated with pulmonary emphysema in smokers [13
]. The (GT)n
polymorphism was also associated with level of lung function in early-onset COPD families and with COPD [19
] but not with measures of functional impairment [20
]. In the LHS, no association was found between (GT)n
repeat classes and the rate of decline of lung function [6
]. In contrast, Guenegou et al.
reported an association between lung function decline and the (GT)n
polymorphism in the European Community Respiratory Health Survey (ECRHS) [21
]. Similar results were reported in Japanese [22
] and Dutch populations [23
Lung function traits were investigated in both the ECRHS [21
] and the LHS [6
]. However, the criteria for recruitment of subjects and the methods of analysis were different between the ECRHS and the LHS. In our study of the LHS we used a nested case control design that allowed comparison of phenotypic extremes within a cohort of smokers selected for evidence of mild/moderate airflow obstruction. Therefore, disease severity genes were investigated in the LHS, while disease susceptibility genes were investigated in the ECRHS. This may be the basis of the discrepant results between the two studies [6
]. Alternatively, our study may be a false negative result due to lower sample size. The previous studies that demonstrated an association of the (GT)n
polymorphism and lung function decline had sample sizes of 749 [21
], 101 [22
], and 1390 [23
] and therefore a larger sample size than that used in our study may be required to show the effect of this repeat polymorphism.
The tagSNP that was nominally associated with lung function decline in this study (rs2071749) was in perfect LD (r2 = 1) with a SNP that was located in the promoter region of the HMOX1 gene (rs3761439). The promoter SNP was predicted to change the sequence of an NF-κB binding site and therefore may be the causal SNP for the association. We hypothesized that the G allele of rs3761439 was associated with relative resistance to lung function decline in smokers because it caused higher expression of the HMOX1 gene in response to specific stimuli.
Therefore, we conducted a functional analysis of the HMOX1
promoter. Two polymorphisms, rs2071746 (designated A-413T by Ono et al.
]) and the (GT)n
polymorphism, both of which were reported as functional [12
], are located between rs3761439 and the transcriptional start site of the gene. Previously Yamada et al.
showed in A549 cells that H2
exposure up-regulated the transcriptional activity of the HMOX1
promoter with S alleles of (GT)n
but not with M or L alleles [13
]. In lymphoblastoid cell lines, HMOX1
mRNA expression, HMOX1 activity, and resistance to oxidant-induced apoptosis were significantly higher in cells with the S/S genotypes than those with L/L [24
]. However, it is still unknown whether the higher activities of HMOX1 were due to the (GT)n
polymorphism, because the polymorphisms in this region are in strong LD with each other. Indeed, Ono et al.
demonstrated that the A allele of rs2071746, but not the S allele of (GT)n
, was associated with higher basal promoter activity in a reporter gene assay using bovine aortic endothelial cells [12
To elucidate the functional consequences of promoter polymorphisms on HMOX1
inducibility by oxidative stress, we made nine plasmid constructs for reporter gene assays, based on the haplotype structure of the (GT)n
, rs2071746, and rs3761439 polymorphisms. We used A549 cells and stimulated them with H2
or hemin. Hemin is known as an inducer of HMOX1, and generates reactive oxygen species. However, we did not detect any significant difference of the promoter activity among the nine plasmids in the transient transfection assays, in contrast to the previous in vitro
functional studies [12
]. Our negative results in the transient transfection assays may be related to regulatory sequences that were not included in the plasmids in this study but were included in the constructs utilized in the previous studies. We did not include a pGL3-basic vector in the experimental design as we were interested in the fold increase in gene expression i.e. comparing before and after H2
and hemin stimulation. This decision was motivated by the previous study by Yamada et al.
] that demonstrated that the up-regulation of a reporter gene in response to H2
differed between promoters of different haplotypes. Nevertheless, this is a limitation of our study design.