Published ahead of print on November 3, 2003, doi:10.1164/rccm.200307-973OC
© 2004 American Thoracic Society Transforming Growth Factor-ß1 Promoter Polymorphism C–509T Is Associated with AsthmaDepartment of Environmental Health, Harvard School of Public Health; Division of Pulmonary and Critical Care Medicine and Channing Laboratory, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts; Department of Epidemiology and Biostatistics, Case Western Reserve University, Cleveland, Ohio; Western Australian Institute for Medical Research and Centre for Medical Research, University of Western Australia, Perth, Australia Correspondence and requests for reprints should be addressed to Eric S. Silverman, M.D., Department of Environmental Health, Harvard School of Public Health, 667 Huntington Avenue, Boston, MA 02115. E-mail: esilverm{at}hsph.harvard.edu
Transforming growth factor-ß1 (TGF-ß1) is increased in the lungs of individuals with asthma and may modulate airway inflammation and remodeling. Some genetic studies have found that a C-to-T single-nucleotide polymorphism (C–509T) in the TGF-ß1 gene promoter may be associated with altered gene expression and asthma phenotype. To build on these data, we performed a case–control association study at this locus involving 527 subjects with asthma and 170 control subjects without asthma. All individuals were white. Genotyping at 49 unlinked polymorphisms indicated that a subset of case subjects and all control subjects were well matched and without evidence of population stratification. Logistic regression was used to model the effects of age, sex, and genotype on case–control status. The diagnosis of asthma was positively associated with the T allele and TT genotype under a codominant model (odds ratio, 2.98; 95% confidence interval, 1.45 to 6.25; p = 0.003). Total serum IgE, eosinophil count, and FEV1% predicted levels were not associated with this polymorphism. Furthermore, we show that the C–509T polymorphism alters TGF-ß1 promoter–reporter activity and promoter interactions with the transcription factor Yin Yang 1. We conclude that the T allele of C–509T is associated with the diagnosis of asthma and may enhance TGF-ß1 gene transcription.
Key Words: airway remodeling • genetics • population stratification • single-nucleotide polymorphism • Yin Yang 1 transcription factor Transforming growth factor-ß1 (TGF-ß1) is a multifunctional cytokine that has been implicated in the pathogenesis of asthma. Levels of TGF-ß1 are increased in the bronchoalveolar lavage fluid of asthmatics as compared with those of nonasthmatic individuals and levels increase after allergen challenge (1–3). TGF-ß1 is expressed by airway epithelial cells, eosinophils, helper T Type 2 lymphocytes, macrophages, and fibroblasts and may be bound and stored in the subepithelial extracellular matrix of the airways (4). TGF-ß1 is important in growth, development, transformation, tissue repair, fibrosis, and the modulation of inflammatory immune responses (5), but its role in asthma is unclear. Studies in rodent models of asthma indicate that TGF-ß1 may have both antiinflammatory and profibrotic effects (6–8). The expression of TGF-ß1 is influenced by polymorphisms in the TGF-ß1 gene, and some of these polymorphisms may be associated with asthma and other diseases (9–12). In particular, there is a C-to-T promoter polymorphism at base pair position –509 (position relative to transcriptional start site defined in GenBank NM_000660 and NT_011109) that alters a Yin Yang 1 (YY1) transcription factor consensus binding site (–CCATCTC/TG–) and is associated with higher circulating concentrations of TGF-ß1 in plasma (9). It has been hypothesized that the T allele enhances the YY1 binding site on the TGF-ß1 promoter and is responsible for increased TGF-ß1 transcription (13), although this has never been demonstrated. Pulleyn and coworkers have shown that the T allele of the C–509T single-nucleotide polymorphism (SNP) is associated with the diagnosis of asthma and asthma severity (14). However, it is not clear whether C–509T is a functional variant or is the variant responsible for the genetic association with asthma. Moreover, this genetic association has not withstood replication in all asthma case–control studies, raising questions about its veracity (15). Several candidate genes and their polymorphisms have been associated with the diagnosis of asthma in case–control studies; however, only a few of these associations have been successfully reproduced in follow-up studies (16). Population stratification, multiple comparisons, and small study populations are among the many confounding factors that may limit case–control association studies and explain inconsistencies in published results (17, 18). It is clear that replication studies with different populations are essential for validation of any asthma genetic association. Therefore, we sought to replicate the association of the T allele of C–509T with asthma in a large, well characterized population of case subjects and control subjects that is without evidence of population stratification. Our results demonstrate a significant association of the T allele of C–509T with asthma. To determine whether altered YY1 binding might be responsible for the association, we examined the effect of C–509T on YY1 binding and promoter function. Our results demonstrate that the T allele enhances YY1 binding and TGF-ß1 promoter activity. Some of the results of these studies have been previously reported in the form of an abstract (19).
Study Populations and Population Stratification Analysis Case DNA was obtained from 527 white patients with asthma diagnosed according to American Thoracic Society criteria (20). Control DNA from 170 white subjects without asthma was obtained from the Environmental Medicine Genome Bank (EMGB) (21). Subjects were assessed for age, sex, history of asthma or exercise-induced bronchospasm, spirometry, and total serum IgE. Both populations are described in Table 1 and the online supplement. All DNA was purified from peripheral blood by standard techniques after subjects provided written informed consent and approval was obtained from the Brigham and Women's Hospital (Boston, MA) Institutional Review Board.
Case subjects and control subjects were well matched genetically without any evidence of population stratification, as previously described in detail (22) and in the online supplement.
Genotyping
Statistical Analysis Serum total IgE levels, eosinophil counts, and FEV1% predicted were analyzed as continuous variables in secondary analyses. IgE levels and eosinophil counts were log transformed to approximate a normal distribution. Hardy–Weinberg equilibrium was tested on a contingency table of observed versus predicted genotype frequencies, using a modified Markov chain random-walk algorithm (24). S-Plus 2000 (Mathsoft, Cambridge MA), Sib-Pair version 0.99.9 (http://www2.qimr.edu.au/davidD/), Arlequin version 2.0 (http://anthro.unige.ch/arlequin), and ldmax (http://www.well.ox.ac.uk/asthma/GOLD/) were used to manage and analyze the data. p Values were derived by empirical simulation when possible. Statistical significance was defined at the standard 5% level.
Cell Culture
Electrophoretic Mobility Shift Assay
Reporter Constructs and Transient Transfection Analysis A549 cells were transiently transfected with SuperFect transfection reagent (Qiagen Sciences, Germantown, MD), according to the manufacturer's instructions, and incubated for 24 hours at 37°C in growth medium. Luciferase assays were performed with reporter lysis buffer and assay reagent (Promega) according to the manufacturer's instructions. Luciferase data were normalized to ß-galactosidase activity. Three independently synthesized plasmids were compared for each allele, and data were pooled. Results are expressed as means ± SEM and compared by Student t-test.
Association Study General characteristics. The study population consisted of 527 case subjects and 170 control subjects; its characteristics are described in Table 1. The sex and age of asthma case subjects and nonasthmatic control subjects were significantly different, reflecting the tendency of young males to enlist in the army. The total serum IgE levels were significantly increased and the FEV1% predicted values were significantly decreased in the asthma case subjects compared with the nonasthmatic control subjects.
Primary outcomes analysis.
The results of the stratification analysis of the 49 unlinked marker SNPs relative to C–509T are shown in Figure 1 as a frequency distribution of  2 values for alleles at 49 unlinked stratification SNPs and the C–509T SNP versus asthma case–control status. For this data set, the overall stratification test statistic was set at 249 = 47.9, p = 0.48 (22). A close concordance was found between expected and observed 2 values across the 49 SNPs, indicating the absence of population stratification.
Secondary outcomes analysis. Bivariate analysis within the case group did not suggest any significant associations between the C–509T genotype and total serum IgE levels (F2, 255 = 0.33, p = 0.72), eosinophil count (F2, 519 = 0.02, p = 0.98), or FEV1% predicted (F2, 494 = 1.53, p = 0.22). Multivariate analyses confirmed the lack of association between C–509T and these secondary outcomes (data not shown).
Effect of C–509T SNP on DNA–Protein Interactions and TGF-ß1 Promoter Activity In the presence of A549 nuclear extracts, four gel-shift bands (a–d) are consistently produced by radiolabeled ODNs of both alleles (Figure 2) . A similar pattern was produced by human bronchial epithelial cells (data shown in Figure E1 of the online supplement). Band d was identified as YY1 by supershift analysis using antibodies to YY1, whether the C allele (Figure 2) or the T allele (data not shown) was present. Attempts to supershift bands a–c with control antibodies to SP1 and AP2 transcription factors failed (Figure 2), and the identity of these nuclear proteins remains unknown.
Relative band densitometry measurements for all bands are shown in Figure 3 . YY1 was consistently more intense with T allele ODNs than with C allele ODNs (p = 0.0002, n = 8 each, two ODN preparations for each allele). Band c was also more intense with T allele ODNs than with C allele ODNs. In contrast, bands a and b were more intense with C allele ODNs than with T allele ODNs. Similar relative intensities were found with human bronchial epithelial cell nuclear extracts (data shown in Figure E1 of the online supplement).
Transient transfection analysis. We hypothesized that YY1 is an activator of the TGF-ß1 promoter and that C–509T affects TGF-ß1 gene transcription by altering the affinity of YY1, and possibly other transcription factors, for the promoter. We further speculated that increased binding of YY1 conferred by the presence of the T allele of C–509T would enhance TGF-ß1 promoter activity. We tested this hypothesis by transient transfection analysis of A549 cells with TGF-ß1 promoter–reporter constructs and overexpression of YY1. The promoter region contained within these constructs has no known sequence variant other than C–508T. This is possible because C–508T is the SNP most proximal to the transcription initiation site. All other known SNPs in the promoter region of TGF-ß1 are upstream (5') of C–508T and were excluded from these constructs so that the data would reflect the functional impact of C–508T. Transfection of reporter constructs containing the C allele TGF-ß1 promoter upstream of the luciferase cDNA increased relative luciferase activity more than 10-fold compared with that of constructs without the promoter (pGL3-Basic) (Figure 4) . Cotransfection with increasing amounts of a YY1 expression construct (pcDNA-YY1) increased luciferase activity in a dose–response fashion (Figure 4). We studied the impact of C–509T on promoter function by comparing the basal activity of two constructs that were identical except for the C versus T allele at the 509 locus. We found that the presence of the T allele consistently increased luciferase activity by about 30% compared with the activity of constructs with the C allele (Figure 4). Overexpression of YY1 increased the activity of both C and T allele constructs proportionately. In contrast, YY1 overexpression did not increase pGL3-Basic luciferase expression, suggesting that YY1 is an activator of the TGF-ß1 promoter. These experiments were repeated with two additional sets of independently synthesized reporter plasmids with similar results (data not shown).
We conducted a case–control association study to see whether the C–509T TGF-ß1 promoter SNP influences asthma susceptibility in a relatively large population of outbred white individuals. In accordance with our primary hypothesis, we found a strong positive association between the T allele and TT genotype and the diagnosis of asthma. Secondary analyses showed no association between C–509T and IgE levels, eosinophil count, or FEV1% predicted. Furthermore, we showed for the first time that C–509T influences the affinity of the promoter for the transcription factor YY1 and other nuclear proteins. Specifically, the T allele augments YY1 binding (about 30%) and basal promoter function (about 30%) in the context of transient transfection analysis with promoter–reporter constructs. These data suggest that C–509T may be the functional variant rather than a variant in linkage disequilibrium with a functional variant. The results of our study are in general agreement with some but not all published results examining this locus in the context of asthma. Because the odds ratio of this association is relatively high for a genetically complex disease (TT versus CC genotype: odds ratio, 2.98; 95% confidence interval, 1.45 to 6.25; p = 0.003) and the C–509T SNP is relatively common in our European–American population (30% minor allele frequency), this association could signify an important and common mechanism leading to the development of asthma. This genetic association study has the following strengths: (1) TGF-ß1 has been implicated in the pathogenesis of asthma and is an excellent asthma candidate gene (4, 6–8); (2) heritability studies indicate that TGF-ß1 levels are under genetic control (9, 10); (3) TGF-ß1 is located on chromosome 19q, a genomic region linked to the diagnosis of asthma in some genome-wide scans, and is orthologous with a region of the mouse genome that has been linked to bronchial hyperresponsiveness in mice (16); (4) our study is a large association study involving almost 700 individuals (17); (5) a subset of asthma case subjects and all nonasthmatic control subjects has been tested for the potential confounding factor of population stratification and determined to be well matched in this regard (18, 28); (6) case subjects and control subjects were carefully assessed from the perspective of phenotype; and finally, (7) TGF-ß1 polymorphisms, and the C–509T SNP in particular, have already been associated with asthma and asthma severity in smaller genetic studies (14). As false-positive results are a common problem in the literature, replication is essential for the validation of any genetic association (17). Our results once again suggest that TGF-ß1 is an asthma gene and that the C–509T SNP may be the functional variant. More than 100 SNPs and other genetic variants have been identified in genes of the TGF-ß1 signaling pathway, and a few of these have been associated with disease (11). C–509T was chosen as the sole candidate SNP because previous studies suggest that it is associated with altered serum levels of TGF-ß1, asthma diagnosis, asthma severity, and serum IgE levels (9, 13, 14). For example, in a study involving 84 monozygous and 86 dizygous twins, Grainger and coworkers showed that TGF-ß1 levels in plasma are under genetic control (heritability estimate, 0.54) with the C–509T SNP responsible for 8.2% of the additive genetic variance (9). The T allele was associated with higher levels of TGF-ß1 than was the C allele, and there was an allele dose effect with highest levels in TT individuals (7.62 ng/ml), intermediate levels in CT individuals (5.06 ng/ml), and lowest levels in CC individuals (3.83 ng/ml) (9). In a study by Pulleyn and coworkers, 122 individuals with severe asthma (FEV1% predicted, 63.5%), 91 individuals with mild asthma (FEV1% predicted, 90.9%), and 122 nonasthmatic control subjects were genotyped at four SNPs in the TGF-ß1 gene (14). Only C–509T was significantly (p = 0.016) stratified among with groups, with the TT genotype found in 13.9% of individuals with severe asthma, 7.7% of individuals with mild asthma, and 2.5% of control subjects. Because the differences between the mild asthma group and control group were smaller than those found between the mild asthma group and the severe asthma group, the authors concluded that C–509T may be associated with asthma severity. Although our secondary outcomes analysis showed no association between FEV1% predicted and genotype in our asthmatics (F2, 494 = 1.53, p = 0.22), the results of our association study are strikingly similar to the findings of Pulleyn and coworkers. TGF-ß1 is a multifunctional cytokine that is increased in the airways of individuals with asthma compared with those without asthma and is further increased in patients with status asthmaticus (1, 3, 29). The increased TGF-ß1 is localized principally in the extracellular connective tissue of the subepithelial space of the airways in association with the binding proteoglycan decorin (4). Airway epithelial cells, eosinophils, T lymphocytes, fibroblasts, and macrophages express TGF-ß1; however, the precise cellular source of increased TGF-ß1 in the airways of individuals with asthma is unknown (2, 3). TGF-ß1 is secreted as a latent complex that must be cleaved via proteases, acid, or reactive oxygen species to become active (5). Although multiple mechanisms are involved in the control TGF-ß1 activity, transcriptional mechanisms are important and regulated by inflammatory cytokines, nitric oxide, and reactive oxygen species found in the airways of individuals with asthma (26). If C–509T impacts the transcription of TGF-ß1, C–509T could have important effects on the activity of TGF-ß1 in the airways. There are at least two possible mechanisms by which TGF-ß1 may impact the development and severity of asthma. Some studies suggest that increased TGF-ß1 has a beneficial role in asthma by suppressing airway inflammation and hyperresponsiveness through the inhibition of T lymphocytes, dendritic cells, eosinophils, and mast cells. In this way, TGF-ß1 may be part of a negative-feedback loop, turning off inflammation that augments its production (7). Under this paradigm, genetic variants that are associated with increased TGF-ß1 activity, such as the T allele of C–509T, would be expected to be associated with decreased asthma prevalence or decreased asthma severity. Other studies suggest that TGF-ß1 has harmful effects in the airways of individuals with asthma. TGF-ß1 is profibrotic, and its sustained elevation may stimulate airway remodeling. Under this paradigm, genetic variants that are associated with increased TGF-ß1 activity, such as the T allele of C–509T, would be expected to be associated with higher asthma prevalence or increased asthma severity. Our study found a positive association between the T allele of C–509T and asthma prevalence, supporting this later paradigm and a harmful role of TGF-ß1 in asthma. At least 20 SNPs in addition to C–509T span the promoter and coding region of the TGF-ß1 gene, and there is extensive linkage disequilibrium among these SNPs (see http://snpper.chip.org) (13). The association of C–509T with asthma might not be causal but rather the result of another polymorphism in linkage disequilibrium with C–509T. Our data suggest that C–509T per se may have a functional impact on TGF-ß1 transcription. We now show that C–509T has an effect on basal promoter function in A549 cells, with the T allele increasing reporter levels by about 30% compared with those of the C allele. Luedecking and coworkers studied the TGF-ß1 promoter in COS-1 cells, using similar promoter–reporter constructs, and found an approximately 24% increase in basal activity of constructs containing the T allele versus constructs with the C allele (30). Furthermore, we now show for the first time that the C allele, in comparison with the T allele, increases the EMSA band intensity of two DNA–protein complexes (bands a and b) and decreases the band intensity of two other DNA–protein complexes (bands c and d). Although the identities of these complexes remain largely unknown, the transcription factor YY1 was identified as a component of band d by supershift analysis. Other investigators have been unable to show YY1 binding to this region of the promoter for reasons that may relate to differences in EMSA binding conditions or methods used to extract nuclear proteins (14). YY1 is a ubiquitously expressed zinc finger transcription factor that may function as both an activator or repressor of gene transcription (31). Because both the YY1 affinity and promoter activity of the T allele are increased about 30% compared with the C allele, it is tempting to speculate that alterations in YY1 affinity are responsible for the association of C–509T with asthma. In conclusion, this case–control association study suggests that the C–509T SNP of the TGF-ß1 gene is an important susceptibility locus for asthma. We speculate that the T allele of C–509T contributes to the development of asthma by increasing basal levels of TGF-ß1 gene transcription in the airways of susceptible individuals by increasing YY1 affinity for the promoter.
The authors thank Drs. Jeffrey M. Drazen and Larry A. Sonna for their insightful comments and suggestions.
Supported by National Institutes of Health grant HL70573. This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Conflict of Interest Statement: E.S.S. has no declared conflict of interest; L.J.P. has no declared conflict of interest; V.S. has no declared conflict of interest; A.H. has no declared conflict of interest; S.M. has no declared conflict of interest; J.V. has no declared conflict of interest; D.S.F. has no declared conflict of interest; T.S. has no declared conflict of interest; B.A.R. has no declared conflict of interest; S.T.W. received a grant for $900,065, Asthma Policy Modeling Study, from AstraZeneca from 1997–2003 and he was a Co-Investigator on a grant from Millennium Pharmaceuticals to pursue asthma genetics in China from 1996–2001 and received a grant from Pfizer to examine diabetes mellitus and its relationship to lung function between 2000–2003 and was a consultant for Schering-Plough and received $5,000 from 1999–2000 and has been a Co-Investigator on a grant from Boehringer Ingelheim to investigate a COPD natural history model which began in 2003 (received no funds for his involvement in this project), and has been a consultant for Variagenics on human subjects issues and received $5,000 in 2003 and has been a consultant to Genome Therapeutics in 2003 and received $1,500 and was a consultant for Merck Frost on asthma genetics in 2002 and received $2,000 and was an advisor to the TENOR Study for Genetech and has received $5,000 for 2002–2003 and grant from Glaxo-Wellcome for $500,000 for genomic equipment from 2002–2003 and was a consultant for Roche Pharmaceuticals in 2000 and received no financial remuneration for this consultancy; S.A.S. was the recipient of a $15,000 grant from Merck to study the effects of IL-9 on airway smooth muscle. Received in original form July 17, 2003; accepted in final form October 28, 2003
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
