Beta-2 Adrenergic Receptor (ADRB2) Gene Polymorphisms and the Risk of Asthma: A Meta-Analysis of Case-Control Studies

Background and Objective A number of studies have assessed the relationship between beta-2 adrenergic receptor (ADRB2) gene polymorphisms and asthma risk. However, the results are inconsistent. A meta-analysis that focused on the association between asthma and all ADRB2 polymorphisms with at least three case-control studies was thus performed. Methods A literature search of the PubMed, Embase, Web of Science, CNKI, and Wangfang databases was conducted. Odds ratios with 95% confidence intervals were used to assess the strength of associations. Results Arg16Gly, Gln27Glu, Thr164Ile, and Arg19Cys single nucleotide polymorphisms (SNPs) were identified in 46 case-control studies. The results showed that not all of the SNPs were associated with asthma in the overall population. Significant associations were found for the Arg16Gly polymorphism in the South American population via dominant model comparison (OR = 1.754, 95% CI = 1.179–2.609, I2 = 16.9%, studies  = 2, case  = 314, control  = 237) in an analysis stratified by ethnicity. For the Gln27Glu polymorphism, a protective association was found in children via recessive model comparison (OR = 0.566, 95% CI = 0.417–0.769, I2 = 0.0%, studies  = 11, case  = 1693, control  =  502) and homozygote genotype comparison (OR = 0.610, 95% CI = 0.434–0.856, I2 = 0.0%, studies  = 11, case  = 1693, control  = 1502), and in adults via dominant model comparison (OR = 0.864, 95% CI = 0.768–0.971, I2 = 46.9%, n = 18, case  = 3160, control  = 3433). Conclusions None of the ADRB2 gene polymorphisms were reproducibly associated with a risk of asthma across ethnic groups in the general population.


Introduction
Asthma, which is characterized by variable airway obstruction caused by bronchial hyper-reactivity and airway inflammation, is one of the most common chronic respiratory diseases worldwide. The prevalence of asthma varies worldwide, ranging from 0.2% in China to 21.0% in Australia [1]. Recent studies show that asthma is a genetically related disease, with heritability estimates varying between 48% and 79% [2]. An increasing number of studies are focusing on asthma genetics research. Therefore, the identification of asthma susceptibility genes contributing to asthma pathogenesis is important. Candidate-gene linkage studies, positional cloning, and genome-wide association studies (GWAS) have already identified a large number of asthma susceptibility genes, and one of these, the beta-2 adrenergic receptor (ADRB2, also known as b2-AR) gene, has been extensively studied.
The b2-AR (ADRB2), a member of the G protein-coupled receptor (GPCR) family, is abundantly expressed on bronchial smooth muscle cells, and specifically binds and is activated by a class of ligands known as catecholamines, and epinephrine in particular [3]. The activation of b2-AR can result in the expansion of the small airways, and thus b2-AR agonists are used in first-line bronchodilator therapy in asthma [4]. The b2-AR, which can directly influence the effect of beta-2 adrenergic bronchodilator, is encoded by an intronless gene located on chromosome 5q31-32 [5]. It has been reported that ADRB2 variants are associated with airway hypersensitivity, asthma severity, and the response to medications [6,7]. Several single nucleotide polymorphisms (SNPs), including Arg16Gly (A46G, rs1042713), Gln27Glu (C79G, rs1042714), and Thr164Ile (C491T, rs1800888) have been identified in the coding region of the ADRB2 gene [8]. Replacement of the base may not only alter the gene expression and function of the b2-AR, it may also alter the response to b2-AR agonist therapies and even increase the risk of asthma.
To date, various case-control studies have been conducted to investigate the relationship between ADRB2 gene polymorphisms and asthma risk in different population groups [9][10][11][12][13], but the results have been conflicting and inconclusive. One reason for this inconsistency may be the typically small sample size of the individual studies, which may mean that there was insufficient

Inclusion and exclusion criteria
Studies that fulfilled the following criteria were incorporated into the meta-analysis: (1) case-control studies that evaluated the association between ADRB2 gene polymorphisms and risk of asthma; (2) the genotype distributions or allele frequency of each study was available or sufficient data could be extracted for calculating the odds ratio (OR) with 95% confidence interval (CI). For overlapping studies, the one with the most suitable data was selected. Studies were only excluded if they did not meet these inclusion criteria.

Data extraction
The basic information extracted for each study was as follows: name of first author, publication year, country and ethnicity of case control, age of case, asthma definition, sample size, and genotype frequencies in cases and controls.

Statistical analysis
Pearson's chi-square test was performed to evaluate whether the genotype distribution deviated from Hardy-Weinberg equilibrium (HWE) in the control group. Significantly deviating samples were re-assessed by 1000 time Montecarlo permutation analysis using the freely available software at http://krunch.med.yale.edu/ hwsim. The OR with 95% CI was used to assess the strength of the association between ADRB2 polymorphism and asthma risk. The pooled OR for ADRB2 polymorphisms and asthma risk was performed for four genetic model comparisons ( In the current study, the aa genotype was a wild-type, while the AA genotype was a mutant. The Q-test and I 2 test were used to assess the effect of heterogeneity. Heterogeneity was considered statistically significant when Q-test (P,0.10) or I 2 .50%. If heterogeneity was indicated, data were combined according to the random-effects model; when the Q-test (P.0.10) or I 2 ,50%, the fixed-effect model was used. Stratified analysis was performed by 1000 time permutation HWE P-value, ethnicity and case age to further explore HWE-specific, ethnicity-specific and age-specific effects. Sensitivity analysis was conducted by sequentially excluding one study at a time to examine the effect of each study on the combined result. Potential publication bias was investigated through the funnel plot and further assessed using Egger's test. A cumulative analysis was conducted after sorting by publication date. All statistical analyses of this meta-analysis were performed using the computer software STATA 11.0 (State Corp., College Station, TX, USA).

Characteristics of included studies
After a comprehensive search of the PubMed, Embase, Web of Science, Wanfang, and CNKI databases, 1154 articles were identified, 948 of which were subsequently excluded because they were not relevant to ADRB2 polymorphisms and asthma risk. Thus, 206 relevant records were identified. Of these, 121 were excluded due to the lack of a case-control design. Of the remaining 85 articles, 26 were excluded due to overlapping data. Therefore, 59 articles were identified for further study. Of these 59 articles, four [17][18][19][20] were excluded as they were conference abstracts, seven [12,[21][22][23][24][25][26] did not report useable data, and one [27] was excluded because the full text was not available. In addition, one article [28] was excluded as it was in Polish. Ultimately, 46 articles [8-11,13,29-69] met the inclusion criteria ( Figure 1). The   Table 1. Of these 46 articles, one [64] contained two independent studies, so the data were extracted accordingly. Furthermore, one article [65] did not provide the genotype distribution or allele frequency data, but these data were obtained from another study [15], so this article [65] was still included. Of these 46 case-control studies, three [51,59,64] only provided data on allele frequency and not on genotype distribution. Further analysis was performed on the ADRB2 polymorphisms that had been reported in at least three case-control studies. A total of four SNPs met the inclusion criteria: Arg16Gly (A46G, rs1042713), Gln27Glu (C79G, rs1042714), Thr164Ile (C491T, rs1800888), and Arg19Cys (T-47C, rs1042711). Some of the included studies only focused on the Chinese population, so a meta-analysis of the Chinese population was performed independently. The genotype and allele distribution for the four SNPs are shown in Tables 2 to 5.

HWE for included studies
The HWE for each included study was calculated by chi-square test. The P-value of the genotype distribution in each control group is shown in Tables 2 to 5. As some of the included studies were not in HWE, a stratified analysis according to the P-value for the Arg16Gly and Gln27Glu polymorphisms was conducted. The results are shown in Table 6.

Meta-analysis of ADRB2 polymorphisms and asthma
Meta-analysis of Arg16Gly variants and asthma. For Arg16Gly, there was no significant association in any of the genetic model comparisons in the overall population (Figures 2 to 5). In the analysis stratified by ethnicity, a significant association was found in the South American population in the dominant model comparison (OR = 1.754, 95% CI = 1.179-2.609, I 2 = 16.9%, studies = 2, case = 314, control = 237), but not in the other genetic comparisons or other ethnic groups. In the Chinese population, there was no significant association in any of the genetic model comparisons. The results are shown in Table 6.
Meta-analysis of Gln27Glu variants and asthma. For Gln27Glu, no evidence of an association with asthma risk was found in the overall population in any of the genetic model comparisons (Figures 6 to 9). In the analysis stratified by case age, a protective association was found in children only in the recessive model comparison (OR = 0.566, 95% CI = 0.417-0.769, I 2 = 0.0%, studies = 11, case = 1693, control = 1502) and homozygote genotype comparison (OR = 0.610, 95% CI = 0.434-0.856, I 2 = 0.0%, studies = 11, case = 1693, control = 1502), and in adults only in the dominant model comparison (OR = 0.864, 95% CI = 0.768-0.971, I 2 = 46.9% n = 18, case = 3160, control = 3433). In the Chinese population, there was no significant association in any of the genetic model comparisons. The results are shown in Table 6.
Meta-analysis of Thr164Ile variants and asthma. For Thr164Ile, only four case-control studies were included, so no stratified analysis was performed. There was no evidence of an association with asthma risk in any of the genetic models in the overall population. The results are shown in Table 6.
Meta-analysis of Arg19Cys variants and asthma. For Arg19Cys, only three case-control studies provided genotype distribution data, therefore no stratified analysis was conducted. No significant association was found in the overall population in any of the genetic models. The results are shown in Table 6.

Cumulative meta-analysis
Cumulative analysis of the association between Arg16Gly and Gln27Glu polymorphisms and the risk of asthma was performed  Table 2. Genotype and allele distributions in the meta-analysis for Arg16Gly (rs1042713).    Table 3. Genotype and allele distributions in the meta-analysis for Gln27Glu (rs1042714).

Sensitivity analysis
Sensitivity analysis was conducted by sequentially excluding individual studies to estimate the stability of the results. After sequentially excluding each study, statistically similar results were found.

Publication bias
Potential publication bias was investigated using the funnel plot and was further assessed using Egger's test. Significant publication bias was detected for the Gln27Glu polymorphism in the dominant model comparison (t = 2.69, P = 0.011). No evidence of publication bias was found for the Arg16Gly, Thr164Ile, or Arg19Cys polymorphism in any of the genetic model comparisons. The results are shown in Table 7.

Discussion
Asthma is a well-known disease of the respiratory system that is characterized by cramps and obstruction of the small bronchus. B2-AR binds specifically to a class of ligands that can lead to the expansion of the small airways. In the present study, the relationship between all related ADRB2 gene polymorphisms and the overall risk of asthma was examined. The purpose of this meta-analysis was to provide more information for asthma candidate gene research, based on the hypothesis that genetic effects vary across different ethnic cohorts.
Four ADRB2 polymorphisms that had been investigated in at least three case-control studies were included in the study. The results indicated that Arg16Gly, Gln27Glu, Thr164Ile, and Arg19Cys were not associated with risk of asthma in the overall population. The findings of the current study are consistent with those of Migita [14] and Contopoulos-Ioannidis [6]. Migita and his colleagues performed a meta-analysis by a random-effects model that showed a non-significant odds ratio for the Arg16Gly and the Gln27Glu polymorphism. Contopoulos-Ioannidis found that polymorphisms of ADRB2 are not major risk factors for the development of asthma. Cumulative analysis further confirmed that there was no significant association between the Arg16Gly polymorphism or the Gln27Glu polymorphism and the risk of asthma, showing that the variants had no effect with the accumulation of more data over time.
In the analysis stratified by case age, a protective effect for the Gln27Glu polymorphism was observed in adults in the dominant model comparison and in children in the recessive model comparison and the homozygote genotype comparison. This finding corroborates the ideas of Ammarin Thakkinstian, who suggested that the Gln/Glu and Glu/Glu genotypes could reduce the risk of asthma [15]. Besides, the pathogenesis of asthma in adults and children may differ, but the exact mechanism remains unknown and needs further detailed research.
In the analysis stratified by ethnicity, an increased risk of asthma was only seen with the Arg16Gly polymorphism in the South American population, and a protective effect was only found with the Gln27Glu polymorphism in the North American population and only in the dominant model comparison. The discrepancies in linkage disequilibrium (LD) structure in Chinese and Europeans may explain these differences: the minor allele of the ADRB2 Table 3. Cont.  Table 4. Genotype and allele distributions in the meta-analysis for Thr164Ile (rs1800888).  Table 5. Genotype and allele distributions in the meta-analysis for Arg19Cys (rs1042711).  Table 6. Main results of pooled ORs in the meta-analysis.                  there was no significant association with the risk of asthma with either the Arg16Gly polymorphism or the Gln27Glu polymorphism in any of the genetic model comparisons, supporting Ni Suiqin's [16] conclusion. In the analysis stratified by HWE according to the P-value for the Arg16Gly and Gln27Glu polymorphisms, a significant association was found in the recessive model comparison and the homozygote genotype comparison for Arg16Gly in the group with P,0.05, but not in the group with P.0.05. For Gln27Glu, a significant association was found in the dominant model comparison in the group with P.0.05. These results therefore need to be interpreted with caution. There are several possible explanations as to why the control group population was not in HWE. First, the population was not characterized by random mating. Second, the locus under consideration exhibited an inconstant fluctuating mutation rate. Third, there was selection for a particular phenotype. Fourth, the population was not sufficiently large or non-random. Fifth, there had been a change in the population structure during the period of study due to migration.
No significant association with the risk of asthma was found for the Thr164Ile and Arg19Cys polymorphisms. Thus, the Thr164Ile and Arg19Cys polymorphisms may not be involved in the pathogenesis of asthma. Further research is needed because, as only four case-controls were included in the study, there might not be sufficient statistical evidence to clarify the association between the Thr164Ile and Arg19Cys polymorphisms and the risk of asthma.
ADRB2 is located on chromosome 5q31-32, encodes 413 amino acids, and is an intronless gene [5]. According to the SNPper database, there are more than 100 SNPs in the promoter region, five SNPs in the 59UTR region and 18 SNPs in the coding region of the gene. The mutation of the two most important SNPs, Arg16Gly and Gln27Glu, which are located at nucleotide positions 46 and 79 of the coding region of the ADRB2 gene, respectively, can cause changes in the amino acid sequence. The altered amino acid sequence can lead to down-regulation of the b2-AR and may cause the desensitization of related reactions [70]. Thr164Ile is also located in the coding region of the ADRB2 gene; a base change from C to T can lead to a change in amino acid from threonine (Thr) to isoleucine (Ile). The missense polymorphisms of Arg16Gly, Gln27Glu, and Thr164Ile may lead to functional changes in ADRB2. Most of the studies relating to ADRB2 and asthma risk have focused on coding region polymorphisms. In recent years, studies on ADRB2 have not been confined to coding region polymorphisms alone, as more and more studies have begun to pay attention to promoter region polymorphisms. Arg19Cys is located in the 59 leader region that harbors an open reading frame (ORF) in the promoter region of the ADRB2 gene; a base change from T to C leads to a change in amino acid from arginine (Arg) to cysteine app:addword:cysteine(Cys). Recent in vivo and in vitro research has demonstrated that this change can impede the translation of ADRB2 mRNA, and thus can regulate cellular expression of the receptor [71]. Further studies are therefore required to assess whether the SNPs in ADRB2 alter signal regulation, gene expression, or the function of its product or not.
There are certain inevitable limitations to the current metaanalysis. First, all available literature should be included in the meta-analysis, but we only included literature published in English and Chinese, thus neglecting studies published in other languages. In addition, most of the included studies just focus on Chinese and Asian, which may result in an inability to detect modest association due to lack of power because of underreporting/lower incidence of asthma in these populations. Second, most original literature only provides a generic asthma definition, and does not describe asthma phenotype(s) and environmental factors in detail, so we cannot supply this information. Third, several studies were not included because they did not provide sufficient data for statistical analysis, which may have biased the result. Fourth, publication bias was only detected for the Gln27Glu polymorphism in the dominant model comparison (t = 2.69, P = 0.011), but not in the other three genetic model comparisons. In fact, positive results or results with ''expected'' findings are more likely to be published. Publication bias may lead to a false positive result. We detected significant publication bias for the Gln27Glu polymorphism in the dominant model, so the results need to be interpreted with caution. Fifth, moderate heterogeneity was found in some genetic models for the Arg16Gly polymorphism. Because no information was available other than the factors we performed a stratified analysis, and thus we were unable to use meta-regression to explore other possible sources of between-group heterogeneity. Furthermore, the result of the sensitivity analysis was stable.
Therefore, the heterogeneity seemed to have no effect on the results, suggesting their reliability.
In conclusion, the current meta-analysis suggests that the Arg16Gly, Gln27Glu, Thr164Ile, and Arg19Cys polymorphisms may not be involved in the risk of asthma in the overall population or the Chinese population. Well-designed, high-quality studies with a larger sample size and various ethnicities should be conducted to confirm these results.

Supporting Information
Checklist S1 PRISMA checklist.