β2 Adrenoceptor Functional Gene Variants, Obesity, and Blood Pressure Level Interactions in the General Population
We investigated the association of β2 adrenoceptor functional gene variants (Arg16Gly, Gln27Glu, and Thr164Ile polymorphisms), obesity phenotypes, and blood pressure levels in a large, ethnically mixed urban population. The individuals (n=1576) were randomly selected for a cross-sectional study of cardiovascular risk factors in Vitória, Brazil. Statistically significant associations among systolic blood pressure and the Arg16Gly and Thr164Ile variants were identified in univariate analysis. The Gly16/Gly16 genotype was still associated with systolic blood pressure (SBP) in multivariate analysis adjusting for age, gender, ethnicity, total cholesterol, diabetes, and body mass index (BMI) (P=0.01). The Arg16 allele was the only genotypic variable associated with BMI, and, in a dominant model, it remained associated with an increased BMI even after adjustment for age, gender, ethnicity, triglycerides, HDL cholesterol, LDL cholesterol, diabetes, and hypertension status (P=0.02). Although the different polymorphisms did not interact in the determination of SBP, a significant interaction with BMI (P=0.02), not through linkage disequilibrium, was identified between the Gln27Glu and the Thr164Ile variants. Furthermore, a significant interaction among the Arg16Gly polymorphism and BMI (P=0.036) and waist-hip ratio (P=0.003) in determining SBP was disclosed by ANOVA factorial modeling, with SBP used as the dependent variable. An interaction between the Thr164Ile polymorphism and waist-hip ratio was also identified (P=0.018). Finally, multiple logistic regression models showed a 1.48-fold increase in the risk of hypertension in individuals harboring the Gly16/Gly16 genotype and a 1.31-fold (P=0.01) and a 1.49-fold (P=0.003) increased risk of obesity in individuals harboring the Gln27/Gln27 genotype or the presence of the Arg16 allele, respectively. Taken together, these data provide evidence for a strong but complex relation between β-adrenoceptor gene variants, hypertension, and obesity.
Both hypertension and obesity are considered complex traits.1,2 The interplay between environmental and genetic factors is a major determinant of either final phenotype. In addition, it is currently accepted that both share not only similar risk factors but possibly similar genetic determinants.3
Despite intense effort, the genetic pathways underlying hypertension or obesity remain elusive. This is largely due to the complexities circumventing these processes, which include age of onset, quantitative variability of blood pressure phenotypes, polygenic inheritance, genetic heterogeneity, incomplete penetrance, unknown mode of action of disease alleles, effect of ethnicity, age, gender, and environmental factors, such as diet, physical activity, or smoking status, to name a few.
In this scenario, the study of components of important physiological control systems that are known to contribute to the regulation of both blood pressure and fat metabolism can offer important insights into the determination of the genetic mechanisms or pathways underlying not only interindividual blood pressure or obesity variation among humans, but, most importantly, may shed light on the complex interrelations of these two common diseases.
The components of the sympathetic nervous system are of particular interest, since they influence the control of vasomotion as well as energy expenditure associated with catecholamine metabolism.4
For instance, the adrenergic receptor β2 (ADRB2) is coupled to a stimulatory G protein that activates protein kinase, which mediates a variety of responses, depending on the cell type. In vascular smooth muscle cells, ADRB2 agonists promote a rise in intracellular cAMP concentration, leading to marked vasodilation. Other potential blood pressure regulating effects of ADRB2 include their action on renal sodium handling and the control of renin release.5 In addition, the lipolytic effects of catecholamines are mediated through members of the ADRB2 family.6
Genetic polymorphisms of ADRB2 have already been associated with obesity, diabetes mellitus, and essential hypertension, with conflicting results.7
In the present study, we have investigated the association of these genetic variants of the ADRB2 with blood pressure–related and obesity-related phenotypes in a large sample of individuals randomly selected from an ethnically mixed urban population.
A cross-sectional study of risk factors for cardiovascular diseases was performed in the urban population of Vitória, Brazil, following the general guidelines of the WHO-MONICA project.8 Ascertainment and characteristics were described previously.9
Blood Pressure Phenotype Determination
Blood pressure was measured by trained technicians, using a standard mercury sphygmomanometer on the left arm after 5 minutes’ rest with the subject in the sitting position. The first and fifth phase of Korotkoff sounds were used for systolic and diastolic pressures, respectively. Systolic and diastolic blood pressures were calculated from two readings taken by two different observers. The two measurements were obtained with a minimal interval of 10 minutes. Hypertension was defined as the mean systolic blood pressure of ≥140 mm Hg and/or diastolic blood pressure of ≥90 mm Hg.10 Pulse pressure was the difference between systolic and diastolic blood pressures.
Assessment of ADRB2 Gene Polymorphism Genotypes
Genomic DNA was extracted from leukocytes in samples of whole blood, following standard techniques. The three studied polymorphisms were detected by polymerase chain reaction–restriction fragment length polymorphism (PCR-RFLP), as previously described.11
Quality control for these assays was assessed by randomly selecting 50 samples to be regenotyped by two independent technicians.
Allele and genotype frequencies among study participants were analyzed by χ2 test and multivariate logistic regression, with the use of the Statistical Package StatView for Windows, version 5.0 (SAS Institute Inc). Correction for multiple comparisons was not performed in any of the analyses in the present study. To test for differences in various characteristics, the Student t test was used for continuous variables and the χ2 test was used for categoric variables. In addition, univariate analysis comparing continuous phenotypes was done by using correlation or simple linear regression when appropriate.
Hardy-Weinberg equilibrium for the distribution of genotypes, linkage disequilibrium statistics, haplotypic frequency estimation, and test of population differentiation were conducted with the use of Arlequin software.12 The odd ratios for different association models were calculated with 95% confidence intervals and 2-tailed probability values. Genetic models of action of the studied variants were constructed by combining genotypes (ie, dominant=heterozygous+homozygous for the polymorphism associated with increased levels of the gene product; recessive=homozygous for the polymorphism associated with increased levels of the gene product).
An unpaired t test was used to study the association between genotype and the different blood pressure phenotypes studied. Correlation between a particularly chosen genetic model and blood pressure phenotype was examined with simple and multiple linear regression.
Logistic regression analysis that allowed for age, gender, smoking status, diabetes mellitus, plasma cholesterol, HDL cholesterol, LDL cholesterol, VLDL cholesterol, and ethnicity explored the association between genotype and risk of hypertension.
In addition, a linear regression–based genetic model (ie, dominant, recessive, or additive/codominant) characterizing the effects of alleles in the studied loci was also determined by assigning simple indicator variables to the individuals, based on genotype information. The best-fitting genetic model of the three for each marker was determined as the model that produced the smallest least square error and was taken to characterize the allelic effects of each loci.13
A value of P<0.05 on a 2-sided test was considered significant.
Demographic and Genetic Structure Data
Table 1 summarizes the demographic data. In Table 2 we present allelic, genotypic, and haplotypic data for the studied polymorphisms. All studied loci allele and genotype frequencies were in accordance with Hardy-Weinberg equilibrium. Pairwise linkage disequilibrium analysis has shown that significant disequilibrium exists between the Arg16Gly and Gln27Glu (P<0.001) and the Arg16Gly and Thr164Ile (P<0.001) polymorphisms but not between Gln27Glu and Thr164Ile polymorphisms (P=0.06). Interestingly, we disclosed an ethnic-related genetic structure (P<0.001) through the use of the exact test of differentiation. The genotypic content related to the studied alleles of white individuals of this population was significantly different from the one of mulatto and black individuals. More specifically, this difference could be explained by a decrease in the allele frequency of both the Ile164 and Glu27 variants and an increase in the frequency of the Gln27 variant in mulatto and black individuals. Neither gender, age, or smoking status was associated with the studied variants in univariate analysis (data not shown).
Univariate and Multivariate Analysis Regarding Blood Pressure and Obesity
Univariate analysis testing for factors associated with systolic blood pressure have shown that gender (P=0.0007), age (P<0.0001), body mass index (BMI) (P<0.0001), waist-hip ratio (WHR) (P<0.0001), total cholesterol (P<0.0001), triglycerides (P<0.0001), LDL cholesterol (P<0.0001), VLDL cholesterol (P<0.0001), glucose (P<0.0001), diabetes status (P<0.0001), and uric acid (P<0.0001) were associated with systolic blood pressure in our population. Neither smoking status, HDL cholesterol, heart rate, or physical activity was associated with systolic blood pressure in univariate analysis (data not shown).
Univariate analysis testing for factors associated with BMI have shown that gender (P=0.004), age (P<0.0001), WHR (P<0.0001), total cholesterol (P<0.0001), triglycerides (P<0.0001), LDL cholesterol (P<0.0001), VLDL cholesterol (P<0.0001), HDL cholesterol (P<0.0001), glucose (P<0.0001), diabetes status (P<0.0001), uric acid (P<0.0001), and hypertension status (P<0.0001) were associated with BMI in our population.
The effect of the studied polymorphisms in blood pressure phenotypes in univariate analysis are shown in Table 3. Statistically significant associations between systolic blood pressure and the Arg16Gly and Thr164Ile functional variants were identified. The best-fitting genetic model for the Arg16Gly polymorphism was considering a recessive mode of action for the Gly16 allele. Regarding the Gln27Glu polymorphism, the best model was the one considering the Glu27 allele in a dominant mode of action.
We have also studied the relation between these β-adrenergic receptor functional variants and obesity-related phenotypes in our population. These data are presented in Table 4. Interestingly, a significant association was also demonstrated between the Arg16Gly polymorphism and BMI, considering a recessive mode of action of the Gly16 allele. In addition, a tendency toward increased BMI in individuals harboring Gln27 and Ile164 was also apparent. In Figure 1, we summarize the findings regarding univariate analysis of both blood pressure–related and obesity-related phenotypes in our population.
Multiple linear regression models were constructed to show the independent relation of the studied polymorphisms and blood pressure variation in this population. Presence of the Gly/Gly genotype for the Arg16Gly polymorphism was significantly associated with systolic (P=0.01) and diastolic blood pressures (P=0.006), even after adjustment for age, gender, ethnicity, total cholesterol, diabetes, and BMI. Presence of the Gln/Gln genotype of the Gln27Glu polymorphism was not associated with blood pressure in our multiple linear regression model (P=0.30). Presence of the Ile164 allele was marginally associated with systolic blood pressure in a model with adjustment for age, gender, ethnicity, total cholesterol, and diabetes (P=0.06), but this marginal association disappeared after inclusion of BMI in the model (P=0.15). This may suggest that the effect of this allele in blood pressure regulation is operant only under interaction with obesity. The same seems not to occur with the Arg16Gly polymorphism: The effect of the Gly/Gly genotype persists even after adjustment for obesity-related variables. We could not disclose any interaction between the different genetic variants studied and blood pressure (data not shown).
In addition, we have created multiple linear regression models by using BMI as the dependent variable. The only genotypic variable significantly associated with BMI in our regression model was the presence of the Arg16 allele. Presence of this allele in a dominant model was still significantly associated with increased BMI even after adjustment for age, gender, ethnicity, triglycerides, HDL cholesterol, LDL -cholesterol, diabetes status, and hypertension status (P=0.02). Interestingly, although no interaction was found between the Arg16Gly polymorphism and any of the other studied variants, a significant interaction was present between the Gln27Glu and the Thr164Ile polymorphism in determining BMI (P=0.02). This interaction is graphically presented in Figure 2A. In the absence of the Ile164 allele, the Gln27/Gln27 genotype has no effect on BMI. On the other hand, when in the presence of the Ile164 allele, there is a statistically significant association between the Gln/Gln genotype and increased BMI.
Interaction Effects Between ADRB2 Polymorphism, Blood Pressure, and Obesity
To better understand the complex relations between β-adrenoceptor gene polymorphism, blood pressure, and obesity phenotypes, we have also constructed ANOVA factorial modeling by using systolic blood pressure as the dependent variable and the best-fitting genetic model for each of the studied polymorphisms, BMI, and WHR. The Arg16Gly polymorphism significantly interacted with BMI (P=0.036) and WHR (P=0.003) to determine systolic blood pressure. As expected, a significant interaction between the Thr164Ile polymorphism and WHR was also disclosed (P=0.018). No interaction between blood pressure and obesity phenotypes was present for the Gln27Glu polymorphism. These interactions are graphically presented in Figure 3.
Because of linkage disequilibrium existence between the Arg16Gly and the Thr164Ile polymorphisms in our population, we have also studied whether one of these particular alleles could be associated with blood pressure because of linkage disequilibrium. The relation between different haplotypes of the Arg16Gly and the Thr164Ile polymorphisms and systolic blood pressure in our population is shown in Figure 2B. As can be seen, the presence of the Ile164 allele increases blood pressure both in the presence of the Arg/Arg genotype or the Gly/Gly genotype, suggesting an independent role in modulating blood pressure in this population. The same can be noticed for the Gly16 allele.
Finally, we have conducted univariate analysis by using dichotomous variables for these phenotypes: hypertension and obesity. Hypertension was significantly associated with the Gly/Gly genotype (P=0.015) and marginally with the Ile164 allele (P=0.06) but not with the Gln/Gln genotype. Obesity, on the other hand, was associated with the Arg16 allele (P=0.02) and with the Gln/Gln genotype (P=0.01) but not with the Ile164 allele. To better characterize such associations, we have constructed multiple logistic regression models, studying the relation between the studied polymorphisms and other risk factors for the development of hypertension and obesity. The only studied genetic variable that remained statistically associated with hypertension after multiple adjustment was the presence of the Gly16/Gly16 genotype. This genotype was associated with a 1.48-fold increase in the risk of presenting hypertension, even after adjustment for ethnicity, age, gender, diabetes, total cholesterol, LDL cholesterol, HDL cholesterol, and BMI. Of importance, even the addition of other studied polymorphisms to the model as independent variables did not significantly change the odds ratio associated with the Gly16/Gly16 genotype. In our multiple logistic regression model for obesity, presence of the Gln27/Gln27 genotype was associated with a 1.31-fold–increased risk of presenting obesity even after adjustment for ethnicity, age, gender, diabetes, triglycerides, total cholesterol, and hypertension status (P=0.01); presence of the Arg16 allele was associated with a 1.49-increased risk of obesity in the same model (P=0.003).
This study investigated 3 genetic variants of the ADRB2 gene as candidates to predispose to hypertension and obesity. Each of the studied polymorphisms is known to be functionally relevant. The Gly16 allele of the Arg16Gly polymorphism has been ascribed a greater degree of agonist-induced downregulation of the receptor in transfected fibroblasts and human bronchial smooth muscle and 5-fold–increased sensitivity to catecholamine-induced lipolysis in human adipocytes.11 The receptor with the Glu27 allele does not appear to be expressed in the fully mature form on Western blots, and in vitro observations have shown that the Gln27Glu substitution reduces agonist-mediated downregulation of the β2 adrenergic receptor, leading to depressed receptor degradation during agonist exposure and presumably resulting in sustained vasodilation. Finally, the 164Ile variant of the ADRB2 is associated with a decreased native adipocyte receptor function.14
Although several previous reports have described associations between these polymorphisms and hypertension or obesity,15–20 few studies have taken into account both variables in the same design. To our knowledge, this is the most extensive work done analyzing these polymorphisms and blood pressure and obesity phenotypes both as continuous and dichotomous variables in the same population. In addition, the design of our study was aimed at reducing selection bias through the enrollment of individuals who are truly representative of the general population, following a MONICA type design.
We show a significant association between the Gly16/Gly16 genotype and blood pressure phenotypes either measured as quantitative or qualitative variables. Presence of this genotype was associated with systolic and diastolic blood pressures in multiple regression models adjusted for confounding variables and increased the chance of hypertension in 48% in a multiple logistic regression model also adjusting for confounding variables. Interestingly, the other allele of this locus, the Arg16 allele, was also significantly associated with obesity phenotypes, both quantitative (BMI) and qualitative (obesity) in multiple regression models. Although not associated with blood pressure, the Gln27 allele was significantly associated with obesity (OR, 1.31 for the Gln/Gln genotype). The same allele was also associated with BMI. Interestingly, this association is operant through an interaction with the Ile164 allele.
Association reports of β2 adrenoceptor variants have been discordant regarding both hypertension and obesity. The relations between the polymorphisms of the ADRB2 gene and hypertension phenotypes have been assessed in European,18 American,19 Japanese,16,17 and African Caribbean20 populations, and the apparent lack of consistency in the results among these studies may be, at least partially, attributable to ethnic differences. Another possible explanation for the discordant data regarding association of these alleles with hypertension is a population-specific role for a particular allele. These could be explained, for example, through different gene-gene or gene-environment interactions. A third possibility is that no causal relation exists between the Arg16Gly polymorphism and hypertension, and the different associations reported are due to linkage disequilibrium with another unstudied, nearby genetic variant.
Our study design can contemplate some of these criticisms. Indeed, our analysis has shown the existence of significant population structure in our ethnically mixed individuals. Although the population structure in our ADRB2 gene haplotypes was not given because of the Arg16Gly allele, this situation could be modeled in our analysis through the adjustment of all multivariate models for ethnicity. In addition, the use of an ethnically mixed population with different genetic backgrounds (African descent versus European descent) has made it possible to analyze blood pressure trends for each genetic background in a population exposed to the same environment. It is reassuring that systolic blood pressure was higher in individuals with the Gly/Gly genotype in all ethnic subgroups of our population (white, 124.3× 126.2 mm Hg; African descent, 128.4×131.3 mm Hg; and other ethnicities, 120.2×129.7 mm Hg). This observation gives support to the idea that the Arg16Gly variant, as shown in in vitro and in vivo intermediate phenotypes, can directly influence blood pressure regulation. Still, particular gene-environmental interactions may be operant and may explain some of the discordant findings in the literature.
The association between ADRB2 polymorphisms and obesity has also been contradictory. The Glu/Glu genotype was associated with obesity in Swedish women11 and obesity and fat distribution in Japanese16,17; however, the Gln/Gln genotype was associated with obesity in French.21 Again, the same possible scenarios already pointed out for blood pressure may explain data on obesity. It is reassuring that the Arg16 allele, not previously shown to be linked to obesity, has been shown to have a 5-fold–decreased sensitivity to catecholamine-induced lipolysis in human adipocytes and, in our data, is associated with a continuous variable and a dichotomous obesity-related variable. Although linkage disequilibrium with the Gln27Glu could be modulating this association, linkage disequilibrium data between these two alleles is only of marginal statistical significance (P=0.06). In addition, ethnicity does not appear to influence these results, since it was adjusted in all multivariate models and an increased BMI was associated with the Arg16 allele despite ethnic subgroup (26.0×25.6 for whites; 26.8×25.9 for African descent). Taken together, our data suggest an association between the Arg16 allele and increased BMI and a higher risk of obesity.
It should also be noted that the effect of this locus was discordant with the known relation between obesity and hypertension. Although the presence of the Gly16 allele is associated with increased blood pressure levels, the presence of the other variant, the Arg16 allele, is associated with increased obesity. It is tempting to suggest that these discordant findings may, at least in part, have decreased the power of previous, smaller, unadjusted studies to identify a relation with blood pressure. This is also an example of how understanding of the pleiotropic actions of a particular gene/allele may shed light on aspects of a complex pathophysiological process.
Concordant with previous findings but discordant with others, we report an increased risk of obesity in individuals with the Gln27/Gln27 genotype. Interestingly, the association between Gln27/Gln27 genotype was significantly influenced by the Thr164Ile variant. This interaction has not been previously described and empowers the hypothesis that haplotype-based analyses may help to explain literature-discordant results for these polymorphisms.
Potential sources of bias when studying different single nucleotide polymorphisms (SNPs) of the same gene, especially when they can all potentially be functionally relevant, are the presence of linkage disequilibrium between SNPs and the different functional effects of the haplotypes created by these SNPs. Unfortunately, we could not define the chromosomal phase of our alleles in every participant to conduct individual haplotype analysis. Nevertheless, we have taken these potential limitations into account by using information of linkage disequilibrium derived from our own population and by testing for interactions between the studied alleles and the phenotypes of interest. This approach was only made possible because of the considerable number of individuals investigated in our study. As a result, we were able to disclose a significant interaction between the Gln27Glu and the Thr164Ile polymorphisms and BMI (Figure 2B) and analyze the association of both the Arg16Gly and the Thr164Ile allele and SBP in the context of significant linkage disequilibrium between these two genetic variants (Figure 2A). Our data generate an interesting hypothesis to be tested not only in human populations but also in the context of computer modeling and in vitro assays.
Finally, our data made it possible to investigate another unstudied aspect of this complex system: whether these genetic variants were not only influencing the interindividual variability of blood pressure and obesity but also if they could be modulating the relation between obesity and blood pressure in the general population. Interestingly, significant interactions could be disclosed between the Arg16Gly polymorphism and BMI and WHR and between the Thr164Ile polymorphism and WHR. These data can have important implications in the creation of algorithms to predict hypertension, based on known risk factors such as obesity.
The current paradigm for the understanding of complex biological systems implies the existence of multilevel interactions between genetic and environmental variables. Genetic association studies, although exploratory regarding causal relations, may have an important role in the generation of hypothesis to be tested in more controlled studies. However, they must take into account not only statistical issues such as power and selection bias but also the possible pleiotropic effect of the studied genes. Our findings add information to the controversy regarding ADRB2 genetic variants, hypertension, and obesity. The design of our study has made possible the study of blood pressure and obesity phenotypes both as quantitative and qualitative variables. As a result, the association of different cardiovascular phenotypes with the studied polymorphisms has disclosed previously undescribed associations and interactions in this multivariable system. Taken together, our data provide further evidence for a role, possibly direct, for the ADRB2 alleles in the homeostatic control of both blood pressure and fat metabolism.
This work was supported by grants from Fundação de Amparo a Pesquisa do Estado de São Paulo (FAPESP 01-00009-0), Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq 520696 and 522733), and Prefeitura de Vitória/Facitec.
- Received May 12, 2003.
- Revision received June 9, 2003.
- Accepted June 27, 2003.
Guimaraes S, Moura D. Vascular adrenoceptors: an update. Pharmacol Rev. 2001; 53: 319–356.
van Baak MA. The peripheral sympathetic nervous system in human obesity. Obes Ver. 2001; 2: 3–14.
Tomaszewski M, Brain NJ, Charchar FJ, Wang WY, Lacka B, Padmanabahn S, Clark JS, Anderson NH, Edwards HV, Zukowska-Szczechowska E, Grzeszczak W, Dominiczak AF. Essential hypertension and beta2-adrenergic receptor gene: linkage and association analysis. Hypertension. 2002; 40: 286–291.
Pereira AC, Mota GF, Cunha RS, Herbenhoff FL, Mill JG, Krieger JE. Angiotensinogen 235T allele “dosage” is associated with blood pressure phenotypes. Hypertension. 2003; 41: 25–30.
Schneider S, Roessli D, Excoffier L. Arlequin ver 2000: a software for population genetics data analysis. Geneva, Switzerland: Genetics and Biometry Laboratory, University of Geneva, Switzerland; 2000.
Shork NJ, Krieger JE, Troilliet MR, Franchini KG, Koike G, Krieger EM, Lander ES, Dzau VJ, Jacob HJ, A biometrical genome search in rats reveals the multigenic basis of blood pressure variation. Genome Res. 1995; S: 164–172.
Mori Y, Kim-Motoyama H, Ito Y, Katakura T, Yasuda K, Ishiyama-Shigemoto S, Yamada K, Akanuma Y, Ohashi Y, Kimura S, Yazaki Y, Kadowaki T. The Gln27Glu beta2-adrenergic receptor variant is associated with obesity due to subcutaneous fat accumulation in Japanese men. Biochem Biophys Res Commun. 1999; 258: 138–140.
Bray MS, Krushkal J, Li L, Ferrell R, Kardia S, Sing CF, Turner ST, Boerwinkle E. Positional genomic analysis identifies the β2-adrenergic receptor gene as a susceptibility locus for human hypertension. Circulation. 2000; 101: 2877–2882.
Kotanko P, Binder A, Tasker J, DeFreitas P, Kamdar S, Clark AJL, Skrabal F, Caulfield M. Essential hypertension in African Caribbeans associates with a variant of the β2-adrenoceptor. Hypertension. 1997; 30: 773–776.
Meirhaeghe A, Luan J, Selberg-Franks P, Hennings S, Mitchell J, Halsall D, O’Rahilly S, Wareham NJ. The effect of the Gly16Arg polymorphism of the beta(2)-adrenergic receptor gene on plasma free fatty acid levels is modulated by physical activity. J Clin Endocrinol Metab. 2001; 86: 5881–5887.