A Study on Association of Renin-Angiotensin System Gene Polymorphisms with Atrial Fibrillation

• Main Authors: Chia-Ti Tsai, 蔡佳醍
• Other Authors: Yung-Zu Tseng
• Format: Others
• Language: zh-TW
• Published: 2002
• Online Access: http://ndltd.ncl.edu.tw/handle/53181996333377479408

碩士 === 國立臺灣大學 === 臨床醫學研究所 === 90 === Introduction The renin-angiotensin (RAS) has been shown to be involved in many cardiovascular disease, which causes myocardial fibrosis and/or hypertrophy in hypertensive heart disease, congestive heart failure, myocardial infarction, and cardiomyopathy. Its effects on ventricular myocardium could possibly be applied to atrial myocardium and causes atrial hypertrophy and fibrosis. Recent reports suggest that atrial fibrillation (Afib) is associated with the activation of the RAS in the atria in humans and in a dog model of Afib. Angiotensin II triggered the mitogen activated protein kinase pathway, which was responsible for the proliferation of fibroblasts and hypertrophy of cardiomyocytes. It has also been shown that inhibition of endogenous angiotensin II prevented atrial effective refractory period shortening during rapid atrial pacing in dogs. This indicates that angiotensin II may be involved in the mechanism of atrial electrical remodeling and that the blockade of angiotensin II may lead to the better therapeutic management of human atrial fibrillation. Pedersen et al. provided the first evidence that angiotensin-converting enzyme (ACE) inhibitor therapy could reduce the occurrence of Afib in patients with left ventricular dysfunction after myocardial infarction. Van den Berg et al. observed that pretreatment with ACE inhibitors reduced the relapse rate of Afib after electrical cardioversion. These results indicate that inhibition of the cardiac RAS by ACE inhibitors or angiotensin receptor antagonists might affect the pathophysiological substrate of Afib and may offer a new therapeutic approach. It has been shown that aging, valvular heart disease, and left ventricular dysfunction are important risk factors for the development of atrial fibrillation. However, atrial fibrillation is a multi-factorial disease. Some patients remain in sinus rhythm despite the presence of significant valvular pathology and/or left ventricular dysfunction while some other patients develop atrial fibrillation in the absence of any known risk factor (lone atrial fibrillation). It is possible that predisposing genetic factors may play important roles in the development of atrial fibrillation. However, so far there have been very few reports addressing the genetic control of Afib. Based upon the aforementioned studies about RAS and Afib, we hypothesized that rennin-angiotensin system genes might be the susceptible genes of Afib, and performed a genetic case-control study to prove it. We totally genotyped 8 single nucleotide polymorphisms among the RAS genes, including ACE gene insertion/deletion (I/D) polymorphism, G-217A, G-152A, A-20C, G-6A, M235T and T174M polymorphisms of the angiotensinogen (AGT) gene, and the A1166C polymorphism of the angiotensin II type I receptor gene (AT1R). Methods Study population Case patients were 110 consecutive patients who were admitted to our adult cardiology ward and had a history of atrial fibrillation. Patients with hyperthyroidism were excluded. For every case patient, a matched control without a history of atrial fibrillation was selected from the same ward. The case and control patients were matched regarding their gender, age, the presence of left ventricular dysfunction (ejection fraction < 55%) and the presence of significant valvular heart disease (at least moderate-severe) (Table 1). More than 80% of the Afib patietns had persistent Afib. Ten percent did not any identifiable cause of Afib, and were defined as having lone Afib. All patients agreed to participate and informed consents were obtained. Clinical assessment The presence of atrial fibrillation was determined by history taking, serial ECG and/or ambulatory ECG monitoring. Patients with palpitations without ECG documentation were excluded from both patient and control groups. Transthoracic echocardiography was performed to measure the left atrial and left ventricular dimensions, left ventricular ejection fraction and to detect significant valvular disease (defined as moderate-severe or severe valvular regurgitation or stenosis). The left ventricular mass was calculated using echocardiographic parameters and the Devereux formula. Identification of diallelic polymorphisms Genomic DNA was extracted by a nonenzymatic method. DNA fragments were amplified by polymerase chain reaction (PCR). Genotyping of ACE gene I/D polymorphism was performed as our previously reported methods. Genotyping of AT1R gene A1166C polymorphism was performed with PCR-restriction fragment length polymorphism method. For genotyping of the AGT gene polymorphisms, we used mini-PCR direct sequencing as in our previously report. In addition to the G-6A and A-20C diallelic polymorphisms in angiotensinogen core-promoter element 1, the functional studies of which have been reported, we also identified 2 polymorphisms at positions —152 and —217 relative to the transcriptional start site (G-152A and G-217A, respectively). Based on our preliminary results of transcriptional activity study, we found more upstream promoter regions other than core-promoter element 1 also played a critical role in transcriptional control. The functional significance of these two polymorphisms important, and are currently in development in our laboratory. Statistical methods The between-group data were compared with Student’s unpaired t test for continuous data and with theχ2 test for categorical data. Allele frequencies were calculated from the genotypes of the subjects. Differences in allele frequencies between cases and controls were compared with theχ2 test or Fisher’s exact test. Deviation from Hardy-Weinberg equilibrium was assessed by the χ2 test. There were three genotypes for a diallelic polymorphism (AA, AB, and BB [A is the wild-type allele and B the mutant allele]).The genotype-phenotype correlation was examined with the following 3 models: additive model (AA vs. AB vs. BB; AA=0, AB=1 and BB=2), dominant model (AA vs. AB+BB; AA=0 and AB+BB=1), and recessive model (AA+AB vs. BB; AA+AB=0 and BB=1). The association between genotypes and Afib were evaluated using logistic regression model. Odds ratios (ORs) were calculated, together with their 95% confidence intervals (CIs). Because left atrial size was not balanced in cases and controls, we also perform the analyses with adjustment for left atrial size. Because the 6 polymorphisms within the AGT gene located on the same chromosome with short distances between each other, they were not segregated independently. Therefore, we used haplotype analysis to see if there were any specific haplotypes which were associated with Afib. We used maximum-likelihood estimation with expectation-maximization algorithm. The methods of the expectation-maximization-based haplotype frequency estimation and permutation-based hypothesis testing procedure were performed based on the work of Fallin, et al. For evaluation of gene-gene interaction, we stratified the patients according to the genotype of one gene, and analyzed the effect of the other gene in different strata defined by the genotype of the former gene. This included tests of homogeneity of odds ratio across strata by Mantel-Haenszel test in single locus analysis, and haplotype analysis of the AGT gene as mentioned above in different stata of different ACE gene I/D or AT1R gene A1166C genotypes. P value less than 0.006 (0.05 / 8) was considered statistically significant after Bonferoni’s correction in the single-locus analyses and tests of Hardy-Weinberg equilibrium, and 0.05 for other analyses. Results Single-locus Analyses The results of the single-locus analyses are shown in Table 2. The G-217A, G-6A and M235T polymorphisms of the AGT gene were associated with Afib with at least one significant P value among the models of analysis. M235T was associated with Afib in most models of the analysis, including allele frequency analyses, additive model analysis, and recessive model analyses, either corrected for LA size or not. Hardy-Weinberg equilibrium tests for the 8 diallelic polymorphisms Tests of Hardy-Weinberg equilibrium were performed for all loci among cases and controls separately. No locus was significantly deviated from the Hardy-Weinberg expectation in cases and controls, except G-6A locus in cases. The genotype distribution of G-6A locus in patients with Afib was significantly deviated from Hardy-Weinberg equilibrium at the significance level P<0.05 (P=0.036), but not after Bonferroni’s correction. Haplotypes of the AGT Gene and their association with Afib Table 3 displays the results of six-locus estimated haplotype frequency analyses for the AGT gene in cases and controls. The omnibus haplotype profile test was significant (χ2= 35.5, p= 0.034), which indicated the overall haplotype frequency profile difference between cases and controls were significant, and thus there might be some disease-predisposing haplotypes existing in patients with Afib. Accordingly, in the individual haplotype analyses, we identified 2 haplotypes (GGAACT and GGAGCT) with significantly higher haplotype frequency in cases than in controls with a significant P value by permutation tests. Judging from the polymorphisms comprising the haplotypes, we found that G-6 and M235 were associated with Afib. Gene-gene interaction Single locus analysis In the single-locus analysis, G-217, G-6, and M235 alleles of the AGT gene were associated with hypertension. No significant gene-gene interaction was noted when the analysis of association of M235T polymorphism with Afib was stratified according to the G-217A genotype (χ2 =2.28; P=0.131 for test of homogeneity of OR; combined M-H OR for recessive model 0.3 [0.2-0.8], P=0.008) or G-217A stratified by M235T polymorphism (χ2 =2.28; P=0.131 for test of homogeneity of OR; combined M-H OR 0.6 (0.3-1.1), P=0.087). AGT gene haplotype analysis The results of gene-gene interaction between the AGT gene and ACE gene using haplotype methods with the stratification method were shown in Table 4. The haplotype analysis of the AGT gene was performed in each of the 2 strata (patients with II genotype as one stratum, and with ID and DD as the other stratum), respectively. The omnibus haplotype profile test was not significant in both strata, which might be due to decreased case number after stratification. However, GGAGCT haplotype was still associated with Afib in patients with ID or DD genotype (P=0.025), and borderline significant in patients with II genotype (P=0.069). GGAACT haplotype was associated with Afib in patients with II genotype (P=0.005), and borderline significant in patients with ID or DD genotype (P=0.090). These results indicated that the association of AGT gene haplotype with Afib was not changed in patients with different ACE I/D genotype, and implied no significant gene-gene interaction between the AGT and ACE gene in the association with Afib. Stratification was not performed according to the AT1R gene polymorphism because most subjects (89%) had AA genotype and the case number of patients with AC and CC genotypes was too small to perform stratification analysis. Discussion We first reported the association between RAS gene polymorphims and Afib. Our findings provide the possibility that the RAS genes are not only the candidate genes for hypertension, but also for Afib. Possible explanation of the association We demonstrated that the M235 allele in exon 2 of the AGT gene, and G-6 and G-217 alleles in the promoter region of the AGT gene were associated with Afib. In our previous functional studies of the promoter polymorphism of AGT gene, we found that G-6 and G-217 were associated with a higher AGT gene transcription activity than A-6 and A-217, respectively. This higher transcription may cause a higher tissue angiotesin II concentration in the atrium under the stimulation of high atrial pressure, which subsequently activates the mitogen activated protein kinase pathway and cause atrial fibrosis, conduction heterogeneity, decreased atrial effective refractory period, and provide the substrates for the development of Afib. There is no functional significance of the M235T polymorphism. The association of M235 allele with Afib may be through its tight linkage with G-6 allele. However, M235T is the most significant locus in our study, even more significant than G-6A locus. Therefore, we could not rule out the possibility that M235T is linked to other loci with functional significance. Study design We used a one-by-one matched case-control study design. The choice of case and control groups in a genetic association study is crucial. In our study, only patients with ECG-documented atrial fibrillation were selected as cases. For control patient selection, only patients with sinus rhythm who denied a history of palpitations were included. Patients with recurrent palpitations and without ECG documentation were excluded from both groups. This may further increase the contrast. As aging, left ventricular dysfunction and valvular heart disease are known major risk factors for atrial fibrillation, the cases and controls were selected from the same ward and were matched for age, sex, left ventricular dysfunction and valvular disease on a one-by-one base. This design therefore excluded the possibilities of age, sex, left ventricular dysfunction and valvular disease as the confounding factors. Furthermore, this study was performed in a pure Taiwanese population, which excluded the possibility of false positive results due to population stratification. It might be argued that why controls did not have Afib was due to they had a smaller mean left atrial size, but not due to the difference of the AGT gene genotypes. If this is the case, there will be no difference of genotypes or haplotypes between cases and controls. However, our results did show difference of the AGT genotypes or haplotypes between cases and controls. This is rather impossible to come from statistical artifacts, because the difference was far from small, and we had corrected the P value for multiple comparisons. There are 3 possibilities to explain why both the left atrial size and the AGT gene structure were different between cases and controls. First, the association of the AGT genetic variations with Afib was through their effect on atrial size. Some genetic variations were associated with a lower atrial tissue angiotensin II level, even under high atrial pressure, and consequently, the progression of atrial enlargement was retarded. Therefore, patients with these genetic variations had less tendency to develop Afib. Second, the left atrial size enlarged after the occurrence of Afib. Then, the imbalance of left atrial size would not affect the validity of our results. In the above 2 conditions, the association of AGT gene polymorphisms with Afib will disappear after correction for left atrial size in the multivariate analyses. Therefore, we concluded that both left atrial size and the AGT gene variations were associated with Afib. Our results showed that the association of AGT genetic variations with Afib persisted after adjustment for left atrial size. Advantages and limitations There were several advantages of our study. First, it has been described that a specific multilocus haplotype, rather than a single-locus allele is the more significant determinant of association. We demonstrated the association of the AGT gene polymorphism with Afib not only by conventional single-locus allele frequency analyses, but also by a new, valid, and well established method of haplotype analysis. Second, this is also the first study of a comprehensive multilocus and multi-gene study of RAS genes in the genetic study of non-familial Afib. Third, our results could be explained by our previous functional studies of the AGT gene promoter polymorphisms. There are also limitations in our studies. First, our results only provide evidence of the association between RAS genes variation and Afib in gene level, but do not demonstrate the direct mechanism how RAS causes Afib. Second, Afib is a heterogeneous disease. It can be predisposed by many environmental factors such as left ventricular dysfunction, valvular heart disease and hypertensive cardiovascular disease. It can also occur without identifiable causes, such lone Afib. It may be more appropriate to conduct a study on a homogeneous group of patients. Conclusions Development of Afib is possibly genetically controlled to some extent. Patients who have a specific genetic variation or polymorphism in the AGT gene may be more liable to develop Afib when exposed to environmental factors which elevate atrial pressure.