Angiotensinogen Gene Promoter Region Variant Modifies Body Size–Ambulatory Blood Pressure Relations in Hypertension
Background— The extent to which genes modify the relationship between risk factors for hypertension and blood pressure (BP) is unclear. As angiotensinogen is expressed in adipose tissue and angiotensinogen (AGT) gene promoter variants influence the production of angiotensinogen, we evaluated the role of AGT gene variants as potential modifiers of body size–BP relations.
Methods and Results— Five hundred twenty-one hypertensives of African origin sampled from a group with a high mean body mass index (BMI) had 24-hour ambulatory BP (ABP) measurements determined off therapy and were genotyped for the AGT –6G→A, –532C→T, –20A→C, and 704T→C (M235T) gene variants. Genotypes were also determined in 547 control subjects of African origin who had a normal clinic BP. The –6A and –532C alleles were concordant with the M235T variant. Although AGT gene variants had no independent effects on either the presence of hypertension or ABP values in hypertensives, the –20A→C polymorphism had a marked influence on the relation between ambulatory systolic BP and BMI. This relation was present in patients homozygous for the –20A allele (n=399, r=0.23, P<0.0001), but absent in those with at least one copy of the –20C allele (n=122, r=0.01, P=0.89). The M235T polymorphism did not impact on the BMI-BP relation. Specificity of the –20A→C polymorphism effect on the BMI-BP relation is further indicated by the lack of effect on the systolic BP–age relation.
Conclusion— An AGT gene promoter region variant is an important modifier of the relation between body size and BP. Hence, these data corroborate the notion that genetic modifiers can produce a profound impact on BP-phenotypic relations.
Received May 6, 2002; revision received June 24, 2002; accepted June 25, 2002.
Body size is associated with blood pressure (BP).1 However, not all obese persons have an increased BP and even in severe obesity 40% of subjects may have a normal BP.2,3⇓ The reasons for these diverse effects of body size on BP have not been identified. Factors that influence the relationship between body size and BP in hypertension require elucidation as they may predict the impact of weight reduction on BP. The BP response to obesity may be influenced by genetic factors, which in turn may depend on modifying alleles in the genetic background.4 Hence, modifier genes have been suggested to have a substantial effect on the impact of body size on BP.4 However, as yet there are no data to show a profound influence of genetic modifiers on body size–BP relations in human studies. Production of angiotensinogen, which is in part derived from adipose tissue,5,6⇓ is influenced by polymorphisms in the angiotensinogen (AGT) gene.7–9⇓⇓ These polymorphisms, notably an exon 2 variant (704T→C [M235T]), which is in linkage disequilibrium with potential functional promoter region polymorphisms, and a promoter region variant (–20A→C) have been variably linked with the risk of hypertension.7,8,10,11⇓⇓⇓ Therefore, the aim of the present study was to assess whether these AGT gene polymorphisms substantially influence the impact of body size on BP in human hypertension.
This study was approved by the Committee for Research on Human Subjects of the University of the Witwatersrand (approval No. M951122).
Study Groups, BP Measurements, and Hypertension Grading
Five hundred fifty-six consecutive hypertensive patients of African ancestry initially screened at district clinics in suburban areas of Johannesburg and referred to tertiary centers for more thorough clinical assessments were recruited if they had mean daytime ambulatory diastolic BPs (DBPs) >90 mm Hg (Spacelabs model 90207) off medication. Patients with auscultatory BPs <200/115 mm Hg had ambulatory measurements performed after at least 2 weeks off medication. A minority of patients (6%) with either first-visit auscultatory BPs ≥200/115 mm Hg and with either target organ damage or two or more additional risk factors for cardiovascular disease had 24-hour BP monitoring performed within a shorter period off medication. To avoid population stratification, only patients of the Nguni, Sotho, and Venda chiefdoms of South Africa were selected. In addition, patients with type I diabetes mellitus, uncontrolled type II diabetes mellitus (defined as hemoglobin A1C of >10%), renal and endocrine disease, and clinically important cardiac pathology (clinically significant arrhythmias, heart failure, valvular disease, ischemic heart disease, previous myocardial infarction, and unstable angina) were excluded. Ambulatory BP (ABP) measurements were performed at least every half hour during the day (6 am to 10 pm) and hourly during the night (10 pm to 6 am), and ambulatory monitors were calibrated using standard techniques. All patients were advised not to smoke, imbibe alcohol, or ingest caffeine during this period. Grading of the severity of hypertension was determined from the mean of daytime ABP measurements on the basis of standard criteria (grade I, 140 to 159/90 to 99; grade II, 160 to 179/100 to 109; grade III, ≥180/100 mm Hg).12 Five hundred twenty-one patients had >90% and the rest >85% of intended ABP recordings obtained.
To evaluate whether the genetic variants examined were associated with the development of BP, a case-control study was also performed. Five hundred forty-seven control subjects of African origin without a family history of hypertension were recruited from suburban areas of Johannesburg and were considered normotensive if auscultatory and oscillometric (Spacelabs model 90207) DBPs were <90 mm Hg after 5 minutes of rest in the seated position and if the subjects had been residents of an urban area for at least 2 years.
DNA was extracted from whole blood by lysing red blood cells and digesting the remaining white cell pellet with proteinase K. Genotyping of an A-to-C transition at nucleotide –20 of the 5′ upstream promoter region of the AGT gene11 and the T-to-C transition at nucleotide 704 in exon 2 of the AGT gene13 was undertaken using polymerase chain reaction–restriction fragment length polymorphism–based techniques with the appropriate primer pairs and restriction enzymes. At least 50% of samples had repeat genotyping performed on them to ensure reproducibility. We also considered assessing the impact of the –6G→A8 and the –532C→T9 promoter region polymorphisms of the AGT gene on body size–BP relations but found in initial genotyping (using direct sequencing techniques) of 300 subjects (150 cases and 150 controls) that the –6G allele only occurred with a frequency of 3% to 4%, and the –532T allele with a frequency of only 11%. Nevertheless, the 235T allele occurred with the –6A allele in 94.3% of subjects and with the –532C allele in 100% of subjects, but with the –20C allele of the –20A→C polymorphism in only 12.3% of subjects (although the M235 allele occurred with the –20A allele in 97% of subjects). Consequently the M235T variant was used as a marker of effects of either the –6G→A or the –532C→T polymorphisms but not of the –20A→C polymorphism.
To test for Hardy-Weinberg equilibrium, the expected genotype numbers were calculated from the allele frequencies and deviation from the observed genotype numbers was determined using a χ2 test. Independent effects of either alleles or genotypes on the presence of hypertension or the grade of hypertension were evaluated using either a Fisher exact test or a χ2 test and probability values were adjusted for multiple genotyping using the Bonferroni method. Multiple logistic regression analysis (with age, gender, and body mass index [BMI] used as covariates) was also performed to assess the independent effects of genotype on the risk of hypertension. The phenotypic factors that influence BP were identified from regression analysis. In contrast to the analysis performed assessing the relationship between allele/genotype and the presence of hypertension, in which all 556 patients were included, regression analysis was only performed in those patients for whom >90% of intended ABP recordings were obtained. The impact of genotype on ABP values was evaluated by using ANCOVA techniques, with age, gender, and BMI used as covariates, and also by assessing the effect of genotype on the regression relations between ABP and other phenotypic parameters. Continuous data are expressed as mean±SEM.
Demographic and Clinical Data
Demographic and clinical data are shown in the Table. Both the case and the control groups had a preponderance of females and individuals with an increased BMI. The preponderance of females reflects the gender distribution of patients attending district clinics rather than a profoundly greater incidence of hypertension in African women compared with men. Except for a higher mean BMI in the case group (Table), the case and control groups were matched according to all other demographic features including the frequency of subjects from different chiefdoms. Although the majority of patients had grades I to II hypertension as determined from mean daytime ABPs, a high percentage of patients with grade III hypertension were also recruited (Table).
Genotype Effects on Risk of Hypertension and ABP
The genotype frequencies of both the –20A→C and M235T polymorphisms were in Hardy-Weinberg equilibrium. Neither variant was independently associated with the presence of hypertension (Figure 1). Further, neither polymorphism showed a quantitative association with 24-hour (Figure 2), day, or night ABP. Similarly, no association between genotype and the grade of hypertension was evident.
Genotype Effects on BP-Phenotypic Relations
Both age and BMI showed significant correlations with systolic BP (SBP) (Figure 3), but not diastolic ABP, in the hypertensive group. The –20A→C gene polymorphism markedly modified the relation between BMI and SBP. In patients with the AA genotype, the relationship was more pronounced, whereas in the AC+CC genotype group a relationship was absent (Figure 3). In contrast, the –20A→C gene polymorphism did not influence the relation between SBP and age (Figure 3). Moreover, the M235T polymorphism failed to influence either the BMI-SBP or the age-SBP relationship (Figure 3).
The main finding of the present study is that a potentially functional14 promoter region variant of the AGT gene has a marked influence on the relation between body size and systolic BP in hypertensives. Because the –20A→C polymorphism did not produce an independent effect on either BP in hypertensives or the risk of development of hypertension when accounting for body size differences between subjects, the effect of the variant reflects a moderating influence of the genotype on the BMI-BP relationship rather than an effect of body size on the relationship between genotype and BP. This distinction is important as it implies that the presence of the risk genotype is insufficient to account for a BP effect alone. Rather, the risk genotype in part determines the overall effect of body size on BP. The presence of the –20C allele abolished the impact of body size on BP, whereas in those patients homozygous for the –20A allele a significant relationship between body size and SBP was evident. Although an angiotensin-converting enzyme gene insertion/deletion polymorphism has previously been shown to influence the slope of the relationship between body size and BP in humans,15 the novelty of the present data is that they show a gene modifier effect that abolishes this relation. This profound context-dependent effect of a genetic polymorphism is in keeping with the complex nature of polygenic traits and could assist in predicting the effect of body size on BP in humans.
A large number of studies attempting to associate AGT gene variants with risk of hypertension have produced highly variable results.10 Sample sizes inadequate to limit the risk of false-positive or false-negative results, poor phenotypic characterization, and population admixture may limit the outcome of case-control studies.16 With respect to these limitations, our study has a sufficiently large sample size (556 cases and 547 controls) to avoid false-negative results (for the –20A→C and the M235T gene variants), and the presence of hypertension has been confirmed in all patients using 24-hour ABP monitoring with patients off medical therapy. Moreover, patient and control groups were not only matched for their usual demographic characteristics, including racial background, but also for more precise ethnic backgrounds (chiefdoms that are historically derived from the same gene pool). Although our present data do not support an independent role for the AGT gene in contributing to either the development or the severity of hypertension in this population, the possibility of epistatic interactions with other genes, moderating the impact of the AGT gene, cannot be excluded.
In the present study, we evaluated genetic effects on BP considered as a continuous trait in a hypertensive group rather than in a cross section of the population. Our reasons for the choice of study sample were 3-fold. First, to improve on the statistical power of detecting genotype-phenotype interactions using ABP measurements, a group sampled with a relatively wide distribution of BPs as opposed to a group sampled with BPs clustered around the median (general population) was necessary. Second, effects of body size on BP are more relevant to hypertensives, as it is to this group that advice regarding weight reduction would be given to assist with BP control. Third, in patients of African origin there is a higher prevalence of more severe hypertension.12 Therefore, equally as important as assessing the contribution of genes to BP in the general population is an approach that assesses the role of genes in contributing to the severity of hypertension in groups with a higher prevalence of grades II and III hypertension. A major strength of our study is the assessment of BP by 24-hour ambulatory monitoring rather than by clinical readings. ABP monitoring has been shown to better correlate with end-organ damage and complications of hypertension.17
Patients with clinically important cardiac pathologies were excluded in the present study because, for their safety, they could not be taken off medical therapy and therefore ABP measurements could not be performed. Although a selection bias may have been introduced as a result of the exclusion of such patients, this bias would have been against ourselves and therefore, if anything, underscores our data. As renin–angiotensin system (RAS) polymorphisms have been reported as possible risk factors for cardiovascular mortality, one would be more likely to observe a significant genetic finding in this population than in one that is limited to patients with only essential hypertension. Thus, our data cannot be attributed to or confounded by the associations previously demonstrated between RAS gene polymorphisms and the risk of cardiovascular mortality.
As the AGT gene –20A→C promoter region polymorphism has no direct effect on either SBP or DBP, the mechanisms by which this polymorphism influences the relation between body size and SBP would therefore seem more complex. In in vitro studies the –20A→C polymorphism has been found to influence basal transcription of angiotensinogen as well as stimulated transcription in response to receptor binding.14 A key site for this transcription effect might be adipose tissue where angiotensinogen is known to be expressed at significant levels5,6⇓ with a subsequent effect on the capacity of adipose tissue to produce angiotensinogen. Hence, we propose that the effect of the –20A→C polymorphism on the BMI-SBP relation may be related to extrahepatic production of angiotensinogen possibly in adipose tissue.
In the present study the M235T variant of the AGT gene was in linkage disequilibrium with both the –6G→A and the –532C→T variants. As a marked predominance of the –6A and –532C alleles was noted in the population studied (which prevented the assessment of the impact of the –6G→A and –532C→T variants on body size–BP relations), the M235T variant was used as a marker of the effects of the –6G→A and –532C→T variants, both of which have previously been shown to influence angiotensinogen concentrations.8,9⇓ As the M235T variant did not influence body size–BP relations, these data would suggest that the –6G→A and –532C→T polymorphisms are not significant modifiers of body size–BP relations. However, further studies in populations with an informative distribution of –6G→A and –532C→T alleles will be required before this conclusion can be drawn with certainty.
At this stage the clinical implications of our findings are speculative. However, if the results are confirmed in other studies and populations, then it is conceivable that the effect of the AGT gene could, for example, influence therapeutic choices. If the AGT –20A→C variant influences the effect of BMI on BP, then pharmacological agents targeting the RAS may be particularly efficacious in obese hypertensive subjects with the deleterious genotype. Moreover, if body size effects on BP are genotype dependent, then one would expect that weight reduction would be beneficial with respect to BP effects in patients with the risk genotype, but not the nonrisk genotype. These hypotheses have yet to be tested.
In summary, the results of the present study indicate that in the absence of an independent effect on BP, an AGT gene promoter region polymorphism markedly influences the body size–BP relationship, at least in hypertensives. These context-dependent effects of the AGT gene on BP could be useful in predicting the effect of changes in body size on BP. These data show that genetic factors impact profoundly on the relation between body size and BP, an effect previously purported to significantly influence BP in human hypertension.4
This work was supported by the University Research Council of the University of the Witwatersrand, The Medical Faculty Research Endowment Fund of the University of the Witwatersrand, the C.V. Hodges Charitable Trust, the H.E. Griffin Charitable Trust, the Southern African Hypertension Society supported by an open grant from Astra-Zeneca, and the Medical Research Council of South Africa. A.D.T. was supported by a grant from the Netherlands Organization for International Cooperation in Higher Education.
- ↵Mark AL, Correia M, Morgan DA, et al. Obesity-induced hypertension: new concepts from the emerging biology of obesity. Hypertension. 1999; 33: 537–541.
- ↵Pratt JH, Ambrosius WT, Tewksbury DA, et al. Serum angiotensinogen concentration in relation to gonadal hormones, body size and genotype in growing young people. Hypertension. 1998; 32: 875–879.
- ↵Paillard F, Chansel D, Brand E, et al. Genotype-phenotype relationships for the renin-angiotensin-aldosterone system in a normal population. Hypertension. 1999; 34: 423–429.
- ↵Ishigami T, Umewara S, Tamura K, et al. Essential hypertension and 5′ upstream core promoter region of human angiotensinogen gene. Hypertension. 1997; 30: 1325–1330.
- ↵Russ AP, Maerz W, Ruzicka V, et al. Rapid detection of the hypertension-association M235-Thr allele of the human angiotensinogen gene. Hum Mol Genet. 1993; 2: 609–610.
- ↵Zhao YY, Zhou J, Narayanan CS, et al. Role of C/A polymorphism at -20 on the expression of human angiotensinogen gene. Hypertension. 1999; 33: 108–115.
- ↵Turner ST, Boerwinkle E, Sing CF. Context-dependent associations of the ACE I/D polymorphism with blood pressure. Hypertension. 1999; 34: 773–778.