Epithelial Sodium Channel Allele T594M Is Not Associated With Blood Pressure or Blood Pressure Response to Amiloride
The T594M allele of the epithelial sodium channel β-subunit has been proposed as a gain-of-function mutation leading to salt-sensitive hypertension in blacks that is particularly responsive to the specific sodium channel antagonist amiloride. However, the positive associations derive from small convenience samples, and the amiloride challenge study lacked a control group. We determined whether the T594M allele was associated with hypertension and blood pressure (BP) response to amiloride in 2 well-characterized random population samples including 3137 Dallas County subjects and 1666 Jamaican blacks. In multivariate models, the T594M allele was not predictive of systolic BP (adjusted odds ratio for hypertension 1.1; 95% confidence interval, 0.7 to 1.8). Amiloride treatment did not lower the BP of 6 T594M heterozygotes significantly more than in 22 control subjects (P=0.8). We conclude that the T594M allele does not contribute significantly to BP in blacks and does not predict a significantly superior response to amiloride therapy.
People of African descent are particularly susceptible to developing salt-sensitive hypertension, but the molecular mechanisms are poorly understood. The renal epithelial sodium channel (ENaC) constitutes the final common pathway for the reabsorption of the final 2% of filtered sodium from the distal nephron. A missense mutation in the β-subunit of the human ENaC gene (T594M allele of SCNN1B or β-ENaC) has been found only in persons of African descent and has been implicated as a gain-of-function mutation causing impaired renal sodium excretion and salt-sensitive hypertension.1–3 The allele was found to be present in 6% of blacks overall and to be highly enriched in a hypertensive clinic population compared with normotensive controls.2 Additional evidence for an association with hypertension was reported in a larger clinic population.4 However, the enthusiasm generated from these positive association studies2,4,5 has been tempered by independent reports that failed to confirm such an association.6,7 The conflicting data surrounding this allele might derive from inadequate sample size, sampling biases inherent in clinic populations, the variability in resting blood pressure (BP) over time, confounding effects of drug treatment, and the multiplicity of environmental exposures and signaling pathways involved in the regulation of resting BP.
The BP responsiveness to certain drugs, which predominately target a single BP-regulating pathway, might link more directly with the genetic underpinnings of hypertension than resting BP alone. The potassium-sparing diuretic amiloride, a specific ENaC antagonist, is normally a weak antihypertensive agent but has been proposed as the ideal treatment for hypertension in blacks carrying the T594M allele.3 In a study of 14 T594M heterozygotes with mild-to-moderate hypertension, amiloride monotherapy was as effective in controlling BP as usual empiric combination therapy with 2 or 3 standard antihypertensive agents.8 However, the conclusions from that study must be considered preliminary, because no control group of black hypertensives lacking the T594M allele was studied for comparison.
Accordingly, the aims of our study were 2-fold. First, we tested for an association between the T594M allele and hypertension in 2 large and independent African-origin population samples, one from a major metropolitan center in the Southwestern United States and the other from rural Jamaica. We collected detailed survey data on environmental exposures known to affect BP and studied a sufficient number of untreated subjects to study BP as a quantitative trait. In a large subset of subjects, we also tested for an association with hypertensive heart disease as assessed by cardiac MRI. Second, we tested for an association of the T594M allele with amiloride responsiveness of BP, including an adequate number of control subjects (blacks who do not carry the T594M allele), using 24-hour ambulatory BP monitoring to reduce within-subject variability.
Subjects derived from the Dallas Heart Study (DHS) in Texas9 and the Spanish Town study in Jamaica. The ethnic category of each subject was self-assigned from the same structured list of categories used in the Third National Health and Nutrition Examination Survey.10 Phlebotomy, clinic examination, and thoracic MRI to measure left ventricular (LV) mass and function11 and aortic compliance12 were obtained as described. The study was approved by the Institutional Review Boards at both the University of Texas Southwestern Medical Center at Dallas and Research Triangle Institute (who managed field staff). Subjects gave informed written consent for data collection and DNA analysis and separately for the amiloride treatment study (below).
Jamaican participants were recruited into the International Collaborative Study of Hypertension in Blacks project.13,14 Participants were recruited from Spanish Town, a working class suburb of Kingston, as described previously. Data collected included anthropometrics [weight, height, body mass index (BMI), waist and hip circumference, and waist/hip ratio], BP, and prevalence of hypertension. The study protocol was reviewed and approved by the Institutional Review Board at the University of the West Indies, Mona, Jamaica, and subjects gave informed written consent.
Measurements of BP
At each of the 3 visits for the DHS subjects, 5 BP and heart rate measurements were taken in the seated position using an automatic oscillometric device (Welch Allyn, Series 52000) that has been validated by the British Society of Hypertension protocol.15 All of the study personnel were trained to use the device, to select an appropriately sized arm cuff, and to start the measurements after 10 minutes of rest with the subject having voided and being seated quietly in a straight-back chair with the arm at heart level. BP is reported as the mean of the last 3 readings at each of the 2 home visits. The median time between visits 1 and 2 was 17 days, and analyses in this article are limited to the black, white, and Hispanic participants ages 30 to 65 who provided DNA samples and participated in both home visits. For the Jamaican subjects, 3 BPs were measured in the sitting position between 1994 and 1996, and the mean of the last 2 readings was used in the analyses. For association studies, hypertension was defined as a systolic BP (SBP) ≥140 mm Hg or diastolic BP ≥90 mm Hg or current treatment with antihypertensive medications.16
Genomic DNA was isolated from peripheral blood using the Puregene kit (Gentra System). Sequencing was performed using an ABI 3730 automated DNA sequencing instrument and Big Dye Version 3.1 dye terminator chemistry (Applied Biosystems). SCNN1B genotyping was performed using a 5′ nuclease assay (TaqMan) designed and manufactured using Applied Biosystems Assay-by-Design custom service (T and M alleles VIC and FAM labeled, respectively). Reactions (7 μL) were assembled in 384-well plates containing TaqMan Master Mix (3.5 μL) with 0.4 mmol/L deoxynucleotide triphosphates, Taq polymerase (0.2 U), and genomic DNA (20 ng). Thermocycling conditions were 50o for 2 minutes and 95o for 10 minutes followed by 40 cycles of 95o for 15 s and 60o for 1 minute. The T and M alleles were scored by the ratio of FAM/VIC fluorescence on an ABI Prism 7900HT Sequence Detection System. PCR amplicons were directly sequenced to confirm genotype for all of the T594M alleles (Figure 1A).
Amiloride Treatment Study
We recruited 6 DHS black subjects heterozygous for the T594M allele, as well as an excess of 22 black control subjects without the T594M allele, to improve the power of the study despite limited numbers of hypertensive T594M heterozygotes. Subjects had a SBP of 120 to 159 mm Hg or a diastolic BP of 80 to 99 mm Hg without antihypertensive therapy. Exclusion criteria included body mass index >45 kg/m2; history of cardiovascular or renal disease, insulin-treated diabetes, or renal artery stenosis; pregnant women or women of childbearing age who do not use contraception; serum potassium >5 mmol/L; and children under the age of 18. After a 2 week wash-in, a 24-hour ambulatory BP monitoring (ABPM) was performed using SpaceLab 90217 oscillatory BP monitors. Amiloride was administered orally twice daily, 5 mg BID for 1 week, and 10 mg BID for another 2 weeks. Another 24-hour ABPM was recorded after 3 weeks, and compliance was monitored by pill count.
Venipuncture was performed at baseline, 1 week, and 3 weeks of amiloride treatment with the subject seated for ≥5 minutes. Blood was drawn into serum separator and disodium EDTA tubes, and plasma samples were centrifuged promptly at room temperature, immediately separated, and frozen at −80 C until assayed. Serum sodium and potassium were measured by flame photometry. Aldosterone and renin were assayed by immunochemiluminescence assay using a Nichols Institute Diagnostics Advantage instrument and reagents.
SAS/STAT software, version 8.2, was used for all of the analyses. For presentations of population estimates, the SURVEYMEANS procedure was used to account for the stratified sampling design and finite population. These estimates incorporate the sample weights, which are estimates of each individual’s representation in the residential population of Dallas County. Explanatory models were based on unweighted data; each observation has equal representation in the models. Models were restricted to non-Hispanic blacks.
In tables comparing wild-type and T594M allele heterozygotes for selected characteristics, P values are based on t tests of null hypotheses of equal means. Reported P values for association between genotype and unadjusted hypertension prevalence derive from Fisher’s exact test. The linear regression models (excluding treated hypertensives) for systolic and diastolic BP test the significance of a main effect from the allele when controlling for age, BMI, gender, smoking and heart rate. Logistic regression models for hypertension test the significance of a main effect from the allele when controlling for age, BMI, diabetes, and smoking.
T tests were used to assess the significance of differences between wild-type and allelic variants in baseline levels of SBP, aldosterone, renin, and in changes of these measures in response to amiloride treatment. Results of the t tests were confirmed with rank-sum tests to guard against violations of normality.
Association of Sodium Channel T594M Allele With Hypertension and LV Mass
The T594M allele was found almost exclusively in blacks, with an allele frequency of 5.8±0.1% versus 0.3±0.3% in Hispanics (2 subjects) and 0% in whites, which is consistent with other studies (Figure 1B).1,2,4,6 Thus, 104 of 1658 blacks were T594M heterozygotes, and 3 blacks were homozygotes for the T594M allele. No significant difference in the prevalence of hypertension was found between those with or without the variant allele (42% versus 49%, respectively). In a multivariate logistic model (Figure 1C), age, diabetes, cigarette smoking, and BMI were significant predictors of hypertension. In contrast, the T594M allele was not predictive of hypertension (adjusted odds ratio, 1.1; 95% CI, 0.7 to 1.8; P=0.65).
Among the 1214 black subjects not receiving prescription BP medication, SBP increased as a function of age, regardless of genotype (Figure 1D). In a linear model that controlled for age, gender, diabetes, smoking, BMI, and heart rate, the regression lines relating SBP to age were identical for the 2 genotypes [main effect, −1.3 mm Hg (−5.1 to +2.4); P=0.49]. Multiple demographic, biochemical, and hemodynamic characteristics of subjects with and without the variant T594M allele were not significantly different (Table 1). Survey measures of poverty, education, and access to health care did not differ by genotype (data not shown).
Hypokalemia is a frequent characteristic of monogenic hypertension because of gain-of-function mutations in ENaC.17 We found no difference in mean serum potassium values between subjects with and without the T594M allelic variant (Table 1). No cases of unprovoked hypokalemia (K+ <3.5 mmol/L) were observed in either the 104 heterozygotes or the 3 homozygotes with the T594M allele. Table 1 also shows that MRI measures of LV mass, LV systolic function, and aortic compliance were indistinguishable between genotypes. Thus, multiple analyses revealed no evidence implicating the T594M allele as a predisposition for hypertension or hypertensive heart disease in blacks.
Association of T594M Allele With Hypertension in Jamaican Sample
To confirm that the lack of association observed in blacks from Dallas County was not unique to this study population, we studied a similar well-characterized population of 1666 subjects from Spanish Town, Jamaica.18 The prevalence of the T594M allele was 6.2% (103 heterozygotes), similar to the frequency observed in the DHS cohort. Mean BP was indistinguishable between the subjects with and without the T594M allele in this population (Table 2). In multivariate models, age and BMI were significant predictors of SBP and diastolic BP (DBP), but the T594M allele was not (P=0.4 and 0.6, respectively).
Response to Amiloride Treatment in T594M Heterozygotes and Controls
Subject characteristics for 6 T594M heterozygous and 22 control subjects recruited for this study are shown in Table 3. No parameters, including baseline aldosterone and renin (or the ratio), were significantly different between the 2 groups. The individual BP response to amiloride, assessed by 24-hour ABPM, was highly reproducible in 7 subjects rechallenged with amiloride a second time (Figure 2). Amiloride lowered SBP and DBP in most of the 22 control subjects and in all of the T594M heterozygotes (Figure 3), and the mean SBP reduction was not different between the 2 groups (12±12.5 and 13.5±7.2 mm Hg, respectively; P=0.8). The magnitude of the SBP response correlated with basal SBP for both T594M heterozygotes and control subjects (Figure 4), a phenomenon observed in other trials of antihypertensive drugs.19 The group difference in relative response to treatment, which might be a more appropriate comparison, was even less significant than the absolute difference (P=0.9).
Aldosterone [253±253 versus 322±481 nmol/L (4.9±4.9 versus 7.4±7.1 ng/dL); P=0.3] and renin (9.1±7.4 versus 11.6±17.3 μU/mL; P=0.8) and the aldosterone/renin ratio at baseline (Table 3) were not different between the control subjects and T594M heterozygotes, respectively. As expected for a diuretic treatment study, aldosterone and renin rose significantly in both T594M heterozygotes and control subjects (P<0.0001 baseline versus 3 weeks). In contrast, neither the rise in aldosterone [+283±281 versus +356±403 nmol/L (+10.2±10.1 versus +12.8±14.5 ng/dL); P=1.0] nor the rise in renin (+17.2±17.0 versus +21.3±20.9 μU/mL; P=0.6) were different between the T594M heterozygotes and control subjects, and hyperkalemia was not observed in any subjects.
The major new findings of this study are 2-fold. First, in 2 independent African-origin population samples, we found no evidence for the previously hypothesized association between the ENaC T594M allele with either hypertension prevalence or untreated BP in blacks. Second, using ambulatory BP monitoring and an adequate number of control subjects lacking the T594M allele, we found that the BP fall induced by the ENaC blocker amiloride is no larger in blacks who carry the T594M allele than in those who do not. In contrast to what has been hypothesized, our data suggest that the T594M allele is neither an important cause of hypertension in blacks nor a genotype that is uniquely sensitive to amiloride therapy.
Previous studies may have been underpowered to accurately assess the association between the T594M allele and hypertension because of methodological shortcomings, including small clinic samples, medication effects, insufficient information about environmental exposures, population stratification, and imprecise phenotyping of BP or of hypertensive target organ damage.20 We used 2 probability-based population samples, which yielded more cases of the T594M allele than all of the previously published studies combined. In the DHS, BP was carefully phenotyped based on multiple in-home measurements on multiple days by ethnically congruent surveyors who also obtained detailed survey data on environmental exposures. In multivariate models that confirmed the prior expectation that age, diabetes, cigarette smoking, and BMI are significantly associated with hypertension and with untreated BP, the T594M allele simply was not informative. In a large subset of participants, hypertensive heart disease was quantitatively assessed using cardiac MRI, and the T594M allele was not informative in models derived from these data either.
The DHS participants might differ significantly in their genetics, diet, or other variables that influence BP from West Indies cohorts in whom a positive association of the T594M allele with hypertension was observed previously. For example, obesity, which is so prevalent in the United States and such a powerful determinant of BP, might obscure an association of the ENaC polymorphism with hypertension. However, we detected no such association in the leaner Jamaican population sample.
Previous pharmacogenetic studies with another diuretic, hydrochlorothiazide, have been limited by a large degree of within-subject variability in the BP response to treatment when a given subject is challenged repeatedly with the same dose of drug.21 Furthermore, specific polymorphisms have accounted for only a small portion of the between-subject variability in BP response to hydrochlorothiazide compared with the large portion explained by a high pretreatment BP,19 a nonspecific effect. By using 24 hour-ABPM rather than standard office measurements, we were able to decrease the within-subject variability22 and demonstrate excellent test-retest reproducibility in individual BP reductions induced by amiloride (Figure 2). Because the T594M allele is rare, occurring in only 6% of blacks, we studied 3 times the number of cases as controls to bracket the range of the normal BP response to amiloride. Although we demonstrated considerable between-subject heterogeneity in these BP reductions, the magnitude of the BP decline did not segregate with the presence or absence of the T594M allele.
Therefore, our data call into question the notion advanced by a previous study in which amiloride monotherapy appeared to have a particularly dramatic BP-lowering effect in black hypertensives with the T594M allele.8 The confirmation of the null hypothesis in our study may be related to several important methodological factors, including: (1) measurement of BP with ambulatory monitoring rather than office measurements; (2) a lower average level of pretreatment BP in our study, which avoids exaggeration of BP response; and, most importantly, (3) inclusion of a control group of black participants lacking the T594M allele. It is possible that a small gene effect was obscured by the number of T594M–positive subjects studied, and we did not assess response to dynamic challenges, such as a high-salt diet. Gene dosage might be important, because the mean BP in the 3 homozygotes identified in our cohort was 151/93; however, the average BMI of these subjects (all female) was 42.3, limiting our conclusions. Nonetheless, it is unlikely that any effect of the allele is sufficiently large to impact clinical decision making. SCNN1B genotyping does not appear to be a useful means of selecting amiloride therapy in blacks.
There are other limitations of our studies. As with all cardiovascular health surveys, the DHS certainly is not immune to nonparticipation biases related to socioeconomic status and other known ethnic disparities.9 However, these biases are relatively small when compared with the marked selection biases inherent in the small clinic-based studies that implicated the T594M allele. Because all of the ethnic designations thus far have been self-assigned, the current analyses theoretically are vulnerable to biases introduced by population stratification.23 The prominent ethnic-specific differences in allele frequencies suggest that such biases were rather small; these biases were additionally minimized by exclusion of nonblacks from all of the explanatory models and by the consistency of findings in 2 geographically independent samples with very different patterns of socialization.
The genetic basis of human hypertension has been established with certainty only for rare familial forms of severe hypertension that are inherited as Mendelian traits. Because in each case the molecular mechanism leads to overactivity of ENaC, an attractive hypothesis is that a less severe but more common mutation in ENaC subunits accounts for some cases of primary hypertension. However, the previously suggested association of the T594M allele with both hypertension (basal BP) and therapeutic responses to ENaC blockade with amiloride was not confirmed in 2 large, well-characterized population samples and by pharmacogenetic studies using appropriate controls. Although these particular studies were negative, these methodologies may prove useful in the identification of other genes or combinations of genes that prove more informative.
We thank Drs Jonathan Cohen and Tommy Hyatt for conducting the high throughput genotyping in the McDermott Core Genotyping Facility and Jody Balko for aldosterone and renin assays. We also thank the DHS investigators (named in reference 8) who contributed to the collection and construction of the database. This research was funded by a center grant from the Donald W. Reynolds Foundation (R.G.V., R.J.A., D.L., J.M.H., J.-L.L., M.G.C.), by grants from the National Heart, Lung, and Blood Institute (HL53353 and HL45508 to R.S.C.), by fellowship grants from the Doris Duke Foundation (D.F.M., D.B.), and partially supported by US Public Health Service General Clinical Research Center grant #M01-RR00633 from National Institutes of Health/National Center for Research Resources-Clinical Research.
- Received August 31, 2005.
- Revision received September 27, 2005.
- Accepted December 6, 2005.
Su YR, Rutkowski MP, Klanke CA, Wu X, Cui Y, Pun RY, Carter V, Reif M, Menon AG. A novel variant of the β-subunit of the amiloride-sensitive sodium channel in African Americans. J Am Soc Nephrol. 1996; 7: 2543–2549.
Baker EH, Dong YB, Sagnella GA, Rothwell M, Onipinla AK, Markandu ND, Cappuccio FP, Cook DG, Persu A, Corvol P, Jeunemaitre X, Carter ND, MacGregor GA. Association of hypertension with T594M mutation in β subunit of epithelial sodium channels in black people resident in London. Lancet. 1998; 351: 1388–1392.
Persu A, Barbry P, Bassilana F, Houot AM, Mengual R, Lazdunski M, Corvol P, Jeunemaitre X. Genetic analysis of the β subunit of the epithelial Na+ channel in essential hypertension. Hypertension. 1998; 32: 129–137.
Nkeh B, Samani NJ, Badenhorst D, Libhaber E, Sareli P, Norton GR, Woodiwiss AJ. T594M variant of the epithelial sodium channel β-subunit gene and hypertension in individuals of African ancestry in South Africa. Am J Hypertens. 2003; 16: 847–852.
Pegoraro RJ, Roberts CB, Rom L, Moodley J. T594M mutation of the epithelial sodium channel β-subunit gene in pre-eclampsia and eclampsia in Black South African women. Br J Obstet Gynecol. 2004; 111: 1012–1013.
Baker EH, Duggal A, Dong Y, Ireson NJ, Wood M, Markandu ND, MacGregor GA. Amiloride, a specific drug for hypertension in black people with T594M variant? Hypertension. 2002; 40: 13–17.
Victor RG, Haley WH, Willett DL, Peshock RM, Vaeth PA, Leonard D, Basit M, Cooper RS, Iannachione VG, Visscher WA, Staab JM, Hobbs HH. The Dallas Heart Study: a population-based probability sample for the multidisciplinary study of ethnic differences in cardiovascular health. Am J Cardiol. 2004; 93: 1473–1480.
National Health and Nutrition Examination Survey III interviewer’s manual. Available at: http://www.cdc.gov/nchs/data/nhanes/nhanes3/cdrom/nchs/manuals/fieldint.pdf. Accessed December 1, 2003.
Mohiaddin RH, Underwood SR, Bogren HG, Firmin DN, Klipstein RH, Rees RS, Longmore DB. Regional aortic compliance studied by magnetic resonance imaging: the effects of age, training, and coronary artery disease. Br Heart J. 1989; 62: 90–96.
Jones CR, Taylor K, Poston L, Shennan AH. Validation of the Welch Allyn ‘Vital Signs’ oscillometric blood pressure monitor. J Hum Hypertens. 2001: 191–195.
Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr, Jones DW, Materson BJ, Oparil S, Wright JT Jr, Roccella EJ. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA. 2003; 289: 2560–2572.
Turner ST, Schwartz GL, Chapman AB, Boerwinkle E. C825T polymorphism of the G protein β3-subunit and antihypertensive response to a thiazide diuretic. Hypertension. 2001; 37: 739–743.
Finkielman JD, Schwartz GL, Chapman AB, Boerwinkle E, Turner ST. Lack of agreement between office and ambulatory blood pressure responses to hydrochlorothiazide. Am J Hypertens. 2005; 18: 398–402.