Background: Salt-sensitive (SS) hypertension affects >30 million Americans and is often associated with low plasma renin activity. We tested the diagnostic validity of several candidate genes for SS and low-renin hypertension.
Methods: In Japanese patients with newly diagnosed, untreated hypertension (n = 184), we studied polymorphisms in 10 genes, including G protein–coupled receptor kinase type 4 (GRK4), some variations of which are associated with hypertension and impair D1 receptor (D1R)-inhibited renal sodium transport. We used the multifactor dimensionality reduction method to determine the genotype associated with salt sensitivity (≥10% increase in blood pressure with high sodium intake) or low renin. To determine whether the GRK4 genotype is associated with impaired D1R function, we tested the natriuretic effect of docarpamine, a dopamine prodrug, in normotensive individuals with or without GRK4 polymorphisms (n = 18).
Results: A genetic model based on GRK4 R65L, GRK4 A142V, and GRK4 A486V was 94.4% predictive of SS hypertension, whereas the single-locus model with only GRK4 A142V was 78.4% predictive, and a 2-locus model of GRK4 A142V and CYP11B2 C-344T was 77.8% predictive of low-renin hypertension. Sodium excretion was inversely related to the number of GRK4 variants in hypertensive persons, and the natriuretic response to dopaminergic stimulation was impaired in normotensive persons having ≥3 GRK4 gene variants.
Conclusions: GRK4 gene variants are associated with SS and low-renin hypertension. However, the genetic model predicting SS hypertension is different from the model for low renin, suggesting genetic differences in these 2 phenotypes. Like low-renin testing, screening for GRK4 variants may be a useful diagnostic adjunct for detection of SS hypertension.
Hypertension (blood pressure ≥140/90 mmHg) affects more than 65 million adult Americans, and prehypertension (blood pressure between 120/80 mmHg and 139/89 mmHg) affects millions more (1). Salt sensitivity, defined as a 10% increase in mean arterial blood pressure after consumption of a diet high in salt (2)(3)(4)(5)(6)(7)(8)(9), affects 58 million Americans without hypertension and can lead to morbidity and mortality rates similar to those of hypertension. New methods for patient identification and risk stratification should be developed for these chronic disease risk factors (10). Low plasma renin activity (PRA) 1 is used as a diagnostic adjunct for determining salt sensitivity despite the fact that salt sensitivity and low PRA do not correlate well (11)(12). The discovery of single-nucleotide polymorphisms (SNPs) in genes that encode proteins involved in sodium regulatory pathways could lead to new diagnostic tools for salt sensitivity with or without hypertension.
Renal dopamine receptors are responsible for more than 50% of sodium excretion during moderate sodium surfeit (13). Renal paracrine inhibition of sodium transport by dopamine is impaired in genetically hypertensive rats and humans with essential hypertension (13)(14)(15) because of decreased dopamine D1 receptor (D1R) function not related to a primary defect of the D1R but rather to its uncoupling from second messengers. This uncoupling is caused by activating variants of the G protein–coupled receptor kinase type 4 (GRK4) (13)(16), a gene involved in the desensitization of the D1R (16)(17)(18)(19). Three variants of GRK4 (R65L, A142V, and A486V) impair D1R stimulation of renal cAMP production (16). Preventing renal GRK4 expression normalizes D1R function in renal proximal tubule cells from hypertensive persons (16) and ameliorates hypertension in spontaneously hypertensive rats (20). In mice, overexpression of human GRK4 142V but not wild-type GRK4 impairs the natriuretic response to D1R stimulation and produces hypertension (16).
To determine whether certain genotypes are predictive of salt sensitivity and/or low-renin hypertension, we studied the association between polymorphisms in the genes GRK4 (16)(21)(22), D1R (23), α-adducin (ADD) (24)(25)(26), G-protein β3 subunit (GNB3) (27), CYP11B2 (aldosterone synthase) (28), angiotensin-converting enzyme (ACE) (26)(29), angiotensinogen (AGT) (21)(25)(30), angiotensin II type 1 receptor (AT1R) (31), and 11β-hydroxysteroid dehydrogenase 2 (11βHSD2)(32), all of which have been reported to be associated with low-renin and/or salt-sensitive (SS) hypertension. Because plasma plasminogen activator inhibitor-1 (PAI-1) concentrations also correlate with low-renin hypertension (33), we also examined variants of this gene in newly diagnosed, untreated hypertensive Japanese patients.
Our analyses included investigation of multilocus genetic models for these 2 phenotypes. As a physiologic correlate to our genetic studies, we measured the renal responses to docarpamine, a dopamine prodrug (34), in normotensive Japanese persons with or without GRK4 variants to test the hypothesis that GRK4 variants are related to the impairment of sodium excretion in response to D1R stimulation even in the absence of expression of the hypertensive phenotype.
The selection of cohorts for genetic studies can have a profound effect on study outcomes (35). We therefore selected a cohort of Japanese persons with a relatively homogeneous genetic background (30).
Patients and Methods
All protocols were carried out at the Fukushima Medical University with the approval of its Institutional Review Board (IRB).
Study participants were randomly selected Japanese patients with newly diagnosed, untreated essential hypertension who were referred to the hospital by office or outpatient-clinic physicians. We verified blood pressure with a mercury sphygmomanometer at least twice before enrollment. After providing informed consent, patients with systolic blood pressures ≥140 mmHg and/or diastolic blood pressures ≥90 mmHg were admitted to the study unit for physical examination, routine urine and plasma laboratory tests, electrocardiography, and a chest x-ray. Patients with diabetes mellitus, renal dysfunction (serum creatinine >11 mg/L; creatinine clearance <70 mL/min; microalbuminuria >30 mg/g of creatinine; and/or abnormal urinalysis), or secondary hypertension were excluded from the study. A total of 184 patients [104 females, 80 males; mean (SE) age, 54.8 (0.8) years; mean (SE) body mass index (BMI), 23.1 (0.2) kg/m2] were admitted into the study. After the diagnosis of essential hypertension was established, the study patients were admitted to the hospital for 5 weeks. After completion of the study of salt sensitivity and before evaluation of the relationship between genotype and phenotype, each patient was started on antihypertensive medication in accordance with the IRB.
We assessed salt sensitivity of blood pressure by determining the blood pressure responses to changes in dietary sodium (36)(37) in 83 study patients who were willing to comply with the dietary protocol to assess salt sensitivity in the hospital. Patients received a normal-sodium diet (153 mmol/day) and 50 mmol potassium/day for 2 weeks, followed by a published dietary protocol (2)(9)(36)(37) consisting of 2 weeks of normal sodium intake followed by a low-sodium diet (51 mmol/day) for 5 days, a high-sodium diet (340 mmol/day) for 5 days, and a normal sodium diet (153 mmol/day) for another 5 days. Adherence to the diet was evident because sodium balance was achieved and sodium intake was matched by sodium excretion (Fig. 1⇓ ). As expected, during the periods of high salt intake the sodium excretion in the SS patients did not match the sodium intake, but sodium balance was restored when normal sodium intake was reinstituted. On the final day of each regimen, we monitored ambulatory blood pressure (Nippon Colin ABPM-630). Patients were considered to be SS if mean arterial pressure increased (≥10%) after the change in sodium intake from low to high; otherwise they were considered to be salt resistant (SR) (2)(9)(11)(37) (see Fig. 1 in the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol52/issue3/ ). Mean arterial pressure was calculated as: diastolic blood pressure + 1/3 (systolic blood pressure − diastolic blood pressure).
Heart rates were not different between SS and SR individuals and were not affected by sodium intake. Therefore, heart rate was not taken into consideration in the calculation of mean arterial pressure.
We measured PRA in 2 groups of patients: 101 patients who remained on a normal sodium diet and 75 of the 83 patients studied for salt sensitivity (43 SR and 32 SS). In both groups, PRA was measured while patients were on a normal sodium diet.
Response to dopaminergic stimulation.
We tested the effect of GRK4 variants on the natriuretic effect of docarpamine (34) in normotensive persons homozygous for the wild-type GRK4 polymorphisms [n = 8; mean (SE) age, 38.0 (2.7) years; BMI, 25.0 (1.1) kg/m2] and those heterozygous for each of the GRK4 polymorphisms [n = 10; age, 32 (1.8) years; BMI, 23.9 (0.8) kg/m2]. These persons had serum creatinine and blood urea nitrogen (BUN) concentrations within the reference intervals, as did the hypertensive patients, but had higher BMIs than the hypertensive patients. Each participant rested in the study unit for 30 min and then ingested 500 mL of distilled water, after which blood pressure and heart rate monitoring were begun. Urine was collected each hour. A total of 5 samples were obtained: 1 before and 4 after the oral ingestion of 750 mg of docarpamine.
genotyping and blood analyses
Variants of GRK4 R65L, GRK4 A142V, GRK4 V247I, GRK4 A253T, GRK4 A486V, GRK4 G562D, AGT M235T, AT1R A1166C, CYP11B2 C-344T, PAI-1 4/5G, GNβ3 C825T, and 11βHSD2 G534A were detected by fluorescence probe melting curves (38). AGT A-6G, ACE I/D, ADD, and D1R were genotyped as reported (23)(26)(39)(40). No variations for 11βHSD2 G534A, GRK4 V247I, GRK4 A253T, or GRK4 G562D were detected; thus, these variants were not included in subsequent analyses.
Patient samples for fasting PRA and plasma aldosterone (determined by RIA) were obtained at 0830 after patients had rested 30 min in the recumbent position. PRA <1 g · L−1 · h−1 was designated as low renin (24). Catechols were analyzed by reversed-phase ion-pair HPLC with electrochemical detection. Routine laboratory analytes were measured with an automated method (Hitachi 7600-120).
Data are reported as the mean (SE). The effects of different sodium diets in SS and SR patients were evaluated by repeated-measures ANOVA. The linear contrast low/normal/high sodium was tested separately within each group and as a first-order interactive contrast in a two-way (factorial by repeated measures) ANOVA. No adjustments for multiple comparisons were applied to P values in Table 1⇓ (also see Table 1 in the online Data Supplement). In survey tables such as Table 1⇓ and Table 1 in the online Data Supplement, P values are descriptive rather than inferential. Thus, P values should be interpreted in light of the possibility of type 1 errors. Nevertheless, significance with Bonferroni correction for multiple comparisons was also noted (Table 1⇓ and Table 1 in the online Data Supplement). We determined differences between groups by t-test or χ2 test. Statistical analyses were performed with SPSS, Release 11.0.1 (SPSS Inc.), or SigmaPlot.
Allele frequency differences between SS and SR individuals and deviations from Hardy–Weinberg equilibrium were determined by TFPGA (http://bioweb.usu.edu/mpmbio/) (41). Associations of multilocus genotypes with salt sensitivity and PRA concentrations were assessed by multifactor dimensionality reduction (MDR). MDR assesses all possible genetic models and provides the best genetic model for differentiating salt sensitivity from salt resistance or low PRA from normal PRA based on both cross-validation consistency and prediction error (42)(43). The best genetic model for salt sensitivity was determined by the highest classification success and the lowest prediction error (42)(43). To determine the single best genetic model, we repeated the MDR analysis with all sets of loci with significant findings in the original analyses. Significance of the models was determined by permutation testing. Classic diagnostic measures of sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic odds ratio were also calculated.
Age [SS group, 59.3 (2.1) years; SR group, 57.1 (1.7) years], sex distribution (SS group, 26 females and 22 males; SR group, 21 females and 14 males), height [SS group, 158.4 (1.5) cm; SR group, 160.1 (1.5) cm], BMI [SS group, 22.7 (0.4) kg/m2; SR group, 22.6 (0.4) kg/m2], weight, and blood pressures (normal sodium diet, 153 mmol/day; see Table 1 in the online Data Supplement) did not differ between SS and SR hypertensive individuals. Increasing sodium intake from low to high increased blood pressures >10% in SS [criterion for salt sensitivity (2)(9)(11)(37)] but not in SR patients (See Table 1 and Fig. 2 in the online Data Supplement). Decreasing sodium intake from high to normal decreased blood pressure in SS but not in SR patients (Fig. 2⇓ ).
The small increase in diastolic blood pressure after the increase in sodium intake in SR patients did not meet the criterion for salt sensitivity; it was also not significant with the Bonferroni correction [a P value of 0.0014 is significant for 36 comparisons (27 comparisons per group and 9 intergroup comparisons)]. In both groups, body weight increased with the high-sodium diet, but the increase was greater in SS patients [1.66 (0.14) kg; 3.02 (0.26)%] than in SR patients [0.24 (0.09) kg; 0.38 (0.16)%]. A high-sodium diet decreased the hematocrit in both groups, but the decrease was greater in SS patients [2.04 (0.36); 5.58 (0.96)%] than in SR patients [0.55 (0.18); 1.55 (0.52)%]. Creatinine clearances were normal and similar in both groups and unaffected by sodium intake (Table 1 in the online Data Supplement). The change in sodium intake from low to high decreased PRA and aldosterone concentrations to a similar degree in both groups (Table 1 in the online Data Supplement).
A low-sodium diet decreased and a high-sodium diet increased sodium excretion in both the SR and SS groups (Fig. 1⇑ ). However, the increase in sodium excretion was more immediate and greater in SR than in SS patients, and when they returned to a normal-sodium intake, sodium excretion decreased more quickly in SR than in SS patients. Sodium excretion converged in the 2 groups 5 days after the decrease in sodium intake, from 340 to 153 mmol/day, but more sodium (81.9 mmol) was still retained in SS patients than in SR patients.
Urinary potassium excretion and the urinary sodium-to-potassium ratios were similar in both groups and positively correlated with sodium intake (data not shown). When evaluated with all the other variables listed in Table 1 in the online Data Supplement, urinary norepinephrine concentrations in both groups were found to be unaffected by sodium intake. However, when we examined the effect of dietary change on urinary norepinephrine separately from the other variables (Table 1 in the online Data Supplement), we found that urinary norepinephrine decreased with high sodium intake in the SR group but not in the SS group, in agreement with other reports (44)(45)(46). In contrast, urinary dopamine increased with sodium intake in both the SR and SS groups (Table 1 in the online Data Supplement), suggesting that decreased renal production of dopamine is not involved in the hypertension of our study participants (13).
The allele and genotype frequencies of ACE, AGT, AT1R, PAI-1, CYP11B2, D1R, GNβ3, and ADD did not differ between SR and SS patients (Table 1⇑ shows variants giving significant differences. The full data are shown in Table 2 in the online Data Supplement). In contrast, in SS patients, variant alleles at all 3 polymorphic GRK4 sites were more common and genotype frequencies differed between groups (P <10−4 for all comparisons; Table 1⇑ ; also see Table 2 in the online Data Supplement). Most sites were in Hardy–Weinberg equilibrium except for AT1R genotypes, which had deficiencies of heterozygotes in both SS and SR patients (P <10−4 and P = 0.0003 for SR and SS patients, respectively), GRK4 A486V in the SS group (P = 0.006), and ACE I/D in the SR group (P = 0.04). The GRK4 A486V deviation is the result of a deficiency in the number of wild-type homozygotes and the ACE I/D deviation is attributable to a deficiency of heterozygotes.
The multisite genetic model that best predicted salt sensitivity included only the GRK4 variants R65L, A142V, and A486V and predicted SS status correctly 94.4% of the time (P <0.001), based on the MDR analysis. The model using the 3 GRK4 variants for Japanese patients with SS hypertension had a sensitivity of 83% and a specificity of 100% with an infinite positive likelihood ratio, a 17% negative likelihood ratio, and an infinite diagnostic odds ratio. Although the presence of 3 or more variants was always associated with high risk for SS status (Fig. 3⇓ ), one should use caution in interpreting the accuracy of the risk assignment in genotypes with small numbers (n = 1). The model including all 3 GRK4 variants was significantly better than any of the models using individual GRK4 variants alone. This model is in agreement with the finding that the presence of 3 GRK4 variants was always associated with salt sensitivity, based on the actual increase in blood pressure with increased sodium intake (Fig. 4⇓ ). Indeed, the number of GRK4 alleles varied inversely with the increase in sodium excretion after the increase in sodium intake (r2 = 0.99; P <0.001; Fig. 4⇓ ).
The findings for allele and genotype frequency distributions between low-renin [mean (SE) age, 54.7 (1.2) years; BMI, 23.2 (0.3) kg/m2; 35 males and 59 females] and normal-renin [age, 53.6 (1.2) years; BMI, 23.1 (0.3) kg/m2; 42 males and 40 females] phenotypes were similar, but not identical, to those for salt sensitivity. All 3 GRK4 sites differed between low- and high-renin individuals (P <10−4 in all cases). However, there were also differences in ADD allele and genotype frequencies between the low- and high-renin groups (P = 0. 006 and 0.01, respectively). A few of the loci were out of Hardy–Weinberg equilibrium, including GRK4 A142V (P = 0.04) and GNB3 (P = 0.04) in the low-renin population and PAI-1 (P = 0.01) and ADD (P = 0.05) in the normal-renin population. However, the single best genetic model for low-renin hypertension included only GRK4 A142V, with an estimated prediction success of 78.4% (P <0.001). All variants that show single-site association, except GRK4 A142V, were not found in the best genetic model, indicating that the strongest effects are for this site. As opposed to the SS model, the GRK4 variants do not interact to increase disease predisposition. A 2-locus model that included both GRK4 A142V and CYP11B2 was also highly statistically significant (P <0.001) with prediction success at 77.8%, indistinguishable from the GRK4 A142V model.
Because salt sensitivity was always found in hypertensive patients with 3 or more GRK4 gene variants (Fig. 4⇑ ), we studied the effect of docarpamine on electrolyte excretion in normotensive patients without any or with all 3 GRK4 gene variants (Fig. 5⇓ ; also see Table 3 in the online Data Supplement). One oral dose of docarpamine (750 mg), which minimally affected blood pressure, induced natriuresis, kaliuresis, and calciuresis in patients without GRK4 gene variants but not in those with all 3 GRK4 gene variants (Fig. 5⇓ ; also see Table 3 in the online Data Supplement).
We report several novel findings. First, 3 of the 6 GRK4 variants (R65L, A142V, and A486V, but not the variants V247I, A253T, and G562D) were more frequent in SS than in SR hypertensive Japanese patients and predicted salt sensitivity 94.4% of the time. Second, there were no differences in the allele frequencies of other genes previously thought to be important in the etiology of salt sensitivity. Third, hypertensive Japanese patients with 3 or more GRK4 gene variant alleles at any of the 3 sites (R65L, A142V, and A486V) were always sensitive to salt, although some patients with <3 GRK4 variants were also salt sensitive. Fourth, the ability to excrete a salt load was inversely related to the number of GRK4 variant alleles (R65L, A142V, and A486V), and the natriuretic effect of a dopaminergic drug was abrogated when 3 GRK4 gene variants were expressed, even in normotensive Japanese persons. Fifth, the same 3 GRK4 gene variants were individually associated with low-renin hypertension, but in addition low- and normal-renin participants showed differences in ADD allele and genotype frequencies. Sixth, the underlying genetic models for salt sensitivity and low-renin hypertension were different, indicating that they have different etiologies and are not clinically equivalent among Japanese. These findings support the notion that even in the event of a single gene being critical in the development of a phenotype, multiple polymorphic sites should be considered in detection of at-risk persons.
Patients with essential hypertension and low PRA are presumed to have SS hypertension. Our SS hypertensive patients had low PRA and aldosterone concentrations, similar to those described for SS hypertensive patients in other populations (25). This profile in SS patients is in keeping with decreased D1R-mediated inhibition of renal sodium transport and sodium and fluid retention in SS hypertension (13)(15). However, our findings and those of others indicate that low specificity precludes the use of PRA concentrations as an effective way to distinguish SS from SR individuals (2)(11)(12).
The usefulness of blood pressure sensitivity to salt intake as an intermediate phenotype has been reported to be reduced by the variability in blood pressure response to increased sodium intake (47), but GRK4 variants were not investigated in that study. In our study, GRK4 variants were not only more frequent in SS than in SR hypertensive Japanese patients, but homozygous gene variants for GRK4 R65L and A142V were seen only in SS hypertensive Japanese patients. Among our hypertensive patients, only SS hypertensive patients had at least 3 GRK4 variant alleles (65L, 142V, and 486V).
Three GRK4 gene variants (R65L, A142V, and A486V; allele and genotype frequencies), but not the other 3 GRK4 gene variants (V247I, A253T, and G562D), were associated with low-renin hypertension in our hypertensive Japanese patients. ADD G460W was also associated with low-renin essential hypertension, in agreement with other reports (24)(25)(26)(48)(49). However, the single best genetic model for low-renin hypertension was GRK4 A142V, with 78.4% prediction accuracy in hypertensive Japanese, a result that differed from that for salt sensitivity. A 2-locus model including both GRK4 A142V and CYP11B2 was also highly statistically significant (P <0.001) with a prediction success of 77.8%, which was indistinguishable from the GRK4 A142V model. These analyses support the conclusion that salt sensitivity and low-renin hypertension are not of identical in genetic makeup or physiologic origins.
Gene–gene interactions are important in the development of hypertension (50). We have reported that in Ghanaian hypertensive patients, GRK4 variants (termed FJ) were in nonrandom association with the variants of ACE, AGT, and AT1R in hypertensive but not in normotensive individuals and that ACE and GRK4 2-locus genotypes were significantly predictive of essential hypertension (not characterized for salt sensitivity or PRA status) (22)(51)(52). SS hypertensive Italians also have increased frequency of GRK4 A486V compared with normotensive Italians (21), and the haplotype combination of GRK4 R65, 142V, and 486V was associated with essential hypertension in another group of Caucasians (53). The latter study also demonstrated a gene dosing effect (i.e., a positive correlation between the number of GRK4 alleles and blood pressure) in hypertensive patients, similar to our findings. Mice overexpressing human GRK4γ A486V on a C57BL/6 background developed hypertension only after sodium chloride intake was increased from 0.4% to 0.8% (54). In contrast, GRK4γ A142V transgenic mice on a C57BL/6 background were hypertensive regardless of salt intake (16)(55). Thus, different variants in the same gene can cause SS or salt-independent hypertension, depending on the gene variant, and may be influenced by the genetic background.
We found that a few of the studied genes were out of Hardy–Weinberg equilibrium, such as AT1R genotypes, which had deficiencies of heterozygotes in both SS and SR patients, GRK4 A486V, which had deficiencies in the number of homozygotes in the SS group, and ACE, which had deficiencies in the number of heterozygotes in the SR group. Deviations from Hardy–Weinberg equilibrium may be caused by several factors, including systematic genotyping error, random chance, and true biological causes. These are not easy to differentiate, but if genotyping errors are systematic, deviations should appear in both cases and controls. Because this was not the case for all deviations, excepting AT1R, we draw no strong conclusions regarding the causation of deviations from Hardy–Weinberg equilibrium among the other variants. Because AT1R did not appear to associate significantly with disease in any analysis, this does not pose a problem in our final interpretations. That GRK4 A486V in the SS group is out of Hardy–Weinberg equilibrium may be of interest because this is one of the variants that strongly associates with the SS phenotype. The deficiency in the number of wild-type GRK4 A486 homozygotes is what would be expected if variation at this site is functional. In addition, ACE I/D, although not associated in the present study with the phenotype, has been previously reported to be associated with essential hypertension in an epistatic model with GRK4 (51). Our current findings may reflect this previously described interaction. Additional testing for Hardy–Weinberg equilibrium in the samples for low-renin hypertension reinforces the conclusion that deviations from Hardy–Weinberg equilibrium are not attributable to genotyping error, because the deviating markers are different from those for SS and SR patients. This difference is unexpected if the deviations were attributable to experimental error. The positive association between certain variants (GRK4 and ADD) for at least one of the phenotypes of interest (salt sensitivity or low renin) indicates that our sample size had enough power to detect certain genes associated with SS and low-renin hypertension.
The high degree of correlation between the number of GRK4 gene variants and the reduced ability to excrete a sodium load may underlie the correlation between the loss of dopaminergic function and hypertension described previously (13)(14)(15)(16). Even normotensive persons with 3 GRK4 variants did not respond to the natriuretic effect of docarpamine. Because these normotensive persons were much younger (32 years) than the hypertensive patients (54.8 years), it is very possible that they will develop salt sensitivity and hypertension later in life, as has been reported for SS Americans (56). Indeed, our SS hypertensive patients tended to be older than the hypertensive patients with low PRA (P = 0.057). To conclusively relate the dopaminergic defect found in hypertensive rodents and humans to the studies presented here, we will have to determine the natriuretic effect of dopaminergic drugs in SS hypertensive patients with ≥3 GRK4 variant alleles. Nevertheless, in SS hypertensive white Americans, the D1-like agonist fenoldopam failed to decrease proximal tubular sodium reabsorption (15); however, the genotypes of these patients were not determined.
Although the number of individuals studied is sufficiently high to warrant confidence in the predictive value of GRK4 for salt sensitivity in the Japanese population, it may not be sufficiently high to rule out the possible association between the other polymorphisms we tested and SS hypertension. This can be determined only by increasing the number of patients analyzed. In addition, the ability to generalize these findings will require 2 steps. First, we will need to replicate the results in another Japanese cohort. Second, the role of GRK4 variation will have to be studied in another population. This would demonstrate that the GRK4 variants underlie the physiologic bases of salt sensitivity. Fortunately, Speirs et al. (53) have studied GRK4 polymorphisms and their role in hypertension with similar results in a cohort of hypertensive Caucasians.
In summary, the genotype and allele frequencies of variants of GRK4 (R65L, A142V, and A486V) are higher in SS than in SR patients, and variations at all 3 sites predict salt sensitivity in hypertensive Japanese. We did not detect any differences in allele frequencies of variants of ACE, AGT, AT1R, PAI-I, CYP11B2, D1R, GNβ3, and ADD genes between SS and SR patients. In contrast, low-renin hypertension was associated not only with GRK4 variants but also with ADD variants, and the genetic model for low-renin hypertension is different and does not include all of the sites detected in single-site association analyses. Therefore, salt sensitivity and low-renin hypertension do not have the same underlying genetic architecture. Moreover, the presence of 3 GRK4 gene variants can impair renal dopamine-induced natriuresis, even in the absence of hypertension. Determining the SNP genotype for selected loci might be considered a model molecular diagnostic screening test for salt sensitivity and hypertension.
This work was supported in part by grants from the National Institutes of Health (DK39308, HL23081, DK52612, HL68686, HL074940, and RR13297), a grant from the Fukushima Society for the Promotion of Medicine (No. 18), and a grant from the Salt Science Research Foundation (No. 05C6). In addition, this work was supported by a Small Technology Transfer Research Award from the NIH (1-R41 HL074526-01).
1 Genotype frequency comparisons are significant for SS vs SR (and low- vs normal-renin) classes at all GRK4 sites in all cases (P <0.001).
2 Allele frequency comparisons are significant for SS vs SR (and low- vs normal-renin) classes at all GRK4 sites in all cases (P <0.001).
↵1 Nonstandard abbreviations: PRA, plasma renin activity; SNP, single-nucleotide polymorphism; D1R, D1-dopamine receptor; GRK, G protein–coupled receptor kinase; ADD, α-adducin; GNβ3, G-protein β3 subunit; CYP11B2, aldosterone synthase; ACE, angiotensin-converting enzyme; AGT, angiotensinogen; AT1R, angiotensin II type 1 receptor; 11βHSD2, 11β-hydroxysteroid dehydrogenase 2; SS, salt sensitive; PAI-1, plasminogen activator inhibitor-1; IRB, Institutional Review Board; BMI, body mass index; SR, salt resistant; BUN, blood urea nitrogen; and MDR, multifactor dimensionality reduction.
- © 2006 The American Association for Clinical Chemistry