Author + information
- Received August 18, 2009
- Revision received November 10, 2009
- Accepted November 24, 2009
- Published online April 6, 2010.
- Kuixing Zhang, MD, PhD*,
- Fangwen Rao, MD*,
- Lei Wang, PhD*,
- Brinda K. Rana, PhD‡,
- Sajalendu Ghosh, PhD*,
- Manjula Mahata, PhD*,
- Rany M. Salem, PhD*,
- Juan L. Rodriguez-Flores, PhD*,
- Maple M. Fung, MD*,
- Jill Waalen, MD, MPH¶,
- Bamidele Tayo, PhD#,
- Laurent Taupenot, PhD*,
- Sushil K. Mahata, PhD*,∥,* ( and )
- Daniel T. O'Connor, MD*,†,§∥,* ()
- ↵*Reprint requests and correspondence:
Dr. Daniel T. O'Connor or Dr. Sushil K. Mahata, Department of Medicine and CHGG, UCSD School of Medicine, 9500 Gilman Drive, La Jolla, California 92093-0838
Objectives The purpose of this study is to understand whether naturally occurring genetic variation in the promoter of chromogranin B (CHGB), a major constituent of catecholamine storage vesicles, is functional and confers risk for cardiovascular disease.
Background CHGB plays a necessary (catalytic) role in catecholamine storage vesicle biogenesis. Previously, we found that genetic variation at CHGB influenced autonomic function, with association maximal toward the 5′ region.
Methods Here we explored transcriptional mechanisms of such effects, characterizing 2 common variants in the proximal promoter, A-296C and A-261T, using transfection/cotransfection, electrophoretic mobility shift assay (EMSA), and chromatin immunoprecipitation (ChIP). We then tested the effects of promoter variation on cardiovascular traits.
Results The A-296C disrupted a c-FOS motif, exhibiting differential mobility shifting to chromaffin cell nuclear proteins during EMSA, binding of endogenous c-FOS on ChIP, and differential response to exogenous c-FOS. The A-261T disrupted motifs for SRY and YY1, with similar consequences for EMSA, endogenous factor binding, and responses to exogenous factors. The 2-SNP CHGB promoter haplotypes had a profound (p = 3.16E-20) effect on blood pressure (BP) in the European ancestry population, with a rank order of CT<AA<<CA<AT on both systolic blood pressure (SBP) and diastolic blood pressure (DBP), accounting for ≈2.3% to ≈3.4% of SBP/DBP variance; the haplotype effects on BP in vivo paralleled those on promoter activity in cella. Site-by-site interactions at A-296C and A-261T yielded highly nonadditive effects on SBP/DBP. The CHGB haplotype effects on BP were also noted in an independent (African ancestry) sample. In normotensive twins, parallel effects were noted for a pre-hypertensive phenotype, BP response to environmental stress.
Conclusions The common CHGB promoter variants A-296C and A-261T, and their consequent haplotypes, alter binding of specific transcription factors to influence gene expression in cella as well as BP in vivo. Such variation contributes substantially to risk for human hypertension. Involvement of the sex-specific factor SRY suggests a novel mechanism for development of sexual dimorphism in BP.
The sympathoadrenal system exerts minute-to-minute control over cardiac output and vascular tone. Genes governing catecholaminergic processes may play a role in the development of hypertension (1). Sympathoadrenal catecholamine secretion is exocytotic (all or none), releasing not only catecholamines but also the acidic proteins with which catecholamines are stored. The chromogranins/secretogranins comprise a family of acidic, soluble proteins that are stored in secretory granules with different hormones, transmitters, and neuropeptides throughout the endocrine and nervous system (2). Chromogranin B (CHGB), first described in the 1980s (3–5), is a major catecholamine storage vesicle core protein and seems to play a necessary role in the biogenesis of catecholamine secretory vesicles (6).
CHGB is differentially expressed in neuroendocrine diseases, and its measurement may serve in the diagnosis and staging of such conditions (7–15). Interaction of CHGB with signaling molecules such as the inositol-1,4,5-trisphosphate-activated calcium channel (16) may influence cytosolic calcium and ultimately risk for such disease states as Alzheimer's disease, epilepsy, or schizophrenia (17). In addition, polymorphisms in the CHGB gene may be associated with schizophrenia in Chinese and Japanese populations (18,19).
Expression of CHGB may mark the action of still poorly characterized trans-quantitative trait locus (QTL) influencing exocytotic sympathoadrenal activity (20,21). CHGB is overexpressed in rodent models of genetic (22,23) as well as acquired (24) hypertension, thus suggesting augmented sympathoadrenal activity in the pathogenesis of these syndromes. Therefore, CHGB might give rise to early, pathogenic “intermediate phenotypes” (25) for exploration of sympathoadrenal activity in human essential hypertension.
Previously, we described genetic variation at the CHGB locus, and concluded that sex and CHGB interact to influence blood pressure (BP) (26). Since the association of CHGB genetic variation to BP was maximal toward the 5′ end of the gene, and such variation predicted quantitative changes in CHGB expression, we turned to potential transcriptional mechanisms. Here, we characterize the proximal promoter region of CHGB, discovering 2 common polymorphisms that disrupt transcription factor binding, giving rise to systemic hypertension. One of these sites recognizes the sex-specific factor, sex-determining region Y (SRY), providing insight into the sexual dimorphism of BP.
Subjects and clinical characterization
Subjects were volunteers, and each gave informed, written consent to protocols approved by local institutional review boards. Recruitment procedures, definitions, and confirmation of subject diagnoses are according to previous reports.
Systematic polymorphism discovery across the CHGB locus
As previously described (26), we resequenced each of CHGB's 5 exons, exon/intron borders, untranslated region, and proximal promoter in n = 160 subjects (2n = 320 chromosomes) of 4 self-identified biogeographic ancestries: white/European (n = 56), black/sub-Saharan African (n = 56), Hispanic/Mexican American (n = 24), and east Asian (n = 24). We used an ABI-3100 capillary system (Applied Biosystems, Foster City, California) to accomplish dideoxy sequencing.
Primary care population with extremes of high and low BP
As previously described (26), we ascertained 951 European-ancestry subjects, approximately half male and half female, from the highest and lowest fifth diastolic blood pressure (DBP) percentiles of a large primary care population in the Kaiser-Permanente Medical Group of Southern California (27). The DBP criterion was chosen because of the heritability of DBP (28). The statistical power of association between biallelic deoxyribonucleic acid (DNA) markers and human quantitative trait loci can be substantially augmented by the sample subjects from opposite (upper and lower) ends of the trait distribution (29–31); and analyses of the quantitative trait in extreme subjects (as opposed to dichotomization of the trait) further enhances power (32). This population sample afforded us >90% power (29,30) to detect genotype association with a trait when the genotype contributes as little as 2.5% to the total variation in males (even at p < 10−8); the power is even higher in females (31). Evaluation included physical examination, blood chemistries, hemogram, and extensive medical history questionnaire. Of the hypertensive group, 40.6% were receiving antihypertensive medications, whereas none in the normotensive group was receiving such drugs. The subjects are described in Online Table 1. Overall, 1.98% of subjects were excluded because of elevated serum creatinine (>1.5 mg/dl).
Black population with blood pressure extremes
Sub-Saharan African Ancestry
Three hundred fifty-seven adult Nigerians (approximately one-half male, one-half female) selected from the highest (n = 191) and lowest (n = 190) 25th percentiles of population BP were included as a replication sample for CHGB promoter variant effects on BP. This population has been described (33).
In studies of the influence of CHGB polymorphism on the pressor response to environmental stress in vivo, 156 twin pairs and 80 siblings (312 individual subjects) were evaluated. The response of BP to cold stress (by immersion of 1 hand in ice water for 1 min) was evaluated as previously described (34); responses wherein DBP increased after cold stress were analyzed. Zygosity (69% monozygotic and 31% dizygotic pairs) was confirmed by extensive microsatellite and single nucleotide polymorphism (SNP) genotyping, as described (35). Twins ranged in age from 15 to 84 years; 10% were hypertensive. Twins in these allelic/haplotype association studies were self-identified as of European (white) ancestry, to guard against potential artifactual effects of population stratification.
Haplotype blocks were visualized in Haploview (36), while haplotype assignments in individual subjects were performed by the HAP algorithm (37) in individual subjects with both A-296C and A-261T genotypes. Chi-square tests were performed to test for deviations from the Hardy-Weinberg equilibrium (HWE). When testing for associations of haplotypes with continuous/quantitative BP traits, sex, age, and body mass index (BMI) were included as covariates in the univariate tests of the general linear model using SPSS version 11.5 software (SPPS Inc., Chicago, Illinois). Both the final measured BP and that BP adjusted for the effects of antihypertensive medication (38) were analyzed. Each factor was then assessed for significance using standard analysis of variance (ANOVA) F-tests (39). Haplotype analyses were both (diploid) individual based as well as chromosome based; here, each haplotype allele (as opposed to a haplotype allele pair) was considered and analyzed separately, using the outcomes and characteristics of the subject carrying that allele (40). Associations between BP status and allele, genotype or haplotype were analyzed in n × 2 tables by either ANOVA or by SHEsis (41). Twin analyses were conducted in 2 ways: twin trait heritability (h2) was estimated in SOLAR (Southwest Foundation for Biomedical Research, San Antonio, Texas) (42); twin descriptive and inferential statistics were computed by using generalized estimating equations in SAS software (SAS Institute, Cary, North Carolina), to account for intrapair correlations (35). A p value of ≤0.05 was considered significant.
Genomic DNA was prepared from leukocytes in ethylenediaminetetra-acetic acid–anticoagulated blood, using PureGene extraction columns (Gentra Systems, Minneapolis, Minnesota).
Transfected CHGB promoter haplotype/luciferase reporter activity, genotyping of CHGB variants, electrophoretic gel mobility shift assays (EMSA), and chromatin immunoprecipitation (ChIP) were performed (for details, see the Online Appendix).
Patterns of linkage disequilibrium across CHGB locus
To visualize association patterns, 16 SNPs (each in HWE) were scored and plotted by Haploview as pairwise linkage disequilibrium (LD) parameter r2 across the ∼14 kbp locus. The proximal promoter (including common variants A-296C and A-261T) was maintained within a single block in both white subjects and black subjects. The allele and genotype frequencies differed between white and black populations (Online Table 2). Although pairwise r2 values were generally higher in white subjects than in black subjects, just 2 LD blocks spanned the locus in each group (Fig. 1A). The original resequencing strategy and SNP discovery have been described (26).
Domains and motifs in the CHGB promoter
Figure 1B diagrams known motifs in the CHGB promoter, and superimposes common variants. Functional domains in the core/proximal promoter (such as the TATA box, cyclic-adenosine monophosphate response element, and G/C-rich regions) were invariant in ∼180 people (∼360 chromosomes) subjected to systematic polymorphism discovery by resequencing.
Two very common SNPs (minor allele frequency [MAF] >30%) occur in the proximal promoter: A-296C and A-261T, whose allele, diploid genotype, and haplotype frequencies differed by ethnicity (Online Table 2). The A-296C variant lies in a c-FOS transcriptional control motif (-298/-291), whereas A-261T lies in recognition motifs for both Yin-Yang 1 (YY1) (-264/-259) and SRY (-265/-260).
CHGB promoter haplotype/reporter activity assays
To probe the functional significance of the 2 common promoter variants for transcriptional efficiency, we inserted each of the 4 haplotypes of the 2 promoter SNPs (A-296C and A-261A) into the luciferase reporter plasmid pGL3-Basic (Online Fig. 1). After transfection into rat chromaffin (PC12) cells, these 4 haplotypes yielded substantially different luciferase reporter activity (Table 1). Site-specific effect analysis of reporter activity showed that both A-296 and A-261 have context-dependent actions. The strengths of combinations of A-296C and A-261T were AT>CT>AA>CT (p = 8.46E-07) (Table 1). By 2-way ANOVA, overall F = 107.2, p = 8.46E-07; A-296C, p = 2.09E-05; A-261T, p = 0.064; and 296-by-261 interaction, p = 3.07E-07.
We evaluated responses of CHGB promoter haplotypes to agents simulating natural secretory stimuli (Fig. 1C). Pituitary adenylate cyclase-activating peptide increased promoter activity by ∼4.1- to 7.7-fold (p = 4.86E-14), while nicotine increased activity by ∼0.6-fold to 1.5-fold (p = 1.34E-8). There were differences among the 4 haplotypes in response to pituitary adenylate cyclase-activating peptide (p = 1.9E-5), although nicotine responses were similar (p = 0.31).
CHGB promoter variant A-296C
The A-296C variant lies in an evolutionarily conserved region among humans, nonhuman primates, and other mammals (Online Fig. 2). During EMSA, a labeled oligonucleotide representing the C allele was shifted by PC12 nuclear proteins; specificity was suggested by displacement with the same (C) allele when unlabeled (Online Fig. 3).
A-296C occurred in a potential recognition site for transcription factor c-FOS, with a 7/8 base consensus match for the A allele (43), declining to 6/8 for the C allele (Fig. 2A). When interrogated by ChIP, endogenous c-FOS binding to the motif was detected in all 4 haplotypes, although unexpectedly more intense for the A-296C than the A-296 allele (Fig. 2A).
When a plasmid expressing c-FOS was cotransfected into PC12 cells with CHGB promoter/reporters, all 4 CHGB haplotypes responded (p = 2.36E-9), although unequally (p = 7.23E-5). On a haplotypic background of the A-261 allele, c-FOS stimulated A-296 and A-296C similarly, although on a background of the -261T allele, the transcriptional response of A-296 was far greater than that of A-296C (Fig. 2B). Thus, the A-296C response to exogenous c-FOS seemed to be context dependent.
CHGB promoter variant A-261T
A-261T variant also occurs in an evolutionarily conserved region (Online Fig. 4). During EMSA, the T allele was more effectively shifted by PC12 nuclear proteins than the A allele (Online Fig. 5). During EMSA, an oligonucleotide spanning the A allele was shifted by PC12 nuclear proteins, while the T allele was shifted to a lesser degree; specificity was suggested, especially for the A allele, by displacement with the same allele when unlabeled (Online Fig. 5).
A-261T occurred in potential recognition motifs for the transcription factors SRY and YY1: the A allele displayed a superior match to both SRY (5 of 6 bases) (44,45) and YY1 (6 of 6 bases) (43) motifs, as compared with the T allele (Fig. 3B). Involvement of endogenous SRY and YY1 was probed by ChIP (Fig. 3A): for both SRY and YY1, the A allele was more effectively bound than the T allele, on either A-296C haplotypic background.
When a plasmid expressing SRY was cotransfected into PC12 cells with CHGB promoter/reporters, all 4 haplotypes showed decreased reporter activity (p = 1.15E-10) (Fig. 3B), although the degree of inhibition depended on A-296C background (p = 4.29E-8).
When cotransfected with a YY1 expression plasmid, CHGB promoter reporter activity increased for each haplotype (p = 7.6E-10), and the effect was more prominent for the A-261 allele than for the A-261T allele, regardless of A-296C context (p = 1.64E-4) (Fig. 3B).
CHGB promoter common variants A-296C and A-261T
Implications for Hypertension in the Population
Here we studied BP trait-extreme subjects of European ancestry, to enhance statistical power (29,30). Chromosome-based haplotype analysis on subjects dichotomized into 2 groups (higher BP vs. lower BP) indicated that subjects with the less common AT or CA haplotypes had a strong tendency to be hypertensive (odds ratio [OR]: 4.898 for AT haplotype, p = 3.19E-11; OR: 3.84 for CA haplotype, p = 3.64E-10). Subjects with the more common haplotype CT had a strong tendency to be normotensive (OR: 0.637, 95% confidence interval: 0.524 to 0.773, p = 4.64E-6). The most common haplotype AA had no effect on blood pressure status. The overall effect of CHGB haplotypes on BP status was substantial (Fig. 4A,Table 2), whether analyzed by chromosome/haplotype (global chi-square = 93.9, p = 3.16E-20), or by diploid haplotype pairs (global chi-square = 75.0, p = 4.92E-13).
Single SNP-based allele tests showed far less power to detect BP associations (Table 2): A-296C (although not A-261T) had an effect on BP status (p = 0.038), with the A-296 allele tending toward hypertension. Of note, when subjects were stratified by sex, A-261T displayed significant effects on systolic blood pressure (SBP)/DBP in males (p < 0.001/p = 0.011), although not in females.
We also pursued association of the quantitative traits (SBP and DBP in mm Hg) with CHGB haplotypes (Fig. 4B), and here we found substantial predictions for both SBP and DBP. With increasing copy number (0, 1, 2) of haplotype AT, SBP increased by ∼29 mm Hg (p = 0.0002), while DBP increased by ∼21 mm Hg (p = 1.59E-5), each in additive fashion. With increasing copy number of haplotype CA, SBP increased by ∼13 mm Hg (p = 0.005) while DBP increased by ∼12 mm Hg (p = 8.49E-6). With increasing copy number of haplotype CT, however, SBP decreased by ∼11 mm Hg (p = 0.0045), with a parallel decrease in DBP by ∼7 mm Hg (p = 0.011). When we adjusted BP values for the effects of antihypertensive medications >in treated hypertensive patients (38), the haplotype effects on SBP/DBP persisted or increased: AT, p = 4.1E-5/p = 6.83E-6; CA, p = 0.008/p = 2.04E-5; and CT, p = 0.002/p = 0.008. Haplotypes AA and AT displayed more prominent effects on BP in females than in males, whereas CT had a greater effect in males.
Promoter polymorphisms A-296C/A-261T interacted nonadditively (p = 1.15E-6/p = 1.07E-10) to influence SBP and DBP (Fig. 4C). On a background of A-296C major allele homozygosity (A/A), the A-261T major (A) allele lowered SBP/DBP by ∼27/∼22 mm Hg, whereas on a background of A-296C minor allele homozygosity (C/C), the A-261T major (A) allele elevated SBP/DBP by ∼21/∼18 mm Hg. When treatment-adjusted (38) BPs were analyzed, the interactions persisted or increased (SBP/DBP: p = 4.41E-7/p = 1.08E-10). The A-296C-by-A-261T (SNP-by-SNP) interaction was confirmed on analyses of the dichotomous BP trait (higher vs. lower): p = 3.71E-15.
The SBP/DBP values predicted by CHGB promoter haplotypes were, in rank order (Fig. 4D, left), CT<AA<<CA<AT (SBP/DBP: p = 6.16E-8/p = 5.18E-12). On ANOVA, CHGB promoter genetic variation accounted for ∼2.3% of SBP variance and ∼3.4% of DBP variance in this primary care population.
Parallel effects of human CHGB promoter haplotypes in vivo and in cella
Of note, the 4 promoter haplotypes display the same rank order for effects on BP in vivo and luciferase reporter activity in chromaffin cells in cella (Fig. 4D, right).
Extension to a second population
Sub-Saharan African Hypertension
An association between CHGB promoter haplotypes and hypertension was also found in a Nigerian population selected for extreme BP values (top and bottom 25th percentiles; p = 0.007 for BP status) (Online Table 3). Here, haplotype -296C/A-261 increased SBP by ∼34 mm Hg (p = 0.002) and DBP by ∼22 mm Hg (p = 3.52E-4) (Fig. 4E, left). The rank order of overall haplotypic variation on blood pressure (Fig. 4E, right) was AA, AT<CT<<CA (SBP/DBP: p = 0.007/p = 0.002). Whereas haplotype CA was associated with higher BP in both black subjects and white subjects, haplotype AT predicted higher BP only in whites. However, CHGB promoter allele and haplotype frequencies differed substantially between black subjects and white subjects (Online Table 2), perhaps contributing to different haplotype effects on BP; for example, haplotype AT is relatively unusual in white persons but far more common in black persons (Fig. 4E, Online Table 2). The CHGB promoter haplotypes did not clearly differ in effects on BP between sexes.
Association of Heritable BP Response to Environmental Stress with CHGB Promoter Variants A-296C and A-261T
In predominantly (∼90%) normotensive white twin pairs (n = 163 pairs), the DBP response to environmental (cold) stress was substantially heritable (h2 = 32 ± 8%, p = 0.0003) (46). The CHGB promoter individual genotypes A-296C (A>C, p = 0.0237) and A-261T (A>T, p = 0.037) predicted change DBP (ΔDBP) in the cold stress test. Because of the modest sample size (2n = 326 chromosomes), we observed only 5 examples of CA haplotypes and 1 AT haplotype, so analyses could only be performed for haplotypes AA and CT. Consistent with basal BP in white BP extremes (Fig. 4A), the CT haplotype decreased the stress ΔDBP, while the AA haplotype increased ΔDBP (Fig. 5); thus, the rank order of effects of the haplotypes on BP was preserved in twins (AA>CT), although the effects of the more prominent BP-increasing haplotypes (CA, AT) could not be quantified in this predominantly normotensive sample.
Patients with hypertension often exhibit increased sympathetic activity (47,48), and hypertension tends to develop in persons with sympathetic overactivity (49,50). Suppression of CHGB expression in chromaffin cells leads to a reduction in the number of catecholamine secretory granules, whereas ectopic expression of CHGB in nonneuroendocrine cells, which normally do not contain regulated secretory machinery, leads to granule biogenesis (6). In light of the emerging secretory biology of CHGB, we undertook the present study to probe how heredity shapes human functional responses in the sympathetic neuroeffector junction, using CHGB as a likely focal point in the pathogenesis of essential hypertension. Recently, we reported CHGB haplotype effects on BP across the CHGB locus, suggesting that the major effect was located in the 5'/promoter region (26), and the effect of CHGB on intermediate traits seems to be quantitative rather than qualitative; thus, we focused here on promoter variation at the CHGB locus. We also noted a sex-dependent effect of CHGB polymorphism on BP (26); here, we defined the effect of genetic variation on gene expression, and found evidence that the response of the gene to the male-determination factor SRY was altered by 1 promoter variant (A-261T), raising a potential mechanism for the well-known sexual dimorphism in BP.
CHGB promoter common variants
Transcriptional mechanisms of action. We identified 2 common variants in the CHGB proximal promoter: A-296C and A-261T. Here, we established that these 2 variants occur in evolutionary conserved regions, and can create or interrupt particular transcriptional control motifs. We probed these processes in 2 ways: by exposing the variants to the exogenous factors, and by testing whether the endogenous factors recognize the motifs.
A-296C Variant: c-FOS Motif
A c-FOS motif was activated by the exogenous factor and bound by the endogenous factor (Fig. 2). The c-FOS, a b-ZIP (leucine zipper) factor of the immediate/early class, may heterodimerize with a variety of other such family members (e.g., c-JUN) to trigger transcription, especially by activating AP-1 sites.
A-261T Variant: YY1 and SRY Motifs
This variant spanned motifs for both YY1 and SRY (Fig. 3). Exogenous YY1 increased CHGB promoter expression (especially for the A-261 allele), whereas exogenous SRY decreased expression; endogenous factors YY1 and SRY each bound the motif, preferentially for the A-261 (major) allele.
The YY1 motif is a widely expressed C2H2 zinc finger transcription factor whose ability to direct local histone modifications within chromatin yields fundamental roles in embryogenesis, differentiation, replication, and proliferation (51).
The SRY protein, testis-determining factor, is a high-mobility group box factor best known as an initiator of male development, in which its major transcriptional targets may include SOX9 (52). Of note for hypertension, we previously found that promoter A-261T exerted a sex-specific effect on population BP, with the effect confined to males (26). Here, we provide a basis for that sex-specific effect, because only males express the SRY factor, encoded by the SRY locus on human chromosome Yp11.31. While we focused on the SRY motif match in the A-261T region of the CHGB promoter (Fig. 3A), and demonstrated its binding (Fig. 3A) and trans-activation (Fig. 3B) by SRY, other members of the SOX transcription factor family, including the SRY target SOX9, may share similar consensus DNA target motifs (53) (i.e., SRY as WACAAW; SOX9 as AACAAT), and hence, both constitute potential CHGB trans-activators.
Interactions A-296C by A-261T
The likelihood of site-by-site interactions within the CHGB promoter (Fig. 4C) is suggested by 2 previous observations. First, the YY1 promoter responds transcriptionally to c-FOS (54). Second, the multifunctional factor YY1 interacts noncovalently with a variety of other transcription factors, including members of the b-ZIP family such as CREB (55). While we have documented an interaction in cis between A-296C and A-261T in transfected CHGB promoter haplotype/luciferase reporter plasmids (296-by-261 interaction, p = 3.07E-07) (Online Table 1), we have not yet explored factor interactions in trans during such transfections.
CHGB promoter common variants and hypertension across populations
The LD analysis across the CHGB locus indicated that the proximal CHGB promoter, including common variants A-296C and A-261T, is maintained within 1 block (Fig. 1A) in both white subjects and black subjects.
We enhanced power to detect genetic associations by using population trait-extreme values (29,30). We then found that haplotype effects upon BP in the population were highly significant (Fig. 4), indeed far more significant than the effects of single SNPs alone (Table 2). Substantially greater effects on BP by the A-296C/A-261T haplotypes (Fig. 4A) than by either SNP alone (Table 2), also speaks toward functional SNP-by-SNP interactions in the CHGB promoter (see transcription factor section in preceding text).
In the European ancestry population, the rank order of haplotype effects on SBP or DBP was AT>CA>>AA>CT (Fig. 4D, left), which is the same pattern of CHGB promoter haplotype activity in cella (Fig. 1C), lending weight to the viewpoint that altered CHGB transcription underlies the BP differences between haplotypic groups.
In an independent (African) population, allele and haplotype frequencies differed substantially from those in subjects of European ancestry (Online Table 2). Even so, CA haplotype copy number (0, 1, 2) influenced BP in the Nigerian sample (Fig. 4E), with the same directional effect found in white subjects (Fig. 4B). Although haplotype AT was found in substantial numbers in Nigerians (134 chromosomes), AT carriers did not display elevated BP, suggesting other factors (such as differences in environment or genetic background) influencing BP in this population, or the inclusion of other variants in the promoter LD block in subjects of African ancestry (Fig. 1A).
A previous large, genome-wide association case/control study, the WTCCC (Wellcome Trust Case Control Consortium) study, did not find association of the CHGB locus to hypertension (56). How is our study different? First of all, the WTCCC study used the Affymetrix 500K gene chip, whose average marker spacing of ∼3 × 109/500 × 103, or ∼6 kbp, did not include the CHGB promoter variants considered here; indeed, the closest CHGB promoter variant studied on the Affymetrix 500K chip in the WTCCC study was rs236129, which is 4,729 bp upstream of the cap site. Second, the WTCCC study employed single point associations, rather than studying haplotype or SNP-by-SNP interaction effects (which were crucial in our analyses). Third, because of relatively sparse marker spacing, the HapMap approach employed in the WTCCC study does not fully capture the full spectrum of potentially causal allelic variation at candidate loci (57,58). Finally, the WTCCC study used unselected/unphenotyped population controls (59); the high population prevalence (∼24%) of hypertension thus greatly diminishes the power of the WTCCC study to detect associations with hypertension. By contrast, our approach (31), using population trait (BP) extremes, offers substantially greater statistical power to detect genetic associations; indeed, we estimate that our sample has >80% power to detect loci contributing as little as ∼2.5% of BP variance.
CHGB Promoter Variants and Heritable BP Response to Environmental Stress in Twin Pairs
In longitudinal studies, the pressor response to environmental (cold) stress is an effective predictor of future development of hypertension (60,61). We therefore evaluated whether CHGB genetic variation in the transcriptional control region might influence this risk predictor.
In a series of twin pairs, CHGB promoter common haplotypes influenced ΔDBP in the cold stress test (Fig. 5), with A-296/A-261 elevating and -296C/-261T diminishing the pressor response, findings that are in rank order with basal BP effects in the population (AA<CT) (Fig. 4A). Haplotypes associated with much higher BP in the population extreme subjects (AT, CA) (Fig. 4A) were too infrequent in this predominantly normotensive twin sample for meaningful conclusions to be drawn.
Common genetic variants in the CHGB proximal promoter seem to exert a powerful, interactive effect on BP. The CHGB promoter variants A-296C and A-261T differed substantially in transcriptional efficiency during luciferase reporter activity assays (Fig. 1C), and such activity paralleled SBP and DBP in the population (Fig. 4D, right). Particular transcription factors (c-FOS at A-296C; YY1 and SRY at A-261T) differed in activity at the variant sites (Figs. 2 and 3); such effects were captured by both cotransfection and ChIP. Differential SRY effects at A-261T suggest a mechanism that might ultimately contribute to the sexual dimorphism of BP in the population. Substantially greater effects on BP of the variants in combination (rather than as individual SNPs) suggest intrapromoter A-296C by A-261T interactions. Finally, CHGB promoter variants predict change in BP in response to environmental stress even in predominantly normotensive subjects, suggesting an early pathway by which hypertension may ultimately be mediated.
Thus, common genetic variation at the CHGB locus, especially in the proximal promoter, influences CHGB expression, and later the early heritable responses to environmental stress, and finally resting/basal BP in the population (Fig. 6). These results point to new molecular strategies for probing autonomic control of the circulation, and ultimately the susceptibility to and pathogenesis of cardiovascular disease states such as hypertension.
For an expanded Statistics section and supplementary tables and figures, please see the online version of this article.
This study is supported by the Department of Veterans Affairs, National Institutes of Health (NIH), the NIH/National Heart, Lung, and Blood Institute (HL58120), the NIH/National Center on Minority Health and Health Disparities-sponsored (MD000220) EXPORT/Comprehensive Research Center in Health Disparities Minority Health Center, and the NIH/National Center for Research Resources-sponsored (RR00827) General Clinical Research Center.
- Abbreviations and Acronyms
- analysis of variance
- body mass index
- blood pressure
- chromogranin B
- chromatin immunoprecipitation
- diastolic blood pressure
- dioxyribonucleic acid
- electrophoretic mobility shift assay
- Hardy-Weinberg equilibrium
- linkage disequilibrium
- minor allele frequency
- odds ratio
- quantitative trait locus
- systolic blood pressure
- single nucleotide polymorphism
- sex-determining region Y
- Yin-Yang 1
- Received August 18, 2009.
- Revision received November 10, 2009.
- Accepted November 24, 2009.
- American College of Cardiology Foundation
- Rosa P.,
- Hille A.,
- Lee R.W.,
- Zanini A.,
- De Camilli P.,
- Huttner W.B.
- Huh Y.H.,
- Jeon S.H.,
- Yoo S.H.
- Stridsberg M.,
- Husebye E.S.
- Greenwood T.A.,
- Cadman P.E.,
- Stridsberg M.,
- et al.
- Greenwood T.A.,
- Rao F.,
- Stridsberg M.,
- et al.
- Takiyyuddin M.A.,
- De Nicola L.,
- Gabbai F.B.,
- et al.
- Lillie E.O.,
- O'Connor D.T.
- Zhang K.,
- Rao F.,
- Rana B.K.,
- Gayen J.R.,
- et al.
- Wessel J.,
- Moratorio G.,
- Rao F.,
- et al.
- Wen G.,
- Wessel J.,
- Zhou W.,
- et al.
- Rao F.,
- Wen G.,
- Gayen J.R.,
- et al.
- Rao F.,
- Zhang L.,
- Wessel J.,
- et al.
- Barrett J.C.,
- Fry B.,
- Maller J.,
- Daly M.J.
- Halperin E.,
- Eskin E.
- Vlieghe D.,
- Sandelin A.,
- De Bleser P.J.,
- et al.
- Mertin S.,
- McDowall S.G.,
- Harley V.R.
- Zhou Q.,
- Gedrich R.W.,
- Engel D.A.
- Ahn S.,
- Olive M.,
- Aggarwal S.,
- Krylov D.,
- Ginty D.D.,
- Vinson C.
- Markovitz J.H.,
- Raczynski J.M.,
- Wallace D.,
- Chettur V.,
- Chesney M.A.