Multi-Center Genetic Study of Hypertension
The Family Blood Pressure Program (FBPP)
The Family Blood Pressure Program (FBPP) consists of 4 independently established multicenter networks of investigators who have complementary approaches to the genetics of blood pressure levels and hypertension. The program has recruited participants from the African American, Mexican American, Asian, and non-Hispanic white populations. Each network utilized study designs, laboratory measurements, and analytic methods that made efficient use of the unique characteristics of their populations and the investigators’ expertise. The individual networks subsequently unified core study components into a single cohesive program. The unified FBPP includes (1) standardized clinic and laboratory protocols for core variables to facilitate direct comparison of results among networks, (2) coordination among laboratories to avoid unnecessary duplication of effort, (3) utilization of a single laboratory for genome-wide marker typing, and (4) a pooled data set containing phenotype and genotype information from >11 000 individuals.
Family studies have demonstrated a significant genetic component underlying interindividual variation in blood pressure levels and the occurrence of hypertension,1 but the identity of the individual genes remains uncertain. Blood pressure control involves a redundancy of traits with balancing pressor and depressor roles,2 and therefore, the impact of any one gene will likely be reduced as its effect is transmitted across intervening levels of a biologic hierarchy. This type of variation is difficult to elucidate, but advances in both laboratory and analytical approaches continue to facilitate progress. The Family Blood Pressure Program (FBPP) was established by the National Heart, Lung, and Blood Institute (NHLBI) to localize and identify genes influencing interindividual blood pressure variation and the occurrence of hypertension in the population-at-large.
Four separate “genetic determinants of high blood pressure” networks were established in 1995 under collaborative agreements with the NHLBI. Each network had multiple field sites collecting data on multiple racial or ethnic groups. In addition, each network included components for biochemical measurements, genotyping, data management, and statistical analyses. Each network had, and continues to have, unique areas of expertise and interests that are reflected in a diversity of study designs. All networks ascertained families having individuals with elevated blood pressure. Most networks ascertained families having individuals with a clinical diagnosis of hypertension, whereas others focused on younger individuals who were in the upper portion of the blood pressure distribution for their age and gender but would not be classified as having overt hypertension.
GenNet sought to address blood pressure as a continuously distributed quantitative phenotype. Non-Hispanic white subjects were recruited from Tecumseh, Mich, and African American subjects were recruited from Maywood, Ill. Probands were defined as individuals age 18 to 50 years with blood pressures in the upper 20% to 25% of the age-/gender-specific blood pressure distribution. Once the proband was identified, an attempt was made to enroll all siblings and parents of the proband, irrespective of their blood pressure or hypertension treatment status. Biochemical measurements included plasma renin activity, serum angiotensinogen concentration, ACE activity, red blood cell lithium-sodium countertransport, and urinary fractional lithium clearance. In addition, GenNet included an animal project in which inbred hypertensive rat strains were used to identify genetic regions to be compared with human orthologs identified from genetic linkage studies to target regions of the human genome for further study.
GENOA (Genetic Epidemiology Network of Atherosclerosis) includes 3 field centers: the field center in Jackson, Miss, recruited African Americans; the field center in Starr County, Tex, recruited Mexican Americans; and the field center in Rochester, Minn, recruited non-Hispanic whites. The field centers in Jackson, Miss, and Rochester, Minn, recruited sibships containing at least 2 individuals with hypertension diagnosed before age 60. Because of the high prevalence of non–insulin-dependent diabetes in the Mexican American population,3 and the resulting confounding that this would create, the field center in Starr County recruited sibships containing at least 2 individuals with adult onset diabetes. All available full biologic siblings of the index sibling pairs were invited to participate in interviews, physical examinations, and phlebotomy.
HyperGEN (Hypertension Genetic Epidemiology Network) has field centers in Birmingham Ala; Forsyth County, NC; Framingham, Mass; Minneapolis, Minn; and Salt Lake City, Utah. African American and non-Hispanic white hypertensive siblings and their parents (when available) and ≥1 untreated adult offspring of some of the siblings were recruited. Preference in ascertainment and recruitment was given to hypertensive sibships in which at least 1 of the subjects was classified as having severe hypertension. In addition to the core examination, stressed blood pressures and a variety of hypertension-related intermediate traits (eg, urinary aldosterone and catecholamines) were also measured.
SAPPHIRe (Stanford Asian Pacific Program in Hypertension and Insulin Resistance) focused its investigation on Asian Pacific populations of Chinese and Japanese origin residing in Taiwan, Hawaii, and California. Two types of sibling pairs were recruited: those concordant for hypertension (both siblings with hypertension) and those with 1 hypertensive individual and 1 individual with low blood pressure. In addition to the core examination, all SAPPHIRe participants underwent an oral glucose tolerance test.
The 4 networks confederated to form the FBPP to facilitate the identification of hypertension susceptibility genes. The program works through a steering committee and 4 subcommittees, in consultation with NHLBI staff and an observational study monitoring board (OSMB), to achieve its stated goals (see Figure and www.hypertensiongenetics.org). The FBPP steering committee serves to facilitate information exchange and to coordinate joint activities. The steering committee members include the directors of the 4 networks, the chairperson of each subcommittee, and 2 NHLBI project officers. The charge of the steering committee is to facilitate the establishment and conduct of multinetwork activities. It also establishes timelines and monitors progress of all joint program activities. The phenotype subcommittee coordinates measurement of core phenotypes by each of the 4 networks in a standardized fashion and disseminates information about the phenotypes collected by each network. It also oversees core laboratory functions and quality control. The DNA subcommittee coordinates activities of each of the 4 networks’ DNA laboratories and with the NHLBI Mammalian Genotyping Service to decrease the duplication of effort by the networks. In addition, it has a major responsibility for prioritizing positional candidate genes for further analysis. The data management and quality control subcommittee coordinates data management, quality control, and medication coding within the 4 networks and in the pooled data set. The charge of the analysis and publications subcommittee is to coordinate data analyses and evaluation of results, in addition to developing, evaluating, and adapting novel approaches to combined data analysis. It also tracks and coordinates manuscript preparation and publication.
Taken together, the program includes 13 field centers enrolling 5 racial groups (African American, Mexican American, Japanese, Chinese, and non-Hispanic white). A hallmark of the FBPP is a standardized clinical protocol that is used in all field centers to collect data on a set of core phenotypes for all study participants. A common set of 95 core phenotypes and standardized measurement methods was approved by each participating network and by the FBPP steering committee (Table 1). Because of the desire to combine information, it was imperative to use a common and reproducible blood pressure measuring protocol at each site (Criticon Inc). The FBPP has standardized the genotyping of all study participants using a common set of highly polymorphic marker loci evenly spaced throughout the genome (CHLC screening set, version 8.0), to facilitate combined genome-wide linkage analyses. To further ensure comparability of genotyping, FBPP study participants were all typed by the NHLBI-sponsored Mammalian Genotyping Service.
During the examination at all field centers, prescription medications taken by the participant during the previous month are recorded. Using a database of all antihypertensive medications available in North America, which includes the generic name, trade name, and a description of the pharmacologic action of each medication, the software assigns each prescription antihypertensive medication a 6-digit code number that corresponds to the first 6 digits of the Medi-Span (First DataBank Inc) generic product identifier that defines pharmaceutically equivalent drugs. These coded records can be readily manipulated to categorize antihypertensive medications into groups with similar modes of action.
The Pooled Dataset
The pooled data set currently contains >120 measured and derived phenotypic variables on 11 357 subjects. In addition, the pooled dataset contains extensive and ever-increasing genotype data on these individuals. Table 2 contains basic descriptive statistics for select variables for participants in the current pooled data set. Several points are worth highlighting. In all networks, more females than males participated in this study. The younger average age and lower prevalence of antihypertensive drug use of participants from the GenNet network is a result of its specific ascertainment scheme and interest in prehypertension. The lower average body mass index in the SAPPHIRe network is the result of their strong recruitment effort in Taiwan, where obesity is less common than in the United States.
Because of its size, the FBPP data set provides excellent statistical power to detect small genetic effects, and permits in-depth analyses within various subsets of study participants, which may help to detect effects that are confounded by nongenetic factors (see below). There are myriad ways in which to present a description of the family structure of the FBPP pooled data set. Because of its common application and efficiency, we have opted to show the structure of sibships participating in the FBPP (Table 3). The sibships described in Table 3 combine to form >11 524 sibling pairs for genetic linkage analyses. Because this resource does contain data that was collected under varying study designs, careful attention must be given to the selection of questions that can be validly asked in the pooled data set and of valid methods of analysis. In addition, combining data from multiple samples may also introduce additional heterogeneity. However, these caveats are far outweighed by the great potential of the resource. The sibships in the FBPP are particularly well suited for affected sibling pair linkage analysis, such as that implemented in GENEHUNTER4; quantitative trait linkage analysis, such as that implemented in SOLAR5; and the transmission disequilibrium test, such as that implemented in FBAT.6
Presenting the statistical power of the FBPP data resource, in general terms, is impossible because such power is dependent on the exact hypothesis being addressed, details of the method being used to test the hypothesis, and other considerations. Because of its simplicity and topical nature, we have calculated the power of the FBPP pooled data resource for an affected sibling pair linkage analysis using the linkage method of Suarez,7 which employs a t test on allele sharing. The results are shown in Table 4. We consider a sample size of 1500 pairs because that is possible within several of the separate racial groups. For each level of population prevalence of hypertension (Kp) and for each genetic model (combination of phenocopies and penetrance), the sample yields >80% power to detect the effects of genes having modest to moderate effects. Although not shown here, power calculations for localizing genes influencing quantitative trait variation using genetic linkage analysis yielded similar levels of power. Importantly, the power to detect genes using association methods was high and bodes well for this method of analysis as the genome single nucleotide polymorphism (SNP) map and candidate gene SNPs become more available.8
One of the most powerful and important uses of the FBPP pooled data resource is the localization of hypertension susceptibility genes by combining evidence across all 4 networks. Although the study of each network was originally designed to achieve a level of power sufficient to detect genes above some minimal threshold of effect, only by combining evidence across networks can the program achieve the power necessary to localize genes of smaller effect. One of the simplest methods for combining results is the method of combining probability values.9 It is based on the observation that if a number (n) of independent tests are made of the same hypothesis resulting in n probability values, then the quantity
is asymptotically distributed as a χ2 with 2n degrees of freedom, which provides a combined probability value for all n tests. In the case of linkage analysis, one can easily work on either the lod score or the probability scale, because there is a simple 1-to-1 correspondence between the two. The n individual tests need not use the same statistic to produce the probability values, and they may each even operate on very different sampling units (eg, pedigrees versus sibling pairs). More sophisticated methods of meta-analysis are also possible,10 but the simplicity and elegance of Fisher’s method makes it an attractive screening tool to search for the most promising regions identified by the combined FBPP resource.
Special Opportunities and Analyses
It is well known that hypertension and its complications are the consequence of interactions among many genetic and environmental factors. Most studies of the genetics of blood pressure levels or hypertension status admit to the complex nature of the trait but fail to take context-dependent effects into account. The linkage and association studies in the FBPP will be performed within each racial/population subgroup and then will also be performed within well-defined subgroups of each sample to permit direct analysis of context-dependent genetic effects of positional candidate genes. Examples of contexts within racial group that are examined include, but are not limited to age, gender, measures of body size, and measures of family history of hypertension and co-morbidities. These in-depth studies are only possible because of the large sample size available in the program.
A fundamental decision in genetic hypertension research is whether to use actual blood pressure as the phenotype or to use the qualitative diagnostic categories of hypertension or normotension. As a practical matter, most individuals with hypertension will be treated with antihypertensive medications by the time they are examined and are considered as a single diagnostic class by the program. However, because many of the features of the physiology and biochemistry of hypertension may be most profitably studied in young prehypertensives, blood pressure levels may be the most relevant phenotype.11 The FBPP has the ability to analyze both blood pressure levels (in young individuals in GenNet, GENOA, and HyperGEN) and hypertension as a diagnostic category (all 4 networks).
One of the challenges of localizing and characterizing genes contributing to hypertension is to dissect this heterogeneous phenotype into subtypes that are sufficiently homogeneous and large enough to lend themselves to genetic analyses. Several approaches can be used to partition the study population into more homogeneous subgroups for analysis of blood pressure levels and hypertension status. These approaches include categorization based on (1) measures of metabolic intermediates (eg, plasma insulin), (2) co-morbidity with diseases of possible common etiology (eg, obesity), (3) degree of specific target organ damage (eg, left ventricular hypertrophy), and (4) differential responses to environmental input (eg, smoking). The large number of individuals phenotyped in the FBPP and the diverse measures employed make such strategies feasible.
Cooperation and coordination of the 4 networks making up the FBPP give rise to valuable opportunities for cost and information sharing not otherwise possible. One of the major opportunities arises from an agreement to distribute the workload of identifying SNP variation in and around candidate genes and to share this information immediately with the other networks. There are 3 types of candidate genes that are the focus of the FBPP. First are biologic candidate genes identified through their presumed role in blood pressure regulation. Some of these biologic candidate genes come to the attention of the FBPP through the work of others, and the FBPP offers the opportunity to perform coordinated analyses to examine the effects of the genes in a variety of ethnic groups and environmental contexts. The FBPP has already published such coordinated biologic candidate gene analyses as simultaneous papers from the 4 networks,12–15⇓⇓⇓ and as a single paper.16 Second, differences in gene expression using genomic arrays17 also provide candidate genes (ie, expressional candidate genes) in which DNA sequence variation may be influencing inherited susceptibility to elevated blood pressure. The third type of candidate gene, and the one the FBPP is poised to take full advantage of, is genes targeted because of on their location within regions identified through genetic linkage analyses (ie, positional candidate genes).
Linkage results from the FBPP genome scans are also a significant component of the FBPP’s decision-making algorithm for prioritization of future genotyping. The overall goal is to take into account and weigh the available information, including knowledge about physiological function of positional candidate genes, the strength and consistency of the homologous linkage evidence in humans, and the strength and consistency of the linkage evidence in rats. The underlying premise is that the more compatible the various sources of information are, the more likely it is that a signal is true (ie, there exists a hypertension susceptibility gene having a measurable effect within the linked region).
The process of identifying susceptibility genes, although daunting only 1 or 2 years ago, is greatly facilitated by the burgeoning human DNA sequence data and sophisticated bioinformatic tools (eg, www.ncbi.nlm.nih.gov). Importantly, successes are beginning to be realized in hypertension18 and other conditions.19
The FBPP has placed an emphasis on animal studies to facilitate the identification of genes responsible for hypertension. The rat provides the ideal physiological platform with its wealth of physiological and biochemical literature and the recent expansion of genomic resources.20 The FBPP Animal Studies Center performs the following tasks: (1) translation of genetic information between rat and human, (2) identification of positional candidate genes within linked regions, and (3) testing of positional candidate genes for differences in gene expression between rat lines, accompanying changes in blood pressure within a line, and after physiological challenges. Linked regions overlapping in orthologous genomic regions between the rat and human provide an unprecedented opportunity to relate the wealth of physiological data in the rat to humans.
The FBPP investigators are cognizant of the importance of collaboration and interdisciplinary research, and the FBPP embodies an important intellectual and tangible scientific resource for blood pressure and cardiovascular disease research. As such, the FBPP investigators welcome the opportunity to extend their collaborative sphere to non-FBPP investigators to maximize the potential of the FBPP resource. Available genomic DNA and stored plasma collected from FBPP participants, however, is limited and resources are not currently available to systematically create and maintain lymphoblast cell lines by transformation. In addition, only small quantities of plasma and serum were collected and aliquoted for measurement of future analytes. Therefore, transfer of genomic DNA or plasma from FBPP laboratories to potential collaborators is not encouraged.
Non-FBPP investigators wishing to establish a collaboration with the FBPP are encouraged to first visit the program’s Web site (www.hypertensiongenetics.org) to identify a FBPP investigator and network with whom to initially collaborate. This investigator will serve as a liaison to the program. The 4 participating networks differ in focus and available populations, so there are a variety of opportunities available for collaboration. It is recommended that the contacted FBPP investigator be a network director. The identified FBPP investigator should be directly contacted to finalize details of the proposed collaboration. The nature of this collaboration should be the same as any other professional relationship, except its content will be shared within the FBPP. The FBPP already has in place a mechanism by which the findings from 1 particular network are brought to the program for verification and further exploration. This mechanism, which includes guidelines for data sharing and publication, will also operate with ideas and results emanating from collaboration between 1 of the networks and a non-FBPP investigator.
Collaboration with the FBPP permits non-FBPP investigators to take advantage of the very large sample and data resource not possibly amassed by any single investigator. Collaboration of the FBPP with non-FBPP investigators broadens the area of its investigation by incorporating information from other ongoing studies. Together, collaboration of the FBPP with non-FBPP investigators will further our understanding of the genetics of hypertension and its complications.
Identifying functional variation in genes contributing to interindividual differences in blood pressure levels and hypertension status will suggest improved interventions to reduce the risk and improve treatment of hypertension in several ways: (1) more accurate and earlier identification of those individuals at increased risk to develop hypertension; (2) improved understanding of the etiology and pathophysiology of hypertension; (3) development of new, more efficacious treatments, and tailoring of particular treatments to patients who are most likely to respond; and (4) a better understanding of the factors contributing to differences in the prevalence of hypertension among populations.
The FBPP Investigators include the following individuals:
GENOA:Eric Boerwinkle (Director), C. Andrew Brown, Maurita Carrejo, Robert Ferrell, Craig Hanis, Richard Hutchinson, Sharon Kardia, Charles Sing, and Stephen Turner.
GenNet: Alan Weder (Director), Aravinda Chakravarti, Richard Cooper, Howard Jacob, and Nicholas Schork.
HyperGEN: Steven Hunt (Director), Donna Arnett, Ingrid Borecki, John Eckfeldt, R. Curtis Ellison, Chi Gu, Gerardo Heiss, Mark Leppert, Albert Oberman, Michael Province, and D.C. Rao.
SAPPHIRe: David Cox and J. David Curb (Directors), Ida Chen, John Grove, Kamal Masaki, Tom Quertermous, Koustubh Ranade, Neil Risch, and Beatriz Rodriguez.
National Heart, Lung, and Blood Institute: Stephen Mockrin, Susan Old, and Peter Savage.
The entire Family Blood Pressure Program is supported by a series of cooperative agreements with the National Heart, Lung, and Blood Institute.
- Received May 10, 2001.
- Revision received May 28, 2001.
- Accepted July 26, 2001.
Ward R. Familial aggregation and genetic epidemiology of hypertension.In: Laragh JH, Brenner BM, eds. Hypertension: Pathophysiology, Diagnosis, and Management. 2nd ed, vol 1. New York, NY: Raven Press; 1995: 67–88.
Guyton A. Abnormal renal function and autoregulation in essential hypertension. Hypertension. 1991; 18 (supplIII): III-49–III-53.
Hanis C, Ferrell R, Barton S, Aguilar L, Garza-Ibarra A, Tulloch B, Garcia C, Schull W. Diabetes among Mexican-Americans in Starr County, Texas. Am J Epidemiol. 1983; 118: 659–672.
Collins F, Guyer M, Charkravarti A. Variations on a theme: cataloging human DNA sequence variation. Science. 1997; 278: 1580–1581.
Fisher R. Statistical Method for Research Workers. 13 ed. London, UK: Oliver and Lloyd Ltd; 1925.
Weder A, Schork N. Adaptation, allometry and hypertension. Hypertension. 1994; 24: 145–156.
Bray M, Krushkal J, Li L, Ferrell R, Kardia S, Sing C, Turner S, Boerwinkle E. Positional genomic analysis identifies the β2-adrenergic receptor gene as a susceptibility locus for human hypertension. Circulation. 2000; 101: 2877–2882.
Horikawa Y, Oda N, Cox N, Li X, Orho-Melander M, Hara M, Hinokio Y, Linder T, Mashima H, Schwarz P, del Bosque-Plata L, Horikawa Y, Oda Y, Yoshiuchi I, Colilla S, Polonsky KS, Wei S, Concannon P, Iwasaki N, Schulze J, Baier LJ, Bogardus C, Groop L, Boerwinkle E, Hanis CL, Bell GI. Genetic variation in the calpain 10 gene (CAPN10) is associated with type 2 diabetes mellitus. Nat Genet. 2000; 26: 163–175.
Stroll M, Kwitek-Black A, Cowley A, Harris E, Harrap S, Krieger J, Printz M, Provoost A, Sassard J, Jacob H. New target regions for human hypertension via comparative genomics. Genome Res. 2000; 10: 473–482.