Genes and environmental factors contribute to an individual's risk of hypertension. Recent advances in DNA sequencing technology have enabled the discovery of new causative genes in inherited forms of hypertension, identifying novel pathways for blood pressure control. Meta-analyses of genome-wide association studies have also identified regions of the genome that are significantly associated with blood pressure control, and these regions may be involved in an individual's response to antihypertensive medication. This article reviews the latest gene discoveries in inherited forms of hypertension, recent meta-analyses of genome-wide association studies, genetic determinants of antihypertensive therapy response, and development of genetic risk scores.
Statistics from Altmetric.com
Hypertension is a major contributor to worldwide morbidity and mortality, with approximately one in three individuals affected by the condition.1 The prevalence increases with age, as does the vulnerability to complications. Risk factors for the development of hypertension include high body mass index (BMI), high dietary sodium intake, low levels of physical activity, alcohol consumption and a poor diet.2 The heritability of hypertension has been the focus of many studies over the past 60 years with early work estimating that approximately 30% of blood pressure variation is due to genetics.3
Genetic linkage studies in families have identified a number of genes that are responsible for rarer Mendelian forms of hypertension.4 The genes which are known to be mutated in these conditions, have provided good candidates for essential hypertension and have helped to determine some of the molecular pathways involved in blood pressure control.5 These candidate genes have been used in association studies of hypertension,6 however, the results have been largely inconsistent due to genetic heterogeneity and differences in environmental factors that contribute to the disease.7 Evidence suggests that the majority of cases of hypertension are under the control of a large number of genes, each of which contributes a small effect, complicating efforts to determine the underlying genetic components of the disease.8
This review will describe more recent efforts of gene discovery in essential hypertension including the application of the latest DNA sequencing technologies to revisit Mendelian forms of hypertension, meta-analyses of genome-wide association studies (GWAS), and the genetic basis of antihypertensive response.
The next generation of hypertension genetics—Mendelian diseases revisited
A number of forms of hypertension are now known to be caused by single gene defects. These conditions include pseudohypoaldosteronism, glucocorticoid remediable aldosteronism, Liddle syndrome, apparent mineralocorticoid excess, glucocorticoid receptor deficiency, autosomal dominant hypertension with brachydactyly and pheochromocytoma syndromes.9 Early genetic work on these conditions involved linkage analysis of large multigenerational pedigrees, with many affected individuals. This enabled researchers to identify the region of the genome that harboured the disease locus, followed by candidate gene screening within this region to identify the disease-causing mutation. The DNA sequencing technology available at the time of the original gene discoveries in these conditions was based on Sanger biochemistry. This technology was costly, and the process of gene discovery was time consuming, with no guarantee of identifying the gene involved in disease pathogenesis.
While there are still many benefits of using traditional linkage studies for disease gene discovery, suitable families are often not available. Over the past decade, new strategies for DNA sequencing, including Next Generation Sequencing, have been developed10 which have reduced the cost of DNA sequencing by several orders of magnitude (figure 1), and decreased the amount of time required to sequence the human genome, thus making large-scale genome sequencing a plausible option for disease gene discovery.
The exome, which includes only the protein-coding regions, constitutes approximately 1% of the human genome, and is predicted to contain 85% of disease-causing mutations.11 Exome sequencing, using new sequencing technology, is a cost-effective and robust alternative to traditional methods of disease gene discovery,12 ,13 and has recently been applied to Mendelian forms of hypertension, namely pseudohypoaldosteronism type 2 (PHAII), and familial hyperaldosteronism.
Pseudohypoaldosteronism type 2
Candidate genomic regions for genes involved in PHAII were first identified by Mansfield et al in 1997,14 and subsequently narrowed to the WNK lysine-deficient protein kinases by Wilson et al in 2001.5 WNK1 and WNK4 encode proteins involved in electrolyte homeostasis, and mutations within these genes result in reduced potassium excretion from the kidney, despite normal renal glomerular filtration. These mutations however, only account for a small fraction of families with PHAII, leaving a large proportion of cases where the genetic basis of disease is unknown.
A landmark study by Boyden et al15 in 2012 studied 52 families with PHAII, of which seven families had mutations in the WNK kinases. Using exome sequencing, the authors identified a further two genes, Kelch-like 3 (KLHL3) and Cullin 3 (CUL3) responsible for PHAII in 41 of the remaining 45 families. The authors showed that both dominant and recessive mutations within KLHL3 cause PHAII, contrasting with mutations in WNK kinases and CUL3, which follow a dominant transmission pattern. An independent study by Louis-Dit-Picard et al16 also used whole exome sequencing combined with linkage analysis to identify KLHL3 mutations in two families with PHAII. They then used direct DNA sequencing to show that a further 14 out of 43 index cases had mutations within KLHL3. Again, these mutations were either dominant or recessive, with recessive cases having an earlier age of onset, and a more severe phenotype, than individuals with dominant mutations. Both KLHL3 and CUL3 express proteins that are members of a complex signalling pathway that regulates ion homeostasis in the distal nephron.16
Individuals with primary aldosteronism constitutively produce aldosterone from the adrenal gland, resulting in hypertension with variable hypokalaemia. This suite of conditions accounts for 8–13% of individuals with hypertension which represents approximately 2% of the general population.17 The most common cause of primary hyperaldosteronism is bilateral hyperplasia, followed by aldosterone-producing adrenal adenomas (APAs), while around 5% of individuals have Mendelian forms of primary aldosteronism.18 A study by Choi et al19 performed whole exome sequencing in tumour tissue and blood pairs from four patients with APA, and identified a small number of somatic mutations in each tumour, of which the KCNJ5 gene was mutated in two of the four tumours. They then showed that mutations within KCNJ5 were present in a further six out of 18 tumours studied, and hypothesized that mutations within this gene could be responsible for Mendelian forms of primary aldosteronism. This was shown to be the case in a family they had previously identified with severe hypertension, aldosteronism and massive adrenal hyperplasia.20
Essential hypertension informed by Mendelian hypertension
Screening individuals with essential hypertension for a panel of genes known to cause familial hypertension, such as those identified above, may uncover a larger number of families with inherited forms of hypertension than previously thought. In addition, exome sequencing of large cohorts of individuals with essential hypertension may further elucidate the genetics of blood pressure control. There are potentially many rare coding variants that alter the function of genes involved in blood pressure control, and these may form the basis of hypertension heritability in families. The challenge is to identify genes that commonly harbour these rare variants in individuals with essential hypertension, and develop targeted therapies to these forms of the disease.
Genome-wide association studies
Hypertension is a complex disease and is known to be influenced by environmental factors, making the study of hypertension genetics challenging. GWAS have been extensively used to identify genes associated with blood pressure control and the pathogenesis of hypertension.21–23 Many GWAS studies have failed to attain genome-wide significance levels (p<5×10−8), and in the past decade, there are still relatively few variants within the genome that are significantly associated with hypertension (figure 2). In addition, genes known to alter blood pressure, such as current drug targets and causative genes in rare Mendelian forms of hypertension, often fail to produce strong signals in the genome-wide scans.23
More recent GWAS have increased coverage across the genome and included larger numbers of individuals, from multiple studies, in order to identify genes involved in blood pressure control and hypertension. The Global BPgen consortium conducted a meta-analysis of 17 cohorts of European ancestry, totalling 34 433 individuals, and reported eight loci that were significantly associated with blood pressure.21 The Cohorts for Heart and Ageing Research in Genome Epidemiology (CHARGE) consortium analysed data from six population-based cohort studies which included 29 136 participants.7 This study identified 13 loci associated with systolic blood pressure, 20 associated with diastolic blood pressure and 10 associated with hypertension, using a significance level of p<4×10−7.
Genetic variation within drug target genes have also been explored within the CHARGE cohorts for their association with blood pressure control and hypertension. A study by Johnson et al24 examined the association between variants in 30 genes that were known antihypertensive drug targets. Variants within two genes, namely rs1801253 in the β-adrenergic receptor (ADRB1), and rs11122587 within angiotensinogen, were both associated with systolic blood pressure, diastolic blood pressure, and hypertension. This was the first study in which these variants had reached genome-wide significance in a meta-analysis.
A meta-analysis, combining data from Global BPgen and CHARGE found one single nucleotide polymorphism (SNP) that was significantly associated with hypertension (rs2681472).21 This SNP is an intonic variant within the ATP2B1 gene, which encodes a Ca2+ ATPase enzyme that removes Ca2+ from cells against a large concentration gradient, and is therefore important in intracellular Ca2+ homeostasis.25 This meta-analysis also identified four SNPs associated with systolic blood pressure and six SNPs associated with diastolic blood pressure. Both the BPgen and CHARGE studies analysed individuals of European ancestry.
A larger meta-analysis has since been published by the International Consortium for Blood Pressure Genome-Wide Association Studies, analysing data from 200 000 individuals of European descent.26 This study identified 29 independent SNPs at 28 loci, that were significantly associated with systolic blood pressure, diastolic blood pressure or both, of which 16 were novel. Many of these loci are plausible candidates for blood pressure control, including the NPPA and NPPB genes that encode for the natriuretic peptides (A and B respectively). These peptides are known to have blood pressure lowering properties and have previously been associated with variation between individuals with regard to blood pressure and hypertension.27 The genes identified in these studies are summarised in table 1.
All the loci identified by these three large meta-analyses of GWAS data have provided new candidates for the molecules that control blood pressure, as well as the possibility of exploring new pathways for therapeutic intervention.
Replication studies have since been carried out to validate the loci associated with blood pressure control in Europeans, and in individuals of both European and non-European ancestry. For instance, studies have been carried out in a Japanese population which replicated the findings for four of the blood pressure susceptibility loci in Europeans (ATP2B1, FGF5, CYP17A1 and CSK).28 Three of the European loci (CYP17A1, CACNB2 and PLEKHA7) were also replicated in a She Chinese population29; four loci (ATP2B1, CSK, CYP17A1 and PLEKHA7) were replicated in a Korean population30; and three loci (SH2B3, TBX3-TBX5 and CSK) were replicated in an African–American population.31 A meta-analysis of individuals from East Asia replicated seven of the blood pressure susceptibility loci identified in Europeans (CASZ1, MTHFR, FGF5, CYP17A1, ATP2B1, CSK and MTHFR) in addition to finding novel loci associated with blood pressure control (ST7L-CAPZA1, FIGN-GRB14, ENPEP, NPR3 and a variant near TBX3).32 In addition, a replication study was carried out in the Women's Genome Health study, a homogenous population of European ancestry which replicated 13 out of 18 SNPs that had previously reached genome-wide significance for an association with blood pressure.33
Combining GWAS and exome sequencing
Albrechtsen et al,34 analysed a Danish population by performing whole exome sequencing in 1000 cases (with type 2 diabetes, BMI>27.5 kg/m2 and hypertension) and 1000 controls, followed-up by SNP genotyping in a further 15 989 Danes. Results were then replicated for 45 SNPs in 63 896 Europeans. This study aimed to identify novel associations of coding variants (>1% minor allele frequency) with metabolic phenotypes, including hypertension. Three associations were reported, namely an amino acid polymorphism in CD300LG was associated with fasting HDL-cholesterol, while common variants within COBLL1 and MACF1 were associated with type 2 diabetes. No SNPs were found to be associated with hypertension, systolic blood pressure or diastolic blood pressure in this study.
Exome sequencing differs from GWAS as it only detects variants within the coding regions of the genome. It therefore has the ability to detect rare variants that would not be genotyped in a GWAS study. One drawback, however, is that it will not detect SNPs located outside the exons that are associated with a trait. Due to the infancy of the technology, there are still a number of challenges associated with the application of exome sequencing to complex traits, as in the study by Albrechtsen et al, including how to account for epigenetic changes in blood pressure control.
Epigenetics is an emerging field in hypertension research which refers to the gene regulatory information that can be transmitted through cell divisions without changing the underlying DNA sequence. Examples of epigenetics include DNA methylation, histone modifications, changes in chromatin structure and the action of small non-coding RNA. There is evidence that there is less DNA methylation in patients with essential hypertension, and the more advanced the stage of hypertension, the less the amount of methylation.35 Histone modification and miRNAs have also been implicated in blood pressure control, largely from animal and cellular studies (reviewed in refs 36 ,37). Further research is required to understand how epigenetic biomarkers alter blood pressure and lead to hypertension. Understanding the epigenetics of hypertension will allow for novel intervention strategies. Clinicians may be able to differentiate the types of hypertension (salt-sensitive, neurogenic and others) based on an individual's epigenome, allowing for a more targeted treatment approach.
Translating genetics into therapy
The aim of antihypertensive medication is to reduce a patient's blood pressure to an optimal range of less than 140 mm/Hg systolic and less than 90 mm/Hg diastolic,38 with a lower target range in patients with diabetes mellitus, chronic kidney disease or heart conditions.39 ,40 Many of the antihypertensive drugs available today inhibit components of the renin-angiotensin aldosterone system (RAAS) which is important in the control of blood volume and systemic vascular resistance. The classes of blood pressure medications used as a first-line therapy, in combination with lifestyle modifications, include β-blockers, ACE inhibitors, diuretics, angiotensin II receptor blockers and calcium channel blockers. A large number of patients, however, will not reach an optimal blood pressure range by using a single drug class, and will require a combination of these therapies to reach an optimal blood pressure range (figure 3).
The decision on which class of antihypertensive to prescribe must take into account the patients age, comorbidities and interactions with other prescribed medications.41 This is ultimately a personalised decision, informed by the characteristics of the individual patient and supported by evidence generated in large clinical trials. This process still involves trial and error in order to achieve optimal blood pressure due to differences between individuals with regard to their response to treatment. Individuals receiving treatment still have barriers to reaching blood pressure targets, with approximately 40–50% of patients remaining without adequate control.42 Prescription decisions informed by an individual's genetics (pharmacogenomics), is one potential approach that would offer a personalised approach to therapy.
Pharmacogenetic studies of antihypertensive medications have been carried out over the past 15 years, with the majority of research focusing on single candidate genes, to determine if SNPs within those genes alter an individual's response to medication.43 ,44 Candidate genes have been sourced from biological pathways known to be involved in blood pressure control, namely, components of the RAAS,45 natriuretic peptides,46 β-adrenergic receptors,47 and renal sodium retention.48 Not all studies show significant associations between genetic variants and response to antihypertensive medications. More recent studies have focused on SNPs identified in GWAS of blood pressure control, to test whether these same variants are associated with blood pressure response to antihypertensive therapy. In addition, these studies have hypothesized that a combination of alleles from various loci would be a better predictor of blood pressure response to therapy than individual alleles, as described below.
Genetic risk scores
Individual variants identified in the large GWAS explain only a very small proportion of variation within systolic and diastolic blood pressure, to the order of 1 mm Hg per allele for systolic blood pressure and 0.5 mm Hg per allele for diastolic blood pressure.7 ,21 These results are consistent with the hypothesis that common variants associated with blood pressure phenotypes have a very small effect, and it is likely that a combination of many genomic variants, each which contribute a small effect, is responsible for blood pressure control. As such, the contribution of an individual's genetics to assess their risk of high blood pressure could be determined by the combined effect of multiple risk alleles, in addition to traditional risk factors, such as age and BMI.
A study by the International Consortium for Blood Pressure GWAS showed there was an association between a genetic risk score, calculated by weighting the average of systolic and diastolic effects for 29 SNPs, identified to contribute to blood pressure control, with the odds of hypertension. Individuals with one SD above the mean genetic risk score were reported to have a 21% increase in the odds of hypertension.26
A more recent study by Fava et al49 used these same 29 SNPs to confirm the utility of a genetic risk score in the field of hypertension. While the study showed that the score was independently associated with hypertension incidence and blood pressure change over time, there was no improvement to risk prediction when including the genetic risk score with the non-genetic traditional risk factors. Obesity and prehypertension status were shown to have a stronger association with hypertension incidence than the genetic risk score. The authors did not rule out the possibility of one day including a genetic risk score in the clinical setting if different SNPs, rare genetic variants, and the interactions of these SNPs with other genes and the environment become clearer.
The application of a genetic risk score to determine an individual's response to antihypertensive therapy is also under investigation. The eight SNPs associated with systolic or diastolic blood pressure in the Global BPgen study21 were genotyped in Swedish participants of the Nordic Diltiazem (NORDIL) antihypertensive study, to test for an association between these variants with response to antihypertensive therapy.50 Of the eight SNPs, two showed nominal evidence of an association with blood pressure control. SNP rs12946454 was associated with a reduction of both systolic and diastolic blood pressure in the calcium channel blocker (diltiazem) group, while SNP rs11191548 was associated with reduction in diastolic blood pressure in the β-blockers/diuretics group. Neither of these associations was significant after p value adjustment for multiple testing. A genetic risk score was also calculated based on all eight SNPs, and no association was found with blood pressure response in any of the antihypertensive groups.
The Pharmacogenomic Evaluation of Antihypertensive Responses (PEAR) study, a randomised multicentre clinical trial, tested the association between 37 SNPs identified via GWAS and meta-analyses in Caucasians, with response to a β-blocker (atenolol) and a thiazide diuretic (hydrochlorothiazide).51 This study found that no SNP alone reached a priori significance, however, SNP rs1458038 near FGF5 reached nominal significance (p<0.05) and was associated with a better response to both antihypertensive therapies, with genotype effects in opposite directions. The authors then tested a genetic risk score (1–6), based on four blood pressure lowering alleles in response to atenolol monotherapy, and found a significant association between the genetic score and blood pressure response to therapy. That is, the greater the number of blood pressure lowering alleles, the greater the blood pressure response to atenolol, with a 4.0 mm Hg reduction in systolic blood pressure in individuals with a score of 1, compared with a 18.6 mm Hg reduction in systolic blood pressure in individuals with a score of 6 (in Caucasians). Similarly, they constructed a genetic risk score (1–6) based on three blood pressure lowering alleles in response to hydrochlorothiazide monotherapy, and found a significant association between the genetic risk score and response to therapy. The authors note that the loci included in these genetic risk scores have a moderate contribution to blood pressure response to antihypertensive therapy.
Due to the variability in results from pharmacogenomic studies of hypertension, further research with large cohorts is required to understand how an individual's genetic make-up influences their response to hypertensive therapy. The utility of a genetic risk score to inform antihypertensive prescribing is still in its infancy. Further discovery of SNPs that alter an individual's response to therapy are required, before pharmocgenomics could be applied to hypertension in a clinical setting.
A great effort, over many decades, has been placed on deciphering the genetic basis of blood pressure. While there is strong evidence that there is a heritable component to blood pressure control, our current knowledge can only explain a fraction of this heritability in the majority of individuals with hypertension.
The advances in DNA sequencing technology and development of cost-effective whole exome sequencing have already discovered genes not previously known to be involved in blood pressure control. These technologies are likely to uncover more variants involved in Mendelian forms of hypertension, such as rare variants clustered within the same gene of multiple affected individuals. Discovery of genes harbouring rare variants is likely to elucidate novel pathways involved in blood pressure control. We predict that inherited forms of hypertension may be more common than previously predicted.
A genetic risk score, based on a number of variants contributing a small effect on blood pressure, holds the greatest potential for translating hypertension genetics into therapy, although more discoveries are required before a genetic risk score can be used in the clinical setting. This holds true when determining an individual's risk of hypertension, as well as their response to antihypertensive therapy.
A number of novel pathways have now been implicated in blood pressure control, as a result of significant SNP associations in GWAS. These include the natriuretic peptides, genes involved in calcium homeostasis, transcription factors and fibroblast growth factor, to name just a few. This knowledge, combined with increasing evidence of the role of epigenetic regulation of genes involved in blood pressure homeostasis, will consume research over the coming years. The ultimate aim of which is to identify new pathways for pharmacological intervention and use genetic testing in the clinical setting.
JML is supported by an Australian National Health and Medical Research Council Fellowship.
Contributors This work is an invited review. Both JML and CLC contributed to the literature research, article writing, and preparation of tables and figures. JML is the guarantor of this work.
Competing interests None.
Provenance and peer review All commissioned and peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.