OPEN A genomic deep ﬁ eld view of hypertension

Blood pressure is regulated by a complex neurohumoral system including the renin-angiotensin-aldosterone system, natriuretic peptides, endothelial pathways, the sympathetic nervous system, and the immune system. This review charts the evolution of our understanding of the genomic basis of hypertension at increasing resolution over the last 5 decades from monogenic causes to polygenic associations, spanning w 30 monogenic rare variants and > 1500 single nucleotide variants. Unexpected early wins from blood pressure genomics include deepening of our understanding of the complex causation of hypertension; re ﬁ nement of causal estimates bidirectionally between blood pressure, risk factors, and outcomes through Mendelian randomization; risk strati ﬁ cation using polygenic risk scores; and opportunities for precision medicine and drug repurposing.

The development of genetics has profoundly affected the field of hypertension, the most important global risk factor for cardiovascular disease.In their illuminating review, Garimella, Padmanabhan, and colleagues provide an overview of the genetic architecture of blood pressure control, ranging from ultrarare highimpact variants to common polymorphisms and their use for pharmacogenomics and risk stratification through polygenic risk scores.They describe how these insights improved our understanding of fundamental processes involved in NaCl handling, its hormonal regulation, and the vascular counterpart of blood pressure control.These advances are balanced by the need to bridge the gap between genome-wide association study variants and actionable practice; to better understand the role of structural variants, epigenetics, and gene-environment interactions; and to address the uneven access to clinical genomic or genetic services.
nervous system, and the immune system. 7Perturbations in any component of this system can arise from behavioral, environmental, or genetic factors or a combination of all, leading to increases or decreases in mean BP.
This review charts the evolution of our understanding of the genomic basis of hypertension at increasing resolution over the last 5 decades.The definitions of key terms are presented in Box 1.0][11][12][13][14] Currently, the genetic architecture of hypertension encompasses w30 monogenic rare variants and >1500 single nucleotide variants comprising both lowfrequency and common variants reflecting the continuous spectrum of risk alleles influencing BP and consequently hypertension.Although monogenic variants implicate causal genes illuminating underpinning mechanisms, the wealth of common variants contrasts with the paucity of success in connecting to causal pathways.The immediate and direct application of GWAS results in clinical practice or translational studies looks unfeasible, and nascent functional studies to follow up GWAS signals require years to generate actionable results.Notwithstanding these, there have been unexpected early wins including expansion of our understanding of the complex causation of hypertension; confirmation and/or refinement of causal estimates bidirectionally between BP, risk factors, and outcomes through Mendelian randomization; understanding opportunities and requirements for risk prediction using polygenic risk scores; and opportunities for precision medicine and drug repurposing.

Sodium pathways
A direct relationship between excess Na þ intake and hypertension has been well recognized for a long time and further confirmed by epidemiological studies and clinical trials. 15his implied that perturbations in physiological pathways that maintain Na þ homeostasis may be the underlying cause of hypertension.A majority of the early monogenic BP syndromes involved mutations in genes in the kidney tubules and adrenal glands with roles in tubular Na þ transport mechanisms, suggesting the importance of Na þ in BP regulation.GWASs have also identified single-nucleotide polymorphisms (SNPs) near genes involved in Na þ pathways with varying levels of functional validation (Figure 1).

Box 1 | Definition of key terms
Linkage analysis.A powerful tool to detect the chromosomal location of disease genes.It is based on the observation that genes that reside physically close on a chromosome remain linked during meiosis.It attempts to locate a disease-causing gene by identifying genetic markers of known chromosomal location that are co-inherited with the trait of interest.It requires a well-defined trait (phenotype), an extensive pedigree of families usually with multiple generations, and genetic markers and maps.Genome-wide association studies (GWASs).This involves testing hundreds of thousands of common genetic variations across the DNA of large numbers of individuals to find those statistically associated with a specific trait or disease.GWAS results have a range of applications, such as gaining insight into a phenotype's underlying biology, estimating its heritability, calculating genetic correlations, making clinical risk predictions, informing drug development, and inferring potential causal relationships between risk factors and health outcomes.Single-nucleotide polymorphisms (SNPs).They are the most common type of naturally occurring genetic variation among people.Each SNP represents a single base substitution in the human genome with a population frequency of >1%.SNPs occur approximately once in every 1000 nucleotides throughout the genome, which means each individual's genome contains w4 to 5 million SNPs.Phenome-wide association studies (PheWAS).This is an unbiased approach to test for associations between a specific single-nucleotide polymorphism (SNP) or a combination of genetic variants across a wide range of phenotypes in large populations.The direction of inference in a PheWAS is from a SNP to multiple phenotypes, whereas in genome-wide association studies it is from one phenotype to multiple SNPs.They are well-suited to facilitate the identification of new associations between SNPs and phenotypes as well as between SNPs and pleiotropy.PheWAS have been proposed to enhance drug development through elucidating mechanisms of action, identifying alternative indications, predicting adverse drug events, and opportunities for drug repurposing.Pleiotropy.A phenomenon whereby a single genetic variant influences $2 apparently unrelated phenotypic traits via independent biological pathways, for instance, because of the effects in different tissues or because the effect of the variant on one trait is causally related to variation in another trait.Polygenic risk score (PRS).A single value estimate of an individual's genetic susceptibility to a phenotype, calculated as a sum of the genome-wide genotypes weighted by corresponding genotype effect size estimates derived from genome-wide association study (GWAS) data.Classic genetic risk scores include only a reduced set of single-nucleotide polymorphisms (SNPs) that fulfill a statistical level of significance.In contrast, PRS include millions of SNPs, explicitly modeling the correlation structure between SNPs without identifying a minimal subset of SNPs for prediction.These risk scores have limited predictive accuracy as they cannot confidently predict the clinical outcome of interest with precision at the individual patient level.PRS-based risk stratification could be of potential utility for diseases that already have population-based screening and prevention programs where additional information from PRS can direct screening toward a more restricted group, which could potentially decrease the risks associated with screening the population overall, and lead to cost savings.Mendelian randomization (MR).MR uses genetic variation as a natural experiment to investigate the causal relationships between potentially modifiable risk factors and diseases or phenotypes in observational data.Major limitations of evidence from observational studies include unmeasured confounding and reverse causality.The idea behind MR is that as genotype is randomly allocated at conception and is invariant over the lifetime, they can be used as genetic proxies unaffected by confounding or reverse causation to infer causality.A valid MR study depends on the genetic variant (instrument) fulfilling 3 key assumptions: they associate with the risk factor of interest; they share no common cause with the outcome; and they do not affect the outcome except through the risk factor.| Genomic landscape of hypertension representing a composite of monogenic and polygenic variants from linkage studies and genome-wide association studies (GWASs) along with the molecular and tissue context of the implicated genes.The genes discovered through linkage analysis or GWASs are depicted in bold within the cells, and the GWAS implicated molecular pathways are in red.The circos plot at the lower right depicts the monogenic (filled red circles) and polygenic (purple, dark green, and light green circles representing, respectively, single-nucleotide polymorphisms [SNPs] associated with systolic blood pressure, diastolic blood pressure, or pulse pressure in GWAS) variants associated with blood pressure.Chromosomes are represented as numbered segments.The monogenic syndromes and causal genes are presented circumferentially along with a selection of GWAS SNPs and their associated genes ðcontinuedÞ that have therapeutic potential.CD, collecting duct; DCT, distal convoluted tubule; NCC, Na þ -Cl À cotransporter; NKCC2, Na þr e v i e w PS Garimella et al.: Genomics of hypertension Monogenic syndromes.Gordon syndrome (pseudohypoaldosteronism type II), which is linked to 4 genetic defects involving the WNK1, WNK4, KLHL3, and CUL3 genes, 16 is the only form of monogenic hypertension that manifests as low-renin hypoaldosteronism with hyperkalemia and acidosis. 17,18Mutations in these genes result in the lack of inhibition of WNK4 kinase by the protein products kelch-like 3, cullin 3, and with-no-lysine kinase 1 (WNK1), thereby leading to with-no-lysine kinase 4 (WNK4) accumulation and overactivity of the thiazide-sensitive Na þ -Cl À cotransporter (NCC), which, in turn, causes hypertension with acidosis and hyperkalemia.Clinically, it is a biochemical and phenotypic "mirror image" of Gitelman syndrome (see below) and treatment is by blocking the Na þ -Cl À cotransporter with thiazide diuretics. 19iddle syndrome is an autosomal dominant low-renin and low-aldosterone hypertension disorder resulting from frameshift mutations in the genes coding the b and g subunits of the epithelial Na þ channel (encoded by SCNN1B or SCNN1G). 20,21Consequently, these regions are unable to bind to the ubiquitin ligase neural precursor cell expressed developmentally down-regulated protein 4-2 (Nedd4-2), resulting in disruption of epithelial Na þ channel internalization and proteasomal degradation with consequent overexpression and increased sodium reabsorption independent of aldosterone. 22The typical clinical features are suppressed plasma renin and aldosterone levels, hypokalemic metabolic alkalosis, and early-onset hypertension.Blockers of the epithelial Na þ channel-amiloride and triamterene (not aldosterone antagonists)-ameliorate the condition.
Like Liddle syndrome, the syndrome of apparent mineralocorticoid excess is also characterized by hypertension, hypokalemia, and metabolic alkalosis but caused by a deficiency of 11b-hydroxysteroid dehydrogenase encoded by the HSD11B2 gene. 23,24The primary role of this enzyme is the peripheral metabolism of cortisol to cortisone, thus preventing its binding to the mineralocorticoid receptor.Lack of the enzyme results in unopposed mineralocorticoid activation and Na þ reabsorption. 25Management is with mineralocorticoid receptor antagonists, K þ supplements, and dietary Na þ restriction.
Geller syndrome is caused by heterozygous mutation of the mineralocorticoid receptor gene (nuclear receptor subfamily 3 group C member 2, NR3C2), 26 resulting in increased Na þ reabsorption and hypertension arising from activation of mineralocorticoid receptors by progesterone.This presents as severe hypertension during pregnancy when progesterone levels are increased.Mineralocorticoid receptor antagonists such as spironolactone paradoxically exacerbate hypertension and electrolyte disturbances and are thus contraindicated.This is because the mutation alters the binding parameters of the ligand-binding domain of the mineralocorticoid receptor, increasing its affinity for spironolactone.
The causal role of Na þ in hypertension is further bolstered by the identification of mutations that result in Na þ wasting and hypotension.Classic Bartter syndrome types 1 and 2 are disorders of the thick ascending limb of loop of Henle (TAL) resulting from variants in SLC12A1 and KCNJ1 genes with the consequent loss of function of Na þ -K þ -Cl À cotransporter 2 (NKCC2) and K þ rectifier channel (KCNJ1), respectively (Figure 1). 27Bartter syndrome types 3, 4a, and 4b are specific to the distal convoluted tubule involving the CLCNKB, BSND, and CLCNKA genes, which encode proteins for the Cl À channels ClC-Kb, barttin, and CLC-Ka, respectively.9][30] Each of these variants is inherited in an autosomal recessive manner, whereas Bartter syndrome type 5 involving the melanoma-associated antigen D2 (MAGED2) gene shows X-linked inheritance and exhibits defects in both the TAL and the distal convoluted tubule. 31Gitelman syndrome (familial hypokalemia hypomagnesemia) is the most common inherited tubulopathy (1 in 40,000), and although similar to Bartter syndrome with respect to hypotension and hypokalemic metabolic alkalosis, it is additionally characterized by hypomagnesemia and hypocalciuria.Gitelman syndrome results from biallelic inactivation of the SLC12A3 gene 32 encoding the Na þ -Cl À cotransporter expressed in the apical membrane of cells lining the distal convoluted tubule.To date, >350 mutations have been identified, with most patients being compound heterozygous for the SLC12A3 gene. 33The use of next-generation sequencing including genes involved in both Bartter and Gitelman syndromes is recommended to distinguish overlapping clinical phenotypes. 32WAS.The GWAS SNP that has the widest range of evidence supporting a causal role in hypertension through Na þ pathways is the uromodulin locus. 34Uromodulin is a protein produced exclusively by the TAL and distal convoluted tubule. 35Evidence from the last decade indicates its role in a novel hypertension pathway.Carriers of the minor G allele of the UMOD promoter SNP rs13333226 have lower levels of urinary uromodulin excretion and a lower risk of hypertension. 34,36This lower risk of hypertension is the result of resistance to sodium-induced elevations in BP, which has been demonstrated using UMOD knockout mice, which show a leftward shift in the pressure-natriuresis curve in response to saline loading. 36Additionally, UMOD overexpression in transgenic mouse models results in a dose-dependent increase in uromodulin excretion and rise in BP, which is mitigated with loop diuretics in both mice and humans homozygous for these alleles. 37Further support for uromodulin influencing Na þ homeostasis through tubular mechanisms comes from cotransporter 2; PPGL, pheochromocytoma paraganglioma; KCNJ1/ROMK, potassium channel, inwardly rectifying subfamily J member 1/renal outer medullary potassium channel; TAL, thick ascending limb of loop of Henle; VSMC, vascular smooth muscle cell.
general population studies where higher urinary uromodulin concentrations have been shown to associate with higher urinary Na þ , Cl À , and K þ excretion and osmolality. 38Uromodulin has been shown to upregulate Na þ -K þ -Cl À cotransporter 2 activity by phosphorylation in the TAL. 39herefore, in states where the rise in BP is dependent on sodium reabsorption in the TAL, blocking this with loop diuretics may provide an effective means of treating hypertension, as is being investigated in a clinical trial 40 (ClinicalTrials.govidentifier NCT03354897).Although evidence from transplantion studies in humans suggests that the presence of a donor T allele at rs12917707 is associated with lower uromodulin levels and a lower risk of incident kidney failure, 41 the presence of neither donor nor recipient T allele of rs12917707 is associated with the risk of hypertension after kidney transplantation. 42The definitive evidence that uromodulin is independently associated with BP comes from Mendelian randomization (MR) studies using urinary uromodulin GWAS SNPs as exposures and BP and kidney function GWAS SNPs as outcomes.Ponte et al. 43 showed that each 1 mg higher genetically predicted urinary uromodulin/ creatinine level was associated with 1 ml/min per 1.73 m 2 lower estimated glomerular filtration rate (eGFR), 6% higher odds of having chronic kidney disease, 0.11 mm Hg higher systolic BP, and 0.09 mm Hg higher diastolic BP (DBP).The independent effect of uromodulin on BP and eGFR was quantified using bidirectional and multivariable MR to show that 28% of uromodulin's total effect on BP was mediated by eGFR with the remainder due to the direct effect. 43P GWAS SNPs located near the natriuretic peptide A and B genes (NPPA/B) 44 and natriuretic peptide receptor 3 (NPR3) 44 implicate natriuretic peptides, which increase mGFR and inhibit kidney Na þ reabsorption by decreasing activity of Na þ /K þ adenosine triphosphatase and Na þ -glucose cotransporter in the proximal convoluted tubule.An association between a low-frequency missense variant rs139491786 in solute carrier family 9, subfamily A, member 3 regulator 2 (SLC9A3R2) and BP has now been reinforced by a large exome sequencing study which found that the burden of rare loss-of-function and missense variants in SLC9A3R2 was strongly associated with a lower risk of hypertension. 11,14LC9A3R2 encodes Na þ /H þ exchange regulatory cofactor 2, which is a scaffolding protein interacting with Na þ /H þ exchanger 3 in kidney and intestinal cells modulating Na þ absorption and thence hypertension.

Adrenal and renin-angiotensin-aldosterone systems
Primary hyperaldosteronism accounts for w10% of all forms of refractory hypertension and includes sporadic (adrenal adenoma and hyperplasia) and familial forms.There is a surfeit of monogenic mutations in genes of the adrenal steroid and renin-angiotensin-aldosterone pathways.][11][12][13][14] Congenital adrenal hyperplasia.Congenital adrenal hyperplasia types IV and V caused by loss-of-function mutations in the genes encoding mutations in 11b-hydroxylase (cytochrome P450 family 11 subfamily B member 1, CYP11B1) and 17a-hydroxylase (cytochrome P450 family 17 subfamily A member 1, CYP17A1), respectively, are the 2 subtypes of congenital adrenal hyperplasia known to cause monogenic hypertension.The loss of 11b-hydroxylase prevents the conversion of deoxycortisone and deoxycortisol into corticosterone and cortisol, respectively, resulting in high levels of deoxycorticosterone, deoxycortisol, and androgens, mainly androstenedione and dehydroepiandrosterone.Elevated deoxycortisol and deoxycorticosterone levels have mineralocorticoid function leading to hypertension and hypokalemia.Loss of 17a-hydroxylase blocks the production of cortisol and sex hormones and shunts all steroid production in the mineralocorticoid pathway and decreases the production of sex hormones.Antihypertensive therapy for both includes suppression of adrenocorticotropic hormone secretion with glucocorticoids to inhibit excess production of steroids and mineralocorticoids, along with spironolactone, amiloride, and calcium channel blockers. 45amilial hyperaldosteronism.Familial hyperaldosteronism type I, also known as glucocorticoid remediable aldosteronism, is an autosomal dominant syndrome due to increased adrenocorticotropic hormone production. 46A chimeric gene formed by the fusion of the 5ʹ regulatory sequence of 11bhydroxylase (CYP11B1) with the distal coding sequences of aldosterone synthase (CYP11B2) leads to the ectopic expression of aldosterone synthase in the zona fasciculata, resulting in continuous aldosterone production under the control of adrenocorticotropic hormone. 47,48In contrast, familial hyperaldosteronism type II is caused by germline mutations in the chloride voltage-gated channel 2 (CLCN2) gene (R172Q), resulting in increased aldosterone production triggered by cellular depolarization from increased Cl À efflux and Ca 2þ influx. 48,49Familial hyperaldosteronism type III is associated with heterozygous germline mutations in the potassium inwardly rectifying channel subfamily J member 5 (KCNJ5) gene (T158A, G151R, and G151E) characterized by significant bilateral adrenal hyperplasia with increased aldosterone synthase and enzymes involved in cortisol synthesis. 48eterozygous germline mutations in the calcium voltagegated channel subunit alpha1 H (CACNA1H) gene (M1549V and M1549I) causes familial hyperaldosteronism type IV characterized by increased Ca 2þ influx and aldosterone production.
Somatic mutations causing primary aldosteronism.Somatic mutations in the K þ channel Kir3.4 (KCNJ5), Ca 2þ channel ClC-2 (CLCN2), b-catenin (catenin beta 1, CTNNB1), and/or G-protein subunits a q/11 (GNAQ/11) are responsible for autonomous aldosterone-producing adenomas and usually present with unilateral adrenal tumors and hypertension.G151R and L168R mutations in KCNJ5 account for >40% of aldosterone-producing adenomas.The identification of some of these mutations in aldosterone-producing (micro)nodules indicates a pathogenic continuum from a de novo mutation in a single cell through nodule to adenoma formation and a clinical continuum from the normal state through subclinical to overt primary aldosteronism. 48renergic/noradrenergic pathways Most genomic signals are monogenic mutations resulting in pheochromocytoma and paraganglioma tumors collectively referred to as pheochromocytoma paraganglioma (PPGL), which represent the second set of tumor syndromes associated with hypertension.Treatment options are summarized in Table 1.][11][12][13][14] Monogenic.PPGLs originate from the chromaffin cells of the embryonic crest.Pheochromocytomas originate from the adrenal medulla, whereas paragangliomas are extra-adrenally located in the abdomen, thorax, pelvis, and neck.They cause hypertension through catecholamine hypersecretion except the head and neck paragangliomas, which arise from the parasympathetic ganglia.PPGLs are due to germline and/or somatic mutations in >20 genes clustered into 3 groups on the basis of the involvement of specific signaling pathways and clinical presentations (Figure 1). 50,51seudohypoxic signaling cluster.Mutations in genes encoding hypoxia-inducible factor 2a (HIF2A), succinate dehydrogenase subunits (SDHA, SDHB, SDHC, and SDHD), succinate dehydrogenase complex assembly factor 2 (SDHAF2), von Hippel-Lindau tumor suppressor (VHL), egl-9 prolyl hydroxylase 1 and 2 (EGLN1/2), fumarate hydratase (FH), malate dehydrogenase 2 (MDH2), and isocitrate dehydrogenase (IDH) activate the hypoxia-inducible factor signaling pathway without hypoxic stimulus and cause an increased production of vascular endothelial growth factor, platelet-derived growth factor, and transforming growth factor a, leading to cell growth, microvascular proliferation, increased tyrosine hydroxylase, and catecholamine overproduction.PPGLs in this cluster are almost all (except von Hippel-Lindau tumor suppressor) extra-adrenal, present with multiple and recurrent tumors that are aggressive and frequently metastatic, and have poor clinical outcomes. 50,51inase signaling cluster.Mutations in rearranged during transfection proto-oncogene (RET), Harvey rat sarcoma viral proto-oncogene (H-RAS), and Kirsten rat sarcoma viral proto-oncogene (K-RAS); neurofibromin 1 (NF1) tumor suppressor; transmembrane protein 127 (TMEM127); Mycassociated factor X (MAX); alpha thalassemia/mental retardation syndrome X-linked (ATRX); and cold shock domain containing E1 (CSDE1) dysregulate phosphatidylinositol-3 0 -kinase (PI3K)/mechanistic target of rapamycin kinase (mTOR) signaling and present as PPGLs, which are mainly adrenal and generally have good clinical outcomes (an exception is the ATRX mutation-related PPGL). 50,51nt signaling cluster.These pheochromocytomas are caused by somatic mutations in CSDE1 and the mastermind like transcriptional coactivator 3 (MAML3) fusion genes (upstream binding transcription factor, RNA polymerase I [UBTF]-MAML3, and transcription factor 4 [TCF4]-MAML3).Wnt-altered tumors exhibit high expression of CHGA, a gene that encodes chromogranin A-a clinical marker of neuroendocrine tumors. 50,51scular More recently, a growing list of monogenic disorders leading to hypertension have been associated with sites of action outside the kidney tubules.
r e v i e w PS Garimella et al.: Genomics of hypertension leading to vasoconstriction. 59The ATP2B1 gene encodes an adenosine triphosphate-dependent Ca 2þ channel critical for vascular contractility and vasodilatation, and the absence of this gene results in hypertension, increased cellular Ca 2þ , and a robust BP response to calcium channel blockers. 60,61henome-wide association studies show that the PHACTR1 SNP rs9349379 is implicated in 5 diseases with vascular components: CAD, migraine, cervical artery dissection, fibromuscular dysplasia, and hypertension.This SNP has been shown to be a distal regulator of EDN1, which encodes endothelin-1 (ET-1), 62 with the G allele associated with higher EDN1 expression, higher ET-1, and lower risk of all diseases mentioned above except CAD.ET-1 can cause both vasoconstriction and hypertension (paracrine) and vasodilation (autocrine) via its actions on vascular smooth muscle cell ET receptor subtypes A and B, respectively.ET-1 induces angiotensin II, and the effects of ET-1 and angiotensin II on vascular reduced nicotinamide adenine dinucleotide phosphate (NADPH) oxidases, particularly reduced nicotinamide adenine dinucleotide phosphate oxidase 1, reduced nicotinamide adenine dinucleotide phosphate oxidase 2, and reduced nicotinamide adenine dinucleotide phosphate oxidase 5, result in sustained BP elevations. 63ET-1 results in vasodilatation via its action on ET receptor subtype B by inducing nitric oxide and prostacyclin release. 64henome-wide association studies are likely explained by the potent vasoconstrictive effect of ET-1 on the coronary circulation, which is devoid of ET receptor subtype B. Thus,  Figure 2 | Causal relationships between genetically determined blood pressure (BP) and a range of traits from Mendelian randomization studies.Increasing BP based on genetic proxies from genome-wide association studies shows a causal effect on the increasing risk of cardiovascular outcomes and phenotypes such as monocytes, neutrophils, and eosinophils (top right).The red arrows denote higher risk of outcomes or increased levels of measured phenotypes in response to BP change.Genetic proxies for the drug effect are used to determine the effect of pharmacological BP lowering on outcomes (left panel).The red arrows denote higher risk of outcomes in response to genetically predicted BP decrease as a marker of drug effect.The panel below the x-axis shows the causal effect of genetically predicted higher levels of a range of measured risk factors on BP (increasing BP: red arrows; decreasing BP: blue arrows).These causal effects represent lifelong influence on the trait, and hence the magnitude of BP effects is small.ACE, angiotensin-converting enzyme; ACEI, angiotensin-converting enzyme inhibitor; AF, atrial fibrillation; ARB, angiotensin receptor blocker; BB, b-blocker; BMI, body mass index; CAD, coronary artery disease; CCB, calcium channel blocker; CKD, chronic kidney disease; CVA, cerebrovascular accident; eGFR, estimated glomerular filtration rate; HbA1C, hemoglobin A1c; HDL, high-density lipoprotein; HF, heart failure; HTN Preg, hypertensive disorders of pregnancy; IGFBP3, insulin-like growth factor binding protein 3; IL-16, interleukin-16; LDL, low-density lipoprotein; NT-proBNP, N-terminal pro-brain natriuretic peptide; T2DM, type 2 diabetes mellitus; TSH, thyroid stimulating hormone; UACR, urine albumin-to-creatinine ratio; WMH, white matter hyperintensity.
unopposed vasoconstrictive ET-1 action on the coronary vasculature is atherogenic and via its action on ET receptor subtype A results in coronary vasospasms. 65

Outcomes
The causal relationships between high BP and cardiovascular disease (CVD) and non-CVD outcomes have been the subject of MR studies (Figure 2).Genetically predicted systolic BP (SBP) was causally associated with hypertension-related CVD such as CAD, stroke, heart failure, atrial fibrillation, and also a range of additional CVDs including aortic aneurysm, aortic stenosis, dilated cardiomyopathy, endocarditis, peripheral vascular disease, and rheumatic heart disease as well as negatively associated with venous thromboembolism. 66The authors extrapolated these results from the UK Biobank to estimate an overall 17%, 31%, and 56% decrease in morbidity for a 5, 10, and 23 mm Hg decrease in SBP at a population level. 667][68][69][70][71] In contrast, MR studies using eGFR as an exposure showed that lower genetically predicted eGFR is associated with higher BP. 72Although SBP and DBP are correlated traits, SBP alone is included in CVD risk prediction.Epidemiologically, DBP is more closely associated with coronary heart disease development in the young whereas in those older than 60 years SBP is more predictive.BP GWAS SNPs predominantly show association with both SBP and DBP, but a minority of SNPs show exclusive association with just 1 trait.An MR study 73 used 3 sets of BP GWAS SNPs-242 independent SNPs associated with both SBP and DBP, 120 SBPexclusive SNPs, and 80 DBP-exclusive SNPs-to unravel the distinct effects of SBP and DBP on hypertension outcomes.This study showed that SBP is the causal driver for CAD, stroke, and ischemic stroke while it is DBP for small vessel stroke.Furthermore, SBP is exclusively associated with heart failure, atrial fibrillation, and type 2 diabetes mellitus. 73

Pharmacogenomics
Genetic variants associated with disease traits have pointed to effective drug targets. 74Examples include HMGCR, which is associated with serum cholesterol levels and is the target for statins 75 ; 27 drug target genes of approved rheumatoid arthritis drugs demonstrated a significant overlap with 98 biological rheumatoid arthritis risk genes from GWASs 76 ; SNPs in NR3C2 is associated with moderately increased albuminuria, and an NR3C2 antagonist, finerenone, is now approved for the treatment of chronic kidney disease. 77issense variants in the tyrosine kinase 2 gene (TYK2) have been associated with systemic lupus erythematosus, and evidence of its interaction with the interferon a/b receptor subunit 1 led to the development of the interferon a/b receptor subunit 1 antagonist anifrolumab for the treatment of systemic lupus erythematosus. 78This demonstrates the potential of using indirect evidence from genetic association to drive drug discovery.By extension, the growing wealth of GWAS data on BP and hypertension should inform the selection of the best targets with a measurable impact on the successful development of new drugs (Table 1).
Another valuable use of GWAS results is to use gene variants corresponding to the targets of common pharmacological agents for hypertension as a proxy for treatment effects in MR (Figure 2).This allows establishing any relationship with adverse events and offers an insight into drug repurposing. 79Such studies have shown that calcium channel blockers have a protective effect on stroke, atrial fibrillation, CAD, and diverticulosis 79 ; b-blockers and thiazide diuretics increase the risk of T2DM 79 ; angiotensin-converting enzyme inhibitors may have an adverse impact on schizophrenia risk 80 and colorectal cancer 81 but reduce the risk of type 2 diabetes mellitus 79 ; and the beneficial effect of antihypertensive drugs on Alzheimer disease risk is due to their effect on SBP. 82NA interference is a natural mechanism by which short strands of RNA, such as small, interfering RNA, cause targeted gene suppression.83 From a hypertension perspective, zilebesiran, an RNA interference therapeutic targeting hepatic angiotensinogen synthesis, is currently in a phase 2 trial (ClinicalTrials.govidentifier NCT05103332) after demonstrating sustained serum angiotensinogen and BP reductions through 6 months in a phase 1 trial.

Polygenic risk scores
GWASs have shown that BP is a polygenic trait influenced by hundreds of DNA variants each of which contributes smallto-moderate effects, and the aggregate effect of these represent the polygenic hypertension risk. 84A BP genetic risk score accounted for w13 mm Hg in variation of BP.However, the BP genetic risk score failed to show a clear predictive link with eGFR, 44 suggesting that BP is not a strong causal risk factor for kidney failure and this is supported by MR studies as noted above.A BP polygenic risk score (PRS) in the top 2.5% conferred a 2.3-fold risk of hypertension and earlier hypertension onset by 10 years and incident CVD. 44,85PRSs are set at conception and can be used earlier in life than lifestyle, agerelated, or other nongenetic risk factors.However, PRSs have limited predictive accuracy, primarily because genetic factors are not the sole risk factors for hypertension and the risk scores contain information only from SNPs that represent a fraction of the genetic contribution to the trait.A number of potential applications are envisaged for PRSs, including costeffective primary prevention and precision medicine. 84A possible application of PRSs in hypertension would be in the early stages of the disease to confirm the diagnosis and prioritize patients for more intensive investigation and follow-up or initiation of treatment.

Conclusions
The opportunities for leveraging genomics in hypertension prediction and management have vastly expanded over the last 15 years.Although challenges remain, parallel advances in gene silencing and polygenic risk scores along with growing recognition of genetic inequity indicate areas where the next r e v i e w PS Garimella et al.: Genomics of hypertension wave of application research is expected.Beyond sequence variations, the dark matter of common disease genomics representing structural and epigenetic variations and geneenvironmental interactions are now tractable by advances in high-throughput sequencing and omic technologies.

DISCLOSURE
PSG has served as a consultant for Otsuka Inc and Dialysis Clinic Inc.All the other authors declared no competing interests.

PS
Figure1| Genomic landscape of hypertension representing a composite of monogenic and polygenic variants from linkage studies and genome-wide association studies (GWASs) along with the molecular and tissue context of the implicated genes.The genes discovered through linkage analysis or GWASs are depicted in bold within the cells, and the GWAS implicated molecular pathways are in red.The circos plot at the lower right depicts the monogenic (filled red circles) and polygenic (purple, dark green, and light green circles representing, respectively, single-nucleotide polymorphisms [SNPs] associated with systolic blood pressure, diastolic blood pressure, or pulse pressure in GWAS) variants associated with blood pressure.Chromosomes are represented as numbered segments.The monogenic syndromes and causal genes are presented circumferentially along with a selection of GWAS SNPs and their associated genes ðcontinuedÞ that have therapeutic potential.CD, collecting duct; DCT, distal convoluted tubule; NCC, Na þ -Cl À cotransporter; NKCC2, Na þ - subunit alpha 1, ATP1A1), plasma membrane Ca 2þ transporting adenosine triphosphatase 3 (ATPase plasma membrane Ca 2þ transporting 3, ATP2B3), Ca 2þ channel Ca v 3.2 (calcium voltage-r e v i e w PS Garimella et al.: Genomics of hypertension gated channel subunit alpha1 H, CACNA1H), Cl À channel

Table 1 |
BP monogenic genes and their specific treatment along with GWAS loci that are near gene targets for known BPlowering medications (approved therapies appear in bold, and others with repurposing potential appear in italics)