|Home | About | Journals | Submit | Contact Us | Français|
Hypertension is a common disorder that affects a large heterogeneous patient population. Subgroups can be identified on the basis of their responses to hormonal and biologic stimuli. These subgroups include low-renin hypertensives and nonmodulators. Aldosterone, the principal human mineralocorticoid, is increasingly recognized as playing a significant role in cardiovascular morbidity, and its role in hypertension has recently been reevaluated with studies that suggest that increased aldosterone biosynthesis (as defined by an elevated aldosterone to renin ratio) is a key phenotype in up to 15% of individuals with hypertension. It was reported previously that a polymorphism of the gene (C to T conversion at position −344) encoding aldosterone synthase is associated with hypertension, particularly in individuals with a high ratio. However, the most consistent association with this variant is a relative impairment of adrenal 11β-hydroxylation. This review explores the evidence for this and provides a hypothesis linking impaired 11β-hydroxylation and hypertension with a raised aldosterone to renin ratio. It is also speculated that there is substantial overlap between this group of patients and previously identified low-renin hypertensives and nonmodulators. Thus, these groups may form a neurohormonal spectrum reflecting different stages of hypertension or indeed form sequential steps in the natural history of hypertension in genetically susceptible individuals.
High BP remains an important and common clinical problem that affects one in four adults in the United States (1). Historically, hypertension has been subdivided into “essential” and “secondary” forms. Essential hypertension (cause unknown) accounts for 95 to 99% of cases and has traditionally been viewed as a consequence of interaction between environmental factors (e.g., sodium intake) and genetic background. However, the identity of the genes that predispose to hypertension in the great majority of patients remains unknown. A smaller proportion of patients are identified as having secondary hypertension, as a consequence of some biochemical or mechanical pathology that is potentially reversible. More recently, however, the notion that secondary hypertension is rare has been challenged by the suggestion that primary aldosteronism (PA; originally thought to be present in only 1% of individuals with hypertension) is present in up to 15% of unselected individuals with hypertension (2). In this review, we explore the concept that altered regulation of aldosterone production is a common feature of essential hypertension and PA, with the implication that there is substantial overlap between the two. In turn, this could lead to a shift in therapeutic strategies and increased awareness of a role for aldosterone in cardiovascular pathophysiology.
Aldosterone, the principal human mineralocorticoid, is produced in the zona glomerulosa of the adrenal gland. The hormone is the product of a series of biosynthetic reactions summarized in Figure 1. The final key steps in aldosterone synthesis are sequential 11-hydroxylation, 18-hydroxylation, and 18-oxidation of the precursor steroid deoxycorticosterone (DOC) in zona glomerulosa cells. A single enzyme, aldosterone synthase, carries out all of these steps and is encoded by the gene CYP11B2. A similar, highly homologous enzyme, 11β-hydroxylase (encoded by the gene CYP11B1), acts in parallel in the zona fasciculata to convert 11-deoxycortisol to cortisol. The two genes lie in tandem on chromosome 8 in humans and are highly homologous; the protein products are also very similar and share ~95% identity of primary sequence (3).
Angiotensin II (Ang II; regulated by the renin-angiotensin system) and plasma potassium are the principal regulators of aldosterone production. Ang II stimulates aldosterone secretion in response to sodium depletion and reduced extracellular fluid volume (4), and even small increments in plasma potassium act as a powerful stimulus for aldosterone production (5). Other factors also influence aldosterone production. In particular, adrenocorticotrophin (ACTH) exerts an acute effect to stimulate aldosterone production (6), although its importance in long-term regulation of aldosterone is uncertain.
Aldosterone binds to the mineralocorticoid receptor (MR), an intracellular receptor that belongs to the steroid/thyroid/retinoid/orphan receptor superfamily (7). Once bound, the ligand/receptor complex translocates to the nucleus and acts as a transcription factor by direct interaction with DNA regulatory elements (the classical genomic effect of aldosterone) (8). As the MR has similar affinities for aldosterone and cortisol, the 11β-hydroxysteroid dehydrogenase system acts as a gatekeeper to prevent activation by much higher available levels of cortisol (9). The type 2 isoform of this enzyme is found in the renal distal nephron and converts cortisol to its inactive metabolite, cortisone, which has no affinity for the MR.
Traditionally, the principal target organ for aldosterone was said to be the kidney; MR are found in high concentration in the renal distal nephron as well as other epithelial sites, such as the colon and ducts of sweat and salivary glands (10). However, MR have also been identified in nonepithelial sites, such as heart, brain, vascular smooth muscle, liver, and peripheral blood leukocytes (10).
The best-characterized physiologic effect of aldosterone is to increase the reabsorption of sodium in the kidney and at other secretory epithelial sites at the expense of potassium and hydrogen ions (4). The major sites of aldosterone-induced sodium and potassium transport are luminal cells of the cortical collecting tubules and the distal convoluted tubule. The apically located epithelial sodium channel (ENaC) is the major determinant of renal sodium reabsorption (11). Its availability in open conformation at the apical membrane of the cell is increased by aldosterone and also by vasopressin, glucocorticoids, and insulin; elevated intracellular levels of calcium and sodium downregulate it (12).
The gene product(s) resulting from the interaction of aldosterone/MR complexes binding to DNA regulatory elements are termed aldosterone-induced protein(s) (AIP). AIP may have effects on the apical membrane, cellular energy production, and/or the basolateral Na/K-ATPase pump, resulting in increased sodium reabsorption and potassium and hydrogen ion excretion (13) (Figure 2).
Recently, a serine/threonine kinase, serum glucocorticoid-regulated kinase (sgk), has been identified as an AIP resulting in an increase in ENaC activity (14). Aldosterone causes phosphorylation and activation of sgk, which in turn increases ENaC activity by an increase in the number of channels at the cell surface (15). The principal ENaC inhibitory accessory protein is Nedd 4 (neuronal precursor cells expressed developmentally downregulated). This ubiquitin protein ligase binds to the C tails of β and γ subunits of ENaC, leading to channel internalization and degradation. Recent work suggests that the stimulatory action of sgk on ENaC is mediated through phosphorylation of serine residues on Nedd4. Such phosphorylation reduces the interaction between Nedd4 and ENaC, leading to elevated ENaC cell surface expression (16).
It is likely that aldosterone affects BP regulation by mechanisms other than or in addition to simple plasma volume expansion and the associated increase in cardiac output as a result of its action on sodium homeostasis. For example, activation of MR in vascular smooth muscle results in alteration in pressor responsiveness to adrenergic stimulation. Moreover, evidence suggests that aldosterone binding by the MR in cardiac tissue regulates collagen formation (17). It is feasible that similar action in peripheral blood vessels might result in remodeling, which could sustain an elevated BP. This is supported by evidence suggesting that aldosterone levels are inversely related to arterial compliance in essential hypertension.
It is now accepted that, as well as classical genomic effects through ligand/receptor binding of DNA regulatory elements, aldosterone exerts rapid, nongenomic effects. This has led to investigation of aldosterone action in tissues other than the kidney. Reports of rapid, nongenomic effects of aldosterone have been described in smooth muscle, skeletal muscle, colonic epithelial cells, and myocardial cells (18). These effects have been linked to the development of increased systemic vascular resistance and so could, theoretically, contribute to hypertension and cardiovascular disease.
Although the underlying pathophysiology remains incompletely understood, it is assumed that essential hypertension has a multifactorial cause and that no single cause exists. However, a number of studies draw attention to the adrenal cortex and its contribution to BP elevation.
Functional abnormalities of the adrenal cortex were suggested as a cause for essential hypertension many years ago (19); indeed, early studies reported that adrenal cortex hyperplasia was a feature of many hypertensive individuals at post mortem examination (20,21). In addition, a number of abnormalities of urinary excretion, plasma levels, and clearance of several adrenal steroids in patients with hypertension have been identified over the years (22,23). No single defect in adrenal corticosteroid biosynthesis has been identified, but it is relevant to consider briefly two rare monogenic syndromes involving 11β-hydroxylase and aldosterone synthase that cause hypertension and help identify candidate mechanisms (Table 1): (1) Glucocorticoid remediable aldosteronism (GRA) is a rare autosomal dominant condition characterized by hypertension and aldosterone excess that is regulated by ACTH rather than Ang II. The molecular basis of this condition was first described in 1992 (24). In GRA, a chimeric gene that contains the 5' promoter sequence of CYP11B1 and functional elements of CYP11B2 is created, resulting in aldosterone production under control of ACTH. (2) 11β-Hydroxylase deficiency is a rare cause of congenital adrenal hyperplasia, accounting for 5 to 8% of cases (25). In this autosomal recessive disorder, mutations in CYP11B1 result in impaired activity of 11β-hydroxylase, leading to accumulation of the steroid precursors 11-deoxycortisol and deoxycorticosterone (Figure 1). This leads to mineralocorticoid hypertension in approximately two thirds of cases.
In light of these rare syndromes, it is of interest that impaired activity of the enzyme 11β-hydroxylase has also been reported to be a feature of patients with essential hypertension. In 1985, de Simone et al. (26) demonstrated that ACTH-stimulated plasma levels of DOC were increased in hypertensive patients compared with control subjects, a finding similar to that of Honda et al. (27) 10 yr earlier. More recently, we observed a similar phenomenon in patients with hypertension from Italy, in whom the ratio of 11-deoxycortisol to cortisol (a marker of 11β-hydroxylase activity) was elevated (28). Although the precise cause of this is unclear, we suggest later that it may be a consequence of variation at the CYP11B1/CYP11B2 loci, which encode 11β-hydroxylase and aldosterone synthase, respectively.
Over the years, there has been a substantial effort to categorize large, heterogeneous groups of individuals with essential hypertension into smaller, homogeneous subgroups on the basis of hormonal responses to biologic stimuli. One of the earliest classifications was that of low-renin hypertension (29). In this form of hypertension, subjects exhibit low plasma renin activity, which does not respond normally to sodium restriction, and maintain basal aldosterone levels, which, although not elevated, are inappropriate for the principal trophin, Ang II. Such a hormonal profile may be due to increased responsiveness of aldosterone to Ang II (30,31), although not all patients share this abnormality (30). Classically, individuals with low renin have sodium-sensitive hypertension, which tends to respond better to diuretics than to agents that block the renin-angiotensin-aldosterone system. Low-renin hypertension is found more frequently among black and elderly populations (32).
A second subset of hypertension was subsequently described in the mid-1980s by William and Hollenberg (33). In this group, changes in sodium intake fail to produce the anticipated reciprocal changes in adrenal (aldosterone) and renal vascular responses to Ang II infusions. Such individuals, who have normal/high renin levels, are termed “nonmodulators.” Non-modulators tend to be older than modulators, and there is evidence to suggest that nonmodulation has a genetic basis. In one study of individuals with hypertension, 81% with a positive family history of hypertension were nonmodulators (34). A previous study of sibling pairs with hypertension showed that nonmodulation tended to aggregate within families and is independent of sodium intake (35). In common with low-renin essential hypertension, nonmodulators also demonstrate salt sensitivity. In contrast to low-renin hypertension, however, nonmodulators clinically respond best to angiotensin-converting enzyme inhibitors as opposed to diuretics (Table 2) (36).
Thus, within the large population of individuals with essential hypertension, two significant, salt-sensitive phenotypes have been identified with different optimal therapeutic agents. We suggest there is an additional subgroup of individuals with hypertension that may be relatively common and that may relate to the previously described low-renin population such that there is a continuum in which a common genetic predisposition is modified by other factors over many years to result in a range of seemingly distinct clinical presentations.
PA can be defined as overproduction of aldosterone independent of its normal chronic regulator, Ang II (37). Classically, Conn’s syndrome was reported in subjects with aldosterone-producing adrenal adenomas and was previously thought to be the most common cause of PA. However, recent studies have led to a reevaluation of its prevalence and suggest that bilateral adrenal hyperplasia (also known as idiopathic hyperaldosteronism) is more common than first thought (38). This change in emphasis has been provoked by wider use of the aldosterone to renin ratio (ARR) as a screening and diagnostic test. Using this, several groups worldwide now have reported the prevalence of PA, as defined by an ARR value above a given cutoff, in unselected individuals with hypertension to be between 5 and 15%. For example, one study of unselected individuals with hypertension in Dundee, Scotland, suggested a prevalence of 10% in both a primary and a secondary care setting (39). Independent groups in Australia (40), the United States (41), South East Asia (42), South Africa (43), and South America (44) have made similar claims. Indeed, such figures are reminiscent of claims by Conn in 1960s that normokalemic PA is present in up to 20% of all patients with hypertension—a claim that was generally discredited at the time (45). Thus, aldosterone excess may be an increasingly significant contributor to cardiovascular morbidity.
The ARR was first introduced as a screening tool for PA in the early 1990s. This ratio seems to be fairly robust and is less affected by day-to-day or diurnal variation and posture than plasma renin activity or aldosterone levels individually (46). However, use of this ratio is not without problems. In particular, the ratio can obviously be affected by concomitant drug administration. Furthermore, in the majority of patients identified by a raised ARR, the real reason is low levels of renin, which dominate the ratio. Thus, in a careful analysis of the mathematical derivation of the ARR, Montori and Young (47) demonstrated that the renin measurement was the determinant of the ratio, so its positive predictive power for PA was relatively low. Indeed, in many patients, the level of aldosterone is within the “normal” range, and it is uncertain whether they would meet earlier criteria for the diagnosis of PA. However, in many such individuals, dynamic tests of the renin-angiotensin system are abnormal, confirming an altered relationship between renin and aldosterone. How this group differs from the subgroup of low-renin hypertension previously described is unclear. In these patients, levels of aldosterone are higher than predicted from the prevailing renin in common with individuals with a raised ARR, and there is likely to be substantial overlap between the two groups. This theory reiterates previous arguments by Padfield et al. (48), who claimed that PA as a result of bilateral adrenal hyperplasia was a variant not of classical Conn’s adrenal adenoma but of low-renin essential hypertension. In keeping with this, it should be noted that nodular change in the adrenal gland with or without hyperplasia is not specific for idiopathic aldosteronism, having been described in individuals with essential hypertension, low-renin hypertension, and even normotension (49). Furthermore, pathologic studies in adrenal tissue removed from individuals with apparent solitary adenomas show that there is often hyperplasia of the adjacent zona glomerulosa and formation of multiple nodules throughout the gland, suggesting that a solitary adenoma may arise in an already abnormal gland (50).
In summary, an elevated ARR can be demonstrated in ~10% of unselected individuals with hypertension, making it the most common cause of secondary hypertension. It is unclear how many of these individuals have “classical” PA with elevated plasma levels of aldosterone, but all clearly have demonstrable abnormalities in renin-angiotensin-aldosterone dynamics, and it may be more appropriate to redesignate such patients as having “aldosterone-associated hypertension.”
Debate over the exact physiologic differences and similarities between subgroups is largely academic. Individuals who are identified by an increased ARR have an inappropriately high aldosterone concentration for its principal trophin, Ang II; whatever gives rise to this disproportionate renin for aldosterone level could clearly have pathogenic and therapeutic implications and deserves further investigation (51).
The corticosteroid biosynthetic pathway and the crucial role of aldosterone synthase (encoded by CYP11B2) and 11β-hydroxylase (CYP11B1) in catalyzing the terminal steps of aldosterone and cortisol biosynthesis, respectively, have already been discussed (see Figure 1). These genes are obvious candidates that might be involved in hypertensive disorders as confirmed by monogenic forms of hypertension described earlier (Table 1). Animal models, in particular the Dahl salt-sensitive rat (51), also suggest a role for this locus (Table 1). In the past few years, studies have focused on two common polymorphisms within the CYP11B2 gene (Figure 3). One is a single nucleotide polymorphism in the 5' promoter region at −344 (C-T) that alters a putative recognition site for the transcription factor SF-1 (52). It is unclear whether this has any physiologic significance: binding of SF-1 is reduced fourfold with the T allele, and there is no detectable effect on gene transcription when studied in vitro (53). The other polymorphism involves intron 2 of CYP11B2, which is partly replaced by the corresponding intron of CYP11B1 (52). These two polymorphisms are in close linkage disequilibrium such that the common haplotypes generated are T/conversion (38%), T/wild type (16%), and C/wild type (45%) (54).
The physiologic significance of these polymorphisms remains controversial. In our study of 138 individuals with hypertension, there was a highly significant excess of TT homozygosity compared with CC homozygosity when compared with individually matched normotensive control subjects (54). Other groups reported similar findings (55). Moreover, subsequent assessment of aldosterone excretion (as its urinary metabolite, tetrahydroaldosterone) in an unrelated population demonstrated a strong association between the T allele and higher excretion rates (54). Further investigations have reported that plasma levels of aldosterone are raised in individuals with the T allele (56), although other groups have failed to confirm these findings (57). In addition, we have reported an association between T allele frequency and elevated ARR with a population of patients with hypertension (Table 3) (58).
In addition to the data on aldosterone and BP, we and others have consistently shown that the T allele and intron 2 conversion are associated with raised basal and ACTH-stimulated levels of the 11-deoxysteroids, DOC, and deoxycortisol (Figure 4) (59). These steroids are converted to corticosterone and cortisol, respectively, by 11β-hydroxylase within the zona fasciculata (Figure 1). This evidence, which initially seems paradoxical, suggests that the T allele of CYP11B2 is associated with impaired activity of the enzyme 11β-hydroxylase, which is encoded by the adjacent gene, CYP11B1. Although the exact molecular mechanism that accounts for this is not understood, we have hypothesized that the −344 T polymorphism in the promoter region of CYP11B2 is in close linkage disequilibrium with a key quantitative trait locus in CYP11B1 adversely affecting its expression or function, resulting in increased levels of 11-deoxysteroids (60). Definitive studies to examine the pattern of variation across the entire locus are currently in progress.
The theory that 11β-hydroxylation may be impaired in individuals with hypertension is not novel (26,27). However, until now, there has been no suggestion that this biochemical abnormality is related to variation at the CYP11B locus. These data also lead us to speculate on the link between a minor change in 11-hydroxylase efficiency and hypertension with aldosterone excess. It is unlikely that the minor increases in DOC and 11-deoxycortisol will have significant biologic effects. However, impaired conversion of deoxycortisol to cortisol as a consequence of reduced 11β-hydroxylase activity should result in a slight reduction in cortisol levels in response to ACTH. In turn, normal feedback regulation should result in a resetting of the hypothalamic-pituitary-adrenal axis such that cortisol levels are maintained. Consequently, there will be a subtle increase in ACTH drive to the adrenal cortex (Figure 5).
Thus, if this were a genetically determined phenomenon, then we would predict that individuals with less efficient cortisol synthesis will maintain a slightly enhanced ACTH drive to the adrenal (effectively, a minor variant of classical 11β-hydroxylase deficiency). In the long term, this is likely to cause hyperplasia of both zona fasciculata and zona glomerulosa of the adrenal cortex, resulting in increased synthetic capacity for both cortisol and aldosterone. Importantly, expression of a number of genes necessary for aldosterone production, including steroidogenic acute regulatory protein, p450 side-chain cleavage (CYP11A), and p450–21-hydroxylase (CYP21), is responsive to ACTH, emphasizing the potential for increased synthetic capacity of aldosterone (61).
Importantly, we do not propose that ACTH is a principal stimulator of excess aldosterone production by the adrenal. Indeed, previous studies have demonstrated the ability of pharmacologic doses of ACTH to stimulate aldosterone production in the short term, but aldosterone production decreases within a few days (62). However, these experiments concentrated on very unphysiologic exposure of the adrenal to grossly excessive amounts of ACTH. In ACTH-dependent Cushing’s disease, in which there is chronic sustained exposure of the adrenal cortex, aldosterone concentrations are not diminished (63). Other proopiomelanocortin (POMC) breakdown products, such as joining peptide and β-endorphin, have also been shown to stimulate aldosterone production by human adrenal cells in vitro. Thus, we suggest that increased levels of a POMC-related product (either ACTH or a related peptide) favor zona glomerulosa hyperplasia and may enhance aldosterone secretion in response to other, more conventional trophins, such as Ang II or potassium. Thus, over a very long period, the genetic change in 11β-hydroxylation efficiency (along with an additional environmental or genetic influence) might result in ACTH-driven adrenal zonal hyperplasia and an alteration (steepening of the dose–response relationship) of the response of aldosterone to Ang II and potassium. It is pertinent that very early studies reported that a proportion of patients with essential hypertension showed a good BP response to low-dose dexamethasone treatment, seemingly supporting the suggestion that ACTH was sustaining production of a hyper-tensinogenic adrenal steroid (64). Furthermore, there are reports of increased levels of dehydroepiandrosterone sulfate (an adrenal androgen driven by ACTH) in patients with hypertension (65). Finally, it is significant that adrenal gland hyperplasia is a common histologic finding in patients with hypertension post mortem and even in patients with seemingly “solitary” adrenal adenomas (20,21).
The concept of “tertiary aldosteronism,” whereby there is sustained and prolonged stimulation of the adrenal by Ang II, is already recognized in renovascular hypertension (38). It has been suggested that the phenotype of hypertension with a raised ARR could be redefined as a form of tertiary aldosteronism, perhaps preceded by low-renin hypertension over a much longer time. Although this theory focuses on a single genetic polymorphism, it is widely accepted that hypertension is a polygenic disorder. It is likely, therefore, that other genes may interact in a synergistic manner to lead to the phenotype of hypertension with an elevated ARR.
In summary, we propose that within the heterogeneous population of essential hypertension, there may be three distinct subgroups, which are currently classified on a hormonal or biochemical basis: individuals with low-renin hypertension, nonmodulators (normal to high renin), and individuals with hypertension and an elevated ARR. These three groups may form a neurohormonal spectrum that reflects differing stages of hypertension, the rate of progression of which depends on other genetic and environmental factors. Hence, the natural history of hypertension may proceed from essential (high to normal renin) hypertension through to low-renin hypertension and finally to tertiary aldosteronism. Clearly, longitudinal studies that compare groups of patients of different ages followed up over time are needed to investigate this theory.
The concept that hyperaldosteronism is a prevalent cause of hypertension has developed at a time of increased interest in the role of aldosterone in a variety of cardiovascular diseases. For example, recent experimental and clinical evidence suggests that excessive circulating levels of aldosterone cause adverse cardiovascular sequelae such as cardiac fibrosis and left ventricular hypertrophy, independent of its effects on BP or plasma volume (66). The high prevalence of congestive cardiac failure, which is very commonly associated with secondary hyperaldosteronism, suggests a significant role for aldosterone excess as a cause of cardiovascular injury. Furthermore, the concept of aldosterone-mediated cardiac injury has led to studies that have explored the use of selective aldosterone receptor antagonists in patients with cardiac disease. The Randomized Aldactone Evaluation Study reported substantial benefit (30% reduction in mortality) in patients who had advanced cardiac failure and were given spironolactone in addition to conventional treatment (67). More recently, the Eplerenone Post Acute Myocardial Infarction Heart Failure Efficacy and Survival Study showed that eplerenone could provide substantial benefit (15% reduction in mortality) in patients after acute myocardial infarction (68). It may be that selective aldosterone receptor antagonists have a wider role in the therapy of hypertension, a concept that merits further evaluation.
As a result of increased use of the ARR as a diagnostic tool, the contribution of aldosterone in hypertension has been more widely recognized. This may be important in up to 10% of unselected individuals with hypertension. There is likely to be a substantial degree of overlap between this group and other well-defined hypertension subgroups (low-renin essential hypertension and nonmodulators), and it may well be that these form sequential steps in the natural history of hypertension in genetically predisposed individuals. As a result, the classical subdivision of hypertension into essential and secondary forms has become diffuse.
There is now greater recognition of the adverse effects of aldosterone on endothelial, renal, cardiac, and central nervous system tissues. Many of these effects may be independent of BP elevation and plasma volume expansion. Thus, development of new, more selective aldosterone receptor antagonists is a major therapeutic challenge to optimize BP control, minimize side effects, and improve cardiovascular morbidity in an increasing number of patients.
J.C. is supported by a Medical Research Council program grant. M.F. is a clinical research fellow funded by the Wellcome Trust (grant no. 069205).