Introduction

Primary aldosteronism was once thought to be an uncommon (<1%) cause of hypertension, relatively benign, and with hypokalemia required for diagnosis. We now know this not to be the case: around the world, primary aldosteronism is now known to account for 8–13% of unselected hypertensives, that it incurs substantially higher morbidity than is seen in essential hypertensives matched for age, sex, and blood pressure, and that hypokalemia is found in only about one quarter of cases. Given the prevalence of hypertension (about 20% of the population in developed countries) primary aldosteronism is thus present in approximately 2% of the population, and as such constitutes a major public health problem. This is particularly the case given its heightened cardiovascular risk profile and our ability of treat it. Unilateral disease, overwhelmingly the result of an aldosterone-producing adenoma, can be treated surgically, and bilateral adrenal hyperplasia can be treated specifically by mineralocorticoid receptor antagonists.

From the viewpoint of pathophysiology, the definition of primary aldosteronism is “autonomous aldosterone secretion.” This definition entails recognition of what is normally involved in elevating aldosterone levels: ACTH, which yokes circadian aldosterone levels to those of cortisol; angiotensin II, which responds to postural change and acute circulating volume depletion; and plasma potassium concentration, which responds to potassium loading or sodium deficiency. Importantly, the autonomy is rarely absolute; for example, most aldosterone-producing adenomas are sensitive to ACTH, and a proportion are sensitive to angiotensin II. For validation of confirmatory testing, potassium repletion is similarly required, and the addition of dexamethasone recently was reported to refine saline suppression testing in the diagnosis of primary aldosteronism [1]. Parenthetically, ascription of “gold standard” status to the fludrocortisone suppression test may in fact reflect its potent glucocorticoid (as well as mineralocorticoid) activity [2] and its relatively low reflection constant at the blood–brain barrier [3].

Primary aldosteronism is diagnosed by confirmation/exclusion testing after screening, and the overwhelming majority of cases appear to be sporadic. Three different varieties of familial hyperaldosteronism (FH) are currently described, known as FH I, FH II, and FH III; the genetic bases of FH I and FH III have been established, with work continuing on the basis of FH II. In addition, various mouse models of primary aldosteronism have been reported, similarly on the basis of germline mutation. To date, however, most sporadic cases of primary aldosteronism appear not to reflect germline defects, although (as described below) a substantial minority of aldosterone-producing adenomas reflect mutation in KCNJ5 gene confined to the adrenal cortex.

Familial Hyperaldosteronism Type I

Initially termed glucocorticoid suppressible hypertension (GSH) or glucocorticoid remediable aldosteronism (GRA), FH I was solved in terms of molecular mechanisms in a landmark study by Lifton et al. [4]. The clinical picture is that of inappropriately high aldosterone levels in response to normal levels of ACTH; at the genomic level, this reflects the operation of a chimeric gene, combining a 5′ moiety of 11βhydroxylase (CYP 11B1) with a 3′ moiety of aldosterone synthase (CYP 11B2). The CYP 11B1 upstream sequence contains the fasciculata-specifying expression sequence, so that the chimeric gene is expressed throughout the fasciculata; in addition, it remains sensitive to ACTH, unlike the normal CYP 11B2 response to ACTH, which is glucocorticoid-suppressible in the short to medium term.

Diagnosis is by long polymerase chain reaction (PCR), which is mandatory for patients with a family history of primary aldosteronism and recommended in patients under the age of 40 or with a family history of cerebrovascular events at a young age. FH I is commonly held to be very rare, which may be a serious underestimate, given that about 99% of individuals with primary aldosteronism are never screened, let alone diagnosed [5]. Given this, plus the combination of small families and geographic mobility, it would be surprising if the true prevalence were not substantially higher.

Familial Hyperaldosteronism Type II

FH II is currently a diagnosis of exclusion, when related patients (with adrenal adenoma or bilateral hyperplasia) are shown to be negative for FH I. Linkage analysis of five cohorts from Australia, Italy, and South America suggests that the mutation(s) involved lie on chromosome 7p22 [6•]. Insertions and deletions of >50 kb appear to be excluded, as do five candidate genes in the locus, selected on the basis of their involvement in cell growth. The true prevalence of FH II may be substantially higher than it currently appears, for the same reasons as for FH I.

Familial Hyperaldosteronism Type III

FH III is the most recently described of the germline causes of primary aldosteronism [7]. The causative molecular mechanism has been shown to be a mutation in the gene coding for KCNJ5, a component of the Kir 3–4 potassium channel [8••]. The mutation in the index case and his two daughters is T158A, with the excess aldosterone production originating from grossly hyperplastic zona fasciculata cells, rather than cells of the zona glomerulosa. All three of the affected patients required bilateral total adrenalectomy to treat the condition. Not yet resolved is whether similar but not identical KCNJ5 germline mutations other than T158A can also produce FH III, and whether a spectrum of phenotypic severity may be seen in the condition, consistent with the clinical history (hypertension, myocardial infarction at age 34) of the index case’s father.

Recent in vitro studies on human adrenal cancer (HAC-15) cells transfected to overexpress the KCNJ5 T158A sequence have yielded puzzling results [9•]. As perhaps might be anticipated, the transfected cells produced fivefold higher levels of aldosterone than HAC cells transfected with empty vector, but counterintuitively showed lower levels of cell proliferation.

Autonomous Aldosterone Secretion: Insights from Animal Models

In the broadest sense, this section of the review might profitably cover the hundreds of studies in which exogenous aldosterone (and by extension deoxycorticosterone) has been administered to experimental animals and the effects documented. In terms of the insights they may provide for clinical primary aldosteronism, perhaps the most important is the absolute necessity of an inappropriate sodium status for the deleterious effects of aldosterone to be seen. In primary aldosteronism (for example in FH I patients with blood pressures within the normal range), even very modestly elevated aldosterone levels can be shown to be damaging [10]. Much higher aldosterone levels, as in chronic sodium deficiency, are homeostatic and have no deleterious effects on the cardiovascular system. The literature on mineralocorticoid/salt-induced experimental hypertension extends back over 70 years and is clearly beyond the scope of this review, which focuses on models of aldosterone excess consequent upon germline genetic manipulation.

TWIK-Related Acid-Sensitive K+-Containing Channels: TASK-1, TASK-3

Outside of the central nervous system, the adrenal cortex (in particular, the zona glomerulosa) is a major site of expression of both TASK-1 and TASK-3. TASK channel activity is inhibited by angiotensin II and acidosis, suggesting that they are responsible for background K+ conductance in glomerulosa cells, and that gene deletion studies may be illuminating as a model for primary aldosteronism. This is indeed the case [11], although such studies have provided more questions than answers.

When both TASK-1 and TASK-3 are deleted (hereafter TASK-/-), the mutant mice have elevated aldosterone levels (despite low renin levels), which are refractory to angiotensin receptor blockade or a high-salt diet. Their adrenal glands show normal zonation and histology, with typical distribution of steroidogenic enzymes (e.g., CYP 11B2 is confined to the glomerulosa). On these criteria, then, drugs that increase channel activity have been suggested [12] as a possible additional therapeutic strategy in primary aldosteronism. Selectivity may be an issue for such an approach, given the major roles for TASK channels in the heart and central nervous system.

If the compound TASK-/- phenotype appears relatively simple, the single TASK1-/- mouse is curiously complex. In young TASK-1-/- mice of either sex, aldosterone is overproduced from glomerulosa cells, not from those of the zona fasciculata [13]. The mice show the classic signs of hyperaldosteronism—suppressed renin, hypokalemia, hypertension—with two caveats. First, the hypersecretion of aldosterone and the hypertension can be corrected by dexamethasone administration, reminiscent of FH I but clearly mechanistically different. Second, and intriguingly, the disorders of zonation and aldosterone hypersecretion are reversed by androgens, so that after puberty, male mice revert to wild-type status, whereas female TASK-1-/- mice remain affected.

Preliminary studies on TASK-3 in mice show its localization uniquely in the zona glomerulosa in female mice [14] but in both glomerulosa and fasciculata in male mice. TASK-3-/- mice have an impaired response of aldosterone secretion to potassium loading; ex vivo, angiotensin II caused a paradoxic hyperpolarization rather than the classic depolarization seen in wild-type cells. Further studies are needed to dissect the gender differences and the interplay between channel subunits in terms both of zonation and steroid secretion.

Cryptochrome-1 and Cryptochrome-2: Cry Null Mice

In 2010, Doi et al. [15] reported that mice with both Cry genes rendered null showed the lack of circadian rhythmicity characteristic of clock-gene disruption, elevated levels of plasma aldosterone, and a hypertensive response to high salt. Hyperaldosteronism was elegantly shown to be due to increased expression of 3βhydroxysteroid dehydrogenase type VI (hsd3b6), which is uniquely expressed in the zona glomerulosa cells of the adrenal cortex and is negatively regulated by the cellular circadian clock. Classically, the rate-limiting steps in aldosterone biosynthesis have been recognized as side chain cleavage and aldosterone synthase; the demonstration by Doi et al. of aldosterone hypersecretion by Cry null mice would appear to add hsd3b6 to that category. The human homologue of mouse hsd3b6 is HSD3B1; its potential role, if any, in primary aldosteronism remains to be established, along with perhaps more likely roles in disorders associated with shift work, jet lag, and other circadian challenges.

β-Catenin: Constitutive Activation

In preliminary studies [16], mice with constitutive activation of β-catenin developed adrenal tumors, and knockdown of β-catenin levels in the adrenal cancer H295R cell line caused marked reductions in basal and angiotensin-stimulated aldosterone synthesis. On immunohistochemistry, abnormal cytoplasmic and/or nuclear accumulation of β-catenin was seen in 19 of 21 aldosterone-producing adenomas, leading to the claim that “these data…establish constitutive β-catenin activation as a prominent cause of primary aldosteronism.” [16] If this claim is supported by further evidence, it may point to a hitherto neglected role for Wnt/β-catenin signaling in the control of aldosterone biosynthesis, and thus possible involvement in the genesis of primary aldosteronism.

Somatic Mutations in the KCNJ5 Potassium Channel

In January 2011, Choi et al. [8••] published a groundbreaking study in which they described finding two different KCNJ5 mutations (G151R, L168R) in 8 of 22 aldosterone-producing adenomas. In a series of electrophysiologic studies, the authors showed that these mutations result in diminished channel selectivity, increased sodium conductance, and consequent membrane depolarization. Interestingly, the simultaneously characterized KCNJ5 mutation (T158A) underlying FH III showed a lesser degree of functional impairment, despite its florid phenotype. The G151R and L168R mutations are somatic, with none of the affected subjects showing the mutation in leukocytes. This paper is a major step in mapping the genetics of primary aldosteronism [17] and has prompted a number of subsequent studies, only one of which [18] is in press at this time.

In this paper, by Boulkroun et al. [18••], a total of 380 aldosterone-producing adenomas from the European Network for the Study of Adrenal Tumors (ENS@T) were screened for KCNJ5 mutations. The tumors came from centers in France, Italy, and Germany, in contrast with those in the initial study, which were from Sweden and the United States. Not surprisingly, as has previously been noted [19], there is a considerable degree of agreement between the two studies. In both studies, all patients with an adrenal KCNJ5 mutation have a normal sequence in leukocytes. In terms of incidence, one or the other mutant KCNJ5 sequence was found in 8 (36%) of 22 subjects by Choi et al. [8••], and in 129 (34%) of 380 subjects by Boulkroun et al. [18]; no different mutations were found in the second study, despite its much larger patient numbers. In both studies, there was a clear female preponderance: in the initial study, 7 of 16 female patients (44%) showed a mutation in KCNJ5, but only 1 of 6 of males (17%); in the second study, the corresponding figures were 49% and 19%. In both studies, KCNJ5 mutant adenomas were significantly more common in younger patients, and in both studies those with KCNJ5 mutations were at the more severely affected end of the spectrum—on the basis of higher aldosterone-to-renin ratios in the initial study [8••], and on the basis of higher plasma aldosterone levels in the study by Boulkroun et al. [18••].

There are, however, salient differences between these two published studies, however. The initial paper focused on large tumors, and in that relatively small group found KCNJ5 tumors to be smaller (mean diameter, 23 mm) than wild-type tumors (32 mm). In the ENS@T cohort, no difference was seen between wild-type (15 mm) and KCNJ5 mutant tumors (16 mm), but it should be noted in passing that even in this group, small (<10-mm) aldosterone-producing adenomas are probably underrepresented. Another major difference between the two studies is the distribution of the two mutant sequences: Choi et al. [8••] found two G151R and six L168R mutations, and Boulkroun et al. [18••] found 76 G15R and 53 L168R mutations.

In the year since the influential publication by Choi et al. [8••], a number of presentations on KCNJ5 mutations have been made, and a number of submitted manuscripts are now in revision. In the pipeline are studies from Japan, a major contributor to the area of primary aldosteronism; detailed histologic studies on KCNJ5 mutant and wild-type aldosterone-producing adenomas, and a clear distinction in terms of cells of origin; further studies of tumor transcriptomes, first addressed by Boulkroun et al. [18]; alignment of angiotensin/posture sensitivity with tumor genotype; and novel mutations in the KCNJ5 hotspot causing primary aldosteronism. Another area to be addressed is the KCNJ5 status of multiple adenomas in the same adrenal or in both adrenals, and whether bilateral adrenal hyperplasia may also involve similar mutations. Even in a cohort as carefully curated as the one studied by Boulkroun et al., a significant minority of patients (about 7%) appear to be unimproved by unilateral adrenalectomy [19], indicating a possible group in which the above questions may be addressed.

Conclusions

In many ways, this brief survey is perhaps a year too early, given the timelines of science and more particularly of publication; the ENS@T study was over 6 months from first submission to online publication. Those both interested and involved in the area—framing their own questions, designing their own studies, analyzing data from others—would be well advised to check PubMed each week over the next 6 to 12 months. Our understanding of the drivers of adrenal zonation and autonomous aldosterone secretion will be far better at the end of 2012 than at the beginning of the year. Inevitably (and mercifully), for every answer, two further questions will spring up: such are the dragon’s teeth of science.