Report

Activating Mineralocorticoid Receptor Mutation in Hypertension Exacerbated by Pregnancy

See allHide authors and affiliations

Science  07 Jul 2000:
Vol. 289, Issue 5476, pp. 119-123
DOI: 10.1126/science.289.5476.119

Abstract

Hypertension and pregnancy-related hypertension are major public health problems of largely unknown causes. We describe a mutation in the mineralocorticoid receptor (MR), S810L, that causes early-onset hypertension that is markedly exacerbated in pregnancy. This mutation results in constitutive MR activity and alters receptor specificity, with progesterone and other steroids lacking 21-hydroxyl groups, normally MR antagonists, becoming potent agonists. Structural and biochemical studies indicate that the mutation results in the gain of a van der Waals interaction between helix 5 and helix 3 that substitutes for interaction of the steroid 21-hydroxyl group with helix 3 in the wild-type receptor. This helix 5–helix 3 interaction is highly conserved among diverse nuclear hormone receptors, suggesting its general role in receptor activation.

Although blood pressure is normally reduced throughout gestation, about 6% of pregnancies are complicated by the development of hypertension, raising the risk of pre-eclampsia, a hypertensive disorder of pregnancy that increases maternal and perinatal mortality (1). The factors responsible for these developments are unknown; however, the prompt resolution of many cases with delivery suggests pregnancy-specific factors.

Mutations that change renal salt reabsorption alter blood pressure (2). For example, heterozygous loss-of-function mutations in the mineralocorticoid receptor (MR; locus symbol NR3C2), a member of the nuclear receptor family, cause pseudohypoaldosteronism type 1 (PHA1), a disease featuring salt wasting and hypotension (3). Normally, activation of MR by the steroid hormone aldosterone raises renal salt reabsorption by increasing activity of the epithelial sodium channel of the distal nephron.

To determine if gain-of-function mutations in MR could cause increased renal salt reabsorption and hypertension, we screenedMR in 75 patients with early onset of severe hypertension (4). A 15-year-old boy with severe hypertension, suppressed plasma renin activity, low serum aldosterone, and no other underlying cause of hypertension was heterozygous for a missense mutation, resulting in substitution of leucine for serine at codon 810 (Fig. 1A). The S810L mutation lies in the MR hormone-binding domain (HBD), altering an amino acid that is conserved in all MRs from Xenopus to human (5) but not found in other nuclear receptors (Fig. 1B). S810L was not detected in 160 control chromosomes.

Figure 1

MRL810cosegregates with early hypertension in kindred AMR 503. (A) Sequence of a novel MR variant identified in K503. The DNA sequence corresponding to the sense strand of codons 808 to 812 of the wild-type (left) and mutant MR (right) are shown (4). The asterisks indicate the position of a C→T substitution, which changes codon 810 in the HBD from serine to leucine. (B) Sequence alignment of steroid receptors. Corresponding segments of the HBDs of the mineralocorticoid (MR), progesterone (PR), glucocorticoid (GR), androgen (AR), and estrogen α (ERα) receptors are shown. MR sequences are fromHomo sapiens (h), the rat Rattus norvegicus (r), the toad Xenopus laevis (x), and the tree shrew Tupaia belangeri (t) (5). S810 is conserved among all known MRs, yet is not seen in other members of the steroid hormone receptor family (5). (C) Cosegregation of early hypertension and MRL810 in kindred 503. Family members diagnosed with hypertension before the age of 20 are shown as filled symbols and those without this trait are shown as unfilled symbols. Individuals of unknown phenotypic status are shown as shaded symbols. The index case is indicated by the black arrow. The SSCP genotype of exon 6 of MR (4) is shown below the symbol corresponding to the appropriate kindred members; the positions of the S810 and L810 variants are indicated by open arrows.

The impact of MR bearing S810L (MRL810) was assessed by clinical and biochemical studies. Twenty-three relatives of the proband were evaluated. Remarkably, 11 had been diagnosed with severe hypertension before age 20, a rare trait in the general population (6), whereas the remainder had unremarkable blood pressures (Fig. 1C). These findings suggest Mendelian segregation of hypertension. MRL810 precisely cosegregated with early-onset hypertension in this family (Fig. 1C), providing strong evidence of linkage. The maximum lod score (logarithm of the odds ratio for linkage) was 5.24 at a recombination fraction of zero (odds of 174,000:1 in favor of linkage) (7).

Comparison of the clinical features of MRL810 carriers and noncarriers revealed a marked increase of blood pressure among carriers even though they were taking antihypertensive medication, as well as suppression of aldosterone secretion (Table 1). There was a nonsignificant trend toward lower serum potassium among carriers, and there were no significant effects of gender or age on phenotypic expression of MRL810. Of note, three deceased pedigree members with early-onset hypertension all died of heart failure before age 50.

Table 1

Clinical features of MRL810 carriers (+) and noncarriers (−) in kindred 503. All values are the mean ± SEM. Statistical comparisons were assessed by the Mann-Whitney U test. Normal values: SBP < 140 mmHg; DBP < 90 mmHg; serum K+: 3.8 to 5.2 mM; serum HCO3 : 23 to 30 mM; serum aldosterone 4 to 15 (ng/dl); and urinary aldosterone: 6 to 31 μg/24 hours. Urinary aldosterone was measured in three MRL810 + and four MRL810 individuals. Abbreviations: HTN, hypertension; SBP, systolic blood pressure; DBP, diastolic blood pressure; HCO3 , bicarbonate.

View this table:

To investigate the functional effects of the S810L mutation, we expressed wild-type MR (MRWT) or MRL810 in Cos-7 (African green monkey kidney) cells and measured induction of luciferase under control of the MR-sensitive mouse mammary tumor virus (MMTV) promoter (8). The activities of MRWT and MRL810 in response to aldosterone were indistinguishable (Fig. 2A). However, in the absence of added steroid, MRL810 showed 27% of maximal activity (P < 0.001 versus MRWT), indicating constitutive activity of the apo-receptor or activation by endogenous compounds (Fig. 2A). Constitutive MRL810 activity was also seen in HeLa cells. This constitutive activity may account for or contribute to the hypertension seen in patients harboring MRL810.

Figure 2

Transcriptional activation by MRWTand MRL810. The ability of MRWTand MRL810 to induce luciferase expressed under control of the MMTV promoter was assessed in Cos-7 cells in the absence or presence of the indicated steroids (8). Luciferase activity is expressed as the percentage of maximal induction of MRWTby aldosterone. In (B), (C), (E), and (F), all steroids were assayed at 1 nM. The Mann-Whitney U test was used to compare the significance of differences between groups (in all cases, similar significance was obtained with the two-tailed Student's t test). (A) Dose-response curve for induction of luciferase by MRWT and MRL810 in response to aldosterone (*P < 0.001 MRL810 versus MRWT). (B) Luciferase induction in response to 21-hydroxylated steroids (DOC, 11-deoxycorticosterone; *P < 0.001 MRL810versus MRWT). (C) Luciferase induction in response to progesterone derivatives (*P < 0.001 versus vehicle control). (D) Dose-response curve for induction of luciferase by MRWT and MRL810 in response to progesterone (*P < 0.001 versus vehicle control). (E) Luciferase induction in response to spironolactone (*P < 0.001 versus vehicle). (F) Progesterone-induced luciferase activity in cells coexpressing varying ratios of MRWT and MRL810. Cells were transfected with a constant amount of receptor plasmid (2 μg), with the proportion of MRWT and MRL810 varying as indicated. Luciferase induction in response to progesterone (▴) is shown and compared to theoretical curves for completely active (○) and completely inactive (•) heterodimers.

We next tested MR activation by a variety of steroids. Steroids bearing 21-hydroxyl groups, such as aldosterone, bind and activate MRWT (9) and similarly activate MRL810 (Fig. 2B). Steroids with 17-keto groups, including estradiol and testosterone, activate neither MRWT nor MRL810. Twenty-one–carbon steroids with neither modification, such as progesterone, normally act as antagonists, binding but not activating MRWT (9). In contrast to their action on MRWT , these steroids were all potent activators of MRL810 (Fig. 2C). For example, the dose-response curve for activation of MRL810 by progesterone is indistinguishable from that seen with MRWTand aldosterone (Fig. 2, A and D). Thus, activation of MRs with the S810L mutation no longer requires a steroid 21-hydroxyl group. Similarly, another MR antagonist, the clinically used drug spironolactone, is also a potent agonist of MRL810 (Fig. 2E); we infer that this medication is contraindicated in MRL810 carriers.

MR activates transcription as a dimer (10). Because MRL810 is heterozygous in affected individuals, we tested progesterone-mediated MR activation in the presence of varying ratios of MRWT and MRL810 (Fig. 2F). Activation was intermediate between that expected for completely active and completely inactive heterodimers, consistent with partial activation of heterodimers.

Progesterone levels normally increase 100-fold in pregnancy, reaching concentrations of 500 nM (11); this suggests that females with MRL810 might develop severe hypertension in pregnancy. Two MRL810 carriers have undergone five pregnancies; all have been complicated by marked exacerbation of hypertension. For example, blood pressure decreased early in the first pregnancy of one carrier, but then rose dramatically, reaching 170/130 mm Hg at 28 weeks despite antihypertensive therapy. This was accompanied by development of low serum potassium with marked renal potassium wasting. Aldosterone levels, which normally increase 10-fold in pregnancy, were undetectable (12). Of note, there were no proteinuria, edema, or neurologic changes, excluding pre-eclampsia. Because of still worsening blood pressure (210/120 mmHg), at 34 weeks a Caesarian section was performed, and the patient gave birth to a healthy son. Two subsequent pregnancies followed similar courses, and the patient was advised to avoid further pregnancy. The other carrier's two pregnancies were both complicated by severe exacerbation of hypertension, precipitating delivery in the sixth and seventh months of gestation, also with advice to avoid further pregnancy. These findings strongly support progesterone's in vivo agonism of MRL810.

The structures of the HBDs of steroid receptors are highly similar (13). The MR and progesterone receptor (PR) differ by only 3 of 18 residues lining the hormone-binding cavity (13,14), permitting modeling of the MR HBD on the basis of the known structure of PR (15) in order to generate testable hypotheses about the mechanism underlying the S810L mutation. In the resulting MRL810 model, the L810 side chain in helix 5 projects into the ligand-binding cavity, potentially forming van der Waals interactions with both A773 of helix 3 and the C19 methyl group of the steroid (Fig. 3A). In contrast, in the model of MRWT, S810 interacts with only the C19 methyl group (Fig. 3B). Of note, the steroid 21-hydroxyl group required for activation of MRWT interacts with N770 in helix 3 (14).

Figure 3

Helix 3–helix 5 interaction in progesterone-mediated activation of MRL810. (A) Structural model of a portion of the HBD of MRL810 bound to aldosterone. Based on the crystal structure of the progesterone receptor (13), a model of the MR LBD was created by substituting MR-specific residues in the ligand-binding cavity for their corresponding residues in PR (15). The side chain of L810 lies in sufficiently close proximity to A773 and the C19 methyl group of the steroid to form van der Waals interactions. (B) Model of MRWT. The side chain of S810 does not interact with A773. (C) Activity of MRs with various amino acid substitutions at residues 810 and 773. Mutant receptors containing the indicated substitutions at positions 810 in helix 5 and 773 in helix 3 were tested for their ability to induce luciferase activity in the presence of 1 nM aldosterone or 19-NP. The length of each side chain is approximated. Each data point represents the mean of nine independent transfections and is expressed as the mean ± SEM of the percentage of luciferase induction by MRWT in response to aldosterone. Abbreviations for the amino acid residues are as follows: A, Ala; G, Gly; L, Leu; M, Met; S, Ser; and V, Val.

We tested the importance of these interactions by modifying both the steroid and the receptor. The progesterone derivative 19-norprogesterone (19-NP), lacking the C19 methyl group, fully activated MRL810, indicating that the L810–C19 methyl interaction is not required for MRL810 activation (Fig. 2C). Substitution of methionine, with a longer side chain, at position 810 retained activation by 19-NP, whereas substitution of amino acids with shorter side chains such as valine or alanine led to progressive loss of receptor activation by 19-NP, but not aldosterone (Fig. 3C). That this effect involves interaction with A773 in helix 3 is supported by substitution at this position. MR with a V810-G773 pair showed no activation by 19-NP, but activation was restored in the V810-A773 pair (Fig. 3C). This second-site complementation supports the importance of helix 3–helix 5 interaction in the activation of MRL810 by steroids lacking 21-hydroxyl groups.

These observations suggest that progesterone binding to MRL810 results in an additional van der Waals interaction; this should increase receptor affinity for progesterone. MRWT and MRL810 show indistinguishable dissociation constants for aldosterone (K d= 1.10 ± 0.09 nM versus 1.06 ± 0.12 nM, respectively) (Fig. 4A) (16). In contrast, progesterone, 19-NP, and spironolactone all show increased competition with aldosterone for binding to MRL810, indicating substantially higher affinity for the mutant receptor (Fig. 4, B to D). For example, the concentration of spironolactone required to compete 50% of aldosterone binding was 10-fold lower for MRL810than for MRWT (Fig. 4D).

Figure 4

MRL810 has increased affinity for progesterone, 19-NP, and spironolactone. (A) Scatchard analysis of the binding of [3H]aldosterone in extracts expressing MRWT (▪) or MRL810 (□). Each data point was assayed in quadruplicate; one representative set of data for binding to MRWT and MRL810 is shown, and the mean K d's are indicated. (B toD) Competition of varying concentrations of the indicated unlabeled steroids and 5 nM [3H]aldosterone for binding to MRWT and MRL810 in Cos-7 cell extracts. Each data point represents the mean ± SEM of at least three independent experiments. (*) Significant differences between binding of MRWT and MRL810; all differences are significant at P < 0.001.

The finding of a helix 5–helix 3 interaction critical for MRL810 activation by progesterone led us to examine other nuclear steroid receptors for similar interactions. Intriguingly, the residues corresponding to MR amino acids 810 and 773 that are found in PR (M759 and G722) (13), the estrogen receptor (L387 and A350) (17), and even the nonsteroid retinoid X receptor (RXR) (L314 and A277) (18) all show van der Waals interactions in their respective crystal structures. Indeed, all steroid hormone receptors and RXRs have either methionine-glycine or leucine-alanine pairs at the corresponding helix 5 and helix 3 positions, respectively (5, 12). Only MRWT violates this rule, and the S810L mutation restores this interaction (Fig. 3B). The conservation of this interaction and its role in MRL810 activation by progesterone suggest its general significance in nuclear receptor activation.

Together, these findings define a new Mendelian form of hypertension featuring marked exacerbation in pregnancy. The mutation demonstrates the ability of a single amino acid substitution to dramatically alter nuclear receptor ligand selectivity and activation. It will be of interest to determine the effects of the S810L mutation on the subcellular distribution of MR and on binding of coactivators and corepressors.

Normal pregnancy is characterized by protean physiological changes. Efforts to identify novel endocrine factors in pregnancy-related hypertension and pre-eclampsia have been unsuccessful (1). The demonstration that one form of pregnancy-related hypertension is attributable to aberrant action of a normal pregnancy-related hormone raises the question of whether similar mechanisms might underlie other forms. It is noteworthy that a substantial fraction of women with pregnancy-related hypertension and pre-eclampsia show suppression of the renin-aldosterone axis like that seen in this kindred.

The mechanism of early-onset hypertension in men in this kindred is unclear, because levels of free progesterone in males are normally low (14). This observation suggests either that the observed constitutive receptor activation is sufficient to cause hypertension or that another steroid acts as an MRL810agonist. Interestingly, free levels of 17-hydroxyprogesterone (17-HP), an agonist of MRL810(Fig. 2C), rival those of aldosterone in men (11), suggesting that gain of 17-HP agonist activity may contribute to hypertension in men with MRL810.

Activation of MRWT requires a steroid 21-hydroxyl group. How does MRL810 dispense with this requirement? Ligand binding induces a number of conformational changes, including the bending of helix 3, which in conjunction with helix 12 forms a surface for coactivator binding (10). We propose that the conserved helix 5–helix 3 interaction is important for helix 3 bending and receptor activation. In MR, the inability of S810 to interact with A773 suggests that helix 3 bending is achieved by other means. Biochemical studies indicate that N770 interacts with the 21-hydroxyl group of aldosterone (14). We propose that this steroid–helix 3 interaction achieves the bending of helix 3 necessary for receptor activation. That steroids lacking a 21-hydroxyl group bind but cannot activate MRWT supports the key role of this interaction. The demonstration that the steroid 21-hydroxyl group becomes dispensable when the helix 5–helix 3 interaction is re-created in MRL810 provides evidence that these two interactions are interchangeable for receptor activation. The phenotype of patients harboring this mutation indicates the biological significance of this effect in vivo. Interference with this interaction could provide a general approach to receptor antagonism.

Why MRL810 shows constitutive activity, whereas other nuclear receptors with similar helix 5–helix 3 interactions do not, is of interest. One possibility is that formation of an active conformation is aided by an endogenous ligand. In this regard, constitutive activity of an RXR mutant has recently been attributed to binding by an endogenous fatty acid (18).

Finally, although the distal nephron is recognized as the major site of action of mineralocorticoids, expression of MR in hippocampus, heart, and endothelium has suggested extrarenal activity (19). The early development of congestive heart failure in this kindred raises the question of whether MRL810 might contribute to additional clinical effects due to expression in other tissues.

  • * Deceased.

  • To whom correspondence should be addressed. E-mail: richard.lifton{at}yale.edu

REFERENCES AND NOTES

View Abstract

Navigate This Article