Autosomal Dominant Pseudohypoaldosteronism Type 1: Mechanisms, Evidence for Neonatal Lethality, and Phenotypic Expression in Adults

Patients

All clinical studies were approved by the Yale Human Investigation Committee. Kindreds PHA15 and PHA16 have been described previously (5). Although not known to be related, these families live in the same village in northwestern Spain and were found to carry the identical R537X disease-causing mutation. We extended these pedigrees by recruiting relatives of known adPHA1 carriers. We prospectively screened individuals in these pedigrees for clinical indices of renin-angiotensin-aldosterone system activity, including BP, serum chemistries, and aldosterone levels. At the time of study, no individuals were taking sodium supplements or BP medications. They subsequently were genotyped for the presence of the R537X mutation. BP was measured in the seated position. Blood for serum chemistry and aldosterone levels was drawn in the seated position. Genomic DNA was prepared, and single-strand conformation polymorphism was performed as described previously (5).

The remaining patients studied were referred for genetic analysis by clinicians who had diagnosed their PHA1. A blood sample for DNA analysis was obtained from each index patient and their immediate family members after they provided informed consent, in accordance with institutional review board approved protocols, and genomic DNA was extracted as described previously (11). Sequence variants were detected as described previously (12). Briefly, coding exons and their flanking intronic regions of MR were amplified using previously described primers (5) and were analyzed using denaturing HPLC (DHPLC) as described previously (12). The identified mutations were not observed by DHPLC in 90 control individuals whose DNA had been referred to our laboratory for evaluation of other genetic diseases; these individuals were primarily American of Caucasian descent with a small number of Hispanic Americans as well. RNA was isolated, and cDNA was prepared from patients’ leukocytes as described previously (13). cDNA was prepared and sequenced by standard methods.

Confirmation of Parental Identity

Biologic parenthood was confirmed by genotyping both parents and affected children with at least 10 unlinked highly polymorphic genetic markers. In each case, the child was found to share one allele with each parent at each marker locus, a result expected to occur by chance with a likelihood of approximately 1.4 × 10⁻⁵ if one of the reported parents were in fact biologically unrelated.

Transcriptional Activation Studies

Cos-7 cells (ATCC, Manassas, VA) were cultured in DMEM (Life Technologies-BRL, Grand Island, NY) supplemented with 10% heat-inactivated FCS. On the night before transfection, cells were plated at 2 × 10⁵ cells/well. Cells in each well were transfected with 1 μg of pMTV-Luc, 2 μg of MR receptor plasmid, and 5 ng of pSV2, using cationic liposomes (Lipofectamine-2000; Life Technologies), after which cells were incubated in DMEM supplemented with 10% FCS overnight. Cells were washed, and then serum-free DMEM that contained aldosterone was added. The cells were incubated for an additional 16 h before assay. Luciferase and β-galactosidase activities were measured as described previously (14). Luciferase activity was normalized to β-galactosidase activity to correct for transfection efficiency and is expressed as a percentage of the wild-type receptor activity at 1 nM aldosterone. All results are the mean of at least nine independent transfections. Statistical comparisons are by the Mann-Whitney U test.

Novel adPHA1-Causing Mutations in MR

We previously demonstrated that adPHA1 is caused by heterozygous mutations in MR. To gain further insight into disease pathogenesis, we identified further patients with PHA1. Clinical features for kindred PHA30 have been described previously (13). Clinical data for other patients in whom MR mutations were identified are summarized in Table 1. Each index case presented in the first month of life with failure to thrive, hyponatremia, and hyperkalemia despite markedly elevated renin and aldosterone levels. Each improved rapidly after the administration of sodium supplementation and resumed normal growth and development. Patient PHA67-1 was remarkable in that, in addition to the PHA1 phenotype, she presented with microcephaly and neurodevelopmental delay; her mother, who was hospitalized as an infant for severe feeding difficulties and vomiting, also has mental retardation with microcephaly.

Coding exons in MR were screened with DHPLC, and identified variants were sequenced directly. In this way, MR were identified in each PHA1 kindred (Figure 1). Patient PHA51-1 is heterozygous for a single base pair insertion in exon 2 that converts Tyr503 to an immediate stop codon (Figure 1A). Patient PHA72-1 (Figure 1B), like affected individuals from kindred PHA30 (data not shown), is heterozygous for a nonsense mutation in exon 3 that alters R590 to an immediate stop codon. The index case in kindred PHA67 and her mother are heterozygous for a nonsense G→T substitution at the final base pair in exon 5 (Figure 1C); although the expected effect of this mutation is conversion of the codon GGA coding for G589 to an immediate TGA stop codon, sequencing of MR cDNA from the index patient and her mother revealed that the actual effect of this mutation is to disrupt normal exon 5 splicing and unveil a cryptic splice site 10 bp 5′ to the normal exon 5 splice donor site, resulting in a frameshift (data not shown). Patient PHA82-1 is heterozygous for a single nonsense mutation in exon 9 that converts Q967 to an immediate stop codon (Figure 1D); this mutation is predicted to truncate helix 12, a crucial domain for steroid hormone receptor function. Affected individuals in kindred PHA76 are heterozygous for an insertion of alanine at nucleotide 2681 that results in a frameshift after H821 and a stop codon 14 amino acids later (Figure 1E). Finally, patient PHA47-1 is heterozygous for a missense mutation in exon 6 that converts serine 818 to leucine (Figure 1F). This mutation lies in the center of the hormone-binding domain of the receptor; to determine whether mutation had a functional effect on the activity of the receptor, we tested the ability of this receptor to activate transcription of a reporter gene in response to aldosterone (see Materials and Methods). In contrast to the wild-type receptor, which showed normal activation in the presence of aldosterone, we saw no activity of the S818L mutant receptor, indicating that this mutation disrupts MR activity, thereby identifying codon 818 as an essential residue for normal receptor function (Figure 1G).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Novel mineralocorticoid receptor (MR) mutations in patients with autosomal dominant pseudohypoaldosteronism type 1 (adPHA1). (A) The index case in kindred PHA51 is heterozygous for a single adenine insertion that converts Tyr503 to a TAA stop codon and results in a frameshift thereafter. Neither parent carries this mutation. (B) The index case in kindred PHA72 and his mother are heterozygous for a C→T nonsense mutation in exon 3 that converts Arg590 to a stop codon. (C) The index case in kindred PHA67 and her mother are heterozygous for a G→T nonsense mutation at the final base pair in exon 5. Neither maternal grandparent carries this mutation. The splice donor site at the end of exon 5 is underlined. (D) The index case in kindred PHA82 is heterozygous for a C→T nonsense mutation that converts Gln967 to a stop codon. Neither parent carries this mutation. (E) Affected individuals in kindred PHA76-1 are heterozygous for an insertion of an alanine at nucleotide 2681 that results in a frameshift after H821 and a stop codon 14 amino acids later. (F) The index case in kindred PHA47 is heterozygous for a C to T substitution in exon 6 that converts serine 818 to leucine. Neither parent carries this mutation. (G) Transactivation analysis of MR_L818 in the presence of rising concentrations of aldosterone. MR_WT is normally activated by aldosterone, but MR_L818 has no significant transactivational activity in the presence of aldosterone.

The identified mutations are summarized in Table 2. Each of these mutations was not identified in 180 control chromosomes, indicating that these are not common polymorphisms. Moreover, four of the six mutations are de novo mutations, as they were absent in the parents of affected individuals. To confirm that these were in fact the biologic parents of the affected patients, we performed genotypic analysis of the parents of the index cases in kindreds PHA47, PHA51, and PHA82 and of the parents of the affected mother in kindred PHA67 using 10 or more unlinked short tandem report markers (see Materials and Methods). In each case, parentage was confirmed, indicating that the mutations in their children are true de novo mutations. The identification of de novo mutations in the affected children of unaffected parents in a gene that previously was linked to the disease offers conclusive evidence that these mutations are disease causing.

Haploinsufficiency of MR Is Sufficient to Cause the adPHA1 Phenotype

PHA30 is an adPHA1 kindred that has been well described previously (13). The kindred is composed of nine affected individuals in three generations with a diagnosis of adPHA1 on the basis of findings of salt wasting and hyperkalemic metabolic acidosis despite markedly elevated serum aldosterone levels (Figure 2A). As noted above, analysis of genomic DNA in this kindred identified a heterozygous C→T substitution that converts R590 to a stop codon in all affected individuals. We gained insight into the underlying mechanisms of adPHA1 through further studies of kindred PHA30. We sequenced cDNA that was prepared from RNA that was freshly isolated from peripheral blood lymphocytes from patient PHA30-1 and compared this with genomic sequence. As expected, genomic DNA was heterozygous for the C1984T substitution. In contrast, the cDNA sequence showed evidence of only the wild-type cytosine (Figure 2B). Although it would have been interesting to reconfirm this finding via analysis of mRNA from an epithelial tissue, such as the skin or salivary glands, from this patient, we did not have access to such tissue. Nevertheless, the absence of the mutant mRNA in lymphocytes suggests that it is degraded by cellular mechanisms that prevent translation of mRNA with stop codons outside the last exon via nonsense mediated decay (15); we therefore infer that MR haploinsufficiency is sufficient to cause the adPHA1 phenotype in this kindred.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Haploinsufficiency is sufficient to cause the adPHA1 phenotype. (A) Structure of the PHA30 pedigree. Affection status was determined via demonstration of the Arg590X mutation, and affected individuals are shown in black. (B) Haploinsufficiency causes the adPHA1 phenotype. Sequence of exon 3 genomic DNA and cDNA in patient 30-1. cDNA was prepared from lymphocyte RNA and amplified using a forward primer from exon 2 and a reverse primer in exon 3. Genomic DNA reveals the presence of a heterozygous C→T substitution that alters CGA (Arg) to TGA (Stop). cDNA reveals only the presence of the wild-type C allele.

Genotype–Phenotype Correlation in adPHA1

The determination that MR haploinsufficiency causes the adPHA1 phenotype suggests that MR receptor quantity can have important effects on renal sodium handling, particularly in the neonatal period. However, there are few data on the effect of MR deficiency in adults. The ability to assign unambiguously affection status on the basis of genotype allowed us to perform genotype–phenotype correlation studies in extended adPHA1 kindreds to gain insight into this question. Kindreds PHA15 and PHA16 have been described previously (5); although not known to be related, these families live in the same village in northwestern Spain and carry the identical R537X mutation in MR. We extended the pedigrees by recruiting relatives of known adPHA1 carriers (see Materials and Methods). In total, 35 family members were studied; 14 individuals carried the mutation, and 21 individuals did not (Figure 3).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Co-segregation of elevated aldosterone levels with the Arg537 mutation in kindreds PHA15 and PHA16. Genotypically affected individuals are indicated in black. Serum aldosterone levels (seated, ng/dl) are indicated beneath each individual in the pedigree. The single-strand conformation polymorphism genotype of each individual for MR exon 2 is shown beneath each living individual in the pedigree. The arrow indicates the position of the Arg537X allele. All individuals with aldosterone levels >30 are heterozygous for the Arg537X allele; all those with aldosterone levels <30 are homozygous for the wild-type allele.

In kindred PHA15, only the proband had been symptomatic with adPHA1, having presented at 8 mo of age with a clinical history of anorexia, vomiting, salt wasting, and failure to thrive. At the time, his weight and height were below the 3rd percentile, and he was started on 3 g/d Na⁺. He remained at the 3rd percentile for 1 yr and then exhibited catch-up growth, reaching the 25th percentile by age 3 and the 50th percentile by age 16 off sodium supplementation. The remaining genotypically affected individuals in PHA15 were asymptomatic.

The proband in kindred PHA16 presented at age 4 mo with signs of dehydration. Her younger siblings each received a diagnosis of adPHA1 as neonates because of family history and poor growth. The remaining five genotypically affected individuals in the pedigree were asymptomatic. Importantly, four at-risk children in kindred PHA16 died at an early age; one child died of burns at the age of 3 yr, and three other children (ages 1 yr, 15 d, and 8 d, respectively) died of unknown causes. No tissue was available for genotypic evaluation of these infants.

We studied various clinical indices of aldosterone activity in affected and unaffected adults in these kindreds (Figure 4). We found no significant difference in systolic or diastolic BP, serum sodium, serum potassium, fractional excretion of sodium, or transtubular potassium gradient between affected individuals and their unaffected relatives. The only biochemical marker of disease was a dramatically elevated serum aldosterone level (Figure 4). Mean serum aldosterone levels were 126 ± 31.2 ng/dl in affected individuals, compared with 8.8 ± 1.4 ng/dl in unaffected individuals (P < 0.0001). Random seated serum aldosterone levels proved to be an excellent marker for affection status: Individuals with aldosterone levels >30 ng/dl carry the R537X mutation, whereas individuals with aldosterone levels <30 ng/dl are genotypically unaffected.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Genotype–phenotype correlations in adPHA1. Clinical characterization of kindreds PHA15 and PHA16. Individuals who carry the Arg537X mutation (○) were compared with their unaffected relatives (□) for clinical indices of aldosterone effect, including systolic BP (SBP; mmHg), diastolic BP (DBP; mmHg), serum Na (mEq/L), serum K (mEq/L), fractional excretion of sodium (FENa), and serum aldosterone (ng/dl). Individual data points are shown, and the means of the recordings for affected and unaffected individuals (±SEM) are indicated by • and ▪, respectively. There was no significant difference between affected and unaffected family members for any of these indices, except for serum aldosterone levels, for which the difference was highly significant (P < 0.001).

Haploinsufficiency for MR Is Sufficient to Cause the adPHA1 Phenotype

Many human diseases that are caused by loss-of-function mutations in nuclear receptors have been described, but the mechanism of disease differs for each receptor. For example, whereas glucocorticoid receptor (NR3C1) mutations cause familial cortisol resistance (OMIM #138040) via receptor haploinsufficiency (18,19), mutations in the thyroid hormone receptor cause generalized thyroid hormone resistance (OMIM #188570) via a dominant-negative mechanism; the mutant receptor interferes with wild-type receptor by forming a transcriptionally inactive dimer (20,21). We found wild-type but not mutant MR mRNA in peripheral blood lymphocytes of a patient with adPHA1. The absence of mutant RNA likely is due to nonsense-mediated decay, a process by which mRNA that bear premature stop codons in nonterminal exons are degraded by the cell (15). This suggests that patients who have adPHA1 and bear premature stop codons in MR, the majority of patients described, are also haploinsufficient. Although we cannot rule out other mechanisms of action by which other MR mutations cause adPHA1 in other families, it is clear that haploinsufficiency is sufficient to cause the adPHA1 phenotype. Furthermore, the finding that MR haploinsufficiency can cause a measurable salt-wasting phenotype suggests that early in life, the kidney cannot upregulate expression of the wild-type MR allele sufficiently to normalize sodium balance and that mineralocorticoid activity is limited in part by receptor levels.

Genotype–Phenotype Correlation in adPHA1

The MR has been ascribed a wide variety of physiologic roles. The study of individuals who lack one of two functional copies of MR provides a unique opportunity to study its role in human physiology. This study represents the first detailed physiologic analysis of patients with adPHA1 confirmed to have a disease-causing mutation in MR. By performing genotype–phenotype correlation studies in extended adPHA1 kindreds, we have gained insight into MR’s role(s) in human physiology. A number of conclusions regarding MR can be drawn from our findings.

adPHA1: A Fatal Neonatal Disorder?

We found striking phenotypic diversity in patients who had adPHA1 that was caused by MR mutation. adPHA1 generally is regarded to be a mild disease, and consistent with this, some of the adult patients whom we studied were entirely asymptomatic with no recollection of significant illness as infants. Conversely, the probands each were hospitalized for salt wasting, growth retardation, and failure to thrive. It is not clear why some infants come to clinical attention and others do not, but we have previously suggested that infants with intercurrent illness that impairs oral intake or produces diarrhea may be more likely to become symptomatic (5). Alternatively, it may be that most infants with adPHA1 do present, and only infants with genetic or environmental advantages survive without active intervention.

It is noteworthy that four of the patients reported here as well as one patient reported previously carry confirmed de novo mutations (5). Furthermore, two previously described patients with adPHA1 have MR mutations that are not carried by either parent, although formal genotyping to prove parental identity was not performed (7,8). Therefore, up to seven of the 22 MR mutations that have been reported may have arisen de novo. The high incidence of de novo mutations suggests impaired reproductive fitness in patients with adPHA1. Analysis of the few extended adPHA1 pedigrees available suggests no obvious impairment of fertility or failure to transmit the disease allele (Figures 1 and 3) (5,6,8–10), suggesting that fertility concerns do not account for the observed impairment in disease transmission. In contrast, we observed a number of deaths in infants who were at risk for adPHA1 (Figure 3B), a finding observed by others as well (22). Although we have no genotypic information on the deceased infants, the frequency of unexplained neonatal death in these kindreds suggests that adPHA1 is potentially lethal in neonates, and we infer that the impaired reproductive fitness most likely is due to neonatal death of these children. We believe that this inference, based on available but admittedly incomplete information, is compelling enough to recommend the prophylactic provision of salt to at-risk infants until a definitive diagnosis can be made. Furthermore, if this hypothesis is correct, we believe that in the modern medical era, more such children will survive to procreate; therefore, the incidence of the disease may well increase, forcing neonatologists to be aware of the potential perinatal dangers. The clinical severity of disease in patients with adPHA1 early in life highlights the essential role of aldosterone-sensitive sodium transport in the neonate.

adPHA1 in Adults

The findings presented also give insight into renin-angiotensin system in adult physiology. Despite an intrinsic salt-wasting predisposition, adults with adPHA1 are clinically indistinguishable from their unaffected relatives. We found no significant differences between these groups with regard to systolic BP, diastolic BP, serum sodium, serum potassium, urine sodium, and urine potassium. These patients maintain normal volume homeostasis via an approximately 14-fold elevation in serum aldosterone levels, suggesting that homeostatic mechanisms that regulate sodium balance remain intact in these individuals at the expense of a marked increase in aldosterone levels. Whether these individuals would struggle to cope with salt deprivation or severe volume depletion is not known.

Angiotensin II has been proposed as a key molecule in the pathogenesis of hypertension, left ventricular hypertrophy, and myocardial fibrosis (23,24). Our patients had markedly elevated renin levels and presumably angiotensin II levels contributing to hyperaldosteronemia, yet we found no evidence of hypertension or congestive heart failure in affected individuals. This suggests that the effect of angiotensin II on long-term BP regulation and on myocardium is at least partially if not completely load dependent and that in the absence of aldosterone-induced volume expansion, no significant hypertension or heart failure is seen. Similarly, much has been written about nongenomic effects of aldosterone, effects of aldosterone not mediated via MR (25). The absence of significant cardiovascular findings in our patients with adPHA1, in whom serum aldosterone levels have been elevated consistently throughout their lives, suggests that nongenomic effects of aldosterone do not lead to clinically relevant cardiovascular sequelae either. We cannot exclude, however, that nongenomic effects of aldosterone affect short-term cardiovascular responses.