| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Special Features |
Pediatric and Reproductive Endocrinology Branch, National Institute of Child Health and Human Development (E.C., S.L.M., M.F.K., G.P.C., D.P.M.), The Warren Grant Magnuson Clinical Center (R.W., D.P.M.), and Clinical Neurocardiology Section, National Institute of Neurological Disorders and Stroke (G.E.), National Institutes of Health, Bethesda, Maryland 20892; and Department of Pediatrics, New York Presbyterian Hospital, Cornell Medical Center (A.C., M.I.N.), New York, New York 10021
Address all correspondence and requests for reprints to: Dr. Evangelia Charmandari, Pediatric and Reproductive Endocrinology Branch, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20892.
Abstract
Classic congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency is characterized by decreased synthesis of glucocorticoids and mineralocorticoids, adrenal hyperandrogenism, and impaired development and function of the adrenal medulla. Although genotype can usually predict phenotype, genotype-phenotype discordance has been described. We investigated the association between adrenomedullary function, disease severity, and genotype in 37 children [22 males and 15 females; age range, 4.7–14.9 yr; 28 salt-wasting (SW) and 9 simple virilizing (SV) CAH] with classic 21-hydroxylase deficiency. Plasma and 24-h urinary catecholamines and their metabolites, and the 21-hydroxylase genotype were determined in all patients. The disease-causing mutations were divided into 4 groups (Null, A, B, and C) according to in vitro 21-hydroxylase activity as previously described. Genotype groups Null (n = 9) and A (n = 15) were predicted to result in SW CAH, group B (n = 8) was predicted to have the SV phenotype, and group C (n = 1) was predicted to have nonclassic CAH. A fifth group, D (n = 4), included patients in whom mutations were detected in only 1 allele.
Plasma total metanephrine (420.1 ± 60.0 vs. 657.7 ± 67.8 pg/ml; P = 0.04) and free metanephrine (13.4 ± 1.7 vs. 24.0 ± 4.1 pg/ml; P = 0.008) concentrations were significantly lower in children with SW CAH than in those with the SV form of the disease. Plasma free metanephrine concentrations best predicted phenotype, with accuracy similar to that of genotype. Concordance rates between genotype and phenotype were higher in the most severely affected patients (Null, 88.9%; A, 93.3%; B, 75%; plasma free metanephrine, <18.5 pg/ml: SW, 92%). The plasma free metanephrine concentration correlated with the expected 21-hydroxylase activity based on genotype, and there was a significant trend for free metanephrine concentrations across the three genotype groups (P < 0.0001).
Our findings indicate that measurement of adrenomedullary function, best assessed by the free metanephrine concentration, is a useful biomarker of disease severity in 21-hydroxylase deficiency. Molecular genotype and plasma free metanephrine concentration predict phenotype with similar accuracy. Both methods are more accurate in the most severe forms of the disease.
CONGENITAL ADRENAL hyperplasia (CAH) due to 21-hydroxylase deficiency is an autosomal recessive condition in which deletions or mutations of the cytochrome P450 21-hydroxylase gene result in decreased synthesis of glucocorticoids and often mineralocorticoids. This leads to increased secretion of CRH and ACTH, adrenal hyperplasia, and excessive production of adrenal androgens and androgen precursors for which 21-hydroxylation is not necessary (1). The clinical spectrum of the disease ranges from most severe to mild forms depending on the degree of 21-hydroxylase activity. Accordingly, three main clinical phenotypes of CAH have been described: classic salt-wasting (SW), classic simple-virilizing (SV), and nonclassic (1, 2, 3). In addition to impaired adrenocortical function, classic 21-hydroxylase deficiency is characterized by compromised adrenomedullary function (4). The latter is due to developmental defects in the formation of the adrenal medulla, leading to depletion of epinephrine stores and decreased production of metanephrine, the O-methylated metabolite of epinephrine.
The 21-hydroxylase gene (CYP21) is located in the human leukocyte antigen gene cluster region on the short arm of chromosome 6 (5). The functional gene (CYP21) and its homolog, the nonfunctional pseudogene (CYP21P), consist of 10 exons each and share 98% sequence homology in exons and approximately 96% sequence homology in introns (6, 7). In most cases the mutations responsible for 21-hydroxylase deficiency (Fig. 1
) are the result of unequal crossing over or gene conversion events between CYP21 and CYP21P (1, 8, 9, 10), whereas novel mutations or unique point mutations not present in the pseudogene have been identified in only a few cases (1).
|
The aim of the present study was to investigate the association between adrenomedullary function, assessed by measurement of plasma catecholamines and their metabolites, disease severity, and genotype in patients with classic 21-hydroxylase deficiency.
Subjects and Methods
Subjects
Thirty-seven children (22 males and 15 females; median age, 9.3 yr; range, 4.7–14.9 yr) with classic 21-hydroxylase deficiency were studied prospectively. The clinical characteristics of all subjects are summarized in Table 1. Patients were classified as having the SW (n = 28) or SV (n = 9) form of the disease according to clinical and endocrinological evaluation at presentation. Criteria for classification into the SW phenotype included history of a salt-losing crisis with documented hyponatremia, hyperkalemia and markedly elevated plasma renin activity. The number of adrenal crises requiring hospitalization and the associated biochemical abnormalities were obtained from parental reports and medical records. Patients in whom early virilization was diagnosed at an older age were classified as having the SV form of the disease.
|
-fludrocortisone (144 ± 66 µg/d) twice daily, whereas the remaining 17 patients were taking reduced doses of hydrocortisone (11.2 ± 3.9 mg/m2/d) and standard doses of 9-fludrocortisone (194 ± 90 µg/d), flutamide, and testolactone (20). Seventeen patients were also receiving GnRH agonists. The study was approved by the institutional review board at the National Institute of Child Health and Human Development, NIH. Written informed consent was obtained in all cases by a parent, and assent was given by children older than 7 yr.
Methods
Patients were seen early in the morning on the day of the study. An indwelling venous catheter for blood sampling was inserted, and all subjects rested in a supine position for a minimum of 30 min before samples were collected. Blood samples for measurement of epinephrine, total metanephrine, free metanephrine, cortisol, 17-hydroxyprogesterone, and ACTH concentrations were obtained at 0800 h before administration of the morning medication. Samples were centrifuged and separated immediately after collection, and were stored at -80 C until assayed. Epinephrine concentrations were also measured in a 24-h urine specimen. All patients underwent DNA analysis for determination of the 21-hydroxylase genotype.
Assays
Plasma and urinary epinephrine concentrations were quantified by liquid chromatography (21, 22). Plasma total and free metanephrine concentrations were determined using a different liquid chromatography procedure after extraction onto solid phase ion exchange columns (23). The intraassay coefficient of variation was 3.0% for epinephrine and 3.3% for metanephrine. The interassay coefficient of variation was 9.9% for epinephrine and 5.1% for metanephrine.
Mutation analysis
The 21-hydroxylase genotype was determined by allele-specific PCR and Southern blotting. The former detects eight-point mutations (24), whereas the latter detects gene deletions of the CYP21 gene and large gene conversions (9). The mutations screened included gene deletions, large gene conversions involving the promoter region, P30L in exon 1, an intron 2 splice site mutation (In2), an 8-bp deletion in exon 3 (8-bp deletion), I172N in exon 4, a cluster of mutations (I236N, V237E, M239K) in exon 6 (E6 cluster), V281L and 1762T insertion in exon 7, and Q318X and R356W in exon 8.
Categorization of mutation groups
Patients were divided into four groups (Null, A, B, and C) according to the mutations detected and the enzymatic activity that they confer as previously described (11, 12, 15). The Null group (n = 9) consisted of patients who were homozygous for mutations previously shown to confer no 21-hydroxylase activity (gene deletions, large gene conversions, 8-bp deletion, E6 cluster, Q318X, and R356W). Group A (n = 15) included patients who were homozygous for the In2 mutation or compound heterozygous carrying an In2 mutation on one allele and a null mutation on the other allele (In2/In2 or In2/Null). The In2 mutation has been shown to result in very low, but detectable, enzymatic activity in vitro (25). Group B (n = 8) included patients homozygous for the less severe I172N mutation or compound heterozygous with more severe mutations (I172N/I172N, I172N/Null, or I172N/In2). The I172N mutation results in approximately 2% of residual enzymatic activity in vitro. Group C (n = 1) consisted of patients who were homozygous for P30L and V281L or compound heterozygous with more severe mutations from other groups. These mutations are estimated to confer approximately 10–20% of 21-hydroxylase activity. A fifth group, D (n = 4), included patients in whom mutation analysis detected a mutation in one allele only.
Genotypes assigned to groups Null and A were expected to result in the SW form of CAH. Those in group B were expected to manifest as SV CAH, and those in group C as nonclassic CAH (11, 12, 15). The accuracy of phenotype prediction by the genotype was evaluated by estimating the positive predictive value in each group.
Statistical analysis
Nonnormally distributed data were logarithmically transformed before statistical analysis. Comparisons between two groups were performed using t test. The positive predictive value (ppv) for each mutation group was calculated as the number of patients with the expected phenotype divided by the number of patients in the given group, and expressed as a percentage.
Cross-validation analysis (26) was used to explore the association between free metanephrine concentrations and clinical phenotype. Because there is only one predictor variable (free metanephrine concentration), a rule for predicting phenotype (SW vs. SV) simply reduces to choosing a cut-point: a person with a free metanephrine value below the cut-point is predicted to be SW, and one above the cut-point is predicted to be SV. Cross-validation analysis is implemented by leaving a single observation out at a time, building the predictor from the remaining observations, applying the predictor to the observation that was left out and then assessing whether it was correctly or incorrectly classified. This method allows for the computation of an almost unbiased estimate of the true error rate of the prediction approach for best discriminating between the SW and SV forms of the disease. The cut-point was defined as the arithmetic average of the minimum and maximum potential cut-points that lead to the minimal number of classification errors. Comparison between the error rates observed when the clinical phenotype was predicted by genotype and by free metanephrine concentrations was performed using McNemar’s test for correlated dichotomous data. The association between free metanephrine concentrations and genotype was evaluated using Jonckheere’s nonparametric test for trend in ANOVA. Values are expressed as the mean ± SEM unless otherwise specified.
Results
Patient characteristics
There were no significant differences in age, body mass index, and hydrocortisone or 9-fludrocortisone replacement doses between children with SW and those with SV CAH (Table 1
).
Adrenomedullary function
Plasma total metanephrine (P = 0.04) and free metanephrine (P = 0.008) concentrations were significantly lower in children with SW CAH than in those with the SV form of the disease. No significant difference in plasma epinephrine or 24-h urinary epinephrine concentrations was observed between groups (Table 1). There was no significant correlation between free metanephrine concentrations and cortisol (r = 0.192; P = 0.26), 17-hydroxyprogesterone (r = -0.045; P = 0.79) or ACTH (r = -0.090; P = 0.6) concentrations. Moreover, despite differences in the hormonal control between the two groups of patients receiving different treatment regimens (hydrocortisone and 9-fludrocortisone vs. hydrocortisone, 9-fludrocortisone, flutamide, and testolactone), plasma catecholamines and their metabolites did not differ between the two groups.
Mutation analysis findings
The combination of allele-specific PCR and Southern blotting allowed the detection of mutations in 70 of the 74 alleles (sensitivity, 94.6%). In 33 patients mutations were detected in both alleles, whereas in 4 patients only 1 mutation was detected. For each subject, the number of mutations identified ranged from 1–4.
The full spectrum of genotypes in our cohort and the corresponding clinical phenotypes are shown in Table 2. Gene deletions were present in 18 alleles (24.3%), and point mutations were found in 52 alleles (70.3%). The most frequent point mutation was the intron 2 splice site mutation (In2), which was found in a total of 37 alleles (50%). Other mutations frequently detected included the P30L mutation, which was present in 12 alleles (16.2%); the I172N mutation, which was present in 10 alleles (13.5%); and the 8bp deletion, which was present in 10 alleles (13.5%). No new mutations were detected.
|
Prediction of phenotype from genotype
Genotype accurately predicted phenotype in 28 of the 33 patients (84.8%) in whom mutations were detected in both alleles (Table 2). The 4 patients in group D in whom mutation analysis detected a mutation in only 1 allele were excluded from this analysis.
In groups Null and A, 8 of 9 (88.9%) and 14 of 15 (93.3%) patients, respectively, had the expected phenotype (SW). In group B, 6 of the 8 (75%) patients had the expected phenotype (SV). The single patient in group C presented with the SV form of CAH (Fig. 2). Accordingly, the positive predictive value for each of the above groups was: ppvnull, 88.9%; ppvA, 93.3%; and ppvB, 75.0%. The error rate observed when clinical phenotype was predicted from genotype was 15.2%.
|
Cross-validation analysis indicated that a free metanephrine value of 18.5 pg/ml best served as a cut-off level in predicting the clinical phenotype. Patients who had free metanephrine concentration equal to or less than 18.5 pg/ml were likely to manifest the SW phenotype, whereas those with free metanephrine concentrations greater than 18.5 pg/ml were likely to present with the SV form of the disease. The error rate observed when the clinical phenotype was predicted from the plasma free metanephrine concentration was 19.4%. Therefore, the positive predictive value of the plasma free metanephrine concentration was 80.6%.
Comparison between the error rates detected when phenotype was predicted from genotype (15.2%) or from free metanephrine concentrations (19.4%) using McNemar’s test indicated that there was no significant difference (P = 0.34) between the two approaches.
Association between free metanephrine concentrations and genotype
Plasma free metanephrine concentrations were lower in children with mutations expected to confer lower 21-hydroxylase activity and rose progressively, from group Null to B, in parallel with an increase in the 21-hydroxylase activity. There was a significant trend in plasma free metanephrine concentrations across the three genotype groups (P < 0.0001; Fig. 3).
|
We found a significant correlation between the level of adrenomedullary impairment and both disease expression and molecular genotype in patients with 21-hydroxylase deficiency. In our study, both the measurement of plasma free metanephrine concentration and the determination of molecular genotype were well correlated with the clinical severity of the disease, and were similarly accurate in predicting clinical phenotype. Moreover, free metanephrine concentrations were significantly correlated with the expected 21-hydroxylase activity based on genotype.
In our patients with CAH, plasma epinephrine and metanephrine concentrations and urinary epinephrine excretion were all well below the normal range (4). However, free metanephrine values best discriminated disease severity. In humans, over 90% of the circulating epinephrine is released by the adrenal medulla directly into the bloodstream and more than 90% of circulating metanephrine, the O-methylated product of epinephrine, is produced from epinephrine that has leaked from storage vesicles into the cytoplasm of chromaffin cells (27, 28). The low plasma and urinary epinephrine concentrations reflect decreased adrenomedullary catecholamine secretion, whereas the decrease in metanephrine concentration indicates decreased adrenomedullary stores of epinephrine. Therefore, the differences in total and free metanephrine concentrations between the SW and SV phenotypes reflect a profound decrease in catecholamine adrenomedullary stores in patients with the most severe form of 21-hydroxylase deficiency. These findings concur with our previous description of adrenomedullary hypofunction in patients with classic 21-hydroxylase deficiency, which was attributed to a combination of decreased intraadrenal cortisol secretion and developmental defects in the formation of the adrenal medulla (4).
In our study we obtained a single blood sample from each subject for measurement of plasma epinephrine and metanephrine. Although plasma epinephrine concentrations display circadian variation, with higher values observed during the day and lower values during the night (29), plasma concentrations of total metanephrine do not fluctuate (30). Exocytotic release of epinephrine can be highly variable and is dependent on the activity of innervating nerves, which may increase with physical activity and upright posture during the day. In contrast, leakage of epinephrine from vesicular storage granules in adrenal medullary cells and subsequent O-methylation to metanephrine is a continuous process that occurs independently of exocytotic release of epinephrine by adrenal medullary cells (31).
The mutations of the 21-hydroxylase gene detected in our patients are known to be the result of either of two types of recombinations between the active CYP21 gene and the CYP21P pseudogene. These two mechanisms are unequal crossing over during meiosis, resulting in deletion of CYP21 (9, 32, 33), or apparent gene conversion events that transfer deleterious mutations normally present in CYP21P to CYP21 (10, 34, 35). The frequency of mutations detected in our study is very similar to that in other large cohorts (12, 14, 15, 36). The most frequent was the In2 mutation (50%), followed by gene deletions (24.3%), the P30L (16.2%) and I172N (13.5%) mutations, and the 8-bp deletion in exon 3 (13.5%). Eleven alleles (14.9%) were found to carry more than one mutation: the P30L, In2, and 8-bp deletion were detected in 10 alleles, and the I172N and cluster of mutations in exon 6 were found in 1 allele. Because these mutations are contiguous on the CYP21P gene, it is likely that they have been transferred to the CYP21 gene as a small conversion. In four (10.8%) patients with SW CAH, only one allele was affected, whereas the other allele did not carry a known or novel mutation.
In our study both measurement of the plasma free metanephrine concentration and determination of molecular genotype were well correlated with the clinical severity of the disease and were similarly accurate in predicting clinical phenotype. Thus, measurement of the free metanephrine concentration may be useful in determining overall disease severity as well as the degree of adrenomedullary hypofunction in patients with classic CAH and may have important clinical implications. It has been proposed that girls with severe forms of CAH could be more easily managed if their adrenals were removed at an early age (37, 38). Consideration of adrenalectomy has been proposed, using genotype to determine disease severity. Our findings suggest that measurement of adrenomedullary function is another useful biomarker of disease severity.
Previous studies reported genotype-phenotype concordance in patients with the most severe and the mildest forms of the disease, and nonconcordance mostly in moderately severe cases, where there may be overlap with the more or less severe phenotypes (11, 12, 13, 14, 15, 16). These differences have been attributed to the variable 21-hydroxylase activity conferred by the In2 and I172N mutations, the variety of mutations often observed in patients who are compound heterozygotes, the possibility of additional not yet identified mutations, and the genetic variations in extraadrenal 21-hydroxylase activity or sensitivity to glucocorticoids (1, 17, 18, 19). As with the prediction of phenotype from genotype, the prediction of phenotype from free metanephrine concentration was more accurate in the most severe than in the moderately severe cases of CAH.
The free metanephrine concentration was significantly correlated with the expected 21-hydroxylase activity based on genotype. The significant trend in free metanephrine concentrations across the three groups of patients with different mutations reflecting varying degrees of 21-hydroxylase activity (Null, A, and B) suggests that the severity of adrenomedullary hypofunction is directly related to the degree of enzymatic deficiency and, hence, adrenocortical impairment. This parallel decrease in adrenomedullary and adrenocortical function in patients with classic CAH represents another example of the close anatomical and functional links between the adrenal cortex and adrenal medulla in humans, and provides additional evidence that an intact intraadrenal cellular communication is crucial for their normal development and function (39, 40, 41). The concept of bidirectional signaling and interaction between the adrenal cortex and adrenal medulla is supported by the fact that glucocorticoids regulate the expression of enzymes participating in the synthesis of catecholamines, such as tyrosine hydroxylase (42, 43), dopamine-ß-hydroxylase (44), and phenylethanolamine-N-methyltransferase (45), while catecholamines enhance adrenocortical steroidogenesis by stimulating the expression of cytochrome P450 enzymes (39, 46, 47, 48). Our findings of more accurate prediction in severely affected cases and the correlation between free metanephrine concentration and 21-hydroxylase activity based on genotype support the idea that a significant portion of the adrenomedullary impairment observed in patients with 21-hydroxylase deficiency is due to developmental defects in the formation of the adrenal medulla. Most likely this is due to decreased in utero exposure to cortisol, unknown effects of the CYP21 gene, or a combination of factors.
We conclude that the accuracy of predicting phenotype in patients with 21-hydroxylase deficiency is similarly performed by determination of molecular genotype and measurement of the free metanephrine concentration. Both methods are more accurate in the most severe forms of the disease. The plasma free metanephrine concentration, a measurement of adrenomedullary function, appears to be a biomarker of disease severity and also reflects expected 21-hydroxylase activity based on genotype. Further studies are necessary to elucidate the role of adrenomedullary hypofunction in phenotype, and understand phenotypic nonconcordance with genotype and adrenomedullary function.
Acknowledgments
We thank the patients and their families for participation in this study, Ms. Donna Peterson for assistance with data management, Ms. Courtney Holmes and Ms. Patricia Sullivan for expert technical assistance, and the 9 West nursing staff of the Warren Grant Magnuson Clinical Center for assistance in the care of these patients.
Footnotes
D.P.M. is a Commissioned officer in the USPHS.
Abbreviations: CAH, Congenital adrenal hyperplasia; ppv, positive predictive value; SV, simple virilizing; SW, salt wasting.
Received December 18, 2001.
Accepted March 22, 2002.
References
This article has been cited by other articles:
 |
|
 |
|
  P. Sartorato, E. Zulian, S. Benedini, B. Mariniello, F. Schiavi, F. Bilora, G. Pozzan, N. Greggio, A. Pagnan, F. Mantero, et al. Cardiovascular Risk Factors and Ultrasound Evaluation of Intima-Media Thickness at Common Carotids, Carotid Bulbs, and Femoral and Abdominal Aorta Arteries in Patients with Classic Congenital Adrenal Hyperplasia due to 21-Hydroxylase Deficiency J. Clin. Endocrinol. Metab., March 1, 2007; 92(3): 1015 - 1018. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. M. K. Volkl, D. Simm, J. Dotsch, W. Rascher, and H. G. Dorr Altered 24-Hour Blood Pressure Profiles in Children and Adolescents with Classical Congenital Adrenal Hyperplasia due to 21-Hydroxylase Deficiency J. Clin. Endocrinol. Metab., December 1, 2006; 91(12): 4888 - 4895. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. E. de Galan, C. J. Tack, J. J. Willemsen, C. G. J. Sweep, P. Smits, and J. W. M. Lenders Plasma Metanephrine Levels Are Decreased in Type 1 Diabetic Patients with a Severely Impaired Epinephrine Response to Hypoglycemia, Indicating Reduced Adrenomedullary Stores of Epinephrine J. Clin. Endocrinol. Metab., May 1, 2004; 89(5): 2057 - 2061. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. W. Speiser Adrenomedullary Function May Predict Phenotype and Genotype in Classic 21-Hydroxylase Deficiency J. Clin. Endocrinol. Metab., July 1, 2002; 87(7): 3029 - 3030. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
