| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Other Original Articles |
The Nemours Children’s Clinic (R.C.O., E.B.M., J.W., J.E.S.), Jacksonville, Florida 32207; Department of Pediatrics (D.I.S.), All Children’s Hospital, St. Petersburg, Florida 33701; and the University of South Florida College of Medicine, Tampa, Florida 36612
Address all correspondence and requests for reprints to: Robert C. Olney, M.D., The Nemours Children’s Clinic, 807 Children’s Way, Jacksonville, Florida 32207. E-mail: rolney{at}nemours.org
Abstract
The most common cause of congenital adrenal hyperplasia is steroid 21-hydroxylase deficiency. The molecular genetics of this disease are such that genotyping is a potentially useful tool in its diagnosis. An assay was developed using real-time, quantitative PCR to detect deletions of the steroid 21-hydroxylase gene (CYP21A2). This assay was able to detect heterozygous gene deletions with an 
error rate of less than 5%, with a power greater than 95%. When combined with allele-specific PCR, genotyping for the nine most common mutations can be completed within hours of blood sampling. This technique was used to study subjects with 21-hydroxylase deficiency in North Florida. Twenty-eight subjects with congenital adrenal hyperplasia, seven first-degree relatives and thirteen normal subjects, were characterized. Of 96 chromosomes, 69 abnormal alleles were identified. Among unrelated abnormal alleles, the frequency of specific mutations was 28% for a gene deletion, 24% for the intron 2 splice mutation, 10% for ile172asn, 8% each for val281leu and the exon 6 cluster, and 6% for gln318x mutations. These frequencies, as well as the genotype/phenotype correlation, were similar to those found in comparable populations. The utility of genotyping in the diagnosis of 21-hydroxylase deficiency is increased by the rapidity of the analysis. With quantitative PCR, the need for more expensive and time consuming Southern blot analysis is reduced and limited to the clarification of certain genotypes. Faster results will allow for more timely initiation of appropriate therapy and limit the exposure of potentially unnecessary therapy.
CONGENITAL ADRENAL HYPERPLASIA (CAH) results from a deficiency in one of several enzymes responsible for cortisol biosynthesis (for review, see Ref. 1). In 90–95% of cases, 21-hydroxylation is impaired in the adrenal cortex (21-hydroxylase deficiency) so that conversion of 17-hydroxyprogesterone (17-OHP) to 11-deoxycortisol is blocked. Because of defective cortisol synthesis, ACTH levels are increased, resulting in overproduction and accumulation of precursors proximal to the block. This accumulation results in excessive production of adrenal androgens, causing virilization. Steroid 21-hydroxylase also catalyzes the conversion of progesterone to 11-deoxycorticosterone in the biosynthetic pathway of aldosterone. Defects in this enzyme variably impair mineralcorticoid synthesis, resulting in renal salt-wasting. There are four recognized clinical forms of 21-hydroxylase deficiency; salt-wasting, simple virilizing, nonclassic (also called late onset, attenuated, or acquired), and cryptic.
The molecular genetics of 21-hydroxylase deficiency have been studied extensively (1). This disease is autosomal recessive and unusual in that the majority of mutations found in the steroid 21-hydroxylase gene (CYP21A2) are found in a closely related pseudogene (CYP21A1P). The relatively high prevalence of this disease may be due to recurrent recombination between these two genes. It is estimated that either gene deletion, gene inactivation by conversion to the pseudogene, or the presence of one of eight point mutations account for 95% of abnormal alleles (2). Although not perfect, there is a good correlation of the genotype of CYP21A2 and the clinical severity of CAH (3, 4).
Several different approaches for genotyping CYP21A2 have been developed as a potential tool in the diagnosis of 21-hydroxylase deficiency (1). Although the point mutations can be detected rapidly and reliably, identifying gene deletions has been more laborious. We present here data showing the use of real-time, quantitative PCR (qPCR) to screen for gene deletions as part of a rapid technique for genotyping CYP21A2.
Subjects and Methods
Experimental subjects
Forty-eight subjects were recruited from the following Florida pediatric endocrinology clinics: the Nemours Children’s Clinics in Jacksonville and Pensacola and All Children’s Hospital, St. Petersburg. Subjects were patients with known CAH due to 21-hydroxylase deficiency (28 subjects), first degree relatives (7 subjects), or normal control subjects (9 subjects). The diagnosis of CAH and the determination of its severity were based on clinical presentation and 17-hydroxyprogesterone levels at presentation or after ACTH stimulation. The diagnosis and severity were confirmed by review of medical records. Classic CAH (salt-wasting or simple virilizing) was defined as onset of symptoms before 5 yr of age and a peak 17-OHP following ACTH stimulation greater than 30,000 pmol/liter (10,000 ng/dl) (4). Salt-wasting was determined by a continued need for mineralcorticoid replacement. Four subjects were enrolled where the diagnosis of CAH was being considered, but not confirmed. These included three infants with abnormal newborn screens for CAH and one school-age boy with precocious puberty and a borderline 17-OHP level on ACTH stimulation test. Subsequent biochemical evaluation showed that none of the four had CAH. Signed informed consent was obtained from the subject or a parent in each case. Table 1
shows the characteristics of the subjects studied.
|
Blood (2.5 ml) was collected in EDTA anticoagulant tubes and placed immediately on ice. Genomic DNA was isolated from a 0.2 ml blood sample using the QIAamp DNA Blood Kit (QIAGEN Inc., Valencia, CA). The manufacturer’s instructions were followed, except the final elution was done with 0.2 ml of 10 mM Tris (pH 8.3). Control DNA was pooled DNA obtained from six normal subjects.
DNA was quantified using the Hoechst Dye 33258 assay (5) and a dedicated fluorometer (TKO 100 Fluorometer, Hoefer Scientific Instruments, San Francisco, CA). This assay is not affected by the presence of RNA or protein (5). Two milliliters of DNA assay buffer (0.1 µg/ml Hoechst Dye 33258, 10 mM Tris, pH 7.5, 200 mM NaCl, 1 mM EDTA) was added to a cuvette and the fluorometer zeroed. Five microliters of sample DNA were added and the fluorescence read. Then 5 µl of control DNA were added and the fluorescence read again. Each sample was assayed in triplicate. The ratio of fluorescence of the sample to the control DNA gives the ratio of DNA concentration between the two. An aliquot of control DNA was then diluted to the same concentration as the sample DNA and the assay repeated, to confirm that the sample and control DNA concentrations were the same for the qPCR analysis. In practice, the ratio between the sample and diluted control DNA was 0.99 ± 0.07 (mean ± SD).
qPCR
All PCR reactions were done on the LightCycler (Roche Molecular Biochemicals, Indianapolis, IN).
qPCR for the ß-globin gene (HBB) was done using the LightCycler-Control Kit DNA (Roche Molecular Biochemicals). Manufacturer’s instructions were followed.
Gene quantification for CYP21A2 was done using primer sequences described by Wedell and Luthman (6). The upstream primer (P55) sequence was 5'-CCTGTCCTTGGGAGACTACT-3' and spans the 8 bp in exon 3 that are deleted in the pseudogene. This eliminated amplification of the pseudogene. The downstream primer (P16) sequence was 5'-GTCCACAATTTGGATGGACCA-3'. These primers amplify a 508-bp product. Reaction conditions were 0.5 µM of each primer, 2.5 mM MgCl2, 0.33x SYBR Green 1 dye (Molecular Probes, Inc., Eugene, OR), 1x LightCycler-DNA Master Hybridization Probes (Roche Molecular Biochemicals), and 2 µl of DNA sample in a final reaction volume of 10 µl. The LightCycler DNA Master was pretreated with TaqStart Antibody (CLONTECH Laboratories, Inc., Palo Alto, CA), 0.08 µl of antibody per 1 µl of LightCycler DNA Master and incubated for 5 min at room temperature. The reaction components (less the DNA sample) were made up as a 1.25x master mix. Eight microliters of the master mix were pipetted into glass capillary tubes, followed by the DNA sample. Sample DNA and control DNA were run in eight identical tubes each and a negative control was included in each run. PCR parameters were 96 C for 2 min, then 35 cycles of 97 C for 0 sec, 60 C for 5 sec, 72 C for 90 sec with fluorescent sampling at the end of the 72 C segment. This was then followed by a melting curve determination between 65 C and 98 C. Under these conditions, a single amplicon was amplified from genomic DNA, with a melting temperature of 92.3 C. PAGE with ethidium bromide staining confirmed that only a single band of 508 bp was present.
After it had been determined that only a single product accumulated in each tube and that no DNA had accumulated in the negative control, the quantitative analysis was done. The LightCycler software (version 3.0) analyzed each of the PCR curves by plotting the accumulation of fluorescence over the course of the run. The peak of the second derivative of the curve was determined and defined as the "threshold cycle" for the DNA sample. The mean threshold cycle for the eight sample tubes were compared with that of the eight control tubes by the t test. If the P value was >0.05, then the sample DNA was classified as having a normal gene complement. If the P value was <0.05 then the sample DNA was classified as having an abnormal gene complement. In this case, if the mean threshold cycle for the sample DNA was greater than that for the control DNA, then the sample DNA was classified as having only a single copy of the CYP21A2 per genome equivalent. If the threshold cycle was less, then the sample was classified as having more than two copies per genome equivalent.
For this study, CYP21A2 gene conversion was defined as the presence of the eight bp deletion found in CYP21A1P. This assay does not discriminate between a complete gene deletion and a gene conversion.
If no product amplified from the reaction (suggesting homozygous gene deletion or gene conversion), a second PCR reaction was done to confirm that the PCR reagents were active. This reaction was a duplex reaction including a primer pair that was common to both CYP21A2 and CYP21A1P (primers P5 and P16) and an additional downstream primer specific for CYP21A2 (primer P48) (6). Reaction conditions were as above, except MgCl2 was 3.0 mM, and primer P16 was 0.15 µM. PCR parameters were 96 C for 2 min, then 30 cycles of 97 C for 0 sec, 60 C for 5 sec, 72 C for 40 sec with fluorescent sampling at the end of the 72 C segment. This was then followed by a melting curve determination between 65 and 98 C. The presence of either CYP21A2 or CYP21A1P was determined by the presence of an amplicon with a melting temperature of 90.9 C and confirmed that the PCR reaction was working. If CYP21A2 was present, a second amplicon with a melting temperature of 87.9 C was found. The absence of this amplicon confirmed the absence of CYP21A2.
Allele-specific PCR
Allele-specific PCR was performed as described by Wedell and Luthman (6), except that the reactions were modified to run on the LightCycler. Using melting curves to identify specific amplicons eliminated the need for gel electrophoresis. Briefly, CYP21A2 was amplified in two sections with primers that spanned the 8-bp region in exon 3 that is deleted in the pseudogene. This first reaction was run using the same conditions as for the qPCR. The resulting reactions were pooled, diluted 1:1,000 with 10 mM Tris (pH 8.3), then subjected to allele-specific PCR, with primers specific for eight known mutations. The mutations screened for were the intron 2 splice mutation (A/C656G, I2 splice), pro30leu, ile172asn, the cluster ile236asn/val237glu/met239lys in exon 6 (E6 cluster), val281leu, leu307insT, gln318x, and arg356trp. The polymorphism (A or C) at position 656 was also determined. The primers used were those described (6), except for the primer specific for the I2 splice site mutation (primer P659G). The upstream primer used for this was 5'-CCTCCAGCCCCCAG-3' to eliminate cross-over of the previously described primer with the wild-type sequence in these conditions. The reaction conditions were the same as listed for the qPCR, above, except the MgCl2 concentration was 1.5 or 2.0 mM (depending on the specific reaction). PCR parameters were 96 C for 2 min, then 23 cycles of 97 C for 0 sec, 54 C for 5 sec, 72 C for 60 sec with fluorescent sampling at the end of the 72 C segment. This was then followed by a melting curve determination between 65 and 98 C. Negative controls for each reaction were run at the same time. For any reaction where a DNA product was amplified, a third round of PCR was done that included reactions for both the wild-type and mutation sequences and positive and negative controls. Positive control DNA for the mutations was kindly provided by Dr. Anna Wedell (Karolinska Hospital, Stockholm, Sweden) (6). If an amplicon was seen in the reaction specific for a mutation, but not in the reaction specific for the corresponding wild-type sequence, the sample was classified as homozygous for that mutation. If an amplicon was present in the reactions specific for both the mutation and the wild-type sequence, the sample was classified as heterozygous for that mutation.
Southern blotting
Genomic DNA was isolated as above. Five micrograms were digested with TaqI restriction endonuclease overnight and resolved on a 1% agarose gel. DNA was transferred to GeneScreen Plus (NEN Life Science Products, Boston MA) and the membrane dried. The CYP21A2 specific gene probe was synthesized from a cloned 2206-bp CYP21A2 gene fragment (nucleotides 703-2908). A control probe was synthesized using a 800-bp cloned fragment of the ß-2 adrenergic receptor (ADRB2, nucleotides -213 to +586), a well described, single copy gene. The probes were labeled by random primer extension with [32P]-dCTP and purified by passing the sample over a G25 spin column. The membrane was hybridized first with the CYP21A2 probe and autoradiography was done. The membrane was then hybridized with the ADRB2 probe and autoradiography repeated. The membrane was not stripped between probings. Expected results were a 3,740-bp fragment for CYP21A2, a 3,204-bp fragment for CYP21A1P, and a 2,123-bp fragment for ADRB2.
Densitometry of the autoradiographs were done by scanning the films (Arcus II, Agfa Corp., Ridgefield Park, NJ) and analyzing the scan with SigmaGel software (version 1.0, SPSS, Inc., Chicago, IL). The density of the CYP21A2 band was normalized to the ADRB2 band from the same lane to control for differences in loading and transfer. DNA samples from six normal controls were run on each gel. A normal range for the normalized density of CYP21A2 was calculated for each Southern blot, defined as 2.5 SD above and below the mean for the six normal controls. A subject was considered to have one gene copy per genome equivalent if the normalized density of CYP21A2 fell below this range and more than two copies if it fell above this range.
Statistical analysis
Two sample means were compared by the t test, paired analysis was done where appropriate. Three sample means were compared by ANOVA with Tukey posthoc comparison, if ANOVA showed significance. Significance was assumed for P < 0.05. These tests were done with SPSS, Inc. for Windows (version 10.0) software. Power calculations were done with GPOWER (version 2.0) software (7).
Results
Blood processing
To determine the optimal conditions for acquiring DNA, several parameters of blood sample storage and DNA isolation were optimized. Fresh, whole blood from six normal subjects was used for DNA isolation using both the Generation DNA Purification System (Gentra Systems, Inc., Minneapolis, MN) and the QIAamp DNA Blood Kit. Yield was identical for both kits. Quality of the DNA was determined by qPCR for the ß-globin gene. Results for the QIAGEN kit showed the ratio of copies present (determined by qPCR) to expected copies (calculated from the amount of DNA in the reaction) was 1.14 ± 0.35 (mean ± SD) and for the Gentra kit was 0.40 ± 0.27 (P < 0.005 by paired t test). The QIAGEN kit was used for all subsequent samples.
The effect of blood storage conditions was explored by drawing blood from four normal subjects and storing aliquots at room temperature, 4 C, and -70 C for 24 h, then isolating DNA using the QIAGEN kit. Yield was not different between the room temperature and 4 C samples, but was 19% higher in the -70 C samples (P < 0.05 by ANOVA). Quality analysis showed that the room temperature samples had 0.23 ± 0.07 ß-globin gene copies present/copies expected, 4 C samples had 0.84 ± 0.19, and -70 C samples had 0.85 ± 0.08 (P < 0.01 by ANOVA, P < 0.01 for room temperature vs. 4 C and for room temperature vs. -70 C by Tukey posthoc comparison). For all subsequent samples, DNA was isolated from blood immediately when possible, or stored or shipped at 4 C for no longer than 24 h.
qPCR
The PCR conditions were optimized so that the qPCR reaction for CYP21A2 gave a single amplicon as determined both by melting temperature and by gel electrophoresis with ethidium bromide staining. The optimal conditions are listed above. A standard curve for the quantification of CYP21A2 (Fig. 1) demonstrated the linear relationship between the threshold cycle and the log of the amount of DNA in the reaction, for DNA amounts used for the assay. The slope of the standard curve showed the efficiency of this PCR reaction was 0.93, or 1.08 ± 0.21 PCR cycles per doubling (mean ± SD, from four experiments on different days). Preliminary work found the mean SD of threshold cycle within identical reactions in a single run was 0.55 cycles. Power calculations showed that to detect a 2-fold difference in gene copy number (1.08 cycles) with this SD (assuming an error of <0.05 and a power of 95%, two-tailed analysis), each sample would require eight replicates. Calculations showed that for an intraassay SD of 0.8 cycles, the power of the assay dropped to 70%. Because of this, assay runs where either the control or sample SD exceeded 0.8 cycles were repeated. Using this criteria, we retrospectively found the mean SD within replicates at the completion of the study was 0.35 cycles (intraassay coefficient of variation of 1.4%). Figure 2 demonstrates this sensitivity. A sample of DNA from a normal control subject was assayed for CYP21A2 along with an aliquot diluted 1:2. The assay showed a significant difference between the diluted aliquot and the control DNA (P < 0.001 by ANOVA and Tukey posthoc comparison). Detecting the presence of three gene copies per genome equivalent (a difference of 0.63 cycles compared with normal) resulted in a drop in the power of the assay to 92%.
|
|
). The figure shows the Southern blot results for two normal subjects and four sibling pairs with CAH. qPCR and densitometry analysis identified the sibling pairs as having greater than two copies (lanes B, Fig. 3), two copies (lanes C), one copy (lanes D), and no copies (lanes E) of CYP21A2 per genome equivalent. The absence of the pseudogene in lanes B suggests that gene conversion has normalized both copies of CYP21A1P. Similarly, the increased density of CYP21A1P in lanes D suggest that one copy of CYP21A2 has been inactivated by conversion to CYP21A1P. In contrast, the density of the pseudogene in lanes E suggest that these siblings have homozygous CYP21A2 gene deletions, rather than a gene conversion.
|
. Of 34 subjects studied by both methods, 28 had agreement of gene copy number. Two of the subjects (who were siblings, the arrowed points, Fig. 4) were classified as one copy per genome by qPCR, but two copies by Southern blotting. Allele-specific PCR confirmed the presence of a CYP21A2 allele with the exon three 8-bp deletion. This demonstrates that this sibling pair had a gene microconversion. A single subject was classified as having two copies by qPCR but only one copy by Southern (circled point, Fig. 4). This subject was a girl diagnosed at birth with salt-wasting CAH. Allele-specific PCR demonstrated the presence of the I2 splice mutation with no allele having the wild-type sequence at this locus. It could not be determined if she was homozygous for the mutation (suggested by qPCR) or compound heterozygous for the mutation and a gene deletion/conversion (suggested by Southern blotting). Three subjects were classified as having two copies by qPCR, but greater than two copies by Southern blotting. One was a boy with nonclassic CAH who was compound heterozygous for val281leu and gln318x mutations. The other two were the unaffected parents of a girl with salt-wasting CAH who was found to have two gene copies by both qPCR and Southern blot. She was heterozygous for the I2 splice mutation; the inactivating mutation of the second allele was unknown. The true copy number in these cases could not be independently determined.
|
Forty-eight subjects (96 chromosomes) were studied by qPCR of CYP21A2. Twenty-six subjects were found to have two gene copies per genome equivalent, 15 had only one gene copy (heterozygous gene deletion/conversion), and 3 had no detectable CYP21A2 genes. Four subjects (two of which were siblings) were found to have more than two copies per genome equivalent. Interestingly, all four had salt-wasting CAH.
In one family, the studied subjects included a family in which the proband was diagnosed with salt-wasting CAH and was found to have no detectable genes copies. Analysis of the parents (obligate heterozygotes) showed that both had only a single gene copy. An unaffected sibling was also found to have only a single copy. These results were confirmed by Southern blot analysis.
Genotyping for point mutations
Allele-specific PCR was used to determine the genotype for the known polymorphism at base 656. At this position, A or C are wild-type alleles, but G results in abnormal splicing of intron 2 (the I2 splice mutation). Of the 79 gene copies detected, 42 were 656A (68% of wild-type alleles), 20 were 656C (32% of wild-type alleles), and 14 were 656G. In the 15 subjects where only a single gene copy was detected by qPCR, none were found to be heterozygous at this locus.
Table 2 shows the mutations detected and their frequency. In this study, no subjects were found to have the pro30leu, leu307insT, or arg356trp mutations. In ten subjects, one allele was known to be abnormal, but no mutations were detected by allele-specific PCR. In a single family, two siblings were heterozygous for the E6 cluster mutation, while a half-sibling (same mother) was heterozygous for the I2 splice mutation. All three had salt-wasting CAH. The mother of the family (an obligate heterozygote) had no mutations detected. In this family, the abnormal allele was linked to 656A. Sequencing of the gene in this family revealed several point mutations, none of which could be predicted apriori to eliminate enzyme function. Another subject with salt-wasting CAH (unrelated to the above family) was also heterozygous for the I2 splice mutation. One subject with nonclassic CAH was heterozygous for val281leu mutation. One subject (salt-wasting CAH) was found to have three gene copies and was heterozygous for the I2 splice mutation and the gln318x mutation. The third allele was linked to 656A but did not have an identifiable mutation. An additional subject with nonclassic CAH had two gene copies present and no mutations were detected. This girl presented at 7 yr of age with precocious adrenarche and had a peak 17-OHP of 54,100pmol/liter (1,787ng/dl) after standard ACTH stimulation testing. Finally, one subject with salt-wasting CAH was found to be heterozygous for a gene deletion/conversion, but no mutation was detected by allele-specific PCR. Sequencing of the gene showed a G to A mutation at nucleotide 2160. This is the first nucleotide of intron 8. A G must be in this position for normal splicing to occur. This mutation is predicted to block the out-splicing of intron 8 and would bring a stop codon into frame, eight codons downstream. Such a premature termination would eliminate exons 9 and 10, which are necessary for normal enzyme function.
|
Genotype/phenotype correlation
The genotype of the subjects with CAH were classified using the system described and modified previously (4, 8, 9). Table 3 shows the genotype classification and the phenotype for these subjects. Data from many studies have shown that patients with the genotypes of the null classification universally have salt-wasting CAH. Genotypes of the A classification are associated with salt-wasting (90%) or simple virilizing (10%) CAH (percentages calculated from data from references 3 and 4). The B classification is associated with salt-wasting (26%) or simple virilizing (74%) CAH. The Ca classification is associated with simple virilizing (10%) or nonclassic (90%) forms. Genotypes of the Cb classification have the most variability of phenotype, being associated with salt-wasting (12%), simple virilizing (53%), or nonclassic (35%) CAH. The results of the current study are in line with these percentages. The two subjects in the B classification with salt-wasting are unrelated. Both were girls noted to be virilized at birth. One was found to have a random 17-OHP of 73,800pmol/liter (2,440 ng/dl) in the neonatal period. At 1 month of age, screening found borderline hyponatremia and a plasma renin activity greater than five times the upper limit of the normal range. The second subject had a random 17-OHP on d 3 of life of 39,000 pmol/liter (1,290 ng/dl). On d 5, she was found to have a sodium of 136 mmol/liter (136 mEq/liter and a potassium of 6.2 mmol/liter (6.2 mEq/liter). Mineralcorticoid replacement was started and continued in both subjects. Neither of these subjects had salt-wasting crisis.
|
A number of methods are now in use for genotyping CYP21A2. However, the most common mutation found in CAH due to 21-hydroxylase deficiency is a gene deletion or gene conversion. The previous method for detecting these was Southern blot analysis, a time consuming approach. We present here a method of detecting these mutations by quantifying gene copy number by qPCR. qPCR has evolved rapidly over the last decade. The first approaches were based on end-point analysis of the amplicon and were problematic for very high variability. A 2-fold difference in gene copy number between two samples could not be detected reliably. However, the advent of thermal cyclers with continuous fluorescent monitoring capabilities has allowed for data collection during the PCR. This, along with the new approaches for data analysis, has reduced the variability substantially. After careful optimization, we have achieved a mean intraassay variability of 1.4% for quantifying CYP21A2 in genomic DNA. With enough replicates, detecting a 2-fold difference (e.g. a deletion of one of the two copies of CYP21A2) is achievable with an error of less than 5% and a power of greater than 95%.
This assay was validated by several approaches. In one family, the child with salt-wasting CAH was found to be homozygous for CYP21A2 deletion/conversion). Thus, his parents (neither of whom had CAH) were both obligate heterozygotes for the gene deletion. When the parents were studied, the assay identified both as having only a single gene copy per genome equivalent. Further, Mendelian genetics predicted that the subject’s unaffected sibling had a 67% chance of also having only a single gene copy (the possible combinations are maternalnormal/paternalnormal, maternalnormal/paternaldeletion, and maternaldeletion/paternalnormal; the maternaldeletion/paternaldeletion combination is eliminated because the sibling does not have CAH). Analysis confirmed that she did in fact have only one copy. For this family, qPCR results were confirmed by classic genetics. Another approach was to analyze the polymorphism at base 656. Our data showed a prevalence of 68% of the A allele and 32% of the C allele. Genetics predicts that in people with two gene copies, 44% should be heterozygous at this locus. Of the 15 subjects found to have only a single gene copy, none were heterozygous at this locus. Further, in subjects with CAH who were found to have only a single gene copy and a point mutation identified on the remaining copy (11 subjects), none proved to be heterozygous for the point mutation. Thus, there were no cases where a subject classified as a having only a single copy of CYP21A2 by qPCR was found to be heterozygous at any position, consistent with their hemizygous status. Finally, the frequency of CYP21A2 deletion/conversion in subjects with CAH found in this study was comparable to the frequency found by Southern blotting in previous large studies. These data show that qPCR is a reliable approach to detecting gene deletions/conversions of CYP21A2.
There were some discrepancies between gene quantification by qPCR and Southern blotting. Because the qPCR assay relied on the wild-type sequence at the locus of the exon three 8-bp deletion, a gene with this mutation in isolation would be classified as a gene deletion or conversion, whereas Southern blotting detected it as a normal gene. Clinically, this distinction is less important as both result in complete gene inactivation. The majority of the qPCR/Southern blot discrepancies arose when discriminating between two and three gene copies per genome. The logarithmic nature of qPCR makes the detection of this difference difficult and there is a drop in the power of the qPCR assay. In these cases, confirmation of the qPCR results by Southern blotting is recommended. In a single case, there was a discrepancy when discriminating between one and two gene copies. Because this situation has significant potential clinical implications, borderline or ambiguous qPCR results must be confirmed by Southern blotting. We suggest that the criteria for ambiguous or borderline qPCR results be: 1) greater than two gene copies are detected; 2) a single copy is detected, but the difference in cycle number between the sample and controls is less than 1.0; and 3) a single copy is detected, but allele-specific PCR identifies two alleles at any of the tested loci. It should be noted that these criteria would not have classified the results from subject with the qPCR/Southern discrepancy as ambiguous. In this particular case, the subject was found to be homozygous for the I2 splice mutation and would have still have been correctly diagnosed as having 21-hydroxylase deficiency.
The advantage of using this method in genotyping CYP21A2 is speed. The very rapid cycling times of the LightCycler and the ability to do melting point analysis of the amplicons (eliminating the need for gel electrophoresis) reduce analysis time substantially. The entire assay, including DNA isolation from blood, qPCR, and allele-specific PCR, can be done in 5 h. There are some disadvantages to this method as well. One is the critical dependence on the DNA quantification. Small errors in the starting amount of DNA can lead to errors in interpretation. We addressed this by using a very sensitive DNA assay that is not effected by the presence of RNA or protein and by checking the DNA concentration before and after dilution. Another disadvantage is the method does not give the information about gene conversions vs. gene deletions that can be had from Southern blot analysis.
The genotype analysis of this study (from a north Florida population) showed no substantial differences in the frequencies of the common mutations from other U.S. populations. The frequency of gene deletions/conversions and the I2 splice mutation is remarkably consistent across all populations studied. The other point mutations have more variable prevalence in different populations and are likely to be more population specific.
There remains controversy about the use of genotyping in the clinical diagnosis of CAH (10, 11). The recent technical report and policy statement on CAH (12) from the American Academy of Pediatrics concluded that molecular genetic analysis of the family is required for accurate prenatal diagnosis and is suggested as an adjunct to newborn screening. On a molecular genetics basis, CAH due to 21-hydroxylase deficiency is one of the best and most widely studied diseases. The unusual pattern of mutations of this gene, with 90–95% of abnormal alleles containing one of only a handful of mutations, suggests that most cases could be diagnosed by genotyping. This would have a sensitivity of 80–90%, given the rare and unique mutations of this gene that would not be identified by any of the rapid genotyping methods. Specificity approaches 100%. Although not perfect, it is comparable to other diagnostic tests in use today. An additional drawback to genotyping for diagnosis of CAH is that it does not allow complete confidence in predicting the phenotype. Genotypes that fall into the null or A categories can be predicted to cause salt-wasting CAH with high confidence. The presence of the ile172asn, val281leu, or pro30leu mutations make prediction more problematic, but reduce the probability of the patient will have salt-wasting. It is clear that genotyping cannot replace biochemical testing for CAH; however, in most cases, it would provide useful diagnostic information.
The value of genotyping is improved if it can be done rapidly. In cases of prenatal diagnosis in families that have opted for maternal dexamethasone treatment, the quicker the genotype can be determined, the less time an unaffected fetus will be exposed. In cases of neonates with positive neonatal screen for CAH, at least some information about the potential for salt-loss can be obtained before problems occur. Genotyping of parents allows for confirmation of the results and adds significant confidence to the analysis. This should be done whenever possible (1).
In summary, we present a method for genotyping CYP21A2 that is reliable and very rapid. Although it will not replace biochemical testing, it is a useful adjunct to diagnosing CAH due to 21-hydroxylase deficiency.
Acknowledgments
We thank Drs. Nelly Mauras, Larry Fox, and Helen Hsiang for help in recruiting subjects and Lynn Everline for technical assistance.
Footnotes
This study was presented at the 82nd Annual Meeting of The Endocrine Society, Toronto, Canada, 2000. This study was supported by a grant from the Nemours Research Program.
Abbreviations: CAH, Congenital adrenal hyperplasia; CYP21A2, steroid 21-hydroxylase gene; qPCR, quantitative PCR.
Received November 8, 2000.
Accepted November 8, 2001.
References
This article has been cited by other articles:
 |
|
 |
|
  S. Parajes, C. Quinterio, F. Dominguez, and L. Loidi A Simple and Robust Quantitative PCR Assay to Determine CYP21A2 Gene Dose in the Diagnosis of 21-Hydroxylase Deficiency Clin. Chem., September 1, 2007; 53(9): 1577 - 1584. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. E.L.M. Vissers, J. A. Veltman, A. G. van Kessel, and H. G. Brunner Identification of disease genes by whole genome CGH arrays Hum. Mol. Genet., October 15, 2005; 14(suppl_2): R215 - R223. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Keen-Kim, J. B. Redman, R. U. Alanes, M. M. Eachus, R. C. Wilson, M. I. New, J. M. Nakamoto, and R. G. Fenwick Validation and Clinical Application of a Locus-Specific Polymerase Chain Reaction- and Minisequencing-Based Assay for Congenital Adrenal Hyperplasia (21-Hydroxylase Deficiency) J. Mol. Diagn., May 1, 2005; 7(2): 236 - 246. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. L Nagpal, Y. Chen, and T. Lin Effects of overexpression of CXCL10 (cytokine-responsive gene-2) on MA-10 mouse Leydig tumor cell steroidogenesis and proliferation J. Endocrinol., December 1, 2004; 183(3): 585 - 594. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Tukel, O. Uyguner, J. Q. Wei, M. Yuksel-Apak, N. Saka, D. X. Song, H. Kayserili, F. Bas, H. Gunoz, R. C. Wilson, et al. A Novel Semiquantitative Polymerase Chain Reaction/Enzyme Digestion-Based Method for Detection of Large Scale Deletions/Conversions of the CYP21 Gene and Mutation Screening in Turkish Families with 21-Hydroxylase Deficiency J. Clin. Endocrinol. Metab., December 1, 2003; 88(12): 5893 - 5897. [Abstract] [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 |
