Duplications of the Functional CYP21A2 Gene Are Primarily Restricted to Q318X Alleles: Evidence for a Founder Effect

Context: Rare haplotypes with Q318X mutations and duplicated CYP21A2 genes have been reported to occur in different populations to a varying extent. Discrimination between a normal (Q318X mutation on one of the duplicated CYP21A2 genes) and a congenital adrenal hyperplasia (CAH, Q318X mutation without duplicated functional gene) allele is of importance, particularly for prenatal diagnosis and the respective genetic counseling. Although methods to differentiate between such alleles have been published only recently, it remains unclear with which frequency Q318X mutations are associated with duplicated CYP21A2 genes and whether these haplotypes have a common ancestry.

Subjects and Methods: Human leukocyte antigen (HLA) typing has been performed in 38 unrelated individuals and in 11 family members detected to carry a Q318X mutation in the course of CYP21 genotyping using sequence, multiplex ligation-dependent probe amplification, and Southern blot analyses.

Results: The majority (n = 32, 84.2%) of the 38 unrelated individuals carrying the Q318X mutation had the trimodular RCCX haplotype, carrying the Q318X mutation on a duplicated CYP21A2 gene. Twenty-two individuals of these 32 (68.8%) were of the rare HLA-B*50-Cw*06 haplotype, suggesting a common ancestry of this haplotype. In five (13.2%) of the 38 subjects, the Q318X mutation was not associated with a duplicated CYP21A2 gene and thus represents a CAH allele. None of these five patients had the above mentioned HLA haplotype.

Conclusion: The majority of individuals in whom Q318X mutations are detected carry a duplicated functional CYP21A2 gene and the rare HLA-B*50-Cw*06 haplotype.

To avoid false positive genotyping, it is of utmost importance to be aware of the high frequency of duplicated CYP21A2-genes in association with Q318X-aberrations.

Congenital adrenal hyperplasia (CAH) due to mutations in the 21-hydroxylase gene CYP21A2 (OMIM+201910) is a recessively inherited disorder characterized by androgen overproduction and impaired cortisol and aldosterone synthesis (1–3).

Lying within the human leukocyte antigen (HLA) major histocompatibility complex on chromosome 6p21, both the functional CYP21A2 and the highly homologous, but functionally inactive pseudogene CYP21A1P (1, 2) form a genetic unit designated as RCCX module with their neighboring genes tenascin TNXA/B, complement C4A/B, and the serine/threonine nuclear protein kinase RP plus small truncated sections of TNX and RP (4, 5). Copy number variations for this RCCX locus have been described, two modules representing the standard and three or four modules being rare cases (4, 5). A trimodular haplotype carrying one copy of the CYP21A1P pseudogene and two copies of the CYP21A2 gene, the latter affected by one (Q318X) or more (Q318X, Intron2splice) mutations has been described (6, 7) to occur in less than 2% of the general population. Recently Parajes et al. (8) found 7% of healthy Spanish individuals to carry a duplicated CYP21A2 gene, one copy affected with a Q318X mutation. In all Spanish (8) and Swedish (6) cases, the duplicated CYP21A2 gene and the Q318X mutation were in linkage disequilibrium with two uncommon genetic variants, a G→A change in intron 2 (IVS2-79G/A) and a C→T change at nucleotide 13 in the 3′-untranslated region (UTR) region (c*C/T) (8), suggesting the same ancestry of this rare haplotype. Due to detection limits of Southern blot analysis (SB), detection of ratios 2:3 as present for CYP21A duplications can be missed. Moreover, complex CYP21 genotypes such as CYP21A2 and CYP21A1P pseudogene duplications in trans can mask CYP21A2 duplications in SB and semiquantitative PCR. Sequence analyses provide such information in subjects carrying three nucleotides at the respective locus. This has been shown for the haplotype carrying the IVS2 splice mutation (G instead of A/C) on one and the polymorphism (A/C) on the two other CYP21A2 genes (9). Due to the limited number of publications (10, 11) focusing on this methodological challenge, it remains speculative by which in-house developed methods (including quantitative and semiquantitative PCR and SB) and to which extent such duplicated CYP21A2 genes are detected in the course of worldwide performed CYP21A2 genotyping. The present study analyzes 38 carriers of Q318X mutations with respect to duplicated CYP21A2 genes using a recently developed commercially available multiplex ligation-dependent probe amplification (MLPA) assay. Additionally, single-nucleotide polymorphisms (SNPs), recently reported to be associated with such rare duplicated CYP21A2 genes (8), as well as HLA haplotypes were determined to evaluate a common ancestry of this duplicated CYP21A2 haplotype.

Subjects and Methods

Subjects

In the course of routine CYP21 genotyping, 38 unrelated individuals were identified carrying a Q318X mutation on one functional CYP21A2 gene. Blood samples from these 38 individuals had been sent to our laboratories for CAH genotyping due to the assigned diagnosis of CAH (n = 17), hirsutism (n = 3), unfulfilled pregnancy (n = 6), and elevated testosterone levels (n = 2). Five individuals were partners of CAH patients, and five subjects were referred in the course of segregation analyses to detect/confirm a CAH carrier status.

In these above mentioned unrelated 38 individuals and in 11 family members with or without the Q318X mutation, HLA typing was performed.

Written informed consent for CYP21 genotyping was obtained from all individuals studied.

HLA typing

Low-resolution typing of HLA-A, -B, and -C alleles was performed by reverse sequence-specific oligonucleotide typing using a commercial kit (Dynal AutoReli; Dynal, Bromborough, UK). In case of ambiguities and if only one allele was detected, samples were additionally typed by sequence-specific primers (Olerup SSP Tray; QIAGEN, Austria, Vienna). When this single allele was confirmed, homozygosity was assumed.

CYP21A2 genotyping

Genomic DNA was extracted from peripheral blood leukocytes according to standard methods (12, 13).

Southern blot analysis

In brief, TaqI- and BglII-digested DNAs, immobilized on nylon membranes, were hybridized with a [³²P]dCTP- or Dig-dUTP-labeled CYP21A probe as published previously (12, 13).

Sequence analyses were performed by direct sequencing of three fragments (I–III), specifically amplified using selective PCR primers differentiating the functional CYP21A2 gene from the CYP21A1P pseudogene by the 8-bp deletion located in exon 3 of CYP21A1P, as described previously (9, 14–16).

MLPA was performed using the Salsa MLPA kit (P050B/B2-FAM; MRC-Holland, Amsterdam, The Netherlands) according to the manufacturer’s instructions. In brief, MLPA-PCR products were separated on an ABI 3130XL automatic sequencer, preanalyzed by Genemapper version 4.0 and evaluated by the Excel-based software CoffalyserV or SequencePilot (jsi), which normalize the data in relation to control peaks from other genes and convert the peak areas into bar graphs. With a normal gene dosage, the ratio of control and test bars lies at 1.0 (variance 0.8–1.2); with a hemizygous duplication of an exon, the corresponding peak is doubled; and with a heterozygous duplication, the peak should have a ratio of approximately 1.5.

Results

CYP21A2 genotyping

Large CYP21A2 gene duplication

In the majority (n = 32; 84.2%) of the 38 unrelated individuals carrying the Q318X mutation, a duplicated CYP21A2 gene was detected by MLPA. In 31 (81.6%) of the 32 individuals carrying the duplicated CYP21A2 gene, the duplication was detected in all exons (1, 3, 4, 6, and 8) covered by MLPA. For two of these Q318X carriers, MLPA analysis showed a duplication of both CYP21A2 genes. For those individuals where family analysis was possible (n = 11), the duplication could be traced to the Q318X-carrying CYP21A2 gene.

SB analysis, performed in 18 of the 32 individuals, confirmed the duplication in 16 individuals. In one patient exhibiting a CYP21A2 gene duplication in exons 1, 2, 4, 6, and 8 as shown by MLPA, such a constellation was not detected by SB (as depicted in Fig. 1). For one patient, in whom MLPA did not detect a duplicated CYP21A2 gene in all exons, it cannot be ruled out that a partial CYP21A2 deletion is present in addition to the Q318X mutation, the Val281Leu mutation, and the duplicated CYP21A1P pseudogene. CYP21A2 genotyping of the parents or siblings could possibly help to reveal such a rare constellation but was not possible. In that patient, reliable interpretation of the complex SB pattern due to the duplicated pseudogene was not possible. A similar problem was encountered in a patient carrying the IVS2, Ile172Asn, Cluster, Val281Leu, and Q318X mutations (indicative for a large conversion) as well as a duplicated CYP21A2 gene (detected by MLPA) and a duplication of the pseudogene (detected by SB and MLPA).

Fig. 1

Pitfall of SB analysis: discordant result in comparative analysis of one subject (S1) with a Q318X mutation by MLPA and SB. As depicted, a duplication of all exons (1, 3, 4, 6, and 8) of CYP21A2 was detected by MLPA, whereas SB analysis using TaqI and BglII restriction enzymes is not able to detect the CYP21A2 duplication but showed a normal pattern with identical intensity of the respective band in comparison with the internal control sample S2, the latter showing heterozygous deletion of the functional CYP21A2 gene (reduced intensity of the 3.7-kb TaqI and the 12-kb BglII band).

In all cases where family members (n = 11) were available, the Q318X mutation was associated with the allele carrying the duplicated gene.

CYP21A2-CAH allele

In five (13.2%) of the 38 subjects, the Q318X mutation was not associated with a duplicated CYP21A2 gene and thus represents a CAH allele.

CYP21A2 SNPs (IVS2-79G/A and 3′-UTR region +12c*C/T)

In the majority (n = 25) of the subjects with a duplicated CYP21A2 gene, the IVS2-79G/A (78.1%) and the +12c*C/T SNP (68.8%) were detectable in heterozygous form (Table 1). In three subjects, this analysis could not be performed. Four subjects (12.5%) carried the wild-type allele for the IVS2-79 SNP and seven (21.9%) that of the +12c*C/T SNP.

Table 1

Patients with Q318X-mutations (n = 38)

MLPA analysis				HLA, B50 Cw06		CYP21A2, IVS2-79G/A		CYP21A2, +12 C/T
Group	CYP21A2	n	%	n	%a	n	%a	n	%a
A	Duplication	32	84.2	22	68.8	25	78.1	22	68.8
						(3 ND)	9.4	(3 ND)	9.4
						(4 WT)	12.5	(7 WT)	21.9
B	No duplication	5	13.2			1	20	1	20

ND, Not determined; WT, wild type.

^a

In relation to the respective group A and B, respectively.

HLA haplotypes

Of the 32 subjects with CYP21A2 duplications, 22 (68.8%) had the HLA haplotype B*50-Cw*06, whereas no such haplotype was found in the five subjects without CYP21A2 duplication. All individuals with a Val281Leu mutation possessed the HLA-B*14-Cw*08 haplotype, being in line with previous reports.

Discussion

The present study shows that the majority (>;80%) of individuals presenting with a Q318X mutation are not carriers of a CAH allele but have the mutation on one of three functional CYP21A2 genes. In all cases where family analysis was feasible, the Q318X mutation could be traced to the duplicated CYP21A2 gene. Functional activity of duplicated CYP21A2 genes with a Q318X mutation has not been proven by in vitro expression studies so far. It is, however, deduced from the clinical observation [this and previous studies (9)] that all subjects with a Q318X mutation on the duplicated gene and a second pathogenic mutation in the third CYP21A2 gene in trans are clinically unaffected.

The first identification (6, 7) of such duplicated CYP21A2 haplotypes was based on complex restriction patterns of TaqI- and BglII-digested DNA hybridized with CYP21A and complement C4 probes (SB). This method is time consuming and requires high yields of DNA, resulting in less use in routine CYP21A2 genotyping. There is only a limited number of publications on alternative quantitative methods as real-time or semiquantitative PCR (10, 11) to assess the extent and conformity of detection of duplications in routine CYP21A2 genotyping.

Genotyping 38 individuals with Q318X aberrations using the commercially available MLPA assay, the present study establishes a link between previous and recent studies (6–8, 10). Parajes et al. (8) found a relatively high frequency (7%) of clinically unaffected Spanish individuals to carry such duplicated CYP21A2 genes with a Q318X mutation, whereas this haplotype was at first (6, 7) regarded as a rare one.

A high frequency of this haplotype in healthy individuals is confirmed by our finding that 84.2% of subjects with Q318X mutations carry this aberration on one allele of a duplicated CYP21A2 gene. Individuals with such a haplotype therefore are easily overlooked in screening studies, because two additional genes would have to be rendered inactive by two additional mutations. This may explain why this haplotype has hitherto been considered to be rare, particularly in CAH patients.

In that context, it is of note that the frequency of duplicated CYP21A2 genes in the subjects (without those carrying a Q318X mutation) so far genotyped in our laboratory is less than 1%, whereas a considerable number of these individuals carry deletions or duplications of the CYP21A1P pseudogene, as previously reported (13).

To avoid false-positive genotyping, it is of utmost importance to be aware of the high frequency of duplicated CYP21A2 genes in association with Q318X aberrations. In case of identification of a Q318X aberration, we suggest MLPA or SB analysis to clear the CYP21A2 gene copy number. If two CYP21A2 genes are found, the Q318X bearing allele represents a CAH allele. If three CYP21A2 genes are present, family analysis has to be performed to determine the localization of the Q318X mutation. If the Q318X mutation is on the allele carrying the duplicated gene, it does not represent a CAH but rather a functional normal allele. Such constellations have to be documented and addressed in the respective genetic reports to provide correct genetic counseling.

Obviously, the described real-time and semiquantitative PCR methods and SB as well as the recently developed commercially available MLPA assay allow identification of such duplicated CYP21A2 haplotypes in the majority of cases.

Nevertheless, the present work shows that in certain constellations of mutations with duplicated pseudogenes (in association with the Val281Leu mutation) or large conversions (recombinations of the functional with the pseudogene), the different methods are difficult to interpret and/or do not give consistent results. In that context, it is of note that SB does not allow exact determination of the extent of the deletion or duplication, whereas due to variations (SNPs) in the individual’s genomes, aberrant hybridization of single MLPA probe cannot be ruled out.

The majority (Table 1) of carriers of the duplicated CYP21A2 gene with the Q318X mutation were heterozygous for the SNPs in intron 2 (IVS2-79) and in the 3′-UTR (+12c*C/T), whereas all Spanish subjects with the duplicated haplotype carried those SNPs (8). Of note, 68.8% of the subjects with the duplicated CYP21A2/Q318X haplotype, but none of the five subjects with the Q318X-CAH allele exhibited the HLA-B*50-Cw*06 haplotype. The frequency of this HLA haplotype has been estimated in Italian and Spanish populations to be 1.4 and 2.3% (17), respectively, and thus represents a relatively rare haplotype in European Caucasoids. Our observation that 68.8% of the subjects with duplicated CYP21A2/Q318X carry HLA-B*50-Cw*06 suggest a common ancestor. Assuming a recombination frequency of 0.5% between the CYP and HLA-B genes and given the approximately 20-fold low frequency of HLA-B*50-Cw*06 haplotype in healthy European populations, this could thus represent an example for a founder effect.

In conclusion, the present study for the first time characterizes a relatively large number of subjects (n = 38) with a Q318X mutation and their family members (n = 11) employing a recently developed commercially available MLPA assay in parallel to sequence and SB analysis. So far, up to eight carriers of the Q318X mutation have been identified and characterized in a single population (8). Thus, to the best of our knowledge, the present work presents the highest number of Q318X carriers studied.

The observation that more than 80% of these individuals are carriers of a duplicated CYP21A2 gene stresses the importance of considering and detection of such haplotypes for genetic CAH diagnosis and counseling.

Acknowledgments

We thank Ulrike Cizek, Corinna Eberle, Ingrid Faé, Angelika Freudenthaler, and Silke Straunik for expert technical assistance.

Disclosure Summary: The authors have nothing to disclose.

Abbreviations

• CAH

  Congenital adrenal hyperplasia

• HLA

  human leukocyte antigen

• MLPA

  multiplex ligation-dependent probe amplification

• SB

  Southern blot

• SNP

  single-nucleotide polymorphism

• UTR

  untranslated region