- Email: [email protected]
21-Hydroxylase deficiency: Mutational spectrum and Genotype–Phenotype relations analyses by next-generation sequencing and multiplex ligation-dependent probe amplification
European Journal of Medical Genetics xxx (xxxx) xxxx
Contents lists available at ScienceDirect
European Journal of Medical Genetics journal homepage: www.elsevier.com/locate/ejmg
21-Hydroxylase deficiency: Mutational spectrum and Genotype–Phenotype relations analyses by next-generation sequencing and multiplex ligationdependent probe amplification Ihsan Turana, Mehmet Tastanb, Duygu D. Bogac, Fatih Gurbuzb, Leman D. Kotanb, Abdullah Tulic, Bilgin Yükselb,∗ a b c
Sanlıurfa Training and Research Hospital, Clinic of Pediatric Endocrinology, Sanlıurfa, Turkey Cukurova University, Faculty of Medicine, Division of Pediatric Endocrinology, Adana, Turkey Cukurova University, Faculty of Medicine, Division of Medical Biochemistry, Adana, Turkey
A R T I C LE I N FO
A B S T R A C T
Keywords: cyp21a2 Congenital adrenal hyperplasia 21-Hydroxylase CAH
Congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency (21OHD) is autosomal recessive disorder of cortisol biosynthesis. Genetic defects in CYP21A2 cause 21OHD. The aim of this study was to determine spectrum of mutations in CYP21A2 in a large cohort and analyze the genotype-phenotype correlation to assess predictive characteristics of genotype. We investigated a total of 113 patients with 21OHD. Next-generation sequencing and Multiplex ligation-dependent probe amplification of the CYP21A2 gene were performed in patients and their parents. The genotypes were categorized into Groups 0, A, B, and C according to the residual 21-hydroxylase activities. In this study, the group A was divided into two subgroups as A1 and A2. Three novel variants were found. The genotype–phenotype correlation of the mutation classification was 91.5%. Positive predictivity of subgroups A1 was higher than groups A and subgroups A2. Our study reports genotype–phenotype correlations in the largest 21OHD cohort in Turkey. This correlation sustained when we analyzed our data in combination with metadata from other published studies. This study confirms that CYP21A2 genotyping with next-generation sequencing and MLPA can accurately and reliably confirm the diagnosis of 21OHD. We propose a new classification by dividing group A into two new subgroups to better predict the phenotype. In light of this very high genotype-phenotype correlation, with their ever-increasing availability, declining cost, and turnaround time, we propose that molecular genetic studies can be more economical and practical alternative to the current initial diagnostic laboratory studies based on assays of intermediary steroid metabolites.
1. Introduction Congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency (21OHD) is autosomal recessive disorder of cortisol biosynthesis. Genetic defects in CYP21A2 cause 21OHD and represents 95–99% of all CAH cases (Arlt et al., 2010; Gidlof et al., 2013; Speiser et al., 2018). Enzymatic deficiency of steroid 21-hydroxylase is characterized by impairment of cortisol biosynthesis (with or without impairment of aldosterone biosynthesis) and excessive androgen synthesis that leads to the clinical manifestations of adrenal insufficiency and hyperandrogenism. Phenotypically, CAH can be divided into classical, severe, and non-classical (NC) forms. Classic CAH occurs in 1:13,000 to 1:15,000 live births and is represented by two phenotypes as simple virilizing (SV) and salt-wasting (SW) (Parsa and New, 2017). The SW
∗
form, the most severe form, is caused by the severe impairment of 21hydroxylase enzyme (< 2% enzymatic activity), leading to renal salt loss due to the lack of aldosterone as well as pre- and postnatal virilization, due to reactive androgen overproduction. In the SV form, patients present with variable degrees of virilization, but, they do not usually have a renal salt loss. Finally, NC form is generally asymptomatic at a neonatal period and leads to pseudo precocious puberty, hirsutism, acne, or subfertility in later childhood or adolescence (Parsa and New, 2017; Nordenstrom and Falhammar, 2018). The 21-hydroxylase gene (CYP21A2), encoding steroid 21-hydroxylase enzyme, is located HLA class III region on the chromosome 6 (6p21.3), close to the nonfunctional pseudogene (CYP21A1P). CYP21A2 and CYP21A1P genes nucleotide sequences are 98% identical in exons and approximately 96% identical in noncoding sequence
Corresponding author. E-mail address: [email protected] (B. Yüksel).
https://doi.org/10.1016/j.ejmg.2019.103782 Received 19 March 2019; Received in revised form 17 September 2019; Accepted 1 October 2019 1769-7212/ © 2019 Published by Elsevier Masson SAS.
Please cite this article as: Ihsan Turan, et al., European Journal of Medical Genetics, https://doi.org/10.1016/j.ejmg.2019.103782
European Journal of Medical Genetics xxx (xxxx) xxxx
I. Turan, et al.
Table 1 Allelic frequency of CYP21A2 gene by clinical form in 113 patients with 21OHD. Nucleotide changea
c.92c > t c.293-13C > G c.518T > A c.332_339del c.844G > T c.850A > G c.955C > T c.1069C > T c.710T > A; c.713T > A; c.719T > A Deletion/Large gen conversion Undetected Novel mutations c.20dup c.254del c.961_963del Alleles with multiple mutations c.1069C > T; c.1360C > T c.844G > T, c.1069C > T 706_713del8; c.844G > T c.955C > T; c.1069C > T c.293-13C > G; c.518T > A; c.955C > T c.293-13C > G; c.955C > T c.844G > T, c.955C > T c.518T > A; c.710T > A; c.713T > A; c.719T > A c.518T > A; c.844G > T c.1019G > A; c.1360C > T Total
Protein changea
Number of alleles (% of alleles)
p.Pro31Leu p.? p.Ile173Asn p.Gly111Valfs*21 p.Val282Leu p.Met284Val p.Gln319Ter p.Arg357Trp p.Ile237Asn; p.Val238Glu; p.Met240Lys (E6 cluster)
p.Leu8Alafs*72 p.Lys85Serfs*59 p.Glu321del p. Arg357Trp; p.Pro454Ser p.Val282Leu; p.Arg357Trp p.Gly111Valfs*21; p.Val282Leu p.Gln319Ter; p.Arg357Trp p.?; p.Ile173Asn; p.Gln319Ter p.?; p.Gln319Ter p.Val282Leu; p.Gln319Ter p.Ile173Asn; p.Ile237Asn; p.Val238Glu; p.Met240Lys (E6 cluster) p.Ile173Asn; p.Val282Leu p. Arg340His; p.Pro454Ser
Total
SW
SV
NC
2 (0.8) 87 (38.4) 18 (7.9) 10 (4.4) 7 (3.0) 1 (0.4) 26 (11.5) 21 (9.2) 4 (1.7) 18 (7.9) 4 (1.7)
– 65 (39.6) 1 (0.6) 10 (6.0) – – 22 (13.4) 20 (12.1) 4 (2.4) 18 (10.9) –
– 22 (45.8) 16 (33.3) – 1 (2.0) – 3 (6.2) 1 (2.0) – – 1 (2.0)
2 – 1 – 6 1 1 – – – 3
2 (0.8) 1 (0.4) 1 (0.4) 24 (10.6) 1 (0.4) 3 (1.3) 1 (0.4) 2 (0.8) 8 (3.5) 3 (1.3) 1 (0.4) 2 (0.8) 2 (0.8) 1 (0.4) 226
2 (1.2) – 1 (0.6) 21 (12.6) 1 (0.6) 3 (1.8) 1 (0.6) 2 (1.2) 8 (4.8) 3 (1.8) 1 (0.6) 2 (1.2) – – 166
– 1 (2.0) – 3 (6.2) – – – – – – –
– – –
2 (4.1) 1 (2.0) 48
(14.2) (7.1) (42.8) (7.1) (7.1)
(21.4)
– – – – – – – – – – 14
a
The codingDNA, and the protein sequence level are shown (RefSeq for DNA:NC_000006.12; cDNA: NM_000500.9; protein: NP_000491.4, most common mutation listed in bold.
patient's medical records by two pediatric endocrinologist (I.T, B.Y) for all cases. The SW form was characterized with salt-wasting crisis (dehydration, vomiting, hyponatremia, and hyperkalemia), ambiguous external genitalia in females, elevated levels of serum 17-hydroxyprogesterone and plasma renin activity (PRA) or plasma renin concentration at diagnosis or during a later evaluation. SV form was defined as signs and symptoms of hyperandrogenism, elevated serum 17hydroxyprogesterone, but without evidence of SW during the newborn period. NC CAH was diagnosed in both sexes diagnosed by precocious pubarche and adrenarche or hirsutism and mildly elevated levels of serum 17OHP (Speiser et al., 2018; Nordenstrom and Falhammar, 2018; White and Speiser, 2000).
(White et al., 1986). Molecular genetic analysis has shown that more than 90% of these mutations result from the transfer of deleterious micro sequences normally present in the CYP21A1P pseudogene into the functional CYP21A2 gene (Speiser et al., 2018). To date, more than 200 mutations have been described around the world. Most recently, Concolino and Costella listed 238 pathogenic variants reported up to now (Concolino and Costella, 2018). Clinical form depending on the degree of impairment of 21-hydroxylase enzyme activity caused by the mutations in the CYP21A2. Variants on CYP21A2 are classified into four groups as Null, A, B and C. Null variants displayed 0% enzyme activity in vitro assays, group A variants, the c.29313C > G mutation, conserve a minimal (< 1%) residual activity. Null and A are associated with the SW form. Finally, group B and C variants show 1–5% and 20–50% enzyme activity, respectively and are related to the SV and NC form of the disease, respectively (Concolino and Costella, 2018; Speiser et al., 1992; Wedell et al., 1994; Loke, 2008). In this study, we present the result of mutation analyses of CYP21A2 in a large cohort of 113 patients with CAH due to 21OHD and their 165 parents. Additionally, the genotype-phenotype correlation was analyzed systematically to determine predictive characteristics of variant groups of CYP21A2.
2.2. Molecular analysis of CYP21A2 DNA samples were obtained with the isolation from 200 μl of blood sample from each individual, by using QIAamp DNA Blood Mini Kit (Qiagen Inc.). Primer designs were carried out for the coding regions of CYP21A2. In designing primers, the 8-bp deletion region in exon 1 was selected to discriminate between CYP21A2 and its pseudogene, so that sequencing results would not contain data from the allele which contains 8-bp deletion and impossible to be a functional allele. The gene was designed to amplify in two amplicons; one expands from the beginning of the gene to exon 3 and the other one from exon 3 to the end of the gene, so that one of the primers of each amplicon resides on the 8-bp deletion region. Two PCRs were carried out on isolated DNA samples, by using the designed primers and the reactions were checked by using 2% agarose gel electrophoresis. PCRs belong to each individual were mixed to obtain PCR pools, which have both of the amplicons of each individual in one tube. While mixing, the amplification efficiency and length of the amplicons were taken into consideration; the volume for each PCR is directly proportional to the length of the amplicon and inversely
2. Material and methods 2.1. Patients We investigated a total of 113 patients with 21OHD from 99 unrelated families. These patients were treated in the Cukurova University, Faculty of Medicine, Division of Pediatric Endocrinology during the period of 2007 through 2017. The Ethics Committee of the Cukurova University Faculty of Medicine approved this study. Informed consents were obtain from all participants or their legal guardians. Clinical forms of CAH due to 21OHD were determined based on clinical and hormonal criteria and a retrospective review of each 2
European Journal of Medical Genetics xxx (xxxx) xxxx
I. Turan, et al.
Table 2 Genotype phenotype correlation. %a
Genotype Allele 1 Group 0 deletion or large gene conversion deletion or large gene conversion deletion or large gene conversion deletion or large gene conversion p.Gly111Valfs*21 p.Gly111Valfs*21 p.Gly111Valfs*21 p.Gly111Valfs*21 p.Gly111Valfs*21; p.Val282Leu p.Gln319Ter p.Gln319Ter p.Gln319Ter p.Gln319Ter p.Gln319Ter; c.293-13C > G; p.Ile173Asn p. Arg357Trp p. Arg357Trp; p.Val282Leu E6 cluster E6 cluster/p.Ile173Asn p. Leu8Alafs*72 Group A Group A1 c.293-13C > Ga c.293-13C > G c.293-13C > G c.293-13C > G Group A2 c.293-13C > G Group B p.Ile173Asn p.Ile173Asn p.Ile173Asn p.Ile173Asn p.Ile173Asn p.Ile173Asn; p.Val282Leu Group C p.Val282Leu p.Val282Leu p.Val282Leu p.Pro31Leu Group D c.293-13C > G p.Val282Leu p.Ile173Asn Group E c.293-13C > G p.Gln319Ter p.Val282Leu
Allele 2
Number of allele
P.P
Total
SW
SV
NC
p.Val282Leu c.293-13C > G p.Ile173Asn p.Pro31Leu
39.8 5.3 0.8 1.7 0.8 2.6 0.8 0.8 0.8 0.8 5.3 0.8 0.8 0.8 3.5 7.9 0.8 1.7 0.8 0.8 39.8 7.9 1.7 0.8 4.4 0.8 31.8 31.8 10.6 3.5 0.8 2.6 0.8 1.7 0.8 3.5 0.8 0.8 0.8 0.8
44 6 1 2 1 3 1 1 1 1 6 1 1 1 4 9 1 2 1 1 45 9 2 1 5 1 36 36 13 4 1 3 1 3 1 4 – – – 1
44 6 1 2 1 3 1 1 1 1 6 1 1 1 4 9 1 2 1 1 36 9 2 1 5 1 27 27 1 – – – – 1 – – – – – –
– – – – – – – – – – – – – – – – – – – – 9 – – – – – 9 9 12 4 1 3 1 2 1 – – – – –
– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – 4 1 1 1 1
p.Glu321del p.Met284Val p.R340H; p.454S
0.8 0.8 0.8
1 1 1
1 – –
– – 1
– 1 –
undetected undetected undetected
0.8 0.8 0.8
1 1 2
– – –
1 – –
– 1d 2e
deletion/large gene conversion p.Gln319Terc p.Gln319Ter; c.293-13C > G p. Arg357Trp; p.Pro454Ser p.Gly111Valfs*21 p.Gln319Ter p.Gln319Ter; c.293-13C > G p.Gln319Ter; p. Arg357Trp p. Arg357Trp; p.Val282Leu p.Gln319Ter p. Arg357Trp p.Gln319Ter; p. Arg357Trp p.Gln319Ter; p.Val282Leu p.Gln319Ter; c.293-13C > G; p.Ile173Asn p. Arg357Trp p. Arg357Trp; p.Val282Leu E6 cluster E6 cluster/p.Ile173Asn p. Leu8Alafs*72
deletion or large gene conversion p.Gly111Valfs*21 p.Gln319Ter p. Arg357Trpc c.293-13C > G p.Ile173Asn p. Arg357Trp p.Gln319Ter p.Lys85Serfs*59 c.293-13C > G p.Ile173Asn; p.Val282Leu
b
100%
80% 100%
75% 92.3%
100%
NA
NA
The codingDNA, and the protein sequence level are shown (RefSeq for DNA:NC_000006.12; cDNA: NM_000500.9; protein: NP_000491.4). NA: not applicable. a % frequency of genotype. b Positive predictivity for expected phenotype. c One allele is de novo mutation. d Patient carrying duplication of CYP21A2. e One of two patient carrying duplication of CYP21A2.
MLPA (Multiplex ligation-dependent probe amplification) was used for copy number and gene conversion, analyses (MRC-Holland b.v., SALSA MLPA P050 CAH probemix). Data from both sequencing and MLPA were used collaboratively to do the interpretation. In any case of contradiction between sequencing and MLPA, sequencing data was prioritized for sequencing variations and MLPA data was prioritized for copy number and gene conversions.
proportional to the efficiency of the reaction, which were estimated with the help of gel electrophoresis. The PCR pools for each individual were purified using NucleoFast® 96 PCR kit (MACHEREY-NAGEL GmbH). The purified pools were quantified using ND1000 (Thermo Fisher Scientific Inc.) micro volume spectrophotometer and standardized to 0,2 ng/ul, which was needed for sample preparation step. The samples were got ready for next-generation sequencing by using NexteraXT sample preparation kit (Illumina Inc.). Next-generation sequencing of the samples were carried out by using Miseq system (Illumina Inc.). The data were analyzed on IGV 2.3 software (Broad Institute). NC_000006.12 (NM_00500.9) was used for reference sequence of CYP21A2.
2.3. Analyses of novel variants We searched for all identified variants in the Human Gene Mutation Database (Stenson et al., 2003), LOVD (Fokkema et al., 2011), and the 3
European Journal of Medical Genetics xxx (xxxx) xxxx
I. Turan, et al.
Table 3 Genotype phenotype correlation as combination with metadata from other published studies. A1
A2
GG
PP
SW
SV
NC
Turkey*
SW
SV
NC
Netherlands (Stikkelbroeck et al., 2003)
SW
SV
NC
SW
SV
NC
Middle Europe (Dolzan et al., 2005)
Germany (Krone et al., 2000)
SW
SV
NC
USA (Finkielstain et al., 2011)
SW
SV
NC
Brazil (de Carvalho et al., 2016)
SW
SV
NC
To
Croatia (Dumic et al., 2017)
Total
SW
SV
NC
%P.
0
0
0
SW
44
–
–
28
1
–
80
1
1
32
–
–
35
5
–
55
4
–
28
–
**
269
257 11
1
95.5
A A A
A/0 0 A
A A1 A2
SW SW SW
35 9 27
9 – 9
– –
21 14 7
1 1 –
– – –
108 44 64
6 1 5
1 1 –
43 25 18
5 3 2
– – –
60 52 8
8 7 1
– – –
66 43 23
36 11 25
– – –
29 13 16
5 3 2
** ** **
434 227 207
363 70 200 26 163 44
1 1 –
83.6 88.1 78.7
B
B/A/0
B
SV
1
12
–
2
10
4
24
65
5
13
37
–
2
23
3
–
53
1
4
12
**
259
46
13
77.2
200
-A1 and A2: mutation group of allele 1 and 2, respectively. GG: group of genotype. PP: predictive phonotype. SW: salt wasting. SV: simple virilization. NC: nonclassic. To: total. %P.: positive predictive value. *this study. **patients with non-classical CAH were not included in Croatia's study. A/0: A or 0. B/A/0: B or A or 0. -Mutation group 0: gene deletions, large gene conversion, E6 cluster, p. Arg357Trp, p.Gln319Ter, p.L308Ffs*6, p.Gly111Valfs*21, novel frameshift mutation and multiple mutations alleles containing any of these mutations. Mutation group A: c.293-13C > G and multiple mutations alleles containing c.293-13C > G but not mutations of group 0. Mutation group B: p.Ile173Asn and promoter conversion + Pro31Leu. Mutation group C: p.Pro31Leu, p.Val282Leu, p.Pro454Ser. -Genotypes that do not fit this classification in the articles were excluded from the estimation.
2.5. Statistical analysis
recent review article by Concolino and Costella (2018). Date of check was march 13th 2019. Novel variants were classified based on the 2015 American College of Medical Genetics and Genomics and Association for Molecular Pathology guidelines (ACMG-AMP) (Richards et al., 2015).
The Pearson Chi-Square tests were used to compare the groups and the odds ratio (OR) with 95% confidence interval (CI) were calculated. The positive predictive values were determined as the percentage of patients with the expected phenotype for each group. The Statistical Package for Social Sciences 20 (SPSS20) was used to carry out the statistical analyses.
2.4. Classification in mutation groups and genotype-phenotype relation To help predict the severity of the disease-causing mutations, genotypes of patients were divided into five groups according to their residual 21-hydroxylase enzyme activity, based on prior in vitro studies (Speiser et al., 1992; Wedell et al., 1994). The group 0 included patients who carry homozygous null mutations that result in completely inactive enzymes (gene deletions, large gene conversion, E6 cluster, p. Arg357Trp, p.Gln319Ter, p.L308Ffs*6, p.Gly111Valfs*21, novel frameshift mutation and multiple mutations alleles containing any of these mutations) (Speiser et al., 1992). The Group A included patients homozygous for c.293-13C > G (which has minimal residual enzymatic activity) (Higashi et al., 1988) or compound heterozygous for c.293-13C > G and a null mutation. Additionally, in this study, the group A was divided into two subgroups as A1 for compound heterozygous mutations and A2 for homozygous mutations. The Group B included the mutation I172N (~2% residual enzymatic activity) (Krone et al., 2005) and the promoter conversion + Pro31Leu, homozygous or compound heterozygous with a null or group A mutation. The group C contained mutations P31L, V282L, and P454S (~20–60% residual enzyme activity) (Speiser and New, 1987), homozygous or compound heterozygous with a null, A, or B mutation. The Group D included patients with novel mutations or variants whose enzymatic activity impairment had not been assessed. Finally, the group E included patients with at least one allele without any identified mutations. In this classification, the mutation with the mildest effect on enzymatic activity (in the compound heterozygous composition) determines the predicted phenotype. In alleles containing multiple mutations, the most deleterious mutation determines the genotype groups (Dolzan et al., 2005). Groups 0 and A were predicted to result in the SW form. The expected phenotype was SV for Group B, and NC for Group C mutations. Positive predictive value was calculated for groups A, B, and C. We concomitantly analyzed six previously published studies similar to our study while examining the genotype-phenotype relations (Dolzan et al., 2005; Stikkelbroeck et al., 2003; de Carvalho et al., 2016; Krone et al., 2000; Finkielstain et al., 2011; Dumic et al., 2017). Mutations that do not fit this classification in the articles were excluded from the estimations.
3. Results 3.1. Phenotypes and mutational analyses of patients In total, from 113 patients with 21OHD (82 SW, 24 SV, and 7 NC), 226 alleles were analyzed for the CYP21A2 gene variants. Seventy-five patients (66%) were homozygous and 34 patients (30%) were compound heterozygous. The parents of all patients carrying compound heterozygous variants (except two) were analyzed. Overall, 160 parents were genotyped and only (1.8%) of them were without variants of their offspring (i.e de novo mutations c.293-13C > G, p. Arg357Trp, p.Gln319Ter). Parents of six of 75 patients (8%) with homozygous mutations denied consanguinity. Four patients whose parents were consanguineous were carried compound heterozygous mutations. The CYP21A2 duplication was detected in 11 patients (9.7%). The homozygous c.293-13C > G was the most common genotype in both classical forms, SW and SV. The most frequent mutation was c.293-13C > G (37.1%) followed by p.Gln319Ter (11.0%), large conversation/deletion (9.2%), p. Arg357Trp (9.2%), and p.Ile173Asn (7.9%). In this study. ten different combinations of point mutations were detected; the allele frequency of this complexity was found in 10.6%. In four patients (one SV, three NC), mutations were detected only in one allele and two of them were CYP21A2 duplication. Our cohort contained the full spectrum of mutations shown in Table 1. 3.2. Novel variants Three variants that have not been described previously were identified. One patient was homozygous for the novel frameshift mutation p. Leu8Alafs*72 (c.20dup) that led to the SW form. A deletion of the glutamic acid at codon 320 or 321 (c.961_963del) was detected in trans with c.293-13C > G in one patient with the SW form. One patient with the SV form was compound heterozygous for p.Lys85Serfs*59 (c.254del) and p.Ile173Asn. All these novel variants were considered ‘likely pathogenic’ based on the 2015 ACMG/AMP guideline. New variants were submitted to ClinVar. Accession numbers is 4
European Journal of Medical Genetics xxx (xxxx) xxxx
I. Turan, et al.
variants: two frameshift variants and one in-frame deletion. Functional analyses of these variants were not performed. The p. Leu8Alafs*72 and the p.Lys85Serfs*59 frameshift variants are likely to cause nonsense mediated decay resulting in no residual enzymatic activity. The p.Glu321del was detected in trans with c.293-13C > G in one patient with the SW form. The glutamic acid at 320 or 321 is highly conserved among different P450 cytochromes in 30 different species; located in a negatively charged sequence, it probably interacts with part of positively charged residues (Tardy et al., 2010; Kent et al., 2002). These may explain why this inframe deletion can significantly affect the enzyme activity. An accurate and reliable genotype-phenotype association in 21OHD will help guide the treatment management such as glucocorticoid and mineralocorticoid requirement. This study largely confirmed that there is a good positive genotype-phenotype correlation in 21OHD. But we note that the prediction of severe genotypes is more reliable than milder genotypes. This classification is useful and acceptable because it provides predictions to 96.4% of the cases (Dolzan et al., 2005; Stikkelbroeck et al., 2003; Krone et al., 2000; Finkielstain et al., 2011; Dumic et al., 2017). The genotype-phenotype correlation was the best for group 0, when we refined it by including other studies (Dolzan et al., 2005; Stikkelbroeck et al., 2003; de Carvalho et al., 2016; Krone et al., 2000; Finkielstain et al., 2011; Dumic et al., 2017). The c.293-13C > G mutation, known to be associated with minimal residual enzymatic activity may cause discordance of phenotype. The prediction was perfect when this splice site mutation was detected in trans with null mutation. In this study, the group A was divided into two subgroups: A1 for compound heterozygous and A2 for homozygous mutations. When the data of a total of 434 patients from seven studies were collectively analyzed, there was a statistically significant difference in the two subgroups for positive predictivity of a distinct and segregated phenotype. These findings led us c.293-13C > G is more deleterious than homozygous genotype when trans with null mutations. Additionally, predictivity of new subgroup A1 is more reliable. This new classification improves genotype-phenotype relationships, therefore enhances the clinical utility of CYP21A2 mutational analyses. This finding may be particularly valuable for ethnic groups where the compound heterozygote genotypes are more common. The developing clarity of genotype-phenotype relationship and increasing availability of molecular genetic studies will decrease unnecessary repeated hormonal tests and unneeded medical treatments. This will have both medical and economic benefits. In conclusion, our study reports genotype–phenotype correlations in the largest 21OHD cohort in Turkey. This study confirms that CYP21A2 genotyping with next-generation sequencing and MLPA can accurately and reliably confirm the diagnosis of 21OHD. The spectrum of mutations in the CYP21A2 gene was investigated in Southern Turkey. We confirmed a good genotype-phenotype correlation in another ethnic group. We propose a new classification by dividing group A into two new genotype subgroups to better predict the phenotype.
VCV000623477, VCV000623473, and VCV000623476 for c.254del, c.20dup, and c.961_963del respectively. 3.3. Genotype-phenotype relationship In this study, the genotype–phenotype correlation of mutation classification was 91.5%. A complete genotype-phenotype concordance was observed in Group 0. Genotypes correctly predicted phenotypes in 79.5% of the patients in group A; 100% in subgroup A1, and 75% in subgroup A2. Genotype-phenotype discordance occurred in 9 patients with group A and all of them included in subgroup A2. The genotypephenotype relation of our cohort was detailed in Table 2. The overall, in seven studies (including our study), the highest positive predictive value among all groups is in the group 0 with 95%. Genotype correctly predicted phenotype in 83.6% of the patients in group A, 88.1% in subgroup A1, and 78.7% in subgroup A2. Positive predictivity of subgroups A1 was higher than groups A and subgroups A2. This difference is statistically significant with p value < 0.05. Between subgroup A1 and A2, OR was 2 (95% CI: 1.19–3.37). The genotype-phenotype relation of seven studies was detailed in Table 3. 4. Discussion How important is the molecular genetic analysis in the diagnosis and management of 21OHD? In clinical practice, starting steroid treatment without a confirmed diagnosis is often inevitable. In some parts of the world baseline steroid levels may not be readily available due to their high costs among other factors. With the ever-increasing availability, reducing cost, and shortening turnaround time, genetic mutation analyses of CYP21A2 can be more practical and economical initial diagnostic modality, provided there is an established and reliable genotype-phenotype correlation. This study was the first large cohort study using next-generation sequencing as the surrogate diagnostic method, which detected 222 of 226 alleles of the affected patients. The diagnostic sensitivity was 98.2% (95.8% in SV and 100% in the SW form). This study confirms that CYP21A2 genotyping with next-generation sequencing and MLPA can accurately and reliably confirm the diagnosis of 21OHD. Homozygous genotype frequency (70%) was markedly higher in our study than other analyzed studies (Stikkelbroeck et al., 2003; de Carvalho et al., 2016; Krone et al., 2000; Finkielstain et al., 2011; Dumic et al., 2017) (21–35%) except for a Middle European (69%) report (Dolzan et al., 2005). This is an expected result for populations with high consanguinity rates. Because both parents are carriers of the same deleterious variant in the CYP21A2 gene. Interestingly, the four patients whose parents were related carried compound heterozygous mutations while the parents of six patients with homozygous mutations denied consanguinity. Although, the history of consanguinity is a valuable clue in genetically inherited diseases like 21OHD, these inconsistencies should be kept in mind. The finding that the c.293-13C > G was the most prevalent mutation in Turkey according to Bas et al. (Bas et al., 2009) was confirmed by our larger cohort study. The most common mutation in other large cohorts was also c.293-13C > G. The only exception was the p.Val282Leu in de Carvalho et al. (de Carvalho et al., 2016) study, which included a very high percentage of 206 (42%) NC patients. Mutation frequencies of affected alleles in this study were closely similar to those published large cohorts study in Turkey (Bas et al., 2009). As differently, a high frequency (10.6%) of alleles with multiple mutations was found. This may be due to superior technique of the genetic analyses used in this study. Compared with other ethnic groups, the frequencies of the p. Arg357Trp and the p.Gln319Ter were higher, therefore, the SW/SV form ratio was higher in our study (Dolzan et al., 2005; Krone et al., 2000; Finkielstain et al., 2011; Dumic et al., 2017; Bas et al., 2009). Genetic analyses of CYP21A2 in this study identified three novel
Acknowledgments This work was supported by the Scientific and Technological Research Council of Turkey (TÜBİTAK, Project No.108S236). We would like to thank Dr. A. Kemal Topaloglu for his valuable discussions on the manuscript. We are grateful to Fatma Dereli (Intergene Genetic Center, Ankara) and Haldun Dogan (Intergene Genetic Center, Ankara). References Arlt, W., Willis, D.S., Wild, S.H., Krone, N., Doherty, E.J., Hahner, S., Han, T.S., Carroll, P.V., Conway, G.S., Rees, D.A., Stimson, R.H., Walker, B.R., Connell, J.M., Ross, R.J., 2010. United Kingdom Congenital Adrenal Hyperplasia Adult Study E: health status of adults with congenital adrenal hyperplasia: a cohort study of 203 patients. J. Clin. Endocrinol. Metab. 95, 5110–5121. Bas, F., Kayserili, H., Darendeliler, F., Uyguner, O., Gunoz, H., Yuksel Apak, M., Atalar, F.,
5
European Journal of Medical Genetics xxx (xxxx) xxxx
I. Turan, et al.
virilizing forms of congenital adrenal hyperplasia due to 21-hydroxylase deficiency. J. Clin. Endocrinol. Metab. 90, 445–454. Loke, K.Y., 2008. Clinical applications of molecular genetics: the model of congenital adrenal hyperplasia. Ann. Acad. Med. Singapore 37 18-14. Nordenstrom, A., Falhammar, H., 2018. Management OF endocrine disease: diagnosis and management of the patient with non-classic CAH due to 21-hydroxylase deficiency. Eur. J. Endocrinol. 180 (3), 127–145. https://doi.org/10.1530/EJE-18-0712. Parsa, A.A., New, M.I., 2017. Steroid 21-hydroxylase deficiency in congenital adrenal hyperplasia. J. Steroid Biochem. Mol. Biol. 165, 2–11. Richards, S., Aziz, N., Bale, S., Bick, D., Das, S., Gastier-Foster, J., Grody, W.W., Hegde, M., Lyon, E., Spector, E., Voelkerding, K., Rehm, H.L., Committee, A.L.Q.A., 2015. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of medical genetics and genomics and the association for molecular Pathology. Genet. Med. 17, 405–424. Speiser, P.W., New, M.I., 1987. Genotype and hormonal phenotype in nonclassical 21hydroxylase deficiency. J. Clin. Endocrinol. Metab. 64, 86–91. Speiser, P.W., Dupont, J., Zhu, D., Serrat, J., Buegeleisen, M., Tusie-Luna, M.T., Lesser, M., New, M.I., White, P.C., 1992. Disease expression and molecular genotype in congenital adrenal hyperplasia due to 21-hydroxylase deficiency. J. Clin. Investig. 90, 584–595. Speiser, P.W., Arlt, W., Auchus, R.J., Baskin, L.S., Conway, G.S., Merke, D.P., MeyerBahlburg, H.F.L., Miller, W.L., Murad, M.H., Oberfield, S.E., White, P.C., 2018. Congenital adrenal hyperplasia due to steroid 21-hydroxylase deficiency: an endocrine society clinical practice guideline. J. Clin. Endocrinol. Metab. 103, 4043–4088. Stenson, P.D., Ball, E.V., Mort, M., Phillips, A.D., Shiel, J.A., Thomas, N.S., Abeysinghe, S., Krawczak, M., Cooper, D.N., 2003. Human Gene Mutation Database (HGMD): 2003 update. Hum. Mutat. 21 (6), 577–581. https://doi.org/10.1002/humu.10212. Stikkelbroeck, N.M., Hoefsloot, L.H., de Wijs, I.J., Otten, B.J., Hermus, A.R., Sistermans, E.A., 2003. CYP21 gene mutation analysis in 198 patients with 21-hydroxylase deficiency in The Netherlands: six novel mutations and a specific cluster of four mutations. J. Clin. Endocrinol. Metab. 88, 3852–3859. Tardy, V., Menassa, R., Sulmont, V., Lienhardt-Roussie, A., Lecointre, C., Brauner, R., David, M., Morel, Y., 2010. Phenotype-genotype correlations of 13 rare CYP21A2 mutations detected in 46 patients affected with 21-hydroxylase deficiency and in one carrier. J. Clin. Endocrinol. Metab. 95, 1288–1300. Wedell, A., Thilen, A., Ritzen, E.M., Stengler, B., Luthman, H., 1994. Mutational spectrum of the steroid 21-hydroxylase gene in Sweden: implications for genetic diagnosis and association with disease manifestation. J. Clin. Endocrinol. Metab. 78, 1145–1152. White, P.C., Speiser, P.W., 2000. Congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Endocr. Rev. 21, 245–291. White, P.C., New, M.I., Dupont, B., 1986. Structure of human steroid 21-hydroxylase genes. Proc. Natl. Acad. Sci. U. S. A. 83, 5111–5115.
Bundak, R., Wilson, R.C., New, M.I., Wollnik, B., Saka, N., 2009. CYP21A2 gene mutations in congenital adrenal hyperplasia: genotype-phenotype correlation in Turkish children. Journal of clinical research in pediatric endocrinology 1, 116–128. Concolino, P., Costella, A., 2018. Congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency: a comprehensive focus on 233 pathogenic variants of CYP21A2 gene. Mol. Diagn. Ther. 22, 261–280. de Carvalho, D.F., Miranda, M.C., Gomes, L.G., Madureira, G., Marcondes, J.A., Billerbeck, A.E., Rodrigues, A.S., Presti, P.F., Kuperman, H., Damiani, D., Mendonca, B.B., Bachega, T.A., 2016. Molecular CYP21A2 diagnosis in 480 Brazilian patients with congenital adrenal hyperplasia before newborn screening introduction. Eur. J. Endocrinol. 175, 107–116. Dolzan, V., Solyom, J., Fekete, G., Kovacs, J., Rakosnikova, V., Votava, F., Lebl, J., Pribilincova, Z., Baumgartner-Parzer, S.M., Riedl, S., Waldhauser, F., Frisch, H., Stopar-Obreza, M., Krzisnik, C., Battelino, T., 2005. Mutational spectrum of steroid 21-hydroxylase and the genotype-phenotype association in Middle European patients with congenital adrenal hyperplasia. Eur. J. Endocrinol. 153, 99–106. Dumic, K.K., Grubic, Z., Yuen, T., Wilson, R.C., Kusec, V., Barisic, I., Stingl, K., Sansovic, I., Skrabic, V., Dumic, M., New, M.I., 2017. Molecular genetic analysis in 93 patients and 193 family members with classical congenital adrenal hyperplasia due to 21hydroxylase deficiency in Croatia. J. Steroid Biochem. Mol. Biol. 165, 51–56. Finkielstain, G.P., Chen, W., Mehta, S.P., Fujimura, F.K., Hanna, R.M., Van Ryzin, C., McDonnell, N.B., Merke, D.P., 2011. Comprehensive genetic analysis of 182 unrelated families with congenital adrenal hyperplasia due to 21-hydroxylase deficiency. J. Clin. Endocrinol. Metab. 96, E161–E172. Fokkema, I.F., Taschner, P.E., Schaafsma, G.C., Celli, J., Laros, J.F., den Dunnen, J.T., LOVD, v, 2011. 2.0: the next generation in gene variant databases. Hum. Mutat. 32, 557–563. Gidlof, S., Falhammar, H., Thilen, A., von Dobeln, U., Ritzen, M., Wedell, A., Nordenstrom, A., 2013. One hundred years of congenital adrenal hyperplasia in Sweden: a retrospective, population-based cohort study. Lancet Diabetes Endocrinol 1, 35–42. Higashi, Y., Tanae, A., Inoue, H., Hiromasa, T., Fujii-Kuriyama, Y., 1988. Aberrant splicing and missense mutations cause steroid 21-hydroxylase [P-450(C21)] deficiency in humans: possible gene conversion products. Proc. Natl. Acad. Sci. U. S. A. 85, 7486–7490. Kent, W.J., Sugnet, C.W., Furey, T.S., Roskin, K.M., Pringle, T.H., Zahler, A.M., Haussler, D., 2002. The human genome browser at UCSC. Genome Res. 12, 996–1006. Krone, N., Braun, A., Roscher, A.A., Knorr, D., Schwarz, H.P., 2000. Predicting phenotype in steroid 21-hydroxylase deficiency? Comprehensive genotyping in 155 unrelated, well defined patients from southern Germany. J. Clin. Endocrinol. Metab. 85, 1059–1065. Krone, N., Riepe, F.G., Grotzinger, J., Partsch, C.J., Sippell, W.G., 2005. Functional characterization of two novel point mutations in the CYP21 gene causing simple
6
Contents lists available at ScienceDirect
European Journal of Medical Genetics journal homepage: www.elsevier.com/locate/ejmg
21-Hydroxylase deficiency: Mutational spectrum and Genotype–Phenotype relations analyses by next-generation sequencing and multiplex ligationdependent probe amplification Ihsan Turana, Mehmet Tastanb, Duygu D. Bogac, Fatih Gurbuzb, Leman D. Kotanb, Abdullah Tulic, Bilgin Yükselb,∗ a b c
Sanlıurfa Training and Research Hospital, Clinic of Pediatric Endocrinology, Sanlıurfa, Turkey Cukurova University, Faculty of Medicine, Division of Pediatric Endocrinology, Adana, Turkey Cukurova University, Faculty of Medicine, Division of Medical Biochemistry, Adana, Turkey
A R T I C LE I N FO
A B S T R A C T
Keywords: cyp21a2 Congenital adrenal hyperplasia 21-Hydroxylase CAH
Congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency (21OHD) is autosomal recessive disorder of cortisol biosynthesis. Genetic defects in CYP21A2 cause 21OHD. The aim of this study was to determine spectrum of mutations in CYP21A2 in a large cohort and analyze the genotype-phenotype correlation to assess predictive characteristics of genotype. We investigated a total of 113 patients with 21OHD. Next-generation sequencing and Multiplex ligation-dependent probe amplification of the CYP21A2 gene were performed in patients and their parents. The genotypes were categorized into Groups 0, A, B, and C according to the residual 21-hydroxylase activities. In this study, the group A was divided into two subgroups as A1 and A2. Three novel variants were found. The genotype–phenotype correlation of the mutation classification was 91.5%. Positive predictivity of subgroups A1 was higher than groups A and subgroups A2. Our study reports genotype–phenotype correlations in the largest 21OHD cohort in Turkey. This correlation sustained when we analyzed our data in combination with metadata from other published studies. This study confirms that CYP21A2 genotyping with next-generation sequencing and MLPA can accurately and reliably confirm the diagnosis of 21OHD. We propose a new classification by dividing group A into two new subgroups to better predict the phenotype. In light of this very high genotype-phenotype correlation, with their ever-increasing availability, declining cost, and turnaround time, we propose that molecular genetic studies can be more economical and practical alternative to the current initial diagnostic laboratory studies based on assays of intermediary steroid metabolites.
1. Introduction Congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency (21OHD) is autosomal recessive disorder of cortisol biosynthesis. Genetic defects in CYP21A2 cause 21OHD and represents 95–99% of all CAH cases (Arlt et al., 2010; Gidlof et al., 2013; Speiser et al., 2018). Enzymatic deficiency of steroid 21-hydroxylase is characterized by impairment of cortisol biosynthesis (with or without impairment of aldosterone biosynthesis) and excessive androgen synthesis that leads to the clinical manifestations of adrenal insufficiency and hyperandrogenism. Phenotypically, CAH can be divided into classical, severe, and non-classical (NC) forms. Classic CAH occurs in 1:13,000 to 1:15,000 live births and is represented by two phenotypes as simple virilizing (SV) and salt-wasting (SW) (Parsa and New, 2017). The SW
∗
form, the most severe form, is caused by the severe impairment of 21hydroxylase enzyme (< 2% enzymatic activity), leading to renal salt loss due to the lack of aldosterone as well as pre- and postnatal virilization, due to reactive androgen overproduction. In the SV form, patients present with variable degrees of virilization, but, they do not usually have a renal salt loss. Finally, NC form is generally asymptomatic at a neonatal period and leads to pseudo precocious puberty, hirsutism, acne, or subfertility in later childhood or adolescence (Parsa and New, 2017; Nordenstrom and Falhammar, 2018). The 21-hydroxylase gene (CYP21A2), encoding steroid 21-hydroxylase enzyme, is located HLA class III region on the chromosome 6 (6p21.3), close to the nonfunctional pseudogene (CYP21A1P). CYP21A2 and CYP21A1P genes nucleotide sequences are 98% identical in exons and approximately 96% identical in noncoding sequence
Corresponding author. E-mail address: [email protected] (B. Yüksel).
https://doi.org/10.1016/j.ejmg.2019.103782 Received 19 March 2019; Received in revised form 17 September 2019; Accepted 1 October 2019 1769-7212/ © 2019 Published by Elsevier Masson SAS.
Please cite this article as: Ihsan Turan, et al., European Journal of Medical Genetics, https://doi.org/10.1016/j.ejmg.2019.103782
European Journal of Medical Genetics xxx (xxxx) xxxx
I. Turan, et al.
Table 1 Allelic frequency of CYP21A2 gene by clinical form in 113 patients with 21OHD. Nucleotide changea
c.92c > t c.293-13C > G c.518T > A c.332_339del c.844G > T c.850A > G c.955C > T c.1069C > T c.710T > A; c.713T > A; c.719T > A Deletion/Large gen conversion Undetected Novel mutations c.20dup c.254del c.961_963del Alleles with multiple mutations c.1069C > T; c.1360C > T c.844G > T, c.1069C > T 706_713del8; c.844G > T c.955C > T; c.1069C > T c.293-13C > G; c.518T > A; c.955C > T c.293-13C > G; c.955C > T c.844G > T, c.955C > T c.518T > A; c.710T > A; c.713T > A; c.719T > A c.518T > A; c.844G > T c.1019G > A; c.1360C > T Total
Protein changea
Number of alleles (% of alleles)
p.Pro31Leu p.? p.Ile173Asn p.Gly111Valfs*21 p.Val282Leu p.Met284Val p.Gln319Ter p.Arg357Trp p.Ile237Asn; p.Val238Glu; p.Met240Lys (E6 cluster)
p.Leu8Alafs*72 p.Lys85Serfs*59 p.Glu321del p. Arg357Trp; p.Pro454Ser p.Val282Leu; p.Arg357Trp p.Gly111Valfs*21; p.Val282Leu p.Gln319Ter; p.Arg357Trp p.?; p.Ile173Asn; p.Gln319Ter p.?; p.Gln319Ter p.Val282Leu; p.Gln319Ter p.Ile173Asn; p.Ile237Asn; p.Val238Glu; p.Met240Lys (E6 cluster) p.Ile173Asn; p.Val282Leu p. Arg340His; p.Pro454Ser
Total
SW
SV
NC
2 (0.8) 87 (38.4) 18 (7.9) 10 (4.4) 7 (3.0) 1 (0.4) 26 (11.5) 21 (9.2) 4 (1.7) 18 (7.9) 4 (1.7)
– 65 (39.6) 1 (0.6) 10 (6.0) – – 22 (13.4) 20 (12.1) 4 (2.4) 18 (10.9) –
– 22 (45.8) 16 (33.3) – 1 (2.0) – 3 (6.2) 1 (2.0) – – 1 (2.0)
2 – 1 – 6 1 1 – – – 3
2 (0.8) 1 (0.4) 1 (0.4) 24 (10.6) 1 (0.4) 3 (1.3) 1 (0.4) 2 (0.8) 8 (3.5) 3 (1.3) 1 (0.4) 2 (0.8) 2 (0.8) 1 (0.4) 226
2 (1.2) – 1 (0.6) 21 (12.6) 1 (0.6) 3 (1.8) 1 (0.6) 2 (1.2) 8 (4.8) 3 (1.8) 1 (0.6) 2 (1.2) – – 166
– 1 (2.0) – 3 (6.2) – – – – – – –
– – –
2 (4.1) 1 (2.0) 48
(14.2) (7.1) (42.8) (7.1) (7.1)
(21.4)
– – – – – – – – – – 14
a
The codingDNA, and the protein sequence level are shown (RefSeq for DNA:NC_000006.12; cDNA: NM_000500.9; protein: NP_000491.4, most common mutation listed in bold.
patient's medical records by two pediatric endocrinologist (I.T, B.Y) for all cases. The SW form was characterized with salt-wasting crisis (dehydration, vomiting, hyponatremia, and hyperkalemia), ambiguous external genitalia in females, elevated levels of serum 17-hydroxyprogesterone and plasma renin activity (PRA) or plasma renin concentration at diagnosis or during a later evaluation. SV form was defined as signs and symptoms of hyperandrogenism, elevated serum 17hydroxyprogesterone, but without evidence of SW during the newborn period. NC CAH was diagnosed in both sexes diagnosed by precocious pubarche and adrenarche or hirsutism and mildly elevated levels of serum 17OHP (Speiser et al., 2018; Nordenstrom and Falhammar, 2018; White and Speiser, 2000).
(White et al., 1986). Molecular genetic analysis has shown that more than 90% of these mutations result from the transfer of deleterious micro sequences normally present in the CYP21A1P pseudogene into the functional CYP21A2 gene (Speiser et al., 2018). To date, more than 200 mutations have been described around the world. Most recently, Concolino and Costella listed 238 pathogenic variants reported up to now (Concolino and Costella, 2018). Clinical form depending on the degree of impairment of 21-hydroxylase enzyme activity caused by the mutations in the CYP21A2. Variants on CYP21A2 are classified into four groups as Null, A, B and C. Null variants displayed 0% enzyme activity in vitro assays, group A variants, the c.29313C > G mutation, conserve a minimal (< 1%) residual activity. Null and A are associated with the SW form. Finally, group B and C variants show 1–5% and 20–50% enzyme activity, respectively and are related to the SV and NC form of the disease, respectively (Concolino and Costella, 2018; Speiser et al., 1992; Wedell et al., 1994; Loke, 2008). In this study, we present the result of mutation analyses of CYP21A2 in a large cohort of 113 patients with CAH due to 21OHD and their 165 parents. Additionally, the genotype-phenotype correlation was analyzed systematically to determine predictive characteristics of variant groups of CYP21A2.
2.2. Molecular analysis of CYP21A2 DNA samples were obtained with the isolation from 200 μl of blood sample from each individual, by using QIAamp DNA Blood Mini Kit (Qiagen Inc.). Primer designs were carried out for the coding regions of CYP21A2. In designing primers, the 8-bp deletion region in exon 1 was selected to discriminate between CYP21A2 and its pseudogene, so that sequencing results would not contain data from the allele which contains 8-bp deletion and impossible to be a functional allele. The gene was designed to amplify in two amplicons; one expands from the beginning of the gene to exon 3 and the other one from exon 3 to the end of the gene, so that one of the primers of each amplicon resides on the 8-bp deletion region. Two PCRs were carried out on isolated DNA samples, by using the designed primers and the reactions were checked by using 2% agarose gel electrophoresis. PCRs belong to each individual were mixed to obtain PCR pools, which have both of the amplicons of each individual in one tube. While mixing, the amplification efficiency and length of the amplicons were taken into consideration; the volume for each PCR is directly proportional to the length of the amplicon and inversely
2. Material and methods 2.1. Patients We investigated a total of 113 patients with 21OHD from 99 unrelated families. These patients were treated in the Cukurova University, Faculty of Medicine, Division of Pediatric Endocrinology during the period of 2007 through 2017. The Ethics Committee of the Cukurova University Faculty of Medicine approved this study. Informed consents were obtain from all participants or their legal guardians. Clinical forms of CAH due to 21OHD were determined based on clinical and hormonal criteria and a retrospective review of each 2
European Journal of Medical Genetics xxx (xxxx) xxxx
I. Turan, et al.
Table 2 Genotype phenotype correlation. %a
Genotype Allele 1 Group 0 deletion or large gene conversion deletion or large gene conversion deletion or large gene conversion deletion or large gene conversion p.Gly111Valfs*21 p.Gly111Valfs*21 p.Gly111Valfs*21 p.Gly111Valfs*21 p.Gly111Valfs*21; p.Val282Leu p.Gln319Ter p.Gln319Ter p.Gln319Ter p.Gln319Ter p.Gln319Ter; c.293-13C > G; p.Ile173Asn p. Arg357Trp p. Arg357Trp; p.Val282Leu E6 cluster E6 cluster/p.Ile173Asn p. Leu8Alafs*72 Group A Group A1 c.293-13C > Ga c.293-13C > G c.293-13C > G c.293-13C > G Group A2 c.293-13C > G Group B p.Ile173Asn p.Ile173Asn p.Ile173Asn p.Ile173Asn p.Ile173Asn p.Ile173Asn; p.Val282Leu Group C p.Val282Leu p.Val282Leu p.Val282Leu p.Pro31Leu Group D c.293-13C > G p.Val282Leu p.Ile173Asn Group E c.293-13C > G p.Gln319Ter p.Val282Leu
Allele 2
Number of allele
P.P
Total
SW
SV
NC
p.Val282Leu c.293-13C > G p.Ile173Asn p.Pro31Leu
39.8 5.3 0.8 1.7 0.8 2.6 0.8 0.8 0.8 0.8 5.3 0.8 0.8 0.8 3.5 7.9 0.8 1.7 0.8 0.8 39.8 7.9 1.7 0.8 4.4 0.8 31.8 31.8 10.6 3.5 0.8 2.6 0.8 1.7 0.8 3.5 0.8 0.8 0.8 0.8
44 6 1 2 1 3 1 1 1 1 6 1 1 1 4 9 1 2 1 1 45 9 2 1 5 1 36 36 13 4 1 3 1 3 1 4 – – – 1
44 6 1 2 1 3 1 1 1 1 6 1 1 1 4 9 1 2 1 1 36 9 2 1 5 1 27 27 1 – – – – 1 – – – – – –
– – – – – – – – – – – – – – – – – – – – 9 – – – – – 9 9 12 4 1 3 1 2 1 – – – – –
– – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – – 4 1 1 1 1
p.Glu321del p.Met284Val p.R340H; p.454S
0.8 0.8 0.8
1 1 1
1 – –
– – 1
– 1 –
undetected undetected undetected
0.8 0.8 0.8
1 1 2
– – –
1 – –
– 1d 2e
deletion/large gene conversion p.Gln319Terc p.Gln319Ter; c.293-13C > G p. Arg357Trp; p.Pro454Ser p.Gly111Valfs*21 p.Gln319Ter p.Gln319Ter; c.293-13C > G p.Gln319Ter; p. Arg357Trp p. Arg357Trp; p.Val282Leu p.Gln319Ter p. Arg357Trp p.Gln319Ter; p. Arg357Trp p.Gln319Ter; p.Val282Leu p.Gln319Ter; c.293-13C > G; p.Ile173Asn p. Arg357Trp p. Arg357Trp; p.Val282Leu E6 cluster E6 cluster/p.Ile173Asn p. Leu8Alafs*72
deletion or large gene conversion p.Gly111Valfs*21 p.Gln319Ter p. Arg357Trpc c.293-13C > G p.Ile173Asn p. Arg357Trp p.Gln319Ter p.Lys85Serfs*59 c.293-13C > G p.Ile173Asn; p.Val282Leu
b
100%
80% 100%
75% 92.3%
100%
NA
NA
The codingDNA, and the protein sequence level are shown (RefSeq for DNA:NC_000006.12; cDNA: NM_000500.9; protein: NP_000491.4). NA: not applicable. a % frequency of genotype. b Positive predictivity for expected phenotype. c One allele is de novo mutation. d Patient carrying duplication of CYP21A2. e One of two patient carrying duplication of CYP21A2.
MLPA (Multiplex ligation-dependent probe amplification) was used for copy number and gene conversion, analyses (MRC-Holland b.v., SALSA MLPA P050 CAH probemix). Data from both sequencing and MLPA were used collaboratively to do the interpretation. In any case of contradiction between sequencing and MLPA, sequencing data was prioritized for sequencing variations and MLPA data was prioritized for copy number and gene conversions.
proportional to the efficiency of the reaction, which were estimated with the help of gel electrophoresis. The PCR pools for each individual were purified using NucleoFast® 96 PCR kit (MACHEREY-NAGEL GmbH). The purified pools were quantified using ND1000 (Thermo Fisher Scientific Inc.) micro volume spectrophotometer and standardized to 0,2 ng/ul, which was needed for sample preparation step. The samples were got ready for next-generation sequencing by using NexteraXT sample preparation kit (Illumina Inc.). Next-generation sequencing of the samples were carried out by using Miseq system (Illumina Inc.). The data were analyzed on IGV 2.3 software (Broad Institute). NC_000006.12 (NM_00500.9) was used for reference sequence of CYP21A2.
2.3. Analyses of novel variants We searched for all identified variants in the Human Gene Mutation Database (Stenson et al., 2003), LOVD (Fokkema et al., 2011), and the 3
European Journal of Medical Genetics xxx (xxxx) xxxx
I. Turan, et al.
Table 3 Genotype phenotype correlation as combination with metadata from other published studies. A1
A2
GG
PP
SW
SV
NC
Turkey*
SW
SV
NC
Netherlands (Stikkelbroeck et al., 2003)
SW
SV
NC
SW
SV
NC
Middle Europe (Dolzan et al., 2005)
Germany (Krone et al., 2000)
SW
SV
NC
USA (Finkielstain et al., 2011)
SW
SV
NC
Brazil (de Carvalho et al., 2016)
SW
SV
NC
To
Croatia (Dumic et al., 2017)
Total
SW
SV
NC
%P.
0
0
0
SW
44
–
–
28
1
–
80
1
1
32
–
–
35
5
–
55
4
–
28
–
**
269
257 11
1
95.5
A A A
A/0 0 A
A A1 A2
SW SW SW
35 9 27
9 – 9
– –
21 14 7
1 1 –
– – –
108 44 64
6 1 5
1 1 –
43 25 18
5 3 2
– – –
60 52 8
8 7 1
– – –
66 43 23
36 11 25
– – –
29 13 16
5 3 2
** ** **
434 227 207
363 70 200 26 163 44
1 1 –
83.6 88.1 78.7
B
B/A/0
B
SV
1
12
–
2
10
4
24
65
5
13
37
–
2
23
3
–
53
1
4
12
**
259
46
13
77.2
200
-A1 and A2: mutation group of allele 1 and 2, respectively. GG: group of genotype. PP: predictive phonotype. SW: salt wasting. SV: simple virilization. NC: nonclassic. To: total. %P.: positive predictive value. *this study. **patients with non-classical CAH were not included in Croatia's study. A/0: A or 0. B/A/0: B or A or 0. -Mutation group 0: gene deletions, large gene conversion, E6 cluster, p. Arg357Trp, p.Gln319Ter, p.L308Ffs*6, p.Gly111Valfs*21, novel frameshift mutation and multiple mutations alleles containing any of these mutations. Mutation group A: c.293-13C > G and multiple mutations alleles containing c.293-13C > G but not mutations of group 0. Mutation group B: p.Ile173Asn and promoter conversion + Pro31Leu. Mutation group C: p.Pro31Leu, p.Val282Leu, p.Pro454Ser. -Genotypes that do not fit this classification in the articles were excluded from the estimation.
2.5. Statistical analysis
recent review article by Concolino and Costella (2018). Date of check was march 13th 2019. Novel variants were classified based on the 2015 American College of Medical Genetics and Genomics and Association for Molecular Pathology guidelines (ACMG-AMP) (Richards et al., 2015).
The Pearson Chi-Square tests were used to compare the groups and the odds ratio (OR) with 95% confidence interval (CI) were calculated. The positive predictive values were determined as the percentage of patients with the expected phenotype for each group. The Statistical Package for Social Sciences 20 (SPSS20) was used to carry out the statistical analyses.
2.4. Classification in mutation groups and genotype-phenotype relation To help predict the severity of the disease-causing mutations, genotypes of patients were divided into five groups according to their residual 21-hydroxylase enzyme activity, based on prior in vitro studies (Speiser et al., 1992; Wedell et al., 1994). The group 0 included patients who carry homozygous null mutations that result in completely inactive enzymes (gene deletions, large gene conversion, E6 cluster, p. Arg357Trp, p.Gln319Ter, p.L308Ffs*6, p.Gly111Valfs*21, novel frameshift mutation and multiple mutations alleles containing any of these mutations) (Speiser et al., 1992). The Group A included patients homozygous for c.293-13C > G (which has minimal residual enzymatic activity) (Higashi et al., 1988) or compound heterozygous for c.293-13C > G and a null mutation. Additionally, in this study, the group A was divided into two subgroups as A1 for compound heterozygous mutations and A2 for homozygous mutations. The Group B included the mutation I172N (~2% residual enzymatic activity) (Krone et al., 2005) and the promoter conversion + Pro31Leu, homozygous or compound heterozygous with a null or group A mutation. The group C contained mutations P31L, V282L, and P454S (~20–60% residual enzyme activity) (Speiser and New, 1987), homozygous or compound heterozygous with a null, A, or B mutation. The Group D included patients with novel mutations or variants whose enzymatic activity impairment had not been assessed. Finally, the group E included patients with at least one allele without any identified mutations. In this classification, the mutation with the mildest effect on enzymatic activity (in the compound heterozygous composition) determines the predicted phenotype. In alleles containing multiple mutations, the most deleterious mutation determines the genotype groups (Dolzan et al., 2005). Groups 0 and A were predicted to result in the SW form. The expected phenotype was SV for Group B, and NC for Group C mutations. Positive predictive value was calculated for groups A, B, and C. We concomitantly analyzed six previously published studies similar to our study while examining the genotype-phenotype relations (Dolzan et al., 2005; Stikkelbroeck et al., 2003; de Carvalho et al., 2016; Krone et al., 2000; Finkielstain et al., 2011; Dumic et al., 2017). Mutations that do not fit this classification in the articles were excluded from the estimations.
3. Results 3.1. Phenotypes and mutational analyses of patients In total, from 113 patients with 21OHD (82 SW, 24 SV, and 7 NC), 226 alleles were analyzed for the CYP21A2 gene variants. Seventy-five patients (66%) were homozygous and 34 patients (30%) were compound heterozygous. The parents of all patients carrying compound heterozygous variants (except two) were analyzed. Overall, 160 parents were genotyped and only (1.8%) of them were without variants of their offspring (i.e de novo mutations c.293-13C > G, p. Arg357Trp, p.Gln319Ter). Parents of six of 75 patients (8%) with homozygous mutations denied consanguinity. Four patients whose parents were consanguineous were carried compound heterozygous mutations. The CYP21A2 duplication was detected in 11 patients (9.7%). The homozygous c.293-13C > G was the most common genotype in both classical forms, SW and SV. The most frequent mutation was c.293-13C > G (37.1%) followed by p.Gln319Ter (11.0%), large conversation/deletion (9.2%), p. Arg357Trp (9.2%), and p.Ile173Asn (7.9%). In this study. ten different combinations of point mutations were detected; the allele frequency of this complexity was found in 10.6%. In four patients (one SV, three NC), mutations were detected only in one allele and two of them were CYP21A2 duplication. Our cohort contained the full spectrum of mutations shown in Table 1. 3.2. Novel variants Three variants that have not been described previously were identified. One patient was homozygous for the novel frameshift mutation p. Leu8Alafs*72 (c.20dup) that led to the SW form. A deletion of the glutamic acid at codon 320 or 321 (c.961_963del) was detected in trans with c.293-13C > G in one patient with the SW form. One patient with the SV form was compound heterozygous for p.Lys85Serfs*59 (c.254del) and p.Ile173Asn. All these novel variants were considered ‘likely pathogenic’ based on the 2015 ACMG/AMP guideline. New variants were submitted to ClinVar. Accession numbers is 4
European Journal of Medical Genetics xxx (xxxx) xxxx
I. Turan, et al.
variants: two frameshift variants and one in-frame deletion. Functional analyses of these variants were not performed. The p. Leu8Alafs*72 and the p.Lys85Serfs*59 frameshift variants are likely to cause nonsense mediated decay resulting in no residual enzymatic activity. The p.Glu321del was detected in trans with c.293-13C > G in one patient with the SW form. The glutamic acid at 320 or 321 is highly conserved among different P450 cytochromes in 30 different species; located in a negatively charged sequence, it probably interacts with part of positively charged residues (Tardy et al., 2010; Kent et al., 2002). These may explain why this inframe deletion can significantly affect the enzyme activity. An accurate and reliable genotype-phenotype association in 21OHD will help guide the treatment management such as glucocorticoid and mineralocorticoid requirement. This study largely confirmed that there is a good positive genotype-phenotype correlation in 21OHD. But we note that the prediction of severe genotypes is more reliable than milder genotypes. This classification is useful and acceptable because it provides predictions to 96.4% of the cases (Dolzan et al., 2005; Stikkelbroeck et al., 2003; Krone et al., 2000; Finkielstain et al., 2011; Dumic et al., 2017). The genotype-phenotype correlation was the best for group 0, when we refined it by including other studies (Dolzan et al., 2005; Stikkelbroeck et al., 2003; de Carvalho et al., 2016; Krone et al., 2000; Finkielstain et al., 2011; Dumic et al., 2017). The c.293-13C > G mutation, known to be associated with minimal residual enzymatic activity may cause discordance of phenotype. The prediction was perfect when this splice site mutation was detected in trans with null mutation. In this study, the group A was divided into two subgroups: A1 for compound heterozygous and A2 for homozygous mutations. When the data of a total of 434 patients from seven studies were collectively analyzed, there was a statistically significant difference in the two subgroups for positive predictivity of a distinct and segregated phenotype. These findings led us c.293-13C > G is more deleterious than homozygous genotype when trans with null mutations. Additionally, predictivity of new subgroup A1 is more reliable. This new classification improves genotype-phenotype relationships, therefore enhances the clinical utility of CYP21A2 mutational analyses. This finding may be particularly valuable for ethnic groups where the compound heterozygote genotypes are more common. The developing clarity of genotype-phenotype relationship and increasing availability of molecular genetic studies will decrease unnecessary repeated hormonal tests and unneeded medical treatments. This will have both medical and economic benefits. In conclusion, our study reports genotype–phenotype correlations in the largest 21OHD cohort in Turkey. This study confirms that CYP21A2 genotyping with next-generation sequencing and MLPA can accurately and reliably confirm the diagnosis of 21OHD. The spectrum of mutations in the CYP21A2 gene was investigated in Southern Turkey. We confirmed a good genotype-phenotype correlation in another ethnic group. We propose a new classification by dividing group A into two new genotype subgroups to better predict the phenotype.
VCV000623477, VCV000623473, and VCV000623476 for c.254del, c.20dup, and c.961_963del respectively. 3.3. Genotype-phenotype relationship In this study, the genotype–phenotype correlation of mutation classification was 91.5%. A complete genotype-phenotype concordance was observed in Group 0. Genotypes correctly predicted phenotypes in 79.5% of the patients in group A; 100% in subgroup A1, and 75% in subgroup A2. Genotype-phenotype discordance occurred in 9 patients with group A and all of them included in subgroup A2. The genotypephenotype relation of our cohort was detailed in Table 2. The overall, in seven studies (including our study), the highest positive predictive value among all groups is in the group 0 with 95%. Genotype correctly predicted phenotype in 83.6% of the patients in group A, 88.1% in subgroup A1, and 78.7% in subgroup A2. Positive predictivity of subgroups A1 was higher than groups A and subgroups A2. This difference is statistically significant with p value < 0.05. Between subgroup A1 and A2, OR was 2 (95% CI: 1.19–3.37). The genotype-phenotype relation of seven studies was detailed in Table 3. 4. Discussion How important is the molecular genetic analysis in the diagnosis and management of 21OHD? In clinical practice, starting steroid treatment without a confirmed diagnosis is often inevitable. In some parts of the world baseline steroid levels may not be readily available due to their high costs among other factors. With the ever-increasing availability, reducing cost, and shortening turnaround time, genetic mutation analyses of CYP21A2 can be more practical and economical initial diagnostic modality, provided there is an established and reliable genotype-phenotype correlation. This study was the first large cohort study using next-generation sequencing as the surrogate diagnostic method, which detected 222 of 226 alleles of the affected patients. The diagnostic sensitivity was 98.2% (95.8% in SV and 100% in the SW form). This study confirms that CYP21A2 genotyping with next-generation sequencing and MLPA can accurately and reliably confirm the diagnosis of 21OHD. Homozygous genotype frequency (70%) was markedly higher in our study than other analyzed studies (Stikkelbroeck et al., 2003; de Carvalho et al., 2016; Krone et al., 2000; Finkielstain et al., 2011; Dumic et al., 2017) (21–35%) except for a Middle European (69%) report (Dolzan et al., 2005). This is an expected result for populations with high consanguinity rates. Because both parents are carriers of the same deleterious variant in the CYP21A2 gene. Interestingly, the four patients whose parents were related carried compound heterozygous mutations while the parents of six patients with homozygous mutations denied consanguinity. Although, the history of consanguinity is a valuable clue in genetically inherited diseases like 21OHD, these inconsistencies should be kept in mind. The finding that the c.293-13C > G was the most prevalent mutation in Turkey according to Bas et al. (Bas et al., 2009) was confirmed by our larger cohort study. The most common mutation in other large cohorts was also c.293-13C > G. The only exception was the p.Val282Leu in de Carvalho et al. (de Carvalho et al., 2016) study, which included a very high percentage of 206 (42%) NC patients. Mutation frequencies of affected alleles in this study were closely similar to those published large cohorts study in Turkey (Bas et al., 2009). As differently, a high frequency (10.6%) of alleles with multiple mutations was found. This may be due to superior technique of the genetic analyses used in this study. Compared with other ethnic groups, the frequencies of the p. Arg357Trp and the p.Gln319Ter were higher, therefore, the SW/SV form ratio was higher in our study (Dolzan et al., 2005; Krone et al., 2000; Finkielstain et al., 2011; Dumic et al., 2017; Bas et al., 2009). Genetic analyses of CYP21A2 in this study identified three novel
Acknowledgments This work was supported by the Scientific and Technological Research Council of Turkey (TÜBİTAK, Project No.108S236). We would like to thank Dr. A. Kemal Topaloglu for his valuable discussions on the manuscript. We are grateful to Fatma Dereli (Intergene Genetic Center, Ankara) and Haldun Dogan (Intergene Genetic Center, Ankara). References Arlt, W., Willis, D.S., Wild, S.H., Krone, N., Doherty, E.J., Hahner, S., Han, T.S., Carroll, P.V., Conway, G.S., Rees, D.A., Stimson, R.H., Walker, B.R., Connell, J.M., Ross, R.J., 2010. United Kingdom Congenital Adrenal Hyperplasia Adult Study E: health status of adults with congenital adrenal hyperplasia: a cohort study of 203 patients. J. Clin. Endocrinol. Metab. 95, 5110–5121. Bas, F., Kayserili, H., Darendeliler, F., Uyguner, O., Gunoz, H., Yuksel Apak, M., Atalar, F.,
5
European Journal of Medical Genetics xxx (xxxx) xxxx
I. Turan, et al.
virilizing forms of congenital adrenal hyperplasia due to 21-hydroxylase deficiency. J. Clin. Endocrinol. Metab. 90, 445–454. Loke, K.Y., 2008. Clinical applications of molecular genetics: the model of congenital adrenal hyperplasia. Ann. Acad. Med. Singapore 37 18-14. Nordenstrom, A., Falhammar, H., 2018. Management OF endocrine disease: diagnosis and management of the patient with non-classic CAH due to 21-hydroxylase deficiency. Eur. J. Endocrinol. 180 (3), 127–145. https://doi.org/10.1530/EJE-18-0712. Parsa, A.A., New, M.I., 2017. Steroid 21-hydroxylase deficiency in congenital adrenal hyperplasia. J. Steroid Biochem. Mol. Biol. 165, 2–11. Richards, S., Aziz, N., Bale, S., Bick, D., Das, S., Gastier-Foster, J., Grody, W.W., Hegde, M., Lyon, E., Spector, E., Voelkerding, K., Rehm, H.L., Committee, A.L.Q.A., 2015. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of medical genetics and genomics and the association for molecular Pathology. Genet. Med. 17, 405–424. Speiser, P.W., New, M.I., 1987. Genotype and hormonal phenotype in nonclassical 21hydroxylase deficiency. J. Clin. Endocrinol. Metab. 64, 86–91. Speiser, P.W., Dupont, J., Zhu, D., Serrat, J., Buegeleisen, M., Tusie-Luna, M.T., Lesser, M., New, M.I., White, P.C., 1992. Disease expression and molecular genotype in congenital adrenal hyperplasia due to 21-hydroxylase deficiency. J. Clin. Investig. 90, 584–595. Speiser, P.W., Arlt, W., Auchus, R.J., Baskin, L.S., Conway, G.S., Merke, D.P., MeyerBahlburg, H.F.L., Miller, W.L., Murad, M.H., Oberfield, S.E., White, P.C., 2018. Congenital adrenal hyperplasia due to steroid 21-hydroxylase deficiency: an endocrine society clinical practice guideline. J. Clin. Endocrinol. Metab. 103, 4043–4088. Stenson, P.D., Ball, E.V., Mort, M., Phillips, A.D., Shiel, J.A., Thomas, N.S., Abeysinghe, S., Krawczak, M., Cooper, D.N., 2003. Human Gene Mutation Database (HGMD): 2003 update. Hum. Mutat. 21 (6), 577–581. https://doi.org/10.1002/humu.10212. Stikkelbroeck, N.M., Hoefsloot, L.H., de Wijs, I.J., Otten, B.J., Hermus, A.R., Sistermans, E.A., 2003. CYP21 gene mutation analysis in 198 patients with 21-hydroxylase deficiency in The Netherlands: six novel mutations and a specific cluster of four mutations. J. Clin. Endocrinol. Metab. 88, 3852–3859. Tardy, V., Menassa, R., Sulmont, V., Lienhardt-Roussie, A., Lecointre, C., Brauner, R., David, M., Morel, Y., 2010. Phenotype-genotype correlations of 13 rare CYP21A2 mutations detected in 46 patients affected with 21-hydroxylase deficiency and in one carrier. J. Clin. Endocrinol. Metab. 95, 1288–1300. Wedell, A., Thilen, A., Ritzen, E.M., Stengler, B., Luthman, H., 1994. Mutational spectrum of the steroid 21-hydroxylase gene in Sweden: implications for genetic diagnosis and association with disease manifestation. J. Clin. Endocrinol. Metab. 78, 1145–1152. White, P.C., Speiser, P.W., 2000. Congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Endocr. Rev. 21, 245–291. White, P.C., New, M.I., Dupont, B., 1986. Structure of human steroid 21-hydroxylase genes. Proc. Natl. Acad. Sci. U. S. A. 83, 5111–5115.
Bundak, R., Wilson, R.C., New, M.I., Wollnik, B., Saka, N., 2009. CYP21A2 gene mutations in congenital adrenal hyperplasia: genotype-phenotype correlation in Turkish children. Journal of clinical research in pediatric endocrinology 1, 116–128. Concolino, P., Costella, A., 2018. Congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency: a comprehensive focus on 233 pathogenic variants of CYP21A2 gene. Mol. Diagn. Ther. 22, 261–280. de Carvalho, D.F., Miranda, M.C., Gomes, L.G., Madureira, G., Marcondes, J.A., Billerbeck, A.E., Rodrigues, A.S., Presti, P.F., Kuperman, H., Damiani, D., Mendonca, B.B., Bachega, T.A., 2016. Molecular CYP21A2 diagnosis in 480 Brazilian patients with congenital adrenal hyperplasia before newborn screening introduction. Eur. J. Endocrinol. 175, 107–116. Dolzan, V., Solyom, J., Fekete, G., Kovacs, J., Rakosnikova, V., Votava, F., Lebl, J., Pribilincova, Z., Baumgartner-Parzer, S.M., Riedl, S., Waldhauser, F., Frisch, H., Stopar-Obreza, M., Krzisnik, C., Battelino, T., 2005. Mutational spectrum of steroid 21-hydroxylase and the genotype-phenotype association in Middle European patients with congenital adrenal hyperplasia. Eur. J. Endocrinol. 153, 99–106. Dumic, K.K., Grubic, Z., Yuen, T., Wilson, R.C., Kusec, V., Barisic, I., Stingl, K., Sansovic, I., Skrabic, V., Dumic, M., New, M.I., 2017. Molecular genetic analysis in 93 patients and 193 family members with classical congenital adrenal hyperplasia due to 21hydroxylase deficiency in Croatia. J. Steroid Biochem. Mol. Biol. 165, 51–56. Finkielstain, G.P., Chen, W., Mehta, S.P., Fujimura, F.K., Hanna, R.M., Van Ryzin, C., McDonnell, N.B., Merke, D.P., 2011. Comprehensive genetic analysis of 182 unrelated families with congenital adrenal hyperplasia due to 21-hydroxylase deficiency. J. Clin. Endocrinol. Metab. 96, E161–E172. Fokkema, I.F., Taschner, P.E., Schaafsma, G.C., Celli, J., Laros, J.F., den Dunnen, J.T., LOVD, v, 2011. 2.0: the next generation in gene variant databases. Hum. Mutat. 32, 557–563. Gidlof, S., Falhammar, H., Thilen, A., von Dobeln, U., Ritzen, M., Wedell, A., Nordenstrom, A., 2013. One hundred years of congenital adrenal hyperplasia in Sweden: a retrospective, population-based cohort study. Lancet Diabetes Endocrinol 1, 35–42. Higashi, Y., Tanae, A., Inoue, H., Hiromasa, T., Fujii-Kuriyama, Y., 1988. Aberrant splicing and missense mutations cause steroid 21-hydroxylase [P-450(C21)] deficiency in humans: possible gene conversion products. Proc. Natl. Acad. Sci. U. S. A. 85, 7486–7490. Kent, W.J., Sugnet, C.W., Furey, T.S., Roskin, K.M., Pringle, T.H., Zahler, A.M., Haussler, D., 2002. The human genome browser at UCSC. Genome Res. 12, 996–1006. Krone, N., Braun, A., Roscher, A.A., Knorr, D., Schwarz, H.P., 2000. Predicting phenotype in steroid 21-hydroxylase deficiency? Comprehensive genotyping in 155 unrelated, well defined patients from southern Germany. J. Clin. Endocrinol. Metab. 85, 1059–1065. Krone, N., Riepe, F.G., Grotzinger, J., Partsch, C.J., Sippell, W.G., 2005. Functional characterization of two novel point mutations in the CYP21 gene causing simple
6