RUNX2 analysis of Danish cleidocranial dysplasia families

© 2010 John Wiley & Sons A/S

Clin Genet 2011: 79: 254–263 Printed in Singapore. All rights reserved

CLINICAL GENETICS doi: 10.1111/j.1399-0004.2010.01458.x

Short Report

RUNX2 analysis of Danish cleidocranial dysplasia families Hansen L, Riis AK, Silahtaroglu A, Hove H, Lauridsen E, Eiberg H, Kreiborg S. RUNX2 analysis of Danish cleidocranial dysplasia families. Clin Genet 2011: 79: 254–263. © John Wiley & Sons A/S, 2010 Cleidocranial dysplasia (CCD) is an autosomal dominant inherited disease caused by mutations in the Runt gene RUNX2. Screening of 19 Danish CCD families revealed 16 pathogenic mutations (84%) representing 8 missense mutations, 2 nonsense mutations, 4 frame-shift mutations and 2 large deletions in the RUNX2 locus. Eight mutations were novel, two were found twice, and polymorphisms were found in the promoter region and in the conserved polyglutamine/polyalanine repeat. A large duplication downstream of RUNX2 found in one patient suggests a possible regulatory RUNX2 element. The CCD phenotypes and genotypes adhere to the large phenotypic variability reported in previous CCD studies. Identification of large chromosome aberrations in or near the RUNX2 locus in 3 of the 19 cases suggests copy number analyses to be included in future RUNX2 mutation analyses.

L Hansena,b , AK Riisc , A Silahtaroglua,b , H Hoved , E Lauridsena,b , H Eiberga and S Kreiborgc,d a Department of Cellular and Molecular Medicine, Panum Institute, University of Copenhagen, DK-2200 Copenhagen N, Denmark, b Wilhelm Johannsen Centre for Functional Genome Research, ICMM, Panum Institute, University of Copenhagen, DK-2200 Copenhagen N, Denmark, c Department of Pediatric Dentistry and Clinical Genetics, School of Dentistry, Faculty of Health Science, University of Copenhagen, DK-2200 Copenhagen, Denmark, and d Department of Clinical Genetics, Unit for Rare Diseases, Juliane Marie Centre, Copenhagen University Hospital at Rigshospitalet, DK-2100 Copenhagen, Denmark

Key words: cleidocranial dysplasia – copy number variation – large deletion – mutation – polyGln/polyAla repeat – RUNX2 Corresponding author: Lars Hansen, Department of Cellular and Molecular Medicine, Panum Institute, University of Copenhagen, Blegdamsvej 3, DK-2200 Copenhagen N, Denmark. Tel.: +45 3532 7825; fax: +45 3532 7845; e-mail: [email protected] Received 8 February 2010, revised and accepted for publication 26 April 2010

Cleidocranial dysplasia (CCD, OMIM #119600) is caused by mutations in the transcription factor RUNX2 (alias CBFA1, chromosome 6p12.3, OMIM #600211). CCD is inherited in an autosomal dominant fashion, and more than 100 mutations have been reported (1–4). CCD, the only disorder associated with RUNX2 mutations, is characterized by persistently open skull sutures with bulging calvaria, hypoplasia or aplasia of 254

the clavicles permitting abnormal opposition of the shoulders, vertebral malformations, wide pubic symphysis, short middle phalanx of the fifth fingers, and dental anomalies, including supernumerary teeth (5, 6). Three RUNX2 transcripts are known, and the mutations represent missense, nonsense, frameshift, and splice mutations as well as large chromosomal deletions and translocations (7). Missense

RUNX2 mutations in a Danish CCD cohort

mutations are solely found in the Runt domain, whereas nonsense mutations and frame-shift mutations are found throughout the gene. CCD is generally accepted to be caused by haploinsufficiency, but dominant negative effects cannot be ruled out. CCD displays clinical heterogeneity as well as variable expression, and stringent genotype–phenotype correlations have not been found despite the large number of mutations. Recessive forms have not been reported. This study reports mutational analyses of a cohort of Danish CCD families and expands the mutation spectra by copy number analysis. Materials and methods

out, and data from several of the CCD patients analyzed have been described previously (8). Control data for adult individuals came from previously published data of two groups of dental students, a male group and a female group (9–11). In addition, the skull films of the patients were assessed for the presence of Wormian bones. The panoramic X-rays were used to assess the dental phenotype, including the presence of supernumerary teeth and timing of tooth eruption. Finally, somatic development was assessed as body height and weight and compared to existing normative data (12). The investigations adhered to the tenets of the Declaration of Helsinki, and all persons gave informed consent for participation in the mutation analyses.

Patients material

Families with CCD were referred to the Department of Pediatric Dentistry and Clinical Genetics, School of Dentistry, University of Copenhagen. Pedigree constructions were made after interviews of the families, and blood samples were collected from one or several individuals of each family. All patients and controls were examined by radiographic techniques, including roentgencephalometry and panoramic X-rays. Cephalometric analyses of craniofacial morphology were carried

DNA extraction, sequence and mutation analyses

DNA was extracted from whole blood using the standard methods, and exons and exon–intron boundaries of RUNX2 (transcript NM_001024630) were bidirectionally sequenced [for polymerase chain reaction (PCR) primers see Table 1]. Briefly, exons 1, 2, 4, 5, 6, 7 and 9 were amplified using HotStartTaq (Qiagen, Bothell, WA) according to the manufacturer’s protocol (15 min, 95◦ C; 40 cycles of 0.5 min, 96◦ C; 0.5 min, 53.5◦ C;

Table 1. List of RUNX2 polymerase chain reaction (PCR) primers Exons 1 and 2 Internal sequence primers for exons 1 and 2 Exon 3 Exon 4 Exon 5 Exon 6 Exon 7 Exon 8 Exon 9 Internal sequence primers for exon 9 Fluorescence three-primer PCR for polyQ/A repeat Fluorescence primer RUNX2_polyQ_F and RUNX2_polyQ PCR extreme long repeat expansions RUNX2-Qrep-F RUNX2_polyQ_repeat_r

GCAAAAAGGCAGAGGTTGAG TGCTATTTGGAAAAGCTAGCAG CTTCATTCGCCTCACAAACA TCACTGACTCTGTTGGTCTCG CGGCCACTTCGCTAACTTGT GCCAAGGCAGGAGGTCTT CATTCCTGTCGGCCATTACT GAAAAACACTCAACTTCATCTGGA GGAGTCCTGCCTCTTGTCTTT ATGCAGATAGCAAAGTCCACAA TCCTTGGCTTAAACTCCCAGT CCCAAAATGTTATCCCCAAGT TTGGGTTGCATGTTTCTAAGG CTTGCCCACATGCCTCTAAT AATGAGGGATGGGAACCTCT GTGCCATGATGTGCATTTGT GGCTTGCTGTTCCTTTATGG GGCTGCAAGATCATGACTGA AAAGCCAGAGTGGACCCTTC ACAGTTGGGGAACTGCTGTG FAM-TGACCGGCAGCAAAATT TGACCGGCAGCAAAATTGGCAAAATGAGCGACGTGAG GGCGATGATCTCCACCAT TGACCGGCAGCAAAATTGCAGCAGCAGCAGCAGCA GGCGATGATCTCCACCAT

255

Hansen et al.

1 min, 72◦ C; post extension 6 min at 72◦ C). Exons 3 and 7 were amplified using Titanium™ Taq (Clonetech, Palo Alto, CA) according to the manufacturer’s protocol, which is supplied with 1.5 M betain and 5 mM MgCl2 (5 min, 95◦ C; 40 cycles of 0.5 min, 96◦ C; 0.5 min, 57◦ C; 1 min, 68◦ C; post extension 6 min at 72◦ C). PCR products were analyzed by 2% agarose gel electrophoresis, 100 mM Tris, 90 mM borate and 2 mM ethylenediaminetetraacetic acid (EDTA) before sequencing. DNA sequencing was carried out using bigdye version 1.1 and analyzed on an ABI3130XL (Applied Biosystems, Foster City, CA). Sequence data were analyzed using standard software and aligned to the RUNX2 transcription variants 1, 2 and 3 (accession nos. NM_001024630, NM_001015051, and NM_004348). Analyses of DNA variations were carried out for a minimum of 60 unrelated individuals of matching ethnic background. Diagnostic restriction enzyme digests of novel mutations were carried out after PCR in a 20-μl volume using 2–5 μl of PCR product according to the manufacturer’s protocols (New England Biolabs, Ipswich, MA), and the digests were analyzed by 2% agarose or 20% acryl amide, 100 mM Tris, 90 mM borate and 2 mM EDTA, gel electrophoresis. Polymorphism analyses of the polyGln/Ala repeat in exon 3

DNA sequence variations were compared with the NCBI SNP database (dbSNP, build 130), and variations in the polyQ/A repeat were studied by fragment analysis using an ABI3130XL, and a fluorescence-labeled three-primer system was developed for detection of the repeat length variations (for primers see Table 1). Briefly, PCR using one carboxyfluorescein fluorescence-labeled primer and two RUNX2 -specific PCR primers was carried out according to the standard protocol using touchdown conditions (Ampliqon III Taq polymerase, Ampliqon, Copenhagen, Denmark). The repeat variations were analyzed in all CCD patients and in 94 normal persons of corresponding ethnic background. Fragment length variations were subcloned (TA Cloning Kit, Invitrogen, Life Technologies Corporation, Carlsbad, CA) and screened by colony PCR, and sequenced and aligned to the reference allele NM_001024630. Analyses of extreme long repeat expansions not detectable under standard PCR conditions were carried out using a redundant PCR primer (Table 1) matching the polyalanine repeat. 256

Copy number variation analyses, fluorescence in situ hybridization and quantitative PCR

DNA (500 ng) from one affected individual each from CCD1, CCD4, CCD8, CCD9 and CCD19 families was analyzed for SNP variation using the Affymetrix Genome-Wide Human SNP Array 6.0 microchip technology (Affymetrix, Santa Clara, ˚ CA, and AROS Applied Biotechnology, Arhus, Denmark) and analyzed using the Genotyping Console (Affymetrix, Santa Clara, CA). Fluorescence in situ hybridization (FISH) was carried out using the bacterial artificial chromosome (BAC) clone RP1-244F24 to verify deletions in the RUNX2 locus. In brief, 250 ng of BACDNA was labeled with biotin-14-dATP by nick translation and hybridized to metaphase chromosomes prepared from peripheral blood lymphocytes. Quantitative PCR (Q-PCR) for copy number state was carried out using the TAG-MAN probe Hs06157701_cn SUPT3 located upstream of RUNX2 in the deleted regions together with a control probe (ribonuclease P RNA component H1) according to the manufacturer’s protocol (Applied Biosystems, Foster City, CA). Results Mutation analyses of Danish CCD families

The RUNX2 gene is represented by three transcripts (Fig. 1a), and both transcripts 1 and 3 have previously been used for exon numbering and mutation nomenclatures (2, 13, 14). Therefore, we suggest transcript 1 as the reference sequence for future mutation and exon nomenclature according to the Human Genome Variation Society (HGVS) recommendations (http://www.hgvs.org/ mutnomen/) (15). The Danish cohort represented a total of 19 families (pedigree data for all 19 families are known but not shown). The CCD phenotype was found to be segregating in 10 of the families, and 9 families represented sporadic cases (Table 2). The mutations in the sporadic cases were considered as de novo mutations. The RUNX2 gene was analyzed by DNA sequencing of all exons and exon–intron boundaries in one affected individual from each family, and single-nucleotide polymorphism array analysis for copy number variations (CNVs) was carried out for five families without a RUNX2 point mutation (Table 2). A pathogenic mutation was found in 16 of the 19 families (Fig. 1b,c and Table 2). The DNA sequence analysis failed to identify a CCD-associated mutation in five families, and SNP array copy number analysis of these families discovered a deletion of the RUNX2 locus

RUNX2 mutations in a Danish CCD cohort

a

b

c

Fig. 1. The genome organization of the RUNX2 locus. (a) The RUNX2 gene has nine exons and three different transcripts (GenBank NM_001024630, NM_001015051 and NM_004348). Transcripts 1 and 2 use the P1 promoter and code for the protein isoforms A and B; transcript 2 lacks exon 8. Transcript 3 uses promoter P2 located in intron 3 and has a different amino terminal and codes for isoform C. The amino terminals for the three isoforms are shown in the one letter amino acid code. (b) The mutations found in the Danish cohort are shown together with protein isoform A. The domains are Q/A, the poly Q/A repeat; Runt, the Runt domain (PFAM accession no. PF00853); NLS, the nuclear leading signal and PST, the proline/serine/threonine-rich domain. The border positions are shown above the protein structure and the exon positions underneath. The exon numbering follows the recommendations of Human Genome Variation Society (15). (c) The cytogenetic positions of the two deletions (CCD1 and CCD19) and the duplication (CCD4) are shown on chromosome 6p. The bacterial artificial chromosome clone RP1-224F24 used for the fluorescence in situ hybridization experiments and the positions of the genes SUPT3H and RUNX2 are shown.

in two families and a 60-kbp large duplication downstream of RUNX2 in one family (Fig. 1 and Table 2). A recessive inherited mechanism cannot be ruled out for the sporadic cases CCD4 and CCD9. RUNX2 point mutations

Six of the point mutations have not been found before (Table 2). Mutations affecting the Runt domain Arg190 codon were found in four families,

and mutations affecting the nuclear leading signal Arg225 codon were found in three families (Fig. 1b and Table 2). One mutation was found in the Ser118 codon in the polyQ/A domain, and two nonsense mutations were found in the Runt and the poly serine and threonine (PST) domains (Fig. 1b and Table 2). Four frame-shift mutations were found: one located between the polyQ/A and the Runt domains and three in the PST domain (Fig. 1b and Table 2). The mRNA 257

Hansen et al. Table 2. RUNX2 mutations in the Danish cohort of cleidocranial dysplasia (CCD) families CCD family

Family history

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Familial Sporadic Familial Sporadic Familial Sporadic Sporadic Familial Sporadic Familial Familial Familial Sporadic Familial Sporadic Familial

17 18 19

Number of affected

Mutation Nucleotide changea Exon type

Restriction enzyme

Protein changea Domainb Reference

4 1 2 1 3 1 1 1 1 2 4 3 1 2 1 4

Deletion (500 kbp) 1–5 Deletion c.568C>T 4 Missense c.568C>T 4 Missense Duplication (60 kbp) — — c.1121delG 9 Frame shift c.569G>A 4 Missense c.577C>T 4 Nonsense Not found — — Not found — — c.868C>T 7 Nonsense c.295insA 3 Frame shift c.354C>A 3 Missense c.569G>C 4 Missense c.821-822delCA 6 Frame shift c.674G>T 5 Missense c.674G>A 5 Missense

— — — — BstNI — — — — — FokI — MspI HpyCH4III — —

— p.Arg190Trp p.Arg190Trp — p.Arg374Sfs p.Arg190Gln p.Arg193Stop — — p.Gln290Stop p.Asn99fs p.Ser118Arg p.Arg190Pro p.Pro274fs p.Arg225Leu p.Arg225Gln

— Runt Runt — PST Runt Runt — — PST — Runt Runt PST NLS NLS

Sporadic Familial

1 2

c.1131delA c.674G>A

BseYI —

p.Ser378fs p.Arg225Gln

PST NLS

Sporadic

1

Deletion (750 kbp)

—

—

—

9 5

Frame shift Missense

1–5 Deletion

This study This study (16) This study (16) — This study This study (16) This study (17) — — This study (18) This study This study (17) This study This study This study This study (17, 19, 20) This study This study (17, 19, 20) This study

a Nomenclature

according to Baumert et al. (14) is based on reference sequence NM_001024630, isoform A. RUNX2 domains: Runt, Runt domain; NLS, nuclear leading sequence; PST, poly serine and threonine domain, domain structure is as per Otto et al. (13).

b

transcripts for the frame-shift mutations p.Asn99fs and p.Pro274fs code for small fusion proteins with the stop codons in exons 4 and 6, respectively, and the transcripts for these two alleles and for the two nonsense mutations are most probably removed by nonsense-mediated decay (NMD) due to the stop codons’ positions. The transcripts for the mutations p.Arg374fs and p.Ser378fs have the stop codons in exon 9 and they probably avoid NMD and result in large fusion proteins.

downstream of the RUNX2 gene (physical position 6:45,724,421-45,785,077; hg18). Family CCD4 represented a sporadic case where a sister and the two parents were healthy. The family refused to participate in further analysis, so segregation analysis and additional gene diagnostics of a possible pathogenic effect of the duplicated region were not carried out. The copy number state 1 for the deleted regions was confirmed by Q-PCR (Fig. 2b) and further proved by FISH (Fig. 3).

Copy number variation

CNV analyses of the CCD families without a RUNX point mutation revealed two large deletions and one duplication in the RUNX2 locus (Table 2 and Fig. 1c). A 499-kbp deletion in family CCD1 encompassed the 3 end of SUPT3H and exons 1–5 of RUNX2 (physical position 6:45,027,25045,527,089; hg18), and a 750-kbp deletion in CCD19 covered the entire SUPT3H gene and again exons 1–5 of RUNX2 (physical position 6:44,818,282-45,568,528; hg18). The function of SUPT3H (Homo sapiens’ suppressor of Ty 3 homolog, NM_003599) is unknown, but copy number polymorphism has been described for the SUPT3H gene region (21). In the third family (CCD4), a 60-kbp large duplication was identified 258

Familial segregation analyses

Segregation analysis was possible for two CCD families. A TAG-MAN probe upstream of RUNX2 was chosen for verification of the deletions, and the Q-PCR analysis revealed that the one affected in CCD19 and all affected in CCD1 carried one copy of the RUNX2 gene and the unaffected in CCD1 carried two copies (Fig. 2a,b). Segregation analysis of a single-nucleotide mutation was possible in family CCD5, and the diagnostic restriction enzyme analysis using BstNI for the detection of c.1325delG deletion showed segregation with the disease trait (Fig. 2c). Segregation analyses of the remaining familial cases were not possible due to lack of DNA.

RUNX2 mutations in a Danish CCD cohort

a

b

c

Fig. 2. The segregation analyses for family CCD1 and CCD5. (a) The pedigree for CCD1 shows that the deletion is a de novo deletion in the first generation and the CCD phenotype segregates with single RUNX2 copy number state 1 (shown in brackets). (b) The Q-PCR analysis shows copy number state 1 for the affected individuals II:2, III:1, IV:1, IV:2 and CCD19, and copy number state 2 for unaffected individuals II:1 and III:2. (c) The pedigree for CCD5 shows a three generation family, and Bst NI restriction enzyme analysis revealed the c.1121delG mutant allele heterozygous (two bands) in the affected individuals and only the wild-type allele (one band) for the unaffected and the control persons.

Nucleotide polymorphism

One novel SNP (c.-546C>A) was found in the RUNX2 promoter in two family genomes. Both families carried an additional RUNX2 mutation in the coding region, and the SNP was judged to be without an effect for the regulation of RUNX2 and was not analyzed in the normal population. Fragment length analyses of the polyA/Q repeat (Fig. 1b) in the 19 CCD families and 94 control persons revealed five different repeat alleles (Table 3). An 18-nucleotide deletion of the polyA repeat [(CAR)23 GAG (GCD)11 ] was found in one CCD patient and nine control persons heterozygous for the wild-type allele [(CAR)23 GAG (GCD)17 ] (Table 3). Two CCD families carried a 21-nucleotide deleted polyQ allele [(CAR)16 GAG (GCD)17 ], not found among the control individuals in addition to the wild-type allele, and finally, two different polyA repeat variations were found

to be heterozygous in two different control persons (Table 3). All the polyQ/A variations were considered without pathogenic effects because they were represented either in the normal persons and/or in the CCD families carrying a pathogenic RUNX2 mutation. The cis or trans location of the repeat variations and the pathogenic mutation were not determined. Analysis for extreme long repeat extension using a specifically designed PCR primer (Table 1) matching five (CAG) codons and CA did not detect any extreme long repeat expansions. Genotype–phenotype correlation

The genotypic and phenotypic results are summarized in Tables 2 and 4. When mean values were compared for the phenotypic data of genotypes representing missense mutations vs nonsense and frame-shift mutations, no major differences 259

260

b Data

a Nomenclature

— — 1/188 (0.005) 1/188 (0.005) [CAR]23 GAG [GCD]16 [CAR]23 GAG [GCD]18 p.Ala73del p.Ala73insAla c.217_219del c.217_insGCG

according to den Dunnen and Antonarakis (15), reference sequence NM_001024630, R represents the nucleotides A or G; D the nucleotides A or G or T. are based on fragment analyses of 94 normal individuals from the Danish population and 19 Danish CCD patients.

This study This study

This study This study This study (2)

Wild-type allele CCD2, CCD23 CCD17, normal individuals Normal individuals Normal individuals 35/38 (0.921) 2/38 (0.053) 1/38 (0.026) 177/188 (0.947) 0/188 (0.0) 9/188 (0.048) — p.Gln58_Gln64del p.Ala73_Ala78del — c.193_213del c.217_235del

[CAR]23 GAG [GCD]17 [CAR]16 GAG [GCD]17 [CAR]23 GAG [GCD]11

Allele frequencies, CCD patientsb Allele frequencies, normal individualsb PolyGln/Ala repeatb Consequencea

were observed and all values were within 1 SD from each other (data not shown). Thus, none of the two genotypes showed a tendency to be more severely affected in some of the phenotypic categories compared with the other. Furthermore, a non-parametric Wilcoxon rank-sum test (data not shown) failed to find any statistically significant differences between the two genotypes in the selected cephalometric measurements. The cephalometric analysis showed that the CCD patients in this study revealed the expected deviations in the cranial base (increased bending), the orbital region (hypertelorism), nasal bone development (reduced), mandibular prognathism (increased) and facial height (decreased) when compared with normative data (data not shown). Wormian bones were present in nearly all cases,

Nucleotide changea

Fig. 3. The hybridization of the bacterial artificial chromosome probe RP1-224F24 to spread metaphase chromosomes from CCD1 person III:1 and CDD19 revealed only signals from the wild-type chromosome 6 and not the deleted chromosome. Arrows indicate the RUNX2 locus on the wild-type chromosome as well as the deleted chromosome 6.

Table 3. RUNX2 polymorphisms found in the Danish cohort of cleidocranial dysplasia (CCD) patients and normal individuals

Origin

Reference

Hansen et al.

RUNX2 mutations in a Danish CCD cohort Table 4. Phenotypic findings CCD family 1

2 3 4 5

6 7 8 9 10

Patient ID

Body heighta

Height/weight centilea

Teeth delayed in eruptionb

Supernumerary teethb

Wormian bonesc

1.1 2.1 3.1 1.1 1.1 2.1 1.1 1.1 2.1 3.1 1.1 1.1

− +++ +++ − +++ +++ +++ +++ +++ + +++ − ++ +++ ++ ++ +++ +++ +++ NA ++ +++ +++ − − − +++ +++ ++ +++ +++ NA +++

+++ +++ +++ +++ NA +++ +++ +++ NA +++ +++ +++ NA +++ NA +++ +++ +++ +++ NA NA NA +++ +++ +++ +++ +++ +++ +++ +++ +++ NA −

NA +++ − ++ +++ +++ + NA + + +++ − +++ + − + +++ +++ ++ + +++ ++ − − ++ − +++ ++ +++ +++ +++ ++ ++

NA + + ++ ++ +++ + NA − ++ + ++ +++ + + ++ ++ ++ ++ ++ + +++ ++ − + + + + ++ +++ ++ ++ ++

NA + NA +++ +++ +++ +++ NA ++ +++ +++ +++ NA +++ +++ ++ +++ +++ +++ − ++ − NA NA ++ +++ NA NA +++ +++ NA NA +++

1.1 1.1 2.1 1.1 1.2 1.3 1.1 2.1 2.2 1.1 1.1 2.1 1.1 1.1 1.2 2.1 2.2 1.1 1.1

11

12

13 14 15 16

17 18 19

CCD, cleidocranial dysplasia; NA, not available. a Criteria

of grouping the somatic data of the CCD patients as compared with the centiles of the control data (10)

Group symbol Body height Weight/height centile b

− ≥25th centile ≥25th centile

+