Introduction

Familial cerebral cavernous malformations (CCMs) (MIM 116860, 603284, 603285) are autosomal dominantly inherited vascular abnormalities with genetic heterogeneity and likely interaction among gene products [15]. Exon-by-exon screening approaches found CCM1 mutations in 43–54% of familial CCMs [2, 12]. Up to 22% were shown to carry a CCM2 mutation [4, 8] and less than 10% a CCM3 mutation [1, 5, 9, 12]. Two large CCM2 deletions [4] and one deletion involving the entire CCM3 gene [1] initially contributed to identification of the CCM2 and CCM3 genes via loss-of-heterozygosity mapping. Since large genomic deletions escape detection by conventional, nonquantitative polymerase chain reaction (PCR)-based mutation analysis, we adopted the multiplex ligation-dependent probe amplification (MLPA) gene dosage assay to screen for large deletions/duplications in the CCM1–3 genes. MLPA allows the relative quantitation of up to 50 different target DNA sequences in a single reaction and has been proven to be a reliable and sensitive method [7, 10]. We here present identification of a large, heterozygous deletion that encompasses the entire CCM1 coding region.

Patients and methods

The index case is an 8-year-old boy (III-1, Fig. 1a) who experienced a generalized tonic-clonic seizure with Todd’s paralysis and headaches at the age of 5. Magnetic resonance imaging (MRI) of the brain showed multiple cavernous malformations, including a right temporomesial lesion (Fig. 1b,c). This symptomatic lesion with a diameter of 3.5 cm was completely excised microsurgically via a pterional approach. The postoperative course of the patient was uneventful. After 6 months, he did not require further antiepileptic medication.

Fig. 1
figure 1

a Pedigree of the German cerebral cavernous malformation (CCM) family (black circle = affected female, black square = affected male). b Axial T2-weighted magnetic resonance image (MRI) shows a large, right temporomesial cavernous malformation (white arrows) in the index case (arrow in a) before microsurgical excision. c Additional asymptomatic cavernous malformations (black arrows) were diagnosed in both cerebral hemispheres by gradient echo MRI. d Axial gradient echo MRI of the boy’s father shows two small asymptomatic cavernous malformations, left frontal and parietal (arrows). e One aunt (II-5 in a) also revealed an asymptomatic cavernous malformation in the left basal ganglia (arrow)

Full size image

Family history revealed a paternal grandfather with multiple intracranial lesions and fatal hemorrhage at age 47. The patient’s father is clinically unaffected, but MRI revealed multiple small cavernous malformations which have not required surgical intervention so far (Fig. 1d). Neuroimaging of three further asymptomatic aunts (II-3, II-5, and II-7; Fig. 1a) revealed a small cavernous malformation in the basal ganglia of aunt II-5 (Fig. 1a,e).

Genetic testing was approved by the local ethics committees (University of Würzburg, Study 21/05; Philipps-University Marburg, Study 149/05). With informed consent, genomic DNA was extracted from peripheral blood lymphocytes. All coding CCM1–3 exons were directly sequenced on a Beckmann CEQ 8800 capillary electrophoresis system according to published protocols, with slight modifications [1, 2, 4]. Screening for large CCM alterations requires two MLPA kits (SALSA MLPA Kits P130 & P131 CCM; MRC Holland, Amsterdam, The Netherlands). The protocol provided by MRC Holland was followed without further optimization. CCM1–3 MLPA analyses of four control individuals in each test and all ten available family members were carried out using an ABI Prism 310 genetic analyzer. Haplotype analyses were performed for the index case and his parents using 19 intragenic single nucleotide polymorphisms (SNPs) (rs975707, 1064819, 1064820, 1064821, 11984192, 17164451, 2027950, 1034575, 10223994, 10282603, 10274699, 6953959, 12113704, 11542682, 1052043, 1063658, 1063659, 11542681, 1063660) and five polymorphic markers flanking the CCM1 locus (D7S2410, D7S1813, D7S2189, D7S646, and noninformative D7S689).

Results

Direct sequencing of all coding CCM1, CCM2, and CCM3 exons and adjacent splice sites did not reveal any pathological intragenic alterations in the index patient. In contrast, only the index case but none of the controls displayed a large deletion encompassing all CCM1 exons when tested by MLPA (Fig. 2b,e,f). CCM2 and CCM3 peaks and ratios did not differ between proband and controls (Fig. 2a,c,d). A second independent MLPA analysis included all ten family members of the second and third generation (Fig. 1a). The heterozygous CCM1 deletion was confirmed in the three affected family members only (data not shown). Thus, the CCM1 deletion was reproducible, segregates with the disease, and was not transmitted to children III-2 and III-3 and uncle II-6 (Fig. 1a), rendering neuroimaging unnecessary for these individuals and their offspring.

Fig. 2
figure 2

Multiplex ligation-dependent probe amplification gene dosage assay (MLPA) data demonstrating a heterozygous deletion of the entire CCM1 gene in the index case. Example of: a normal, and b pathological MLPA raw data (SALSA MLPA Kit P130). The electropherograms show reduced peaks for all CCM1 exons in the proband (asterisks), whereas CCM2 peaks (dots) are comparable between control and patient. Peaks from internal controls are not highlighted. e, f Quantitative analyses demonstrate that the relative peak areas are decreased to approximately 50% in the patient’s two noncoding and 15 coding CCM1 exons analyzed (black) when compared with internal (white) and c, d external controls and to e CCM2 (grey) and f CCM3 (grey) probes. Numbers below the columns in c, d indicate the CCM1–3 exons analyzed according to the manufacturer’s protocol

Full size image

To further confirm the deletion detected by MLPA, haplotype analyses were performed with 19 intragenic CCM1 SNPs, none of which was found to be heterozygous in the patient and his affected father. Only three SNPs turned out to be informative. While the patient’s mother (II-2) is homozygous for G at rs975707, the father (II-1) is homo- or hemizygous for C (c.1-3078G > C). Since their son did not inherit a paternal C allele (Fig. 3), he is hemizygous for the maternal G allele. Similarly, the mother carries a homozygous C at rs2027950 and a homozygous T at rs6953959 (c.989 + 4389C > T), whereas the father’s sequence revealed a G (c.989 + 63C > G) and a C, respectively. The proband only shows the maternal C and T alleles. On the basis of the order of microsatellite markers linked to the disease locus and intragenic SNPs as D7S2410-D7S1813-D7S2189-rs975707-rs2027950-rs6953959-D7S646, the proband and his mother share the haplotype 1-2-2-G-C-T-1. The son inherited the disease haplotype 2-3-3-del-del-del-2 from his father, and this haplotype clearly lacks a second allele for three intragenic CCM1 SNPs (Fig. 3).

Fig. 3
figure 3

Haplotype analysis of the proband (III-1), his father (II-1), and his mother (II-2). While heterozygosity (i.e., the existence of two different alleles, a maternal and a paternal) for microsatellite markers flanking the CCM1 locus (D7S2410, D7S1813, D7S2189, D7S646) could be demonstrated for the index patient and his father, the proband carried only the maternal intragenic single nucleotide polymorphism (SNP) haplotype G-C-T, indicating hemizygosity (presence of only one copy of the respective DNA sequence) at the CCM1 locus for both father and son. The haplotype associated with the cerebral cavernous malformation (CCM) phenotype was determined as 2-3-3-del-del-del-2 (boxed) and lacks a second allele for three informative SNPs within the CCM1 gene (rs975707, rs2027950, rs6953959) (del = deletion), thus confirming the Multiplex ligation-dependent probe amplification gene dosage assay (MLPA) results

Full size image

Discussion

The CCM1 deletion was found in a total of five CCM families in which four novel intragenic CCM mutations had been identified by direct sequencing ([11] and unpublished data). Based on microsatellite genotyping and cDNA analyses, previous publications reported that two out of ten CCM2 mutations [4] and one out of eight CCM3 mutations [1] were large deletions. An additional CCM3 mutation was described as possibly being due to a deletion of the genomic region that encompasses exon 5 [1]. Furthermore, a genomic deletion involving the 3’ end of CCM1 exon 18 and part of intron 18 was detected [6]. We anticipate that a significant proportion of CCM patients will display large deletions or duplications that remain undetected by direct sequencing of genomic DNA.

MLPA is a suitable and efficient method for identifying such CCM gene alterations. MLPA has been reported to be more precise, accurate, and time effective than real-time PCR [3]. Furthermore, deletions larger than the entire coding region, such as the CCM1 deletion presented in this report, would escape detection by RNA-based reverse transcriptase (RT)-PCR analysis using primers from the coding region, as has been described for deletions of the NF1 gene causing neurofibromatosis type 1 [14]. Fluorescence in situ hybridization (FISH) of chromosomes is much more laborious than is MLPA and would not be sensitive enough to detect single- or multiexon deletions/duplications. However, if sequencing of genomic DNA and MLPA fail to detect a mutation, RNA-based analysis is a complementary method that allows detection of splice mutations caused by, e.g., point mutations that activate a cryptic splice motif [13] or alterations located deep within a large intron. Thus, the application of MLPA and, in some instances RT-PCR, in addition to sequencing of genomic DNA is important for improving the mutation detection rate in CCM patients which is the basis for predictive testing of at-risk relatives.