CCM1 and CCM2 variants in patients with cerebral cavernous malformation in an ethnically Chinese population in Taiwan

Cerebral cavernous malformation (CCM) is a vascular malformation characterized by clustered enlarged capillary-like channels in the central nervous system. The genes harboring variants in patients with CCM include CCM1/Krev interaction trapped-1, CCM2/MGC4607, and CCM3/programmed cell death protein 10. We aimed to identify pathogenic variants in an ethnic Chinese population in Taiwan. We recruited 95 patients with multiple CCMs or a single lesion with a relevant family history. Sanger sequencing was performed for 41 patients. Variants were identified using sequence alignment tools, and the clinical significance of these variants was determined using American College of Medical Genetics and Genomics standards and guidelines. Several pathogenic variants were found in six patients, including three unrelated patients and three affected members of one family. Two novel pathogenic variants leading to early truncation comprised a deletion variant in exon 18 of CCM1 (c.1846delA; p.Glu617LysfsTer44) and an insertion variant in exon 4 of CCM2 (c.401_402insGCCC; p.Ile136AlafsTer4). One novel pathogenic splice site variant was c.485 + 1G > C at the beginning of intron 8 of CCM1. In this study, we identified novel variants related to CCM in an ethnically Chinese population in Taiwan.

DNA analysis. Of the 95 patients, we collected blood samples from 41 (43.1%), which comprised 29 patients with unrelated sporadic cases, 9 with unrelated familial cases, and 3 who were affected members of one family. The pedigree of this family is depicted in Fig. 1. We could not obtain blood samples from the remaining 54 patients because of loss to follow-up (47) or patients' refusal (7). Five patients and one patient had pathogenic CCM1 and CCM2 variants, respectively; additionally, one patient possessed a CCM2 variant of uncertain significance. The clinical manifestations and gene analysis results are listed in Table 2. Among the patients with variants, www.nature.com/scientificreports www.nature.com/scientificreports/ one (Patient 20) underwent both surgical removal and radiotherapy for hemisphere lesions and another (Patient 54) received only radiotherapy because of symptomatic CCMs. One patient (Patient 20) experienced another episode of paraplegia, and a follow-up MRI study revealed a cavernous hemangioma in the spinal cord.
Radiological finding. All the patients with pathogenic CCM1 variants had multiple lesions in the bilateral hemisphere ( Fig. 2A, C-F), and one patient had lesions in the spinal cord, as observed in the follow-up MRI study (Fig. 2B). Not only the patient (Patient 54) with one pathogenic CCM2 variant but also the one with one CCM2 variant of uncertain significance (Patient 69) possessed a prominent pontine lesion (Fig. 2G,H,I). Those images were reviewed by experienced radiologists based on susceptibility-weighted images 21 . The number of lesions among the six patients with pathogenic variants was not affected by genes (p = 0.77), sex (p = 0.827), or age (p = 0.22).
Variant analysis and prediction. In the variant analysis, Patient 20 had two heterozygous CCM1 variants including one novel missense variant of uncertain significance (c.1844G > C, p.Ser615Thr) and one novel one-base pair (bp) pathogenic deletion variant (c.1846delA, p.Glu617LysfsTer44) simultaneously in the exon 18 (  (Fig. 3E). All previously described missense variants were located in the evolutionary conservation of sequences among functional domains (Fig. 4).

Discussion
In familial CCM, pathogenic variants in three genes have been reported [3][4][5] . Our results revealed five patients with CCM1 variants and two patients with CCM2 variants in 41 ethnically Chinese patients with multiple CCMs or a single CCM with a relevant family history. All the patients (100%) with variants had multiple lesions. The size and number of lesions in the brain MRI were smaller and lower, respectively, in patients with CCM2 variants than in those with CCM1 variants. This finding is compatible with those of previous studies, indicating that CCM2 variant carriers may have a milder phenotype than CCM1 variant carriers 22 .
In our study, for the detection rate of variants in familial CCM (7.6% [1 of 13]; three affected members of one family among 15 patients) and sporadic (3.75% [3 of 80]; 3 sporadic cases of 80 patients) was lower than www.nature.com/scientificreports www.nature.com/scientificreports/ that in previous studies. Studies have reported a variable detection rate of variants ranging from 16% to 60% in sporadic cases and from 70% to 90% in familial cases [23][24][25] . The rate may be underestimated because DNA analysis in our study could be performed only in 41 patients (43.1%). The DNA analysis technique is essential because Sanger sequencing may miss large deletions, large insertions, or duplications. Moreover, pathogenic variants of other genes causing CCM may be another factor. A case report showed that a balanced translocation between chromosome 3 and chromosome X caused decreased zona pellucida-like domain-containing protein 1 (ZPLD1) expression and led to multiple cerebral cavernous hemangiomas 26 . The function of ZPLD1 in pathogenesis of CCMs is unknown.
Among the patients with CCM1 variants, one sporadic case (Patient 20) had a combination of a novel missense variants (c.1844G > C, p.Ser615Thr) and a novel pathogenic deletion variant (c.1846delA, p.Glu617LysfsTer44), leading to early truncation. One SNP was present at the same location with different base pair substitutions (rs780608959: c.1844G > A, p.Ser615Asn). No information was available regarding the clinical significance of this SNP in dbSNP; therefore, this missense variant was still considered to be a variant of uncertain significance. The patients with familial CCM (Patients 57, 66, and 67) had one deletion variant (c.1255-4_1255-2delGTA), which had been reported as a pathogenic splice site variant in ClinVar. One case series revealed that six affected members of one family with the same variant presented with seizure and cutaneous vascular lesions, whereas our patients only had headache or visual disturbance. The heterogeneity of clinical symptoms in patients with the same variant may relate to underlying diseases 27 , immune responses 28 , or post-transcriptional modifications 29 . Another novel splice site variant (c.485 + 1G > C), which belonged to one patient (Patient 40), affected the invariant splice donor consensus sequence, leading to abnormal splicing products.
In patients with CCM2 variants, one patient with sporadic CCM (Patient 54) had a first reported insertion variant, leading to a frameshift and early truncation (c.401_402insGCCC, p.Ile136AlafsTer4). The other patient (Patient 69) had a CCM2 missense variant (c.970G > A, p.Glu324Lys) of uncertain significance, which was already recorded in the ExAC databank.
In total, 14 studies have investigated CCM1 and CCM2 variants in ethnically Chinese populations. Detailed information is listed in Table 3 [6][7][8][9][10][11][12][13][14][15][16][17][18][19][20] , and the locations of variants are shown in Fig. 4. Among those studies conducted in the ethnic Chinese population, most were case reports. One case series recruited five families and reported three novel variants causing early truncation 15 . Only two studies performed molecular screening of patients with CCMs from their brain MRI data 8,9 . These studies were conducted in 2004 and 2005, and one of the two reported only missense variants. In our study, we enrolled the largest number of patients from an ethnically Chinese population (41 patients) for molecular screening.
Our study has several limitations. First, because this was a retrospective study, information bias may have been present. Second, DNA analysis was performed for less than half of our recruited patients; variants may not be detected among those without DNA analysis. Third, Sanger sequencing may fail to detect large deletions, large insertions, or duplications; therefore, other detection methods, such as multiplex ligation-dependent probe amplification (MLPA), can be applied to our recruited patients to detect more variants in future studies. Finally, there were two identified variants of uncertain significance according to ACMG standards and guidelines; therefore, biological methods such as genome editing or genotyping were necessary to verify the clinical significance of those variants and to clarify the pathogenesis of familial CCMs.
In conclusion, we identified one novel pathogenic deletion and one pathogenic splice site variant in CCM1 and one novel pathogenic insertion variant in CCM2. The findings expand the knowledge related to variants present in patients with CCM, especially in the ethnic Chinese population.

Methods
Ethics approval and consent to participate. This study was approved by the Institutional Review Board of Chang Gung Memorial Hospital (CGMH) in Taiwan (IRB 96-1772B and 100-1666C). All examinations were performed after written informed consent was obtained. All samples were collected and analyzed in accordance with the relevant guidelines and regulations. Patient recruitment. We retrospectively reviewed patients who had received a diagnosis of CCM on the basis of brain MRI findings from 1998 to 2006 at CGMH. We recruited patients who had multiple lesions or a single lesion with a relevant family history. Afterward, we collected radiological and clinical assessment data including age at onset, initial presentation, underlying diseases, family history, and surgical or radiosurgical interventions.
DNA extraction and sequencing. Venous blood samples were collected and DNA was routinely extracted using a DNA extraction kit (Stratagene La Jolla, California, United States). The extraction was monitored quantitatively and qualitatively through UV spectrophotometer absorption (ND-1000) (Nanodrop, Wilmington, United States) to verify the purity of the sample depending on the absorption at 260 nm. www.nature.com/scientificreports www.nature.com/scientificreports/ Thereafter, to selectively amplify a specific DNA fragment through polymerase chain reaction (PCR), suitable primers were designed according to Primer Express version 2.0 (Applied Biosystems, United States) and online Primer3 (http://fokker.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi). Our primers were short synthesized oligonucleotides, ranging from 18 to 25 bp. We used a pair of forward and reverse primers for each exon in the template DNA strand of those three genes. The primers and amplicon size of each exon are listed in Supplementary  Table 1. PCR was performed using the Fast-Run Taq Master kit (Pro Tech, Taipei, Taiwan). A typical 50-μL solution contained (1) 10 × Taq Master Mix, (2) 0.5 µM each of forward and reverse primers, (3) 100 ng of a genomic DNA template, (4) dimethyl sulfoxide for optimal performance, and (5)    missense variant was searched in gnomAD or ExAC databases. Variants were considered to be pathogenic according to American College of Medical Genetics and Genomics (ACMG) standards and guidelines 30 . Furthermore, we analyzed evolutionary conservation in MutationTaster. Variants in evolutionary conservation may be more inclined to have clinical significance.
We followed the Human Genome Variation Society's (HGVS) recommendations to describe sequence variants, and we used the position of coding DNA (cDNA) sequences to exhibit the position of variants. For example, a cDNA sequence with a first nucleotide corresponded to A of ATG (translation initiation codon).

Statistical methods and data analysis.
Because of the small sample size, we used Mann-Whitney U test to compare the two groups. Correlations between two continuous variables were analyzed using Spearman rank correlation. We performed all analyses using SPSS version 24 (IBM, Armonk, NY, USA). A p value of < 0.05 was considered significant.