Screening Mutations of MYBPC3 in 114 Unrelated Patients with Hypertrophic Cardiomyopathy by Targeted Capture and Next-generation Sequencing

Hypertrophic cardiomyopathy (HCM) is a major cause of sudden cardiac death. Mutations in the MYBPC3 gene represent the cause of HCM in ~35% of patients with HCM. However, genetic testing in clinic setting has been limited due to the cost and relatively time-consuming by Sanger sequencing. Here, we developed a HCM Molecular Diagnostic Kit enabling ultra-low-cost targeted gene resequencing in a large cohort and investigated the mutation spectrum of MYBPC3. In a cohort of 114 patients with HCM, a total of 20 different mutations (8 novel and 12 known mutations) of MYBPC3 were identified from 25 patients (21.9%). We demonstrated that the power of targeted resequencing in a cohort of HCM patients, and found that MYBPC3 is a common HCM-causing gene in Chinese patients. Phenotype-genotype analyses showed that the patients with double mutations (n = 2) or premature termination codon mutations (n = 12) showed more severe manifestations, compared with patients with missense mutations (n = 11). Particularly, we identified a recurrent truncation mutation (p.Y842X) in four unrelated cases (4/25, 16%), who showed severe phenotypes, and suggest that the p.Y842X is a frequent mutation in Chinese HCM patients with severe phenotypes.

MYBPC3 mutations. This technology can analyze large genomic regions at lower cost and faster time than conventional Sanger sequencing. It is also regarded as the most advantageous technology in finding almost all types of mutations including small indel mutations [10][11][12] . Here we report the mutation spectrum of MYBPC3 in a large cohort of unrelated Chinese patients with HCM using this approach, and explore the clinical characteristics and their correlation with different MYBPC3 genotypes.

Results
Demographic and clinical characteristics. A total of 114 unrelated patients with HCM were recruited from 2012 to 2013. Detailed demographic and clinical characteristics of the patients are summarized in Table 1. Particularly, 29 patients (25.4%) had familial history of HCM, and 12 patients (10.5%) had a family history of unexplained SCD. None of these patients had history of genetic counseling or gene diagnosis. Sequence analysis. Overall, the mean read depth for the exome sequence of MYBPC3 was > 300× .
For each subject, more than 78% of the targeted bases were sequenced at least 20 times. The filtering process of targeted capture and sequencing data from 114 HCM patients were shown in Fig. 1. A total of 145 potential mutations (e.g., nonsynonymous, nonsense, and splice-site mutations) and 7 InDels were identified in the 114 patients. By filtering multiple databases (dbSNP137, HapMap, 1000 Genome, ESP6500, and 300 Chinese Han exome in-house database), we identified 8 novel mutations (2 missense, 1 nonsense, and 5 InDel-induced frameshift mutations) as well as 12 known causative mutations (as shown in the HGMD database). All of these 20 mutations were predicted to be probably pathogenic by SIFT and Polyphen-2, and further confirmed by Sanger sequencing analyses (the forward sequencing data are shown in Fig. 2b; the reverse sequencing results are included in Supplementary Fig. S2  were identified with 20 different mutations in MYBPC3 (Table 2), including 3 nonsense, 6 InDels, and 11 missense mutations. Here we show the 8 novel mutations and one recurrent mutation on the schematic protein structure (Fig. 2a) and the confirmation of the mutations by Sanger sequencing (Fig. 2b). Four mutations (i.e., p.G37X, p.I49S, p.I603V and p.A741fs) were located in the immunoglobulin (Ig) domains, two mutations (p.P1208fs and p.P1245fs) located in the Ig C2 domain, and only one mutation (p.Y842X) located in the FN3 domain. All of the amino acid residues that are mutated are highly conserved across multiple species from zebrafish to human ( Supplementary Fig. S1).
We further analyzed the correlation of the phenotype-genotype in the patients. Among the 12 patients carrying premature stop codon (PTC) mutations, two patients (R0513 and R0585) carrying novel frame shift mutations (p.P330fs and p.P1245fs) had familial SCD history, severe ventricular arrhythmias or history of receiving septum myectomy due to severe hypertrophy (Table 3). Of note, the p.Y842X, producing a PTC in exon 24, was identified in four unrelated patients showing severe manifestations of hypertrophy requiring surgery or symptomatic arrhythmia. In addition, one patient with the p.S139X mutation was hospitalized for receiving implantable cardioverter-defibrillator. Two patients with severe manifestations of the disease were detected carrying double mutations (p.G37X and p.R160W; p.I603V and p.R810H). Notably, the p.G37X mutation is the closest premature stop mutation to the starting codon identified thus far, which is predicted to produce a full-truncated protein. And yet, its combination with p.R160W in the patient R0248 may have led to the onset of the disease at younger age, severe clinical symptoms as well as severe hypertrophy for receiving septum myectomy. It was not determined whether these double mutations were compound mutations or simply occurred in the same allele (cis). PCR and Sanger sequencing were performed on the available DNA samples from patient (R0164)'s mother and child. The fact that no mutations (p.I603V and p.R810H) were detected in the family members ( Supplementary Fig. S3) suggests that the double mutations were derived, as single allele, from the deceased father. However, the possibility of a de novo mutation that may occur in either allele in germline cells of fertilized egg or somatic cells of embryonic tissues could not be excluded.
Together, our results have shown that the patients with double mutations (n = 2) or PTC mutations (n = 12) are correlated with more severe manifestations requiring invasive therapies, compared with patients with missense mutations (n = 11) (p = 0.01). However, no correlation was found between other clinical indications and missense mutation group, PTC group and double mutations group.

Discussion
This study for the first time provided exome sequence analysis of MYBPC3 in Chinese patients by targeted capture and next-generation sequencing. Mutational screening yielded a genetic diagnosis of 25 patients (21.9%) in 114 unrelated HCM patients, suggesting MYBPC3 is the predominant HCM-causing gene in Chinese cases. The prevalence is similar to the Finnish patients 13 , but higher than that of 12.3% in Denmark 14 or 15% in European 15 HCM cases with MYBPC3 mutations.
Although the spectrum of clinical phenotypes was broad, severe manifestations that required invasive therapies were more frequently found in our patients with disruptive mutations or double mutations, compared to the missense mutations. This finding is consistent with that in 110 consecutive, unrelated patients of European descent 15 . However, Nonsense mutations of MYBPC3 were also found to be associated with relatively benign clinical course in Japanese and French families [16][17][18] , suggesting the existing of genetic modifiers for HCM.
It is worth mentioning that among all mutations, the p.Y842X mutation was identified in 4 patients, indicating it is a frequent mutation in Chinese cases. Remarkably, all the patients with p.Y842X showed severe phenotype and ventricular arrhythmias, and mostly requiring invasive therapies. The p.Y842X was also observed in another Chinese pedigree, leading to severe hypertrophy and diastolic dysfunction 19 . In Caucasian, however, only one HCM case was previously reported with p.Y842X 14 . Nevertheless, our data for the first time showed strong association between p.Y842X mutation and severe phenotype in this cohort, indicating this nonsense mutation may be used as a predictor for earlier clinical intervention.
In summary, we have successfully applied the targeted capture and next-generation sequencing technique as a diagnostic tool and identified multiple novel mutations in MYBPC3, including the recurrent Y842X mutations, in Chinese patients with HCM. The detection of MYBPC3 mutation, especially the PTC mutation and double-mutation, may serve as a molecular marker for clinical risk stratification of HCM.

Methods
Patient selection. 114 unrelated Han Chinese patients, diagnosed as HCM, were consecutively recruited at Anzhen hospital, Capital Medical University, Beijing. Clinically, typical HCM was determined by maximal left ventricular wall thickness (LVWT) ≥15 mm in echocardiography, while atypical HCM was considered by the wall thickness of 13 to 14 mm in the presence of other compelling information (e.g., family history of HCM) 1 . Other loading conditions such as hypertension or aortic valve stenosis were excluded during the screening. Experimental protocol was approved by the Ethics Committee of the Anzhen hospital, and was carried out in accordance with the approved guidelines. Informed consent for genetic analysis was obtained from each of the patients.

Bioinformatics analysis.
High-quality reads were retrieved by filtering low quality reads and adaptor sequences using the Solexa QA package and the cut adapt program (http://code.google.com/p/cutadapt/), respectively. SOAP aligner program was used to align the clean read sequences to the human reference genome (hg19). After PCR duplicates were removed by the Picard software, SNPs was firstly identified using the SOAPsnp program (http://soap.genomics.org.cn/soapsnp.html). Subsequently, the selected reads were realigned to the reference genome using BWA. Insertions and deletions (InDels) were identified using the GATK program (http://www.broadinstitute.org/gsa/wiki/index.php/Home_Page). Identified SNPs and InDels were annotated using the Exome-assistant program  Table S1) was applied to confirm the identified mutations 20 .