1. Introduction
Hypertrophic cardiomyopathy (HCM) is a common genetic cardiac disease that affects one out of 500 individuals from the general population [
1]. It is a clinically variable and genetically heterogeneous disease. In fact, more than 20 genes were related with HCM and a total number of about 1400 distinct mutations were identified in affected patients [
2]. The most frequently encountered mutations fall within myosin heavy chain 7 (
MYH7) and myosin binding protein C (
MBPC3) [
3,
4]. Sequencing of sarcomere protein genes in patients fulfilling clinical diagnostic criteria identifies a disease-causing mutation in only 35% to 60% of cases [
5,
6,
7,
8]. Identification of an HCM-causing mutation is an important step in the disease’s clinical management, not only to better support the clinical diagnosis in the proband but also to either exclude or confirm the presence of disease-causing mutations in other family members.
Considering the extreme genetic heterogeneity of the disease and the cost of genetic testing, several attempts were made to identify the clinical predictors of an underlying mutation [
9,
10,
11]. In a large study of HCM patients genotyped for mutations in nine genes, the presence of a set of five clinical markers, including age at diagnosis <45 years, accounted for an 80% likelihood of positive genetic testing [
11].
In addition, more reliable, precise, and possibly not time-consuming molecular diagnostic approaches are needed. In this regard, Next-Generation Sequencing (NGS), which has already been applied for the diagnosis of hereditary cardiovascular conditions as well as of other diseases [
12,
13,
14,
15,
16], may represent a suitable tool. Targeted gene panels were shown to generate results with analytical quality identical to Sanger sequencing, and to have the advantage of being faster and cheaper with better coverage and sensitivity than that used in more expanded analyses.
The purpose of the present study was to analyse the yield of NGS applied to the genetic screening of a well-phenotyped Italian HCM cohort, composed of patients with both early- and late-onset diagnosis, also including patients with positive family history, and to explore the ability of NGS to accomplish the molecular diagnostic needs in clinical practice.
3. Discussion
This report describes the results of a genetic screening obtained through NGS approach in an Italian population of unrelated and clinically well characterized HCM cases, divided into two groups according to age at diagnosis. Our population included a good percentage of patients with a family history of HCM. As expected, the prevalence of familial forms was higher in the EO group, whereas the prevalence of sporadic forms was higher in the LO group.
The key finding of our investigation was the higher yield of mutation detection rate in the EO group and in patients with a family history of disease, with 90.5% of cases carrying an identified mutation. The overall yield of genetic testing was close to 50%, and, as previously reported in the literature [
4,
7,
8,
9,
11], mutations in
MYBPC3 and
MYH7 accounted for about 65% of all variants. Other mutations were found in six additional sarcomeric genes (
TNNT2,
CAV3,
MYH6,
TNNI3,
MYL2, and
MYL3) and in one non-sarcomeric gene (
GLA). Approximately a quarter of all variants were novel, most of them belonging to
MYH7. The pathogenicity of novel mutations was verified through appropriate software for analysis.
HCM is a disease characterized by a relevant heterogeneity of both morphological and clinical features. For this reason, despite the growing knowledge on its genetic basis, the establishment of a more precise genotype–phenotype correlation has been difficult to achieve.
The main original aspect of our investigation was to test through NGS a wide range of HCM-causing genes (14 sarcomeric and three non-sarcomeric) while comparing the extreme ages of disease onset and evaluating the impact of familial occurrence of the disease even in patients with late diagnosis. Due to the small sample size of the population, our study could not address the issue of a relationship between genetic variants and phenotypic characteristics of different HCM onset patients. Notably, the presence of double and triple mutations was detected mostly among younger patients, and one of them showed a more severe form of the disease.
The different rate of pathogenic mutations found in HCM patients with early and late onset of the disease was consistent with the literature [
17,
18,
19], confirming that some mutations can be found mainly in young HCM patients (
TNNT2) whereas other mutations are detected exclusively in the elderly (
MYH6) [
17,
18,
19].
In our study, a majority of patients with young age at diagnosis had a positive genetic testing (80% of cases), four-fold higher than that of the elderly and sporadic HCM cases. These data, together with previous observations, reinforce the concept that age at HCM diagnosis is a powerful predictor of positive genetic testing [
11,
17,
18,
19]. We also support the notion that family history of HCM has a key role in appropriately addressing the genetic test. In fact, among HCM patients with a late diagnosis, those with a family history of the disease had a higher rate of mutation detection (75%).
We used an expanded panel of 17 genes in the attempt to improve the mutation detection rate. With this approach we mostly confirmed the type of mutations and the mutation distribution already described in the literature for HCM. In particular, the most frequent sarcomeric gene mutations, namely those in
MYBPC3 and
MYH7, accounted for the majority of the positive findings. Moreover, six of the seven novel mutations identified in our patients were in the main sarcomeric genes (three in
MYH7, one in
MYH6, one in
MYBPC3, and one in
TNNI3). In this regard, the limitations of using a wide diagnostic panel for HCM genetic testing have been recently highlighted in one of the largest clinical genetic studies ever reported for HCM [
20]. Consistently, a panel designed only for the main HCM genes (
n = 9), was able to successfully screen a large cohort of HCM patients [
21]. Our findings support the choice of a limited, well-selected panel of HCM genes as the best tool for diagnostic purposes.
4. Materials and Methods
4.1. Patient Selection
Seventy patients with clinical diagnosis of HCM were included in the study. We selected 35 patients with early diagnosis of the disease (≤25 years, EO-early onset) and 35 patients with a late diagnosis (≥65 years, LO-late onset). All patients underwent a cardiologic evaluation as well as genetic counselling. Clinical data for each patient included a detailed personal and family history and a thorough scrutiny of the age at which HCM was first diagnosed. Both electrocardiographic and echocardiographic examinations were performed at the time of inclusion into the study. The echocardiographic parameters included both structural measurements and resting LV outflow tract gradients derived from the continuous-wave Doppler velocities. The clinical diagnosis of HCM was based on the echocardiographic demonstration of a hypertrophied and not dilated left ventricle (wall thickness >15 mm in adults, or the equivalent wall thickness relative to body surface area in children) in the absence of another cardiac or systemic disease that could produce comparable left ventricular hypertrophy [
22,
23].
The mutational status for MYH7, MYBPC3, TNNI3, TNNT2, TPM1, and MYL2 genes was already known in 22/70 patients (8 EO and 14 LO patients). All coding exons (±20 bp) of the six genes were previously analysed by Sanger sequencing. The 22 samples were included in our study as positive and negative controls for the six genes also present in our NGS panel. The seven positive controls carried mutations in MYBPC3 (EO7, EO29, EO33, EO35), MYH7 (LO13), TNNI3 (EO30), and MYL2 (EO20). The 15 negative controls for the six genes were: EO10, EO15, LO5, LO6, LO9, LO12, LO19, LO21, LO22, LO25, LO27, LO28, LO29, LO32, and LO33.
This study was conducted in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards (The approval identification number: 42 of 28 September 2007). A signed informed consent for blood sampling was obtained from all patients included in the study.
4.2. DNA Extraction and Quantification
Genomic DNA was extracted from peripheral whole blood using a commercially available kit (Invitrogen, Milan, Italy), and then quantified using Qubitds DNA HS Assay Kit on Qubit 2.0 Fluorometer (Invitrogen).
4.3. Sequencing
Seventeen genes known to be causative of HCM phenotype were selected for targeted sequencing (
Table 2). A custom panel for coding DNA (+/−25 bp of intronic flanking regions) analysis of selected genes was designed online using Ion AmpliSeq Designer 2.0.3 (
https://www.ampliseq.com/browse.action) [
24]. The final custom panel was composed of 358 amplicons divided into two primer pools for a total of 61.89 kb of DNA. The panel covered 96.47% of regions of interest (ROI). Libraries were prepared using Ion AmpliSeq Library Kit v2.0 (Life Technologies, Carlsbad, CA, USA), according to the manufacturer’s instructions. One of 16 barcodes of the Ion Xpress Barcode Adapters1-16 Kit (Thermo Fisher Scientific Life Sciences Solutions, Carlsbad, CA, USA) was added to each sample. Libraries were quantified with Qubit dsDNA HS Assay Kit on Qubit 2.0 Fluorometer (Molecular Probes, Eugene, OR, USA) and equimolar amounts of each library were used to prepare template for clonal amplification. Emulsion PCR with Ion PGM Template OT2 200 Kit (Life Technologies, Carlsbad, CA, USA) was performed on OneTouch2 Systems (Life Technologies, Carlsbad, CA, USA). Templates were enriched using Ion OneTouch ES (Life Technologies, Carlsbad, CA, USA) and prepared for 316v2 chip loading (Life Technologies, Carlsbad, CA, USA). Groups from 12 to 16 sample libraries were sequenced on each chip. Sequencing runs were performed on Ion Torrent Personal Genome Machine (PGM, Life Technologies) using Ion PGM Sequencing 200 Kit v2, according to the manufacturer’s instructions.
4.4. Alignment
Data analysis was performed using the Torrent Suite Software v.4.0.2. (Life Technologies, Carlsbad, CA, USA). Reads were aligned to human reference genome hg19 from UCSC Genome Browser [
25] and to a designed bed file from Ion AmpliSeq Designer results. Alignments were visually verified with Integrative Genomics Viewer IGV v.2.3, Broad Institute [
26].
4.5. Coverage Analysis
The average read depth and the percentage of reads that mapped on ROI out of the total number of reads (reads on target) was calculated using Coverage Analysis plug-in (Life Technologies, Carlsbad, CA, USA). For each sample the percentage of ROI covered by at least 100× and 20× using amplicon coverage matrix file was calculated.
4.6. Variant Analysis
Variant calling was performed with Variant Caller plug-in configured with germ line-low stringency parameters. Variants were annotated using Ion Reporter 4.0 software (Carlsbad, CA, USA) [
27]. Common single nucleotide variants (minor allele frequency MAF>5%, source 1000 Genomes), exonic synonymous variants, and intronic variants were removed from the analysis, while exonic non-synonymous, splice-site, and loss-of-function variants were analysed. The novel variants were analysed by means of three types of prediction software (SIFT, POLYPHEN, and PROVEAN) and classified based on the concordance of the prediction between the three types: “likely pathogenic,” “likely benign” (3/3 concordance), or “uncertain significance” (2/3 concordance).
4.7. Variant Validation
The identified variants were validated by Sanger sequencing using standard protocols. Specific primers were designed for the analysis. Polymerase Chain Reaction (PCR) products were directly sequenced by using the BigDye Terminator v3.1 Cycle Sequencing Kit (Life Technologies Corporation, Carlsbad, CA, USA). Sample analysis was performed on an ABI PRISM 3130xl Genetic Analyser (Applied Biosystems, Carlsbad, CA, USA).
4.8. Statistical Analysis
Statistical analysis was performed with SPSS statistical software (SPSS Inc., Chicago, IL, USA, version 17.0). Continuous variables are expressed as mean±SD. Comparisons between the two groups were performed using a Student’s t-test. The association between the mutational status and the clinical features of the two patient groups was evaluated using Chi-square and Fisher’s exact tests. A p value was considered statistically significant when <0.05.