Genomic DNA was isolated from white blood cells using standard techniques (DNA Extraction Kit, Qiagen, Hilden, Germany). The DNA fragments were ligated with adaptors and two paired-end DNA libraries with an insert size of 100 – 900 bp. After enrichment by polymerase chain reaction (PCR), the DNA libraries were prepared for cardiac gene enrichment re-sequencing on a HiSeq 4000 next-generation sequencing system according to the manufacturer’s instructions (TruSeq 3000/4000 SBS Kit v3 — protocol HiSeq 3000/4000 system user guide, part # 15066496, Rev. A HSC 3.3.20). The patient was screened for variants in 142 CM-associated genes using the TruSeq™ Cardiomyopathy gene panel, which resulted in 150-bp paired-end sequencing reads and at least a 100-fold average sequencing depth for each sample.

The low-quality reads were discarded from raw data to generate clean reads. The clean reads were then aligned to the human reference genome UCSC Genome Build hg19. The Illumina automated workflow was used for the alignment of the reads and base calling: the Burrows-Wheeler Aligner (BWA), Genome Analysis Toolkit (GATK), and Sequence Alignment Map (SAMtools).

For the annotation of variants, we used the ANNOVAR and VEP software programs. The variants were then computationally compared with a list of reported pathogenic variations from the Human Gene Mutation Database (HGMD, professional version). The extremely rare variants located in the exon and consensus splice sequences (± 1 and 2 positions) were filtered using a global minor allele frequency (MAF) of ≤ 0.0005. The variants that were not in the HGMD and synonymous and intronic variants that were greater than 15 bp from the exon boundaries were not considered.

The guidelines of the American College of Medical Genetics and Genomics (ACMG) were used for variant classification to assess dominant single-gene mutations with clear clinical impact on CMs (cardiomyopathies). The Exome Aggregation Consortium (ExAC) browser was used as a reference for the allele frequencies. The missense variant was assessed using a variety of prediction tools, including deleterious annotation of genetic variants using neural networks (DANN), Genomic Evolutionary Rate Profiling (GERP), MutationTaster, Protein Variation Effect Analyzer (PROVEAN), and Sorting Intolerant from Tolerant (SIFT). The degree of conservation of the affected residue was measured by multiple ortholog alignment using Alamut software (Interactive Biosoftware, Rouen, France).

All variants of interest were confirmed using Sanger sequencing. The primers for the identified variants were designed using Primer3. The PCR primer pair used to amplify the products were CALM2-F: 5’-CGCCATGTGATGACAAACCT-3’ and CALM2-R: 5’-GTTCCCTAACAAAGAGCCTC-3’. Direct sequencing of the PCR products was performed in an ABI3130XL sequencer (Applied Biosystems) according to the manufacturer’s protocol.

Figure 1

The subject of this study was a newborn baby girl who suffered from recurrent syncope and QT interval prolongation on a resting 12-lead surface ECG. The patient, who was born to healthy parents, was the family's only child. Her family history was negative for SCD, and both parents were asymptomatic regarding CVD and showed normal ECGs (data not shown).

A fetal echocardiogram at 29 weeks detected signs of congenital heart disease (CHD), including sinus arrhythmia with slowing of the heart rate, macrocardia, and highly inconsistent echocardiographic grading of aortic stenosis (AS) which was represented by bicuspid aortic valve (BAV) disease (Figure 1 A and B).

On June 10, 2018, the primigravida mother had a premature rupture of membranes (PROM) at 36 weeks of gestation. After amniorrhexis, a baby girl weighing 2,300 grams, who was in a breech presentation, was delivered in good condition by emergency caesarean section.

The neonate's diagnosis at the time of hospital admission showed a conscious state; however, her sinus rhythm was slow at a rate of 60 beats per minute (bpm).

During hospitalization, the low-birth-weight infant was well breastfed and reached a weight of 2,450 grams after one week. However, the electrocardiography showed a markedly prolonged QTc interval of 700 ms. Sinus arrhythmia was also included in the clinical record as the heart rate ranged from 60 to 100 bpm. Moreover, an inconsistent T wave alternans (TWA) signal was detected using the standard clinical leads. Echocardiograms showed a 2:1 atrioventricular (AV) block, left ventricular non-compaction cardiomyopathy (LVNC) symptoms, and a cardiac ejection fraction (EF) that ranged from 54 to 60.6% (Figure 1 C and D). Based on these clinical symptoms, the cardiologists strongly suspected LQTS type 3. The patient was then subdermally implanted with a single-chamber pacemaker beneath the collarbone which was connected to the heart by one wire. A beta-blocker and mexiletine were administered to the patient daily.

On June 29, 2018, the patient suffered from sudden cardiac arrest (SCA) and required cardiopulmonary resuscitation (CPR) for 15 minutes to provide oxygenation and circulation to the body. After this intensive medical care, propranodol was administered to the patient. A low heart rate of 50 to 60 bpm was identified. During the subsequent days, the patient required a single-chamber pacemaker again and was continually medicated with propranodol. The heart rate increased to approximately 70 bpm.

Twelve days later, on July 10, 2018, the patient presented to the emergency department due to choking (on milk), cyanosis, and cardiac arrest, and required electrical shock (ES) and respiratory devices. The heart ceased to beat due to torsade de pointes (TdP), which occurred twice and lasted for 3 seconds each time. An ECG Holter monitor showed the occurrence of polymorphic ventricular tachycardia and the heart rate ranged from 80 to 115 bpm. A digital X-ray image showed cardiac enlargement (Figure 1 E). Moreover, the patient's QTc had increased to 720 ms (Figure 1 F). Thus, the patient was medicated with adrenaline and magnesium sulfate. Two weeks later, the patient suffered from TdP for the third time and required ES and endotracheal intubation (EI). An endotracheal tube was placed through the mouth into the trachea to facilitate breathing. Unfortunately, the patient expired at 3 months of age on September 10, 2018, due to serious inflammatory issues.

Genetic testing could help to increase the diagnostic accuracy and facilitate the clinical assessment and appropriate therapy of patients who have heart failure or arrhythmias, thus, preventing SCD. Therefore, in the present study, WES was performed on the patient to identify genetic variants.

The genetic data obtained for the patient are shown in Table 1. The Illumina sequencer yielded 27,514,276 reads with an average of 151 bp in length, and the bioinformatics analyses identified a total of 88,187 SNP and 182 Indel. The GC content was 49%.

To further analyze the mutations and identify potential genes, the potential candidates were condensed and included 142 genes that have been linked to CM. These genes were further filtered based on the mapping quality MQ ≥ 40, prediction of protein alteration, missense variants, and absence from the SNP database of ADS. As a result, four variants in different genes were identified (Table 2).

The results demonstrated that the WES identified a novel heterozygous missense variant in the CALM2 gene (NM_001305626). The variant c.280G > T (p.Asp94Tyr) belonged to the exon 4, where a single nucleotide substitution occurred: G was replaced by T at position 280 in the cDNA, leading to the replacement of aspartic acid (Asp, D) by tyrosine (Tyr, Y) at the position 94 on the calmodulin protein (CaM) (Figure 2). (p.Asp94Tyr) produces the substitution of an acidic amino acid with a polar non-charged one. The physicochemical properties of Asp and Tyr are very different.

In this case, the only potentially pathogenic variant observed in the patient was c.280G > T (p.Asp94Tyr) in CALM2 (Table 2). c.280G > T (p.Asp94Tyr) is not listed in public databases, nor has it been identified in genotyping projects that included the general population and were comprised of several thousands of subjects (Exome Variant Server, ExAC). The pathogenic effect of the variant (pathogenicity) was graded according to its presence in a previously associated or candidate gene. Therefore, we concluded, given its extremely low MAF, that c.280G > T (p.Asp94Tyr) was a mutation.

The human CaM is a small 149-amino acid protein that comprises two Ca²⁺-binding lobes at its N- and C-terminus, each containing two EF-hand structural motifs with Ca²⁺-binding properties. The novel c.280G > T (p.Asp94Tyr) mutation is located at the central EF-hand calcium-binding 3 domain (Figure 2 A). Importantly, the C-terminals EF-hands 3 and 4 possess a higher Ca²⁺-binding affinity than those in the N-terminus and cluster various calmodulinopathy-causing mutations that express multiple arrhythmic phenotypes in the young, suggesting the functional importance of specific topological domains that appear to be intolerant to genetic variation20.

Table 3

Detected variant	Coding impact	ACMG classification	In silico prediction
CALM2:NM_001305626:exon4:c.G280T:p.D94Y	Nonsynonymous SNV	Likely Pathogenic (PM1, PM2, PM5, PP3)	DANN (0.9899) GERP (5.7399) MutationTaster (D) PROVEAN (D) SIFT (D)

SNV: Single Nucleotide Variant, ACMG: The American College of Medical Genetics and Genomics, DANN: Deleterious annotation of genetic variants using neural networks, GERP: Genomic Evolutionary Rate Profiling, PROVEAN: Protein Variation Effect Analyzer, SIFT: Sorting Intolerant from Tolerant, D: Deleterious effect of a variant

To investigate the in silico predicted impact on the calmodulin protein, the 2015 ACMG criteria, as well as widely used software scores, were taken into account. The five protein function prediction bioinformatics tools chosen were DANN, GERP, MutationTaster, PROVEAN, and SIFT. These software programs classified this variant as likely pathogenic (Table 3). Thus, it was extremely likely that this mutation was pathogenic and disease-causing in the patient.

To assess the possibility that the mutation c.280G > T (p.Asp94Tyr) caused SCD in the proband, WES was also performed on blood samples of the healthy parents (data not shown). Both parents were asymptomatic and did not show phenotypic abnormalities on the baseline or stress test ECGs. Additionally, they both showed a normal resting heart rate in sinus rhythm and normal exercise-induced tachycardia (data not shown). Therefore, we assumed that the mutation was not present in the patient's parents and that the mutation c.280G > T (p.Asp94Tyr) in the CaM was a de novo cause of the SCD in the patient. Consistent with the high-throughput sequencing, the results of the Sanger sequencing revealed that the patient carried a point mutation changing aspartic acid residue (Figure 3 B) while her parents did not (Figure 3 C and D). The clinical status was also indicated by the family pedigree of the patient (Figure 3 E).

The novel mutation c.280G > T (p.Asp94Tyr) altered the highly conserved aspartic acid residue that directly chelates Ca²⁺ ions in EF-hand domain 3 at position Y in the pentagonal bipyramidal coordination sphere21, predicting a high pathogenic effect. Additionally, the correlation between the patient’s symptoms and the known involvement of CaM in the heart [1,22] showed the effect of the mutation on the protein function. Lastly, c.280G > T (p.Asp94Tyr) affects a highly conserved aspartic acid at position 94. This amino acid mutation point is naturally conservative. Moreover, a comparison of the protein sequence encoded by the CALM2 gene of humans with that of other species indicated that the amino acid of the mutation was not changed (Figure 3 F).