A Splice Variant of the MYH7 Gene Is Causative in a Family with Isolated Left Ventricular Noncompaction Cardiomyopathy

Variants of the MYH7 gene have been associated with a number of primary cardiac conditions, including left ventricular noncompaction cardiomyopathy (LVNC). Most cases of MYH7-related diseases are associated with such variant types as missense substitutions and in-frame indels. Thus, truncating variants in MYH7 (MYH7tv) and associated mechanism of haploinsufficiency are usually considered not pathogenic in these disorders. However, recent large-scale studies demonstrated evidence of the significance of MYH7tv for LVNC and gave rise to an assumption that haploinsufficiency may be the causal mechanism for LVNC. In this article, we present a family with isolated LVNC and a heterozygous splice variant of the MYH7 gene, analyze possible consequences of this variant and conclude that not all variants that are predicted truncating really act through haploinsufficiency. This study can highlight the importance of a precise assessment of MYH7 splicing variants and their participation in the development of LVNC.


Introduction
Left ventricular noncompaction cardiomyopathy (LVNC) is characterized by the presence of a "spongy" hypertrabeculated endocardial layer of the heart wall. During recent years, along with the improvement of instrumental diagnostics, the detectability of LVNC has been growing. First by discovering the high variability of its clinical presentation, and second by highlighting the need for genetic testing in order to avoid excessive diagnostics. It is now recognized that genetic diagnostics is necessary for correct clinical management and risk stratification in patients with LVNC [1]. At the same time, extreme genetic and phenotypic overlapping of LVNC and other primary cardiomyopathies makes this task challenging.
Many genes have been reported in connection with LVNC [2], but only few of them show robust association, confirmed by multiple studies. Almost all such genes encode sarcomere proteins and are associated with other types of primary cardiomyopathies [1]. One such gene, MYH7 (OMIM #160760), is responsible for hypertrophic cardiomyopathy (HCM), dilated cardiomyopathy (DCM), left ventricular noncompaction (LVNC), Laing distal myopathy, and myosin storage myopathy [3][4][5]. 2 of 10 During recent years, growing efforts are being made to distinguish the LVNC-causing genetic variants from those responsible for other types of cardiomyopathies. Recent results in this area of genetic research draw attention to predicted truncating variants in the MYH7 gene (MYH7tv), which include nonsense substitutions, frameshifts, and splice donor/acceptor variants [6].
Some splice variants do not lead to a premature stop codon and, therefore, do not result in haploinsufficiency. These splice variants cause exon skipping without frame shift [7] and lead to a dominant negative effect identical to that of missenses and in-frame indels, the prevalent types of disease-causing alterations in MYH7 [8,9]. In-frame exon skipping was previously described in MYH7-related myopathic conditions [10], but not in primary cardiomyopathies. However, more recent studies demonstrate that the role of MYH7tv in LVNC pathogenesis may be underestimated.
In this article, we present a family of two generations with isolated LVNC and a heterozygous variant altering a canonical splice site in the MYH7 gene. Bioinformatics analysis showed that the most likely outcome of the canonical donor splice site variant, skipping of exon 10, would result in translation of an aberrant protein product missing 33 amino acid residues, rather than premature truncation. This study presents an example of this pathogenetic mechanism in LVNC and will highlight the importance of correct estimation of genetic variants that are predicted truncation.

Clinical Investigation of the Patients
Two generations of a family with LVNC were admitted at the National Medical Research Center for Therapy and Preventive Medicine (Moscow, Russia). Family members underwent clinical examination, which included blood sample collection, biochemical and general examination, electrocardiography using 24-h Holter monitoring electrocardiogram (HM-ECG), cardiac magnetic resonance imaging (cMRI), and echocardiography (ECHO) with contrast. ECHO and cMRI imaging criteria of LVNC were applied, as previously suggested by Jenni et al. [11] and Petersen et al. [12]. Our study was conducted in accordance with the Declaration of Helsinki in its current form and approved by the Institutional Review Boards of the National Medical Research Center for Therapy and Preventive Medicine (Moscow, Russia). Every participant and/or their legal representative gave their written informed consent to be involved in this study.

Cardiac Magnetic Resonance Imaging
cMRI was performed with a 1.5-T imager (Magnetom Avanto, Siemens, Munich, Germany) using a standard protocol. Breath-hold cine MRI was performed using ECGgated segmented true fast imaging with steady-state free-precession (SSFP). Cine MRI was acquired in long-axis and short-axis planes covering the whole LV and right ventricle (RV). Late gadolinium enhancement (LGE) images were acquired in the same planes 15 min after IV injection of the gadolinium contrast agent (Gd-DTPA-BMA, Omniscan, GE Healthcare Inc., Chicago, IL, USA) in a dose of 0.15 mmol/kg using inversion-recovery turbo fast low-angle shot (FLASH) pulse sequence.

Whole Genome Sequencing and Bioinformatic Analysis
Molecular genetic analysis was performed in the Centre for Strategic Planning and Management of Biomedical Health Risks (Moscow, Russia) as detailed by Meshkov et al. [13]. We extracted DNA from whole blood, then performed whole genome sequencing (WGS) on Novaseq 6000 (Illumina, San Diego, CA, USA) up to 30× coverage. After that, SNPs and long structural variants were called via Dragen Bio-IT platform (Illumina, San Diego, CA, USA) and Smoove software, respectively. Clinical interpretation of potentially clinically relevant findings was performed in accordance with current guidelines, considering the modified frameworks provided by ClinGen [14,15]. Validation of NGS results was carried out at the National Medical Research Center for Therapy and Preventive Medicine (Moscow, Russia) by Sanger sequencing using the Applied Biosystem 3500 Genetic Analyzer (Thermo Fisher Scientific, Waltham, MA, USA) in accordance with the manufacturer's protocol. For verification by Sanger sequencing the following oligonucleotides were used: 5 AACCAATGGCCAGCGTCTTA-3 and 5 -TCCTTGTGCCCAAACCCTAA-3 .

Results
The clinical features of the investigated patients are summarized in the Figure 1 and Table 1 (see below).
SNPs and long structural variants were called via Dragen Bio-IT platform (Illumina, San Diego, CA, USA) and Smoove software, respectively. Clinical interpretation of potentially clinically relevant findings was performed in accordance with current guidelines, considering the modified frameworks provided by ClinGen [14,15]. Validation of NGS results was carried out at the National Medical Research Center for Therapy and Preventive Medicine (Moscow, Russia) by Sanger sequencing using the Applied Biosystem 3500 Genetic Analyzer (Thermo Fisher Scientific, Waltham, MA, USA) in accordance with the manufacturer's protocol. For verification by Sanger sequencing the following oligonucleotides were used: 5′ AACCAATGGCCAGCGTCTTA-3′ and 5′-TCCTTGTGCCCAAACCCTAA-3′.

Results
The clinical features of the investigated patients are summarized in the Figure 1 and Table 1 (see below).    In the proband (III-4 Table 1) mitral valve prolapse has been observed since childhood. At the age of 17, he underwent a medical examination. ECHO investigation revealed the signs of myocardial noncompaction (Jenni, Chin, and Stollberger criteria): end diastolic volume (EDV) of 157 mL, ejection fraction (EF) of 48%, and end diastolic diameter (EDD) of 5.8 cm. cMRI ( Figure 2) showed two-layered structure of myocardium-the thickness of the non-compact layer was 30 mm and the thickness of the compact layer was 5 mm. The indexed left ventricular EDV was 97 mL/m 2 . In blood tests, all parameters were within normal values. B-type natriuretic peptide (BNP) level was 100 pg/mL. ECG showed sinus rhythm. According to the Holter ECG, he had a sinus rhythm with the heart rate between 45-69-130 per min without arrhythmia. The patient undergoes dynamic examination, constantly takes bisoprolol 2.5 mg/day, spironolactone 25 mg/day, and torasemide 10 mg/day. 5 mm. The indexed left ventricular EDV was 97 mL/m 2 . In blood tests, all parameters were within normal values. B-type natriuretic peptide (BNP) level was 100 pg/mL. ECG showed sinus rhythm. According to the Holter ECG, he had a sinus rhythm with the heart rate between 45-69-130 per min without arrhythmia. The patient undergoes dynamic examination, constantly takes bisoprolol 2.5 mg/day, spironolactone 25 mg/day, and torasemide 10 mg/day. The proband's father (II-5), 53 y.o., underwent a cardiological examination. According to the ECHO, the heart chambers were not expanded, with an EDD of 5.4 cm, an EF of 52%, and signs of non-compact myocardium (Jenni and Chin criteria) in the area of the apex and lateral wall. According to the HM-ECG, there were no data that suggested lifethreatening cardiac arrhythmias. According to the cMRI, the heart chambers were not dilated. Signs of non-compact myocardium were present.
The proband's 52-year-old mother (II-6) and 17-year-old brother (III-5) underwent a cardiological examination. According to the ECHO, the heart chambers were not dilated and there were no signs of non-compact myocardium.
To establish the genetic diagnosis, WGS and bioinformatic analysis were performed for the proband and his three relatives (mother, father, and brother). We found the canonical splice site variant in the MYH7 gene (t.NM_000257.4:c.895+1G>A; The proband's father (II-5), 53 y.o., underwent a cardiological examination. According to the ECHO, the heart chambers were not expanded, with an EDD of 5.4 cm, an EF of 52%, and signs of non-compact myocardium (Jenni and Chin criteria) in the area of the apex and lateral wall. According to the HM-ECG, there were no data that suggested life-threatening cardiac arrhythmias. According to the cMRI, the heart chambers were not dilated. Signs of non-compact myocardium were present.
The proband's 52-year-old mother (II-6) and 17-year-old brother (III-5) underwent a cardiological examination. According to the ECHO, the heart chambers were not dilated and there were no signs of non-compact myocardium.
To establish the genetic diagnosis, WGS and bioinformatic analysis were performed for the proband and his three relatives (mother, father, and brother). We found the canonical splice site variant in the MYH7 gene (t.NM_000257.4:c.895+1G>A; g.hg38.chr14:23430900C>T) in the proband's and his father's genome, but not in the genomes of other examined relatives ( Figure 3). This variant is absent in gnomAD and other population databases but is reported in dbSNP [17] and in ClinVar [18]. The ClinVar entry reports it as the variant of uncertain significance (VUS), but no phenotype data are provided by the submitter. We performed the in silico splicing prediction analysis via MaxEntScan v1.0 (Christopher Burge Laboratory, Department of Biology, Massachusetts Institute of Technology, Cambridge, MA, USA) [19] and SpliceAI v1.3.1 (Kishore Jaganathan, Illumina, Inc., San Diego, CA, USA) [20] software and discovered a strong probability of donor splice site loss (see Table 2). A functional RNA study that could establish the impact of the variant on the gene and its product was not conducted due to family refusal.
Institute of Technology, Cambridge, MA, USA) [19] and SpliceAI v1.3.1 (Kishore Jaganathan, Illumina, Inc., San Diego, CA, USA) [20] software and discovered a strong probability of donor splice site loss (see Table 2). A functional RNA study that could establish the impact of the variant on the gene and its product was not conducted due to family refusal.

Discussion
Currently, the prognosis of the disease in patients with LVNC varies from favorable to fulminant course requiring heart transplantation [21]. Here we present a family with favorable clinical course harboring the splicing variant in the MYH7 gene.
To date, haploinsufficiency of MYH7 is generally considered to be tolerated. ClinGen Expert Panel recommendations discourage the use of PVS1 ("very strong") criterion of pathogenicity for predicted truncating variants in MYH7 because such variants are thought to have only moderate impact. Thus, heterozygous predicted MYH7tvs, per se, are generally understood as non-causative [22].
The discovered variant c.895+1G>A leads to the disruption of the canonical donor splice site after exon 10 (out of 40, of which 1st and 2nd are non-coding). The analysis of the gene sequence by Mutalyzer software [23] showed that this disruption is unlikely to lead to the frame shift. Exon 10′s borders are in the same reading frame (its length is a multiple of three base pairs). We speculate that the most likely outcome of the canonical

Discussion
Currently, the prognosis of the disease in patients with LVNC varies from favorable to fulminant course requiring heart transplantation [21]. Here we present a family with favorable clinical course harboring the splicing variant in the MYH7 gene.
To date, haploinsufficiency of MYH7 is generally considered to be tolerated. ClinGen Expert Panel recommendations discourage the use of PVS1 ("very strong") criterion of pathogenicity for predicted truncating variants in MYH7 because such variants are thought to have only moderate impact. Thus, heterozygous predicted MYH7tvs, per se, are generally understood as non-causative [22].
The discovered variant c.895+1G>A leads to the disruption of the canonical donor splice site after exon 10 (out of 40, of which 1st and 2nd are non-coding). The analysis of the gene sequence by Mutalyzer software [23] showed that this disruption is unlikely to lead to the frame shift. Exon 10 s borders are in the same reading frame (its length is a multiple of three base pairs). We speculate that the most likely outcome of the canonical donor splice site variant, skipping of exon 10, would rather result in translation of an aberrant protein product missing 33 amino acid residues than in premature truncation ( Figure 4) [7]. The affected codons encode myosin motor domain and are critically important for proper function of myosin fibrils [14]. Based on all of the above, we apply the modified PVS1_Strong pathogenicity criterion to our variant, as established in [15], and eventually evaluate this variant as likely pathogenic with the following criteria: PVS1_Strong, PM2 (low populational frequency), PP1 (familial cosegregation with the phenotype). errant protein product missing 33 amino acid residues than in premature truncation (Figure 4) [7]. The affected codons encode myosin motor domain and are critically important for proper function of myosin fibrils [14]. Based on all of the above, we apply the modified PVS1_Strong pathogenicity criterion to our variant, as established in [15], and eventually evaluate this variant as likely pathogenic with the following criteria: PVS1_Strong, PM2 (low populational frequency), PP1 (familial cosegregation with the phenotype). A recent large-scale study by Mazzarotto et al. [6] showed a significant prevalence of predicted truncating (i.e., nonsense, frameshift, and donor/acceptor splicing) MYH7 variants in the cohorts of patients with LVNC. The authors compared frequencies of rare variants in six LVNC cohorts (a total of 840 patients) with population control frequencies derived from gnomAD and revealed a 20-fold enrichment of predicted MYH7tv in the LVNC cases versus controls. Mazzarotto et al. also found that the predicted MYH7tv were enriched specifically in the LVNC cases, and not enriched in the HCM or DCM cohorts. Furthermore, in ostensibly healthy controls harboring predicted MYH7tv, myocardial hypertrabeculation was observed. These findings suggest haploinsufficiency to be the causative mechanism for isolated noncompaction, and attract attention of clinicians to such variants. Nevertheless, it should be noted that the conclusion about the prevalence of MYH7 haploinsufficiency in the mentioned study implied the assignment of all splicing variants to MYH7tv. Variation of the donor splicing region after exon 8 (c.732 splice region) was repeatedly revealed in different LVNC cohorts [6]. This region was considered A recent large-scale study by Mazzarotto et al. [6] showed a significant prevalence of predicted truncating (i.e., nonsense, frameshift, and donor/acceptor splicing) MYH7 variants in the cohorts of patients with LVNC. The authors compared frequencies of rare variants in six LVNC cohorts (a total of 840 patients) with population control frequencies derived from gnomAD and revealed a 20-fold enrichment of predicted MYH7tv in the LVNC cases versus controls. Mazzarotto et al. also found that the predicted MYH7tv were enriched specifically in the LVNC cases, and not enriched in the HCM or DCM cohorts. Furthermore, in ostensibly healthy controls harboring predicted MYH7tv, myocardial hypertrabeculation was observed. These findings suggest haploinsufficiency to be the causative mechanism for isolated noncompaction, and attract attention of clinicians to such variants. Nevertheless, it should be noted that the conclusion about the prevalence of MYH7 haploinsufficiency in the mentioned study implied the assignment of all splicing variants to MYH7tv. Variation of the donor splicing region after exon 8 (c.732 splice region) was repeatedly revealed in different LVNC cohorts [6]. This region was considered a "hotspot" for MYH7-dependent LVNC with haploinsufficiency as a supposed pathogenetic mechanism. However, exon 8 is in-frame (Figure 4), and alterations of its donor splice site likely cause skipping of a fragment of 31 amino acids, not premature termination. This suggests that the possible mechanism of disease in c.732 splice region variants is dominant negative effect (gain of function), not haploinsufficiency ( Figure 5). a "hotspot" for MYH7-dependent LVNC with haploinsufficiency as a supposed pathogenetic mechanism. However, exon 8 is in-frame (Figure 4), and alterations of its donor splice site likely cause skipping of a fragment of 31 amino acids, not premature termination. This suggests that the possible mechanism of disease in c.732 splice region variants is dominant negative effect (gain of function), not haploinsufficiency ( Figure 5). The MYH7 gene has 40 exons (38 coding), and 14 of these ( Figure 4) can be skipped without disrupting the reading frame. To date, 100 donor and acceptor splice site variants in MYH7 are present in dbSNP, with the highest minor allele frequency of 0.001647% (according to ExAC [24]), which fits the PM2 population criterion of pathogenicity of the MYH7 variants [14]. Of them, 39 are adjacent to one of the in-frame exons and can likely lead to production of an aberrant protein product (Supplementary Table S1). However, The MYH7 gene has 40 exons (38 coding), and 14 of these ( Figure 4) can be skipped without disrupting the reading frame. To date, 100 donor and acceptor splice site variants in MYH7 are present in dbSNP, with the highest minor allele frequency of 0.001647% (according to ExAC [24]), which fits the PM2 population criterion of pathogenicity of the MYH7 variants [14]. Of them, 39 are adjacent to one of the in-frame exons and can likely lead to production of an aberrant protein product (Supplementary Table S1). However, only twelve of them are reported in ClinVar, including three pathogenic, four likely pathogenic and five VUS. Meanwhile, as evidence of possible functional roles of some variants/variant types accumulates, ClinVar entries may serve as a useful source of information for clinical interpreters, especially when there are several recordings made by different submitters. In view of emerging evidence of possible contribution of predicted MYH7tv (including splice variants) to LVNC, accumulation of new and existing observations in open-source databases seems essential.
The MYH7-associated LVNC is characterized by a relatively benign clinical course in comparison to other genetic forms of the disease. The 2018 study by Waning et al. that included 327 patients with LVNC [25] showed that the risk of decreased systolic function was higher for genotype-positive patients compared to negative ones (p = 0.024), but patients with variants in MYH7 had lower probability of major adverse cardiac events compared to those with variants in other genes (p = 0.03). This conclusion was confirmed in the previously published work of Sedaghat-Hamedani et al., who revealed the relationship between poor prognosis in patients and variants in the LMNA and RBM20 genes [26]. In patients with variants in the TTN and MYH7 genes, such a correlation was not observed [26]. In our earlier study we presented a family with LVNC where the proband and his father, both harboring the novel MYH7 missense variant, had moderate signs of systolic dysfunction and remodeling of the left ventricle, in the absence of intramyocardial fibrosis, which allowed us to classify those patients with a relatively good prognosis [27]. All these results represent the clinical features of the disease caused by MYH7 dominant negative variants, and the phenotypes of affected family members in the present study are consistent with it. Although in absence of the functional study we cannot confidently confirm that in our case splicing variant in MYH7 definitely leads to dominant negative effect, the concordance of clinical features with previous observations testifies to this. Co-segregation of the variant and the LVNC phenotype in two generations of the family serves as an important additional source of evidence for pathogenicity and causality of our finding.
Our clinical observations support the previously reported genotype-phenotype correlations of the MYH7 variants and attract attention to functional diversity of splicing alterations that now appear to be underestimated. Currently, the clinical assessment of the variants in the positions corresponding to the canonical splice sites seems to be predominantly based on the widespread presumption that splicing interruption always leads to loss of one allele. Since this mechanism is not proven to date to be significant in the MYH7-related disorders, splicing variants tend to be interpreted as VUSes without going deep into clinical features, role of the affected exons, etc. This can be seen from the majority of the existing ClinVar entries which often contain only the postulate of non-pathogenicity of MYH7tv. Moreover, recent study still defines the mechanism of pathogenicity of the alterations in c.732 splice region, adjacent to the in-frame exon of MYH7, as "unclear" [28]. We propose that the splicing variants in MYH7, both known and novel, need to be systematically evaluated in terms of possible consequences of exon skipping and/or other alternative splicing events. This approach could help distinguish the truly non-causative variants from previously ignored ones.

Conclusions
Precise genetic differentiation of mechanisms underlying LVNC and its overlapping conditions is still challenging and, at the same time, important for the clinical management of affected families. Recent results suggest that truncating variants in the MYH7 gene should not be ignored, although their role is still not entirely clear. Continuing this theme, our finding emphasizes the importance of the distinction between truncation and frame-preserving splice site variations. We believe that accurate analysis of possible molecular consequences in the patients with splicing variants in MYH7 will expand the spectrum of known genotype-phenotype correlations and improve our knowledge of LVNC pathogenesis.

Limitations of the Study
The main limitation of the study was that the presented genetic data were based only on DNA-level experiments and in silico analysis. We were not able to validate our predictions with RNA sequencing or to conduct any functional study of alternative splicing