Novel frameshift variant in MYL2 reveals molecular differences between dominant and recessive forms of hypertrophic cardiomyopathy

Abstract

Hypertrophic cardiomyopathy (HCM) is characterized by thickening of the ventricular muscle without dilation and is often associated with dominant pathogenic variants in cardiac sarcomeric protein genes. Here, we report a family with two infants diagnosed with infantile-onset HCM and mitral valve dysplasia that led to death before one year of age. Using exome sequencing, we discovered that one of the affected children had a homozygous frameshift variant in Myosin light chain 2 (MYL2:NM_000432.3:c.431_432delCT: p.Pro144Argfs*57;MYL2-fs), which alters the last 20 amino acids of the protein and is predicted to impact the most C-terminal of the three EF-hand domains in MYL2. The parents are unaffected heterozygous carriers of the variant and the variant is absent in control cohorts from gnomAD. The absence of the phenotype in carriers and the infantile presentation of severe HCM is in contrast to HCM associated with dominant MYL2 variants. Immunohistochemical analysis of the ventricular muscle of the deceased patient with the MYL2-fs variant showed a marked reduction of MYL2 expression compared to an unaffected control. In vitro overexpression studies further indicate that the MYL2-fs variant is actively degraded. In contrast, an HCM-associated missense variant (MYL2:p.Gly162Arg) and three other MYL2 stop-gain variants (p.E22*, p.K62*, p.E97*) that result in loss of the EF domains are stably expressed but show impaired localization. The degradation of the MYL2-fs can be rescued by inhibiting the cell's proteasome function supporting a post-translational effect of the variant. In vivo rescue experiments with a Drosophila MYL2-homolog (Mlc2) knockdown model indicate that neither the MYL2-fs nor the MYL2:p.Gly162Arg variant supports normal cardiac function. The tools that we have generated provide a rapid screening platform for functional assessment of variants of unknown significance in MYL2. Our study supports an autosomal recessive model of inheritance for MYL2 loss-of-function variants in infantile HCM and highlights the variant-specific molecular differences found in MYL2-associated cardiomyopathy.

Conflict of interest statement

The authors have declared that no competing interests exist.

Figures

Fig 1. Identification of a novel homozygous frameshift variant in a proband with infantile hypertrophic cardiomyopathy.

(A) Pedigree of the family with multiple infant deaths due to early-onset cardiac disease. The consanguinity and age of the parents are also shown. The proband (fs/fs) was homozygous for the MYL2-fs allele while the parents were found to be heterozygous (fs/wt) carriers. The status of the MYL2 allele in siblings is unknown (Unk.). Proband is marked by an arrowhead. (B) Table summarizing key clinical findings of the four siblings (IV:1 –IV:4). (C) Dissection of the post-mortem heart from the proband shows severe biatrial dilatation along with hypertrophy of the right ventricle (left). Severe mitral valve dysplasia and hypertrophy of the left ventricle with a small ventricular cavity is also noted (right). RA- right atrium; TV- tricuspid valve; RV- right ventricle; LA- left atrium; MV- mitral valve; LV- left ventricle (D-top) The genomic locus of the MYL2 gene shows the frameshift variant identified in the proband (highlighted in red), which is located in the last exon of the gene. (D-bottom) Sequence chromatograms from Sanger sequencing of control (MYL2 wt), the proband (homozygous fs variant: MYL2:NM_000432.3:c.431_432delCT) and the parents (heterozygous fs).

Fig 2. Analysis of the ventricular myocardium of the proband shows characteristic features of hypertrophic cardiomyopathy (HCM) and molecular consequences of the variant.

(A) Histochemical analysis of the ventricular myocardium from the proband by H&E staining shows myocyte disarray as compared to control. Masson’s trichrome staining of the proband’s ventricular myocardium shows increased fibrosis (stained with blue). (B) Immunohistochemical analysis of MYL2 expression in the proband’s ventricular myocardium shows a remarkable reduction in the expression of MYL2 as compared to control. The antibody used targets residues at the N-terminal region of the protein (S4 Fig). Cardiac troponin I (TNNI3) was detected at comparable levels in the control and the proband myocardium. The unstained areas correspond to fibrosis. (C) Flow diagram showing the steps used to analyze MYL2 mRNA expression in the ventricular myocardium of the proband. An RT-PCR product of the expected size of MYL2 mRNA is detected in the control and proband tissue, and not in the negative (No-RT) controls. Sanger sequencing of the product from the proband shows the MYL2 mRNA transcript with the dinucleotide deletion.

Fig 3. In vitro analysis of stability and localization of MYL2 variants.

(A) Schematic of MYL2 protein showing N-terminal and C-terminal EF-hand domains along with Serine 15 residue that is target of MLCK phosphorylation, three stop-gain variants (red octagons), the frameshift variant identified in the proband (red block showing extension of C-terminal end) and missense variant associated with HCM (orange oval) in the C-terminal domain. The frameshift mutation changes the canonical MYL2 protein sequence (‘Ref.’, black text) from residue 144 onwards, modifying the last 20 amino acids of the protein in addition to adding 36 non-canonical amino acids to the C-terminus (‘fs’, red text). (B) Western blot image showing overexpression of EGFP-tagged MYL2 wildtype (wt) and variant proteins. The expression of the protein harboring the frameshift variant is significantly reduced while the other stop-gain variants do not show a significant change in expression. Analysis of mCherry signal by western blot does not show any significant changes between MYL2 wt and tested variants. Loading controls: mCherry and Actin. (C) In H9c2 cells, immunofluorescence analysis of EGFP tagged MYL2 wt and tested variants (green) that overexpress mCherry from the same transcript (red) is shown. EGFP-tagged MYL2-wt localizes to the cytoskeleton, while stop-gain variants (E22*, K62*, E97*) do not localize to the cytoskeleton and are observed in a diffuse pattern. EGFP signal from the frameshift variant (fs) is significantly reduced. Localization of the missense (G162R) variant is in a pattern similar to the wt. mCherry signal from the transfected cells does not change between variants. (D) Immunofluorescence images of EGFP-tagged MYL2-wt, MYL2-fs, and MYL2-G162R in MG-132 treated H9c2 cells show the rescue of MYL2-fs variant signal following MG-132 treatment (western blot of the rescue shown in S3B Fig). Scale bar indicates 50 μm length. Quantitation of signal from the western and immunofluorescence experiments is provided in S3B Fig.

Fig 4. In vivo functional analysis of MYL2 variants using Drosophila Mlc2 substitution assay in the heart.

(A) Schematic shows Hand-enhancer driven GAL4 (Hand-GAL4) that was used to knockdown Mlc2 expression by RNAi in the heart and to overexpress human MYL2-wt and variant cDNAs for functional substitution. In parallel, the Hand-GAL4 drives the expression of UAS-CD8.mCherry, which facilitates the visualization of the heart. (B-i) Immunofluorescence image of the Drosophila heart from UAS-CD8.mCherry. Hand-GAL4 animals show expression of mCherry in the heart chambers and aorta. (B-ii) Magnified image of the A7-A8 posterior denticle band region showing expression of mCherry in both cardiomyocytes (arrowheads) as well as pericardial cells (arrows). The expression of pericardin is shown (green, B-iii). Actin expression (blue, B-iv) shows neighboring skeletal muscles (actin is stained with phalloidin). (B-v) Merged image of the three channels. No expression of mCherry was detected in the skeletal muscle. (C) Graph showing percent of progeny that emerge as adults compared to expected numbers following Mlc2 RNAi knockdown using Hand-GAL4, and rescue by overexpression of human MYL2-wildtype (wt) and variants. Partial rescue of the lethality is observed with wt MYL2 cDNA overexpression while no significant difference is observed for the tested variants. Significance is analyzed using Brown-Forsythe and Welch ANOVA across all samples followed by multiple comparisons using Games-Howell multiple comparison test. Significance is indicated using * adj. P-Value < 0.05, ** adj. P-Value < 0.01, **** adj. P-Value < 0.0001. (D) Image showing the rhythmic pattern of cardiac contraction that was observed using UAS-CD8-mCherry reporter driven by Hand-GAL4. Diastolic width is shown by ‘D’, and systolic width by ‘S’. The formula used to calculate fractional shortening is shown. (E) Interleaved scatter plot showing fractional shortening across different genotypes (S1–S5 Movies). Fractional shortening was significantly reduced due to Mlc2 RNAi knockdown which is partially rescued by overexpression of wt human MYL2. The missense and fs variants showed no significant improvement in fractional shortening compared to the wt. Significance was calculated using two-way ANOVA followed by Tukey’s multiple comparison tests. Multiplicity adjusted P-Values of Tukey’s comparison are indicated as * adj. P-Value < 0.05, ** adj. P-Value < 0.01, *** adj. P-Value < 0.001, **** adj.P-Value < 0.0001.