Multiplexed Functional Assessments of MYH7 Variants in Human Cardiomyocytes

Abstract

Background: Pathogenic autosomal-dominant missense variants in MYH7 (myosin heavy chain 7), which encodes the sarcomeric protein (β-MHC [beta myosin heavy chain]) expressed in cardiac and skeletal myocytes, are a leading cause of hypertrophic cardiomyopathy and are clinically actionable. However, ≈75% of MYH7 missense variants are of unknown significance. While human-induced pluripotent stem cells (hiPSCs) can be differentiated into cardiomyocytes to enable the interrogation of MYH7 variant effect in a disease-relevant context, deep mutational scanning has not been executed using diploid hiPSC derivates due to low hiPSC gene-editing efficiency. Moreover, multiplexable phenotypes enabling deep mutational scanning of MYH7 variant hiPSC-derived cardiomyocytes are unknown.

Methods: To overcome these obstacles, we used CRISPRa On-Target Editing Retrieval enrichment to generate an hiPSC library containing 113 MYH7 codon variants suitable for deep mutational scanning. We first established that β-MHC protein loss occurs in a hypertrophic cardiomyopathy human heart with a pathogenic MYH7 variant. We then differentiated the MYH7 missense variant hiPSC library to cardiomyocytes for multiplexed assessment of β-MHC variant abundance by massively parallel sequencing and hiPSC-derived cardiomyocyte survival.

Results: Both the multiplexed assessment of β-MHC abundance and hiPSC-derived cardiomyocyte survival accurately segregated all known pathogenic variants from synonymous variants. Functional data were generated for 4 variants of unknown significance and 58 additional MYH7 missense variants not yet detected in patients.

Conclusions: This study leveraged hiPSC differentiation into disease-relevant cardiomyocytes to enable multiplexed assessments of MYH7 missense variants for the first time. Phenotyping strategies used here enable the application of deep mutational scanning to clinically actionable genes, which should reduce the burden of variants of unknown significance on patients and clinicians.

Keywords: cardiomyopathy, hypertrophic; deep mutational scanning; induced pluripotent stem cells; myosin heavy chain 7.

Figure 1.

Sarcomeric protein abundance in human hearts and in isogenic hiPSC-CMs. A Experimental strategies to quantify β-MHC protein abundance in human hearts and isogenic human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs). B Representative Western blots for β-MHC (220 kDa) and cardiac troponin I (cTnI; 24 kDa) protein expression in normal (n = 1 heart sample) and HCM (MYH7^E848G/+, n = 6 samples from the same patient) human heart samples. Quantifications of β-MHC (C) and cTnI (D) protein expression from B normalized to normal. E Representative flow cytometry histogram overlaying eGFP intensities of age-matched (T42) MYH7^WT-eGFP/+ (red) and MYH7^E848G-eGFP/+ (blue) isogenic cardiomyocytes. WT = wildtype. T = days after the activation of differentiation. F Quantifications of eGFP geometric means (GM) from E normalized to MYH7^WT-eGFP/+ (n = 3 replicates). Error bars indicate standard error of the mean. P values calculated using unpaired t test; p < 0.05 considered significant. Dashed red line in C, D, and F indicates normalized mean. See Supplemental Figure 1 for complete and stain-free Western blots.

Figure 2.

β-MHC abundance in a library of heterozygous MYH7 missense variant hiPSC-CMs. A Live flow cytometry time course analysis of Variant-β-MHC-mTagBFP2 (x-axis) and WT-β-MHC-eGFP (y-axis) expression from a single, cardiac-directed differentiation of biallelically-edited hiPSCs containing a library of heterozygous MYH7 variant hiPSC-CMs (MYH7^{WT-eGFP/Variant-mTagBFP2}; generated in Friedman et al.), left, or age-matched, WT hiPSC-CMs (MYH7^{WT-eGFP/WT-mTagBFP2}), right. T = days after the activation of differentiation. Low = mTagBFP2^low/eGFP⁺ cells. B Experimental strategy to quantify β-MHC-mTagBFP2 abundance in MYH7^{WT-eGFP/Variant-mTagBFP2} hiPSC-CMs using variant abundance by massively parallel sequencing (VAMP-seq). Low, Mid, and High = sorting bins based on mTagBFP2 intensity. NGS = next generation sequencing. C Mean normalized β-MHC-mTagBFP2 abundance scores for synonymous (Syn; white), pathogenic/likely pathogenic (P/LP; blue), and all variants (black) (n = 2 replicates). Variants of unknown significance (VUS) scores in All = green dots. Upper dotted line = lowest normal abundance score; lower dotted line = highest reduced abundance score; scores between normal and reduced are uncertain. Note broken y-axis. D Position map displaying categorical β-MHC-mTagBFP2 abundance scores for each mutagenized residue (vertical) relative to amino acid substitutions (horizontal). Grey = no data. Black dots = WT residues. See Supplemental Figure 2 for pooled analyses of MYH7 variant hiPSC-CM library mTagBFP2 intensity and FSCA. See Supplemental Figure 3 for VAMP-seq replicate data.

Figure 3.

β-MHC abundance in clonal heterozygous MYH7 variant hiPSC-CMs. A Representative flow cytometry plots of Variant-β-MHC-mTagBFP2 (x-axis) and WT-β-MHC-eGFP (y-axis) expression in biallelically-edited T47 hiPSC-CMs containing a single heterozygous MYH7 variant (left to right; MYH7^{WT-eGFP/WT-mTagBFP2} (WT), MYH7^{WT-eGFP/S866F-mTagBFP2}, and MYH7^{WT-eGFP/S866W-mTagBFP2}). B Quantifications of mTagBFP2 geometric means (GM) from A normalized to WT (n = 4 replicates). C Simple linear regression comparing mean normalized β-MHC abundance scores from VAMP-seq (x-axis; Fig. 2C) to mean normalized mTagBFP2 GMs from corresponding clonal heterozygous MYH7 variant hiPSC-CMs (y-axis). Solid line = best-fit line. Dotted lines = 95% confidence intervals. D Quantifications of eGFP GMs from A normalized to WT (n = 4 replicates). E Quantifications of FSCA GM in mTagBFP2^low versus mTagBFP2^high hiPSC-CMs, normalized to the corresponding measure in WT hiPSC-CMs (n = 2–3 replicates). Dashed red lines in B and D indicate normalized WT means and red dots correspond to plots shown in A. Error bars indicate standard error of the mean. P values calculated using repeated measure one-way ANOVA (B and D) or one-way ANOVA (E); p < 0.05 considered significant; p > 0.05 considered not significant (NS). See Supplemental Figure 4 for confocal microscopy of clonal heterozygous MYH7 variant hiPSC-CMs.

Figure 4.

Survival in a library of heterozygous MYH7 missense variant hiPSC-CMs. A Variant frequency of MYH7^{WT-eGFP/Variant-mTagBFP2} hiPSC-CMs with single synonymous (Syn; white) or pathogenic/likely pathogenic (P/LP; blue) MYH7 variants assessed at T47 (n = 2 replicates) relative to starting frequencies as hiPSCs (n = 1 replicate) (hiPSC dataset from Friedman et al.). T = days after the activation of differentiation. B Mean normalized T47 survival scores for synonymous (Syn; white), pathogenic/likely pathogenic (P/LP; blue), and all variants (black) (n = 2 replicates). Variants of unknown significance (VUS) scores in all column = green dots. Upper dotted line = lowest ‘Normal’ survival score; lower dotted line = highest ‘Reduced’ survival score; scores between normal and reduced are uncertain. C Position map displaying T47 normalized categorical survival scores (relative to hiPSCs) for each mutagenized residue (vertical) relative to amino acid substitutions (horizontal). Grey = no data. Red box = present in T25 hiPSC-CMs and hiPSCs but ‘Lost’ at T47. Black dots = WT residues. Dashed red line in A indicates normalized starting variant frequencies in hiPSCs. See Supplemental Figure 5 for T25 and T47 MYH7 variant hiPSC-CM survival data and T47 replicate data.

Figure 5.

Comparison of β-MHC abundance to survival across a library of heterozygous MYH7 missense variant hiPSC-CMs. Plot displaying mean normalized β-MHC-mTagBFP2 abundance scores (x-axis) relative to mean normalized hiPSC-CM survival scores (T47; y-axis) for synonymous (Syn; white), pathogenic/likely pathogenic (P/LP; blue), variants of unknown significance (VUS; green), and unlabeled variants (black) (n = 2 replicates). The white domain indicates normal variant effect in both assays, the orange domain indicates abnormal variant effects in both assays, the yellow domains indicate conflicting effects between assays, and the grey domain indicates uncertain effects in at least one assay.