Novel Adult-Onset Systolic Cardiomyopathy Due to MYH7 E848G Mutation in Patient-Derived Induced Pluripotent Stem Cells

Visual Abstract


METHODS
Details of the materials and methods are available in the Supplemental Material. Briefly, all patient recruitments followed the protocol approved by the institutional review board. Dermal fibroblasts from FCM and normal healthy patients were episomally reprogrammed to iPSCs and differentiated into cardiomyocytes with our well-established monolayer differentiation protocol (7). Patient-specific iPSC-CMs were replated as single cells or in a three-dimensional co-culture to form engineered heart tissues (EHTs) to assess contractile function. The Clustered Regularly Interspaced Short Palindromic Repeats/CRISPR-Associated Protein 9 system was used to knock out MYH7 (MYH7 KO) to create an isogenic iPSC line.
Yeast two-hybrid analysis was used to test the protein-protein interaction between MYH7 S2 and cMyBP-C C1C2 domains.   (MYH7, MYBPC3, TNNT2,   TNNI3, ACTC, MYL2, MYL3, LAMP2, and PRKAG2) identified a novel heterozygous missense mutation that caused a glutamic acid-to-glycine mutation at amino acid position 848 (E848G) on MYH7 in the proband (Patient Ia). The offspring were subsequently specifically tested for the E848G allele only.
This mutation occurs at a residue conserved from mouse to humans (Supplemental Table 3 Table 1). Systolic wall thickening is a regional measure of systolic function (8,9) and has been shown to be sensitive for the detection of regional contractile abnormalities despite globally preserved left ventricular ejection fraction (8).  (10). We reasoned that aligning the myofibrils of our iPSC-CMs would minimize noise in contractile measurements due to myofibril disarray.
We plated patient-derived iPSC-CMs onto nano-  Table 4). There was no difference in cell size between groups (Supplemental Figure 2).
Overall, a nanopatterned surface structurally matured the iPSC-CMs and transformed their contractile cytoskeletons into a more adult cardiomyocyte-like phenotype.   Posts were designed to be 5 mm apart to form loops on either end of the EHT ( Figure 3B). Constructs were conditioned with static external stress for 2.5 to 3 weeks to increase cellular alignment and hypertrophy. Interestingly, EHTs with WT hiPSC-CMs exhibited significantly greater cardiomyocyte alignment than did those from FCM hiPSC-CMs (p < 0.05) To study isometric contractile function of the patient-specific EHTs, the loops in the EHTs were mounted between a force transducer and a rod whose position was precisely controlled by a length      Figure 4B).
Twitch kinetics and amplitude comparisons are shown in Figures 4C and 4D. Strikingly, the maximal active twitch force per area of the FCM IIb EHT was reduced by >75% compared with WT Ib EHT (p < 0.001) ( Figure 4E). When normalized to total force per twitch, both the time from electrical stimulation to maximal force production and to 50% relaxation were shorter in the FCM IIb EHTs compared with the WT Ib EHTs (both, p < 0.05) (Figures 4F and 4G).
There was no difference in time from 50% relaxation to time of 90% relaxation ( Figure 4H). Systolic Cardiomyopathy due to MYH7 E848G As expected for a negative control, no growth was seen with unfused BD and AD peptides ( Figure 6A), whereas the positive control ( Figure 6B) and S2 WT -AD and S2 WT -BD resulted in robust growth, confirming S2 homodimerization ( Figure 6C). Interestingly, the S2 E848G -AD and S2 WT -BD condition did not have any growth, indicating that the E848G mutation weakens the dimerization of myosin S2 ( Figure 6D). S2 WT -AD and C1C2 WT -BD resulted in growth, confirming the interactions between MYH7 and cMyBP-C ( Figure 6E). The S2 WT -AD and C1C2 E258K -BD combination did not have any growth, confirming that this mutation in cMyBP-C prevents its binding to the WT S2 domain ( Figure 6F) (12). The key finding of this yeast-two hybrid experiment, however, was the lack of growth with the S2 E848G -AD and C1C2 WT -BD fusion proteins ( Figure 6G). As expected, the doublemutation experiment involving S2 E848G -AD and C1C2 E258K -BD did not result in any growth ( Figure 6H). Taken together, these experiments support the hypothesis that the E848G mutation results in disruption of the myosin S2 and cMyBP-C C1C2 protein-protein interaction. Although WT and mutant hiPSC-CMs replated on nanopatterned surfaces had very similar myofibril alignment, it was interesting that there was significantly less alignment in FCM EHTs compared with WT EHTs. This lack of alignment correlates with the myofiber disarray noted in the pathology report of patient Ic, who had the FCM phenotype but whose genotype is unknown. The pathogenesis of myofiber disarray is unknown, but it is interesting to speculate that it could relate to the impaired homodimerization of the S2 domain of MYH7 we observed by using the yeast two-hybrid analysis. Another possibility is that the alignment in engineered tissues may depend on the overall stress that is applied to the construct.
Because the WT Ib EHTs generate more active force, this action may be providing a stronger alignment stimulus parallel to the long-axis of EHT. It is notable that hiPSC-CMs seeded on the nanopatterned surfaces exhibited no genotype-dependent differences in myofibril alignment, which suggests that the 2-dimensional system generates stronger external alignment stimuli than in the 3-dimensional tissue engineering.
A contractile impairment was observed in day 50 FCM-CMs that was not present in the day 30 FCM-CMs. We suspect this finding is in part likely due to the increased expression of MYH7 that occurs in older day 50 hiPSC-CMs compared with the younger day 30 FCM-CMs (review elsewhere [21]) ( Figure 5D).  (14). Given that both MYH7 E848G and cMyBP-C E258K mutations result in decreased force production, this finding suggests that the MYH7 S2 and cMyBP-C C1C2 protein-protein interaction is necessary for the preservation of systolic function; further studies are needed to confirm this hypothesis, however.
Yang et al.