Comparing the Efficacy of Myosin Inhibition Versus Thin Filament Calcium Desensitization for Treatment of Pediatric Restrictive Cardiomyopathy Using a Patient-derived hiPSC Model

Restrictive Cardiomyopathy (RCM) is a rare form of cardiomyopathy characterized by profound, isolated diastolic dysfunction, normal or near-normal ventricular dimensions and wall thickness, and preserved systolic function^1,2. RCM patients have notably poor outcomes, with studies showing increased morbidity and mortality compared to other forms of cardiomyopathy². Targeting the sarcomere with myosin inhibitors such as mavacamten improves outcomes in hypertrophic cardiomyopathy (HCM)³. However, whether this strategy, or other sarcomere-targeted strategies, could be efficacious in RCM is unknown.

Original data and detailed methods are available upon request. We identified a pediatric patient with RCM who carried a pathogenic R94C variant of TNNT2. After obtaining written informed consent from their parents and with approval from the Stanford School of Medicine Institutional Review Board (IRB), we derived human induced pluripotent stem cells (hiPSCs) from patient PBMCs and created an isogenic control line in which the mutation was corrected by CRISPR/Cas9 gene editing, then differentiated these lines into cardiomyocytes (CMs). We assessed systolic and diastolic function using kinetic imaging of spontaneously beating cardiomyocytes by plating cells in a monolayer on polyacrylamide gels containing fluorescent beads and measuring forces using traction force microscopy (Panel A). Contractile kinetics were normalized to the beat rate to remove confounding effects of different rates across cell lines and differentiation batches. Total contraction times and relaxation times were prolonged relative to the contractile period, demonstrating a kinetic defect in relaxation (Panel B,C). The contraction rise time was also prolonged (Panel D). Measurements of contractile amplitude revealed that patient cells generated greater forces during contraction (Panel E). To determine whether these cells generated more diastolic tension, cells were removed from the gels and bead positions compared to those in diastole (Panel F). Converting the bead displacements to applied force revealed significantly increased diastolic tension in TNNT2 R94C cells compared to corrected controls (Panel G).

We next examined the molecular processes involved in these diastolic defects. To assess if calcium sensitivity was altered in the TNNT2 R94C cells, we combined measurements of cardiomyocyte calcium handling using a fluorescent calcium-sensitive dye with our contractility measurements using high-speed interleaved imaging (Panel H-J). TNNT2 R94C cells exhibited prolonged total calcium duration relative to beat rate, in keeping with the prolonged contractile kinetics (Panel K). The near-simultaneous imaging of calcium flux and force generation allowed us not only to measure calcium kinetics but also to construct the calcium-contractility relationship by plotting contractility versus the calcium fluorescence throughout the contractile cycle. Due to the rapid kinetics of calcium inflow during systole, the myofilament response is delayed compared to the cytosolic calcium influx, but the sarcomere reaches a dynamic equilibrium with cytoplasmic calcium during diastole^4,5. Thus, as in previous studies^4,5, analyses of calcium binding were performed during diastole, when the myofilament has equilibrated with the cytosolic calcium measured by the calcium dye. By extracting and averaging the diastolic portion of these curves, we observed increased absolute force generation at all cellular calcium levels (Panel L). Furthermore, by normalizing these traces to average maximal force, we determined that the calcium fluorescence required to achieve 50% of maximum force was decreased in patient cells (Panel M,N). Taken together, these findings indicate an increased calcium sensitivity in TNNT2 R94C cardiomyocytes.

Lastly, we evaluated possible therapeutic options to treat the observed diastolic pathology of TNNT2 R94C cardiomyocytes. Specifically, we examined therapeutic effects on diastolic dysfunction and potentially detrimental effects on systolic function of two logical approaches: decreasing sarcomeric contractility with the clinically approved drug mavacamten versus decreasing sarcomeric calcium sensitivity with the compound W7. Treatment with both mavacamten and W7 improved diastolic dysfunction in TNNT2 R94C cardiomyocytes (Panel O,P). However, mavacamten caused a severe drop in systolic function (Panel Q,R) that increased in proportion to the effect on diastolic function compared to W7 (Panel S,T).

In this study, we combined a new cellular model of pediatric RCM with a novel high-content measurement system for the determination of calcium sensitivity, mechanical contraction, and diastolic tension. We used our RCM model to compare two potential therapeutic strategies targeted at the sarcomere, namely myosin inhibition and thin filament calcium desensitization. Our data suggest that both strategies can improve diastolic function, but that myosin inhibition may have a greater relative effect on systolic function than calcium desensitization, at least in this genetic background. This difference may be due to their respective molecular mechanisms. Myosin inhibitors, especially at higher doses, could decrease the maximum possible force production at all calcium levels due to a decreased number of active myosin motors available to bind to actin. In contrast, calcium desensitizers will shift the calcium response curve without directly altering myosin availability, thus maintaining the maximum force generated at high, systolic calcium levels. Our data suggest that therapeutically targeting thin filament calcium sensitivity rather than myosin activity has the potential to decrease unwanted effects on systolic function, at least in RCM with thin filament mutations. In addition, the study provides a methodological paradigm for the investigation of approaches to treat diastolic dysfunction.

Figure:

(A) hiPSC-CMs were seeded in multi-well plates containing 8 kPa polyacrylamide gels with embedded fluorescent beads. Bead position was tracked using video microscopy and converted to applied force via Traction Force Microscopy. Normalized (B) Total Contraction Duration (C) Contraction Relaxation Time and (D) Contraction Rise Time were increased in the RCM patient cells compared to their Isogenic Control. (E) Systolic force was increased compared to that of the Isogenic Control. (F) Cells were removed from the gel, the gels were re-imaged, and bead position was compared to that in diastole to calculate Diastolic Tension. (G) Diastolic Tension was increased in RCM cells. (H) hiPSC-CMs on gels with fluorescent beads were stained with the calcium sensing dye Cal520. High speed (66 fps total, 33 fps/channel) videos were recorded with every other frame measuring bead movement or calcium probe intensity. (I) Force and calcium curves were extracted from the raw movies, and (J) force/calcium relationships were plotted to obtain calcium sensitivity. (K) Calcium duration is increased in RCM cells. (L) Diastolic Force-calcium plots demonstrated that RCM cells generated more force at any given calcium level and (M) RCM cells reached half-max contractility at lower calcium levels, indicating increased calcium sensitivity. (N) Quantification of (M) showing increased calcium sensitivity of RCM cells as measured by normalized fluorescence at 50% max force. (O) Mavacamten and (P) W7 both decrease diastolic tension in a dose-dependent manner. (Q) Mavacamten severely decreases systolic function while (R) W7 has a lesser effect. (S) The ratio of the effect on systole to the effect on diastole is greater in Mavacamten compared to (T) W7. Colors indicate separate batches of cells. Data are presented as median and interquartile range. Significance was calculated using linear mixed models treating differentiation start as a random effect to account for batch-to-batch variation. TNNT2 R94C and control differentiations in panels B-N were differentiated in parallel and analyzed as pairs matched by differentiation start date. A Holm-Bonferroni correction was performed to correct for multiple testing, taking into account the multiple functional variables and, in panels O-T, the multiple drug treatments, examined. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001

Nonstandard Abbreviations and Acronyms

RCM

Restrictive Cardiomyopathy

HCM

Hypertrophic Cardiomyopathy

hiPSC

Human induced pluripotent stem cell

CM

Cardiomyocyte

PBMC

Peripheral Blood Mononuclear Cells

IRB

Institutional Review Board