Interplay of Genotype and Substrate Stiffness in Driving the Hypertrophic Cardiomyopathy Phenotype in iPSC-Micro-Heart Muscle Arrays

Abstract

Introduction

In clinical and animal studies, Hypertrophic Cardiomyopathy (HCM) shares many similarities with non-inherited cardiac hypertrophy induced by pressure overload (hypertension). This suggests a potential role for mechanical stress in priming tissues with mutation-induced changes in the sarcomere to develop phenotypes associated with HCM, including hypercontractility and aberrant calcium handling. Here, we tested the hypothesis that heterozygous loss of function of Myosin Binding Protein C (MYBCP3^+/−, mutations in which account for almost 50% of inherited HCM) combines with environmental stiffness to drive HCM phenotypes.

Methods

We differentiated isogenic control (WTC) and MYBPC3^+/− iPSC into cardiomyocytes using small molecule manipulation of Wnt signaling, and then purified them using lactate media. The purified cardiomyocytes were seeded into “dog bone” shaped stencil molds to form micro-heart muscle arrays (μHM). To mimic changes in myocardial stiffness stemming from pressure overload, we varied the rigidity of the substrates μHM contract against. Stiffness levels ranged from those corresponding to fetal (5 kPa), healthy (15 kPa), pre-fibrotic (30 kPa) to fibrotic (65 kPa) myocardium. Substrates were embedded with a thin layer of fluorescent beads to track contractile force, and parent iPSC were engineered to express the genetic calcium indicator, GCaMP6f. High speed video microscopy and image analysis were used to quantify calcium handling and contractility of μHM.

Results

Substrate rigidity triggered physiological adaptation for both genotypes. However, MYBPC3^+/− μHM showed a lower tolerance to substrate stiffness with the peak traction on 15 kPa, while WTC μHM had peak traction on 30 kPa. MYBPC3^+/− μHM exhibited hypercontractility, which was exaggerated by substrate rigidity. MYBPC3^+/− μHM hypercontractility was associated with longer rise times for calcium uptake and force development, along with higher overall Ca²⁺ intake.

Conclusion

We found MYBPC3^+/− mutations cause iPSC-μHM to exhibit hypercontractility, and also a lower tolerance for mechanical stiffness. Understanding how genetics work in combination with mechanical stiffness to trigger and/or exacerbate pathophysiology may lead to more effective therapies for HCM.

Supplementary Information

The online version contains supplementary material available at (10.1007/s12195-021-00684-x).

Introduction

Hypertrophic cardiomyopathy (HCM) is the most common inherited heart disease affecting about 1 in 500 individuals,35,40,41 and is the most frequent cause of sudden cardiac death in the young (especially athletes).42,57 Other clinical manifestations include heart failure, stroke, and atrial fibrillation.41,47 Tremendous progress in genetic studies of inherited heart disease has linked HCM to mutations in the contractile apparatus (sarcomere) of heart muscle cells (cardiomyocytes). The most frequent HCM-associated mutations are found in Myosin Binding Protein C (MYBPC3), which contributes to sarcomere structural properties and contractile function, accounting for almost 50% of HCM cases.5,22,44

While genetic studies have been crucial to deciphering the specific mutations that correlate with HCM, it is also well-established that the penetrance of HCM disease phenotypes in genotype-positive individuals is highly variable.35,41 This suggests that non-genetic, environmental factors may act as triggers for this disease.17,28 Cardiac sarcomeres play a critical role in allowing these cells to adapt to systemic mechanical stress on the heart induced by blood pressure, exercise, and adrenergic tone.2 Importantly, in the general population (without HCM-associated genotypes), excessive blood pressure induced hypertrophic remodeling in the heart is similar to what is observed in symptomatic HCM patients. This suggests that mechanical stress on the heart in general, and mechanical overload caused by blood pressure and myocardial stiffness in particular, could act as triggers for development of HCM disease phenotypes in individuals with HCM mutations. Consistent with a potential role for mechanical overload in HCM disease pathogenesis, clinical studies show increased mortality for HCM patients with hypertension.37,48

To study potential synergy between mechanical loading and HCM mutations in a controlled experimental setting, mouse models of HCM, including HCM linked to MYBPC3 mutations, have been developed.52,59In vivo, mechanical overload is induced through surgical procedures, most commonly transverse aortic constriction (TAC), which artificially increases blood pressure.1 In one representative study, Barefield et al. studied cardiac structural and functional properties between wild type and HCM mutant mice that harbored HCM-associated MYBPC3 heterozygous knockout mutations. Importantly, at baseline there was little structural or functional difference observed, but only when TAC was applied (higher pressure overload), increased hypertrophy and reduced ejection fraction were observed in MYBPC3^+/− mice.1

While animal models are crucial for studying disease mechanisms and testing the efficacy of new therapeutics, the difficulty in precisely controlling overload in vivo, together with differences in physiology between humans and mice,46 highlights the need for developing in vitro physiological models using human cardiomyocytes. Primary human cardiomyocytes can only be obtained using invasive procedures, making it infeasible to obtain these cells from HCM genotype positive, disease negative individuals. This makes human-induced pluripotent stem cell (iPSC) derived cardiomyocytes (iPSC-CM) the optimal cell source for in vitro modeling.65 Previous iPSC based HCM models have identified a subset of HCM phenotypes in vitro, typically using 2D monolayer or single-cell formats.4,10,30,45 In these studies, different degrees of HCM-associated Ca²⁺ handling and contractility abnormalities have been reported, with sometimes conflicting results in whether HCM mutations cause hypercontractility or contractile deficits, especially when MYBPC3 mutation was involved.4,7,39,50 For example, contractile defects were reported in single iPSC-CM harboring loss-of-function mutations in MYBPC3, or shRNA knockdown of this protein.4,50 In contrast, studies by Cohn et al. using 3D-engineered heart tissues (EHT) suggested that MYBPC3 loss-of-function mutations led to higher twitch force compared to isogenic control tissues.7

In addition to differences in results between previous 2D and 3D studies, few existing models have investigated the synergistic effect of mechanical stress and genetics in triggering the HCM phenotypes. We previously found that MYBPC3^−/− cardiac micro-tissues formed on fibrous scaffolds had similar contractility compared to isogenic controls when working against compliant fibers. In contrast, we observed a contractile deficiency in MYBPC3^−/− micro-tissues when the fiber bending stiffness increased.39 Compared to 2D systems, 3D EHT models can mimic native tissue structure while modulating mechanical rigidities. However, limited throughput is hindering the data collection using 3D models.

Here, we leveraged our recently developed medium-throughput iPSC-micro heart muscle (iPSC-μHM) model to study the combined effects of substrate rigidity and MYBPC3 loss of function mutation in triggering contractility and calcium-handling phenotypes associated with HCM.18,24 We grafted iPSC-μHM on elastomeric substrates with defined stiffness to simulate the mechanical rigidity changes in the myocardium caused by changes in blood pressure during development and with hypertension.

To model HCM, we used iPSC-CM with a physiologically relevant, MYBPC3 heterozygous knockout mutation, as the majority of HCM-associated MYBPC3 mutations are loss-of-function mutations.21 We observed hypercontractility of MYBPC3^+/− μHM. The high traction force generation was associated with longer traction rise time and Ca²⁺ upstroke duration, regardless of matrix elasticity. Mechanical stiffness exaggerated the contractility in both genotypes. However, MYBPC3^+/− μHM exhibited a lower tolerance to substrate rigidity, with peak traction generated at relatively lower stiffness 15 kPa, whereas isogenic control μHM exhibited a more robust adaptation at stiffer 30 kPa condition. Understanding the mechanism through which mechanical stress contributes to HCM pathogenesis can help unveil disease mechanisms leading to new therapeutics.

Methods

iPSC Culture

The WTC iPSC encoded with GCaMP6f was reprogrammed from a healthy 30-year-old male of Japanese descent with a normal electrocardiogram and no family history of heart disease (Coriell Repository GM25256).25 MYBPC3^+/− iPSC harboring the GCaMP6f reporter were generously provided by Bruce Conklin and Mohammed Mandegar (Gladstone Institute, of Cardiovascular Disease, San Francisco, CA; Fig. 1a). MYBPC3 knockout mutations were previously made using genome editing to insert stop codons and a Puromycin selection cassette into first exon of MYBPC3.39,50

Figure 1.

iPSC-CM differentiation and characterization. a Illustration and timeline of generating WTC and MYBPC3^+/− iPSC-CM. b Representative sarcomeric α-actinin images of purified WTC (left) and MYBPC3^+/− (right) iPSC-CM. c Characterization of MYBPC3 gene expression. d Representative flow cytometry quantification of purified WTC (left) and MYBPC3^+/− (right) iPSC-CM. e CM purity of WTC and MYBPC3^+/− iPSC-CM. Scale bar: 25 μm b. Error bar: standard deviation (SD), n = 4–7 differentiation batches. ** indicates p < 0.01.

Both WTC and MYBPC3^+/− iPSC were differentiated into iPSC-CM and purified as described previously using small molecule control over the Wnt signaling pathway.18,33 Briefly, iPSC were first cultured in Essential 8 Media (Thermo Scientific; Waltham, MA). Media on confluent iPSC monolayers (differentiation day 0) was changed to RPMI with B27 supplement without insulin (RPMI-i; Thermo Scientific), which was supplemented with 6 μM CHIR99021 (Wnt agonist; Peprotech; Rocky Hill, NJ) and 150 μg/mL L-ascorbic acid (Sigma; St. Louis, MO). After 48 h (differentiation day 2), the cells were treated with RPMI-i containing 5 μM IWP-2 (Wnt antagonist; Peprotech) and 150 μg/mL L-ascorbic acid for 48 h (days 2–4). Media was changed to RPMI-i and 150 μg/mL L-ascorbic acid on day 4 and then RPMI with B27 containing insulin (RPMI-c) on day 6, with media changed every 2–3 days thereafter (Fig. 1a).

Spontaneously beating iPSC-CM were observed by differentiation day 10. At day 15 of differentiation, beating iPSC-CM were dissociated using 0.25% trypsin EDTA and replated at 250,000 cells/cm² for purification using glucose depleted, lactate enriched media.60 After lactate purification, the cardiomyocytes were changed to RPMI-c for at least 2–3 days for recovery before making iPSC-μHM. The iPSC-CM were used to form μHM at day 28–35 of differentiation.

The purity of cardiomyocytes was quantified by flow cytometry for cardiac troponin T (TNNT2). At least 1 million cells were collected and fixed as a cell pellet using 4% paraformaldehyde. Mouse anti cardiac troponin T (cTnT, antibody clone 13-11, Thermo Scientific) and goat anti mouse Alexa Fluor 488 were used to label cardiomyocytes. ACEA NovoCyte flow cytometer (San Diego, CA) was used to sort for cTnT positive cells.18 Qdot 800A signal was used as a reference signal (Fig. 1d).

Molecular Analysis

Expression levels of MYBPC3 in iPSC-CM were assessed with quantitative RT-PCR. Purified cardiomyocytes were dissociated. For each sample, a pellet of 1 million cells wase collected and washed with dPBS before performing cell lysis and column-based RNA recovery using the RNAqueous®-Micro Kit (Thermo Scientific). Following RNA recovery and DNAase treatment with reagents provided with the RNAqueous® Micro-Kit, RNA concentration was measured using Nanodrop (Nanodrop One, Thermo Scientific). At least 320 ng RNA per sample was used to synthesize cDNA using a Super Script III Reverse Transcriptase Kit (Thermo Scientific), cDNA was then probed for expression levels of Sarcomeric Alpha-Actinin (ACTN2) and MYBPC3 using SYBR green compatible primers (Sigma) with sequence show in below, qPCR was performed on an Applied Biosystems Step One Plus Light Cycler instrument using SYBR green master mix (Thermo Scientific).

Primer name	Forward sequence	Reverse sequence
ACTN2	GCTTCTACCACGCTTTTGCG	CATTCCAAAAGCTCACTCGCT
MYBPC3	GGCATGCTAAAGAGGCTCAA	TCTTGTGGCCTTTGCTCAC

Substrate Preparation

μHM substrates were fabricated following our previous developed method.18 In summary, two PDMS grades: Sylgard 184 (stiff) and Sylgard 527 (soft) were mixed at specific ratios to control substrate mechanical properties.18,49 A PDMS base layer (~ 200 μm thick) and a PDMS/fluorescent bead (1 μm diameter green fluorescence beads, Polysciences, Inc.; Warrington, PA) top thin layer (~ 5 μm thick) were spin coated and cured at 60 °C sequentially onto 22 mm × 22 mm coverslips.18,67 Substrates with 5, 15, 30 and 65 kPa elasticity were fabricated for this study. TFM substrates were then attached to the 6 wall plates using very small amount of Sylgard 184 as a “glue” on the edges of the coverslips.

Robust mechanical coupling between micro-tissues and substrates was achieved by covalently grafting ECM protein fibronectin to the substrates using amino silane chemistry and glutaraldehyde (GA) crosslinker.18 Upon plasma oxidation, substrates were reacted with (3-aminopropyl)triethoxysilane silane (APTES; 5% v/v in methanol; Sigma) for 1 h followed by GA reaction. We observed that grafting GA to the substrates immediately following a APTES grafting resulted in strong autofluorescence of the substrates (Fig. 2b). This reduces the dynamic range for measuring μHM GCaMP6f signals. To reduce substrate autofluorescence, following the APTES modification and the three subsequent washes in methanol, the substrates were next incubated in Dulbecco’s phosphate buffered saline (dPBS) for 24 h at room temperature before GA reaction.

Figure 2.

iPSC-μHM device fabrication and optimization. a Schematic of generating iPSC-μHM and physiological data collection timeline. b Characterization of autofluorescence from surface chemistry. GA reacting immediately after APTES modification resulted in significant higher autofluorescence (20 μg/mL FN). c Fibronectin binding affinity between different surfaces. The hFN 7.1 fluorescent intensity increased with increased hFN concentration, no significant difference was found between two surfaces (p = 0.259). Scale bar: 1000 μm a. Error bar: SD, * indicates p < 0.05, n = 3.

Following APTES grafting and methanol washes (with or without a subsequent 24 h dPBS incubation), substrates were reacted with 2.5% GA in dPBS at room temperature for 2 h, followed by 3 dPBS washes. Bovine FN (15 μg/mL, Sigma) was then applied to the surface for at least 30 min on the shaker followed by 3 dPBS washes and then 3 DI water washes. The substrates were then reacted with 2.5% w/v ethanolamine hydrochloride (Alfa Aesar; Haverhill, MA) in dPBS to quench unreacted GA followed by 3 dPBS washes.18

To ensure the 24 h dPBS incubation did not alter the bioactivity of substrate-grafted fibronectin, we probed substrates coupled with human fibronectin using antibody clone hFN7.1 (Developmental Studies Hybridoma Bank; Iowa City, IA). The hFN 7.1 antibody has previously been shown to predict the bio-availability of integrin-binding epitopes within substrate-adsorbed FN.14 hFN7.1 binding assay was performed on substrates that were human FN (Sigma)-grafted following either GA added immediately or 24 h upon APTES reaction. Following primary antibody incubation (1-hour at room temperature), substrates were thoroughly washed and probed with 100 ng/mL Li-COR IRDye 800cw secondary antibody (LI-COR Biosciences; Lincoln, NE).18 The stained surfaces were then scanned with Li-COR Odyssey CLx scanner (LI-COR Biosciences). Mean intensity values were measured and normalized to substrates modified with the original procedure (GA added immediately after APTES grafting).

Micro tissues were formed by seeding iPSC-CM into “dog bone” shaped PDMS stencils molds applied to substrates18,24 (Fig. 2a). To avoid tissue attachment to stencils, stencil molds were first soaked in 10% F68 Pluronic at room temperature for 45 min followed by 3 dPBS washes.18 The stencil molds were then soaked into methanol, before gently pressed onto the FN grafted substrates. The assembly were then incubated at 60 °C for 15 min for methanol evaporation and stencil attachment to the substrates. The assembled devices were disinfected with 70% sterile ethanol for at least 3 h (preferably overnight). Upon disinfection, the surfaces were then washed 3 times with dPBS and stored at 4 °C. Devices were used for seeding tissue within 48 h upon disinfection.

iPSC-μHM Formation

Purified iPSC-CM were dissociated using 0.25% trypsin EDTA and resuspended in Embryoid Body 20 (Dulbecco’s Modified Eagle Media, DMEM, containing 20% Qualified Fetal Bovine Serum; EB 20) supplemented with 10 μM Y27632 (Peprotech), 150 μg/mL L-ascorbic acid, 4 μg/mL vitamin B12 and 3.2 μg/mL penicillin. To fill each “dog bone” shaped mold, 3 μL of 5 × 10⁷ cell/mL cell mixture was used.18 After seeding, the plate was centrifuged at 100 g for 5 min. The cells were then incubated in 37 °C with 5% CO₂ for 2 h before adding supplemented EB 20 media to the well containing μHM. Uniaxial contraction was typically observed within 24 h of forming μHM (Supplemental Video 1). 48 h after seeding, the EB 20 media were changed to RPMI-c with 150 μg/mL L-ascorbic acid, 4 μg/mL vitamin B12 and 3.2 μg/mL penicillin. The media was then changed every 4 days. Data in this study are derived from μHM from 6 independent differentiations of isogenic control WTC and 4 independent differentiations of MYBPC3^+/− iPSC-CM.

Fluorescent High-Speed Imaging

Following tissue formation, the physiology of iPSC-μHM at day 5, 10 and 15 (corresponding to days 35, 40 and 45 for μHM formed from day 30 iPSC-CM, respectively) was analyzed using high speed fluorescent microscopy. For these studies, μHM were imaged with an epifluorescence microscope (Nikon Eclipse Ts2R; Tokyo, Japan) equipped with a digital CMOS camera (Hamamatsu ORCA-Flash4.0 V2; Japan), and Aura II Epifluorescence source with single LED light sources (450, 488, 540 and 647 nm).18 Videos of spontaneously contracting iPSC-μHM were taken at 100 frames per second for 5 s. Samples were warmed using a thermal plate (Tokai Hit; Shizuoka, Japan) set to a constant 37 °C temperature during measurements.

iPSC-μHM Traction Force Kinetics

iPSC-μHM contractility was measured by quantifying substrate embedded green fluorescent beads motion using open source TFM software50 (beads motion: see Supplemental Video 2). Although TFM method was originally used for 2D studies, previous experimental studies and finite element simulation have validated TFM is an adequate way to quantify tissue-level traction forces.18 To prevent intense calcium signals effects on traction analyses, green fluorescent videos were first post-processed using unsharp filter to emphasize fluorescent beads before running TFM software (Supplemental Fig. 1). The analyzed force-time traces were used to assess peak contraction force and traction kinetics using a custom MATLAB code previously developed to analyze calcium handling kinetics.18,39 Kinetic parameters measured include beat rate, traction rise time (time required for relaxed tissues to reach peak contraction; Fig. 4a), and traction relaxation 30, 50 and 75 times (times required for traction to decrease 30, 50 and 75% from peak back to relaxed state; Fig. 4a). To overcome confounding effects of spontaneous beat rate on kinetics of traction, the time duration was then corrected using Fridericia’s formula: T_CF = $\frac{T}{\sqrt[3]{\mathit{RR}}}$, and RR is the interval between spontaneous beats.13

Figure 4.

Traction kinetics of iPSC-μHM. a Illustration of traction force kinetics measurement, traction rise time, time required for 30%, 50% and 75% relaxation from peak. Beat rate corrected traction rises time of WTC and MYBPC3^+/− μHM on day 5 (b), 10 (c) and 15 (d). At day 5 (b), substrate rigidity did not affect WTC or MYBPC3^+/− μHM rise time. MYBPC3^+/− μHM had significantly longer contraction rise time compared to WTC. At day 10 (c), For WTC, 15 kPa μHM had significantly shorter rise time compared to other substrate stiffness conditions, while substrate rigidity did not affect MYBPC3^+/− μHM traction rise time. MYBPC3^+/− μHM had significant longer rise time on 5 and 15 kPa compared to WTC. At day 15 (d), substrate rigidity did not affect traction rise time for both genotypes. Compared to WTC, MYBPC3^+/− μHM had significantly longer rise time on 5 and 15 kPa. Beat rate corrected traction relaxation time of WTC and MYBPC3^+/− μHM on day 5 (e), 10 (f) and 15 (g). At day 5 (e), two-way ANOVA indicated both genotype and substrate rigidity were not affecting traction relaxation. At day 10 (f), WTC μHM had significant longer relaxation on 30 kPa substrate compared to 15 kPa. MYBPC3^+/− μHM relaxation did not change on different stiffness conditions. However, on 30 kPa condition, WTC had longer decay compared to MYBPC3^+/− μHM. At day 15 (g), compared to WTC, MYBPC3^+/− μHM had significant longer relaxation on 15 kPa substrate. h Summary of statistical analysis between different stiffness conditions within WTC or MYBPC3^+/− μHM at different time points. Error bar: SD. *, **, ***, and **** indicates p < 0.05, 0.01, 0.001, 0.0001.

iPSC-μHM Calcium Transient

To study iPSC-μHM calcium-contraction coupling, the same GFP channel videos that were used for quantifying traction forces were also used to analyze Ca²⁺ transients. To reduce fluorescent beads motion effects, three representative bead-free regions of the tissue were selected to quantify the Ca²⁺ transient peak amplitude and kinetics. Waveforms were analyzed to assess beat rate, Ca²⁺ upstroke duration (time required to reach peak Ca²⁺), background corrected Ca²⁺ peak amplitude (∆F/F₀), and Ca²⁺ decay times. The time interval was corrected using Fridericia’s formula as described above.18 To calculate the Ca²⁺ upstroke velocity, we divided background corrected Ca²⁺ peak amplitude (∆F/F₀) by upstroke duration.

Statistics

GraphPad 8.4.3 was used for statistics. For data between 2 groups, parametric unpaired t-test with Welsh correction (assuming non-equal mean values) was used. For data between multiple groups, one-way or two-way analysis of variance (ANOVA) was used. Following ANOVA results, multiple comparison between different groups was performed using Holm Sidak’s test. The p value less than 0.05 was considered a statistically significant difference.

Results

Generation of WTC and MYBPC3^+/− iPSC-CM

Following differentiation and lactate media-based purification shown in Fig. 1a, both WTC and MYBPC3^+/− iPSC were able to generate iPSC-CM with clear sarcomere structures (Fig. 1b). Baseline levels of MYBPC3 transcript expression in the MYBPC3^+/− iPSC-CM prior to tissue formation were significantly lower (~ 32%) than the levels expressed in isogenic control iPSC-CM (p = 0.008, Fig. 1c). Flow cytometry results indicated similar purity of WTC and MYBPC3^+/− iPSC-CM (p = 0.364; Figs. 1d and 5e). These observations are consistent with the finding that heterozygous MYBPC3 mutations are linked to HCM but not congenital heart malformation, suggesting that MYBPC3 mutations do not interfere with basal cardiomyocytes differentiation.40,42

Figure 5.

Calcium intake of iPSC-μHM. Representative Ca²⁺ transient curves for WTC (a) and MYBPC3^+/− (b) μHM for 15 kPa condition at day 15. MYBPC3^+/− μHM exhibited prolonged plateau phase at peak Ca²⁺. Background corrected Ca²⁺ amplitude for WTC and MYBPC3^+/− μHM on day 5 (c), 10 (d) and 15 (e). MYBPC3^+/− μHM had significantly higher Ca²⁺ intake compared to WTC. f Summary of statistical analysis between different stiffness conditions within WTC or MYBPC3^+/− μHM at different time points. For MYBPC3^+/− μHM, 30 kPa had significantly higher Ca²⁺ amplitude compared to 15 and 65 kPa. Error bar: SD. *, **, ***, and **** indicates p < 0.05, 0.01, 0.001, 0.0001.

Device Fabrication and Optimization

Glutaraldehyde-crosslinking of proteins to APTES-modified surfaces has been widely reported.16,27,36 However, autofluorescence resulting from this grafting approach has been noted.31 We found that a prolonged incubation in dPBS between APTES grafting and GA reaction substantially reduced autofluorescence (p = 0.023, Fig. 2b). Despite the lowered substrate autofluorescence, FN antibody binding assays indicated that this dPBS incubation did not affect FN bioactivity (p = 0.259, Fig. 2c). It is likely that this prolonged dPBS incubation period reduces autofluorescence by washing out unreacted APTES.

Substrate Rigidity Enhances MYBPC3^+/− iPSC-μHM Hypercontractility

Tissue physiology, including spontaneous beat rate and contractility, was analyzed 5, 10 and 15 days upon forming the tissue (Fig. 3). Overall, the MYBPC3^+/− μHM exhibited hypercontractile property compared to isogenic controls at all time points. Figure 3a shows two representative day 5 TFM contraction waveforms from WTC (top) and MYBPC3^+/− (bottom) μHM at 15 kPa condition with relatively similar spontaneous beat rates. Compared to WTC μHM, we observed much higher traction force and longer traction rise time for MYBPC3^+/− μHM. Moreover, MYBPC3^+/− μHM exhibited longer duration at peak traction (with a steep rise at the beginning and then a slower rise phase).

Figure 3.

Substrate rigidity enhances MYBPC3^+/− iPSC-μHM hypercontractility. a Representative traction force curves of WTC (top) and MYBPC3^+/− (bottom) μHM on 15 kPa condition at day 5. A more ‘rounded’ peak was seen in MYBPC3^+/− μHM with a slower rise phase close to the peak. Spontaneous beat rate of WTC and MYBPC3^+/− μHM at day 5 (b), 10 (c) and 15 (d). On day 5 (b), MYBPC3^+/− μHM had significantly lower beat rate compared to WTC. At day 10 (c), spontaneous beat rate was not affected by substrate rigidity or genotype. At day 15 (d), MYBPC3^+/− μHM had significantly lower beat rate on 15 kPa. Contractility of WTC and MYBPC3^+/− μHM at day 5 (e), 10 (f) and 15 (g). At day 5 (e), WTC μHM traction force increased with enhanced stiffness and peak at 30kPa condition. 30 kPa tissue had significantly higher force than 5 and 15 kPa but stay similar compared to 65 kPa. MYBPC3^+/− μHM generated the highest traction force at a lower threshold, 15 kPa condition, which was significantly higher compared to 5 kPa; but 15 kPa was similar compared to 30 and 65 kPa conditions. Compared to WTC, MYBPC3^+/− μHM generated significantly higher force on 5, 15, and 65 kPa. Similar trend was also observed at day 10 and 15 (g). However, at day 15, WTC 30 kPa tissue contractility decreased significantly and therefore did not differ from other conditions. (h) Summary of statistical analysis between different stiffness conditions within WTC or MYBPC3^+/− μHM at different time points. Error bar: SD. *, **, ***, and **** indicates p < 0.05, 0.01, 0.001, 0.0001, ns indicates non-significant p > 0.05.

In terms of spontaneous beat rate, at day 5, WTC μHM at 30 kPa had significant faster beat rate compared to 5 kPa (p = 0.014); however, substrate rigidity did not affect MYBPC3^+/− μHM spontaneous beat rate (Fig. 3h). Compared to WTC, MYBPC3^+/− tissues had significantly lower beat rate compared to WTC, with WTC beat rate slightly above 1 Hz and MYBPC3^+/− beat rate around 0.7 Hz (Fig. 3b); yet, at 65 kPa, beat rate was similar between two genotypes (around 0.7 Hz). At day 10, beat rate remained similar across different genotypes and stiffness conditions with an average of 1 Hz, comparable to human beat rate (Fig. 3c). At day 15, for WTC, 30 kPa tissues contracted much slower compared to 15 kPa. MYBPC3^+/− μHM beat rate was not affected by substrate stiffness (slightly below 1 Hz). However, compared to WTC, MYBPC3^+/− μHM contracted at a significantly lower rate on 15 kPa substrates (Fig. 3d).

For contractility, at day 5, substrate rigidity triggered increase in traction force for both genotypes (p < 0.0001). For WTC, 30 kPa tissues generated the most traction (~ 60 μN), which is significantly higher compared to 5 kPa (~ 20 μN, p < 0.0001) and 15 kPa (~ 33 μN, p = 0.028); while 30 kPa generated similar force compared to 65 kPa (51 μN, p = 0.545). Meanwhile, MYBPC3^+/− μHM had a lower peak threshold with 15 kPa tissue generated the most traction (~ 100 μN), which was significantly higher compared to 5 kPa (~ 49 μN, p = 0.0003); however, 15 kPa force was similar compared to 30 kPa (~ 73 μN, p = 0.078) and 65 kPa (90 μN, p = 0.486, Fig. 3h). Compared to WTC, MYBPC3^+/− μHM exhibit hypercontractility on 5 kPa (p = 0.015), this hypercontractility was further enhanced on 15 kPa condition with p < 0.0001, Fig. 3e).

Similar trends were observed on day 10: WTC-μHM exhibited peak contractility on 30 kPa substrates (~ 45 μN). In contrast, MYBPC3^+/− μHM exhibited peak contractility at 15 kPa (Fig. 3f). Compared to WTC, MYBPC3^+/− μHM had significantly higher traction at all levels of stiffness. However, the magnitude of difference between the MYBPC3^+/− and WTC μHM was markedly higher at softer substrates (5 and 15 kPa conditions; Fig. 3f).

At day 15, for WTC, overall trends in traction forces remained similar. However, μHM formed on substrates with stiffness similar to the early stages of fibrosis (30 kPa)18 exhibited a marked decrease in contractility from day 10 to day 15. For MYBPC3^+/− μHM, tissue traction still peaked at 15 kPa and remained similar on 30 and 65 kPa, which were significantly more traction compared soft 5 kPa (Fig. 3h). Compared to WTC, MYBPC3^+/− μHM still exhibited hypercontractility (Fig. 3g).

Traction Kinetics of iPSC-μHM

Traction force kinetics, specifically, the time required for reaching peak contraction and relaxation time were evaluated for both WTC and MYBPC3^+/− μHM at different stiffness conditions.39 Hypercontractility of MYBPC3^+/− μHM was correlated with significantly longer traction rise time. These longer rise time of MYBPC3^+/− μHM were seen at day 5 for all stiffness conditions. However, at day 10 and 15, significantly longer rise time were seen only on 5 and 15 kPa conditions. In contrast to our relatively frequent observation of prolonged contraction rise times in MYBPC3^+/− μHM, longer traction relaxation time of MYBPC3^+/− μHM was only observed at 15 kPa at later time point day 15 (Fig. 4).

Calcium Handling of iPSC-μHM

Cardiac contraction is strongly linked to Ca²⁺ handling, and contractile dysfunction may be linked to mishandling of this key ion.3 To investigate the calcium-contraction coupling of the iPSC-μHM, Ca²⁺ transient was evaluated. Notably, we observed a prolonged plateau phase at the peak Ca²⁺ for MYBPC3^+/− μHM, as shown in Figs. 5a and 5b. Correlated with the higher traction force generation in MYBPC3^+/− μHM, these tissues exhibited significantly higher Ca²⁺ amplitude (∆F/F₀) at all 3 time points (genotype factor: p <0.0001; Figs. 5c to 5e). MYBPC3^+/− μHM on 30 kPa substrates had significantly higher background-corrected peak Ca²⁺ intensity compared to 15 and 65 kPa at all three time points (p < 0.01, Fig. 5f).

Consistent with the observation of longer contraction rise times in MYBPC3^+/− μHM, we observed significantly longer Ca²⁺ upstroke duration in these tissues compared to controls (genotype factor: p < 0.0001, Figs. 6a to 6c), which may partially result from prolonged plateau phase at Ca²⁺ peak. We found substrate stiffness did not alter Ca²⁺ upstroke duration for WTC μHM over time (Fig. 6j). For MYBPC3^+/− μHM, substrate stiffness did not change Ca²⁺ upstroke at day 5 and day 10. However, at day 15, MYBPC3^+/− μHM had much longer upstroke time on 65 kPa compared to 30 kPa (p = 0.028, Fig. 6j).

Figure 6.

Calcium transient of iPSC-μHM. beat rate corrected Ca²⁺ upstroke duration for WTC and MYBPC3^+/− μHM on day 5 (a), 10 (b) and 15 (c). MYBPC3^+/− μHM had significantly longer upstroke compared to WTC. Ca²⁺ upstroke velocity for WTC and MYBPC3^+/− μHM on day 5 (d), 10 (e) and 15 (f). At day 5 (d), MYBPC3^+/− μHM had significant higher Ca²⁺ upstroke velocity on 30 kPa. Substrate rigidity did not affect Ca²⁺ upstroke velocity for both genotypes. At day 10, MYBPC3^+/− μHM had higher Ca²⁺ upstroke velocity on 5 kPa. For both genotypes, 65 kPa had lower Ca²⁺ upstroke velocity than 30 kPa. At day 15, similar upstroke velocity was observed at day 15 (f) between genotypes. For both genotypes, 65 kPa had lower Ca²⁺ upstroke velocity than 30 kPa. beat rate corrected Ca²⁺ decay for WTC and MYBPC3^+/− μHM on day 5 (g), 10 (h) and 15 (i). At day 5 (g), Ca²⁺ decay was similar across different substrate rigidity and genotype. At day 10 (h), compared to WTC, MYBPC3^+/− μHM had significantly shorter Ca²⁺ decay on 30 and 65 kPa. At day 15 (i), Compared to WTC, MYBPC3^+/− μHM had significantly longer decay on 15 kPa condition, while had significantly shorter decay on 30 kPa condition. j Summary of statistical analysis between different stiffness conditions within WTC or MYBPC3^+/− μHM at different time points. Error bar: SD. *, **, ***, and **** indicates p < 0.05, 0.01, 0.001, 0.0001.

Despite the prolonged Ca²⁺ upstroke duration observed in MYBPC3^+/− μHM, the average upstroke velocity (Ca²⁺ amplitude (∆F/F₀) divided by upstroke duration) was similar in most conditions between WTC and MYBPC3^+/− μHM (Figs. 6d to 6f). Nevertheless, higher upstroke velocity for MYBPC3^+/− μHM was seen on 30 kPa at day 5, as well as 5 kPa at day 10 (Figs. 6d to 6f). Moreover, significantly lower upstroke velocity was observed on stiff 65 kPa condition for both genotypes (Fig. 6j).

Ca²⁺ decay time remained similar across all substrate stiffness and genotype conditions at day 5 (Fig. 6g). At day 10 (Fig. 6h), WTC had significantly longer Ca²⁺ decay on stiffer 30 and 65 kPa substrates compared to 15 kPa. Compared to WTC, MYBPC3^+/− μHM had significantly shorter Ca²⁺ decay on 30 kPa (p = 0.002) and 65 kPa (p = 0.005) substrates. At the terminal timepoint day 15, WTC 30 kPa tissues had much longer decay on 30 kPa compared to 15 kPa (p < 0.0001) while MYBPC3^+/− μHM had much longer decay on 15 kPa compared to 30 kPa (p = 0.0004, Fig. 6j). At this day 15 timepoint, notably, longer decay times appeared to be correlated with peak contraction in each genotype. Interestingly, compared to WTC, MYBPC3^+/− μHM had significantly longer Ca²⁺ decay on 15 kPa (p = 0.013) but significantly shorter Ca²⁺ decay on 30 kPa (p < 0.0001, Fig. 6i) substrates.

Discussion

We optimized our previously developed model to study the synergistic effects of substrate rigidity and genetic mutations in affecting the development of HCM phenotypes. Physiological studies in iPSC-μHM indicated that (1) MYBPC3^+/− μHM exhibited hypercontractility compared to WTC; and substrate rigidity exacerbated the hypercontractile phenotype. (2) WTC μHM traction peaked at 30 kPa condition, while MYBPC3^+/− μHM had peaked at a softer, 15 kPa condition. (3) The hypercontractility of MYBPC3^+/− μHM was correlated with higher intercellular Ca²⁺, longer Ca²⁺ transients with a prolonged Ca²⁺ plateau phase at peak. These findings are consistent with the observation that although patients without an HCM-associated genotype can develop hypertrophic remodeling and cardiac hypercontractility in the setting of high afterload (caused by high blood pressure and linked to stiffnesses of the myocardium),19 patients with an HCM-positive genotype can develop such symptoms, including hypercontractility, at blood pressure levels that are benign in genotype-negative individuals.54 MYBPC3 mutations account for almost half the cases of inherited HCM, and clinical studies suggest that the phenotypes of HCM patients with MYBPC3 mutations do not differ significantly from patients with mutations in thick or thin filament proteins (e.g. Myosin Heavy Chain Beta (MHC-β, MYH7).40,63 Nevertheless, our findings remain to be confirmed in iPSC-based models of other HCM associated mutations, such as those in MYH7.

WTC and MYBPC3^+/− iPSC-CM

Spontaneous beating cardiomyocytes with clear sarcomere structures indicated successful CM differentiation and purification from iPSC for both genotypes (Fig. 1). Although MYBPC3 resides in the sarcomere, heterozygous deletion of this protein did not change the gross sarcomere structure at baseline (Fig. 1b). This finding is consistent with our previous 3D micro tissue study: even with complete knock out of MYBPC3, similar sarcomere structure was seen in MYBPC3^−/− tissues compared to isogenic control tissues.39In vivo, clear sarcomere patterns were also reported in MYBPC3 knock out mice heart sections.20 These results indicate MYBPC3 mutations alone may not be sufficient to cause sarcomere structure disarray. Instead, sarcomere disarray may arise from a combination of sarcomere mutations with prolonged mechanically and/or hormonally induced stress.15

Contractility of iPSC-μHM

In this study, we observed hypercontractility of MYBPC3^+/− μHM, which is consistent with clinical and in vivo animal studies, where patients and mice with an HCM-linked MYBPC3 loss of function mutations exhibit hypercontractility.55,59 Moreover, a similar hypercontractility phenotype was previously observed in other in vitro EHT systems using human or mice CM.7,56 However, these previous studies used EHT that are optimized to exhibit a large degree of tissue-induced motion, hence the level of mechanical resistance (overload) in these models is low. Thus, these previous studies reported a modest effect of MYBPC3 mutations on EHT contractility, which is similar to what we observed at low substrate rigidity levels.

In contrast, we observed that hypercontractility of MYBPC3^+/− μHM was further exacerbated with substrate rigidity. Interestingly, whereas isogenic controls exhibited increased contractility when substrate stiffness approached 30 kPa, MYBPC3^+/− μHM exhibited peak contractility at 15 kPa condition. This suggests an increased sensitivity to mechanical stiffness. Furthermore, we observed lower beat rates in MYBPC3^+/− μHM at an early time point day 5, but it remained similar compared to WTC on day 10. Compared to studies where a similar or slightly higher beat rate was reported in HCM mutations,11,56,66 the slower beat rate of MYBPC3^+/− μHM could potentially be due to a differential adaptation to the substrate rigidity: previous studies also advised that mechanical overload can cause slower spontaneous contraction in isogenic control engineered cardiac tissues.18,39

Notably, whereas 3D EHT studies,7 including our present data, along with clinical studies,55 suggest hypercontractility in HCM-mutant heart muscle, studies with 2D human iPSC-CM with the same MYBPC3^+/− mutation we studied here indicated deficits in contractility.4,50 These suggest that genotype-phenotype relationships in iPSC-CM can be fundamentally different in 2D and 3D studies, highlighting the importance of physiologically relevant in vitro model systems.

In our previous work with MYBPC3 null iPSC-CM, we observed no difference between contractility of MYBPC3^-/- and isogenic control cardiac micro-tissues contracting against compliant vs. rigid fibrous substrates.39 In contrast, in the current study, compared to WTC, MYBPC3^+/− μHM exhibited hypercontractility at all levels of stiffness. However, while the rigid fibers in our previous work caused tissue contraction forces within the range we observed here (> 10 μN), the most compliant fibers were contracted by significantly smaller forces (1 μN).39 It is thus possible that using even softer substrates (< 5 kPa), corresponding to the stiffness experienced in early stages of embryonic heart development, would lead to convergence of contractility generated by MYBPC3^+/− and control tissues.

In the present study, we observed the hypercontractility of MYBPC3^+/− μHM was associated with a longer traction rise time, and a more ‘rounded’ traction peaks as shown in Fig. 3a. In particular, the initial phase of the rise was not slower with MYBPC3 mutation; instead, the rise seems to be slowed near the peak plateau phase, which may reflect the intracellular Ca²⁺ accumulation caused by increased late Na⁺ currents associated with HCM.9,62 In contrast, many previous studies with 3D human or mouse EHT models have reported similar or even faster contraction rise-times with HCM mutations.7,11,21,56 Nevertheless, longer contraction rise-times have been reported in primary cardiomyocytes isolated from myectomy surgeries performed on patients with MYBPC3-linked HCM.26 Similarly, one study using iPSC-CM associated with MHC mutation showed longer contraction with less force generation.45 Clinically, HCM patients often exhibit impaired mechanical relaxation,6 which has been observed in some studies with iPSC-HCM EHT models.7 Here, we observed that MYBPC3^+/− μHM had similar or slight shorter relaxation time at earlier time points day 5 and 10, but longer relaxation on 15 kPa condition at later time point. Existing studies have observed shorter relaxation in both mice56,59 and EHT models.11 While, longer relaxation has also been reported.7,8,39 Altogether, these data suggest that impairment of mechanical relaxation is not solely caused by HCM mutations, and may instead require changes in mechanical stress, culture duration, and/or secondary changes in tissue structure (e.g., fibrosis) downstream of initial changes in cardiomyocyte function.

Calcium Dynamics of iPSC-μHM

The sarcomeric contractile apparatus not only responds to calcium, but also plays a role in intracellular Ca²⁺ handling.8,30,43 Here, we found substantially higher Ca²⁺ intensity and upstroke duration were associated with hypercontractility in MYBPC3^+/− μHM. In particular, a prolonged plateau phase at Ca²⁺ peak was identified in MYBPC3^+/− μHM as shown in Figs. 5a and 5b, which likely contributed to the longer Ca²⁺ upstroke (Figs. 6a to 6c). This prolonged Ca²⁺ peak may reflect that MYBPC3 loss of function mutations induce faster cross-bridge turn over and incomplete Ca²⁺ release from the sarcomeres.44 However, secondary effects of HCM on electrophysiology are also possible. For example, clinically, HCM mutations are associated with increased late Na⁺ current.9,62 This could also lead to the prolonged “plateau” in the calcium transients we observed.

Potential Ca²⁺ mishandling of HCM caused by sarcomere mutations can lead to diastolic dysfunction.9,12,38,43 Interestingly, however, while we saw an overall increase in systolic time for both contractility and calcium transients, when we quantified the decay time (amount of time required for calcium amplitude to decay from its peak value to either 70, 50 or 25% of that amplitude), we did not observe an time increase, whereas prolonged Ca²⁺ decay times have been reported in some other studies.9,53 Moreover, we observed a faster Ca²⁺ decay on stiffer substrates (high mechanical resistance) in MYBPC3^+/− μHM; this result is similar to reports of faster Ca²⁺ transient decay times in some animal12 and in vitro39 HCM models. Altogether, these results suggest that mechanical rigidity may exacerbate genetically-associated abnormalities in Ca²⁺ handling. Importantly, due to the prolonged Ca²⁺ plateau phase at peak in MYBPC3^+/− μHM (Fig. 5b), the true peak identified from the signal (red circles in Fig. 5b) may underestimate the Ca²⁺ decay times. Future studies may include specifying different traction and Ca²⁺ rise phases (steep and slower rises) and decay times right after the steep phase.

Substrate Rigidity Enhanced 3D iPSC-μHM Model

Our iPSC-μHM platform allows a relatively pure cardiomyocyte population (> 90%; Fig. 1e) to form tissue, which allowed us to focus mainly on the cardiomyocyte behavior from different genotypes. While non-cardiomyocyte cell types, including fibroblasts,58 also play a role in HCM pathogenesis, our findings in 3D μHM tissues suggest that crosstalk between substrate rigidity and HCM genotypes may not require action of these other cell types. Compared to existing 3D-engineered heart muscle technologies using micro tugs to modulate material elasticity to change mechanical resistance,23,32,39 our system involves a simpler device fabrication process, increasing assay throughput.

There are several limitations to the present study. First, we captured both traction force (green fluorescence beads motion) and calcium (GCaMP6f intensity) dynamics together in the same videos. While this allows a direct examination of calcium-contraction coupling, it substantially increases the time required for image analysis, as there was a need to manually select bead-free regions for analysis of Ca²⁺ waveforms. Second, despite similar clinical phenotypes among HCM patients with different sarcomere mutations, different phenotype severity was also reported,34,64 to further delineate genotype-phenotype relations using iPSC-μHM model, different sarcomere mutation (e.g., MYH7) may be investigated in future studies. Third, unlike 3D bio-printed cardiac organoid that subject to hemodynamic loadings,29 the stiffness of the substrate where tissue contract against is not a true mechanical force heart experience in vivo. Nevertheless, multiple systems have demonstrated that change of material elasticity where cell or tissue formed on can induce phenotypes like increased cell size and contractility, similar to what was observed in vivo with increased pressure overload (afterload).23,32,51,61

In summary, we have demonstrated that HCM-associated mutations synergize with substrate rigidity to trigger hypercontractility in an iPSC-cardiac micro-tissue model. Tissues harboring an HCM-related mutation exhibited hypercontractility and hypersensitivity to mechanical stiffness. These results account for the potential crosstalk between genotype and mechanical loading, which may be essential for future studies on HCM pathogenesis, and in developing more robust therapies for this disease.

Electronic supplementary material

Below is the link to the electronic supplementary material.