Modelling diastolic dysfunction in induced pluripotent stem cell-derived cardiomyocytes from hypertrophic cardiomyopathy patients

Abstract

Aims

Diastolic dysfunction (DD) is common among hypertrophic cardiomyopathy (HCM) patients, causing major morbidity and mortality. However, its cellular mechanisms are not fully understood, and presently there is no effective treatment. Patient-specific induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) hold great potential for investigating the mechanisms underlying DD in HCM and as a platform for drug discovery.

Methods and results

In the present study, beating iPSC-CMs were generated from healthy controls and HCM patients with DD. Micropatterned iPSC-CMs from HCM patients showed impaired diastolic function, as evidenced by prolonged relaxation time, decreased relaxation rate, and shortened diastolic sarcomere length. Ratiometric Ca²⁺ imaging indicated elevated diastolic [Ca²⁺]_i and abnormal Ca²⁺ handling in HCM iPSC-CMs, which were exacerbated by β-adrenergic challenge. Combining Ca²⁺ imaging and traction force microscopy, we observed enhanced myofilament Ca²⁺ sensitivity (measured as dF/Δ[Ca²⁺]_i) in HCM iPSC-CMs. These results were confirmed with genome-edited isogenic iPSC lines that carry HCM mutations, indicating that cytosolic diastolic Ca²⁺ overload, slowed [Ca²⁺]_i recycling, and increased myofilament Ca²⁺ sensitivity, collectively impairing the relaxation of HCM iPSC-CMs. Treatment with partial blockade of Ca²⁺ or late Na⁺ current reset diastolic Ca²⁺ homeostasis, restored diastolic function, and improved long-term survival, suggesting that disturbed Ca²⁺ signalling is an important cellular pathological mechanism of DD. Further investigation showed increased expression of L-type Ca²⁺channel (LTCC) and transient receptor potential cation channels (TRPC) in HCM iPSC-CMs compared with control iPSC-CMs, which likely contributed to diastolic [Ca²⁺]_i overload.

Conclusion

In summary, this study recapitulated DD in HCM at the single-cell level, and revealed novel cellular mechanisms and potential therapeutic targets of DD using iPSC-CMs.

Translational perspective

Diastolic dysfunction (DD) is commonly associated with hypertrophic cardiomyopathy (HCM). DD can progress into diastolic heart failure, the major cause of morbidity and mortality due to HCM. However, the cellular pathogenic mechanisms of DD remain unclear, and thus an effective drug therapy has yet to be developed. The present study establishes human induced pluripotent stem cell-derived cardiomyocyte (iPSC-CM) model that recapitulates impaired diastolic function in HCM patients at the single cell level. Functional imaging analysis and drug testing based on this HCM iPSC-CM DD model provided new mechanistic insights into the cellular pathophysiology and potential therapeutic strategies of DD.

Introduction

Hypertrophic cardiomyopathy (HCM) is characterized by abnormal thickening of the ventricular wall and increased risk of arrhythmia, sudden death, and heart failure.¹ Many familial HCM cases are caused by mutations in sarcomere genes.² Diastolic dysfunction (DD), manifested by slowed or incomplete ventricular relaxation during diastole, is a prominent clinical feature in HCM patients.³ Without treatment, DD can progress into heart failure, which leads to significant morbidity and mortality.^4–6 However, the underlying cellular mechanisms of DD in HCM are not well understood, which greatly impedes the development of specific and effective therapies.

To study the mechanisms of DD in HCM, various animal models (from rodents to large animals) and disease-induction methods have been used.^7,8 However, significant interspecies differences make it difficult to extrapolate animal results to humans, and there has been no investigation of DD mechanism using human-origin cardiomyocytes. Human induced pluripotent stem cell (iPSC) technology has enabled patient-specific disease modelling of various cardiovascular diseases,^9–13 which offered a unique opportunity to establish human cell modelling platforms of DD in HCM.

Here, we generated induced pluripotent stem cell-derived cardiomyocyte (iPSC-CM) models of DD that carry different mutations in three most commonly involved sarcomeric proteins in familial HCM [i.e. myosin heavy chain 7 (MYH7), myosin binding protein C3 (MYBPC3), and cardiac troponin T2 (TNNT2)]. These iPSC-CM models of DD recapitulated impaired diastolic function at the single-cell level. Based on various functional assays, we demonstrated that both elevated diastolic Ca²⁺ and increased myofilament Ca²⁺ sensitivity contribute to the cellular mechanisms of DD in HCM iPSC-CMs. To rescue impaired relaxation, we rebalanced Ca²⁺ homeostasis in HCM iPSC-CMs with optimized doses of Ca²⁺ blockers (verapamil, diltiazem) and late Na⁺ current blockers (ranolazine, eleclazine), all of which improved cellular diastolic performance. Our research provides a patient-specific iPSC-CM platform for modelling and elucidating cellular mechanisms of DD in HCM.

Methods

Peripheral blood mononuclear cell isolation and reprogramming

HCM patients with DD and healthy control individuals were recruited based on the most updated guidelines.^14,15 All recruitment and consenting procedures were carried out according to Stanford Institutional Review Board (IRB) approved protocol. Patient-specific iPSC lines were generated through over-expression of 4-Yamanaka factors in peripheral blood mononuclear cells (Supplementary material online, Method).

Maintenance and differentiation of induced pluripotent stem cells

Human iPSCs were maintained in Essential 8™ Medium (Gibco^®, Life Technology) and were differentiated into beating cardiomyocytes using monolayer differentiation protocol¹⁶ (Supplementary material online, Method).

Cell micropatterning on the hydrogel

The single iPSC-CMs were micropatterned according to the adult CM-like morphology on 10 kPa polyacrylamide hydrogel, which was pre-coated with Matrigel (Corning, 1:10 dilution in PBS) in a defined geometry relying on the stencil-based microfabrication techniques, as described previously.¹⁷

Statistical analysis

All data are presented as mean ± standard deviation. The statistical analysis was performed using SigmaPlot 12.5 and GraphPad Prism 8 (Supplementary material online, Method).

Results

Human induced pluripotent stem cell-derived cardiomyocytes carrying hypertrophic cardiomyopathy mutations exhibit diastolic dysfunction

We recruited six familial HCM mutation carriers who exhibited clinical records of diastolic dysfunction (DD) (Supplementary material online, Figure S1A). These patients carry MYH7 R663H (n = 2), MYBPC3 V321M (n = 1), MYBPC3 V219L (n = 1), and TNNT2 R92W (n = 2) mutations. We also recruited one HCM patient without the DD phenotype, and who carried a mutation in MYBPC3 (c.2905+1G>A). Healthy individuals (n = 2) without HCM mutations were used as wild type controls (Supplementary material online, Table S1). Peripheral blood mononuclear cells of all groups were isolated and reprogrammed to iPSCs. The quality of all derived iPSC lines was tested by immunostaining of pluripotent markers such as Sox2, Nanog, SSEA-4, and Oct4 (Supplementary material online, Figure S1B). Human iPSC-CMs were differentiated using standard monolayer methods.¹⁶ The purity of iPSC-CMs was >90% as evaluated by fluorescence-activated cell sorting (FACS) and immunostaining of cardiac sarcomere protein markers troponin T and α-actinin (Supplementary material online, Figure S1C and D).

To measure the diastolic function, beating iPSC-CMs from each group were dissociated into single cells and seeded on a micropatterned matrix with a 1:7 width/length ratio (Figure 1A).¹⁷ After recovery, micropatterned iPSC-CMs exhibited well-aligned sarcomere structures along the longitudinal axis (Figure 1A, bottom panel). To mark the sarcomere structure in live cells, a GFP-human actin fusion protein was expressed in the iPSC-CMs (Figure 1B). Sarcomere shortening was recorded using confocal microscopy by line scanning along the myofilaments (Figure 1B and C). Our results show that the diastolic sarcomere length is significantly shorter in all three HCM iPSC-CM groups (1.788 ± 0.113, 1.844 ± 0.111, and 1.720 ± 0.130 µm for MYH7, MYBPC3, and TNNT2 groups, respectively) compared with the Ctrl cells (1.985 ± 0.138 µm) (Figure 1D). Interestingly, the maximum relaxed sarcomere lengths of Ctrl and HCM iPSC-CMs, measured in Ca²⁺-free Tyrode’s solution, were comparable (Figure 1E), indicating that the shorter diastolic sarcomere length in HCM iPSC-CMs was not due to structural remodelling of the sarcomere. Hypertrophic cardiomyopathy iPSC-CMs also exhibited a decreased maximum relaxation rate and prolonged normalized relaxation duration when compared with Ctrl iPSC-CMs (Figure 1F and G), whereas the beating rate and maximum contraction velocity remained unchanged (Supplementary material online, Figure S2A–D). These results suggest that iPSC-CM model exhibits a distinct DD phenotype at the single-cell level.

Figure 1.

Recapitulation of diastolic dysfunction in patient-specific hypertrophic cardiomyopathy (HCM) induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs). (A) Representative immunostaining image of micropatterned iPSC-CMs, showing aligned sarcomere structure. (B and C) Sarcomeres in single iPSC-CMs were indicated by GFP-actin over-expression, and the sarcomere shortening were captured by live-cell confocal line-scanning. (D–G) Results show shorter diastolic sarcomere length (D), unchanged maximum relaxation sarcomere length (E), slower relaxation velocity (F), and prolonged normalized relaxation duration (G) in all HCM iPSC-CM groups compared with iPSC-CMs by one-way ANOVA (Tukey method). N = 42, 36, 40, and 44 cells in Ctrl, MYH7, MYBPC3, and TNNT2 groups, respectively. For each group, data were generated from two different iPSC lines and three batches of differentiation.

Abnormal Ca²⁺ handling properties in hypertrophic cardiomyopathy induced pluripotent stem cell-derived cardiomyocytes with diastolic dysfunction

Given the central role of Ca²⁺ in regulating cardiac excitation-contraction coupling, we sought to understand the mechanisms of DD by examining the Ca²⁺ handling in Ctrl and HCM iPSC-CMs with Fura-2 imaging (Figure 2A). Field stimulation was most stable at 0.5 HZ (Supplementary material online, Fig. S3A–D), which was used in the comparisons of all groups. Diastolic [Ca²⁺]_i of all HCM groups was significantly higher than that in Ctrl iPSC-CMs (Figure 2B), while the normalized decay time of HCM iPSC-CMs was markedly prolonged compared with the Ctrl (Figure 2C). Moreover, caffeine-induced Ca²⁺ transients showed higher sarcoplasmic reticulum (SR) Ca²⁺ load and slightly increased caffeine transient decay Tau in HCM iPSC-CMs compared with Ctrl iPSC-CMs (Figure 2D–F), which suggests reduced Na⁺/Ca²⁺ exchange (NCX) function in HCM iPSC-CMs. We also characterized the iPSC-CMs from the HCM patient without clinical DD (Supplementary material online, Figure S4A), which showed normal diastolic Ca²⁺ and relaxation functions (Supplementary material online, Figure S4B–I).

Figure 2.

Ratiometric Fura-2 AM imaging reveals elevated diastolic Ca²⁺ in hypertrophic cardiomyopathy (HCM) induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs). (A) Representative traces of Fura-2 ratiometric Ca²⁺ imaging. (B and C) HCM iPSC-CMs showed elevated diastolic Ca²⁺ (B) and prolonged Ca²⁺ recycling (C) compared with Ctrl iPSC-CMs by one-way ANOVA (Tukey method). N = 51, 42, 43, and 72 cells in Ctrl, MYH7, MYBPC3, and TNNT2 groups, respectively. For each group, data were generated from two different iPSC lines and three batches of differentiation. (D) Representative traces of SR Ca²⁺ load measurement by caffeine-induced Ca²⁺ transient. (E and F) Caffeine-induced transient showed higher SR Ca²⁺ content (E) and slightly decreased caffeine-transient decay-Tau (F) in HCM iPSC-CMs compared with Ctrl iPSC-CMs by one-way ANOVA (Tukey method). N = 22, 28, 25, and 26 cells in Ctrl, MYH7, MYBPC3, and TNNT2 groups, respectively. For each group, data were generated from two different iPSC lines and two batches of differentiation.

Isogeneic induced pluripotent stem cell lines confirmed the pathogenicity of TNNT2 R92W mutation in diastolic dysfunction

To confirm the pathogenicity of the sarcomeric mutations in the observed DD phenotype, we next generated isogenic Ctrl and diseased iPSC lines of TNNT2 R92W mutation using CRISPR/Cas9 (Supplementary material online, Figure S5A–E).¹⁸ Functional assays showed that both heterozygous and homozygous mutation introduction led to elevated diastolic Ca²⁺ and prolonged Ca²⁺ transient decay time. Moreover, the mutation-carrying isogenic iPSC-CMs exhibited shorter diastolic sarcomere length, slower relaxation, and prolonged relaxation duration (Supplementary material online, Figure S5F–L). On the other hand, mutation-corrected patient iPSC-CMs showed improved diastolic Ca²⁺ homeostasis and diastolic function compared with the diseased cells (Supplementary material online, Figure S5F–L). Collectively, these results confirmed the TNNT2 R92W HCM mutation is pathogenic and causes DD in iPSC-CM models.

Functional imaging indicates enhanced risk of diastolic dysfunction in hypertrophic cardiomyopathy induced pluripotent stem cell-derived cardiomyocytes

To evaluate how elevated diastolic [Ca²⁺]_i affects the relaxation in HCM iPSC-CMs, we next combined traction force microscopy (TFM) and [Ca²⁺]_i recording to simultaneously assess Ca²⁺ handling and contractility (Figure 3A–D). While the peak Δ[Ca²⁺]_i amplitude was unchanged, the maximum contractile velocity was significantly faster in HCM compared with Ctrl iPSC-CMs (Figure 3E and F). Based on previous study on the [Ca²⁺]_i and contractile force relationship,¹⁹ we defined the ratio of the contraction rate to the Ca²⁺ transient amplitude (dF/Δ[Ca²⁺]_i) as an index of myofilament Ca²⁺ sensitivity, which was significantly increased in all HCM groups compared with Ctrl (Figure 3G).

Figure 3.

Functional imaging indicates enhanced risk of diastolic dysfunction in hypertrophic cardiomyopathy (HCM) induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs). (A–C) Representative traces of simultaneous recording of Ca²⁺ transient and contractile stress. (D) Measurement of Ca²⁺ and contraction parameters from a single beating episode. (E–H) HCM iPSC-CMs showed unchanged Ca²⁺ transient amplitude (E), faster contractile rate (F), increased risk of DD as indicated by Ca²⁺ sensitivity index (G), and prolonged relaxation duration (H) compared with Ctrl iPSC-CMs by one-way ANOVA (Tukey method). N = 77, 190, 192, and 64 cells in Ctrl, MYH7, MYBPC3, and TNNT2 groups, respectively. For each group, data were generated from two different iPSC lines and three batches of differentiation.

To further validate the Ca²⁺ sensitivity index, we also measured dF/Δ[Ca²⁺]_i in dilated cardiomyopathy (DCM) iPSC-CMs that carry a TNNT2 R173W mutation known to reduce the myofilament Ca²⁺ sensitivity.²⁰ As expected, DCM iPSC-CMs showed unchanged diastolic [Ca²⁺]_i, decreased transient amplitude, weaker contractile force, slower contractile rate, and reduced dF/Δ[Ca²⁺]_i (Supplementary material online, Figure S6A–F). Taken together, HCM iPSC-CMs concomitantly exhibited elevated diastolic [Ca²⁺]_i and enhanced myofilament Ca²⁺ sensitivity, both of which contribute to the increased diastolic tension and impaired relaxation, as reflected by prolonged relaxation duration of HCM iPSC-CMs in TFM measurement (Figure 3H).

Short-term β-adrenergic challenge exaggerates diastolic dysfunction in hypertrophic cardiomyopathy induced pluripotent stem cell-derived cardiomyocytes

β-adrenergic signalling is the most important lusitropic regulation in the intact heart.²¹ To evaluate the effect of β-adrenergic activation in diastolic function regulation, we next measured the function of Ctrl and HCM iPSC-CMs in response to β-agonist isoproterenol (ISO, 1 μM) treatment (Figure 4A, E). Our results showed that ISO did not alter diastolic [Ca²⁺]_i, but did increase the Ca²⁺ transient amplitude in Ctrl iPSC-CMs. However, in HCM iPSC-CMs, ISO significantly increased diastolic [Ca²⁺]_i but not Ca²⁺ transient peak, which resulted in a reduction of transient amplitude (Figure 4B, Supplementary material online, Figure S7A). Additionally, the normalized Ca²⁺ transient decay times were significantly increased after ISO treatment in HCM but not Ctrl iPSC-CMs (Figure 4C). β-adrenergic stimulation increased the SR load in Ctrl iPSC-CMs by approximately 20%, but this increment was absent in HCM iPSC-CMs (Supplementary material online, Figure S7B). As diastolic [Ca²⁺]_i was further elevated, the abnormal arrhythmia-like Ca²⁺ transients were also significantly increased in ISO-treated HCM iPSC-CMs (Figure 4D). In sarcomere shortening measurement, the beating rates of Ctrl and HCM iPSC-CMs were significantly accelerated by ISO (Figure 4F). However, HCM iPSC-CMs consistently showed shorter diastolic and peak sarcomere lengths, and slower relaxation velocity after ISO treatment compared with Ctrl iPSC-CMs (Figure 4G–H, Supplementary material online, Figure S7C). Moreover, both normalized relaxation and contraction duration 90 were further prolonged in ISO treated HCM iPSC-CMs compared with untreated HCM cells (Figure 4I, Supplementary material online, Figure S7D). Collectively, these results strongly suggest that short-term β-adrenergic challenge exaggerate DD in HCM iPSC-CMs.

Figure 4.

Short term β-adrenergic challenge exaggerates diastolic dysfunction in hypertrophic cardiomyopathy (HCM) induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs). (A) Representative traces of Fura-2 AM Ca²⁺ imaging of Ctrl and HCM iPSC-CMs with or without isoproterenol (ISO) treatment. Red arrow indicates abnormal Ca²⁺ handling events. (B and C) ISO challenge further enhanced diastolic Ca²⁺ concentrations and significantly prolonged the normalized transient decay times in HCM iPSC-CMs compared with non-ISO groups by two-way ANOVA (Holm–Sidak method). N > 31 cells in each group. (D) Percentage of abnormal Ca²⁺ handling events. N > 28 cells in each group. All groups were compared with non-ISO groups by z-test. (E) Representative traces of sarcomere shortening measurement in Ctrl and HCM iPSC-CMs with or without ISO treatment. (F–I) ISO increased beating rate in both Ctrl and HCM iPSC-CMs (F), yet HCM iPSC-CMs still showed shorter diastolic sarcomere length (G), slower relaxation velocity (H), and further prolonged relative relaxation duration (relaxation duration 90 normalized to the beating interval) in HCM iPSC-CMs after ISO treatment (I). All groups were compared with non-ISO group by two-way ANOVA (Holm–Sidak method). N > 28 cells in each group. For all the groups, data were generated from two different iPSC lines and three batches of differentiation.

Ca²⁺ and late Na⁺ current blockers rebalance Ca²⁺ homeostasis and restore diastolic function

Ca²⁺ channel blockers have shown beneficial effects when prescribed to HCM patients with DD.^22,23 In addition, late Na⁺ current (I_NaL) has been reported to play a role in the pathogenesis of DD, and late Na⁺ channel blockers improved diastolic function in both HF patient CMs and DD animal models.^24,25 However, the therapeutic effects of Ca²⁺ current (I_Ca) and I_NaL inhibition in DD are not clear. To better understand the therapeutic mechanisms of these agents, we treated HCM iPSC-CMs with Ca²⁺ blockers (verapamil and diltiazem) and late Na⁺ channel blockers (ranolazine and eleclazine) (Figure 5A, Supplementary material online, Figure S8A). The drug doses (10–200 nM for verapamil, 10–500 nM for diltiazem, 1–10 µM for ranolazine, and 200 nM to 1 µM for eleclazine) were applied according to previous studies and clinical prescriptions (Supplementary material online, Figure S8A).¹¹ Our results show that the treatment at an optimized dose of these drugs have restored diastolic [Ca²⁺]_i and transient decay time in HCM iPSC-CMs (Figure 5B and C, Supplementary material online, Figure S8B–D). Sarcomere shortening measurements also confirmed that optimized dose of Ca²⁺ and late Na⁺ blockers partially recovered diastolic sarcomere lengths (Figure 5D, Supplementary material online, Figure S8E), increased relaxation rate, and reduced relaxation duration in HCM iPSC-CMs (Figure 5 and F, Supplementary material online, Figure S8F–G). Sacubitril/valsartan (neprilysin inhibitor and angiotensin receptor blockers) has been shown to be effective for treating HF patients with DD.²⁶ However, when tested in our DD model, sacubtril/valsartan has no significant effect, probably because our single-cell iPSC-CM model does not reflect the signalling complexity that exists at the tissue/systemic level (Supplementary material online, Figure S8A–G).

Figure 5.

Ca²⁺ and late Na⁺ channel blockers rescued prolonged diastolic process in hypertrophic cardiomyopathy (HCM) induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) by re-balancing Ca²⁺ homeostasis in diseased cells. (A) Representative traces of Fura-2 AM Ca²⁺ imaging of Ctrl and HCM iPSC-CMs with and without Ca²⁺ and late Na⁺ blocker treatment. (B and C) Treatment by Ca²⁺ and late Na⁺ blocker significantly reduced diastolic Ca²⁺ concentration (B) and shortened Ca²⁺ transient decay duration (C). N > 27 cells in each groups. All groups were compared with untreated HCM iPSC-CMs by two-way ANOVA (Tukey method). (D–F) Using the sarcomere shortening assay, treatment by all three drugs retained the diastolic sarcomere length (D), increased relaxation rate (E), and reduced the normalized relaxation duration in HCM iPSC-CMs compared with untreated baseline (F). N > 28 cells for each groups. All groups were compared with untreated HCM iPSC-CMs by two-way ANOVA (Tukey method). For each group, data were generated from two different iPSC lines and three batches of differentiation.

Improve the viability of hypertrophic cardiomyopathy induced pluripotent stem cell-derived cardiomyocytes by targeting long-term Ca²⁺ overload

To understand the mechanisms of diastolic Ca²⁺ overload in HCM iPSC-CMs, we quantified the mRNA expression of key Ca²⁺ handling proteins. qPCR results show increased Cav1.2, MYH7, and stress marker (NPPA, NPPB) and decreased MYH6, Casq2, and Cav3 mRNA levels in HCM iPSC-CMs compared with Ctrl iPSC-CMs (Figure 6A, Supplementary material online, Figure S9A). Our results also showed that transient receptor potential cation channels (TRPCs) 1 and 3, the main cardiac TRPC subtypes, were both increased in HCM iPSC-CMs (Supplementary material online, Figure S9B, Figure 6E). As TRPCs mediate cation entry and contribute to cardiac hypertrophic remodelling,^27,28 we next evaluated the TRPC-mediated Ca²⁺ influx via store-operated Ca²⁺ entry (SOCE) measurement (Supplementary material online, Figure S9C–E). Our results showed that Ctrl and HCM iPSC-CMs exhibited comparable basal [Ca²⁺]_i in Ca²⁺-free buffer (Supplementary material online, Figure S9F), while HCM iPSC-CMs showed higher TRPC-dependent SOCE amplitude and rising rate compared with the Ctrl group (Supplementary material online, Figure S9G and H). To validate the role of TRPCs during the pathogenesis of DD, we also treated HCM iPSC-CMs with the newly developed TRPC3 specific blocker Pyr3, which improved Ca²⁺ handling and diastolic function in the DD model (Supplementary material online, Figure S8A–G).

Figure 6.

Viability of hypertrophic cardiomyopathy (HCM) induced pluripotent stem cell-derived cardiomyocytes (iPSC-CMs) under long-term β-adrenergic activation is improved by rebalancing Ca²⁺ homeostasis. (A) Real time PCR expression profiles of key Ca²⁺ related proteins. All groups compared with Ctrl with Student’s t-test, data from two lines of Ctrl iPSC-CMs and two lines of HCM iPSC-CMs from at least two differentiation batches. (B) Western blot analysis shows increased CaMKIId activation (pThr286) in HCM iPSC-CMs compared with Ctrl iPSC-CMs, which was recovered by Ca²⁺ and late Na⁺ channel blocker treatment. All groups compared with HCM group with Student’s t-test. Data from two lines of each iPSC-CMs from at least two differentiation batches. (C and D) Long-term isoproterenol (ISO) treatment further enhanced the diastolic Ca²⁺ overload and prolonged Ca²⁺ transient decay time in HCM iPSC-CMs, which were partially rescued by Ca²⁺ and late Na⁺ channel blockers. N > 50 cells for each group. All groups compared with ISO treated HCM iPSC-CMs by one-way ANOVA (Tukey method). (E and F) Cell viability assay indicates that Ca²⁺ channel blockers partially restored viability of HCM iPSC-CMs after long-term ISO challenge. β-blockers were used as treatment control and doxorubicin was used as cell death control. All groups compared with baseline (white bar) by two-way ANOVA (Tukey method). For each group, data were generated from two different iPSC lines and two batches of differentiation.

Take home figure.

iPSC-CM model reveals a novel cellular mechanism of diastolic dysfunction in hypertrophic cardiomyopathy.

A recent study reported that CaMKIId activation contributes to diastolic Na⁺ and Ca²⁺ overload and impaired diastolic function.²⁹ Indeed, our western blot analysis showed that CaMKIId autophosphorylation was significantly up-regulated by 308 ± 42% in HCM iPSC-CMs compared with Ctrl iPSC-CMs, which was partially suppressed by I_Ca or I_NaL blockers (Figure 6B). To confirm the role of CaMKIId in DD, we sought to over-express (OE) the constitutively-active form of CaMKIId T287D in Ctrl iPSC-CMs, and to knock-down (KD) CaMKIId in HCM iPSC-CMs (Supplementary material online, Figure S10A–C). Our functional assays demonstrated CaMKIId gain-of-function lead to typical DD phenotype in Ctrl iPSC-CMs, while knock-down of CaMKIId in HCM iPSC-CMs partially recovered Ca²⁺ homeostasis, diastolic function, and long-term survival (Supplementary material online, Figure S10D–K). These results confirm that CaMKIId activation plays an importing role during the pathogenesis of DD in HCM iPSC-CMs.

To evaluate the long-term beneficial effects of I_Ca and I_NaL blockers in DD, we next compared the function and viability of HCM and Ctrl iPSC-CMs under chronic β-adrenergic challenge (100 nM ISO, 7 days) with and without treatments of β-blockers, I_Ca and I_NaL blockers, and doxorubicin (positive control for cell death). Our results showed long-term ISO treatment impaired diastolic function in HCM iPSC-CMs in terms of elevated diastolic [Ca²⁺]_i and prolonged Ca²⁺ transient decay time, which were alleviated by I_Ca and I_NaL inhibition (Figure 6C and D). Long-term ISO treatment also activated apoptotic signalling, as both Ctrl and HCM iPSC-CMs exhibited increased caspase activity (Figure 6E and F), while β-blockers repressed the pro-apoptotic effect. Interestingly, I_Ca and I_NaL blockers elicited significant pro-survival effects only in HCM iPSC-CMs, presumably because diastolic Ca²⁺ handling was only altered in HCM iPSC-CMs.

Taken together, these results suggest that by rebalancing Ca²⁺ homeostasis, I_Ca and I_NaL blockers exert beneficial effects on HCM iPSC-CMs both in short-term functional performance and long-term signalling integrity.

Discussion

As a growing cause of congestive heart failure, DD has been widely modelled and investigated in various systems, including rodents, canine, swine, and newly isolated or cultured cardiomyocytes.^7,8 However, few studies have been based in human cardiomyocytes, and none are patient-specific. Here, we established a novel platform to investigate DD using patient-specific HCM iPSC-CMs. Working with multiple functional imaging technologies, we systematically measured the relaxation function of diseased cells and demonstrated that diastolic Ca²⁺ overload and increased myofilament Ca²⁺ sensitivity both contribute to DD. This is the first demonstration that patient-specific iPSC-CM model can recapitulate DD disease phenotype at the single-cell level. Moreover, we showed that the diastolic function in HCM iPSC-CMs was improved by the treatment of commonly used drugs, such as I_Ca and I_NaL blockers, through rebalancing of Ca²⁺ homeostasis. These findings suggest that iPSC-CMs are a suitable platform for disease modelling, mechanism study, and drug testing of DD in HCM.

Hypertrophic cardiomyopathy mutations are known to sensitize myofilaments in cardiomyocytes.³⁰ To measure the myofilament Ca²⁺ sensitivity, we established a novel iPSC-CM-based imaging method that combines [Ca²⁺]_i imaging and TFM. Compared with traditional Ca²⁺ sensitivity assays in permeabilized muscle tissue or skinned cardiomyocytes, our method offers exceptional simplicity in usage, faster reading and analysis, and higher throughput potential.³¹ With this technology, we defined a novel parameter (dF/Δ[Ca²⁺]_i) that reflects the myofilament Ca²⁺ sensitivity and indicates the risk of DD in our single-cell iPSC-CM model. Given our ability to recruit HCM patients with sarcomeric variants and generate their iPSCs, the functional imaging assays validated here show great promise as a potential diagnostic platform to predict the risk and severity of DD in carriers of known HCM mutations who lack clinical symptoms, as well as in individuals carrying variants of unknown significance.³²

Ca²⁺ homeostasis is critical for the contractile function of cardiomyocytes,³³ and Ca²⁺ mishandling has been implicated in the pathogenesis of various cardiac diseases.^10,11,34 Our study showed that diastolic [Ca²⁺]_i in HCM iPSC-CMs was significantly elevated, which contributes to increased basal tension and impaired diastolic function. Based on our observations, we believe that two mechanisms contribute to the diastolic [Ca²⁺]_i overload in HCM iPSC-CMs:

First, Ca²⁺ removal via NCX is reduced in HCM iPSC-CMs, as evidenced by slower [Ca²⁺]_i decline of caffeine-induced Ca²⁺ transients (Figure 2F). A possible reason for reduced NCX function with unaltered gene expression may be elevated [Na⁺]_i, which limits the [Na⁺] gradient that fuels Ca²⁺ extrusion. Impaired Ca²⁺ extrusion through NCX loads Ca²⁺ in HCM iPSC-CMs and drives the observed higher SR Ca²⁺ content. Interestingly, ISO treatment greatly increased SR Ca²⁺ content in Ctrl iPSC-CMs but not in HCM iPSC-CMs, which indicates the SR Ca²⁺ load has reached a pump-leak balance in diseased cells. As SR Ca²⁺ leak is steeply dependent on SR Ca²⁺ content,³⁵ the secondary rise in SR Ca²⁺ leak would be expected to promote the local activation of CaMKIId, which is concentrated at SR Ca²⁺ release sites. CaMKIId activation further increases Ca²⁺ influx via I_Ca, promotes arrhythmogenic SR Ca²⁺ leak, and increases I_NaL. All of these excess Ca²⁺ and Na⁺ fluxes further contribute to overload of diastolic [Ca²⁺]_i.³⁶ This positive inherent feedback cycle can be broken by partial blockade of I_Ca and I_NaL, which restored diastolic [Ca²⁺]_i in HCM iPSC-CMs.

Second, HCM mutations increased myofilament Ca²⁺ sensitivity and elevate diastolic active force development and basal tension. As previous reports showed that pressure overload in heart leads to increased expression of TRPC channels that promote cardiac hypertrophy,^27,28 TRPC up-regulation may contribute to DD pathogenesis in HCM iPSC-CM model. Indeed, qPCR profiling and SOCE measurements confirmed that the expression and activity of LTCC, and TRPC1/3 were up-regulated in HCM iPSC-CMs. Increased Ca²⁺ influx through these channels promoted diastolic Ca²⁺ overload, and impaired the relaxation of HCM iPSC-CMs, which was partially rescued by TRPC3 specific blocker Pyr3.

One important outcome of long-term diastolic Ca²⁺ overload in HCM iPSC-CMs is the activation of CaMKIId (Figure 6B). Chronic CaMKIId activation are commonly seen in heart failure. However, our gain- and loss-of-function study based on iPSC-CM models has clearly confirmed the critical role of CaMKIId activation during the pathogenesis of DD, which exacerbates the Ca²⁺ and Na⁺ disturbances, induces apoptosis, and influences transcriptional regulation.³⁶ This finding indicate two possible working mechanisms through which I_Ca or I_NaL inhibition relieve symptoms in patient with DD:^22–25 (i) I_Ca and I_NaL blockers limit overall Ca²⁺ and Na⁺ influx, which stabilize the diastolic and improve the ability of NCX to extrude Ca²⁺; (ii) I_Ca and I_NaL blockers maintain the Ca²⁺ homeostasis and restrict the CaMKIId activity. Indeed, re-balancing diastolic [Ca²⁺]_i in HCM iPSC-CMs exhibited the beneficial effects in both short-term diastolic function and long-term cell viability. Therefore, all the key factors in this positive feedback loop during DD pathogenesis, including increased diastolic [Ca²⁺]_i, CaMKIId activity, I_NaL, TRPCs, and SR leakage, could be potential therapeutic targets of DD at the cellular level (Supplementary material online, Figure S11).

One limitation of our study is that our single-cell iPSC-CM model cannot capture tissue or systemic level contributions of increased wall thickness and extracellular matrix changes,³⁷ which may also contribute to DD in HCM. Other factors, such as changes in myocardium stiffness, conduction blockade, aging, and increased fibrosis cannot be recapitulated in our current model. The morphological and physiological differences of iPSC-CMs compared with adult ventricle cardiomyocytes also pose challenges to clinical applications.³⁸ Future studies incorporating engineered heart tissues may enable further systematic investigations of DD at a more integrated level.

In summary, we report a human iPSC-based cellular model of HCM that recapitulates DD at the single-cell level. Functional imaging analysis highlighted diastolic Ca²⁺ overload and enhanced myofilament Ca²⁺ sensitivity, both of which contribute to the increased diastolic tension and impaired relaxation seen in HCM iPSC-CMs. Moreover, perturbed Ca²⁺ homeostasis in HCM iPSC-CMs exacerbates the DD, providing a window of potential therapeutic intervention by I_Ca, I_NaL, and CaMKII inhibitors which may restore diastolic function and prevent apoptosis during long-term adrenergic stress. Our findings here augment our knowledge of the pathogenesis in DD and the mechanisms underlying the beneficial effects of currently available drugs. Moreover, our work suggests that the iPSC-CM model provide a suitable platform for mutation-specific mechanistic studies and drug screening for DD in HCM.

See page 3696 for the editorial comment on this article (doi: 10.1093/eurheartj/ehz447)