|Home | About | Journals | Submit | Contact Us | Français|
Dilated cardiomyopathy (DCM) is the most common cardiomyopathy, characterized by ventricular dilatation, systolic dysfunction, and progressive heart failure. DCM is the most common diagnosis leading to heart transplantation and places a significant burden on healthcare worldwide. The advent of induced pluripotent stem cells (iPSCs) offers an exceptional opportunity for creating disease-specific models, investigating underlying mechanisms, and optimizing therapy. Here we generated cardiomyocytes (CMs) from iPSCs derived from patients of a DCM family carrying a point mutation (R173W) in the gene encoding sarcomeric protein cardiac troponin T. Compared to the control healthy individuals in the same family cohort, DCM iPSC-CMs exhibited altered Ca2+ handling, decreased contractility, and abnormal sarcomeric α-actinin distribution. When stimulated with β-adrenergic agonist, DCM iPSC-CMs showed characteristics of failure such as reduced beating rates, compromised contraction, and significantly more cells with abnormal sarcomeric α-actinin distribution. β-adrenergic blocker treatment and over-expression of sarcoplasmic reticulum Ca2+ ATPase (Serca2a) improved DCM iPSC-CMs function. Our study demonstrated that human DCM iPSC-CMs recapitulated to some extent the disease phenotypes morphologically and functionally, and thus can serve as a useful platform for exploring molecular and cellular mechanisms and optimizing treatment of this particular disease.
Dilated cardiomyopathy (DCM) is a cardiac disease characterized by ventricular dilatation and systolic dysfunction (1). DCM is the most common cause of heart failure after coronary artery disease and hypertension, and is the leading indication for heart transplantations (2–3). The cost for management of DCM in the US alone has been estimated at between $4 and $10 billion (3). Mutations in genes encoding sarcomeric, cytoskeletal, mitochondrial, and nuclear membrane proteins, as well as proteins involved in Ca2+ metabolism, are linked to approximately a third to half the cases of DCM (4–6). Cardiac troponin T (cTnT) is one of the 3 subunits of the troponin complex (Troponin T, C, and I) that regulate the sarcomeric thin filament activity and muscle contraction in cardiomyocytes (CMs). cTnT is essential for sarcomere assembly, contraction, and force production (7). Mutations in the cardiac troponin T gene (TNNT2) often lead to DCM (8) and are frequently expressed as a malignant phenotype with sudden cardiac death and heart failure at an early age (9–10). In vitro biochemical studies have found that decreased Ca2+ sensitivity and/or ATPase activity, which impair force production, may be the underlying mechanisms for certain TNNT2-mutation induced DCM (9,11–14). Mouse models of TNNT2 mutations recapitulate the human DCM phenotype, providing extensive insight into the possible mechanisms of the disease (11–12). The contribution of mouse models in the overall understanding of DCM has been enormous. However, several important differences exist between the mouse and human models. For example, the mouse resting heart rate is approximately 10-fold faster than human’s. The electrical properties, ion channel contributions, and cardiac development of mouse CMs are all differ from those of human. Unfortunately, cardiac tissues from DCM patients are difficult to obtain and do not survive in long-term culturing. With the advent of induced pluripotent stem cells (iPSCs) (15–16), functional CMs can be differentiated from human iPSCs (17–18). Patient-specific iPSC-CMs such as the long QT syndrome, Leopard syndrome, and Timothy syndrome have been shown to recapitulate human cardiovascular diseases and enable the testing and optimization of empirical therapies (19–22). Thus, a human DCM iPSC-CM model would be an important complement to mouse models for understanding the cellular and physiological processes of DCM as well as for drug screening in human cells.
Here we generated iPSC-CMs from a three-generation family of DCM patients carrying a point mutation (R173W) in exon 12 of the TNNT2 gene. We studied the morphology and function of DCM iPSC-CMs, recorded their electrophysiology with patch clamping and microelectrode arrays (MEAs), assessed Ca2+ handling with Ca2+ transient analysis, and quantified contractile force production using atomic force microscopy (AFM). Compared to the controls, iPSC-CMs derived from DCM patients displayed a consistent increased heterogeneous sarcomeric organization at early stage post differentiation, consisting of a more severe punctate distribution of sarcomeric α-actinin. Individual DCM iPSC-CMs also exhibited altered Ca2+ handling compared to controls. β-Adrenergic stimulation increased the number of iPSC-CMs with abnormal sarcomeric α-actinin distribution, compromised contractility, and induced failure of spontaneous contraction. Importantly, we demonstrated that treatment with β1-selective β-blocker (metoprolol) improved the sarcomeric organization, whereas over-expression with sarcoplasmic reticulum Ca2+ ATPase (Serca2a) markedly increased the contractile force and improved Ca2+ handling in DCM iPSC-CMs. In summary, our results show that DCM iPSC-CMs to some extent recapitulated the disease phenotype, and therefore can be a useful tool for investigating the disease mechanisms involved understanding DCM and drug screening, as well as optimizing medical management.
We recruited a cohort of seven individuals from a DCM proband carrying an autosomal dominant point mutation on exon 12 of the gene TNNT2, which causes an Arginine (R) to Tryptophan (W) mutation at amino acid position 173 in the protein cTnT. The potential causal effect for DCM of this particular point mutation was confirmed by genetic screening of a panel of 17 primary DCM associated genes, in silico analysis (Table S1), and genetic co-segregation studies (Fig. S1). A mutation at the same amino acid position (R173G) was also reported in a completely unrelated Belgian family with DCM (23), suggesting a strong association of this particular locus with the disease. The seven recruited individuals covered 3 generations (I, II, and III) (Fig. 1A). Four patients (Ia, IIa, IIb, and IIIa) were confirmed to carry the TNNT2 R173W mutation in one of the two alleles by PCR amplifying the genomic locus of TNNT2 and DNA sequencing, while the other 3 individuals (Ib, IIc, and IIIb) were confirmed normal and served as controls in the subsequent studies (Fig. 1B and Fig. S1). All four patients who carry the specific R173W mutation manifested clinical DCM symptoms with dilated left ventricle and decreased ejection fraction, and were treated medically (Table S2). A 14-year-old diseased patient (IIIa) had an orthotopic heart transplant due to severe clinical symptoms.
To generate patient-specific iPSCs, skin fibroblasts were expanded from skin biopsies taken from each individual (Fig. 1C) and reprogrammed with lentiviral Yamanaka 4 factors (Oct4, Sox2, Klf4, and c-MYC) under feeder-free condition. Colonies with TRA-1-60+ staining and human embryonic stem cell (hESC)-like morphology (Fig. 1, D and E) were selected, expanded, and established as individual iPSC lines. For each individual, 3–4 iPSC lines were established for subsequent analyses. All of the DCM iPSC lines were confirmed to contain the specific R173W mutation by genomic PCR and DNA sequencing (Fig. S2). All established iPSC lines expressed the pluripotency markers Oct4, Nanog, TRA-1-81, and SSEA-4, and were positive for alkaline phosphatase (Fig. 1F). Microarray analyses indicated these iPSC lines were distinct from the parental skin fibroblasts, expressing a global gene pattern more similar to hESCs (Fig. S3A). Quantitative bisulphite sequencing showed that the promoter regions of Oct4 and Nanog were hypomethylated in all the tested iPSC lines, indicating active transcription of the pluripotency genes (Fig. 1G). The established iPSC lines maintained a normal karyotype after extended passage (Fig. S3B), with the majority of them exhibiting silencing of exogenous transgenes and re-expression of endogenous Nanog (Fig. S4). iPSC lines with incomplete transgene silencing were removed from the subsequent studies. These patient-specific iPSCs were able to differentiate in vitro into cells of all three germ layers (Fig. S5) and subsequently formed teratomas upon injection into the kidney capsules of immunodeficient mice (Fig. 1H).
We next differentiated the DCM iPSCs into the cardiovascular lineage using a well-established 3D differentiation protocol developed by Yang et al. (24). Two iPSC lines from each individual were selected for differentiation into spontaneous beating EBs and for subsequent functional analyses (Table S3). Spontaneous beating was observed as early as day 8 post differentiation. The efficiency of differentiation to cardiac lineage varied among different lines (Fig. S6A, Videos S1 and S2). Beating EBs derived from control and patient iPSCs contained approximately 50–60% cTnT-positive CMs (Fig. S6, B and C). Allele-specific reverse transcriptase (RT)-PCR of beating EBs derived from three iPSC clones of 3 DCM patients indicated bi-allelic expression of the wild type and mutant (R173W) TNNT2 gene (Fig. S7). The beating EBs from the control and DCM iPSCs 18–48 days post differentiation were seeded on multi-electrode array (MEA) probe (Fig. S8A and Video S3) and their electrophysiological properties recorded (Fig. S8B). Both control (n=45) and DCM (n=57) iPSC-derived beating EBs exhibited comparable beat frequencies, field potentials, interspike intervals, and field potential durations (FPD) at baseline (Table S4 and Fig. S8C).
We next dissociated the beating EBs into small beating clusters and single beating CMs for further analysis (Videos S4 to S7). The organization of myofibrils in the iPSC-CMs was assessed by immunocytochemistry. Both control and DCM iPSC-CMs expressed sarcomeric proteins cTnT, sarcomeric α-actinin, and myosin light chain 2a (MLC2a), as well as the cardiac gap junction protein connexin 43 (Fig. S9). However, compared to control iPSC-CMs (n=368) at day 30 post differentiation, a significant higher percentage of DCM iPSC-CMs (n=391) showed a punctate distribution of sarcomeric α-actinin over one fourth of the total cellular area (p=0.008) (Fig. 2A, 2B and Fig. S10, A to C). There were no significant differences in cell size between control and DCM iPSC-CMs (Fig. 2C) at this stage. This phenotype was consistently observed in two different DCM iPSC lines each from the 4 DCM patients, suggesting a homogeneous correlation to the disease-associating R173W mutation. Sarcomeric α-actinin is an excellent marker for sarcomeric integrity and degeneration and was used for evaluation of sarcomeric organizations in heart tissues from human patients with DCM (25). These results therefore suggest that, compared to the controls, an increased number of DCM iPSC-CMs had a more disturbed sarcomeric organization at this stage. Notably, CMs clusters always showed a much less disturbed sarcomeric pattern (Fig. S10, D and E). The majority of CMs with punctate sarcomeric α-actinin distribution were single cells or cells at the very edge of CMs clusters (Fig. S10, D and E), suggesting a higher tendency for single TNNT2 R173W DCM iPSC-CMs to malfunction in maintaining sarcomere integrity. To further assess the myofibrillar organization in detail, we performed transmission electron microscopy (TEM) on both control and DCM iPSC-CMs at day 30 post differentiation. Well organized myofibrils with aligned Z-lines and recognizable A-bands and I bands were found in both control (n=11) and DCM iPSC-CMs (n=12) (Fig. 2D and Fig. S10F), although mitochondria and sarcoplasmic reticulum were still immature in both groups at this stage (Fig. S10, G and H). However, compared to controls, DCM iPSC-CMs exhibited an increased variability in the degree of sarcomeric organization, with a higher number of less well-aligned Z-lines and scattered patterns of condensed Z-bodies (Fig. 2D and Fig. S10F). Overall, these results are consistent with the sarcomeric α-actinin immunostaining in DCM CMs shown in Fig. 2A and Fig. S10, A to C.
Positive inotropic stress can induce DCM phenotype in transgenic mouse models of DCM (26–27) and aggravate the disease in clinical patients (28). We next examined whether treatment with positive inotropic reagent, such as β-adrenergic agonists, can expedite the phenotypic response of DCM iPSC-CMs. Indeed, we found that 10 μM norepinephrine (NE) treatment induced an initial positive chronotropic effect that later became negative, eventually leading to failure of spontaneous contraction in DCM iPSC-derived beating EBs (n=14) as reflected by MEA recording. By contrast, the control iPSC-derived beating EBs (n=14) exhibited prolonged positive chronotropic activity (Fig. 3A). One week of 10 μM NE treatment in vitro markedly increased the number of CMs with punctate sarcomeric α-actinin distribution from DCM iPSC clones, with almost 80–90% of the DCM iPSC-CMs found to have the disorganized sarcomeric pattern (Fig. 3B, 3C, and Fig. S11, A to E). A few single DCM iPSC-CMs showed complete degeneration of myofilaments after prolonged NE treatment (Fig. 3B and Fig. S11A), which was not observed in control iPSC-CMs. TEM analyses indicated that, compared to controls (n=6), NE-treated DCM iPSC-CMs (n=7) exhibited a more severe scattered distribution of Z-bodies (Fig. 3D and Fig. S11F), which was consistent with the markedly increased heterogeneous pattern of sarcomeric α-actinin staining after NE treatment. Tracking with video imaging of individual beating clusters of both control and DCM iPSC-CMs treated with 10 μM NE over time showed distinct outcomes. Decreased inotropic and chronotropic activities were often observed in the DCM iPSC-CMs (Fig. 3E, 3F, and Videos S8 to S15). These results suggest that DCM iPSC-CMs are more susceptible to the stress of β-adrenergic stimulation.
Biomechanical stress generated by pressure or volume overload resulted from hypertension or myocardial injury often induce DCM and heart failure, and tends to aggravate the disease (29). We next examined the effect of mechanical stress on iPSC-CMs by cyclic stretch. Prolonged stretching led to significant thickening and loss of obvious striation of the myofibrils in both control and DCM iPSC-CMs (Fig. S12, A and B). Mechanical strain also significantly increased the number of cells with relative disorganized sarcomeric pattern in both control and DCM iPSC-CMs. However, compared to controls, an increased heterogeneity in sarcomeric pattern was observed in DCM iPSC-CMs (Fig. S12C). These results suggest that DCM iPSC-CMs are more susceptible to biomechanical stress.
CM contraction starts from the electrical excitation of the myocytes, as reflected by the membrane action potentials (APs) (30). To investigate the possible underlying etiology, we assessed whether the DCM-associating R173W mutation in TNNT2 affects the electrical excitation of the CMs. We examined the electrical activities of the dissociated single beating iPSC-CMs by patch clamping. Three types of spontaneous APs (ventricular-like, atrial-like, and nodal-like) were observed in both control and DCM iPSC-CMs (Fig. 4A). DCM ventricular-like myocytes (n=17) exhibited normal APs that were comparable to control (n=18) (Fig. 4B). The average action potential duration at 90% repolarization (APD90) of the DCM iPSC-CMs was not significantly different from that seen in control iPSC-CMs (Fig. 4C). The average AP frequency, peak amplitude, and resting potential were also very similar between the 2 groups (Fig. 4D to 4F). These results indicate that the electrical activities of individual control and DCM iPSC-CMs were normal at baseline, consistent with the results obtained by MEA analysis of beating EBs.
To further investigate the underlying DCM disease mechanism, we measured the Ca2+ handling properties at the excitation-contraction coupling level by fluorescent Ca2+ imaging. DCM iPSC-CMs (n=40, 5 lines from 3 DCM patients, Table S3) exhibited rhythmic frequency and timing comparable to those of the control iPSC-CMs (n=87, 5 lines from 3 control individuals, Table S3) (Fig. 5A, 5B, 5C, 5E, and 5F). However, DCM iPSC-CMs exhibited significantly smaller [Ca2+]i transient amplitudes compared to those of the control iPSC-CMs (p=0.002) (Fig. 5D), indicating the [Ca2+]i available for each contraction of DCM iPSC-CMs was significantly lower. The smaller [Ca2+]i transients of CMs were consistently observed in all examined DCM iPSC lines, suggesting weaker force production in DCM iPSC-CMs. To further analyze the Ca2+ handling properties of iPSC-CMs, both control and DCM iPSC-CMs were subjected to caffeine treatment to induce the Ca2+ release from the sarcoplasmic reticulum (SR) through ryanodine receptor (RyR) Ca2+ channels (Fig. 5, G–J). Compared to controls, DCM iPSC-CMs exhibited relatively smaller amplitudes, prolonged time to peak, and delayed decay time (Fig. 5, H–J), indicating that DCM iPSC-CMs have a relatively lower SR Ca2+ storage and altered function of Ca2+ related molecules such as SR Ca2+ release channels and Ca2+ pumps in the plasma and SR membranes.
Deficiency in contractile force production is one of the most important mechanisms responsible for inducing DCM and heart failure (4). To investigate this further, we next measured the contraction force of iPSC-CMs using atomic force microscopy (AFM), which has been used to measure contraction of cultured chicken embryonic CMs (31). The AFM allowed us to probe the contractile properties at the single cell level (Fig. S13 and Video S16). Compared to single control iPSC-CMs (n=13), the DCM iPSC-CMs (n=17) showed similar beat frequency and duration (Fig. S14) but significantly weaker contraction forces (Fig. 6A, 6B, and Table S5). There was no correlation between the cell size and contraction force for each single cell measured by AFM (Fig. S15).
Previous studies have shown that Serca2a over-expression, a treatment investigated in a pre-clinical trial (32), mobilized intracellular Ca2+ and restored contractility of cardiomyocytes in failing human hearts and improved failing heart functions in animal models (33–35). Given our results showing smaller Ca2+ transients and compromised contractility in the DCM iPSC-CMs, we hypothesized that over-expression of Serca2a may rescue the phenotypes of DCM iPSC-CMs. Transduction of DCM iPSC-CMs with adenoviruses carrying Serca2a co-expressing GFP (Ad.Seca2a) (see Material and Methods section) at a multiplicity of infection (MOI) of 100 led to over-expression of Serca2a in these cells (Fig. 6C). Co-expression of GFP along with Serca2a allowed us to recognize the individual transduced cells and measure their contractile forces by AFM (Fig. 6D and Videos S17 to S20). After 48 hours of transduction, over-expression of Serca2a (n=12) restored the contractile force of single DCM iPSC-CMs to a level similar to that seen in control iPSC-CMs (Fig. 6A, 6B, and Table S5). Ca2+ imaging using the red fluorescent Ca2+ indicator Rhod-2 AM (Fig. S16) indicated that DCM iPSC-CMs transduced with Ad.Serca2a co-expressing GFP (n=22) had significantly increased global [Ca2+]i transients compared to cells transduced with Ad.GFP only (n=14) (Fig. 6, E and F) (p=0.04), which is consistent with the restored force production. Although Rhod-2 Ca2+ dye is not ideal to quantify cytoplasmic Ca2+ level (since it is difficult to calibrate its level using the indicator in live cells), the kinetics of Ca2+ transients and sarcomeric organization in Serca2a-transduced DCM iPSC-CMs was not significantly changed (Fig. 6, G and H). On the other hand, over-expression of Serca2a in control iPSC-CMs failed to produce a statistically significant increase in contractility (Fig. S17), likely because the endogenous amount and function of SERCA were already at a relatively high level in control cells. Altogether, these results demonstrate that over-expression of Serca2a increased the [Ca2+]i transients and contraction force of DCM iPSC-CMs and improved their function to some extent.
Although Serca2a gene therapy is now in clinical trial, the overall mechanism of individual CM cellular response after Serca2a gene therapy has not been extensively studied previously (36). Hence we set out to investigate the mechanisms by which Serca2a delivery might repair defects in DCM iPSC-CMs. Gene expression profiling of Serca2a-transduced control and DCM iPSC-CMs showed different sets of genes had greater than 2-fold expression changes, indicating different responses to Serca2a over-expression (Table S6 and S7). There were 191 genes (65 upregulated and 126 downregulated) with greater than 2 fold expression changes in DCM iPSC-CMs over-expressed with Serca2a that were rescued to an expression level similar to those in control iPSC-CMs (Fig. S18A). Enriched pathways analysis indicated that several previously known pathways, such as Ca2+ signaling, protein kinase A signaling, and G-protein coupled receptor signaling, are significantly involved in rescuing the DCM phenotype by Serca2a over-expression (37–38). Interestingly, several pathways not previously linked to DCM, including factors promoting cardiogenesis, integrin and cytoskeletal signaling, and ubiquitination pathway, were also found to participate in rescuing the DCM CM function (Fig. S18B and Table S8).
Clinical studies have shown that metoprolol, a β1-selective β-adrenergic blocker, has a beneficial effect on the clinical symptoms and hemodynamic status of DCM patients (39–40). We thus tested whether in vitro metoprolol treatment has a beneficial effect on the TNNT2 R173 DCM iPSC-CMs as well. We found that 10 μM metoprolol treatment for one week significantly decreased the number of single DCM iPSC-CMs with disorganized sarcomericα-actinin staining (Fig. S19A). Although not statistically significant, metoprolol treatment on DCM iPSC-CMs resulted in a relatively reduced chronotropic effect and increased global Ca2+ transients on DCM iPSC-CMs (Fig. S19, B to D). Metoprolol treatment also significantly prevented aggravation of the DCM iPSC-CMs that is induced by NE treatment (Fig. S19E). No significant effect on sarcomeric α-actinin distribution in control iPSC-CMs treated with metoprolol was observed (Fig. S19F). These results suggest that blockade ofβ-adrenergic pathway helped DCM iPSC-CMs resist mechanical deterioration and improved their myofilament organization.
We have generated patient-specific iPSCs from a DCM family carrying a single point mutation R173W in the sarcomeric protein cTnT and derived CMs from these iPSCs. This has allowed us to generate, for the first time, a large number of human DCM-specific iPSC-CMs and to analyze their functional properties, explore the potential underlying etiology, and test effective therapies. Although the TNNT2 R173W mutation does not seem to affect other cells from cardiovascular lineage (Fig. S20), we observed significant phenotypic differences between the control and DCM iPSC-CMs.
In this study, a higher tendency of disturbance in sarcomeric organization was observed in DCM iPSC-CMs. An increased number of DCM iPSC-CMs exhibited punctate sarcomericα-actinin staining in immunocytochemistry and a more scattered distribution pattern of Z-bodies in transmission electron microscopy. Notably, this phenotype was more frequently observed in single cells or cells at the very edges of a cluster than in cells within the inner side of a cluster. The heterogeneous presentation of sarcomeric organization in iPSC-CMs could be explained by the following hypotheses. First, individual CMs and CM clusters have different architectural matrices and physical properties in tolerating mechanical forces generated by spontaneous contractions, leading to heterogeneous sarcomeric organization. As in the analogy to breaking chopsticks, a bundle of chopsticks will tolerate much stronger breaking force than just one. Second, CMs seeded on culture dishes confront different environmental factors, such as the topology of attaching surfaces and paracrine factors from surrounding cells, leading to heterogeneous myofilament organization. It is actually not unusual to observe heterogeneous sarcomeric organization in cultured rat neonatal CMs as shown by previous studies (41–42). The increased heterogeneous presentation of sarcomeric organization in DCM iPSC-CMs could be explained by their higher susceptibility to stress. Indeed, both β-agonist stimulation and cyclic stretch markedly increased the heterogeneity of sarcomeric organization in DCM iPSC-CMs, indicating they were more susceptible to positive inotropic stress.
Overall, our data are consistent to some extent with previous studies showing that muscle LIM protein (MLP)-deficient mice neonatal DCM CMs and zebrafish embryonic DCM heart tissues with mutations in nexilin were more susceptible to mechanical stress (43–44). These results suggest that DCM iPSC-CMs are less capable of maintaining their sarcomeric integrity compared to control iPSC-CMs, and more susceptible to positive inotropic and chronotropic stress.
While the baseline electrophysiological activities of the DCM iPSC-CMs were not significantly different from those of the controls, abnormal Ca2+ transients were found in the DCM iPSC-CMs. These results suggest that DCM iPSC-CMs have impaired Ca2+ handling associated with a lower contractility. Gene expression profiling using microarray analysis also indicates that DCM iPSC-CMs express a lower level of Ca2+-related key molecules (CASQ, TMEM38, NFAT, and NECAB) which is consistent with the compromised Ca2+ handling properties observed on Ca2+ imaging. Finally, AFM analyses indicate that individual DCM iPSC-CMs manifest decreased contractile force compared to controls, which is consistent with the smaller [Ca2+]i transients observed. Interestingly, previous studies have also shown altered Ca2+ handing in CMs isolated from human patients with heart failure (45). These CMs represent a very late stage of the disease, and it is still not clear whether the altered Ca2+ handing is the primary factor that contributes to the disease or merely a secondary consequence of the disease progression. Our current model of DCM using iPSC-CMs represents a very early stage of heart development, showing that abnormalities in Ca2+ handing can occur at very early stage.
Although there is no biochemical data showing how the particular cTnT R173W mutation impact the Ca2+ sensitivity or ATPase activity of the myofibers in the current literature, a major outcome of most of the cTnT mutations affecting DCM is a decreased Ca2+ sensitivity in the myofilaments (9–10,13–14). Decreased Ca2+ sensitivity usually suggests a decreased contractility in the myofibers at the physiological cytosolic Ca2+ concentration. Indeed, our AFM data indicated that the R173W cTnT DCM iPSC-CMs had decreased contractility compared to controls. It is likely that decreased contraction attenuates maturation of the DCM iPSC-CMs as well as the mechanical stretch-induced gene expression of molecules associated with myofilament and Ca2+ handling in DCM cells (Fig. S21), resulting in an increased heterogeneity in sarcomeric organization under inotropic and mechanical stress. Altered Ca2+ handling also possibly induced a reduction in Ca2+ transients under the condition without in vivo remodeling, which further decreased the contractility of DCM iPSC-CMs in vitro. This could form a negative cycle which eventually compromises the overall CMs function. Although a direct relationship between the R173W cTnT mutation and the abnormal Ca2+ handling has not been well-established, the defects in gene expression in our study suggest that the contractility and Ca2+ handling phenotypes were secondary consequences of the R173W mutation in DCM iPSC-CMs. Further biochemical and molecular studies are required in the future to understand the disease progression and the overall DCM mechanisms induced by the R173W mutation in the cTnT gene.
We have demonstrated that prolonged treatment of β-blocker metoprolol had a beneficial effect on the DCM iPSC-CMs’ abnormal sarcomeric phenotype by decreasing the number of single CMs with abnormal sarcomeric α-actinin staining. Metoprolol treatment led to a relatively negative chronotropic effect and an improved global Ca2+ transient on DCM iPSC-CMs. These results are consistent with a previous study showing that metoprolol treatment on cultured neonatal rat CMs induced both negative chronotropic and positive inotropic effect on fast beating cells (46). These results also indicate that metoprolol can reduce the number of contraction in a given time, possibly reflecting an improved contraction force; they may represent some of the beneficial factors contributing to the protective effects on DCM iPSC-CMs. In addition, metoprolol treatment on cultured neonatal rat CMs have been shown to upregulate protein levels of cardiac gap junction channels (47), allowing a better connection and communication between cells. Altogether, these may exert beneficial effects on the DCM iPSC-CMs observed in our study.
We have also shown in this study that over-expression with Serca2a, a novel gene therapy treatment for heart failure currently in clinical trials (48), can significantly improve the contractile function of DCM iPSC-CMs. Delivery of exogenous Serca2a could prevent the possible negative cascades induced by the R173W mutation in cTnT (Fig. S21). It also overcame the phenotype of lower SR Ca2+ storage and contractility in DCM iPSC-CMs, thus rescuing the compromised contractility. Gene expression profiling further identified several novel pathways that are involved in Serca2a rescue, including ubiquitination and integrin signaling pathways. These results could also help guide further investigations of other molecular mechanisms underlying DCM and in finding potential targets to treat DCM in the future.
In summary, our data indicate that the TNNT2 R173W mutation caused impairment in myofilament regulation, Ca2+ handling, and force production of individual CMs, which might be the primary reason for the eventual appearance of the DCM clinical phenotype in patients. Despite some limitations in the current iPSC-CM platform (e.g., CM immaturity and lack of an in vivo environment for possible disease remodeling), our overall findings demonstrate that the iPSC platforms are useful to investigate disease mechanisms at an early stage of genetic diseases and to discover novel therapeutic targets. Patient-specific human iPSC-CMs model of DCM could be an important complement to the biochemical and mouse DCM models to help us understand the complex etiology and disease mechanisms. In the future, we anticipate more studies will use this platform for exploring mechanisms and treatments of the different genetic mutations responsible for familial DCM as well as other hereditary cardiovascular disorders.
All of the protocols for this study were approved by the Stanford University Human Subjects Research Institutional Review Board. Generation of patient-specific iPSC lines were performed as previously described (49).
Dissociated iPSC-CMs were seeded in gelatin-coated 4-well LAB-TEK II chambers (Nalge Nunc International) and were loaded with 5 μM Fluo-4 AM or 2 μM Rhod-2 AM (for cells expressing GFP) and 0.02 % Pluronic F-127 (all from Molecular Probes) in the Tyrodes solution (140 mM NaCl, 5.4 mM KCl, 1 mM MgCl2, 10 mM glucose, 1.8 mM CaCl2, and 10 mM HEPES pH 7.4 with NaOH at 25°C) for 15 min at 37°C. Cells were then washed three times with the Tyrodes solution. Ca2+ imaging was conducted with a confocal microscope (Carl Zeiss, LSM 510 Meta) with a 63x lens (NA=1.4) using Zen software. Spontaneous Ca2+ transients were acquired at room temperature using line scan mode at a sampling rate of 1.92 ms/line. A total of 10,000 lines were acquired for 19.2 s recoding. For Measurement of caffeine-evoked Ca2+ release, caffeine (20 mM) in Ca2+ free solution (Tyrodes solution containing 5 mM EGTA instead of CaCl2) was used to evoke SR/ER Ca2+ transients in iPSC-CMs.
iPSC-CMs were seeded on glass bottom petri dishes before each experiment, and switched from culture media to warm Tyrode’s solution. Cells were maintained at 36°C for the entire experiment. Beating cells were interrogated by AFM (MFP-3D Bio, Asylum Research) using a silicon nitride cantilever (spring constants ~0.04 N/m, BudgetSensors). To measure forces, cells were gently contacted by the cantilever tip with 100 pN of force, with a typical cellular indentation of around 100–200 nm, with the cantilever tip remained in the position without Z-piezo feedback for multiple sequential two minute intervals while deflection data were collected at a sample rate of 2 kHz. Typical noise during these measurements was around 20 pN. Deflection data were converted to force by multiplying by the spring constant. Typically, 100–400 beats were collected for each single cell, and statistics were calculated for the forces, intervals between beats, and duration of each contraction. Forces across cells were compared using two tailed Student’s t-test.
First-generation type 5 recombinant adenoviruses carrying cytomegalovirus (CMV) promoter driving Serca2a plus a separate CMV promoter driving GFP (Ad.Serca2a) and adenoviruses carrying CMV promoter driving GFP only (Ad.GFP) as control were used (35). Dissociated iPSC-CMs were transduced at MOI 100 overnight and then refreshed with culture medium (DMEM supplemented with 10% FBS). Cells were used for subsequent experiments 48 hours after transduction.
Data were analyzed using either Excel or R. Statistical differences among two groups were tested using two tailed Student’s t-tests. Statistical differences among more than two groups were analyzed using one-way ANOVA tests followed by Tukey’s Multiple Comparison Test. Significant differences were determined when p value is less than 0.05.
Fig. S1. Co-segregation of the R173W mutation with the DCM patients in the family.
Fig. S2. R173W mutation in the iPSCs derived from DCM patients in the family.
Fig. S3. Characterization of iPSC lines.
Fig. S4. Quantitative-PCR of relative expression levels of total versus endogenous Yamanaka reprogramming factors.
Fig. S5. Patient-specific iPSCs can differentiate into cells from the 3 germ layers in vitro.
Fig. S6. Relative cardiac differentiation efficiency of the patient-specific iPSCs.
Fig. S7. Allele-specific PCR of wild type (Wt) and mutant (R173W) TNNT2 expression in DCM and control iPSC-CMs.
Fig. S8. Multi-electrode arrays (MEA) examining electrophysiologic properties of iPSC-derived beating EBs.
Fig. S9. iPSC-CMs expressed cardiac-specific proteins.
Fig. S10. DCM iPSC-CMs exhibited an increased heterogeneous sarcomeric organization.
Fig. S11. NE treatment significantly aggravated myofilament organization in DCM iPSC-CMs.
Fig. S12. Enhanced susceptibility of DCM iPSC-CMs to prolonged mechanical strain.
Fig. S13. Atomic force microscopy (AFM) measurement of contraction force of iPSC-CMs.
Fig. S14. Beat frequency and duration of single iPSC-CMs measured by AFM.
Fig. S15. Dot plots of relative cell size versus contraction force for each single cell measured by AFM.
Fig. S16. Ca2+ imaging of iPSC-CMs transduced with Ad.Serca2a or Ad.GFP adenoviruses with red fluorescent Ca2+ indicator Rhod-2 AM.
Fig. S17. Contractility of control iPSC-CMs transduced with Ad.Serca2a or Ad.GFP as measured by AFM.
Fig. S18. Gene expression profiling of DCM iPSC-CMs with Serca2a over-expression identified enriched pathways that may function in rescuing the DCM phenotype.
Fig. S19. Metoprolol treatment improved sarcomeric organization of DCM iPSC-CMs and alleviate the aggravation effect of NE treatment.
Fig. S20. Similar functional properties of DCM iPSC-ECs and control iPSC-ECs.
Fig. S21. Schematic of potential mechanisms by Serca2a gene therapy in DCM iPSC-CM.
Table S1. Genetic screening of the DCM gene panel by next generation sequencing (Illumina)
Table S2. Clinical characteristics of the R173W DCM family
Table S3. Spread sheet of iPSC lines used in this study and analyzed by each assay
Table S4. Baseline electrophysiological parameters of iPSC-derived beating EBs obtained via MEA Recordings
Table S5. Parameters of single DCM iPSC-CMs measured by AFM
Table S6. List of genes with greater than 2-fold expression changes in DCM iPSC-CMs over-expressed with Serca2a compared to DCM iPSC-CMs only.
Table S7. List of genes with greater than 2-fold expression changes in control iPSC-CMs over-expressed with Serca2a compared to control iPSC-CMs only.
Table S8. Selected enriched pathways for rescued genes after Serca2a over-expression in DCM iPSC-CMs
Table S9. Primers used for real time quantitative-PCR and allelic-PCR
Video S1. iPSC-derived beating EBs 1.
Video S2. iPSC-derived beating EBs 2.
Video S3. iPSC-derived beating EBs seeded on a MED64 MEA probe.
Video S4. Single beating DCM iPSC-CMs.
Video S5. Single beating control iPSC-CMs.
Video S6. A beating cluster of DCM iPSC-CMs.
Video S7. Beating clusters of control iPSC-CMs.
Video S8. Tracking morphological, inotropic, and chronotropic changes of a beating cluster of control iPSC-CMs over time after NE stimulation, day 0.
Video S9. Tracking morphological, inotropic, and chronotropic changes of a beating cluster of control iPSC-CMs over time after NE stimulation, day 2.
Video S10. Tracking morphological, inotropic, and chronotropic changes of a beating cluster of control iPSC-CMs over time after NE stimulation, day 4.
Video S11. Tracking morphological, inotropic, and chronotropic changes of a beating cluster of control iPSC-CMs over time after NE stimulation, day 6.
Video S12. Tracking morphological, inotropic, and chronotropic changes of a beating cluster of DCM iPSC-CMs over time after NE stimulation, day 0.
Video S13. Tracking morphological, inotropic, and chronotropic changes of a beating cluster of DCM iPSC-CMs over time after NE stimulation, day 2.
Video S14. Tracking morphological, inotropic, and chronotropic changes of a beating cluster of DCM iPSC-CMs over time after NE stimulation, day 4.
Video S15. Tracking morphological, inotropic, and chronotropic changes of a beating cluster of DCM iPSC-CMs over time after NE stimulation, day 6.
Video S16. Probing a single beating iPSC-CMs using AFM.
Video S17. A fluorescent single beating DCM iPSC-CM transduced with Ad.Serca2a.
Video S18. Phase contrast video of the same cell shown in Video S17.
Video S19. A fluorescent single beating DCM iPSC-CM transduced with Ad.GFP.
Video S20. Phase contrast video of the same cell shown in Video S19.
We thank Dr. Shinji Komazaki at the Department of Anatomy, Saitama Medical University for his advice on transmission electron microscopic analyses, and Dr. Beth Pruitt at the Department of Biochemistry, Stanford University for her help with stretch experiments.
Funding: Supported by NIH New Innovator Award DP2OD004437, RC1 AG036142, R33 HL093172, CIRM RB3-05129 (JCW), RC1 HL100490 (MTL, JCW), U01 HL099776 (RCR), P01GM099130 (MPS), Swiss National Science Foundation PBBEP3_129803 (VSF), AHA WSA Postdoctoral Fellowship 10POST3730079 (SH), AHA WSA Postdoctoral Fellowship 10POST3870063 (MY), and the Oak Foundation Cardiovascular Fellowship (NS).
Author contributions: N.S. performed reprogramming, established iPSCs, characterizations, differentiation, MEA assays, Ca2+ imaging, project planning, experimental design, data analysis, and preparation of manuscript; M.Y. performed patch clamping, Ca2+ imaging, experimental design, data analysis, and preparation of manuscript; E.G.N. performed MEA assays, data analysis, and preparation of manuscript; V.S-F., S.H., L.W., L.H., R.C., A.L., and O.J.A. performed experimental work; J.L. performed AFM; R.E.D, A.P., and M.B. analyzed data and prepared manuscript; E.A.A, R.J.H., M.T.L., R.C.R. and J.C.W. designed experiments and prepared manuscript.
Competing interests: The authors have no competing interests.
Accession numbers: The microarray data can be found at GEO with accession number GSE35108.