Cardiomyocytes Obtained From Induced Pluripotent Stem Cells With Long-QT Syndrome 3 Recapitulate Typical Disease-Specific Features In VitroNovelty and Significance
Rationale: Current approaches for the investigation of long-QT syndromes (LQTS) are mainly focused on identification of the mutation and its characterization in heterologous expression systems. However, it would be extremely helpful to be able to characterize the pathophysiological effects of mutations and to screen drugs in cardiomyocytes.
Objective: The aim of this study was to establish as a proof of principle the disease-specific cardiomyocytes from a mouse model with LQTS 3 by use of induced pluripotent stem (iPS) cells and to demonstrate that the mutant cardiomyocytes display the characteristic pathophysiological features in vitro.
Methods and Results: We generated disease-specific iPS cells from a mouse model with a human mutation of the cardiac Na+ channel that causes LQTS 3. The control and LQTS 3–specific iPS cell lines were pluripotent and could be differentiated into spontaneously beating cardiomyocytes. Patch-clamp measurements of LQTS 3–specific cardiomyocytes showed the biophysical effects of the mutation on the Na+ current, with faster recovery from inactivation and larger late currents than observed in controls. Moreover, LQTS 3–specific cardiomyocytes had prolonged action potential durations and early afterdepolarizations at low pacing rates, both of which are classic features of the LQTS 3 mutation.
Conclusions: We demonstrate that disease-specific iPS cell–derived cardiomyocytes from an LQTS 3 mouse model with a human mutation recapitulate the typical pathophysiological phenotype in vitro. Thus, this method is a powerful tool to investigate disease mechanisms in vitro and to perform patient-specific drug screening.
Long-QT syndrome (LQTS) is a severe disorder of the electric activity of the heart. It is caused by delayed repolarization of cardiomyocytes, which leads to abnormally long action potential (AP) durations that result in a prolonged QT interval in the ECG.1 The lifespan of patients with LQTS is often limited because of the development of ventricular tachycardia and sudden cardiac death.1,2 Generally, LQTS are caused by loss-of-function mutations, but they can also be caused by gain-of-function mutations, the most common being LQTS 3. The most frequent LQTS 3 mutation is the deletion of the amino acids lysine-proline-glutamine (ΔKPQ) in the intracellular loop between domains III and IV of the cardiac Na+ channel.1,3 This results in reactivation of Na+ channels during the late phase of the AP, which leads to prolongation of the AP duration and QT interval, as well as dangerous early afterdepolarizations (EADs) at slow heart rates.3 Because of the strong frequency dependence of this effect, patients with LQTS 3 have lethal arrhythmias preferentially at rest and during sleep.1,2
The common pathophysiological feature of LQTS is the prolonged AP duration in cardiomyocytes. This cannot be investigated directly in vitro because sufficient numbers of human ventricular cardiomyocytes cannot be harvested from patients. Therefore, the properties of mutated ion channels are preferentially investigated in heterologous expression systems that lack the typical cell biological and physiological features of cardiomyocytes.2,4
Induced pluripotent stem (iPS) cells can be propagated to an unlimited extent and have been shown to serve as a source of cardiomyocytes in vitro.5 Moreover, it has been shown recently that iPS cells derived from patients with LQTS 1 or 2 can be used to obtain disease-specific cardiomyocytes.6,7 In the present study, we provide evidence that cardiomyocytes differentiated from disease-specific iPS cells do recapitulate the typical frequency-dependent features of the LQTS 3 phenotype in vitro.
An expanded Methods section is available in the Online Data Supplement at http://circres.ahajournals.org.
iPS cells were generated by retroviral transduction of murine embryonic fibroblasts (MEFs) from Scn5aΔ/+ mice (mice heterozygous for the KPQ deletion of Scn5a) and wild-type littermates either with the 3 factors Oct4, Sox2, and Klf4 or additionally with the fourth factor c-Myc, as described previously.8,9 Teratoma assay was performed by injection of undifferentiated iPS cells into SCID mice. For in vitro differentiation of iPS cells, the hanging drop method was used to generate embryoid bodies as described previously.10 Immunostainings were performed according to standard protocols,10 and TaqMan assays (Applied Biosystems, Foster City, CA) were used for quantitative polymerase chain reaction.
Na+ currents were recorded and biophysically characterized from differentiated cardiomyocytes by use of the patch-clamp technique as described previously.3,4,10,11 For recording of frequency-dependent AP durations, APs were evoked at various pacing periods, and the slope of the AP duration at 90% of repolarization (APD90)–pacing period relationship was analyzed for each individual cell by a linear regression.
Generation of Disease-Specific iPS Cell Lines
iPS cells were generated from MEFs derived from wild-type and Scn5aΔ/+ littermates. For this purpose, MEFs were prepared from single embryos, genotyped (Figure 1B), and infected with retroviruses to express specific “stemness” factors.8 Reprogramming efficiency was enhanced by the addition of extracellular signal-regulated kinase and glycogen synthase kinase 3 inhibitors.9 Embryonic stem (ES) cell–like colonies appeared after 14 days and were selected on the basis of their ES cell–like morphology. Wild-type and Scn5aΔ/+ iPS cell lines were propagated on irradiated MEF layers and retained the morphology of undifferentiated ES cells (Figure 1A). Both wild-type and Scn5aΔ/+ iPS cells expressed the stem cell markers Oct4 and SSEA1 (Figure 1C). As expected, genotyping of iPS cell lines revealed the mutation in the Scn5aΔ/+ but not in the wild-type iPS cells (Figure 1B).
Pluripotency of Wild-Type and Scn5aΔ/+ iPS Cells
The pluripotency of established iPS cell lines was proven by in vivo teratoma formation and in vitro differentiation. After injection of wild-type and Scn5aΔ/+ iPS cells into SCID mice, teratoma developed with tissues from all 3 germ layers (Figure 1D). We also analyzed the in vitro differentiation characteristics and stained embryoid bodies at day 10 for markers of cells from the 3 different germ layers. Both wild-type and Scn5aΔ/+ iPS cell lines showed Troma-1–positive cells, which indicates endodermal differentiation (Figure 2A). Mesodermal differentiation (Figure 2A) was proven on the basis of platelet endothelial cell adhesion molecule-1–positive endothelial cells and α-actinin–positive cardiomyocytes. The presence of nestin- and βIII-tubulin–positive cells highlighted ectodermal differentiation (Figure 2A). Wild-type and Scn5aΔ/+ embryoid bodies showed typical spontaneous beating areas starting around day 7 of differentiation (supplementary Videos I and II). Relative gene expression analysis by quantitative polymerase chain reaction revealed a similar expression of stem cell markers (Oct4 and Nanog) in wild-type and Scn5aΔ/+ iPS cells that decreased on differentiation (Figure 2B). Expression of the cardiac-specific markers α-myosin heavy chain (Myh6), Nav1.5 Na+ channel (Scn5a), hyperpolarization-activated cyclic nucleotide-gated channel 4 (Hcn4), Kv2.1 delayed rectifier K+ channel (Kcnb1), Cav1.2 L-type Ca2+ channel (Cacna1c), and α2-actinin (Acnt2) increased to a similar extent during differentiation of wild-type and Scn5aΔ/+ iPS cells (Figure 2C). Interestingly, expression of the Na+ channel (Scn5a) further increased with ongoing differentiation from day 9 to day 14.
Biophysical Characterization of Na+ Currents in Disease-Specific and Wild-Type Cardiomyocytes From iPS Cells
To measure Na+ currents and APs in single wild-type and Scn5aΔ/+ iPS cell-derived cardiomyocytes, single cells were isolated from beating areas of embryoid bodies and investigated by use of the patch-clamp technique. Cardiomyocytes from both lines were spontaneously beating and showed similar Na+ channel distribution and well-organized sarcomeric structures with Scn5a and α-actinin staining (Figure 3A).
Because LQTS 3 is caused by mutated Na+ channels, we investigated the functional expression of the voltage-dependent Na+ current at early (day 12) and late (late developmental stage, days 19 to 22) stages of differentiation using voltage ramps (data not shown). The percentage of cells with Na+ currents was lower in the early developmental stage (66.6% of wild-type [n=12] and 66.6% of Scn5aΔ/+ [n=15] cells) than in the late developmental stage (92.8% of wild-type [n=14] and 76.4% of Scn5aΔ/+ [n=17] cells). This is in line with gene expression data of the Na+ channel (Figure 2C), and therefore, only cardiomyocytes from the late developmental stage were investigated further. The capacitance of these cells was very similar (Table 1), which excludes the possibility that functional differences were related to cell size. Na+ current density at −10 mV was similar in wild-type and Scn5aΔ/+ cells (Figures 3B and 3C; Table 1). We also did not observe differences between genotypes in the steady state activation and inactivation of the Na+ current (Figure 3D) or in the potential for half-maximal activation and inactivation (Figure 3E; Table 1). The time constants for the kinetics of deactivation did not differ at −40 mV and at 0 mV between wild-type and Scn5aΔ/+ cardiomyocytes (Table 1). However, recovery from inactivation of the Na+ current was significantly faster in Scn5aΔ/+ than in wild-type cells (Figures 3F and 3G; Table 1). Similarly, the tetrodotoxin-sensitive late Na+ current was significantly larger in Scn5aΔ/+ than in wild-type iPS cell–derived cardiomyocytes (Figures 3H and 3I; Table 1). These data revealed that the classic biophysical features of the mutated Scn5a Na+ current could be observed in iPS cell–derived late-stage cardiomyocytes.
Characterization of APs From Wild-Type and Scn5aΔ/+ iPS Cell–Derived Cardiomyocytes
Because prolonged duration of APs, especially at low frequencies, is the typical hallmark of LQTS 3, we evoked APs in the current clamp mode and focused primarily on AP duration at different pacing frequencies. Amplitude, upstroke velocity, and resting membrane potential were similar in both genotypes (Table 2); however, clear differences were observed when we analyzed frequency-dependent AP durations. Wild-type cardiomyocytes had similar AP durations at fast and slow pacing rates (Figure 4A). In contrast, Scn5aΔ/+ cardiomyocytes showed prolonged AP durations when paced at slower rates (Figure 4B).
Because of the high variability of AP duration at APD90 between individual cells, only a tendency toward prolongation of averaged APD90 was noted in Scn5aΔ/+ cells (Figure 4C). Because of the cell-to-cell variability, we quantified the dependency of AP duration of pacing periods (APD90–pacing period) in individual cardiomyocytes using linear regression analysis (examples in Figure 4D). In wild-type cells, APD90 was hardly affected by pacing frequency, therefore yielding a flat slope in the APD90–pacing period relationship (Figure 4E; Table 2). In contrast, this was significantly different in Scn5aΔ/+ cardiomyocytes, in which the slope was steeply positive (Figure 4E; Table 2). In addition, approximately half of the Scn5aΔ/+ cardiomyocytes developed EADs at low pacing rates (Figures 4F and 4G), which were not observed in any of the wild-type cells (Figures 4F and 4G).
The aim of the present study was to investigate whether reprogramming of fibroblasts harvested from a representative LQTS 3 model enabled reproduction of the characteristic electrophysiological features of the disease in vitro. For this purpose, we used a well-established LQTS 3 mouse model carrying the human ΔKPQ mutation of the Na+ channel (Scn5aΔ/+) that is known to display typical electrophysiological features of LQTS 3, including specific frequency-dependent changes of the AP duration.3,12
We demonstrated that iPS cells can be generated from wild-type and Scn5aΔ/+ MEFs. The pluripotency of iPS clones was confirmed by gene expression analysis, teratoma formation, and in vitro differentiation assays. Most importantly, wild-type and Scn5aΔ/+ iPS cells could be differentiated into functional intact cardiomyocytes. We investigated different stages of iPS cell differentiation and found that compared with cells in the early developmental stage, late developmental stage cells have higher Scn5a gene expression, and a larger percentage of late developmental stage cardiomyocytes expressed functional Na+ currents, which is similar to ES cell–derived cardiomyocytes.11 Therefore, late developmental stage spontaneously beating single cardiomyocytes were analyzed, and the biophysical properties between wild-type and Scn5aΔ/+ cardiomyocytes were compared. Cell capacitance, Scn5a channel distribution, and Na+ peak current densities were similar between the 2 genotypes; however, in Scn5aΔ/+ cardiomyocytes, Na+ currents had faster recovery from inactivation and larger amplitudes of the late component of the Na+ current. These biophysical properties are typical of the ΔKPQ mutation and have been described previously in heterologous expression systems4 and in cardiomyocytes from the LQTS 3 mouse model.3 Other biophysical features of the Na+ current, such as steady state activation and inactivation, were unaltered, which is in accordance with previous reports on the ΔKPQ mutation.3,4 When measuring APs, we observed a tendency toward longer APD90 in Scn5aΔ/+ cardiomyocytes, but this did not reach statistical significance because of high intercellular variability. This variability is most likely because in contrast to investigations of the adult heart in previous studies,3 the developmental stages of iPS and ES cell–derived cardiomyocytes is not identical between individual cells. We therefore determined the APD90 at different pacing rates in the same cell and found that all Scn5aΔ/+ iPS-derived cardiomyocytes had prolonged APD90 at lower stimulation rates and a high incidence of EADs. Thus, cardiomyocytes with the ΔKPQ mutation were characterized by a steep APD90–pacing period ratio, and this clearly differed from iPS cell–derived wild-type cardiomyocytes. The steep positive frequency dependence of AP or QT duration is, in contrast to LQTS 1 and 2, a phenotypic hallmark of LQTS 3 and has been reported in mice3 and patients.12
We were able to analyze APD90 in LQTS3-specific iPS cell–derived cardiomyocytes within a broad range (0.5 to 6 seconds) of pacing periods. Although these were substantially lower rates than observed in adult mice, the large pacing range enabled investigations of the frequency-dependent adverse effects of ion channel mutations and of novel compounds, in particular with the use-dependent block of ion channels.
The generation of disease-specific cardiomyocytes from iPS cells represents an important step, because human cardiomyocytes cannot be harvested and expanded in sufficient numbers from patients. Because of this limitation, previous investigations of LQTS used heterologous expression in noncardiomyocytes to unravel the consequences of a mutation on the biophysical properties of the affected ion channel. However, because these cells lack the potential for AP generation, such systems can only predict the possible consequences related to AP shape and duration in cardiomyocytes.4,13 In addition, heterologous systems do not reflect the influence of potential compensatory pathways, the impact of known and unknown accessory channel subunits, and the variability of cell-specific trafficking aspects, all effects that may be present in functional cardiomyocytes. Although mouse models displaying human LQTS have been generated with success and recapitulate at least in part the phenotype in cardiomyocytes,2,3 they are of limited relevance because of striking differences in heart rate2 and because they cannot be used to generate unlimited amount of cardiomyocytes for high-throughput screenings in vitro.
Recently, LQTS 1 and 2 disease–specific cardiomyocytes were obtained from reprogrammed fibroblasts of patients.6,7 Here, we provide evidence that LQTS 3 disease–specific iPS cells can also be generated and differentiated into cardiomyocytes that maintain the functional hallmarks of the mutation in vitro. This outcome could not be predicted by previous studies, because iPS-derived cardiomyocytes are not terminally differentiated. At the early developmental stage, ES cell–derived cardiomyocytes have primarily Ca2+-driven APs, and only at later stages of differentiation does the functional Na+ current density increase and the AP become Na+ dependent.11 In addition, beating frequencies of iPS- and ES-derived cardiomyocytes are similar to fetal hearts, which are characterized by much lower beating rates than adult cardiomyocytes. It was therefore unclear whether the functional phenotype of the ΔKPQ mutation could be reproduced despite these obvious physiological differences between adult mouse and iPS cardiomyocytes. The present data clearly demonstrate that the pathognomonic functional features of the ΔKPQ mutation, namely, the long AP duration at slower pacing rates and the occurrence of EADs, are identified in LQTS 3–specific iPS-derived cardiomyocytes. In future studies, it should be possible to measure these effects with a high-throughput patch clamp or extracellular field potential recordings. Such approaches could enable pharmacological screening in vitro on cardiomyocytes, sparing animal experiments. Moreover, it is very likely that the use of iPS cells will allow functional analysis and patient-specific drug development even in patients in whom the specific mutation underlying the disease has not yet been identified. The feasibility of generating iPS cells from skin biopsy samples, keratinocytes, or somatic cell types of humans will enable the establishment of a wide range of relevant human disease models in the culture dish.
Sources of Funding
This work was supported by a grant from the German Ministry of Education and Research (01GN0813) to P.S. and B.K.F.
We thank P. Carmeliet, Flanders Interuniversity Institute for Biotechnology (VIB), Leuven, Belgium, for providing the Scn5aΔ/+ mouse line; H.R. Schoeler and H. Zaehres, Max Planck Institute for Molecular Biology, Muenster, Germany, for the help with the iPS technology; and R. Schneider, Rhenish-Westphalian Technical University (RWTH Aachen), Aachen, Germany, for help with the teratoma analysis.
In June 2011, the average time from submission to first decision for all original research papers submitted to Circulation Research was 14.48 days.
Non-standard Abbreviations and Acronyms
- action potential
- action potential duration at 90% of repolarization
- early afterdepolarization
- embryoid body
- ES cells
- embryonic stem cells
- iPS cells
- induced pluripotent stem cells
- murine embryonic fibroblast
- mice heterozygous for KPQ deletion of Scn5a
- Received July 13, 2011.
- Revision received February 23, 2011.
- Revision received July 13, 2011.
- Accepted July 14, 2011.
- © 2011 American Heart Association, Inc.
- Nuyens D,
- Stengl M,
- Dugarmaa S,
- Rossenbacker T,
- Compernolle V,
- Rudy Y,
- Smits JF,
- Flameng W,
- Clancy CE,
- Moons L,
- Vos MA,
- Dewerchin M,
- Benndorf K,
- Collen D,
- Carmeliet E,
- Carmeliet P
- Mauritz C,
- Schwanke K,
- Reppel M,
- Neef S,
- Katsirntaki K,
- Maier LS,
- Nguemo F,
- Menke S,
- Haustein M,
- Hescheler J,
- Hasenfuss G,
- Martin U
- Kolossov E,
- Lu Z,
- Drobinskaya I,
- Gassanov N,
- Duan Y,
- Sauer H,
- Manzke O,
- Bloch W,
- Bohlen H,
- Hescheler J,
- Fleischmann BK
- Maltsev VA,
- Wobus AM,
- Rohwedel J,
- Bader M,
- Hescheler J
- Schwartz PJ,
- Priori SG,
- Locati EH,
- Napolitano C,
- Cantu F,
- Towbin JA,
- Keating MT,
- Hammoude H,
- Brown AM,
- Chen LS
Novelty and Significance
What Is Known?
The pathophysiological consequences of ion channel mutations that cause long-QT syndrome (LQTS) cannot be analyzed directly in cardiomyocytes from patients.
Induced pluripotent stem (iPS) cells can be generated from skin biopsy samples of patients and differentiated into cardiomyocytes.
What New Information Does This Article Contribute?
Disease-specific iPS cells can be generated from murine fibroblasts that carry a human mutation of the Na+ channel that causes LQTS 3.
Cardiomyocytes can be differentiated in the culture dish from LQTS 3–specific iPS cells and show the known biophysical features of the cardiac Na+ channel mutation.
Action potential durations of LQTS 3 cardiomyocytes were found to be prolonged at slow heart rates, which is the pathognomonic feature of LQTS 3.
LQTS are characterized by severe, potentially life-threatening cardiac arrhythmias and are caused by ion channel mutations that lead to prolongation of the cardiac action potential duration. The pathophysiological features of LQTS have been investigated by identifying the underlying mutation, analyzing mutated ion channels in nonexcitable cells, and predicting the functional consequences on action potential duration in cardiomyocytes. A much more promising approach would be the direct analysis of cardiomyocytes from LQTS patients. Because it is technically not feasible to harvest sufficient numbers of cells from cardiac biopsy samples, we suggest an alternative approach using cardiomyocytes differentiated from iPS cells that can be obtained from skin biopsy samples by reprogramming. We demonstrate as a proof-of-principle that iPS cell–derived cardiomyocytes can be harvested from a mouse model with a human Na+ channel mutation that causes LQTS 3. Most importantly, for the first time, we analyzed action potentials of LQTS 3–specific cardiomyocytes from iPS cells and identified prolonged action potential durations and early afterdepolarizations at low pacing rates, which is the typical feature of LQTS 3. We suggest that cardiomyocytes differentiated from disease-specific iPS cells are a powerful tool for the investigation of disease mechanisms in vitro and for patient-specific drug screenings.