From a paper presented at the International Symposium on Embryonic Stem Cells – Prospects for Human Health, at the University of Sheffield, UK, 10 September 2001.
Cardiomyocyte differentiation of mouse and human embryonic stem cells*
Article first published online: 25 APR 2002
Journal of Anatomy
Volume 200, Issue 3, pages 233–242, March 2002
How to Cite
Mummery, C., Ward, D., Van Den Brink, C. E., Bird, S. D., Doevendans, P. A., Opthof, T., De La Riviere, A. B., Tertoolen, L., Van Der Heyden, M. and Pera, M. (2002), Cardiomyocyte differentiation of mouse and human embryonic stem cells. Journal of Anatomy, 200: 233–242. doi: 10.1046/j.1469-7580.2002.00031.x
- Issue published online: 25 APR 2002
- Article first published online: 25 APR 2002
- Accepted for publication 18 January 2002
- cardiomyocyte markers;
Ischaemic heart disease is the leading cause of morbidity and mortality in the western world. Cardiac ischaemia caused by oxygen deprivation and subsequent oxygen reperfusion initiates irreversible cell damage, eventually leading to widespread cell death and loss of function. Strategies to regenerate damaged cardiac tissue by cardiomyocyte transplantation may prevent or limit post-infarction cardiac failure. We are searching for methods for inducing pluripotent stem cells to differentiate into transplantable cardiomyocytes. We have already shown that an endoderm-like cell line induced the differentiation of embryonal carcinoma cells into immature cardiomyoctyes. Preliminary results show that human and mouse embryonic stem cells respond in a similar manner. This study presents initial characterization of these cardiomyocytes and the mouse myocardial infarction model in which we will test their ability to restore cardiac function.
Pluripotent stem cells from various sources could be used to replace specialized cells lost or malfunctioning as a result of disease. Transplantation of cardiomyocytes could thus be a treatment for cardiac failure. However, much basic research will be required before cell transplantation therapy is transformed from a media item into serious clinical practice. For example, for most potential clinical applications, it is entirely unclear what the most suitable source of pluripotent stem cells for the derivation of transplantable cells might be. Human embryonic stem cells are one option but it will be necessary to generate large numbers of functionally mature cells either in vitro or after transplantation in vivo for cardiac function (for example) to be restored.
The first step is to control differentiation. Molecular pathways that lead to specification and terminal differentiation of cardiomyocytes from embryonic mesoderm during development are not entirely clear but data largely derived from amphibian and chick, and more recently from mouse, have suggested that signals emanating from endoderm in the early embryo may be involved in both processes. Tissue recombination experiments have shown that, for example, in chick, primitive hypoblast (endoderm) induces cardiogenesis in posterior epiblast (ectoderm) (Yatskievych et al. 1997), while in Xenopus, endoderm and Spemann organizer synergistically induce cardiogenesis in embryonic mesoderm undergoing erythropoiesis (Nascone & Mercola, 1995). In addition, the zebrafish mutant casanova, which lacks endoderm, also exhibits severe heart anomalies.
We have previously isolated various cell lines by cloning from a culture of P19 EC cells treated as aggregates in suspension (embryoid bodies) with retinoic acid, before replating (Mummery et al. 1985). One of these cell lines, END-2, has characteristics of visceral endoderm (VE), expressing alpha fetoprotein and the cytoskeletal protein ENDO-A. When undifferentiated P19 EC cells are plated on to a confluent monolayer of these END-2 cells, they aggregate spontaneously and within a week differentiate to cultures containing areas of beating muscle at high frequency (Mummery et al. 1991). This was not observed in co-cultures with other differentiated clonal cell lines from P19 EC that did not express characteristics of VE. Differentiation to beating muscle, however, was observed when aggregates of P19 EC cells were grown in conditioned medium from END-2 cells, although not in the absence of conditioned medium. This effect was inhibited by activin A (van den Eijnden-van Raaij et al. 1991). More recently, we have observed similar effects on mouse embryonic stem (mES) cells, as reported previously by Rohwedel et al. (1994) using the same END-2 cells. In addition, Dyer et al. (2001) have shown that END-2 cells also induce the differentiation of epiblast cells from the mouse embryo to undergo hematopoiesis and vasculogenesis and respecify prospective neurectodermal cell fate, an effect they show to be largely attributable to Indian hedgehog (Ihh), a factor secreted by END-2 cells and VE from the mouse embryo.
Here we present preliminary results showing similar effects in hES cells (Reubinoff et al. 2000) co-cultured with END-2 cells, including the appearance of beating muscle. By contrast, in a pluripotent hEC cell line, GCT27X (Pera et al. 1989), aggregation takes place in the co-culture, but we found no evidence of beating muscle. Characterization of the hES-derived cardiomyocytes has been initiated. Although the spontaneous differentiation to cardiomyocytes of a subclone of one other hES cell line in embryoid bodies has also recently been reported (Kehat et al. 2001), we believe that the present study is the first description of the induction of cardiomyocyte differentiation in hES cells. The myocardial infarction (MI) model in mice, in which we will test cardiomyocyte function after transplantation, is described.
Materials and methods
END-2 cells, P19 EC, hEC and hES cells were cultured as described previously (Mummery et al. 1985, 1991; van den Eijnden-van Raaij et al. 1991; Slager et al. 1993; Reubinoff et al. 2000). The hES2 cell line from ESI (Reubinoff et al. 2000) was used in all experiments. To initiate co-cultures, mitogenically inactive END-2 cell cultures, treated for 1 h with mitomycin C (10 µg mL−1), as described previously (Mummery et al. 1991), replaced mouse embryonic fibroblasts (MEFs) as feeders for hEC, mES and hES. Co-cultures with P19EC, which are feeder-independent, were initiated and maintained as described previously (Mummery et al. 1991). Cultures were then grown for 2–3 weeks and scored for the presence of areas of beating muscle from 5 days onwards.
Isolation of primary human adult cardiac cells
Human atrial cells from surgical biopsies served as a control for antibody staining, electrophysiology and characterization of ion channels by RT-PCR. Cardiac tissue was obtained with consent from patients undergoing cardiac surgery. Atrial appendages routinely removed during surgery were immediately transferred to ice cold Krebs–Ringer (KR) saline solution. Tissues were trimmed of excess connective and adipose tissue and washed twice with sterile KR solution. Myocardial tissue was minced with sterile scissors, then dissociated to release individual cells by a three-step enzymatic isolation procedure using published methods (Peeters et al. 1995). The first step involved a 15-min incubation with 4.0 U mL−1 protease type XXIV (Sigma, St Louis, MO, USA) at 37 °C. Tissues were then transferred to a solution consisting of collagenase 1.0 mg mL−1 and hyaluronidase 0.5 mg mL−1, followed by three further incubations with collagenase (1.0 mg mL−1) for 20 min each at 37 °C. Tissue extracts were combined and the calcium concentration restored to 1.79 mmol L−1. Cardiomyocytes were transferred to tissue culture medium M199 enriched with 10% FBS, penicillin (100 U mL−1)/streptomyocin (100 µg mL−1), 2.0 mmol L−1 L-carnitine, 5.0 mmol L−1 creatine, 5.0 mmol L−1 taurine and seeded directly on to glass cover-slips coated with 50 µg mL−1 poly L-lysine and cultured overnight.
Attached primary cardiomyocytes, mES (E14 and R1) and hES-derived cardiomyocytes were fixed with 3.0% paraformaldehyde in PBS with Ca2+ and Mg2+ for 30 min at room temperature, then permeabilized with 0.1% triton X 100 in PBS for 4 min. Immunocytochemistry was performed by standard methods using monoclonal antibodies directed against sarcomeric proteins including α-actinin and tropomyosin (Sigma). Antibodies specific for isoforms of myosin light chain (MLC2a/2v) were used to distinguish between atrial and ventricular cells (gift of Dr Ken Chien) (Table 1). Secondary antibodies were from Jackson Immunoresearch Laboratories. Cultured cardiac fibroblasts served as a negative control for sarcomeric proteins and cells were visualized using a Zeiss Axiovert 135M epifluorescence microscope (Carl Zeiss, Jena GmbH, Germany). Images were pseudocoloured using image processing software.
|Primary antibody||Dilution||Secondary antibody||Dilution|
|Mouse anti-α-actinin IgG||1 : 800||Goat anti-mouse IgG-cy3/FITC conjugated||1 : 250|
|Mouse anti-tropomyosin IgG||1 : 50||Goat anti-mouse IgG-cy3 conjugated||1 : 250|
|Polyclonal rabbit anti-mouse mlc-2a (atrial)||1 : 500||Goat anti-rabbit IgG-cy3 conjugated||1 : 250|
|Hoechst (nucleic acid)||1 : 500|
Semi-quantitative RT-PCR for ion channel expression
P19EC cells were differentiated into beating muscle by the aggregation protocol in the presence of 1% dimethyl sulphoxide (Rudnicki & McBurney, 1987). After 16 days in these culture conditions, beating areas were excised and RNA was isolated using Trizol (Gibco) and reversed transcribed using M-MLV-RT (Gibco). Primers for cardiac actin (Lanson et al. 1992), MLC2v (Meyer et al. 2000), ERG (Lees-Miller et al. 1997) and Kir2.1 (Vandorpe et al. 1998), were used as described previously. Primers for mouse l-type calcium channel subunit α1c (sense 5′-CCAGATGAGACCCGCAGCGTAA; antisense 5′-TGTCTGCGGCGTTCTCCATCTC; GenBank accession no. L01776; product size 745 bp), Scn5a (sense 5′-CTTGGCCAAGATCAACCTGCTCT; antisense 5′-CGGACAGGGCCAAATACTCAATG; AJ271477; 770 bp) and β-tubulin (sense 5′-TCACTGTGCCTGAACTTACC; antisense 5′-GGAACATAGCCGTAAACTGC; X04663; 319 bp) were designed using VectorNTI software (InforMax, North Bethesda, MD, USA).
Experiments were performed at 33 °C, using the whole cell voltage clamp configuration of the patch-clamp technique. After establishment of the gigaseal the action potentials were measured in the current clamp mode. The data were recorded from cells in spontaneously beating areas using an Axopatch 200B amplifier (Axon Instruments Inc., Foster City, CA, USA). Output signals were digitized at 2 kHz using a Pentium III equipped with an AD/DAC LAB PC+ acquisition board (National Instruments, Austin, TX, USA). Patch pipettes with a resistance between 2 and 4 MΩ were used. Composition of the bathing medium was 140 mm NaCl, 5 mm KCl, 2 mm CaCl2, 10 mm HEPES, adjusted to pH 7.45 with NaOH. Pipette composition: 145 mm KCl, 5 mm NaCl, 2 mm CaCl2, 10 mm EGTA, 2 mm MgCl2, 10 mm HEPES, adjusted to pH 7.30 with KOH.
Myocardial infarction model in mice
In order to test the ability of stem-cell-derived cardiomyocytes to restore cardiac function, a MI model is being developed in mice. In pentobarbital anaesthetized adult mice, the chest is opened through a midsternal approach. The anterior descending branch is identified and ligated. Successful procedures induce a discoloration of the distal myocardium. The chest is closed with three sutures and the animal is allowed to recover. In total, 17 animals have been operated on to date. Seven received a sham procedure including positioning of the suture and 10 were ligated. Four weeks after MI the mice were anaesthetized again using the same medication by intraperitoneal injection. For the haemodynamic study the animals were intubated, and connected to a rodent respirator (Hugo Sachs Electronics, March – Hugstetten Germany). Instrumentation was performed with the chest closed by introducing a catheter into the jugular vein. A 1.4 French conductance-micromanometer (Millar Instruments, Houston, TX, USA) was delivered to the left ventricle through the carotid artery. Pressure and conductance measurements were recorded using Sigma SA electronic equipment (CDLeycom, Zoetermeer, the Netherlands) and stored for offline analysis. A typical pressure volume (PV) loop recorded in a normal heart is presented in Fig. 5(a). From the PV-loops many haemodynamic parameters can be deduced including the end-systolic PV relationship (ESPVR) and preload recruitable stroke work (PRSW).
mEC – END-2 co-cultures
Two days after initiation of co-cultures with END-2 cells, P19 EC cells aggregated spontaneously and 7–10 days later many of the aggregates contained areas of beating muscle (Fig. 1a), as described previously (Mummery et al. 1991). Electrophysiology and RT-PCR showed that functional ion channels characteristic of embryonic cardiomyocytes were expressed in these cells (Fig. 2a, Table 2).
|EC||CMC||Heart||Ion channel and current|
|cardiac actin||+||+ +||+ + + + +|
|MLC2v||+||+ + +||+ + + + +|
|α1c||+||+ + +||+ + + +||l-type calcium channel, ICa|
|Scn5a||+||+||+ + + + +||Heart specific sodium channel, INa|
|ERG||+ +||+ +||+ + + +||Delayed rectifier potassium channel, IKr|
|Kir2.1||–||+||+ + + +||Voltage-gated potassium channel, IK1|
|Tubulin||+ + +||+ + +||+ + +|
mES – END-2 co-cultures
Two independent mouse ES cell lines (E14 and R1) were tested for their response to co-culture conditions. The cultures were initiated as single cell suspensions, but within 3 days large aggregates were evident for both cell lines (Fig. 1b,c). Almost simultaneously, extensive areas of spontaneously beating cardiomyocytes were evident in the R1 ES cell cultures, although only 7 days later were (smaller) areas of beating muscle found in the E14 ES cells. Cells in beating areas exhibited the characteristic sarcomeric banding pattern of myocytes when stained with α-actinin (see Fig. 4d).
hEC – and END-2 co-cultures
The human EC cell line GCT27X is a feeder-dependent, pluripotent EC cell line, with characteristics similar to human ES cells (Pera et al. 1989). In co-culture with END-2 cells, formation of large aggregates was observed (Fig. 1d). However even after 3 weeks, there was no evidence of beating muscle.
hES – END-2 co-cultures
During the first week of co-culture, the small aggragates of cells gradually spread and differentiated to cells with mixed morphology but with a relatively high proportion of epithelial-like cells. By the second week, these swelled to fluid-filled cysts (not shown). Between these, distinct patches of cells become evident which begin to beat a few days later. Between 12 and 21 days, more of these beating patches appear (e.g. Fig. 1e). Overall, 15–20% of the wells contains one or more areas of beating muscle. Beating rate is approximately 60 min−1 and is highly temperature sensitive, compared with mouse ES-derived cardiomyocytes. These cells stain positively with α-actinin, confirming their muscle phenotype (Fig. 4e). In contrast to mES and P19EC-derived cardiomyocytes, however, the sarcomeric banding patterns were poorly defined but entirely comparable with primary human cardiomycytes grown for only 2 days in culture (Figs 3 and 4a–c). It is clear that while primary human cardiomycytes initially retain the sarcomeric structure, standard culture conditions result in its rapid deterioration (Fig. 3). It may be assumed that hES culture conditions are not optimal for cardiomyocytes so that the hES-derived cardiomyocytes similarly exhibit a deterioration in their characteristic phenotype. It will be essential to optimize these conditions to obtain fully functional cardiomyocytes from stem cells in culture.
Despite deterioration in sarcomeric structure, hES-derived cardiomyocytes continued to beat rhythmically over several weeks and action potentials were detectable by current clamp electrophysiology (Fig. 2b), performed by inserting electrodes into aggregates, as shown in Fig. 2(c). However, carrying out electrophysiology in this manner, i.e. in aggregates rather than single cells, yields action potentials that are the accumulated effects of groups of cells. They are therefore difficult to interpret and to attribute to either ventricular, atrial or pacemaker cells. Work is currently in progress to dissociate and replate aggregates to allow single cell determinations.
Cardiac ion channel expression during stem cell differentiation
The order in which ion currents, responsible for the subsequent phases of the adult action potential, appear during heart development has been established in electrophysiological studies (Davies et al. 1996; Yasui et al. 2001). Inward l-type Ca2+ currents play a dominant role during early cardiac embryogenesis, whereas inward Na2+ currents increase only just before birth (An et al. 1996; Davies et al. 1996). Mouse ES and P19 EC cells display similar timing in ion current expression (Maltsev et al. 1994; Wobus et al. 1994). To unravel the sequence of ion-channel expression at the molecular level during differentiation of P19 EC cells, we performed RT-PCR on RNA isolated from undifferentiated and 16-day-old beating clusters of P19-derived cardiomyocytes and compared these with expression in adult mouse heart (Table 2). Expression of cardiac markers cardiac actin and MLC2v is detected in EC cell cultures but increases in P19-derived cardiomyocytes. At the protein level, no MLC2v is detected in EC cells, while prominent expression is found in P19-derived cardiomyocytes (data not shown). Likewise, RNA for calcium, sodium and potassium channels can be detected in EC cells. Calcium channel RNA level (α1c) is up-regulated in P19-derived cardiomyocytes, while sodium channel RNA (Scn5a) remains at initial levels. Furthermore, RNA for the delayed rectifier potassium channel (ERG) also remains unchanged. Inward rectifier Kir2.1 RNA cannot be detected in EC cells, but becomes expressed after cardiomyocyte differentiation, albeit at low levels compared to adult heart. These data are compatible with the action potential of day 16 P19 cardiomyocytes (Fig. 2a), i.e. a low upstroke velocity (ICa mainly instead of INa), a relatively positive resting membrane potential between –40 and –60 mV (little to no IK1). These results indicate that day 16 P19 cardiomyocytes resemble fetal cardiomyocytes with respect to ion channel expression, as has been described previously for mES-derived cardiomyocytes (Doevendans et al. 1998, 2000).
MI model in mice
In order to test the effectiveness of cardiomyocyte transplantation in vivo, it is important to have a reproducible animal model with a measurable parameter of cardiac function. The parameters used should clearly distinguish control and experimental animals (see for example Palmen et al. 2001) so that the effects of transplantation can be adequately determined. PV relationships are a measure of the pumping capacity of the heart and could be used as a read-out of altered cardiac function following transplantation. Here, we have tested aspects of a procedure towards establishing an MI model in immunodeficient mice as a ‘universal acceptor’ of cardiomyocytes from various sources. The infarct size obtained through occluding the anterior descending coronary artery encompassed 30–50% of the left ventricular circumference. The septum involvement can be neglected in mice, resulting in more than 94% survival after 4 weeks (one infarcted animal died). A 30% increase in left ventricular volume was recorded with conserved contractility post MI. There were significant differences in ESPVR (22.6 sham vs. 10 post MI, P < 0.05), and PRSW (81.2 sham vs. 43.5 post MI, P < 0.05). This is easily visualized by comparing the shapes of the ‘PV loops’ in sham-operated vs. infarcted mice in Fig. 5(a,b).
Despite maintained left ventricular contractile function, this mouse MI model provides a reproducible system for studying left ventricular remodelling, making it feasible to assess the extent of cardiac repair following transplant interventions.
The results of the work described here show that VE-like cells induce/promote differentiation of pluripotent cells to cardiomyocytes. These cells include pluripotent mouse EC cells, mouse ES as well as human ES cells, which are now documented for the first time to respond to inductive cues derived from cells similar to those normally adjacent to the region of heart development in the embryo. The results also showed that the capacity of different mES cell lines to differentiate to cardiomyocytes is variable. We have so far tested only one hES cell line for endoderm co-culture responses and it is not unlikely that different hES cell lines equally show variable capacities for differentiation. This remains to be tested. Kehat et al. (2001) have also showed that subclone 9.2 of H9 hES cells (Thomson et al. 1998; Amit et al. 2000) will also differentiate to cardiomyocytes. The H9.2 cells, however, form embryoid bodies when grown as aggregates in suspension, rather like mES cells but unlike hES2 cells. The significance of differences between these cell lines remains to be resolved. An interesting experiment would be to subject the H9 cell line and/or its subclones to our cardiomyocyte inductive conditions.
Pluripotent mouse stem cells differentiate to cardiomyocytes with embryonic rather than mature characteristics. It is unclear what the phenotype of hES-derived cardiomyocytes precisely is and, indeed, the phenotypic characteristics of primary human cardiomyocytes (ventricular vs. atrial, fetal vs. adult) are insufficiently detailed for comparisons to be made directly between various human sources. Mouse and rat cardiomycytes are at present the best reference tissues. We have also shown here that primary adult human cardiomyocytes have a sharply defined morphology and sarcomeric banding pattern in culture but within a few days this deteriorates, presumably because culture conditions are suboptimal for these highly sensitive cells. Likewise, hES-derived cardiomycytes express appropriate markers and display action potentials but under present conditions have a relatively poor morphology. This contrasts with mouse ES-derived cardiomyocytes, which maintain morphology under standard culture conditions. An immediate aim is then to optimize culture conditions for primary cardiomyocytes and apply these to the hES derivatives. The MI model in mice provides a means of functional analysis of cardiomyocytes after transplantation. Transfer to immunodeficient mice will make it suitable for comparing human and mouse cardiomyocytes from different sources for their ability to restore cardiac function after infarct. Once the efficiency of cardiomyocyte differentiation and culture conditions for cardiomyocytes have been improved, we will use this model to evaluate effects on cardiac function in vivo.
Part of this study (M.vdH.) is financed by NWO-MW, Embryonic Stem Cell International (S.D.B.), Netherlands Interuniversity Cardiology Institute (D.W.). hES cells (lines 1 and 2) were supplied by Embryonic Stem Cell International. We thank Daniel Lips (University of Maastricht) and Teun de Boer for Fig. 5(b) and contributions to the electrophysiology, respectively.
- 2000) Clonally derived human embryonic stem cell lines maintain pluripotency and proliferative potential for prolonged periods of culture. Dev. Biol. 200, 271–278. , , , , , , et al. (
- 1996) Developmental changes in beta-adrenergic modulation of 1-type Ca2+ channels in embryonic mouse heart. Circ. Res. 78, 371–378. , , , , , , et al. (
- 1996) Developmental changes in ionic channel activity in the embryonic murine heart. Circ. Res. 78, 15–25. , , , , , (
- 1998) Cardiovascular phenotyping in mice. Cardiovasc. Res. 39, 34–49. , , , (
- 2000) Differentiation of cardiomyocytes in floating embryoid bodies is comparable to fetal cardiomyocytes. J. Mol. Cell Cardiol. 32, 839–851. , , , , , (
- 2001) Indian hedgehog activates hematopoiesis and vasculogenesis and can respecify prospective neurectodermal cell fate in the mouse embryo. Development 128, 1717–1730. , , , , (
- 1991) Differentiation of aggregated murine P19 embryonal carcinoma cells is induced by a novel visceral endoderm-specific FGF-like factor and inhibited by activin A. Mech Dev. 33, 157–165. , , , , , , et al. (
- 2001) Human em, bryonic stem cells can differentiate into myocytes with structural and functional properties of cardiomyocytes. J. Clin. Invest. 108, 407–414. , , , , , , et al. (
- 1992) Gene expression and atrial natriuretic factor processing and secretion in cultured AT-1 cardiac myocytes. Circulation 85, 1835–1841. , , , , (
- 1997) Electrophysiological characterization of an alternatively processed ERG K+ channel in mouse and human hearts. Circ. Res. 81, 719–726. , , , (
- 1994) Cardiomyocytes differentiated in vitro from embryonic stem cells developmentally express cardiac specific genes and ionic currents. Circ. Res. 75, 233–244. , , , , (
- 2000) A fluorescent reporter gene as a marker for ventricular specification in ES-derived cardiac cells. FEBS Lett. 478, 151–158. , , , , (
- 1985) Clonal variants of differentiated P19 embryonal carcinoma cells exhibit epidermal growth factor receptor kinase activity. Dev. Biol. 109, 402–410. , , , , (
- 1991) Visceral-endoderm-like cell lines induce differentiation of murine P19 embryonal carcinoma cells. Differentiation 46, 51–60. , , , , , , et al. (
- 1995) An inductive role for the endoderm in Xenopus cardiogenesis. Development 121, 515–523. , (
- 2001) Cardiac remodeling but not cardiac function is impaired in IGF-1 deficient mice after chronic myocardial infarction. Cardiovasc. Res. 50, 516–524. , , , , (
- 1995) Method for isolation of human ventricular myocytes from single endocardial and epicardial biopsies. Am. J. Physiol. 268, H1757–H1764. , , , , , , et al. (
- 1989) Isolation and characterization of a multipotent clone of human embryonal carcinoma cells. Differentiation 42, 10–23. , , , (
- 2000) Embryonic stem cell lines from human blastocysts: somatic differentiation in vitro. Nat. Biotechnol. 18, 399–404. , , , , (
- 1994) Muscle cell differentiation of embryonic stem cells reflects myogenesis in vivo: developmentally regulated expression of myogenic determination genes and functional expression of ionic currents. Dev. Biol. 164, 87–101. , , , , , (
- 1987) Cell culture methods of induction of differentiation of embryonal carcinoma cell lines. Teratocarcinomas and Embryonic Stem Cells, a Practical Approach (ed. ), pp. 19–49. Oxford: IRL Press. , (
- 1993) Transforming growth factor-beta in the early mouse embryo. implications for the regulation of muscle formation and implantation. Dev. Genet. 14, 212–224. , , , , (
- 1998) Embryonic stem cell lines derived from human blastocysts. Science 282, 1145–1147. , , , , , , et al. (
- 1998) cDNA cloning and functional characterization of the mouse Ca2+-gated K+ channel, mIK1. J. Biol. Chem. 273, 21542–21553. , , , , , , et al. (
- 1994) Cardiomyocyte-like cells differentiated in vitro from embryonic carcinoma cells P19 are characterized by functional expression of adrenoceptor and Ca2+ channels. In Vitro Cell. Dev. Biol. 30A, 425–434. , , , (
- 2001) If current and spontaneous activity in mouse embryonic ventricular myocytes. Circ. Res. 88, 536–542. , , , , , , (
- 1997) Induction of cardiac myogenesis in avian pregastrula epiblast: the role of the hypoblast and activin. Development 124, 2561–2570. , , (