- Email: [email protected]
Embryonic stem cells differentiate in vitro into cardiomyocytes representing sinusnodal, atrial and ventricular cell types
Mechanisms of Development, 44 (1993) 41-50
41
© 1993Elsevier ScientificPublishers Ireland, Ltd. 0925-4773/93/$06.00 MOD 191
Embryonic stem cells differentiate in vitro into cardiomyocytes representing sinusnodal, atrial and ventricular cell types Victor A. Maltsev a, Jiirgen Rohwedel a, Jiirgen Hescheler b and A n n a M. Wobus a,, a Institut fiir Pflanzengenetik und Kulturpflanzenforschung, Corrensstr. 3, D-06466 Gatersleben, Germany and b Institut far Pharmakologie, Freie Universitiit Berlin, D-14195 Berlin, Germany
(Received 1 June 1993;revisionreceived 3 August1993;accepted6 August1993)
Pluripotent embryonic stem cells (ESC, ES ceils) of line D3 were differentiated in vitro via embryo-like aggregates (embryoid bodies) of defined cell number into spontaneously beating cardiomyocytes. By using RT-PCR technique, a- and /3-cardiac myosin heavy chain (MHC) genes were found to be expressed in embryoid bodies of early to terminal differentiation stages. The exclusive expression of the/3-cardiac MHC gene detected in very early differentiated embryoid bodies proved to be dependent on the number of ES cells developing in the embryoid body. Cardiomyocytes enzymatically isolated from embryoid body outgrowths at different stages of development were further characterized by immunocytological and electrophysiological techniques. All cardiomyocytes appeared to be positive in immunofluorescence assays with monoclonal antibodies against cardiac-specific a-cardiac MHC, as well as muscle-specific sarcomeric myosin heavy chain and desmin. The patch-clamp technique allowed a more detailed characterization of the in vitro differentiated cardiomyocytes which were found to represent phenotypes corresponding to sinusnode, atrium or ventricle of the heart. The cardiac cells of early differentiated stage expressed pacemaker-like action potentials similar to those described for embryonic cardiomyocytes. The action potentials of terminally differentiated cells revealed shapes, pharmacological characteristics and hormonal regulation inherent to adult sinusnodal, atrial or ventricular cells. In cardiomyocytes of intermediate differentiation state, action potentials of very long duration (0.3-1 s) were found, which may represent developmentally controlled transitions between different types of action potentials. Therefore, the presented ES cell differentiation system permits the investigation of commitment and differentiation of embryonic cells into the cardiomyogenic lineage in vitro.
Cardiomyocyte; Embryonic stem cell differentiation; Cardiac-specific gene expression; Action potential
Introduction
A considerable amount of data has been accumulated on the morphology and physiology of adult heart ceils, but only limited information is available about the origin, commitment and differentiation of the cells that will make up the myocardium (Litvin et al., 1992). In contrast to the well-described process of skeletal muscle cell development (Buckingham, 1992) the temporally and spatially controlled events of commitment and differentiation of cardiomyocytes remain to be
* Corresponding author. Tel: (+ 49)39482-5256. Fax: (+ 49)39482280.
investigated. In this respect it is relevant to emphasize that these two muscle cell types arise from different progenitor cells. Skeletal muscle cells develop from somites, but cardiac cells from anterior lateral plate mesoderm at an earlier stage of embryonic development. Recent data suggest that a very short interval exists between mesodermal commitment and phenotypic manifestation of differentiated cardiomyocytes (Han et al., 1992). One of the most important obstacles to study commitment and differentiation of cardiac myocytes is the lack of permanent cell lines to model and analyze the earliest stages of cardiomyogenesis. Physiological studies of developing heart cells have been mainly carried out on organ-cultured chick hearts (Sperelakis, 1982; Sperelakis and Shigenobu, 1974),
42 precardiac areas of blastoderm (McLean et al., 1987; Rosenquist and de Haan, 1966), and cultured mammalian heart cells of embryonic and neonatal origin (Wollenberger 1984; 1985). These studies on primary cardiac cell cultures are limited since cell properties and normal cardiogenesis apparently are disturbed during cultivation. For example, when 3-day-old chick embryonic heart cells are placed into culture, the cells do not continue to differentiate with respect to their electrical and morphological properties (Sperelakis, 1978). Differentiated ventricular cells from 16-day-old chick embryos cultured as monolayers revert back to the young embryonic state (Sperelakis and Pappano, 1983). Cultivated adult ventricular myocytes lose myofibrillar proteins and highly organized morphology (Nag et al., 1983; Claycomb and Palazzo, 1980). They retain only a few properties of the differentiated heart (e.g. lack of automaticity) during a limited culture period (Bugaisky and Zak, 1989). Up to now, only a few permanent cell lines (e.g. H9c2 from embryonic rat heart) expressing some cardiac-specific properties including L-type Ca 2+ current were characterized (Hescheler et al., 1991; Sipido and Marban, 1991). A new approach may be offered by using cardiomyocytes differentiated in vitro from pluripotent mouse embryonic stem cells (ES ceils, ESC) (Doetschmann et al., 1985; Wobus et al., 1991). These permanent cell lines derived by in vitro culturing undifferentiated cells of early embryos are capable to take part in the embryonic development in vivo after retransfer into blastocysts (e.g. Bradley et al., 1984). In vitro, ES cells have been shown to spontaneously differentiate into derivatives of all three primary germ layers, endoderm, ectoderm and mesoderm (Evans and Kaufman, 1981; Martin, 1981; Wobus et al., 1984; 1988). Our previous results with ES cell-derived cardiomyocytes (lines D3 and Bl17) revealed the functional expression of adrenoceptors, cholinoceptors and L-type Ca 2 ~ channels (Wobus et al., 1991). In the present study we used the ES cell differentiation system to investigate the cardiomyocyte differentiation in vitro with respect to expression of tissue-specific genes, proteins and the developmentally controlled functional expression of electrophysiological properties. The expression of cardiac-specific genes was demonstrated by RT-PCR analysis, the formation of cardiac-specific intracellular proteins was detected by indirect immunofluorescence, and the development of action potentials was characterized by patch-clamp technique. Our data proved that ES cell-derived cardiomyocytes expressed cardiac-specific genes and proteins. Furthermore, a more detailed characterization of cardiomyocytes differentiating into sinusnodal, atrial and ventricular cells was only possible by electrophysiological measurements of action potentials.
Results
Expression of cardiac-.V)eciJic genes i.s del'elopmenta/b" regula ted Because the differentiation capacity is clearly dependent on the number of cells differentiating in the embryoid bodies (see Smith et al., 1987: Wobus ct al., 1991), all experiments were done with a standardized differentiation protocol, i.e. cultivation of cmbryoid bodies with a defined cell number in hanging drops (Fig. 1). During different stages of cultivation single embryoid bodies were analyzed by RT-PCR wilh oligonucleotide primers specific for (v- and /3-cardiac and embryonic skeletal MHC genes (Fig. 2). The first transcripts of /3-cardiac MHC could be detected in "4 d' embryoid bodies originated from 1000 cells but not from 400 cells. At this stage beating areas in embryoid bodies originated from 400 ES cells wcrc never observed.
Embryonic stem cells cultivated on feeder layer
1 Cultivation of 400 or 1000 cells/2Olal medium in hanging drops
/ Incubation 2 days
"~J
\.a. j
"-L- j
<.It; '
77 ~ 7 7 1 Transfer of embryoid bodies and cultivation in suspension .....
Incubation for 5 days
Plating of embryoid bodies
to 24-well tissue culture plates
IIU
UUUII
Incubation up to 20 days
1.PCR analysis 2.Enzymatic dissociation of cardiomyocytes 3.1mmunofluorescence 4.Patch-clamp studies Fig. 1. S c h e m a t i c p r e s e n t a t i o n of the e x p e r i m e n t a l p r o t o c o l u s e d for ES cell d i f f e r e n t i a t i o n into c a r d i o m y o c y t e s .
43 diomyocytes could be discriminated by their cell size of 40-100 /~m, larger than stem ceils (20-30 /xm) but smaller than epithelial-like cells (150-200 /zm). Cardiomyocytes of all morphological types were used in our study. Basic electrophysiological properties seemed to be expressed independently of cellular morphology. ES cell-derived cardiomyocytes developed musclespecific sarcomeric myosin heavy chain as well as desmin intermediate filament proteins detected by monoclonal antibody (MAb) MF20 (Bader et al., 1982, Fig. 3a) and MAb anti-desmin (Fig. 3b), respectively. As already suggested from the RT-PCR experiments, the cardiac-specific a-cardiac MHC detected with MAb BA-G5 (Rudnicki et al., 1990) was found to be specific for ESC-derived cardiomyocytes (Fig. 3c-g). The immuno-positive cells were identified as contracting before fixation. Striations by sarcomeric structures were visualized by anti-MHC MAbs MF-20 and BA-G5. The sarcomeric length was estimated to be in the range of 1.9-2.1 Izm and corresponds to that of adult mammalian heart (Capogrossi and Lakato, 1989).
The first pulsating cells in embryoid body outgrowths were found at day '7', when '5 d' embryoid bodies were plated (at '5 + 2 d'). At this time all investigated embryoid bodies with beating areas showed strong expression of both, a- and /3-cardiac MHC genes. We also detected a significant lower amount of a- and /3-cardiac MHC gene transcripts in about half of embryoid bodies without beating areas (not shown). From day '7' throughout the further differentiation the MHC gene expression was independent on the number of ES cells originating the embryoid bodies. In embryoid body outgrowths from day '7 + 5' and thereafter, in addition to a- and /3-cardiac MHC transcripts, also embryonic skeletal MHC transcripts were found. Up to terminal stages of differentiation ('7 + 12 d') all three. MHC transcripts were expressed but embryonic skeletal MHC always with a lower rate.
ES cell-dericed cardiomyocytes exhibit cardiac-specific intracellular proteins Cardiomyocytes enzymatically isolated from beating areas of embryoid body outgrowths spontaneously contracted at a rate ranging between 0.5 and 5 Hz. Cardiomyocytes were distinguished morphologically from epithelioid- and fibroblastoid-like cells and stem cells and revealed cell types similar to previously described cardiomyocytes of mouse embryos (Lane et al., 1977) and of adult myocardium (Bugaisky and Zak, 1989; Eppenberger et al., 1990). Typical cardiomyocytes were spindle-, round-, as well as tri- or multiangular-shaped shortly (1-3 days) after dissociation (Fig. 3a-f). During prolonged cultivation (1-2 weeks) the ceils developed into fiat polymorphic shaped cells with multiple pseudopodia-like processes (Fig. 3g). Furthermore, car-
4d(400)r4d(1000)l M ~x
ES cells differentiate into cardiomyocytes expressing action potentials of sinusnodal, atrial and ~,entricular type All cardiomyocytes of early differentiation stage ('7 + 2 d' to '7 + 4 d') revealed the same type of spontaneous pacemaker-like action potentials synchronous to the contractions (Fig. 4a). Action potentials were generated with a frequency in the range of 1 to 5 Hz, amplitudes of 60 to 70 mV, durations of 70 to 100 ms and upstroke velocities of 1 to 10 V / s . The most negative membrane potential attained was about - 5 0 mV and the threshold for action potential upstroke of near - 3 5 mV. During further differentiation car-
7d
I 7+4dl
7+5d17+12d es
Fig. 2. Expression of a-cardiac (a),/3-cardiac (b) and embryonicskeletal (es) MHC genes amplified by RT-PCR during the development of ES cell-derivedembryoidbodies (investigatedfrom '2 d' to '7+ 12 d'). Expressionpatterns of embryoidbodies differentiated for '4 d', '7 d', '7+4 d', '7+5 d' and '7+ 12 d' are shown. '4 d' old embryoid bodies originated from 1000 cells showed expression of the/3-cardiac MHC gene, whereas those originated from 400 cells did not. No differences in the MHC gene expression pattern dependent on the cell number were observed at further developmental stages. In case of /3-cardiac MHC, in addition to processed transcripts also unprocessed transcripts were amplified, generating fragments of approximately800 bp.
~ ~
~ ~i ¸
45 2
--
+60
a OmV
4
b
iilI
-8o
b OmV
..................
c
•
I
I
100
ma
c
8
__
-80
j
d
OmV
e
--
f
+ 8 0 "[
-
OmV
.80 j
12
-] -
.80 j
control (185)
........
Bay K (197)
I - - - - ~ - - - - ~ - -
J/
~
~
,
..........
~1
I
I~radlplne (176) =
O,5S
V "7+n" clays Fig. 4. Diversity of action potentials during FS cell-derived cardiomyocyte differentiation. (a) Pacemaker-like action potentials in early differentiated cells ('7+2 d' to ' 7 + 4 d'); (b,c) Action potentials of higher amplitude and pacemaker-like action potentials in cells of intermediate stage ('7+5 d' to ' 7 + 8 d'); d-f: action potentials at terminal stage ('7+9 d' to '7+12 d'): atrial-, ventricular- and sinusnode-like action potentials. The action potentials shown in (d,e) were evoked by pulse current stimulation in current-clamp, all others were generated spontaneously.
diomyocytes appeared to diverse into ceils exhibiting different types of specialized action potentials. Cardiomyocytes of intermediate developmental stages ('7 + 5 d' to '7 + 8 d') generated action potentials with higher upstroke velocity and more negative resting potentials, as well as pacemaker-like action potentials with slower upstroke velocity and more positive resting potentials (Fig. 4b,c). The higher upstroke velocity (10-50 V/s) action potentials were 100-200 ms long and exhibited amplitudes ranging from 80 to 100 mV. The slower upstroke velocity (2-10 V/s) action potentials had a duration of about 70 ms and an amplitude of about 70 mV. During further differentiation into terminal differentiation stage ('7 + 9 d' to '7 + 12 d') higher upstroke velocity action potentials obviously developed into atrial- and ventricular-like types (Fig. 4d,e). These terminally differentiated cardiomyocytes had a stable resting potential (-74.1 + 5.1 mV, n = 47) and lost their diastolic depolarization. The action potentials elicited
Fig. 5. Pharmacological modulation of action potentials measured in ventricular-like cardiomyocytes of terminally differentiated ES cells ('7+ 12 d'). Values of the determined upstroke velocity are given in parenthesis. (a) Superfusion of tetrodotoxin (TTX, 20 #M) caused a marked slowing of the upstroke velocity. (b) Cd 2+ (100/xM) slowed the upstroke velocity and strongly reduced the plateau amplitude. (c) Isradipine (2 #M) and BayK 8644 (1 /zM) reduced and prolonged the duration of action potential, respectively. Both dihydropyridines did not affect the upstroke velocity.
by external current stimuli, were characterized by high amplitudes (124 _+13 mV, n = 46), faster upstroke velocities in the range of 130-330 V / s (232_+ 62 V/s, n = 46) and an overshoot phase. The atrial- and ventricular-like types of action potentials differed by their shape of plateau and duration, the former being triangle-shaped with durations of 50-270 ms (124 _+60 ms, n = 21), the latter exhibiting more pronounced plateau phases and longer durations of 90-290 ms (148 _+56 ms, n = 26). The upstroke velocity of both these cell types proved to be similar: 227 + 55 V / s (n = 21) and 235 _+68 V / s (n--27) for atrial- and ventricular-like action potentials, respectively. The upstroke velocity was sensitive to TFX and Cd 2÷, the latter also suppressed the plateau phase (Fig. 5a,b). BayK 8644, a specific opener of L-type Ca 2÷ channels, considerably increased the plateau phase thus prolongating the action potentials (Fig. 5c). In contrast, isradipine, a specific blocker of L-type C a 2 + channels, completely suppressed the plateau phase without appreciable effect on the upstroke velocity of action potentials.
Fig. 3. Morphological characterization of enzymatically isolated ES cell-derived cardiomyocytes by indirect immunofluorescence. Detection of sarcomeric myosin heavy chain (MF-20, (a)), intermediate filament protein desmin (b) and a-cardiac myosin heavy chain (BA-G5, (c-g)); the major morphologies of cardiomyocytes were spindle-shaped (d), round (f), triangular (a-c), multiangular (e) after short cultivation, and flat polymorphic-shaped with multiple pseudopodia-like processes (g) after prolonged cultivation. Bar = 10/~m.
46 a
a
+60
b
~_~
+60
100
ms
adr
Or~V
control
~ ii l
I
~
-80
i
control ~
'" "\ \\
"\ ~
~
carb
-80 b
t
500 ms
,
c
C +60]
d
~
adr '~ carb
-60J
d +60]
200 ms r
~]
i
2 0 0 pA
0 Fig. 6. Unusual long action potentials seen in cardiomyocytes differentiated from ES cells for ' 7 + 5 d' to ' 7 + 12 d'. (a-b) Spontaneously generated action potentials of durations between 500 to 1000 ms. (c) Long action potential triggered from a stable resting potential by a short depolarizing current pulse. (d) In some cardiomyocytes a 10 ms long depolarizing current pulse (200 pA amplitude, lower panel) was enough to induce spontaneous action potentials from a stable holding potential.
Fig. 7. Hormonal modulation of action potentials measured in atrial (a,b) and ventricular (c,d) cardiomyocytes of terminally differentiated ES cells ('7+ 12 d'). Adrenaline (adr, 1 #M) prolonged the duration and increased the plateau amplitude in both atrial and ventricular action potentials (a,c). The muscarinergic cholinoceptor agonist carbachol (carb, 10 >M) shortened the duration and hyperpolarized the resting potential of atrial (b) but not of ventricular action potentials (d). In ventricular cardiomyocytes, carbachol antagonized the adrenaline effect (c).
potential could be induced to generate action potentials by short current pulses (Fig. 6d). In these cells, the generation of action potentials could be blocked by a long hyperpolarizing current pulse (not shown).
1 g + 40
a
The slower upstroke velocity action potentials of the intermediate developmental stage seemed to be precursors of sinusnodaMike action potentials (Fig. 4f). The sinusnodal-like action potentials exhibited more pronounced linear diastolic depolarizations and faster upstroke velocities (5-20 V/s; 9.3 -+_4.7 V/s, n = 11) as compared to those inherent to early differentiation stages. An unexpected finding was the detection of action potentials of very long duration (0.3-1 s) and high plateau phases (+ 30 to + 40 mV) suggesting a transitional state between different phenotypes of cardiomyocytes during development. These action potentials (Fig. 6) were observed at the intermediate and terminal differentiation stage ('7 + 5 d' to '7 + 12 d'). Action potentials generated spontaneously revealed upstroke velocities in the range of 30-100 V / s (Fig. 6a,b), those elicited by stimulation from stable resting potentials had a higher upstroke velocity (100-300 V/s) (Fig. 6c). In addition, some cells having stable level of membrane
contro~
r
1
OmV - -
-60 +40 ]
b
~ carbachol
OmV . . . . . . . . . . . . . . .
-60
'I~
~
adrenaline +40
C
OmV
-60
Fig. 8. Hormonal modulation of action potentials measured in sinusnodal cardiomyocytes of terminally differentiated KS ceils ('7 + 12 d'). Shown are spontaneous pacemaker action potentials of one cell under control conditions (a), after superfusion of 10 /xM carbachol (b) and of 1 # M adrenaline (c).
47
Action potentials of ES cell-derived cardiomyocytes are hormonally modulated depending on the cardiac phenotype In our previous study on multicellular beating areas of ES cell-derived embryoid body outgrowths we found positive and negative chronotropic effects of/3-adrenoceptor and cholinoceptor agonists, respectively (Wobus et al., 1991). This prompted us to investigate the effects of these agonists on action potentials of isolated cardiomyocytes at terminal stage of differentiation, i.e., action potentials of atrial, ventricular and sinusnodal types (see Fig. 4d,e,f, respectively). On superfusion of adrenaline (1 tzM) the atrial- (Fig. 7a) and ventricular-like (Fig. 7c) action potentials showed an enhancement of the plateau phase and a prolongation of action potential duration. Carbachol (10 ~zM) had no significant effect on ventricular-like action potentials (Fig. 7d), but induced hyperpolarization and shortening of atrial-like action potentials (Fig. 7b). However, carbachol reduced the duration of ventricular-like action potentials when previously stimulated by adrenaline (Fig. 7c). Sinusnodal-like action potentials were modulated by both, carbachol and adrenaline, which slowed and enhanced the diastolic depolarization resulting in negative and positive chronotropic effects, respectively (Fig. 8a-c).
Discussion
By using three different approaches we could demonstrate that pluripotent mouse ES cells were capable to differentiate into cardiomyocytes expressing major features of mammalian heart. While PCR analysis and immunofluorescence staining revealed the expression of heart-specific genes and proteins, the patch-clamp technique allowed a more detailed characterization of cells into different phenotypes corresponding to sinusnode, atrium and ventricle of the heart. The expression pattern of the MHC genes in D3 ESC-derived embryoid bodies described in this paper is in agreement with the expression pattern previously described by Sfinchez et al. (1991) and reflects the situation in the developing mouse embryo (Lompr6 et al., 1984). Moreover, our standardized differentiation system allowed the conclusion that early expression of the/3-cardiac MHC gene is influenced by cell number originating the embryoid body, indicating the significance of cell-cell interactions for gene expression. This suggests that during spontaneous differentiation a minimum number of ES cells in the embryoid body is necessary to produce a threshold concentration of factor(s) regulating commitment and differentiation (see
Smith et al., 1987; Wobus et al., submitted) into the cardiomyogenic lineage. Coexpression of a- and B-cardiac MHC genes is a characteristic for ventricular as well as atrial cells during early embryonic heart development (Lompr6 et al., 1984). Coexpression was also demonstrated for the myosin light chain (MLC) genes MLC1A and MLC1V, which were found to be differentially expressed only around birth and thereafter (Lyons et al., 1990). A difference in gene expression pattern between atrial and ventricular cells in early embryonal stages was observed for /3-cardiac MHC, which is not further expressed in atrial cells of the mouse embryo from day 10 p.c. During ES cell differentiation, both, ventricular and atrial cells are differentiated throughout the embryoid body development. Because of the mixed population of different phenotypes of cardiac cells in the embryoid body outgrowths, the identification of different types of cardiomyocytes would be difficult only on the basis of gene expression. However, electrophysiological measurements allowed us to characterize the major cardiac cell types. The present study using patch-clamp technique on single cells isolated from beating areas of embryoid bodies demonstrated the development of cardiomyocytes exhibiting different shapes of action potentials during differentiation. The pacemaker-like action potentials of cardiomyocytes at early differentiation stage proved to be similar t o those described in cardiomyocyte aggregates of embryoid body outgrowths (Wobus et al., 1991). Similar characteristics have been previously described for early differentiated chick (Sperelakis and Pappano, 1983) and mammalian embryonic heart cells (Kamino, 1991). In contrast to the homogeneity of action potential shapes observed in early differentiated cardiomyocytes, the cardiomyocytes of terminal differentiation stage revealed diversity of action potentials inherent to cells from different parts of adult heart as described for sinusnode, atrial and ventricular cells (Irisava, 1989; for developing chick heart, see Sperelakis and Pappano, 1983; for neonatal rat heart, see Kilborn and Fedida, 1990). The cells proved to express all significant pharmacological properties and cardiac-specific signal transduction pathways. Both, TTX- and cadmium sensitivity of upstroke velocities suggests involvement of cardiac-specific Na+channels (sensitive to cadmium, see Backx et al., 1992) in generation of upstroke in action potentials at terminal stage of cell differentiation. High sensitivity of plateau phase of action potentials to dihydropyridines as well as to cadmium reflects that the plateau is supported by L-type Ca 2+ current. As far as the hormonal control of action potentials is concerned, similar modulations of action potentials by adrenaline and carbachol were previously described for adult cardiac tissue (Trautwein and Hescheler, 1990).
48 Carbachol had no significant effect on vcntricular cells as it cannot down regulate the basal adenylyl cyclase activity. However, its effect could be realized when adenylyl cyclase was stimulated by pretreatment with adrenaline (see also Hescheler et al.. 1986). In contrast, action potentials of atrial- and sinusnode-like cells were strongly and directly modulated by carbachol indicating the functional expression of muscarinic cholinoceptor coupling to inwardly rectiI~cing K +and I t currents (Belardinelli et al., 1989; DiFrancesco and Tromba, 1989). This was confirmed by our voltageclamp experiments on ES cell-derived cardiomyocytes (Maltsev et al., submitted). Our data on expression of various shapes of action potentials together with our previous voltage-clamp studies (Maltsev et al., submitted) suggest differential expression of ionic channels during cardiogenesis. The pacemaker-like action potentials in early differentiated cardiomyocytes may be generated by only two main channel types (Ic~,2+ and lto) expressed at this stage. However, cells differentiating further into ventricular and atrial-like cells in addition may express Na + and inwardly rectifying K + channels resulting in more negative resting potentials and faster upstroke velocity (Trube, 1989). This idea may also be confirmed by our finding of rhythmic action potentials triggered from stable level of resting potential as well as of extremely long action potentials. This may suggest the presence of developmental stages of cells expressing different sets of ionic channels. A developmental increase in the transient outward K+current expression were found to result in progressive shortening of the newborn rat action potential in the period up to 10 days of age (Kilborn and Fedida, 1990). Developmentally regulated expression of different ionic channels may provide a new basis for gene expression studies applying PCR techniques to genes encoding different types of ionic channels. In conclusion, our data demonstrated that in contrast to all other previously described cultivation systems for cardiac cells (e.g. Sperelakis 1978; Bugaisky et al., 1989), the ES cell differentiation system for the first time permits the investigation of commitment and differentiation of embryonic cells into the cardiomyogenic lineage in vitro.
Experimental Procedures Cell culture The embryonic stem cell line D3 (Doetschmann et al., 1985) cultivated in undifferentiated state on Mitomycin C-inactivated feeder layer of primary cultures of mouse embryonic fibroblasts (Wobus et al., 1984) was
used throughout the study. Cells were cultivated on gelatin (0.1%)-coated plastic petri dishes (Falcon) in DMEM (Gibco BRL, Life Technologies, Germany) supplemented with 15¢} fetal calf serum (FCS, selected batches, Gibco), l-glutaminc (Gibco, 2 raM), /3mercaptoethanol (Scrva, Heidelberg, final concentration 5 × 10 5M) and non-essential amino acids (Gibco, stock solution diluted 1: 10l)) as described (Wobus et al., 1991). For differentiation into cardiomyocytes, cells were cultivated in hanging drops (see Fig. 1) as aggregates ('embryoid bodies') in medium containing 20c~; FCS. A definite number of ES cells (400 or 1000 cells) in 20 #1 cultivation medium was placed in drops on the lids of pctri dishes filled with PBS. After 2 days cultNation in hanging drops the embryoid bodies were resuspended in 5 ml medium in bacteriological petri dishes and cultivated for further 5 days in suspension. The 7 days ('7 d') old embryoid bodies were plated separately onto gelatin-coated 24 well-microwell plates and were investigated at several days after plating as indicated (only for RT-PCR analysis at "7 d': embryoid bodies were cultivated '5 d" in suspension and '2 d' after plating). One day after attachmcnt of the '7 d" cmbryoid bodies to the substratc the first areas of spontaneously beating cardiomyocytcs appeared between epithelial-like cells of the embryoid body outgrowths ('7 + 1 d'; Wobus et al., 1991 ). The [k)llowing developmental stages were investigated: (i) early differentiation, shortly after initiation of contractions ('7 4-2 d' to '7 + 4 d'), (ii) an intermediate stage ('7 + 5 d' to "7 + 8 d') and (iii) a terminal differentiation stage ('7 + 9 d' to '7 + 12 d').
Detection of cardiac-specific myosin transcripts by RTPCR analysis Ten embryoid bodies or outgrowths of cmbryoid bodies were collectcd in lysis buffer (4.0 M guanidinium thiocyanate, I),l M Tris-HCl pH 7.5, 1% ¢3mercaptoethanol) and the cell extract affinity purified for poly(A) ÷ mRNA according to Sheardown (1992). Poly(A) + mRNA was reverse transcribed and amplified using oligonucleotide primer complementary and identical to the a- and /3-cardiac and embryonic skeletal myosin heavy chain transcripts. The oligonucleotide primers previously described by Robbins et al. (1990) were used. The reverse transcription and amplification reactions were carried out with rTth-Polymerase (Perkin Elmer, USA) following the protocol supplied by the manufacturer. The products of the reverse transcription reactions were denatured for 2 min at 95°C, followed by 39 cycles of amplification (1 min denaturation at 95°C, 31) s annealing at 65°C and 1 rain elongation at 72°C). One fifth of each RT-PCR reaction was electrophoretically separated on 2% agarose gels.
49
Single cell preparations Single cardiomyocytes were prepared from embryoid body outgrowths originated from 400 ES ceils by modification of the procedure described by Isenberg and Kl6ckner (1982). The following solutions were used: (1) Low Ca2+-medium: (in mM) NaC1 120, KC1 5.4, MgSO 4 5, Na-pyruvate 5, glucose 20, taurine 20, HEPES 1 0 / N a O H pH 6.9 at 24°C. (2) Enzyme medium: low Ca 2+ medium (above) supplemented with 1 mg/mI collagenase (collagenase B, Boehringer, Mannheim, Lot 12086223-05) and 30 # M CaCI 2. (3) KB medium (in mM): KC1 85, K2HPO 4 30, MgSO 4 5, EGTA 1, Na2ATP 2, Na-pyruvate 5, creatine 5, taurine 20, glucose 20, pH 7.2 at 24°C. Beating areas of 5-10 embryoid bodies were mechanically isolated with a microscalpel and washed in low CaZ+-medium. Subsequently, tissue fragments were incubated in enzyme medium for 25-35 rain at 37°C. The dissociation of the tissue was completed in KI3 medium by gentle shaking at room temperature for 1 h. The isolated cells were resuspended in cultivation medium supplemented with 20% FCS and incubated at 37°C. After 6-12 h of incubation cardiomyocytes resumed spontaneous rhythmic contractions. These beating ceils were used for immunostaining and electrophysiological measurements. At the terminal differentiation stage (9-12 days after embryoid body plating) cells without spontaneous contractions were selected by their morphology (see Results) similar to those of beating cardiomyocytes.
Indirect immunofluorescence assay Single cell preparations of beating aggregates were plated onto gelatin-coated glass-coverslips placed in tissue culture dishes and incubated over night. Ceils were rinsed 2 times with PBS, fixed with methanol : aceton = 3 : 1 at - 20°C, incubated with the respective monoclonal antibodies for 45 min at 37°C in a humid chamber, rinsed 3 times with PBS and incubated with the FITC-labelled anti-mouse IgG (Dianovo, Hamburg, Germany) diluted in PBS (1:2000) for 45 min at 37°C. After rinsing 3 times with PBS, cover slips were embedded in Tris-glycerin buffer (Walter et al., 1984) and analyzed with the fluorescence microscope Axioskop (Zeiss, Oberkochen, Germany). The following antibodies were used: MF-20 (Bader et al., 1982), anti-desmin (Boehringer, Mannheim, Germany), and anti-a-cardiac MHC (Rudnicki et al., 1990).
Electrophysiology Action potentials were measured with a List L M / EPC-7 amplifier under current clamp conditions in the
whole-cell configuration of the patch-clamp technique (Hamill et al., 1981). The micropipettes were pulled from Jencons glass capillaries (Leighton Buzzard, UK). The pipette solution contained (in mM): K-aspartate 80, KC1 50, MgC12 1, Mg-ATP 3, EGTA 10, HEPES 10/KOH. Pipette resistances ranged from 3 to 4 MI2. During measurement cells were perfused with a bath solution containing (in mM): NaC1 140, CaC12 1.8, MgC12 1.0, KC1 5.4, glucose 10, HEPES 10 (pH 7.4 at 37°0. The bath temperature was kept constant at 37°C. The duration of action potential evoked by stimulation was measured at the level of 80% recovery of resting potential after depolarization. For pacemakerlike action potentials it was measured at the level of the upstroke. Data analysis was accomplished by EPCMENU software (Cambridge) and Sigma-Plot (5.0) programs. All electrophysiological data are presented as mean + S.D. for n cells. Isradipine was a gift from Sandoz (Basel, Switzerland), all other substances were purchased from Sigma (St.Louis, USA).
Acknowledgements The authors wish to thank Dr. R. Kemler, MaxPlanck-Institute of Immunobiology, Freiburg/Br., Germany, for supplying us with the D3 line, Dr. S. Ausoni, Department of Biology, University of Padova, Italy, for the generous gift of monoclonal antibody BA-G5, and Dr. H.- H. Arnold, University of Braunschweig, Germany, for MF-20. The skilful technical assistance of Mrs. Oda WeiB and Mrs. Sabine Sommerfeld is gratefully acknowledged. The research was supported by DFG, Ministry of Science and Research of Sachsen-Anhalt and Zebet, Bundesgesundheitsamt of Germany. The authors wish to thank also the Fonds der Chemischen Industrie for financial support.
References Backx, P.H., Yue, D.T., Lawrence, J.H., Marban, E. and Tomaselli, G.F. (1992) Science 257, 248-251. Bader, D, Masaki, T. and Fischman, D.A. (1982) J. Cell. Biol. 95, 763-770. Belardinelli, L., Kl6ckner, U. and lsenberg, G. (1989) In: Piper H.M., Isenberg G. (eds.) Electrophysiology and Contractile Function, CRC Press, Boca Raton, FL, Vol. 2, pp. 155-180. Bradley, A., Evans, M., Kaufman, M.H. and Robertson E. (1984) Nature 309, 255-256. Buckingham, M. (1992) Trends Genet. 8, 144-149. Bugaisky, L.B. and Zak, R. (1989) Circ. Res. 64, 493-500. Capogrossi, M.C. and Lakato, E.G. (1989) In: Piper, H.M., Isenberg, G. (eds.), Electrophysiology and Contractile Function, CRC Press, Boca Raton, FL, Vol. 2, pp. 183-212. Claycomb, W.C. and Palazzo, M. (1980) Dev. Biol. 80, 466-482. DiFrancesco, D. and Tromba, C. (1989) In: Piper, H.M., Isenberg, G. (eds.) Electrophysiology and Contractile Function, CRC Press, Boca Raton, FL, Vol. 2, pp. 97-115.
50 Doetschmann, T.C., Eistetter, H.R., Katz, M., Schmidt, W. and Kemler, R. (1985) J. Embryol. Exp. Morphot. 87, 27 45. Eppenberger, H.M., Messerli, M., Mi.iller, M., Schwarb, P. and Eppenberger-Eberhardt. M.E. (1990) In: Pette, D. (ed.) The Dynamic State of Muscle Fibers, De Gruyter, Berlin, pp. 193-2114. Evans, M.J. and Kaufman, M.H. (19811 Nature 291, 85-10(/. Hammil, O.P., Marty, A., Neher, E., Sakmann, B. and Sigworlh, J. (19811 Pfliigers Arch. 391, 85-100. Han, Y., Dennis, J.E., Cohen-Gould, L., Bader, D.M. and Fischman, D.A. (19921 Dev. Dynamics 193, 257 265. Hescheler, J., Kameyama, M. and Trautwein. W. (19861 Pfliigers Arch. 407, 182-189. Hescheler, J., Meyer, R., Plant, S., Krautwurst, D., Rosenthal, W. and Schu]tz, G. (1991) Circ. Res. 69, 1476-1486. lrisawa, H. (1989) In: Piper, H.M., lsenberg, G. (eds.), Isolated Adult Cardiomyocytes, CRC Press, Boca Raton, FL, Vol. 2, pp.1-12. Isenberg, G. and Kl6ckner, U. (1982) Pfliigers Arch. 395, 6-18. Kamino, K. (1991) Phys. Rev. 71, 53-91. Kilborn, M.J. and Fedida, D. (1990) J. Physiol. 430, 37-60. Lane, M.-A., Sastre, A., Law, M. and Salpeter, M.M. (1977) Dev. Biol. 57, 254-269. LiWin, J., Montgomery, M., Gonzalez-Sanchez, A., Bisaha, J.H. and Bader, D. (19921 Trends Cardiovasc. Med. 2, 27-32. Lomprf, A.M., Nadal-Ginard, B. and Mahdavi. V. (1984) J. Biol. Chem. 259, 6437-6446. Lyons, G.E., Schiaffino, S., Sassoon, D., Barton, P. and Buckingham, M. (1990)J. Cell. Biol. 111, 2427-2436. Maltsev, V.A., Wobus, A.M. and Hescheler, J. (19941 (submitted). Martin, G. (19811 Proc. Natl. Acad. Sci. USA 78, 7634-7638. McLean, M.J., Renaud, J.F. and Sperelakis, N. (1987) Differentiation 11, 13-17. Nag, A.C., Cheng, M., Fischman, D.A. and Zak, R. (1983) J. Mol. Cell. Cardiol. 15, 301-317. Robbins, J., Gulick, J., Sanchez, A., Howles> P. and Doetschmann, T. (1990) J. Biol. Chem. 265, 119115-11909. Rosenquist, G. and DeHaan, R.L. (1966) Contrib. Embryol. Carnegie Inst. Wash. 263, 113-121.
Rudnicki. M.A., Jackowski, (i., Saggin, L. and McBurney, M?,\. (19901 l)ev. Biol. 138, 348 358. Sanehez, A., Jones, W.K., Gulick, J., l)oetschmann, rl. and Robbins, J. (19911 J. Biol. Chem. 266, 22419 22426. Sheardown. S.A. (1992) Trends Genct. 8, 121. Sipido, K.R. and Marban, t.(19tJl)('irc. Res.(~9, 1487 1499. Smith, S.('., Reuhl. K.R., ('raig, J. and McBurney, M.W. (1987} J. Cell. Physiol. 131, 74-.84. Sperelakis, N. (1978) Env. llcalth Persp. 2(). 243-267. Sperehikis. N. (1982) In: Van Stee, E.W. (cd.). Cardiovascular-l'~lxicology, Raven, New York, NY. pp. 57-108. Sperelakis, N. and Pappano, ,,\.J, (1983} Pharmacol. Ther. 22, 1-39. Sperelakis, N. and Shigenobu. K. (1974) J. Mol. Cell. Cardio]. ¢~, 449 471. Trautwein, W. and Heschelcr, ,I. (l~J91)) Annu. Rex. f)llysiol. 52, 257 274. Trube, G. (1989) In: Piper, t].M., lsenberg, G. (eds.), Eleclrophysiology and contractile function, ('R(' Press, Boca Raton, FL, Vol. 2, pp. 75-95. Walter, G., lntek, A., Wobus, A.M. and Sch(:,neich, ,1. (1984)Cc]l Diff. 15, 147-151. Wobus, A.M., Holzhausen, 11., Jiikel. P. and Sch(hleich, J. (lOiN4} Exp. ('ell Res. 152. 212-21cj. Wobus, A.M., Grosse, R. and Sch6neich, ,1. (19881 Biomed. Biochim. Acta 47, 965 973. Wobus, A.M., Wallukal. G. and l leschclcr. ,I. (19911 Differentiation 48, 173-182. Wobus, A.M., Kleppisch, T., Maltse~, V. and llescheler. J. (submitred}. Wollcnberger, A. (1984) In: Abe. tt.. lto, Y., Tada, M., Opic, I,.tt. (eds.). Regulation of cardiac function. Molecular. cellular and pathophysiological aspects, Japan Scientific Soc. Press, T(~k~,o. VNU Science Press. BV Utrecht, pp. 269-284. Wollenberger. A. (1985) Basic Res. Cardiol. 81), Suppl. 2, 9 14.
41
© 1993Elsevier ScientificPublishers Ireland, Ltd. 0925-4773/93/$06.00 MOD 191
Embryonic stem cells differentiate in vitro into cardiomyocytes representing sinusnodal, atrial and ventricular cell types Victor A. Maltsev a, Jiirgen Rohwedel a, Jiirgen Hescheler b and A n n a M. Wobus a,, a Institut fiir Pflanzengenetik und Kulturpflanzenforschung, Corrensstr. 3, D-06466 Gatersleben, Germany and b Institut far Pharmakologie, Freie Universitiit Berlin, D-14195 Berlin, Germany
(Received 1 June 1993;revisionreceived 3 August1993;accepted6 August1993)
Pluripotent embryonic stem cells (ESC, ES ceils) of line D3 were differentiated in vitro via embryo-like aggregates (embryoid bodies) of defined cell number into spontaneously beating cardiomyocytes. By using RT-PCR technique, a- and /3-cardiac myosin heavy chain (MHC) genes were found to be expressed in embryoid bodies of early to terminal differentiation stages. The exclusive expression of the/3-cardiac MHC gene detected in very early differentiated embryoid bodies proved to be dependent on the number of ES cells developing in the embryoid body. Cardiomyocytes enzymatically isolated from embryoid body outgrowths at different stages of development were further characterized by immunocytological and electrophysiological techniques. All cardiomyocytes appeared to be positive in immunofluorescence assays with monoclonal antibodies against cardiac-specific a-cardiac MHC, as well as muscle-specific sarcomeric myosin heavy chain and desmin. The patch-clamp technique allowed a more detailed characterization of the in vitro differentiated cardiomyocytes which were found to represent phenotypes corresponding to sinusnode, atrium or ventricle of the heart. The cardiac cells of early differentiated stage expressed pacemaker-like action potentials similar to those described for embryonic cardiomyocytes. The action potentials of terminally differentiated cells revealed shapes, pharmacological characteristics and hormonal regulation inherent to adult sinusnodal, atrial or ventricular cells. In cardiomyocytes of intermediate differentiation state, action potentials of very long duration (0.3-1 s) were found, which may represent developmentally controlled transitions between different types of action potentials. Therefore, the presented ES cell differentiation system permits the investigation of commitment and differentiation of embryonic cells into the cardiomyogenic lineage in vitro.
Cardiomyocyte; Embryonic stem cell differentiation; Cardiac-specific gene expression; Action potential
Introduction
A considerable amount of data has been accumulated on the morphology and physiology of adult heart ceils, but only limited information is available about the origin, commitment and differentiation of the cells that will make up the myocardium (Litvin et al., 1992). In contrast to the well-described process of skeletal muscle cell development (Buckingham, 1992) the temporally and spatially controlled events of commitment and differentiation of cardiomyocytes remain to be
* Corresponding author. Tel: (+ 49)39482-5256. Fax: (+ 49)39482280.
investigated. In this respect it is relevant to emphasize that these two muscle cell types arise from different progenitor cells. Skeletal muscle cells develop from somites, but cardiac cells from anterior lateral plate mesoderm at an earlier stage of embryonic development. Recent data suggest that a very short interval exists between mesodermal commitment and phenotypic manifestation of differentiated cardiomyocytes (Han et al., 1992). One of the most important obstacles to study commitment and differentiation of cardiac myocytes is the lack of permanent cell lines to model and analyze the earliest stages of cardiomyogenesis. Physiological studies of developing heart cells have been mainly carried out on organ-cultured chick hearts (Sperelakis, 1982; Sperelakis and Shigenobu, 1974),
42 precardiac areas of blastoderm (McLean et al., 1987; Rosenquist and de Haan, 1966), and cultured mammalian heart cells of embryonic and neonatal origin (Wollenberger 1984; 1985). These studies on primary cardiac cell cultures are limited since cell properties and normal cardiogenesis apparently are disturbed during cultivation. For example, when 3-day-old chick embryonic heart cells are placed into culture, the cells do not continue to differentiate with respect to their electrical and morphological properties (Sperelakis, 1978). Differentiated ventricular cells from 16-day-old chick embryos cultured as monolayers revert back to the young embryonic state (Sperelakis and Pappano, 1983). Cultivated adult ventricular myocytes lose myofibrillar proteins and highly organized morphology (Nag et al., 1983; Claycomb and Palazzo, 1980). They retain only a few properties of the differentiated heart (e.g. lack of automaticity) during a limited culture period (Bugaisky and Zak, 1989). Up to now, only a few permanent cell lines (e.g. H9c2 from embryonic rat heart) expressing some cardiac-specific properties including L-type Ca 2+ current were characterized (Hescheler et al., 1991; Sipido and Marban, 1991). A new approach may be offered by using cardiomyocytes differentiated in vitro from pluripotent mouse embryonic stem cells (ES ceils, ESC) (Doetschmann et al., 1985; Wobus et al., 1991). These permanent cell lines derived by in vitro culturing undifferentiated cells of early embryos are capable to take part in the embryonic development in vivo after retransfer into blastocysts (e.g. Bradley et al., 1984). In vitro, ES cells have been shown to spontaneously differentiate into derivatives of all three primary germ layers, endoderm, ectoderm and mesoderm (Evans and Kaufman, 1981; Martin, 1981; Wobus et al., 1984; 1988). Our previous results with ES cell-derived cardiomyocytes (lines D3 and Bl17) revealed the functional expression of adrenoceptors, cholinoceptors and L-type Ca 2 ~ channels (Wobus et al., 1991). In the present study we used the ES cell differentiation system to investigate the cardiomyocyte differentiation in vitro with respect to expression of tissue-specific genes, proteins and the developmentally controlled functional expression of electrophysiological properties. The expression of cardiac-specific genes was demonstrated by RT-PCR analysis, the formation of cardiac-specific intracellular proteins was detected by indirect immunofluorescence, and the development of action potentials was characterized by patch-clamp technique. Our data proved that ES cell-derived cardiomyocytes expressed cardiac-specific genes and proteins. Furthermore, a more detailed characterization of cardiomyocytes differentiating into sinusnodal, atrial and ventricular cells was only possible by electrophysiological measurements of action potentials.
Results
Expression of cardiac-.V)eciJic genes i.s del'elopmenta/b" regula ted Because the differentiation capacity is clearly dependent on the number of cells differentiating in the embryoid bodies (see Smith et al., 1987: Wobus ct al., 1991), all experiments were done with a standardized differentiation protocol, i.e. cultivation of cmbryoid bodies with a defined cell number in hanging drops (Fig. 1). During different stages of cultivation single embryoid bodies were analyzed by RT-PCR wilh oligonucleotide primers specific for (v- and /3-cardiac and embryonic skeletal MHC genes (Fig. 2). The first transcripts of /3-cardiac MHC could be detected in "4 d' embryoid bodies originated from 1000 cells but not from 400 cells. At this stage beating areas in embryoid bodies originated from 400 ES cells wcrc never observed.
Embryonic stem cells cultivated on feeder layer
1 Cultivation of 400 or 1000 cells/2Olal medium in hanging drops
/ Incubation 2 days
"~J
\.a. j
"-L- j
<.It; '
77 ~ 7 7 1 Transfer of embryoid bodies and cultivation in suspension .....
Incubation for 5 days
Plating of embryoid bodies
to 24-well tissue culture plates
IIU
UUUII
Incubation up to 20 days
1.PCR analysis 2.Enzymatic dissociation of cardiomyocytes 3.1mmunofluorescence 4.Patch-clamp studies Fig. 1. S c h e m a t i c p r e s e n t a t i o n of the e x p e r i m e n t a l p r o t o c o l u s e d for ES cell d i f f e r e n t i a t i o n into c a r d i o m y o c y t e s .
43 diomyocytes could be discriminated by their cell size of 40-100 /~m, larger than stem ceils (20-30 /xm) but smaller than epithelial-like cells (150-200 /zm). Cardiomyocytes of all morphological types were used in our study. Basic electrophysiological properties seemed to be expressed independently of cellular morphology. ES cell-derived cardiomyocytes developed musclespecific sarcomeric myosin heavy chain as well as desmin intermediate filament proteins detected by monoclonal antibody (MAb) MF20 (Bader et al., 1982, Fig. 3a) and MAb anti-desmin (Fig. 3b), respectively. As already suggested from the RT-PCR experiments, the cardiac-specific a-cardiac MHC detected with MAb BA-G5 (Rudnicki et al., 1990) was found to be specific for ESC-derived cardiomyocytes (Fig. 3c-g). The immuno-positive cells were identified as contracting before fixation. Striations by sarcomeric structures were visualized by anti-MHC MAbs MF-20 and BA-G5. The sarcomeric length was estimated to be in the range of 1.9-2.1 Izm and corresponds to that of adult mammalian heart (Capogrossi and Lakato, 1989).
The first pulsating cells in embryoid body outgrowths were found at day '7', when '5 d' embryoid bodies were plated (at '5 + 2 d'). At this time all investigated embryoid bodies with beating areas showed strong expression of both, a- and /3-cardiac MHC genes. We also detected a significant lower amount of a- and /3-cardiac MHC gene transcripts in about half of embryoid bodies without beating areas (not shown). From day '7' throughout the further differentiation the MHC gene expression was independent on the number of ES cells originating the embryoid bodies. In embryoid body outgrowths from day '7 + 5' and thereafter, in addition to a- and /3-cardiac MHC transcripts, also embryonic skeletal MHC transcripts were found. Up to terminal stages of differentiation ('7 + 12 d') all three. MHC transcripts were expressed but embryonic skeletal MHC always with a lower rate.
ES cell-dericed cardiomyocytes exhibit cardiac-specific intracellular proteins Cardiomyocytes enzymatically isolated from beating areas of embryoid body outgrowths spontaneously contracted at a rate ranging between 0.5 and 5 Hz. Cardiomyocytes were distinguished morphologically from epithelioid- and fibroblastoid-like cells and stem cells and revealed cell types similar to previously described cardiomyocytes of mouse embryos (Lane et al., 1977) and of adult myocardium (Bugaisky and Zak, 1989; Eppenberger et al., 1990). Typical cardiomyocytes were spindle-, round-, as well as tri- or multiangular-shaped shortly (1-3 days) after dissociation (Fig. 3a-f). During prolonged cultivation (1-2 weeks) the ceils developed into fiat polymorphic shaped cells with multiple pseudopodia-like processes (Fig. 3g). Furthermore, car-
4d(400)r4d(1000)l M ~x
ES cells differentiate into cardiomyocytes expressing action potentials of sinusnodal, atrial and ~,entricular type All cardiomyocytes of early differentiation stage ('7 + 2 d' to '7 + 4 d') revealed the same type of spontaneous pacemaker-like action potentials synchronous to the contractions (Fig. 4a). Action potentials were generated with a frequency in the range of 1 to 5 Hz, amplitudes of 60 to 70 mV, durations of 70 to 100 ms and upstroke velocities of 1 to 10 V / s . The most negative membrane potential attained was about - 5 0 mV and the threshold for action potential upstroke of near - 3 5 mV. During further differentiation car-
7d
I 7+4dl
7+5d17+12d es
Fig. 2. Expression of a-cardiac (a),/3-cardiac (b) and embryonicskeletal (es) MHC genes amplified by RT-PCR during the development of ES cell-derivedembryoidbodies (investigatedfrom '2 d' to '7+ 12 d'). Expressionpatterns of embryoidbodies differentiated for '4 d', '7 d', '7+4 d', '7+5 d' and '7+ 12 d' are shown. '4 d' old embryoid bodies originated from 1000 cells showed expression of the/3-cardiac MHC gene, whereas those originated from 400 cells did not. No differences in the MHC gene expression pattern dependent on the cell number were observed at further developmental stages. In case of /3-cardiac MHC, in addition to processed transcripts also unprocessed transcripts were amplified, generating fragments of approximately800 bp.
~ ~
~ ~i ¸
45 2
--
+60
a OmV
4
b
iilI
-8o
b OmV
..................
c
•
I
I
100
ma
c
8
__
-80
j
d
OmV
e
--
f
+ 8 0 "[
-
OmV
.80 j
12
-] -
.80 j
control (185)
........
Bay K (197)
I - - - - ~ - - - - ~ - -
J/
~
~
,
..........
~1
I
I~radlplne (176) =
O,5S
V "7+n" clays Fig. 4. Diversity of action potentials during FS cell-derived cardiomyocyte differentiation. (a) Pacemaker-like action potentials in early differentiated cells ('7+2 d' to ' 7 + 4 d'); (b,c) Action potentials of higher amplitude and pacemaker-like action potentials in cells of intermediate stage ('7+5 d' to ' 7 + 8 d'); d-f: action potentials at terminal stage ('7+9 d' to '7+12 d'): atrial-, ventricular- and sinusnode-like action potentials. The action potentials shown in (d,e) were evoked by pulse current stimulation in current-clamp, all others were generated spontaneously.
diomyocytes appeared to diverse into ceils exhibiting different types of specialized action potentials. Cardiomyocytes of intermediate developmental stages ('7 + 5 d' to '7 + 8 d') generated action potentials with higher upstroke velocity and more negative resting potentials, as well as pacemaker-like action potentials with slower upstroke velocity and more positive resting potentials (Fig. 4b,c). The higher upstroke velocity (10-50 V/s) action potentials were 100-200 ms long and exhibited amplitudes ranging from 80 to 100 mV. The slower upstroke velocity (2-10 V/s) action potentials had a duration of about 70 ms and an amplitude of about 70 mV. During further differentiation into terminal differentiation stage ('7 + 9 d' to '7 + 12 d') higher upstroke velocity action potentials obviously developed into atrial- and ventricular-like types (Fig. 4d,e). These terminally differentiated cardiomyocytes had a stable resting potential (-74.1 + 5.1 mV, n = 47) and lost their diastolic depolarization. The action potentials elicited
Fig. 5. Pharmacological modulation of action potentials measured in ventricular-like cardiomyocytes of terminally differentiated ES cells ('7+ 12 d'). Values of the determined upstroke velocity are given in parenthesis. (a) Superfusion of tetrodotoxin (TTX, 20 #M) caused a marked slowing of the upstroke velocity. (b) Cd 2+ (100/xM) slowed the upstroke velocity and strongly reduced the plateau amplitude. (c) Isradipine (2 #M) and BayK 8644 (1 /zM) reduced and prolonged the duration of action potential, respectively. Both dihydropyridines did not affect the upstroke velocity.
by external current stimuli, were characterized by high amplitudes (124 _+13 mV, n = 46), faster upstroke velocities in the range of 130-330 V / s (232_+ 62 V/s, n = 46) and an overshoot phase. The atrial- and ventricular-like types of action potentials differed by their shape of plateau and duration, the former being triangle-shaped with durations of 50-270 ms (124 _+60 ms, n = 21), the latter exhibiting more pronounced plateau phases and longer durations of 90-290 ms (148 _+56 ms, n = 26). The upstroke velocity of both these cell types proved to be similar: 227 + 55 V / s (n = 21) and 235 _+68 V / s (n--27) for atrial- and ventricular-like action potentials, respectively. The upstroke velocity was sensitive to TFX and Cd 2÷, the latter also suppressed the plateau phase (Fig. 5a,b). BayK 8644, a specific opener of L-type Ca 2÷ channels, considerably increased the plateau phase thus prolongating the action potentials (Fig. 5c). In contrast, isradipine, a specific blocker of L-type C a 2 + channels, completely suppressed the plateau phase without appreciable effect on the upstroke velocity of action potentials.
Fig. 3. Morphological characterization of enzymatically isolated ES cell-derived cardiomyocytes by indirect immunofluorescence. Detection of sarcomeric myosin heavy chain (MF-20, (a)), intermediate filament protein desmin (b) and a-cardiac myosin heavy chain (BA-G5, (c-g)); the major morphologies of cardiomyocytes were spindle-shaped (d), round (f), triangular (a-c), multiangular (e) after short cultivation, and flat polymorphic-shaped with multiple pseudopodia-like processes (g) after prolonged cultivation. Bar = 10/~m.
46 a
a
+60
b
~_~
+60
100
ms
adr
Or~V
control
~ ii l
I
~
-80
i
control ~
'" "\ \\
"\ ~
~
carb
-80 b
t
500 ms
,
c
C +60]
d
~
adr '~ carb
-60J
d +60]
200 ms r
~]
i
2 0 0 pA
0 Fig. 6. Unusual long action potentials seen in cardiomyocytes differentiated from ES cells for ' 7 + 5 d' to ' 7 + 12 d'. (a-b) Spontaneously generated action potentials of durations between 500 to 1000 ms. (c) Long action potential triggered from a stable resting potential by a short depolarizing current pulse. (d) In some cardiomyocytes a 10 ms long depolarizing current pulse (200 pA amplitude, lower panel) was enough to induce spontaneous action potentials from a stable holding potential.
Fig. 7. Hormonal modulation of action potentials measured in atrial (a,b) and ventricular (c,d) cardiomyocytes of terminally differentiated ES cells ('7+ 12 d'). Adrenaline (adr, 1 #M) prolonged the duration and increased the plateau amplitude in both atrial and ventricular action potentials (a,c). The muscarinergic cholinoceptor agonist carbachol (carb, 10 >M) shortened the duration and hyperpolarized the resting potential of atrial (b) but not of ventricular action potentials (d). In ventricular cardiomyocytes, carbachol antagonized the adrenaline effect (c).
potential could be induced to generate action potentials by short current pulses (Fig. 6d). In these cells, the generation of action potentials could be blocked by a long hyperpolarizing current pulse (not shown).
1 g + 40
a
The slower upstroke velocity action potentials of the intermediate developmental stage seemed to be precursors of sinusnodaMike action potentials (Fig. 4f). The sinusnodal-like action potentials exhibited more pronounced linear diastolic depolarizations and faster upstroke velocities (5-20 V/s; 9.3 -+_4.7 V/s, n = 11) as compared to those inherent to early differentiation stages. An unexpected finding was the detection of action potentials of very long duration (0.3-1 s) and high plateau phases (+ 30 to + 40 mV) suggesting a transitional state between different phenotypes of cardiomyocytes during development. These action potentials (Fig. 6) were observed at the intermediate and terminal differentiation stage ('7 + 5 d' to '7 + 12 d'). Action potentials generated spontaneously revealed upstroke velocities in the range of 30-100 V / s (Fig. 6a,b), those elicited by stimulation from stable resting potentials had a higher upstroke velocity (100-300 V/s) (Fig. 6c). In addition, some cells having stable level of membrane
contro~
r
1
OmV - -
-60 +40 ]
b
~ carbachol
OmV . . . . . . . . . . . . . . .
-60
'I~
~
adrenaline +40
C
OmV
-60
Fig. 8. Hormonal modulation of action potentials measured in sinusnodal cardiomyocytes of terminally differentiated KS ceils ('7 + 12 d'). Shown are spontaneous pacemaker action potentials of one cell under control conditions (a), after superfusion of 10 /xM carbachol (b) and of 1 # M adrenaline (c).
47
Action potentials of ES cell-derived cardiomyocytes are hormonally modulated depending on the cardiac phenotype In our previous study on multicellular beating areas of ES cell-derived embryoid body outgrowths we found positive and negative chronotropic effects of/3-adrenoceptor and cholinoceptor agonists, respectively (Wobus et al., 1991). This prompted us to investigate the effects of these agonists on action potentials of isolated cardiomyocytes at terminal stage of differentiation, i.e., action potentials of atrial, ventricular and sinusnodal types (see Fig. 4d,e,f, respectively). On superfusion of adrenaline (1 tzM) the atrial- (Fig. 7a) and ventricular-like (Fig. 7c) action potentials showed an enhancement of the plateau phase and a prolongation of action potential duration. Carbachol (10 ~zM) had no significant effect on ventricular-like action potentials (Fig. 7d), but induced hyperpolarization and shortening of atrial-like action potentials (Fig. 7b). However, carbachol reduced the duration of ventricular-like action potentials when previously stimulated by adrenaline (Fig. 7c). Sinusnodal-like action potentials were modulated by both, carbachol and adrenaline, which slowed and enhanced the diastolic depolarization resulting in negative and positive chronotropic effects, respectively (Fig. 8a-c).
Discussion
By using three different approaches we could demonstrate that pluripotent mouse ES cells were capable to differentiate into cardiomyocytes expressing major features of mammalian heart. While PCR analysis and immunofluorescence staining revealed the expression of heart-specific genes and proteins, the patch-clamp technique allowed a more detailed characterization of cells into different phenotypes corresponding to sinusnode, atrium and ventricle of the heart. The expression pattern of the MHC genes in D3 ESC-derived embryoid bodies described in this paper is in agreement with the expression pattern previously described by Sfinchez et al. (1991) and reflects the situation in the developing mouse embryo (Lompr6 et al., 1984). Moreover, our standardized differentiation system allowed the conclusion that early expression of the/3-cardiac MHC gene is influenced by cell number originating the embryoid body, indicating the significance of cell-cell interactions for gene expression. This suggests that during spontaneous differentiation a minimum number of ES cells in the embryoid body is necessary to produce a threshold concentration of factor(s) regulating commitment and differentiation (see
Smith et al., 1987; Wobus et al., submitted) into the cardiomyogenic lineage. Coexpression of a- and B-cardiac MHC genes is a characteristic for ventricular as well as atrial cells during early embryonic heart development (Lompr6 et al., 1984). Coexpression was also demonstrated for the myosin light chain (MLC) genes MLC1A and MLC1V, which were found to be differentially expressed only around birth and thereafter (Lyons et al., 1990). A difference in gene expression pattern between atrial and ventricular cells in early embryonal stages was observed for /3-cardiac MHC, which is not further expressed in atrial cells of the mouse embryo from day 10 p.c. During ES cell differentiation, both, ventricular and atrial cells are differentiated throughout the embryoid body development. Because of the mixed population of different phenotypes of cardiac cells in the embryoid body outgrowths, the identification of different types of cardiomyocytes would be difficult only on the basis of gene expression. However, electrophysiological measurements allowed us to characterize the major cardiac cell types. The present study using patch-clamp technique on single cells isolated from beating areas of embryoid bodies demonstrated the development of cardiomyocytes exhibiting different shapes of action potentials during differentiation. The pacemaker-like action potentials of cardiomyocytes at early differentiation stage proved to be similar t o those described in cardiomyocyte aggregates of embryoid body outgrowths (Wobus et al., 1991). Similar characteristics have been previously described for early differentiated chick (Sperelakis and Pappano, 1983) and mammalian embryonic heart cells (Kamino, 1991). In contrast to the homogeneity of action potential shapes observed in early differentiated cardiomyocytes, the cardiomyocytes of terminal differentiation stage revealed diversity of action potentials inherent to cells from different parts of adult heart as described for sinusnode, atrial and ventricular cells (Irisava, 1989; for developing chick heart, see Sperelakis and Pappano, 1983; for neonatal rat heart, see Kilborn and Fedida, 1990). The cells proved to express all significant pharmacological properties and cardiac-specific signal transduction pathways. Both, TTX- and cadmium sensitivity of upstroke velocities suggests involvement of cardiac-specific Na+channels (sensitive to cadmium, see Backx et al., 1992) in generation of upstroke in action potentials at terminal stage of cell differentiation. High sensitivity of plateau phase of action potentials to dihydropyridines as well as to cadmium reflects that the plateau is supported by L-type Ca 2+ current. As far as the hormonal control of action potentials is concerned, similar modulations of action potentials by adrenaline and carbachol were previously described for adult cardiac tissue (Trautwein and Hescheler, 1990).
48 Carbachol had no significant effect on vcntricular cells as it cannot down regulate the basal adenylyl cyclase activity. However, its effect could be realized when adenylyl cyclase was stimulated by pretreatment with adrenaline (see also Hescheler et al.. 1986). In contrast, action potentials of atrial- and sinusnode-like cells were strongly and directly modulated by carbachol indicating the functional expression of muscarinic cholinoceptor coupling to inwardly rectiI~cing K +and I t currents (Belardinelli et al., 1989; DiFrancesco and Tromba, 1989). This was confirmed by our voltageclamp experiments on ES cell-derived cardiomyocytes (Maltsev et al., submitted). Our data on expression of various shapes of action potentials together with our previous voltage-clamp studies (Maltsev et al., submitted) suggest differential expression of ionic channels during cardiogenesis. The pacemaker-like action potentials in early differentiated cardiomyocytes may be generated by only two main channel types (Ic~,2+ and lto) expressed at this stage. However, cells differentiating further into ventricular and atrial-like cells in addition may express Na + and inwardly rectifying K + channels resulting in more negative resting potentials and faster upstroke velocity (Trube, 1989). This idea may also be confirmed by our finding of rhythmic action potentials triggered from stable level of resting potential as well as of extremely long action potentials. This may suggest the presence of developmental stages of cells expressing different sets of ionic channels. A developmental increase in the transient outward K+current expression were found to result in progressive shortening of the newborn rat action potential in the period up to 10 days of age (Kilborn and Fedida, 1990). Developmentally regulated expression of different ionic channels may provide a new basis for gene expression studies applying PCR techniques to genes encoding different types of ionic channels. In conclusion, our data demonstrated that in contrast to all other previously described cultivation systems for cardiac cells (e.g. Sperelakis 1978; Bugaisky et al., 1989), the ES cell differentiation system for the first time permits the investigation of commitment and differentiation of embryonic cells into the cardiomyogenic lineage in vitro.
Experimental Procedures Cell culture The embryonic stem cell line D3 (Doetschmann et al., 1985) cultivated in undifferentiated state on Mitomycin C-inactivated feeder layer of primary cultures of mouse embryonic fibroblasts (Wobus et al., 1984) was
used throughout the study. Cells were cultivated on gelatin (0.1%)-coated plastic petri dishes (Falcon) in DMEM (Gibco BRL, Life Technologies, Germany) supplemented with 15¢} fetal calf serum (FCS, selected batches, Gibco), l-glutaminc (Gibco, 2 raM), /3mercaptoethanol (Scrva, Heidelberg, final concentration 5 × 10 5M) and non-essential amino acids (Gibco, stock solution diluted 1: 10l)) as described (Wobus et al., 1991). For differentiation into cardiomyocytes, cells were cultivated in hanging drops (see Fig. 1) as aggregates ('embryoid bodies') in medium containing 20c~; FCS. A definite number of ES cells (400 or 1000 cells) in 20 #1 cultivation medium was placed in drops on the lids of pctri dishes filled with PBS. After 2 days cultNation in hanging drops the embryoid bodies were resuspended in 5 ml medium in bacteriological petri dishes and cultivated for further 5 days in suspension. The 7 days ('7 d') old embryoid bodies were plated separately onto gelatin-coated 24 well-microwell plates and were investigated at several days after plating as indicated (only for RT-PCR analysis at "7 d': embryoid bodies were cultivated '5 d" in suspension and '2 d' after plating). One day after attachmcnt of the '7 d" cmbryoid bodies to the substratc the first areas of spontaneously beating cardiomyocytcs appeared between epithelial-like cells of the embryoid body outgrowths ('7 + 1 d'; Wobus et al., 1991 ). The [k)llowing developmental stages were investigated: (i) early differentiation, shortly after initiation of contractions ('7 4-2 d' to '7 + 4 d'), (ii) an intermediate stage ('7 + 5 d' to "7 + 8 d') and (iii) a terminal differentiation stage ('7 + 9 d' to '7 + 12 d').
Detection of cardiac-specific myosin transcripts by RTPCR analysis Ten embryoid bodies or outgrowths of cmbryoid bodies were collectcd in lysis buffer (4.0 M guanidinium thiocyanate, I),l M Tris-HCl pH 7.5, 1% ¢3mercaptoethanol) and the cell extract affinity purified for poly(A) ÷ mRNA according to Sheardown (1992). Poly(A) + mRNA was reverse transcribed and amplified using oligonucleotide primer complementary and identical to the a- and /3-cardiac and embryonic skeletal myosin heavy chain transcripts. The oligonucleotide primers previously described by Robbins et al. (1990) were used. The reverse transcription and amplification reactions were carried out with rTth-Polymerase (Perkin Elmer, USA) following the protocol supplied by the manufacturer. The products of the reverse transcription reactions were denatured for 2 min at 95°C, followed by 39 cycles of amplification (1 min denaturation at 95°C, 31) s annealing at 65°C and 1 rain elongation at 72°C). One fifth of each RT-PCR reaction was electrophoretically separated on 2% agarose gels.
49
Single cell preparations Single cardiomyocytes were prepared from embryoid body outgrowths originated from 400 ES ceils by modification of the procedure described by Isenberg and Kl6ckner (1982). The following solutions were used: (1) Low Ca2+-medium: (in mM) NaC1 120, KC1 5.4, MgSO 4 5, Na-pyruvate 5, glucose 20, taurine 20, HEPES 1 0 / N a O H pH 6.9 at 24°C. (2) Enzyme medium: low Ca 2+ medium (above) supplemented with 1 mg/mI collagenase (collagenase B, Boehringer, Mannheim, Lot 12086223-05) and 30 # M CaCI 2. (3) KB medium (in mM): KC1 85, K2HPO 4 30, MgSO 4 5, EGTA 1, Na2ATP 2, Na-pyruvate 5, creatine 5, taurine 20, glucose 20, pH 7.2 at 24°C. Beating areas of 5-10 embryoid bodies were mechanically isolated with a microscalpel and washed in low CaZ+-medium. Subsequently, tissue fragments were incubated in enzyme medium for 25-35 rain at 37°C. The dissociation of the tissue was completed in KI3 medium by gentle shaking at room temperature for 1 h. The isolated cells were resuspended in cultivation medium supplemented with 20% FCS and incubated at 37°C. After 6-12 h of incubation cardiomyocytes resumed spontaneous rhythmic contractions. These beating ceils were used for immunostaining and electrophysiological measurements. At the terminal differentiation stage (9-12 days after embryoid body plating) cells without spontaneous contractions were selected by their morphology (see Results) similar to those of beating cardiomyocytes.
Indirect immunofluorescence assay Single cell preparations of beating aggregates were plated onto gelatin-coated glass-coverslips placed in tissue culture dishes and incubated over night. Ceils were rinsed 2 times with PBS, fixed with methanol : aceton = 3 : 1 at - 20°C, incubated with the respective monoclonal antibodies for 45 min at 37°C in a humid chamber, rinsed 3 times with PBS and incubated with the FITC-labelled anti-mouse IgG (Dianovo, Hamburg, Germany) diluted in PBS (1:2000) for 45 min at 37°C. After rinsing 3 times with PBS, cover slips were embedded in Tris-glycerin buffer (Walter et al., 1984) and analyzed with the fluorescence microscope Axioskop (Zeiss, Oberkochen, Germany). The following antibodies were used: MF-20 (Bader et al., 1982), anti-desmin (Boehringer, Mannheim, Germany), and anti-a-cardiac MHC (Rudnicki et al., 1990).
Electrophysiology Action potentials were measured with a List L M / EPC-7 amplifier under current clamp conditions in the
whole-cell configuration of the patch-clamp technique (Hamill et al., 1981). The micropipettes were pulled from Jencons glass capillaries (Leighton Buzzard, UK). The pipette solution contained (in mM): K-aspartate 80, KC1 50, MgC12 1, Mg-ATP 3, EGTA 10, HEPES 10/KOH. Pipette resistances ranged from 3 to 4 MI2. During measurement cells were perfused with a bath solution containing (in mM): NaC1 140, CaC12 1.8, MgC12 1.0, KC1 5.4, glucose 10, HEPES 10 (pH 7.4 at 37°0. The bath temperature was kept constant at 37°C. The duration of action potential evoked by stimulation was measured at the level of 80% recovery of resting potential after depolarization. For pacemakerlike action potentials it was measured at the level of the upstroke. Data analysis was accomplished by EPCMENU software (Cambridge) and Sigma-Plot (5.0) programs. All electrophysiological data are presented as mean + S.D. for n cells. Isradipine was a gift from Sandoz (Basel, Switzerland), all other substances were purchased from Sigma (St.Louis, USA).
Acknowledgements The authors wish to thank Dr. R. Kemler, MaxPlanck-Institute of Immunobiology, Freiburg/Br., Germany, for supplying us with the D3 line, Dr. S. Ausoni, Department of Biology, University of Padova, Italy, for the generous gift of monoclonal antibody BA-G5, and Dr. H.- H. Arnold, University of Braunschweig, Germany, for MF-20. The skilful technical assistance of Mrs. Oda WeiB and Mrs. Sabine Sommerfeld is gratefully acknowledged. The research was supported by DFG, Ministry of Science and Research of Sachsen-Anhalt and Zebet, Bundesgesundheitsamt of Germany. The authors wish to thank also the Fonds der Chemischen Industrie for financial support.
References Backx, P.H., Yue, D.T., Lawrence, J.H., Marban, E. and Tomaselli, G.F. (1992) Science 257, 248-251. Bader, D, Masaki, T. and Fischman, D.A. (1982) J. Cell. Biol. 95, 763-770. Belardinelli, L., Kl6ckner, U. and lsenberg, G. (1989) In: Piper H.M., Isenberg G. (eds.) Electrophysiology and Contractile Function, CRC Press, Boca Raton, FL, Vol. 2, pp. 155-180. Bradley, A., Evans, M., Kaufman, M.H. and Robertson E. (1984) Nature 309, 255-256. Buckingham, M. (1992) Trends Genet. 8, 144-149. Bugaisky, L.B. and Zak, R. (1989) Circ. Res. 64, 493-500. Capogrossi, M.C. and Lakato, E.G. (1989) In: Piper, H.M., Isenberg, G. (eds.), Electrophysiology and Contractile Function, CRC Press, Boca Raton, FL, Vol. 2, pp. 183-212. Claycomb, W.C. and Palazzo, M. (1980) Dev. Biol. 80, 466-482. DiFrancesco, D. and Tromba, C. (1989) In: Piper, H.M., Isenberg, G. (eds.) Electrophysiology and Contractile Function, CRC Press, Boca Raton, FL, Vol. 2, pp. 97-115.
50 Doetschmann, T.C., Eistetter, H.R., Katz, M., Schmidt, W. and Kemler, R. (1985) J. Embryol. Exp. Morphot. 87, 27 45. Eppenberger, H.M., Messerli, M., Mi.iller, M., Schwarb, P. and Eppenberger-Eberhardt. M.E. (1990) In: Pette, D. (ed.) The Dynamic State of Muscle Fibers, De Gruyter, Berlin, pp. 193-2114. Evans, M.J. and Kaufman, M.H. (19811 Nature 291, 85-10(/. Hammil, O.P., Marty, A., Neher, E., Sakmann, B. and Sigworlh, J. (19811 Pfliigers Arch. 391, 85-100. Han, Y., Dennis, J.E., Cohen-Gould, L., Bader, D.M. and Fischman, D.A. (19921 Dev. Dynamics 193, 257 265. Hescheler, J., Kameyama, M. and Trautwein. W. (19861 Pfliigers Arch. 407, 182-189. Hescheler, J., Meyer, R., Plant, S., Krautwurst, D., Rosenthal, W. and Schu]tz, G. (1991) Circ. Res. 69, 1476-1486. lrisawa, H. (1989) In: Piper, H.M., lsenberg, G. (eds.), Isolated Adult Cardiomyocytes, CRC Press, Boca Raton, FL, Vol. 2, pp.1-12. Isenberg, G. and Kl6ckner, U. (1982) Pfliigers Arch. 395, 6-18. Kamino, K. (1991) Phys. Rev. 71, 53-91. Kilborn, M.J. and Fedida, D. (1990) J. Physiol. 430, 37-60. Lane, M.-A., Sastre, A., Law, M. and Salpeter, M.M. (1977) Dev. Biol. 57, 254-269. LiWin, J., Montgomery, M., Gonzalez-Sanchez, A., Bisaha, J.H. and Bader, D. (19921 Trends Cardiovasc. Med. 2, 27-32. Lomprf, A.M., Nadal-Ginard, B. and Mahdavi. V. (1984) J. Biol. Chem. 259, 6437-6446. Lyons, G.E., Schiaffino, S., Sassoon, D., Barton, P. and Buckingham, M. (1990)J. Cell. Biol. 111, 2427-2436. Maltsev, V.A., Wobus, A.M. and Hescheler, J. (19941 (submitted). Martin, G. (19811 Proc. Natl. Acad. Sci. USA 78, 7634-7638. McLean, M.J., Renaud, J.F. and Sperelakis, N. (1987) Differentiation 11, 13-17. Nag, A.C., Cheng, M., Fischman, D.A. and Zak, R. (1983) J. Mol. Cell. Cardiol. 15, 301-317. Robbins, J., Gulick, J., Sanchez, A., Howles> P. and Doetschmann, T. (1990) J. Biol. Chem. 265, 119115-11909. Rosenquist, G. and DeHaan, R.L. (1966) Contrib. Embryol. Carnegie Inst. Wash. 263, 113-121.
Rudnicki. M.A., Jackowski, (i., Saggin, L. and McBurney, M?,\. (19901 l)ev. Biol. 138, 348 358. Sanehez, A., Jones, W.K., Gulick, J., l)oetschmann, rl. and Robbins, J. (19911 J. Biol. Chem. 266, 22419 22426. Sheardown. S.A. (1992) Trends Genct. 8, 121. Sipido, K.R. and Marban, t.(19tJl)('irc. Res.(~9, 1487 1499. Smith, S.('., Reuhl. K.R., ('raig, J. and McBurney, M.W. (1987} J. Cell. Physiol. 131, 74-.84. Sperelakis, N. (1978) Env. llcalth Persp. 2(). 243-267. Sperehikis. N. (1982) In: Van Stee, E.W. (cd.). Cardiovascular-l'~lxicology, Raven, New York, NY. pp. 57-108. Sperelakis, N. and Pappano, ,,\.J, (1983} Pharmacol. Ther. 22, 1-39. Sperelakis, N. and Shigenobu. K. (1974) J. Mol. Cell. Cardio]. ¢~, 449 471. Trautwein, W. and Heschelcr, ,I. (l~J91)) Annu. Rex. f)llysiol. 52, 257 274. Trube, G. (1989) In: Piper, t].M., lsenberg, G. (eds.), Eleclrophysiology and contractile function, ('R(' Press, Boca Raton, FL, Vol. 2, pp. 75-95. Walter, G., lntek, A., Wobus, A.M. and Sch(:,neich, ,1. (1984)Cc]l Diff. 15, 147-151. Wobus, A.M., Holzhausen, 11., Jiikel. P. and Sch(hleich, J. (lOiN4} Exp. ('ell Res. 152. 212-21cj. Wobus, A.M., Grosse, R. and Sch6neich, ,1. (19881 Biomed. Biochim. Acta 47, 965 973. Wobus, A.M., Wallukal. G. and l leschclcr. ,I. (19911 Differentiation 48, 173-182. Wobus, A.M., Kleppisch, T., Maltse~, V. and llescheler. J. (submitred}. Wollcnberger, A. (1984) In: Abe. tt.. lto, Y., Tada, M., Opic, I,.tt. (eds.). Regulation of cardiac function. Molecular. cellular and pathophysiological aspects, Japan Scientific Soc. Press, T(~k~,o. VNU Science Press. BV Utrecht, pp. 269-284. Wollenberger. A. (1985) Basic Res. Cardiol. 81), Suppl. 2, 9 14.