Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Development 121, 2439-2450 (1995) Printed in Great Britain © The Company of Biologists Limited 1995 2439 Initiation of cardiac differentiation occurs in the absence of anterior endoderm Maureen Gannon* and David Bader† Department of Cell Biology and Anatomy, Cornell University Medical College, 1300 York Avenue, New York, NY 10021, USA *Department of Biochemistry, 420 Henry Mall, University of Wisconsin, Madison, WI 53706, USA †Author for correspondence at present address: Vanderbilt University School of Medicine, MRBII, Nashville, TN 37232, USA SUMMARY Anterior endoderm has been proposed to be a specific inducer of cardiac differentiation in vertebrates (reviewed in Jacobson and Sater, Development 104, 341-359, 1988). The ability of cardiogenic mesoderm to differentiate in a minimal culture system was examined using cardiacspecific gene expression as an assay. Anterior lateral plate mesoderm was explanted from chick embryos with and without associated endoderm at developmental stages from just after gastrulation (stage 4; Hamburger and Hamilton, J. Morph. 88, 49-67, 1951) to just prior to contraction (stage 9). At all stages examined, cardiogenic mesoderm expressed a profile of cardiac-specific mRNAs after two days in minimal medium independent of the presence of endoderm. Our studies indicate that endoderm is necessary for the generation of stable sarcomeric protein expression, organized myofibrils and beating tissue from stage 4-6. Subsequent to this stage, an interaction with anterior endoderm is no longer required. Examination of cardia bifida embryos from which anterior endoderm had been unilaterally removed also showed a stage-dependent effect of endoderm on beating, while cardiac gene expression and heart morphogenesis were unaffected. These results demonstrate that anterior endoderm does not induce or maintain cardiac gene expression, nor is it required for terminal differentiation. Endoderm does appear to be necessary for a short period of time between initiation of cardiac gene expression and the onset of contraction. INTRODUCTION 1993). Prior to gastrulation in avian embryos, prospective heart cells reside in the primitive streak, just caudal to Henson’s node, with the anterior endoderm arising from cells within the node itself (Selleck and Stern, 1991). In the chick, both prospective cardiac mesoderm and anterior endoderm have completed gastrulation by Hamburger and Hamilton (1951) stage 4 (Selleck and Stern, 1991; Garcia-Martinez and Schoenwolf, 1993). Studies in amphibian and chick embryos have shown that prospective cardiac mesoderm is specified to differentiate as heart by the time these cells have completed gastrulation (Sater and Jacobson, 1989; Gonzalez-Sanchez and Bader, 1990), yet they do not differentiate for several hours. Recent results demonstrate that cardiac primordia differentiate several hours before the initiation of contractions, expressing muscle- and cardiac-specific mRNAs and proteins (Bisaha and Bader, 1991; Han et al., 1992; Logan and Mohun, 1993: Yutzey et al., 1994). Thus, the differentiated, non-contractile cardiomyocyte is a distinct cell type within this lineage. Concomitant with the onset of differentiation at stage 7-8, cardiogenic mesoderm becomes a true epithelium and is separated from the more dorsal somatic lateral plate mesoderm by the formation of the coelom (Linask, 1992). Striated myofibrils do not become apparent until stage 9+, which is just prior to the first contractions (Han et al., 1992). Thus, the process of heart development occurs in several distinct steps. Single cell analysis of isolated stage 4 cardiogenic cells has Induction is the process by which the developmental fate of one cell population is regulated by signals provided by a different cell or tissue type. This signal could be mediated by direct cell-cell contact or by a diffusible molecule. In the absence of such a signal, the responding tissue fails to differentiate according to its predicted fate (for reviews see Jacobson and Sater, 1988; Slack, 1993). A prevailing concept in the literature is that the anterior or pharyngeal endoderm is a specific inducer of heart differentiation. This system has been used as an example to study inductive interactions during vertebrate development (see Jacobson and Sater, 1988). A role for pharyngeal endoderm in heart development is based on studies in amphibians and avians where, prior to neurulation, removal of the endoderm overlying the cardiogenic mesoderm resulted in a lack of beating heart tissue later in development (Jacobson, 1960; Orts-Llorca and Gil, 1965). However, the cellular and molecular mechanisms underlying this proposed inductive interaction are as yet unknown. The vertebrate heart tube is a mesodermally derived structure that arises from initially paired primordia located on either side of the embryonic midline. In vertebrate embryos, cells that will give rise to the heart are some of the first to gastrulate, taking up positions within the anterior lateral plate mesoderm (Rosenquist, 1985; Keller, 1976; Stanier et al., Key words: heart development, cardiac gene expression, induction, chick 2440 M. Gannon and D. Bader shown that newly gastrulated mesoderm is capable of myocyte differentiation in culture (Gonzalez-Sanchez and Bader, 1990). These results suggest that cardiac differentiation can occur in the absence of specific cell-cell interactions, but are complicated by the fact that these cells were grown in a serum-rich medium which may substitute for the proposed inductive effect of endoderm. Recent studies using cultured explants of cardiogenic mesoderm from avian embryos have supported an inductive role for endoderm in cardiac differentiation using beating and tissuespecific protein expression as an assay (Sugi and Lough, 1994; Antin et al., 1994). In addition, the recent isolation of novel homeobox genes expressed specifically in early heart primordia and the overlying pharyngeal endoderm, suggests a relationship between these two tissues (Lints et al., 1993; Tonissen et al., 1994). Still, the effect of pharyngeal endoderm on specific stages within the cardiogenic cell lineage is unresolved. While it is apparent that endoderm influences the process of heart development, it has never been shown that endoderm is essential for cardiomyogenic gene expression. We therefore examined the role of endoderm in the initiation of cardiacspecific gene expression as an assay for cardiac differentiation. In the chick embryo, cardiac gene expression begins at stage 7+, ten hours before the initial contractions at stage 10 (Bisaha and Bader, 1991; Han et al., 1992). In the present study, we examined the effect of endoderm removal on cardiac gene expression at all developmental stages from just after gastrulation of cardiac mesoderm (stage 4) to just prior to contraction (stage 9). Analysis of cardiomyogenic differentiation was conducted in a minimal culture system to eliminate the potential inductive effects of added factors. In contrast to previously published reports, we observed differentiation of cardiac mesoderm at high frequency in the absence of endoderm as measured by the expression of heart-specific genes. At all stages examined, endodermless explants expressed a profile of cardiac mRNAs. Still, corresponding proteins were found to assemble into sarcomeric structures only in endodermless explants removed after stage 6. We also observed a stage-dependent effect of endoderm on the ability of cardiogenic mesoderm to beat in vitro and in vivo. These results demonstrate that endoderm does not induce cardiac gene expression and that sustained contact with endoderm is not necessary to maintain gene expression. In addition, our data demonstrate that the requirement for endoderm in cardiomyogenesis is transient in that after stage 6 it is not necessary for terminal differentiation. Thus, the interaction between cardiogenic mesoderm and endoderm is not needed for initiation of cardiac-specific gene expression or terminal differentiation, but instead may play a role in maturation of this cell lineage. MATERIALS AND METHODS Embryos and cultures Fertilized White Leghorn chicken eggs (Truslow Farms, Chesterton, Md.) were incubated at 37°C with high humidity for 23-28 hours. Staging was according to Hamburger and Hamilton (1951). Embryos were removed from the egg on filter paper rings, rinsed with PBS [150 mM NaCl, 10 mM Na2HPO4, 10 mM NaH2PO4 (pH 7.4)] and placed ventral side up in culture dishes prior to dissection. Dissections were performed using needles made from glass capillary tubes (1.0 mm) drawn on a vertical pipette puller (Kopf Instruments). Location of cardiac primordia was based on published fate maps (Rawles, 1943; DeHaan, 1963). It has previously been shown that cardiac differentiation begins at stage 7+ and proceeds in a rostral to caudal direction (Bisaha and Bader, 1991; Han et al., 1992; Yutzey et al., 1994). Only those progenitors that had not begun to differentiate were removed for analysis. For mesoderm/endoderm explants, the endodermal and mesodermal cell layers were cut through without cutting the underlying ectoderm. The top two layers were then peeled off together, leaving the ectoderm behind. For endodermless explants, the endoderm was first mechanically peeled away as an intact sheet from the underlying mesoderm and, in most cases, discarded. In some cases, the endodermal sheet was removed for culture. The mesoderm and ectoderm were then removed together. Mesoderm removed from embryos at stages 4-6 dissociates into single cells if cultured alone and thus requires the presence of ectoderm in order to survive in a minimal culture system. This has been reported by others (Yamazaki and Hirakow, 1991; Antin et al., 1994). In cases where the endoderm did not come off as an intact sheet of cells, the sample was discarded. In all cases, explants were transferred in PBS to a chamber slide containing M199 defined minimal medium (Sigma). Permanox Lab-Tek chamber slides (Nunc) were coated with 0.01% collagen I (Sigma). Medium consisted of 1× M199 plus 100 µg/ml penicillin and 100 u/ml streptomycin (Sigma). No other supplements were added. Tissue explants were incubated at 37°C with high humidity and 5% CO2 for approximately 48 hours. All explants were assayed for beating at 24 and 48 hours. Incubations of up to 4 days did not yield different results and thus all explants were cultured for 48 hours. After 48 hours, explants were rinsed three times with sterile PBS and fixed as below. For whole embryo culture, removal of endoderm from one side of the cardiogenic region was performed as described above. All dissections were photographically monitored to confirm the embryonic stage at the time of dissection as well as the location of the dissection in the embryo. Embryos were then cultured ventral side up on paper rings until they reached stage 7 (approximately 5 hours), at which time cardia bifida was generated by clipping the anterior intestinal portal with a glass needle. Embryos were cultured in 50:50; M199 (plus 100 µg/ml penicillin and 100 u/ml streptomycin):thin egg albumin for 24 hours at 37°C with high humidity and 5% CO2. After 24 hours, embryos were examined for the presence of two heart tubes and beating of both tubes. Embryos were then rinsed three times with sterile PBS, photographed and fixed as below. In situ analysis of mRNA expression and sectioning of embryos Explants or embryos were fixed in 4% paraformaldehyde (Prill, Electron Microscopy Sciences) in PBS at room temperature for 2030 minutes then transferred to 70% ethanol at −20°C and stored for later use (up to two weeks). Prior to in situ hybridization, embryos were dissected away from extraembryonic membranes. Digoxigenin UTP-labelled sense and antisense riboprobes were prepared using the Boehringer Mannheim Genius 4 system. Ventricular myosin heavy chain (VMHC1; Bisaha and Bader, 1991) and atrial myosin heavy chain (AMHC1; Yutzey et al., 1994) probes were prepared as described in Yutzey et al., 1994. Cardiac troponin I (cTnI; Hastings et al., 1991) probes were generated using the SP6 and T7 promoters of pGEM4 (Promega) vectors containing the 5′, 680 bp of the quail cTnI cDNA linearized using EcoRI or BamHI. Cardiac C-protein probes were generated using the T7 and T3 promoters of pBluescript II (Stratagene) vectors containing the 3′, 1.4 kb of the chicken cDNA linearized using EcoRI and HindIII (a gift of Dr Takashi Obinata). VMHC1 is expressed at the onset of cardiomyocyte differentiation (Bisaha and Bader, 1991), as is cTnI (Gannon and Bader, unpublished observations). AMHC1 is expressed in a subpopulation of cardiac myocytes (Yutzey et al., 1994). Cardiac C-protein is expressed in the heart just prior to the initiation of contractions (Gannon and Bader, unpublished observations). In situ hybridizations on whole embryos were performed as Cardiac differentiation without endoderm 2441 described previously (Continho et al., 1992, 1993; Yutzey et al., 1994). In situ hybridizations on tissue explants were performed with the following modifications. The xylene clearing step was omitted and substituted with a 30 minute treatment with 100% ethanol. Proteinase K treatment was reduced to 3 µg/ml for 3-5 minutes at room temperature. Explants were hybridized in 150-200 µl probe at 63°C for 1820 hours in a humid chamber. The probe concentration was initially 5-10 µg/ml and probes were reused up to eight times. Incubation with anti-digoxigenin antibody was 2 hours at room temperature and wash times were halved prior to color development. Following in situ hybridization, all embryos and explants were photographed on a Nikon dissecting scope. Subsequent to in situ analysis, whole chick embryos were dehydrated in an ethanol series, transferred to xylene and embedded in paraplast. Serial sections were cut at 8 µm and mounted on Superfrost/Plus slides (Fisher). Sections were deparaffinized, rehydrated and stained in a 10 µg/ml solution of ethidium bromide to visualize nuclei. RT-PCR analysis Total RNA was isolated from single explants of stage 4-6 cardiogenic mesoderm with or without associated endoderm immediately upon explantation or following 48 hours of culture in minimal medium as in Chomczynski and Sacchi (1987). Single-stranded cDNA was synthesized using oligo(dT) primer and the Superscript Preamplification System (Gibco BRL). First-strand cDNA was divided into two samples of equal volume and amplified for 25 cycles using Taq DNA polymerase (Boehringer-Mannheim) with VMHC1-specific primers as in Bisaha and Bader (1991) or primers specific for glyceraldehyde phosphate dehydrogenase [GAPD 5′: 5′-GGGTCTTATGACCACTGTCC-3′ (nt566-585); GAPD 3′: 5′-GTAAGCTTCCCATTCAGCTCAG-3′ (nt736-715)]. Control experiments determined that accumulation of amplified products was still within the linear range over 30 cycles. Amplification products were separated on a 1% agarose gel and analyzed by Southern blot (Maniatis et al., 1982) using a radiolabeled probe consisting of either a 251 bp fragment of the VMHC1 cDNA (Bisaha and Bader, 1991) or the entire quail GAPD cDNA (a gift of Dr Charles P. Emerson). Hybridization and washing conditions were as in Bisaha and Bader (1991). A band of 208 bp is predicted for the VMHC1 transcript, corresponding to the 3′ end of the cDNA. A band of 170 bp is predicted for the GAPD transcript and was used to control for equal amounts of first strand cDNA used in the amplification reaction. As a control, reactions were performed in the absence of reverse transcriptase and amplified using VMHC1 or GAPD primers. Amplification products were never detected in reactions without reverse transcriptase. Immunohistochemistry Whole embryos or tissue explants were fixed in 70% methanol for 1 hour at room temperature, rinsed several times with PBS and then blocked with 1%BSA in PBS 1 hour at room temperature. Embryos or explants were incubated in MF-20 hybridoma supernatant (Bader et al., 1982) undiluted overnight at 4°C. Specimens were washed all day with several changes of PBS and then incubated overnight at 4°C in goat anti-mouse IgG conjugated to either FITC or Texas Red (Jackson Labs) reconstituted according to the manufacturer’s instructions and diluted 1:40 (FITC) or 1:50 (Texas Red) in 1% BSA in PBS. For immunohistochemistry on explants following in situ hybridization, explants still attached to chamber slides were rinsed several times with PBS and then blocked with 1% BSA in PBS for 30 minutes at room temperature. Explants were incubated in MF-20 for 1 hour at room temperature, rinsed several times with PBS and then incubated in Texas Red-conjugated secondary antibody as above for 1 hour at room temperature. All embryos and explants were fixed in neutral buffered formalin 30 minutes at room temperature following removal of secondary antibody. Specimens were mounted using 50:50, glycerol: PBS and photographed under epi-fluorescence. RESULTS Cardiogenic mesoderm differentiates reliably in vitro Anterior lateral plate mesoderm used for the analysis of cardiac differentiation in vitro has been shown by fate mapping to incorporate into the heart tube in vivo (Gonzalez-Sanchez and Bader, 1990). Thus, we first examined whether cardiogenic mesoderm differentiates as reliably in vitro in minimal medium. Undifferentiated cardiogenic mesoderm plus its associated anterior endoderm was removed from either side of the cardiogenic crescent from embryos of stages 4-9. Fig. 1 shows the location of the cardiac primordia at stage 4, the earliest stage at which cardiogenic lateral plate mesoderm can be removed. After 48 hours of culture, explants from early embryonic stages (4-6) were observed to beat in 43% of the cases, and explants removed from later stages (7-9) beat in 54% of the cases (Table 1). These results suggested that explants of cardiogenic mesoderm were unable to differentiate terminally and function predictably in a minimal culture system. As an alternative to a functional assay, and to assess the differentiative capacity of cardiogenic mesoderm further, explants were fixed after 48 hours and analyzed for sarcomeric myosin heavy chain protein expression using the MF-20 antibody (Bader et al., 1982). In addition, explants were examined for the expression of one of four cardiac-specific mRNAs using digoxigenin-labelled riboprobes. These results are shown in Fig. 2 and Table 1. Regardless of their ability to beat, explants of cardiogenic mesoderm plus endoderm expressed myosin heavy chain protein (Fig. 2B), and this protein was localized to cross striations (Fig. 2G). In situ hybridization analysis revealed that these explants express the mRNAs of cardiac-specific myosin heavy chain isoforms, VMHC1 and AMHC1 (Fig. 2C,D), as well as mRNAs for the cardiac isoforms of troponin I and myosin binding protein-C (Fig. 2E,F). In situ hybridization with sense probes revealed no detectable background staining while the same sample showed a high level of myogenic differentiation as detected with the MF20 antibody (Fig. 3). Table 1 summarizes the analysis of cardiac-specific protein and mRNA expression in cultures of mesoderm plus endoderm, showing that both are expressed in Fig. 1. Location of cardiac primordia. Schematic diagram of a stage 4 embryo. One side of the precardiac mesoderm is indicated by the stippled region. Boxed area indicates the region removed for explant cultures. 2442 M. Gannon and D. Bader Table 1. Analysis of cardiac differentiation in vitro Explant Mesoderm/Endoderm Mesoderm/Ectoderm Mesoderm/Endoderm Mesoderm/Ectoderm Embryo stage Cardiac mRNA % (n)* Myosin protein % (n) Striations % (n) Beaing % (n) 4-6 4-6 7-9 7-9 86 (35) 67 (27) 88 (35) 95 (22) 88 (8)† 12 (8)† 100 (16) 100 (22) 81 (16)‡ 0 (8)‡ 100 (4) 75 (4) 44 (43)§ 0 (35)§ 53 (51) 28 (35) Comparisons were made between groups of explants of the same stages with or without endoderm. Only those P values for statistically different groups are shown. Statistical analysis to compare the difference between two groups was performed with the proportion test using normal approximation with Yates’ continuity correction. P value is given for significant differences between groups where P<0.01. *Percent positive with total number of explants examined (n) is indicated. †P=0.01. ‡P=0.001. §P=0.001. >85% of the explants. These results demonstrate that undifferentiated cardiogenic precursors are able to differentiate consistently in a minimal defined culture medium and express a profile of cardiac-specific genes in the presence of anterior endoderm. It is clear from this data that beating does not accurately reflect whether these cells express cardiac-specific genes and muscle structural proteins and, thus, whether the cells have differentiated. A molecular analysis of differentiation is therefore a more reliable indicator of the presence of differentiated cardiomyocytes than assaying for function. Cardiac-specific gene expression is activated in the absence of endoderm The anterior or pharyngeal endoderm has been proposed to be a specific inducer of cardiac differentiation as indicated by the lack of formation of beating tissue and striated myofibrils when endoderm is removed early in development (reviewed in Jacobson and Sater, 1988; Yamakazi and Hirakow, 1991; Sugi and Lough, 1994; Antin et al., 1994). As cardiac gene expression is a more reliable indicator of the differentiated state, we determined whether cardiogenic mesoderm was capable of initiating tissue-specific gene expression in the absence of anterior endoderm. In order to characterize the possible inductive effect of endoderm on cardiac differentiation, we examined all developmental stages from the time just after gastrulation (stage 4) to just before contractions begin in vivo (stage 9). Endoderm overlying the cardiogenic mesoderm was removed as an intact sheet of cells as shown in Fig. 4A. To determine whether endoderm dissections were contaminated with underlying cardiogenic mesoderm, the endodermal cell layer was cultured and assayed for VMHC1 mRNA expression or myosin heavy chain protein. Such cultures were routinely negative for myocytes (Fig. 4C). Following endoderm removal, undifferentiated cardiogenic mesoderm was removed and cultured in minimal medium for 48 hours along with its associated ectoderm. Cultured endodermless explants were assayed for cardiac differentiation by analyzing the expression of VMHC1, AMHC1, cTnI or cardiac C-protein mRNAs using in situ hybridization. As shown in Fig. 5, explants of cardiogenic mesoderm express all four cardiac-specific genes in the absence of endoderm, even when removed from the embryo just after these cells have gastrulated at stage 4. While these explants were not beating at the time of fixation, 73% expressed cardiac-specific transcripts (Table 1). These results were confirmed using RT-PCR analysis and assaying for VMHC1 expression (Fig. 6). Explants of cardiogenic mesoderm removed at stage 4-6 do not express the VMHC1 message either in the absence or presence of endoderm (left panel, lanes 1, 3). At long exposures, a faint band was detectable in reactions from uncultured explants, but this hybridization signal was equivalent to that seen in reactions lacking RT (data not shown). This confirms that explants contained only undifferentiated progenitors at the time of their removal. Following 48 hours of culture in minimal medium, VMHC1 mRNA was expressed in endodermless explants (lane 2) as well as those containing endoderm (lane 4). Identical results were obtained in four individual experiments. Expression of GAPD mRNA was used as a control to show roughly equivalent input of first strand cDNA from each explant (Fig. 6, right panel). These data show unequivocally that prospective cardiac mesoderm is able to differentiate in minimal culture medium in the absence of any inductive influences from the endoderm at the earliest stages at which this mesodermal tissue can be removed from the embryo. Furthermore, this differentiation includes the expression of several genes normally expressed in the heart and whose protein products are localized to both thick and thin filaments within the sarcomere. Endoderm is required transiently during cardiac differentiation Once we had determined that cardiogenic mesoderm was capable of tissue-specific gene transcription in the absence of endoderm, we examined myosin heavy chain protein expresson in stage 4-9 explants. While 73% of cardiogenic mesodermal explants removed at stages 4-6 expressed cardiac-specific mRNAs, myosin heavy chain protein expression was detected in only 12% of these cultures (Table 1). Fig. 7B shows an example of a stage 4 explant cultured for two days and reacted with the MF-20 antibody. This explant demonstrated the strongest protein staining observed at stages 4-6. The protein appeared punctate and diffuse within the cytoplasm and was not organized into striated myofibrils (Fig. 7C). In contrast, 100% of endodermless explants removed at later stages (7-9) exhibit strong protein expression (Fig. 7E) which is organized into cross striations similar to those seen in cultures with endoderm (Fig. 7F). These results demonstrate that prior to stage 7, an interaction with endoderm appears to be essential in order for striated myofibrils to form in cultured cardiogenic Cardiac differentiation without endoderm 2443 Fig. 2. Explanted cardiogenic mesoderm plus endoderm expresses the cardiomyogenic phenotype. Undifferentiated explants containing mesoderm and endoderm were removed from the embryo at stages 4-7 and analyzed for cardiac differentiation after 48 hours culture in minimal medium. Explant removed at stage 4 (A) expresses sarcomeric myosin heavy chain protein (B) as detected by the MF-20 antibody. Higher magnification showed that this protein is organized into cross striations (G). Explants from stage 7 (C,D) express ventricular (C) and atrial (D) myosin heavy chain mRNAs. Explants from stage 5 (E) and stage 6 (F) express cardiac TnI (E) and cardiac Cprotein (F) mRNAs. Transcripts were detected using digoxigenin labelled riboprobes. All explants were beating at the time of fixation. In C-F, phase-contrast images are shown with insets taken under bright-light illumination. 2444 M. Gannon and D. Bader mesoderm as sarcomeric proteins detected are in a non-myofibrillar pattern. However, endoderm is not required during terminal differentiation of cardiac mesoderm removed at stages 7-9 (ie. generation of striated myofibrils). Previously published studies in amphibians and avians revealed a role for anterior endoderm in heart morphogenesis and contractility prior to neurulation (Jacobson, 1960; Orts- Llorca and Gil, 1965; Yamakazi and Hirakow, 1991; Sugi and Lough, 1994; Antin et al., 1994). In agreement with these studies, removal of endoderm from the cardiogenic region had a drastic effect on the ability of cultured cardiogenic mesoderm to beat (Table 1). This effect is most apparent in explants of stage 4-6 mesoderm where, in this culture system, endoderm is absolutely required for beating. Cardiogenic mesoderm explanted at stages 7-9 and cultured without endoderm shows an increased incidence of beating (28%). The striated pattern of myosin expression observed in cardiac mesoderm cultures from stage 7-9 embryos occurs independent of beating as 75% of these cultures contained striated myofibrils (Table 1). Thus, the present results confirm that the presence of anterior endoderm is required for beating to occur prior to stage 7 and also suggest that the interaction between endoderm and mesoderm is not necessary for beating after stage 6. To determine whether endoderm has the same effect on heart differentiation and organogenesis in the context of the whole embryo, we examined cardiac differentiation in vivo by analyzing bifid hearts from which the anterior endoderm had been unilaterally removed (see Methods and Fig. 8). We used Fig. 3. Sense riboprobes show no detectable staining in differentiated explants. Stage 9 cardiac mesoderm plus endoderm was removed from the embryo and cultured for 48 hours in minimal medium. The explant showed no reactivity with the VMHC1 sense riboprobe (A). This explant was beating at the time of fixation and expressed myosin heavy chain protein as detected by the MF20 antibody (B). Fig. 4. Dissection of anterior endoderm is free of underlying mesoderm. Endoderm can be separated from cardiogenic mesoderm as an intact sheet as early as stage 4 (A). This endoderm is free of contaminating cardiac mesoderm as shown by a lack of expression of VMHC1 mRNA (arrow in C). Ecto, ectoderm; meso, mesoderm; endo, endoderm; HN, Henson’s node; PS, primitive streak. Cardiac differentiation without endoderm 2445 Fig. 5. Cardiogenic mesoderm explanted just after gastrulation expresses cardiac-specific genes in the absence of endoderm. Undifferentiated explants containing mesoderm plus ectoderm were removed from the embryo at stage 4 (A,B) or stage 5 (C,D) and analyzed for cardiacspecific transcripts by in situ hybridization after 48 hours culture in minimal medium. These explants express ventricular (A) and atrial (B) myosin heavy chains as well as cardiac TnI (C) and cardiac C-protein (D) mRNAs. a protocol that insured that the endoderm removed did not contain any contaminating mesoderm (see Fig. 4C). Fig. 9A,D,G depicts three embryos at the time of dissection. Embryos were cultured as described in the Methods for 24 hours at which time they have reached the equivalent of stage 10-11. Sectioning of embryos showed that the endoderm had not grown back on the operated side in the majority of cases (Fig. 10). In all cases in which cardia bifida was successfully generated (n=61), heart tubes were apparent on both the operated and unoperated side of the embryo (Fig. 9B,E,H), suggesting that removal of endoderm did not prevent heart tube formation. This was true at all stages examined. Control heart tubes were beating after 24 hours in 73% of the cases from stage 4-6 embryos (n=22) and 97% of the cases from stages 79 (n=39). In agreement with the results obtained in vitro, heart tubes on the operated side failed to beat if the endoderm was removed prior to stage 7 (9% of the cases were beating), but showed an increased ability to beat if the endoderm was removed after this stage (59%). These in vivo results also demonstrate the transient nature of endoderm’s effect on heart development in that its presence is not required during terminal differentiation. Culturing the embryos for an additional 24 Fig. 6. RT-PCR analysis of VMHC1 and GAPD expression in cardiogenic mesoderm cultured in the absence of endoderm. Total RNA isolated from single explants of stage 4-6 cardiogenic mesoderm does not contain VMHC1 mRNA at the time of explantation in the absence (left panel, lane 1) or presence (lane 3) of anterior endoderm. Explants cultured for 48 hours express VMHC1 message in the absence of endoderm (lane 2) as do explants cocultured with endoderm (lane 4). Analysis of GAPD expression in the same explants demonstrates that this message is present in roughly comparable amounts under all conditions examined (right panel). 2446 M. Gannon and D. Bader Fig. 7. Endoderm is not required for myosin heavy chain protein expression in explants of cardiogenic mesoderm but is necessary for myofibril formation prior to stage 7. Undifferentiated explants containing mesoderm and ectoderm were removed at stage 4 (A-C) or stage 7 (D-F) and assayed for sarcomeric myosin heavy chain protein expression using the MF-20 antibody. At stage 4, myosin heavy chain could be detected in a small percentage of explants (B). Higher magnification showed this expression was punctate and diffuse in the cytoplasm (C). At stage 7, strong myosin heavy chain staining is seen (E) and this protein is localized to cross striations (F). hours did not increase the incidence of beating on the operated side demonstrating that these heart tubes were not merely delayed in their development. As an additional control, in some embryos, the endoderm was mechanically peeled away from one side and then immediately replaced. These embryos developed beating heart tubes on both sides showing that the observed results are not due to the operation itself damaging the cardiogenic mesoderm. When experimental embryos were assayed for myosin heavy chain expression, heart tubes on the operated and unoperated sides demonstrated the presence of both protein and mRNA in 100% of the embryos examined (n=21) (Fig. 9C,F,I). In contrast to the results obtained in vitro, heart tubes lacking endoderm showed a level of protein expression comparable to that of the unoperated heart tube (Fig. 9C; discussed below). In both heart tubes, expression of AMHC1 mRNA was restricted to the posterior cardiac myocytes where it is normally expressed (Fig. 9I; Yutzey et al., 1994). The larger domain of AMHC1 expression seen on the operated side of the embryo in Fig. 9I is most likely due to the greater population of preatrial cells located within this side of the cardiogenic crescent (Stalsberg and DeHaan, 1969) and not an effect of endoderm removal, as control cardia bifida embryos show the same expression pattern Fig. 8. Diagram showing unilateral removal of anterior endoderm and generation of cardia bifida. Following removal of endoderm from one side of the cardiogenic region at stage 4, embryos were allowed to develop to stage 7 (approximately 5 hours) and cardia bifida generated by clipping the anterior intestinal portal. Embryos were then cultured for 24 hours until controls reached stage 10-11, at which time they were assayed for: (1) the presence of two heart tubes, (2) beating of both tubes and (3) cardiac differentiation using MF-20 or in situ hybridization. Cardiac differentiation without endoderm 2447 (data not shown). These results demonstrate that removal of endoderm in vivo does not inhibit cardiac-specific mRNA and protein expression or heart tube formation, but does have a stage-dependent affect on beating. Taken together with the in vitro data, we have shown that endoderm is not required for initiation of cardiac gene expression, nor is it necessary during the terminal stages of heart differentiation which include myofibrillogenesis and contraction. DISCUSSION The interaction between endoderm and mesoderm in heart differentiation has been proposed as a model for embryonic induction although the cellular and molecular mediators of this proposed induction have yet to be identified. An inductive role for endoderm in the process of cardiac differentiation is based on the fact that beating occurs only after sustained contact between these two tissues. We have examined this process using the initiation of cardiac-specific gene expression as the first indicator of cardiac mesoderm differentiation. Our data show that sustained contact between anterior endoderm and cardiogenic mesoderm is not necessary for the activation of heart-specific gene transcription. Thus, endoderm is not an inducer of cardiac differentiation. Our present results show a brief requirement for endoderm at approximately stage 7, influencing maturation of the cardiac cell lineage in order to subsequently generate organized striated myofibrils and beating heart tissue. After this time, contact of cardiac mesoderm with endoderm is not necessary to undergo terminal differentiation. We and others have observed that undifferentiated cardiogenic mesoderm is capable of terminal differentiation when cultured with its associated endoderm (Gonzalez-Sanchez and Bader, 1990; Yamakazi and Hirakow, 1991; Sugi and Lough, 1994; Antin et al., 1994). The defined minimal medium used in the present study contains the necessary nutrients to support cardiomyocyte differentiation from previously undifferentiated anterior lateral plate mesoderm in contact with endoderm. Although the percentage of explants observed to beat does not exceed 54%, greater than 85% express myosin heavy chain protein organized in the cross striations characteristic of a mature myocyte. In addition, >85% of these cultures express mRNAs for muscle- and cardiac-specific genes, regardless of their ability to beat. The four transcripts detected in these explants have been shown to be expressed in the differentiating heart in vivo (Bisaha and Bader, 1991; Yutzey et al., 1994; Gannon and Bader, unpublished observations). The high incidence of differentiation observed in vitro is not surprising as these cells will reliably incorporate into the definitive heart tube in vivo (Gonzalez-Sanchez and Bader, 1990). These results demonstrated that detection of cardiac-specific gene expression is a more reliable assay for differentiation than beating alone and that anterior lateral plate mesoderm consistently differentiates into cardiomyocytes in culture. Previous studies have led to the concept that anterior endoderm is essential for induction of cardiomyogenic differentiation (Jacobson, 1960; Orts-Llorca and Gil, 1965; Yamakazi and Hirakow, 1991; Sugi and Lough, 1994). In order to assess the role of anterior endoderm in cardiac differentiation, we assayed for the ability of cardiogenic mesoderm, removed at specific stages after gastrulation, to activate cardiac- specific gene expression in culture in the absence of associated endoderm. The explants of cardiogenic mesoderm used in this study, contained only undifferentiated progenitors. Thus, any synthesis of cardiac mRNAs or proteins detected in these explants was initiated during the culture period. Expression of four known muscle- and/or cardiac-specific mRNAs, was detected in endodermless explants of cardiac mesoderm from all developmental stages examined (stages 4-9). Because cardiac gene expression normally commences at stage 7-8 within the anterior cardiogenic mesoderm in the intact embryo (Bisaha and Bader, 1991; Han et al., 1992; Yutzey et al., 1994), stage 4-6 mesodermal explants are removed from the influence of endoderm and other embryonic structures hours before they begin to differentiate, yet are still capable of initiating the cardiac differentiation program. Still, cultures of stage 4-6 explants do not contain striated myofibrils and fail to beat. The role of endoderm in generating striations and beating of differentiated heart tissue in cultured explants appears to be stagedependent. Cultured explants from stage 7-9 embryos contain functional striated myofibrils in the absence of endoderm (see Table 1). A stage-dependent effect of endoderm on beating in vitro has been observed by others (Smith and Armstrong, 1990; Antin et al., 1994). Striated cardiomyocytes are first detected in vivo at stage 9+, just prior to the initiation of contraction (Han et al., 1992). Thus, sustained contact between endoderm and mesoderm is not necessary for the activation of cardiac genes, nor is it required to establish functional myofibrils. A brief interaction is required at approximately stage 7 possibly to stabilize or potentiate terminal differentiation. This stage-dependent effect of endoderm on beating of cardiac mesoderm was also observed in vivo. While there have been conflicting reports in the methods used to produce endodermless bifid hearts (Orts-Llorca, 1963; DeHaan, 1964), our data clearly show that heart tube formation and expression of cardiac genes and protein occurs in the absence of endoderm. It is possible that the structure of the heart tube accounts for the stable expression of sarcomeric proteins not observed in vitro. Still, as seen in vitro, there is a transient dependence on anterior endoderm after differentiation and prior to contraction in order for beating to occur. This interaction is no longer necessary after stage 6, several hours before the initiation of contraction. These results show that differentiation and organogenesis do not require sustained contact with endoderm in vivo, but that endoderm specifically affects beating of differentiated cardiac tissue. Expression of cardiac muscle genes without contraction should not be unexpected. Several recent studies have shown that muscle-specific gene expression precedes contractility by several hours (Bisaha and Bader, 1991; Han et al., 1992; Logan and Mohun, 1993; Yutzey et al., 1994) and that the initial localization of myosin heavy chain protein is non-myofibrillar (Han et al., 1992). Thus, the non-contractile differentiated myocyte is a distinct member of the cardiogenic cell lineage. Studies in the axolotl have led to the suggestion that cardiomyocyte differentiation is a two-step process (Smith and Armstrong, 1990). The authors proposed that, in the first inductive step, contractile proteins are expressed but not organized into functional myofibrils. Organized myofibrils were thought to be promoted or stabilized by the presence of a second inducing substance. The present data suggest that endoderm is not involved with initial differentiation (i.e. activation of cardiac 2448 M. Gannon and D. Bader Cardiac differentiation without endoderm 2449 genes). If endoderm provides the proposed second signal for later events in cardiogenesis, this signal is transmitted during a brief time frame at or around stage 7, as sustained contact is not required for terminal differentiation. Upon examining the expression of myosin heavy chain protein in cultures of differentiated cardiac mesoderm, it was found that organized cross striations were present only in those explants removed from stage 7-9 embryos. In contrast, explants removed from stage 4-6 embryos showed diffuse cytoplasmic localization of sarcomeric myosin heavy chain protein. Previous studies in which endodermless cultures of cardiogenic mesoderm were assayed for the presence of cardiac structural proteins such as sarcomeric alpha-actin and light meromyosin, showed a lack of protein expression in explants removed from embryos prior to stage 6 (Sugi and Lough, 1994; Antin et al., 1994). Upon close examination, however, what was interpreted to be negative staining may in fact be positive protein expression. Although this protein was not organized into the cross striations characteristic of a functional cardiomyocyte, definitive non-myofibrillar staining was apparent. The unstable protein expression observed in endodermless explants can be rescued with the addition of serum suggesting that the mediator of these later inductive events may be a soluble factor (Gonzalez-Sanchez and Bader, 1990; Yamakazi and Hirakow, 1991; Antin et al, 1994). The results of the present study demonstrate that, in the absence of exogenous factors or tissue interactions, specifically, anterior endoderm, cardiogenic mesoderm is able to differentiate into a non-contractile cardiomyocyte, but that additional factors, possibly provided by the endoderm, are necessary to stabilize translated structural proteins and generate a striated contractile myocyte. The influence of anterior endoderm on cardiac development seems to occur subsequent to the initiation of cardiac gene expression and structural protein synthesis, and to be more involved in later aspects of cardiomyocyte maturation and function such as stabilization of structural proteins for organization into functional myofibrils. The formation of the coelom and epithelialization of the cardiogenic mesoderm occurs during stages 6-8 (Linask, 1992), the time when cardiogenic mesoderm becomes independent of the requirement for endoderm in order to form striated myofibrils and contract. These events may therefore be dependent on diffusible factors from, or cell-cell contact with, the anterior endoderm. Removal of endoderm from undifferentiated stage 7-9 cardiogenic mesoderm does not effect protein stability or organization into striated myofibrils. Therefore cardiogenic mesoderm does not require a sustained signal emanating from the anterior endoderm after stage 7. This suggests that the time frame for generation of and/or response to an endodermal signal is short-lived. Fig. 9. Removal of endoderm does not affect heart tube formation or differentiation in vivo. Endoderm was unilaterally removed from the cardiogenic region at stage 4 (arrowheads in A,D) or stage 7 (arrowhead in G) and, following cardia bifida, embryos were allowed to develop 24 hours to the equivalent of stage 10-11 (B,E,H). In each case, heart tubes were visible on both the unoperated and operated sides (arrows in B,E,H). Heart tubes on the unoperated side were beating at the time of fixation, those on the operated side were not. Both heart tubes expressed myosin heavy chain protein as detected by the MF-20 antibody (C), and ventricular (F) and atrial (I) myosin heavy chain mRNAs as detected by whole-mount in situ hybridization. HN, Henson’s node; PS, primitive streak; S, somite. Fig. 10. Section through a cardia bifida embryo showing failure of endoderm to regenerate on the operated side. Endoderm was removed unilaterally at stage 4 from the right side of the embryo. Cardia bifida was generated at stage 7 and the embryo allowed to develop an additional 24 hours. At the time of fixation, the operated heart tube on the right was not beating while the control heart tube on the left was beating. Both heart tubes expressed the AMHC1 mRNA as detected by whole-mount in situ hybridization (not shown in this section). (A) Ethidium bromide staining of nuclei in a crosssection through the heart region of this embryo showing the location of embryonic structures. (B) Drawing of same section showing limit of endoderm on the right side and failure of this tissue to extend beneath the heart tube as seen on the unoperated side. NT, neural tube; n, notochord; hm, head mesenchyme; Ec, ectoderm; En, endoderm; H, heart tube. While the present results clearly show that endoderm is not necessary for initiation of cardiac gene expression, it remains possible that endoderm exerts an effect earlier in development than we have been able to examine. For example, endoderm may play a role in the initial commitment of lateral plate mesodermal cells to the cardiac lineage. While we can not exclude this 2450 M. Gannon and D. Bader possibility in the present study, removal of the anterior endoderm at stage 4 results in these two cell layers being in contact for only a brief period. We are able to remove cardiogenic precursors at early stage 4, just after these cells have gastrulated. Prior to this stage, both the anterior lateral plate mesoderm and the endoderm reside in the primitive streak (Selleck and Stern, 1991; Garcia-Martinez and Schoenwolf, 1993). These two cell populations are located in non-overlapping regions within the primitive streak (Selleck and Stern, 1991), making an interaction between them unlikely during gastrulation. In support of this, it has been shown in Xenopus that the superficial pharyngeal endoderm is not required for heart mesoderm specification during gastrulation (Sater and Jacobson, 1989). These results were complicated by the fact that the deep endoderm remained in contact with prospective heart mesoderm. It has recently been demonstrated that deep endoderm enhances the formation of beating hearts in Xenopus and that this function is also transient in nature (Nascone and Mercola, 1995). The process of cardiac development occurs in several distinct steps, which are becoming increasingly more defined. These include commitment to the cardiac cell lineage, initiation of cardiac gene expression, epithelialization of the cardiogenic mesoderm, structural protein expression, myofibril formation and contraction. The current study demonstrates that cardiogenic precursors can be isolated that differentiate to particular points along this developmental pathway. While our present data cannot exclude a role for anterior endoderm in the commitment of lateral plate mesoderm to the cardiac cell lineage, it is clear that endoderm is not required for the conversion of commited cardiac precursors to differentiated cardiomyocytes as determined by the expression of cardiacspecific genes. The point at which endoderm is involved in later aspects of cardiac differentiation and function is yet to be determined. We are grateful to Dr Takashi Obinata for the cardiac C-protein cDNA and to Dr Charles P. Emerson for the GAPD plasmid. We thank K. E. Yutzey and members of the Bader lab for valuable discussions, and T. Gallagher for assistance in preparation of figures. This work was funded by NIH grants HL34318 and HL37675 to D. B. REFERENCES Antin, P. B., Taylor, R. G. and Yatskievych, T. (1994). Precardiac mesoderm is specified during gastrulation in quail. Dev. Dynamics 200, 144-154. Bader, D., Masaki, T. and Fischman, D. A. (1982). Immunochemical analysis of myosin heavy chain during avian myogenesis in vivo and in vitro. J. Cell Biol. 95, 763-770. Bisaha, J. G. and Bader, D., (1991). Identification and characterization of a ventricular-specific avian myosin heavy chain, VMHC1: Expression in differentiating cardiac and skeletal muscle. Dev. Biol. 148, 355-364. Chomczynski, P. and Sacchi, N. (1987). Single-step method of RNA isolation by guanidinium-thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162, 156-159. Coutinho, L. L., Morris, J. and Ivarie, R. (1992). Whole mount in situ detection of low abundance transcripts of the myogenic factor qmf1 and myosin heavy chain protein in quail embryos. Biotechniques 13, 722-724. Coutinho, L. L., Morris, J., Marks, H. L., Buhr, R. J. and Ivarie, R. (1993). Delayed somite formation in a quail line exhibiting myofiber hyperplasia is accompanied by delayed expression of myogenic regulatory factors and myosin heavy chain. Development 117, 563-569. DeHaan, R. L. (1963). Organization of the cardiogenic plate in the early chick embryo. Acta Embryol. et Morph. Exp. 6, 26-38. DeHaan, R. L. (1964). Cell interactions and oriented movements during development. J. Exp. Zool. 157, 127-138. Garcia-Martinez, V. and Schoenwolf, G. C. (1993). Primitive streak origin of the cardiovascular system in avian embryos. Dev. Biol. 159, 706-719. Gonzalez-Sanchez, A. and Bader, D. (1990). In vitro analysis of cardiac progenitor cell differentiation. Dev. Biol. 139, 197-209. Hamburger, V. and Hamilton, H. L. (1951). A series of normal stages in the development of the chick embryo. J. Morph. 88, 49-67. Han, Y., Dennis, J. E., Cohen-Gould, L., Bader, D. M. and Fischman, D. A. (1992). Expression of sarcomeric myosin in the presumptive myocardium of chicken embryos occurs within six hours of myocyte commitment. Dev. Dynamics 193, 257-265. Hastings, K. E. M., Koppe, R. I., Marmor, E., Bader, D., Shimada, Y. and Toyota, N. (1991). Structure and developmental expression of troponin I isoforms. cDNA clone analysis of avian cardiac troponin I mRNA. J. Biol. Chem. 266, 19659-19665. Jacobson, A. G. (1960). Influences of ectoderm and endoderm on heart differentiation in the newt. Dev. Biol. 2, 138-154. Jacobson, A. G. and Sater, A. K. (1988). Features of embryonic induction. Development 104, 341-359. Keller, R. E. (1976). Vital dye mapping of the gastrula and neurula of Xenopus laevis II. Prospective areas and morphogenetic movements of the deep layer. Dev. Biol. 51, 118-137. Linask, K. K. (1992). N-cadherin localization in early heart development and polar expression of Na+, K+-ATPase, and integrin during pericardial coelom formation and epithelialization of the differentiating myocardium. Dev. Biol. 151, 213-224. Lints, T. J., Parsons, L. M., Hartley, L., Lyons, I. and Harvey, R. P. (1993). Nkx-2.5: a novel murine homeobox gene expressed in early heart progenitor cells and their myogenic descendents. Development 119, 419-431. Logun, M. and Mohun, T. (1993). Induction of cardiac muscle differentiation in isolated animal pole explants of Xenopus laevis embryos. Development 118, 865-875. Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982). ‘Molecular Cloning: A Laboratory Manual’. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Nascone, N. and Mercola, M. (1995). An inductive role for the endoderm in Xenopus cardiogenesis. Development 121, 515-523. Orts-Llorca, F. (1963). Influence of the endodermon on heart differentiation during the early stages of development of the chicken embryo. Wilhelm Roux’ Arch. Entwicklungsmech. Org. 154, 533-551. Orts-Llorca, F. and Gil, D. R. (1965). Influence of the endoderm on heart differentiation. Wilhelm Roux’ Archiv. Entwicklungsmech. Org. 156, 368370. Rawles, M. E. (1943). The heart forming regions of the early chick blastoderm. Physiol. Zool. 16, 22-42. Rosenquist, G. C. (1985). Migration of precardiac cells from their origin in epiblast until they form the definitive heart in the chick embryo. In: Cardiac Morphogenesis Ferrans, Rosenquist, Weinstein, editors. Sater, A. K. and Jacobson, A. G. (1989). The specification of heart mesoderm occurs during gastrulation in Xenopus laevis. Development 105, 821-830. Selleck, M. A. J. and Stern, C. D. (1991). Fate mapping and cell lineage analysis of Hensen’s node in the chick embryo. Development 112, 615-626. Slack, J. M. W. (1993). Embryonic induction. Mech. Dev. 41, 91-107. Smith, S. C. and Armstrong, J. B. (1990). Heart induction in wild-type and cardiac mutant axolotls (Ambystoma mexicanum). J. Exp. Zool. 254, 48-54. Stalsberg, H. and DeHaan, R. L. (1969). The precardiac areas and formation of the tubular heart in the chick embryo. Dev. Biol. 19, 128-159. Stanier, D. Y. R., Lee, R. K. and Fishman, M. C. (1993). Cardiovascular development in the zebrafish I. Myocardial fate and heart tube formation. Development 119, 31-40. Sugi, Y. and Lough, J. (1994). Anterior endoderm is a specific effector of terminal cardiac myocyte differentiation of cells from the embryonic heart forming region. Dev. Dynamics 200, 155-162. Tonissen, K. F., Drysdale, T. A., Lints, T. J., Harvey, R. P. and Krieg, P. A. (1994). XNkx-2.5 a Xenopus gene related to Nkx-2.5 and tinman: Evidence for a conserved role in cardiac development. Dev. Biol. 162, 325-328. Yamakazi, Y. and Hirakow, R. (1991). Factors required for differentiation of chick precardiac mesoderm cultured in vitro. Proc. Japan Acad. 67, 165-169. Yutzey, K. E., Rhee, J. T. and Bader, D. (1994). Expression of the atrialspecific myosin heavy chain AMHC1 and the establishment of anteroposterior polarity in the developing chicken heart. Development 120, 871-883. (Accepted 5 May 1995)