SEARCH

SEARCH BY CITATION

Keywords:

  • cardiomyocytes;
  • embryoid bodies;
  • gene expression;
  • human embryonic stem cells;
  • immunochemistry

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

During early embryogenesis, the cardiovascular system is the first system to be established and is initiated by a process involving the hypoblastic cells of the primitive endoderm. Human embryonic stem (hES) cells provide a model to investigate the early developmental stages of this system. When removed from their feeder layer, hESC create embryoid bodies (EB) which, when plated, develop areas of beating cells in 21.5% of the EB. These spontaneously contracting cells were demonstrated using histology, immunostaining and reverse transcription–polymerase chain reaction (RT-PCR), to possess morphological and molecular characteristics consistent with cardiomyocytic phenotypes. In addition, the expression pattern of specific cardiomyocytic genes in human EB (hEB) was demonstrated and analyzed for the first time. GATA-4 is the first gene to be expressed in 6-day-old EB. Alpha cardiac actin and atrial natriuretic factor are expressed in older hEB at 10 and 20 days, respectively. Light chain ventricular myosin (MLC-2V) was expressed only in EB with beating areas and its expression increased with time. Alpha heavy chain myosin (α-MHC) expression declined in the pulsating hEB with time, in contrast to events in EB derived from mice. We conclude that human embryonic stem cells can provide a useful tool for research on embryogenesis in general and cardiovascular development in particular.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

During early embryogenesis, the first system to be established is the cardiovascular system and this process is initiated by interaction with the hypoblastic cells of the primitive endoderm (Icardo 1996). Embryonic stem (ES) cells isolated from the inner cell mass of blastocysts can be maintained as continuously growing cell lines (Evans & Kaufman 1981; Thomson et al. 1998). ES cells retain their self-replicative ability in an undifferentiated state for long periods and have the capacity to form tissues derived from all three germ layers (Evans & Kaufman 1981; Thomson et al. 1998). Mouse ES cells have provided a model to investigate the early developmental stages of the heart (Grepin et al. 1997; Hescheler et al. 1997; Narita et al. 1997) and recent studies have indicated that this is also the case regarding human ES (hES) cells (Kehat et al. 2001; Kehat et al. 2002; Xu et al. 2002; He et al. 2003; Mummery et al. 2003). Removal of undifferentiated cells from their feeder layer and their culture in suspension leads to the formation of embryo-like aggregates, known as embryoid bodies (EB) (Itskovitz-Eldor et al. 2000), which may spontaneously differentiate into cells with cardiomyocytic characteristics.

When plated, EB may spontaneously develop pulsatile or beating areas (Kehat et al. 2001; Xu et al. 2002; He et al. 2003), with cells displaying structural and functional properties of early stage cardiomyocytes (Kehat et al. 2001). Enhanced differentiation of cardiomyocytes from hES cells may be obtained by either lowering the density of plated EB (Xu et al. 2002), or by treating the cells with 5-aza-2′-doxycytidine (Xu et al. 2002). In addition, cardiomyocyte differentiation of hES cells can be induced by co-culture with visceral endoderm-like cells (Mummery et al. 2003). Such a co-culture system does not require EB formation for the generation of cardiomyocytes.

Cardiac gene expression in differentiating EB has been studied extensively using mouse ES cells. ES cell-derived cardiomyocytes express cardiac gene mRNA products in a developmentally controlled manner. As in early myocardial development, mRNA encoding GATA-4 and nkx2.5 transcription factors appear in mouse EB before mRNA encoding atrial natriuretic factor (ANF), myosin light chain atrial (MLC-2a), myosin light chain ventricular (MLC-2v), α-myosin heavy chain (α-MHC) and β-myosin heavy chain (β-MHC) (see Boheler et al. 2002 for review). However, little is known about the genes and molecular signals involved in human embryonic heart development (Olson & Srivastava 1996).

The recent availability of hES cell lines (Thomson et al. 1998; Reubinoff et al. 2000) has opened exciting new possibilities to explore human heart development and the expression of new genes during embryonic development. In particular, research has stimulated exploration of the potential of hES cells to create new cardiac tissues for therapeutic purposes. These cells may help clarify the complex processes involved in human heart development and provide a suitable cellular system to study the commitment and differentiation of tissues under in vitro conditions; it may also eventually lead to the ability to generate differentiated human cells for transplantation therapy.

This study presents an in vitro model to follow early human heart development and the timing of expression of specific candidate and cardiac gene expression during cardiogenesis.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Cell culture

hES cell clone H9.2 at passages 45–80 (Amit et al. 2000) were grown on mouse embryonic fibroblasts (MEF) in 80% knockout Dulbecco's modified Eagle's medium, 20% knockout serum replacement, 4 ng/mL bFGF, 1 mm glutamine, 0.1 mmβ-mercaptoethanol, and 1% nonessential amino acid stock (all from Gibco-BRL, Gaithersburg, MD, USA). To induce EB formation, hES cells were transferred to Petri dishes with 1 mg/mL type IV collagenase (Gibco-BRL), to allow their aggregation. Resultant EB were grown in 80% knockout Dulbecco's modified Eagle's medium, 20% defined fetal bovine serum (HyClone, Logan, UT, USA), 1 mm glutamine, and 1% nonessential amino acid stock. EB were then cultured in suspension for 7 days and plated on gelatin-coated (0.1%; Sigma Chemicals, St. Louis, MO, USA) 24-well plates (Nunc, Roskilde, Denmark), one EB per well. Daily microscopic observations were conducted to detect EB with spontaneous pulsatile areas and to determine the beating rate.

Reverse transcription–polymerase chain reaction (RT-PCR) analysis

Total RNA was extracted using the TRI reagent kit (Sigma Chemicals) according to manufacturer's instructions. cDNA was synthesized from either 1 µg total RNA or a single EB by using superscript reverse transcriptase (Gibco-BRL, Gaithersburg, MD, USA). cDNA samples were subjected to PCR amplification with human cardiac gene primers (Kehat et al. 2001) and human mesodermal gene primers (Gokhale et al. 2000; Gerecht-Nir et al. 2005). The absence of DNA contamination in RNA samples was confirmed with PCR primers flanking an intron. Products were size-fractionated by 2% agarose gel electrophoresis and the number of cycles was calibrated for each set of primers to reflect the linear condition.

Expression of cardiac genes during EB development

The expression pattern of several cardiac-specific genes was analyzed by RT-PCR of RNA samples taken at five time points during the development of EB in suspension (days 1, 6, 10, 15 and 20 or 21 of culture). For genes whose expression began in pulsatile EB, RNA was prepared on days 3 and 45 after initiation of pulsations. At each time point RNA was prepared from at least nine individual EB. The PCR reaction was linear and the results were normalized to reduced glyceraldehyde-phosphate dehydrogenase (GAPDH) expression in the same samples. Results were also analyzed statistically by a variance test (SD) and the Scheffe post-hoc test with a value of P < 0.05 as definitive variance. Nkx2.5 expression was examined in RT-PCR samples transcribed from 1 µg total RNA taken from EB in suspension (days 2, 4, 7, 8, 10 and 14 of culture).

Histology

Beating EB grown for 4–6 weeks in suspension were fixed in 3% glutaraldehyde in 0.1% cacodylate buffer (pH 7.4) at room temperature for 24 h, rinsed in 0.1 m sodium cacodylate and 7.5% sucrose and postfixed in 1% osmium tetroxide for 1 h. The samples were dehydrated in graded ethanols and embedded in Epon resin heat-polymerized for 18 h in 60°C. Thin (1 µm) ultramicrotome sections were stained in 0.1% toluidine blue in 1% borax prior to examination by brightfield microscopy.

Immunostaining

Pulsatile regions of EB were mechanically dissected using a sterile needle. These dissected regions were then enzymatically dispersed using trypsin ethylenediamine tetraacetic acid (0.55 trypsin, 0.53 mm EDTA; Gibco-BRL) for 15 min at 37°C. Cells were plated on gelatin-coated glass coverslips, incubated for 48 h, fixed using 4% paraformaldehyde with sucrose, and permeated using 0.5% Triton X-100 (Sigma). Cells were then treated with 1% H2O2 in methanol and rehydrated by incubating successively in 100%, 95% and 70% ethanol, and finally incubated with primary antibodies antibody diluent (Dako, Carpinteria, CA, USA) for one hour in the dark. The primary antibodies used were: monoclonal α-actinin 1:20 (ICN Biomedical, Costa Mesa, CA, USA), monoclonal mouse anti-human desmin 1:20 (Dako) and monoclonal mouse anti-human heavy chain β-myosin 1:10 (Chemicon, Temecula, CA, USA). The cells were washed with Tris-buffered saline (TBS) and incubated for 15 min in the dark with secondary antibody linked to biotin followed by Streptavidin-peroxidase and diaminobenzidine (DAB). Slides were counterstained with hematoxylin and dehydrated with ascending ethanols prior to mounting with DPX (Sigma). As negative controls, the reaction was performed without the primary antibody or using dispersed cells from non-beating EB. Preparations were photographed using an Olympus (Tokyo, Japan) CH30 brightfield microscope with SC35 camera. For immunofluorescence, EB or cells from contracting regions were plated on gelatin-coated glass coverslips, incubated for 48 h, and fixed using 4% paraformaldehyde with sucrose. Cells were blocked with 10% BSA and incubated with primary antibodies overnight at 4°C. The primary antibodies used were: monoclonal mouse anti-natriuretic peptide (ANP) 1:250, monoclonal mouse anti-troponin I (cTnI) 1:500 and monoclonal mouse anti-troponin T (cTnT) 1:500 (all from Chemicon). The antibodies were diluted in antibody diluent (Dako). The cells were washed with PBS and incubated for 1 h in the dark at room temperature with secondary antibody mouse Cy3 (Jackson ImmunoResearch Laboratories, West Grove, PA, USA) 1:100. As negative controls, the reaction was performed without the primary antibody or using dispersed cells from non-beating EB. Nuclei were counterstained with the DNA dye TO-PRO-3 (Molecular Probes, Eugene, OR, USA). Preparations were examined using a confocal microscope.

Western blot analysis

Pulsating 5-day-old EB, and 1-month-old pulsating and non-pulsating EB were lyzed using lysis buffer (Gu et al. 1994). Lysates were centrifuged at 10 000 g for 10 min, and the supernatant (cytosol) was stored at −70°C for protein quantification and Western blot analysis. Protein concentrations were determined using the Bradford protein assay (Bio-Rad, Hercules, CA, USA). Samples for Western analysis were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) on 4–20% gradient SDS-PAGE gels (Gradipore, Frenchs Forest, Australia) and transferred to nitrocellulose membranes (Schleicher & Schuell, Dassel, Germany). After 1 h blocking with 1 × TBS, 0.1% Tween-20, 5% non-fat milk, membranes were incubated with a suitable antibody overnight at 4°C. Antibodies used were: mouse anti-heavy chain myosin α/β (1:10; Chemicon), mouse anti-α actin (1:20; ICN Biomedical), or rabbit anti-Oct4 (1:100; Santa Cruz Biotechnology, Santa Cruz, CA, USA). Anti-mouse or anti-rabbit peroxidase-labeled antibodies were used as second antibody. The membrane was incubated for 1 h at room temperature with the peroxidase-labeled antibody (1:1000). Detection was performed using the ECL Western blotting analysis system (Amersham Pharmacia Biotech, Piscataway, NJ, USA) and visualized using ImageMaster VDS-CL (Amersham Pharmacia Biotech, Bucks, England).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Human ES cell differentiation into cardiomyocytes

EB grown in suspension for 7 days were transferred to 24-well gelatin-coated plates, one EB per well. Each EB consisted on average 10 000 cells. Daily observations of the pulsating EB revealed that 3% of them showed spontaneous beating areas 1 day after plating (8 days from the start of the experiments). Most EB developed pulsations by day 17 (10 days after seeding). Initiation was observed in some EB up to day 28 (Fig. 1A). Contracting regions appeared in 21.5% of the EB (Fig. 1B), reaching up to 40% in some experiments. We noticed increased percentages of contracting EB generated from hES cells, grown for more than 10 passages following thawing (up to 40%) compared to 5% in the first few passages. As hES cells are known for their very demanding culture requirements (Thomson et al. 1998; Amit et al. 2000) and relatively low survival rates after cryopreservation (Reubinoff et al. 2001), this phenomenon of late initiation of beating may possibly be attributable to slow recovery of the physiological properties of the cells after thawing. Pulsating regions had typical diameters of 0.1–1.2 mm and were located close to the periphery. The onset of contraction ranged from 4 to 27 days, followed by rhythmic stabilization 2–3 days later. The contracting areas continued to pulsate vigorously in culture for up to 4 months (the longest period studied) at an average frequency of 99 ± 8 beats per minute, but the contracting region did not grow further. Overgrowth of other cells differentiated from the EB occasionally interfered with the contraction. In order to be able to characterize the contracting areas over time, they were mechanically isolated and replated every 7–14 days.

image

Figure 1. (A) Day of onset of spontaneous pulsatile areas of embryoid bodies (EB) after embryonic stem (ES) cells were plated as single cell suspensions on Petri dishes to allow EB formation. Values are expressed as percentage of contractile EB screened. In 3% of pulsating EB the earliest contractile activity occurred on day 8 after plating. Day 17 was the mean day for initiation of pulsations in EB with values ranging from day 8–28. (B) Cumulative percentage of EB containing spontaneously contracting areas as a function of days after plating.

Download figure to PowerPoint

Cell morphology of the contracting areas

The histology of the contracting regions of EB grown for over 4 weeks in suspension (Fig. 2A–C) was determined in thin Epon sections stained with alkaline toluidine blue. Even at low magnification the beating regions of the EB located close to the periphery of the EB revealed well-differentiated aggregates of cardiomyocyte-like cells embedded in mesenchymal areas and containing primary blood vessels. These aggregates of cardiomyocyte-like cells were composed of relatively small cells (10–20 µm in longest dimension) with elongated morphology, one or two regular oval nuclei and prominent nucleoli.

image

Figure 2. Morphology of pulsating EB. (A) Detail of pulsating cystic EB in culture. Bar, 141 µm. (B,C) 1 µm epon resin sections of pulsating EB stained with alkaline toluidine blue. (B) Low power view showing the pulsating area (in box) revealed as an aggregate of well-differentiated cardiomyocyte-like cells located within a layer of mesenchymal cells, adjacent to the primary blood vessels (thin arrow). The external epithelial layer is shown in the upper part of the micrograph (broad arrow). A more detailed view of the cell aggregate is shown in (C). Bar, 67 µm. (C) The aggregate of cardiomyocyte-like cells is composed of relatively small elongated cells (10–20 µm in longest dimension), containing one or two rounded or oval nuclei and prominent nucleoli. Bar, 12 µm. Immunostaining of contracting regions in day 10 and 14 EB (D & E, respectively) for anti-cTnI antibodies (red) and TO-PRO-3 (nuclear staining, blue). Bar, 100 µm. (F) reverse transcription–polymerase chain reaction (RT-PCR) analysis for Nkx2.5 and glyceraldehyde-phosphate dehydrogenase (GAPDH) expression in EB in suspension (days 2, 4, 7, 8, 10 and 14 of culture). (–), PCR reaction without RT.

Download figure to PowerPoint

The presence of cardiac-specific proteins and their spatial organization were studied in dispersed cells from contracting EB using immunocytochemical techniques (as described in Materials and Methods). EB grown in suspension for 10 and 14 days were plated and stained with cTnI antibody in order to detect cardiomyocytes in the EB (Fig. 2D–E, respectively). The cells from the contracting areas were stained with α-actinin, desmin, β-heavy chain myosin, and ANP, cTnI and cTnT antibodies (Fig. 3). The dispersed cells showed a histological appearance similar to that of mouse ES cardiomyocytes (Lane et al. 1977; Bugaisky & Zak 1989; Hescheler et al. 1997), including small, round, elongated, branched and triangular cells with one or two oval nuclei and prominent nucleoli (Fig. 3).

image

Figure 3. Immunostaining of ES-cell-derived cardiomyocytes. Pulsatile regions of EB were mechanically dissected using a sterile needle. These dissected regions were enzymatically dispersed and fixed (as described in Materials and Methods). (A) Positive staining with anti-α-actin antibody in cells with triangular or elongated morphology, oval nucleus and prominent nucleolus. Bar, 46 µm. (B) Staining with desmin antibody in triangular cell with round nucleus and prominent nucleolus. Bar, 15 µm. Triangular shape cell stained with β-heavy chain myosin with two rounded nuclei (bar, 5 µm) (C), or with clumped cytoplasmic staining pattern (bar, 16 µm) (D). Positive staining with anti-cTnI antibody (E) and (F), anti-cTnT antibody (G) and anti-ANP antibodies (H). Bar, 10 µm.

Download figure to PowerPoint

Expression of cardiac genes during EB development

The expression of the early cardiomyocyte marker Nkx2.5 was examined in EB grown in suspension at different time points (2, 4, 7, 8, 10 and 14 days of differentiation). Nkx2.5 expression was first noticed in 4-day-old EB and was significantly increased by day 14 (Fig. 2F). The expression of other cardiac genes in individual EB was examined in order to analyze whether human EB can be used as an in vitro model for gene expression during human heart development. Nine EB grown in suspension were collected at specified time points (1, 6, 10, 15 and 20 days of culture). RNA was prepared followed by RT-PCR for each individual EB. For semiquantitative analysis, the PCR results were measured by densitometry and normalized to gapdh generated from the same sample. The mRNA expression of several cardiac-specific proteins and transcription factors and some early mesodermal markers was similarly examined. The first gene to display enhanced expression was GATA-4, a transcription factor known for its significant role in regulating different genes involved in the differentiation and development of the heart (Durocher & Nemer 1998). EB in suspension exhibited a dramatic increase in the expression of GATA-4 from day 6, remaining high during the following days, and decreasing only on day 20 (Fig. 4). A Scheffe post-hoc test showed a significant low expression of GATA-4 1 day after plating, in comparison with its expression at days 6–20. The decreased GATA-4 expression on day 20 is in accordance with a decrease in the early heart cell mass and an increase in the differentiating heart cells. As stated before (Kehat et al. 2001; Mummery et al. 2003), there is a shift from an immature phenotype manifested by disorganized myofibrillar stacks in early stage EB to a more organized sarcomeric structure in later stage EB, indicating a correlation between the expression of GATA-4 and the morphology of the cardiomyocytes.

image

Figure 4. Total RNA from nine individual EB from 1, 6, 10, 15 and 20 day differentiation cultures was analyzed by RT-PCR. PCR was performed under linear conditions in order to reflect the original amount of the specific transcript. Gene expression was normalized to gapdh expression. Variable expression in developing EB was measured for α-cardiac actin, GATA-4, MLC-2a and ANF. The α-cardiac actin increasing expression in single EB at five time points of differentiation as it appears on the gel.

Download figure to PowerPoint

Alpha-cardiac actin increased with time in the developing EB (Fig. 4). A Scheffe post-hoc test showed a significant increase in its expression from days 10–20. The earliest expression of alpha-cardiac actin during the development of the human EB corresponded with its expression pattern in the mouse. As its expression preceded GATA-4, it was found useful as an early marker of cardiomyocytes (Gunning et al. 1983). As the cardiomyocytes at this stage of differentiation are known to proliferate (Auda-Boucher et al. 2000), the increased expression of α-cardiac actin may probably result from the increase in heart muscle cells in the EB.

Flk1 (Vegfr-2/KDR) was chosen as a cell marker for the endothelial cells of the vascular system and their precursors. EB in suspension exhibited an increase in its expression as the culture developed. On day 1, only 25% of the EB expressed this marker. This percentage was markedly increased to 80% by day 7 and to 100% by day 21 (Fig. 5).

image

Figure 5. Total RNA from nine individual EB from 1, 6, 10, 15 and 21 day differentiation cultures was analyzed by RT-PCR. PCR was performed under linear conditions in order to reflect the original amount of the specific transcript. Gene expression was normalized to gapdh expression. Variable expression in developing EB was measured for CD34, brachyury and flk-1. Bottom graph represents total expression of the different markers, as a percentage of the whole EB population at the different time points.

Download figure to PowerPoint

Brachyury is a marker for recognition of mesodermal differentiation (Herrmann et al. 1990), which is expressed at gastrulation and is restricted to the primitive streak and mesoderm emerging from the streak. Later, it is found in the notochord (Herrmann et al. 1990; Wilkinson et al. 1990). Brachyury was significantly increased by day 21 of the EB in suspension (Fig. 5).

CD34 is expressed by endothelial and hematopoietic stem and progenitor cells (Nakayama et al. 1998). Similar to flk-1 and brachyury as the culture developed its expression increased (Fig. 5). The differences in the expression of CD34, flk1 and brachyury were not significant. This may be due to the fact that the RT-PCR was performed on single EB, and therefore some of them may have been spontaneously differentiating through other lineages.

This system may prove beneficial in identifying atrial markers (atrial natriuretic factor, atrial heavy chain myosin, atrial light chain myosin) and ventricular (ventricular light chain myosin) of cardiomyocytes. ANF and MLC2a were both expressed even prior to the onset of contractions, whereas myosin light chain ventricular (MLC-2v) was expressed only in the contracting EB. MLC-2a expression was initially observed in EB as young as 1-day-old, without significant differences in the following days (Fig. 4). It has previously been demonstrated that human embryos express large amounts of MLC-2a in the whole heart and in skeletal muscle (Price et al. 1980). ANF expression increased markedly on day 20 compared to days 1–15 (Fig. 4). In mice, ANF is first detected in 8-day-old embryos. Later, in adult mice and humans it is expressed only in the atria (Zeller et al. 1987; Larsen 1990). Because MLC-2v was found to be expressed only in contracting EB, a comparison was made between RNA from 3-day-old EB and 45-day-old pulsating EB. The former were found to express higher levels of MLC-2v (Fig. 6A). A resemblance was found between the onset of MLC-2v expression in human and mouse EB. In both cases, expression was detectable only after pulsations occurred (Miller-Hance & Chien 1993). The expression of α-MHC was also examined in pulsating EB and a significant decrease was observed in 45-day-old pulsating EB compared to those of 3 days of pulsations (Fig. 6B). These results were further verified by examining protein expression. Immunoblotting of proteins produced from early contracting EB (5 days after the initiation of contraction) was compared to proteins produced from 30 to 1-day-old pulsating EB and non-pulsating EB. The results showed a marked decrease in the levels of α-MHC from 5 to 30-day-old pulsating EB, similar to those observed using RT-PCR (Fig. 7). Alpha-actin expression was noticed in both the 5 and 30-day-old pulsating and non-pulsating EB, while on the other hand, α-MHC was noticed only in pulsating EB. The expression of OCT4 was noticed in 5-day-old pulsating EB and disappeared in the 30-day-old pulsating and non-pulsating EB (Fig. 7). As OCT4 is a marker of undifferentiated cells this phenomenon reflects the disappearance of undifferentiated cells as the differentiation process continues.

image

Figure 6. Variable expression of ventricular light chain myosin 2, and α MHC as pulsating EB matured. Total RNA from 10 individual EB, that were contracting 3 and 45 days in culture prior to isolation, was analyzed by RT-PCR for MLC-2v (A) α-MHC (B). Gene expression was normalized to gapdh expression.

Download figure to PowerPoint

image

Figure 7. Western blot analysis of total protein content from 5- and 30-day-old pulsating EB vs 30-day-old non-pulsating EB. The presence of α-actin, heavy chain myosin α/β and OCT4 was examined in these samples.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The ability to obtain functional cardiomyocytes from differentiating hES cells has previously been demonstrated (Kehat et al. 2001; Xu et al. 2002; He et al. 2003). The potential therapeutic use of these cardiomyocytes is enormous, including transplantation to improve or cure myocardial dysfunction. These cells provide a novel tool to explore the differentiation of cardiomyocytes and the genes and factors involved in this process. Using H9.2 hES cells, an average of 21.5% of the EB developed contracting areas. It appears that ‘old’ cultures grown for more than 10 passages after thawing and low density plating of EB contribute to the high percentage of pulsating areas within the EB as was previously reported by Metzger et al. (1995) in the mouse. In the present study, following 7 days of culture in suspension, individual EB were plated separately on gelatin-coated culture dishes. One to seven beating areas appeared on days 8–21 of culture. These areas continued to beat vigorously in the culture for up to 4 months (the longest period studied) at an average frequency of 99 ± 8 beats per minute. The underlying reason for such improved results are yet to be determined.

The cells were routinely checked for karyotype change and differences in surface markers such as SSEA-3, -4, TRA-1-60, -1-80 and Oct4. The percentage of those markers was not altered with the continuous growth of the culture (Carpenter et al. 2003; Rosler et al. 2004). Therefore, further work is needed to identify the profile of genes expressed in early versus late passages of embryonic stem cells, which might contribute to these phenomena.

Using RT-PCR, immunostaining and morphological characterization of the pulsating tissue, the spontaneously contracting cells exhibit morphological features similar to those of cardiomyocytes, previously obtained from mouse ES cells (Hescheler et al. 1997). In addition, this study demonstrates for the first time the expression pattern of specific cardiomyocyte genes in developing EB. Our results show that cells derived from the contracting areas of hEB exhibit cardiac intracellular proteins and sarcomeric organization. Positively immunostained single cells (Fig. 3) and the cardiomyocyte-like cells in sections of the contracting areas (Fig. 2) exhibited cardiomyocyte-like histology. Round or oval nuclei, occasionally binucleate, were typically centrally positioned. The positively immunostained isolated cells also exhibit cardiomyocyte cell structure with spindle, round, triangular or multinuclear appearance (Fig. 3) similar to cardiomyocytes of mouse embryos (Lane et al. 1977), mouse adult myocardium (Bugaisky & Zak 1989) and mouse pulsating EB (Hescheler et al. 1997).

Contracting areas derived from hES cells express several cardiomyocyte-specific genes that recapitulate the development of cardiomyocytes from very early cardiac precursor cells (Nkx2.5, α-cardiac-actin and GATA-4) to terminally differentiated cells. These genes express phenotypes of atrial (atrial natriuretic factor, α-heavy chain myosin, atrial light chain myosin) and ventricular cardiomyocytes (ventricular myosin light chain).

In this study, RT-PCR of several cardiomyocyte-specific genes were performed on single developing EB. Analyzing the expression pattern of these genes revealed that MLC-2a appeared in EB as early as 1-day-old. GATA-4 was expressed at significantly high levels in 6-day-old EB. Alpha-cardiac-actin was the second gene to show a significantly enhanced expression in 10-day-old EB. The mesodermal markers Flk1, CD34 and Brachyury increased as the differentiation proceeded, similar to their expression in the mouse. Mesoderm-derived populations identified within the developing EB included hematopoietic, endothelial, cardiac, skeletal muscle and adipocyte lineages. Brachyury is a marker of early mesoderm formation (Kispert & Herrmann 1993; Lacaud et al. 2004). Brachyury is expressed transiently in the nascent and early migrating mesoderm (Herrmann 1991). Flk1 and CD34 were both identified as markers for the early mouse vasculature (Vittet et al. 1996; Drake & Fleming 2000). It was previously shown that the levels of CD34 increased human EB differentiation, reaching a maximum on days 13–15. Flk1 expression plays an important part in the vascular development of mouse embryos. It has been shown that the expression pattern in the mouse embryo is consistent with Flk1, marking very early progenitors of the endothelial cell lineages as well as the differentiating endothelial cells themselves (Yamaguchi et al. 1993). Flk1 was shown to be expressed in undifferentiated hES cells (Kaufman et al. 2001; Levenberg et al. 2002), and to increase very slightly during differentiation (Levenberg et al. 2002). Using DNA microarray to assess the progression of gene expression with differentiation, in hEB a significant increase in the CD34 expression, but not in the Flk1 expression (Gerecht-Nir et al. 2005), was observed. Later on, atrial natriuretic factor was expressed at high levels in 20-day-old EB. Finally, the expression of MLC-2v and α-MHC was initiated after the onset of spontaneous contractions. A similar pattern of expression was also detected during mouse embryogenesis in the uterus and murine systems (Miller-Hance & Chien 1993; Boheler et al. 2002; Fijnvandraat et al. 2003; Sachinidis et al. 2003). This data suggests that developing EB follow a developmental pattern similar to the in vivo expression of these cardiac genes, from early precardiac mesodermal cells (GATA-4) to the terminally differentiated atrial (ANF and MLC-2a) and ventricular (MLC-2v) cells, and thus provide a useful system for studying early mammalian cardiogenesis. Markers such as Nkx2.5 and Mesp1 & 2 are early markers of cardiomyocytes. We and others have previously shown the expression of Nkx2.5 in cardiomyocytes from human embryonic stem cells (Kehat et al. 2001; Xu et al. 2002). Unfortunately, their detection was possible only when several EB were pulled together. As shown in Figure 2F, Nkx2.5 was first noticed in 4-day-old EB, and its expression arose on day 14. These results resemble other works on cardiomyocytes originating from mouse ES cells (Dell’Era et al. 2003; Terami et al. 2004).

The similarities between human and mouse cardiac gene expression is not surprising. Nevertheless, there are significant differences between human and mouse development as shown by this study and reflected by the negative correlation between the α-heavy chain myosin phenotype and the timing of onset of contractile activity of ES cell-derived cardiomyocytes. In mouse embryogenesis as well as in mouse ES system, α-MHC expression increased as the contractile period extended, representing the transition from fetal isoform β-MHC to adult isoform α-MHC during murine cardiogenesis in vivo (Nakao et al. 1997; Ritter et al. 1999). In human ES cells, the opposite process occurred where α-MHC expression declined as the contractile period extended. This observation suggests the α-MHC dominance in human embryos at mid-pregnancy and the restriction of the gene and low expression in the human atrium (Nakao et al. 1997; Ritter et al. 1999). These observations illustrate that the human ES cell system differs from murine systems and may improve our understanding of processes in human cardiogenesis. The inconsistent contractile rhythms in the first days of contraction and the change into rapid contractions associated with the light expression of α-MHC suggests that the latter may have a role in regulating the contractile activities of differentiating cardiomyocytes of human EB.

Establishment of a hES cell-derived cardiomyocytic differentiating system may serve as a potent new tool for the understanding of cardiac development and function, as well as for designing possible novel therapeutic strategies. hES cells also provide an attractive source for cell transplantation. In the mouse, ES cells have been transfected with a vector containing α-MHC promoter linked to neomycin resistance (Klug et al. 1996); forming cells were transplanted into a mouse heart and shown to improve its function. Knowledge derived from this current study may contribute to the planning of vectors for selection of cardiomyocytes and to the evaluation of the specific timing needed for such selection. For example, using constructs with α-MHC as used by Klug et al. (1996) requires a shorter selection time and may demand an earlier selection, perhaps even before the contractions are detectable. On the other hand, for effective use of constructs with an MLC-2v promoter (Muller et al. 2000), prolonged selection may prove more effective.

Elimination of the undifferentiated ES cells is one of the major obstacles in future transplantation of hES cell-derived cells, because undifferentiated ES cells can grow as teratomas. Careful examination of any transplanted cell should be done to avoid any presence of undifferentiated cells. One of the methods suggested is the use of the suicide gene in the hES cells (Schuldiner & Benvenisty 2003).

In conclusion, cardiomyocytes obtained in culture from hES cells follow the developmental course of human embryos and may therefore be used as an effective model system for understanding the developmental processes and functioning of the human heart. They may also provide potential tools in fields such as drug development and testing, early heart development and clinical treatment based on cell transplantation and tissue engineering.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The authors thank Mrs Kochava Shariki and Mr Sivan Osenberg for technical assistance, Mrs Ludmilla Mazor for technical assistance with the histology and Mrs Hadas O’Neill for editing the manuscript. The authors thank Professor Ofer Binah for critical reading of the manuscript. This research was supported by a grant from the Israel Science Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References