Genomic profiling of developing cardiomyocytes from recombinant murine embryonic stem cells reveals regulation of transcription factor clusters
Abstract
Cardiomyocytes derived from pluripotent embryonic stem cells (ESC) have the advantage of providing a source for standardized cell cultures. However, little is known on the regulation of the genome during differentiation of ESC to cardiomyocytes. Here, we characterize the transcriptome of the mouse ESC line CM7/1 during differentiation into beating cardiomyocytes and compare the gene expression profiles with those from primary adult murine cardiomyocytes and left ventricular myocardium. We observe that the cardiac gene expression pattern of fully differentiated CM7/1-ESC is highly similar to adult primary cardiomyocytes and murine myocardium, respectively. This finding is underlined by demonstrating pharmacological effects of catecholamines and endothelin-1 on ESC-derived cardiomyocytes. Furthermore, we monitor the temporal changes in gene expression pattern during ESC differentiation with a special focus on transcription factors involved in cardiomyocyte differentiation. Thus, CM7/1-ESC-derived cardiomyocytes are a promising new tool for functional studies of cardiomyocytes in vitro and for the analysis of the transcription factor network regulating pluripotency and differentiation to cardiomyocytes.
cell transplantation was discussed recently as an alternative treatment for heart failure. The transplantation of adult stem cells of different sources has been performed in the clinical setting with limited success (14). Moreover, the transdifferentiation of adult bone marrow stem cells to cardiomyocytes (23) following cell transplantation into the heart was shown to be a rare event (29, 33, 54) and beneficial effects on cardiac performance appeared to be low and not sustained (24). Embryonic stem cells (ESC) are pluripotent cells derived from the inner cell mass of the mammalian blastocyst. They are capable of self-renewing and can differentiate into a variety of cell lineages in vitro, including cardiomyocytes (50), which is manifested by the appearance of spontaneously and rhythmically contracting cell clusters. Cardiac myocytes derived from ESC were also used for experimental cell therapy and transplanted into infarcted mouse ventricular myocardium (7, 16, 36). However, ESC-derived cardiomyocytes have not been used in clinical trials yet due to an increased tumor risk, immunological issues, and ethical concerns. However, ESC cultures provide a valuable source for the analysis of pluripotency regulation and cardiomyocyte differentiation.
We demonstrated previously that confluent cardiomyocyte monolayers could be obtained from the genetically engineered murine ES cell line CM7/1, which carries a construct comprising the cardiac α-myosin heavy chain promoter driving the expression of the neomycin-phosphotransferase gene mediating neomycin resistance (55). Antibiotic-driven cardiomyocyte enrichment leads to synchronously contracting cells with myocyte-specific immunohistochemistry of >99% purity without the need of contaminating feeder layer cells (52). CM7/1 cells can also be cultivated in spinner flask cultures or bioreactors (42, 53). Thus, CM7/1 cells provide an excellent source for the analysis of cellular networks leading to cardiomyocyte specification and differentiation. We therefore focused our analysis also on the transcription factors regulated during differentiation of the CM7/1 mESC.
However, little is known on the regulation of the CM7/1-transcriptome during the differentiation to cardiomyocytes. Here we provide data on changes of the human transcriptome during the differentiation from undifferentiated pluripotent CM7/1 toward cardiomyocytes. We compared our data to gene expression patterns of isolated adult murine cardiomyocytes (mAC) and murine left ventricular myocardium (mHeart). Temporal changes in the gene expression pattern of the ESC cultures during the differentiation process were analyzed by self-organizing maps (SOM). Quantitative real-time PCR (qRT-PCR) was used to verify expression profiles determined by SOM analysis of microarray data. Furthermore, we functionally characterized the differentiated ESC-derived cardiomyocytes and found positive chronotropic effects of catecholamines and endothelin-1.
MATERIALS AND METHODS
Cell culture and tissue preparation.
The genetically engineered mouse ESC line CM7/1 was derived from J1 ESC cells (strain 129S4/Jae; ATCC# SCRC-1010) carrying a construct consisting of the cardiac α-myosin heavy chain promoter driving the neomycin-resistance gene as previously reported (55). ESC-CM7/1 cells were cultivated with leukemia inhibitory factor (LIF)-containing medium to preserve pluripotency until confluence (43, 49). Withdrawal of LIF-induced ESC differentiation and embryoid body (EB) formation in spinner cultures after 3 days. Selection with geniticin G418 (Invitrogen/GIBCO, Carlsbad, CA) after 11 days leads to formation of beating cardiac bodies (CB), from which isolated cardiomyocytes can be derived by trypsinization. Chronotropy of isolated cardiomyocytes was analyzed by addition of increasing amounts of isoprenaline and terbutaline and counting the number of beats per minute. Nonlinear regression was performed using Prism 5.0 (GraphPad Software, San Diego, CA) for the analysis of concentration response relationships.
mAC were isolated by retrograde perfusion of mouse hearts derived from the 129S1 strain with an enzyme solution as described previously (38). We isolated left ventricular myocardium from 129S1 mice after cardiac puncture and rinsing the organs twice with PBS.
RNA isolation and Affymetrix microarrays.
Total RNA of either cultured cells or myocardial tissues was isolated with the Qiagen RNeasy kit. The quality of the obtained RNA was checked using the Bioanalyzer 2100 (Agilent).
Reverse transcription of 1 μg of total RNA and subsequent in vitro transcription of resulting cDNA as well as biotin labeling and microarray hybridization was carried out according to standard protocols of the manufacturer (Affymetrix User Guide). We used Affymetrix MOE430A arrays for whole genome transcriptome analysis. Microarray expression data were analyzed using Genedata Expressionist Analyst Pro Software version 4.0.5 and R ( http://www.r-project.org). Normalization was performed using the robust multiarray average (RMA) algorithm as implemented in Bioconductor ( http://www.bioconductor.org). There are four stages to RMA. First, probe set data from all arrays are simultaneously normalized with quantile normalization, which eliminates systematic differences between GeneChips, without significantly altering the relative intensity of probes within a GeneChip. Second, mean optical background level for each array is estimated, and the intensity for each probe is adjusted to remove this. Third, the normalized, background-corrected data are transformed to the log2 scale. Finally, a median-polish procedure is used to combine multiple probes into a single measure of expression for each gene on each array.
Multivariate principal component analysis (PCA) was used to describe the association between gene expression and sample phenotype. PCA was used as a mathematical technique that exploits these factors to pick out patterns in the data, while reducing the effective dimensionality of gene expression space without significant loss of information. Thus, PCA is a technique that provides a projection of complex data sets onto a reduced, easily visualized space.
Cluster analysis.
Significantly deregulated genes were identified in an unsupervised fashion based on the coefficient of variation. RMA-normalized expression values were taken as the basis with which to compute the coefficient. Subsequently, genes with the highest coefficients were selected. Gene clusters presenting similar time-dependent expression profiles were identified using unsupervised, SOM clustering with a maximum of 50 iterations. SOM clustering can be used to partition expression profiles into clusters of similar items. It is more powerful than hierarchical clustering in cases where many profiles have to be grouped into clusters. It uses a deterministic iterative algorithm that starts with a random and arbitrary number of cluster centers. We have chosen 3 × 3 clusters because this was the maximum number of clusters that still allowed interpreting all clusters biologically. A smaller number would have omitted clusters with interpretable profiles. We refrained from using a larger number of clusters, because this introduced additional clusters, which could not be explained biologically. Distance metrics were calculated using Pearson correlation coefficients based on RMA-normalized expression values. SOM analysis focuses attention on the patterns rather than on absolute levels of gene expression. Using Genedata Expressionist software (Genedata, Basel, Switzerland), we chose a SOM 3 × 3 grid to minimize both the variance within individual clusters and the redundancy of similar clusters while maintaining profiles that reflect altered gene expression patterns throughout differentiation. We tried to identify time-shifted profiles by taking the representative profile from the 3 × 3 SOM clustering and looking for similar profiles (shifted by t within [−3; +3]). Similarity of the time-shifted profiles was judged according to Pearson correlation. However, we could not identify any additional profile that was not yet a member of the cluster corresponding to the representative profile. We refrained from using Fourier-transform bases analysis techniques, because such an analysis would assume periodicity within the data. The given experiment, however, does not have periodic character but drives stem cells into a cardiomyocyte-like state (directional character).
We performed hierarchical clustering to partition the profiles based on the results. K-means clustering was tried as well but did not yield satisfactory results.
Using the coefficient of variation, 50 transcription factors (TF) with the strongest variation over time and stimuli were selected and used for hierarchical clustering.
RT-PCR.
The differential expression of selected genes identified by oligonucleotide arrays was confirmed by qRT-PCR with the TaqMan protocol using an ABI Prism 7700 sequence detection instrument. Briefly, 1 μg of total RNA was digested with DNase I and reverse transcribed into cDNA using Superscript-II RT-PCR kit (Invitrogen). 2.5% of each cDNA was taken for amplification using the TaqMan Universal Master Mix (Eurogentec, Cologne, Germany) and the following PCR protocol: denaturation 15 s at 94°C, annealing and extension 1 min at 60°C, 40 cycles. DNA sequences of PCR primers and labeled probes were designed by Primer Express 1.5 software (Applied Biosystems, Foster City, CA) and available from the authors upon request. Concentration of primers was 300 nmol/l and of labeled probes 150 nmol/l, respectively. Comparable amplification efficiencies for all primer/probe sets were checked by standard dilution curves. Threshold cycle (Ct value) correlates inversely with target mRNA level. Ct values were corrected for ribosomal protein L32, β2-microglobulin, and β-actin mRNA levels to exclude different starting amounts of total RNA. Expression was calculated from the formula: expression = 2(20−dCt), where dCt is the difference in Ct value between the gene of interest and the reference gene. We observed that the calculated expression level between different samples was independent of the housekeeping gene used, indicating that the mRNA expression level of ribosomal protein L32, β2-microglobulin, and β-actin is not significantly different between samples from ESC cultures, primary mAC, and mHeart, respectively. Therefore all mRNA expression data were normalized to the mRNA of ribosomal protein L32. The gene expression data are given as representative result of three independent experiments.
RESULTS
We have developed cardiomyocytes from murine ESC and compared their pharmacological to rat adult cardiomyocytes and human cardiac muscle preparations published previously.
Pharmacological characterization of ESC-derived cardiomyocytes.
We differentiated the murine cell line CM7/1 according to the procedure published previously (Fig. 1) (55). After 7 days in culture, beating EB were observed. Addition of G418 to the culture medium at day 11 led to enrichment of cardiomyocytes and the formation of beating CB. After trypsinization isolated cardiomyocytes were plated on cell culture dishes. Adherent cells were pharmacologically characterized with increasing concentrations of endothelin-1, the nonselective β-adrenoceptor agonist isoprenaline, and the β2-adrenoceptor selective ligand terbutaline (Fig. 2). We found positive chronotropic effects for endothelin-1 and the β-adrenoceptor agonists, which were fit by nonlinear regression (Fig. 2). The log-EC50 for endothelin was −8.1 and for isoprenaline −8.7 (Hill slope 0.68), which is in the range of previously reported data on murine ESC-derived cardiomyocytes (−7.3) (1), isolated adult rat cardiomyocytes (−8.4) (13), or human atrial trabeculae (−8.6) (25), respectively. The log-EC50 for terbutaline was −5.34, which is comparable to pD2 values found in human ventricles (5.8) (41) (pD2 being defined as the negative logarithm of the molar concentration of an antagonist necessary to inhibit the maximal concentration by 50%).
Fig. 1.Procedure for the cultivation and differentiation of murine embryonic stem cells (mESC) CM7/1 into cardiomyocytes. Phase contrast microscopy of CM7/1 (right). d, Day. White bars represent 100 μm.
Fig. 2.Concentration-response curve for the positive chronotropic effect of endothelin-1, isoprenaline, and the β2-adrenergic agonist terbutaline on CM7/1-derived cardiomyocytes.
Temporal gene expression pattern during differentiation of ESC to cardiomyocytes.
In a time-course experiment total RNA was isolated from CM7/1 cells at seven different time points during differentiation and hybridized to Affymetrix MOE430A arrays (culture days given in parenthesis): (d0) pluripotent CM7/1 cells, (d3) EB 3 days after LIF withdrawal, (d7) day 7 EB, (d10) day 10 beating EB before G418 selection, (d12) CB immediately after G418 selection, (d14) CB comprising enriched cardiomyocytes, (d17) seeded, isolated cardiomyocytes. Data were registered in the EBI ArrayExpress database under the accession number E-MEXP-1405.
The transcriptome of ESC differentiated into cardiomyocytes was compared with those of adherent mAC and mHeart. We used a PCA to compare the overall similarity of the gene expression patterns at various stages of the ESC culture with those of mAC and mHeart (Fig. 3A). The PCA revealed that the gene expression pattern of mESC-derived cardiomyocytes approaches very closely to that of mAC during the course of differentiation and selection. Surprisingly, the gene expression pattern of differentiated ESC-derived cardiomyocytes appeared to be highly similar to that of isolated adult cardiomyocytes and even to murine myocardium.
Fig. 3.Whole genome temporal gene expression profiling of mESC differentiating toward cardiomyocytes and compared with primary mouse adult cardiomyocytes (mAC) and ventricular myocardium (mHeart), respectively. A: principal component analysis (PCA) of robust multiarray average (RMA)-normalized microarray data using all probe sets. Gene expression signatures from various time points of the differentiating mESC culture is marked in red, whereas the transcriptome pattern of mAC and murine myocardium (mHeart) is shown in green and yellow, respectively. Data sets for meta-analysis of cardiac precursor cells were derived from Christoforou et al. (5) and given in blue. The resultant dots in the 3-dimensional space characterize the similarity of the transcription patterns of the respective samples. The more similar the transcriptional patterns, the closer the dots are, and hence the transcriptional state of cell culture samples. The PCA reveals that the transcriptome pattern of our data and data from Ref. 5 becomes more and more similar to the mAC and the murine heart samples. B: 3 × 3 self-organized map (SOM) clustering of top 1,000 genes filtered by coefficient of variation. The number of genes represented by the clusters and the corresponding figures of the heat maps are given. The red line indicates the time point of antibiotics selection.
However, some striking differences were found between mESC-derived cardiomyocytes, adult cardiomyocytes, and murine myocardium, respectively. SOX17, which was found to be essential for specification of the cardiac mesoderm, was detectable in the myocardium, whereas we did not find evidence for its expression in mAC or isolated ESC-derived cardiomyocytes, respectively. On the other hand, activating transcription factor (ATF) 3 and FosB were detected in isolated murine adult and ESC-derived cardiomyocytes but not in the myocardium. ATF3 was identified as a stress response gene (34) and negative regulator of the Toll-like receptor and was reported to play a cardioprotective role (32). FosB was previously reported to be involved in hypertrophy regulation and matrix metalloproteinase-2 transcription (3, 39). Interestingly, CDH11, belonging to the cadherin family of cell-cell adhesion molecules, was detectable only in the myocardium and ESC bodies. Obviously, the disruption of the cell contacts by protease digestion switched off the transcription of CDH11.
We next investigated by identification of SOM analysis clusters of genes with similar regulation during ESC differentiation. We found clusters of genes that were upregulated, downregulated, or transiently expressed during the course of differentiation (Fig. 3B). Cluster 7 (Fig. 3B) of the SOM represents genes with upregulated expression during ESC differentiation, which were also present in adult mAC and mHeart. The top 50 genes from this cluster are shown as a heat map (Fig. 4A) and include typically a number of cardiac-specific genes including TTN, MYH6, ATP2A2, ACTN2, and TNNT2. These genes were previously described as being upregulated during the differentiation of an human ESC line to a cardiomyocyte-like phenotype (2).
Fig. 4.Clustering of microarray data. A: top 50 genes from SOM cluster 7 (see Fig. 3B) containing genes that are upregulated during mESC differentiation into cardiomyocytes. B: top 50 genes from SOM cluster 1 (see Fig. 3B) containing genes that are downregulated during mESC differentiation. C: top 50 genes from SOM cluster 8 (see Fig. 3B) containing genes that are transiently expressed throughout ESC differentiation.
Cluster 1 contains genes with downregulated expression during ESC culture. This cluster comprises genes that are known to regulate ESC pluripotency like NANOG, OCT4 (POU5F1) (20), and SOX2 (48) or were found to be enriched in undifferentiated ESC such as DNMT3B (4), E-cadherin (CDH1) (8), and LIN28 (11) (Fig. 4B).
Cluster 8 of the SOM contains genes with a transient expression pattern during ESC differentiation, whereas their expression was almost undetectable in undifferentiated ESC, ESC-derived cardiomyocytes, mAC, and mHeart, respectively. The top 50 genes of this cluster are shown in Fig. 4C. Typical genes in this cluster are DLK1 and RBP4, which are found in adipocytes (for details see online supplemental material1 1The online version of this article contains supplemental material.
TF involved in ESC differentiation.
TF are essential for the regulation of the genome during differentiation. For the analysis of TF regulation, we used the Pfam-based classification of the human genome and identified those TF within the top 1,000 genes showing the highest coefficient of variation. Hierarchical clustering of microarray data by the top 50 TF revealed four characteristic groups, which appear expressed in the early (culture d0–d3; cluster C1), midterm (d7–d10; cluster C2), midterm with sustained expression (d7–d10; cluster C3), or late (d7–d17; cluster C4) phase of ESC cardiac differentiation and cardiomyocyte enrichment (Fig. 5) and which were confirmed by real-time RT-PCR analysis (Fig. 6C). TF found in the late phase of the ESC differentiation during antibiotic selection were also detectable in adult cardiomyocytes and murine myocardium, respectively. Dissociation of cardiac bodies and cardiomyocyte seeding resulted in the downregulation of TF typically expressed in the developing heart (HAND2, MRG1) or in cardiac precursor cells (ISL1) (27), respectively. As expected, TF that regulate pluripotency, like OCT4 (POU5F1) and NANOG, were downregulated early when differentiation was induced by LIF withdrawal (Fig. 5B). In parallel MYBL2, an MYB-family member of TF that is involved cell cycle progression (40) was downregulated within 7 days of differentiation. Typical TF that are transiently expressed during formation of EBs are ISL1 and HEY2. Whereas ISL1 was found to be present in cardiac precursor cells, HEY2/CHF1 was found to be involved in cardiac hypertrophy regulation (51), and its functional interruption resulted in defects during heart development (37). Interestingly, FOXA1, FOXA2, and FOXA3 were found to be upregulated in parallel. They are expressed in tissues of endodermal origin and belong to the hepatocyte nuclear TF (15, 18). Thus, at this stage of differentiation, multiple cell lineages were present in the culture, and notably endoderm-related factors were found to be upregulated.
Fig. 5.Transcription factors (TF) and cardiomyocyte differentiation. A: clustering of top 50 transcription factors in top 1,000 genes filtered based on the coefficient of variation from microarray data. Mapping of gene expression data reveals 4 clusters of TF during CM7/1 differentiation to cardiomyocytes depicted as C1–C4. B: clustering of TF associated with regulation of undifferentiated ESC and reprogramming of fibroblasts to induced pluripotent cells (45). C: model for the TF network derived from TF clustering and TF onset of transcription during mESC culture.
Fig. 6.Cardiac-specific gene expression. A: differentially expression of cardiac marker genes during the differentiation of CM7/1 to cardiomyocytes and their expression in primary mAC and mHeart, respectively. B: analysis of selected cardiac genes by quantitative real-time PCR during differentiation of CM7/1, mAC and mHeart (right). Dotted line indicates start of antibiotics selection. C: analysis of TF representing clusters C1–C4 identified by hierarchical cluster algorithm. Dotted line indicates start of antibiotics selection.
HAND2 and MEF2C, which are characteristic for the differentiation of cardiomyocytes and heart development, are present in a late cluster of TF (44). Notably, we found that under our ESC culture conditions a pattern of TF was expressed that has clear similarities to the TF network found during embryonic heart development (35) (Fig. 5). Due to the series of TF expression in our cell culture model, we propose a TF network as shown in Fig. 5C.
Myocardial genes.
Genes responsible for cardiomyocyte function are expected to be upregulated before day 10 of differentiation, since beating EB were observed at this stage. Addition of antibiotics for the enrichment of cardiomyocytes on day 11 may further increase the expression levels of respective genes. Accordingly, we found that the expression of early cardiomyocyte-specific TF, such as GATA4 and NKX2.5, can be detected at day 7, whereas the expression of their respective target genes, i.e., atrial natriuretic peptide and cardiac troponin C, is strongly induced after day 12 (Fig. 6, A and B).
In accordance with the positive chronotropic effect of isoprenaline on fully differentiated mESC, β1-adrenergic receptor expression (ADRB1) increases during differentiation with a peak level at day 14 when expression is even higher than in primary adult murine cardiomyocytes (Fig. 6B). Surprisingly, the β2-adrenergic receptor (ADRB2) is expressed constitutively at a lower level during differentiation of ESC.
In comparison, genes involved in the electromechanical coupling of cardiomyocytes including SERCA2A (ATP2A2) and connexin 43 (GJA1) and 45 (GJA7) are detectable even in undifferentiated ESC. The highest expression level of SERCA2A was found after day 12, when also the ryanodine receptor 2 (RYR2), triadin (TRDN), the Na+-Ca2+-exchanger (SLC8A1), and the L-type Ca2+ channel α1c-subunit (CACNA1C) show their highest expression levels. Thus, an orchestrated expression of components of the excitation-contraction coupling appeared to be induced in seeded, selected cardiomyocytes (Fig. 6, A and B).
DISCUSSION
The differentiation of pluripotent ESC to cardiomyocytes is well established. However, differentiation of human ESC to cardiomyocytes depends on feeder layer cocultures. The murine CM7/1 cell line investigated here can be differentiated by simple cell culture techniques and therefore provides a means to study ESC differentiation without contaminating non-ESC cocultures. CM7/1 can be kept in an undifferentiated state by LIF, as was reported before for other murine ESC (43, 49). However, this cell line can be used for the enrichment of ESC-derived cardiomyocytes since antibiotic selection allows the cultivation of almost pure cultures of cardiomyocytes even in bioreactors, which are pharmacologically competent to respond to catecholamines and endothelin-1 stimuli. Thus, CM7/1 might be regarded as a model system of the ideal ESC-derived cardiomyocyte for cell replacement therapies or drug target screening, respectively.
However, little is known on the mechanisms of mouse ESC-differentiation to cardiomyocytes and especially on the regulation at the transcriptional level. Therefore, in this study we analyzed the changes of the whole transcriptome during differentiation of CM7/1 to cardiomyocytes. We also isolated mAC and murine ventricular myocardium and compared their gene expression patterns to the pattern of isolated terminal differentiated cardiomyocytes derived from CM7/1. The overall gene expression patterns analyzed by PCA revealed a striking change of the transcription patterns from undifferentiated ESC to differentiated, enriched cardiomyocytes. The PCA also indicated that the gene expression profile of cardiomyocytes derived from ESC is highly similar to that of isolated mAC. Surprisingly, on this level of analysis the transcriptome of mAC was closely related to the gene expression pattern found in murine myocardium, although the myocardium consists of multiple other lineages, such as fibroblasts, endothelial cells, and smooth muscle cells, and only ∼30–35% of the cell nuclei are actually found within cardiomyocytes (30). The overall interpretation of this finding is that the contribution of endothelial cells and cardiac fibroblasts, the two other major cell types within the heart, to the myocardial transcriptome is rather low compared with the contribution of the cardiomyocyte-derived transcripts. This notion is further supported by the fact that expression levels of typical endothelial marker genes like von Willebrand factor or PECAM or cardiac fibroblast markers like vimentin are surprisingly low in our microarray data. Thus, the transcriptome of the myocardium appears to be dominated by the cardiomyocyte mRNA of the ventricles.
We looked at changes of the transcriptome during significant time points of the CM7/1. We analyzed probes at the undifferentiated state (d0), after LIF withdrawal on cell culture plates (d3), after development of EB in spinner culture (d7), before (d10) and after (d12) induction of antibiotic selection, in mature CB (d14), and finally after trypsinization of CB in isolated plated CM7/1-derived cardiomyocytes (d17).
ESC are pluripotent cells giving rise to all three germ layers. The gene expression patterns analyzed in our study represent therefore the development of the transcriptome of different cell populations, which are differentiated in EB. However, cardiomyocyte specification and differentiation occurred at early stages of the culture since contracting EB were already observed as early as day 7 of the culture. In addition, it is known that the induction of the cardiomyocyte fate in mesodermal precursors needs the presence of endodermal cells (28). Recently, it was shown that specification and differentiation of cardiomyocytes in the anterolateral mesoderm during embryonic development are essential for the differentiation of cardiac myocytes and dependents on BMP and FGF signaling (for a review see Ref. 21). Based on our TF clusters we propose a TF network that leads to cardiomyocyte differentiation within the EB (Fig. 5C). The network reveals some similarities with pathways analyzed in embryonic development (35). We identified TF of the core regulatory network for cardiac development in clusters 2–4 (Fig. 5A); however, in contrast to previous reports in embryonic development the TF NKX2.5 did not appear early during cell culture but late in cluster 4. Thus, from our data we conclude that this TF does not belong to the upstream activators for cardiomyocyte differentiation in mESC.
Blocking of the canonical Wnt pathway was shown to be essential for the induction of cardiogenesis (9). In addition, SOX17 was found to be essential for the specification of mesodermal cells to cardiomyocytes (19). Thus, differentiation of cardiomyocytes from mesodermal precursors depends on the interaction with endodermal cell types. These cells might indeed be present in our EB, since we identified at culture days 7–12 the expression of the transcription factors FOXA1 and FOXA2, which are required for the development of endoderm-derived cell types (10) and especially hepatic specification (18). SOX17 was transiently expressed in the second cluster of TF (see below) but was switched off during selection and isolation of ESC-derived cardiomyocytes. It was therefore not detectable in isolated mAC but in murine myocardium, indicating that SOX17-expression might be restricted to nonmyocardial cells of the myocardium.
We then focused our analysis on the gene expression patterns of TF in the developing cell culture of CM7/1 and identified four different clusters of TF, which are associated with undifferentiated ESC, EB formation, and enriched ESC-derived cardiomyocytes, respectively. From our data we conclude that in the early undifferentiated state of ESC a set of TF is transiently upregulated, which is mandatory for the maintenance of pluripotency and distinct from those expressed during EB formation and in differentiated cardiomyocytes. Typically OCT4 (POUF5) and SOX2 expression (48) is reduced in our cell culture system after LIF withdrawal. Both TF are known to be expressed in ESC and were recently shown to be essential to reinduce pluripotency in fibroblasts (45). However, OCT4 and SOX2 were also reported to be found in cancers of breast, pancreas, colon (26), adenocarcinomas (46), and prostate (12). In our culture system these TF were switched down to undetectable levels after day 14 in differentiation culture, indicating that undifferentiated cells might still be present even 3 days after antibiotic-driven enrichment of cardiomyocytes.
ESC cultures were claimed to be of relevance for developmental biology (31). In 1985 Doetschman et al. (6) reported that all three germ layers are present in EB derived from ES cells, and some morphological similarities between EBs and day 6–8 mouse embryos have been described. It was also reported that in EB a considerable number of cells (∼30%) spontaneously develop a cardiomyocyte phenotype. We found in our ESC cultures clusters of TF, which represent the core regulatory network of cardiac development (35). The development of the mammalian heart is dependent on the differentiation of mesodermal cells to cardiomyocytes and complex looping processes after formation of the primary and secondary heart fields. TF were reported to be involved in both processes. For instance the knockout mice for the TF GATA4 (17) and Nkx2.5 (22) do not prevent cardiomyocyte differentiation but do affect heart morphogenesis. In our culture system FoxH1 appears to be upstream of Isl1, which is in contrast to in vivo studies (47). In addition, in vivo the upstream activators appear to be different between the primary and secondary heart field (35) giving rise to the left or right ventricles, respectively. These differences might be due to the fact that ESC cultures inevitably are models for cardiomyocyte differentiation, whereas morphogenic processes cannot be simulated in EB. Therefore, the transcriptome analysis in this study provides further candidates of pluripotency regulation and cardiomyocyte differentiation in a recombinant murine ESC line. Thus, the transgenic line CM7/1 is an efficient tool for the analysis of the transcriptional network leading to cardiomyocyte specification and differentiation.
GRANTS
This work was supported by grants from the Erich und Hanna Klessmann-Stiftung (Guetersloh, Germany).
REFERENCES
- 1 Ali NN, Xu X, Brito-Martins M, Poole-Wilson PA, Harding SE, Fuller SJ. Beta-adrenoceptor subtype dependence of chronotropy in mouse embryonic stem cell-derived cardiomyocytes. Basic Res Cardiol 99: 382–391, 2004.
Crossref | PubMed | ISI | Google Scholar - 2 Beqqali A, Kloots J, Ward-van Oostwaard D, Mummery C, Passier R. Genome-wide transcriptional profiling of human embryonic stem cells differentiating to cardiomyocytes. Stem Cells 24: 1956–1967, 2006.
Crossref | PubMed | ISI | Google Scholar - 3 Bergman MR, Cheng S, Honbo N, Piacentini L, Karliner JS, Lovett DH. A functional activating protein 1 (AP-1) site regulates matrix metalloproteinase 2 (MMP-2) transcription by cardiac cells through interactions with JunB-Fra1 and JunB-FosB heterodimers. Biochem J 369: 485–496, 2003.
Crossref | PubMed | ISI | Google Scholar - 4 Brandenberger R, Wei H, Zhang S, Lei S, Murage J, Fisk GJ, Li Y, Xu C, Fang R, Guegler K, Rao MS, Mandalam R, Lebkowski J, Stanton LW. Transcriptome characterization elucidates signaling networks that control human ES cell growth and differentiation. Nat Biotechnol 22: 707–716, 2004.
Crossref | PubMed | ISI | Google Scholar - 5 Christoforou N, Miller RA, Hill CM, Jie CC, McCallion AS, Gearhart JD. Mouse ES cell-derived cardiac precursor cells are multipotent and facilitate identification of novel cardiac genes. J Clin Invest 118: 894–903, 2008.
PubMed | ISI | Google Scholar - 6 Doetschman TC, Eistetter H, Katz M, Schmidt W, Kemler R. The in vitro development of blastocyst-derived embryonic stem cell lines: formation of visceral yolk sac, blood islands and myocardium. J Embryol Exp Morphol 87: 27–45, 1985.
PubMed | Google Scholar - 7 Ebelt H, Jungblut M, Zhang Y, Kubin T, Kostin S, Technau A, Oustanina S, Niebrugge S, Lehmann J, Werdan K, Braun T. Cellular cardiomyoplasty: improvement of left ventricular function correlates with the release of cardioactive cytokines. Stem Cells 25: 236–244, 2007.
Crossref | PubMed | ISI | Google Scholar - 8 Fok EY, Zandstra PW. Shear-controlled single-step mouse embryonic stem cell expansion and embryoid body-based differentiation. Stem Cells 23: 1333–1342, 2005.
Crossref | PubMed | ISI | Google Scholar - 9 Foley AC, Mercola M. Heart induction by Wnt antagonists depends on the homeodomain transcription factor Hex. Genes Dev 19: 387–396, 2005.
Crossref | PubMed | ISI | Google Scholar - 10 Friedman JR, Kaestner KH. The Foxa family of transcription factors in development and metabolism. Cell Mol Life Sci 63: 2317–2328, 2006.
Crossref | PubMed | ISI | Google Scholar - 11 Gerecht-Nir S, Dazard JE, Golan-Mashiach M, Osenberg S, Botvinnik A, Amariglio N, Domany E, Rechavi G, Givol D, Itskovitz-Eldor J. Vascular gene expression and phenotypic correlation during differentiation of human embryonic stem cells. Dev Dyn 232: 487–497, 2005.
Crossref | PubMed | ISI | Google Scholar - 12 Gu G, Yuan J, Wills M, Kasper S. Prostate cancer cells with stem cell characteristics reconstitute the original human tumor in vivo. Cancer Res 67: 4807–4815, 2007.
Crossref | PubMed | ISI | Google Scholar - 13 Harding SE, Vescovo G, Kirby M, Jones SM, Gurden J, Poole-Wilson PA. Contractile responses of isolated adult rat and rabbit cardiac myocytes to isoproterenol and calcium. J Mol Cell Cardiol 20: 635–647, 1988.
Crossref | PubMed | ISI | Google Scholar - 14 Janssens S, Dubois C, Bogaert J, Theunissen K, Deroose C, Desmet W, Kalantzi M, Herbots L, Sinnaeve P, Dens J, Maertens J, Rademakers F, Dymarkowski S, Gheysens O, Van Cleemput J, Bormans G, Nuyts J, Belmans A, Mortelmans L, Boogaerts M, Van de Werf F. Autologous bone marrow-derived stem-cell transfer in patients with ST-segment elevation myocardial infarction: double-blind, randomised controlled trial. Lancet 367: 113–121, 2006.
Crossref | PubMed | ISI | Google Scholar - 15 Kaestner KH, Hiemisch H, Luckow B, Schutz G. The HNF-3 gene family of transcription factors in mice: gene structure, cDNA sequence, and mRNA distribution. Genomics 20: 377–385, 1994.
Crossref | PubMed | ISI | Google Scholar - 16 Kolossov E, Bostani T, Roell W, Breitbach M, Pillekamp F, Nygren JM, Sasse P, Rubenchik O, Fries JW, Wenzel D, Geisen C, Xia Y, Lu Z, Duan Y, Kettenhofen R, Jovinge S, Bloch W, Bohlen H, Welz A, Hescheler J, Jacobsen SE, Fleischmann BK. Engraftment of engineered ES cell-derived cardiomyocytes but not BM cells restores contractile function to the infarcted myocardium. J Exp Med 203: 2315–2327, 2006.
Crossref | PubMed | ISI | Google Scholar - 17 Kuo CT, Morrisey EE, Anandappa R, Sigrist K, Lu MM, Parmacek MS, Soudais C, Leiden JM. GATA4 transcription factor is required for ventral morphogenesis and heart tube formation. Genes Dev 11: 1048–1060, 1997.
Crossref | PubMed | ISI | Google Scholar - 18 Lee CS, Friedman JR, Fulmer JT, Kaestner KH. The initiation of liver development is dependent on Foxa transcription factors. Nature 435: 944–947, 2005.
Crossref | PubMed | ISI | Google Scholar - 19 Liu Y, Asakura M, Inoue H, Nakamura T, Sano M, Niu Z, Chen M, Schwartz RJ, Schneider MD. Sox17 is essential for the specification of cardiac mesoderm in embryonic stem cells. Proc Natl Acad Sci USA 104: 3859–3864, 2007.
Crossref | PubMed | ISI | Google Scholar - 20 Loh YH, Wu Q, Chew JL, Vega VB, Zhang W, Chen X, Bourque G, George J, Leong B, Liu J, Wong KY, Sung KW, Lee CW, Zhao XD, Chiu KP, Lipovich L, Kuznetsov VA, Robson P, Stanton LW, Wei CL, Ruan Y, Lim B, Ng HH. The Oct4 and Nanog transcription network regulates pluripotency in mouse embryonic stem cells. Nat Genet 38: 431–440, 2006.
Crossref | PubMed | ISI | Google Scholar - 21 Lough J, Sugi Y. Endoderm and heart development. Dev Dyn 217: 327–342, 2000.
Crossref | PubMed | ISI | Google Scholar - 22 Lyons I, Parsons LM, Hartley L, Li R, Andrews JE, Robb L, Harvey RP. Myogenic and morphogenetic defects in the heart tubes of murine embryos lacking the homeo box gene Nkx2-5. Genes Dev 9: 1654–1666, 1995.
Crossref | PubMed | ISI | Google Scholar - 23 Makino S, Fukuda K, Miyoshi S, Konishi F, Kodama H, Pan J, Sano M, Takahashi T, Hori S, Abe H, Hata J, Umezawa A, Ogawa S. Cardiomyocytes can be generated from marrow stromal cells in vitro. J Clin Invest 103: 697–705, 1999.
Crossref | PubMed | ISI | Google Scholar - 24 Meyer GP, Wollert KC, Lotz J, Steffens J, Lippolt P, Fichtner S, Hecker H, Schaefer A, Arseniev L, Hertenstein B, Ganser A, Drexler H. Intracoronary bone marrow cell transfer after myocardial infarction: eighteen months' follow-up data from the randomized, controlled BOOST (BOne marrOw transfer to enhance ST-elevation infarct regeneration) trial. Circulation 113: 1287–1294, 2006.
Crossref | PubMed | ISI | Google Scholar - 25 Milting H, Janssen PM, Wangemann T, Kogler H, Domeier E, Seidler T, Hakim K, Grapow M, Zeitz O, Prestle J, Zerkowski HR. FK506 does not affect cardiac contractility and adrenergic response in vitro. Eur J Pharmacol 430: 299–304, 2001.
Crossref | PubMed | ISI | Google Scholar - 26 Monk M, Holding C. Human embryonic genes re-expressed in cancer cells. Oncogene 20: 8085–8091, 2001.
Crossref | PubMed | ISI | Google Scholar - 27 Moretti A, Caron L, Nakano A, Lam JT, Bernshausen A, Chen Y, Qyang Y, Bu L, Sasaki M, Martin-Puig S, Sun Y, Evans SM, Laugwitz KL, Chien KR. Multipotent embryonic isl1+ progenitor cells lead to cardiac, smooth muscle, and endothelial cell diversification. Cell 127: 1151–1165, 2006.
Crossref | PubMed | ISI | Google Scholar - 28 Mummery C, Ward-van Oostwaard D, Doevendans P, Spijker R, van den Brink S, Hassink R, van der Heyden M, Opthof T, Pera M, de la Riviere AB, Passier R, Tertoolen L. Differentiation of human embryonic stem cells to cardiomyocytes: role of coculture with visceral endoderm-like cells. Circulation 107: 2733–2740, 2003.
Crossref | PubMed | ISI | Google Scholar - 29 Murry CE, Soonpaa MH, Reinecke H, Nakajima H, Nakajima HO, Rubart M, Pasumarthi KB, Virag JI, Bartelmez SH, Poppa V, Bradford G, Dowell JD, Williams DA, Field LJ. Haematopoietic stem cells do not transdifferentiate into cardiac myocytes in myocardial infarcts. Nature 428: 664–668, 2004.
Crossref | PubMed | ISI | Google Scholar - 30 Nag AC. Study of non-muscle cells of the adult mammalian heart: a fine structural analysis and distribution. Cytobios 28: 41–61, 1980.
PubMed | Google Scholar - 31 Nishikawa SI, Jakt LM, Era T. Embryonic stem-cell culture as a tool for developmental cell biology. Nat Rev Mol Cell Biol 8: 502–507, 2007.
Crossref | PubMed | ISI | Google Scholar - 32 Nobori K, Ito H, Tamamori-Adachi M, Adachi S, Ono Y, Kawauchi J, Kitajima S, Marumo F, Isobe M. ATF3 inhibits doxorubicin-induced apoptosis in cardiac myocytes: a novel cardioprotective role of ATF3. J Mol Cell Cardiol 34: 1387–1397, 2002.
Crossref | PubMed | ISI | Google Scholar - 33 Nygren JM, Jovinge S, Breitbach M, Sawen P, Roll W, Hescheler J, Taneera J, Fleischmann BK, Jacobsen SE. Bone marrow-derived hematopoietic cells generate cardiomyocytes at a low frequency through cell fusion, but not transdifferentiation. Nat Med 10: 494–501, 2004.
Crossref | PubMed | ISI | Google Scholar - 34 Okamoto Y, Chaves A, Chen J, Kelley R, Jones K, Weed HG, Gardner KL, Gangi L, Yamaguchi M, Klomkleaw W, Nakayama T, Hamlin RL, Carnes C, Altschuld R, Bauer J, Hai T. Transgenic mice with cardiac-specific expression of activating transcription factor 3, a stress-inducible gene, have conduction abnormalities and contractile dysfunction. Am J Pathol 159: 639–650, 2001.
Crossref | PubMed | ISI | Google Scholar - 35 Olson EN. Gene regulatory networks in the evolution and development of the heart. Science 313: 1922–1927, 2006.
Crossref | PubMed | ISI | Google Scholar - 36 Roell W, Lu ZJ, Bloch W, Siedner S, Tiemann K, Xia Y, Stoecker E, Fleischmann M, Bohlen H, Stehle R, Kolossov E, Brem G, Addicks K, Pfitzer G, Welz A, Hescheler J, Fleischmann BK. Cellular cardiomyoplasty improves survival after myocardial injury. Circulation 105: 2435–2441, 2002.
Crossref | PubMed | ISI | Google Scholar - 37 Sakata Y, Kamei CN, Nakagami H, Bronson R, Liao JK, Chin MT. Ventricular septal defect and cardiomyopathy in mice lacking the transcription factor CHF1/Hey2. Proc Natl Acad Sci USA 99: 16197–16202, 2002.
Crossref | PubMed | ISI | Google Scholar - 38 Sambrano GR, Fraser I, Han H, Ni Y, O'Connell T, Yan Z, Stull JT. Navigating the signalling network in mouse cardiac myocytes. Nature 420: 712–714, 2002.
Crossref | PubMed | ISI | Google Scholar - 39 Santalucia T, Christmann M, Yacoub MH, Brand NJ. Hypertrophic agonists induce the binding of c-Fos to an AP-1 site in cardiac myocytes: implications for the expression of GLUT1. Cardiovasc Res 59: 639–648, 2003.
Crossref | PubMed | ISI | Google Scholar - 40 Saville MK, Watson RJ. B-Myb: a key regulator of the cell cycle. Adv Cancer Res 72: 109–140, 1998.
PubMed | ISI | Google Scholar - 41 Schafers RF, Adler S, Daul A, Zeitler G, Vogelsang M, Zerkowski HR, Brodde OE. Positive inotropic effects of the beta 2-adrenoceptor agonist terbutaline in the human heart: effects of long-term beta 1-adrenoceptor antagonist treatment. J Am Coll Cardiol 23: 1224–1233, 1994.
Crossref | PubMed | ISI | Google Scholar - 42 Schroeder M, Niebruegge S, Werner A, Willbold E, Burg M, Ruediger M, Field LJ, Lehmann J, Zweigerdt R. Differentiation and lineage selection of mouse embryonic stem cells in a stirred bench scale bioreactor with automated process control. Biotechnol Bioeng 92: 920–933, 2005.
Crossref | PubMed | ISI | Google Scholar - 43 Smith AG, Heath JK, Donaldson DD, Wong GG, Moreau J, Stahl M, Rogers D. Inhibition of pluripotential embryonic stem cell differentiation by purified polypeptides. Nature 336: 688–690, 1988.
Crossref | PubMed | ISI | Google Scholar - 44 Srivastava D, Olson EN. A genetic blueprint for cardiac development. Nature 407: 221–226, 2000.
Crossref | PubMed | ISI | Google Scholar - 45 Takahashi K, Yamanaka S. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 126: 663–676, 2006.
Crossref | PubMed | ISI | Google Scholar - 46 Tsukamoto T, Mizoshita T, Mihara M, Tanaka H, Takenaka Y, Yamamura Y, Nakamura S, Ushijima T, Tatematsu M. Sox2 expression in human stomach adenocarcinomas with gastric and gastric-and-intestinal-mixed phenotypes. Histopathology 46: 649–658, 2005.
Crossref | PubMed | ISI | Google Scholar - 47 Von Both I, Silvestri C, Erdemir T, Lickert H, Walls JR, Henkelman RM, Rossant J, Harvey RP, Attisano L, Wrana JL. Foxh1 is essential for development of the anterior heart field. Dev Cell 7: 331–345, 2004.
Crossref | PubMed | ISI | Google Scholar - 48 Wang ZX, Teh CH, Kueh JL, Lufkin T, Robson P, Stanton LW. Oct4 and Sox2 directly regulate expression of another pluripotency transcription factor, Zfp206, in embryonic stem cells. J Biol Chem 282: 12822–12830, 2007.
Crossref | PubMed | ISI | Google Scholar - 49 Williams RL, Hilton DJ, Pease S, Willson TA, Stewart CL, Gearing DP, Wagner EF, Metcalf D, Nicola NA, Gough NM. Myeloid leukaemia inhibitory factor maintains the developmental potential of embryonic stem cells. Nature 336: 684–687, 1988.
Crossref | PubMed | ISI | Google Scholar - 50 Wobus AM, Wallukat G, Hescheler J. Pluripotent mouse embryonic stem cells are able to differentiate into cardiomyocytes expressing chronotropic responses to adrenergic and cholinergic agents and Ca2+ channel blockers. Differentiation 48: 173–182, 1991.
Crossref | PubMed | ISI | Google Scholar - 51 Xiang F, Sakata Y, Cui L, Youngblood JM, Nakagami H, Liao JK, Liao R, Chin MT. Transcription factor CHF1/Hey2 suppresses cardiac hypertrophy through an inhibitory interaction with GATA4. Am J Physiol Heart Circ Physiol 290: H1997–H2006, 2006.
Link | ISI | Google Scholar - 52 Xu RH, Peck RM, Li DS, Feng X, Ludwig T, Thomson JA. Basic FGF and suppression of BMP signaling sustain undifferentiated proliferation of human ES cells. Nat Meth 2: 185–190, 2005.
Crossref | PubMed | ISI | Google Scholar - 53 Zandstra PW, Bauwens C, Yin T, Liu Q, Schiller H, Zweigerdt R, Pasumarthi KB, Field LJ. Scalable production of embryonic stem cell-derived cardiomyocytes. Tissue Eng 9: 767–778, 2003.
Crossref | PubMed | Google Scholar - 54 Ziegelhoeffer T, Fernandez B, Kostin S, Heil M, Voswinckel R, Helisch A, Schaper W. Bone marrow-derived cells do not incorporate into the adult growing vasculature. Circ Res 94: 230–238, 2004.
Crossref | PubMed | ISI | Google Scholar - 55 Zweigerdt R, Burg M, Willbold E, Abts H, Ruediger M. Generation of confluent cardiomyocyte monolayers derived from embryonic stem cells in suspension: a cell source for new therapies and screening strategies. Cytotherapy 5: 399–413, 2003.
Crossref | PubMed | ISI | Google Scholar
