Temporal gene expression during differentiation of human embryonic stem cells and embryoid bodies

Tamar Dvash1,*, Yoav Mayshar1,*, Henia Darr1,*, Michael McElhaney2, Douglas Barker2, Ofra Yanuka1, Karen J. Kotkow2, Lee L. Rubin2, Nissim Benvenisty1,3 and Rachel Eiges1,3

1 Department of Genetics, The Institute of Life Sciences, The Hebrew University, Jerusalem 91904, Israel and 2 Curis Incorporation, 45 Moulton Street, Cambridge, Massachusetts 02138, USA

3 To whom correspondence should be addressed. or Email: rachela{at}mail.ls.huji.ac.il


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
BACKGROUND: The aim of this study was to characterize human embryonic stem (ES) cells at the molecular level by performing large-scale complementary DNA (cDNA) analysis using DNA micro-arrays. METHODS: The transcription profile of human ES cells was determined by comparing it to 2, 10 and 30-day old embryoid bodies (EBs) using Affymetrix Genechip human micro-arrays (U133). RESULTS: According to this analysis we demonstrate that two human ES cell lines are more close to each other than to their differentiated derivatives. We also show the spectrum of cytokine receptors that they express, and demonstrate the presence of five genes that are highly specific to human ES cells and to germ cells. Moreover, by profiling different stages in the differentiation of human embryoid bodies, we illustrate the clustering of five sets of temporally expressed genes, which could be related to the sequential stages of embryonic development. Among them are known genes that are involved in early pattern formation. CONCLUSIONS: The present study provides a molecular basis for the identity of human ES cells and demonstrates that during their in vitro differentiation they express embryonic specific genes in a stage specific manner.

Key words: differentiation/ES cells/gene expression/micro-arrays/pluripotency


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Human embryonic stem (ES) cells, isolated from the inner cell mass (ICM) of early human embryos, were suggested to be true pluripotent stem cells (Thomson et al., 1998Go; Reubinoff et al., 2000Go). As such, they are an important tool for the study of human embryonic development and cell-based therapies (Schuldiner and Benvenisty, 2001Go). Human ES cells may be differentiated in vitro and in vivo into many different cell types. In vitro, they can be triggered to undergo spontaneous differentiation by forming embryoid bodies (EBs) (Itskovitz-Eldor et al., 2000Go). Initially they generate simple bodies composed of densely packed cells that are surrounded by a single cell layer of endodermal like cells. These become cavitated and eventually accumulate fluid. The resulting cystic EBs, which are a product of spontaneous differentiation, are considered as mature EBs and contain many different cell types (Itskovitz-Eldor et al., 2000Go). Differentiation of human ES cells in vitro may be partially directed by addition of growth factors to the culture media (Schuldiner et al., 2000Go). Indeed, differentiation of human ES cells produces cells that express genes specific to differentiated cells such as neurons (Itskovitz-Eldor et al., 2000Go; Reubinoff et al., 2000Go, 2001Go; Schuldiner et al., 2000Go, 2001Go; Carpenter et al., 2001Go; Zhang et al., 2001Go) cardiomyocytes (Itskovitz-Eldor et al., 2000Go; Schuldiner et al., 2000Go; Kehat et al., 2001Go; Mummery et al., 2002Go) and hematopoietic cells (Itskovitz-Eldor et al., 2000Go; Schuldiner et al., 2000Go; Kaufman et al., 2001Go). In addition, methods for genetically modifying human ES cells now allow labeling and sorting of cells at different stages of their differentiation (Eiges et al., 2001Go).

While it is obvious that human ES cells could become a useful system to study the differentiation of human embryonic cells, presently little is known about the genes they express. Human ES cells do express several markers suggested to reflect human ICM cells (Henderson et al., 2002Go). Recently, mouse ES cells were compared to hematopoietic and neuronal stem cells using DNA micro-arrays (Ivanova et al., 2002Go; Ramalho-Santos et al., 2002Go). These analyses showed that embryonic stem cells have some features in common with adult stem cells, allowing the identification of a cluster of genes which is associated with the ‘stemness’ character of all stem cells (Ivanova et al., 2002Go; Ramalho-Santos et al., 2002Go). Global transcription profiles for undifferentiated human ES cells were independently published by several groups (Sato et al., 2003Go; Sperger et al., 2003Go; Richards et al., 2004Go). In the present study we chose to carry out a large-scale transcription analysis to profile human ES cells at different stages during their differentiation in vitro, as an initial step towards understanding the genetic control of human embryonic differentiation.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Cell culture
Human ES cells (passage 42–47, normal karyotype) and EBs were cultured as previously described (Itskovitz-Eldor et al., 2000Go; Schuldiner et al., 2000Go) with minor modifications. Undifferentiated ES cells (H9 and H13 cell lines) were cultured in 80% KnockOutTM Dulbecco's modified Eagle's medium (Gibco-BRL), supplemented with 20% KnockOutTM SR—a serum-free formulation (Gibco-BRL), 1 mM glutamine (Gibco-BRL), 0.1 mM {beta}-mercaptoethanol (Sigma), 1% non-essential amino acids stock (Gibco-BRL), penicillin (50 U/ml), streptomycin (50 µg/ml) and 4 ng/ml of basic fibroblast growth factor (bFGF), on a mitomycin-C treated mouse embryonic fibroblast (MEF) feeder layer. To reduce the presence of feeder cells in the culture (to <5%), human ES cells were grown on 0.1% gelatin (Merck) coated plates for a single passage. Under these conditions >95% of the cells in the colonies are still undifferentiated as shown by OCT4 staining (see our website: http://www.ls.huji.ac.il/~nissimb/gene_profiling.html). High-density cultures of undifferentiated cells were trypsinized and used either for RNA extraction or for EB formation by allowing cells to aggregate in suspension on plastic petri bacterial dishes. EBs were collected for analysis following 2, 10, 20 and 30 days of cell aggregation in culture.

DNA micro-array analysis
Total RNA was extracted from populations of undifferentiated (confirmed by REX1 and OCT4 expression) and differentiated cell derivatives of human ES cells. RNA extraction was performed according to the manufacturer's protocol (Affymetrix). Hybridization to the DNA micro-arrays, washing and scanning were performed according to the manufacturer's protocols, and compared for expression pattern using Affymetrix U133 DNA chip micro-arrays. Affymetrix DNA micro arrays are provided with human specific probes each of which consists of 20 different sequence combinations, in order to overcome hybridization efficiency differences. Moreover, we have normalized signal value for each probe through dividing by the average signal of the hybridization in each experiment, to reduce differences in signal levels between experiments. Analysis of the results was performed using the GENE SPRING and GO ANNOTATION programs.

The U133 GeneChip human micro-array, contains nearly 45 000 probe sets (representing 33 000 human genes). This DNA micro-array was used to compare the global expression of H9 human ES cells and 2, 10 and 30-day old EBs. Each differentiation stage was assayed in three independent experiments to determine the variability of the system.

ES specific gene expression analysis
Average signal value was calculated for each probe in ES cells and EBs, and ordered according to the ratio of ES/EB value. Complementary (cDNA) probes in which expression was at least as high as the average signal level of the chip (value of 100 following normalization) and for which the ratio between ES and EBs was >20 (total of 50 sequences: see our website: http://www.ls.huji.ac.il/~nissimb/gene_profiling.html) were further examined for tissue distribution by searching in databanks (dbESTs-NCBI and Source-USCS).

Databases
We have used the following databases: Entrez (NCBI)—the text-based search system used at NCBI for its major databases. Unigene (NCBI)—a database which partitions GeneBank sequences into a non-redundant set of gene-oriented clusters, each of which contains sequences that represent a unique gene and its related information i.e. map location and tissues where it is expressed. Blast (NCBI)—a sequence similarity search tool. LocusLink (NCBI)—provides a descriptive information regarding genetic loci (e.g. information on official nomenclature, aliases and sequence accession numbers etc.) and AceView (NCBI)—offers an integrated view of the human genes as reconstructed by alignment of all publicly available mRNAs and Expressed Sequence Tags (ESTs) on the genome sequence. SOURCE (Stanford University)—collects and compiles data about genetics and molecular biology of genes from human and other species genomes. BLAT (UCSC)—contains reference sequences for human and other species and integrate map location and various types of annotation.

RT–PCR analysis
Total RNA was extracted as described (Chomczynski and Sacchi, 1987Go) and 1 µg of RNA was reverse transcribed by random hexamer priming using EZ-First Strand cDNA Synthesis Kit (Biological Industries). Amplification was performed on the cDNA using Takara Ex TaqTM, in the presence of X1 Ex TaqTM Buffer, 200 µM dNTPs each, and 2.5 mM Mg2 + . PCR conditions include a first step of 3 min at 94°C, a second step of 20–30 cycles of 30 s at 94°C, 30 s annealing step at 60–64°C, and 45 s at 72°C, and a final step of 5 min at 72°C. Several markers were examined: OCT4 as a marker for undifferentiated ES cells, GAPDH as a house keeping gene, LEFTY A as a transiently expressed gene and NODAL, LEFTY B and PITX2 as genes related to the LEFTY A pathway. A full description of primers, annealing temperature and size of final products is described in Table I. Final products were assessed by gel electrophoresis on 2% agarose ethidium-bromide stained gels and their identity was verified by direct sequencing.


View this table:
[in this window]
[in a new window]
 
Table I. Primers, annealing temperature and size of final products

 

    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
A molecular characterization of human ES cells
Human ES cells and their differentiated progeny were profiled using Affymetrix GeneChip micro-array. Profiles were obtained for two independent human ES cell lines (H9 and H13) and EBs (from H9 cells) at different stages of differentiation. (Figure 1A). By normalizing signal values and comparing expression levels of all probes between the different samples, as described above, we could demonstrate a significant difference between fully differentiated EBs (30 days) (Itskovitz-Eldor et al., 2000Go) and all other cell types, ES cells and partially differentiated EBs (Figure 1A). Different samples of ES cells or ES cells from different lines (H9 and H13) show a very similar pattern of gene expression (Figure 1B). These data are shown as scatter plots (Figure 1B). Comparing different ES cell lines (H9 and H13) with 30-day old EBs (H9), demonstrates that variation within ES samples is smaller than variation between ES cells and EBs from the same cell line.



View larger version (23K):
[in this window]
[in a new window]
 
Figure 1. Gene profile of human embryonic stem (ES) cell lines and human embryoid bodies (EBs). (A) A graphic illustration demonstrating the degree of similarity between different ES cell lines (H9 and H13) and between EBs at different stages of their differentiation (2, 10 and 30-day old EBs derived from H9 ES cells), using the U133A Affymetrix DNA chip micro-array. Hierarchical clustering was performed as described above. (B) A scatter plot analysis of the degree of similarity in gene expression, based on the U133A Affymetrix micro-array, between two different human ES cell lines (H9 and H13) as compared to the correlation between a given ES cell line (H9) and it's differentiated cell derivatives (30-day old EBs).

 
Self-renewal and a broad developmental potential are important characteristics of stem cells. These two properties are likely to depend on surface molecules, including receptors, which interact with cytokines and other signaling factors. Human ES cells were analyzed for the presence of cytokine receptors and their related factors by examining expression levels (Figure 2A). Of the 74 receptors tested (members of 21 different families), 28 were found to be present in ES cells (Figure 2). Members of the protein tyrosine phosphatase (PTP), fibroblast growth factor (FGF), insulin-like growth factor (IGF), bone morphogenetic protein (BMP), ACTIVIN and tumour necrosis factor (TNF) families of receptors are expressed at relatively high levels whereas transforming growth factor-beta receptor (TGF{beta}R), MET and gp130 are expressed at a low level. In an attempt to identify possible autocrine loops, we examined the expression of cytokines that correspond to the displayed receptors (total of 67). Fifteen of these cytokines are expressed in ES cells (Figure 2B). Interestingly, a rough correlation can be found in the relative level of expression between the receptors and their associated ligands (e.g. see PTP receptors and their growth factors midkine and pleiotrophin, or FGF receptors and their cytokines FGF2, FGF4 and FGF13). By performing hierarchical clustering of the different tissues with respect to their receptors, ES cells show the highest similarity to ovary and uterus; equally so. Yet, EBs more closely resemble placenta (data not shown).



View larger version (20K):
[in this window]
[in a new window]
 
Figure 2. Analysis of expression of cytokines and their receptors in human ES cells. (A) Shown are the relative expression levels of the different cytokine receptors, which are expressed by human ES cells. Levels were normalized relative to the average expression level of the entire micro-array. Full and empty bars represent expressed and non-expressed receptors, respectively, as defined by Affymetrix analysis. Receptor families are color-coded and sorted according to their relative expression levels. The values are given with SE. (B) Shown are cytokines that may signal via an autocrine loop in human ES cells. They are color-coded using the same key as their respective receptors. Only cytokines with their receptors expressed in human ES cells are listed. The values are given with SE.

 
Identification of ES cell specific genes
In order to isolate genes specific to ES cells, a pair wise comparison in gene expression between ES cells and 30-day old EBs was carried out. Of 50 sorted sequences, five were selected since they were highly expressed in germ cells and ES cells (Figure 3). These include four ESTs as well as OCT4. Apart from OCT4, which is associated with pluripotency in mice (Nichols et al., 1998Go), all others which display expression pattern similar to OCT4, have unknown functions but are predicted to be transcription factors, based on their DNA sequence. To verify the DNA micro-array results, the selected sequences were examined for expression levels by RT–PCR in H9 human ES cells and in 20-day old EBs. All show down-regulation upon differentiation. Expression of these markers in undifferentiated cells was also verified in two additional human ES cell lines (data not shown).



View larger version (16K):
[in this window]
[in a new window]
 
Figure 3. Identification of Human ES-specific genes. (A) H9 human ES cells and 30-day old EB (three replicas each) were analyzed using the U133 Affymetrix DNA chip micro-array. Average signal levels were calculated for each probe and ordered according to the ratio of their expression levels in ES/EB. Sequences in which the ratio was >20 (total of 50 sequences), were examined for tissue distribution in available databanks, and selected for further investigation if they were highly abundant in germ cells (total of five sequences). To verify results, the selected genes were analyzed in H9 human ES cells and their differentiated counterparts, 20-day old EB by RT–PCR. Final products, obtained by the amplification of only cDNA sequences, were detected by gel electrophoresis on 2% ethidium-bromide stained gels. Samples were examined for GAPDH as reference control. Expression levels were confirmed in two other human ES cells lines, H13 and BGN1 (data not shown). (B) The sequences were searched for known protein motifs using Aceview database. Line represents untranslated region, open box represents open reading frame. (Domains – Coiled: coiled coil; NLS: nuclear localization signal; Homeo: homeobox; LZ: leucine zipper; Prx: peroxisomal; CSD: cold shock domain; ZF: zinc finger; ER: ER membrane.)

 
Temporal gene expression during differentiation of EBs
To follow differentiation of EBs in vitro, we identified genes that are transiently expressed at different stages of EB formation. Normalized expression levels were compared for each probe between ES cells, 2, 10 and 30-day old EBs (three independent experiments each). These stages represent simple, cavitated and cystic EBs, respectively. The temporally expressed genes, identified by U133 chip analysis, clustered into three groups. Using GENE SPRING software. We classify them as early, mid or late expressed genes (Figure 4A).



View larger version (25K):
[in this window]
[in a new window]
 
Figure 4. Patterns of temporal gene expression during human EB differentiation. (A) Clustering of genes into three major groups, early (red), mid (green) and late (blue) expressed genes. Classification was performed according to the temporal expression pattern of the transiently expressed genes in undifferentiated ES cells, 2, 10 and 30-day old EBs. Expression levels are displayed by the log scale of normalized signal intensity values. (B) Subdividing the analyzed genes into early, early-mid, mid, mid-late and late expression patterns, as demonstrated by the following representative genes: LECTIN (galactoside-binding, soluble 1), OCT4, LEFTY A, {alpha}-FETOPROTEIN and SURFACTANT-D, respectively. The values are given with SE for three independent experiments. (C) Expression patterns of the genes involved in the Nodal signaling pathway, NODAL (early), LEFTY A and LEFTY B (mid) and PITX2 (late), as determined by the DNA micro-array analysis. The values are given with SE for three independent experiments. (D) Validation of chip results by RT–PCR for NODAL, LEFTY A, LEFTY B and PITX2 expression in ES, 2, 10 and 30-day old EBs. All four samples were examined for GPDH as a reference control.

 
A more refined analysis shows that the process of in vitro differentiation may be further subdivided into five groups: early, early-mid, mid, mid-late and late expressed genes, which are either down-regulated immediately or gradually, or begin to up-regulate in maturing or fully matured EBs. Representative gene markers which correspond to the following groups include LECTIN GALACTOSIDASE-BINDING PROTEIN (early), OCT4 (early-mid), LEFTY A (mid), {alpha}-FETOPROTEIN (mid-late) and SURFACTANT-D (late), as demonstrated in Figure 4B. In addition, it is possible to follow differentiation by studying the temporal expression pattern of a set of genes, known to be involved in a given developmental pathway, and that are expressed successively during embryogenesis. As an example, we show temporal expression of the NODAL pathway, that plays a major role in the determination of embryonic axes (left–right, dorsal–ventral, anterior–posterior) as well as in mesoderm induction during early development (Hamada et al., 2002Go). Thus, during EB differentiation we could demonstrate transient expression of NODAL, LEFTY A and LEFTY B growth factors, followed by induction of their target, the PITX2 transcription factor (Figure 4C). This was confirmed by RT–PCR, at non-saturating conditions, for all four genes (Figure 4D).


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
The wide developmental potential and capacity for self-renewal suggests that human ES cells might greatly differ from all other adult somatic tissues in terms of gene expression. Recently, global transcription profiles for hES cells were published using micro-array (Sato et al., 2003Go; Sperger et al., 2003Go), SAGE (Richards et al., 2004Go) and cDNA library analysis (Brandenberger et al., 2004Go). All demonstrate the existence of gene clusters that are expressed at significantly higher levels in hES cells as compared to fully differentiated cells. We wished to further expand these studies by profiling hES cells at different stages of their differentiation (undifferentiated cells, 2, 10 and 30-day old EBs). Such an analysis should allow to follow functional pathways that take part in the process of differentiation as it proceeds during embryo development.

Profiling human ES cells
Hierarchical cluster analysis with different ES cell lines or different cultures of the same line show high similarity. Variation is most probably due to spontaneous differentiation that occurs during ES cell propagation. The difference between ES cells and fully matured EBs is dramatically larger than the variance between two different ES cell lines, representing transcriptional changes that would be expected to accompany differentiation.

As part of our attempt to establish a genetic bar code for ES cells at the transcriptional level, we characterized cytokine receptors and their related growth factors. Of 74 receptors examined, 28 are expressed in ES cells, with a relatively high expression of members of the PTP, FGF, IGF and BMP, activin and TNF receptor families. TGF{beta}R, MET and gp130 seem to be present at a very low level. These results may explain why leukemia inhibitory factor (LIF), which is based on the activation of STAT3 by the gp130-LIF receptor pathway, may not play a role in sustaining undifferentiated growth in human ES cells, as it does in mouse ES cells (Thomson et al., 1998Go; Reubinoff et al., 2000Go). In addition, the high expression level of FGF receptors in ES cells may support the need of bFGF in the growth media (Thomson et al., 1998Go). Since a good correlation exists between the relative level of the receptors and their associated ligands, it should be interesting to investigate the role of these factors in human ES cells.

Pluripotent specific genes
There is presently little information regarding genes that are directly associated with pluripotency and self-renewal. A short list of molecular markers that rapidly down-regulate upon differentiation is available in mouse (for a review, see Eiges and Benvenisty, 2002Go). Most of these are transcription factors that are also expressed by the ICM. However none are exclusively expressed by pluripotent cells. Oct4 is the only gene that has been shown to be directly involved in the maintenance of the undifferentiated state of the cells in vivo (Nichols et al., 1998Go). By comparing the expression profile of human ES cells to their differentiated derivatives, we identified a list of candidate genes which may be associated with the pluripotent nature of the cells. Sequences that are highly expressed in ES cells and down regulated upon differentiation are predicted to be good marker genes for determining the state of differentiation of ES cells in vitro. However, most are expressed by other differentiated cell types. Since the DNA micro-array analysis contains probes of only previously cloned sequences, it is not possible to isolate unique sequences, which are truly ES-specific. Under the assumption that genes associated with pluripotency should be expressed by both early embryonic cells and germ cells, we examined the tissue distribution of highly expressed genes in databanks and identified genes expressed exclusively in ES cells and germ cells (enriched in ovary, teratoma, testis and pure populations of germ cells). Using these constraints, five sequences were defined as ES-specific: OCT4 and four other uncharacterized ESTs. RT–PCR showed down-regulation of all five sequences upon differentiation of ES cells and most have DNA-binding motifs, suggesting a role as transcription factors. The presence of OCT4 strongly supports the strategy used to find these genes. The function of these genes and their targets remains to be determined. In addition, it should be interesting to examine their expression in the ICM of human blastocyst-stage embryos.

Recently, the transcriptional programs in mouse embryonic and adult stem cells were compared to identify transcripts enriched in embryonic, neuronal and hematopoietic stem cells (Ivanova et al., 2002Go; Ramalho-Santos et al., 2002Go). These transcribed genes may characterize the ‘stemness’ of the different types of murine stem cells. We wished to examine whether orthologs of these mouse genes are also enriched in undifferentiated human ES cells, and whether our selected genes appear among them. Although the annotation of probe sets now allows a fairly comprehensive comparison, there are technical limitations (resulting from the construction of different probe sets in the two DNA chips), which prevent us from drawing definitive conclusions. Nonetheless, this analysis suggests that genes implicated in mouse ‘stemness’ of embryonic and adult cells differ from the set of genes we have identified as enriched in undifferentiated human ES cells. We suggest that these two sets of genes may complement one another. In the present study we compared the profile of gene expression in ES cells to that of EBs, their immediate differentiated cell derivatives. These EBs are composed of mature stem cells, progenitor cells and fully differentiated cells. Thus, we aim to identify genes specific to the pluripotent state of human ES cells. This is in contrast to previous studies searching for "stemness genes" in which genes unique, but common, to different types of stem cells (embryonic and adult) were identified. It would be interesting to evaluate in the future the hierarchy and inter-relationship between these two sets of genes.

Human EB differentiation as a model for early development
It is possible to trigger differentiation of ES cells in vitro by growth in suspension culture, resulting in formation of EBs. The EBs, which contain mesodermal, ectodermal and endodermal cells, undergo spontaneous differentiation. This is accompanied by morphological alterations which, similar to mouse, include cavitation (beginning by day 5) and continuous expansion, resulting in a fluid-filled cystic EB (by 2 weeks) (Itskovitz-Eldor et al., 2000Go). It was proposed that expanding EBs mimic, to some extent, early embryonic development. Indeed, in mice it has been shown that some temporal and spatial relationships between developmentally regulated genes which exist in vivo, are recapitulated in vitro (Leahy et al., 1999Go). Thus, we wished to use DNA micro-array analysis in order to characterize and follow the differentiation program of growing human EBs. A detailed analysis shows that in developing EBs some expressed genes are temporally expressed and may be classified into three distinct groups: early, mid and late expressed genes.

A more refined analysis shows that these three groups (ES, early/mid EBs and late EBs) may be further subdivided into: early, early-mid, mid, mid-late, and late expressed genes. Perhaps these subgroups represent sequential stages in embryo development: (i) blastocyst and ICM specific genes, (ii) ICM and primitive ectoderm, (iii) gastrulation, (iv) early organogenesis and (v) late organogenesis. Indeed, LECTIN (galactoside binding, soluble 1) may be a marker for ICM cells; OCT4 is known to be expressed in the mouse in the ICM and the primitive ectoderm (Pelton et al., 2002Go); LEFTY A is a marker for gastrulation (Hamada et al., 2002Go); {alpha}-FETOPROTEIN is expressed in early organogenesis (Gillespie and Uversky, 2000Go); and SURFACTANT-D is expressed in late organogenesis (Crouch, 1998Go) (Figure 4B). Furthermore, such analysis may allow the isolation of new developmentally regulated genes and could serve as an in vitro model for studying aspects of human gastrulation and organogenesis. Nevertheless, since differentiation in EBs is largely disorganized, it remains impractical to study pattern formation in vitro.

Finally, we demonstrate that it is possible to follow differentiation by studying the temporal expression pattern of a cascade of genes that are involved in a given pathway but are active in succession. As an attempt to determine how well in vitro differentiation of EBs correlates with early embryogenesis, we have studied the expression pattern of the Nodal signaling pathway, which holds a major role in the determination of embryonic axes. In the mouse, Nodal induces the expression of Lefty A and Lefty B, which restrict its expression and its downstream target, PITX2, to the left side of the embryo by acting as midline barriers and feedback inhibitors (Hamada et al., 2002Go). By comparing the expression level of NODAL, LEFTY A, LEFTY B and PITX2, in early, mid and fully matured human EBs (2, 10 and 30-day old EBs), we were able to demonstrate their transient expression at different time courses during EB formation, and recover the molecular pathway at the cellular level. These results support the impression that human EBs may model, at least to some extent, early human embryogenesis and encourages us to believe that large-scale cDNA comparisons can provide new insights into the stages of normal human embryo development, which are otherwise inaccessible for research.


    Acknowledgements
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
We thank Professor Joseph Itskovitz-Eldor at the Rambam Medical Center for kindly providing us with the human ES cells as collaboration. We thank Yevgeny Doctor for his assistance in the bioinformatic analysis. This research was partially supported by funds from the Ministry of Science, Israel (R.E.), Herbert Cohn Chair (N.B.), Israel Science Foundation (N.B.) and by an NIH grant (D.A.M.).


    Notes
 
* These authors contributed equally to this work Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Acknowledgements
 References
 
Brandenberger R, Wei H, Zhang S et al. (2004) Transcriptome characterization elucidates signaling networks that control human ES cell growth, differentiation. Nat Biotechnol 22, 707–716.[CrossRef][ISI][Medline]

Carpenter MK, Inokuma MS, Denham J, Mujtaba T, Chiu CP and Rao MS (2001) Enrichment of neurons and neural precursors from human embryonic stem cells. Exp Neurol 172, 383–397.[CrossRef][ISI][Medline]

Chomczynski P and Sacchi N (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162, 156–159.[CrossRef][ISI][Medline]

Crouch EC (1998) Structure, biologic properties, and expression of surfactant protein D (SP-D). Biochim Biophys Acta 1408, 278–289.[ISI][Medline]

Eiges R and Benvenisty N (2002) A molecular view on pluripotent stem cells. FEBS Lett 529, 135.[CrossRef][ISI][Medline]

Eiges R, Schuldiner M, Drukker M, Yanuka O, Itskovitz-Eldor J and Benvenisty N (2001) Establishment of human embryonic stem cell-transfected clones carrying a marker for undifferentiated cells. Curr Biol 11, 514–518.[CrossRef][ISI][Medline]

Gillespie JR and Uversky VN (2000) Structure and function of alpha-fetoprotein: a biophysical overview. Biochim Biophys Acta 1480, 41–56.[ISI][Medline]

Hamada H, Meno C, Watanabe D and Saijoh Y (2002) Establishment of vertebrate left-right asymmetry. Nat Rev Genet 3, 103–113.[CrossRef][ISI][Medline]

Henderson JK, Draper JS, Baillie HS, Fishel S, Thomson JA, Moore H and Andrews PW (2002) Preimplantation human embryos and embryonic stem cells show comparable expression of stage-specific embryonic antigens. Stem Cells 20, 329–337.[Abstract/Free Full Text]

Itskovitz-Eldor J, Schuldiner M, Karsenti D, Eden A, Yanuka O, Amit M, Soreq H and Benvenisty N (2000) Differentiation of human embryonic stem cells into embryoid bodies compromising the three embryonic germ layers. Mol Med 6, 88–95.[ISI][Medline]

Ivanova NB, Dimos JT, Schaniel C, Hackney JA, Moore KA and Lemischka IR (2002) A stem cell molecular signature. Science 298, 601–604.[Abstract/Free Full Text]

Kaufman DS, Hanson ET, Lewis RL, Auerbach R and Thomson JA (2001) Hematopoietic colony-forming cells derived from human embryonic stem cells. Proc Natl Acad Sci USA 98, 10716–10721.[Abstract/Free Full Text]

Kehat I, Kenyagin-Karsenti D, Snir M, Segev H, Amit M, Gepstein A, Livne E, Binah O, Itskovitz-Eldor J and Gepstein L (2001) Human embryonic stem cells can differentiate into myocytes with structural and functional properties of cardiomyocytes. J Clin Invest 108, 407–414.[Abstract/Free Full Text]

Leahy A, Xiong JW, Kuhnert F and Stuhlmann H (1999) Use of developmental marker genes to define temporal and spatial patterns of differentiation during embryoid body formation. J Exp Zool 284, 67–81.[CrossRef][ISI][Medline]

Mummery C, Ward D, van den Brink CE et al. (2002) Cardiomyocyte differentiation of mouse and human embryonic stem cells. J Anat 200, 233–242.[CrossRef][ISI][Medline]

Nichols J, Zevnik B, Anastassiadis K, Niwa H, Klewe-Nebenius D, Chambers I, Scholer H and Smith A (1998) Formation of pluripotent stem cells in the mammalian embryo depends on the POU transcription factor Oct4. Cell 95, 379–391.[ISI][Medline]

Pelton TA, Sharma S, Schulz TC, Rathjen J and Rathjen PD (2002) Transient pluripotent cell populations during primitive ectoderm formation: correlation of in vivo and in vitro pluripotent cell development. J Cell Sci 115, 329–339.[Abstract/Free Full Text]

Ramalho-Santos M, Yoon S, Matsuzaki Y, Mulligan RC and Melton DA (2002) ‘Stemness’ transcriptional profiling of embryonic and adult stem cells. Science 298, 597–600.[Abstract/Free Full Text]

Reubinoff BE, Itsykson P, Turetsky T, Pera MF, Reinhartz E, Itzik A and Ben-Hur T (2001) Neural progenitors from human embryonic stem cells. Nat Biotechnol 19, 1134–1140.[CrossRef][ISI][Medline]

Reubinoff BE, Pera MF, Fong CY, Trounson A and Bongso A (2000) Embryonic stem cell lines from human blastocysts: somatic differentiation in vitro. Nat Biotechnol 18, 399–404.[CrossRef][ISI][Medline]

Richards M, Tan SP, Tan JH, Chan WK and Bongso A (2004) The transcriptome profile of human embryonic stem cells as defined by SAGE. Stem Cells 22, 51–64.[Abstract/Free Full Text]

Sato N, Sanjuan IM, Heke M, Uchida M, Naef F and Brivanlou AH (2003) Molecular signature of human embryonic stem cells and its comparison with the mouse. Dev Biol 260, 404–413.[CrossRef][ISI][Medline]

Schuldiner M and Benvenisty N (2001) Human embryonic stem cells: From human embryogenesis to cell therapy. Recent Res Devel Mol Cell Biol 2, 223–231.

Schuldiner M, Eiges R, Eden A, Yanuka O, Itskovitz-Eldor J, Goldstein RS and Benvenisty N (2001) Induced neuronal differentiation of human embryonic stem cells. Brain Res 913, 201–205.[CrossRef][ISI][Medline]

Schuldiner M, Yanuka O, Itskovitz-Eldor J, Melton DA and Benvenisty N (2000) From the cover: effects of eight growth factors on the differentiation of cells derived from human embryonic stem cells. Proc Natl Acad Sci USA 97, 11307–11312.[Abstract/Free Full Text]

Sperger JM, Chen X, Draper JS, Antosiewicz JE, Chon CH, Jones SB, Brooks JD, Andrews PW, Brown PO and Thomson JA (2003) Gene expression patterns in human embryonic stem cells and human pluripotent germ cell tumors. Proc Natl Acad Sci USA 100, 13350–13355.[Abstract/Free Full Text]

Thomson JA, Itskovitz-Eldor J, Shapiro SS, Waknitz MA, Swiergiel JJ, Marshall VS and Jones JM (1998) Embryonic stem cell lines derived from human blastocysts. Science 282, 1145–1147.[Abstract/Free Full Text]

Zhang SC, Wernig M, Duncan ID, Brustle O and Thomson JA (2001) In vitro differentiation of transplantable neural precursors from human embryonic stem cells. Nat Biotechnol 19, 1129–1133.[CrossRef][ISI][Medline]

Submitted on February 5, 2004; resubmitted on June 7, 2004; accepted on August 26, 2004.