Comparative Transcriptional Analysis of Embryoid Body Versus Two-Dimensional Differentiation of Murine Embryonic Stem Cells

Abstract

Understanding the process of ex vivo embryonic stem (ES) cell differentiation is important for generating higher yields of desirable cell types or lineages and for understanding fundamental aspects of ES differentiation. We used DNA microarray analysis to investigate the differentiation of mouse ES cells cultured under three differentiation conditions. Embryoid body (EB) formation was compared to differentiation on surfaces coated with either gelatin (GEL) or matrigel (MAT). Based on the transcriptional patterns of a list of literature-based “stemness” genes, ES cell differentiation on the two coated surfaces appeared similar but not identical to EB differentiation. A notable difference was the GEL and MAT upregulation but EB downregulation of nine such stemness genes, which are related to cell adhesion and epithelial differentiation. Further, GEL and MAT differentiation showed higher expression of bone formation–related genes (Spp1, Csf1, Gsn, Bmp8b, Crlf1). Gene ontology analysis shows an increase in the expression of genes related to migration and cell structure in all three conditions. Overall, GEL and MAT conditions resulted in a more similar to each other transcriptional profile than to the EB condition, and such differences are apparently related to higher nutrient and metabolite gradients and limitations in the EB versus the GEL or MAT cultures.

Introduction

Embryonic stem (ES) cells derive from the inner cell mass of the blastocyst, and can both self-renew and differentiate into many different cell lineages of the endoderm, mesoderm, and ectoderm germ layers.¹ ES cells have potential clinical applications in the context of regenerative medicine if only their differentiation is sufficiently understood and properly controlled. Mouse ES cells can sustain totipotentiality in vitro either by using feeder layers or by the addition of leukemia inhibitory factor (LIF).² Various cell lineages have been obtained under different culture conditions to produce cell types such as heart,^3,4 blood,^5–7 muscle,⁸ insulin-producing,^9,10 and brain cells.¹¹ Several approaches have been employed to study the behavior of such differentiation processes. Studies have focused on using growth factors, extracellular matrix (ECM) proteins, or chemicals in order to promote differentiation and study the characteristics of the resulting cell populations. Most studies have employed immunocytochemistry techniques, Western blots, and polymerase chain reaction (PCR) amplification of genes of interest. Relatively few studies have used DNA microarray analysis to examine the transcriptional program of ES differentiation. Microarray approaches have been used as a tool to identify a genetic signature for “stemness,”^12–15 and a paper has used transcriptional profiling to identify key steps in general ES differentiation.¹⁶ Several studies have been, however, published on the transcriptional analysis of specialized tissue-specific differentiation, such as pancreas development,¹⁷ angiogenesis,¹⁸ and neuronal differentiation.¹⁹

A standard approach in ES cell differentiation is the formation of embryoid bodies (EBs), which are cell aggregates derived from ES cells. This approach carries some limitations: notably, that of nutrient and metabolite transport limitations from/to the core of the EB to/from the culture medium. This may result in necrotic cells in the core of the EB during the first 5 days prior to vascularization.²⁰ Another limitation is the lack of control over the concentration of growth factors/differentiation agents inside EBs. The latter would result in production of a more heterogeneous cell population as a culture output. Thus, using surface differentiation protocol, whereby all cells can be exposed to similar concentrations of nutrients and differentiation agents, would likely result in a more homogeneous culture outcome.

Here, using DNA microarray analysis, we characterize and compare three differentiation conditions: the standard EB differentiation protocol and two surface differentiation protocols, one using gelatin-coated plates (GEL) and the other Matrigel-coated plates (MAT). Previous studies have shown the ability to differentiate ES cells into hematopoietic and endothelial lineages using a surface (gelatin- or collagen-coated plates) differentiation protocol, whereby formation of EBs is avoided.²¹ Other studies have shown that when using ECM components such as Matrigel, differentiation outputs include cells of the endothelial,²² pancreatic,²³ vascular, and endothelial²⁴ lineages. None of these studies used global gene expression analysis to examine the transcriptional programs underlying differentiation.

Materials and Methods

Chemicals and materials

High-glucose DMEM, gelatin, and penicillin/streptomycin were purchased from Sigma (St. Louis, MO); nonessential amino acids, sodium pyruvate, and fetal calf serum from Gibco-Invitrogen (Carlsbad, CA); and LIF from Chemicon Int. (Temecula, CA). Matrigel matrix Growth Factor Reduced (MATGFR) was purchased from BD Biosciences (San Diego, CA). Total RNA isolation mini kit, Low RNA input fluorescent linear amplification kit, mouse oligo microarrays (G4121A), and hybridization kits were purchased from Agilent Technologies (Wilmington, DE).

Embryonic cell cultures

Murine ES R1 cells²⁵ (a gift from Dr. Peter Zandstra, University of Toronto) were cultured on gelatin-coated flasks in ES-MEM medium (DMEM high glucose, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 100 mM 2-ME, 2 mM L-glutamine, 15% fetal bovine serum [FBS], penicillin/streptomycin 50 mg/mL each, and 1000 U LIF/mL). Cultures were carried out in a humidified incubator at 37ºC with 5% carbon dioxide. For the differentiation protocols, the medium used was ES-Diff (same as ES-MEM but without sodium pyruvate and without LIF). Three conditions were used for cell differentiation: EB formation, where the cells were cultured in nonadherent plastic dishes; GEL plates; and MAT plates. GEL plates were coated with a solution at 0.1% gelatin, the standard gelatin concentration for undifferentiated state maintenance of ES cells. Gelatin coating was used after trying to differentiate the cells on tissue culture–treated plates with no success (the cells still formed some EBs). MAT plates were treated with a 0.1% Matrigel protein (MATGFR). This concentration was chosen to achieve the same protein concentration as the GEL plates. Since cell growth is fast during the first day of differentiation, seeding densities were adjusted to obtain sufficient cell numbers for cell counting and RNA extraction at the day of harvest and to avoid cell replating for plates processed on days 5 and 7. The medium was changed every other day. Cell samples from days 0, 3, 5, and 7 were flash frozen in liquid nitrogen for subsequent RNA isolation. Three independent biological replicate experiments were carried out.

DNA microarray analysis

Total RNA was extracted using Agilent's Total RNA isolation mini kit. RNA integrity was evaluated using a Bioanalyzer 2100 (Agilent, Palo Alto, CA). Microarray analysis was based on the reference design,²⁶ whereby all samples are hybridized against a common reference RNA (see below). This allows for any desirable comparative analyses (here, time course analysis or comparison between differentiation conditions) by examining appropriate expression ratios. The reference RNA, which is meant to contain as many possible transcripts as possible such that meaningful expression ratios can be measured, was a pool of RNA from one experiment, whereby samples were taken daily from an EB experiment for 7 days. Total RNA was extracted and equal amounts from each day were pooled together. Reference RNA was labeled with Cy5 and samples with Cy3 using Agilent's Low RNA input fluorescent linear amplification kit. Agilent microarrays were used according to manufacturer's instructions. Slides were scanned using the Agilent Microarray Scanner G2565BA. Agilent's Feature Extraction software (G2567AA, version 7.2) was used to identify spot and feature outliers. Spot quality was assessed, and if not discarded, multiple spots of the same gene were averaged (geometric mean). The SNNLERM algorithm²⁷ was used to calculate normalized signal intensity ratios and confidence levels. Calculations of expression ratios for the various comparisons (time course or between conditions) were performed with Microsoft Excel. Differentially expressed genes were determined using ANOVA²⁸ considering three biological replicates. After the selection of differentially expressed genes, the biological replicates were averaged and were analyzed using hierarchical clustering²⁹ and Gene Ontology (GO) analysis tools of MultiExperiment Viewer (MeV)³⁰ (The Institute for Genomic Research [http://www.tigr.org/software/tm4/menu/TM4_Biotechniques_2003.pdf]). GO analysis is based on the Expression Analysis Systematic Explorer (EASE) program (version 1.21) (http://david.niaid.nih.gov/david/ease.htm), which identifies overrepresented gene groups within a set of genes using the EASE score.³¹ The Self-Organizing Tree Algorithm (SOTA)^32,33 implemented in MeV was used to cluster groups of genes judged as statistically significant by ANOVA analysis. Raw and normalized data were deposited in the Gene Expression Omnibus (GEO; GSE8574; http://www.ncbi.nlm.nih.gov/geo/).

Quantitative reverse transcription PCRs (Q-RT-PCRs)

Total RNA was reverse transcribed to cDNA using the High Capacity cDNA Archive Kit (Applied Biosystems [ABI], Foster City, CA). PCR reactions were based on the ABI Assay on Demand kit for each gene. Ten μg of RNA of each sample was processed according to manufacturer's directions. Primers (ABI) for the following functionally diverse set of genes were used: GATA binding protein 4 (Gata4), Fibroblast Growth Factor 5 (Fgf5), POU domain, class 5, transcription factor 1 (Pou5f1/Oct-4), and Brachyury (T). Genes were chosen to reflect differentially expressed genes between culture conditions and time points of a wide range of microarray signal intensities and as representatives of each of the germ layers: Gata4 (primitive endoderm³⁴), Fgf5 (ectoderm³⁵), and Brachyury (mesoderm³⁶). We used Gapdh as a normalization (housekeeping) gene.³⁷ PCR reactions were carried out²⁶ using the TaqMan Universal PCR Master Mix (ABI) in 96-well plates on a Bio-Rad iCycler (Bio-Rad, Hercules, CA). Q-RT-PCR assays for the four selected genes were carried out for four time points for each of the three culture conditions. Microarray data are available for almost all of the conditions and time points. Triplicate reactions for each sample were carried out, and the results averaged. Further processing was performed with the iCycler software to obtain Ct values. A dilution curve prepared with reference RNA used in the microarray experiments was used to calibrate the resulting values to correct for variations in RNA amounts used.

Flow cytometry analysis

Tie-2 and its secondary antibody (FITC) were purchased from Chemicon; Ter119 (FITC), CD45 (PerCP), CD31 (PE), CD34 (PE), and ACE (FITC) were purchased from BD Biosciences. Flow cytometry was carried out as described.³⁸ Briefly, cells were dissociated by incubating for 2 min in a PBS buffer with 7.5 mM EDTA, and then the cells were washed twice in a PBS buffer with 4% FBS. Incubation with antibodies was carried out for 30 min at 4°C in the dark. Afterward, cells were washed twice with PBS with 4% FBS and resuspended in the same buffer. The 10,000-gated events from each tube were acquired using a FACScan cytometer (BD Biosciences) and analyzed using CellQuest (BD Biosciences).

Results

Characteristics and reproducibility of ES differentiation protocols and the associated global transcriptional patterns

Three differentiation protocols were used to examine the transcriptional profile of differentiating ES cells. One of the differentiation protocols was the standard EB formation, and the other two employed two different coated surfaces (GEL, MAT) without formation of three dimensional cell structures. The surface-coated differentiation protocols were chosen as a means for better control of the concentration of metabolites, nutrients, and growth factors, or signals produced by the differentiating cells. The expectation was that they would provide as an output a more homogeneous cell population. Three biological replicates were carried out for each condition. For each condition, growth rates, glucose consumption rates, and gene expression profiles for each of the replicate cultures were similar, and thus the cultures and conditions were deemed as very reproducible (data not shown).

Transcriptional patterns of stemness genes

Several studies have examined the stemness genes that define the pluripotency in different stem cell types, from ES cells to adult hematopoietic, neural, and other stem cells.^{12–15,39,40} Among these, three studies have carried out microarray analysis in order to identify gene candidates for stemness: Ivanova et al. (2002), Ramalho-Santos et al. (2002), and Fortunel et al. (2003). Based on these three studies, only one gene was common to all different stem cells. However, for murine ES cells, that number increases to 320 genes. Two hundred and fourteen of those genes (Supplemental Table 1 and Supplemental Fig. 1, available online at www.liebertonline.com/ten) are present on the Agilent microarrays used here. If the aforementioned genes correspond to the stemness state, their expression profile should show a decrease as differentiation progresses. In our study, most of these 214 genes show a decreasing profile over time, in particular, Nanog and Oct-4, which are the genes identified as characterizing the undifferentiated state in mouse ES cells (we will revisit these two genes later). However, there is a group of 31 genes whose expression increases during the 7 days of differentiation examined in our study. These genes were split into two groups: one made up of those whose expression increased in all differentiation conditions (21 genes; group 1 in Supplemental Table 2, available online at www.liebertonline.com/ten), and the other made up of genes whose expression decreased in the EB condition but increased in the other two differentiation conditions (9 genes; group 2 of Supplemental Table 2). The expression pattern of group 1 would disqualify them from being classified as stemness genes. Consistent with that assessment, in vivo, expression of most of the genes in this first group reaches a peak around Theiler stages 10 to 20.^41,42 There is no in vivo expression data for two of the 31 genes with profiles increasing over time, Ercc5 and Sca10 (Atxn10). We observed a common functional theme in the genes of the second group (marked in yellow in Supplemental Fig. 1, and listed as group 2 in Supplemental Table 2), whose expression increases in conditions GEL and MAT but decreases in EB. This theme is related to cell adhesion and epithelial differentiation: Perp has a potential function in epithelial structures and development.⁴³ Lamc2 codes for an ECM glycoprotein that participates in differentiation and cell migration,⁴⁴ and facilitates the basal membrane assembly; expression of Sprr2a has been described to play a role in epithelial differentiation^45,46; Igfbp2 induces the expression of integrin a5,⁴⁷ and its interaction is essential for its function in cellular migration⁴⁸; Tjp1 is part of the “Tight junctions” and forms a scaffold with F-actin; Spint1(Hai-1) expression is correlated with the expression of epithelial cell markers⁴⁹ as well as with a function in branching during embryonic development and morphogenesis⁵⁰; Cldn6 is present in tight junctions and has a function in epitheliazation,⁵¹ while its overexpression causes a defect in the “epidermal permeability barrier”⁵²; when the expression levels of Csrp2 decrease, cell migration increases⁵³; Ddit4 (Redd1) is activated by Hif-1 and was shown to be essential for prostatic cancer invasiveness.⁵⁴

It is worth noting the subtleties of the decreasing expression patterns of the two genes widely accepted as stemness genes, Pou5f1 (Oct-4) and Nanog. The Nanog expression patterns and levels are quite comparable between the three conditions. The Pou5f1 (Oct-4) profiles are somewhat different however. In EB, the expression decreases much faster between days 3 and 5, while the decrease in its expression in GEL and MAT is accelerated between days 5 and 7, but in all three conditions, the final levels of expression were similar (Fig. 1). Such expression differences could be explained by the known mode of function of Oct-4 in differentiation. Niwa et al. (2000) reported that depending on the level of expression of Oct-4, stem cells' fate was different. A small increase in its expression promotes differentiation toward primitive endoderm and mesoderm, repression of Oct-4 results in loss of pluripotency and differentiation toward trophectoderm, and midlevel expression allows the maintenance of pluripotency.⁵⁵

FIG. 1.

Expression profiles of Pou5f1 (Oct-4) (A) and Nanog (B) in the three differentiating conditions, embryoid body formation (EB), gelatin-coated plate (GEL), and Matrigel-coated plate (MAT).

Q-RT-PCR versus microarray data

Q-RT-PCR analysis was carried out on samples of different conditions and time points. The selected genes for this analysis were Oct-4, the standard differentiation indicator, Gata4, Brachyury, and Fgf5 as representatives of differentiation of each germ layer, endoderm, mesoderm, and ectoderm, respectively. We used Gapdh as a housekeeping gene.³⁷ Ratios between each gene and Gapdh were used to compare the expression values with the microarray data. Data were normalized to day 0 to better assess differences between the three culture conditions. While we find directional agreement (which is the crucial point here) between the Q-RT-PCR and microarray data for all four genes (Supplemental Fig. 2, available online at www.liebertonline.com/ten), we observe substantial quantitative differences for a few time points, whereby Q-RT-PCR ratios are typically higher than microarray expression ratios (a rather typical finding^26,56), with a few exceptions for Oct-4 and Gata4. Gapdh data from microarrays corresponding to day3 in EB condition and day 5 for MAT condition were flagged and were not considered for the analysis. The Q-RT-PCR data showed large intrasample variability (data not shown), and generally, Q-RT-PCR data cannot be assumed to be any more accurate than good microarray data.

ANOVA analysis of gene expression

A two-factor ANOVA analysis was performed on the averaged gene expression data on days 3, 5, and 7, which were normalized by the corresponding value on day 0. Each factor had three levels: the “Time” factor (days 3, 5, and 7) (Supplemental Fig. 3, available online at www.liebertonline.com/ten), and the “Condition” factor (EB, GEL, and MAT) (Supplemental Fig. 4, available online at www.liebertonline.com/ten). Subsequently, SOTA clustering was applied to the genes judged as significant by ANOVA analysis for each factor. Two main clusters were identified in the Time factor ANOVA analysis, one with genes whose expression increased over time and the other with expression decreasing over time, marked in yellow and blue, respectively (Supplemental Fig. 3).

During differentiation, the downregulated gene cluster is highly enriched in genes related to metabolism, DNA binding, biosynthesis, cell cycle, and DNA replication

Seven hundred twenty three genes identified by ANOVA Time factor analysis show a decrease in expression over time, while 842 genes show increased expression (Supplemental Fig. 3). These patterns suggest that as they differentiate, cells abandon certain cellular programs and capabilities downregulating a large number of genes, and acquire new specialized functions by expressing an even larger number of genes. In order to obtain a global picture as to which programs and functions are abandoned by the differentiating cells, we carried GO analysis. The main categories with EASE scores between 6.37 × 10⁻⁷ and 9.3 × 10⁻⁵ are related to metabolism, biosynthesis, cell cycle, and proliferation (Table 1). Differentiation is typically associated with reduced metabolism and cycling, so these findings make sense and strengthen the validity and value of our microarray analysis. Some genes are worth noting. Dppa5 (Esg1) expression has been correlated with Oct-4 and Nanog expression, suggesting that Dppa5 is important in the maintenance of the pluripotential state of ES cells⁵⁷; yet, it has been shown that said Dppa5 is dispensable in embryonic development.⁵⁸ Another interesting gene is Gdf3, a member of the transforming growth factor β/bone morphogenetic protein (TGF β/BMP) pathway that works by inhibiting the signaling cascade through BMP and is able to modulate, depending on its level of expression, the pluripotency state or the start of a differentiation process.⁵⁹ Several genes coding for DNA-binding proteins clustered together: Gdf3, Phf17 (Jade-1), Sycp3, Slc7a3, Pipox, Dppa5, Cbr3, Plekha4 (Pepp1), Ndg2, transcription factors Nr5a2 and Zfp40, and many Riken sequences (mouse full cDNA clone sequences from the Riken Institute, Japan, http://fantom2.gsc.riken.go.jp/). Many of these genes code for zinc finger proteins typically localized in the nucleus. Many of the Riken sequences also correspond to zinc finger and metal ion–binding proteins in the nucleus. For example, Phf17 codes for a protein that has two zinc fingers and stabilizes von Hippel-Lindau protein (pVHL), a tumor suppressor in renal cancer,^60,61 and the Sycp3 protein is involved in the meiosis process.⁶² Slc7a3 codes for a specific cationic amino acid transporter in the brain,⁶³ and Zfp40 for a zinc finger protein expressed in germ cells in neuronal embryonic tissue.⁶⁴ In summary, GO analysis suggests that the genes whose expression decreases are associated with cell metabolism and cycling, and that many of these genes code for proteins that are localized in the nucleus and contain zinc fingers.

Table 1.

Gene Ontology–Significant Terms in the Cluster for the Time Factor in the ANOVA Analysis with Increasing and Decreasing Profiles Over Time

Time profile	Accession number	GO term	No. of genes	EASE score
Decrease	GO:0008152	Metabolism	276	6.34E-07
	GO:0009058	Biosynthesis	63	1.02E-05
	GO:0003676	Nucleic acid binding	133	1.32E-05
	GO:0007049	Cell cycle	44	1.41E-05
	GO:0006260	DNA replication	15	3.12E-05
	GO:0000084	S phase of mitotic cell cycle	15	3.52E-05
	GO:0000278	Mitotic cell cycle	24	4.02E-05
	GO:0008283	Cell proliferation	53	5.15E-05
	GO:0006412	Protein biosynthesis	38	7.96E-05
	GO:0009059	Macromolecule biosynthesis	38	7.96E-05
	GO:0006259	DNA metabolism	33	7.98E-05
	GO:0006139	Nucleobase\, nucleoside\, nucleotide, and nucleic acid metabolism	119	9.28E-05
	GO:0004386	Helicase activity	14	1.68E-04
	GO:0008026	ATP-dependent helicase activity	13	1.78E-04
	GO:0016070	RNA metabolism	23	3.35E-04
	Increase	GO:0045892	Negative regulation of transcription\, DNA-dependent	12	3.61E-04
GO:0004674		Protein serine/threonine kinase activity	39	8.00E-04
GO:0004672		Protein kinase activity	43	1.03E-03
GO:0016773		Phosphotransferase activity\, alcohol group as acceptor	47	1.32E-03
GO:0005515		Protein binding	100	3.06E-03
GO:0008565		Protein transporter activity	23	3.35E-03
GO:0016481		Negative regulation of transcription	12	3.58E-03
GO:0006996		Organelle organization and biogenesis	38	3.77E-03
GO:0006468		Protein amino acid phosphorylation	37	3.86E-03
GO:0016772		Transferase activity\, transferring phosphorus-containing groups	53	3.97E-03
GO:0016301		Kinase activity	53	4.20E-03
GO:0007242		Intracellular signaling cascade	53	4.42E-03
GO:0016740		Transferase activity	84	4.61E-03
GO:0000122		Negative regulation of transcription from Pol II promoter	9	5.84E-03
GO:0016310		Phosphorylation	38	5.94E-03

Increased expression of genes related to regulation of transcription and also to cell migration and cell structure–related functions

Upon GO analysis, genes with an increasing expression profile (Supplemental Fig. 3) show enrichment in genes related to negative regulation of transcription (EASE score 3.6 × 10⁻⁴), kinase activity (8 × 10⁻⁴), and protein binding (3.05 × 10⁻³) (Table 1). For example, in the category “regulation of transcription,” Zmynd11 (BS69) codes for a zinc finger protein that has been postulated as a tumor suppressor.⁶⁵ Stat3 codes for a widely recognized transcription factor in different signaling functions being activated by a variety of growth factors and interleukins. It is involved in the Jak/Stat signaling cascade through gp130,⁶⁶ which is also used by Lif. Irf2 is induced by IFNγ and is implicated in epithelial organization.⁶⁷ Hipk2 is implicated in UV-dependent apoptosis and participates in the Wnt/Nlk signaling cascade.⁶⁸ Snai2 is involved in the formation of the neural crest⁶⁹ and has also an important role in metastasis of melanoma cells.⁷⁰ Many of the genes in this category (regulation of transcription) are related to cell migration and structure maintenance.

Bone formation–related genes present higher expression in surface differentiation protocols compared to EB formation

SOTA clustering was performed on the list that was significant for the Condition factor in ANOVA analysis. One of those clusters showed a mostly higher expression in surface differentiation protocols compared to EB differentiation. This cluster (Fig. 2A) includes Pdgfrα (coding for the platelet-derived growth factor receptor alpha, PDGFRα) which is expressed during the G0 phase of the cell cycle.⁷¹ In embryo development, the PDGFR system is essential for the development of lung and connective tissues, the kidney, cardiovascular system, and brain.⁷² Generally, the PDGFR system is located primarily on mesenchymal cells and transduces signals for cell survival, proliferation, and chemotaxis.⁷³ Also of some interest are the genes Spp1, Csf1, Gsn, Bmp8b, and Crlf1, which display higher levels of expression in GEL and MAT compared to EB. These genes are related with the differentiation and formation of bone, by osteoclastogenesis and bone marrow maintenance (Csf1⁷⁴), cartilage and bone formation (Bmp8b/Osteogenic protein 3), osteoblast organization (Gsn⁷⁵), or expression in osteosarcomas and osteoclasts (Spp1/Osteopontin⁷⁶).

FIG. 2.

(A) Eisen plot (heat map) of gene expression showing bone formation–related genes presenting higher expression in surface differentiation protocols compared to EB formation. Ratios shown are against day 0 expression levels. (B) Eisen plot (heat map) of gene expression showing genes related with negative regulation of cell cycle presenting higher expression in EB compared to surface differentiation protocols. Ratios shown are against day 0 expression levels.

Based on these results we examined the possibility of confirming by immunostaining techniques the bone differentiation characteristic cartilage and mineral deposition. However, based on the literature, these events occur at more advanced stages during development.^77,78 Hence, we decided this would not add useful information. Thus, we sought if more general characteristics, namely, endothelial and hematopoietic markers, would provide other means to discriminate among the three culture conditions, since bone formation, endothelium development, and hematopoiesis are all related as they are part of mesoderm development. We performed flow cytometry assays with CD45, Ter119 (hematopoietic markers), Tie2, ACE (endothelial markers), CD31, and CD34 (hematopoietic and endothelial markers). CD31 dynamics on GEL and MAT followed a similar trend, reaching a peak on day 7 (Table 2). CD31 expression was significantly higher (especially in the first 7 days) in EB cells compared to the GEL and MAT. With respect to Tie-2, all three conditions show a low and decreasing level of expression, although the EB condition shows an increase around days 10–12 (Table 2).

Table 2.

Cell Percentages Shown Positive for CD31 and Tie2 Using Flow Cytometry

	CD31			Tie2
Day	EB	GEL	MAT	EB	GEL	MAT
2	9.20	5.40	3.95	4.80	1.68	3.01
3	6.79	5.24	10.46	4.23	4.01	6.90
4	11.37	0.05	8.27	2.16	0.25	1.73
5	25.65	4.48	8.65	6.15	0.98	2.55
6	9.96	4.06	4.93	0.45	0.13	0.08
7	31.79	14.95	18.01	1.69	0.77	0.73
8	10.96	11.37	11.61	0.69	0.19	0.23
10	8.13	7.13	9.04	1.83	0.53	0.22
12	10.03	2.62	3.28	5.79	0.06	0.08

CD45 and Ter119 were not detected until day 6, when a substantial increase was detected for EB and a more modest increase for the other two conditions (Table 3), although all levels were in the low single digits. This behavior is consistent with previous reports to the effect that CD45 and Ter119 expression is detected later in development.^41,42 CD34 and ACE were not detected during the time period of this study in any of the conditions, which is consistent with the reported late expression for ACE.^41,42 These data suggest that endothelium development is higher under the EB versus the GEL and MAT conditions, that hematopoietic development is delayed compared to endothelial development, and that the various markers capture endothelial or hematopoietic development by different expression patterns.

Table 3.

Cell Percentages Shown Positive for CD45 and Ter119 Using Flow Cytometry

	CD45			Ter119
Day	EB	GEL	MAT	EB	GEL	MAT
2	N/D	N/D	N/D	N/D	N/D	N/D
3	N/D	N/D	N/D	N/D	N/D	N/D
4	N/D	N/D	N/D	N/D	N/D	N/D
5	N/D	N/D	N/D	N/D	N/D	N/D
6	0.54	0.17	0.10	0.14	0.13	0.10
7	0.24	0.00	0.00	0.26	0.07	0.00
8	2.26	0.08	0.04	0.05	0.02	0.04
10	3.20	0.52	0.24	0.44	0.20	0.26
12	2.82	0.16	0.19	1.10	0.13	0.25

Another interesting group of genes (Fig. 2A) with higher expression levels in GEL and MAT is related with the maintenance/remodeling of intracellular or extracellular structure. Spp1 is involved in formation of cell aggregates; the product of Pcolce2 improves the procollagen I processing through Bmp1, Krt2-8, and Krt2-7; the Krt1-19 protein forms intracellular filaments and participates in differentiation. The Gsn protein organizes the nucleation during different stages of cell division, and Klf5 participates in tissue remodeling at the angiogenesis level. The Serpine1 protein inhibits the degradation of ECM and participates in the remodeling of tissue, Msln participates in cell shape changes, Perp participates in functions of cell adhesion to the extracellular matrix, S100a6 participates in metastasis processes in some cancer, and Wnt10a has been associated to ECM.

Among the genes expressed at higher levels in EB compared to surface GEL or MAT, only Cdh11 and Slit2 (Fig. 2A) are related with structure and tissue remodeling. Cdh11 is mainly expressed in EB: its function is in cell–cell adhesion, is involved in metastasis of some human tumors,⁷⁹ and has some function in osteoblast development⁸⁰ and in differentiation of hippocampus neurons.⁸¹ Slit2 participates in the definition of neuronal patterns. Nevertheless, the number of genes related to osteogenesis is higher in surface differentiation conditions compared to EB formation.

Genes related with negative regulation of cell cycle are expressed higher in EBs compared to surface differentiation

By GO analysis, the best represented categories among genes expressed higher in EB are nucleotide metabolism, cell cycle regulation, transcription, metabolism, and cell proliferation. Genes with higher expression in surface differentiation protocols are enriched in GO categories such as structural activity molecules, receptor binding, cytoskeleton constituents, and cytokine and chemoattractant activities.

By SOTA clustering of genes identified as significant for the Condition factor in the ANOVA analysis, a group of genes expressed higher in EB was identified to be associated with negative regulation of the cell cycle (Fig. 2B). In this group of genes, Gas1, Ccnd2, and Sesn3 block the entrance to S, while Ccng2 and Pot1 are negative regulators of proliferation and cell cycle progression. This overall upregulation of cell cycle–negative regulators could be connected with the downregulation of cell cycle–related genes we described above. A second group of genes with higher expression in EBs is related with organ formation (Fig. 2B). Such genes represent organ differentiation from the three different germ layers: heart (Gata6, Gas1, Chd3, Adcy6, and Pcnx), nervous system (Zfp521, Mxd4, and Adcy6), and various internal organs (Chd3, Pdgfrα, and Gas1). This contrasts with the surface differentiation protocols since GEL and MAT conditions show higher expression of genes related to bone formation primarily, which is derived from the mesoderm germ layer.

The two surface differentiation conditions (GEL and MAT) display a more similar pattern of gene expression to each other than to EB differentiation

Among the 571 genes identified by ANOVA analysis as being significant for the Condition factor, about half (green cluster) were expressed overall higher in EB than in GEL or MAT, and the other half the other way around (Supplemental Fig. 4). It is notable that the expression patterns among the two surface differentiation conditions (GEL and MAT) were more similar to each other than to EB differentiation. The obvious interpretation is that these differences derive from the presumed differences between the EB and the surface differentiation protocols in the gradients of nutrients (including oxygen),^82,83 metabolites, and extracellular signaling molecules as already discussed. Such differences likely result in substantial direct and indirect changes in gene expression patterns as clearly displayed in Supplemental Figure 4 (available online at www.liebertonline.com/ten).

We also assessed the similarity and differences of the three protocols by the distribution properties of cell size (forward scatter, FSC) and inner complexity/granularity (side scatter, SSC) characteristics using flow cytometry. As shown in Supplemental Figure 5 (available online at www.liebertonline.com/ten), the population from the EB condition shows a broader distribution in terms of SSC on day 12 compared to the other two conditions (GEL and MAT), which show only a slight change in SSC and a more pronounced change in FSC. These data show a reduction in cell size in all three conditions but with a more pronounced reduction in EB. The broader distribution of SSC properties in the EB condition indicates a more heterogeneous population in terms of inner organelles and cellular granularity, in contrast to the more homogeneous population in the case of GEL and MAT. Again, based on the FSC and SSC properties, the GEL and MAT cells appear quite similar and very distinct from the EB cells, which is identical to the conclusion drawn from the microarray analysis.

Discussion and Conclusions

A practical aim of ES cell culture is to achieve lineage specific differentiation. To achieve this goal, efforts have focused on identifying genes that make this possible, in order to design protocols for directing differentiation toward a particular cell lineage. In this study we used microarray analysis to compare three different differentiation protocols. The first and frequently used one, formation of EBs, was compared to two surface differentiation (GEL and MAT) protocols. In the past, microarray analysis had been used mainly with pure or isolated cell types, that is, specific cells expressing certain surface markers. Our approach is based on analyzing the whole differentiating population, and the analysis is done by examining the enrichment of genes associated with a cell lineage or program or stemness genes. This approach has the advantage of examining a particular differentiation protocol in terms of enrichment for genes of both expected and unexpected lineages or cellular programs and functions. Here we assessed the impact of the three protocols on the maintenance of stemness, and found similarities but also differences, thus suggesting that each protocol impacts differentiation in a complex multigenic manner that cannot be easily captured by traditional low-throughput methods.

We found that expression of many genes related ontologically to cell metabolism decreases over time for all three protocols. This is consistent with the decrease in the proliferation rate present in all three conditions. It is also worth noting that many of the proteins coded by genes whose expression decreased over time have a nuclear localization and that many of them are DNA-binding proteins containing a zinc finger.

A notable category of upregulated genes for all three differentiation protocols is related to cell migration and structure, and includes genes coding for proteins participating in forming junctional plaques or tight junctions. Genes displaying higher expression in the surface differentiation protocols compared to EB in the stemness category, and upregulated genes of the “negative regulation of transcription” and “bone formation” are also related to cell migration and structure maintenance/organization, suggesting a balance between cell migration and cellular reorganization during differentiation.

While genes with higher expression in the GEL and MAT protocols are related to relatively few strongly overrepresented cellular categories (e.g., bone formation), genes with higher expression in the EB condition are enriched for multiple categories (heart, nervous system, and various internal organs). These differences in expression patterns (displayed in Supplemental Fig. 4) are most likely due to differences in concentration gradients of nutrients and metabolites: EB cultures have more and higher such gradients versus the surface protocols.

During development, bone is formed from cells originating from distinct embryonic lineages. Cranial neural crest cells, paraxial mesoderm, and lateral plate mesodermal cells migrate to the place where bone will be formed, form mesenchymal condensations, and finally differentiate into osteoclasts or chondrocytes.⁸⁴ From this point, bone development can follow one of the two paths: direct differentiation to intramembranous bone or differentiation to chondrocytes that will form cartilage.⁸⁴ Finally, cartilage will undergo endochondral ossification that will result in functional bone tissue.⁸⁴ The osteogenesis-related genes we identified as expressed higher under the GEL and MAT conditions, namely, Csf1, Gsn, Spp1, Bmp8b, and Crlf1, all code for proteins that apparently participate in the endochondral pathway of osteogenesis or in migration. For example, the proteins coded by Spp1 and Gsn participate in the assembly of podosomes binding to α4 β3 integrin.⁷⁵ Together they provide a mobility function to osteoclasts. The Csf1 protein has been described as a osteoclast growth factor providing differentiation of mononuclear hematocytes to osteoclast.⁸⁵ Also worth noting is Crlf1, which is expressed in all precartilagenous tissues and is known to participate in endochondral differentiation⁸⁶ and Bmp8b, which has been described as a cartilage marker.^41,42

The way bone is developed presents multiple challenges to tissue engineering, both in terms of coordination of signaling and of the supporting structure. The first point requires complex mixtures of growth factors and cytokines in order to provide cells with osteogenic potential. Our data show that the surface differentiation approaches, the MAT condition in particular, give rise to cells expressing bone-related genes. Significantly, such genes are expressed in differentiating cells even in the absence of a three-dimensional structure. This could be further investigated and refined to obtain specialized cell types such as osteoblasts or osteoclasts. The second point to consider in terms of tissue-engineering challenges is the matrix-supporting cells. This requires use of an appropriate biodegradable matrix. Recent studies⁸⁷ show the use of a new approach called micromass technology. This technology considers cell reaggregation in cellular spheres. Such spheres are cultured until an extracellular-supporting matrix is developed allowing manipulation of the generated microtissue. This technology requires a source of cells with osteogenic potential. It is likely that our protocols can be further extended to provide cells for this new technology.

Footnotes

Acknowledgments

We acknowledge the financial support from a Fulbright fellowship (D.E.S.), the Millennium Scientific Initiative (Institute grant P-05-001-F), Fondecyt 1061119, the National Institutes of Health (USA) (grant R01-GM065476), and the Northwestern University Lurie Cancer Center toward the purchase of the Agilent microarray scanner, and the use of the Keck Biophysics facilities.

Work was carried out at the Department of Chemical and Biological Engineering at Northwestern University, Evanston, Illinois, and at the University of Chile and University of Delaware at the addresses above.

References

Wobus

A.M.

, Boheler

K.R.

Embryonic stem cells: prospects for developmental biology and cell therapy. Physiol Rev, 85:635. 2005.

Murray

, Edgar

The regulation of embryonic stem cell differentiation by leukaemia inhibitory factor (LIF)

Differentiation, 68:227. 2001.

Doetschman

T.C.

, Eistetter

, Katz

, Schmidt

, Kemler

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. 1985.

Zandstra

P.W.

, Bauwens

, Yin

, Liu

, Schiller

, Zweigerdt

, Pasumarthi

K.B.

, Field

L.J.

Scalable production of embryonic stem cell-derived cardiomyocytes. Tissue Eng, 9:767. 2003.

Palacios

, Golunski

, Samaridis

In vitro generation of hematopoietic stem cells from an embryonic stem cell line. Proc Natl Acad Sci USA, 92:7530. 1995.

Potocnik

A.J.

, Nielsen

P.J.

, Eichmann

In vitro generation of lymphoid precursors from embryonic stem cells. EMBO J, 13:5274. 1994.

Choi

, Kennedy

, Kazarov

, Papadimitriou

J.C.

, Keller

A common precursor for hematopoietic and endothelial cells. Development, 125:725. 1998.

Goldring

, Partridge

, Watt

Muscle stem cells. J Pathol, 197:457. 2002.

Assady

, Maor

, Amit

, Itskovitz-Eldor

, Skorecki

K.L.

, Tzukerman

Insulin production by human embryonic stem cells. Diabetes, 50:1691. 2001.

10.

Lumelsky

, Blondel

, Laeng

, Velasco

, Ravin

, McKay

Differentiation of embryonic stem cells to insulin-secreting structures similar to pancreatic islets. Science, 292:1389. 2001.

11.

Lee

S.H.

, Lumelsky

, Studer

, Auerbach

J.M.

, McKay

R.D.

Efficient generation of midbrain and hindbrain neurons from mouse embryonic stem cells. Nat Biotechnol, 18:675. 2000.

12.

Ivanova

N.B.

, Dimos

J.T.

, Schaniel

, Hackney

J.A.

, Moore

K.A.

, Lemischka

I.R.

A stem cell molecular signature. Science, 298:601. 2002.

13.

Fortunel

N.O.

, Otu

H.H.

, Ng

H.H.

, Chen

, Mu

, Chevassut

, Li

, Joseph

, Bailey

, Hatzfeld

J.A.

, Hatzfeld

, Usta

, Vega

V.B.

, Long

P.M.

, Libermann

T.A.

, Lim

Comment on “ ‘Stemness': transcriptional profiling of embryonic and adult stem cells” and “a stem cell molecular signature”

Science, 302:393; author reply 393. 2003.

14.

Ramalho-Santos

, Yoon

, Matsuzaki

, Mulligan

R.C.

, Melton

D.A.

“Stemness”: transcriptional profiling of embryonic and adult stem cells. Science, 298:597. 2002.

15.

Bhattacharya

, Miura

, Brandenberger

, Mejido

, Luo

, Yang

A.X.

, Joshi

B.H.

, Ginis

, Thies

R.S.

, Amit

, Lyons

, Condie

B.G.

, Itskovitz-Eldor

, Rao

M.S.

, Puri

R.K.

Gene expression in human embryonic stem cell lines: unique molecular signature. Blood, 103:2956. 2004.

16.

Brandenberger

, Wei

, Zhang

, Lei

, Murage

, Fisk

G.J.

, Li

, Xu

, Fang

, Guegler

, Rao

M.S.

, Mandalam

, Lebkowski

, Stanton

L.W.

Transcriptome characterization elucidates signaling networks that control human ES cell growth and differentiation. Nat Biotechnol, 22:707. 2004.

17.

Wells

J.M.

Genes expressed in the developing endocrine pancreas and their importance for stem cell and diabetes research. Diabetes Metab Res Rev, 19:191. 2003.

18.

Nakagawa

, Abe

, Yamazaki

, Miyashita

, Niwa

, Kokubun

, Sato

HEX acts as a negative regulator of angiogenesis by modulating the expression of angiogenesis-related gene in endothelial cells in vitro. Arterioscler Thromb Vasc Biol, 23:231. 2003.

19.

Ahn

J.I.

, Lee

K.H.

, Shin

D.M.

, Shim

J.W.

, Kim

C.M.

, Kim

, Lee

S.H.

, Lee

Y.S.

Temporal expression changes during differentiation of neural stem cells derived from mouse embryonic stem cell. J Cell Biochem, 93:563. 2004.

20.

Wartenberg

, Donmez

, Ling

F.C.

, Acker

, Hescheler

, Sauer

Tumor-induced angiogenesis studied in confrontation cultures of multicellular tumor spheroids and embryoid bodies grown from pluripotent embryonic stem cells. FASEB J, 15:995. 2001.

21.

Nishikawa

S.I.

, Nishikawa

, Hirashima

, Matsuyoshi

, Kodama

Progressive lineage analysis by cell sorting and culture identifies FLK1+VE-cadherin + cells at a diverging point of endothelial and hemopoietic lineages. Development, 125:1747. 1998.

22.

Ruhnke

, Ungefroren

, Zehle

, Bader

, Kremer

, Fandrich

Long-term culture and differentiation of rat embryonic stem cell-like cells into neuronal, glial, endothelial, and hepatic lineages. Stem Cells, 21:428. 2003.

23.

Gao

, Ustinov

, Pulkkinen

M.A.

, Lundin

, Korsgren

, Otonkoski

Characterization of endocrine progenitor cells and critical factors for their differentiation in human adult pancreatic cell culture. Diabetes, 52:2007. 2003.

24.

Gerecht-Nir

, Ziskind

, Cohen

, Itskovitz-Eldor

Human embryonic stem cells as an in vitro model for human vascular development and the induction of vascular differentiation. Lab Invest, 83:1811. 2003.

25.

Nagy

, Rossant

, Nagy

, Abramow-Newerly

, Roder

J.C.

Derivation of completely cell culture-derived mice from early-passage embryonic stem cells. Proc Natl Acad Sci USA, 90:8424. 1993.

26.

Ramsborg

C.G.

, Papoutsakis

E.T.

Global transcriptional analysis delineates the differential inflammatory response interleukin-15 elicits from cultured human T cells. Exp Hematol, 35:454. 2007.

27.

Yang

, Haddad

, Tomas

, Alsaker

, Papoutsakis

E.T.

A segmental nearest neighbor normalization and gene identification method gives superior results for DNA-array analysis. Proc Natl Acad Sci USA, 100:1122. 2003.

28.

Zar

J.H.

Biostatistical Analysis. NJ: Prentice Hall, 1999.

29.

Eisen

M.B.

, Spellman

P.T.

, Brown

P.O.

, Botstein

Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA, 95:14863. 1998.

30.

Saeed

A.I.

, Sharov

, White

, Li

, Liang

, Bhagabati

, Braisted

, Klapa

, Currier

, Thiagarajan

, Sturn

, Snuffin

, Rezantsev

, Popov

, Ryltsov

, Kostukovich

, Borisovsky

, Liu

, Vinsavich

, Trush

, Quackenbush

TM4: a free, open-source system for microarray data management and analysis. Biotechniques, 34:374. 2003.

31.

Dennis

Jr. , Sherman

B.T.

, Hosack

D.A.

, Yang

, Gao

, Lane

H.C.

, Lempicki

R.A.

DAVID: Database for Annotation, Visualization, and Integrated Discovery. Genome Biol, 4:P3. 2003.

32.

Dopazo

, Carazo

J.M.

Phylogenetic reconstruction using an unsupervised growing neural network that adopts the topology of a phylogenetic tree. J Mol Evol, 44:226. 1997.

33.

Herrero

, Valencia

, Dopazo

A hierarchical unsupervised growing neural network for clustering gene expression patterns. Bioinformatics, 17:126. 2001.

34.

Arceci

R.J.

, King

A.A.

, Simon

M.C.

, Orkin

S.H.

, Wilson

D.B.

Mouse GATA-4: a retinoic acid-inducible GATA-binding transcription factor expressed in endodermally derived tissues and heart. Mol Cell Biol, 13:2235. 1993.

35.

Haub

, Goldfarb

Expression of the fibroblast growth factor-5 gene in the mouse embryo. Development, 112:397. 1991.

36.

Herrmann

B.G.

Expression pattern of the Brachyury gene in whole-mount TWis/TWis mutant embryos. Development, 113:913. 1991.

37.

Murphy

C.L.

, Polak

J.M.

Differentiating embryonic stem cells: GAPDH, but neither HPRT nor beta-tubulin is suitable as an internal standard for measuring RNA levels. Tissue Eng, 8:551. 2002.

38.

Eisen

M.B.

, Spellman

P.T.

, Brown

P.O.

, Botstein

Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA, 95:14863. 1998.

39.

Sato

, Sanjuan

I.M.

, Heke

, Uchida

, Naef

, Brivanlou

A.H.

Molecular signature of human embryonic stem cells and its comparison with the mouse. Dev Biol, 260:404. 2003.

40.

Rao

R.R.

, Calhoun

J.D.

, Qin

, Rekaya

, Clark

J.K.

, Stice

S.L.

Comparative transcriptional profiling of two human embryonic stem cell lines. Biotechnol Bioeng, 88:273. 2004.

41.

Eppig

J.T.

, Bult

C.J.

, Kadin

J.A.

, Richardson

J.E.

, Blake

J.A.

the Mouse Genome Database Group.The Mouse Genome Database (MGD): from genes to mice—a community resource for mouse biology. Nucleic Acids Res, 33:D471. 2005.

42.

Hill

D.P.

, Begley

D.A.

, Finger

J.H.

, Hayamizu

T.F.

, McCright

I.J.

, Smith

C.M.

, Beal

J.S.

, Corbani

L.E.

, Blake

J.A.

, Eppig

J.T.

, Kadin

J.A.

, Richardson

J.E.

, Ringwald

The mouse Gene Expression Database (GXD): updates and enhancements. Nucleic Acids Res, 32:D568. 2004.

43.

Ihrie

R.A.

, Marques

M.R.

, Nguyen

B.T.

, Horner

J.S.

, Papazoglu

, Bronson

R.T.

, Mills

A.A.

, Attardi

L.D.

Perp is a p63-regulated gene essential for epithelial integrity. Cell, 120:843. 2005.

44.

Salo

, Haakana

, Kontusaari

, Hujanen

, Kallunki

, Tryggvason

Laminin-5 promotes adhesion and migration of epithelial cells: identification of a migration-related element in the gamma2 chain gene (LAMC2) with activity in transgenic mice. Matrix Biol, 18:197. 1999.

45.

Mischke

, Korge

B.P.

, Marenholz

, Volz

, Ziegler

Genes encoding structural proteins of epidermal cornification and S100 calcium-binding proteins form a gene complex (“epidermal differentiation complex”) on human chromosome 1q21. J Invest Dermatol, 106:989. 1996.

46.

Brembeck

F.H.

, Opitz

O.G.

, Libermann

T.A.

, Rustgi

A.K.

Dual function of the epithelial specific ets transcription factor, ELF3, in modulating differentiation. Oncogene, 19:1941. 2000.

47.

Wang

, Wang

, Shen

, Huang

, Hu

, Ramdas

, Zhou

Y.H.

, Liao

W.S.

, Fuller

G.N.

, Zhang

Insulin-like growth factor binding protein 2 enhances glioblastoma invasion by activating invasion-enhancing genes. Cancer Res, 63:4315. 2003.

48.

Wang

G.K.

, Hu

, Fuller

G.N.

, Zhang

An interaction between insulin-like growth factor binding protein 2 and integrin alpha 5 is essential for IGFBP2-induced cell mobility. J Biol Chem, 281:14085. 2006.

49.

Oberst

, Anders

, Xie

, Singh

, Ossandon

, Johnson

, Dickson

R.B.

, Lin

C.Y.

Matriptase and HAI-1 are expressed by normal and malignant epithelial cells in vitro and in vivo. Am J Pathol, 158:1301. 2001.

50.

Tanaka

, Nagaike

, Takeda

, Itoh

, Kohama

, Fukushima

, Miyata

, Uchiyama

, Uchinokura

, Shimomura

, Miyazawa

, Kitamura

, Yamada

, Kataoka

Hepatocyte growth factor activator inhibitor type 1 (HAI-1) is required for branching morphogenesis in the chorioallantoic placenta. Mol Cell Biol, 25:5687. 2005.

51.

Turksen

, Troy

T.C.

Claudin-6: a novel tight junction molecule is developmentally regulated in mouse embryonic epithelium. Dev Dyn, 222:292. 2001.

52.

Turksen

, Troy

T.C.

Permeability barrier dysfunction in transgenic mice overexpressing claudin 6. Development, 129:1775. 2002.

53.

Wei

, Gorman

T.E.

, Liu

, Ith

, Tseng

, Chen

, Simon

D.I.

, Layne

M.D.

, Yet

S.F.

Increased neointima formation in cysteine-rich protein 2-deficient mice in response to vascular injury. Circ Res, 97:1323. 2005.

54.

Schwarzer

, Tondera

, Arnold

, Giese

, Klippel

, Kaufmann

REDD1 integrates hypoxia-mediated survival signaling downstream of phosphatidylinositol 3-kinase. Oncogene, 24:1138. 2005.

55.

Niwa

, Miyazaki

, Smith

A.G.

Quantitative expression of Oct-3/4 defines differentiation, dedifferentiation or self-renewal of ES cells. Nat Genet, 24:372. 2000.

56.

Fuhrken

P.G.

, Chen

, Miller

W.M.

, Papoutsakis

E.T.

Comparative, genome-scale transcriptional analysis of CHRF-288-11 and primary human megakaryocytic cell cultures provides novel insights into lineage-specific differentiation. Exp Hematol, 35:476. 2007.

57.

Western

, Maldonado-Saldivia

, van den Bergen

, Hajkova

, Saitou

, Barton

, Surani

M.A.

Analysis of Esg1 expression in pluripotent cells and the germline reveals similarities with Oct4 and Sox2 and differences between human pluripotent cell lines. Stem Cells, 23:1436. 2005.

58.

Amano

, Itakura

, Maruyama

, Ichisaka

, Nakagawa

, Yamanaka

Identification and targeted disruption of the mouse gene encoding ESG1 (PH34/ECAT2/DPPA5)

BMC Dev Biol, 6:11. 2006.

59.

Levine

A.J.

, Brivanlou

A.H.

GDF3, a BMP inhibitor, regulates cell fate in stem cells and early embryos. Development, 133:209. 2006.

60.

Zhou

M.I.

, Foy

R.L.

, Chitalia

V.C.

, Zhao

, Panchenko

M.V.

, Wang

, Cohen

H.T.

Jade-1, a candidate renal tumor suppressor that promotes apoptosis. Proc Natl Acad Sci USA, 102:11035. 2005.

61.

Panchenko

M.V.

, Zhou

M.I.

, Cohen

H.T.

von Hippel-Lindau partner Jade-1 is a transcriptional co-activator associated with histone acetyltransferase activity. J Biol Chem, 279:56032. 2004.

62.

Yuan

, Liu

J.G.

, Hoja

M.R.

, Wilbertz

, Nordqvist

, Hoog

Female germ cell aneuploidy and embryo death in mice lacking the meiosis-specific protein SCP3. Science, 296:1115. 2002.

63.

Hosokawa

, Sawamura

, Kobayashi

, Ninomiya

, Miwa

, Masaki

Cloning and characterization of a brain-specific cationic amino acid transporter. J Biol Chem, 272:8717. 1997.

64.

Noce

, Fujiwara

, Ito

, Takeuchi

, Hashimoto

, Yamanouchi

, Higashinakagawa

, Fujimoto

A novel murine zinc finger gene mapped within the tw18 deletion region expresses in germ cells and embryonic nervous system. Dev Biol, 155:409. 1993.

65.

Ladendorff

N.E.

, Wu

, Lipsick

J.S.

BS69, an adenovirus E1A-associated protein, inhibits the transcriptional activity of c-Myb. Oncogene, 20:125. 2001.

66.

Daheron

, Opitz

S.L.

, Zaehres

, Lensch

W.M.

, Andrews

P.W.

, Itskovitz-Eldor

, Daley

G.Q.

LIF/STAT3 signaling fails to maintain self-renewal of human embryonic stem cells. Stem Cells, 22:770. 2004.

67.

Hida

, Ogasawara

, Sato

, Abe

, Takayanagi

, Yokochi

, Sato

, Hirose

, Shirai

, Taki

, Taniguchi

CD8(+) T cell-mediated skin disease in mice lacking IRF-2, the transcriptional attenuator of interferon-alpha/beta signaling. Immunity, 13:643. 2000.

68.

Hofmann

T.G.

, Moller

, Sirma

, Zentgraf

, Taya

, Droge

, Will

, Schmitz

M.L.

Regulation of p53 activity by its interaction with homeodomain-interacting protein kinase-2. Nat Cell Biol, 4:1. 2002.

69.

Savagner

, Yamada

K.M.

, Thiery

J.P.

The zinc-finger protein slug causes desmosome dissociation, an initial and necessary step for growth factor-induced epithelial-mesenchymal transition. J Cell Biol, 137:1403. 1997.

70.

Gupta

P.B.

, Kuperwasser

, Brunet

J.P.

, Ramaswamy

, Kuo

W.L.

, Gray

J.W.

, Naber

S.P.

, Weinberg

R.A.

The melanocyte differentiation program predisposes to metastasis after neoplastic transformation. Nat Genet, 37:1047. 2005.

71.

Lih

C.J.

, Cohen

S.N.

, Wang

, Lin-Chao

The platelet-derived growth factor alpha-receptor is encoded by a growth-arrest-specific (gas) gene. Proc Natl Acad Sci USA, 93:4617. 1996.

72.

Jones

A.V.

, Cross

N.C.

Oncogenic derivatives of platelet-derived growth factor receptors. Cell Mol Life Sci, 61:2912. 2004.

73.

Ingram

J.L.

, Bonner

J.C.

EGF and PDGF receptor tyrosine kinases as therapeutic targets for chronic lung diseases. Curr Mol Med, 6:409. 2006.

74.

Niida

, Kondo

, Hiratsuka

, Hayashi

, Amizuka

, Noda

, Ikeda

, Shibuya

VEGF receptor 1 signaling is essential for osteoclast development and bone marrow formation in colony-stimulating factor 1-deficient mice. Proc Natl Acad Sci USA, 102:14016. 2005.

75.

Chellaiah

, Kizer

, Silva

, Alvarez

, Kwiatkowski

, Hruska

K.A.

Gelsolin deficiency blocks podosome assembly and produces increased bone mass and strength. J Cell Biol, 148:665. 2000.

76.

Kusumi

, Nishi

, Tanaka

, Tsuchida

, Kudo

A murine osteosarcoma cell line with a potential to develop ossification upon transplantation. Jpn J Cancer Res, 92:649. 2001.

77.

Shukunami

, Shigeno

, Atsumi

, Ishizeki

, Suzuki

, Hiraki

Chondrogenic differentiation of clonal mouse embryonic cell line ATDC5 in vitro: differentiation-dependent gene expression of parathyroid hormone (PTH)/PTH-related peptide receptor. J Cell Biol, 133:457. 1996.

78.

zur Nieden

N.I.

, Kempka

, Rancourt

D.E.

, Ahr

H.J.

Induction of chondro-, osteo- and adipogenesis in embryonic stem cells by bone morphogenetic protein-2: effect of cofactors on differentiating lineages. BMC Dev Biol, 5:1. 2005.

79.

Tanihara

, Sano

, Heimark

R.L.

, St. John

, Suzuki

Cloning of five human cadherins clarifies characteristic features of cadherin extracellular domain and provides further evidence for two structurally different types of cadherin. Cell Adhes Commun, 2:15. 1994.

80.

Okazaki

, Takeshita

, Kawai

, Kikuno

, Tsujimura

, Kudo

, Amann

Molecular cloning and characterization of OB-cadherin, a new member of cadherin family expressed in osteoblasts. J Biol Chem, 269:12092. 1994.

81.

Togashi

, Abe

, Mizoguchi

, Takaoka

, Chisaka

, Takeichi

Cadherin regulates dendritic spine morphogenesis. Neuron, 35:77. 2002.

82.

Chow

D.C.

, Wenning

L.A.

, Miller

W.M.

, Papoutsakis

E.T.

Modeling pO(2) distributions in the bone marrow hematopoietic compartment. II. Modified Kroghian models. Biophys J, 81:685. 2001.

83.

Chow

D.C.

, Wenning

L.A.

, Miller

W.M.

, Papoutsakis

E.T.

Modeling pO(2) distributions in the bone marrow hematopoietic compartment. I. Krogh's model. Biophys J, 81:675. 2001.

84.

Heng

B.C.

, Cao

, Stanton

L.W.

, Robson

, Olsen

Strategies for directing the differentiation of stem cells into the osteogenic lineage in vitro. J Bone Miner Res, 19:1379. 2004.

85.

Wei

, Dai

X.-M.

, Stanley

E.R.

Transgenic expression of CSF-1 in CSF-1 receptor-expressing cells leads to macrophage activation, osteoporosis, and early death. J Leukoc Biol, 80:1445. 2006.

86.

Alexander

W.S.

, Rakar

, Robb

, Farley

, Willson

T.A.

, Zhang

J.G.

, Hartley

, Kikuchi

, Kojima

, Nomura

, Hasegawa

, Maeda

, Fabri

, Jachno

, Nash

, Metcalf

, Nicola

N.A.

, Hilton

D.J.

Suckling defect in mice lacking the soluble haemopoietin receptor NR6. Curr Biol, 9:605. 1999.

87.

Handschel

, Depprich

, Kubler

, Wiesmann

H.-P.

, Ommerborn

, Meyer

Prospects of micromass culture technology in tissue engineering. Head Face Med, 3:4. 2007.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

1.28 MB

0.00 MB