Stem Cells
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

First published online February 28, 2008
Stem Cells Vol. 26 No. 5 May 2008, pp. 1186 -1201
doi:10.1634/stemcells.2007-0821; www.StemCells.com
© 2008 AlphaMed Press

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
2007-0821v1
26/5/1186    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Reprints/Permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Huang, T.-S.
Right arrow Articles by Wang, H.-W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Huang, T.-S.
Right arrow Articles by Wang, H.-W.

STEM CELL GENOMICS AND PROTEOMICS

Functional Network Reconstruction Reveals Somatic Stemness Genetic Maps and Dedifferentiation-Like Transcriptome Reprogramming Induced by GATA2

Tse-Shun Huanga, Jui-Yu Hsieha, Yu-Hsuan Wua, Chih-Hung Jenb, Yang-Hwei Tsuangc, Shih-Hwa Chioud,e, Jukka Partanenf, Heidi Andersonf, Taina Jaatinenf, Yau-Hua Yue,g, Hsei-Wei Wanga,b,h

Institutes of aMicrobiology and Immunology,
gOral Biology, and
dClinical Medicine and
bVeteran General Hospital-Yang Ming Genome Center, National Yang-Ming University, Taipei, Taiwan;
eDepartment of Medical Research and Education, Taipei Veterans General Hospital, Taipei, Taiwan;
fFinnish Red Cross Blood Service, Helsinki, Finland;
Departments of cOrthopedics and
hEducation and Research, Taipei City Hospital, Taipei, Taiwan

Key Words. CD133+ stem cell • Somatic stem cell • Systems biology • GATA2 • Dedifferentiation

Correspondence: Correspondence: Hsei-Wei Wang, Ph.D., Institute of Microbiology and Immunology, National Yang-Ming University, No. 155, Sec 2, Li-Nong Street, Taipei 112, Taiwan. Telephone: 886-2-2826-7109; Fax: 886-2-2821-2880; e-mail: hwwang{at}ym.edu.tw

Received on October 3, 2007; accepted for publication on February 14, 2008.

First published online in STEM CELLS EXPRESS  February 28, 2008.

   ABSTRACT
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
 Acknowledgments
 References
 
Somatic stem cell transplantation holds great promise in regenerative medicine. The best-characterized adult stem cells are mesenchymal stem cells (MSCs), neural stem cells (NSCs), and CD133+ hematopoietic stem cells (HSCs). The applications of HSCs are hampered since these cells are difficult to maintain in an undifferentiated state in vitro. Understanding genes responsible for stem cell properties and their interactions will help on this issue. The construction of stem cell genetic networks will also help to develop rational strategies to revert somatic cells back to a stem-like state. We performed a systemic study on human CD133+ HSCs, NSCs, MSCs, and embryonic stem cells and two different progenies of CD133+ HSCs, microvascular endothelial cells (MVECs) and peripheral blood mononuclear cells. Genes abundant in each or in all three somatic stem cells were identified. We also observed complex genetic networks functioning in postnatal stem cells, in which several genes, such as PTPN11 and DHFR, acted as hubs to maintain the stability and connectivity of the whole genetic network. Eighty-seven HSC genes, including ANGPT1 and GATA2, were independently identified by comparing CD34+CD33–CD38– hematopoietic stem cells with CD34+ precursors and various matured progenies. Introducing GATA2 into MVECs resulted in dedifferentiation-like transcriptome reprogramming, with HSC genes (such as ANGPT1) being up and endothelial genes (such as EPHB2) being down. This study provides a foundation for a more detailed understanding of human somatic stem cells. Expressing the newly discovered stem cell genes in matured cells might lead to a global reversion of somatic transcriptome to a stem-like status.

Disclosure of potential conflicts of interest is found at the end of this article.


   INTRODUCTION
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
 Acknowledgments
 References
 
Adult stem cells are required for a lifelong sustenance of matured cell replacement and hold great promise for future therapeutic applications. In vitro as well as in vivo studies have established that hematopoietic and endothelial cells develop from a common postnatal progenitor, the hemangioblast [1, 2]. The hematopoietic stem cell (HSC) is currently been used in clinical stem cell transplantation for the treatment of leukemia [3]. It is also used in the treatment of many nonhematological disorders, such as autoimmune diseases and metabolism disorders [3]. However, the application of HSCs is hindered by their expandability, whereas cell dose is a major determinant of survival after HSC transplantation [3]. Further characterization of hemangioblasts will be critical for a better understanding of the molecular events involved in stem cell properties, as well as for using this cell population for clinical applications.

Currently, hemangioblast or HSC isolation is performed by recognizing the CD133 antigen [4, 5]. Until recent years, the gene expression pattern in human CD133+ HSC was characterized by microarray analyses in several studies, whereby genes involved in self-renewal, differentiation, and lineage choice were revealed [5–7]. These expression analyses helped to uncover genetic programs accompanying the cascade in hematopoiesis or angiogenesis. Transcription factors have drawn much attention since they very often play key roles in stemness maintenance or fate determination during differentiation. For example, GATA1 plays an essential role in the promotion of hematopoietic cell differentiation [8]. GATA2, on the other hand, strongly blocks hematopoietic cell differentiation and stimulates immature cell proliferation [9].

Mesenchymal stem cells (MSCs), found in many adult tissues, are also attractive somatic stem cell sources for the regeneration of damaged tissues. Currently, MSCs can be isolated from various human sources, including bone marrow, umbilical cord, cord blood, adipose tissue, and muscle [10–14]. MSCs from different sources are all able to self-renew with a high proliferative capacity, and all possess a mesodermal differentiation potential, including osteogenic, chondrogenic, and adipogenic differentiation [15]. The gene expression profiles of MSCs of different sources have been widely characterized by microarray analyses [16, 17]. A very recent report compared the gene expression pattern of cord blood CD133+ HSCs with that of bone marrow-derived MSCs in hypoxia, and genes commonly expressed in both stem cells have been highlighted [7]. More functional studies are still needed to characterize the roles of identified MSC and HSC genes under different physiological or culture conditions.

The genetic signature of another somatic stem cell type, neural stem cell (NSC), has also been studied. In mouse, a portion of the genetic program of hematopoietic stem cells is shared with embryonic stem cells (ESCs) and NSCs. These common gene products (283 genes) represent a molecular signature of stem cells [18, 19]. ESCs and NSCs are largely similar at the transcriptional level [19]. Gene expression profiles of human NSC during temporal changes of priming and differentiation, as well as neuronal precursors from human brain, have also been well studied [20, 21]. However, no study has yet addressed a systematic comparison of human ESCs and somatic stem cells. Gene signatures provided by such a study will provide a foundation for a profound understanding of human stem cell biology.

Although gene expression profiling can reveal differentially expressed genes, the challenge remains to assign the biological significance of these genes into a complete biological system. There is increasing recognition that genes do not act as individuals but collaborate as a module, where cellular functions are carried out by many modules in overlapping networks [22, 23]. Disrupted signaling crosstalk among biological modules can be a hallmark of cancer [24]. With the accumulation of gene functional annotations and molecular interaction data, dynamic mapping of gene expression data to a particular pathway or a genetic network is possible. Comprehension of the value of such knowledge-based analysis in stem cells has been partly supported by the finding that stem cells and cancer cells share many features and pathways [25]. Functional network analysis has been applied to analyze the transcriptomes of embryonic stem cells and bone marrow MSCs [17, 26]. No similar research using systems biology tools has been conducted on CD133+ HSCs or NSCs. Moreover, the genetic network revealed from previous studies is incomplete because of the limitation of human-curated knowledge-based databases [17, 27]. The global genetic networks among stemness genes still need to be constructed via a data-driven approach.

In this study, we applied gene expression microarray and systems biology tools to obtain genes involved in stemness in cord blood CD133+ HSCs, NSCs, and bone marrow MSCs and to provide a global genetic network for each somatic stem cell type. Genes common in all three somatic cells, as well as those also abundant in ESCs, were revealed. Novel GATA2-regulated genes were identified by further microarray experiment. Dedifferentiation-like transcriptome reprogramming was observed in GATA2-expressing endothelial cells, indicating the critical role of GATA2 in stem cell properties. Our data contribute new insights to a refined molecular picture and a better understanding of somatic stem cells. Manipulating the steady expression of stem cell genes may eventually facilitate the clinical use of CD133+ HSCs, NSCs, and MSCs. Introducing critical stemness factors into matured cells to convert their genetic networks to a stem cell state may eventually produce stem-like cells directly from somatic sources.


   MATERIALS AND METHODS
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 Conclusion
 Disclosure of Potential...
 Acknowledgments
 References
 
CD133+ Stem Cells, MSCs, NSCs, and Primary Microvascular Endothelial Cells
Human CD133+ stem cells of healthy individuals were isolated from umbilical cord blood as published before [5]. Bone marrow MSCs and NSCs were isolated from healthy individuals (Poietics, Lonza Inc., Conshohocken, PA, http://www.lonzabioscience.com/Lonza_CatNav.asp?oid=867). After isolation, CD133+ HSCs were subjected directly to RNA extraction. MSCs were cultured in MesenCult medium (StemCell Technologies, http://www.stemcell.com) for fewer than five passages. NSCs were cultured in the Neural Progenitor Maintenance BulletKit (Poietics), and differentiation was induced by culturing them in the Neural Progenitor Differentiation BulletKit (CC-3229, Poietics). Human primary microvascular endothelial cells (MVECs; Clonetics, Lonza Inc.) and a human endothelial cell line, HMEC1, were cultured in EGM-2 MV BulletKit medium (Clonetics).

Array Probe Preparation, Data Analysis, Group Distance Calculation, and Function Network Analyses
Total RNA collection, cRNA probe preparation, array hybridization, and data analysis were done as previously described [28–30]. More details are given in the supplemental online Materials and Methods. The average linkage distance was used to assess similarity between two groups of gene expression profiles, as described previously [28]. The difference in distance between two groups of sample expression profiles and a third group was assessed via the comparison of corresponding average linkage distances (the mean of all pairwise distances [linkages] between members of the two groups concerned). The error on such a comparison was estimated by combining the SEs (the SD of pairwise linkages divided by the square root of the number of linkages) of the average linkage distances involved [28]. Classic multidimensional scaling (MDS) was performed using the standard function of the R program to provide a visual impression of how the various sample groups are related. Principal component analysis (PCA), a technique similar to MDS, was performed using Partek Genomics Suite software (Partek, Inc., St. Louis, http://www.partek.com).

Gene annotation was performed by the ArrayFusion Web tool (http://microarray.ym.edu.tw/tools/arrayfusion/) [31]. Differential gene expression profiles were imported into the Ingenuity Pathways Analysis (IPA) software (Ingenuity Systems, Redwood City, CA, http://www.ingenuity.com) to obtain functional regulatory networks. The knowledge base behind IPA was built upon scientific evidence, manually curated from thousands of journal articles, textbooks, and other data sources. After a list of signature genes was uploaded, interaction among focus genes and interaction among interacting genes and molecules from the knowledge base were used to combine genes into networks according to their probability of having more focus genes than expected by chance. Networks are scored on the basis of the number of uploaded signature genes they contain. The network score is based on the hypergeometric distribution and is calculated with the right-tailed Fisher's exact test. The score is the negative log of this p value. The higher the score, the lower the probability of finding the observed number of uploaded signature genes in a given network by random chance.

Microarray Expression Data Sets
Affymetrix U133 Plus 2.0 microarray data of human ESCs were from the Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/geo/; GSE7896 [NCBI GEO] and GSE6561 [NCBI GEO] ) (Affymetrix, Santa Clara, CA, http://www.affymetrix.com). To extract HSC-enriched genes, we used 57 Affymetrix U133A array data produced by us [28] or from publicly accessible array databases, including GEO, ArrayExpress (http://www.ebi.ac.uk/arrayexpress), the Medical University of South Carolina (MUSC) DNA Microarray Database (http://proteogenomics.musc.edu/pss/home.php), and The Genomics Institute of the Novartis Research Foundation (GNF) SymAtlas (http://symatlas.gnf.org/SymAtlas/) (Fig. 4A). The in-house data sets comprised samples from human umbilical cord vein endothelial cell (HUVEC), aortic or uterine smooth muscle cell, lymphatic endothelial cell, and blood vessel endothelial cell (ArrayExpress E-MEXP-66). The downloaded data sets comprised CD34+ HSCs (CD34+CD38–CD33–KIT+Rhohigh and CD34+CD38–CD33–Rholow [GSM51391 to GSM51408 [NCBI GEO] from GEO GSE2666 [NCBI GEO] ]), CD34+ precursor cells (including pre-B cells, pro-B cells [ArrayExpress E-MEXP-384], CD34+CD33+ myeloid cells, and CD34+CD71+ early erythroid cells [from GNF SymAtlas]), and a variety of matured progeny cells (including human endothelial cells of artery and vein origin [HUVEC, MUSC data set 010504, 040204, 070102, GEO GSE973 [NCBI GEO] ] and hematopoietic cells [GEO GSE1140 [NCBI GEO] ]). For Figure 4B and supplemental online Figure 1A and 1B, a total of 256 Affymetrix U133A array data points for normal human tissues were an extension of the 57-array data set and were collected from GNF SymAtlas and GEO (GSE1140 [NCBI GEO] , GSE2248 [NCBI GEO] , GSE2361 [NCBI GEO] , and GSE2666 [NCBI GEO] ). All of the array data are available on our web site (http://infobio.ym.edu.tw/).

Promoter Analysis
Transcription factor binding sites in a given promoter region (base pairs –3,000 to +1,000) were analyzed by the Patch program (http://www.gene-regulation.com/pub/programs.html#patch) from the BioBase Biological Databases (Beverly, MA) and the MatInspector program from Genomatix (Munich, Germany, http://www.genomatix.de/). BioTapestry software (http://www.biotapestry.org) was used to represent the relationships between stemness genes.

Plasmid Construction and Lentivirus Transduction
Lentivirus production and infection were performed as described [32, 33]. Plasmid plenti4-GATA2, which was used to express FLAG-tagged GATA2, was constructed by polymerase chain reaction (PCR) using the follow primer pair: 5'-VGAAcggtccgCTGCACCCAGACCCTGAG (one CpoI site lowercased) and 3'-VGAActcgagGTCCTCGACGTCCATCTGTT (one XhoI site lowercased). The PCR products were TA cloned into the pGEMT-easy vector (Promega, Madison, WI, http://www.promega.com), sequence-verified, and then cloned in-frame into the FLAG-tagged pLenti4 lentiviral vector (a gift kindly provided by Dr. Su-Fang Lin, National Health Research Institute, Znunan, Taiwan) by cutting at the CpoI and XhoI sites.


   RESULTS
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 Conclusion
 Disclosure of Potential...
 Acknowledgments
 References
 
Isolation and Characterization of Somatic Stem Cells and Primary Cells
To access molecular signature genes for somatic stem cells, a set of primary human somatic stem or matured cells, including cord blood CD133+ HSCs, NSCs, MSCs derived from bone marrow, dermal MVECs, and peripheral blood mononuclear cells (PBMCs), were collected. Of all the matured cell types, only PBMCs and MVECs were collected, since we focused more on CD133+ HSCs, the ancestor of these two somatic cell types. MVECs were positive for endothelial cell (EC) markers CD31 and von Willebrand factor (Fig. 1A). MSCs were positive for CD44 and CD73 (both as mesenchymal markers) and negative for CD34 and CD45 (both as hematopoietic markers) (Fig. 1B). Isolated MSCs could also differentiate into cells of osteogenic, adipogenic, and chondrogenic lineages (not shown). NSCs could be maintained in an undifferentiated state as neurospheres (Fig. 1C, upper panel) and were positive for nestin, a NSC marker (not shown). Cultured NSCs could be induced into neurogenic and glial lineages, since differentiated cells expressed the neuron marker β-tubulin III or the glial cell marker glial fibrillary acidic protein (an intermediate filament protein that is found in glial cells, such as astrocytes) (Fig. 1C, lower panel, green and red, respectively).


Figure 1
View larger version (47K):
[in this window]
[in a new window]

  		Figure 1. Gene expression microarray analysis of three somatic stem cells, PBMCs, and primary MVECs. (A): Characteristics of isolated MVECs. All MVECs (passage 6) express CD31 and vWF. (B): Immunophenotype of MSCs by flow cytometric analysis. Representative histograms are demonstrated, and their respective isotype controls are shown by filled blue areas. MSCs were positive for CD44 and CD73 and negative for CD34 and CD45. (C): Characteristics of NSCs. Top panel, undifferentiated NSCs formed neurospheres; middle and lower panels, NSCs were induced into differentiation for 3 and 12 d, respectively. At d 12 postinduction, differentiated cells were stained for β-tubulin III (a neuron marker; green), glial fibrillary acidic protein (for glial cells such as astrocytes; red) and nuclear DNA (Hoechst 33258; blue).(D): A heat map shows genes enriched in CD133+ stem cells, in MSCs, in NSCs, or in MVECs. Genes in red, increased expression; in blue, decreased. (E): An MDS plot shows the discrimination ability of the obtained molecular signatures of cell groups. Each spot represents a single array sample. Each cell group exhibited a significantly distinct global gene expression profile. (F): Venn diagram detailing shared and distinct gene expression among human HSCs, NSCs, and MSCs. (G): Validation of CD133+ HSC genes by real-time reverse transcription-polymerase chain reaction. Mean expression levels of target genes were compared with that of GAPDH control. Results are expressed as the mean ± SD. Abbreviations: d, days; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; HSC, hematopoietic stem cell; MDS, multidimensional scaling; MSC, mesenchymal stem cell; MVEC, microvascular endothelial cell; NSC, neural stem cell; PBMC, peripheral blood mononuclear cell; vWF, von Willebrand factor.

 
The gene expression profiles of all collected cell types were implemented at least in triplicate by using the Affymetrix HG-U133 Plus 2.0 whole-genome chip. Genes differentially expressed between cell types (the molecular signatures) were identified according to a statistical pipeline we used [28]. A gene expression heat map for these genes indicated their unique expression patterns among each cell type (Fig. 1D), and their discrimination ability was also assessed by MDS (Fig. 1E). Compared with PBMCs and MVECs, 1,572 probe sets were abundantly overexpressed in CD133+ cells (with a positive false discovery rate [pFDR] threshold of q < 0.0001), 1,456 probe sets were overexpressed in MSCs (q < 0.001), and 3,252 probe sets were overexpressed in NSCs (q < 0.001) (Fig. 1F). Abundant expression of some genes (including GATA2, MCM3, MLF1IP, MYB, KIT, and FLT3) in CD133+ cells was verified by quantitative PCR (qPCR) (Fig. 1G). qPCR results showed a high degree of correlation with microarray results (R > 0.95).

Molecular Signatures of Somatic Stem Cells
The top 100 genes most strongly expressed in CD133+ HSC and the top 60 genes in MSC and NSCs are listed in Tables 1GoGo–3. In HSCs, PROM1 (CD133) and CD34, two hematopoietic and endothelial precursor markers [4, 34], were among those 100 genes (Table 1Go, underlined). BAALC, another novel marker of hematopoietic progenitor cells, also appeared [35] (Table 1Go, underlined). Also shown in Table 1Go are FLT3, KIT, HLF, ITGA9, LMO2, MLLT3, and MYB, which have all been shown to be enriched in pluripotent hematopoietic stem cells [5, 6, 9, 36]. These consistent findings demonstrate the reliability of our gene list. In MSCs, SOX9, SNAI2 (Slug), and TWIST1, which regulate epithelial-mesenchymal transition and cell migration in early neural crest development or in tumor metastasis [37, 38], were present (Table 2, underlined). In NSCs, known neuronal precursor markers (such as PTN and FOXG1) were present (Table 3, underlined). PTN mRNA is highly expressed in neural stem cells of mouse ventral mesencephalon, and PTN can promote the production of DAergic neurons from embryonic stem (ES) cell-derived nestin-positive cells [39]. FOXG1 is known to constitutively suppress the generation of the Cajal-Retzius cell, the earliest-born neuron [40]. Details for these genes and all other genes can be found in supplemental online Table 1.


View this table:
[in this window]
[in a new window]

  		Table 1. Top 100 genes in CD133+ HSC

 


View this table:
[in this window]
[in a new window]

  		Table 1. (continued)

 


View this table:
[in this window]
[in a new window]

  		Table 2. Top 60 genes in MSC

 


View this table:
[in this window]
[in a new window]

  		Table 3. Top 60 genes in NSC

 
One hundred eighty-three probe sets were commonly expressed in all three somatic stem cells, indicating their critical roles in stem cell properties (Fig. 1F; Table 4Go). We also checked which of those 104 genes were also abundant in human ESCs. A gene expression heat map for genes differentially expressed among all stem cells (three somatic stem cells and ESCs) and matured cells (MVEC and PBMC) indicated 34 genes abundantly expressed in all stem cells (q < 10–6; Fig. 2A). PCA plots proving their discrimination ability are also shown (Fig. 2B). Details of those 34 genes are given in Table 4Go (asterisks). Pre-B-cell leukemia transcription factor 1 (PBX1), a homeodomain transcription factor that was originally identified as the product of a proto-oncogene in acute pre-B-cell leukemia, is a global regulator of embryonic and B cell development [41, 42]. The role of PBX1 in HSCs is also evidenced by the reduced numbers and impaired functions of committed hematopoietic progenitors in the fetal liver that result in inadequate maintenance of definitive hematopoiesis and severe anemia [43]. EPDR1, also known as MERP1, is expressed in a hematopoietic stem cell-enriched population but is downregulated with proliferation and differentiation [44]. The presence of those known stem cell genes strengthens the reliability of our gene list. Most of the 34 stem cell genes, as well as most of the somatic stem cell genes, are novel and worthy of being investigated further.


View this table:
[in this window]
[in a new window]

  		Table 4. Genes highly expressed in all 3 somatic stem cells

 


View this table:
[in this window]
[in a new window]

  		Table 4. (continued)

 


Figure 2
View larger version (72K):
[in this window]
[in a new window]

  		Figure 2. Supervised hierarchical analysis of human somatic and embryonic stem cells. (A): A heat map shows genes differentially expressed between stem cells and matured cells. Genes in red, increased expression; blue, decreased expression. (B): Principal component analysis using global transcriptome (upper panel) or genes in (A) (lower panel). Each spot represents a single array sample. x-, y-, and z-axes represent three major PCs. Stem cells formed one cluster, and matured cells (PBMCs and two different types of endothelial cells) formed another. Abbreviations: BEC, blood vessel endothelial cell; ESC, embryonic stem cell; HSC, hematopoietic stem cell; LEC, lymphatic endothelial cell; MSC, mesenchymal stem cell; NSC, neural stem cell; PBMC, peripheral blood mononuclear cell; PC, principal component.

 
The relative expression levels of these common stemness genes across the whole human body were examined, and most of them (such as ANGPT1 and NPR3) were unique in stem cells (supplemental online Fig. 1A, 1B). One of these genes is angiopoietin-1 (ANGPT1; Tables 1Go, 2, underlined). The role of ANGPT1 secreted by CD133+ HSCs and MSCs in angiogenesis is supported by the finding that defective vascular remodeling in RUNX1 homozygous mutant mice could be rescued by addition of HSCs or ANGPT1 [45]. ANGPT1 was abundant in several somatic stem cell types, including MSCs and long-term self-renewing (LT) and short-term self-renewing (ST) HSCs (supplemental online Fig. 1A), and its high expression level in another somatic stem cell type, limbus stem cells, was revealed by qPCR (supplemental online Fig. 1C). ESCs, on the other hand, do not express high levels of ANGPT1 (supplemental online Fig. 1A).

Genetic Networks of CD133+ Stem Cells, MSCs, and NSCs
Increasing evidence shows that genes do not act as individuals but collaborate in genetic networks [24]. To better understand how genes enriched in somatic stem cells are related to each other, we performed genetic network analysis for signature genes. Signature probe sets were input in the IPA software to construct network modules. The knowledge base behind IPA summarizes known molecular interactions evidenced in published literature (described in Materials and Methods). The term "network" in IPA is not the same as a biological or canonical pathway with a distinct function (i.e., angiogenesis) but a reflection of all interactions of a given protein as defined in the literature.

In CD133+ stem cells a major network consisting of 215 genes was identified (Fig. 3A). This network included most of the known stemness-related or pro-proliferating genes. Among those genes, STAT5A/B and ESR1 are novel markers for CD133+ HSCs, and HOXA9, GATA2, KIT, MPL, MYB, and MYCN can support self-renewal and keep HSCs in undifferentiated status [9, 36]. CDK6 is a factor promoting G1 phase of HSC [46]. CD34 is a well-known marker for hematopoietic precursors, and ANGPT1 is crucial for angiogenesis [45]. Besides reproducing what have already known for CD133+ cells, novel genes that may play crucial roles in self-renewal and differentiation were also revealed. Genes without previous implication in hematopoiesis or endothelial differentiation but with evidence in the development of other organs, including NCL, THRB, NR1I2 (PXR), and TRH [47, 48], were also in this main network (supplemental online Table 2).


Figure 3
View larger version (46K):
[in this window]
[in a new window]

  		Figure 3. Interaction network analysis as a framework for the interpretation of stem cell biology. (A): A functional genetic network composed of multiple CD133+ hematopoietic stem cell (HSC) genes. This network is displayed graphically as nodes (gene products) and edges (biological relationships between nodes) mapped by the Ingenuity Pathway Analysis tool. The intensity of the node color indicates the degree of upregulation. Nodes are displayed using various shapes that represent the functional class of the gene product (right panel). (B): Selected core regions of the HSC gene network, highlighting several hub genes. Their corresponding locations in the genetic network are indicated by numbers. Key hub genes are labeled in yellow. (C): Locations of stem cell-common genes in the CD133+ genetic network. Genes commonly expressed in CD133+ cells, neural stem cells, and mesenchymal stem cells (Fig. 2; Table 4Go) are shown in red in the right panels.

 
This network also revealed genes with significant biological roles in CD133+ cells. Some genes, regarded as "hubs," had higher connectivity to others or resided in a position among submodules in the major network (Fig. 3B). Dysregulation of hubs may eventually lead to the disruption of the genetic network and the malfunction of cells [24]. NFE2L2, GNAQ, and MYCN were hubs connecting different submodules in the major network component (Fig. 3B, 3B2). Central to the network, there were significant hubs, including GATA2, HOXA9, SPP1, ESR1, IL1B, KIT, MPL, MYB, IGF1R, PTPN11, and STAT5A/B (Fig. 3B). Of these hub genes, PTPN11, TIMP3, and DHFR were commonly expressed in CD133+ cells, NSCs, and MSCs (Table 4Go; Fig. 3C). We got similar results—DHFR, TIMP3, and especially PTPN11 with higher connectivity—when conducting functional network analysis on MSC genes (supplemental online Fig. 2A). When a similar function network analysis was conducted on NSC-enriched genes, DHFR was once again a hub (supplemental online Fig. 2B), suggesting its critical role in maintaining the integrity of genetic networks and in stem cell properties.

Universal Genes of CD133+ HSCs
To further compare our gene signature to published ones, we collected a set of array data for HSCs (both LT-HSCs and ST-HSCs), CD34+ precursors, and terminal differentiated progeny cells of HSCs. These array data (57 arrays in total) were implemented using the Affymetrix U133A chip, which contains ~22,200 probe sets. By comparing the gene expression profiles of HSCs and precursors/matured cells, we acquired another list of stem cell signature genes (q < 10–3). This list was then compared with the one obtained from the U133 Plus 2.0 chip analysis in Figure 1. As a result, a total of 87 genes were consistent in both signatures (Fig. 4A). These genes therefore represent the most likely stemness genes in CD133+ HSC population. Several of them have been proved or suggested before, including ANGPT1, GATA2, HLF, ITGA9, KIT, NPR3, and PROM1 (CD133) (Fig. 4A) [5, 49, 50].


Figure 4
View larger version (48K):
[in this window]
[in a new window]

  		Figure 4. Narrowing down HSC genes by comparing CD34+ HSCs with CD34+ precursors and a variety of matured progeny cells of HSCs. (A): A heat map showing 87 stemness genes enriched in CD34+ HSCs. U133A array data for each cell type were collected from publicly accessible databases (described in Materials and Methods). This collection comprised 18 arrays for CD34+ HSCs (both CD34+CD38–CD33–KIT+Rhohigh and CD34+CD38–CD33–Rholow HSCs), 8 arrays for CD34+ precursor cells (including pre-B cells, pro-B cells, CD34+CD33+ myeloid cells, and CD34+CD71+ early erythroid cells), and 31 arrays for matured progeny cells (including endothelial cells of artery, vein, microvascular blood vessel, or lymphatic vessel origin; hematopoietic cells; and smooth muscle cells). Genes underlined are transcription factors. Genes in red, increased expression; green, decreased expression. (B): GATA2 gene expression distribution among various types of normal human tissues. A total of 256 Affymetrix U133A array data were normalized together (described in Materials and Methods). Abbreviations: BEC, blood vessel endothelial cell; ESC, embryonic stem cell; GI, gastrointestinal; HAEC, human artery endothelial cell; HBEC, human bronchial epithelial cell; HSC, hematopoietic stem cell; HUVEC, human umbilical vein endothelial cell; LEC, lymphatic endothelial cell; LT, long-term self-renewing; MSC, mesenchymal stem cell; PBMC, peripheral blood mononuclear cell; ST, short-term self-renewing.

 
The relative expression levels of those 87 genes across the whole human body were examined by checking the genes' relative hybridization signals in another 256-chip microarray data set. The abundant expression of GATA2 in both LT- and ST-HSCs was observed (Fig. 4B), indicating its essential role in HSCs.

Dedifferentiation-Like Transcriptome Reprogramming Induced by GATA2 in Human Endothelial Cells
Using the knowledge-based strategy that we applied, only a small fraction (215/1,572 = 13.7%) of CD133+ cell-enriched genes were involved in network formation (Fig. 3A). This may be due to the fact that only few molecular processes of stem cells had been unveiled. To improve our understanding of stem cell biology, we next aim to find out novel interactions between signature genes of CD133+ HSC. Since GATA2 is uniquely expressed in CD133+ cells (Fig. 4) and is essential for the proliferation/survival of early hematopoietic cells [51, 52], we explored the downstream target genes for this critical transcription factor.

GATA2 was introduced into the endothelial cell line HMEC1 or primary MVECs by lentivirus transduction, and gene expression microarray analysis was performed to examine the effects of GATA2 overexpression. The expression of 3,013 probe sets was affected by GATA2 in HMEC1 (with a pFDR threshold q < 0.2). The gene expression profile of GATA2-transduced ECs was then compared with that of vector-transduced cells. We found that GATA2 overexpression resulted in the upregulation of CD133+ genes (such as ANGPT1 and NPR3) but the downregulation of EC genes (such as EPHB2, a key receptor involved in angiogenesis [53]) (Fig. 5B, 5C, respectively). This scenario is similar to that of somatic cell dedifferentiation or EC dysfunction [54–56]. Among those 45 GATA2-regulated HSC genes, 13 genes (28.9%) (ANGPT1, BBX, DPY19L3, GSTM1, HELLS, MEIS1, NRIP1, OBSL1, PDE1A, PHYH, SMAD5, SPAG9, and TPR) were present in all three somatic stem cells, suggesting an upstream and critical role of GATA2 in stem cells.


Figure 5
View larger version (36K):
[in this window]
[in a new window]

  		Figure 5. Constitutive expression of GATA2 in matured endothelial cells induced dedifferentiation-like transcriptome reprogramming. (A): Expression of GATA2 protein in HMEC1. FLAG-tagged GATA2 was transduced into a human endothelial cell line HMEC1 by lentivirus infection at a multiplicity of infection of 100, four times. Two days after the fourth infection, cell lysates were collected and GATA2 protein was detected by western blotting with an anti-FLAG antibody. A 56-kDa GATA2 band was observed (indicated by an arrow). (B): A heat map showing the upregulated CD133+ genes by GATA2. The gene expression profiles between GATA2- or vector-transduced HMEC1 were compared to reveal the impact of GATA2 on endothelial cell (EC) transcriptome. (C): A heat map showing the downregulated EC genes by GATA2. Again, the gene expression profiles between GATA2- and vector-transduced HMEC1 were compared. (D): Transcriptome distance analysis for CD133+ hematopoietic stem cells (HSCs), GATA2-transduced ECs, and empty lentivirus-transduced ECs. Average linkage distances between transcriptomes were calculated as described [70] using 4,454 probe sets distinguishing CD133+ HSCs and HMEC1 (q < 10–4). GATA2-overexpressing cells exhibited a gene expression pattern closer to that of CD133+ stem cells than was the profile of vector-transduced ones. (E): Initial gene regulatory network relations for GATA2 and its downstream targets. All these genes contain at least 1 GATA motif in their promoter regions (base pairs –3,000 to +1,000). Genes in yellow, transcription factors or cofactors; blue, signaling-related genes according to Gene Ontology database; green, signaling-related genes (also acting as transcription factors or cofactors); in black, genes with other functions. Genes commonly expressed in CD133+ HSC, neural stem cells, and mesenchymal stem cells (according to Table 4Go) are indicated by asterisks. Abbreviation: EC, endothelial cell.

 
To provide more quantitative evidence, we calculated the average linkage distances between CD133+ and GATA2-transduced ECs and between CD133+ and vector-transduced ECs. We used an average linkage distance analysis to assess the similarity between two groups of gene expression profiles, as described previously [28]. As shown in Figure 5D, the distance between CD133+ HSCs and GATA2-transduced ECs was smaller than that between HSCs and vector-transduced ECs. The genetic profile of GATA2-expressing ECs was therefore closer to that of CD133+ cells, supporting a dedifferentiation-like transcriptome reprogramming induced by GATA2 overexpression.

We also checked which GATA2-regulated CD133+ genes might be the direct targets of GATA2. Promoter regions of GATA2-regulated genes were examined for putative GATA-binding sites, and such motifs could be found in the promoter regions of 30 genes (Fig. 5E). One of them is ANGPT1, a cytokine commonly expressed in CD133+ HSCs, NSCs, and MSCs (Table 4Go; supplemental online Fig. 1). There were quite a few putative GATA-binding motifs in the promoter region of ANGPT1 (supplemental online Fig. 3A). Semiquantitative PCR confirmed that the ANGPT1 level was increased after GATA2 overexpression (supplemental online Fig. 3B). The relationships between GATA2 and its regulating genes, including ANGPT1, are summarized in Figure 5E.


   DISCUSSION
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
Conclusion
 Disclosure of Potential...
 Acknowledgments
 References
 
Stem cells have recently drawn immense research interest because of their unique biological behaviors and potential clinical uses. An improved understanding of HSCs can help the ex vivo expansion of them or the in vivo control of their differentiation directions, thereby furthering their potential in therapeutic applications. In this study we encompassed an extensive comparative transcriptome and gene network analysis of HSCs, NSCs, MSCs, ESCs, and the progeny populations of HSCs. Moreover, we developed methods in systems biology to reveal the interactions between signature genes. Genetic networks of stem cells provide in-depth information (such as key hub genes and novel GATA2 targets) that could not be readily extracted from one-dimensional gene list analysis. Novel relationships between GATA2 and other stemness genes were also mapped and confirmed in this study (Fig. 5; supplemental online Fig. 3A, 3B). For the first time, we found that somatic stem cell genes, like some embryonic stemness genes, also hold the potential of dedifferentiate matured cells: GATA2 overexpression induced a dedifferentiation-like transcriptome reprogramming in endothelial cells. These findings might pave the way to unraveling the myth of somatic stem cells, especially HSC, and, furthermore, to contributing to cell-based therapy.

We narrowed down our stem cell genes by comparing our gene list with those of previous work. We did a meta-analysis on several public microarray data sets, which comprised CD34+CD38–CD33– HSCs (both Rhohigh LT-HSCs and Rholow ST-HSCs), CD34+ precursor cells (including pre-B cells, pro-B cells, CD34+CD33+ myeloid cells, and CD34+CD71+ early erythroid cells) and a variety of matured progeny cells of HSCs (including endothelial cells of artery, vein, microvascular blood vessel, or lymphatic vessel; hematopoietic cells; and smooth muscle cells). For the identification of HSC-enriched genes, this collection was very comprehensive. By overlapping HSC signatures from two different sources (i.e., from our own Affymetrix U133 Plus 2.0 data set and the publicly accessible U133A array collection), a total of 87 genes were then disclosed. Among those genes, there were known hemangioblast markers (ANGPT1, CRHBP, GATA2, HLF, KIT, MEIS1, NPR3, and PROM1 [5–7]), supporting results generated using proposed strategy. These 87 genes should serve as candidate targets for future research.

The gene expression profile of human CD133+ cells has been reported by a few groups [5–7]. Nevertheless, finding significant targets and related genes is always a daunting task. Analyzing gene signatures by dividing them into functional subgroups (e.g., by gene set enrichment analysis [57]) or network modules is an efficient way to provide more insights into gene lists [57]. A systemic approach is mandatory to view the overall molecular events as a biological system for a given biological process, where we can find important genes as controllers [22, 23]. Those key genes very often serve as hubs to maintain the stability of a genetic module or to connect modules within a major network. By applying systems biology tools we identified a major functional network in CD133+, cells as well as in two other somatic stem cell types, and hub genes such as DHFR were identified via this systemic approach. Several CD133+ hub genes, such as PTPN11 (protein tyrosine phosphatase, nonreceptor type 11) and DHFR (dihydrofolate reductase), were also hubs in MSCs (Fig. 3D). PTPN11, also known as Shp-2, is required for embryonic development, as mice homozygous for the mutant allele die in utero at midgestation [58]. Mice mutant for EGFR and PTPN11 have defective cardiac semilunar valvulogenesis [59], supporting the hypothesis that PTPN11/Shp-2 is required for EGFR signaling in vivo. In our array analysis, EGFR was abundantly expressed in both MSC and NSC (supplemental online Table 1). Of note, in the network maps (Fig. 3; supplemental online Fig. 2), the number of common stem cell genes varies in each gene network. This is due to the fact that we applied a knowledge-based strategy to construct genetic networks, and only genes with known interactions will be included.

An important task in systems biology is to identify novel interactions and networks for signature genes. Using a knowledge-based strategy, only 13.7% of CD133+ genes were involved in networks (Fig. 3A). Although several crucial hub genes were mapped, many more key genes, such as GATA2, are still waited to be unveiled. We extended this network core through the assistance of further microarray array experiment, promoter analysis, and wetlab confirmation. By overexpressing GATA2 in MVECs, we expanded this known genetic network and our knowledge on stem cells by identifying novel GATA2 targets, including ANGPT1. The robustness of this strategy can be sped up by incorporating more systems biology algorithms, such as coexpression correlation deduced by Pearson correlation coefficient, or more dynamic algorithms, such as liquid association [22, 23, 26, 60]. Applying such analyses to stem cells is the notion of correlated expression patterns of genes with related functions or regulator-target relationships, a high-level self-organization in gene expression networks, and a scale-free topology of such networks in cells [23, 61].

We found that GATA2 could induce transcriptome reprogramming in human endothelial cells. Previous studies had shown that differentiated adult cells could be transformed, or dedifferentiated, into pluripotent cells when fused with ES cells or by exposing them to extracts of ES cells [54, 55, 62]. This suggested that factors found in ES cells might be essential to conferring pluripotency on other cells. The artificially induced pluripotent stem cells could be generated directly from human or mouse skin fibroblast cultures by the introducing just four defined stemness factors: Oct3/4, Sox2, cMyc, and Klf4 [56, 63–67]. However, it is not clear yet whether genes from somatic stem cells have a similar ability. For HSCs, one of the critical factors might be GATA2. GATA2 is involved in the restriction of hematopoietic cell differentiation. Forced expression of GATA2 in erythroid precursors provoked their proliferation while blocking their differentiation [9, 52]. In adipose tissues, constitutive expression of GATA2 in brown adipocytes suppressed genes expressed in matured adipocytes, whereas disruption of a GATA2 allele resulted in significantly elevated differentiation of preadipocytes [68]. These data, together with our transcriptome distance measurement, indicated that key factors of somatic stem cells also hold the potential to reinduce multipotency. The combination of GATA2 with other critical somatic stemness genes may eventually reprogram the genetic network of endothelial cells back to a stem cell-like state in vitro. Manipulating the steady expression of GATA2, as well as other discovered stem cell genes, may also help the ex vivo expansion of CD133+ stem cells.

We evaluated the impact of GATA2 overexpression in matured endothelial cells, but no significant phenotypic changes could be observed (data not shown). This may due to the fact that it will take more than one gene to induce a significant stem cell-like phenotype [56, 63–67]. GATA2 alone, therefore, may not be enough to induce a clear stem-like phenotype. However, a global transcriptome change did occur in GATA2-expressing ECs (Fig. 5), implying that they had started their journey back to the ancestor stem cell state. These data further imply that detecting gene expression changes is a more sensitive tactic than analyzing cellular function in studies such as somatic cell dedifferentiation or transformation. A recent report on the transformation of human MSCs supported this point: even though clear malignancy phenotypes (such as grow in soft agar or form tumors in nude mice) could be observed only after all five oncogenes (hTERT, E6, E7, small t Ag, and Ras) were introduced into primary MSCs, clear transcriptome changes could be detected whenever an oncogene was added [69].


   CONCLUSION
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
Disclosure of Potential...
 Acknowledgments
 References
 
Stemness and differentiation are most complex processes governed by a highly coordinated regulation of distinct genetic programs. The biological function of genes enriched in CD133+ cells, NSCs, or MSCs remain mostly unknown. More comprehensive, integrated studies enabling the determination of all interactions will offer additional insights into how such a complex interaction map may contribute to unique stem cell behaviors. Thus, this study provides novel strategies for additional genetic network studies on stem cells and on somatic cell dedifferentiation research. We also provide a novel bioinformatics approach, via calculating the transcriptome distances, to study dedifferentiation induced by stem cell genes. Lastly, we believe that the approaches taken here can also serve as a model for other complex biological processes.


   DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
Acknowledgments
 References
 
The authors indicate no potential conflicts of interest.


   ACKNOWLEDGMENTS
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
 Acknowledgments
References
 
We thank Li-Li Li and Dr. Oscar Kuang-Sheng Lee for critical reading of the manuscript. We acknowledge the technical services provided by Gene Expression Analysis Core Facility of the Veteran General Hospital-Yang Ming Genome Center (VYMGC), National Yang-Ming University. The Gene Expression Analysis Core Facility is supported by National Research Program for Genomic Medicine (NRPGM), National Science Council. This work was supported by the National Science Council (NRPGM, NSC96-3112-B-010-009 and NSC96-2320-B-010-026), the Yen Tjing Lin Medical Foundation (CI-94-10 and CI-96-11), the Taipei City Hospital (95002-62-086), and a grant from Ministry of Education Aim for the Top University Plan.


   REFERENCES
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Conclusion
 Disclosure of Potential...
 Acknowledgments
 References
 

  1. Pelosi E, Valtieri M, Coppola S et al. Identification of the hemangioblast in postnatal life. Blood 2002;100:3203–3208.[Abstract/Free Full Text]

  2. Gläsker S, Li J, Xia JB et al. Hemangioblastomas share protein expression with embryonal hemangioblast progenitor cell. Cancer Res 2006;66:4167–4172.[Abstract/Free Full Text]

  3. Brunstein CG, Baker KS, Wagner JE. Umbilical cord blood transplantation for myeloid malignancies. Curr Opin Hematol 2007;14:162–169.[Medline]

  4. Salven P, Mustjoki S, Alitalo R et al. VEGFR-3 and CD133 identify a population of CD34+ lymphatic/vascular endothelial precursor cells. Blood 2003;101:168–172.[Abstract/Free Full Text]

  5. Jaatinen T, Hemmoranta H, Hautaniemi S et al. Global gene expression profile of human cord blood-derived CD133+ cells. STEM CELLS 2006;24:631–641.[Abstract/Free Full Text]

  6. He X, Gonzalez V, Tsang A et al. Differential gene expression profiling of CD34+ CD133+ umbilical cord blood hematopoietic stem progenitor cells. Stem Cells Dev 2005;14:188–198.[CrossRef][Medline]

  7. Martin-Rendon E, Hale SJ, Ryan D et al. Transcriptional profiling of human cord blood CD133+ and cultured bone marrow mesenchymal stem cells in response to hypoxia. STEM CELLS 2007;25:1003–1012.[Abstract/Free Full Text]

  8. Kawabata H, Germain RS, Ikezoe T et al. Regulation of expression of murine transferrin receptor 2. Blood 2001;98:1949–1954.[Abstract/Free Full Text]

  9. Minegishi N, Suzuki N, Yokomizo T et al. Expression and domain-specific function of GATA-2 during differentiation of the hematopoietic precursor cells in midgestation mouse embryos. Blood 2003;102:896–905.[Abstract/Free Full Text]

  10. Mitchell KE, Weiss ML, Mitchell BM et al. Matrix cells from Wharton's jelly form neurons and glia. STEM CELLS 2003;21:50–60.[Abstract/Free Full Text]

  11. Lee OK, Kuo TK, Chen WM et al. Isolation of multipotent mesenchymal stem cells from umbilical cord blood. Blood 2004;103:1669–1675.[Abstract/Free Full Text]

  12. Panepucci RA, Siufi JL, Silva WA Jr et al. Comparison of gene expression of umbilical cord vein and bone marrow-derived mesenchymal stem cells. STEM CELLS 2004;22:1263–1278.[Abstract/Free Full Text]

  13. Wang HS, Hung SC, Peng ST et al. Mesenchymal stem cells in the Wharton's jelly of the human umbilical cord. STEM CELLS 2004;22:1330–1337.[Abstract/Free Full Text]

  14. Chang YJ, Shih DT, Tseng CP et al. Disparate mesenchyme-lineage tendencies in mesenchymal stem cells from human bone marrow and umbilical cord blood. STEM CELLS 2006;24:679–685.[Abstract/Free Full Text]

  15. Pittenger MF, Mackay AM, Beck SC et al. Multilineage potential of adult human mesenchymal stem cells. Science 1999;284:143–147.[Abstract/Free Full Text]

  16. Wagner W, Wein F, Seckinger A et al. Comparative characteristics of mesenchymal stem cells from human bone marrow, adipose tissue, and umbilical cord blood. Exp Hematol 2005;33:1402–1416.[CrossRef][Medline]

  17. Tsai MS, Hwang SM, Chen KD et al. Functional network analysis on the transcriptomes of mesenchymal stem cells derived from amniotic fluid, amniotic membrane, cord blood, and bone marrow. STEM CELLS 2007;25:2511–2523.[Abstract/Free Full Text]

  18. Ivanova NB, Dimos JT, Schaniel C et al. A stem cell molecular signature. Science 2002;298:601–604.[Abstract/Free Full Text]

  19. Ramalho-Santos M, Yoon S, Matsuzaki Y et al. "Stemness": Transcriptional profiling of embryonic and adult stem cells. Science 2002;298:597–600.[Abstract/Free Full Text]

  20. Cai Y, Wu P, Ozen M et al. Gene expression profiling and analysis of signaling pathways involved in priming and differentiation of human neural stem cells. Neuroscience 2006;138:133–148.[CrossRef][Medline]

  21. Yu S, Zhang JZ, Xu Q. Genes associated with neuronal differentiation of precursors from human brain. Neuroscience 2006;141:817–825.[CrossRef][Medline]

  22. Barabási AL, Oltvai ZN. Network biology: Understanding the cell's functional organization. Nat Rev Genet 2004;5:101–113.[CrossRef][Medline]

  23. Janes KA, Yaffe MB. Data-driven modelling of signal-transduction networks. Nat Rev Mol Cell Biol 2006;7:820–828.[CrossRef][Medline]

  24. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100:57–70.[CrossRef][Medline]

  25. Morrison SJ, Kimble J. Asymmetric and symmetric stem-cell divisions in development and cancer. Nature 2006;441:1068–1074.[CrossRef][Medline]

  26. Woolf PJ, Prudhomme W, Daheron L et al. Bayesian analysis of signaling networks governing embryonic stem cell fate decisions. Bioinformatics 2005;21:741–753.[Abstract/Free Full Text]

  27. Calvano SE, Xiao W, Richards DR et al. A network-based analysis of systemic inflammation in humans. Nature 2005;437:1032–1037.[CrossRef][Medline]

  28. Wang HW, Trotter MW, Lagos D et al. Kaposi sarcoma herpesvirus-induced cellular reprogramming contributes to the lymphatic endothelial gene expression in Kaposi sarcoma. Nat Genet 2004;36:687–693.[CrossRef][Medline]

  29. Storey JD, Tibshirani R. Statistical significance for genomewide studies. Proc Natl Acad Sci U S A 2003;100:9440–9445.[Abstract/Free Full Text]

  30. Li C, Wong WH. Model-based analysis of oligonucleotide arrays: Expression index computation and outlier detection. Proc Natl Acad Sci U S A 2001;98:31–36.[Abstract/Free Full Text]

  31. Yang TP, Chang TY, Lin CH et al. ArrayFusion: A web application for multi-dimensional analysis of CGH, Snp and microarray data. Bioinformatics 2006;22:2697–2698.[Abstract/Free Full Text]

  32. Godfrey A, Anderson J, Papanastasiou A et al. Inhibiting primary effusion lymphoma by lentiviral vectors encoding short hairpin RNA. Blood 2005;105:2510–2518.[Abstract/Free Full Text]

  33. Lagos D, Trotter MW, Vart RJ et al. Kaposi sarcoma herpesvirus-encoded vFLIP and vIRF1 regulate antigen presentation in lymphatic endothelial cells. Blood 2007;109:1550–1558.[Abstract/Free Full Text]

  34. Phillips RL, Ernst RE, Brunk B et al. The genetic program of hematopoietic stem cells. Science 2000;288:1635–1640.[Abstract/Free Full Text]

  35. Baldus CD, Tanner SM, Kusewitt DF et al. BAALC, a novel marker of human hematopoietic progenitor cells. Exp Hematol 2003;31:1051–1056.[Medline]

  36. Ogawa M, Kizumoto M, Nishikawa S et al. Expression of alpha4-integrin defines the earliest precursor of hematopoietic cell lineage diverged from endothelial cells. Blood 1999;93:1168–1177.[Abstract/Free Full Text]

  37. Sakai D, Suzuki T, Osumi N et al. Cooperative action of Sox9, Snail2 and Pka signaling in early neural crest development. Development 2006;133:1323–1333.[Abstract/Free Full Text]

  38. Yang J, Mani SA, Donaher JL et al. Twist, a master regulator of morphogenesis, plays an essential role in tumor metastasis. Cell 2004;117:927–939.[CrossRef][Medline]

  39. Jung CG, Hida H, Nakahira K et al. Pleiotrophin mRNA is highly expressed in neural stem (progenitor) cells of mouse ventral mesencephalon and the product promotes production of dopaminergic neurons from embryonic stem cell-derived nestin-positive cells. FASEB J 2004;18:1237–1239.[Abstract/Free Full Text]

  40. Hanashima C, Li SC, Shen L et al. Foxg1 suppresses early cortical cell fate. Science 2004;303:56–59.[Abstract/Free Full Text]

  41. Kim SK, Selleri L, Lee JS et al. Pbx1 inactivation disrupts pancreas development and in Ipf1-deficient mice promotes diabetes mellitus. Nat Genet 2002;30:430–435.[CrossRef][Medline]

  42. Sanyal M, Tung JW, Karsunky H et al. B-cell development fails in the absence of the Pbx1 proto-oncogene. Blood 2007;109:4191–4199.[Abstract/Free Full Text]

  43. DiMartino JF, Selleri L, Traver D et al. The Hox cofactor and proto-oncogene Pbx1 is required for maintenance of definitive hematopoiesis in the fetal liver. Blood 2001;98:618–626.[Abstract/Free Full Text]

  44. Gregorio-King CC, McLeod JL, Collier FM et al. MERP1: A mammalian ependymin-related protein gene differentially expressed in hematopoietic cells. Gene 2002;286:249–257.[CrossRef][Medline]

  45. Takakura N, Watanabe T, Suenobu S et al. A role for hematopoietic stem cells in promoting angiogenesis. Cell 2000;102:199–209.[CrossRef][Medline]

  46. Leibundgut K, Schmitz NM, Hirt A. Catalytic activities of G1 cyclin-dependent kinases and phosphorylation of retinoblastoma protein in mobilized peripheral blood CD34+ hematopoietic progenitor cells. STEM CELLS 2005;23:1002–1011.[Abstract/Free Full Text]

  47. Forrest D, Erway LC, Ng L et al. Thyroid hormone receptor beta is essential for development of auditory function. Nat Genet 1996;13:354–357.[CrossRef][Medline]

  48. White R, Leonardsson G, Rosewell I et al. The nuclear receptor co-repressor nrip1 (RIP140) is essential for female fertility. Nat Med 2000;6:1368–1374.[CrossRef][Medline]

  49. Loges S, Fehse B, Brockmann MA et al. Identification of the adult human hemangioblast. Stem Cells Dev 2004;13:229–242.[CrossRef][Medline]

  50. Suzuki N, Ohneda O, Minegishi N et al. Combinatorial Gata2 and Sca1 expression defines hematopoietic stem cells in the bone marrow niche. Proc Natl Acad Sci U S A 2006;103:2202–2207.[Abstract/Free Full Text]

  51. Tsai FY, Orkin SH. Transcription factor GATA-2 is required for proliferation/survival of early hematopoietic cells and mast cell formation, but not for erythroid and myeloid terminal differentiation. Blood 1997;89:3636–3643.[Abstract/Free Full Text]

  52. Ikonomi P, Rivera CE, Riordan M et al. Overexpression of GATA-2 inhibits erythroid and promotes megakaryocyte differentiation. Exp Hematol 2000;28:1423–1431.[CrossRef][Medline]

  53. Zou JX, Wang B, Kalo MS et al. An Eph receptor regulates integrin activity through R-Ras. Proc Natl Acad Sci U S A 1999;96:13813–13818.[Abstract/Free Full Text]

  54. Taranger CK, Noer A, Sorensen AL et al. Induction of dedifferentiation, genomewide transcriptional programming, and epigenetic reprogramming by extracts of carcinoma and embryonic stem cells. Mol Biol Cell 2005;16:5719–5735.[Abstract/Free Full Text]

  55. Grinnell KL, Yang B, Eckert RL et al. De-differentiation of mouse interfollicular keratinocytes by the embryonic transcription factor Oct-4. J Invest Dermatol 2007;127:372–380.[CrossRef][Medline]

  56. Takahashi K, Yamanaka S. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 2006;126:663–676.[CrossRef][Medline]

  57. Subramanian A, Tamayo P, Mootha VK et al. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proc Natl Acad Sci U S A 2005;102:15545–15550.[Abstract/Free Full Text]

  58. Saxton TM, Henkemeyer M, Gasca S et al. Abnormal mesoderm patterning in mouse embryos mutant for the SH2 tyrosine phosphatase Shp-2. EMBO J 1997;16:2352–2364.[CrossRef][Medline]

  59. Chen B, Bronson RT, Klaman LD et al. Mice mutant for Egfr and Shp2 have defective cardiac semilunar valvulogenesis. Nat Genet 2000;24:296–299.[CrossRef][Medline]

  60. Li KC, Liu CT, Sun W et al. A system for enhancing genome-wide coexpression dynamics study. Proc Natl Acad Sci U S A 2004;101:15561–15566.[Abstract/Free Full Text]

  61. Yook SH, Oltvai ZN, Barabasi AL. Functional and topological characterization of protein interaction networks. Proteomics 2004;4:928–942.[CrossRef][Medline]

  62. Tada M, Takahama Y, Abe K et al. Nuclear reprogramming of somatic cells by in vitro hybridization with ES cells. Curr Biol 2001;11:1553–1558.[CrossRef][Medline]

  63. Okita K, Ichisaka T, Yamanaka S. Generation of germline-competent induced pluripotent stem cells. Nature 2007;448:313–317.[CrossRef][Medline]

  64. Wernig M, Meissner A, Foreman R et al. In vitro reprogramming of fibroblasts into a pluripotent ES-cell-like state. Nature 2007;448:318–324.[CrossRef][Medline]

  65. Park IH, Zhao R, West JA et al. Reprogramming of human somatic cells to pluripotency with defined factors. Nature 2008;451:141–146.[CrossRef][Medline]

  66. Takahashi K, Tanabe K, Ohnuki M et al. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 2007;131:861–872.[CrossRef][Medline]

  67. Yu J, Vodyanik MA, Smuga-Otto K et al. Induced pluripotent stem cell lines derived from human somatic cells. Science 2007;318:1917–1920.[Abstract/Free Full Text]

  68. Tsai J, Tong Q, Tan G et al. The transcription factor GATA2 regulates differentiation of brown adipocytes. EMBO Rep 2005;6:879–884.[CrossRef][Medline]

  69. Funes JM, Quintero M, Henderson S et al. Transformation of human mesenchymal stem cells increases their dependency on oxidative phosphorylation for energy production. Proc Natl Acad Sci U S A 2007;104:6223–6228.[Abstract/Free Full Text]

  70. Wang L, Wakisaka N, Tomlinson CC et al. The Kaposi's sarcoma-associated herpesvirus (KSHV/HHV-8) K1 protein induces expression of angiogenic and invasion factors. Cancer Res 2004;64:2774–2781.[Abstract/Free Full Text]





This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
2007-0821v1
26/5/1186    most recent
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Reprints/Permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Huang, T.-S.
Right arrow Articles by Wang, H.-W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Huang, T.-S.
Right arrow Articles by Wang, H.-W.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
STEM CELLS THE ONCOLOGIST CME ALPHAMED PRESS JOURNALS