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


     

First published online February 1, 2007
Stem Cells Vol. 25 No. 5 May 2007, pp. 1298 -1306
doi:10.1634/stemcells.2006-0660; www.StemCells.com
© 2007 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:
2006-0660v1
25/5/1298    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 HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Shin, S.
Right arrow Articles by Chesnut, J. D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Shin, S.
Right arrow Articles by Chesnut, J. D.

TISSUE-SPECIFIC STEM CELLS

Whole Genome Analysis of Human Neural Stem Cells Derived from Embryonic Stem Cells and Stem and Progenitor Cells Isolated from Fetal Tissue

Soojung Shina, Yu Sunb, Ying Liua, Hanita Khanerc, Smita Svantd, Jingli Caie, Qin Xiu Xuf, Bruce P. Davidsonf, Steven L. Sticeg, Alan K. Smithd, Steven A. Goldmanh, Benjamin E. Reubinoffc, Ming Zhanb, Mahendra S. Raoa, Jonathan D. Chesnuta

aStem Cells and Regenerative Medicine, Invitrogen, Carlsbad, California, USA;
bBioinformatics Unit, Branch of Research Resources, NIA, NIH, Baltimore, Maryland, USA;
cThe Hadassah Human Embryonic Stem Cell Research Center, The Goldyne Savad Institute of Gene Therapy, Department of Gynecology, Hadassah University Hospital, Jerusalem, Israel;
dTheradigm Inc., Baltimore, Maryland, USA;
eThomas Jefferson University, Philadelphia, Pennsylvania, USA;
fES Cell International Pte Ltd, Helios, Singapore;
gUniversity of Georgia, Athens, Georgia, USA;
hUniversity of Rochester, Rochester, New York, USA

Key Words. Large scale genomic analysis • Neural stem cells • Human embryonic stem cells • Adult stem cells • Differentiation Lineage restricted precursor cells

Correspondence: Jon D. Chesnut, Ph.D., 1610 Faraday Ave, Carlsbad, California 92008, USA. Telephone: 760-603-7253; Fax: 760-602-6553 e-mail: jon.chesnut{at}invitrogen.com

Received October 27, 2006; accepted for publication January 19, 2007.
First published online in STEM CELLS EXPRESS   February 1, 2007.


   ABSTRACT
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 Disclosure of Potential...
 Acknowledgments
 References
 
Multipotent neural stem cells (NSC) have been derived from human embryonic stem cells (hESC) as well as isolated from fetal tissues. However, there have been few exclusive markers of NSC identified to date, and the differences between NSC from various sources are poorly understood. Although cells isolated from these two sources share many important characteristics, it is not clear how closely they are related in terms of gene expression. Here, we compare the gene expression profiles of 11 lines of NSC derived from hESC (ES_NSC), four lines of NSC isolated from fetus (F_NSC), and two lines of restricted progenitors in order to characterize these cell populations and identify differences between NSC derived from these two sources. We showed that ES_NSC were clustered together with high transcriptional similarities but were distinguished from F_NSC, oligodendrocyte precursor cells, and astrocyte precursor cells. There were 17 genes expressed in both ES_NSC and F_NSC whose expression was not identified in restricted neural progenitors. Furthermore, the major differences between ES_NSC and F_NSC were mostly observed in genes related to the key neural differentiation pathways. Here, we show that comparison of global gene expression profiles of ES_NSC, F_NSC, and restricted neural progenitor cells makes it possible to identify some of the common characteristics of NSC and differences between similar stem cell populations derived from hESCs or isolated from fetal tissue.

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


   INTRODUCTION
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 Disclosure of Potential...
 Acknowledgments
 References
 
Embryonic stem cells (ESC), derived from the inner cell mass of preimplantation mammalian embryos, are a unique population of pluripotent cells that can differentiate into the embryonic precursors of all corresponding adult tissues both in vitro and in vivo [1, 2]. Perhaps the neural phenotype is the easiest to obtain from ESC and, indeed, several investigators have suggested it should be the default differentiation fate of ESC [3, 4]. Several methods to obtain neural differentiation have been reported. These include embryoid body formation and selection in NSC medium [5–7], sorting with CD133 after differentiation [8], selecting neurosphere forming cells after plating cells in low density [9], a direct differentiation of adherent culture in defined medium [10, 11], or enhancing/biasing neural differentiation by altering bone morphogenic protein/transforming growth factor-β signaling or altering Notch signaling [12–14]. In all of these processes, a number of nestin-positive dividing populations can be obtained that express neural markers, and these populations can be induced to differentiate into neurons, astrocytes, and immature oligodendrocytes. Neurons are generally born first with glial precursors taking longer time periods to differentiate, similar to normal fetal development [15, 16].

NSC have also been isolated from fetal tissues. At least two types of NSC have been identified. The neuroepithelial NSC reside in the ventricular zone and present themselves relatively early in development [17–19] and are followed by the later-appearing neurosphere forming NSC [20] that are likely to reside in the more rostral subventricular zone regions. These NSC have been propagated for prolonged time periods and subjected to large-scale analysis by multiple methods [21, 22]. Each of these NSC populations has the ability to self-renew to some extent and can differentiate into neurons, astrocytes, and oligodendrocytes. It is, however, unclear whether the properties of these cells are similar to those of NSC derived from hESC (ES_NSC). To our knowledge, no direct comparisons between these populations in humans have been made.

Several additional multipotent populations and dividing progenitors with a more restricted phenotype have also been identified in both mouse and human tissue [23–25]. Goldman and colleagues, for example, have isolated a glial progenitor population that has the ability to differentiate into oligodendrocytes and astrocytes when transplanted into a shiverer mouse model [26]. Neuronal and glial restricted populations have also been derived from ESC cultures and their properties described [24, 27, 28]. Whole genome analysis of these populations has not been reported, but markers that clearly distinguish these cells from more undifferentiated stem cells and more differentiated mature cells of the central nervous system have been identified [25].

To assess the properties of NSC, we employed a whole genome bead-based technology developed by Illumina Inc. (San Diego, http://www.illumina.com) that combines the sensitivity and low cost of a focused array with the coverage of a large-scale array [29]. We used this bead array platform to analyze multiple NSC populations that were derived from hESC or directly harvested from human fetal tissue samples. By comparing NSC populations derived from different sources, we identified a set of core similarities and specific differences. Comparison with data sets developed by profiling ESC, embryoid bodies, and differentiated precursor cells showed clear differences, and a series of unique NSC markers could be identified. Overall, our data suggest that common neural stem cell features can be identified, whereas ES_NSC and fetal-derived NSC (F_NSC) differ in their use of critical pathways such as Wnt, fibroblast growth factor (FGF), and leukemia inhibitory factor (LIF) signaling pathways.


   MATERIALS AND METHODS
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 Disclosure of Potential...
 Acknowledgments
 References
 
Cell Culture and Characterization for Sample Verification Before Global Analysis
Nineteen cell samples were selected from multiple laboratories for expression profiling as indicated in Table 1. Cells were prepared according to the protocol of each laboratory and references for detailed information are listed in Table 1. For immunocytochemistry, cells were fixed with 4% paraformaldehyde and blocked for 30 minutes with phosphate-buffered saline containing host serum of secondary antibody before incubation in primary antibody. The antibodies and dilutions used were as follows: Nestin (1:200; R&D Systems Inc., Minneapolis, http://www.rndsystems.com), Sox2 (1:2,000; R&D systems), βIII tubulin (1:2,000; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), 5-bromo-2'-deoxyuridine (1:5; Developmental Studies Hybridoma Bank, Iowa City, IA, http://www.uiowa.edu/~dshbwww), Glast (1:1,000; kind gift from Dr. Maragakis), and galactocerebroside (GalC) (1:5; kind gift from Dr. B. Ranscht). Selected neuronal and non-neuronal gene expression profiles were examined using reverse transcription-polymerase chain reaction (RT-PCR) using the following cycle parameters: 94°C for 1 minute, 55°C for 1 minute, 72°C for 1 minute for 33 cycles (primer sequences are available upon request).


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

  		Table 1. Samples used in the paper

 
BeadArray and Gene Expression Analysis
Total cellular RNA was isolated from cultured cells using the TRIzol (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) or Qiagen RNeasy kit (Qiagen, Hilden, Germany, http://www1.qiagen.com) and used further both for BeadArray and polymerase chain reaction (PCR) analysis. Sample amplification was performed with 100 ng of total RNA using the Illumina RNA Amplification kit and labeling was achieved by incorporation of biotin-16-UTP (PerkinElmer Life and Analytical Sciences, Boston, http://www.perkinelmer.com) at a ratio of 1:1 with unlabeled UTP. Labeled, amplified material (700 ng per array) was hybridized to Illumina BeadChips according to the manufacturer's instructions. Arrays were scanned with an Illumina BeadArray Reader confocal scanner according to the manufacturer's instructions. Array data processing and normalization were performed using Illumina BeadStudio software. Results were normalized to internal background controls and the raw data set is available for download as an Excel spreadsheet from (ftp://resources.invitrogen.com/pub.stemcells2007). TIFF images are available upon request, and detailed analysis of the sensitivity and quality control tests used in array manufacture and algorithm used in the BeadStudio software used for analysis is available from Illumina Inc.

Samples were run over four different time points with three different versions of slide formats (pilot version of HumRef8, HumRef8, and WG6). There were 22,200 genes shared by all three slide formats, and analysis was undertaken for these genes. To access the robustness and compatibility of methods, the genes whose detection scores were over either 0.99 or 0.95 were examined and listed in Table 1. Variation from different time points was observed in samples run in May such that a higher signal value was required to obtain a detection score of ≥0.99. Given the reduced sensitivity of detection (300 or higher signal intensity for a gene with a score of ≥0.99) and the information from supplemental online Table 1 indicating a similar gene distribution from the May through July samples, we determined that a fixed cutoff of 0.99 would give rise to false negative detection of the genes for samples run in May. To compromise the observed discrepancy, we applied two different criteria—a 0.95 cutoff for the slides run in May and a 0.99 cutoff for the other set to decide the presence of or absence of the expression of the selected genes.

Identification of Differentially Expressed Genes and Clustering Analysis
To identify differentially expressed genes, we chose eight samples (N6–N9 with duplicates) of ES_NSC and four samples (F1–F2 with duplicates) of F_NSC. Genes differentially expressed between ES_NSC and F_NSC samples were identified using a t test with p < .01 and the fold change ≥2. Hierarchical clustering of these differentially expressed genes was conducted using the Cluster 3.0 software with centroid linkage and visualized using TreeView (Fig. 4) (http://bonsai.ims.u-tokyo.ac.jp/~mdehoon/software/cluster/software.htm). Principle component analysis was conducted using the Bioconductor packages (http://www.bioconductor.org). Gene ontology biological processes enriched in differentially expressed genes were identified using Fisher's exact test as implemented in the program GoMiner (http://discover.nci.nih.gov/gominer) [30]. The top 15 marker genes that can distinguish ES_NSC and F_NSC samples were identified using a prediction analysis of microarray (PAM) (Fig. 4) [31].


   RESULTS
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 Disclosure of Potential...
 Acknowledgments
 References
 
Obtaining and Characterizing Samples
Nineteen cell samples were selected from multiple laboratories (Table 1). Eleven ES_NSC samples were obtained from five different laboratories and their differentiation potential to neurons, astrocytes, and oligodendrocytes and marker expression profiles as provided by each laboratory are summarized by the representative micrographs shown in Figure 1B. Like other NSC, the NIH sample (N5) consisted of a homogenous population of proliferating cells that expressed known NSC markers such as Sox2 and Nestin. When these cells were differentiated, they expressed βIII tubulin (neuronal marker), Glast (glial marker), or GalC (glial marker). In addition, the expression profiles of known NSC genes were examined for all ES_NSC samples used in this study. As seen in Figure 1A, the overall expression pattern of the selected genes was quite similar among the ten samples tested. Although there seems to be some variation in expression level, all ES_NSC samples displayed expression of the generally accepted as markers of NSC Sox1, Sox2, CD133, Nestin, and Musashi1. Samples were also assayed for the presence of differentiated cells and contaminating non-neural populations using RT-PCR. Although ESC markers were not detected in most of the ES_NSC samples, all ten samples showed expression of neurofilament and some non-neuronal genes such as the endodermal gene AFP and the mesodermal gene Eomes. These results indicated that there was some contamination of non-neuronal cells, which suggests that care should be taken in interpreting the data. This being the case, we did not detect expression of more mature markers (choline acetyltransferase, tyrosine hydroxylase [TH]) or those of late glial markers (glial fibrillary acidic protein [GFAP], myelin-associated glycoprotein [MAG], oligodendrocyte transmembrane [OSP]), suggesting that the neural cells in the sample were indeed NSC.


Figure 1
View larger version (58K):
[in this window]
[in a new window]

  		Figure 1. Characterization of sample populations. ES_NSC (N1–N10) shared gene expression patterns for known neural stem cell genes (A). ES_NSC proliferate, express neural stem cell phenotype markers, and give rise to neurons, astrocytes, and oligodendrocytes. Representative immunocytochemistry results are presented. Nuclei are stained with 4,6-diamidino-2-phenylindole and appear blue in the micrograph. Bar = 50 µm (B). Abbreviations: APC, astrocyte precursor cell; BrdU, 5-bromo-2'-deoxyuridine; EBC, embryoid bodies; ES_NSC, hESC-derived NSC; F_NSC, fetal-derived NSC; GaLC, galactocerebroside; OPC, oligodendrocyte precursor cell.

 
As a further comparison, the following samples were analyzed in parallel. Four samples, identified as fetal tissue NSC (F_NSC), were obtained from two different sources. These samples expressed NSC genes and did not express mature neuron, oligodendrocyte, or non-neuronal lineage genes. However, two samples showed expression of GFAP, indicating some glial cell contamination. Given the presence of contaminating cells, we ran two additional controls, which included astrocyte precursor cells (APC) and oligodendrocyte precursor cells (OPC), that were isolated and purified as previously described [25, 26]. These two populations were selected, as they represent dividing differentiated progeny of NSC and are likely contaminants of propagated fetal NSC. ESC and embryoid bodies were prepared and profiled as previously described, and previously published data sets were used for this comparison [32]. This large data set was considered adequate to identify ES_NSC markers, despite some contaminating populations after subtraction of expression profiles of the control samples.

BeadArray Data Acquisition and Verification
All samples were analyzed using the Illumina BeadArray platform, which includes oligonucleotide sets representing 22,200 unique transcripts. Intensity results of the transcripts were examined and normalized to exclude background signals and reported in arbitrary units. Before detailed analysis of the BeadArray data was done, several quality control tests were performed on the array data sets. Some samples were run in duplicate on arrays processed at different time points to assess array-to-array variability. No array results were accepted if the correlation coefficient (R2 value) between technical replicates across arrays was less than 0.95 (supplemental online Table 2). We also examined the overall pattern of gene expression based on signal intensity (supplemental online Table 1). In order for array data to be accepted for analysis, the overall gene expression patterns had to be similar with no wide divergence between samples. Large variations indicate either hybridization errors or problems with the quality of the RNA used. The number of genes detected at different signal intensities for each sample is shown in supplemental online Table 1. Overall gene numbers were shown to be consistent across the samples so that approximately 40% of genes were observed with signal intensity greater than 100. However, samples in one run (N11, E, U) showed consistently lower signal intensity (26.4–35.6), which suggests that caution should be taken in making comparisons between these slides and analyzing expression changes or pooling results. For this reason, we decided to exclude N11 expression data and did not pool it to ES_NSC samples. Based on the pattern of gene expression across all samples, we chose an arbitrary signal intensity of 100 as a cutoff for our initial analysis. To further assess the quality of the arrays used in this study, we examined the expression values for genes selected from Figure 1, for which PCR results had been obtained (supplemental online Table 5). Although there was good correlation between most samples analyzed, three genes showed some discrepancy. For example, Sox1 and Sox2 were shown to be expressed in these samples by PCR but were not detected by the Illumina BeadArray. In addition, false-positive values were obtained for MAG, which has numerous splice variants that are expressed early in development [33]. These discrepancies emphasize the importance of independent confirmation of microarray results. In summary, we conclude that by using the proper controls and taking the appropriate precautions in analyzing the data, this type of analysis can be a useful tool to compare multiple samples.

Correlation Among Populations and Comparison of ES_NSC with Other Populations
Next, we examined the degree of similarity among samples in terms of overall correlation coefficient (supplemental online Table 2). A cluster analysis was conducted and a dendrogram analysis, presented in Figure 2, was prepared using BeadStudio, a software program specifically designed by Illumina for this analysis. Using this analysis, we found that ES_NSC samples clustered together and could be discriminated from other populations analyzed in this study. Furthermore, the dendrogram showed that samples N1–N10 clustered to form one group of ES_NSC, which could easily be separated from F_NSC, OPC, or APC. Interestingly, samples N1–N4 and N6–N9 clustered very closely, potentially as a result of these samples being derived using similar methodologies (albeit by two different groups). In addition, the N8–N10 ES_NSC samples had a correlation value higher than 0.862, which was also higher than that of F_NSC, OPC, ESC, or EB. Among the populations compared here, APC clustered furthest away from any other group (0.567–0.709). From all these data, we concluded that the overall gene expression profile seen in ES_NSC could be discerned from that seen in F_NSC or from the differentiated progenies of OPC and APC.


Figure 2
View larger version (23K):
[in this window]
[in a new window]

  		Figure 2. Clustering populations using hierarchical analysis using BeadStudio. Hierarchical clustering dendrogram of relative gene expression in different populations was generated using the BeadStudio software. ES_NSC (N1–N10) clustered together and are distinguished from F_NSC (F1–F4), oligodendrocyte precursor cells, and astrocyte precursor cells. Abbreviations: a, replicate; A, astrocyte precursor cells; ES_NSC, hESC-derived NSC; F_NSC, fetal-derived NSC; O, oligodendrocyte precursor cells.

 
Genes Differentially Expressed in ES_NSC
We next wanted to scrutinize relatedness between the embryonic stem cell-derived and fetal-derived NSC populations. To accomplish this, we evaluated the gene expression profiles of ES_NSC (N1–N7) based on their detection scores. Genes with a detection value equal to or higher than 0.95 for both replicates were considered to be present in the population under analysis. The total number of genes expressed by each cell line (obtained from different labs) approximated 8,446 genes for N1–N4 (Reubinoff), 8,476 genes for N5 (NIH), and 8,711 genes for N6–N9 (ESI, ES Cell International, Helios, Singapore). A Venn diagram was prepared to show relative differences and commonalities in genes expressed between ES_NSC cell populations (Fig. 3). Among the expressed genes, 7,371 (75.9%) were shared by all ES_NSC examined. The list of genes included Nestin, Musashi1, and CD133 and was devoid of non-neuroectodermal genes such as AFP, Eomes, and Hand1, whose expressions were detected (Fig. 1) by subsets of ES_NSC samples.


Figure 3
View larger version (27K):
[in this window]
[in a new window]

  		Figure 3. Number and percentage of genes shared by compared populations. ES_NSC (N1–N7) from three separate laboratories (N1–N4, N5, N6–N7) shared 7,371 (75.9%) expressed genes (A). Among these genes, 5,840 genes were shared with both APC and OPC, and 90 genes were exclusively expressed by ES_NSC (B). Abbreviations: APC, astrocyte precursor cells; ES_NSC, hESC-derived NSC; OPC, oligodendrocyte precursor cells.

 
We then compared the expression patterns of ES_NSC shared genes (7,371 genes) with those of restricted progenitors of OPC and APC in an attempt to find genes exclusively expressed by ES_NSC. As shown in Figure 3, APC and OPC expressed 8,446 and 8,794 genes, respectively, and 55.4% of all genes analyzed were shared by all of the three populations of ES_NSC, APC, and OPC. This shared pool was highly enriched with homeostasis and ribosomal protein related genes. In addition, 90 genes (0.9%) were shown to be exclusively expressed by ES_NSC (Table 2). When we searched expressed sequence tag (EST) sets imported from the National Center for Biotechnology Information for these genes, 80 out of 82 genes (for eight genes, data were not available) had EST counts in brain tissue. When their EST counts were broken down by developmental stage, 17 genes showed restricted expression to embryonic stages of development at a time when stem and progenitor cells were much more abundant (data not shown). In addition, the gene PAQR6, which is a progestin and adipoQ receptor family member VI, showed a restricted expression pattern so that more than half of all contributing ESTs came from the brain. When these genes were examined for their expression in ESC, PDE7A, LOC63920, and LOC374823 (lectin) were selected as exclusive genes for ES_NSC, whose expressions were not shown for any of other compared populations (ESC, APC, OPC, F_NSC). Among the 90 genes, there were 17 genes expressed in F_NSC (highlighted in Table 2) that are identified as common stem cell markers both for ES_NSC and F_NSC.


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

  		Table 2. Genes expressed in hESC-derived NSC without detection in oligodendrocyte precursor cells or astrocyte precursor cells

 
Difference Between Origin of Neural Stem Cells (ES_NSC vs. F_NSC)
From the previous section, we could observe that NSC derived from ESC possessed a different global expression pattern than neurosphere-forming cells isolated from fetal tissue. For an in-depth comparison, we compared four samples of ES_NSC (N6–N9) with F_NSC (F1–F2). The principle component analysis showed clear separation between ES_NSC and F_NSC (Fig. 4A). There were 2,041 genes that had more than a twofold change in expression levels and were significantly differentially expressed (p < .01) between ES_NSC and F_NSC samples identified by analysis of variance. Of them, 817 had relatively higher expression in ES_NSC and 1,224 had lower expression in ES_NSC. The heat map analysis of relative expression values for these genes in ES_NSC and F_NSC samples is shown in Figure 4B. Furthermore, we conducted PAM to identify genes that can serve as biological markers to distinguish ES_NSC from F_NSC samples. The top 15 genes identified by PAM are shown in Figure 4C. Three of these genes (H2AFX, BLMH, and CDC 20) had higher expression in ES_NSC samples, whereas the others had higher expression in F_NSC samples.


Figure 4
View larger version (44K):
[in this window]
[in a new window]

  		Figure 4. Differences observed between ES_NSC and F_NSC. Principal component analysis plot of ES_NSC (blue) and F_NSC samples (green) (A). Heat map analysis of relative expression levels of genes differentially expressed in ES_NSC and F_NSC samples identified by analysis of variance. Genes with higher expression than the mean appear in red, whereas those lower appear in green. If the level is equal to the mean, it appears in black. The samples are: 1–N6; 2–N6a; 3–N7; 4–N7a; 5–N8; 6–N8a; 7–N9; 8–N9a; 9–F1; 10–F1a; 11–F2; 12–F2a (B). The top 15 marker genes identified by prediction analysis of microarray distinguish ES_NSC samples from F_NSC samples (C). The gene names (or RefSeq ID) are indicated above the gene expression profile. ES_NSC samples appear in red. F_NSC samples appear in green. The y-axis indicates the normalized expression level. Abbreviations: ES_NSC, hESC-derived NSC; F_NSC, fetal-derived NSC.

 
Analysis of Signal Transduction Pathways in ES_NSC
Next, we analyzed gene expression based on a known signaling pathway. The Fisher's exact test used in this study showed that the genes involved in biological processes such as WNT signaling, FGF signaling, and cell proliferation were significantly enriched (p < .05) in genes upregulated in ES_NSC cells, whereas genes for the MAPKKK cascade, apoptosis, and Janus tyrosine kinase (JAK)-signal transducer and activator of transcription (STAT) cascade were significantly downregulated (Table 3).


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

  		Table 3. Change of gene related in biological process in hESC-derived NSC compared with fetal-derived NSC

 
Data from gene expression analysis of ES_NSC (N1–N10) and F_NSC (1–4) (supplemental online Table 3) indicated, for instance, that LIF was active in fetal-derived and long-term cultured hESC-derived NSC [10, 21]. Furthermore, fetal-derived NSC showed higher gene expression values for LIF (890 vs. 92), IL11 (1368 vs. 67.8), and receptor LIFR (77.9 vs. 22.4) compared with ES_NSC. Accordingly, the expression of downstream genes such as JAK1, STAT1, STAT3, and SOCS2 was also higher in F_NSC compared with ES_NSC. These results would indicate that the LIF pathway is more active in F_NSC than in ES_NSC.

The Illumina microarray used in this study also included 20 subtypes of the FGF family. Among these, FGF9, FGF13, and FGF18 were shown to be expressed in ES_NSC and/or F_NSC. Among the four types of known FGF receptors in this group, three were shown to be highly expressed in ES_NSC compared with that seen in F_NSC (supplemental online Table 3).

Wnt signaling has been implicated in various biological events such as cell fate determination and pattern formation during development. The Fisher's exact test used here showed that there are also significant changes in expression of components of the Wnt signaling pathway in ES_NSC compared with F_NSC. Among 18 WNT genes examined, nine were shown to be expressed in ES_NSC at all levels more than twofold higher than F_NSC. In addition, seven out of nine Frizzled WNT signaling receptors were expressed in ES_NSC. In contrast, the inhibitor molecule (an antagonist of Wnt signaling) DKK1 was depleted in ES_NSC compared with F_NSC (174.3 vs. 815.4), possibly indicating that Wnt signaling is more active in ES_NSC compared with F_NSC. Interestingly, downstream pathway members relating to self-renewal and cell cycle progression such as MYC, TP53, CCND2, CCND3, and CREBBP did not show a concomitant upregulation. Instead, genes related to cell differentiation and transcription regulatory molecules such as FRAT and LEF1 were highly expressed in ES_NSC, indicating that the active Wnt signaling may be more involved in the developmental process as opposed to cell proliferation.

Extracellular Matrix Expression in NSC
Among the specific genes for NSC, those associated with extracellular matrix (ECM) can be valuable as genetic markers and for selection or purification of NSC. Among the examined ECM molecules, six types of collagen were observed to be expressed in ES_NSC (supplemental online Table 4). Collagens represent a class of major basement molecules, and their distribution and dimension is known to be important in cell function. When we compared specific expression profiles of these genes between ES_NSC and F_NSC, we observed several clear differences. First, high levels of Col1A2 expression were observed across all samples tested, whereas the other five collagen subtypes were enriched in ES_NSC but not detected in F_NSC. Furthermore, neither Col2A1 nor Col4A6 expression was detected in the more differentiated progenitors OPC and APC. Laminin family gene expression was high in ES_NSC compared with that seen in F_NSC. Also, a similar differential expression pattern was seen when we examined the RELN (Reeln) family, whose function in neural stem cells has been reported [34].


   DISCUSSION
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
Disclosure of Potential...
 Acknowledgments
 References
 
Undifferentiated human ESC and embryoid bodies that differentiate from them have been analyzed by EST scan, MPSS, serial analysis of gene expression, and microarray [35–37]. Fetal tissue-derived neural stem cells (F_NSC) likewise have been analyzed using these methods [21, 22, 38]. These global expression assessments have highlighted certain critical requirements for generating a high-quality expression profile data set. First, samples must be tested prior to their expression analysis to ensure the quality of input data. Given lab-to-lab variability and allelic variability among human populations, it is important to compare lines from many labs to ensure the significance of interpretations gleaned from particular data sets. The data must be carefully curated and comparisons must be limited to data sets generated using the same assay format. The assay format must be sensitive, reproducible, and affordable and should not require a large amount of sample material. We have found that the Illumina BeadArray system fulfils these criteria and have used it to generate data sets for ten different ES_NSC populations obtained from six different hESC lines. Samples were derived in multiple locations using different techniques and by multiple investigators, and most samples were run in duplicate. These samples were compared with four F_NSC, which had been independently tested for their self-renewal ability and ability to differentiate into neurons, astrocytes, and oligodendrocytes and for the expression of key NSC markers. The entire data set is comprised of 32 samples and is available for download at (ftp://resources.invitrogen.com/pub/stemcells2007).

It is impossible to provide a detailed analysis of all results generated in this study, and in this manuscript we have focused on a few major points. Perhaps the most important finding is the limited overall similarity between NSC derived from hESC (ES_NSC) and those isolated from fetal tissue (F_NSC). Interestingly, although these samples differ in their stage of development, they do share many properties including several well known neural stem cell markers, as shown in Figure 1. However, recent studies have begun to illustrate the differences between ES_NSC and fetal-derived neurosphere-forming cells. Chung et al. compared mouse ESC-derived cells with ventral mesencephalon (VM) derived neural progenitors and their differentiation potential to dopaminergic neurons, a neural cell fate of VM that occurs in vivo [39]. They observed that ES_NSC appear to be able to differentiate into dopaminergic neurons, whereas neural progenitor cells isolated from the VM failed to differentiate down this neuronal lineage, suggesting they are unable to retain their endogenous differentiation potential in vitro. In addition, Colombo et al. compared mouse ES_NSC with fetal-derived NSC and showed several similarities and differences between them [40]. Clearly, gene expression profiles are not completely conserved between mouse and human [35]. However, some patterns can be observed in human, as we see similar profiles as described in mouse for a subset of genes. In the results presented here, ES_NSC express many of the same genes seen in F_NSC, but differences in expressed genes were also observed. For example, both NSC populations actively expressed members of the Wnt receptor family (Frizzled 1–5 and 8–9). However, the expression of downstream genes involved in Wnt signaling such as LEF1, SFRP2, and MYC was significantly different between the two NSC populations, suggesting a differential role for Wnt signaling. Also, expression of some specific Wnt subtypes was detected in ES_NSC with Wnt7b expression identified solely in ES_NSC (in contrast with that seen in mice). In the human model, the expression pattern of Wnt8b in the embryo has been localized to the neuroepithelium of developing brain including the hippocampus, a site where adult neural stem cells are believed to be located later in development [41]. The high expression of Wnt molecules, including Wnt8b, shown in ES_NSC compared with that seen in F_NSC may partially explain the broader differentiation potential of ES_NSC as compared with F_NSC. In addition, the expression of genes involved in the LIF signaling pathway was also observed to follow the same pattern. F_NSC expressed high levels of LIF in accordance with elevated JAK and STAT intracellular signaling components, whereas the expression of these genes in ES_NSC was at lower levels. In contrast, ES_NSC seem to be dependent on the FGF pathways as indicated by high levels of FGFR expression. This differential gene expression between the two NSC populations was also apparent when examining the differential expression of ECM molecules, which showed higher expression levels in ES_NSC when compared with F_NSC. Interestingly, we found that the collagen subtype expression profile in human ES_NSC and F_NSC was consistent with published mouse NSC expression profiles, although underlying functional relevance of this has yet to be determined. In our study, we have also used a bioinformatics analysis algorithm (PAM) to identify a set of key genes that can be used to discriminate ES_NSC from F_NSC, and although we show that NSC from different sources (ESC or fetal tissue) share commonalities, they can be distinguished from each other.

Comparing ES_NSC with more lineage-restricted progenitors (OPC and APC) also allowed us to identify several markers that could be used to categorize ES_NSC from other NSC progenitor populations. As shown in Table 2, 21 novel genes could be identified as specific biomarkers for ES_NSC, four of which also continued to be expressed in F_NSC. The examination of the EST scan of these four genes showed expression in brain for three genes. Furthermore, this global comparison made it possible to identify the receptor molecules GPR23, HRH3, and P2RY5 as well as other signaling pathway molecules GLI2, IL11RA, NKD2, PRKCG, and RBL1.

In summary, our initial efforts toward a global gene expression profiling of ES_NSC and F_NSC have yielded useful insights and allowed the identification of key regulatory pathways as well as novel genes that are likely to play a role in regulating NSC fate. We have analyzed the gene expression profiles of the 15 lines of ES_NSC and F_NSC and, although samples were generated through different methods in multiple labs, distinct ES_NSC expression profiles could be discerned not only from more lineage-restricted progenitors but also from F_NSC. The overall comparison of ES_NSC and F_NSC populations with the gene expression of more restricted progenitors indicated that genes shared by both ES_NSC and F_NSC, not by OPC or APC, could be identified and may serve as the useful biomarkers of NSC. Although ES_NSC shared the expression of many of the genes expressed in F_NSC, they did differ in several key regulatory pathways such as LIF, FGF, and Wnt that could impact cell fate and specification.

The data sets obtained in this study, in addition to the rapidly expanding data from many other studies, should serve as an important resource and can be mined with currently available tools. Our efforts to provide this data in a readily accessible format will also allow even uninitiated researchers to be able to examine the pattern of expression of their favorite genes and compare them across multiple sample sets.


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


   ACKNOWLEDGMENTS
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 Disclosure of Potential...
 Acknowledgments
References
 
This work was supported by NIH, CNS foundation, ALS-Packard foundation, and Illumina Inc. We thank all members of our laboratories for constant stimulating discussions. M.S.R. acknowledges the contributions of Dr. S. Rao that made undertaking this project possible.


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

  1. Thomson JA, Itskovitz-Eldor J, Shapiro SS et al. Embryonic stem cell lines derived from human blastocysts. Science 1998;282:1145–1147.[Abstract/Free Full Text]

  2. Mitalipova M, Calhoun J, Shin S et al. Human embryonic stem cell lines derived from discarded embryos. STEM CELLS 2003;21:521–526.[Abstract/Free Full Text]

  3. Smukler SR, Runciman SB, Xu S et al. Embryonic stem cells assume a primitive neural stem cell fate in the absence of extrinsic influences. J Cell Biol 2006;172:79–90.[Abstract/Free Full Text]

  4. Vallier L, Reynolds D, Pedersen RA. Nodal inhibits differentiation of human embryonic stem cells along the neuroectodermal default pathway. Dev Biol 2004;275:403–421.[CrossRef][Medline]

  5. Wichterle H, Lieberam I, Porter JA et al. Directed differentiation of embryonic stem cells into motor neurons. Cell 2002;110:385–397.[CrossRef][Medline]

  6. Okabe S, Forsberg-Nilsson K, Spiro AC et al. Development of neuronal precursor cells and functional postmitotic neurons from embryonic stem cells in vitro. Mech Dev 1996;59:89–102.[CrossRef][Medline]

  7. Zhang SC, Wernig M, Duncan ID et al. In vitro differentiation of transplantable neural precursors from human embryonic stem cells. Nat Biotechnol 2001;19:1129–1133.[CrossRef][Medline]

  8. Uchida N, Buck DW, He D et al. Direct isolation of human central nervous system stem cells. Proc Natl Acad Sci U S A 2000;97:14720–14725.[Abstract/Free Full Text]

  9. Tropepe V, Hitoshi S, Sirard C et al. Direct neural fate specification from embryonic stem cells: A primitive mammalian neural stem cell stage acquired through a default mechanism. Neuron 2001;30:65–78.[CrossRef][Medline]

  10. Shin S, Mitalipova M, Noggle S et al. Long-term proliferation of human embryonic stem cell-derived neuroepithelial cells using defined adherent culture conditions. STEM CELLS 2006;24:125–138.[Abstract/Free Full Text]

  11. Ying QL, Stavridis M, Griffiths D et al. Conversion of embryonic stem cells into neuroectodermal precursors in adherent monoculture. Nat Biotechnol 2003;21:183–186.[CrossRef][Medline]

  12. Watanabe K, Kamiya D, Nishiyama A et al. Directed differentiation of telencephalic precursors from embryonic stem cells. Nat Neurosci 2005;8:288–296.[CrossRef][Medline]

  13. Pera MF, Andrade J, Houssami S et al. Regulation of human embryonic stem cell differentiation by BMP-2 and its antagonist noggin. J Cell Sci 2004;117:1269–1280.[Abstract/Free Full Text]

  14. Ben-Hur T, Idelson M, Khaner H et al. Transplantation of human embryonic stem cell-derived neural progenitors improves behavioral deficit in Parkinsonian rats. STEM CELLS 2004;22:1246–1255.[Abstract/Free Full Text]

  15. Abney ER, Bartlett PP, Raff MC. Astrocytes, ependymal cells, and oligodendrocytes develop on schedule in dissociated cell cultures of embryonic rat brain. Dev Biol 1981;83:301–310.[CrossRef][Medline]

  16. Frederiksen K, McKay RD. Proliferation and differentiation of rat neuroepithelial precursor cells in vivo. J Neurosci 1988;8:1144–1151.[Abstract]

  17. Mujtaba T, Han SS, Fischer I et al. Stable expression of the alkaline phosphatase marker gene by neural cells in culture and after transplantation into the CNS using cells derived from a transgenic rat. Exp Neurol 2002;174:48–57.[CrossRef][Medline]

  18. Cai J, Wu Y, Mirua T et al. Properties of a fetal multipotent neural stem cell (NEP cell). Dev Biol 2002;251:221–240.[CrossRef][Medline]

  19. Roy NS, Benraiss A, Wang S et al. Promoter-targeted selection and isolation of neural progenitor cells from the adult human ventricular zone. J Neurosci Res 2000;59:321–331.[CrossRef][Medline]

  20. Svendsen CN, Caldwell MA, Shen J et al. Long-term survival of human central nervous system progenitor cells transplanted into a rat model of Parkinson's disease. Exp Neurol 1997;148:135–146.[CrossRef][Medline]

  21. Wright LS, Li J, Caldwell MA et al. Gene expression in human neural stem cells: Effects of leukemia inhibitory factor. J Neurochem 2003;86:179–195.[CrossRef][Medline]

  22. Cai J, Shin S, Wright L et al. Massively parallel signature sequencing profiling of fetal human neural precursor cells. Stem Cells Dev 2006;15:232–244.[CrossRef][Medline]

  23. Pevny L, Rao MS. The stem-cell menagerie. Trends Neurosci 2003;26:351–359.[CrossRef][Medline]

  24. Mayer-Proschel M, Kalyani AJ, Mujtaba T et al. Isolation of lineage-restricted neuronal precursors from multipotent neuroepithelial stem cells. Neuron 1997;19:773–785.[CrossRef][Medline]

  25. Liu Y, Han SS, Wu Y et al. CD44 expression identifies astrocyte-restricted precursor cells. Dev Biol 2004;276:31–46.[CrossRef][Medline]

  26. Windrem MS, Nunes MC, Rashbaum WK et al. Fetal and adult human oligodendrocyte progenitor cell isolates myelinate the congenitally dysmyelinated brain. Nat Med 2004;10:93–97.[CrossRef][Medline]

  27. Carpenter MK, Inokuma MS, Denham J et al. Enrichment of neurons and neural precursors from human embryonic stem cells. Exp Neurol 2001;172:383–397.[CrossRef][Medline]

  28. Mujtaba T, Rao MS. Isolation of lineage-restricted neural precursors from cultured ES cells. Methods Mol Biol 2002;185:189–204.[Medline]

  29. Gunderson KL, Kruglyak S, Graige MS et al. Decoding randomly ordered DNA arrays. Genome Res 2004;14:870–877.[Abstract/Free Full Text]

  30. Zeeberg BR, Feng W, Wang G et al. GoMiner: a resource for biological interpretation of genomic and proteomic data. Genome Biol 2003;4:R28.[CrossRef][Medline]

  31. Tibshirani R, Hastie T, Narasimhan B et al. Diagnosis of multiple cancer types by shrunken centroids of gene expression. Proc Natl Acad Sci U S A 2002;99:6567–6572.[Abstract/Free Full Text]

  32. Liu Y, Shin S, Zeng X et al. Genome wide profiling of hESC, EC and their progenies using bead arrays to develop a base profile of gene expression of pre-August 9th human ES cell lines. BMC Dev Biol 2006;6:20.[CrossRef][Medline]

  33. Tropak MB, Johnson PW, Dunn RJ et al. Differential splicing of MAG transcripts during CNS and PNS development. Brain Res 1988;464:143–155.[Medline]

  34. Kim HM, Qu T, Kriho V et al. Reelin function in neural stem cell biology. Proc Natl Acad Sci U S A 2002;99:4020–4025.[Abstract/Free Full Text]

  35. Zhan M, Miura T, Xu X et al. Conservation and variation of gene regulation in embryonic stem cells assessed by comparative genomics. Cell Biochem Biophys 2005;43:379–405.[CrossRef][Medline]

  36. Richards M, Tan SP, Chan WK et al. Reverse serial analysis of gene expression (SAGE) characterization of orphan SAGE tags from human embryonic stem cells identifies the presence of novel transcripts and antisense transcription of key pluripotency genes. STEM CELLS 2006;24:1162–1173.[Abstract/Free Full Text]

  37. Miura T, Luo Y, Khrebtukova I et al. Monitoring early differentiation events in human embryonic stem cells by massively parallel signature sequencing and expressed sequence tag scan. Stem Cells Dev 2004;13:694–715.[CrossRef][Medline]

  38. Shin S, Rao MS. Large-scale analysis of neural stem cells and progenitor cells. Neurodegener Dis 2006;3:106–111.[CrossRef][Medline]

  39. Chung S, Shin BS, Hwang M et al. Neural precursors derived from embryonic stem cells, but not those from fetal ventral mesencephalon, maintain the potential to differentiate into dopaminergic neurons after expansion in vitro. STEM CELLS 2006;24:1583–1593.[Abstract/Free Full Text]

  40. Colombo E, Giannelli SG, Galli R et al. Embryonic stem-derived versus somatic neural stem cells: A comparative analysis of their developmental potential and molecular phenotype. STEM CELLS 2006;24:825–834.[Abstract/Free Full Text]

  41. Lako M, Lindsay S, Bullen P et al. A novel mammalian wnt gene, WNT8B, shows brain-restricted expression in early development, with sharply delimited expression boundaries in the developing forebrain. Hum Mol Genet 1998;7:813–822.[Abstract/Free Full Text]

  42. Reubinoff BE, Itsykson P, Turetsky T et al. Neural progenitors from human embryonic stem cells. Nat Biotechnol 2001;19:1134–1140.[CrossRef][Medline]




This article has been cited by other articles:


Home page
 Hum Mol GenetHome page
G. Esposito, J. Imitola, J. Lu, D. De Filippis, C. Scuderi, V. S. Ganesh, R. Folkerth, J. Hecht, S. Shin, T. Iuvone, et al.
Genomic and functional profiling of human Down syndrome neural progenitors implicates S100B and aquaporin 4 in cell injury
Hum. Mol. Genet., February 1, 2008; 17(3): 440 - 457.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental Data
Right arrow All Versions of this Article:
2006-0660v1
25/5/1298    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 HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Shin, S.
Right arrow Articles by Chesnut, J. D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Shin, S.
Right arrow Articles by Chesnut, J. D.

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