|Home | About | Journals | Submit | Contact Us | Français|
Christopher S. Navara: Provision of study material, Data analysis and interpretation,
Carrie J. Redinger: Collection and/or assembly of data,
Jocelyn D. Mich-Basso: Collection and/or assembly of data,
Carlos A. Castro: Collection and/or assembly of data,
Stacie Oliver: Collection and/or assembly of data
Lara J. Chensny: Administrative support, Collection and/or assembly of data
Thomas J. Richards: Data analysis and interpretation,
Naftali Kaminski: Conception and design, Data analysis and interpretation, manuscript writing,
Gerald Schatten: Conception and design, Data analysis and interpretation, manuscript writing, financial support, Final approval of manuscript.
While human embryonic stem cells (hESCs) are predisposed towards chromosomal aneploidities on 12, 17, 20 and X, rendering them susceptible to transformation, the specific genes expressed are not yet known. Here, by identifying the genes over expressed in pluripotent rhesus ESCs (nhpESCs) and comparing them to both their genetically-identical differentiated progeny (teratoma fibroblasts) as well as genetically-related differentiated parental cells (parental skin fibroblasts from whom gametes were used for ESC derivation), we find that some of those over expressed genes in nhpESCs cluster preferentially on rhesus chromosomes 16, 19, 20 and X, homologues of human chromosomes 17, 19, 16 and X respectively. Differentiated parental skin fibroblasts display gene expression profiles closer to nhpESC profiles than to teratoma cells, which are genetically identical to the pluripotent nhpESCs. Twenty over and under expressed pluripotency modulators, some implicated in neurogenesis, have been identified. The over expression of some of these genes discovered using pedigreed nhpESCs derived from prime embryos generated by fertile primates, which is impossible to perform with the anonymously donated clinically-discarded embryos from which hESCs are derived, independently confirms the importance of chromosome 17 and X regions in pluripotency and suggests specific candidates for targeting differentiation and transformation decisions.
Human embryonic stem cell (hESC) research is invaluable in discovering early developmental mechanisms and fundamental principles in pluripotency and differentiation. In addition, hESCs hold great promise for clinically relevant insights into the causes of diseases and perhaps even improved treatment strategies [1–7]. Notwithstanding the great significance of the current basic and translational investigations, the origins of, and strength of evidence for hESCs differ strikingly from their murine counterparts. Mouse ESCs were derived from blastocysts obtained from genetically pedigreed fertile mice [8–10]. The pluripotency of mESCs was unequivocally demonstrated by mESC contributions to mouse offspring, including germ-line transmission [9, 10] through the transfer of chimeric embryos generated by mixing mESCs with mouse morulae or blastocysts . In contrast, existing human ESCs were derived from embryos surplus to requirements for assisted reproduction [2, 3, 11, 12]. To ensure patient confidentiality, the established hESC lines are anonymous, so neither pedigree analysis nor investigations between or among closely related lines have been possible.
Pluripotency with hESCs has been demonstrated by detection of specific markers as well as in vitro and in vivo differentiation in teratomas; intra-specific chimera investigations, on the other hand, raise significant ethical concerns [12–15]. However, though hESCs are similar in their global characteristics, heterogeneity exists in gene expression among the various hESC lines . Regardless of these hESC limitations, vital insights into the particular genes expressed during pluripotency and the expression profile switches during differentiation are swiftly emerging with profound implications for induced pluripotency (iPS) cells [17–21]. Recently, patient-specific pluripotent stem cells (PSCs) have been established , providing the foundation for both in vitro clinical disease models as well as potentially immune matched lines for transplantation.
Recently, we established ten pedigreed non-human primate ESC (nhpESC) lines from prime embryos conceived using gametes from fertile breeders . In contrast to hESCs but similar to mESCs, this panel of nhpESCs are extremely homogenous in their gene expression profiles (>97% identity between and among lines). These pedigreed, partially inbred, well-characterized nhpESCs were then used to assess a more comprehensive list of stemness genes without the interference of unknown or heterogeneous genetic backgrounds. Due to our knowledge of the parental pedigrees of these ESCs, we were able to isolate skin fibroblast from the monkeys from which gametes were obtained for establishing these nhpESCs.
hESCs have been found to display predispositions towards aneploidities on chromosomes 12, 17, 20 and X [24–29], though the expressed genes are not yet known. These predispositions render hESCs susceptible to transformation, which jeopardizes their utility both as biological resources for in vitro investigations and undermines their clinical reliability in patients after transplantation. This investigation sought to identify the genes on the homologous non-human primate (NHP) chromosomes. Using a battery of ESCs from pedigreed fertile non-human primates, in this study we compared the gene expression profiles of these fully differentiated cells to nhpESCs as well as to fibroblasts isolated from teratomas derived from these nhpESC, i.e. genetically identical fully differentiated cells. We found that some of the genes over expressed clustered preferentially on rhesus chromosomes 16, 19, 20 and X, homologues of human chromosomes 17, 19, 16 and X respectively. The overabundance of over expressed genes on chromosomes 17 and X independently confirms the importance of these chromosome regions in pluripotency and suggests specific candidates for targeting transformation decisions.
The overall strategy to identify the genes overexpressed and underexpressed in pluripotent stem cells and, if clustered on chromosomes, to localize those chromosomes is depicted in Figure 1. Pluripotent pedigreed rhesus ESCs (female ESC line C3806 – black circle; male ESC line C3106 – black square) were compared to genetically identical differentiated progeny grown from their resultant teratomas (five male teratomas and three female teratomas: `T', i.e. TA31, TB31, TC31, TD31, TE31 and TA38, TB38, TC38 respectively) as well as to differentiated cells from their macaque parents (`M').
Stem cell pluripotency potentials were assessed by teratoma formation in SCID mice. Non-human primate ESCs (nhpESCs) were isolated by scraping colonies with good morphology and using a brief 5 minute treatment with 0.25% Trypsin/EDTA to break up large fragments. Cells were pelleted at 800 × g for 10 min, washed twice in sterile PBS, and approximately 5 × 105 - 5 × 106 cells were injected into the testis of 8–12 week old NOD-SCID mice (Jackson Labs) by modification of efferent duct injection . The injection pipette was advanced along the efferent duct and through the rete testes into the interstitial space where cells were injected (using Eppendorf Femtojet). The injection itself was monitored with Trypan blue added to the cell suspension. Tumor growth was determined by palpation and was typically detectable 10–12 weeks after injection. Teratomas were dissected from euthanized mice placed in tissue culture (for the isolation of outgrowths) or fixed (for pathological examination). For tissue culture (skin or teratoma), cells were minced into small fragments and maintained in fibroblast media (DMEM containing 10% FBS, penicillin/streptomycin, L-glutamine, nonessential amino acids) for 7 days before media replacement to insure attachment of cells. In all fibroblast experiments (both skin fibroblast as well as teratoma fibroblast) cells were cultured for >4 passages before RNA isolation. All RNA was isolated when cells were ~80% confluent to ensure that cells were proliferating, and therefore unaffected by contact inhibitor. For pathological examination, teratomas were fixed in 4% paraformaldehyde, paraffin embedded and examined histologically after Hematoxylin and Eosin (H&E) staining. Isolated tumors were identified as teratomas if tissues derived from all three germ layers were identified in the sections.
Cytogenetic analysis was performed as described previously .
Total RNA was extracted using the Trizol protocol (100μl /10×104 −105 cells)  and purified using the Qiagen RNeasy Micro Kit (Valencia, CA) according to the manufacturer's recommendations. RNA quantity and quality were determined using a Nanodrop spectrophotometer and Agilent Bioanalyzer .
One microgram of total RNA was used to start the manual target preparation using the Codelink Expression Bioarray System (Amersham Biosciences/GE Healthcare). Briefly, double stranded cDNA synthesis was performed with a T7oligo (dT) primer, followed by purification. This cDNA was used as a template for in vitro transcription with biotin labeled nucleotides. Fifteen micrograms of the labeled cRNA were hybridized to Affymetrix Rhesus Macaque genome 49 format Arrays (GeneChip® Rhesus Macaque Genome Array, cat# 900656, Affymetrix, Santa Clara, CA, USA), followed by washing and staining with Streptavidin Phytoerythrin (SAPE) as recommended by the manufacturer. Arrays were scanned on an Affymetrix GeneChip® 3000 Scanner. The arrays contain 52,866 probe sets that represent ~30,000 human orthologues and ESTs. Affymetrix GCOS software was used for the scanning of the probe arrays and the probe intensity analysis and normalization was performed using RMA express .
The gene expression analysis protocol has been previously described . Statistical analysis was performed using the Scoregene gene expression package (http://www.cs.huji.ac.il/labs/compbio/scoregenes), and data visualization was performed using Genomica (http://genomica.weizmann.ac.il) , Spotfire Decision Site 8.0 (Spotfire Inc. Göteborg, Sweden), Treeview (http://jtreeview.sourceforge.net) and BRB ArrayTools (http://linus.nci.nih.gov/BRB-ArrayTools.html). The RMA output for every gene was divided by the geometric mean of all the values for the same gene and was log based 2 transformed. In the analysis we only included transcripts with locus link numbers. To determine the differentially expressed genes we used t-test or Significance Analysis of Microarrays (SAM). In the clustering we included only genes that had a q-value=0 in both scoring methods. This criterion was also employed for assessing over and under abundant expression. FDR analysis was carried out as described.
Cells were collected and pelleted. Total RNA was extracted using Trizol reagent (Invitrogen) and treated with RNase-free kit (Ambion, Austin, TX). PCR reaction mixtures were prepared using an IMPROM- II reverse transcription system (Promega, Madison WI) following manufacturer's instructions. qPCR was performed using an ABI Prism 7700 (Applied Biosystems Incorporated, Foster City, CA). Taqman gene expression assays (Applied Biosystems) were used for NANOG, OCT-4 and β actin. Water and no RT samples were used as control. All samples were run in triplicate. TaqMan® Array human stem cell Pluripotency Panel (Applied Biosystems) was used following manufacturer's instructions. Real-time PCR expression data from 96 genes was analyzed. Six genes were endogenous controls, and 30 genes were not included in our analysis due to very poor amplification in any sample. mRNA fold changes were calculated using the −ΔΔCt method and normalized using β actin expression as endogenous control.
Paraffin imbedded tissue was deparaffinized and hydrated by incubating sections in three washes of xylene for 5 minutes each, followed by two washes of 100% ethanol for 10 minutes each, two washes of 95% ethanol for 10 minutes each and two washes in dH2O for 5 minutes each. Antigens were retrieved by boiling slides in 10 mM sodium citrate buffer pH 6.0, maintained at a sub-boiling temperature for 10 minutes and cooled for 30 minutes. Staining: sections were washed in dH2O three times for 5 minutes each, Incubated in 3% hydrogen peroxide for 10 minutes and washed in dH2O twice for 5 minutes each followed by washing buffer (1X TBS/0.1% Tween-20 (1X TBST)) for 5 minutes. Sections were blocked with 400 μl blocking solution (TBST/5% normal goat serum: to 5ml 1X TBST add 250 μl normal goat serum) for 1 hour at room temperature. Primary antibody was added (OCT4; cell signaling technologies) overnight at 4°C. Antibodies were removed and sections were washed in wash buffer three times for 5 minutes each. 400 μl biotinylated secondary antibody was added to each section and Incubated 30 minutes at room temperature. Sections were washed three times with wash buffer for 5 minutes each. DAB was added to each section to develop slides.
To overcome the genetic background heterogeneity which might mask some aspects of the actual stemness signature, we utilized a newly established bank of pedigreed nhpESCs  to generate a unique non-human primate “stemness gene” list generated by the comparison of nhpESC gene expression to two different sources of genetically-identical or genetically-related differentiated fibroblasts. The first were grown out of skin explants taken from the nonhuman primate parents from which gametes were used to generate the stem cell lines. The second were teratoma explants. The pedigree of the different types of cells is depicted in Figure 1A. For Figures 1B and 1C, RNA was isolated from all three types of cells: 1. Skin fibroblasts; these samples are indicated by the letter M (monkey) before the monkey identification number and sample label (a–f). 2. nhpESCs denoted by C (cell line) followed by the cell line and sample. 3. Fibroblast explants from teratomas generated by the injection of nhpESC into immune compromised mice. Since there were a number of teratomas for every stem cell line, we designated the nomenclature as follows: T (teratoma) - number of teratoma from the line - line name - sample. Therefore, TA31b denotes that this is the second sample of the first teratoma from line 3106.
The accuracy of findings in stem cell research is subject to considerations regarding the purity of the cellular populations. We verified the quality of our cellular populations in several ways. First, we showed that the nhpESC lines can generate teratomas that consist of tissues from all three germ layers (supplementary Figure 1). We also analyzed the teratoma sections for OCT-4 detection; the observations that they did not react with OCT-4 antibodies, while control seminomas did, demonstrates that they did not harbor pluripotent cells. Were they to contain pluripotent stem cells as a contaminant, our ability to detect pluripotent gene expression would have been compromised. We never identified OCT-4 positive cells within teratomas derived from line 3106, and very infrequently (< 1% and only in one minor cell type) could we find Oct-4 positive cells in teratomas derived from line 3806 (supplementary figure 2). Furthermore we examined cellular karyotypes, to ensure that the cells we investigated were euploid. Supplementary figures 3 (line 3106) and 4 (line 3806) show that both the nhpESC lines and the teratoma fibroblasts have normal karyotypes and that the teratoma karyotypes fit the nhpESC lines from which they have been derived. While the teratoma and nhpESC have an identical genome, it is only 50% identical to the parental skin fibroblast. We investigated both male (3106) and female (3806) nhpESC lines to eliminate genetic background, as well as to examine the effects of sex on the differences in gene expression.
After mRNA isolation and hybridization to Affymetrix rhesus arrays, we examined the gene expression profiles of these cells. A heat map (Figure 1B) was generated using cluster analysis and, as expected, the nhpESC displayed a highly different gene expression profile from both types of fibroblasts. When clustering the single samples using all ~53,000 probes on the array (Figure 1C), we found that the three cell types clustered together with their cellular subtype: pluripotent nhpESCs were significantly different from both types of differentiated fibroblasts. In addition, we found that the skin fibroblasts clustered together and separately from the teratoma-derived lines in two distinct nodes. These results indicate that although both were fibroblasts, they were not identical cells. In addition, Figure 1C shows the internal quality of our results since the three independent samples from each cell sample usually clustered together with its replicates. Out of the triplicates used for each experiment, we omitted one outlier of the ES samples (C3806c) since its gene expression differed extensively from all other nhpESC samples, as shown in the heat map illustrated in Supplementary Figure 5.
We next analyzed the genes that are significantly different between nhpESC and both types of fibroblasts. We found that out of the 28,895 annotations with an entrez gene ID, 8,252 annotations differ between the two groups. No significant changes were detected between male and female cells (data not shown) indicating that the differences were greater between cell types (nhpESC versus fibroblast) then between sexes. As expected, we found Oct-4, Nanog and Sox-2 that have been associated with pluripotency were among the 20 most up regulated genes in nhpESC compared to both types of fibroblasts (depicted in Table 1 and Figure 2A). However, many of the genes that were either up regulated or down regulated are still not characterized in the rhesus, and their study will lead to further insights into pluripotency characteristics. In addition, to our surprise, we found that 15% of the most up regulated genes were associated to neurogenesis (according to OMIM). No apoptosis-associated proteins were observed, and consequently it seems unlikely that contact inhibition or apoptosis were significant differences between the cell types.
We confirmed these differences in pluripotent markers between cell lines using quantitative PCR (qPCR; Figure 2). For these experiments, we used the same RNA extracted for the gene arrays for consistency. We used two samples from each cell type and averaged results of the samples are shown in Figure 2B. While Nanog was not detected in either group of fibroblasts, OCT-4 was expressed in low levels, and SOX2 was expressed in low levels in the teratoma fibroblasts; this is similar to the gene array expression differences between both types of fibroblasts (Figure 2A).
Using Ingenuity software we found that a large number of pathways were differentially expressed between stem cells and both types of fibroblast. Ingenuity constructs systems networks by analyzing our data together with its scientific literature-based software. Among the networks which are differentially expressed between pluripotent and differentiated cells are mitochondria dysfunction gene networks, as well as those involved in the G2/M transition (Supplementary Figure 6). It has been established that stem cells are better maintained in a low oxygen environment . In addition, stem cells appear to have less stringent cell cycle checkpoints. Both phenomena might be explained by the enriched pathways in stem cells compared to fibroblast depicted in Supplementary Figure 6.
To test whether genes over- or under-expressed in nhpESCs lie more--or less--often on certain macaque chromosomes than expected by chance we used chi-squared to test the goodness of fit (GOF) of the observed counts of differentially expressed genes on all 22 chromosomes to the expected counts based on the distribution among chromosomes of all probes on the chip with chromosomal attributions. Supplementary Table 1 shows a representative analysis of confidence values for the respective chromosomes using Chi squared `goodness of fit (GOF)' with a p- value of 10−4. Additional statistically analyses using other models and p- values were also performed. Having obtained statistically significant results with the omnibus GOF test, we used Bonferroni to control the familywise error rate at 0.05 in binomial proportions testing to detect over- or under-abundances of differentially expressed genes on chromosomes (Figure 3A). We found that there is an overabundance of over expressed genes on chromosomes 16, 19, 20 and X that correlate to the human chromosomes 17, 19, 16 and X respectively. Interestingly, of these 4 chromosomes 2 have previously been shown to undergo aneploidities in prolonged hESC tissue culture conditions. Rhesus chromosomes 6 and 12 were shown to be under expressed in our experiments (Figure 3B). These results are in agreement with previous published results showing chromosomal instability in rhesus preimplantation embryos [36, 37]. When global gene expression patterns were characterized, we found that the nhpESC gene expression was closer to skin fibroblasts rather than to the teratoma fibroblasts, although genetically they share their genome with teratoma fibroblasts and only half with skin fibroblasts.
We identified a stem cell signature that would define a pluripotent cell by its relative average gene expression for all genes. We examined the average log relative gene expression of both the 29,000 genes with a known gene Entrez gene ID (Figure 4A) as well as the 8,200 genes that differed between nhpESC and fibroblasts (Data not shown). We found that the overall relative gene expression of the fibroblast was indeed significantly higher then in the nhpESCs. Interestingly, the relative expression of the skin fibroblast was closer to the nhpESC than to the teratoma fibroblasts. Using BRB ArrayTools we could visualize a 3-D relationship among the 3 groups of cells illustrated in Figure 4B, in which nhpESC were closer in gene expression to skin fibroblasts then to teratoma fibroblasts. This 3-D analysis permits rotation of the datasets so that their inter-relationships can be viewed from various perspectives. These effects were independent of sex.
Finally, we have compiled system networks inferred to regulate the pluripotent differentiation transition in nonhuman primate stem cells (Figure 5). It is important to note that this list is generated totally independently of the list of `stemness genes' which have been identified in the more heterogeneous human ESCs. Anticipated candidate genes include SOX2, OCT-4 and NANOG, while unanticipated genes include up regulation of IL17. Interestingly, we found that many genes, such as members of LHX and DLX families of genes were down regulated in nhpESCs. Both gene families have been shown to be involved in fetal development [38–40].
Pluripotent stem cell (PSC) research, which emerged from the intensive investigations on embryonic stem cells, has tremendous importance for understanding fundamental concepts of early development, differentiation decisions and cell proliferation controls. Now with the intense focus on human PSCs, this previously fundamental science is rapidly moving towards more clinically relevant arenas. These fields include PSC differentiation for cell replacement, PSC progenitors of cancer, patient-specific and/or disease-specific induced PSCs for mechanistic and drug discoveries. These proposed medical implications of pluripotent stem expression patterns in human PSCs now demand increased scrutiny to assure that the discoveries are indeed accurate and unaffected by inherent biological and genetic limitations, i.e. the heterogeneity of the existing hESC lines; their sub-optimal origins from clinically-discarded embryos conceived in vitro by infertility patients; their genetic anonymity; their extensive propagation in vitro; among others. To address these concerns, we have conducted a completely independent analysis of `stemness' gene networks using pedigreed rhesus ESCs derived from fertile primates . Because these datasets do not rely on the existing hESC stemness gene expression patterns, it is noteworthy that the most prominent pluripotency genes, i.e. SOX2, OCT-4 and NANOG, emerge as dominant regulatory factors. In addition, unexplored candidate genes may be worthy targets for further research attention.
To insure that we did not have contaminations of differentiated cells within our undifferentiated cultures, or undifferentiated cells within our differentiated cultures, we routinely stained our undifferentiated cultures with “stemness markers”, i.e. OCT-4, NANOG, SSEA-3 and -4, as well as neuronal markers (data not shown). In addition, as depicted in supplementary figure 2, we could not find OCT-4 positive cells within the teratomas generated from line 3106, and only very rarely from those generated from line 3806. Since a possibility of “contamination” exists, a pure population of FACS sorted cells would be the ideal population for these experiments. However, since most nhpESC die in a single cell population, we refrained from FACS sorting. In addition, we grew both skin and teratoma fibroblasts for >4 passages before extracting their RNA to insure as pure a population of fibroblasts as possible.
When global gene expression patterns were characterized, we found that the nhpESC gene expression was closer to skin fibroblasts then to teratoma fibroblasts, although genetically they share their genome with the teratoma fibroblast and only half with skin fibroblasts. These results might indicate that the cells isolated from the skin and characterized as fibroblasts might harbor less well differentiated cells closer in their gene expression to authentic undifferentiated cells. Perhaps this result might shed light on the success rate of skin fibroblast to form iPS cells as compared to other types of cells [17–21].
Although most of the genes depicted in Table 1 are unidentified, some of the differentially expressed genes are associated to pluripotency, such as Oct-4 and Sox-2. In addition, others be found in future research to play roles in regulating the pluripotency <-> differentiation transitions. For example, TACSTD1, also called CD44 or Ep-CAM is an epithelial adhesion molecule that was originally identified as a marker of carcinomas[41, 42]. We found this gene to be the most differentially expressed gene between stem cells and fibroblasts, indicating that it might have other functions in signaling rather then solely adhesion. In contrast, decorin (DCN) has an important role in maintaining the balance in Peyronie's disease (PD) where fibrotic plaques are formed. DCN neutralizes one of the causes for plaque formation - transforming-growth-factor β1. Consequently it is perhaps not surprising that DCN was under expressed in the pluripotent stem cells we examined.
To identify “stemness genes” a comparison between gene expression of pluripotent cells to their in vitro differentiated counterparts was carried out [44, 45]. While genetic background is identical, the presence of some undifferentiated or partially differentiated cells can not be ruled out. In contrast, a range of undifferentiated cells can be used for the common denominator[s] that defines the pluripotent state [44, 46–51], or the comparison of gene expression between and among different stem cell lines. These experiments usually showed that hESCs are quite different from each other [16, 23, 52, 53]. A more comprehensive study showed that although closely related, the 59 ESC lines showed heterogeneity in gene expression . Interestingly, variations in gene expression were found not only in genes correlated with the pluripotent state or differentiation, but also in the expression of housekeeping genes . Therefore, interactions among many genes forming an active network that will allow the pluripotent state to be maintained is likely at work.
The first two studies that characterized the “stemness gene” list [49, 50] identified about 250 putative genes involved in mESC pluripotency, and many other genes are being studied today [16, 46–48]. Stemness gene lists have been generated by comparing different types of pluripotent cells  or by comparing differentiated cells to differentiated cells [47, 48, 58–62]. It is interesting to note that OCT-4 is on chromosome 6p21.33, SOX-2 3q26.3 and NANOG 12p13.31, so none of this important pluripotency triad are represented on the over expressed chromosomes found here. Further analysis is underway to identify these pluripotency regulatory clusters.
It has been well established that a number of key proteins play an important role in the maintenance of pluripotency both in mESCs and hESCs, such as OCT-4, NANOG and SOX2 [44, 45] as well as the existence of a regulatory network in mESC  for SOX2 [64, 65], OCT-4  and NANOG . Similar results were shown in hESCs since hESCs treated with RNAi against SOX2 cells readily differentiated . Although there is much information on the cooperation activity of OCT-4, NANOG and SOX2, we still lack information regarding other key players in the maintenance of the pluripotent state.
To better understand the extensive list of genes involved in the differences between cell lines, research has been conducted on specific pathways that are differentially expressed. Rho et al  found that the Wnt, Hh, Notch pathways but not the JAK/STAT had over-abundant transcripts in ESC compared to EBs. Therefore they concluded that self renewal is a coordinated signaling-specific mechanism. In addition, Li et al  identified a transcriptome involved in this differentiation process. Still, most genes that are differentially expressed have yet to be identified (ESTs) or have not been correlated with pluripotency . Studying them will enhance our understanding of the pluripotent state.
In order to achieve an orchestrated pathway in development, a well coordinated machinery must be activated to turn off pluripotent genes and turn on the expression of differentiation genes. This process is usually carried out by transcription factors as shown in mESCs  and hESCs . These transcription factors can together form a hierarchy in a complex network that maintains the pluripotent state  and include reprogramming factors used for iPS (OCT-4, SOX2, KLF-4 and c-MYC) as well as NANOG, DAX-1, REX-1, ZPF281 and NAC-1. Other DNA remodeling proteins have been associated with pluripotency in mESC  and both the similarities and differences between the mouse networks generated by Zhou et al and ours are informative. Both found genes not yet appreciated as regulators of pluripotency.
Recently, new mathematical algorithms have been developed to help identify pluripotent genes . Analyses using them showed that every hESC has its unique signature but shares many genes with other hESC lines that help maintain the pluripotent state[76–78]. However, that analysis was based on the comparison of three hESC lines that, although similar, still contain many differences.
While many studies have been carried out on mESC and hESC, very few have searched for pluripotent genes in nhpESC. A single study has compared rhesus ESC to EBs . They identified 367 genes that were expressed in 5 nhpESC lines and include CCNB1, GDNF3, LeftB, OCT-4 and NANOG. These 367 genes may represent the stemness core allowing the cells to maintain the pluripotent state.
The prominent stemness triad of SOX2/OCT-4/NANOG found in pluripotent human stem cells is the dominant regulatory network in primate PSCs. This confirmation is reassuring since it suggests that the hESC lines' heterogeneous origins – embryos discarded by anonymous infertility patients – do not compromise the fundamental utility of the existing lines. Needless to say, newer, more robust, and more diverse lines would accelerate research discoveries. Differences between gene expression profiles in human lines compared with primate ones may prove instrumental in discovering the genetic basis of certain forms of infertility which might have been transmitted to the resultant hESCs. Also, since the majority of these lines came from Indian rhesus monkeys, the sub-speciation of this group could be further analyzed. Finally, the preferred clustering of over expressed genes on the NHP homologues of human chromosomes 17, 19, 16 and X highlights the centrality of these regions to pluripotency and suggests that investigating the role of these genes may further enhance our understanding of the mechanisms that determine stemness and pluripotency.
We thank Angela Palermo and Andre Tartar for their technical assistance in reformatting the manuscript.
Funding: This work was supported by grant P01 HD 47675 from the National Institutes of Health
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.