Systems biology discoveries using non-human primate pluripotent stem and germ cells: novel gene and genomic imprinting interactions as well as unique expression patterns

The study of pluripotent stem cells has generated much interest in both biology and medicine. Understanding the fundamentals of biological decisions, including what permits a cell to maintain pluripotency, that is, its ability to self-renew and thereby remain immortal, or to differentiate into multiple types of cells, is of profound importance. For clinical applications, pluripotent cells, including both embryonic stem cells and adult stem cells, have been proposed for cell replacement therapy for a number of human diseases and disorders, including Alzheimer's, Parkinson's, spinal cord injury and diabetes. One challenge in their usage for such therapies is understanding the mechanisms that allow the maintenance of pluripotency and controlling the specific differentiation into required functional target cells. Because of regulatory restrictions and biological feasibilities, there are many crucial investigations that are just impossible to perform using pluripotent stem cells (PSCs) from humans (for example, direct comparisons among panels of inbred embryonic stem cells from prime embryos obtained from pedigreed and fertile donors; genomic analysis of parent versus progeny PSCs and their identical differentiated tissues; intraspecific chimera analyses for pluripotency testing; and so on). However, PSCs from nonhuman primates are being investigated to bridge these knowledge gaps between discoveries in mice and vital information necessary for appropriate clinical evaluations. In this review, we consider the mRNAs and novel genes with unique expression and imprinting patterns that were discovered using systems biology approaches with primate pluripotent stem and germ cells.

the gene regulatory networks in these cells will enhance our knowledge of the regulation of the earliest stages of development.
Currently, we lack much information on specifi c stages in development. As described, all tissues of the embryo arise from the same cells; however, they are diff erent from each other, not only in their morphology and function, but also in their total DNA content. While somatic cells are diploid, gametes are only haploid. We have almost no knowledge on how these initial cells are selected and diff erentiated.
Another area that would benefi t from the delineation of pluripotent gene networks is the better understanding of reproductive mechanisms. While the sperm meets the egg to form the zygote, what gene expression in the egg allows the fi rst stages of development to proceed? Since in vitro fertilization (IVF) and other assisted reproductive technologies are so commonly used, can better understanding of these fi rst stages increase the effi ciency of these technologies?
However, while ESC lines contribute to a chimerawhich is when cells injected into a new blastocyst contribute to all tissues in the newborn -epiblast lines do not. We have no understanding of the mechanisms that underline the diff erence between these two types of cells.
ESCs have been studied not only for cell replacement therapy or basic stem cell biology, but also as a tool for development of better and safer drugs. Since the ESC lines resemble, in many aspects, the developing fetus, they can be used as a fi rst and quick tool for drug screening without exposing pregnant mothers and their babies to harmful drugs. Th erefore, better understanding of gene regulatory networks that control these cells or allow them to diff erentiate under specifi c signals will allow the development of new therapies. Th ese therapies will be based on the fi nding of new targets and hence the development of treatments specifi cally for those targets and the processes they control.
Certain cancers have been suggested to have a stem cell origin. Treatment today is usually directed to the amplifying cell rather than to the source of the cancer. Hence, understanding of the gene networks that have changed from those in stem cells and have led to cancer will allow the development of new treatments for the cancer, and as described before, the development of specifi cally targeted molecules for these pathologies.
To summarize, understanding of the gene regulatory networks that enable a cell to maintain its pluripotent phenotype are of great interest today. Better understanding of these networks will lead to better understanding of basic biology questions, the control of the specifi c diff erentiation of stem cells into target cells for cell replacement therapy, and the development of new drugs and treatments for cancer, among other diseases. Th is review summarizes our current knowledge of gene expres sion networks in non-human primates, which resem bles the human model and also has greater advantages.

Embryonic stem cells
ESCs are defi ned as a population of cells capable of selfrenewing while maintaining their pluripotency. ESCs diff erentiate and give rise to cells from all three germ layers: ectoderm, mesoderm and endoderm, including the pancreas [16]. Th ese cells are derived from the inner cell mass of mammalian blastocyst stage embryos [17,18]. While mESCs were cloned over three decades ago, we have only recently celebrated our fi rst decade of hESC derivation. Hence, our knowledge on hESCs is more limited than that on mESCs. While the two types of ESC share many features, such as the expression of the pluripotent marker OCT-4, they also diff er from each other, such as the dependency of mESCs on leukemia inhibitory factor (LIF) to maintain pluripotency. Presumably, therefore, pathways that contribute to both pluripotency as well as specifi c diff erentiation might diff er between mESCs and hESCs. Th us, our knowledge on gene expression in mESCs as well as their utilization for therapy must be verifi ed in hESCs.

Nuclear transfer
While ESCs can be diff erentiated into beta cells, utilizing them in human therapy poses a problem since they do not identically match the patient. Human ESC lines today were derived from 'leftover' blastocysts from IVF clinics. Hence, they might cause an immune reaction when trans planted into human patients. One possibility to overcome this problem is the generation of genetically similar ESCs to the patient. Th is process would require the use of the patient's genome. While routinely carried out in mice, this process has been found to be more diffi cult in other species. Nuclear transfer (NT; also termed therapeutic cloning) requires the enucleation of a donor egg and the removal of the spindle DNA. Into this enucleated egg a single diff erentiated cell, usually a fi broblast, is placed. Th e two cells fuse together, resulting in a single cell carrying the correct number of chromosomes. Th is cell then begins its development and will cleave until a blastocyst is formed, which contains an inner cell mass from which ESCs can be derived (termed NT-ESCs). For this process to succeed, a large number of processes are carried out within the cell, including reprogramming of the genome to an undiff erentiated state. It is this process that lowers the effi ciency of NTderived blastocysts. Much is unknown about the results of reprogramming, and many theories have been developed, which are summarized by Yang and colleagues [19].
While the best known successful NT was Dolly the Sheep, derived by Sir Ian Wilmut [20], NT has been successful in many species, including dog, cat, mouse, cow, goat and others [21][22][23]. Recently, primate cells have been cloned, though with low effi ciency [24]. Th ese results support the contention that human cells could also be generated by NT, though very signifi cant bioethical challenges remain. Th e reasons for the low effi ciency of primate cloning are not clear, and better methods for cloning are being investigated.
A recent study showed that it would be diffi cult to carry out intraspecies cloning [25]. In this method, enucleated animal eggs would be used to house the diff erentiated cells, thereby overcoming one of the major Genes that were diff erentially expressed in both of two previous studies [1,117] could be identifi ed in this representative pathway. For this analysis we compared all genes with a known Entrez gene ID.
hurdles of this process: the lack of suffi cient numbers of donor eggs. However, the control experiments in this study were lacking. While the authors were able to demon strate that there was no development in the cloned intraspecies cells, they did not show their ability to clone human cells, raising questions about the effi ciency of cloning in this study [25].
ESCs and NT-ESCs have been proposed as possible treatments for diseases such as Parkinson's, Alzheimer's, juvenile diabetes and others. Specifi cally, NT-ESCs as cell replacement therapy -that is, using NT to derive ESCshave a great advantage when compared to fertilized ESCs. Patients could provide their own fi broblasts for the derivation of NT-ESCs, and the resulting line would be patient-specifi c. However, since this process uses donated eggs that contain mitochondrial DNA, these cells are not identical to the donor patient [26]. Th erefore, the newly derived ESCs are mostly genetically similar to the donor, and hence might be rejected when fully diff eren tiated cells are transplanted. In contrast, many patients could not (for gender reasons) and/or should not (for ethical reasons) [26,27] donate eggs for the derivation of fertilized ESCs, and even then the newly derived cells would only be similar but not identical to the fi broblast from which they were derived, while fertilized ESCs can be closely related to the donor only if the egg donor is the patient herself or a fi rst degree relative.
NT-ESCs and fertilized ESCs have been shown to have many similar properties, including the ability to diff erentiate into cells from all three germ layers [28]. Since most diff erentiation procedures used for NT-ESCs will follow the same methods from fertilized ESCs, much emphasis is placed on generat ing NT-ESCs more effi ciently, and other studies are carried out to improve the diff erentiation procedures. As proof of principle, a number of publications have shown that NT-ESCs can be used for therapy. Th is was achieved in a Parkinson's model [29] as well as a diabetic model in which NT-ESCs were diff erentiated into beta cells [30], thereby bridging the gap between the two routes of research. Th e method of diff er entiation for these NT-ESCs was based on protocols described previously [30]. Th ese cells were able to maintain normal glucose levels after transplantion into a diabetic mouse [30]; however, an increase in blood glucose levels was seen 8 weeks after transplantation, presumably because the cells were not fully diff erentiated beta cells.

Induced pluripotent stem cells
One major goal of stem cell research is the generation of patient-specifi c stem cells. While successful in mice [31], the derivation of genetically matched, patient-specifi c human ESCs using somatic cell NT (SCNT) has not yet been accomplished. Furthermore, the use of donor oocytes or pre-implantation embryos to derive patient-specifi c stem cells using techniques such as SCNT, cell fusion and parthenogenesis elicits ethical concerns. Th e recent advance in reprogramming adult somatic cells into ESlike cells, termed induced pluripotent stem (iPS) cells, provides another avenue for generating patient-specifi c stem cells without the ethical concerns of other methodologies. Th us, iPS cell derivation is the latest innovation for generating large pools of patient-specifi c stem cells that can be used to treat a wide range of human diseases.
In 2006, Takahashi and Yamanaka [32] demonstrated that mouse embryonic fi broblasts and adult tip fi broblasts could be reprogrammed into a pluripotent, ES-like state by transducing these cells with four transcription factors (Oct4, Sox2, Klf4 and c-Myc) along with a knockin Fbx15 neomycin-resistant reporter gene. After 2 weeks of culture, these mouse iPS cells exhibited similar characteristics to ESCs, such as alkaline phosphatase activity, expression of SSEA-1 and Nanog (two pluripotency markers), and the ability to diff erentiate into all three germ layers (endoderm, ectoderm and mesoderm) via in vitro diff erentiation or teratoma formation in immuno defi cient mice. While this ground-breaking research provided a proof-of-principle for the reprogram ming of adult somatic cells to ES-like cells, these iPS cells diff ered from ESCs in the genomic expression of several genes and the inability of the iPS cells to fully chimerize with donor mouse embryos. Shortly after the publication of this work, three groups showed that dispens ing with the reactivation of the Fbx15 reporter gene generated iPS cells that yielded fully chimerized pups following blastocyst injection and could contribute to germ cell transmission [33][34][35].
Recently, this work has been extended to human cells, as three groups originally demonstrated that human iPS cells could be generated from embryonic, neonatal and adult fi broblasts [36][37][38]. Like ESCs, these human iPS cells exhibited alkaline phosphatase activity, expressed SSEA-3, TRA-1-60, Oct4 and Nanog (human pluripotency markers) from endogenous loci, and exhibited a genomic profi le more similar to ESCs than the originating fi broblasts. Interestingly, two diff erent combinations of retrovirally introduced transcription factors were used: Oct4, Sox2, Klf4 and c-Myc were utilized by two groups [36,37], while Yu and colleagues [38] generated human iPS cells using Oct4, Sox2, Nanog and Lin28, an RNA binding protein that regulates synthesis of the let-7 family of microRNAs (miRNAs). Currently, this work has been repeated in several other labs and has included reprogramming of more terminally diff erentiated cells, such as pancreatic beta cells [39][40][41][42].
Th e potential for iPS cells to be utilized therapeutically has recently been examined [43]. In this study, researchers were able to ameliorate a mouse sickle-cell anemia model by diff erentiating iPS cells into hematopoietic stem cells and re-introducing these cells back into the animal model [43]. Similarly, human iPS cells have been directly diff erentiated into motor neurons and insulin-secreting islet-like clusters in vitro [44,45]. While these results highlight the great promise for iPS cells in a clinical setting, transplanting iPS cells or diff erentiated iPS cells into humans carries a high risk. Around 20% of chimeric mice generated from iPS cells developed tumors within a 2-to 10-month time frame [34]. Th is fi nding is most likely due to the reactivation of c-Myc in iPS cells. While dispensing with c-Myc in iPS cell formation reduces the overall effi ciency of obtaining iPS cell colonies, chimeric mice derived from these cells appear to be free of tumors [46,47]. However, aberrant expression of Oct4, Sox2, Klf4 and/or Nanog has been observed in a number of human malignancies [48][49][50]. Likewise, the random integration of the retroviruses could induce tumorigenesis by activating other oncogenic factors [51]. Because of these concerns, current research has been targeted at deriving iPS cells in a more clinically-friendly manner. Such protocols include using adenoviruses instead of retroviruses, and using miRNAs or a combination of chemical and genetic modifi cations to induce reprogramming [52][53][54]. While iPS cell derivation is years away from being utilized in a clinical setting, the proofof-principal results shown thus far indicate that iPS cells have the promise of treating a wide range of human disorders without the concern of immuno-rejection. Also, without requiring donor oocytes or pre-implan tation embryos, iPS cell technology reduces the ethical concerns about generating patient-specifi c, pluripotent stem cells (PSCs).

Germ cells
Another potential source of stem cells is derived from precursor germ cells. In early embryonic development, a subset of pluripotent cells diff erentiate into primordial germ cells (PGCs) [55,56]. Th ese cells migrate, proliferate and colonize the genital ridge and represent a population of cells that will eventually further diff erentiate to form gametes. Initially discovered in mice, failure of PGCs to mitotically arrest following colonization leads to the formation of teratomas, tumors that contain cells representing all three germ layers: ectoderm, mesoderm, and endoderm [57]. Th e fi rst isolations and cultures of these proliferating PGCs yielded a multipotent cell line termed embryonal carcinoma cells. Th ese cells are capable of being diff erentiated in culture into various cell types, including neurons and cardiomyocytes [58,59]. It was also shown that culture of isolated PGCs prior to genital ridge colonization resulted in germ cell colonies that express numerous pluripotency markers akin to those of ESCs, such as OCT-4 [60,61]. Th ese unique cells, termed embryonic germ cells (EGCs) were shown to be highly pluripotent. EGCs have previously been an interesting cell source for studying gametogenesis in vitro because mouse EGCs appear to follow similar diff erentiation patterns as observed in in vivo gametogenesis [62]. However, ethical concerns about obtaining human EGCs have tamed interest in this fi eld.
Several groups have shown the ability of mouse, nonhuman primate and human ESCs to diff erentiate into germ cell lineages, specifi cally in vitro-derived PGCs (invPGCs) [63][64][65][66][67][68][69][70][71][72][73][74][75][76][77][78][79]. However, three groups in particular have demonstrated three diff erent methodologies for faithfully deriving invPGCs from ESCs at higher effi ciencies [63][64][65]. Yamauchi and colleagues [64] successfully diff erentiated cynomolgus monkey ESCs into germ cells by forming embryoid bodies (EBs) with retinoic acid and culturing these EBs for 28 days. At day 28, germ cells could be identifi ed by positive immunostaining for SSEA1, VASA or DAZL. Furthermore, these researchers showed up-regulation of germ cell gene expression for CXCR4, NANOS1, NANOS2, NANOS3, VASA, PIWIL1 and TEKT1 upon EB formation with retinoic acid for 28 days. Likewise, this group was able to demonstrate that day 28 EBs grown in retinoic acid or bone morphogenetic protein (BMP)-4 elevated expression of the meiotic marker SCP1 but not SCP3. Kee and colleagues [63] showed that adherent diff erentiation with a BMP cocktail (BMP4, BMP7 and BMP8b) induced diff eren tiation of hESCs into invPGCs in 7 to 14 days. Using a green fl uorescent protein (GFP) transgene driven by the VASA promoter, these researchers showed that diff eren tiation medium supplemented with BMPs resulted in increased expres sion of two PGC markers in diff eren tiating hESCs: VASA and DAZL. Kee and colleagues also demonstrated that VASA-GFP+ cells could be isolated and cultured on mouse embryonic fi broblasts for 7 days to form invPGC colonies. Th ese cultured cells also exhibited hypomethylation of the H19 locus, suggesting that these cells, like in vivo PGCs, undergo de-methy lation prior to gametogenic progression [63]. More importantly, Kee and colleagues demonstrated that over expression of DAZ family members (DAZ, DAZL and BOULE) in cultured invPGCs induces meiotic progres sion as determined by immunofl uorescence staining for SCP3 and γH2AX [63]. Even more striking, they demonstrated haploid formation by overexpression of the DAZL family members by the appearance of a small 1N peak in their propidium iodide FACS analysis and the expression of acrosin in a small fraction of cells. Th is remarkable discovery highlights the potential of driving gametogenesis in vitro from PSCs [63].
More recently, Amander Clark's group demonstrated a novel approach for rapidly and more effi ciently diff erentiating hESCs into PGCs. Park and colleagues [65] showed that diff erentiation of hESCs on human fetal gonadal stromal cells signifi cantly improved germ cell diff erentiation. Strikingly, these researchers showed that c-kit+/SSEA1+/VASA+ invPGCs (5% of the total popu lation of cells) could be isolated from diff erentiated hESCs as early as 3 days of culture on human fetal gonadal stromal cells. Similarly to Kee and colleagues, Park and colleagues demonstrated that invPGCs exhibit imprint erasures and show expression of a wide range of germ cell markers [63,65]. Th e work of Park and colleagues demonstrates progress towards a highly effi cient methodology for generating PGCs from ESCs in vitro [65]. Furthermore, Park and colleagues are the fi rst group to diff erentiate human iPS cells into early germ cell lineages. Th ese exciting results combined with the work of Kee and colleagues [63] and Yamauchi and colleagues [64] high light the similarities between in vivo PGCs and invPGCs illustrate the possibility of treating infertility by diff erentiating patient-matched ESCs into gametes or male germline stem cells for transplantation.
Th e ability to generate transplantable male germline stem cells or haploid gametes in culture has signifi cant therapeutic implications for couples with infertility [80,81]. Th e appeal of these approaches is enhanced by iPS cell and NT technologies, which would theoretically enable men to derive germline stem cells or sperm from their own skin cells in vitro. Th us, it is hypothetically possible for a man who is rendered infertile by toxic treatment for cancer (chemotherapy or radiation), and who did not cryopreserve semen prior to treatment, to father his own genetic children from germ cells derived from NT-ESCs or iPS cells. Th is potential can only be realized after extensive feasibility and safety studies are conducted, ideally in nonhuman primate models that are relevant to human physiology. Th ere is a lack of consensus among species regarding the potential of PGCs to undergo spermatogenesis when introduced into seminiferous tubules (mouse PGCs can [82] and rat PGCs cannot (K Orwig, unpublished)). However, there is con sen sus in rodents and several large animal species that gono cytes and spermatogonia from neonate, pup and adult testes undergo spermatogenesis when trans planted into the testes of infertile recipients [82][83][84][85][86][87]. Human PSCs can be diff erentiated into PGCs in the context of EBs [70,88] or adherent diff erentiation cultures [63,65,67]. Similarly, two groups have reported macaque PSC to PGC diff eren tiation in EBs [64,89]. Th ere are no reports of PSC to spermatogonial stem cell (SSC) diff erentiation, but several studies have reported PSC diff erentiation to haploid germ cells [63,77,78], suggesting a transient transition through an SSC-like intermediate. Th us, direct diff erentiation of PSCs to SSCs would provide a source of transplantable cells that could be used to ask important questions about the safety and effi cacy of PSC-derived cells.
Interestingly, the postnatal mammalian testis itself may provide an alternative source of PSCs that bypasses the need for an embryonic intermediate or genetic manipulation. Several groups have shown the ability of germ cells in the mouse postnatal testis to produce PSCs in vitro [90][91][92][93][94][95][96]. Several recent studies have also provided evidence for PSCs derived from the adult human testis [97][98][99][100]. Th ese cells arise in vitro from spermatogonia and can give rise to tissues of all three embryonic germ layers. Given that germ cells are responsible for initiating embryogenesis, it seems possible that germ cell factors could infl uence their ability to become pluripotent (for example, including expression of genes associated with pluripotency). Among the genes that are thought to form a core regulatory network in ESCs (OCT4, SOX2, and NANOG) [101], only OCT-4 is expressed by a few postnatal germ cells or cultured SSCs. Several reports have described a relatively small group of normal mouse spermatogonia that express OCT-4, including those in the adult testis, which could potentially be those that have the capacity to produce PSCs in vitro [102][103][104][105][106][107]. In human spermatogonia, though, only a few postnatal spermatogonia retain embryonic-expressed OCT-4, and this expression is lost after the fi rst few months of infant life except in pathological conditions [108,109]. In cultured SSCs, Oct-4 mRNA and protein can be detected, albeit at substantially lower levels than in ESCs [90,95,110,111], and this feature may be required for long-term survival of SSCs in culture [112]. Th us, the mechanisms that predispose spermatogonia (presumably SSCs) to acquire a pluripotent phenotype in a culture dish are unclear, but may involve similar gene expression features with other pluripotent cells (ESCs).

Gene expression in non-human primate ESCs
Th ere are great challenges working with human ESC lines rather than ESC lines from pedigreed animals. While working with human ESC lines involves some ethical as well as religious issues, use of non-human primate ESC (nhpESC) lines alleviates many of these concerns. Our two groups have recently derived a large number of nhpESC lines, including from rhesus macaques and baboon, using both fertilized blastocysts and NT [9,[113][114][115][116][117][118]. In addition, others have derived other monkey ESC lines, such as from marmoset [119]. Th ere are many other advantages in working with nhpESC compared to hESC lines: namely, the use of prime quality embryos as compared to left over human IVF embryos donated to science that usually either have genetic disorders [120] or lag in their development. In addition, since the donors of primate gametes are not anonymous, we have much information on the parents, as opposed to the anonymity of human donors. For these reasons we have closely examined the factors that defi ne a stem cell line, including genes expressed, and compared them to the literature on gene and mRNA expression of human ESCs.
Using the newly derived nhpESC lines, we examined their gene expression [9] and found that they are very similar (>97%) to each other. We next compared the gene expression of these lines to that of two types of fi broblast: skin fi broblasts from the parents from which gametes were taken to derive the lines; and fi broblasts derived from teratomas generated by the injection of nhpESCs into severe combined immunodefi ciency (SCID) mice [1]. We found a unique set of genes that is diff erentially expressed between the nhpESCs and the two types of fi broblast. Interestingly, many of these genes were membrane-bound proteins and receptors [1] (Table 1). We have also shown that there are chromosomes that show an overabundance of over-expressed genes, such as chromosomes 16, 19 and X, which correlate to human chromosomes 17, 20 and X, respectively.
Our second study [114] has shown that there is indeed a unique set of genes in nhpESCs that maintains pluripotency, and that these diff erentially expressed genes are involved in many pathways. As expected, both studies revealed the over-expression of the 'classic' stem cell factors OCT-4, Nanog and Sox-2 (Table 1). Interestingly, when imprinting was examined, the nhpESCs were found to have aberrant imprinting, the implications of which are unknown [121].
When the two studies are compared, as depicted in Table 1, many of the top diff erentially expressed genes are similar. Taking into account that these were two separate studies, the resemblance between the two gene lists is striking. While Ben-Yehudah and colleagues [1] compared the gene expression of nhpESCs to fi broblasts and generated a list of genes over-expressed in ESCs, Mitalipov and colleagues [117] generated a list of genes that were highly expressed in a number of nhpESC lines. Out of the top 25 genes over-expressed in nhpESCs in both studies, 6 (24%) were found on both lists. As expected, genes known to be involved in maintenance, such as OCT-4 and Nanog, can be found on both lists. Additionally, PTPRZ1 is found on both lists; this gene has been shown to be expressed by hESCs and is downregulated upon diff erentiation. Depletion of PTPRZ1 resulted in decreased colony formation and lower recovery of hESCs. However, the lists include genes that have yet to be associated with pluripotency, such as TACSTD1. TACSTD1, also called Ep-CAM, is an epithelial adhesion molecule that was originally identifi ed as a marker of carcinomas and is also expressed by rat SSCs [122][123][124]. We found this gene to be the most diff erentially expressed gene between stem cells and fi broblasts, indicating that it might have other functions in signaling rather than solely adhesion. In addition, it should be pointed out that 40% of the genes on the list in Ben-Yehudah and colleagues [1] are hypo thetical; these genes may also play a signifi cant role in pluripotency.
When we compiled the data in Ben-Yehudah and colleagues [1] using Ingenuity software to identify system networks responsible for the regulation of the pluripotent state in nhpESCs, we were able to create many gene networks. Some of these networks contained anticipated candidate genes, including SOX2, OCT-4 and NANOG, as we have shown previously [1]. In addition we could identify networks that have been shown to be diff eren tially expressed between stem cells and fi broblasts, as depicted in Table 1. Th ese genes participate in networks that have yet to be associated with pluripotency. Although most of the genes depicted in Figure 1 and Table 1 are unidentifi ed or have not been associated with pluripotency, some were found to play roles in regulating the transition from pluripotency to diff erentiation; for example, the gene TACSTD1 is included in both Figure 1 and Table 1.
Since ESCs can serve as a method of studying development [125], much research has been carried out to understand the mechanisms that underlie regulation of this specifi c process, such as the gene regulatory networks that control pluripotency. Th ese regulatory networks have been studied in mice [126] and have revealed the importance of key regulators of the pluripotent state, including OCT-4, Sox-2 and Nanog. A comprehensive review described similar fi ndings in humans [127,128] and has also been discussed by us [1]. It should be pointed out that although many genes have been implicated in the networks controlling pluripotency, little is known about the networks controlling this process. An exception is the OCT-4/Sox-2/Nanog network, which has been shown to be invaluable for maintaining pluripotency. In our hands, we could identify the pluripotent genes and networks [1,24], but could not fi t all the diff erentially expressed genes into these networks or form new ones.

Imprinted genes in nhpESCs
While genes involved in pluripotency can be identifi ed and even gene regulatory networks can be described, other mechanisms controlling the expression of genes in pluripotent cells can be established -for example, epigenetic mechanisms that control gene expression in pluripotent cells. One such epigenetic mechanism is DNA methylation, which is considered a key factor in the formation of cellular memory and identity [129]. A comprehensive review summarized the key features of the regulatory mechanisms that control the trans crip tional regulatory features in hESCs [130,131], which complemented their work with ChIP-chip in mESCs [132].
Th e rhesus monkey is the only primate in which SCNT has been successful so far [133]. Th erefore, this model can help answer questions on the epigenetic state of cells that undergo reprogramming -for example, whether they are closer to ESCs or to the somatic cell from which they originate. Th e answer to this question might shed light on why it is very diffi cult for primate cells to undergo NT compared to mice and other animals. Th is could lead to improvements in primate NT.
We have recently compared DNA methylation in native ESCs, fi broblasts, and ESCs generated by SCNT [129]. We wished to examine if the SCNT cells undergo changes in methylation state that would resemble a stem cell rather than a somatic cell. We have identifi ed and compared epigenome programming and reprogramming. Based on our previous knowledge, we have characterized hundreds of regions that are hyper-or hypomethylated in fi broblasts compared to native ESCs. We found that these regions are conserved in human cells and tissues. When ESCs were compared to the SCNT cells, we found to our surprise that the vast majority of these regions were reprogrammed in SCNT ESCs. Th e meaning of these phenomena is that these cells do indeed undergo reprogramming of their DNA methylation during SCNT. Th is reprogramming leads to an almost perfect corre lation between the epigenomic profi les of the native (ESC) and reprogrammed (NTSC) lines.
We also found that at least 58% of these changes are correlated in cis to transcription changes, Polycomb repressive complex-2 occupancy, or binding by the CTCF insulator [129].
As expected, since the process of adding or removing a methyl group from the DNA must be a complex process, we found that while epigenomic repro gramming is extensive and globally accurate, the effi ciency of adding and stripping DNA methylation during reprogramming is regionally variable. In several cases, this variability results in regions that remain methylated in a fi broblast-like pattern even after reprogramming [129].

Small RNAs and other RNAs
We have carried out many systems analyses using Ingenuity, as described in Figure 1. Th ese usually identifi ed networks not directly associated with pluripotency. However, we could occasionally identify a connection between genes associated with pluripotency and unique genes. One example is depicted in Figure 2. While Sox2 and Nanog are expressed in pluripotent cells (red), they are also associated with the gene NCRNA00094. Th is gene has been shown previously to be a non-coding RNA with unknown activity that is expressed in ESCs [134]. NCRNA00094 is an example of a large number of noncoding RNAs that might play a role in maintaining pluripotency. One subtype of non-coding RNAs that has been shown to participate in this process is the miRNAs. miRNAs are short non-coding RNA sequences that control gene expression by inhibiting the translation of specifi c mRNAs or causing their degradation [135,136]. Hence, several studies have been carried out to fi nd miRNAs involved in the maintenance of pluripotency in ESCs. Using mathematical and statistical tools, a recent study [137] identifi ed miRNAs that might be involved in pluripotency. A similar study was carried out in mESCs [138,139] and hESCs [140][141][142]. When diff erent types of human adult and ESCs are compared, a number of miRNAs seem to be involved in this process, such as miR302 [53,136,[142][143][144]. Specifi cally, one study showed Figure 2. Non-coding RNAs can be identifi ed as members of pluripotent pathways using Ingenuity. Ingenuity analysis shows that genes known to be involved in stemness (Nanog and Sox-2) are over-expressed (red) in non-human primate embryonic stem cells (nhpESCs) compared to fi broblasts. In addition, we could identify non-coding RNAs that were over-expressed in nhpESCs and associated with Nanog. For this analysis we compared all genes with a known Entrez gene ID. a connection between the 'stem cell factors' and miRNAs that inhibit them [145]. However, more work has to be done to identify their specifi c targets and actions, such as the crosstalk between stem cell factors and miRNAs, as in the case of Lin-28 and Let7, for example [139].
A comprehensive comparison of gene and RNA profi les of mouse fertilized and SCNT lines has been carried out recently [146]. Th ey found that the two types of ESCs have similar miRNA and protein expression profi les. Th ey conclude that this phenomenon is consistent with their similar developmental potentials and might result from their similar transcriptional profi les.
While much research has been conducted on miRNA involvement in pluripotency in hESCs and mESCs, this has been little studied in the monkey. A recent study by some of us [137] has computationally searched the rhesus genome to identify novel miRNAs involved in pluripotency by homology to human miRNAs. Th is study identifi ed 383 novel miRNAs: 173 have 100% homology to human miRNAs and 281 have >90% homology in the seed sequence of the miRNAs [137]. Th is study also identifi ed miRNAs that are involved in human ESC pluripotency, such as miR302, as described above.

Conclusions
In this review we have summarized our and other results from the past decade on the generation of nhpESCs, SCNT, the generation of iPS cells and our work on primordial germ cells. All these fi elds of research cumulatively enhance our understanding of the early stages of human development. Th ese exciting results, together with our results on gene expression in rhesus macaques and other primates, open the possibility of studying the gene expres sion and its control by miRNAs that results in the undiff erentiated state of ESCs. Moreover, these studies may lead to better understanding of the mecha nisms behind processes such as induced pluripotency, the knowledge of which could be used to test cellular therapies in nonhuman primates before introduction in humans.

Competing interests
The authors declare that they have no competing interests.