The homeobox transcription factor Nanog has been proposed to play a crucial role in the maintenance of the undifferentiated state of murine embryonic stem cells. A human counterpart, NANOG, has been identified, but its function and localization have not hitherto been described. We have used a combination of RNA interference and quantitative real-time polymerase chain reaction to study NANOG in human embryonic stem and embryonic carcinoma cells. Transfection of NANOG-specific small interfering RNAs reduced levels of NANOG transcript and protein and induced activation of the extraembryonic endoderm-associated genes GATA4, GATA6, LAMININ B1, and AFP as well as upregulation of trophectoderm-associated genes CDX2, GATA2, hCG-alpha, and hCG-beta. Immunostaining of preimplantation human embryos showed that NANOG was expressed in the inner cell mass of expanded blastocysts but not in earlier-stage embryos, consistent with a role in the maintenance of pluripotency. Taken together, our findings suggest that NANOG acts as a gatekeeper of pluripotency in human embryonic stem and carcinoma cells by preventing their differentiation to extraembryonic endoderm and trophectoderm lineages.
Embryonic stem (ES) cells are able to differentiate into several different cell types, and, as such, they represent an excellent source of cells for continuous cell replacement therapies, drug discovery, and basic biology studies. A key to unlocking this potential is the understanding of the critical pathways and factors that are involved in the maintenance of pluripotency and self-renewal as well as differentiation toward specific lineages. Although the molecular basis of tissue differentiation in the mouse embryo is well appreciated , the molecular mechanisms of ES pluripotency are poorly understood. The LIF/STAT3 pathway is involved in ES cell renewal in the mouse, but its role is only facultative in vivo and it is not fundamental to human ES cell pluripotency [2–4], suggesting the existence of alternative pathways.
Screens for a critical factor that could maintain ES cell pluripotency independently of the LIF/STAT3 signaling pathway recently identified a new transcription factor, Nanog [5, 6]. A human counterpart, NANOG, has been identified and is also expressed in human ES and embryonic carcinoma (EC) cells . The human NANOG gene family contains a duplication pseudogene, NANOGP1, and 10 processed pseudogenes, whereas only two Nanog pseudogenes are evident in mouse; the high human pseudogene number may be attributable to comparatively strong expression in the germ line .
Nanog expression in the mouse embryo is restricted to the founder cells of the inner cell mass (ICM), and its absence from the ICM results in failure to generate epiblasts [5, 6]. Nanog expression is high in ES and EC cells and germ cell tumors  but is downregulated upon differentiation, concomitant with loss of pluripotency, and, at least in human EC cells, this has been partially explained by the sequential methylation of CpG residues in the promoter region of human NANOG and OCT4 genes . In the murine system, downregulation of Nanog expression has been partly attributed to suppression by p53, which was shown to bind to the Nanog promoter region .
Depletion of Nanog from murine ES cells results in loss of pluripotency, reduction in cell growth, and differentiation into extraembryonic endoderm lineage [6, 12]. It was suggested, therefore, that Nanog could act as a transcriptional repressor for proteins such as Gata4 and Gata6, which are expressed in primitive endoderm . However, it has been noted that elevated expression of Nanog resulted in clonal expansion of murine ES cells, maintenance of Oct4 expression, and resistance to differentiation induced in monolayers , perhaps by activating genes that are critical for the maintenance of pluripotency. In support of this observation, Pan and Pei  identified two unusually strong activation domains embedded in the C terminus of the protein, indicating that Nanog can act as a strong transcriptional activator.
Human and mouse ES cells show differences in morphology, cell-surface marker expression, signaling pathways, and differentiation ability, and thus one cannot assume a priori conservation in gene function between the human and mouse Nanog. In the present study, we have used a combination of small interfering (si) RNAs in human ES and EC cells with relative quantification analysis of data from quantitative real-time polymerase chain reaction (PCR) to identify the role of NANOG during early human embryonic development.
Materials and Methods
Human ES cells were maintained as previously described . Two fully characterized cell lines, hES-NCL1  and H1 (WiCell Inc., Madison, WI, http://www.wicell.org), and one EC cell line, TERA2.cl.SP-12 , were used in this study. Throughout this manuscript, we will refer to the EC cell line as NT2-SP12. EC cells were cultured in Dulbecco's modified Eagle's medium (DMEM) containing high glucose and 10% fetal bovine serum. Twenty-four hours before transfection with siRNA, EC cells were trypsinised and seeded onto 24-well plates at 4 × 104 cells per well, while human ES cells were plated on Matrigel in the presence of mouse embryonic fibroblast–conditioned medium as previously described . Bright field images of cell morphology were obtained using a Zeiss microscope and AxioVision software (Carl Zeiss, Jena, Germany, http://www.zeiss.com).
siRNAs and Transfection
siRNAs were obtained from Ambion (Austin, TX, http://www.ambion.com). The siRNAs were designed in accordance with Reynolds et al.  for maximum silencing efficiency. Transfection with a green fluorescent protein (GFP)-specific siRNA or a mutated version of the NANOG siRNA was used as negative control. The target mRNA sequences for the siRNAs were as follows: GFP: AAACUACCUGUUCCAUGGCCA (accession no. U87624, 174–194); NANOG: AACCAGACCUGGAACAAUUCA (accession no. NM_024865, 808–828 and AY455283, 453–473); and mutated NANOG: AACGAGACCAUGAACGAUUCA. Optimization of siRNA/Lipofectamine 2000 ratio was carried out as recommended by the manufacturer. EC and ES cells were transfected with 100 nM of the siRNA duplex using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) at a ratio of 1:2. siRNA-Lipofectamine 2000 complexes were prepared in Opti-MEM I reduced-serum medium. Cells were incubated with the complexes for 24 hours in mouse embryonic fibroblast–conditioned medium for the human ES cells and DMEM with 10% fetal bovine serum containing medium for the EC cells. Cells were transfected again at 24 hours. To assess transfection efficiencies, rhodamine-labeled NANOG, mutated NANOG, and GFP siRNAs were synthesized from Ambion and transfected in the cells as described above. Three fields of view (more than 100 cells) were counted to obtain quantitative data on transfection efficiencies.
Reverse Transcription–PCR Analysis
Total RNA was extracted using TRIzol Reagent (Invitrogen) according to the manufacturer's instructions for small-scale quantities of cells, and RNA concentration was assessed using a spectrophotometer (ND-1000; NanoDrop Technologies, Wilmington, DE, http://www.nanodrop.com). Genomic DNA was degraded by incubation with RQ1 RNase-Free DNase (Promega, Madison, WI, http://www.promega.com) for 30 minutes at 37°C. One microgram of total RNA was reverse-transcribed using Moloney murine leukemia virus reverse transcription (RT) and random hexamers following the manufacturer's instructions (Promega). PCRs were carried out using Taq DNA polymerase (Promega) and the primers shown in supplemental online Table 1. Negative controls samples, which lacked RT during the preparation of cDNA, were included to monitor genomic contamination. PCR products were resolved on 2% agarose gels, stained with ethidium bromide, and visualized in a transilluminator.
Real-Time RT-PCR Analysis for Telomerase RT Quantification
Real-time RT-PCR analysis for quantification of telomerase RT (TERT) was carried out using the LightCycler TeloTAGGG hTERT quantification kit according to the manufacturer's instructions (Roche Diagnostics, Indianapolis, http://www.roche-applied-science.com). For each reaction, 200 ng total RNA was used. All reactions were performed in duplicate and normalized to porphobilinogen deaminase in the same run.
LightCycler Real-Time PCR Analysis
Real-time PCR analysis was carried out using QuantiTect SYBR Green PCR Master mix (Qiagen, Valencia, CA, http://www1.qiagen.com/Default.aspx?). The primers used are shown in supplemental online Table 1. The reaction was performed with 1 μl cDNA per 20 μl reaction, and each reaction was performed in triplicate. The Light-Cycler experimental run protocol used was as follows: PCR activation step (95°C for 15 minutes), amplification with data acquisition repeated 50 times (94°C for 15 seconds, annealing temperature for primers for 30 seconds, 72°C for 20 seconds with a single fluorescence data collection), melting curve (60°C to 95°C with a temperature transition rate of 0.1°C per second and continuous fluorescence data collection), and finally cooling to 40°C. The crossing point (CP) for each transcript was determined using second derivative maximum method in the LightCycler software v3.5.3 (Roche Diagnostics). GAPDH CP for each sample was used as the internal control of these real-time analyses. The data were analyzed using the LightCycler relative quantification software v1.01 (Roche Diagnostics). For each gene, the value for the cells treated with control GFP siRNA was set to 1 (100%), and all other values were calculated with respect to this. Statistical significance of the results was assessed using Student's t-test (p<.05).
Lysates were electrophoresed on a 10% SDS-PAGE gel and electrophoretically transferred to a polyvinylidene difluoride membrane (Hybond-P; Amersham Biosciences, Piscataway, NJ, http://www4.amershambiosciences.com). Membranes were blocked in Tris-buffered saline with 5% milk and 0.1% Tween. The blots were probed with NANOG (1:1,000; R&D Systems, Minneapolis, http://www.rndsystems.com), GATA6 (1:200; Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com), or GAPDH antibody (1:2,000; Abcam, Cambridge, U.K., http://www.abcam.com) overnight and revealed with horseradish peroxidase–conjugated secondary antibodies, anti-goat (1:2,000; DakoCytomation, Glostrup, Denmark, http://www.dakocytomation.com) for NANOG, and anti-rabbit (1:20,000; Amersham Biosciences) for GAPDH and GATA6. Antibody/antigen complexes were detected using ECL reagent (Amersham Biosciences) and GeneSnap software (version 4.00.00; SYNGENE, Cambridge, U.K., http://www.syngene.com) with GeneGnome (SYNGENE). Bands were analyzed using the GeneTools software (version 3.00.22; SYNGENE).
Immunostaining and Fluorescence Microscopy
Cells were fixed with 4% paraformaldehyde for 10 minutes and permeabilized/blocked with 0.1% Triton X-100 and 5% fetal calf serum (FCS) in phosphate-buffered saline (PBS) at room temperature for 45 minutes. After blocking, the cells were incubated at room temperature for 1 hour with either NANOG (1:20; R&D Systems), GATA6 (1:100; Santa Cruz), CDX2 (1:50; Novocastra, Newcastle upon Tyne, U.K., http://www.novocastra.co.uk), or AFP (1:500; Novocastra) antibody. The cells were washed with 5% FCS and PBS before addition of the appropriate secondary antibody: rhodamine-conjugated anti-goat immunoglobulin G (IgG) (1:100; Jackson ImmunoResearch, West Grove, PA, http://www.jacksonimmuno.com) for NANOG, fluorescein isothiocyanate (FITC)–conjugated anti-rabbit immunoglobulins (1:100; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) for GATA6, and FITC-conjugated anti-mouse IgG (1:100; Sigma-Aldrich) for AFP and CDX2. The cells were washed again before fluorescence microscopy. For double staining, cells were fixed, permeabilized, and blocked as described before incubation with the NANOG antibody (R&D Systems, 1:20) for 1 hour. The cells were washed with 5% FCS and PBS and then incubated with the GATA6 antibody (1:100; Santa Cruz Biotechnology) for an additional hour. The cells were again washed with 5% FCS and PBS before addition of rhodamine-conjugated anti-goat IgG (1:100; Chemicon, Temecula, CA, http://www.chemicon.com) and FITC-conjugated anti-rabbit IgG (1 in 100; Chemicon) for 30 minutes. The cells were washed before fluorescence microscopy.
Immunostaining of Human Embryos
Spare human oocytes and embryos from in vitro fertilization treatment were collected in accordance with the criteria set by the Human Fertilisation and Embryology Authority and after patient consent. The zona pellucida of the embryos was removed using Acid Tyrodes solution (MediCult, Jyllinge, Denmark, http://www.medicult.com) before fixation with 4% paraformaldehyde for 10 minutes and permeabilization with PBS Tween-20 (PBST) (0.2% Triton X-100 in PBS) for 30 minutes. After blocking with 5% nonfat dried milk in PBST for 30 minutes, the embryos were incubated with 1:20 dilution of the NANOG antibody in 1% non-fat dried milk in PBST for 2 hours. The embryos were washed with 1% nonfat dried milk in PBST before addition of the secondary antibody, Rhodamine-conjugated anti-goat IgG (1:100; Jackson ImmunoResearch). The cells were washed again before fluorescence microscopy in 5% FCS in PBS. Due to scarcity of donated human material, only four embryos from each stage and four oocytes were used for this analysis. The sample was considered to lack or have NANOG expression only if the results were consistent between all samples. In our observations, lack of NANOG expression was defined by its absence or undetectability in all four samples of each developmental stage.
NANOG Expression Analyses
The presence of a duplication pseudogene, NANOGP1, and 10 processed pseudogenes that are closely related to NANOG [7, 8] could provide, in principle, competing transcript targets in siRNA knockdown studies. By designing specific primers for each of the 10 processed pseudogenes, we found that none of them was expressed in either human ES or EC cells (data not shown). However, like NANOG, NANOGP1 seems to be expressed in both human ES and EC cells (Fig. 1A). Only one band, at the expected molecular weight for NANOG (34.6 kDa), was detected by Western blot analysis (Fig. 1B); the absence of a band corresponding to a full-length NANOGP1 protein (expected at ∼27 kDa) is consistent with the prediction by Booth and Holland  that NANOGP1 is unable to be translated into a full-length protein. Immunofluorescence analyses with a NANOG antibody revealed that NANOG is localized in the nuclei of EC and ES cells (Fig. 1C), consistent with bioinformatic analyses that predict a nuclear localization signal (PVKKQKT) at amino acids 92–98 of NANOG protein.
Immunostaining of human preimplantation embryos and oocytes shows that the NANOG protein was absent from unfertilized oocytes, 2–16 cell embryos, and early morulae (Fig. 1D); however, very restricted expression was observed in the ICM of blastocysts, as has been described for mouse Nanog [5, 6, 12].
siRNA-Mediated NANOG Knockdown in Human EC and ES Cells
Previous studies in murine ES cells have suggested that stable downregulation of Nanog using the RNA interference (RNAi) technique has resulted in an inability to generate stable clones, which may reflect a role for Nanog in survival and renewal of ES cells . In view of this, we used short double-stranded RNA oligonucleotides (corresponding to a NANOG-specific region outside the homeodomain-encoding and SMAD4 domain–encoding regions; see Materials and Methods) to investigate the role of NANOG in human ES and EC cells. To rule out the possibility that effects of NANOG siRNA observed in this study are due to increased toxicity to the cells or increased contamination of this particular siRNA during chemical synthesis, we used several controls, and these included a mutated version of the NANOG siRNA (Fig. 2B). Previous reports have indicated that RNAi with control genes such as GFP or β2-microglobulin in human EC or ES cells does not result in nonspecific effects, changes in cell-surface antigen, or differentiation [18, 19]; therefore, GFP siRNA in addition to untreated and mock-transfected cells was used as additional negative control in this study (Fig. 2B). The most effective short double-stranded RNA corresponded to a sequence shared by NANOG and the duplication pseudogene, NANOGP1. This would suggest that NANOGP1 transcript, despite not being translated into a protein, would be downregulated as result of the RNAi approach. To ensure maximal silencing, ES and EC cells were transfected twice over a 48-hour interval. Evaluation of transfection efficiency 48 hours after the second transfection and using a Rhodamine-labeled NANOG siRNA (Fig. 2A) revealed a relatively high efficiency (in EC cells, 70% ± 1.53%; in ES cells, 65% ± 2.08%). Similar results were obtained with Rhodamine-labeled GFP and mutated NANOG siRNA (data not shown). This is similar to the transfection efficiency achieved by lentiviral delivery of siRNA in human ES cells . Real-time PCR analysis confirmed statistically significant downregulation of NANOG and NANOGP1 in cells transfected with NANOG siRNA compared with all control groups (Fig. 2B), which do not show significant downregulation of NANOG. There was a 63-, 34-, and 18-fold reduction of NANOG in the hES-NCL1, H1, and NT2-SP12 cells treated with siRNA to NANOG, respectively, compared with cells treated with GFP siRNA; corresponding reductions in NANOGP1 transcript levels were 90-, 46-, and 15-fold in the hES-NCL1, H1, and NT2-SP12 cell lines, respectively. There were no statistically significant changes in NANOG expression between cells treated with mutated NANOG siRNA, GFP siRNA, or mock-transfected cells, which indicates that they could be used as reliable negative controls. In a few instances (3 of 60 pairwise comparisons), we did detect statistically significant changes in NANOG expression between untreated cells and the other control groups. Although this occurred at low frequency, it did indicate that untreated cells might not provide the best negative control for these experiments. It is for this reason that we have used cells treated with GFP siRNA as negative control for most of the experiments, as described below. To further confirm NANOG downregulation, we carried out Western blotting analysis, which showed reduction of NANOG protein in both EC and ES cells (Fig. 2C). Furthermore, NANOG downregulation was maintained up to 6 days after transfection (supplemental online Fig. 1).
Nanog Knockdown Induces Differentiation in Human ES and EC Cells
The cell lines hES-NCL1, H1, and NT2-SP12 were transfected with either NANOG or GFP siRNA and monitored over 6 days. EC cells transfected with either mutated NANOG or GFP siRNA as well as mock-transfected cells remained undifferentiated for up to 6 days after transfection. In contrast, 2 days after transfection with NANOG siRNA, the cells began to appear spindle-shaped, and by 6 days, the proportion of spindle-shaped cells had increased (Fig. 2D). Human ES cells transfected with GFP siRNA, mutated NANOG siRNA, and mock-transfected cells remained largely undifferentiated (Fig. 2D), although a small proportion of morphologically differentiated cell types began to appear by 4 days after transfection. In contrast, 2 days after transfection with NANOG siRNA, enlarged flattened cells began to appear with an increased proportion of differentiated cells by days 4 and 6 (Fig. 2D). We analyzed the transfected cells by flow cytometry for the expression of cell-surface antigens typical for the human ES and EC cells, such as stage-specific embryonic antigen (SSEA)-4 and tumor-rejection antigen (TRA)-1-60, and as can be seen from Figure 2E, cells treated with NANOG siRNA had downregulated SSEA-4 expression (mean intensity, 66.78) compared with GFP siRNA–treated cells 4 days after transfection (mean intensity, 135.20), suggesting loss of pluripotency in human ES and EC cells as result of NANOG downregulation.
We sampled RNA transcripts at 2-day periods and conducted RT-PCR analysis for markers of pluripotency, such as SOX2, OCT4, and REX1, and of differentiation, such as PAX6 (neuroepithelium), BRACHYURY (mesoderm), fibroblast growth factor (FGF)-5 (primitive ectoderm), GATA6 and GATA4 (primitive endoderm), and CDX2 and GATA2 (trophectoderm) (data not shown). We noticed that in ES and EC cells transfected with NANOG siRNA, GATA4, GATA6, GATA2, and CDX2 were upregulated compared with cells transfected with GFP siRNA from day 4 after transfection. To enable a better analysis of these differentiation events, we used real-time PCR for all of the marker genes mentioned above (Fig. 3A). Real-time PCR analysis for the mesoderm marker BRACHYURY, primitive ectoderm marker FGF5, and neuroectoderm marker PAX6 showed no significant differences between the NANOG and GFP siRNA–transfected cells 4 days after transfection (data not shown), suggesting that NANOG downregulation is unlikely to induce differentiation along these lineages.
As assessed by real-time RT-PCR, NANOG downregulation induced OCT4 downregulation at day 4 after transfection, indicating loss of pluripotency in human ES cells; however, no downregulation of SOX2 or REX1 (with the single exception of REX1 in EC cells) was observed at this time. Sox2 expression has been found in trophoblast stem cells and shown to play a critical role in the development of the placenta . Rex1 was also shown to be critical for early placenta and preimplantation development in bovine embryos . NANOG downregulation induces the expression of early trophectodermal markers CDX2 and GATA2 as well as later markers hCG-alpha and hCG-beta (Fig. 3A) as early as day 4, consistent with induction of trophectoderm differentiation. The maintenance of SOX2 and REX1 expression could therefore be indicative of this type of differentiation and could not be used as estimate of pluripotency in these cells. In line with this observation, we noticed that SOX2 and REX1 expression was maintained after three passages of NANOG siRNA–transfected cells, whereas the expression of extraembryonic endoderm and trophectoderm markers was significantly upregulated in NANOG siRNA–treated cells compared with GFP siRNA–treated cells (Fig. 3B). Our own work has shown that the RT unit of telomerase, TERT, is downregulated upon differentiation of human ES and EC cells and can therefore serve as a biomarker for human ES/EC cells (Walter and Lako, unpublished observations). We did not observe significant changes in the expression of TERT at 4 days after transfection as a result of NANOG downregulation (Fig. 3A); however, significant downregulation was noticed after three passages (Fig. 3B). These results suggest that investigation of pluripotency and differentiation markers should be carried out at least at two time points after transfections because the kinetics of their regulation during differentiation process might be slightly different.
Expression of extraembryonic endoderm transcription factors GATA4 and GATA6, parietal endoderm marker LAMININ B1, and visceral endoderm marker AFP was significantly upregulated by real-time RT-PCR analysis 4 days after transfection with NANOG siRNA compared with GFP siRNA, consistent with the induction of extraembryonic endoderm differentiation (Fig. 3A). It should be noted that no AFP was detected in either the GFP or NANOG siRNA–transfected NT2-SP12 cells; however, this is perhaps due to monolayer culture conditions as opposed to aggregates, which show AFP upregulation . These findings are consistent with the proposed role of Nanog as suppressor of primitive endoderm differentiation in murine embryos and ES cells [5, 6]. These results obtained by quantitative PCR were confirmed by Western blot analysis (Fig. 4A) or immunocytochemistry on NANOG siRNA–transfected and GFP siRNA–transfected cells (Figs. 4B, 4C). AFP was only detected in the NANOG siRNA–transfected H1 and hES-NCL1 cells and was absent from human ES cells transfected with GFP siRNA (Fig. 4B). Some cells positive for GATA6 (Figs. 4A, 4C), GATA4, and CDX2 (data not shown) were detected at the periphery of the colonies of GFP siRNA–transfected H1, hES-NCL1, and NT2-SP12 cell lines, and this is consistent with a small amount of spontaneous differentiation occurring in these cultures as result of the length of siRNA experiment. However, many more cells positive for GATA6 (Figs. 4A, 4C), GATA4, and CDX2 were observed at human ES and EC cells transfected with NANOG siRNA. To correlate NANOG reduction with differentiation in individual cells, costaining of NANOG siRNA–treated cells with NANOG antibody and GATA6 was carried out (Fig. 4D). This analysis showed that by day 4 after transfection, cells with GATA6 expression had much reduced NANOG expression, providing further evidence of differentiated state in these cells induced by NANOG downregulation.
To investigate whether the effects of NANOG downregulation are irreversible, the transfected cells were maintained for three passages in culture. Real-time RT-PCR analysis at this stage indicated that NANOG expression was downregulated 1.8-fold and 3.2-fold in hES-NCL1 and NT2-SP12 cells, respectively (Fig. 3B), compared with a 63- and 18-fold decrease detected at day 4 after transfection. Despite this, the expression of trophectodermal marker hCG-beta and extraembryonic endoderm LAMININ B1 was upregulated in both ES and EC cells transfected with NANOG siRNA compared with cells treated with GFP siRNA (Fig. 3B). As other reports have suggested, the observed effect is likely to result from overgrowth of ES cells compared with more slowly growing differentiated cells . The presence of differentiated cells that express trophectodermal and endodermal markers after multiple passages suggests that the effects of NANOG downregulation are irreversible; however, rescue-type experiments combined with selection strategies need to be carried out.
The cellular differentiation events that occur just before and after implantation of the embryo are crucial for the viability of the embryo and lineage determination during normal development. At this narrow time window, trophectoderm is clearly delineated from the ICM. The other extraembryonic cell lineage, the primitive endoderm, differentiates into two further cell types, the parietal and visceral endoderm, and these both are major fetal components of the yolk sac . Understanding the molecular events that underpin these processes has been more difficult, especially in humans, given the lack of sufficient embryonic material. Human ES cells can open new avenues for understanding some of the very early lineage determination events that occur during embryogenesis.
In this work, we investigated the role of NANOG and were able to show that NANOG protein was localized to the nuclei of both human ES and EC cells, consistent with its suggested transcription factor activity. Using immunostaining, we were also able to determine that NANOG was absent from human oocytes, 2–16 cell embryos, and early morulae but present in the ICM of expanded blastocysts. This is in complete accord with the expression of Nanog in murine preimplantation embryos and corresponds to the time window from which the human ES cells have been derived so far . The derivation of human ES cells is highly inefficient and in most cases results in differentiation of ICMs to extraembryonic lineages such as trophectoderm and endoderm (Lako and M. Stojkovic, unpublished observations). The presence of NANOG-positive cells in the ICM can therefore be a good indication of the likelihood of a normal blastocyst from which human ES cell lines can be derived more efficiently.
Stable downregulation of Nanog has been attempted in murine ES cells without success probably because of loss of cell renewal in the clones lacking Nanog expression or cell differentiation . This necessitates other methods to investigate NANOG downregulation in ES cells. Using short double-stranded RNAs, we were able to achieve relatively high transfection efficiencies (65% in human ES and 70% in human EC cells), which resulted in significant downregulation of NANOG in both cell types. Downregulation of NANOG led to significant downregulation of OCT4 and loss of ES/EC cell-surface antigens, consistent with loss of pluripotency. This resulted in differentiation toward extraembryonic endodermal lineages, as shown by upregulation of several marker genes, GATA4, GATA6, LAMININ B1, and AFP [25–27], and trophectoderm specification, as indicated by upregulation of CDX2, GATA2, hCG-alpha, and hCG-beta . We were able to detect in both human ES cell lines an increase in AFP expression indicative of further differentiation from primitive endoderm toward visceral endoderm. In both the ES and the EC cell line, we were able to detect a significant increase in LAMININ B1 expression, which is associated with further differentiation to parietal endoderm. Our results are very similar to the findings reported recently by Zaehres et al. . This group used viral systems to deliver inhibitory RNAs under the control of human U6 promoter. Despite differences in the delivery system, cell lines used in this study (H9 was used instead of H1 and hES-NCL1) and NANOG target sequence, similar differentiation of human ES cells toward extraembryonic lineages was reported, providing further evidence on the role of NANOG in maintenance of pluripotency in human ES cells. Similar to data reported by Zaehres et al. , we observed that NANOG downregulation, albeit still present, was attenuated after several passages in culture, which suggests that the undifferentiated cells remaining in culture are likely to overgrow the more slowly dividing differentiated cells. Nevertheless, expression of trophectodermal and primitive endoderm markers was significantly upregulated after several passages, suggesting that effects of NANOG downregulation are unlikely to be reversible.
Absence of Nanog in murine ES cells also resulted in extra-embryonic endoderm differentiation [5, 6], whereas twofold downregulation of Nanog resulted in endoderm, mesoderm, and ectoderm differentiation . The discrepancies in the results can be attributed to differences in function of NANOG between human and mouse ES cells. Perhaps, and more likely, they can be attributed to a gene dosage effect, which suggests that stoichiometric levels of NANOG are required to specify different cell fates similarly to Oct4 in murine ES cells.
Expression studies in murine embryos have shown that Oct4 has a broad expression pattern at the compact morula stage, at which the first lineage specification events are likely to occur . In contrast, murine Nanog is localized to the inside cells of morulae in a reciprocal expression pattern to Gata6 and Cdx2, which indicates that Nanog is much more likely to determine the pluripotency of ICM by repressing transcription factors such as Gata6 and Cdx2, which result in extraembryonic lineage determination . Our results have shown localization of NANOG to the ICM of expanded blastocysts and have identified a role for NANOG in preventing differentiation of human ES cells toward extraembryonic lineages. Identifying NANOG targets and careful expression studies in early human embryos will enable us to investigate the first lineage specification events and molecular pathways that underline these processes.
We are grateful to Dr. Meena Choudhary for help with the collection of embryos. This work was supported by BBSRC, the Leukaemia Research Foundation, Medical Research Council UK, One North East, and the UK Department of Health (Life Knowledge Park).