Embryonic stem cells (ESCs), by virtue of their ability to continually self-renew and to differentiate into any cell type, hold great promise for the field of regenerative medicine. These cells also provide a system for studying the underlying molecular mechanisms that control early development. A central challenge is to develop methods to direct differentiation of ESCs in a controlled manner to produce individual populations of specific cell types. Elucidation of molecular pathways that define ESC pluripotency, self-renewal, and differentiation will be critical to achieving this goal.
In recent years, RNA interference (RNAi) has emerged as a powerful tool to silence gene expression in vitro. RNAi describes the phenomenon by which short double-stranded RNAs, most commonly referred to as short-interfering RNAs (siRNAs), elicit degradation of a homologous target mRNA [1, 2]. In this pathway, the siRNA duplex is believed to assemble with a series of proteins, referred to as the RNA-induced silencing complex, that guides hybridization of the siRNA antisense strand to its complementary target sequences and initiates cleavage of the target mRNA (reviewed by [3, 4]). Although RNAi holds great promise as a tool to study the basic biology of stem cells or to direct differentiation in a specific manner, barriers to achieving effective gene silencing can severely limit the utility of RNAi. Achieving good transfection efficiency or identifying highly active and specific siRNA sequences  are common problems for stem cell researchers using RNAi.
In this study, we used chemically modified RNAi compounds to silence genes responsible for key developmental pathways, to direct differentiation of mouse ESCs (mESCs). We targeted OCT4 and Nanog, two key transcription factors that function as major regulators of pluripotency and self-renewal in ESCs. It has been established that levels of OCT4 expression govern ESC fate. In mESCs, upregulation is associated with commitment to extraembryonic endoderm and mesoderm, whereas downregulation leads to formation of trophectoderm . RNAi-mediated inhibition of OCT4 in mESCs and human ESCs (hESCs) indicates a conserved role for OCT4 in maintaining pluripotency by preventing differentiation toward extraembryonic lineages [7–10]. The more recently discovered Nanog [11, 12] is thought to maintain pluripotency in part, by repressing genes that activate differentiation to lineages associated with extraembryonic endoderm [12, 13]. And unlike OCT4, overexpression of Nanog can override the requirements of leukemia inhibitory factor (LIF)-mediated signaling for maintaining the pluripotent state in mESCs in culture .
Here, we show that RNAi compounds can be used to specifically and effectively downregulate OCT4 or Nanog expression in mESCs, leading to phenotypic changes associated with differentiation to extraembryonic lineages.
Materials and Methods
P19 cells (CRL-1825; American Type Culture Collection [ATCC], Manassas, VA, http://www.atcc.org) were maintained in alpha minimum essential medium (12571-048; Gibco, Grand Island, NY, http://www.invitrogen.com) supplemented with 7.5% bovine calf serum, 2.5% fetal bovine serum, 1.5 g/l sodium bicarbonate, and penicillin/streptomycin. D3 (CRL-1934; ATCC) mESCs were maintained on mouse embryonic fibroblast (MEF) feeder layers treated previously with mitomycin C. mESCs were cultured in 12-well plates in serum-free medium (KnockOut Serum Replacement [KSR] Complete Medium) consisting of KnockOut Dulbecco's modified Eagle's medium (Gibco) supplemented with L-glutamine, nonessential amino acids, β-mercaptoethanol, penicillin/streptomycin, LIF (ES-GRO; Chemicon, Temecula, CA, http://www.chemicon.com), and 15% KSR (Gibco). LIF was present in the media throughout cell culture and transfections.
Stealth RNAi duplexes (Invitrogen Corporation, Carlsbad, CA, http://www.invitrogen.com) were designed to OCT4 (NM_013633) or Nanog (AY278951.1) using the BLOCK-iT RNAi Designer at http://www.invitrogen.com/rnaidesigner. Stealth RNAi compounds are 25-mer dsRNA molecules containing proprietary chemical modifications that enhance nuclease stability and reduce off-target effects by limiting sense strand activity . Sequence information for the Stealth RNAi duplexes and corresponding Stealth controls used in this study is provided in Table 1.
P19 cells were plated in 48-well plates at 10,000 cells per well 24 hours prior to transfection with 100 nM Stealth RNAi or 100 nM BLOCK-iT Fluorescent Oligo complexed with 2 μg/ml Lipofectamine 2000 (all from Invitrogen Corporation) according to the manufacturer's instructions. BLOCK-iT Fluorescent Oligo is a highly stable, fluorescein-labeled, nontargeted dsRNA compound that allows for visual monitoring of transfection efficiency. Transfections were performed in triplicate for each treatment. After 24 hours, transfection efficiency was assessed by visualizing uptake of the BLOCK-iT Fluorescent Oligo using fluorescence microscopy. Poly-A+ mRNA was isolated from individual wells of P19 cells using the mRNA Catcher (Invitrogen Corporation).
mESCs were transfected with complexes consisting of 50 nM BLOCK-iT Fluorescent Oligo or 50 nM Stealth RNAi and 2 μg/ml Lipofectamine 2000. Just prior to transfection, mESCs maintained on feeder layers were trypsinized, pelleted by centrifugation, and resuspended by trituration into a single-cell suspension. mESCs were then combined with freshly prepared transfection complexes and plated down onto fresh monolayers of MEF (175,000/well) at a total of 300,000 ESCs per well of a 12-well dish. Transfections were carried out for 24 hours in duplicate for each treatment and time point. To assess transfection efficiency, uptake of the BLOCK-iT Fluorescent Oligo was visualized by fluorescence microscopy at 24 hours. For the 48-hour time point samples, transfection media was replaced at 24 hours with fresh growth media and incubation continued for an additional 24 hours. At 24 and 48 hours after transfection, total RNA was harvested from individual wells of cells using the PureLink Micro-to-Midi Total RNA Purification System (Invitrogen Corporation). Total RNA was quantified using a spectrophotometer (Spectramax PLUS384; Molecular Devices, Sunnyvale, CA, http://www.moleculardevices.com). At 48 hours, corresponding wells of transfected cells were harvested for Western analysis. Also, at 48 hours after transfection, additional duplicate wells of cells were trypsinized, counted, and re-plated onto gelatin-coated six-well dishes at 100,000 cells per well for extended culture out to 6 days. Culture media containing LIF was changed daily. At day 6, total RNA was harvested from duplicate wells corresponding to each treatment, and cellular phenotype was verified by immunostaining and microscopic evaluation.
Quantitative Polymerase Chain Reaction Analysis of Gene Expression
Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) analysis on equivalent amounts of mRNA or DNAse-treated total RNA was performed using the SuperScript III Platinum Two-Step qRT-PCR Kit according to the manufacturer's instructions (Invitrogen Corporation). Primers used for qPCR reactions are shown in supplemental online Table 1. Real-time PCR was performed using an ABI Prism 7900 sequence detection system (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com). Equivalent amounts of cDNA generated from RT reactions were used as a template for PCR. Reactions were performed in triplicate for each sample. The thermal profile for PCR consisted of activation steps (50°C for 2 minutes, 95°C for 2 minutes), then 50 cycles of denaturation at 95°C for 15 seconds, followed by annealing and extension at 60°C for 1 minute. For each sample, expression of target and marker genes was normalized to the expression of GAPDH (glyceraldehyde 3-phosphate dehydrogenase) or cyclophilin. Data are expressed as the fold change in expression relative to no treatment.
P19 cells were grown on glass coverslips for 24 hours prior to transfection with Stealth RNAi. After 24 hours, transfection media was replaced with fresh growth media and incubation continued for another 24 hours. Cells were then washed, fixed in paraformaldehyde, permeabilized with Triton X-100, blocked, and incubated with primary anti-OCT4 polyclonal antibody (1:100, sc-8628; Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com) followed by secondary antibody (1:200, Alexa Fluor 488 conjugated rabbit anti-goat immunoglobulin G [IgG], A11078; Molecular Probes, Eugene, OR, http://probes.invitrogen.com). Coverslips were fixed with Prolong Gold and nuclei stained with 4′,6-diamidino-2-phenylindole (DAPI; P36935; Molecular Probes). Immunocytochemical localization of OCT4 protein was visualized by fluorescence microscopy using a Nikon Eclipse TE 200 microscope (Nikon, Tokyo, http://www.nikon.com) fitted with an real time color camera using SPOT RT imaging software (Diagnostic Instruments, Inc., Sterling Heights, MI, http://www.diaginc.com).
Trophoblast giant cells generated after 6 days following knockdown of OCT4 in mESCs were cultured on gelatin-coated glass coverslips. Cells were then washed, fixed in paraformaldehyde, permeabilized with Triton X-100, blocked, and incubated with TROMA-1 (anti-cytokeratin 8) monoclonal antibody (1:50; Developmental Studies Hybridoma Bank, Iowa City, IA, http://www.uiowa.edu/∼dshbwww) followed by secondary labeling with Alexa Fluor 488 conjugated goat anti-rat IgG (1: 100, A11006; Molecular Probes). Coverslips were fixed with Prolong Gold (P36930), and nuclei were stained with DAPI (P36935) for imaging by immunofluorescence microscopy or with TO-PRO-3 dye (S33025) for confocal imaging. Fixing and nuclear staining reagents were all from Molecular Probes.
Imaging of giant trophoblast cells by fluorescence microscopy was performed on a Zeiss Axiovert 200M inverted microscope (Carl Zeiss, Jena, Germany, http://www.zeiss.com) equipped with a C-Plan Apochromat ×40, 1.2 numerical aperture water immersion lens, and a CoolSnap HQ 12 Bit CCD (charge-coupled device) camera (Photometrics, Tucson, AZ, http://www.photomet.com). Images were acquired and processed using Metamorph software (Molecular Devices Corp., Sunnyvale, CA, http://www.moleculardevices.com). Corresponding DIC (differential interference contrast) images were acquired also. Confocal fluorescence microscopy was performed on a Zeiss LSM 510 META confocal mounted on a Zeiss Axioplan 2 upright microscope equipped with a C-Plan Apochromat ×40, 1.2 numerical aperture water immersion lens. Images were acquired using LSM5 software (Carl Zeiss) as three-dimensional image stacks and processed by deconvolution using Huygens Pro software (Scientific Volume Imaging B.V., Hilversum, The Netherlands, http://www.svi.nl). Visualization and image generation were performed using Imaris software (Bitplane AG, Zurich, Switzerland, http://www.bitplane.com).
Cells were washed twice in phosphate-buffered saline and lysed in ice-cold RIPA buffer containing protease inhibitors. Cell lysates were incubated on ice for 30 minutes and then clarified by centrifugation for 30 minutes at 5°C. Equivalent amounts of protein were fractionated on 10% Bis-Tris gels (NuPage NP0315; Invitrogen Corporation) in reducing conditions and electroblotted onto PVDF (polyvinylidene difluoride) membrane. Membranes were blocked with Western Breeze blocking buffer (46-7003, 46-7004; Invitrogen Corporation) and probed with anti-OCT4 (1:500, sc-5279; Santa Cruz Biotechnology) or anti-Nanog (1:5,000, A300–397A; Bethyl Laboratories, Inc., Montgomery, TX, http://www.bethyl.com) and anti-α-tubulin (1:1,000, 32–2500; Invitrogen Corporation) and then developed using the Western Breeze kit (anti-mouse [WP7104] or anti-rabbit [WP7106]) according to the manufacturer's instructions (Invitrogen Corporation).
Visualization of Transfection Efficiency
Transfection conditions for P19 cells and mESCs were optimized using cellular internalization of a fluorescently labeled short dsRNA as an indicator of transfection efficiency. Transfections were carried out for 24 hours using the BLOCK-iT Fluorescent Control Oligo complexed with Lipofectamine 2000. Strong intracellular fluorescent signal was observed by fluorescence microscopy in a large proportion of the P19 cells after transfection (Fig. 1A–1C). Transfection efficiency was determined by counting the total number of cells (nuclei) per field versus the number of corresponding cells with visible fluorescent uptake and averaging across three fields. The transfection efficiency for P19 cells was routinely observed to be at least 80%. Likewise, mESCs appeared to be efficiently transfected at 24 hours after co-plating of mESCs and transfection complexes onto feeder layers (Fig. 1D, 1E). The majority of mESC colonies were visibly transfected, with strong fluorescent signal localized within individual cells in each colony. Multilayer growth of the mESC colonies along with cells contributed from the feeder layer precluded precise calculation of transfection efficiency for mESCs, although we estimate it to be at least 50% transfected.
Validation of RNAi Compounds Targeting OCT4 and Nanog
After optimization of transfection conditions in P19 cells and mESCs, we first conducted screens in P19 cells to identify and validate active Stealth RNAi duplexes targeting OCT4 and Nanog. We designed and tested a total of five Stealth duplexes per target gene. Sequence information for the Stealth duplexes is presented in Table 1. We used P19 embryonal carcinoma (EC) cells to test efficacy of the Stealth compounds because these cells are relatively easy to culture and transfect and because they express genes related to pluripotency, such as OCT4 and Nanog. Each of five Stealth RNAi duplexes targeting OCT4 inhibited expression by at least 70% relative to the control, as determined by RT-qPCR analysis (Fig. 2A). The two best duplexes (1 and 3) targeting nonoverlapping sites (hereafter referred to as site 1 and site 2, respectively) were carried forward in experiments to downregulate OCT4 expression in mESCs. To confirm down-regulation of OCT4 expression at the protein level, we performed immunocytochemical analysis on P19 cells transfected with OCT-targeted Stealth RNAi (site 1) or the corresponding Control Stealth sequence. At 48 hours after transfection, OCT4 protein was largely undetectable in cells transfected with OCT4-targeted Stealth duplexes but was visibly localized to nuclei in cells transfected with nontargeted Control Stealth duplexes or in untreated cells (Fig. 2B).
For Stealth RNAi targeting Nanog, each of five duplexes tested in P19 cells inhibited expression of Nanog by at least 70% relative to the control, as determined by RT-qPCR analysis (Fig. 2C). Corresponding Western analysis of two duplexes (1 and 5) targeting nonoverlapping sites indicated that, at 48 hours after transfection, Nanog protein was reduced relative to control or no-treatment samples (Fig. 2D). These two duplexes targeting Nanog (hereafter referred to as site 1 and site 2) were carried forward into mESCs.
Inhibition of OCT4 in mESCs
For experiments targeting OCT4 or Nanog in mESCs, we used each of two Stealth sequences that we identified as having good efficacy based on screening in P19 cells. These sequences were also selected to target nonoverlapping sites, in order to confirm a specific RNAi effect. The sequences are provided in Table 1 and identified as site 1 and site 2 for each target gene. We also included chemistry and sequence composition-matched Stealth negative controls, as further verification of silencing specificity and to control for any effects related to delivery (transfection).
Expression of OCT4 mRNA was reduced by 80% and 67%, respectively, versus corresponding controls at 24 hours after a single transfection of mESCs with each of two Stealth duplexes (sites 1 and 2) (Fig. 3A). By 48 hours, levels of OCT4 inhibition had decreased to 59% and 58%, respectively, relative to controls. No appreciable change in OCT4 expression was detected in control samples relative to no treatment. MEF did not contribute to OCT4 expression, as expression was not detectable in these cells (data not shown). Western analysis (Fig. 3D) revealed that, at 48 hours, OCT4 protein was reduced in those treatments corresponding to inhibition of OCT4 mRNA (lanes 1 and 2) relative to controls (lanes 3 and 4) or no treatment (lane 5). The levels of reduction of OCT4 protein were consistent with corresponding levels of mRNA inhibition for each target site. At 24 and 48 hours, induction of expression of two markers of trophectoderm differentiation (Cdx2 and Hand1) was detected only in those samples in which OCT4 expression was inhibited (Fig. 3B and 3C, respectively). Induction of expression of Cdx2 and Hand1 ranged from approximately 20- to more than approximately 50-fold at 24 or 48 hours. To avoid overgrowth of the cultures and for longer-term analysis of differentiation, at 48 hours, duplicate wells of cells corresponding to each treatment were re-plated into six-well plates and cultured out to 6 days. At day 6, expression of placental lactogen (PL-1), a marker for trophoblast giant cells, was detected only in those wells corresponding to inhibition of OCT4 (Fig. 3E). Immunocytochemical staining of day-6 cultures revealed the presence of cells positive for TROMA-1 antibody, specific for trophoblast giant cells, in those treatments corresponding with inhibition of OCT4 and expression of PL-1. Along with staining for TROMA-1, giant cells were characterized by swollen nuclei and extensive cytoplasmic spreading (Fig. 4A–4G).
Inhibition of Nanog in mESCs
Expression of Nanog was reduced by 33% and 54% (sites 1 and 2, respectively) versus corresponding controls at 24 hours after transfection of mESC with Stealth RNAi (Fig. 5A). At 48 hours, levels of Nanog inhibition had decreased to 25% and 49% (sites 1 and 2, respectively) versus controls. For each time point, no appreciable change in Nanog expression was detected in control samples relative to no treatment (Fig. 5A). MEF did not contribute to Nanog expression, as expression was not detectable in these cells (data not shown). Western analysis (Fig. 5D) revealed that, at 48 hours, levels of Nanog protein were reduced in treatments corresponding to inhibition of Nanog mRNA (lanes 1 and 2) and relative to controls (lanes 3 and 4) or no treatment (lane 5). The reduction of Nanog protein corresponded with reduction in levels of Nanog mRNA for each target site (Fig. 5A, 5D). An approximately one- to fourfold upregulation of expression of extraembryonic endoderm markers GATA6 and GATA4 was detected in response to downregulation of Nanog by each of two Stealth duplexes (Fig. 5B and 5C, respectively). GATA6 expression was upregulated as early as 24 hours, whereas upregulation of GATA4 was not detected until 48 hours. To avoid overgrowth of the cultures and for longer-term analysis of differentiation, at 48 hours, duplicate wells of cells corresponding to each treatment were re-plated into six-well plates and cultured out to 6 days. In day-6 cultures, GATA4 expression was upregulated by an approximately one- to twofold whereas GATA6 was upregulated by approximately 0.5- to 1.5-fold relative to controls (Fig. 5E). Also in day-6 cultures, expression of laminin B1, a marker of parietal endoderm , was upregulated by approximately one- to twofold in treatments corresponding with inhibition of Nanog (Fig. 5F). In cultures in which upregulation of laminin B1 was detected, groups of dispersed cells with stellate shape and multiple branching filopodia were seen, indicative of parietal endoderm phenotype (Fig. 6A, 6B). Cells with a similar phenotype were not detected in control or no treatment wells (Fig. 6C, 6D).
The transcription factors OCT4 and Nanog are major regulators of the pluripotent state in ESCs. The mechanisms by which these factors function to support and maintain ESC pluripotency and self-renewal are of major interest and have not been fully elucidated. Methods to regulate expression of stem cell genes in a specific and efficient manner will facilitate efforts to elucidate ESC signaling pathways or to generate populations of differentiated cell types.
Prior to conducting gene-silencing experiments in mESCs, we first used P19 cells to conduct screens to identify highly active Stealth compounds that could then be carried forward into ESCs. P19 cells are an EC cell line that is derived from the stem cell component of a teratocarcinoma. Although these cells are malignant and have limited capacity to differentiate [16, 17], they do express genes related to pluripotency and can be used to validate efficacy of RNAi reagents. We have found it useful to validate RNAi reagents using cell lines with high transfection efficiency prior to attempting to achieve gene silencing in cell types that might prove more challenging to transfect, such as ESCs. This helps eliminate variation in efficacy of gene silencing that may be attributable to poor target site selection.
Target site selection is a critical first step in conducting RNAi experiments, and it is well known that there is a strong effect of the position of target site on the efficacy of siRNAs [18, 19]. Depending on the method or algorithm used to identify target sites, it may be necessary to screen several sequences in order to identify one or more with good efficacy. From screens conducted in P19 cells, we determined that all 10 of the Stealth RNAi sequences selected by the online design algorithm gave good efficacy against the target gene (70% inhibition or greater). Therefore, no additional sequences were tested.
We then carried forward a subset of these validated Stealth sequences (two per target gene) in experiments to inhibit expression of OCT4 and Nanog in mESCs and monitored the effect on differentiation over time. After a single transfection with Stealth RNAi, we were able to achieve potent and specific inhibition of OCT4 transcript and protein, with corresponding induction of trophoblast genes Cdx2, Hand1, and PL-1 and subsequent formation of cells with a trophoblast giant cell phenotype. These results are in line with previous reports demonstrating that differentiation toward trophectoderm is a consistent feature of OCT4 suppression in mESCs [6, 7, 9]. Likewise, induction of trophectoderm markers after RNAi-mediated inhibition of OCT4 in hESCs [7, 8, 10] indicates a functionally conserved role for OCT4 in maintaining the pluripotent state.
We were also able to reduce levels of Nanog transcript and protein after transfection of mESCs with Nanog-targeted Stealth RNAi, with corresponding induction of extraembryonic endoderm markers GATA4, GATA6, and laminin B1, and with generation of groups of cells resembling parietal endoderm after 6 days in culture.
In the case of both OCT4 and Nanog, transcript expression was inhibited by 24 hours after transfection, although by 48 hours, levels of inhibition were somewhat attenuated. Loss of inhibition could be attributable to recovery of gene expression after initial knockdown. However, at 48 hours we found that it was necessary to re-plate cells at a lower density to avoid overgrowth of the cultures, indicating that rapidly dividing untransfected cells may contribute largely to perceived loss of silencing activity. This effect has been consistently noted by others performing RNAi knockdown in ESCs [8–10, 20]. We also found that upregulation of differentiation markers was sustained over time and detected at day 6, suggesting that the effects of OCT4 and Nanog downregulation may be more permanent. However, longer-term analysis of differentiated cell populations, aided by selection strategies to isolate and propagate stable cell clones, will be required to more fully demonstrate the permanence of gene silencing.
Finally, we observed by immunocytochemical analysis of OCT4 knockdown in P19 cells that by 48 hours, OCT4 protein is largely undetectable within individual cells. Taken together, our results are consistent with observations that OCT4 and Nanog are rapidly downregulated upon initiation of differentiation in ESCs [6, 11, 12, 21–23].
To confirm an RNAi effect, we used two independent Stealth sequences per target that produced similar levels of mRNA knockdown (at least 70%) in P19 cells. When Stealth RNAi was transfected into mESCs, we observed some differences in the levels of target mRNA knockdown between each of the targeted sequences. These differences could reflect differences in transfection efficiency between the two cell types. Based on uptake of the BLOCK-iT Fluorescent Oligo, we routinely observed high levels (at least 80%) of transfection efficiency in P19 cells, which we cultured and transfected in a feeder-free monolayer format. However, we were able to obtain only an estimate of transfection efficiency (at least 50%) in the D3 mESCs, which we transfected as a single-cell suspension during plating and multilayer colony formation on a bed of feeder cells. The lower levels of target gene inhibition obtained in the D3 mESCs most likely reflect the greater challenge in our ability to obtain consistently high levels of transfection efficiency in these cells. Nonetheless, with each RNAi sequence, we were able to demonstrate a reduction of target gene transcript and protein expression leading to a corresponding induction of marker genes and phenotypic changes associated with differentiation under culture conditions that support maintenance of the undifferentiated phenotype. Our results indicate that a modest reduction in expression of OCT4 or Nanog in mESCs is sufficient to trigger differentiation, which supports the model that these transcription factors function in a dose-dependent manner to influence cell fate by repressing genes that activate differentiation [6, 11, 12, 24, 25].
We found that downregulation of Nanog by approximately 30%–50% was sufficient to induce differentiation toward extraembryonic endoderm, indicated by upregulation of marker genes GATA4, GATA6, and laminin B1. The appearance of dispersed groups of cells with stellate shape and multiple branching filopodia by day 6, along with corresponding upregulation of laminin B1, was evidence of further differentiation to parietal endoderm. This phenotype was very similar to that observed after forced expression of GATA4 or GATA6  or complete ablation of Nanog expression  in mESCs which results in exclusive differentiation to extraembryonic endoderm. Conversely, partial reduction in Nanog expression (∼50%) after disruption of a single allele in mESC leads to induction of markers, indicating differentiation to endodermal, mesodermal, and ectodermal derivatives . The differences in these observations seem to suggest that Nanog may function in a dose-dependent manner to influence cell fate. Although the dominant differentiated phenotype we observed in day 6 cultures resembled parietal endoderm, it remains to be determined whether reduction in Nanog expression triggered by transient RNAi in mESCs leads also to induction of markers of additional differentiation pathways.
Reports demonstrating RNAi-mediated inhibition of Nanog in hESCs [10, 20] indicate that Nanog functions to maintain pluripotency and prevents differentiation primarily to extraembryonic lineages. Using a vector-based approach, Zaehres et al.  achieved potent inhibition of Nanog in hESCs with subsequent induction of endodermal and trophectodermal markers. Likewise, Hyslop et al.  found that siRNA-mediated inhibition of Nanog expression in hESCs and human EC cells resulted in upregulation of markers corresponding to extraembryonic lineages, including the endodermal markers GATA4, GATA6, laminin B1, and alfa-fetoprotein.
By sampling different time points, we were able to analyze the kinetics of marker gene induction in response to inhibition of Nanog. We observed early upregulation of GATA6 in response to downregulation of Nanog, followed later by upregulation of GATA4, which supports the role of GATA6 as upstream activator of GATA4, as determined previously in knockout mice in which GATA6 expression is required for expression of GATA4 [26, 27]. One question that has been raised is whether differing thresholds of induction of GATA factors are required for initiation and for promotion of differentiation . Interestingly, we observed that a relatively modest induction (∼one- to fourfold) of GATA factors was associated with formation of an extraembryonic endoderm phenotype. In hESCs, similar levels of induction of GATA factors were observed, along with changes in cell morphology after siRNA-directed silencing of Nanog . Overall, our results are consistent with reports in both mESCs and hESCs [10 –13, 20] which indicate a role for Nanog in maintaining the pluripotent state by preventing differentiation toward extraembryonic lineages.
The data presented here suggest that transient inhibition of gene expression mediated by RNAi in ESCs can trigger differentiation pathways that lead to formation of differentiated cell populations. This represents an attractive alternative to methods that rely on permanent genetic modification (such as with short-hairpin RNAs expressed from stably integrated vectors) to direct gene silencing. The coculture conditions used in this study allowed us to monitor the effects of RNAi-induced differentiation in a serum-free system. Interestingly, our observations after silencing of OCT4 in D3 mESCs were similar to those reported by Velkey et al. , in which reduction of OCT4 expression in feeder-free D3 mESC cultured in serum with LIF resulted in differentiation to trophectoderm, suggesting that efficient differentiation can be achieved in culture systems with or without feeders. However, some differences in induction of differentiation after depletion of OCT4 or Nanog in mESCs have been noted [7, 13] and attributed to various factors, including differences in cell lines, culture conditions, or the level of gene silencing achieved (dosage effect). Although serum factors and feeder cells can be a confounding issue, we believe that the most important of these is dosage–which is in agreement with other reports [6, 7, 9, 10, 12, 20, 21, 24] indicating that Oct4 and Nanog have an intracellular threshold level below which differentiation can occur. Here, we have reported the use of highly potent RNAi duplexes combined with high-efficiency delivery in ESCs. We believe that RNAi-induced differentiation of ESCs, under various conditions, will be achievable if siRNAs can reduce target gene protein concentrations below the critical threshold.