Human embryonic stem cells (hESCs) are derived from the inner cell mass of preimplantation embryos; they can be cultured indefinitely and differentiated into many cell types in vitro. These cells therefore have the ability to provide insights into human disease and provide a potential unlimited supply of cells for cell-based therapy. Little is known about the factors that are important for maintaining undifferentiated hESCs in vitro, however. As a tool to investigate these factors, transfected hES clonal cell lines were generated; these lines are able to express the enhanced green fluorescent protein (EGFP) reporter gene under control of the OCT4 promoter. OCT4 is an important marker of the undifferentiated state and a central regulator of pluripotency in ES cells. These OCT4-EGFP clonal cell lines exhibit features similar to parental hESCs, are pluripotent, and are able to produce all three embryonic germ layer cells. Expression of OCT4-EGFP is colocalized with endogenous OCT4, as well as the hESC surface antigens SSEA4 and Tra-1-60. In addition, the expression is retained in culture for an extensive period of time. Differentiation of these cells toward the neural lineage and targeted knockdown of endogenous OCT4 expression by RNA interference downregulated the EGFP expression in these cell lines, and this correlates closely with the reduction of endogenous OCT4 expression. Therefore, these cell lines provide an easy and noninvasive method to monitor expression of OCT4 in hESCs, and they will be invaluable for studying not only OCT4 function in hESC self-renewal and differentiation but also the factors required for maintenance of undifferentiated hESCs in culture.
Embryonic stem (ES) cells are derived from the inner cell mass of preimplantation embryos and retain the developmental potency of embryonic founder cells, being able to differentiate into cells and tissues of all three germ layers in vitro and in vivo. Therefore, ES cells have important implications for providing insights into basic developmental biology. So far, most studies on ES cells have been carried out in mouse ES cells, as they were the only ES cells available until recently, when human ES cells (hESCs) were generated from human blastocysts. The establishment of hESCs provides resources not only for studying basic human developmental biology but also for their potential clinical applications. Subsequently, much attention has focused on directing hESC differentiation along specific developmental lineages and improving the conditions for maintaining these cells in culture.
Although both human and mouse ES cells were originally isolated and maintained by coculture on a mitotically inactivated mouse embryonic fibroblast (MEF) feeder layer, they may require different signals from the feeder cells for retaining their undifferentiated status. This is suggested by several studies [1–3]. In cultured mouse ES (mES) cells, the essential function of the feeder layer is the provision of the cytokine, leukemia inhibitory factor (LIF), and mES cells can be propagated and maintained in culture provided that the appropriate amount of LIF is added . Stat3 activation by LIF is required to sustain self-renewal [5, 6]. This pathway is thought to function in combination with additional signals that include OCT4 , Nanog [8, 9] and bone morphogenetic protein (BMP) signaling  to maintain mES cells in the undifferentiated state. However, in hESCs, the components involved in self-renewal are poorly defined. In the presence of serum, LIF is not sufficient to support the self-renewal of hESCs, and, even growing on MEF feeders, hESCs require the addition of basic fibroblast growth factor (bFGF) [1, 3]. A feeder-free method has been developed whereby hESCs can be maintained on matrigel-coated plates in culture medium conditioned from mitotically inactivated MEFs and supplemented with bFGF . It remains unclear, however, what factors in the conditioned medium act to support the self-renewal of hESCs. The generation of hES reporter cell lines that contain a specific marker for undifferentiated hESCs would provide invaluable tools for investigating the factors that control self-renewal and allow much needed improvements to be made to current culture conditions.
OCT4 (also known as OCT3, OCT3/4) belongs to the POU (Pit-Oct-Unc) family of transcription factors  and has been widely used as an important marker of undifferentiated ES cells [10, 12]. OCT4 expression is associated specifically with embryonic carcinoma cells, embryonic germ cells, and ES cells [11, 13]. It is essential for the development of the pluripotent inner cell mass (ICM) in mouse embryogenesis since targeted disruption of OCT4 in mice resulted in embryos devoid of an ICM . Quantitative analysis of OCT4 in mES cells revealed that higher levels of OCT4 induce differentiation of ES cells toward endodermal and mesodermal lineages, whereas lower levels result in trophectoderm differentiation, defining OCT4 as a key regulator of stem cell pluripotency and differentiation [14, 15]. OCT4 is expressed at high levels in undifferentiated hESCs and is downregulated upon differentiation , demonstrating that OCT4 in hESCs has a similar role in maintaining pluripotency. Therefore, OCT4 is a good candidate gene for directing reporter gene expression.
Enhanced green fluorescent protein (EGFP) is an autofluorescent protein that provides the opportunity to monitor expression in living cells and can be quantified by flow cytometry, confocal microscopy, and fluorimetric assays. In addition, EGFP-expressing cells in a heterogenous culture can be enriched by fluorescence-activated cell sorter (FACS) analysis. There are two principal strategies to generate OCT4 -EGFP reporter cell lines. One is to use homologous recombination to generate “knock-in” cell lines to introduce a selectable marker and the EGFP reporter into the endogenous OCT4 locus . The other approach is the transgenic method, whereby a plasmid containing a selectable marker and the EGFP reporter under control of the OCT4 promoter can be introduced into hESCs to generate clonal cell lines. The latter approach was used in the work described here since the OCT4 promoter fragment has been extensively characterized and known to drive tissue-specific expression [17–22].
Here we show that stably transfected clonal human ES cell lines with the OCT4 -EGFP plasmid exhibit high expression of EGFP in the undifferentiated state, which is downregulated during differentiation. The expression of EGFP correlates well with endogenous OCT4 gene expression in addition to hESC surface markers. The OCT4 -EGFP cell lines, like their parental cell line, have similar developmental potential and are able to generate cell types of all three germ layers.
Materials and Methods
Construction of phOCT4-EGFP
The human OCT4 promoter (hOCT4pr, from 67539 to 71490 in human DNA sequence with accession number AP000509) was amplified by polymerase chain reaction (PCR) with primers: hOCT4pr-F (5′– TT CCC ATG TCA AGT AAG TGG GGT GG-3′) and hOCT4pr-R (5′-CGA GAA GGC AAA ATC TGA AGC CAG G-3′) using human genomic DNA (Promega G3041; Promega, Southampton, U.K., http://www.promega.com) as a template. The fragment was cloned into TOPO vectors (Invitrogen, Paisley, U.K., http://www.invitrogen.com), and the fidelity of the DNA sequence was confirmed by bi-directional sequencing. The correct hOCT4 promoter was subsequently cloned into vector pEGFP1 (BD Biosciences, Oxford, U.K., http://www.bdbiosciences.com) by insertion into HindIII and AglI restriction enzyme sites upstream of EGFP.
Maintenance and Transfection of hES Cells
H1 cells  provided by Geron Corp. (Menlo Park, CA, http://www.geron.com) were cultured in medium conditioned by mitotically inactivated MEF conditioned medium (MEF–CM) supplemented with 8 ng/ml bFGF on matrigel-coated plates, as previously described . Cells were routinely passaged at a 1:3 dilution by treatment with 200 U/ml collagenase IV (Invitrogen).
H1 cells were split 1:3 by 0.5 mM EDTA treatment 24 hours before transfection and seeded onto a matrigel-coated six-well tissue culture dish. ApaLI linearized phOCT4-EGFP plasmid (10 μg) was tranfected into H1 cells using Fugene 6 transfection reagent (Roche, Indianapolis, http://www.roche-applied-science.com) at a DNA : Fugene ratio of 1:1.5, according to the manufacturer's instructions. G418 selection was applied 48 hours after transfection at 200 μg/ml; after 3 weeks in selection, the surviving colonies were picked individually to a 24-well plate and expanded.
Embryoid Body Formation and Neural Differentiation
Embryoid bodies (EBs) were generated as previously described . Briefly, confluent ES cells in a six-well plate were treated with 200 U/ml collagenase IV; then, small clumps of cells were cultured in suspension in EB differentiation medium (knockout-Dulbecco's modified Eagle's medium [KO-DMEM; Invitrogen], 20% fetal calf serum, 2 mM L-glutamine, and 100 × nonessential amino acids). The culture was maintained in suspension for 5 days, and EBs were collected and plated onto 0.5% gelatin-coated chamber slides. The attached EBs were continuously cultured in EB medium for another 4 days prior to fixation for immunocytochemistry. For neural differentiation, confluent ES cells were treated similarly as making EBs, but they were cultured in N2/B27 medium . The cell aggregates were grown in suspension for 28 days, then trypsinized and plated onto poly-L-lysine and laminin–coated coverslips in N2/B27 medium for a further 5 days before fixation for staining.
Immunocytochemistry and Flow Cytometry
Cells were either stained directly after phosphate-buffered saline (PBS) washing (for cell surface antigen) or fixed at room temperature with 4% paraformaldehyde for 10 minutes (EB for 20 minutes) and permeated with 100% ethanol for 2 minutes after washing with PBS. Cells were then incubated with 10% goat serum to block nonspecific binding, followed by primary antibody for 1 hour. Secondary antibody was applied for 30 minutes, and cells were washed with PBS before mounting with Mowiol (Calbiochem, San Diego, http://www.calbiochem.com). Monoclonal antibodies against OCT4 (1:250; Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com), SSEA4 (1:5; Developmental Studies Hybridoma Bank of Iowa University, Iowa City, IA, http://www.uiowa.edu/∼dshb), and Tra-1-60 (1:12; gifts kindly provided by Prof. P. Andrews, University of Sheffield, Sheffield, U.K.) were used as markers for undifferentiated hESCs. The rabbit polyclonal antibody against GFP (Molecular Probes, Invitrogen) was used at a dilution of 1:250. For differentiation, the following antibodies were applied: monoclonal antibodies against β-tubulin III (1:1000; Sigma-Aldrich Company, Dorset, U.K., http://www.sigma-aldrich.com), α-fetoprotein (1:500; Sigma), and muscle-specific actin (1:50; Dako, Glostrup, Denmark, http://www.ump.com/dako.html); polyclonal antibody against nestin (1:200; Chemicon International, Temecula, CA, http://www.chemicon.com).
Secondary antibodies used were goat anti-mouse tetra-methylrhodamine isothiocyanate (1:100; Sigma), goat anti-rabbit fluorescein isothiocyanate (FITC), goat anti-mouse FITC, and goat anti-rabbit Texas Red (all at 1:400; Jackson Laboratories, West Grove, PA, http://www.jacksoni-mmuno.com). Immunofluorescence was visualized and captured using Nikon Eclipse TC2000-U or Digital Pixel image analysis system.
Cells were prepared similarly for FACS analyses. Ten to twenty thousand cells were acquired for each sample using a FACScan (BD Biosciences) and analyzed with CELL-QUEST software.
Karyotyping and Telomere Length Assay
Cells were incubated with 1 mg/ml colcemid for 3 hours, then trypsinized, lysed with hypotonic buffer, and fixed in glacial : acetic acid (1:3). Metaphase spreads were stained with Giemsa, using standard methods. For each cell line, 10 metaphase spreads were analyzed. Telomere lengths were checked by telomere restriction fragment Southern blotting described previously .
OCT4 Knockdown by RNA Interference
ES cells were passaged by EDTA treatment onto matrigel-coated six-well tissue culture plates at low density. Cells were transfected with double-stranded small interfering RNA (siRNA) oligonucleotides specific for either human or mouse OCT4 at both 24 and 48 hours after plating, as described . Cells were analyzed for GFP and OCT4 expression 24 hours after the second transfection.
Total RNA was extracted using Qiagen RNAeasy kit according to the manufacturer's instructions (Qiagen, Valencia, CA, http://www1.qiagen.com). First-strand cDNA was synthesized as described previously , and 1/10 of the cDNA reaction mix was subjected to PCR amplification with DNA primers selective for genes OCT4 (forward, 5′-CTTGCTGCAGAAGTGGGTGGAGGAA-3′; reverse 5′-CTGCAGTGTGGGTTTCGGGCA-3′), GATA-6 (forward, 5′-GCAATGCATGCGGTCTCTAC-3′; reverse 5′-CTCTTGGTAGCACCAGCTCA-3′), EGFP (forward, 5′-GTAAACGGCCACAAGTTCAGC-3′; reverse 5′-GGC-GGATCTTGAAGTTCA-3′), and internal control ß-actin (forward, 5′-GATCAACTCACCGCCAACAGC-3′; reverse 5′ CTCCTCCTCCAGCGACTCAATCT-3′). PCR cycles consisted of an initial denaturation step at 94°C for 5 minutes, followed by 26 amplification cycles at 95°C for 30 seconds, 61°C for 30 seconds, and 72°C for 45 seconds. A final step of 72°C for 5 minutes was included. DNA contamination was excluded by PCR on the RNA template without reverse transcription.
The human OCT4 promoter fragment amplified from human genomic DNA by PCR spans base pairs −3917 to +55, relative to the transcriptional start site, and is highly homologous to the well-characterized mouse OCT4 fragment . Studies of the mouse promoter revealed that the elements important for OCT4 tissue-specific gene expression reside in this promoter region [17–22]. These include the distal enhancer that drives expression in the ICM and primordial germ cells, and the proximal enhancer that activates OCT4 in the epiblast . The fidelity of the human OCT4 promoter PCR fragment was confirmed by sequencing and was then inserted into the pEGFP-1 reporter construct upstream of EGFP. The resulting construct, phOCT4-EGFP, was linearized and transfected into the H1 hESC line. Transfection was performed using Fugene 6 transfection reagent in two wells of a six-well tissue culture plate and, following selection, nine resistant colonies survived. All the colonies expressed EGFP and were subsequently expanded to establish clonal cell lines. Transgene integration was confirmed by PCR and Southern blotting. Single-copy integration was found in all but one clonal cell line (T9, data not shown). Three of these OCT4-EGFP clonal lines (T5, T7, and T8) were characterized more extensively.
The OCT4-EGFP clonal cell lines grew normally, as did their parental H1 cells, in the MEF-CM, exhibiting typical undifferentiated hES colonies surrounded by differentiated stromal cells (Fig. 1D). EGFP expression in these cell lines was associated specifically with colonies of undifferentiated hESCs and was absent in the surrounding differentiated stromal cells (Figs. 1A, 1D). Antibody staining against glycoprotein antigen Tra-1-60 (Figs. 1B, 1E) and glycolipid antigen SSEA4 (Figs. 1C, 1F) showed that the OCT4-EGFP hESC lines also expressed these surface markers specifically in the undifferentiated hESC colonies that were in correspondence with EGFP expression. FACS analyses were carried out in the T5 cell line and showed that, on average (n = 5), 70% of the cells expressed high levels of EGFP (Figure 2A) and almost all GFP-positive cells exhibited high levels of SSEA4 and Tra-1-60 expression, which are 99% and 97%, respectively (Figure 2C and 2D). SSEA4 and TRA-1-60 expression patterns in OCT4-EGFP cells are very similar to the H1 parental cell line. OCT4-EGFP clones have been propagated in culture for approximately 1 year (50 passages) and have maintained EGFP expression, normal hESC morphology, and markers. They also exhibited stable telomere lengths, which are similar to the parental H1 cells (Fig. 3A). It has been shown that these cell lines retained the ability to generate all three germ layers following EB formation and differentiation (Figs. 1G–I) and are karyotypically normal (Fig. 3B).
Together, these data indicate that clonal hESC lines expressing OCT4-EGFP have been generated and propagated successfully without affecting the developmental potential and viability of hESCs. In addition, expression of EGFP does not appear to be toxic to these cells, and no silencing of the integrated transgene has been observed after long-term culture (over a year).
Since the OCT4-EGFP transgene integrated into the genome randomly, it may not necessarily reflect endogenous OCT4 gene expression and act as a good marker for undifferentiated hESCs. To determine if EGFP expression driven by the OCT4 promoter in these cell lines correlates with endogenous OCT4 gene expression, we carried out a series of co-immunostaining with both antibodies against OCT4 protein and EGFP during differentiation of OCT4-EGFP cell lines toward the neural lineage. The OCT4-EGFP cells were cultured in N2/B27 medium as cellular aggregates in suspension for 28 days. Initially, all cellular aggregates expressed similar levels of EGFP. After 21 days, EGFP expression differed between and within cell aggregates. Some were EGFP-positive, some were EGFP-negative, and some exhibited patchy EGFP expression. When aggregates were dissociated and plated onto poly-L-lysine and laminin–coated plates after 28 days differentiation, a mixture of groups of EGFP positive and negative cells was observed. FACS analysis showed approximately 70% of cells became EGFP-negative, whereas only 6% cells remained EGFP-positive and 24% exhibited reduced levels of EGFP (Fig. 2B). Immunostaining of these cells showed that the majority of the cells were EGFP-negative but positive for β-tubulinIII and nestin, the neural lineage markers; those cells remaining EGFP-positive were negative for β-tubulin III and nestin but positive with OCT4 antibody staining (Fig. 4). These results suggest that EGFP driven by the OCT4 promoter faithfully represents expression of endogenous OCT4 in undifferentiated ES cells and during their differentiation.
To further establish the specificity of EGFP expression driven by the OCT4 promoter, RNA interference was employed to specifically knock down OCT4 mRNA in these cells. It has been reported that the introduction of short double-stranded RNA into mammalian cells can inhibit endogenous gene expression  and that transfec-tion of hESCs with siRNA specific for human OCT4 results in the downregulation of OCT4 expression and, hence, differentiation of hESCs .
OCT4-EGFP cells were transfected with siRNA oligo-nucleotides specific for human or mouse OCT4. It has previously been shown that the human OCT4 siRNA (hOCT4) specifically targets the human OCT4 gene, whereas the functional mouse OCT4 siRNA (mOCT4) exerts no effect on human OCT4 expression . To increase knockdown of OCT4 in OCT4-EGFP cells, transfections were carried out at both 24 and 48 hours after plating. A marked reduction in EGFP expression was observed in those cells transfected with the hOCT4 oligonucleotide 24 hours after the second transfection (Figs. 5C, 5D, 5G, 5H), but not observed in those transfected with mOCT4 (Figs. 5A, 5B, 5E, 5F). The decrease in EGFP expression was also shown to correlate with a change in cell morphology, whereby cells transfected with hOCT4 (Fig. 5C) were found to have a more flattened appearance than the mOCT4-transfected cells had (Fig. 5A). Antibodies against EGFP and OCT4 showed that both EGFP and OCT4 protein levels are reduced in hOCT4-transfected cells (Figs. 5G, 5H) but not in mOCT4-transfected cells (Figs. 5E, 5F) and that OCT4 and EGFP are generally colocalized. Some cells, however, retained EGFP proteins. One explanation could be that reduction of OCT4-EGFP was not a direct effect of OCT4 siRNA transfection but an indirect result of OCT4 downregulation, which induced cell differentiation. This reduction in protein levels was also likely due to the extended half-life of EGFP, which is usually greater than 24 hours. This was further supported by RT-PCR results, examining the mRNA levels of OCT4 and EGFP. Both OCT4 and EGFP mRNAs were decreased at similar rates, about twofold lower in hOCT4 than in mOCT4 (Fig. 5I). In contrast to the reduction of OCT4 and EGFP mRNA levels, GATA-6, a marker associated with trophectoderm in early mouse development and later mesoderm and endoderm development , was clearly induced in hOCT4-transfected cells (Fig. 5I). Control RT-PCR reactions were performed in which no reverse transcriptase was added to ensure true mRNA amplification rather than contaminating genomic DNA (data not shown).
Together, these results confirm that EGFP expression driven by the OCT4 promoter in OCT4-EGFP clonal hESC lines truly reflects the endogenous OCT4 gene expression, and that this correlates with the undifferentiated status of the cells.
We have successfully generated OCT4 reporter human ES cell lines by plasmid transfection and shown that these hES clonal cell lines retain the normal hESC characteristics of self-renewal and pluripotency. This correlates with the results of early reports from other laboratories that clonally derived hESC lines can be propagated for prolonged periods in culture [28, 29]. Differentiating OCT4-EGFP cell lines to neuronal cell types and specific targeting of endogenous OCT4 downregulated EGFP expression further confirming that the 4-kb OCT4 promoter contains appropriate regulatory elements to drive developmentally specific EGFP expression that correlates closely with endogenous OCT4 gene expression. In addition, the OCT4-EGFP cell lines retained features associated with normal undifferentiated hESCs after long-term propagation, including normal karyotype, immunophenotype, and the ability to generate all three germ layers following differentiation.
Similar experiments have been reported to generate hESC reporter cell lines for undifferentiation [16, 29]. Eiges et al.  reported the generation of human reporter cell lines in which EGFP expression was under the control of the Rex1 promoter. The Rex-EGFP was expressed in undifferentiated hESCs and was downregulated during differentiation. In this report, however, the expression of transgenic EGFP and the endogenous Rex1 gene was not closely examined to ensure that the transgene expression was truly representative of the endogenous gene, or whether transfected cell lines were karyotypically normal and capable of long-term EGFP expression. In another study, OCT4-EGFP reporter hESC lines were created from a clonal H1.1 cell line, in which the EGFP sequences were inserted into the 3′ untranslated region of the OCT4 gene by homologous recombination. The targeted cells exhibited EGFP expression only in undifferentiated ES colonies, and the expression was downregulated after differentiation . Similarly, this work did not address the correlation between EGFP expression and endogenous OCT4 expression and other undifferentiated markers before and after differentiation. In our studies, we generated OCT4-EGFP reporter cell lines from parental H1 cells and characterized the transfected clonal cell lines in more details, showing stable EGFP expression in hES colonies after long-term culture and that EGFP expression is representative of endogenous OCT4 gene expression. In our experiments, all colonies exhibited EGFP expression after selection, though at various levels; no transgene silencing was observed, presumably because the selectable marker is linked to the EGFP transgene.
Since the first generation of human ES cells, much attention has been paid to improving the culture conditions for hESCs, with particular emphasis on defining the factors that are sufficient for maintaining hESC self-renewal and pluripotency so that no animal feeder, matrix, or conditioned medium is required [30–33]. It is known that OCT4 is important for self-renewal, and its expression is tightly regulated in mES cells [7, 14, 15]. In addition to OCT4, other factors have been implicated inself renewal of mES cells, including LIF/STAT3 [5, 6], nanog , and BMP4 signaling . However, little is known about which factors are important in hESCs self-renewal. The OCT4-EGFP cell lines will be a powerful model for such studies providing easy insight to the state of undifferentiation and differentiation. This will enable other factors to be examined in hESC lines and determine if they play a role in hESC self-renewal.
OCT4 has also been reported recently to be important for neurogenesis  and endoderm development . The OCT4-EGFP cell lines provide a marker that will allow the OCT4 expression to be followed, quantified, and selected in vitro. This marker will be an extremely valuable tool for studying the function of OCT4 protein, not only in maintaining hESC self-renewal but also during their differentiation. In summary, the availability of these OCT4-EGFP cell lines will provide important insight into improving the culture conditions of hESCs and investigating the function of OCT4 in both self-renewal and differentiation of hESCs.
We thank Prof. Peter Andrews, University of Sheffield, for providing us with the Tra-1-60 antibody and Ms. Judy Fletcher and Ms.Susan Craigmile for helping us with karyo-typing as well as technical assistance with the FACS analyses. We also thank the other members in the lab for their support; this work could not be done without their help. The work was sponsored by Biotechnology and Biological Science Research Council and Geron Corporation.