SEARCH

SEARCH BY CITATION

Keywords:

  • Epiblast stem cells;
  • Pluripotency;
  • Developmental biology;
  • Cell biology

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Epiblast stem cells (EpiSCs) are pluripotent stem cells derived from mouse postimplantation embryos at embryonic day (E) 5.5–E7.5 at the onset of gastrulation, which makes them a valuable tool for studying mammalian postimplantation development in vitro. EpiSCs can also be reprogrammed into a mouse embryonic stem cell (mESC)-like state. Some reports have shown that the reversion of EpiSCs requires transcription factor overexpression, whereas others have suggested that use of stringent mESC culture conditions alone is sufficient for the reversion of EpiSCs. To clarify these discrepancies, we systematically compared a panel of independent EpiSC lines. We found that—regardless of the embryonic day of derivation—the different EpiSC lines shared a number of defining characteristics such as the ability to form teratomas. However, despite use of standard EpiSC culture conditions, some lines exhibited elevated expression of genes associated with mesendodermal differentiation. Pluripotency (Oct4) and mesodermal (Brachyury) marker genes were coexpressed in this subset of lines. Interestingly, the expression of mesendodermal marker genes was negatively correlated with the cells' ability to efficiently undergo neural induction. Moreover, these mesodermal marker gene-expressing cell lines could not be efficiently reverted to an mESC-like state by using stringent mESC culture conditions. Conversely, Brachyury overexpression diminished the reversion efficiency in otherwise Brachyury-negative lines. Overall, our data suggest that different EpiSC lines may undergo self-renewal into distinct developmental states, a finding with important implications for functional readouts such as reversion of EpiSCs to an mESC-like state as well as directed differentiation. STEM CELLS 2011;29:1496–1503


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Pluripotency is the potential of stem cells to undergo differentiation into all somatic lineages of an organism. Different types of pluripotent stem cells can be derived from the mouse embryo. Mouse embryonic stem cells (mESCs) are derived from the inner cell mass (ICM) of preimplantation embryos. These cells can be indefinitely cultured in vitro without loss of developmental potency as demonstrated by their ability to form chimeras when injected into blastocysts of host embryos. Epiblast stem cells (EpiSCs) are derived from postimplantation mouse embryos from embryonic day (E) 5.5 to E7.5. Like mESCs, EpiSCs are considered to be pluripotent, as they form teratomas upon injection into immunocompromised mice. However, EpiSCs do not efficiently colonize host embryos when injected into blastocysts.

mESCs and EpiSCs also differ in morphology and the signaling factors required for self-renewal. For example, mESCs undergo self-renewal in the presence of leukemia inhibitory factor (LIF) and inhibitors of fibroblast growth factor (FGF) signaling [1]. In contrast, EpiSCs require FGF and Activin signaling to sustain self-renewal [2–4]. Intriguingly, human embryonic stem cells (hESCs), although derived from preimplantation embryos, appear to share more characteristics with EpiSCs than with mESCs. Therefore, a better characterization of the pluripotent EpiSC state should enhance our understanding of the nature of hESCs.

Several reports have described the characteristics and properties of EpiSCs. They present conflicting views on their ability to revert to a state similar to that of mESCs (a mESC-like state). Some studies suggest that transcription factor overexpression is required for the reversion of EpiSCs [5–8]. Others report the successful reversion of EpiSCs by simply culturing the cells under stringent mESC culture conditions [4, 9, 10]. We tentatively attributed these conflicting observations to the presence of different EpiSC lines in those experiments.

In this study, we systematically characterized a panel of independent EpiSC lines derived in several different laboratories to determine whether there are functional differences. We found that some basic characteristics of pluripotent stem cells were conserved across the different cell lines. However, we also found expression of marker genes normally involved in mesendodermal differentiation in some cell lines in the undifferentiated ground state. Interestingly, this gene expression signature was negatively correlated with the potential of EpiSCs to efficiently differentiate into the ectodermal lineage. Furthermore, we found that the expression of mesendodermal genes was also negatively correlated with the cells' ability to undergo medium-based reversion to a state similar to that of mESCs. Conversely, overexpression of the mesodermal transcription factor Brachyury/T diminished the reversion efficiency in otherwise Brachyury/T-negative lines.

Overall, our data support a model stipulating that different EpiSC lines may undergo self-renewal into slightly different cellular states of pluripotency. These states resemble different stages of early postimplantation development and may account for the different characteristics observed with the different cell lines.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Cell Culture

New EpiSC lines were derived as previously described [4]. Briefly, E5.5 embryos were plated on irradiated mouse embryonic fibroblast (MEF) cells seeded at low density. Outgrowths were picked and expanded manually. Basal medium consisted of knockout DMEM supplemented with 20% knockout serum replacement, nonessential amino acids, L-glutamine, and β-mercaptoethanol (all from Gibco). For cultivation on MEF cells, 5 ng/ml basic FGF (bFGF; Peprotech, Rocky Hill, New Jersey, www.peprotech.com) was added to the medium to yield EpiSC medium. For preparation of MEF-conditioned medium, another 5 ng/ml bFGF was added to EpiSC medium [conditioned medium (CM)] on inactivated MEFs. For feeder-free culture, EpiSCs were seeded onto fetal calf serum (FCS)-coated dishes. EpiSCs were routinely passaged using 1 mg/ml collagenase IV (Invitrogen). mESCs were cultured in standard culture conditions, namely knockout DMEM supplemented with 15% knockout serum replacement (both from Gibco), 5% FCS (Biowest, East Sussex, United Kingdom, www.biowest.net), nonessential amino acids, L-glutamine, and β-mercaptoethanol (all from Gibco), and either homemade or commercially available LIF (ESGRO; Millipore, Billerica, Massachusetts, www.millipore.com) at 1,000 U/ml. Some experiments were carried out in chemically defined N2B27 medium [11] in the presence of 10 ng/ml bFGF. For reversion experiments, EpiSCs were harvested using accutase (Invitrogen) and seeded as a single cell suspension at approximately 16,000 cells per well per six-well plate (Sarstedt, Nuembrecht, Germany, www.sarstedt.com) in CM. After 24 hours, the medium was changed to PD-CHIR-LIF medium (basal medium, supplemented with LIF [1,000 U/ml], PD0325901 [Axon Medchem, Groningen, The Netherlands, www.axonmedchem.com; 1 μM], and CHIR99021 [Axon Medchem; 3 μM]). After 48 hours of treatment, the reversion medium was further supplemented with 10 μM SB431542 (Sigma, St. Louis, Missouri, www.sigmaaldrich.com) to eliminate any remaining EpiSCs.

Immunocytochemistry and Alkaline Phosphatase Staining

Cells were fixed/permeabilized by using 4% paraformaldehyde/0.1% Triton X-100 and blocked using 5% FCS/2% bovine serum albumin (BSA)/2% glycine in 0.1% phosphate buffered saline–Tween 20 (PBS-T). Primary antibodies were applied in 0.5% BSA/0.5% glycine in 0.1% PBS-T overnight at 4°C. Primary antibodies used were Oct3/4 (Santa Cruz, Santa Cruz, California, www.scbt.com, N-19, sc-8628, or Santa Cruz, C-10, sc-5279), Sox2 (Santa Cruz, Y-17, sc-17320), Brachyury (Santa Cruz, C-19, sc-17745), β-Tubulin III (TUJ1; Covance, Princeton, New Jersey, www.covance.com), and smooth muscle actin (α-SMA) (Clone 1A4; Dako, Glostrup, Denmark, www.dako.com). All primary antibodies were used at a 1:300 dilution. Fluorescent visualization was carried out using suitable Alexa Fluor-conjugated secondary antibodies (1:400) together with 4′,6-diamidino-2-phenylindole (1:400) in 0.5% BSA/0.5% glycine in 0.1% PBS-T for 1 hour at room temperature.

For alkaline phosphatase (AP) staining, cells were fixed and incubated with a 25:1 mixture of Fast Red chromogen (1 mg/ml, Sigma) and naphthol phosphate solution (0.25% [w/v], Sigma) for about 15 minutes.

qPCR and Microarrays

RNA isolation was performed using RNeasy kits (QIAGEN, Hilden, Germany, www.qiagen.com) with on-column DNase digestion. Reverse transcription for real-time quantitative polymerase chain reaction (RT-qPCR) was performed using MMLV reverse transcriptase [USB (Affymetrix), Santa Clara, California, www.affymetrix.com] and Oligo-dT15 priming at 42°C for 1 hour and at 60°C for 10 minutes. A cDNA equivalent of 50 ng total RNA was used as template in a total reaction volume of 20 μl with Power SYBR Green PCR mix (ABI/Ambion/Invitrogen, Carlsbad, California, www.invitrogen.com) on an ABI 7300 cycler. Primers were added at 0.375 μM each. Calculations were based on the ΔΔCt method using two housekeeping genes for normalization. Primer sequences are given in Supporting Information Table S1. For the gene expression microarray analysis, cRNA samples were prepared with the linear TotalPrep RNA Amplification Kit (Ambion). Hybridizations on mouse-8 V2 chips (Illumina, San Diego, California, www.illumina.com) were carried out as recommended by the manufacturer. For the microarray data processing, the bead intensities were mapped to gene information using BeadStudio 3.2 (Illumina). Background correction was performed using the Affymetrix robust multiarray analysis background correction model [12]. Variance stabilization was performed using log2 scaling, and gene expression normalization was calculated with the method implemented in the lumi package of R-Bioconductor. Data postprocessing and graphics were performed with in-house developed functions in Matlab. Hierarchical clustering of genes and samples was performed with one minus correlation metric and the unweighted average distance (unweighted pair group method with arithmetic mean), (also known as group average) linkage method.

Sextyping PCR

Approximately 5 × 105 cells were used for sextyping PCR on genomic DNA. To this end, cells were lysed in lysis buffer (20 mM Tris-HCl, pH 8 [Applichem, Darmstadt, Germany, www.applichem.com], 0.2 mM EDTA, 0.1% Tween 20, 0.1% NP-40 [all Sigma]) plus proteinase K (Applichem) at 65°C for 15 minutes. Proteinase K was heat inactivated at 95°C for 5 minutes. Subsequently, 0.5 μl of lysis solution was used as a template for PCR. PCR reactions were performed with 2X PCR Mastermix (Fermentas, Waltham, Massachusetts, www.fermentas.com) in standard PCR conditions. Primers for Sry and IL-2 (Supporting Information Table S1) were used at 0.25 μM each.

Teratoma Formation

Undifferentiated EpiSC colonies were collected using 1 mg/ml collagenase IV (Invitrogen). After short centrifugation, approximately 2 × 106 cells were injected subcutaneously into the dorsal flank of severe combined immune deficiency (SCID) mice. After 4–6 weeks, tumors were collected and fixed in Bouin's solution (Sigma). Paraffin-embedded tissues were sectioned and stained with H&E.

Transient Transfection

For overexpression experiments, Brachyury cDNA was cloned into the pMSCV vector (Clontech, Mountain View, California, www.clontech.com) using standard cloning techniques. Red fluorescent protein cDNA cloned into pMSCV was used as negative control vector. Brachyury knockdown constructs were obtained from Sigma (SHGLY-NM_009309). Scrambled small hairpin RNA (Addgene plasmid 1864) was used as negative control. For transient transfection experiments, EpiSCs were dissociated using accutase and replated as very small clumps or single cells. The next day, cells were transfected using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. Two days after transfection, the cells were seeded for reversion experiments.

Luciferase Assays

Luciferase constructs containing the proximal enhancer (PE) and the distal enhancer (DE) of the Oct4 promoter [4, 13] were transfected into EpiSCs and mESCs, along with renilla normalization constructs, using Lipofectamine 2000 (Invitrogen). Cells were harvested 48 hours after transfection and processed according to the DualGlo Luciferase protocol (Promega, Madison, Wisconsin, www.promega.com). Relative Luciferase activity was normalized to the activity of the empty vector control.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Characterization of EpiSC Lines

Several different EpiSC lines from different genetic backgrounds were derived in-house. Additionally, we assayed two EpiSC lines generated and described elsewhere [2, 3]. Five lines were derived on E5.5, whereas one line, C1a1, was derived on E6.5. Sextyping of EpiSC lines via PCR on genomic DNA for Sry as well as analysis for expression of the Y-chromosome related gene Eif2s3y revealed that all lines except E3 were female lines (Supporting Information Figure S1D). All EpiSC lines could be stably cultured either on a feeder layer of mitotically inactivated MEFs or on FCS-coated dishes using MEF-conditioned medium. These lines formed typical large, flat colonies that were morphologically different from the domed-shaped mESC colonies and closely resembled hESC cell colonies (Fig. 1A; Supporting Information Fig. S1A). Quantitative PCR revealed that expression of the core pluripotency markers Oct4, Sox2, and Nanog was similar to that of mESCs. Additionally, expression of the epiblast marker Fgf5 was clearly induced, whereas Rex1, a marker for mESCs was expressed at much lower levels in EpiSC lines compared with mESCs (Fig. 1B). Immunofluorescence showed nuclear localization of Oct4 and Sox2 (Fig. 1C). Luciferase assays confirmed higher activity of the epiblast-specific PE of Oct4 in all EpiSC lines and lower activity of the ICM-specific DE compared with mESCs (Fig. 1D). Global gene expression analysis revealed that all EpiSC lines clustered together and differed from mESCs (Fig. 1E). Scatter plots showed variable degrees of differences in the gene expression of these aforementioned genes among the EpiSC lines (Supporting Information Fig. S1C). Overall, these differences, however, were smaller than those between EpiSCs and mESCs, confirming that all the EpiSC lines were EpiSCs and not mESC-like cells. Upon injection into SCID mice, all previously uncharacterized EpiSC lines (OG2.1, OG2.2, and E5) gave rise to teratomas that contained tissues of all three germ layers, confirming their pluripotent properties (Supporting Information Fig. S1B). Lines E3, T9, and C1a1 have previously been characterized using this assay [2–4]. Taken together, these data demonstrate that these six cell lines were pluripotent EpiSC lines.

thumbnail image

Figure 1. Characterization of epiblast stem cell (EpiSC) lines. (A): Colony morphology of EpiSC lines compared with human embryonic stem cells and mouse embryonic stem cells (mESCs) grown on mouse embryonic fibroblasts. (B): Quantitative real-time RT-PCR analysis of marker gene expression in EpiSC lines compared with mESCs. Bars reflect normalization errors. (C): Immunocytochemistry analysis of expression of Oct4 and Sox2 in EpiSC lines. (D): Luciferase assay showing Oct4 enhancer activity in EpiSC lines compared with mESCs. (E): Dendogram of the microarray global gene expression based on linear correlation coefficients (r). Abbreviations: DAPI, 4′,6-diamidino-2-phenylindole; DE, distal enhancer; hESC, human embryonic stem cell; mESC, mouse embryonic stem cell; PE, proximal enhancer; RT-PCR, real-time polymerase chain reaction.

Download figure to PowerPoint

EpiSC Lines Exhibit Diverse Expression Patterns of Mesendodermal Marker Genes

Some reports have described the expression of mesendodermal marker genes as a characteristic of EpiSCs [8, 13, 14]. One prominent example is the primitive streak and mesoderm-associated Brachyury/T gene. Therefore, we analyzed the expression of Brachyury, along with some other mesendodermal markers, in the different EpiSC lines. Strikingly, Brachyury was detected in some, but not all EpiSC lines by immunocytochemistry (Fig. 2A). Notably, all lines had been grown under standard EpiSC culture conditions and showed homogeneous colony expression of Oct4. Virtually all colonies of line C1a1 exhibited Brachyury protein expression. Lines OG2.1 and OG2.2 also showed high expression, but the colonies appeared to have fewer Brachyury-positive cells. Line T9 exhibited the weakest expression of Brachyury compared with EpiSC lines C1a1, OG2.1, and OG2.2. Finally, lines E3 and E5 did not exhibit any Brachyury expression at all. These data were confirmed by RT-qPCR. Line C1a1 exhibited the highest expression of Brachyury, followed by lines OG2.1 and OG2. Line T9 showed only slightly higher expression compared with lines E3 and E5. Similarly, the primitive streak and mesendoderm markers Mixl1, Goosecoid (Gsc), Sox17, and Eomes were expressed at higher levels in lines C1a1 and OG2.1 than in lines T9, E3, and E5, with lowest levels found in the latter two lines (Fig. 2B). This gene expression signature appears to be stable, as Brachyury/T and Mixl1 expression was similar after multiple passages (Supporting Information Fig. S2A). A similar signature was also found in two distinct induced EpiSC lines (Fig. 2B). These results suggest that expression of mesendodermal markers may be significant in some, but not all, EpiSC lines and should therefore not be considered a general hallmark of the EpiSC state.

thumbnail image

Figure 2. Expression of early mesendodermal markers in different epiblast stem cell (EpiSC) lines is negatively correlated with neural induction potential. (A): Immunocytochemistry showing colocalization of Oct4 and Brachyury expression in EpiSC lines. (B): Quantitative real-time RT-PCR analysis of early mesodermal and endodermal marker gene expression in different EpiSC as well as two distinct induced EpiSC lines compared with E3. Bars reflect normalization errors. (C): Quantitative real-time RT-PCR analysis of gene expression of early neural markers after 48 hours of neural induction compared with fibroblast growth factor 2 control. Bars reflect normalization errors. Abbreviations: FGF, fibroblast growth factor; iEpiSC, induced epiblast stem cell; RT-PCR, real-time polymerase chain reaction.

Download figure to PowerPoint

Correlation Between Mesendodermal Gene Expression Signature and Germ Lineage–Specific Differentiation Potential

We next sought to analyze whether the mesendodermal gene expression signature in the ground state may have functional consequences such as influencing the differentiation propensity of a given EpiSC line into particular germ layer lineages. To this end, we used a standard mesodermal differentiation protocol based on FCS stimulation. We observed rapid induction of early mesendodermal markers in Brachyury/T-negative lines within 2 days of differentiation, but no difference in induction of myogenic and cardiac differentiation markers (Pax3 and Nkx2.5, respectively) (Supporting Information Figure S2B). Similarly, we found no difference in mesodermal differentiation propensity, as determined by α-SMA expression 14 days after differentiation by immunocytochemistry analysis (Supporting Information Fig. S2C). These results suggest that all EpiSCs lines exhibit the potential to differentiate into the mesodermal lineage, regardless of whether the EpiSC lines exhibited a mesendodermal marker gene expression signature in the undifferentiated state.

Next, we determined whether the mesendodermal gene expression signature of EpiSCs in the undifferentiated state reduces the induction efficiency of different EpiSC lines into the ectodermal lineage. Therefore, EpiSCs were cultured in conditions known to efficiently induce neural differentiation in EpiSCs [4], that is, by blocking FGF and SMAD signaling. Interestingly, when compared with standard EpiSC culture conditions, the highest induction of early neural markers Sox1 and Pax6 was seen in lines E3 and E5, which were also the ones that had shown the lowest levels of mesendodermal gene expression in the ground state (Fig. 2C). In contrast, those lines that had shown significant Brachyury expression in the undifferentiated state displayed significantly weaker induction of early neural markers. After 4 days of differentiation, expression of neural markers was analyzed by immunocytochemistry. All lines showed expression of the neural marker β-III-Tubulin in at least some cells, thus ruling out all likelihood of a general failure of lines to differentiate along the neural lineage in vitro (Supporting Information Fig. S2D). However, the lines strongly differed in their efficiency for neuroectodermal induction. Interestingly, a comparison of the different lines revealed that the expression of mesendodermal markers in the undifferentiated ground state correlated negatively with the potential of early neuroectodermal induction.

Negative Correlation Between Mesendodermal Gene Expression Signature and Reversion Potential

Next, we sought to determine whether expression of mesendodermal markers under conditions that promote self-renewal might also influence the potential of EpiSCs to revert to an mESC-like cell state. In contrast to mESCs, EpiSC lines showed only weak or even no AP activity whatsoever in the ground state. Therefore, initiation of AP activity, along with dome-shaped mESC-like colony morphology, can be used as a convenient readout of successful reversion [4, 13]. Small domed colonies appeared 4 days after using stringent mESC culture conditions (MEK inhibitor PD PD0325901, CGSK3β inhibitor CHIR99021, and LIF) for EpiSC lines. However, the different EpiSC lines varied drastically in the number of such colonies (Fig. 3A, 3B). Line E3 had the highest number of AP-positive colonies, whereas lines E5 and T9 had several-fold fewer. Lines OG2.1 and OG2.2 had only very few AP-positive colonies, whereas line C1a1 showed none.

thumbnail image

Figure 3. Reversion efficiency varies between epiblast stem cell (EpiSC) lines. (A): Staining for alkaline phosphatase (AP) activity in EpiSC lines after 4 days of reversion treatment. Arrows indicate AP-positive colonies, whereas open arrowheads show brown colonies, which were not considered to be AP positive. (B): Comparison of reversion efficiency based on relative numbers of AP-positive colonies. Bars reflect SD between three biological replicates. Abbreviations: AP, alkaline phosphatase; PCL, PD-CHIR-LIF medium.

Download figure to PowerPoint

The EpiSC lines OG2.1 and OG2.2 carry a transgenic green fluorescent protein (GFP) construct under the control of an ICM-specific Oct4 promoter fragment (GOF18ΔPE) in which the PE has been deleted [15]. While OG2 embryos derived at E5.5 still showed residual GFP expression, the GFP signal rapidly disappeared upon establishing the EpiSC lines (Supporting Information Fig. S3) [13, 16]. Spontaneous GFP reactivation was not detected by routine monitoring using fluorescence-activated cell sorting (data not shown). Therefore, to complement the less stringent AP/morphology assay, reversion of OG2 EpiSCs to mESC-like cells was also monitored for ICM-specific GFP reactivation. In independent experiments, we observed reactivation of GFP in OG2.1 and OG2.2 EpiSCs cultured under stringent mESC conditions for 4 days (Supporting Information Fig. S3), suggesting that reversion was successfully achieved. We consistently found that these two lines exhibited a very low relative number of colonies showing GFP reactivation, similar to the low number of AP-positive colonies obtained previously.

Taken together, these results confirm that the reversion of EpiSCs to an mESC-like state by switching culture conditions can be induced in many—although not all—EpiSC lines. However, reversion efficiency, varied among the cell lines. Mesendodermal gene expression signature in EpiSCs in the ground state was negatively correlated with the reversion potential of a given line. EpiSC lines with little or no expression of Brachyury and other primitive streak markers were readily induced to revert to mESC-like cells by modifying the culture conditions only. In contrast, lines with high mesendodermal expression could only be induced to revert at a much lower efficiency, if at all.

Modulation of the Mesendodermal Gene Expression Signature Modifies the Reversion Potential of EpiSC Lines

To determine whether the reversion inefficiency of some lines is due to their mesendodermal gene expression signature, we overexpressed Brachyury/T in EpiSC lines E3, E5, and T9. When Brachyury expression was upregulated by about 10-fold, a slight change in the expression of other mesendodermal marker genes—Mixl1, Sox17, Gsc, and Eomes—was observed only in line T9, whereas no changes were found in lines E3 and E5 (Fig. 4A). In addition, when EpiSC lines E3, E5, and T9 were cultured in the presence of 15% FCS for 18 hours to induce mesendodermal marker gene expression, we found even higher induction of all mesendodermal marker genes, with no change in Oct4 expression (Fig. 4B), suggesting successful mesendodermal differentiation without alteration of pluripotency. Reversion experiments of treated EpiSCs revealed a reduction of more than 50% in the number of AP-positive colonies after transient FCS exposure. Line E5 revealed the most pronounced reduction—a 75% decrease in the number of AP-positive colonies (Fig. 4C). These results suggest that overexpression of the mesendodermal marker gene Brachyury/T diminishes the reversion potential of EpiSCs.

thumbnail image

Figure 4. Modulation of the mesendodermal gene expression signature modifies the reversion potential of epiblast stem cell (EpiSC) lines. (A): Quantitative real-time RT-PCR analysis of early mesendodermal marker genes after Brachyury overexpression in EpiSC lines E3, E5, and T9 compared with EpiSCs transfected with RFP vector control. Bars reflect normalization errors. (B): Quantitative real-time RT-PCR analysis of early mesendodermal marker genes after 18 hours of serum-induced differentiation in EpiSC lines E3, E5, and T9 compared with fibroblast growth factor control. Bars reflect normalization errors. (C): Percentage of alkaline phosphatase-positive colonies of treated EpiSCs (Brachyury overexpression or serum induction) after reversion. Dashed line represents 100%. Bars reflect SD of three biological replicates. Abbreviations: AP, alkaline phosphatase; FGF, fibroblast growth factor; OE, overexpression; RFP, red fluorescent protein; RT-PCR, real-time polymerase chain reaction; SI, serum induction.

Download figure to PowerPoint

We also performed the reciprocal experiment of silencing Brachyury/T expression in EpiSC lines OG2.1, OG2.2, T9, and C1a1 using RNAi. Reduction of Brachyury expression in these EpiSC lines did not appear to positively influence the reversion efficiency of EpiSCs to mESC-like cells, but we cannot rule out the possibility that the knockdown attained was insufficient for such an outcome (data not shown).

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

EpiSCs are an attractive model system for studying early mammalian development in vitro and for better defining hESCs, which are closely related to EpiSCs, as these cell lines are derived from embryos just before or at onset of gastrulation [14]. However, to be able to harness the full potential of these cells, it is imperative that we understand the in vivo developmental state to which EpiSCs correspond in vitro. Consistent with the time frame of their derivation, our comparison of a panel of different lines confirms that EpiSCs share a number of epiblast-like characteristics, such as Fgf5 expression, lack of ICM marker expression, preferential usage of the Oct4 PE, and teratoma formation ability (Tab. S1).

However, published reports have suggested a less homogeneous profile. For example, some reports proposed the expression of early differentiation markers as a characteristic of EpiSCs [3, 17], whereas others reported the absence of these markers. Furthermore, some studies found that transcription factor overexpression was required for reversion of EpiSCs to an mESC-like state [5–8], whereas others showed reversion could be achieved by somewhat milder means just using stringent mESC culture conditions [4, 9, 10].

We examined a panel of independent EpiSC lines to investigate these apparent discrepancies. We found that different lines could be distinguished by the expression of primitive streak or mesendoderm-associated genes. The Brachyury gene is frequently used as a marker to track early cellular differentiation into the mesodermal lineage in vitro [18, 19]. However, the very early (E5.5) pregastrulation epiblast is negative for Brachyury expression. The heterogeneous expression of Brachyury in different EpiSC lines and the association between Brachyury expression and the mesoderm lineage in vivo led us to propose that Brachyury and other mesendoderm markers are not suitable as universal markers for EpiSCs.

Our data demonstrate that the mesendodermal marker expression signature found in some lines is negatively correlated with both the neural induction efficiency and the propensity of EpiSCs to be reverted to an mESC-like state through the use of stringent mESC culture conditions. In this regard, some EpiSC lines appear to behave as “primed” pluripotent cells [20], in that they are biased in their commitment toward the mesodermal lineage and exhibit reduced propensity to differentiate or dedifferentiate into other lineages.

However, this was not the case for some other lines we investigated. Therefore, we conclude that different EpiSC lines may undergo self-renewal into slightly different developmental states, corresponding to earlier and later stages of postimplantation development [21–23]. This conclusion is also based on the observation that different levels of Brachyury are expressed in embryos of distinct early postimplantation stages [24]. Therefore, we conclude that the EpiSC lines expressing lower levels of Brachyury represent a somewhat earlier developmental stage compared with those EpiSC lines with elevated expression. In addition, the earlier EpiSC lines might represent cells that are not biased to differentiate into a certain germ cell lineage in vivo, whereas the later EpiSC lines might be traced back to a developmental stage subsequent to the earliest mesendodermal differentiation step (Fig. 5).

thumbnail image

Figure 5. Different epiblast stem cell (EpiSC) lines resemble distinct developmental ground states. Timeline of early mouse postimplantation development and the in vitro cell lines corresponding to these developmental stages. The red color marks the Oct4 expression in the pluripotent cells of the epiblast. EpiSC lines (blue) are ordered according to the in vivo ground state. The staging of the postimplantation embryos is based on Pfister et al. [21]. Abbreviation: DVE, distal visceral endoderm.

Download figure to PowerPoint

EpiSC lines corresponding to early epiblast can readily be reverted to an mESC-like state, as they are developmentally similar to mESCs. The opposite appears to apply to lines that show high expression of mesendodermal genes (lines OG2.1, OG2.2, and C1a1) (Fig. 5). Such lines may require overexpression of mESC-specific transcription factors, that is, harsher treatments, to achieve efficient reversion to an mESC-like state. Also, the EpiSC lines produced from somatic cells by direct reprogramming may fall into the latter category, as they show relatively higher expression of mesendodermal marker genes and cannot undergo reversion under stringent mESC conditions [25]. We have previously suggested an equilibrium model describing the transition of mESCs to EpiSCs and vice versa [4]. In light of the apparent differences among the different EpiSC lines revealed by this study, this equilibrium model should be restricted to the “easy-to-revert” EpiSCs (lines on the left side in Fig. 5).

The expression of Brachyury has been postulated as a characteristic of EpiSCs. However, we showed that its expression is correlated with a somewhat diminished propensity of EpiSCs to undergo reversion to an mESC-like cell state. The overexpression of Brachyury leads to a marked decrease of reversion potential. Thus, the expression of Brachyury, along with other mesendodermal marker genes, might be useful for evaluating the developmental state of a given EpiSC line that has been acquired in vitro.

A number of factors may account for the variable degree of mesendodermal gene expression in different EpiSC lines. The routine culture environment may be ruled out in this context, as all the lines used were cultured under the same conditions. The embryonic day of derivation can also be ruled out, as there was no clear correlation between the day of derivation and the other features shown in Figure 5. For example, lines OG2.1 and OG2.2 were derived as early as on E5.5, yet they exhibited high Brachyury expression. Other potentially relevant factors, which cannot be ruled out in this study, include the derivation procedure, embryo strain background, or the gender of the cell line. A recent study showed that transcriptional and epigenetic variations are common among a large set of other pluripotent stem cell lines: hESCs and human induced pluripotent stem cells [26]. The authors showed that those variations influence the propensity of a given cell line to undergo efficient differentiation. The lines used in this study may not be diverse enough to account for these different possibilities. However, considering that EpiSCs can also be derived from preimplantation blastocysts [27], a diverse array of EpiSCs can be obtained to assess the candidate determinants of EpiSC heterogeneity in a systematic manner.

CONCLUSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

We successfully demonstrated that different EpiSC lines share a panel of characteristics, regardless of their embryonic day of derivation. However, we also showed that different EpiSCs exhibit a set of divergent features. We surmised that these differences stem from line-specific distinct developmental ground states. These divergent developmental stages might be useful in the study of mammalian postimplantation development and should be taken into consideration in the functional analysis of EpiSCs with a view to directed differentiation.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

We thank Dr. Jeanine Müller-Keuker for assistance in figure composition, Drs. Gabrielle Brons and Paul Tesar for providing EpiSC lines C1a1 and T9, respectively, Martina Bleidissel and Ingrid Gelker for technical assistance, and Gerrit Fischedick, Dr. Göran Key, and Dr. Tobias Cantz for fruitful discussions. This work was supported by the Max Planck Society.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. CONCLUSION
  8. Acknowledgements
  9. DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
  10. REFERENCES
  11. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
STEM_709_sm_SuppFigure1.tif4757KFigure S1: Characterization of EpiSC lines (A) Colony morphology of EpiSC lines compared with hESCs and mESCs grown feeder-free on FCScoated dishes (EpiScs), Matrigel (hESCs), or 0.1% gelatin (mESCs). (B) H&E-stained sections of teratomas generated by EpiSC lines OG2.1, OG2.2, and E5, showing differentiation into three germ layers: ectoderm, mesoderm, and endoderm. (C) Pair wise scatter plots of EpiSC lines and mESCs compared with line E3 (reference line). The black lines are the boundaries of the 2-fold changes in gene expression levels between the paired samples. Genes upregulated in ordinates samples compared with abscissas samples are shown in red dots; those downregulated are shown in green. The color bar to the right indicates the scattering density — the higher the scattering density, the darker the blue color. Gene expression levels are log2 scaled. Indicated in the bottom-right corners are the numbers of differentially expressed genes between the lines (using a fold change of 2 in log2 scale). (D) Sextyping of EpiSC lines. Upper panel: PCR for Sry DNA showing male gender. IL-2 served as control for genomic DNA. Lower panel: Microarray expression of Y-chromosome–related gene Eif2s3y.
STEM_709_sm_SuppFigure2A.tif2347KFigure S2: Expression of early mesendodermal markers in different EpiSC lines is negatively correlated with neural induction potential but not mesodermal differentiation (A) Quantitative real-time RT-PCR analysis of gene expression of different markers in EpiSC lines after multiple passages. Bars reflect normalization errors. (B) Quantitative real-time RT-PCR analysis of early mesendodermal marker genes (Brachyury, Mixl1, Gsc, Eomes, and Sox17) as well as myogenic (Pax3) and cardiac (Nkx2.5) markers in EpiSC lines after 2 days of FCS stimulation compared with FGF control. Bars reflect normalization errors. (C) Immunocytochemistry showing α-SMA–positive cells after 14 days of mesodermal differentiation. (D) Immunocytochemistry showing β-Tubulin-III–positive cells after 4 days of neural induction (left panels: 20× right panels: 40×).
STEM_709_sm_SuppFigure2B.tif3997KFigure S2: Expression of early mesendodermal markers in different EpiSC lines is negatively correlated with neural induction potential but not mesodermal differentiation (A) Quantitative real-time RT-PCR analysis of gene expression of different markers in EpiSC lines after multiple passages. Bars reflect normalization errors. (B) Quantitative real-time RT-PCR analysis of early mesendodermal marker genes (Brachyury, Mixl1, Gsc, Eomes, and Sox17) as well as myogenic (Pax3) and cardiac (Nkx2.5) markers in EpiSC lines after 2 days of FCS stimulation compared with FGF control. Bars reflect normalization errors. (C) Immunocytochemistry showing α-SMA–positive cells after 14 days of mesodermal differentiation. (D) Immunocytochemistry showing β-Tubulin-III–positive cells after 4 days of neural induction (left panels: 20× right panels: 40×).
STEM_709_sm_SuppFigure3.tif1272KFigure S3: Reversion of OG2 EpiSCs Fluorescence and phase-contrast images of OG2 embryo (E5.5), established OG2 EpiSCs (in CM) and reverted OG2 EpiSCs (in PCL) that show reactivation of GFP. Established EpiSC lines never show spontaneous reactivation of GFP (data not shown).
STEM_709_sm_SuppTable1.doc57KTable S1: Comparison of different EpiSC lines Basic characteristics and functional differences among EpiSC lines and mESCs.
STEM_709_sm_SuppTable2.doc74KTable S2: Primer sequences.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.