SEARCH

SEARCH BY CITATION

Keywords:

  • Induced pluripotent stem cells;
  • Nuclear reprogramming;
  • Pluripotency;
  • Nanog

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

Somatic cells can be reprogrammed to induced pluripotent stem (iPS) cells by ectopic expression of the four factors Oct4, Klf4, Sox2, and Myc. Here, we investigated the role of Gata4 in the reprogramming process and present evidence for a negative role of this family of transcription factors in the induction of pluripotency. Coexpression of Gata4 with Oct4, Klf4, and Sox2 with or without Myc in mouse embryonic fibroblasts greatly impaired reprogramming and endogenous Nanog expression. The lack of Nanog upregulation was associated with a blockade in the transition from the initiation phase of reprogramming to the full pluripotent state characteristic of iPS cells. Addition of Nanog to the reprogramming cocktail blocked the deleterious effects observed with Gata4 expression. Downregulation of endogenous Gata4 by short hairpin RNAs during reprogramming both accelerated and increased the efficiency of the process and augmented the mRNA levels of endogenous Nanog. Using comparative genomics, we identified a consensus binding site for Gata factors in an evolutionary conserved region located 9 kb upstream of the Nanog gene. Using chromatin immunoprecipitation, gel retardation, and luciferase assays, we found that Gata4 bound to this region and inhibited Nanog transcription in mouse embryonic stem cells. Overall, our results describe for first time the negative effect of Gata4 in the reprogramming of somatic cells and highlight the role of Gata factors in the transcriptional networks that control cell lineage choices in the early embryo. STEM Cells2013;31:71–82


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

Reprogramming mouse embryonic fibroblasts (MEF) into induced pluripotent stem (iPS) cells by ectopic expression of the four reprogramming factors (Oct4, Klf4, Sox2, and Myc) is a stochastic and inefficient process characterized by an organized sequence of events that begins with a mesenchymal to epithelial transition (MET) associated with the downregulation of fibroblast markers [1, 2]. Simultaneously, cells activate p53-mediated stress response that leads to senescence, which constitutes the primary barrier to reprogramming [3]. Consistent with the notion that loss of replicative potential provides a barrier for reprogramming, inhibition of senescence effectors such as p16Ink4a or p19Arf or genetic ablation of p53 significantly increases the yield of iPS cell colonies [3, 4]. The cells that successfully overcome this barrier begin to express the transcriptional program characteristic of reprogrammed cells, followed by upregulation of essential pluripotent genes such as Nanog or Lin28 during the maturation and stabilization steps of the process. It is noteworthy that Nanog-deficient cells undergo the initiation phase of reprogramming but fail to acquire full reprogramming to pluripotency [5]. In fact, Nanog is essential to achieve full reprogramming to ground state pluripotency by coordinating the pluripotency networks that allow self-sustained configuration of iPS cells. Reprogramming to a full pluripotent state causes the resetting of epigenetic marks in the somatic cells to a state similar to that of embryonic stem cells (ESCs) [6].

Reprogramming to pluripotency can be accelerated by facilitating the overcoming of roadblocks to pluripotency. In this regard, Bone Morphogenetic Proteins agonists and TGFβ inhibitors increase reprogramming efficiency by favoring MET [2, 7]. Reprogramming can also be accelerated by the coexpression of transcription factors that favor the acquisition of pluripotency like Glis1 or Nanog [8, 9]. Also, combined inhibition of GSK3-beta and mitogen-activated protein kinase signaling pathways not only accelerates reprogramming but also induces partially reprogrammed cells into full pluripotency [10]. Conversely, chromatin modifying molecules such as histone deacetylases inhibitors or DNA methyltransferases enhance reprogramming, suggesting that the complete erasure of epigenetic marks could also be considered as barrier to pluripotency [11]. In conclusion, cell reprogramming efficiency can be manipulated by different strategies that target the essential roadblocks of the process.

Hhex, Foxa2, and Gata4/6 are essential for proper definitive endoderm formation [12–14]. Hhex is required for normal development of ventrally derived endoderm-related organs such as thyroid and liver. Foxa2-deficient embryos rescued for the embryonic-extraembryonic constriction, lack foregut and midgut endoderm. Gata4/6 are highly conserved, functionally redundant, and capable of binding identical nucleotide sequences in genomic DNA to regulate gene expression [15–17]. Gata4/6 are key players for early cell fate decisions in the mouse inner cell mass (ICM) when cells segregate into epiblast and primitive endoderm. In fact, formation of these embryonic layers is governed by the activity of the transcription factors Nanog and Gata4/6 [18]. Nanog protein is initially found throughout the early embryo [19] and then becomes restricted to a subset of cells in the ICM, the epiblast progenitors. It is proposed that induction of Gata4/6 by Fgf signaling through Grb2 may promote primitive endoderm in a “salt and pepper” distribution of Nanog-positive and Gata-positive cells in the ICM. Nanog and Gata4/6 are not coexpressed in any cell throughout the E3.5 blastocyst [19], suggestive of reciprocal inhibition. The mechanisms exerted by Gata4/6 to downregulate Nanog expression and override epiblast commitment are still unclear.

Here, we investigated the role of Gata factors in the reprogramming of somatic cells and found that these factors constitute a negative control of the process. We show for the first time that Gata4 binds to a distal enhancer 9 kb upstream of the Nanog gene and inhibits its transcription. Our studies support the hypothesis that Gata4 expression in ICM constitutes a mechanism to inhibit Nanog gene expression, allowing the segregation of primitive endoderm in the early embryo.

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

Plasmids and Sequence Analysis

The retroviral construct pMIGR1-Hhex was generated by subcloning the Hhex cDNA into the XhoI-NotI restriction sites of pMIGR1 (a kind gift from Dr. Pellicer, NYU Medical Center). pPYCAG-IP-HAGata4 was obtained by PCR using the primers described in supporting information Table S2 and pBABE-Gata4 (a kind gift from Dr. Ken Zaret, University of Pennsylvania) as template. pBABE-Foxa2 and pPYCAG-IP were kind gifts from Dr. Ken Zaret and Dr. Austin Smith (Wellcome Trust Centre for Stem Cell Research, U.K.), respectively. To generate pBabe-Gata6, mouse Gata6 was amplified by PCR from mouse lung cDNA and subcloned into the BamHI-SalI sites of pBabe-puro. To obtain the pGL3-Nanog-enh/prom reporter plasmid, the enhancer region of mouse Nanog was amplified by PCR from mouse genomic DNA (supporting information Fig. S10) and subcloned into the KpnI-SacI sites of the pGL3-Nanog vector (a kind gift from Dr. Paul Robson [20]). All constructs were sequence verified. The plasmids encoding the reprogramming factors pMXs-Oct4, pMXs-Sox2, pMXs-Klf4, pMXs-Myc, pMXs-Nanog, and pMXs-RFP were from Addgene (Cambridge, MA; www.addgene.com).

Evolutionary conserved regions (ECR) within the Nanog gene were extracted by multiple alignments between mouse (mm9), human (hg19), monkey (rheMac2), opossum (monDom4), pufferfish (fr2), junglefowl (galGal3), and frog (xenTro2) genomes available using the ECR Browser web-based tool (http://ecrbrowser.dcode.org) [21]. We searched the ECR identified above for transcription factor binding sites that showed more than 85% sequence similarity using rVISTA 2.0 [22].

Cell Culture and Cell Imaging

CCE1.19 mouse ESCs were a kind gift of Dr. M. Gassman [23]. Cells were grown in ESC medium as described [24]: Glasgow minimum essential medium supplemented with 10% fetal bovine serum (FBS), 0.1 mM β-mercaptoethanol, 1× nonessential amino acids, 1× sodium pyruvate, 1× penicillin/streptomycin (Life-Technologies, Carlsbad, CA, www.invitrogen.com), and 2 mM glutamine in the presence of Leukemia Inhibitory Factor. MEF were isolated from E13 mouse embryos from pregnant CD1 and Tg(Nanog-GFP:Puro) mice [25] by trypsin digestion. MEF were maintained in Dulbecco's modified Eagle's medium with Glutamax, supplemented with 10% FBS and 1× penicillin/streptomycin (Life-Technologies, Carlsbad, CA, www.invitrogen.com). When indicated, colonies were hand-picked and maintained in 2i+LIF medium, that is, ESC medium containing 3 μM of the GSK3 inhibitor CHIR99021and 1 μM of the MAPK inhibitor PD0325901 [26]. All cells were maintained at 37°C with 5% CO2 and were regularly examined with an Olympus CKX41 microscope. Images were taken with a Leica DFC350 FX digital camera connected to a Leica DMI 4000 B fluorescence microscope using Leica Application Suite software.

Generation of Virus, Viral Infections, and Reprogramming

Ecotropic retroviruses were generated in 293T cells by cotransfection of 10 μg of shuttle vector with 10 μg of pCL-eco (replication-incompetent helper vector pCL-eco) by the calcium phosphate method. Supernatants containing the retroviral particles were collected after 48 hours, filtered through a 0.45 μm nitrocellulose filter, supplemented with 8 μg/ml polybrene, and immediately used. Reprogramming of somatic cells was done as described [27].

For the knockdown of the endogenous Gata4, a collection of target sequences of short hairpin RNAs (shRNA) in vector pLKO.1 (RMM4534) was purchased from Open Biosystems (Thermo Scientific Molecular Biology, Waltham, MA, http://www.thermoscientificbio.com/openbiosystems). A shRNA vector against luciferase (pLKO.1-shLuc) was used as a negative control. For lentivirus production, 293T cells were cotransfected with 10 μg of pLKO.1 shuttle vector, 7.5 μg of psPAX2, and 5 μg of pMD2.G. Supernatants containing the viruses were collected 48 hours post-transfection, filtered through a 0.45 μm nitrocellulose filter, supplemented with 8 μg/ml polybrene, and immediately used.

Reverse Transcription and Real-Time Quantitative RT-PCR

Total RNA was extracted using the RNeasy mini kit (Qiagen, Hilden, Germany, www.qiagen.com) and reverse-transcribed using Moloney Murine Leukemia Virus reverse transcriptase (Life-Technologies, Carlsbad, CA, www.invitrogen.com) according to the manufacturer's protocol. PCR amplification was performed using expand high fidelity PCR system (Roche, Basel, Switzerland; www.roche-applied-science.com) following the manufacturer's instructions. Quantitative RT-PCR (qRT-PCR) was run on a Light Cycler 480 II Real-Time PCR System (Roche, Basel, Switzerland; www.roche-applied-science.com) using the Light Cycler 480 SYBR Green I Master (Roche, Basel, Switzerland; www.roche-applied-science.com). The PCR reaction consisted of 7.5 μl of SYBR Green Master I, 0.75 μl of 6 μM of forward primer, 0.75 μl of 6 μM reverse primer, 3 μl water, and 3 μl template cDNA (1/20) in a total volume of 15 μl. Cycling was performed 10 minutes at 95°C, followed by 40 rounds of 10 seconds at 95°C, 10 seconds at 57°C–64°C and 72°C for 20 seconds, and extension at 72°C for 5 minutes. The relative expression of each gene was normalized against mouse Beta-actin (Actb). The specificity of the amplified PCR products was confirmed by analysis of the melting curve and agarose gel electrophoresis. Primers used for the qRT-PCR are shown in supporting information Table S2.

Immunoblotting

Protein extracts (total, nuclear, or cytoplasmic) were obtained by M-PER or NE-PER protein extraction reagent following the manufacturer's instructions (Pierce-Thermo Fisher Scientific, Waltham, MA, www.piercenet.com). Samples were quantified with Coomassie Plus Protein Assay (Pierce-Thermo Fisher Scientific, Waltham, MA, www.piercenet.com). Equal quantities of protein were loaded for each sample and subjected to polyacrylamide gel electrophoresis. Gels were subsequently transferred to polyvinylidene fluoride membranes using an iBlot gel transfer system (Life-Technologies, Carlsbad,CA, www.invitrogen.com). Polyvinylidene fluoride membranes were blocked in TBS-T containing 5% nonfat milk powder and incubated with primary antibody overnight at 4°C. Next day, membranes were washed followed by incubation with the secondary antibody for 45 minutes at RT. Immunoreactive proteins were visualized using the ECL Western Blotting substrate (Pierce). Commercial antibodies used were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, www.scbt.com): anti-Gata4 (sc#9053), anti-HA(sc#57592), and anti-Tubulin (sc#8035).

Luciferase Reporter Assay

Cells were transfected using Lipofectamine 2000 (Life-Technologies, Carlsbad, CA, www.invitrogen.com) following manufacturer's protocol. Renilla luciferase plasmid (pRL-TK from Promega, Fitchburg, WI, www.promega.com) was cotransfected to correct variations in transfection efficiency. The luciferase activity in the lysate was measured using the Dual-Luciferase reporter assay system (Promega, Fitchburg, WI, www.promega.com) in a 96-well luminometer reader (Berthold detection systems, Bad Wildbad, Germany, www.berthold.com).

Low Cell Number Chromatin Immunoprecipitation

Chromatin immunoprecipitation (ChIP) using low cell numbers was done as described [28]. Primers are described in supporting information Table S2. Specificity of the Gata4 and HA antibody was validated by immunoblotting as shown in supporting information Figure S11.

Bromodeoxyuridine Incorporation Assay

Cell proliferation was assessed using the Cell Proliferation ELISA, BrdU (colorimetric) kit (Roche, Basel, Switzerland, www.roche-applied-science.com) following manufacturer's instructions.

Alkaline Phosphatase Staining

Alkaline phosphatase activity was assessed using the Alkaline Phosphatase Detection Kit from EMD Millipore (Billerica, MA, www.millipore.com) following manufacturer's instructions.

Electrophoretic Mobility Shift Assays

Nuclear extracts from CCE1.19 mouse ESCs were obtained using NE-PER protein extraction reagent following the manufacturer's instructions. Probe labeling, gel shift reaction, transfer, and detection were done using the DIG Gel Shift kit, second Generation (Roche, Basel, Switzerland, www.roche-applied-science.com), with minor modifications. DNA binding reactions consisted 15 μg of nuclear extract, 157 fmol labeled probe, 1 μg/μl poly(L-lysine), and 1 μg/μl poly(dIdC) in a final volume of 20 μl buffer (12 mM HEPES pH 7.9, 4 mM Tris-HCl pH 8.0, 12 mM glycerol, 5 mM MgCl2, 60 mM KCl, 0.6 mM EDTA pH 7.9, 0.28 mg/ml bovine serum albumin, and 0.6 mM dithiothreitol). When indicated, 1 μM unlabeled double-stranded competitor was included prior to the addition of nuclear extracts. Binding reactions were incubated for 20 minutes at room temperature. Where indicated, 2 μl of anti-Gata4 antibody (Santa Cruz Biotechnology) was added 10 minutes after the addition of the nuclear extracts and left to proceed for 10 minutes more; after this time, reactions were stopped and further processed as indicated below. Binding reactions were resolved on a prerun 6% nondenaturing Tris-borate-EDTA (TBE)-polyacrylamide gel (16.5 × 20 cm2) in 0.5× TBE at 20 mA (constant current). Gel was electroblotted in 0.4% TBE buffer using Hybond N + transfer membrane during 2 hours at 400 mA with a semi-dry blotter. DNA was crosslinked by UV-light (306 nm) in a Gel-Logic 200 transilluminator for 12 minutes.

Microarray Hybridization and Analysis

Total RNA was extracted from mouse ESCs, MEF, and isolated clones of MEF infected with distinct retroviral combinations using RNeasy extraction kit (Qiagen, Hilden, Germany, www.qiagen.com). RNA was then quantified and analyzed using the Agilent 2100 bioanalyzer. RNA Integrity number (RIN) values range from 10 (intact) to 1 (totally degraded), and our samples had a RIN >9. Gene expression analysis was performed using Affymetrix Mouse Gene 1.0ST array GeneChips containing probes for more than 28,000 genes. Total RNA (300 ng per sample) was labeled using the Affymetrix Gene Chip WT cDNA Synthesis and Amplification kit protocol, and hybridized to the arrays as described by the manufacturer (Affymetrix, Santa Clara, CA, http://www.affymetrix.com). The complementary RNA hybridization cocktail was incubated overnight at 45°C while rotating in a hybridization oven. After 16 hours of hybridization, the cocktail was removed and the arrays were washed and stained in an Affymetrix Gene Chip fluidics station 450, according to the Affymetrix-recommended protocol. Arrays were scanned on an Affymetrix Gene Chip Scanner 3000 7G. Three different samples per group were examined. Data (.CEL files) were normalized using the robust multiarray algorithm [29]. Next, we used a conservative probe-filtering step excluding probes not reaching a log2 expression value of 5 in at least one sample, which resulted in the selection of a total of 19,621 probes from the original set of 28,853. To identify genes differentially expressed between the different microarray study groups, we used linear models for Microarray Data R-package [30].

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

Foxa2 and Gata4 Block Nanog Upregulation During Reprogramming

We sought to investigate the role of endodermal transcription factors in the reprogramming of somatic cells. We selected three transcription factors essential for proper differentiation of the definitive endoderm, Foxa2, Gata4, and Hhex [12–14]. MEF were subjected to a reprogramming protocol by introducing these endoderm-specific transcription factors together with the reprogramming factors Oct4, Klf4, and Sox2 in the presence or absence of Myc (OKSM or OKS, respectively) by retroviral transduction (supporting information Table S1). The specific time point in which colonies were readily detected in the cultures was different for each condition. In particular, the combination of FoxA2 and Gata4 with OKSM (OKSM+F+G) delayed the process significantly. Infection of MEF with the OKS or OKSM combination of retroviruses rendered colonies with a distinct edge, containing cells of typical mouse ESC morphology with a round shape, large nucleoli, and scant cytoplasm. However, the addition of Foxa2 together with Gata4 to the OKS or OKSM mixtures of retroviruses resulted in granular colonies with irregular and nondistinctive borders, composed of small cells resembling fibroblasts rather than ESCs (supporting information Fig. S1A). When colonies acquired a critical size they were hand-picked and expanded independently for molecular analysis (six colonies per group).

Nanog expression is a hallmark of pluripotency [5, 31]. To assess which clones had achieved full reprogramming, we determined the levels of Nanog mRNA by qRT-PCR after four consecutive passages. Clones obtained with the OKSM combination (OKSM-clones) showed the highest levels of Nanog mRNA (Fig. 1A). Addition of Hhex alone (H) or in combination with Foxa2 (H+F) to the reprogramming retroviral cocktails rendered clones with either similar or slightly diminished Nanog mRNA expression compared to their OKS- or OKSM-clone counterparts, suggesting that both factors do not severely limit fibroblast reprogramming. Remarkably, addition of Foxa2 and Gata4 (F+G) to the OKS or OKSM reprogramming cocktails rendered colonies with no expression of Nanog mRNA (Fig. 1A), suggesting that these two factors impair the reprogramming of MEF. Nonetheless, the absence of Nanog mRNA could also be the consequence of the differentiation of early-stage Nanog-positive cells into Nanog-negative cells. To explore this possibility, we performed a time course analysis of Nanog mRNA expression during the first 4 weeks after the initial transduction of the cells with the retroviral mixtures (Fig. 1B). In the cultures infected with OKSM, Nanog mRNA levels became detectable by day 7 (inset in Fig. 1B) and increased steadily until day 28. However, Nanog expression levels remained below the detection limit in MEF infected with OKSM+F+G at all time points. These results suggest that the addition of Foxa2 in combination with Gata4 impairs reprogramming and the expression of Nanog in MEF subjected to OKS/M transduction.

thumbnail image

Figure 1. Gata4 and Foxa2 block Nanog upregulation during reprogramming without alternative cell commitment. (A): MEF were subjected to a reprogramming protocol using the indicated combinations of transcription factors (supporting information Table S1). Nanog mRNA levels were quantified by quantitative RT-PCR (qRT-PCR) in three clones isolated from each group. (B): MEF were infected with the indicated combination of transcription factors and the levels of Nanog mRNA measured at specific time points by qRT-PCR. Data are represented as the average ± SD of three replicates relative to the Nanog mRNA levels found in clon #1 from OKSM in (A) and in OKSM after 28 days in (B). (C): Heat map depicting relative expression levels of selected mRNAs across MEF, mESCs, and five clones isolated from MEF reprogrammed with OKS in the presence of Foxa2- and Gata4-expressing retroviruses. Genes were selected based on their expression profiles in specific embryonic tissue layers. Black: not expressed; red: higher expression; green: lower expression. Microarrays data have been deposited in NCBI's Gene Expression Omnibus [32] and are accessible through GEO Series accession number GSE37548 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc = GSE37548). Abbreviations: MEF, mouse embryonic fibroblasts; mESC, mouse embryonic stem cells; OKSM, Oct4, Klf4, and Sox2 with Myc.

Download figure to PowerPoint

Foxa2 and Gata4 Do Not Reprogram MEF into Endoderm-Like Lineages

ESC differentiation is accompanied by Nanog downregulation [24]. Given the pivotal role of Foxa2 and Gata4 in embryonic endoderm formation [12, 13, 33-35], we speculated that isolated clones from OKS/M+F+G were in fact endodermal progenitor cells or cells committed toward an embryonic endodermal lineage, that is, definitive, visceral, parietal, extraembryonic, or primitive. To test this hypothesis, we performed Affymetrix gene expression profiling of the isolated clones obtained after MEF reprogramming in presence of Foxa2 and Gata4, using MEF and mouse ESCs as controls. We compiled the relative expression levels of specific markers for each embryonic layer into a color-coded graph focusing in endoderm-related genes (Fig. 1C). The expression levels of Lamb1, Sparc, Sox7, Afp, and Ttr were validated by qRT-PCR (supporting information Fig. S1B). Sox7, Afp, and Ttr, which are highly expressed in visceral endoderm, were not detected. Likewise, expression levels of markers for all other endodermal cell types indicate that the cell clones obtained with the OKS+F+G combination are not committed to an endodermal fate.

Next, the pluripotent state of the OKS+F+G-clones was investigated using global gene profiling. To do so, we extended our gene profile analysis to clones isolated from OKSM infection that showed very low or no expression of Nanog by qRT-PCR. We used hierarchical cluster analysis to sort the different mRNA samples based on expression of 19,621 probes expressed above background level (supporting information Fig. S2). OKS+F+G-clones were clustered together with Nanog-negative OKSM-clones, indicating that their expression patterns are more similar than those of nonreprogrammed cell types (MEF and mouse ESCs).

To get a clearer picture of the pluripotent status of our OKS+F+G-clones, we performed an additional analysis of the microarray data based on a selected battery of genes. In this regard, Mikkelsen et al. dissected the transition from MEF to iPS cells by an integrative genomic approach [36]. They found that stable partially reprogrammed cell lines show the reactivation of a distinctive subset of stem-cell-related genes and an incomplete repression of lineage-specifying transcription factors, suggesting that some cells may become trapped in partially reprogrammed states due to incomplete repression or activation of specific transcription factors. In agreement with these findings, the analysis of the gene expression profile of the Nanog-negative OKSM-clones showed their identity as partially reprogrammed iPS (pr-iPS) cells (Fig. 2A). Conversely, OKS+F+G-clones did not downregulate structural genes such as Col6a2 or growth arrest mediators such as Gas1. In addition, upregulation of genes directly linked to early pluripotency such as Zic3 and Fgf4, did not take place in the OKS+F+G-clones while the upregulation of several pluripotency markers such as Utf1, Gdf3, or Nodal, was very limited. Our results show that although the OKS+F+G-clones displayed the expression of several pluripotency markers at low levels, they are not pr-iPS cells. A new hierarchical analysis based on this subset of genes, clustered the five OKS+F+G-clones together and at a lower relative distance to MEF than to the Nanog-negative OKSM clones, indicating that their expression patterns are similar to nonreprogrammed MEF (Fig. 2B).

thumbnail image

Figure 2. MEF reprogrammed in the presence of Gata4 and Foxa2 are not partially reprogrammed induced pluripotent stem cells. (A): Heat map depicting relative expression levels of selected mRNAs across MEF, mESC, Nanog-negative clones from OKSM-reprogrammed MEF, and five clones isolated from MEF reprogrammed with OKS in the presence of Foxa2- and Gata4-expressing viruses (OKS+F+G). Genes were selected based on a previous report [36]. (B): Dendogram plot based on Euclidean distance and average linkage showing hierarchical clustering or closeness in expressed mRNAs shown in (A) (52 probes). (C): Phase-contrast images of a representative clone isolated from MEF reprogrammed with the indicated combination of transcription factors growing for 5 and 10 days in 2i+LIF medium (original magnification ×10) and induced to differentiate as EB (original magnification ×20). Abbreviations: EB, embryoid bodies; Leukemia Inhibitory Factor; MEF, mouse embryonic fibroblasts; mESC, mouse embryonic stem cells; OKSM, Oct4, Klf4, and Sox2 with Myc.

Download figure to PowerPoint

Pluripotency of reprogrammed cells can be definitively assessed by culturing the iPS cell-like clones in a selective serum-free medium containing inhibitors of the mitogen-activated protein kinase and GSK3 kinases in the presence of Leukemia Inhibitory Factor (2i+LIF) [26]. In this medium, partially reprogrammed cells will be promoted to full pluripotency while null-pluripotent cells will eventually die [10]. Nanog-negative OKSM- and OKS-clones grew in the presence of 2i+LIF (Fig. 2C), acquired a typical ESC morphology, expressed pluripotency markers such as Lin28, Nanog, Oct4, or Prdm14, and formed uniform spheres when induced to differentiate as embryoid bodies, all characteristics of fully pluripotent cells. OKS/M+F+G-clones did not survive in this selective media. In conclusion, addition of Foxa2 in combination with Gata4 to the OKS/M reprogramming cocktail results in the formation of proliferative clones that show low expression of pluripotency genes and are different from the previously described pr-iPS cells.

Gata4 But Not Foxa2 Expression Abolishes Nanog Induction and Clonal Expansion During Reprogramming

Reprogramming blockade by Foxa2 and Gata4 is specific, since the addition of Hhex alone or in combination with Foxa2 did not affect the expression of Nanog induced by infection of MEF with the OKS/M mixture of viral supernatants (Fig. 1A). To identify the transcription factor responsible for the impairment of Nanog induction in the reprogramming assays, we infected MEF with OKSM alone or together with Foxa2 (OKSM+F) or Gata4 (OKSM+G) and assessed the induction of pluripotency by analyzing Nanog expression. Nanog mRNA was induced in cells transduced with OKSM or OKSM+F, while no expression of this pluripotency marker was observed with the addition of Gata4 (Fig. 3A). In agreement with the observed Nanog expression profile, the addition of Foxa2 to the reprogramming cocktail induced the formation of cell aggregates that resembled ESC colonies and displayed expression of pluripotency markers after extensive culture (n = 6) (data not shown). Cell colonies obtained upon transduction of MEF with the OKSM+G showed a fibroblast-like morphology and could not be isolated or further expanded in culture (n = 20). Moreover, transduction of MEF derived from the Tg (Nanog-GFP:Puro) mice [25] with OKSM+F activated the Nanog-GFP transgene, whereas reporter activation was not observed in the presence of Gata4 in the retroviral mixture (Fig. 3B). To evaluate pluripotency, we cultured MEF infected with the OKSM+G or OKSM+F retroviral mixtures in 2i+LIF medium and assessed their morphology for 10 days after switching to the selective media. Cells infected with OKSM+G died within 7 days (Fig. 3C). In sharp contrast, 2i+LIF medium favored the growth of multiple colonies in the cells infected with OKSM+F.

thumbnail image

Figure 3. Foxa2 significantly increases Nanog mRNA levels during reprogramming. (A): MEF were infected with OKSM together with Foxa2- (OKSM+F) or Gata4-expressing viruses (OKSM+G) and Nanog mRNA levels were analyzed at the specified time points by quantitative RT-PCR. Data are represented as the average ± SD of three replicates relative to the Nanog mRNA levels found in OKSM at day 31 of the procedure. (B): MEF from Tg(Nanog-GFP:Puro) transgenic reporter mice [25] were infected with OKSM+F or OKSM+G. Phase-contrast (left panels) and fluorescence (right panels) images of representative colonies are shown (original magnification ×10). No GFP-positive colonies were observed upon OKSM+G infection of Nanog-GFP:puro MEF. (C): Phase-contrast images of a representative clone isolated from MEF reprogrammed with OKSM+F or OKSM+G grown for the indicated time points in 2i+LIF selective medium (original magnification ×10). Abbreviations: GFP, green fluorescent protein; Leukemia Inhibitory Factor; MEF, mouse embryonic fibroblasts; OKSM, Oct4, Klf4, and Sox2 with Myc.

Download figure to PowerPoint

Our results suggest that Gata4 is a negative regulator of somatic cell reprogramming. In agreement with this hypothesis, it has been described that Gata4 expression is induced at the initial stages of reprogramming and maintained until the last stages of reprogramming. In addition, iPS cells have been shown to express Gata4 mRNA at lower but detectable levels [2]. Therefore, we reasoned that downregulation of Gata4 should increase the efficiency of cell reprogramming. To test this, we generated lentiviral vectors expressing shRNAs specific for mouse Gata4 and validated their ability to downregulate Gata4 expression in 3T3 cells (supporting information Fig. S3). All four lentiviral constructs reduced the amount of Gata4 protein, and the shRNA4 construct was selected for subsequent experiments due to its specificity in downregulating the expression of this Gata protein in comparison with a negative control shRNA against Luciferase (shLuc). We next assayed the effect of Gata4 downregulation on the efficiency of reprogramming induced by infection of MEF with OKSM using the induction of alkaline phosphatase activity and Nanog expression as readouts (Fig. 4). In agreement with our previous results, forced expression of Gata4 impaired the appearance of alkaline phosphatase-positive colonies (Fig. 4A) and Nanog mRNA expression. Compared to the shLuc-negative control, downregulation of Gata4 expression during MEF reprogramming (shGata4) accelerated the appearance of colonies during reprogramming. The ratio of colonies for control (shLuc) and shGata4 was 1:3 at day 8, respectively (Fig. 4A) and yielded clones with a statistically significant increase in Nanog mRNA (Fig. 4B). Altogether, our results suggest that Gata4 prevents Nanog upregulation and impairs MEF reprogramming.

thumbnail image

Figure 4. Downregulation of Gata4 favors cell reprogramming. (A): MEF were infected with the indicated combination of transcription factors and lentiviral vectors expressing shRNAs for Luciferase (shLuc) or Gata4 (shGata4) and stained for alkaline phosphatase activity. Photographs show an increase in the number of colonies positive for alkaline phosphatase staining in plates that included lentiviruses expressing shGata4. (B): Bar diagram showing the quantitation of Nanog mRNA levels in isolated clones (three from each condition) from MEF transduced as indicated in (A). Downregulation of Gata4 during reprogramming significantly increased Nanog expression. Data are represented as the average ± SD of three replicates relative to the Nanog mRNA levels found in OKSM-derived clones. Abbreviations: MEF, mouse embryonic fibroblasts; OKSM, Oct4, Klf4, and Sox2 with Myc.

Download figure to PowerPoint

Gata4 and Gata6 play semiredundant roles in liver and cardiac development partially by regulating the expression of an overlapping set of genes [37–39]. To test whether Gata6 could also limit cell reprogramming, we coexpressed Gata6 with the OKSM reprogramming factors in MEF and assessed the formation of iPS cell-like colonies by both morphology and an alkaline phosphatase assay. Gata6 induced the formation of cell aggregates lacking ESC morphology which did not express alkaline phosphatase (supporting information Fig. S4), suggesting that both Gata factors are negative regulators of somatic cell reprogramming.

Gata4 Triggers a Selective Blockade of Reprogramming Cornerstones

We reasoned that inclusion of Gata4 in the reprogramming cocktail could induce MEF differentiation into embryonic endoderm. As previously seen with the reprogramming of MEF in the presence of Foxa2 and Gata4 (Fig. 1C and supporting information Fig. S1B), we did not detect expression of critical markers of visceral and parietal endoderm such as Ttr, Sox7, and Afp and the expression of the endodermal markers Lamb1 and Sparc remained below that observed in MEF (supporting information Fig. S1C). These results indicate that in the presence of Gata4, MEF do not differentiate toward endoderm. Conversely, the expression of Gata4 could limit the G1-S transition during reprogramming and hence proliferation [40]. To investigate this possibility, we performed a BrdU incorporation assay in MEF control and infected with Gata4. We did not observe an effect on BrdU incorporation upon Gata4 expression (supporting information Fig. S5), excluding a possible G1-S blockade.

Previous reports have characterized the main cornerstones of MEF reprogramming before Nanog is upregulated [41]. Such cornerstones include retroviral silencing, overcoming senescence and initiation of MET. To monitor retroviral silencing, we included a retrovirus expressing green fluorescent protein (GFP) in the initial reprogramming cocktail and checked the expression of both GFP and the exogenous factors Oct4, Klf4, and Myc. Compared with the OKSM control, the inclusion of Gata4 failed to silence the retroviral GFP reporter (Fig. 5A). Also, cultures infected with OKSM+G did not switch off the expression of the exogenous Oct4, Klf4, and Myc retroviral constructs after 31 days (Fig. 5B). In agreement with the observed failure in retroviral silencing, endogenous pluripotency markers such Oct4, Nanog, or Lin28 were not induced (Fig. 5C). Senescence has been considered as an initial barrier that should be overcome to initiate reprogramming [3], therefore we assessed whether Gata4 interfered with this process. Upon infection, we observed a Gata4-independent upregulation of p16Ink4a and p19Arf [42] mRNA between days 6 and 8 (supporting information Fig. S6). However, we found that the mRNA levels of p16Ink4a were significantly higher at day 17 in the OKSM+G-infected MEF, when reprogramming is already entering its maturation phase. MET is also considered a key event for the initiation of reprogramming [1, 2]. We analyzed the mRNA levels of the epithelial markers Cdh1 and Ocln1 and the mesenchymal markers Snai1 and Zeb1 in the cultures subjected to a reprogramming assay in the absence (OKSM) or presence of Gata4 (OKSM+G). Upregulation of Cdh1 and Ocln1 in parallel with downregulation of Snai1 and Zeb1 confirms MET initiation in OKSM and OKSM+G cultured MEF (supporting information Fig. S7). In summary, Gata4 blocks somatic cell reprogramming without significantly interfering with some of the initial molecular cornerstones of reprogramming such as G1-S transition, senescence, or MET.

thumbnail image

Figure 5. Gata4 blocks somatic cell reprogramming. (A): MEF were infected with OKSM in the absence or in the presence of Gata4-expressing virus (OKSM+G) and a retrovirus expressing GFP to monitor retroviral silencing. Representative fluorescence (GFP) and PhC images of the cultures at the indicated time points are shown (original magnification ×10). Retroviral silencing (GFP-negative) in the OKSM-infected cultures is already visible in some colonies at day 8 (red arrowhead). No GFP-negative colonies were observed in OKSM+G-infected MEF. (B): Bar diagrams showing transcript levels of the indicated exogenous reprogramming factors in MEF reprogrammed in the absence (OKSM) or presence (OKSM+G) of Gata4-expressing virus. mRNA levels in MEF infected with OKSM and cultured for 5 days were set to 1. (C): Bar diagrams displaying the transcript levels of the indicated endogenous pluripotency genes by quantitative RT-PCR (qRT-PCR) in MEF reprogrammed as in (B). (D): Bar diagrams showing the analysis by qRT-PCR of Nanog mRNA levels in MEF reprogrammed with OKSM and reinfected with GFP- or Gata4-expressing viruses as specified in the text. Data are represented as the average ± SD of three replicates. Values are relative to the transcript levels of the indicated exogenous cDNAs found in MEF infected with OKSM and cultured for 5 days in (B), the specified endogenous genes in iPS cells in (C), and the Nanog gene in MEF infected with OKSM and cultured for 31 days in (D). Abbreviations: EB, embryoid bodies; iPSC, induced pluripotent stem cell; MEF, mouse embryonic fibroblasts; mESC, mouse embryonic stem cells; OKSM, Oct4, Klf4, and Sox2 with Myc.

Download figure to PowerPoint

To gain insight into the mechanism of impairment of cell reprogramming by Gata4, we investigated whether Gata4 was also effective when introduced during the maturation phase of reprogramming. OKSM-infected MEF were split as described in detail elsewhere (supporting information Fig. S8A). Reinfection efficiency was assessed 48 hours later by measuring expression of exogenous Gata4 (Gata4-exo) and GFP/red fluorescence protein (RFP) by qRT-PCR and fluorescence microscopy, respectively (supporting information Fig. S8B). We could not detect expression of Nanog 21 days after reinfection (a total of 28 days after the initial transduction of MEF with OKSM) when Nanog mRNA levels are readily detected in control plates (Fig. 5D). We conclude that interference of Gata4 with cell reprogramming is effective when introduced at the maturation stage of the process, at approximately day 8 of reprogramming [2].

Rescue of the Gata4-Imposed Blockade of Reprogramming by Nanog

The results shown above suggest that Gata4 exerts a direct effect over an essential step of the maturation phase of reprogramming, not initiation. Acquisition of pluripotency during reprogramming can be monitored by measuring the upregulation of pluripotency- and not pluripotency-related genes such as Syne2, Perp, Slc38a5, or Gdf3 [36]. mRNA levels of Syne2, Perp, Slc38a5, and Gdf3 were initially upregulated in OKSM and OKSM+G-treated MEF (supporting information Fig. S9 and Fig. 6C lower left). However, OKSM+G-infected MEF failed to downregulate Slc38a5 or upregulate Gdf3 8 days after the viral transduction of the cultures, as observed in OKSM-infected cultures and in bona fide iPS cells [2]. These results support the notion that Gata4 impairs reprogramming by targeting the maturation stage.

thumbnail image

Figure 6. Gata4 inhibition of cell reprogramming is rescued by forced expression of Nanog. (A): Bar diagram showing the time course expression of exogenous Nanog mRNA in MEF infected with OKSM together with Gata4- and Nanog-expressing viruses (OKSM+G+N). Bar in mESC refers to endogenous Nanog. (B): Phase-contrast images of representative clones isolated from cells transduced with the indicated combination of viruses and grown for the indicated time points in 2i+LIF medium and induced to differentiate as EB (right most panels) (original magnification ×10). (C): Bar diagram showing the quantitative RT-PCR analysis of the mRNA levels of the indicated endogenous pluripotency genes in Nanog-rescued OKSM+G cells. Data are represented as the average ± SD of three replicates relative to the transcript levels of Nanog found in undifferentiated ESCs in (A) and the specified endogenous genes found in iPS cells (C). Abbreviations: EB, embryoid bodies; iPSC, induced pluripotent stem cell; MEF, mouse embryonic fibroblasts; mESC, mouse embryonic stem cells; OKSM, Oct4, Klf4, and Sox2 with Myc; PhC, phase-contrast.

Download figure to PowerPoint

Nanog is initially dispensable during reprogramming but becomes essential for dedifferentiated intermediates to transit to a stable ground state of pluripotency [5, 31]. We hypothesized that regulation of Nanog gene expression was the target for the inhibitory effect of Gata4 in the reprogramming of somatic cells. We asked whether exogenous expression of Nanog could reverse the blockade to reprogramming induced by Gata4 expression. To test this, we first infected MEF with the OKSM+G combination in the absence or presence (OKSM+G+N) of Nanog-expressing viruses and monitored colony formation and upregulation of pluripotency markers (Fig. 6). Initial exogenous Nanog (Nanog-exo) mRNA levels were comparable to that found in undifferentiated ESCs (Fig. 6A). Transduction of MEF with the OKSM+G mixture rendered no colonies (Fig. 6B, upper panels) and did not upregulate the expression of pluripotency markers (Fig. 6C). Remarkably, addition of Nanog to the OKSM+G retroviral cocktail yielded iPS cell-like colonies within 10 days that were able to form embryoid bodies when induced to differentiate as cell aggregates (Fig. 6B, lower panels). Moreover, OKSM+G+N-infected MEF upregulated the expression of the pluripotency markers Oct4, Lin28, and Gdf3 and enabled transcriptional activation of the endogenous Nanog locus (Fig. 6C). Consistent with Nanog being negatively regulated by Gata4, Nanog expression rescued OKSM+G infected fibroblasts.

Gata4 Represses Nanog Transcription by Binding to a Distal Enhancer

We searched for ECR within the mouse Nanog gene and identified four ECRs located upstream of the transcription start site (supporting information Fig. S10A). ECRs I to III are conserved in mouse, human, and monkey. ECR I (97 bp in length) contains a composite Oct-Sox cis-regulatory element essential for Nanog expression in pluripotent cells [20]. In silico analysis of the ECR II (180 bp) and III (635 bp) regions revealed a scarcity of known transcription binding sites. ECR IV (732 bp), located approximately 9,000 bp upstream of the transcription start site, is conserved among all species included in the analysis. Analysis of the ECR IV region using the rVista2.0 software revealed the presence of 66 transcription factor binding motifs, suggesting a possible role as a distal enhancer. Among these binding motifs, we identified a consensus binding site for Gata factors, core sequence (A/T)GATA(A/G), located at −9,180 bp from the transcription start site, within the longest uninterrupted homologous sequence of 14 bp (supporting information Fig. S10B).

We investigated whether Gata4 binds to this region of the Nanog gene and regulates its transcription using luciferase and electrophoretic mobility shift assays (EMSA). We constructed a luciferase reporter plasmid containing the complete sequence of the distal enhancer ECR IV upstream the Nanog proximal promoter (ECR I). The chimeric Nanog promoter/enhancer was active in mouse ESCs and its activity decreased in a dose-dependent manner to 50% when cotransfected with an episomal vector expressing Gata4 (pPYCAG-HAGata4) (Fig. 7A).

thumbnail image

Figure 7. Gata4 binds to a novel distal enhancer in the Nanog gene and represses its transcription. (A): Luciferase assays in undifferentiated embryonic stem cells (ESCs) transfected with the pGL3-Nanog-enh/prom reporter and empty vector or pPYCAG-HAGata4. Data are represented as the means ± SEM of the firefly/renilla ratios of three experiments conducted in duplicate relative to cells transfected with the empty vector. Upper diagram, representative immunoblot analysis of the expression of Gata4 in cell extracts used for the luciferase experiments. (B): Electrophoretic mobility shift assay (EMSA) using nuclear extracts from ESCs transfected with pPYCAG (Cont) or pPYCAG-HAGata4 (Gata4). Lanes 1–5 show the binding of Gata4 to the Nanog enhancer identified in this study. Lanes 6–9 display the weak binding of an endogenous Gata factor to the same enhancer sequence. Lanes 10–12 show the binding of Gata4 to an oligo based on the proximal promoter of the atrial natriuretic factor included as a positive control. A representative experiment is shown. (C): Bar diagrams of chromatin immunoprecipitation (ChIP) assays showing the analysis by quantitative RT-PCR of the binding of Gata4 to the indicated DNA regions using the specified antibodies. ECR III and Cyp17a1 promoter were used as negative and positive control, respectively. Data are represented as the average of four independent experiments conducted in duplicate. (D): A representative immunoblot to detect Gata4 expression in nuclear extracts and cytoplasmic extracts from the samples used in EMSA and ChIP experiments. Tubulin is only detected in cytoplasmic extracts. Coomassie blue staining of the membrane is shown below. Abbreviation: ECR, evolutionary conserved regions.

Download figure to PowerPoint

In vitro binding of Gata4 to the ECR IV region was assessed using a 33 bp probe spanning the Gata4 binding site in the Nanog putative enhancer by an EMSA assay using nuclear extracts from control-transfected or Gata4-transfected ESCs (Fig. 7B). Two complexes could be detected: an upper complex (more retarded) that is more intense in Gata4-expressing ESCs and an unspecific lower complex (less retarded) that was equally present in control and Gata4-expressing ESCs. The intensity of the slower migrating complex was reduced to control levels when incubated with the unlabeled probe (lane 3), but not with an unlabeled mutated probe, where Gata4 binding site TGATAG was changed to GCGCGC (lane 4). To gain insight into the nature of the weaker upper complex present in control ESCs (lane 1), we conducted similar competition experiments using nuclear extracts from control-transfected ESCs (lanes 6–9). The upper complex disappeared when the extracts were incubated with unlabeled probe (lane 7) or with an anti-Gata4 antibody (lane 9), suggesting that it might correspond to the binding of an endogenous Gata factor expressed in ESCs. Finally, a probe designed in the proximal promoter of the atrial natriuretic factor was used as a positive control for Gata4 binding (lanes 10–12) [43]. Our in silico analysis revealed the presence of a novel and evolutionary conserved enhancer in the Nanog gene located 9 kbp upstream from the transcription start site. Also, our in vitro results indicate that this distal enhancer contains a functional Gata binding site and that binding of Gata4 to this sequence negatively regulates gene expression.

Finally, we evaluated whether Gata4 could bind to the identified Nanog distal enhancer in intact cells. We carried out ChIP assays in nontransfected and HA-Gata4-transfected ESCs. First, we validated the antibodies used for the immunoprecipitation and confirmed their specificity by immunoblotting (supporting information Fig. S11). Immunoprecipitated DNA fragments were quantitated by qRT-PCR. Three different genomic loci were analyzed, a 149 bp region located in ECR III, a 143 bp domain containing the Gata4 binding motif in ECR IV, and a 269 bp stretch located in the Cyp17a1gene as a positive control [44]. We observed an approximate 20-fold and fourfold enrichment of the ECR IV enhancer in Gata4-expressing ESC samples precipitated with the anti-Gata4 and anti-HA antibodies, respectively, as opposed to the IgG controls (Fig. 7C, middle bar diagram). We did not detect amplification of the ECR III control region in the anti-Gata4 or anti-HA precipitates. In agreement with previously published results [44], immunoprecipitation with anti-HA or anti-Gata4 antibodies resulted in a 28-fold enrichment of the Cyp17a1 promoter region relative to the IgG controls. In conclusion, our data indicate that Gata4 binds to a distal enhancer of the Nanog gene and negatively regulates the expression of this pluripotency gene.

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

In this article, we have shown that Gata4 blocks OKSM-induced cell reprogramming by repressing Nanog gene expression through its direct binding to a novel distal enhancer (supporting information Fig. S12). When Gata4 is expressed together with OKSM in MEF, cells become trapped into a metastable cell type that is unable to both activate the robust expression of pluripotent genes such Nanog and Lin28 and form ESC-like colonies in 2i+LIF medium. Similar results were obtained when the closely related factor Gata6 was expressed in the reprogramming assays instead of Gata4. Gata4 does not interfere with the initial phase of reprogramming but rather with the maturation phase of the process, when Nanog expression becomes essential to reach pluripotency. This phenotype is highly similar to that observed in the reprogramming of Nanog-deficient cells [5] and, in fact, the blockade of cell reprogramming imposed by Gata4 overexpression was rescued by the introduction of exogenous Nanog.

While we have observed that endodermal transcription factors, such as Hhex or Foxa2, may facilitate iPS cell formation (unpublished observations; R. Bort), the expression of Gata4 caused a severe impairment of the reprogramming process, rendering loose colonies that did not upregulate Nanog, a hallmark of full reprogramming [5]. Conversely, knockdown of Gata4 increased the number of early iPS cell colonies and favored Nanog upregulation in MEF transduced with OKS/M (Fig. 4). When Foxa2 was coexpressed with Gata4, colonies could be isolated and clonally expanded, but still Nanog mRNA levels remained low (Figs. 1A, 2C). This partial rescue of the negative effect of Gata4 in cell reprogramming is not completely unexpected, since it has been proposed that Foxa2 promotes iPS cell generation in cooperation with Glis1 [8].

Reprogramming is a gradual process that can be divided in three stages based on the expression of molecular markers [2]: initiation (0–8 days), maturation (8–16 days), and stabilization phases (16–21 days). The fully reprogrammed phenotype is adopted when senescence overcame, the cells enter MET, upregulate initial pluripotent genes, silence the retroviral constructs, and show a robust expression of Nanog (reviewed in [45]). Gata4 expression did not hinder the initial upregulation of senescence or MET markers such p16Ink4a and p19Arf or Cdh1 and Ocln1, respectively (supporting information Figs. S6, S7). Similarly, markers of the first stage of reprogramming such as Gdf3, Slc38a5, Syne2, or PERP were upregulated in the presence of Gata4. However, the downregulation of Slc38a5 and the maintenance of Gdf3 mRNA levels during the maturation phase of the process were impaired. These data indicate that Gata4 does not block reprogramming at the initiation phase, but it impairs the transition of the cells to the maturation phase. In agreement with this possibility, expression of Gata4 during the maturation stage (8 days after the initial transduction of MEF with the OKSM viral mixture) also resulted in a lack of Nanog expression and iPS cell-like colonies (Fig. 5D).

We therefore speculated that Gata4 directly limited Nanog induction during reprogramming based on the following observations: (a) expression of Gata4 together with the Oct4, Klf4, Sox2, and Myc (OKSM) induced the formation of colonies that resembled those obtained with the reprogramming of Nanog-null neural stem cells [5]; (b) the initiation phase of reprogramming was not altered, similarly to the published results obtained when reprogramming assays were carried out using Nanog-deficient MEF [2]; (c) addition of Gata4 after the initiation phase of the reprogramming impaired the generation of iPS cells and Nanog mRNA induction during the process; (d) knockdown of Gata4 rendered iPS cell clones that expressed significantly higher levels of Nanog mRNA than the reprogrammed control clones; and (e) forced Nanog expression rescued the failure to obtain iPS cells in the presence of Gata4 (Fig. 6).

In silico search for Gata4 binding sites in ECR present in the Nanog gene led us to the identification of a Gata binding site located in a novel distal enhancer in this gene, ECR IV. Using EMSA and ChIP assays we found that Gata4 was able to bind in vitro in intact cells to this site. In addition, we found that Gata4 inhibited the expression of a luciferase reporter driven by the distal enhancer ECR IV, suggesting that the upregulation of the Nanog gene during cell reprogramming in the presence of Gata4 is inhibited by the binding of this transcription factor to the ECR IV enhancer. Then, the negative regulation of the Nanog gene by Gata4 will disable the molecular mechanisms necessary to acquire induced pluripotency. Interestingly, EMSA assays suggested that an endogenous Gata4 [46, 47] is bound to this site (Fig. 7C, lanes 6–9). The physiological significance of the binding of Gata factors to the distal enhancer in the Nanog gene is currently unknown, but given the roles of Gata4 as a pioneering factor in different cell types [48–50], we hypothesize that binding of Gata4 to the ECR IV enhancer might contribute to the differentiation of the ICM into extraembryonic endoderm when other factors or increased Gata4 levels are into play.

Exogenous expression of Oct4, Klf4, Sox2, and Myc is necessary and sufficient for triggering the initial steps of reprogramming causing global gene expression changes that will eventually lead to pluripotency, successfully achieved by the action of a Nanog-dependent transcriptional network [31]. By introducing Gata4 together with OKSM (OKSM+G) or 8 days after OKSM infection, Nanog expression was not induced and cell reprogramming was impaired. OKSM+G-infected cells could not grow and expand in 2i+LIF medium and lacked the ability to form embryoid bodies during aggregation-induced differentiation and they cannot therefore be considered pr-iPS cells. Our findings indicate that the lack of reprogramming observed by forced expression of Gata4 together with the four reprogramming factors was not the final result, but rather the consequence of the repression of the Nanog gene caused by the direct binding of Gata4 to the novel distal enhancer ECR IV that we have identified in this study. Given the important role of Gata4 and Nanog in cell differentiation, the results presented here have broad implications for deciphering the inter-regulatory networks controlling cell fate during the segregation of the primitive endoderm and the epiblast from the ICM during early embryogenesis.

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 demonstrated that Gata4 blocks somatic cell reprogramming by limiting the expression of Nanog. Gata4 binds to a conserved distal enhancer of Nanog and impedes Nanog induction during reprogramming.

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 Cristina Corchero for assistance in qRT-PCR, Eva Serna and Juan José Lozano for microarray analysis. We are indebted to Dr. Lisa Sevilla for critical reading of the manuscript. This study was supported by the Ministry of Science and Innovation, Grants SAF2010-15376 2007-64414 and SAF2011-29718 to R.B. and the Instituto de Salud Carlos III, Grant PS09/00248 to J.T. CIBEREHD (CIBER de Enfermedades Hepaticas y Digestivas) is funded by the Instituto de Salud Carlos III, Spain.

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
sc-12-0421_sm_SupplFigure1.pdf578KSupplementary Figure 1. Analysis of cells reprogrammed in the presence of Foxa2 and Gata4. (A) Phase contrast images of MEF infected with OKSM together with Foxa2- and Gata4-expressing viruses (OKSM+F+G) after 10 and 15 days in culture or after isolation and culture (lower panels) (original magnification x10). (B) Bar diagrams showing the qRT-PCR analysis of transcript levels of Lamb1 and Sparc in clones isolated from MEF reprogrammed with OKSM+F+G. Sox7, Ttr and Afp mRNA were not detected. (C) Bar diagrams showing a time-course analysis of the transcript levels of Lamb1 and Sparc by qRT-PCR in MEF infected with OKSM+G. Data are represented as the average ± SD of three replicates relative to the mRNA levels found in pooled mouse embryonic mRNA.
sc-12-0421_sm_SupplFigure2.pdf185KSupplementary Figure 2. Gene expression profiling of clones isolated from MEF transduced with OKS+F+G. Heat map depicting relative expression levels of expressed mRNA from gene array data (19621 probes) across MEF, mouse ES cells (mESC), Nanog-negative clones from OKSM-reprogrammed MEF and five clones isolated from MEF reprogrammed with OKS in the presence of Foxa2- and Gata4- expressing retroviruses. Red: higher expression; blue: lower expression. Microarrays data have been deposited in NCBI's Gene Expression Omnibus50and are accessible through GEO Series accession number GSE37548 (http://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE37548). A dendogram plot based on Euclidean distance and average linkage showing hierarchical clustering or closeness in expressed mRNA levels is shown on the right. Gata4 represses Nanog
sc-12-0421_sm_SupplFigure3.tif2919KSupplementary Figure 3. Validation of lentiviral vectors expressing shRNA for mouse Gata4. 3T3 cells were infected with lentiviruses expressing a shRNA for luciferase (shLuc) or 4 different shRNAs for mouse Gata4 and the efficiency to knockdown the expression of the Gata4 protein assessed by immunoblotting using antibodies against Gata4 and Tubulin.
sc-12-0421_sm_SupplFigure4.pdf418KSupplementary Figure 4. Gata6 blocks cell reprogramming. (A) Phase-contrast images of MEF subjected to a reprogramming assay with OKSM in the presence of Gata4 or Gata6 expressing retroviruses at the indicated time-points (original magnification x10). (B) Alkaline Phosphatase activity staining of MEF infected as in (A) at days 8 and 21 after coinfection.
sc-12-0421_sm_SupplFigure5.tif1925KSupplementary Figure 5. BrdU incorporation in MEF overexpressing Gata4. Bar diagram showing the incorporation of BrdU in MEF transduced with a GFP- (control) or with a Gata4-exppressing retrovirus 48 hours postinfection. Data are represented as the average ± SD of three replicates relative to the control.
sc-12-0421_sm_SupplFigure6.pdf201KSupplementary Figure 6. Initial induction of senescence markers during reprogramming is not altered by Gata4. MEF were infected with OKSM in the absence or presence of Gata4-expressing retroviruses (OKSM+G) and the expression of p16Ink4a and p19Arf were measured by qRT-PCR. Mock infected MEF were used as controls. Data are represented as the average ± SD of three replicates relative to the mRNA levels found in MEF undergoing senescence (> 5 passages). Gata4 represses Nanog
sc-12-0421_sm_SupplFigure7.pdf338KSupplementary Figure 7. Mesenchymal to epithelial transition during reprogramming is not altered by Gata4 expression. Bar diagrams showing the timecourse expression of the indicated MET related markers during the reprogramming of MEF in the absence (OKSM) or presence (OKSM+G) of Gata4-expressing viruses. Data are represented as the average ± SD of three replicates relative to the mRNA levels found in pooled mouse embryonic mRNA.
sc-12-0421_sm_SupplFigure8.pdf439KSupplementary Figure 8. MEF can be reinfected during reprogramming. (A) Diagram showing the experimental design of reinfection of MEF subjected to a reprogramming assay with Gata4-, GFP- or RFP-expressing retroviruses. (B) Upper bar diagram, qRT-PCR analysis of exogenous Gata4 mRNA levels in cells transduced with Gata4-expressing viruses. Grey bars represent plates infected with Gata4 after day 8 as depicted in A. Lower panels show fluorescence images of MEF undergoing reprogramming with RFP- or GFP-expressing retroviruses (original magnification x10). These photographs showed that cells undergoing reprogramming are efficiently reinfected and that reinfection occurs predominantly in cells forming colonies.
sc-12-0421_sm_SupplFigure9.pdf245KSupplementary Figure 9. Initial upregulation of reprogramming markers are not altered by Gata4 expression. Bar diagrams showing the time-course expression by qRT-PCR of the indicated genes in MEF subjected to reprogramming in the absence (OKSM) or presence (OKSM+G) of Gata4-expressing viruses. Grey bars correspond to samples infected with Gata4 after day 8. Data are represented as the average ± SD of three replicates relative to the levels found in pooled mouse embryonic mRNA. Gata4 represses Nanog
sc-12-0421_sm_SupplFigure10.pdf390KSupplementary Figure 10. Identification of the Nanog distal enhancer ECR IV. (A) Alignment of genomic sequences located upstream the Nanog transcription start site from different species. Four ECR (I to IV) were identified. (B) Pairwise alignment of the sequences within the ECR IV enhancer. The consensus binding site for Gata factors is indicated (red line).
sc-12-0421_sm_SupplFigure11.pdf123KSupplementary Figure 11.Validation of antibodies used in the ChIP assays by immunoblotting. Immunoblot of lysates from HAGata4 overexpressing ES cells using the indicated antibodies. Molecular markers are shown at the left of each membrane.
sc-12-0421_sm_SupplFigure12.pdf101KSupplementary Figure 12. Model depicting the effect of Gata4 expression in cell reprogramming. In this model, we show that Gata4 interferes with OKSM-induced reprogramming by repressing Nanog upregulation during the maturation stage of the process.
sc-12-0421_sm_SupplTable1.pdf8KSupplementary Table 1
sc-12-0421_sm_SupplTable2.pdf17KSupplementary Table 2

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.