SEARCH

SEARCH BY CITATION

Keywords:

  • E-cadherin;
  • Embryonic stem cells;
  • β-catenin;
  • Activin;
  • FGF;
  • Nodal

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

We have previously demonstrated that differentiation of embryonic stem (ES) cells is associated with downregulation of cell surface E-cadherin. In this study, we assessed the function of E-cadherin in mouse ES cell pluripotency and differentiation. We show that inhibition of E-cadherin-mediated cell–cell contact in ES cells using gene knockout (Ecad−/−), RNA interference (EcadRNAi), or a transhomodimerization-inhibiting peptide (CHAVC) results in cellular proliferation and maintenance of an undifferentiated phenotype in fetal bovine serum-supplemented medium in the absence of leukemia inhibitory factor (LIF). Re-expression of E-cadherin in Ecad−/−, EcadRNAi, and CHAVC-treated ES cells restores cellular dependence to LIF supplementation. Although reversal of the LIF-independent phenotype in Ecad−/− ES cells is dependent on the β-catenin binding domain of E-cadherin, we show that β-catenin null (βcat−/−) ES cells also remain undifferentiated in the absence of LIF. This suggests that LIF-independent self-renewal of Ecad−/− ES cells is unlikely to be via β-catenin signaling. Exposure of Ecad−/−, EcadRNAi, and CHAVC-treated ES cells to the activin receptor-like kinase inhibitor SB431542 led to differentiation of the cells, which could be prevented by re-expression of E-cadherin. To confirm the role of transforming growth factor β family signaling in the self-renewal of Ecad−/− ES cells, we show that these cells maintain an undifferentiated phenotype when cultured in serum-free medium supplemented with Activin A and Nodal, with fibroblast growth factor 2 required for cellular proliferation. We conclude that transhomodimerization of E-cadherin protein is required for LIF-dependent ES cell self-renewal and that multiple self-renewal signaling networks subsist in ES cells, with activity dependent upon the cellular context. STEM CELLS 2009;27:2069–2080


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

Embryonic stem (ES) cells are pluripotent cells derived from preimplantation embryos and exhibit a number of remarkable qualities. They can be grown indefinitely under appropriate conditions while retaining the ability to differentiate into cells representative of the three primary germ layers. These unique ES cell properties have the potential to revolutionize medicine by offering treatment options for a wide range of diseases/disorders for which no treatments currently exist. In addition, ES cells are an excellent model system for elucidating mechanisms involved in development and disease [1, 2]. In recent years, significant insights into the understanding of the biological properties of ES cells have been achieved, with the identification of signaling pathways associated with both self-renewal and differentiation. For example, mouse ES cells can be maintained in an undifferentiated state by culture in defined medium containing bone morphogenetic proteins (BMPs), leukemia inhibitory factor (LIF), and N2/B27 supplements [3]. By contrast, human ES cells self-renew via fibroblast growth factor (FGF)-2 and Activin/Nodal signaling pathways [4, 5]. BMPs, Activin, and Nodal belong to the transforming growth factor (TGF)-β ligand superfamily and bind to activin receptor-like kinase (Alk) receptors (Alk3/6, Alk4, and Alk4/7, respectively). These TGF-β ligands signal via distinct members of the Smad family (Smad2/3 for Activin/Nodal and Smad1/5/8 for BMPs), although Smad4 is required as a cofactor in both cascades [6]. Downstream of these pathways, the expression of core intrinsic factors, such as Nanog, are essential for the maintenance of pluripotency in both species [7]. Recently, it was reported that mouse ES cells exhibit an innate program for self-renewal, allowing removal of LIF and BMP ligands from the culture medium [8]. However, several novel pluripotent stem cell lines have been isolated that self-renew via LIF-/BMP-independent mechanisms (EpiSC and FAB-SC), suggesting that pluripotent cells may possess multiple self-renewal signaling pathways [9–11]. Indeed, repression of E-cadherin in “stimulated” FAB-SC cells resulted in their rapid differentiation, demonstrating a potential role for this protein in the maintenance of pluripotency.

E-cadherin is a member of the classical cadherin family and is expressed on most epithelial cells, including ES cells [2, 12]. The extracellular domain of E-cadherin interacts in a homotypic calcium-dependent manner with E-cadherin molecules on neighboring cells, thereby facilitating cell–cell contact [13]. E-cadherin is essential for embryogenesis because E-cadherin null (Ecad−/−) embryos fail to develop beyond the blastocyst stage [12], reflecting a loss in epithelial integrity in both the trophectoderm and inner cell mass [12, 14]. Ecad−/− ES cells exhibit a significantly lower ability to form chimeric mice and fail to form any organized structures in teratomas [12]. Loss of cell surface E-cadherin is a defining characteristic of epithelial–mesenchymal transition (EMT), which is required for ingression of epiblast cells within the primitive streak during early embryonic development [13, 15] and is associated with tumor cell metastasis [2, 16]. The cytoplasmic region of E-cadherin binds to β-catenin, allowing interaction with the actin cytoskeleton via the intermediate protein α-catenin, and is required for the maintenance of adherens junctions in epithelial cells [17]. In addition, β-catenin also functions as a transactivating factor via the canonical Wnt-signaling pathway, leading to upregulation of several target genes [17, 18]. β-catenin null (β-cat−/−) mice die soon after gestation as a result of defects in the development of the embryonic ectoderm and failure to form mesoderm [19].

Recently, we demonstrated that an EMT event occurs during both mouse and human ES cell differentiation and that this exhibits a striking similarity to processes associated with tumor cell metastasis [2, 16]. For example, the ES cell EMT event is characterized by an E- to N-cadherin switch, upregulation of E-cadherin repressor molecules (Snail and Slug proteins), gelatinase activity (matrix metalloproteinase [MMP]-2 and MMP-9), and increased cellular motility [2]. Therefore, we hypothesized that cell surface E-cadherin may function to regulate the differentiation of ES cells. In this study we used Ecad−/− cells, wild-type ES cells exhibiting reversible RNA interference (RNAi) targeting of E-cadherin, and a peptide inhibitor of E-cadherin to investigate the role of this protein in the self-renewal of ES cells. We demonstrate that abrogation of E-cadherin in ES cells results in an undifferentiated phenotype in the absence of LIF and identify an alternative mechanism of self-renewal via Activin A and Nodal signaling, with FGF-2 required for cellular proliferation.

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

Culture of Mouse ES Cells

Mouse ES cells were cultured on gelatin-treated plates in knockout Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 2 mM L-glutamine, nonessential amino acids (NEAAs) (100×, 1:100 dilution), and 50 μM 2-mercaptoethanol, all from Invitrogen Corporation, Paisley, U.K., http://www.invitrogen.com), and 1,000 units/ml LIF (ESGRO; Millipore, Watford, U.K., http://www.millipore.com) at 37°C and 5% CO2 unless otherwise stated. The medium was replenished every 24 hours and the cells were passaged before confluence (2 days). Gelatin-treated plates were prepared as described previously [20]. All ES cell lines exhibit a normal karyotype [21]. To assess the role of TGF-β signaling, ES cells were cultured in basal medium (knockout DMEM), L-glutamine (2 mM), 2-mercaptoethanol (50 μM), NEAAs (100×, 1:100 dilution), and serum replacement (10%), all from Invitrogen, and supplemented with FGF-2 (12 ng/ml), Activin A (20 ng/ml), TGF-β (1 ng/ml), and Nodal (50 ng/ml), all from R&D Systems (Oxfordshire, U.K., http://www.rndsystems.com). Cells were treated with dimethylsulfoxide (DMSO) (control) and inhibitors of phosphatidylinositol 3′ kinase (PI3K) (LY294002; Cell Signaling Technology, Beverly, MA, http://www.cellsignal.com), mitogen-activated protein kinase/extracellular signal–related kinase kinase (MEK)-1/2 (PD98059; Tocris Bioscience, Ellisville, MO, http://www.tocris.com), FGFR-1 (SU5402; Tocris) and Alk (SB431542; Tocris) signaling pathways (all 10 μM) for 5 days (cells passaged every 2 days). The E-cadherin peptide inhibitor CHAVC was prepared according to the procedure reported by Blaschuk et al. [22] and solubilized in DMSO (final concentration, 2.5 mg/ml). Four micromolar peptide or an equal volume of DMSO (control) was added to standard ES cell medium and cells were grown as described above. Peptide and medium were replenished every 24 hours. Prior to removal of LIF, cells were grown in the presence of CHAVC peptide for two passages.

Immunofluorescent Imaging of ES Cells

ES cells were cultured in six-well gelatin-treated tissue culture-grade plates, fixed in 4% paraformaldehyde (Sigma-Aldrich, Dorset, U.K., http://www.sigmaaldrich.com), and stained in situ, as previously described [21]. Primary antibodies were as follows: anti-Oct-4 (1:100 dilution; Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com), anti-Nanog (1:100 dilution; Chemicon, Temecula, CA, http://www.chemicon.com), anti-E-cadherin and anti-β-catenin (both 1:100 dilution; BD Biosciences, Oxford, U.K., http://www.bdbiosciences.com), and anti-Smad2 and anti-phospho-Smad2 (both 1:100 dilution; Cell Signaling Technology, http://www.cellsignal.com). Secondary antibodies conjugated with Alexa Fluor 488 or Alexa Fluor 546 were used (1:500 dilution; Molecular Probes Inc., Eugene, OR, http://probes.invitrogen.com), and all samples were mounted using 4′,6-diamidino-2-phenylindole (DAPI) Vector shield (Vector Laboratories, Peterborough, U.K., http://www.vectorlabs.com). The cells were viewed on a Leica DM5000 fluorescence microscope (Leica, Heerbrugg, Switzerland, http://www.leica.com).

Fluorescent Flow Cytometry Analysis of ES Cells

ES cells were exposed to cell dissociation buffer (Invitrogen), washed once in 900 μl phosphate-buffered saline (PBS), resuspended in 100 μl of 0.2% bovine serum albumin in PBS (fluorescence-activated cell sorting [FACS] buffer) containing the primary antibody, and incubated for 30 minutes on ice. Primary antibodies were as follows: anti-mouse E-cadherin (DECMA-1, 1:500 dilution; Sigma) or anti-mouse stage-specific embryonic antigen (SSEA)-1 (1:500 dilution; Santa Cruz). Cells were washed once in 900 μl PBS, resuspended in 100 μl FACS buffer containing the appropriate phycoerythrin-conjugated secondary antibody (all 1:100 dilution; Santa Cruz), and incubated for 30 minutes on ice. The cells were washed once in 900 μl PBS and fixed in 400 μl 1% w/v paraformaldehyde in PBS. Cell fluorescence was analyzed using a Becton Dickinson FACScalibur cell sorter (Becton, Dickinson and Company, Franklin Lakes, NJ, http://www.bd.com). Viable cells were gated using forward and side scatter and all data represent cells from this population.

Reverse Transcription-Polymerase Chain Reaction

Total RNA was extracted from cells using RNAzol B according to the manufacturer's instructions (Biogenesis, Dorset, U.K., http://www.biogenesis.co.uk/), treated with DNase (Promega, Madison, WI, http://www.promega.com) and phenol/chloroform purified. Synthesis of cDNA was performed as described previously [20]. Reverse transcription-polymerase chain reaction (RT-PCR) was performed using 1 μl cDNA solution and amplified for 35 cycles at an optimal annealing temperature. Samples were separated on 2% agarose gels containing 400 ng/ml ethidium bromide and visualized using an Epi Chemi II Darkroom and Sensicam imager with Labworks four software (UVP, Upland, CA, http://www.uvp.com). Primer sequences are shown in supporting information Table S1.

Amaxa Transfection of Mouse ES Cells

ES cells were cultured as described above, trypsinized, and washed twice in PBS. Cells were electroporated as described in the manufacturer's instructions using an Amaxa Biosystems NucleofectorII and ES cell electroporation kit (Amaxa Biosystems, Cologne, Germany, http://www.amaxa.com). Briefly, 2 × 106 ES cells were suspended in Amaxa ES cell solution with a total of 2 μg of the appropriate plasmid (control, full length [FL] E-cadherin, Δβcat, Δβcat–Δp120ctn E-cadherin, or hairpin loop RNAi vectors) and electroporated using program A-30 on the Amaxa NucleofectorII.

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

E-Cad−/− ES Cells Maintain Expression of Pluripotency Markers in the Absence of LIF

E-cad−/− ES cells cultured in medium containing FBS and LIF in gelatin-treated plates exhibited decreased cell–cell contact and a more mesenchymal-like phenotype (Fig. 1A, Ecad−/−) than wild-type D3 (wtD3) cells (Fig. 1A, wtD3). Ecad−/− ES cells exhibited a similar expression profile of the cell surface pluripotency marker SSEA-1 to wtD3 cells (Fig. 1B, i and ii, respectively), demonstrating the undifferentiated phenotype of the cells. To assess whether the absence of E-cadherin could affect differentiation of ES cells, we cultured Ecad−/− and wtD3 ES cells in the absence of LIF under a normal passaging regimen. To our surprise, Ecad−/− ES cells failed to differentiate under these conditions, with Oct-4 protein expression observed in >95% of the cells after 12 days (Fig. 1C, 1D). wtD3 cell numbers rapidly decreased following withdrawal of LIF, reflecting differentiation and greater cell death under these conditions, with cells expressing high levels of Oct-4 representing <10% of the population at 12 days (Fig. 1C, 1D). Ecad−/− ES cells exhibited similar cellular proliferation rates irrespective of LIF supplementation (Fig. 1E), and cell cycle analysis revealed no difference between cells cultured under these conditions and wtD3 cells grown in the presence of LIF (data not shown). To further confirm that Ecad−/− ES cells maintained an undifferentiated phenotype in the absence of LIF under a normal passaging regimen, we assessed expression of the pluripotency markers Oct-4 and Nanog and early differentiation markers brachyury (T), ζ-globin (ZG), and transthyretin (TTR) by RT-PCR (Fig. 1F, i and ii). Both wtD3 and Ecad−/− ES cells cultured in the presence of LIF exhibited expression of pluripotent transcripts, but lacked the differentiation-associated markers (Fig. 1F, LIF). Following culture of wtD3 cells in the absence of LIF, upregulation of early differentiation markers was observed (Fig. 1F, i, −LIF). In contrast, Ecad−/− ES cells cultured in the absence of LIF maintained expression of pluripotent transcripts and lacked expression of all differentiation markers (Fig. 1F, ii, −LIF). Ecad−/− ES cells were cultured for >150 days in the absence of LIF and maintained expression of both Oct-4 and Nanog transcripts and proteins, and exhibited a normal karyotype (data not shown). These data demonstrate that Ecad−/− ES cells do not differentiate in the absence of LIF under a normal passaging regimen.

thumbnail image

Figure 1. E-cad−/− ES cells do not differentiate in the absence of LIF. (A): wtD3 and Ecad−/− ES cells were cultured in medium supplemented with FBS and LIF and assessed for phenotype (phase contrast) and E-cadherin expression using immunofluorescence microscopy (DAPI shows the number of cells in the field of view). (B): Fluorescent flow cytometry analysis of SSEA-1 expression in wtD3 (i) and Ecad−/− (ii) ES cells. SSEA-1, green profile; isotype control antibodies, purple profile. (C): wtD3 and Ecad−/− ES cells were cultured in the absence of LIF for 12 days under a normal passaging regimen and assessed for Oct-4 protein expression using immunofluorescence microscopy. (D): Quantitative Oct-4 protein expression in wtD3 (hatched bar) and Ecad−/− (black bar) ES cells cultured in the presence (+LIF) or absence (−LIF) of LIF for 12 days under a normal passaging regime. Error bars show the SD of three independent experiments quantified by counting the number of Oct-4+ and Oct-4 cells in five fields of view. *p > .05, **p < .05, unpaired t-test. (E): In total, 1 × 105 Ecad−/− ES cells were cultured for 6 days in the presence (♦) or absence (▪) of LIF in gelatin-treated six-well plates and cumulative cell numbers were assessed. Error bars show the SD of three independent experiments. Note that there is no difference in the proliferation rate between the two conditions. (F): wtD3 (i) and Ecad−/− (ii) ES cells were cultured in gelatin-treated six-well plates for 12 days under a normal passaging regimen in the presence (+LIF) or absence (−LIF) of LIF and assessed for expression of βt, Oct, Nan, T, ZG, and TTR transcripts using reverse transcription-polymerase chain reaction (35 cycles). Abbreviations: βt, β-tubulin; DAPI, 4′,6-diamidino-2-phenylindole; E-cad−/−; E-cadherin null; ES, embryonic stem; FBS, fetal bovine serum; LIF, leukemia inhibitory factor; Nan, Nanog; Oct, Oct-4; SD, standard deviation; SSEA, stage-specific embryonic antigen; T, brachyury; TTR, transthyretin; wt, wild-type; ZG, ζ-globin.

Download figure to PowerPoint

Inhibition of E-Cadherin Using RNA Interference Prevents Differentiation of ES Cells in the Absence of LIF

To ensure that the observations with Ecad−/− ES cells were not a result of isolation of an aberrant clone, we transfected MESC20 ES cells with hairpin loop RNAi vectors targeted against E-cadherin (EcadRNAi ES cells). EcadRNAi ES cells cultured in the presence of LIF exhibited loss of cell–cell contact and a more mesenchymal phenotype (Fig. 2A, i, phase contrast), similar to that observed for Ecad−/− ES cells, and lacked cell surface E-cadherin protein expression (Fig. 2A, i, E-cadherin). EcadRNAi ES cells expressed the cell surface pluripotency marker SSEA-1 (Fig. 2A, i) and Oct-4 and Nanog proteins (Fig. 2A, ii, +LIF), demonstrating the undifferentiated phenotype of the cells. Culture of EcadRNAi ES cells in the absence of LIF for 12 days resulted in a similar phenotype to cells cultured in the presence of LIF (data not shown) and expression of Oct-4 and Nanog proteins in the majority of the cell population (Fig. 2A, ii, −LIF).

thumbnail image

Figure 2. Inhibition of E-cadherin in wild-type ES cells using RNA interference prevents differentiation in the absence of LIF. (A): MESC20 ES cells (i) exhibiting RNA inhibition of E-cadherin (EcadRNAi) were assessed for morphology (phase contrast) and cell surface E-cadherin and SSEA-1 expression using fluorescent flow cytometry analysis. SSEA-1/E-cadherin, green profile; isotype control antibodies, purple profile. EcadRNAi ES cells (ii) were cultured in the presence (+LIF) and absence (−LIF) of LIF for 12 days under a normal passaging regimen and assessed for Oct-4 and Nanog protein expression using immunofluorescence microscopy (DAPI shows the number of cells in the field of view). (B): Reversal of E-cadherin RNA inhibition in EcadRNAi ES cells was performed by transfection of these cells with a Tet repressor vector, resulting in the cell line EcadRNAiR. EcadRNAiR ES cells (i) were assessed for morphology (phase contrast) and expression of cell surface E-cadherin and SSEA-1 using fluorescent flow cytometry analysis. SSEA-1/E-cadherin, green profile; isotype control antibodies, purple profile. EcadRNAiR ES cells (ii) were cultured in the presence (+LIF) and absence (−LIF) of LIF for 12 days under a normal passaging regimen and assessed for Oct-4 and Nanog protein expression using immunofluorescence microscopy. (C): Expression of βt (loading control), Oct, Nan, and Sox-2 in EcadRNAi cells cultured for 12 days in the absence of LIF (i) and EcadRNAiR cells cultured in the presence of LIF (ii). (D): Expression of various differentiation markers in EcadRNAi cells (i) and EcadRNAiR ES cells (ii) cultured for 12 days in the absence of LIF. In order: F5, fibroblast growth factor 5; MS1, Musashi 1; β3t, βIII-tubulin; ENO2, neural-specific enolase 2; GFAP, glial fibrillary acidic protein; βt (loading control); FL1, Flk1; RX1, Runx1; 133, CD133; ML2, myosin light chain 2; MH6, myosin heavy chain 6; MH7, myosin heavy chain 7; NX2, Nkx 2.5; AFP, α-fetoprotein; TTR, transthyretin; TFR, transferrin receptor; GT4, GATA4; GT6, GATA6. Abbreviations: βt, β-tubulin; DAPI, 4′,6-diamidino-2-phenylindole; ES, embryonic stem; LIF, leukemia inhibitory factor; Nan, Nanog; Oct, Oct-4; SSEA, stage-specific embryonic antigen.

Download figure to PowerPoint

RNA interference of E-cadherin in EcadRNAi ES cells was reversed by transfection of the cells with a Tet repressor vector (EcadRNAiR ES cells) (Fig. 2B). Reversal of RNA interference was observed by restoration of cell–cell contacts in cells cultured in the presence of LIF (Fig. 2B, i, phase contrast) and detection of cell surface E-cadherin protein (Fig. 2B, i, E-cadherin). EcadRNAiR ES cells expressed the cell surface pluripotency marker SSEA-1 (Fig. 2B, i) and the Oct-4 and Nanog proteins (Fig. 2B, ii, +LIF), demonstrating the undifferentiated phenotype of the cells. Culture of EcadRNAiR ES cells in the absence of LIF for 12 days under a normal passaging regimen resulted in loss of Oct-4 and Nanog protein expression in the majority of the population (Fig. 2B, ii, −LIF). RT-PCR analysis demonstrated expression of Oct-4, Nanog, and Sox-2 transcripts in EcadRNAi ES cells cultured in the absence of LIF for 12 days (Fig. 2C, i) and EcadRNAiR ES cells cultured in the presence of LIF (Fig. 2C, ii). In addition, culture of EcadRNAiR ES cells in the absence of LIF for 12 days resulted in upregulation of transcript markers representative of the three germ layers (Fig. 2D, ii). These results confirm that inhibition of E-cadherin in mouse ES cells results in reversible LIF-independent self-renewal.

Abrogation of E-Cadherin Protein-Mediated Cell–Cell Contact in ES Cells Results in LIF-Independent Self-Renewal

To determine whether abrogation of E-cadherin-mediated cell–cell contact also allows culture of wild-type mouse ES cells in the absence of LIF, we used a cyclic peptide (CHAVC) previously reported to inhibit E-cadherin transhomodimerization in epithelial cells [23]. The predicted chemical structure of the cyclic peptide CHAVC is shown in Figure 3A. Treatment of wtD3 ES cells with CHAVC resulted in loss of cell–cell contact and a more mesenchymal phenotype (Fig. 3B, +CHAVC) than in control-treated cells (Fig. 3B, control). Both control- and CHAVC-treated cells cultured in the presence of LIF exhibited expression of Oct-4 protein in >95% of the cell population (Fig. 3C). Culture of control- and CHAVC-treated cells in the absence of LIF for 12 days resulted in a differentiated phenotype in the former whereas the latter exhibited a phenotype similar to cells cultured in the presence of LIF (Fig. 3D, +CHAVC). In addition, CHAVC-treated ES cells cultured in the absence of LIF maintained expression of Oct-4 protein in >95% of the population whereas control-treated cells exhibited <5% Oct-4 protein expression (Fig. 3E). Removal of CHAVC from the cells described in Figure 3D and 3E resulted in a differentiated phenotype (Fig. 3F) and significantly lower Oct-4 protein expression (Fig. 3G). RT-PCR analysis demonstrated expression of Oct-4, Nanog, and Sox-2 transcripts in CHAVC-treated wtD3 ES cells cultured in the absence of LIF for 12 days (Fig. 3H, ii). Removal of CHAVC from wtD3 ES cells and their culture in the absence of LIF for 12 days resulted in upregulation of transcript markers representative of the three germ layers (Fig. 3J, ii). These results demonstrate that inhibition of E-cadherin-mediated cell–cell contact at the protein level in mouse ES cells results in reversible LIF-independent self-renewal.

thumbnail image

Figure 3. Inhibition of E-cadherin in wt ES cells using a transhomodimerization-inhibiting peptide prevents differentiation of the cells in the absence of LIF. (A): Predicted chemical structure of the transhomodimerization inhibiting peptide CHAVC. (B): Phenotype of wtD3 ES cells treated under control conditions (DMSO) or with CHAVC peptide in the presence of LIF. (C): Immunofluorescence analysis showing Oct-4 protein expression in wtD3 ES cells treated under control conditions (DMSO) or with CHAVC peptide in the presence of LIF. (D): Phenotype of wtD3 ES cells treated with control (DMSO) or CHAVC peptide in the absence of LIF for 12 days under a normal passaging regimen. (E): Immunofluorescence microscopy analysis showing Oct-4 protein expression in wtD3 ES cells treated with control (DMSO) or CHAVC peptide in the absence of LIF for 12 days under a normal passaging regimen. (F): Phenotype of wtD3 ES cells treated with CHAVC peptide for 12 days in the absence of LIF and subsequent removal of CHAVC for 12 days under a normal passaging regimen. (G): Immunofluorescence microscopy analysis showing Oct-4 protein expression in wtD3 ES cells treated with CHAVC peptide for 12 days in the absence of LIF and subsequent removal of CHAVC for 12 days under a normal passaging regimen. DAPI shows the number of cells in the field of view. (H): Expression of βt (loading control), Oct, Nan, and Sox-2 in wtD3 ES cells cultured in the presence of LIF (i) and CHAVC-treated wtD3 ES cells cultured in the absence of LIF (ii) for 12 days. (J): Expression of various differentiation markers in wtD3 ES cells cultured in the absence of LIF (i) and wtD3 ES cells treated with CHAVC peptide for 12 days in the absence of LIF and subsequent removal of CHAVC for 12 days (ii). In order: F5, fibroblast growth factor 5; MS1, Musashi 1; β3t, βIII-tubulin; ENO2, neural-specific enolase 2; GFAP, glial fibrillary acidic protein; βt (loading control); FL1, Flk1; RX1, Runx1; 133, CD133; ML2, myosin light chain 2; MH6, myosin heavy chain 6; MH7, myosin heavy chain 7; NX2, Nkx 2.5; AFP, α-fetoprotein; TTR, transthyretin; TFR, transferrin receptor; GT4, GATA4; GT6, GATA6. Abbreviations: βt, β-tubulin; DAPI, 4′,6-diamidino-2-phenylindole; DMSO, dimethylsulfoxide; ES, embryonic stem; LIF, leukemia inhibitory factor; Nan, Nanog; Oct, Oct-4; wt, wild-type.

Download figure to PowerPoint

The β-Catenin–E-Cadherin Complex Is Required for Differentiation of Ecad−/− ES Cells

We next investigated which region of the E-cadherin protein was required for differentiation of ES cells in the absence of LIF. Ecad−/− ES cells were transiently transfected with either FL E-cadherin (E-cad FL), E-cadherin lacking the β-catenin binding region (Δβcat), or E-cadherin lacking the β-catenin and p120ctn binding regions (Δβcat–Δp120ctn). Fluorescent flow cytometry analysis demonstrated cell surface protein expression 24 hours after transfection of Ecad−/− ES cells with the E-cad FL, Δβcat, or Δβcat–Δp120ctn vectors (Fig. 4A). Only Ecad−/− ES cells transfected with FL E-cadherin cDNA exhibited restoration of cell–cell contact (Fig. 4A, i, phase contrast) and plasma membrane localization of β-catenin protein (Fig. 4A, i, β-catenin). Ecad−/− ES cells transfected with the E-cadherin vectors and cultured for 3 days in the absence of LIF were assessed for expression of Oct-4, Nanog, and the early differentiation markers T, ZG, and TTR (Fig. 4B). Only Ecad−/− ES cells transfected with the FL E-cadherin showed expression of the early differentiation markers (Fig. 4B, E-cad FL), demonstrating reversal of the differentiation block in these cells. Both Δβcat and Δβcat–Δp120ctn E-cadherin cDNA failed to induce upregulation of these markers (Fig. 4B), demonstrating that differentiation of ES cells in the absence of LIF is dependent on the E-cadherin–β-catenin complex.

thumbnail image

Figure 4. The β-catenin–E-cadherin complex is required for cell–cell contact and differentiation of Ecad−/− ES cells in the absence of LIF. (A): Ecad−/− ES cells were transfected with either (i) FL E-cadherin (E-cad FL), (ii) E-cadherin lacking the β-catenin binding region (Δβcat), or (iii) E-cadherin lacking the β-catenin and p120ctn binding regions (Δβcat–Δp120ctn) and assessed for cell surface protein expression by fluorescent flow cytometry analysis (E-cadherin, green profile; isotype control antibody, purple profile), morphology (phase contrast), and β-catenin expression using immunofluorescence microscopy. (B): Ecad−/− ES cells were transfected with either FL, Δβcat, or Δβcat–Δp120ctn E-cadherin vectors and cultured for 3 days in the absence of LIF in a monolayer culture. Reverse transcription-polymerase chain reaction analysis was performed to determine expression of the pluripotent markers Oct and Nan and the early differentiation markers T, ZG, and TTR; βt, loading control. (C): wtD3 and β-catenin null (βcat−/−) ES cells were cultured in FBS in the presence of LIF and assessed for β-catenin expression using immunofluorescence microscopy. (D) wtD3 and βcat−/− ES cells were cultured in FBS plus LIF and assessed for cell surface expression of E-cadherin and SSEA-1 using fluorescent flow cytometry. wtD3, pink profile; βcat−/−, green profile); isotype control antibodies, closed population. (E): βcat−/− ES cells were cultured in the presence of LIF in FBS under a normal passaging regimen for 12 passages and assessed for cellular morphology (phase contrast) and expression of Oct-4 protein using immunofluorescence microscopy. (F): βcat−/− ES cells were cultured in the absence of LIF in FBS under a normal passaging regimen for 12 passages and assessed for cellular morphology (phase contrast) and expression of Oct-4 protein using immunofluorescence microscopy. Abbreviations: βcat−/−, β-catenin null; βt, β-tubulin; DAPI, 4′,6-diamidino-2-phenylindole; E-cad−/−; E-cadherin null; ES, embryonic stem; FBS, fetal bovine serum; FL, full length; LIF, leukemia inhibitory factor; Nan, Nanog; Oct, Oct-4; SD, standard deviation; SSEA, stage-specific embryonic antigen; T, brachyury; TTR, transthyretin; wt, wild-type; ZG, ζ-globin.

Download figure to PowerPoint

These results suggest that β-catenin signaling may be responsible for the undifferentiated phenotype of Ecad−/− ES cells cultured in the absence of LIF. To confirm the function of β-catenin signaling in LIF-independent self-renewal of ES cells, we used βcat−/− ES cells. βcat−/− ES cells cultured in medium containing FBS and LIF in gelatin-treated plates lacked β-catenin protein (Fig. 4C, βcat−/−), whereas wtD3 cells exhibited β-catenin localization at the cell membrane (Fig. 4C, wtD3). βcat−/− ES cells exhibited similar levels of cell surface E-cadherin and SSEA-1 to wtD3 ES cells (Fig. 4D, wtD3, green profile; βcat−/−, pink profile). In addition, βcat−/− ES cells cultured in the presence of LIF and FBS maintained expression of the pluripotency marker protein Oct-4 (Fig. 4E, Oct-4) and exhibited a morphology similar to that observed in Ecad−/− ES cells (Fig. 4E, phase contrast). βcat−/− ES cells cultured in the absence of LIF for 12 days under a normal passaging regimen exhibited expression of the pluripotent protein Oct-4 in >95% of the population (Fig. 4F, Oct-4) and a phenotype similar to cells cultured in the presence of LIF (Fig. 4F, phase contrast). These observations demonstrate that abrogation of the E-cadherin–β-catenin complex allows self-renewal of ES cells in the absence of LIF and that β-catenin signaling is unlikely to play a part in Ecad−/− ES cell self-renewal under these conditions.

ES Cells Lacking E-Cadherin Exhibit Altered Growth Factor Response

To determine alternative pathways for self-renewal in Ecad−/− ES cells, we exposed these cells to small molecule inhibitors of signaling pathways associated with the self-renewal of human ES cells and epiblast-derived mouse cells [5, 9, 10, 24] (Fig. 5A). Treatment of wtD3 ES cells with the small molecule inhibitors of PI3K (LY294002), mitogen-activated protein kinase/MEK (PD98059), Alk-4, Alk-5, and Alk-7 (SB431542), and FGFR-1 (SU5402) did not induce differentiation of the cells, as assessed by immunofluorescence microscopy analysis of Nanog expression (Fig. 5A). In contrast, Ecad−/− ES cells exhibited differentiated morphology and loss of Nanog protein expression in >95% of the population when exposed to SB431542 (Fig. 5A, 5B). The effect of these inhibitors on Ecad−/− ES cells was similar in both the presence and absence of LIF (data not shown). These results suggest that Ecad−/− ES cells may self-renew via the TGF-β signaling pathway, similar to what is described for human ES cells and epiblast-derived mouse cells [5, 9, 10]. We next assessed whether forced expression of E-cadherin cDNA in Ecad−/− ES cells could reverse the differentiation-inducing effect of SB431542 (Fig. 5C). Forced expression of FL E-cadherin cDNA in Ecad−/− ES cells restored cell–cell contact and resulted in lower numbers of differentiated cells in the presence of SB431542, as evidenced by the expression of Nanog protein (Fig. 5C, Ecad−/− + E-cad FL). Ecad−/− ES cells transfected with control plasmid (Fig. 5C, Ecad−/− control) exhibited a differentiated morphology and loss of Nanog expression following treatment with SB431542.

thumbnail image

Figure 5. Abrogation of E-cadherin in ES cells results in altered growth factor response compared with wt ES cells. (A): wtD3 (black bar) and Ecad−/− (hatched bar) ES cells were cultured in gelatin-treated six-well plates in the presence of LIF and treated with DMSO (control) and inhibitors of PI3K (LY294002), MEK1/2 (PD98059), Alk (SB431542), or FGFR1 (SU5402) for 3 days and assessed for Nanog expression using immunofluorescence microscopy. Error bars show the standard deviation of three independent experiments quantified by counting the number of Nanog-positive cells within the total population of five fields of view. *p > .05, **p < .05 unpaired t-test. (B): wtD3 and Ecad−/− ES cells were cultured in the presence of LIF and the Alk inhibitor SB431542 for 3 days and assessed for morphology (phase contrast) and expression of Nanog protein using immunofluorescence microscopy (DAPI shows total cell nuclei in the field of view). (C): Ecad−/− ES cells were transfected with full length E-cadherin (E-cad FL) or a control vector (control) and cultured in the presence of LIF and the Alk inhibitor SB431542. Cells were assessed for morphology (phase contrast) and expression of Nanog following treatment with SB431542 for 3 days. (D): EcadRNAi and EcadRNAiR cells were cultured in the presence of LIF and the Alk inhibitor SB431542 for 3 days and assessed for morphology (phase contrast) and expression of Nanog protein using immunofluorescence microscopy. (E): wtD3 ES cells were treated with DMSO (control) or E-cadherin inhibiting peptide CHAVC (+CHAVC) in the presence of LIF and the Alk inhibitor SB431542 for 3 days and assessed for morphology (phase contrast) and expression of Nanog protein using immunofluorescence microscopy. (F): wtD3 ES cells were treated with the peptide CHAVC for 7 days and the peptide was then removed from the cells for 5 days. Cells were assessed for morphology (phase contrast) and expression of Nanog protein using immunofluorescence microscopy following treatment with SB431542 for 3 days. Abbreviations: Alk, activin receptor-like kinase; DAPI, 4′,6-diamidino-2-phenylindole; DMSO, dimethylsulfoxide; E-cad−/−; E-cadherin null; EcadRNAi, MESC20 ES cells exhibiting RNA inhibition of E-cadherin; EcadRNAiR, EcadRNAi cells with reversal of E-cadherin RNA inhibition; ES, embryonic stem; FGFR, fibroblast growth factor receptor; LIF, leukemia inhibitory factor; MEK, mitogen-activated protein kinase/extracellular signal–related kinase kinase; PI3K, phosphatidylinositol 3′ kinase; wt, wild-type.

Download figure to PowerPoint

To further confirm the observations in Ecad−/− ES cells, we assessed exposure of EcadRNAi and EcadRNAiR ES cells to SB431542 (Fig. 5D). EcadRNAi ES cells (lacking E-cadherin) exhibited a differentiated phenotype and loss of Nanog protein expression following treatment with SB431542 (Fig. 5D, EcadRNAi). In contrast, EcadRNAiR ES cells (in which E-cadherin protein expression had been subsequently activated) maintained an undifferentiated colony morphology and the majority of the cells expressed Nanog protein (Fig. 5D, EcadRNAiR). wtD3 ES cells treated with cyclic peptide (CHAVC) also exhibited a differentiated morphology when exposed to SB431542 and significantly lower Nanog expression (Fig. 5E, wtD3 + CHAVC), compared with control cells. Removal of CHAVC from wtD3 ES cells cultured in the presence of LIF for 5 days and subsequent exposure of the cells to SB431542 resulted in a characteristic undifferentiated colony morphology and expression of Nanog protein in the majority of the cells (Fig. 5F).

To assess whether βcat−/− ES cells also self-renew via the TGF-β signaling pathway, we exposed βcat−/− ES cells to SB431542 (Fig. 6A). Under these conditions, βcat−/− ES cells exhibited a differentiated morphology and loss of Nanog protein expression in the majority of the population. Furthermore, forced expression of β-catenin cDNA in βcat−/− ES cells restored cell–cell contact and resulted in lower numbers of differentiated cells in the presence of SB431542, as evidenced by expression of Nanog protein (Fig. 6B, βcat FL). Analysis of total Smad2 protein, part of the TGF-β family signaling pathway, demonstrated relatively low levels of diffusely distributed protein in wtD3 ES cells, whereas the majority of Ecad−/− and βcat−/− ES cells exhibited strong nuclear and weak cytoplasmic expression of Smad2 (Fig. 6C, i, Smad2). As expected, wtD3, Ecad−/− and βcat−/− ES cells exhibited nuclear localization of phosphorylated Smad2 protein (Fig. 6C, ii, pSmad2).

thumbnail image

Figure 6. Smad2 expression in wt, E-cad−/− and β-cat−/− ES cells. (A): wtD3 and β-cat−/− ES cells were cultured in the presence of LIF and the Alk inhibitor SB431542 for 3 days and assessed for morphology (phase contrast) and expression of Nanog protein using immunofluorescence microscopy (DAPI shows total cell nuclei in the field of view). (B): βcat−/− ES cells were transfected with full length β-catenin (βcat FL) or a control vector (control) and cultured in the presence of LIF and the Alk inhibitor SB431542 for 3 days. Cells were assessed for morphology (phase contrast) and Nanog expression as described above. (C): Immunofluorescence microscopy analysis of (i) total Smad2 and (ii) phosphorylated (P)Smad2 expression in wtD3, Ecad−/−, and βcat−/− ES cells. DAPI shows total cell nuclei in the field of view. Abbreviations: Alk, activin receptor-like kinase; βcat−/−, β-catenin null; DAPI, 4′,6-diamidino-2-phenylindole; E-cad−/−; E-cadherin null; ES, embryonic stem; LIF, leukemia inhibitory factor; wt, wild-type.

Download figure to PowerPoint

E-Cad−/− and β-Cat−/− ES Cells Maintain an Undifferentiated Phenotype in Serum-Free Medium Supplemented With FGF-2, Activin A, and Nodal

To determine exogenous factor requirements for the self-renewal of Ecad−/− ES cells in the absence of LIF, we assessed Nanog protein expression in cells cultured in serum-replacement medium (basal medium) supplemented with FGF-2, Activin A, Nodal, and TGF-β1 (Fig. 7A, 7B). Ecad−/− ES cells cultured in basal medium and FGF-2 alone had the lowest proportion of cells exhibiting Nanog expression (<10%) (Fig. 7A). Supplementation of basal medium with Activin A and FGF-2 resulted in >60% of Ecad−/− cells expressing Nanog protein, and this was increased to >90% of cells expressing Nanog when Nodal was included (Fig. 7A, 7B). TGF-β1 is unlikely to be required for self-renewal of Ecad−/− ES cells because medium containing FGF-2, Activin, Nodal, and TGF-β1 did not increase the proportion of Nanog-expressing cells in the population (Fig. 7A, 7B). Culture of Ecad−/− ES cells in basal medium and Activin A resulted in a similar proportion of cells expressing Nanog to those grown in FGF-2 and Activin (∼60%), although cellular proliferation was compromised. Culture of Ecad−/− ES cells in basal medium supplemented with Nodal resulted in approximately 30% of the cells expressing Nanog protein. Although Ecad−/− ES cells cultured in basal medium supplemented with Activin A and Nodal exhibited Nanog protein expression in >90% of the population, proliferation of the cells was significantly less than in medium containing Activin A, Nodal, and FGF-2 (Fig. 7C). Therefore, the optimal medium for the maintenance of Ecad−/− ES cells in an undifferentiated state represents a combination of FGF-2, Activin A, and Nodal, with the former required for cellular proliferation and Activin A and Nodal essential for maintaining an undifferentiated phenotype. Exposure of Ecad−/− ES cells cultured in basal medium and FGF-2, Activin A, and Nodal to SB431542 induced differentiation of the cells and significantly reduced Nanog protein expression (Fig. 7D). βcat−/− ES cells also maintained an undifferentiated phenotype when cultured in basal medium supplemented with FGF-2, Activin A, and Nodal (Fig. 7E, i), although proliferation was lower. Exposure of these cells to SB431542 induced a differentiated phenotype and significantly reduced Nanog protein expression (Fig. 7E, ii).

thumbnail image

Figure 7. E-cad−/− and β-cat−/− ES cells maintain an undifferentiated proliferative phenotype in serum-free medium supplemented with FGF-2, Activin A, and Nodal. (A): Ecad−/− ES cells were cultured in serum-free basal medium supplemented with various combinations of FGF-2, ActA, Nod, and TGF-β1, and assessed for Nanog expression using immunofluorescence microscopy. FGF/ActA/Nod versus FGF/ActA/Nod/TGF versus ActA/Nod, not significant (p > .05, paired t-test); FGF/ActA versus ActA, not significant (p > .05, paired t-test); all other comparisons, p < .05. (B): Cellular morphology (phase contrast) and Nanog protein expression in Ecad−/− ES cells cultured in serum-free basal medium supplemented with FGF and ActA or FGF, ActA, and Nod or FGF, ActA, Nod, and TGF. (C): Cellular proliferation rates of Ecad−/− ES cells cultured in serum-free basal medium supplemented with ActA and Nodal (♦) or FGF, ActA, and Nodal (▪). p < .05 at days 2, 4, and 6. (D): Cellular morphology (phase contrast) and Nanog protein expression in Ecad−/− ES cells cultured in serum-free basal medium supplemented with FGF-2, ActA, and Nod and treated with the Alk inhibitor SB431542 for 3 days. (E): Cellular morphology (phase contrast) and Nanog protein expression in βcat−/− ES cells cultured in serum-free basal medium supplemented with FGF-2, ActA, and Nod (i) and cellular morphology (phase contrast) and Nanog protein expression in βcat−/− ES cells cultured in serum-free basal medium supplemented with FGF-2, ActA, and Nod and treated with the ActA receptor inhibitor SB431542 for 3 days (ii). Abbreviations: ActA, Activin A; Alk, activin receptor-like kinase; βcat−/−, β-catenin null; DAPI, 4′,6-diamidino-2-phenylindole; E-cad−/−; E-cadherin null; ES, embryonic stem; FGF-2, fibroblast growth factor-2; Nod, Nodal; TGF, transforming growth factor β1.

Download figure to PowerPoint

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

Our study provides novel insights into the function of E-cadherin in ES cells. We demonstrate that the abrogation of E-cadherin-mediated cell–cell contact, using genetic depletion or transhomodimerization inhibition, allows proliferation of these cells in the absence of LIF while maintaining an undifferentiated phenotype. Furthermore, we show that β-catenin signaling is unlikely to play a part in this phenotype because β-cat−/− ES cells proliferate and remain undifferentiated in the absence of LIF. These data demonstrate that abrogation of the E-cadherin–β-catenin complex can alter the responsiveness of ES cells to external ligands, which may provide further insight into the function of this complex in ES cell self-renewal and, potentially, tumor cell metastasis.

Several studies have suggested that β-catenin signaling is required for the self-renewal of mouse and human ES cells [25, 26]. However, while modulation of β-catenin activity in mouse ES cells may result in an altered response of the cells to LIF, our studies in β-cat−/− ES cells demonstrate unequivocally that this protein is not obligatory for maintenance of the undifferentiated phenotype of ES cells cultured in FBS and LIF. However, this phenotype is likely to reflect the presence of FGF-2, Activin A, and Nodal at appropriate concentrations within FBS-supplemented medium, thereby permitting LIF-independent self-renewal. The cytoplasmic region of E-cadherin binds to β-catenin, allowing interaction with the actin cytoskeleton via the intermediate protein α-catenin [13]. Gene trap analysis in which the function of α-catenin was ablated in mice demonstrated a similar phenotype to E-cad−/− embryos [27], with defects observed in the trophectoderm epithelium leading to a developmental block at the blastocyst stage. In addition, inner cell mass outgrowths from α-catenin gene trap embryos exhibit altered cell–cell contact, similar to that observed in E-cad−/− ES cells. Therefore, it remains a possibility that inhibition of a single component of the E-cadherin–β-catenin–α-catenin complex will result in altered self-renewal properties of ES cells.

Our data [2] suggest that the function of the E-cadherin–β-catenin interaction in ES cells is likely to be more complex than simply maintaining epithelial integrity via cell–cell contact. For example, the promigratory 5T4 antigen is translocated from the cytoplasm to the cell surface upon abrogation of E-cadherin in wt ES cells [2]. Furthermore, expression of Eph receptors and ephrins have been shown to be differentially regulated by E-cadherin [28]. Chou et al. [11] recently demonstrated derivation of FAB-SCs from mouse blastocysts cultured in medium supplemented with FGF-2, Activin A, Glycogen Synthase Kinase-3 inhibitor (BIO), and a blocking antibody against LIF. Although FAB-SCs expressed markers of pluripotency, they did not form teratomas and failed to expand when cultured as embryoid bodies. However, culture of FAB-SCs for 7 days in LIF/BMP-supplemented medium resulted in chimera formation, and this was associated with greater expression of E-cadherin. Therefore, our observations and those of Chou et al. [11] demonstrate that E-cadherin is likely to play a key role in the LIF-dependent self-renewal of pluripotent cells.

Self-renewal of E-cad−/− ES cells via FGF-2 and Activin/Nodal signaling is akin to that observed in the late epiblast in vivo [29], and in pluripotent mouse stem cells derived from epiblast (EpiSCs) [9]. Therefore, E-cad−/− ES cells may represent an epiblast-like phenotype. Although this is partly supported by our observation that E-cad−/− ES cells exhibit very low chimera formation efficiency [12] and express low levels of Rex-1, there are clear differences in transcript expression between these cell lines (supporting information; supporting information Table S2).

Our data suggest that distinct self-renewal signaling networks subsist within mouse ES cells, with activity dependent upon the cellular context. For example, whereas E-cad−/− ES cells cultured in serum-containing medium in the presence or absence of LIF self-renew via Activin/Nodal, these cells can also be cultured in serum-free medium supplemented with LIF and BMP4 (data not shown). This demonstrates that E-cad−/− ES cells possess a functional “ground state” signaling pathway (as previously described by Ying et al. [8]) as well as the ability to circumvent this pathway using FGF, Activin A, and Nodal. Indeed, our observation that wild-type ES cells treated with an E-cadherin homodimerization inhibiting peptide exhibit reversible LIF-independent self-renewal demonstrates that both Activin/Nodal-dependent and LIF/BMP-dependent signaling pathways are functional within wt ES cells. This suggests that Activin/Nodal signaling represents a default self-renewal pathway in E-cad−/− ES cells and, where these factors are absent, the cells can revert to LIF/BMP self-renewal. The ability of E-cad−/− ES cells to self-renew in the absence of LIF in serum-supplemented medium is likely to reflect the presence of FGF-2, Activin A, and Nodal within the serum. We have assessed various batches of ES cell-screened serum (e.g., from Invitrogen and PAA Laboratories, Linz, Austria, http://www.paa.com) and have found that all maintain E-cad−/− ES cells in an undifferentiated state in the absence of LIF. Ogawa et al. [30] demonstrated that treatment of wt ES cells with SB431542 dramatically decreased ES cell proliferation. Although we did not observe this effect in wt ES cells, this is likely a reflection of the different cell seeding densities (2,000 cells/well in Ogawa et al. [30], versus 1 × 105 cells/well in our study) and lengths of exposure to SB431542 (7 days versus 3 days, respectively) used in these two studies.

We recently demonstrated that cell surface E-cadherin protein is rapidly downregulated as part of the differentiation process of mouse [2] and human [16] ES cells, with this event exhibiting striking similarity to EMT events during both embryogenesis and tumor cell metastasis. Therefore, our results demonstrating that E-cad−/− and β-cat−/− mouse ES cells do not differentiate in the absence of LIF suggest that the E-cadherin–β-catenin complex plays a critical function in early differentiative events in ES cells. We hypothesize that the rapid downregulation of E-cadherin during ES cell differentiation leads to altered growth factor response of the cells, potentially leading to lineage-specific differentiation. What remains unclear, however, is the kinetics of loss of E-cadherin during ES cell differentiation and the way that this event integrates with other changes at the cell membrane that may influence cell signaling. For example, our observation that E-cad−/− ES cells do not differentiate in the absence of LIF suggests that loss of E-cadherin during wt ES cell differentiation would lead to a block in differentiation, and this clearly does not occur. Therefore, the function of the E-cadherin–β-catenin complex in ES cell differentiation is likely to be a multifaceted event dependent on spatiotemporal expression of these components and their associated effect on the localization of various cell surface molecules.

Although EMT events have been correlated with epithelial tumor cell metastasis in various organs, it is still unclear whether EMT is a prerequisite for tumor cell spread or is simply associated with this process. For example, there is recent evidence that tumor cells can spread in the absence of EMT [31, 32]. In addition, we previously showed that abrogation of E-cadherin in mouse or human ES cells leads to greater cellular motility in the absence of a characteristic EMT event [2]. Therefore, our observations of altered growth factor response in ES cells following abrogation of E-cadherin-mediated cell–cell contacts is unlikely to reflect a true EMT event. It is possible that the correlation between the loss of E-cadherin and a more aggressive tumor phenotype in vivo reflects a requirement for the cells to escape growth factor responses that are inhibitory to cell growth and proliferation, rather than greater cell motility per se.

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 have demonstrated that transhomodimerization of E-cadherin is required for LIF-dependent self-renewal of mouse ES cells. Abrogation of either E-cadherin or β-catenin results in loss of cell–cell contact and self-renewal via Activin A and Nodal signaling, with FGF-2 required for cellular proliferation. We conclude that β-catenin signaling is dispensable for the self-renewal of ES cells in medium supplemented with LIF and FBS because β-cat−/− ES cells maintained an undifferentiated phenotype under these conditions. Furthermore, we showed that LIF-independent self-renewal of ES cells is a reversible process, suggesting that multiple active self-renewal pathways subsist in these cells. Our results provide a novel function for E-cadherin in ES cells and may provide further elucidation of the mechanisms associated with the function of this protein in ES cells and during tumor cell metastasis.

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

C.M.W., A.M.E., S.R., F.S., and L.M. are supported by grants from the Association for International Cancer Research, the Biotechnology and Biological Sciences Research Council, the Royal Society, the Technology Strategy Board, and the Engineering and Physical Sciences Research Council.

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 available online.

FilenameFormatSizeDescription
STEM_134_sm_suppinfo.doc34KSupporting Information
STEM_134_sm_suppinfotable1.tif33446KSupporting Information Table 1
STEM_134_sm_suppinfotable2.tif15462KSupporting Information 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.