Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Epithelial–mesenchymal transition process in human embryonic stem cells cultured in feeder-free conditions

Epithelial–mesenchymal transition process in human embryonic stem cells cultured in feeder-free... Abstract Feeder-free human embryonic stem cell (hESC) culture is associated with the presence of mesenchymal-like cells appearing at the periphery of the colonies. The aim of this study was to identify this early differentiation process. Long-term feeder-free hESC cultures using matrigel and conditioned medium from mouse and from human origin revealed that the appearance of mesenchymal-like cells was similar regardless of the conditioned medium used. Standard characterization confirmed the preservation of hESC properties, but the feeder-free cultures could not be maintained longer than 37 passages. The early differentiation process was characterized in the short term after switching hESCs cultured on feeders to feeder-free conditions. Transmission electron microscopy showed an epithelium-like structure inside the hESC colonies, whereas the peripheral cells revealed the acquisition of a rather mesenchymal-like phenotype. Immunochemistry analysis showed that cells at the periphery of the colonies had a negative E-cadherin expression and a positive Vimentin expression, suggesting an epithelial–mesenchymal transition (EMT). Nuclear staining of β-catenin, positive N-cadherin and negative Connexin 43 expression were also found in the mesenchymal-like cell population. After RT–PCR analysis, Slug and Snail, both EMT-related transcription factors, were detected as up-regulated in the mesenchymal-like cell population. Taken together, our data suggest that culturing hESCs in feeder-free conditions enhances an early differentiation process identified as an EMT. human embryonic stem cells, feeder-free culture, matrigel, differentiation, epithelial–mesenchymal transition Introduction Human embryonic stem cell (hESC) lines present long-term self-renewal capacity and other characteristics of undifferentiated cells in specific culture conditions (Thomson et al., 1998). Because hESCs may be an unlimited source of material for future cell therapy due to their capacity to differentiate into any cell type, an important research topic is to characterize their early and spontaneous differentiation patterns in different culture conditions. At present, different methods of hESC culture exist. Most of them use mouse embryonic fibroblasts or human fibroblasts of different origin as feeder layers (Richards et al., 2002; Amit et al., 2003; Cheng et al., 2003). An alternative method is to culture hESCs on extracellular matrices, without any direct contact with feeder layers, in the presence of a conditioned medium obtained from fibroblast cultures (Xu et al., 2001). These latter culture conditions are very attractive with a view to the therapeutic use of hESCs. Long-term feeder-free cultures of hESCs have indeed been developed successfully to avoid animal cell contamination from, for example, fibroblasts to limit risks of cross-transfer of pathogens from xenogenic or allogenic feeders (Amit et al., 2004; Carpenter et al., 2004; Rosler et al., 2004). Important steps in the development of complete animal-free culture conditions were recently achieved with the derivation of new hESC lines in feeder-free conditions using a defined culture medium (Klimanskaya et al., 2005; Ludwig et al., 2006) and the report of an efficient autogenic feeder cell system using fibroblast-like cells spontaneously derived from the hESC colonies plated on matrigel and cultured in feeder-free conditions (Stojkovic et al., 2005). The hESC feeder-free culture system is known to induce spontaneous differentiation of hESCs (Xu et al., 2001; Rosler et al., 2004; Stojkovic et al., 2005; Ludwig et al., 2006). After plating hESC colonies on extracellular matrix-coated dishes in the presence of conditioned medium, fibroblast-like cells appear and surround the undifferentiated hESC colonies. So far, the nature of this spontaneous differentiation process leading to these fibroblast-like cells has not been studied. The aim of this study was to characterize the early differentiation process enhanced in feeder-free culture conditions. The first part of this study consisted of the maintenance of long-term hESC feeder-free cultures using matrigel matrix and conditioned medium of either mouse or human origin to evaluate the influence of the type of conditioned medium on the presence of the fibroblast-like cells and to determine whether the unique properties of the hESCs were preserved. In the second part of the work, the early differentiation process of three different hESC lines was characterized in the short term by morphological studies, immunostaining experiments and relative quantitative real-time RT–PCR analysis of specific transcripts. Materials and methods Culture of hESCs Three previously described hESC lines (VUB01, VUB03_DM1 carrying the mutation for myotonic dystrophy type 1 and VUB04_CF carrying a mutation in the CFTR gene) were at passage P53, P47 or P65 and P35, respectively, when used for our experiments (Mateizel et al., 2006). The three hESC lines had been maintained on mouse embryonic fibroblasts feeder layers with the hESC medium, which consisted of KnockOut Dulbecco’s modified Eagle’s medium (KO-DMEM) supplemented with 20% KnockOut Serum Replacement, 1 mM glutamine, 1% non-essential amino acids, 0.1 mM β-mercaptoethanol and 4 ng/ml human recombinant basic fibroblast growth factor (hbFGF), and passaged using a mechanical technique (Mateizel et al., 2006). For the long-term feeder-free culture studies, the VUB01 hESC line was switched to feeder-free culture conditions at passage 53. The hESCs were plated on 6-well plates (Nunc, Roskilde, Denmark) coated with a growth factor reduced-matrigel™ matrix (BD, Bedford, MA, USA; http://www.bdbiosciences.com, 1:30 dilution in KO-DMEM) and incubated at 37°C in 10% CO2. The hESC medium was conditioned on cultures of either mitomycin C inactivated mouse embryonic fibroblasts feeder layers, isolated from 14-days post-coitus fetuses (CF1 breed) and cultured until maximum passage 5, or with mitomycin C inactivated human fetal skin fibroblasts, isolated from skin biopsies of human aborted fetuses and maintained in culture for a maximum of 10 passages. For this purpose, 23.5 ml of hESC medium was left 24 h in a T-175 flask confluent with mouse or human fibroblasts. After collection, both types of conditioned medium were filtered and supplemented with 4 ng/ml hbFGF before feeding the hESCs. Passaging of the confluent hESC colonies grown in feeder-free conditions was carried out every 6 days after incubation with collagenase IV solution (1 mg/ml in KO-DMEM, Invitrogen) for 5–10 min. Phase contrast images of cell morphology were obtained using a Nikon Eclipse TE2005 and Eclipse Net software (Nikon instruments Belgium, Waver, Belgium). For the short-term culture studies, clumps of hESC colonies (VUB01, VUB03_DM1 and VUB04_CF) grown on feeder layers were obtained after cutting and careful collection with a Pasteur pipet to avoid contamination with mouse fibroblasts and then equally divided between hESC culture on mouse embryonic fibroblast feeder layers and feeder-free culture using matrigel™ with mouse embryonic fibroblast conditioned medium. Short culture periods were necessary to avoid confluent cell culture, which could interfere with the study of the early differentiation process. For real-time RT–PCR analysis, hESCs were cultured over a period from day 2 to day 4. For immunostaining experiments, the hESC colonies were plated on mouse embryonic fibroblasts- or on matrigel™-coated 4-well chamber slides (Sonic Seal Nalgene Nunc International, Rochester, NY, USA; http://www.nalgenunc.com) and allowed to expand until day 1 or day 3. For transmission electron microscopy experiments, the hESC colonies were plated on matrigel™-coated 4-well chamber slides and allowed to expand until day 3. Isolation of epithelium-like structures by partial trypsin digestion The isolation of the epithelium-like structures from the upper layers of hESC colonies cultured for 3 days in feeder-free conditions relied on short incubations in trypsin solution. E-cadherin-positive epithelial-like cells with strong specialized cell junctions have been shown to be protected from short trypsin digestion (Alvi et al., 2003; Grundemann et al., 2006). After 30 s of trypsin digestion (Invitrogen, 0.025% in phosphate-buffered saline [PBS]), the hESC colonies were disaggregated into a large fragment corresponding to an epithelium-like structure and single cells which were in suspension. The epithelium-like structure was collected by gentle pipetting, washed in hESC medium and allowed to adhere and grow in feeder-free conditions on matrigel™- coated 4-well chamber slides for 0, 12, 24 and 36 h. Standard characterization tests The karyotyping, the alkaline phosphatase activity test, the immunostaining experiments for SSEA-3, SSEA-4, TRA-1-60, TRA-1-81, the RT–PCR for NANOG, HPRT, REX-1 and SOX-2 genes, the in vitro and in vivo pluripotency tests on samples of hESCs cultured under long-term feeder-free conditions were performed as described previously (Mateizel et al., 2006). Light and transmission electron microscopy Sample preparation for microscopy studies was carried out as follows: hESCs were fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate, 1 mM CaCl2 (pH 7.4) and postfixed in 1% osmiumteroxide. The samples were embedded in Spurr’s resin, and for transmission electron microscopy, ultra-thin sections (90 nm) were prepared. After staining with lead citrate and uranyl acetate, the samples were examined in a Philips Tecnai 10 electron microscope (Eindhoven, The Netherlands). For light microscopy, semi-thin plastic sections (1 µm) were mounted on glass coverslips and were stained with 1% toluidine blue. Immunocytochemistry analysis To test for the presence of POU5F1_iA and of NANOG, the hESCs were fixed in 4% formaldehyde for 10 min and then permeabilized with 0.1% Triton X-100 for 10 min at room temperature. Following washing steps with PBS (3 × 5 min), incubation in 3% bovine serum albumin [(BSA) fraction V fatty acid-poor, nuclease and protease-free, Calbiochem, Darmstadt, Germany; http://www.calbiochem.com] was applied for 30 min to avoid non-specific reactions. Incubation with the mouse primary antibody for POU5F1_iA (IgG2b OCT-4 (C-10) sc-5279, Santa Cruz Biotechnology Inc., Heidelberg, Germany; http://www.scbt.com, 1:50 dilution in 1.5% BSA) and with the rabbit polyclonal antibody for NANOG (IgG, ab21624, Abcam, Cambridge, UK, http://www.abcam.com, 1:50 dilution in 1.5% BSA) was performed overnight at 4°C. The primary antibody for POU5F1_iA and for NANOG was detected with fluorescein-conjugated F(ab′)2 fragment of goat anti-mouse IgG and IgM (DakoCytomation, Glostrup, Denmark; 1:150 dilution in 1.5% BSA) and with Texas Red-conjugated goat anti-rabbit IgG (H+L) (Invitrogen, 1:100 dilution in 1.5% BSA), respectively. For the E-cadherin/Vimentin, E-cadherin/N-cadherin and E-cadherin/β-catenin immunostaining experiments, a fixation with cold methanol (at 4°C) was applied for 1 min. After blocking with 3% BSA, incubations with a mouse primary antibody for Vimentin (IgG1 clone V9, Sigma; Saint-Louis, USA; http://sigma-aldrich.com, 1:30 dilution in 1.5% BSA), with a mouse monoclonal antibody for N-cadherin (IgG1 clone 32, Becton Dickinson Transduction Laboratories, Lexington, KY, USA; http://www.bdbiosciences.com, 1:50 dilution in 1.5% BSA) and with a mouse monoclonal antibody for β-catenin (IgG3 clone 196618, R and D systems, Minneapolis, MN, USA; http://www.rndsystems.com, 1:100 dilution in 1.5% BSA) were carried out at room temperature for 1 h. The secondary antibody Texas Red-conjugated F(ab′)2 fragment of goat anti-mouse IgG (H+L) (Jackson ImmunoResearch Laboratories, West Grove, PA, USA; http://www.jacksonimmuno.com, 1:100 dilution in 1.5% BSA) was applied for 1 h after washing steps. A final incubation was performed with fluorescein-labelled E-cadherin antibody (mouse IgG2a, clone 36, Becton Dickinson Transduction Laboratories, 1:50 dilution in 1.5% BSA) for 1 h. For the E-cadherin/Connexin 43 immunostaining experiment, the same procedure was followed, but the detection of the primary antibody for Connexin 43 (polyclonal rabbit IgG, C6219, Sigma, 1:100 dilution in 1.5% BSA) was done with the secondary antibody Texas Red-conjugated goat anti-rabbit IgG (H+L) (Invitrogen, 1:100 dilution in 1.5% BSA). Negative controls were performed by omitting the primary antibodies or replacing them with a mouse immunoglobulin isotype used at the same dilution as the primary antibodies: HLA-DR IgG2b antibody (clone TÜ36, BD Biosciences) for POU5F1_iA, HLA-G IgG1 antibody (MEM-G/9, Exbio, Prague, Czech Republic; http://www.exbio.cz) for Vimentin and for N-cadherin, HLA-DR IgG3 (HL39, Exbio) for β-catenin, fluorescein-labelled IgG2a antibody (clone X39, BD Biosciences) for fluorescein isothiocyanate (FITC)-labelled E-cadherin and rabbit IgG (Invitrogen) for Connexin 43 and for NANOG served as negative controls. The images were scanned by confocal microscopy with an Argon–Krypton laser (488/568) (Fluoview IX70; Olympus, Belgium). RNA isolation and relative quantitative real-time RT–PCR A collagenase IV solution (1 mg/ml in KO-DMEM) was applied for 5–10 min to hESCs cultured under feeder-free conditions and for 1 h to hESCs cultured on feeder layers to ascertain that all the hESCs present in each dish were collected. For some experiments, the distinct cell populations of the hESC colonies within a feeder-free culture, that is, cells from the centre (epithelial-like) or from the periphery (mesenchymal-like), were collected separately using a Pasteur pipette. RNA extraction was carried out with the RNeasy kit (Qiagen, Hilden, Germany) and then treated with the RNase-free DNase kit (Qiagen). Five micrograms of total RNA was reverse-transcribed by using the first strand cDNA synthesis kit (Amersham Biosciences, Buckinghamshire, UK; http://www.amershambiosciences.com) with the NotI-d(T)18 primer, according to the manufacturer’s instructions. Relative quantitative real-time RT–PCR was performed on the ABI 7500 real-time PCR system (Applied Biosystems, Foster City, CA, USA; http://www.appliedbiosystems.com). The final reaction volume of 25 µl contained 12.5 µl of 2× TaqMan Universal Master Mix (Applied Biosystems), 1.25 µl of 20× Assays-on-demand Gene Expression assay mix (Applied Biosystems) and 10–100 ng complementary DNA (cDNA) in 11.25 µl nuclease-free water. The primers and the probes for the gene expression analysis of E-cadherin, Vimentin, Snail (Snail1), Slug (Snail2), POU5F1_iA (OCT-3A) and GAPDH were purchased from Applied Biosystems (Assays-on-demand gene expression products, Applera International Inc., Pleasanton, CA, USA). The following conditions were used: 2 min at 50°C, 10 min at 95°C, 40 cycles of 15 s at 95°C and 1 min at 60°C. All experiments were run in triplicate. Relative quantification of gene expression between multiple samples was achieved by normalization against the endogenous control GAPDH using the ΔΔCt method of quantification. Fold changes were calculated as 2−ΔΔCt. Statistical analysis Each experiment was repeated at least three times. The presented values are given as means ± SD. Differences between groups were determined by paired t-test. A P-value of less than 0.05 was considered significant. Results Characterization of the long-term hESC feeder-free culture The VUB01 hESC line that had originally been derived on mouse embryonic fibroblasts feeder layers was switched to feeder-free culture systems using matrigel with mouse conditioned medium or human conditioned medium to investigate the effect of the type of conditioned medium used on the early differentiation process and on the pluripotent capacities of hESCs. We did not find any difference regarding the standard characterization results between both types of conditioned medium (Table I). The hESCs showed a stable normal karyotype, even after long-term culture under feeder-free conditions. Colonies of hESCs grown in feeder-free conditions showed an alkaline phosphatase activity and a positive immunostaining for the cell surface markers SSEA-3, SSEA-4, TRA-1-60 and TRA-1-81 that were previously described as hESC specific (Thomson et al., 1998) (Figure 1A–D). As revealed by RT–PCR, VUB01 hESCs maintained under feeder-free culture conditions also expressed several pluripotent stem cell markers such as NANOG, SOX-2 and REX-1 (data not shown), whereas their pluripotent capacity was proven in vitro by the formation of embryoid bodies (data not shown) and in vivo by the ability to form a teratoma-containing cells from the three germ layers after intramuscular injection of hESCs in a severely compromised immunodeficient (SCID) mouse (Figure 1E–G). Table I. Summary of standard characterization analysis tested during long-term feeder-free culture of VUB01 at different passages (P) using matrigel and two different conditioned media: on mouse embryonic fibroblast conditioned medium (MEF CM) and on human fetal fibroblast conditioned medium (HFF CM) Long-term feeder-free growth of VUB01  MEF CM  HFF CM  Karyotypes          Early passage (P8)  46 XY  46 XY      Late passages (P27 and P35)  46 XY  46 XY  Alkaline phosphatase          Early passage (P8)  +  +      Late passage (P33)  +  +  Immunostaining of hESC-specific surface markers SSEA-3, SSEA-4 and TRA-1-60, TRA-1-81          Early passage (P10)  +  +      Late passage (P27)  +  +  RT–PCR pluripotency associated genes expressions          Early passage (P8)  +  +      Late passage (P27 and P35)  +  +  Embryoid bodies formation          Early passage (P10)  +  +      Late passage (P28)  +  +  Teratoma formation (P17)  +  +  Long-term feeder-free growth of VUB01  MEF CM  HFF CM  Karyotypes          Early passage (P8)  46 XY  46 XY      Late passages (P27 and P35)  46 XY  46 XY  Alkaline phosphatase          Early passage (P8)  +  +      Late passage (P33)  +  +  Immunostaining of hESC-specific surface markers SSEA-3, SSEA-4 and TRA-1-60, TRA-1-81          Early passage (P10)  +  +      Late passage (P27)  +  +  RT–PCR pluripotency associated genes expressions          Early passage (P8)  +  +      Late passage (P27 and P35)  +  +  Embryoid bodies formation          Early passage (P10)  +  +      Late passage (P28)  +  +  Teratoma formation (P17)  +  +  hESC, human embryonic stem cell. View Large Figure 1. View largeDownload slide Standard characterization tests of human embryonic stem cells (hESCs) cultured under long-term feeder-free conditions: expression of cell surface markers (VUB01 at passage 27): SSEA-3 (A), SSEA-4 (B), TRA-1-60 (C) and TRA-1-81 (D). In vivo differentiation (VUB01 at passage 17): immunohistochemistry analysis of teratoma sections with neurofilament 200K antibody shows neuronal tissue (ectoderm) (E), with actin antibody shows smooth muscle (mesoderm) (F) and with cytokeratin 18 antibody shows primitive epithelium (endoderm) (G). All positive-stained structures are marked with arrows. Original magnification: (A–G) ×100. Figure 1. View largeDownload slide Standard characterization tests of human embryonic stem cells (hESCs) cultured under long-term feeder-free conditions: expression of cell surface markers (VUB01 at passage 27): SSEA-3 (A), SSEA-4 (B), TRA-1-60 (C) and TRA-1-81 (D). In vivo differentiation (VUB01 at passage 17): immunohistochemistry analysis of teratoma sections with neurofilament 200K antibody shows neuronal tissue (ectoderm) (E), with actin antibody shows smooth muscle (mesoderm) (F) and with cytokeratin 18 antibody shows primitive epithelium (endoderm) (G). All positive-stained structures are marked with arrows. Original magnification: (A–G) ×100. During the long-term hESC feeder-free cultures, mesenchymal-like cells appeared at the periphery of the hESC colonies, and these cells were similarly observed regardless of the conditioned medium used (mouse or human origin). Around the 30th passage, a progressive enrichment of these cells within the two feeder-free culture systems was observed with the complete differentiation and the further loss of the hESC colonies around the 37th passage (9 months of culture). Morphology of the short-term hESC feeder-free culture The switching of the colonies of three different hESC lines (VUB01, VUB03_DM1 and VUB04_CF) from culture on mouse embryonic fibroblasts feeder layers to feeder-free conditions led to the appearance of a characteristic cell population with a flattened, mesenchymal-like phenotype at the periphery of the colonies (Figure 2A). These cells were observed under phase contrast light microscopy 24 h after switching culture conditions, and they further expanded around the hESC colonies until they reached confluency (Figure 2A–C). Such mesenchymal-like cells were rarely observed in parallel control cultures on feeder layers (Figure 2D–F). Figure 2. View largeDownload slide Morphology of human embryonic stem cell (hESC) colonies cultured under feeder-free conditions and on feeder layers (VUB03_DM1 at passage 47). Phase contrast images of hESC colonies grown under feeder-free conditions using matrigel and mouse embryonic fibroblast conditioned medium on day 2 (A), day 3 (B) and day 4 (C) and images of hESC colonies cultured on mouse embryonic fibroblast feeder layers on day 2 (D), day 3 (E) and day 4 (F). Arrows mark the periphery of the hESC colonies where mesenchymal-like cells are emerging. Original magnification: (A–F) ×100. Figure 2. View largeDownload slide Morphology of human embryonic stem cell (hESC) colonies cultured under feeder-free conditions and on feeder layers (VUB03_DM1 at passage 47). Phase contrast images of hESC colonies grown under feeder-free conditions using matrigel and mouse embryonic fibroblast conditioned medium on day 2 (A), day 3 (B) and day 4 (C) and images of hESC colonies cultured on mouse embryonic fibroblast feeder layers on day 2 (D), day 3 (E) and day 4 (F). Arrows mark the periphery of the hESC colonies where mesenchymal-like cells are emerging. Original magnification: (A–F) ×100. To analyse the morphology of these structures in more detail, we made semi-thin and ultra-thin sections of plastic-embedded feeder-free cultures of three hESC lines (VUB01, VUB03_DM1 and VUB04_CF). Analysis of sections revealed that the cell morphology changed depending on the location within the colony and that three different zones could be defined (Figure 3A–C). In the central part of the colonies, a single upper layer of columnar cells facing the medium (zone 1) covered a multilayered core of small rounded cells (zone 2), and at the periphery of the colonies, a monolayer of squamous cells was observed (zone 3). The columnar cells (zone 1) showed epithelial-like characteristics such as a clear polarization with a basal cell nucleus and apical microvilli (Figure 3A–C). At the ultrastructural level, the microvilli were associated with prominent bundles of microfilaments running parallel to the cell surface, and cell junctions were present at the apical–lateral side of the columnar cells (Figure 3D–F). The small rounded cells showing a high nuclear–cytoplasm ratio (zone 2) were closely apposed and had an electron-lucent cytoplasm with relatively few cell organelles but with prominent polyribosomes (Figure 3G). As previously described by other groups (Sathananthan et al., 2002; Ginis et al., 2004), such a multilayered structure with epithelial-like characteristics was also observed in the colonies formed by hESC cultured on feeder layers. Figure 3. View largeDownload slide Light microscopy images and transmission electron microscopy images of human embryonic stem cell (hESC) colonies grown under feeder-free conditions (VUB03_DM1 at passage 47). Graphical presentation of a transverse section through a hESC colony plated on matrigel showing three zones with different hESC populations (zones 1, 2 and 3) (A). At the junction (marked by bold curved line) between the colony (zone 1 and zone 2) and the periphery of the colony (zone 3), some mesenchymal-like cells still displayed cell junctions (A). A red line demarcates the upper layer of the colony showing epithelial-like columnar cells with an apico-basal polarization (zone 1) from the central multilayered part of the colony with rounded cells with a high nucleus to cytoplasm ratio (zone 2) (B). The border of the hESC colony showed the transition from central parts towards the periphery consisting of a monolayer of mesenchymal-like cells (zone 3) (C). At the ultrastructure level, columnar cells within the upper layer (zone1) show epithelial-like features (D) such as microvilli at the apical surface, specialized cell junctions at the apical sides (marked with double black arrows) (F) and prominent bundles of microfilaments parallel to the cell surface at the apical surface (E). Small rounded cells within the central parts of the hESC colony (zone 2) show an electron-lucent cytoplasm with few organelles (G). The mesenchymal-like cells at the periphery of the colony (zone 3) show a squamous morphology with a low nucleus–cytoplasm ratio and do not have epithelial-like features (H). Original magnification: (B–C) ×400, (D) ×3700, (E) ×12 500, (F) ×30 000, (G) ×10 500 and (H) ×3500. Figure 3. View largeDownload slide Light microscopy images and transmission electron microscopy images of human embryonic stem cell (hESC) colonies grown under feeder-free conditions (VUB03_DM1 at passage 47). Graphical presentation of a transverse section through a hESC colony plated on matrigel showing three zones with different hESC populations (zones 1, 2 and 3) (A). At the junction (marked by bold curved line) between the colony (zone 1 and zone 2) and the periphery of the colony (zone 3), some mesenchymal-like cells still displayed cell junctions (A). A red line demarcates the upper layer of the colony showing epithelial-like columnar cells with an apico-basal polarization (zone 1) from the central multilayered part of the colony with rounded cells with a high nucleus to cytoplasm ratio (zone 2) (B). The border of the hESC colony showed the transition from central parts towards the periphery consisting of a monolayer of mesenchymal-like cells (zone 3) (C). At the ultrastructure level, columnar cells within the upper layer (zone1) show epithelial-like features (D) such as microvilli at the apical surface, specialized cell junctions at the apical sides (marked with double black arrows) (F) and prominent bundles of microfilaments parallel to the cell surface at the apical surface (E). Small rounded cells within the central parts of the hESC colony (zone 2) show an electron-lucent cytoplasm with few organelles (G). The mesenchymal-like cells at the periphery of the colony (zone 3) show a squamous morphology with a low nucleus–cytoplasm ratio and do not have epithelial-like features (H). Original magnification: (B–C) ×400, (D) ×3700, (E) ×12 500, (F) ×30 000, (G) ×10 500 and (H) ×3500. Although the monolayer of squamous cells present at the periphery of the colonies, showing a flattened mesenchymal-like morphology (zone 3), appeared to be specific for the feeder-free culture conditions, these cells had a low nuclear–cytoplasm ratio and lacked regular cell–cell contacts, microvilli and apico-basal polarity (Figure 3H). At the junction between the colony and the periphery of the colonies, some mesenchymal-like cells still displayed cell junctions (Figure 3C). Characterization of the short-term hESC feeder-free culture by immunostaining In the light of our morphological observations, we performed immunostaining analysis for E-cadherin and Vimentin, chosen as specific markers for epithelial cells and for mesenchymal cells, respectively. E-cadherin is a cell–cell adhesion protein expressed in epithelial cells (Bloor et al., 2002; Lee et al., 2006). Vimentin is a type III intermediate filament normally expressed in mesenchymal cells but was also described in epithelial cells migrating during embryological, organogenesis or pathological processes (Guarino, 1995; Gilles et al., 1999; Lee et al., 2006). When hESCs (VUB01, VUB03_DM1 and VUB04_CF) were cultured on feeder layers, E-cadherin was strongly expressed as a web-like structure in the membrane of hESCs found in the upper layers of the colonies and largely disappeared in the lower layers of the colonies after scanning with a confocal microscope (Figure 4A). A similar distribution of E-cadherin was found 24 h after switching to feeder-free conditions, although the staining was clearly less intense (Figure 4D). After 3 days in feeder-free conditions, the E-cadherin web-like expression pattern was less structured, and the intensity of the staining had further decreased (Figure 4G). In contrast, at the periphery of the colonies, E-cadherin was barely detectable in the large population of mesenchymal-like cells (Figure 4I). Figure 4. View largeDownload slide E-cadherin/Vimentin immunostaining of human embryonic stem cell (hESC) colonies grown under feeder-free conditions, on feeder layers and of isolated epithelium-like structures. (VUB03_DM1 at passage 47): images of a hESC colony grown on mouse embryonic fibroblast feeder layers and fixed after 3 days of culture (A–C). Images of hESC colonies switched to feeder-free culture conditions and fixed after 24 h of culture (D–F) and after 3 days of culture (focus on the border of the colony) (G–I). (VUB03_DM1 at passage 65): images of isolated epithelium-like structures fixed and stained after 0, 12, 24 and 36 h (J–M). Images scanned by confocal microscopy after immunostaining with E-cadherin antibody (green) (A with an additional section through the middle of the hESC colony, D and G), with Vimentin antibody (red) (B, E and H) and merged E-cadherin and Vimentin immunostaining images (C, F, I, J–M). Original magnification: (A–C, G–M) ×200 and (D–F) ×400. Figure 4. View largeDownload slide E-cadherin/Vimentin immunostaining of human embryonic stem cell (hESC) colonies grown under feeder-free conditions, on feeder layers and of isolated epithelium-like structures. (VUB03_DM1 at passage 47): images of a hESC colony grown on mouse embryonic fibroblast feeder layers and fixed after 3 days of culture (A–C). Images of hESC colonies switched to feeder-free culture conditions and fixed after 24 h of culture (D–F) and after 3 days of culture (focus on the border of the colony) (G–I). (VUB03_DM1 at passage 65): images of isolated epithelium-like structures fixed and stained after 0, 12, 24 and 36 h (J–M). Images scanned by confocal microscopy after immunostaining with E-cadherin antibody (green) (A with an additional section through the middle of the hESC colony, D and G), with Vimentin antibody (red) (B, E and H) and merged E-cadherin and Vimentin immunostaining images (C, F, I, J–M). Original magnification: (A–C, G–M) ×200 and (D–F) ×400. Vimentin staining was restricted to the borders of the hESC colonies grown on feeder layers (Figure 4B and C). In agreement with our morphological observations, a large population of mesenchymal-like cells stained positive for Vimentin at the periphery of the hESC colonies, 24 h after switching to feeder-free conditions (Figure 4E and F). After 3 days in feeder-free culture, Vimentin-positive mesenchymal-like cells were found more abundantly at the periphery of the hESCs (Figure 4H and I). These changes in marker expression suggested that the upper layer of the colony could be considered as an epithelium undergoing an epithelial–mesenchymal transition (EMT) process after switching from the hESCs culture on feeder layers to feeder-free culture. To confirm that the mesenchymal-like cells are derived from E-cadherin-positive epithelial-like cells through an EMT process, we performed short-term cultures (0–36 h) of isolated epithelium-like structures from the upper layer of hESC colony cultured under feeder-free conditions (Figure 4J–M). Direct fixation and staining of the isolated fragment (t 0 h, Figure 4J) showed only E-cadherin-positive cells. In contrast, culturing this fragment under feeder-free conditions demonstrated that the E-cadherin-positive epithelial-like cells progressively underwent an EMT. Indeed, the samples stained at 12 and 24 h (Figure 4K and L) displayed a progressive decrease of E-cadherin-positive cells and an increase of Vimentin-positive mesenchymal-like cells. After 36 h of culture (Figure 4M), only Vimentin-positive mesenchymal-like cells were observed. Aiming at a better characterization of the phenotypes found in our hESC colonies cultured under feeder-free conditions, we next performed immunostaining analysis with the markers, Connexin 43, N-cadherin and β-catenin. Connexin 43 is a protein forming functional gap junctions that mediate signals from one cell to the adjacent cells and is known to be expressed inside hESC colonies cultured on feeder layers (Wong et al., 2004). N-cadherin is an integral cell-to-cell adhesion protein, and β-catenin is a cadherin-associated protein involved in the regulation of the membrane structure and of the cell adhesion. An E-cadherin down-regulation, a concomitant N-cadherin up-regulation inside the mesenchymal-like population and a β-catenin accumulation in the nucleus are recognized as typical changes in marker expressions during an EMT process (Muller et al., 2002; Maeda et al., 2005; Lee et al., 2006). After E-cadherin/Connexin 43 immunostaining analysis, the E-cadherin-positive upper layer cells and central layer cells of a hESC colony cultured under feeder-free conditions were positive for Connexin 43 detected at the membrane level (Figure 5A–C). On the contrary, the mesenchymal-like cells present at the periphery of the hESC colonies cultured under feeder-free conditions were showing a weaker Connexin 43 expression and an intracytoplasmic pattern of staining instead of a transmembrane localization (Figure 5B). Figure 5. View largeDownload slide E-cadherin/Connexin 43, E-cadherin/N-cadherin and E-cadherin/β-catenin immunostaining of human embryonic stem cell (hESC) colonies grown under feeder-free conditions (VUB03_DM1 at passage 47). Images of the border of hESC colonies switched to feeder-free culture conditions and fixed after 3 days of culture (A–I). Images scanned by confocal microscopy after immunostaining with E-cadherin antibody (green) (A, D, G), with Connexin 43 antibody (red) (B), N-cadherin antibody (red) (E) and β-catenin antibody (red) (H) and merged E-cadherin/Connexin 43 (C), E-cadherin/N-cadherin (F) and E-cadherin/β-catenin (I) immunostaining images. Arrows mark the periphery of the hESC colonies where mesenchymal-like cells are emerging. Double white arrows indicate some of the nuclei stained by β-catenin antibody. Original magnification: (A–I) ×200. Figure 5. View largeDownload slide E-cadherin/Connexin 43, E-cadherin/N-cadherin and E-cadherin/β-catenin immunostaining of human embryonic stem cell (hESC) colonies grown under feeder-free conditions (VUB03_DM1 at passage 47). Images of the border of hESC colonies switched to feeder-free culture conditions and fixed after 3 days of culture (A–I). Images scanned by confocal microscopy after immunostaining with E-cadherin antibody (green) (A, D, G), with Connexin 43 antibody (red) (B), N-cadherin antibody (red) (E) and β-catenin antibody (red) (H) and merged E-cadherin/Connexin 43 (C), E-cadherin/N-cadherin (F) and E-cadherin/β-catenin (I) immunostaining images. Arrows mark the periphery of the hESC colonies where mesenchymal-like cells are emerging. Double white arrows indicate some of the nuclei stained by β-catenin antibody. Original magnification: (A–I) ×200. After E-cadherin/N-cadherin immunostaining analysis, the E-cadherin-positive staining of the upper layer of epithelial-like cell population was found to be switched at the periphery of the hESC colonies cultured under feeder-free conditions to an N-cadherin web-like structure (Figure 5D–F). A strongly positive N-cadherin staining pattern was detected in the membranes of the mesenchymal-like cell population (Figure 5E). Finally, after E-cadherin/ β-catenin immunostaining analysis, a β‐catenin-positive staining was found at the membrane level in the E‐cadherin-positive epithelial-like cell population and in the central layer cells of the hESC colonies cultured under feeder-free conditions (Figure 5G–I), whereas a strong positive β-catenin staining was detected in the nuclei of the mesenchymal-like cell population present at the periphery (Figure 5I). Immunostaining analysis for stem cell markers of pluripotency POU5F1_iA (formerly called OCT-3A) and NANOG was carried out on hESC colonies (VUB01, VUB03_DM1 and VUB04_CF) switched to feeder-free conditions to see whether the early differentiation process influenced their expression (Figure 6) (Hyslop et al., 2005; Cauffman et al., in press). Expression of these nuclear proteins was found in the central parts of the hESC colonies as well as in the mesenchymal-like cells at the periphery (Figure 6A–F). Immunostaining for both POU5F1_iA and NANOG identified, within the mesenchymal-like cell population, larger, irregular nuclei with a decreased intensity (Figure 6B and E) compared with the small, rounded nuclei of the central parts of the hESC colonies (Figure 6A and D). Figure 6. View largeDownload slide POU5F1_iA and NANOG immunostaining of human embryonic stem cell (hESC) colonies grown under feeder-free conditions (VUB04_CF at passage 35). Images of central part (A, D), at the periphery (B, E) and at the transition zone between central parts to the periphery (C, F) of a hESC colony switched to feeder-free conditions and fixed after 3 days of culture. Images scanned by confocal microscopy after immunostaining with POU5F1_iA antibody (A–C) and NANOG antibody (D–F). Arrows mark the periphery of the hESC colonies where mesenchymal-like cells are emerging. Original magnification: (A, B, D and E) ×400, (C) ×100 and (F) ×200. Figure 6. View largeDownload slide POU5F1_iA and NANOG immunostaining of human embryonic stem cell (hESC) colonies grown under feeder-free conditions (VUB04_CF at passage 35). Images of central part (A, D), at the periphery (B, E) and at the transition zone between central parts to the periphery (C, F) of a hESC colony switched to feeder-free conditions and fixed after 3 days of culture. Images scanned by confocal microscopy after immunostaining with POU5F1_iA antibody (A–C) and NANOG antibody (D–F). Arrows mark the periphery of the hESC colonies where mesenchymal-like cells are emerging. Original magnification: (A, B, D and E) ×400, (C) ×100 and (F) ×200. Characterization of the short-term hESC feeder-free culture by relative quantitative real-time RT–PCR The mRNA levels of E-cadherin, Vimentin, Snail, Slug and POU5F1_iA were studied by relative quantitative real-time RT–PCR in samples from hESCs (VUB01, VUB03_DM1 and VUB04_CF) cultured on feeder layers or from hESCs switched to feeder-free culture (Figure 7A). Snail (Snail1) and Slug (Snail2) belong to the Snail genes family and are commonly used as EMT molecular markers. Both are zinc finger protein transcription factors shown to function as direct repressors of E-cadherin and inducers of the EMT process (Nieto et al., 1994; Savagner et al., 1997; Hajra et al., 2002; Lee et al., 2006). Figure 7. View largeDownload slide Relative quantification real-time RT–PCR result graphics. (A) Relative quantification of E-cadherin, Snail, Slug, Vimentin and POU5F1_iA mRNA levels in samples from human embryonic stem cells (hESCs) cultured on feeder layers and from hESCs cultured under feeder-free conditions. The samples were collected on days 2, 3 and 4. The samples from hESCs culture on feeder layers on day 2 were considered as a reference (value of 1), and data are represented as means ± SD. *, P < 0.05 compared with feeder layers at day 2. α, P < 0.05 compared with mRNA relative level of E-cadherin on MG at day 2. β, P < 0.05 compared with mRNA relative level of POU5F1_iA on MG at day 2. (B) Relative quantification of E-cadherin, Snail, Slug, Vimentin and POU5F1_iA mRNA levels of epithelial-like and mesenchymal-like human embryonic stem cell (hESC) populations located within the central parts and at the periphery of the hESC colonies, respectively, and both collected on day 3 of culture. The mRNA levels of each studied gene in the mesenchymal-like hESC population were compared with the epithelial-like hESCs population considered as reference (value of 1), and data are represented as means ± SD. *, P < 0.05 compared with the epithelial-like hESCs population. These experiments (A and B) were repeated at least three times with VUB03_DM1 at passage 47, and the analyses are shown here. These experiments were also carried out with VUB01 and VUB04_CF hESC lines and showed similar results (data not shown). Figure 7. View largeDownload slide Relative quantification real-time RT–PCR result graphics. (A) Relative quantification of E-cadherin, Snail, Slug, Vimentin and POU5F1_iA mRNA levels in samples from human embryonic stem cells (hESCs) cultured on feeder layers and from hESCs cultured under feeder-free conditions. The samples were collected on days 2, 3 and 4. The samples from hESCs culture on feeder layers on day 2 were considered as a reference (value of 1), and data are represented as means ± SD. *, P < 0.05 compared with feeder layers at day 2. α, P < 0.05 compared with mRNA relative level of E-cadherin on MG at day 2. β, P < 0.05 compared with mRNA relative level of POU5F1_iA on MG at day 2. (B) Relative quantification of E-cadherin, Snail, Slug, Vimentin and POU5F1_iA mRNA levels of epithelial-like and mesenchymal-like human embryonic stem cell (hESC) populations located within the central parts and at the periphery of the hESC colonies, respectively, and both collected on day 3 of culture. The mRNA levels of each studied gene in the mesenchymal-like hESC population were compared with the epithelial-like hESCs population considered as reference (value of 1), and data are represented as means ± SD. *, P < 0.05 compared with the epithelial-like hESCs population. These experiments (A and B) were repeated at least three times with VUB03_DM1 at passage 47, and the analyses are shown here. These experiments were also carried out with VUB01 and VUB04_CF hESC lines and showed similar results (data not shown). The mRNA levels of E-cadherin, Vimentin, Snail, Slug and POU5F1_iA in hESCs grown on feeder layers remained stable from day 2 until day 4 (Figure 7A). The expression levels of Snail, Slug and Vimentin in hESCs at the second day of feeder-free culture were up-regulated 2.9-fold, 7.8-fold and 10.4-fold, respectively, compared with those of hESCs grown on feeder layers (Figure 7A). Time course experiments until day 4 revealed a maintenance of the mRNA levels of Snail, Slug and Vimentin in hESCs cultured under feeder-free conditions, whereas the mRNA levels of E-cadherin significantly decreased from 1.8-fold at day 2 to 0.2-fold at day 4 (Figure 7A). Similarly, POU5F1_iA mRNA levels decreased from 0.9-fold at day 2 to 0.2-fold at day 4 (Figure 7A). To exclude a possible contamination with murine mRNA in the real-time RT–PCR experiments, samples with cDNA derived from mouse fibroblasts were included in the different assays. These samples were found to be positive for Vimentin only. Although contamination with murine mRNA is very unlikely because care was taken to collect hESCs only, contamination with murine mRNA in samples grown on feeder layers would lead to an underestimation of the Vimentin mRNA level in hESCs cultured under feeder-free conditions. To analyse E-cadherin, Snail, Slug, Vimentin and POU5F1_iA mRNA levels in samples enriched with mesenchymal-like and in samples enriched with epithelial-like cells cultured in feeder-free conditions, relative quantitative real-time RT–PCR analysis were carried out (Figure 7B). The results showed a 0.3-fold down-regulation of E-cadherin and a 0.6-fold down-regulation of POU5F1_iA concomitant with a 2.1-fold up-regulation of Snail, a 2.9-fold up-regulation of Slug and a 3.1-fold up-regulation of Vimentin in the mesenchymal-like cell population when compared with the epithelial-like cell population (Figure 7B). Discussion In the first part of this work, we showed that feeder-free culture using matrigel and conditioned medium of mouse or human origin could be an appropriate method for maintaining the self-renewal and pluripotency of hESCs, which were derived on feeder layers, whereas it is known that the use of non-conditioned hESC medium in feeder-free culture conditions gives a total differentiation of the hESC colonies after 48 h (Xu et al., 2005). We were unable to maintain the hESC feeder-free cultures for longer than 37 passages (9 months), whereas the maintenance of the VUB01 hESC line on mouse embryonic fibroblasts feeder layers is still ongoing after more than 100 passages (Mateizel et al., 2006). The majority of previously published results describe in vitro conditions for feeder-free hESC growth for less than 9 months (Carpenter et al., 2004; Rosler et al., 2004). We did not detect any chromosomal anomalies, but karyotypic instabilities were detected after feeder-free derivations and cultures of hESC lines in defined conditions (Ludwig et al., 2006). The difference in efficiency between the two methods in maintaining the hESC colonies in culture may be because of the difference in the passage technique. Clumps of hESCs cultured on feeder layers were passaged using a mechanical technique, which enabled the selection of ‘nice-looking’ hESCs, whereas an enzymatic treatment was necessary to passage the hESCs grown in feeder-free conditions, thereby excluding any selection of cells. Another explanation may be that feeder-free culture conditions force hESC colonies to undergo an early differentiation into mesenchymal-like cells. This phenomenon was previously mentioned in several manuscripts, but the nature of this process has not yet been identified (Xu et al., 2001; Carpenter et al., 2004; Rosler et al., 2004; Ludwig et al., 2006). The results of the second part of this study suggested that this differentiation process was associated with an EMT phenomenon. Indeed, the conversion of an epithelial cell to a mesenchymal cell requires the down-regulation of adhesion protein E-cadherin, the concurrent expression of Vimentin and the acquisition of migratory properties by the cells. These events mark the initiation of EMT (Lee et al., 2006). The EMT phenomenon occurs during different stages of human embryogenesis such as the migration of epiblast cells when they are allocated through the formation of the primitive streak to the three germ layers or the turning of an epithelial somite into a sclerotome mesenchyme (Nieto et al., 1994; Barrallo-Gimeno and Nieto, 2005; Hay, 2005). During these crucial steps, E-cadherin protein expression on the cell surface is decreased (Ciruna and Rossant, 2001; Gumbiner, 2005). Different signalling pathways induce several transcription factors such as Slug and Snail, which are direct repressors of E-cadherin (Ciruna and Rossant, 2001; Hajra et al., 2002; Lee et al., 2006). Coinciding with the loss of epithelial markers, cells undergoing EMT start to express mesenchymal markers and can migrate (Guarino, 1995; Lee et al., 2006). EMT also occurs in cancer cells and is implicated in the progression of primary tumours towards metastasis (Barrallo-Gimeno and Nieto, 2005; Lee et al., 2006). EMT is also seen during physiological situations such as wound healing or in pathological processes such as tumour invasion (Barrallo-Gimeno and Nieto, 2005; Lee et al., 2006) Taking all the data together, the observed morphological features and correlated changes in marker expression (recapitulated in Table II) suggest that the mesenchymal-like population emerging at the periphery of hESC colonies cultured under feeder-free conditions spontaneously replicated an EMT process (Lee et al., 2006). Supporting this analysis, EMT was recently described as being involved during directed differentiation of hESCs into endoderm (D’Amour et al., 2005). Moreover, a culture model system of rhesus monkey ESCs was also reported to replicate an EMT process (Behr et al., 2005). Table II. Summary of characterization analysis tested of three cell population: the upper layer cells (zone 1), the central multilayered part (zone 2) and the periphery (zone 3) of hESC colony in short-term feeder-free culture   Zone 1  Zone 2  Zone 3  Morphological studies  Epithelial-like cells  Undifferentiated rounded cells  Mesenchymal-like cells  Presence of cell-to-cell junctions  Specialized cell junctions and gap junctions  Gap junctions  No cell-to-cell junction  Immunostaining analysis of EMT markers            E-cadherin expression  Positive membrane  Negative  Negative      Vimentin expression  Negative  Negative  Positive      Connexin 43 expression  Positive membrane  Positive membrane  Negative cytoplasmic      N-cadherin expression  Negative  Negative  Positive membrane      β-catenin expression  Positive membrane  Positive membrane  Nuclear accumulation  Immunostaining analysis of pluripotency markers            POU5F1_iA  Positive  Positive  Positive but weaker      NANOG  Positive  Positive  Positive but weaker  RT–PCR analysis            Snail expression  Low mRNA levels  Low mRNA levels  mRNA levels up-regulated      Slug expression  Low mRNA levels  Low mRNA levels  mRNA levels up-regulated    Zone 1  Zone 2  Zone 3  Morphological studies  Epithelial-like cells  Undifferentiated rounded cells  Mesenchymal-like cells  Presence of cell-to-cell junctions  Specialized cell junctions and gap junctions  Gap junctions  No cell-to-cell junction  Immunostaining analysis of EMT markers            E-cadherin expression  Positive membrane  Negative  Negative      Vimentin expression  Negative  Negative  Positive      Connexin 43 expression  Positive membrane  Positive membrane  Negative cytoplasmic      N-cadherin expression  Negative  Negative  Positive membrane      β-catenin expression  Positive membrane  Positive membrane  Nuclear accumulation  Immunostaining analysis of pluripotency markers            POU5F1_iA  Positive  Positive  Positive but weaker      NANOG  Positive  Positive  Positive but weaker  RT–PCR analysis            Snail expression  Low mRNA levels  Low mRNA levels  mRNA levels up-regulated      Slug expression  Low mRNA levels  Low mRNA levels  mRNA levels up-regulated  EMT, epithelial–mesenchymal transition; hESC, human embryonic stem cell. View Large With the morphological studies, we observed that the hESC colonies grown in feeder-free conditions showed a central multilayered part of the hESC colonies with small rounded undifferentiated cells (zone 2) and an upper epithelium-like structure (zone 1) which had a more specialized phenotype, most probably because these cells are in direct contact with the culture medium. These epithelial-like features were shown to be lost inside the mesenchymal-like cell population present at the periphery of the hESC colonies (zone 3). For the immunostaining analyses, E-cadherin was chosen as an epithelial lineage-specific marker. Although E-cadherin is expressed in epithelial cells such as trophectoderm cells, inner cell mass cells and ectodermal cells (Alikani, 2005; D’Amour et al., 2005), some authors consider E-cadherin as a hESC-specific marker because it is found in undifferentiated hESC colonies but not in differentiated embryoid bodies (Cai et al., 2005). In this present study, a positive E-cadherin expression was shown in the upper layer of the hESC colonies (zone 1). In contrast, E-cadherin was barely detectable in the central multilayered part (zone 2) confirming the lack of epithelial features in this zone. The mesenchymal-like cell population emerging at the periphery of the hESC colonies (zone 3) did not show any E-cadherin expression, whereas a strong expression of mesenchymal lineage-specific Vimentin protein was detected within this last cell population. After switching to feeder-free culture, the changes of E-cadherin and Vimentin markers were confirmed at the mRNA level by real-time RT–PCR. The expression of Snail and Slug, both transcription factors known as key regulators of EMT (Nieto et al., 1994; Savagner et al., 1997; Lee et al., 2006), was also increased in the mesenchymal-like cell population present at the periphery of the hESC colonies compared with the epithelial-like cell population. The isolation of the upper epithelium-like structure after a short trypsin digestion and its further feeder-free culture for 36 h demonstrated that the E-cadherin-positive cells gave rise to Vimentin-positive mesenchymal-like cells, adding further evidence that the epithelial-like cells underwent an EMT. On the contrary, all morphological results, immunostaining results and real-time RT–PCR results concerning hESC cultured on feeder layers confirmed that the hESC culture method on feeder layers did not induce such a differentiation process. A further characterization of the different cell populations present in hESC feeder-free cultures also confirmed an EMT process. A weak cytoplasmic Connexin 43 expression was found in the mesenchymal-like cell population probably because no functional gap junctions exist between these adjacent cells. The E-cadherin-positive upper layer cells and central layer cells of a hESC colony cultured under feeder-free conditions were coupled via gap junctions as suggested by the positive immunostaining for Connexin 43 detected at the membrane level, as previously described for hESC cultured on feeders (Wong et al., 2004). A switch in cadherin expression was also found with a strong positive N-cadherin pattern in the mesenchymal-like cells, whereas the epithelial-like cells were positive for E-cadherin. An EMT process is indeed characterized by typical changes in markers such as an E-cadherin down-regulation and a concomitant N-cadherin up-regulation inside the mesenchymal-like population (Maeda et al., 2005; Lee et al., 2006). Finally, β-catenin, another typical EMT marker, was found in the nuclei of the mesenchymal-like cell population (Lee et al., 2006). The pluripotent stem cell marker POU5F1_iA showed a positive nuclear expression in all the cells of the hESC colonies cultured under feeder-free conditions, but the intensity of the staining of the mesenchymal-like cells was slightly decreased in comparison with the rest of the hESC colonies. The immunostaining results were corroborated by the RT–PCR analysis showing a slight down-regulation of POU5F1_iA transcripts within the mesenchymal-like cell population, and this is in agreement with the literature about POU5F1. A decrease of this pluripotent hESC-specific marker is indeed generally observed upon differentiation (Matin et al., 2004; Gerrard et al., 2005). The positive POU5F1_iA and NANOG expressions in the spontaneously differentiated mesenchymal-like cells are probably related to the precocity of the event, which explains why these nuclear proteins may still be persistent. However, to determine the pluripotent capacity of the distinct hESC populations found in feeder-free culture, further experiments should be carried out such as in vitro or in vivo pluripotency tests. The loss of close contact with feeder layers closely surrounding the colonies and the new contact with matrigel matrix are factors that could play a role in enhancing this EMT phenomenon after switching hESC colonies to feeder-free conditions. Matrigel was indeed shown to promote spontaneous differentiation of stem cells (Philp et al., 2005; Ludwig et al., 2006). The authors suggest that extracellular matrices provide a different protein support and therefore other biological signals that can lead the cells into a defined differentiation pathway (Philp et al., 2005). In support of these observations, we observed that the appearance of the mesenchymal-like cells was not influenced by the use of conditioned medium of different origin. This minimizes the role of soluble factors present in the conditioned medium and points to the role of the extracellular matrix matrigel in enhancing the EMT process in feeder-free conditions. In conclusion, hESC colonies cultured in feeder-free conditions using matrigel and conditioned medium consist of epithelium-like structures with mesenchymal-like cells at the periphery. Our results suggested that hESC colonies spontaneously underwent an EMT process. The EMT process could deteriorate the quality of the culture and partly explain the lower efficiency of this type of hESC culture on long-term maintenance as compared with culture on feeder layers. Further studies are needed to provide optimal, stable long-term feeder- and animal component-free culture conditions avoiding karyotypic anomalies and hESC differentiation. This study was presented orally at the 22nd Annual Meeting of the European Society of Human Reproduction and Embryology (ESHRE) in Prague (18–21 June 2006) (abstract O-236). Acknowledgements We gratefully acknowledge the assistance of Ileana Mateizel for her critical corrections and Nele De Temmerman for maintaining the hESC lines in culture; Prof Miriam Marichal, Dr Ellen Degreef and Nicole Buelens from the Pathology Department for the immunohistological analysis of the lines; Montse Urbina from the Cytogenetics Laboratory for karyotyping and Prof Michèle Dramaix and Dr Olivier Vandenberg from the Epidemiological and Statistical Department, Public Health School of the Free University of Brussels for reviewing statistical analysis. The authors are also grateful to Walter Meul for computing pictures and to Lindsey Van Haute for her critical corrections concerning real-time analysis. K. Sermon is a postdoctoral fellow, and U. Ullmann is a research assistant at the Fund for Scientific Research, Flanders (FWO-Vlaanderen). Michael Whitburn of the Language Education Centre proofread the manuscript. References Alikani M ( 2005) Epithelial cadherin distribution in abnormal human pre-implantation embryos. Hum Reprod  20, 3369–3375. Alvi A, Clayton H, Joshi C, Enver T, Ashworth A, Vivanco M, Dale T and Smalley M ( 2003) Functional and molecular characterisation of mammary side population cells. Breast Cancer Res  5, R1–R8. Amit M, Margulets V, Segev H, Shariki K, Laevsky I, Coleman R and Itskovitz-Eldor J ( 2003) Human feeder layers for human embryonic stem cells. Biol Reprod  68, 2150–2156. Amit M, Shariki C, Margulets V and Itskovitz-Eldor J ( 2004) Feeder layer- and serum-free culture of human embryonic stem cells. Biol Reprod  70, 837–845. Barrallo-Gimeno A and Nieto MA ( 2005) The Snail genes as inducers of cell movement and survival: implications in development and cancer. Development  132, 3151–3161. Behr R, Heneweer C, Viebahn C, Denker HW and Thie M ( 2005) Epithelial-mesenchymal transition in colonies of rhesus monkey embryonic stem cells: a model for processes involved in gastrulation. Stem Cells  23, 805–816. Bloor DJ, Metcalfe AD, Rutherford A, Brison DR and Kimber SJ ( 2002) Expression of cell adhesion molecules during human preimplantation embryo development. Mol Hum Reprod  8, 237–245. Cai J, Olson JM, Rao MS, Stanley M, Taylor E and Ni H ( 2005) Development of antibodies to human embryonic stem cell antigens. BMC Dev Biol  5, 26. Carpenter MK, Rosler ES, Fisk GJ, Brandenberger R, Ares X, Miura T, Lucero M and Rao MS ( 2004) Properties of four human embryonic stem cell lines maintained in a feeder-free culture system. Dev Dyn  229, 243–258. Cauffman G, Liebaers I, Van Steirteghem A and Van de Velde H (in press) POU5F1 isoforms show different expression patterns in human embryonic stem cells and preimplantation embryos. Stem Cells . Cheng L, Hammond H, Ye Z, Zhan X and Dravid G ( 2003) Human adult marrow cells support prolonged expansion of human embryonic stem cells in culture. Stem Cells  21, 131–142. Ciruna B and Rossant J ( 2001) FGF signaling regulates mesoderm cell fate specification and morphogenetic movement at the primitive streak. Dev Cell  1, 37–49. D’Amour KA, Agulnick AD, Eliazer S, Kelly OG, Kroon E and Baetge EE ( 2005) Efficient differentiation of human embryonic stem cells to definitive endoderm. Nat Biotechnol  23, 1534–1541. Gerrard L, Zhao D, Clark AJ and Cui W ( 2005) Stably transfected human embryonic stem cell clones express OCT4-specific green feeder fluorescent protein and maintain self-renewal and pluripotency. Stem Cells  23, 124–133. Gilles C, Polette M, Zahm JM, Tournier JM, Volders L, Foidart JM and Birembaut P ( 1999) Vimentin contributes to human mammary epithelial cell migration. J Cell Sci  112, 4615–4625. Ginis I, Luo Y, Miura T, Thies S, Brandenberger R, Gerecht-Nir S, Amit M, Hoke A, Carpenter MK, Itskovitz-Eldor J et al. ( 2004) Differences between human and mouse embryonic stem cells. Dev Biol  269, 360–380. Grundemann C, Bauer M, Schweier O, von Oppen N, Lässing U, Saudan P, Becker K, Karp K, Hanke T, Bachmann M et al. ( 2006) Cutting edge: identification of E-cadherin as a ligand for the murine killer cell lectin-like receptor G1. J Immunol  176, 1311–1315. Guarino M ( 1995) Epithelial-to-mesenchymal change of differentiation. From embryogenetic mechanism to pathological patterns. Histol Histopathol  10, 171–184. Gumbiner BM ( 2005) Regulation of cadherin-mediated adhesion in morphogenesis. Nat Rev Mol Cell Biol  6, 622–634. Hajra KM, Chen DY and Fearon ER ( 2002) The SLUG zinc-finger protein represses E-cadherin in breast cancer. Cancer Res  62, 1613–1618. Hay ED ( 2005) The mesenchymal cell, its role in the embryo, and the remarkable signalling mechanisms that create it. Dev Dyn  233, 706–720. Hyslop L, Stojkovic M, Armstrong L, Walter T, Stojkovic P, Przyborski S, Herbert M, Murdoch A, Strachan T and Lako M ( 2005) Downregulation of NANOG induces differentiation of embryonic stem cells to extraembryonic lineages. Stem Cells  23, 1035–1043. Klimanskaya I, Chung Y, Meisner L, Johnson J, West MD and Lanza R ( 2005) Human embryonic stem cells derived without feeder cells. Lancet  365, 1636–1641. Lee JM, Dedhar S, Kallouri R and Thompson EW ( 2006) The epithelial-mesenchymal transition: new insights in signaling, development, and disease. J Cell Biol  172, 973–981. Ludwig TE, Levenstein ME, Jones JM, Berggen WT, Mitchen ER, Frane JL, Crandall LJ, Daigh CA, Conard KR, Piekarczyck MS et al. ( 2006) Derivation of human embryonic stem cells in defined conditions. Nat Biotechnol  24, 185–187. Maeda M, Johnson KR and Wheelock MJ ( 2005) Cadherin switching: essential for behavioural but not morphological changes during an epithelium-to-mesenchyme transition. J Cell Sci  118, 873–887. Mateizel I, De Temmerman N, Ullmann U, Cauffman G, Sermon K, Van de Velde H, De Rycke M, Degreef E, Devroey P, Liebaers I et al. ( 2006) Derivation of human embryonic stem cell lines from embryos obtained after IVF and after PGD for monogenic disorders. Hum Reprod  21, 503–511. Matin MM, Walsh JR, Gokhale PJ, Draper JS, Bahrami AR, Morton I, Moore HD and Andrews PW ( 2004) Specific knockdown of Oct4 and beta2-microglobulin expression by RNA interference in human embryonic stem cells and embryonic carcinoma cells. Stem Cells  22, 659–668. Muller T, Bain G, Wang X and Papkoff J ( 2002) Regulation of epithelial cell migration and tumor formation by beta-catenin signalling. Exp Cell Res  280, 119–133. Nieto MA, Sargent MG, Wilkinson DG and Cooke J ( 1994) Control of cell behavior during vertebrate development by Slug, a zinc finger gene. Science  264, 835–839. Philp D, Chen SS, Fitzgerald W, Orenstein J, Margolis L and Kleinman HK ( 2005) Complex extracellular matrices promote tissue-specific stem cell differentiation. Stem Cells  23, 288–296. Richards M, Fong CY, Chan WK, Wong PC and Bongso A ( 2002) Human feeders support prolonged undifferentiated growth of human inner cell masses and embryonic stem cells. Nat Biotechnol  20, 933–936. Rosler ES, Fisk GJ, Ares X, Irving J, Miura T, Rao MS and Carpenter MK ( 2004) Long-term culture of human embryonic stem cells in feeder-free conditions. Dev Dyn  229, 259–274. Sathananthan H, Pera M and Trounson A ( 2002) The fine structure of human embryonic stem cells. Reprod Biomed Online  4, 56–61. Savagner P, Yamada KM and Thiery JP ( 1997) The zinc-finger protein slug causes desmosome dissociation, an initial and necessary step for growth factor-induced epithelial-mesenchymal transition. J Cell Biol  137, 1403–1419. Stojkovic P, Lako M, Stewart R, Przyborski S, Armstrong L, Evans J, Murdoch A, Strachan T and Stojkovic M ( 2005) An autogeneic feeder cell system that efficiently supports growth of undifferentiated human embryonic stem cells. Stem Cells  23, 306–314. Thomson JA, Itskovitz-Eldor J, Shapiro SS, Waknitz MA, Swiergiel JJ, Marshall VS and Jones JM ( 1998) Embryonic stem cell lines derived from human blastocysts. Science  282, 1145–1147. Wong RC, Pebay A, Nguyen LT, Koh KL and Pera MF ( 2004) Presence of functional gap junctions in human embryonic stem cells. Stem Cells  22, 883–889. Xu C, Inokuma MS, Denham J, Golds K, Kundu P, Gold JD and Carpenter MK ( 2001) Feeder-free growth of undifferentiated human embryonic stem cells. Nat Biotechnol  19, 971–974. Xu RH, Peck RM, Li DS, Feng X, Ludwig T and Thomson JA ( 2005) Basic FGF and suppression of BMP signaling sustain undifferentiated proliferation of human ES cells. Nat Methods  2, 185–190. Author notes 1Research Centre Reproduction and Genetics, University Hospital and Medical School of the Vrije Universiteit Brussel (VUB, Free University of Brussels), 2Experimental Pathology Department, Medical School of the VUB, 3Laboratory of Developmental and Tumour Biology, University of Liège, 4Centre for Medical Genetics, University Hospital of the VUB and 5Centre for Reproductive Medicine, University Hospital of the VUB, Brussels, Belgium © The Author 2006. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Molecular Human Reproduction Oxford University Press

Epithelial–mesenchymal transition process in human embryonic stem cells cultured in feeder-free conditions

Loading next page...
 
/lp/oxford-university-press/epithelial-mesenchymal-transition-process-in-human-embryonic-stem-tcBRYg9Rdm

References (41)

Publisher
Oxford University Press
Copyright
© The Author 2006. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org
ISSN
1360-9947
eISSN
1460-2407
DOI
10.1093/molehr/gal091
pmid
17090644
Publisher site
See Article on Publisher Site

Abstract

Abstract Feeder-free human embryonic stem cell (hESC) culture is associated with the presence of mesenchymal-like cells appearing at the periphery of the colonies. The aim of this study was to identify this early differentiation process. Long-term feeder-free hESC cultures using matrigel and conditioned medium from mouse and from human origin revealed that the appearance of mesenchymal-like cells was similar regardless of the conditioned medium used. Standard characterization confirmed the preservation of hESC properties, but the feeder-free cultures could not be maintained longer than 37 passages. The early differentiation process was characterized in the short term after switching hESCs cultured on feeders to feeder-free conditions. Transmission electron microscopy showed an epithelium-like structure inside the hESC colonies, whereas the peripheral cells revealed the acquisition of a rather mesenchymal-like phenotype. Immunochemistry analysis showed that cells at the periphery of the colonies had a negative E-cadherin expression and a positive Vimentin expression, suggesting an epithelial–mesenchymal transition (EMT). Nuclear staining of β-catenin, positive N-cadherin and negative Connexin 43 expression were also found in the mesenchymal-like cell population. After RT–PCR analysis, Slug and Snail, both EMT-related transcription factors, were detected as up-regulated in the mesenchymal-like cell population. Taken together, our data suggest that culturing hESCs in feeder-free conditions enhances an early differentiation process identified as an EMT. human embryonic stem cells, feeder-free culture, matrigel, differentiation, epithelial–mesenchymal transition Introduction Human embryonic stem cell (hESC) lines present long-term self-renewal capacity and other characteristics of undifferentiated cells in specific culture conditions (Thomson et al., 1998). Because hESCs may be an unlimited source of material for future cell therapy due to their capacity to differentiate into any cell type, an important research topic is to characterize their early and spontaneous differentiation patterns in different culture conditions. At present, different methods of hESC culture exist. Most of them use mouse embryonic fibroblasts or human fibroblasts of different origin as feeder layers (Richards et al., 2002; Amit et al., 2003; Cheng et al., 2003). An alternative method is to culture hESCs on extracellular matrices, without any direct contact with feeder layers, in the presence of a conditioned medium obtained from fibroblast cultures (Xu et al., 2001). These latter culture conditions are very attractive with a view to the therapeutic use of hESCs. Long-term feeder-free cultures of hESCs have indeed been developed successfully to avoid animal cell contamination from, for example, fibroblasts to limit risks of cross-transfer of pathogens from xenogenic or allogenic feeders (Amit et al., 2004; Carpenter et al., 2004; Rosler et al., 2004). Important steps in the development of complete animal-free culture conditions were recently achieved with the derivation of new hESC lines in feeder-free conditions using a defined culture medium (Klimanskaya et al., 2005; Ludwig et al., 2006) and the report of an efficient autogenic feeder cell system using fibroblast-like cells spontaneously derived from the hESC colonies plated on matrigel and cultured in feeder-free conditions (Stojkovic et al., 2005). The hESC feeder-free culture system is known to induce spontaneous differentiation of hESCs (Xu et al., 2001; Rosler et al., 2004; Stojkovic et al., 2005; Ludwig et al., 2006). After plating hESC colonies on extracellular matrix-coated dishes in the presence of conditioned medium, fibroblast-like cells appear and surround the undifferentiated hESC colonies. So far, the nature of this spontaneous differentiation process leading to these fibroblast-like cells has not been studied. The aim of this study was to characterize the early differentiation process enhanced in feeder-free culture conditions. The first part of this study consisted of the maintenance of long-term hESC feeder-free cultures using matrigel matrix and conditioned medium of either mouse or human origin to evaluate the influence of the type of conditioned medium on the presence of the fibroblast-like cells and to determine whether the unique properties of the hESCs were preserved. In the second part of the work, the early differentiation process of three different hESC lines was characterized in the short term by morphological studies, immunostaining experiments and relative quantitative real-time RT–PCR analysis of specific transcripts. Materials and methods Culture of hESCs Three previously described hESC lines (VUB01, VUB03_DM1 carrying the mutation for myotonic dystrophy type 1 and VUB04_CF carrying a mutation in the CFTR gene) were at passage P53, P47 or P65 and P35, respectively, when used for our experiments (Mateizel et al., 2006). The three hESC lines had been maintained on mouse embryonic fibroblasts feeder layers with the hESC medium, which consisted of KnockOut Dulbecco’s modified Eagle’s medium (KO-DMEM) supplemented with 20% KnockOut Serum Replacement, 1 mM glutamine, 1% non-essential amino acids, 0.1 mM β-mercaptoethanol and 4 ng/ml human recombinant basic fibroblast growth factor (hbFGF), and passaged using a mechanical technique (Mateizel et al., 2006). For the long-term feeder-free culture studies, the VUB01 hESC line was switched to feeder-free culture conditions at passage 53. The hESCs were plated on 6-well plates (Nunc, Roskilde, Denmark) coated with a growth factor reduced-matrigel™ matrix (BD, Bedford, MA, USA; http://www.bdbiosciences.com, 1:30 dilution in KO-DMEM) and incubated at 37°C in 10% CO2. The hESC medium was conditioned on cultures of either mitomycin C inactivated mouse embryonic fibroblasts feeder layers, isolated from 14-days post-coitus fetuses (CF1 breed) and cultured until maximum passage 5, or with mitomycin C inactivated human fetal skin fibroblasts, isolated from skin biopsies of human aborted fetuses and maintained in culture for a maximum of 10 passages. For this purpose, 23.5 ml of hESC medium was left 24 h in a T-175 flask confluent with mouse or human fibroblasts. After collection, both types of conditioned medium were filtered and supplemented with 4 ng/ml hbFGF before feeding the hESCs. Passaging of the confluent hESC colonies grown in feeder-free conditions was carried out every 6 days after incubation with collagenase IV solution (1 mg/ml in KO-DMEM, Invitrogen) for 5–10 min. Phase contrast images of cell morphology were obtained using a Nikon Eclipse TE2005 and Eclipse Net software (Nikon instruments Belgium, Waver, Belgium). For the short-term culture studies, clumps of hESC colonies (VUB01, VUB03_DM1 and VUB04_CF) grown on feeder layers were obtained after cutting and careful collection with a Pasteur pipet to avoid contamination with mouse fibroblasts and then equally divided between hESC culture on mouse embryonic fibroblast feeder layers and feeder-free culture using matrigel™ with mouse embryonic fibroblast conditioned medium. Short culture periods were necessary to avoid confluent cell culture, which could interfere with the study of the early differentiation process. For real-time RT–PCR analysis, hESCs were cultured over a period from day 2 to day 4. For immunostaining experiments, the hESC colonies were plated on mouse embryonic fibroblasts- or on matrigel™-coated 4-well chamber slides (Sonic Seal Nalgene Nunc International, Rochester, NY, USA; http://www.nalgenunc.com) and allowed to expand until day 1 or day 3. For transmission electron microscopy experiments, the hESC colonies were plated on matrigel™-coated 4-well chamber slides and allowed to expand until day 3. Isolation of epithelium-like structures by partial trypsin digestion The isolation of the epithelium-like structures from the upper layers of hESC colonies cultured for 3 days in feeder-free conditions relied on short incubations in trypsin solution. E-cadherin-positive epithelial-like cells with strong specialized cell junctions have been shown to be protected from short trypsin digestion (Alvi et al., 2003; Grundemann et al., 2006). After 30 s of trypsin digestion (Invitrogen, 0.025% in phosphate-buffered saline [PBS]), the hESC colonies were disaggregated into a large fragment corresponding to an epithelium-like structure and single cells which were in suspension. The epithelium-like structure was collected by gentle pipetting, washed in hESC medium and allowed to adhere and grow in feeder-free conditions on matrigel™- coated 4-well chamber slides for 0, 12, 24 and 36 h. Standard characterization tests The karyotyping, the alkaline phosphatase activity test, the immunostaining experiments for SSEA-3, SSEA-4, TRA-1-60, TRA-1-81, the RT–PCR for NANOG, HPRT, REX-1 and SOX-2 genes, the in vitro and in vivo pluripotency tests on samples of hESCs cultured under long-term feeder-free conditions were performed as described previously (Mateizel et al., 2006). Light and transmission electron microscopy Sample preparation for microscopy studies was carried out as follows: hESCs were fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate, 1 mM CaCl2 (pH 7.4) and postfixed in 1% osmiumteroxide. The samples were embedded in Spurr’s resin, and for transmission electron microscopy, ultra-thin sections (90 nm) were prepared. After staining with lead citrate and uranyl acetate, the samples were examined in a Philips Tecnai 10 electron microscope (Eindhoven, The Netherlands). For light microscopy, semi-thin plastic sections (1 µm) were mounted on glass coverslips and were stained with 1% toluidine blue. Immunocytochemistry analysis To test for the presence of POU5F1_iA and of NANOG, the hESCs were fixed in 4% formaldehyde for 10 min and then permeabilized with 0.1% Triton X-100 for 10 min at room temperature. Following washing steps with PBS (3 × 5 min), incubation in 3% bovine serum albumin [(BSA) fraction V fatty acid-poor, nuclease and protease-free, Calbiochem, Darmstadt, Germany; http://www.calbiochem.com] was applied for 30 min to avoid non-specific reactions. Incubation with the mouse primary antibody for POU5F1_iA (IgG2b OCT-4 (C-10) sc-5279, Santa Cruz Biotechnology Inc., Heidelberg, Germany; http://www.scbt.com, 1:50 dilution in 1.5% BSA) and with the rabbit polyclonal antibody for NANOG (IgG, ab21624, Abcam, Cambridge, UK, http://www.abcam.com, 1:50 dilution in 1.5% BSA) was performed overnight at 4°C. The primary antibody for POU5F1_iA and for NANOG was detected with fluorescein-conjugated F(ab′)2 fragment of goat anti-mouse IgG and IgM (DakoCytomation, Glostrup, Denmark; 1:150 dilution in 1.5% BSA) and with Texas Red-conjugated goat anti-rabbit IgG (H+L) (Invitrogen, 1:100 dilution in 1.5% BSA), respectively. For the E-cadherin/Vimentin, E-cadherin/N-cadherin and E-cadherin/β-catenin immunostaining experiments, a fixation with cold methanol (at 4°C) was applied for 1 min. After blocking with 3% BSA, incubations with a mouse primary antibody for Vimentin (IgG1 clone V9, Sigma; Saint-Louis, USA; http://sigma-aldrich.com, 1:30 dilution in 1.5% BSA), with a mouse monoclonal antibody for N-cadherin (IgG1 clone 32, Becton Dickinson Transduction Laboratories, Lexington, KY, USA; http://www.bdbiosciences.com, 1:50 dilution in 1.5% BSA) and with a mouse monoclonal antibody for β-catenin (IgG3 clone 196618, R and D systems, Minneapolis, MN, USA; http://www.rndsystems.com, 1:100 dilution in 1.5% BSA) were carried out at room temperature for 1 h. The secondary antibody Texas Red-conjugated F(ab′)2 fragment of goat anti-mouse IgG (H+L) (Jackson ImmunoResearch Laboratories, West Grove, PA, USA; http://www.jacksonimmuno.com, 1:100 dilution in 1.5% BSA) was applied for 1 h after washing steps. A final incubation was performed with fluorescein-labelled E-cadherin antibody (mouse IgG2a, clone 36, Becton Dickinson Transduction Laboratories, 1:50 dilution in 1.5% BSA) for 1 h. For the E-cadherin/Connexin 43 immunostaining experiment, the same procedure was followed, but the detection of the primary antibody for Connexin 43 (polyclonal rabbit IgG, C6219, Sigma, 1:100 dilution in 1.5% BSA) was done with the secondary antibody Texas Red-conjugated goat anti-rabbit IgG (H+L) (Invitrogen, 1:100 dilution in 1.5% BSA). Negative controls were performed by omitting the primary antibodies or replacing them with a mouse immunoglobulin isotype used at the same dilution as the primary antibodies: HLA-DR IgG2b antibody (clone TÜ36, BD Biosciences) for POU5F1_iA, HLA-G IgG1 antibody (MEM-G/9, Exbio, Prague, Czech Republic; http://www.exbio.cz) for Vimentin and for N-cadherin, HLA-DR IgG3 (HL39, Exbio) for β-catenin, fluorescein-labelled IgG2a antibody (clone X39, BD Biosciences) for fluorescein isothiocyanate (FITC)-labelled E-cadherin and rabbit IgG (Invitrogen) for Connexin 43 and for NANOG served as negative controls. The images were scanned by confocal microscopy with an Argon–Krypton laser (488/568) (Fluoview IX70; Olympus, Belgium). RNA isolation and relative quantitative real-time RT–PCR A collagenase IV solution (1 mg/ml in KO-DMEM) was applied for 5–10 min to hESCs cultured under feeder-free conditions and for 1 h to hESCs cultured on feeder layers to ascertain that all the hESCs present in each dish were collected. For some experiments, the distinct cell populations of the hESC colonies within a feeder-free culture, that is, cells from the centre (epithelial-like) or from the periphery (mesenchymal-like), were collected separately using a Pasteur pipette. RNA extraction was carried out with the RNeasy kit (Qiagen, Hilden, Germany) and then treated with the RNase-free DNase kit (Qiagen). Five micrograms of total RNA was reverse-transcribed by using the first strand cDNA synthesis kit (Amersham Biosciences, Buckinghamshire, UK; http://www.amershambiosciences.com) with the NotI-d(T)18 primer, according to the manufacturer’s instructions. Relative quantitative real-time RT–PCR was performed on the ABI 7500 real-time PCR system (Applied Biosystems, Foster City, CA, USA; http://www.appliedbiosystems.com). The final reaction volume of 25 µl contained 12.5 µl of 2× TaqMan Universal Master Mix (Applied Biosystems), 1.25 µl of 20× Assays-on-demand Gene Expression assay mix (Applied Biosystems) and 10–100 ng complementary DNA (cDNA) in 11.25 µl nuclease-free water. The primers and the probes for the gene expression analysis of E-cadherin, Vimentin, Snail (Snail1), Slug (Snail2), POU5F1_iA (OCT-3A) and GAPDH were purchased from Applied Biosystems (Assays-on-demand gene expression products, Applera International Inc., Pleasanton, CA, USA). The following conditions were used: 2 min at 50°C, 10 min at 95°C, 40 cycles of 15 s at 95°C and 1 min at 60°C. All experiments were run in triplicate. Relative quantification of gene expression between multiple samples was achieved by normalization against the endogenous control GAPDH using the ΔΔCt method of quantification. Fold changes were calculated as 2−ΔΔCt. Statistical analysis Each experiment was repeated at least three times. The presented values are given as means ± SD. Differences between groups were determined by paired t-test. A P-value of less than 0.05 was considered significant. Results Characterization of the long-term hESC feeder-free culture The VUB01 hESC line that had originally been derived on mouse embryonic fibroblasts feeder layers was switched to feeder-free culture systems using matrigel with mouse conditioned medium or human conditioned medium to investigate the effect of the type of conditioned medium used on the early differentiation process and on the pluripotent capacities of hESCs. We did not find any difference regarding the standard characterization results between both types of conditioned medium (Table I). The hESCs showed a stable normal karyotype, even after long-term culture under feeder-free conditions. Colonies of hESCs grown in feeder-free conditions showed an alkaline phosphatase activity and a positive immunostaining for the cell surface markers SSEA-3, SSEA-4, TRA-1-60 and TRA-1-81 that were previously described as hESC specific (Thomson et al., 1998) (Figure 1A–D). As revealed by RT–PCR, VUB01 hESCs maintained under feeder-free culture conditions also expressed several pluripotent stem cell markers such as NANOG, SOX-2 and REX-1 (data not shown), whereas their pluripotent capacity was proven in vitro by the formation of embryoid bodies (data not shown) and in vivo by the ability to form a teratoma-containing cells from the three germ layers after intramuscular injection of hESCs in a severely compromised immunodeficient (SCID) mouse (Figure 1E–G). Table I. Summary of standard characterization analysis tested during long-term feeder-free culture of VUB01 at different passages (P) using matrigel and two different conditioned media: on mouse embryonic fibroblast conditioned medium (MEF CM) and on human fetal fibroblast conditioned medium (HFF CM) Long-term feeder-free growth of VUB01  MEF CM  HFF CM  Karyotypes          Early passage (P8)  46 XY  46 XY      Late passages (P27 and P35)  46 XY  46 XY  Alkaline phosphatase          Early passage (P8)  +  +      Late passage (P33)  +  +  Immunostaining of hESC-specific surface markers SSEA-3, SSEA-4 and TRA-1-60, TRA-1-81          Early passage (P10)  +  +      Late passage (P27)  +  +  RT–PCR pluripotency associated genes expressions          Early passage (P8)  +  +      Late passage (P27 and P35)  +  +  Embryoid bodies formation          Early passage (P10)  +  +      Late passage (P28)  +  +  Teratoma formation (P17)  +  +  Long-term feeder-free growth of VUB01  MEF CM  HFF CM  Karyotypes          Early passage (P8)  46 XY  46 XY      Late passages (P27 and P35)  46 XY  46 XY  Alkaline phosphatase          Early passage (P8)  +  +      Late passage (P33)  +  +  Immunostaining of hESC-specific surface markers SSEA-3, SSEA-4 and TRA-1-60, TRA-1-81          Early passage (P10)  +  +      Late passage (P27)  +  +  RT–PCR pluripotency associated genes expressions          Early passage (P8)  +  +      Late passage (P27 and P35)  +  +  Embryoid bodies formation          Early passage (P10)  +  +      Late passage (P28)  +  +  Teratoma formation (P17)  +  +  hESC, human embryonic stem cell. View Large Figure 1. View largeDownload slide Standard characterization tests of human embryonic stem cells (hESCs) cultured under long-term feeder-free conditions: expression of cell surface markers (VUB01 at passage 27): SSEA-3 (A), SSEA-4 (B), TRA-1-60 (C) and TRA-1-81 (D). In vivo differentiation (VUB01 at passage 17): immunohistochemistry analysis of teratoma sections with neurofilament 200K antibody shows neuronal tissue (ectoderm) (E), with actin antibody shows smooth muscle (mesoderm) (F) and with cytokeratin 18 antibody shows primitive epithelium (endoderm) (G). All positive-stained structures are marked with arrows. Original magnification: (A–G) ×100. Figure 1. View largeDownload slide Standard characterization tests of human embryonic stem cells (hESCs) cultured under long-term feeder-free conditions: expression of cell surface markers (VUB01 at passage 27): SSEA-3 (A), SSEA-4 (B), TRA-1-60 (C) and TRA-1-81 (D). In vivo differentiation (VUB01 at passage 17): immunohistochemistry analysis of teratoma sections with neurofilament 200K antibody shows neuronal tissue (ectoderm) (E), with actin antibody shows smooth muscle (mesoderm) (F) and with cytokeratin 18 antibody shows primitive epithelium (endoderm) (G). All positive-stained structures are marked with arrows. Original magnification: (A–G) ×100. During the long-term hESC feeder-free cultures, mesenchymal-like cells appeared at the periphery of the hESC colonies, and these cells were similarly observed regardless of the conditioned medium used (mouse or human origin). Around the 30th passage, a progressive enrichment of these cells within the two feeder-free culture systems was observed with the complete differentiation and the further loss of the hESC colonies around the 37th passage (9 months of culture). Morphology of the short-term hESC feeder-free culture The switching of the colonies of three different hESC lines (VUB01, VUB03_DM1 and VUB04_CF) from culture on mouse embryonic fibroblasts feeder layers to feeder-free conditions led to the appearance of a characteristic cell population with a flattened, mesenchymal-like phenotype at the periphery of the colonies (Figure 2A). These cells were observed under phase contrast light microscopy 24 h after switching culture conditions, and they further expanded around the hESC colonies until they reached confluency (Figure 2A–C). Such mesenchymal-like cells were rarely observed in parallel control cultures on feeder layers (Figure 2D–F). Figure 2. View largeDownload slide Morphology of human embryonic stem cell (hESC) colonies cultured under feeder-free conditions and on feeder layers (VUB03_DM1 at passage 47). Phase contrast images of hESC colonies grown under feeder-free conditions using matrigel and mouse embryonic fibroblast conditioned medium on day 2 (A), day 3 (B) and day 4 (C) and images of hESC colonies cultured on mouse embryonic fibroblast feeder layers on day 2 (D), day 3 (E) and day 4 (F). Arrows mark the periphery of the hESC colonies where mesenchymal-like cells are emerging. Original magnification: (A–F) ×100. Figure 2. View largeDownload slide Morphology of human embryonic stem cell (hESC) colonies cultured under feeder-free conditions and on feeder layers (VUB03_DM1 at passage 47). Phase contrast images of hESC colonies grown under feeder-free conditions using matrigel and mouse embryonic fibroblast conditioned medium on day 2 (A), day 3 (B) and day 4 (C) and images of hESC colonies cultured on mouse embryonic fibroblast feeder layers on day 2 (D), day 3 (E) and day 4 (F). Arrows mark the periphery of the hESC colonies where mesenchymal-like cells are emerging. Original magnification: (A–F) ×100. To analyse the morphology of these structures in more detail, we made semi-thin and ultra-thin sections of plastic-embedded feeder-free cultures of three hESC lines (VUB01, VUB03_DM1 and VUB04_CF). Analysis of sections revealed that the cell morphology changed depending on the location within the colony and that three different zones could be defined (Figure 3A–C). In the central part of the colonies, a single upper layer of columnar cells facing the medium (zone 1) covered a multilayered core of small rounded cells (zone 2), and at the periphery of the colonies, a monolayer of squamous cells was observed (zone 3). The columnar cells (zone 1) showed epithelial-like characteristics such as a clear polarization with a basal cell nucleus and apical microvilli (Figure 3A–C). At the ultrastructural level, the microvilli were associated with prominent bundles of microfilaments running parallel to the cell surface, and cell junctions were present at the apical–lateral side of the columnar cells (Figure 3D–F). The small rounded cells showing a high nuclear–cytoplasm ratio (zone 2) were closely apposed and had an electron-lucent cytoplasm with relatively few cell organelles but with prominent polyribosomes (Figure 3G). As previously described by other groups (Sathananthan et al., 2002; Ginis et al., 2004), such a multilayered structure with epithelial-like characteristics was also observed in the colonies formed by hESC cultured on feeder layers. Figure 3. View largeDownload slide Light microscopy images and transmission electron microscopy images of human embryonic stem cell (hESC) colonies grown under feeder-free conditions (VUB03_DM1 at passage 47). Graphical presentation of a transverse section through a hESC colony plated on matrigel showing three zones with different hESC populations (zones 1, 2 and 3) (A). At the junction (marked by bold curved line) between the colony (zone 1 and zone 2) and the periphery of the colony (zone 3), some mesenchymal-like cells still displayed cell junctions (A). A red line demarcates the upper layer of the colony showing epithelial-like columnar cells with an apico-basal polarization (zone 1) from the central multilayered part of the colony with rounded cells with a high nucleus to cytoplasm ratio (zone 2) (B). The border of the hESC colony showed the transition from central parts towards the periphery consisting of a monolayer of mesenchymal-like cells (zone 3) (C). At the ultrastructure level, columnar cells within the upper layer (zone1) show epithelial-like features (D) such as microvilli at the apical surface, specialized cell junctions at the apical sides (marked with double black arrows) (F) and prominent bundles of microfilaments parallel to the cell surface at the apical surface (E). Small rounded cells within the central parts of the hESC colony (zone 2) show an electron-lucent cytoplasm with few organelles (G). The mesenchymal-like cells at the periphery of the colony (zone 3) show a squamous morphology with a low nucleus–cytoplasm ratio and do not have epithelial-like features (H). Original magnification: (B–C) ×400, (D) ×3700, (E) ×12 500, (F) ×30 000, (G) ×10 500 and (H) ×3500. Figure 3. View largeDownload slide Light microscopy images and transmission electron microscopy images of human embryonic stem cell (hESC) colonies grown under feeder-free conditions (VUB03_DM1 at passage 47). Graphical presentation of a transverse section through a hESC colony plated on matrigel showing three zones with different hESC populations (zones 1, 2 and 3) (A). At the junction (marked by bold curved line) between the colony (zone 1 and zone 2) and the periphery of the colony (zone 3), some mesenchymal-like cells still displayed cell junctions (A). A red line demarcates the upper layer of the colony showing epithelial-like columnar cells with an apico-basal polarization (zone 1) from the central multilayered part of the colony with rounded cells with a high nucleus to cytoplasm ratio (zone 2) (B). The border of the hESC colony showed the transition from central parts towards the periphery consisting of a monolayer of mesenchymal-like cells (zone 3) (C). At the ultrastructure level, columnar cells within the upper layer (zone1) show epithelial-like features (D) such as microvilli at the apical surface, specialized cell junctions at the apical sides (marked with double black arrows) (F) and prominent bundles of microfilaments parallel to the cell surface at the apical surface (E). Small rounded cells within the central parts of the hESC colony (zone 2) show an electron-lucent cytoplasm with few organelles (G). The mesenchymal-like cells at the periphery of the colony (zone 3) show a squamous morphology with a low nucleus–cytoplasm ratio and do not have epithelial-like features (H). Original magnification: (B–C) ×400, (D) ×3700, (E) ×12 500, (F) ×30 000, (G) ×10 500 and (H) ×3500. Although the monolayer of squamous cells present at the periphery of the colonies, showing a flattened mesenchymal-like morphology (zone 3), appeared to be specific for the feeder-free culture conditions, these cells had a low nuclear–cytoplasm ratio and lacked regular cell–cell contacts, microvilli and apico-basal polarity (Figure 3H). At the junction between the colony and the periphery of the colonies, some mesenchymal-like cells still displayed cell junctions (Figure 3C). Characterization of the short-term hESC feeder-free culture by immunostaining In the light of our morphological observations, we performed immunostaining analysis for E-cadherin and Vimentin, chosen as specific markers for epithelial cells and for mesenchymal cells, respectively. E-cadherin is a cell–cell adhesion protein expressed in epithelial cells (Bloor et al., 2002; Lee et al., 2006). Vimentin is a type III intermediate filament normally expressed in mesenchymal cells but was also described in epithelial cells migrating during embryological, organogenesis or pathological processes (Guarino, 1995; Gilles et al., 1999; Lee et al., 2006). When hESCs (VUB01, VUB03_DM1 and VUB04_CF) were cultured on feeder layers, E-cadherin was strongly expressed as a web-like structure in the membrane of hESCs found in the upper layers of the colonies and largely disappeared in the lower layers of the colonies after scanning with a confocal microscope (Figure 4A). A similar distribution of E-cadherin was found 24 h after switching to feeder-free conditions, although the staining was clearly less intense (Figure 4D). After 3 days in feeder-free conditions, the E-cadherin web-like expression pattern was less structured, and the intensity of the staining had further decreased (Figure 4G). In contrast, at the periphery of the colonies, E-cadherin was barely detectable in the large population of mesenchymal-like cells (Figure 4I). Figure 4. View largeDownload slide E-cadherin/Vimentin immunostaining of human embryonic stem cell (hESC) colonies grown under feeder-free conditions, on feeder layers and of isolated epithelium-like structures. (VUB03_DM1 at passage 47): images of a hESC colony grown on mouse embryonic fibroblast feeder layers and fixed after 3 days of culture (A–C). Images of hESC colonies switched to feeder-free culture conditions and fixed after 24 h of culture (D–F) and after 3 days of culture (focus on the border of the colony) (G–I). (VUB03_DM1 at passage 65): images of isolated epithelium-like structures fixed and stained after 0, 12, 24 and 36 h (J–M). Images scanned by confocal microscopy after immunostaining with E-cadherin antibody (green) (A with an additional section through the middle of the hESC colony, D and G), with Vimentin antibody (red) (B, E and H) and merged E-cadherin and Vimentin immunostaining images (C, F, I, J–M). Original magnification: (A–C, G–M) ×200 and (D–F) ×400. Figure 4. View largeDownload slide E-cadherin/Vimentin immunostaining of human embryonic stem cell (hESC) colonies grown under feeder-free conditions, on feeder layers and of isolated epithelium-like structures. (VUB03_DM1 at passage 47): images of a hESC colony grown on mouse embryonic fibroblast feeder layers and fixed after 3 days of culture (A–C). Images of hESC colonies switched to feeder-free culture conditions and fixed after 24 h of culture (D–F) and after 3 days of culture (focus on the border of the colony) (G–I). (VUB03_DM1 at passage 65): images of isolated epithelium-like structures fixed and stained after 0, 12, 24 and 36 h (J–M). Images scanned by confocal microscopy after immunostaining with E-cadherin antibody (green) (A with an additional section through the middle of the hESC colony, D and G), with Vimentin antibody (red) (B, E and H) and merged E-cadherin and Vimentin immunostaining images (C, F, I, J–M). Original magnification: (A–C, G–M) ×200 and (D–F) ×400. Vimentin staining was restricted to the borders of the hESC colonies grown on feeder layers (Figure 4B and C). In agreement with our morphological observations, a large population of mesenchymal-like cells stained positive for Vimentin at the periphery of the hESC colonies, 24 h after switching to feeder-free conditions (Figure 4E and F). After 3 days in feeder-free culture, Vimentin-positive mesenchymal-like cells were found more abundantly at the periphery of the hESCs (Figure 4H and I). These changes in marker expression suggested that the upper layer of the colony could be considered as an epithelium undergoing an epithelial–mesenchymal transition (EMT) process after switching from the hESCs culture on feeder layers to feeder-free culture. To confirm that the mesenchymal-like cells are derived from E-cadherin-positive epithelial-like cells through an EMT process, we performed short-term cultures (0–36 h) of isolated epithelium-like structures from the upper layer of hESC colony cultured under feeder-free conditions (Figure 4J–M). Direct fixation and staining of the isolated fragment (t 0 h, Figure 4J) showed only E-cadherin-positive cells. In contrast, culturing this fragment under feeder-free conditions demonstrated that the E-cadherin-positive epithelial-like cells progressively underwent an EMT. Indeed, the samples stained at 12 and 24 h (Figure 4K and L) displayed a progressive decrease of E-cadherin-positive cells and an increase of Vimentin-positive mesenchymal-like cells. After 36 h of culture (Figure 4M), only Vimentin-positive mesenchymal-like cells were observed. Aiming at a better characterization of the phenotypes found in our hESC colonies cultured under feeder-free conditions, we next performed immunostaining analysis with the markers, Connexin 43, N-cadherin and β-catenin. Connexin 43 is a protein forming functional gap junctions that mediate signals from one cell to the adjacent cells and is known to be expressed inside hESC colonies cultured on feeder layers (Wong et al., 2004). N-cadherin is an integral cell-to-cell adhesion protein, and β-catenin is a cadherin-associated protein involved in the regulation of the membrane structure and of the cell adhesion. An E-cadherin down-regulation, a concomitant N-cadherin up-regulation inside the mesenchymal-like population and a β-catenin accumulation in the nucleus are recognized as typical changes in marker expressions during an EMT process (Muller et al., 2002; Maeda et al., 2005; Lee et al., 2006). After E-cadherin/Connexin 43 immunostaining analysis, the E-cadherin-positive upper layer cells and central layer cells of a hESC colony cultured under feeder-free conditions were positive for Connexin 43 detected at the membrane level (Figure 5A–C). On the contrary, the mesenchymal-like cells present at the periphery of the hESC colonies cultured under feeder-free conditions were showing a weaker Connexin 43 expression and an intracytoplasmic pattern of staining instead of a transmembrane localization (Figure 5B). Figure 5. View largeDownload slide E-cadherin/Connexin 43, E-cadherin/N-cadherin and E-cadherin/β-catenin immunostaining of human embryonic stem cell (hESC) colonies grown under feeder-free conditions (VUB03_DM1 at passage 47). Images of the border of hESC colonies switched to feeder-free culture conditions and fixed after 3 days of culture (A–I). Images scanned by confocal microscopy after immunostaining with E-cadherin antibody (green) (A, D, G), with Connexin 43 antibody (red) (B), N-cadherin antibody (red) (E) and β-catenin antibody (red) (H) and merged E-cadherin/Connexin 43 (C), E-cadherin/N-cadherin (F) and E-cadherin/β-catenin (I) immunostaining images. Arrows mark the periphery of the hESC colonies where mesenchymal-like cells are emerging. Double white arrows indicate some of the nuclei stained by β-catenin antibody. Original magnification: (A–I) ×200. Figure 5. View largeDownload slide E-cadherin/Connexin 43, E-cadherin/N-cadherin and E-cadherin/β-catenin immunostaining of human embryonic stem cell (hESC) colonies grown under feeder-free conditions (VUB03_DM1 at passage 47). Images of the border of hESC colonies switched to feeder-free culture conditions and fixed after 3 days of culture (A–I). Images scanned by confocal microscopy after immunostaining with E-cadherin antibody (green) (A, D, G), with Connexin 43 antibody (red) (B), N-cadherin antibody (red) (E) and β-catenin antibody (red) (H) and merged E-cadherin/Connexin 43 (C), E-cadherin/N-cadherin (F) and E-cadherin/β-catenin (I) immunostaining images. Arrows mark the periphery of the hESC colonies where mesenchymal-like cells are emerging. Double white arrows indicate some of the nuclei stained by β-catenin antibody. Original magnification: (A–I) ×200. After E-cadherin/N-cadherin immunostaining analysis, the E-cadherin-positive staining of the upper layer of epithelial-like cell population was found to be switched at the periphery of the hESC colonies cultured under feeder-free conditions to an N-cadherin web-like structure (Figure 5D–F). A strongly positive N-cadherin staining pattern was detected in the membranes of the mesenchymal-like cell population (Figure 5E). Finally, after E-cadherin/ β-catenin immunostaining analysis, a β‐catenin-positive staining was found at the membrane level in the E‐cadherin-positive epithelial-like cell population and in the central layer cells of the hESC colonies cultured under feeder-free conditions (Figure 5G–I), whereas a strong positive β-catenin staining was detected in the nuclei of the mesenchymal-like cell population present at the periphery (Figure 5I). Immunostaining analysis for stem cell markers of pluripotency POU5F1_iA (formerly called OCT-3A) and NANOG was carried out on hESC colonies (VUB01, VUB03_DM1 and VUB04_CF) switched to feeder-free conditions to see whether the early differentiation process influenced their expression (Figure 6) (Hyslop et al., 2005; Cauffman et al., in press). Expression of these nuclear proteins was found in the central parts of the hESC colonies as well as in the mesenchymal-like cells at the periphery (Figure 6A–F). Immunostaining for both POU5F1_iA and NANOG identified, within the mesenchymal-like cell population, larger, irregular nuclei with a decreased intensity (Figure 6B and E) compared with the small, rounded nuclei of the central parts of the hESC colonies (Figure 6A and D). Figure 6. View largeDownload slide POU5F1_iA and NANOG immunostaining of human embryonic stem cell (hESC) colonies grown under feeder-free conditions (VUB04_CF at passage 35). Images of central part (A, D), at the periphery (B, E) and at the transition zone between central parts to the periphery (C, F) of a hESC colony switched to feeder-free conditions and fixed after 3 days of culture. Images scanned by confocal microscopy after immunostaining with POU5F1_iA antibody (A–C) and NANOG antibody (D–F). Arrows mark the periphery of the hESC colonies where mesenchymal-like cells are emerging. Original magnification: (A, B, D and E) ×400, (C) ×100 and (F) ×200. Figure 6. View largeDownload slide POU5F1_iA and NANOG immunostaining of human embryonic stem cell (hESC) colonies grown under feeder-free conditions (VUB04_CF at passage 35). Images of central part (A, D), at the periphery (B, E) and at the transition zone between central parts to the periphery (C, F) of a hESC colony switched to feeder-free conditions and fixed after 3 days of culture. Images scanned by confocal microscopy after immunostaining with POU5F1_iA antibody (A–C) and NANOG antibody (D–F). Arrows mark the periphery of the hESC colonies where mesenchymal-like cells are emerging. Original magnification: (A, B, D and E) ×400, (C) ×100 and (F) ×200. Characterization of the short-term hESC feeder-free culture by relative quantitative real-time RT–PCR The mRNA levels of E-cadherin, Vimentin, Snail, Slug and POU5F1_iA were studied by relative quantitative real-time RT–PCR in samples from hESCs (VUB01, VUB03_DM1 and VUB04_CF) cultured on feeder layers or from hESCs switched to feeder-free culture (Figure 7A). Snail (Snail1) and Slug (Snail2) belong to the Snail genes family and are commonly used as EMT molecular markers. Both are zinc finger protein transcription factors shown to function as direct repressors of E-cadherin and inducers of the EMT process (Nieto et al., 1994; Savagner et al., 1997; Hajra et al., 2002; Lee et al., 2006). Figure 7. View largeDownload slide Relative quantification real-time RT–PCR result graphics. (A) Relative quantification of E-cadherin, Snail, Slug, Vimentin and POU5F1_iA mRNA levels in samples from human embryonic stem cells (hESCs) cultured on feeder layers and from hESCs cultured under feeder-free conditions. The samples were collected on days 2, 3 and 4. The samples from hESCs culture on feeder layers on day 2 were considered as a reference (value of 1), and data are represented as means ± SD. *, P < 0.05 compared with feeder layers at day 2. α, P < 0.05 compared with mRNA relative level of E-cadherin on MG at day 2. β, P < 0.05 compared with mRNA relative level of POU5F1_iA on MG at day 2. (B) Relative quantification of E-cadherin, Snail, Slug, Vimentin and POU5F1_iA mRNA levels of epithelial-like and mesenchymal-like human embryonic stem cell (hESC) populations located within the central parts and at the periphery of the hESC colonies, respectively, and both collected on day 3 of culture. The mRNA levels of each studied gene in the mesenchymal-like hESC population were compared with the epithelial-like hESCs population considered as reference (value of 1), and data are represented as means ± SD. *, P < 0.05 compared with the epithelial-like hESCs population. These experiments (A and B) were repeated at least three times with VUB03_DM1 at passage 47, and the analyses are shown here. These experiments were also carried out with VUB01 and VUB04_CF hESC lines and showed similar results (data not shown). Figure 7. View largeDownload slide Relative quantification real-time RT–PCR result graphics. (A) Relative quantification of E-cadherin, Snail, Slug, Vimentin and POU5F1_iA mRNA levels in samples from human embryonic stem cells (hESCs) cultured on feeder layers and from hESCs cultured under feeder-free conditions. The samples were collected on days 2, 3 and 4. The samples from hESCs culture on feeder layers on day 2 were considered as a reference (value of 1), and data are represented as means ± SD. *, P < 0.05 compared with feeder layers at day 2. α, P < 0.05 compared with mRNA relative level of E-cadherin on MG at day 2. β, P < 0.05 compared with mRNA relative level of POU5F1_iA on MG at day 2. (B) Relative quantification of E-cadherin, Snail, Slug, Vimentin and POU5F1_iA mRNA levels of epithelial-like and mesenchymal-like human embryonic stem cell (hESC) populations located within the central parts and at the periphery of the hESC colonies, respectively, and both collected on day 3 of culture. The mRNA levels of each studied gene in the mesenchymal-like hESC population were compared with the epithelial-like hESCs population considered as reference (value of 1), and data are represented as means ± SD. *, P < 0.05 compared with the epithelial-like hESCs population. These experiments (A and B) were repeated at least three times with VUB03_DM1 at passage 47, and the analyses are shown here. These experiments were also carried out with VUB01 and VUB04_CF hESC lines and showed similar results (data not shown). The mRNA levels of E-cadherin, Vimentin, Snail, Slug and POU5F1_iA in hESCs grown on feeder layers remained stable from day 2 until day 4 (Figure 7A). The expression levels of Snail, Slug and Vimentin in hESCs at the second day of feeder-free culture were up-regulated 2.9-fold, 7.8-fold and 10.4-fold, respectively, compared with those of hESCs grown on feeder layers (Figure 7A). Time course experiments until day 4 revealed a maintenance of the mRNA levels of Snail, Slug and Vimentin in hESCs cultured under feeder-free conditions, whereas the mRNA levels of E-cadherin significantly decreased from 1.8-fold at day 2 to 0.2-fold at day 4 (Figure 7A). Similarly, POU5F1_iA mRNA levels decreased from 0.9-fold at day 2 to 0.2-fold at day 4 (Figure 7A). To exclude a possible contamination with murine mRNA in the real-time RT–PCR experiments, samples with cDNA derived from mouse fibroblasts were included in the different assays. These samples were found to be positive for Vimentin only. Although contamination with murine mRNA is very unlikely because care was taken to collect hESCs only, contamination with murine mRNA in samples grown on feeder layers would lead to an underestimation of the Vimentin mRNA level in hESCs cultured under feeder-free conditions. To analyse E-cadherin, Snail, Slug, Vimentin and POU5F1_iA mRNA levels in samples enriched with mesenchymal-like and in samples enriched with epithelial-like cells cultured in feeder-free conditions, relative quantitative real-time RT–PCR analysis were carried out (Figure 7B). The results showed a 0.3-fold down-regulation of E-cadherin and a 0.6-fold down-regulation of POU5F1_iA concomitant with a 2.1-fold up-regulation of Snail, a 2.9-fold up-regulation of Slug and a 3.1-fold up-regulation of Vimentin in the mesenchymal-like cell population when compared with the epithelial-like cell population (Figure 7B). Discussion In the first part of this work, we showed that feeder-free culture using matrigel and conditioned medium of mouse or human origin could be an appropriate method for maintaining the self-renewal and pluripotency of hESCs, which were derived on feeder layers, whereas it is known that the use of non-conditioned hESC medium in feeder-free culture conditions gives a total differentiation of the hESC colonies after 48 h (Xu et al., 2005). We were unable to maintain the hESC feeder-free cultures for longer than 37 passages (9 months), whereas the maintenance of the VUB01 hESC line on mouse embryonic fibroblasts feeder layers is still ongoing after more than 100 passages (Mateizel et al., 2006). The majority of previously published results describe in vitro conditions for feeder-free hESC growth for less than 9 months (Carpenter et al., 2004; Rosler et al., 2004). We did not detect any chromosomal anomalies, but karyotypic instabilities were detected after feeder-free derivations and cultures of hESC lines in defined conditions (Ludwig et al., 2006). The difference in efficiency between the two methods in maintaining the hESC colonies in culture may be because of the difference in the passage technique. Clumps of hESCs cultured on feeder layers were passaged using a mechanical technique, which enabled the selection of ‘nice-looking’ hESCs, whereas an enzymatic treatment was necessary to passage the hESCs grown in feeder-free conditions, thereby excluding any selection of cells. Another explanation may be that feeder-free culture conditions force hESC colonies to undergo an early differentiation into mesenchymal-like cells. This phenomenon was previously mentioned in several manuscripts, but the nature of this process has not yet been identified (Xu et al., 2001; Carpenter et al., 2004; Rosler et al., 2004; Ludwig et al., 2006). The results of the second part of this study suggested that this differentiation process was associated with an EMT phenomenon. Indeed, the conversion of an epithelial cell to a mesenchymal cell requires the down-regulation of adhesion protein E-cadherin, the concurrent expression of Vimentin and the acquisition of migratory properties by the cells. These events mark the initiation of EMT (Lee et al., 2006). The EMT phenomenon occurs during different stages of human embryogenesis such as the migration of epiblast cells when they are allocated through the formation of the primitive streak to the three germ layers or the turning of an epithelial somite into a sclerotome mesenchyme (Nieto et al., 1994; Barrallo-Gimeno and Nieto, 2005; Hay, 2005). During these crucial steps, E-cadherin protein expression on the cell surface is decreased (Ciruna and Rossant, 2001; Gumbiner, 2005). Different signalling pathways induce several transcription factors such as Slug and Snail, which are direct repressors of E-cadherin (Ciruna and Rossant, 2001; Hajra et al., 2002; Lee et al., 2006). Coinciding with the loss of epithelial markers, cells undergoing EMT start to express mesenchymal markers and can migrate (Guarino, 1995; Lee et al., 2006). EMT also occurs in cancer cells and is implicated in the progression of primary tumours towards metastasis (Barrallo-Gimeno and Nieto, 2005; Lee et al., 2006). EMT is also seen during physiological situations such as wound healing or in pathological processes such as tumour invasion (Barrallo-Gimeno and Nieto, 2005; Lee et al., 2006) Taking all the data together, the observed morphological features and correlated changes in marker expression (recapitulated in Table II) suggest that the mesenchymal-like population emerging at the periphery of hESC colonies cultured under feeder-free conditions spontaneously replicated an EMT process (Lee et al., 2006). Supporting this analysis, EMT was recently described as being involved during directed differentiation of hESCs into endoderm (D’Amour et al., 2005). Moreover, a culture model system of rhesus monkey ESCs was also reported to replicate an EMT process (Behr et al., 2005). Table II. Summary of characterization analysis tested of three cell population: the upper layer cells (zone 1), the central multilayered part (zone 2) and the periphery (zone 3) of hESC colony in short-term feeder-free culture   Zone 1  Zone 2  Zone 3  Morphological studies  Epithelial-like cells  Undifferentiated rounded cells  Mesenchymal-like cells  Presence of cell-to-cell junctions  Specialized cell junctions and gap junctions  Gap junctions  No cell-to-cell junction  Immunostaining analysis of EMT markers            E-cadherin expression  Positive membrane  Negative  Negative      Vimentin expression  Negative  Negative  Positive      Connexin 43 expression  Positive membrane  Positive membrane  Negative cytoplasmic      N-cadherin expression  Negative  Negative  Positive membrane      β-catenin expression  Positive membrane  Positive membrane  Nuclear accumulation  Immunostaining analysis of pluripotency markers            POU5F1_iA  Positive  Positive  Positive but weaker      NANOG  Positive  Positive  Positive but weaker  RT–PCR analysis            Snail expression  Low mRNA levels  Low mRNA levels  mRNA levels up-regulated      Slug expression  Low mRNA levels  Low mRNA levels  mRNA levels up-regulated    Zone 1  Zone 2  Zone 3  Morphological studies  Epithelial-like cells  Undifferentiated rounded cells  Mesenchymal-like cells  Presence of cell-to-cell junctions  Specialized cell junctions and gap junctions  Gap junctions  No cell-to-cell junction  Immunostaining analysis of EMT markers            E-cadherin expression  Positive membrane  Negative  Negative      Vimentin expression  Negative  Negative  Positive      Connexin 43 expression  Positive membrane  Positive membrane  Negative cytoplasmic      N-cadherin expression  Negative  Negative  Positive membrane      β-catenin expression  Positive membrane  Positive membrane  Nuclear accumulation  Immunostaining analysis of pluripotency markers            POU5F1_iA  Positive  Positive  Positive but weaker      NANOG  Positive  Positive  Positive but weaker  RT–PCR analysis            Snail expression  Low mRNA levels  Low mRNA levels  mRNA levels up-regulated      Slug expression  Low mRNA levels  Low mRNA levels  mRNA levels up-regulated  EMT, epithelial–mesenchymal transition; hESC, human embryonic stem cell. View Large With the morphological studies, we observed that the hESC colonies grown in feeder-free conditions showed a central multilayered part of the hESC colonies with small rounded undifferentiated cells (zone 2) and an upper epithelium-like structure (zone 1) which had a more specialized phenotype, most probably because these cells are in direct contact with the culture medium. These epithelial-like features were shown to be lost inside the mesenchymal-like cell population present at the periphery of the hESC colonies (zone 3). For the immunostaining analyses, E-cadherin was chosen as an epithelial lineage-specific marker. Although E-cadherin is expressed in epithelial cells such as trophectoderm cells, inner cell mass cells and ectodermal cells (Alikani, 2005; D’Amour et al., 2005), some authors consider E-cadherin as a hESC-specific marker because it is found in undifferentiated hESC colonies but not in differentiated embryoid bodies (Cai et al., 2005). In this present study, a positive E-cadherin expression was shown in the upper layer of the hESC colonies (zone 1). In contrast, E-cadherin was barely detectable in the central multilayered part (zone 2) confirming the lack of epithelial features in this zone. The mesenchymal-like cell population emerging at the periphery of the hESC colonies (zone 3) did not show any E-cadherin expression, whereas a strong expression of mesenchymal lineage-specific Vimentin protein was detected within this last cell population. After switching to feeder-free culture, the changes of E-cadherin and Vimentin markers were confirmed at the mRNA level by real-time RT–PCR. The expression of Snail and Slug, both transcription factors known as key regulators of EMT (Nieto et al., 1994; Savagner et al., 1997; Lee et al., 2006), was also increased in the mesenchymal-like cell population present at the periphery of the hESC colonies compared with the epithelial-like cell population. The isolation of the upper epithelium-like structure after a short trypsin digestion and its further feeder-free culture for 36 h demonstrated that the E-cadherin-positive cells gave rise to Vimentin-positive mesenchymal-like cells, adding further evidence that the epithelial-like cells underwent an EMT. On the contrary, all morphological results, immunostaining results and real-time RT–PCR results concerning hESC cultured on feeder layers confirmed that the hESC culture method on feeder layers did not induce such a differentiation process. A further characterization of the different cell populations present in hESC feeder-free cultures also confirmed an EMT process. A weak cytoplasmic Connexin 43 expression was found in the mesenchymal-like cell population probably because no functional gap junctions exist between these adjacent cells. The E-cadherin-positive upper layer cells and central layer cells of a hESC colony cultured under feeder-free conditions were coupled via gap junctions as suggested by the positive immunostaining for Connexin 43 detected at the membrane level, as previously described for hESC cultured on feeders (Wong et al., 2004). A switch in cadherin expression was also found with a strong positive N-cadherin pattern in the mesenchymal-like cells, whereas the epithelial-like cells were positive for E-cadherin. An EMT process is indeed characterized by typical changes in markers such as an E-cadherin down-regulation and a concomitant N-cadherin up-regulation inside the mesenchymal-like population (Maeda et al., 2005; Lee et al., 2006). Finally, β-catenin, another typical EMT marker, was found in the nuclei of the mesenchymal-like cell population (Lee et al., 2006). The pluripotent stem cell marker POU5F1_iA showed a positive nuclear expression in all the cells of the hESC colonies cultured under feeder-free conditions, but the intensity of the staining of the mesenchymal-like cells was slightly decreased in comparison with the rest of the hESC colonies. The immunostaining results were corroborated by the RT–PCR analysis showing a slight down-regulation of POU5F1_iA transcripts within the mesenchymal-like cell population, and this is in agreement with the literature about POU5F1. A decrease of this pluripotent hESC-specific marker is indeed generally observed upon differentiation (Matin et al., 2004; Gerrard et al., 2005). The positive POU5F1_iA and NANOG expressions in the spontaneously differentiated mesenchymal-like cells are probably related to the precocity of the event, which explains why these nuclear proteins may still be persistent. However, to determine the pluripotent capacity of the distinct hESC populations found in feeder-free culture, further experiments should be carried out such as in vitro or in vivo pluripotency tests. The loss of close contact with feeder layers closely surrounding the colonies and the new contact with matrigel matrix are factors that could play a role in enhancing this EMT phenomenon after switching hESC colonies to feeder-free conditions. Matrigel was indeed shown to promote spontaneous differentiation of stem cells (Philp et al., 2005; Ludwig et al., 2006). The authors suggest that extracellular matrices provide a different protein support and therefore other biological signals that can lead the cells into a defined differentiation pathway (Philp et al., 2005). In support of these observations, we observed that the appearance of the mesenchymal-like cells was not influenced by the use of conditioned medium of different origin. This minimizes the role of soluble factors present in the conditioned medium and points to the role of the extracellular matrix matrigel in enhancing the EMT process in feeder-free conditions. In conclusion, hESC colonies cultured in feeder-free conditions using matrigel and conditioned medium consist of epithelium-like structures with mesenchymal-like cells at the periphery. Our results suggested that hESC colonies spontaneously underwent an EMT process. The EMT process could deteriorate the quality of the culture and partly explain the lower efficiency of this type of hESC culture on long-term maintenance as compared with culture on feeder layers. Further studies are needed to provide optimal, stable long-term feeder- and animal component-free culture conditions avoiding karyotypic anomalies and hESC differentiation. This study was presented orally at the 22nd Annual Meeting of the European Society of Human Reproduction and Embryology (ESHRE) in Prague (18–21 June 2006) (abstract O-236). Acknowledgements We gratefully acknowledge the assistance of Ileana Mateizel for her critical corrections and Nele De Temmerman for maintaining the hESC lines in culture; Prof Miriam Marichal, Dr Ellen Degreef and Nicole Buelens from the Pathology Department for the immunohistological analysis of the lines; Montse Urbina from the Cytogenetics Laboratory for karyotyping and Prof Michèle Dramaix and Dr Olivier Vandenberg from the Epidemiological and Statistical Department, Public Health School of the Free University of Brussels for reviewing statistical analysis. The authors are also grateful to Walter Meul for computing pictures and to Lindsey Van Haute for her critical corrections concerning real-time analysis. K. Sermon is a postdoctoral fellow, and U. Ullmann is a research assistant at the Fund for Scientific Research, Flanders (FWO-Vlaanderen). Michael Whitburn of the Language Education Centre proofread the manuscript. References Alikani M ( 2005) Epithelial cadherin distribution in abnormal human pre-implantation embryos. Hum Reprod  20, 3369–3375. Alvi A, Clayton H, Joshi C, Enver T, Ashworth A, Vivanco M, Dale T and Smalley M ( 2003) Functional and molecular characterisation of mammary side population cells. Breast Cancer Res  5, R1–R8. Amit M, Margulets V, Segev H, Shariki K, Laevsky I, Coleman R and Itskovitz-Eldor J ( 2003) Human feeder layers for human embryonic stem cells. Biol Reprod  68, 2150–2156. Amit M, Shariki C, Margulets V and Itskovitz-Eldor J ( 2004) Feeder layer- and serum-free culture of human embryonic stem cells. Biol Reprod  70, 837–845. Barrallo-Gimeno A and Nieto MA ( 2005) The Snail genes as inducers of cell movement and survival: implications in development and cancer. Development  132, 3151–3161. Behr R, Heneweer C, Viebahn C, Denker HW and Thie M ( 2005) Epithelial-mesenchymal transition in colonies of rhesus monkey embryonic stem cells: a model for processes involved in gastrulation. Stem Cells  23, 805–816. Bloor DJ, Metcalfe AD, Rutherford A, Brison DR and Kimber SJ ( 2002) Expression of cell adhesion molecules during human preimplantation embryo development. Mol Hum Reprod  8, 237–245. Cai J, Olson JM, Rao MS, Stanley M, Taylor E and Ni H ( 2005) Development of antibodies to human embryonic stem cell antigens. BMC Dev Biol  5, 26. Carpenter MK, Rosler ES, Fisk GJ, Brandenberger R, Ares X, Miura T, Lucero M and Rao MS ( 2004) Properties of four human embryonic stem cell lines maintained in a feeder-free culture system. Dev Dyn  229, 243–258. Cauffman G, Liebaers I, Van Steirteghem A and Van de Velde H (in press) POU5F1 isoforms show different expression patterns in human embryonic stem cells and preimplantation embryos. Stem Cells . Cheng L, Hammond H, Ye Z, Zhan X and Dravid G ( 2003) Human adult marrow cells support prolonged expansion of human embryonic stem cells in culture. Stem Cells  21, 131–142. Ciruna B and Rossant J ( 2001) FGF signaling regulates mesoderm cell fate specification and morphogenetic movement at the primitive streak. Dev Cell  1, 37–49. D’Amour KA, Agulnick AD, Eliazer S, Kelly OG, Kroon E and Baetge EE ( 2005) Efficient differentiation of human embryonic stem cells to definitive endoderm. Nat Biotechnol  23, 1534–1541. Gerrard L, Zhao D, Clark AJ and Cui W ( 2005) Stably transfected human embryonic stem cell clones express OCT4-specific green feeder fluorescent protein and maintain self-renewal and pluripotency. Stem Cells  23, 124–133. Gilles C, Polette M, Zahm JM, Tournier JM, Volders L, Foidart JM and Birembaut P ( 1999) Vimentin contributes to human mammary epithelial cell migration. J Cell Sci  112, 4615–4625. Ginis I, Luo Y, Miura T, Thies S, Brandenberger R, Gerecht-Nir S, Amit M, Hoke A, Carpenter MK, Itskovitz-Eldor J et al. ( 2004) Differences between human and mouse embryonic stem cells. Dev Biol  269, 360–380. Grundemann C, Bauer M, Schweier O, von Oppen N, Lässing U, Saudan P, Becker K, Karp K, Hanke T, Bachmann M et al. ( 2006) Cutting edge: identification of E-cadherin as a ligand for the murine killer cell lectin-like receptor G1. J Immunol  176, 1311–1315. Guarino M ( 1995) Epithelial-to-mesenchymal change of differentiation. From embryogenetic mechanism to pathological patterns. Histol Histopathol  10, 171–184. Gumbiner BM ( 2005) Regulation of cadherin-mediated adhesion in morphogenesis. Nat Rev Mol Cell Biol  6, 622–634. Hajra KM, Chen DY and Fearon ER ( 2002) The SLUG zinc-finger protein represses E-cadherin in breast cancer. Cancer Res  62, 1613–1618. Hay ED ( 2005) The mesenchymal cell, its role in the embryo, and the remarkable signalling mechanisms that create it. Dev Dyn  233, 706–720. Hyslop L, Stojkovic M, Armstrong L, Walter T, Stojkovic P, Przyborski S, Herbert M, Murdoch A, Strachan T and Lako M ( 2005) Downregulation of NANOG induces differentiation of embryonic stem cells to extraembryonic lineages. Stem Cells  23, 1035–1043. Klimanskaya I, Chung Y, Meisner L, Johnson J, West MD and Lanza R ( 2005) Human embryonic stem cells derived without feeder cells. Lancet  365, 1636–1641. Lee JM, Dedhar S, Kallouri R and Thompson EW ( 2006) The epithelial-mesenchymal transition: new insights in signaling, development, and disease. J Cell Biol  172, 973–981. Ludwig TE, Levenstein ME, Jones JM, Berggen WT, Mitchen ER, Frane JL, Crandall LJ, Daigh CA, Conard KR, Piekarczyck MS et al. ( 2006) Derivation of human embryonic stem cells in defined conditions. Nat Biotechnol  24, 185–187. Maeda M, Johnson KR and Wheelock MJ ( 2005) Cadherin switching: essential for behavioural but not morphological changes during an epithelium-to-mesenchyme transition. J Cell Sci  118, 873–887. Mateizel I, De Temmerman N, Ullmann U, Cauffman G, Sermon K, Van de Velde H, De Rycke M, Degreef E, Devroey P, Liebaers I et al. ( 2006) Derivation of human embryonic stem cell lines from embryos obtained after IVF and after PGD for monogenic disorders. Hum Reprod  21, 503–511. Matin MM, Walsh JR, Gokhale PJ, Draper JS, Bahrami AR, Morton I, Moore HD and Andrews PW ( 2004) Specific knockdown of Oct4 and beta2-microglobulin expression by RNA interference in human embryonic stem cells and embryonic carcinoma cells. Stem Cells  22, 659–668. Muller T, Bain G, Wang X and Papkoff J ( 2002) Regulation of epithelial cell migration and tumor formation by beta-catenin signalling. Exp Cell Res  280, 119–133. Nieto MA, Sargent MG, Wilkinson DG and Cooke J ( 1994) Control of cell behavior during vertebrate development by Slug, a zinc finger gene. Science  264, 835–839. Philp D, Chen SS, Fitzgerald W, Orenstein J, Margolis L and Kleinman HK ( 2005) Complex extracellular matrices promote tissue-specific stem cell differentiation. Stem Cells  23, 288–296. Richards M, Fong CY, Chan WK, Wong PC and Bongso A ( 2002) Human feeders support prolonged undifferentiated growth of human inner cell masses and embryonic stem cells. Nat Biotechnol  20, 933–936. Rosler ES, Fisk GJ, Ares X, Irving J, Miura T, Rao MS and Carpenter MK ( 2004) Long-term culture of human embryonic stem cells in feeder-free conditions. Dev Dyn  229, 259–274. Sathananthan H, Pera M and Trounson A ( 2002) The fine structure of human embryonic stem cells. Reprod Biomed Online  4, 56–61. Savagner P, Yamada KM and Thiery JP ( 1997) The zinc-finger protein slug causes desmosome dissociation, an initial and necessary step for growth factor-induced epithelial-mesenchymal transition. J Cell Biol  137, 1403–1419. Stojkovic P, Lako M, Stewart R, Przyborski S, Armstrong L, Evans J, Murdoch A, Strachan T and Stojkovic M ( 2005) An autogeneic feeder cell system that efficiently supports growth of undifferentiated human embryonic stem cells. Stem Cells  23, 306–314. Thomson JA, Itskovitz-Eldor J, Shapiro SS, Waknitz MA, Swiergiel JJ, Marshall VS and Jones JM ( 1998) Embryonic stem cell lines derived from human blastocysts. Science  282, 1145–1147. Wong RC, Pebay A, Nguyen LT, Koh KL and Pera MF ( 2004) Presence of functional gap junctions in human embryonic stem cells. Stem Cells  22, 883–889. Xu C, Inokuma MS, Denham J, Golds K, Kundu P, Gold JD and Carpenter MK ( 2001) Feeder-free growth of undifferentiated human embryonic stem cells. Nat Biotechnol  19, 971–974. Xu RH, Peck RM, Li DS, Feng X, Ludwig T and Thomson JA ( 2005) Basic FGF and suppression of BMP signaling sustain undifferentiated proliferation of human ES cells. Nat Methods  2, 185–190. Author notes 1Research Centre Reproduction and Genetics, University Hospital and Medical School of the Vrije Universiteit Brussel (VUB, Free University of Brussels), 2Experimental Pathology Department, Medical School of the VUB, 3Laboratory of Developmental and Tumour Biology, University of Liège, 4Centre for Medical Genetics, University Hospital of the VUB and 5Centre for Reproductive Medicine, University Hospital of the VUB, Brussels, Belgium © The Author 2006. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Journal

Molecular Human ReproductionOxford University Press

Published: Nov 7, 2006

There are no references for this article.