Author contributions: K.H.: conception and design, collection and assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript; L.M.: manuscript writing, provision of study material, and final approval of manuscript; S.R.: conception and design and manuscript writing; C.M.W.: conception and design, financial support, provision of study material, collection and assembly of data, data analysis and interpretation, manuscript writing, and final approval of manuscript; C.L.R.M.: conception and design and manuscript writing.
Disclosure of potential conflicts of interest is found at the end of this article.
First published online in STEM CELLSEXPRESS June 13, 2012.
We have recently shown that loss of E-cadherin in mouse embryonic stem cells (mESCs) results in significant alterations to both the transcriptome and hierarchy of pluripotency-associated signaling pathways. Here, we show that E-cadherin promotes kruppel-like factor 4 (Klf4) and Nanog transcript and protein expression in mESCs via STAT3 phosphorylation and that β-catenin, and its binding region in E-cadherin, is required for this function. To further investigate the role of E-cadherin in leukemia inhibitory factor (LIF)-dependent pluripotency, E-cadherin null (Ecad−/−) mESCs were cultured in LIF/bone morphogenetic protein supplemented medium. Under these conditions, Ecad−/− mESCs exhibited partial restoration of cell–cell contact and STAT3 phosphorylation and upregulated Klf4, Nanog, and N-cadherin transcripts and protein. Abrogation of N-cadherin using an inhibitory peptide caused loss of phospho STAT3, Klf4, and Nanog in these cells, demonstrating that N-cadherin supports LIF-dependent pluripotency in this context. We therefore identify a novel molecular mechanism linking E- and N-cadherin to the core circuitry of pluripotency in mESCs. This mechanism may explain the recently documented role of E-cadherin in efficient induced pluripotent stem cell reprogramming. Stem Cells2012;30:1842–1851
Understanding the molecular mechanisms that govern pluripotency in embryonic stem cells (ESCs) is key to the derivation of novel stem cell types (e.g., induced pluripotent stem [iPS] cells), the use of these cells as models of disease , and the application of such cells in clinical therapies . Pluripotency can be maintained in vitro in mouse ESCs (mESCs) by leukemia inhibitory factor (LIF) and bone morphogenetic proteins  (BMPs), whereas human ESCs (hESCs) and human iPS (hiPS) cells require Activin/Nodal and Fibroblast Growth Factor 2 to maintain this state [4, 5]. Despite the variation in exogenous cues supporting pluripotency between the two species, the core circuitry of pluripotency, Oct4, Sox2, and Nanog, is conserved. Recent examination of the relationship of LIF to this core circuitry in mESCs has revealed two parallel-functioning pathways (Jak/STAT3 and PI3K/Akt) that maintain Oct4, Sox2, and Nanog expression via Kruppel-like factor 4 (Klf4) and T-box factor 3. LIF has also been shown to stimulate the MAPK/ERK pathway in these cells . Critically, artificial activation of STAT3 has been shown to maintain self-renewal of mESCs in the absence of LIF, demonstrating the importance of this protein in mESC pluripotency .
In recent years, E-cadherin has emerged as a key factor in the maintenance of pluripotency. For example, E-cadherin has been implicated in the regulation of signaling pathway hierarchy in mouse stem cells [8, 9], enhanced iPS cell generation [10–12], and in promoting pluripotent hESC expansion when used as a substratum . Moreover, we have recently shown that E-cadherin null (Ecad−/−) mESCs exhibit alterations in more than 2,000 gene transcripts when compared to wild-type (wt) mESCs and that these changes are associated with a wide range of cellular processes . E-cadherin is a well-characterized member of the classical cadherin family. Structurally, it is a single-pass transmembrane glycoprotein with a HAV motif within its extracellular region and a β-catenin binding domain within its cytoplasmic region, the latter facilitating interaction with the actin cytoskeleton via α-catenin . Ecad−/− embryos fail to form a trophectodermal epithelium or to undergo compaction due to loss of cell–cell contact , thus demonstrating the importance of this adhesion molecule for embryogenesis. mESCs derived from Ecad−/− embryos or β-catenin null (βcat−/−) embryos exhibit a single-celled mesenchymal phenotype  and maintain pluripotency marker expression independently of LIF, instead using the Activin/Nodal pathway. As such, Ecad−/− and βcat−/− exhibit more similarity to hESCs, hiPS cells, FAB-SCs , and EpiSCs [17, 18] than wt mESCs.
N-cadherin is also a member of the classic cadherin family and, as such, exhibits a high degree of structural homogeneity to E-cadherin. N-cadherin possesses a histidine-alanine-valine (HAV) motif within its extracellular region and has been shown to form a complex with β-catenin . However, unlike E-cadherin, N-cadherin is not expressed in ESCs, instead it is rapidly upregulated in both hESCs and mESCs upon induction of differentiation in a process akin to an epithelial-to-mesenchymal transition (EMT) event [20, 21]. Forced expression of N-cadherin in Ecad−/− mESCs has been shown to rescue cell–cell contact and increase the frequency of chimera generation . To date, however, there is no evidence that N-cadherin can compensate for the role of E-cadherin in maintaining LIF-dependent mESC pluripotency.
In this study, we have investigated the function of E-cadherin in LIF-dependent pluripotency. We show that E-cadherin positively regulates Klf4 and Nanog transcript and protein expression via STAT3 phosphorylation in mESCs and that its β-catenin binding domain is critical for this function. Furthermore, culture of Ecad−/− mESCs in LIF/BMP-supplemented medium in the absence of Activin/Nodal leads to restoration of cell–cell contact, STAT3 phosphorylation, and enhancement of Klf4 and Nanog transcript and protein expression. This is likely to be via an N-cadherin-dependent mechanism in Ecad−/− mESCs and an E-cadherin-dependent mechanism in wt mESCs since abrogation of these proteins in the respective cells when cultured in LIF/BMP medium results in loss of phospho STAT3 (pSTAT3) and decreased Klf4 and Nanog expression. Therefore, either E-cadherin or N-cadherin can facilitate LIF-dependent pluripotency in mESCs and in their absence no other factor acts to fulfil this function.
MATERIALS AND METHODS
Culture of mESCs
All cell lines used in this study have been described previously . Unless otherwise specified, mESCs were cultured on 0.1% gelatin (Sigma)-coated plates in knockout Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (FBS), 50 μm 2-mercaptoethanol (all Invitrogen, Carlsbad, CA, http://www.invitrogen.com), 2 mM L-glutamine, 1% nonessential amino acids (both PAA, Linz, Austria, http://www.paa.at) and 1,000 U/ml LIF (Chemicon International, Temecula, CA, http://www.chemicon.com) at 37°C, 5% CO2. Cells were passaged every 2 days, when a maximum confluence of 70%–80% was reached. Cells were acclimatized to culture in ESGRO Complete Plus Clonal Grade Medium (LIF/BMP medium; Millipore, Billerica, MA, http://www.millipore.com), or 2i medium (iSTEM Embryonic Stem Cell Culture Medium, Stem Cells Inc., http://www.stemcellsinc.com/) supplemented with 1,000 U/ml LIF, by passage in increasing concentrations of these media within FBS-containing medium (Supporting Information Table S1). At days 0–4, cells were detatched from plates using Trypsin/EDTA (PAA), thereafter cell dissociation buffer (Gibco, Grand Island, NY, http://www.invitrogen.com) was used. Appropriate cells were treated with 1 mM E- and N-cadherin peptide inhibitor  (E/N+) or a control peptide  (E/N−), 10 μM activin receptor-like kinase inhibitor (SB431542; Tocris) or Jak inhibitor I (Calbiochem, San Diego, http://www.emdbiosciences.com) for 8 days. Controls were treated with DMSO.
Amaxa Transfection of mESCs
After 24 hours of passaging, 2 × 106 cells were transfected as previously described  with FL-Ecad, Δβcat, Δβcat-Δp120ctn or empty control plasmids or Nanog promoter regions within the pGL3 basic plasmid.
Colony Forming Assays
Cells were trypsinized, and serial dilutions were plated in a 96-well plate in the appropriate media. Wells containing single cell were identified, and when 70%–80% confluence was reached, cells from these wells were passaged into 24-, 12- and 6-well plates for analysis.
Fluorescent Flow Cytometry
Cells were harvested using cell dissociation buffer, centrifuged, and washed with PBS before being resuspended in 100 μl DECMA-1 (1:100 dilution; Sigma) in fluorescence-activated cell sorting (FACS) buffer (0.2% bovine serum albumin, 0.1% sodium azide in PBS) for 30 minutes on ice. Cells were then washed with PBS, resuspended in phycoerythrin-conjugated secondary antibody (1:100 dilution; Santa Cruz) in FACS buffer, and incubated for 30 minutes on ice. Cells were then washed and fixed in 1% w/v paraformaldehyde (PFA) in PBS. Cell fluorescence was then quantified using a Becton Dickinson FACS Calibur. Viable cells were gated using forward and side scatter, and all data represents cells from this population. For intracellular flow cytometry, cells were fixed in 1% PFA for 10 minutes at room temperature before the membrane was permeablized by incubating cells in ice-cold 70% methanol at −20°C overnight. Cells were then washed with PBS and stained according to the above protocol using an Oct4, Nanog (Abcam, Cambridge, U.K., http://www.abcam.com), or Klf4 antibody (R&D Systems, Minneapolis, http://www.rndsystems.com, all 1:100).
Immunofluorescent Cell Imaging
Cells were fixed in 4% w/v PFA in PBS for 15 minutes before being prepared as previously described  using the following primary antibodies: N-cadherin, (1:200 dilution; Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com) Klf4, Nanog, and Oct4 (all 1:200; Abcam) followed by the appropriate Alexa-Fluor 488-conjugated secondary antibody (1:500; Invitrogen). Images were visualized using a fluorescence microscope and processed using Adobe Photoshop version 6.0.
Reverse Transcription Polymerase Chain Reaction
RNA was harvested using TRI reagent (Sigma), treated with DNase (Promega, Madison, WI, http://www.promega.com) and phenol/chloroform purified. cDNA was synthesized as previously described . Products were amplified for 25 cycles at the optimal primer annealing temperature (Supporting Information Table S2).
Quantitative Real-Time Polymerase Chain Reaction
Triplicate samples containing appropriate forward and reverse primers (100 pmol/μl; Supporting Information Table 2) were analyzed as previously described  with 18S ribosomal RNA as the endogenous control. All results show the SEM of three independent replicates.
Western Blot Analysis
Cells were trypsinized and lysed using RIPA buffer (Sigma) containing protease inhibitor cocktail tablets (Roche Applied Science, Basel, Switzerland, http://www.roche-applied-science.com). The lysate from 4 × 105 cells was incubated on ice for 30 minutes and boiled in reducing conditions. Proteins were then separated using SDS-PAGE and electrotransferred onto Hybond-enchanced chemiluminescence (ECL) nitrocellulose membrane (Amersham Biosciences, Piscataway, NJ, http://www.amersham.com). The membrane was incubated in blocking buffer (5% dry milk in TBS [10 mM TRIS, 15 mM NaCl in PBS]) with the exception of the phospho-ERK1/2 (pERK1/2; Thr202/Tyr204) antibody, which was blocked in 5% BSA in TBS at room temperature for 30 minutes. The membrane was then incubated in primary antibody (α-tubulin [Sigma; 1:2,000], KLF4 [R&D systems], STAT3 [Santa Cruz], pSTAT3; Tyr705, phospho-Akt [pAkt; Ser473], Akt, pERK1/2, ERK1/2 [all Cell Signaling Technology, Beverly, MA, http://www.cellsignal.com], N-cadherin and Nanog [Abcam], all 1:200) in blocking buffer at 4°C overnight. The membrane was washed three times with TBST (TBS + 10% Tween) before being incubated in the appropriate HRP-conjugated secondary antibody (Dako, Glostrup, Denmark, http://www.dako.com; all 1:2,000) in blocking buffer for 1 hour at room temperature. The membrane was then washed as previously described and developed using ECL (Amersham Biosciences). For phospho-Westerns, the medium was replenished for two 30-minute periods prior to sample collection.
Nanog Promoter Analysis
Regions upstream of the Nanog transcriptional start site (530, 1,087, 3,832, 4,432, 4,547 [minus STAT3], and 4,963 bp) were amplified by polymerase chain reaction (PCR) and ligated into the pGL3 basic plasmid (Promega). Vectors were amplified in DH5α competent cells (Invitrogen) and sequenced prior to use. Vectors were transfected into mESCs using an Amaxa Nucleofactor II system (Amaxa Biosystems; program A30). After 24 hours of transfection, cells were lysed using a Dual Luciferase Reporter Assay System (Promega) according to the manufacturer's instructions. Lysates were analyzed for luciferase activity using an Orion L Microplate Luminometer (Berthold Detection Systems). Luciferase activity was normalized using a vector containing the Renilla gene, and mean values were plotted graphically. All results reflect the SEM of three independent replicates.
Klf4 and Nanog Are Positively Regulated by E-cadherin
The current understanding of Nanog regulation via LIF signaling in mESCs is shown in Figure 1A. To explore the role of E-cadherin in this transduction network, Ecad−/− mESCs and the parental cell line wtD3 mESCs were analyzed along with MESC20 mESCs exhibiting RNA interference of E-cadherin (EcadRNAi mESCs) and a clone of these cells in which E-cadherin inhibition has been reversed (EcadRNAiR mESCs). Ecad−/− and EcadRNAi mESCs lack cell surface E-cadherin as evidenced by flow cytometry analysis (Fig. 1B). Reverse transcription PCR (RT-PCR) analysis of these cells cultured in FBS-containing medium supplemented with LIF (FBS-containing medium) confirmed the expression of Oct4, Nanog and Klf4 transcripts (Fig. 1C), and Klf4 and Nanog proteins (Fig. 1D) in wtD3 and EcadRNAiR mESCs. However, Ecad−/− and EcadRNAi mESCs exhibited decreased Nanog and Klf4 transcript (Fig. 1C) and protein (Fig. 1D) levels in comparison to controls. Following the observation of decreased Nanog protein levels in Ecad−/− and EcadRNAi mESCs in comparison to the control cell lines, a possible role for E-cadherin in the regulation of Nanog transcription was explored in further detail. In agreement with the RT-PCR data, quantitative PCR (qPCR) analysis of Nanog transcript expression levels indicated decreased transcript expression in Ecad−/− and EcadRNAi mESCs in comparison to control cell lines (Fig. 1E).
Ecad−/− mESCs Lack pSTAT3 Activation of the Nanog Promoter
To further investigate the regulation of Nanog by E-cadherin, Ecad−/− and wtD3 mESCs were transfected with a luciferase reporter gene transcriptionally regulated by various truncated fragments of the Nanog promoter region. Results indicated low levels of transcriptional activation of Nanog in response to transfection with each of the vectors in Ecad−/− mESCs whereas wtD3 mESCs exhibited a significant increase in luciferase activity when transfected with the 5-kb vector in comparison to when transfected with the 4.5-kb vector (Fig. 2A; p < .05). It has previously been shown that Nanog transcript expression is upregulated in mESCs by STAT3 and Brachyury (T) binding to regions 4,956 and 4,912 bp upstream of the Nanog promoter, respectively . Deletion of the STAT3 and T binding elements within the Nanog promoter vector resulted in significantly decreased promoter activity in wtD3 mESCs whereas Ecad−/− mESCs exhibited similar activity compared to the 5-kb vector (Fig. 2A; minus STAT3; p < .05). RT-PCR analysis of T transcripts demonstrated that T is not expressed in wtD3 mESCs cultured in FBS-containing medium (Fig. 2B) and therefore cannot be responsible for the differences in luciferase activity observed between Ecad−/− and wtD3 mESCs. However, as observed previously , T was upregulated in wtD3 mESCs upon withdrawal of LIF for 8 days (Fig. 2B). These findings suggest a requirement for STAT3 activation of the Nanog promoter in wtD3 mESCs, indicating that the low activity of the Nanog promoter observed in Ecad−/− mESCs may be due to a lack of STAT3 transcriptional enhancement in these cells. To test this hypothesis, the levels of pSTAT3 in wtD3 and Ecad−/− mESCs were assessed by Western blot analysis. While total STAT3 demonstrated comparable levels between Ecad−/− and wtD3 mESCs (Fig. 2B; STAT3), Ecad−/− mESCs lacked pSTAT3 (Fig. 2B; pSTAT3). Together, these results suggest that the decreased Nanog expression observed in Ecad−/− mESCs reflects a lack of pSTAT3 activation of the Nanog promoter in these cells.
To explore the role of E-cadherin in the activation of other branches of the LIF pathway, we also assessed Akt and ERK1/2 phosphorylation by Western blotting. This technique demonstrated that whereas pERK1/2 is absent in Ecad−/− mESCs (Fig. 2C), pAkt is present in Ecad−/− mESCs but absent in wtD3 mESCs (Fig. 2C). However, since these pathways are activated by a wide range of alternative pathways (whereas STAT3 is activated by LIF alone), we did not pursue this analysis further.
The β-Catenin Binding Region of E-cadherin Is Required for STAT3 Phosphorylation
To determine the region of E-cadherin required for STAT3 activation in mESCs, Ecad−/− mESCs were transiently transfected with cDNA encoding full length E-cadherin (FL-Ecad), E-cadherin lacking the β-catenin binding domain (Δβcat) and E-cadherin lacking the entire cytoplasmic domain  (Δβcat-Δp120ctn). Whereas all three expression vectors conferred an increase in cell surface E-cadherin protein on Ecad−/− mESCs, as evidenced by flow cytometry analysis, only FL-Ecad restored cell–cell contact in these cells (Fig. 3A). In addition, Western blot analysis revealed that only FL-Ecad was able to restore STAT3 phosphorylation in Ecad−/− mESCs and that βcat−/− mESCs also lack pSTAT3 (Fig. 3B). Together, these results indicate that the β-catenin binding domain of E-cadherin is crucial for its role in STAT3 activation.
Ecad−/− mESCs Can Maintain Pluripotency via LIF/BMP
We have previously stated that Ecad−/− mESCs maintain viability upon culture in LIF/BMP medium . Ecad−/− mESCs cultured under these conditions exhibited partial restoration of cell–cell contact, and Oct4 expression was maintained in >95% of the wtD3 and Ecad−/− mESC populations, including when exposed to the Activin/Nodal receptor inhibitor SB431542 (Fig. 4A). Conversely, Ecad−/− mESCs cultured in FBS-containing medium lost Oct4 expression under these latter conditions (Fig. 4A), as previously shown . Furthermore, Oct4 expression was lost in both wtD3 and Ecad−/− mESCs when cultured in LIF/BMP medium supplemented with a Jak inhibitor (Fig. 4B), confirming the requirement for Jak/STAT3 signaling to maintain pluripotency under these conditions. In addition, Ecad−/− mESCs exhibited partial restoration of Klf4 and Nanog transcripts (Fig. 4C, 4D) and protein (Fig. 4E) when cultured in LIF/BMP medium in comparison to when grown in FBS-containing medium. Immunofluorescent and flow cytometry analysis of Klf4 and Nanog expression in Ecad−/− mESCs confirmed these results and demonstrated that transfection of Ecad−/− mESCs with FL-Ecad increased Klf4 and Nanog protein expression (Supporting Information Fig. 1A, 1B). Western blot analysis also demonstrated partial restoration of STAT3 phosphorylation in Ecad−/− mESCs cultured in LIF/BMP (Fig. 4F). Together, these results demonstrate that Ecad−/− mESCs possess functional LIF/BMP signaling pathways that can be activated to support pluripotency marker expression in the absence of Activin/Nodal signaling.
N-cadherin Promotes Ecad−/− mESC Pluripotency Marker Expression in LIF/BMP Medium
We hypothesized that the restoration of cell–cell contact observed in Ecad−/− mESCs cultured in LIF/BMP medium was due to de novo expression of N-cadherin, since N-cadherin has previously been shown to rescue cell–cell adhesion in Ecad−/− mESCs . N-cadherin transcripts were not detected in wtD3 or Ecad−/− mESCs cultured in FBS-containing medium, however, Ecad−/− mESCs cultured in LIF/BMP medium exhibited expression of N-cadherin transcripts (Fig. 5A) and protein (Fig. 5B). Furthermore, N-cadherin protein expression was localized at cell–cell contacts (Fig. 5C) and treatment of the cells with an E- and N-cadherin peptide inhibitor (E/N+) abrogated cell–cell contact in both wtD3 and Ecad−/− mESCs (Fig. 5D). These results suggest that N-cadherin maintains cell–cell contact between Ecad−/− mESCs when cultured in LIF/BMP medium. qPCR analysis of Nanog transcript expression levels in cells grown in LIF/BMP medium demonstrated decreased expression in both wtD3 and Ecad−/− mESCs when treated with the E/N+ peptide compared to untreated cells or cells exposed to the control peptide (E/N−; Fig. 5E). Treatment of wtD3 and Ecad−/− mESCs cultured in LIF/BMP medium with the E/N + peptide had no effect on Oct4 expression as determined by flow cytometry (Fig. 5F), however, exposure to this peptide resulted in the loss of Klf4 transcripts whereas cells treated with the E/N− peptide and untreated cells maintained Klf4 transcript and protein expression (Fig. 5G). Furthermore, STAT3 phosphorylation was abolished in both wtD3 and Ecad−/− mESCs following treatment with the E/N + peptide for 48 hours (Fig. 5H). Together, these results indicate that N-cadherin can compensate for E-cadherin-dependent STAT3 phosphorylation in mESCs. Interestingly, similar results were obtained in Ecad−/− mESCs cultured in 2i medium (Supporting Information Fig. 3).
In this study, we have dissected a novel signaling pathway that connects E- and N-cadherin to the core circuitry of pluripotency in mESCs. First, we show that E-cadherin is required for STAT3 phosphorylation and subsequent enhancement of Klf4 and Nanog transcript and protein expression. Interestingly, blockade of E-cadherin causes direct inhibition of STAT3 phosphorylation since treatment of wtD3 mESCs with the E/N+ peptide for 3 hours resulted in loss of STAT3 phosphorylation (Supporting Information Fig. 1C). Importantly, the observation of decreased pSTAT3, Klf4, and Nanog in Ecad−/− mESCs is not due to phenotypic drift during prolonged culture or isolation of an aberrant clone since we have repeated these results in EcadRNAi mESCs and clones from single-cell low passage Ecad−/− mESCs (Supporting Information Fig. 2). Furthermore, we demonstrate that β-catenin and its binding domain within the cytoplasmic region of E-cadherin is required for STAT3 activation (Fig. 6A). These results support our recent finding that LIF supplementation exerted little effect on the trascriptome of Ecad−/− mESCs , presumably as a result of the lack of STAT3 phosphorylation in these cells. In the absence of E-cadherin, we show that N-cadherin can compensate for the role of E-cadherin in STAT3 phosphorylation and positive regulation of Klf4 and Nanog. To our knowledge, this is the first demonstration that N-cadherin can function in mESCs to maintain pluripotency marker expression.
There remains some controversy within the literature regarding the requirement for E-cadherin in maintaining mESC pluripotency. For example, Redmer et al.  suggest that abrogation of E-cadherin expression in wt mESCs leads to their differentiation. This is inconsistent with our findings; wt mESCs can be cultured in the presence of the E-cadherin inhibiting antibody DECMA-1 while retaining pluripotency marker expression and the ability to form cells representative of the three primary germ layers . Therefore, while loss of E-cadherin leads to decreased Nanog expression levels, which could be interpreted as differentiation, the ability to reverse this phenotype demonstrates that differentiation has not occurred. Fundamentally, the ability of N-cadherin to maintain elevated levels of Nanog protein expression in Ecad−/− mESCs demonstrates that loss of E-cadherin alone is insufficient to induce differentiation in mESCs, at least under the conditions described in this study. Instead, we suggest that loss of E-cadherin in mESCs leads to a reversible pluripotent epiblast-like phenotype. This is supported by recent evidence that EpiSCs can be reprogrammed to a mESC-like naïve pluripotent state , suggesting that LIF stimulation, under appropriate conditions, reprograms Ecad−/− mESCs to a state of naïve pluripotency, as observed in FAB-SCs .
The ability of N-cadherin to promote STAT3 phosphorylation and subsequent Klf4 and Nanog expression in the absence of E-cadherin in mESCs was unexpected and provides a novel function for this protein in the maintenance of mESC pluripotency. E-cadherin expression is shown here to be dominant over N-cadherin expression in wt mESCs since the latter is not upregulated in these cells upon LIF stimulation. Therefore, it appears that N-cadherin only functions as a regulator of STAT3 phosphorylation in the absence of E-cadherin. It is likely that this function of N-cadherin is limited to in vitro culture of mESCs since E-cadherin does not exhibit redundancy during embryogenesis , despite the ability of N-cadherin to rescue the chimera-forming abilities of Ecad−/− mESCs to some extent . Our data suggests that only E- and N-cadherin can promote STAT3 phosphorylation in mESCs. While it is possible that VE-cadherin could also function in a similar manner, since this protein also contains a β-catenin binding domain, treatment of wt mESCs with an E- and N-cadherin inhibitory peptide was sufficient to abrogate STAT3 phosphorylation.
There is a clear hierarchy in mESC pluripotent pathway regulation in both the presence and absence of E-cadherin expression (Fig. 6B). First, the dominant pluripotency pathway/cadherin regulator in wt mESCs is LIF/E-cadherin. In the absence of E-cadherin, mESCs can bypass the LIF pathway, instead using Activin/Nodal to maintain pluripotency marker expression, albeit with decreased Klf4 and Nanog expression. Alternatively, when stimulated with LIF in the absence of Activin/Nodal, mESCs can maintain STAT3 phosphorylation and LIF-dependent pluripotency via N-cadherin. Our data suggest that the Activin/Nodal pathway exhibits priority over the LIF/BMP pathway in the absence of E-cadherin since Ecad−/− mESCs maintain pluripotency marker expression via the former when cultured in FBS-containing medium supplemented with LIF. Ecad−/− mESCs also maintain pluripotency marker expression when cultured in GSK3 and MAPK inhibitors (2i medium; Supporting Information Fig. 3), as previously shown , suggesting these cells also possess a “ground state” of self-renewal. Interestingly, these cells exhibit N-cadherin-dependent STAT3 phosphorylation and upregulation of Klf4 and Nanog in a similar manner to when cultured in LIF/BMP medium.
During ESC differentiation, the cells undergo an EMT-like event which is associated with an E- to N-cadherin switch and increased cell motility [20, 21]. However, the upregulation of N-cadherin that occurs in Ecad−/− mESCs cultured in LIF/BMP medium is not a true EMT event as it leads to increased Nanog expression and restored cell–cell contact, which is in contrast to the function of N-cadherin in EMT. To our knowledge, this is the first description of the ability of N-cadherin to promote pluripotency marker expression, albeit in a mutant (Ecad−/− mES) cell line.
In summary, we demonstrate that E-cadherin is required for STAT3 phosphorylation in mESCs, resulting in positive regulation of Klf4 and Nanog transcripts and protein and, ultimately, facilitating LIF-dependent maintenance of pluripotency. Interestingly, this may explain the requirement for E-cadherin in iPS cell reprogramming since STAT3 phosphorylation has recently been shown to be a limiting factor within this process . In addition, we show for the first time that N-cadherin can compensate for this function of E-cadherin in mESCs cultured in LIF/BMP medium, thus providing a novel function for this adhesion protein in mESC pluripotency. Redmer et al.  have recently demonstrated that expression of E-cadherin can replace Oct4 in the derivation of mouse iPS cells, and it may be interesting to determine whether N-cadherin can also exhibit such an effect. The findings presented here contribute to our understanding of the role of cadherins in potentiating pluripotency networks and may aid our understanding of iPS cell derivation.
Ecad−/− mESCs were generously donated by Prof. Rolf Kemmler, Max-Plank Institute of Immunobiology, Germany. This work was funded by grants from the Biotechnology and Biological Sciences Research Council (BBSRC) and the Engineering and Physical Sciences Research Council. K.H. was funded by a BBSRC PhD studentship.