E-Cadherin-Mediated Cell–Cell Contact Is Critical for Induced Pluripotent Stem Cell Generation§

Authors

  • Taotao Chen,

    1. Laboratory of Molecular Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Detian Yuan,

    1. Laboratory of Molecular Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Bin Wei,

    1. Laboratory of Molecular Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Jing Jiang,

    1. Laboratory of Molecular Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Jiuhong Kang,

    1. Laboratory of Molecular Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    2. Shanghai Key Laboratory of Signaling and Disease Research, School of Life Sciences and Technology, Tongji University, Shanghai, China
    Search for more papers by this author
  • Kun Ling,

    1. Department of Biochemistry and Molecular Biology, Mayo Clinic Cancer Center, Rochester, Minnesota, USA
    Search for more papers by this author
  • Yijun Gu,

    1. Laboratory of Molecular Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Jinsong Li,

    1. Laboratory of Molecular Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Lei Xiao,

    1. Laboratory of Molecular Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
  • Gang Pei

    Corresponding author
    1. Laboratory of Molecular Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    2. Shanghai Key Laboratory of Signaling and Disease Research, School of Life Sciences and Technology, Tongji University, Shanghai, China
    • Laboratory of Molecular Cell Biology, Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
    Search for more papers by this author
    • Tel: +86-21-54921373; fax: +86-21-54921372


  • Author contributions: T.C.: conception and design, collection and/or assembly of data, data analysis and interpretation, and manuscript writing; D.Y.: conception and design, collection and/or assembly of data, data analysis and interpretation, and manuscript writing; B.W.: data analysis and interpretation, manuscript writing; J.J.: collection and/or assembly of data, data analysis and interpretation; J.K.: data analysis and interpretation; K.L.: data analysis and interpretation; Y.G.: collection and/or assembly of data; J.L.: data analysis and interpretation; L.X.: data analysis and interpretation; G.P.: conception and design, data analysis and interpretation, manuscript writing, financial and administrative support, and final approval of manuscript. T.C. and D.Y. contributed equally to this work.

  • Disclosure of potential conflicts of interest is found at the end of this article.

  • §

    First published online in STEM CELLS EXPRESS June 2, 2010.

Abstract

The low efficiency of reprogramming and genomic integration of virus vectors obscure the potential application of induced pluripotent stem (iPS) cells; therefore, identification of chemicals and cooperative factors that may improve the generation of iPS cells will be of great value. Moreover, the cellular mechanisms that limit the reprogramming efficiency need to be investigated. Through screening a chemical library, we found that two chemicals reported to upregulate E-cadherin considerably increase the reprogramming efficiency. Further study of the process indicated that E-cadherin is upregulated during reprogramming and the established iPS cells possess E-cadherin-mediated cell–cell contact, morphologically indistinguishable from embryonic stem (ES) cells. Our experiments also demonstrate that overexpression of E-cadherin significantly enhances reprogramming efficiency, whereas knockdown of endogenous E-cadherin reduces the efficiency. Consistently, abrogation of cell–cell contact by the inhibitory peptide or the neutralizing antibody against the extracellular domain of E-cadherin compromises iPS cell generation. Further mechanistic study reveals that adhesive binding activity of E-cadherin is required. Our results highlight the critical role of E-cadherin-mediated cell–cell contact in reprogramming and suggest new routes for more efficient iPS cell generation. STEM CELLS 2010;28:1315–1325

INTRODUCTION

Induced pluripotent stem (iPS) cells can be generated by direct reprogramming of somatic cells through ectopic expression of defined transcription factors, classically a combination of four factors, Oct3/4, Sox2, c-Myc, and Klf4 [1–3]. iPS cells represent an intriguing new source for patient-specific pluripotent stem cells [4–6]; therefore, many efforts have been taken to make iPS cells more amenable for therapeutic application. To improve the iPS generation efficiency and avoid the oncogenic effect of viral vectors, reduced number of factors as well as small molecules was used to optimize the iPS generation technology [7, 8]. However, the reprogramming process is still highly inefficient [9]. A better understanding of the mechanisms underlying the reprogramming process might help us identify other pathways or cooperative factors to improve the reprogramming efficiency.

Small chemicals that can enhance the generation of iPS cells are of great value because potentially they may replace the oncogenic factors used in the current reprogramming protocol. So far, several chemicals were reported to improve reprogramming efficiency. Among them, some were chromatin-modifying agents, such as valproic acid (VPA), 5-aza-2′-deoxycitydine (AZA), and BIX-01294 [7, 10, 11]; others were mainly inhibitors or activators of certain signal transduction pathways (including mitogen activated protein kinase (MAPK), glycogen synthase kinase 3 beta (GSK3β), and transforming growth factor beta (TGFβ) pathways) [10, 12–14]. Through a high throughput screening, we found two chemicals that reported to upregulate E-cadherin expression had positive effect on the generation of iPS cells. E-cadherin belongs to a large family of transmembrane proteins that mediate specific cell–cell adhesion in a Ca2+-dependent manner [15–17]. Each type of cells expresses certain subsets of cadherins, which is essential to establish the basis of cell-specific identity [18]. In ES cells, it is currently understood that E-cadherin is the main cadherin molecule and regulates cell growth not only through its adhesive binding activity but also through its signal transducing activity [19–21]. Our previous ChIP-chip studies investigated the genome-wide promoter occupancy of the “Yamanaka factors” and identified 16 signaling pathways regulated by these factors in iPS cells [22]. Interestingly, among these, there are four pathways, adhesion junction, gap junction, focal adhesion, and tight junction, which regulate complicated intercellular cross talk. Taken together, these results lead us to further investigate the role of E-cadherin molecule in the reprogramming process.

Here we report that two small molecules with reported activity to upregulate E-cadherin expression can increase the reprogramming efficiency. We find that E-cadherin expression is upregulated in the early stage of reprogramming and the cell–cell contact in iPS cells is mainly mediated by E-cadherin. Enhanced iPS cell generation can be achieved by overexpression of E-cadherin, whereas knockdown of endogenous E-cadherin or abrogation of cell–cell contact by E-cadherin inhibition results in reduced reprogramming efficiency. Further mechanistic study demonstrates that the adhesive binding activity is essential for the reprogramming process.

MATERIALS AND METHODS

Derivation of mouse embryonic fibroblast (MEF) Cells

MEF cells were isolated from e13.5 embryos heterozygous for the Oct4::GFP transgenic allele, as previously described [7]. Gonads and internal organs were removed when the embryos were processed for isolation of MEF cells. Isolated MEF cells in early passages (up to passage 4) were used for further experiments.

Chemical Screening

MEF cells at passage 2 or 3 were seeded onto six-well plates at a density of 200,000 per well in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FBS and transduced with pMXs based retroviruses carrying the four factors. Transduced cells were split onto feeders in 96-well plates at a density of 5,000 per well in KO-DMEM (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) supplemented with 1000 U/ml leukemia inhibitory factor (LIF) and 15% KSR (Invitrogen) on day 3 and 1 μM of each chemical was added to the culture on day 4. The medium was changed every other day. VPA (1 mM), was added from day 4 to 9 and served as a positive control. The other chemicals were added till the end of the experiment. On day 18, the cells were fixed and GFP-positive colonies were counted under a fluorescent microscope (Zeiss Axio Observer Z1). The compounds we used in the initial screening were provided by Chinese National Compound Resource Center. The library contains approximately 500 natural compounds and >1,500 synthetic compounds with known biological activities.

Construction of Retroviral Vectors

For overexpression of the E-cadherin protein, coding sequence of full length human E-cadherin or its mutated forms was cloned into the pMXs-IG retroviral vector, which was a kind gift from Dr. Toshio Kitamura from the University of Tokyo. To analyze the reprogramming efficiency by GFP+ colony numbers, we constructed another set of vectors which do not express the GFP protein by cutting off the GFP sequence from the pMXs-IG backbone.

The short hairpin RNA (shRNA) targeting mouse E-cadherin mRNA or the control shRNA was cloned into the pMKO.1 retroviral vector, which was obtained from Addgene (Addgene, Cambridge, MA, http://www.addgene.org). The sequences for the shRNA targeting mouse E-cadherin and for the control shRNA are shown in Supporting Information Fig. S5.

Retrovirus Production and Generation of iPS Cells

Retrovirus were produced by transfection of plat-E cells with pMXs retroviral vectors containing coding sequences of mouse Oct4, Sox2, Klf4, and c-Myc (which were obtained from Addgene), using FugeneHD transfaction reagent (Roche). Virus containing supernatants were added into plates of MEF cell cultures and spinned at 2,000 rpm for 90 minutes to ensure their transduction. Polybrene was used at a final concentration of 4 μg/ml to facilitate the transduction. Medium was changed immediately after virus transduction and this day is termed as “Day 0.” Two days later, medium was changed into mES medium (DMEM supplemented with 15% FBS, L-glutamine, nonessential amino acids, β-mercaptoethanol, and 1,000 U/ml LIF). On day 4 of post-virus transduction, transduced MEF cultures were digested into single cell cultures and reseeded in 6-well plates at a density of 20,000 cells per well on irradiated MEF feeders for further analysis of reprogramming efficiency.

For quantification of alkaline phosphatase-positive (AP+ colonies, four factors transduced MEF cell cultures on day 10 were fixed and stained for alkaline phosphatase activity using an alkaline phosphatase detection kit (Sigma, catalog No. 85L3R). The plates were then photographed and the numbers of AP+ colonies were counted independently according to the photos, and at least three independent experiments completed by different lab workers in a blind manner were performed to ensure the accuracy of the final results. For quantification of reprogramming efficiency using Oct4::GFP reporter, four factors transuded MEF cell cultures on day 16 were scanned under a fluorescence microscope and scored for GFP+ colonies. When three factors were used, the quantification for AP+ and GFP+ were carried out on day 16 and day 20, respectively.

To generate fully reprogrammed iPS cells, ES-like colonies were picked from transduced MEF cell cultures on day 16, digested into single cell suspensions, and plated in 96-well plates with mES medium. Oct4::GFP positive colonies with homogenous morphologies were further expended for characterization.

Embryoid Body Formation of iPS Cells

The iPS cells and mouse ES cells were grown to approximately 80% confluence and dissociated into single cells in the differentiation medium (DMEM supplemented with 1000U/ml leukemia inhibitory factor (LIF) and 15% FBS, L-glutamine, nonessential amino acids, and β-mercaptoethanol) and seeded at a density of 1 × 105 cells per milliliter into uncoated Petri dishes. The medium was changed every 2 days. Embryoid bodies were collected on day 6 for real-time PCR analysis.

FACS Analysis

MEF cells were retroviral transduced with wild-type E-cadherin or its mutants described above and incubated in DMEM with 15% FBS for 4 days. For surface staining, cells were trypsinized, washed once in PBS, and resuspended in 500 μl of 0.2% BSA in PBS (FACS buffer) containing a rat anti-mouse E-cadherin antibody (DECMA-1; 1:500 dilution; Abcam, Cambridge, U.K., http://www.abcam.com). The cell suspensions were incubated on ice for 30 minutes and then washed three times in PBS, and the cells were resuspended in 500 l of FACS buffer containing Cy3-conjugated anti-rat IgG antibody (1:200 dilution), and incubated on ice for 30 minutes. The cells were washed three times in PBS and then fixed in 1% w/v paraformaldehyde in PBS. Cell fluorescence was analyzed using a Becton Dickinson FACScalibur cell sorter (BD Biosciences, San Diego, http://www.bdbiosciences.com). Viable cells were gated using forward and side scatter and all data represent cells from this population.

RT-PCR and Real-Time PCR

Total RNAs were extracted from cells using Trizol reagent according to the manufacturer's instructions (Sigma). RNA was reverse-transcribed using random hexamers and M-MLV Reverse Transcriptase (Promega, Madison, WI, http://www.promega.com). For semiquantitative PCR analysis, the cDNA solution was amplified for 30 cycles at an optimal annealing temperature. For the real-time quantitative PCR analysis, 2× PCR Mix (Sigma) and Eva Green (Biotium) were used, with an MX3000P Stratagene PCR machine. The relative expression values were normalized against the inner control. Primer sequences for this section were listed in Supporting Information Table S1.

Western Blot

Cell lysates were extracted, separated on reduced SDS-PAGE, and transferred onto nitrocellulose membrane. Primary antibodies were as follows: mouse anti-E-cadherin, mouse anti-N-cadherin, mouse anti-b-catenin (all 1:500 dilution; BD Biosciences), mouse anti-E-cadherin, rabbit anti-Nanog (both 1:200 dilution; Santa Cruz), mouse anti-β-actin (1:1,000; Sigma). The membrane was detected by secondary fluorescent antibody CW800 (Rockland) using an infrared imaging system (Odyssey LI-COR).

Antibody Blocking and Peptide Inhibition

On day 4 of post-viral transduction, MEF cells were reseeded on irradiated feeders and cultured in mES medium as described above. The next day, rat anti-E-cadherin antibody DECMA-1 and the E-cadherin peptide inhibitor SHAVSA were added at a final concentration as described above, respectively. For antibody blocking, heat denatured DECMA-1 was used as control. The peptide SHAVSA and its counterpart SHGVSA were prepared as reported [23]. The antibody or peptide was reapplied every other day with each media change. Plates were scored for AP+ colonies after 5 days of treatment.

Teratoma Production and Analysis

Approximately 1 × 106 iPS cells were suspended in 100 μl mES medium and injected into NOD-SCID mice to form teratomas. Three weeks after injection, teratomas were harvested, fixed overnight with 4% paraformaldehyde, embedded in paraffin, sectioned, HE stained, and analyzed.

Production of Chimeric Mice

Zygotes were isolated from superovulated Female ICR mice and iPS cells (with a C57/BL6 background) were injected into the resulting blastocysts. Chimeras were produced by implantation of injected blastocysts into pseudopregnant ICR mice.

RESULTS

Chemicals That Increase E-Cadherin Expression Enhance the Generation of iPS Cells

We established a 96-well-plate based chemical screening system for the four-factor-induced reprogramming. MEF cells carrying a transgenic Oct4 promoter-driven GFP were used in the chemical screening. More than 2,000 chemicals were screened and 10 chemicals were found to enhance the formation of GFP-positive colonies (data not shown). Interestingly, we noticed that two of the 10 positive chemicals, Apigenin and Luteolin, had been reported to elevate the expression of E-cadherin in some cell lines and animal models [24, 25]. However, S-allyl-L-cysteine (SAC), another chemical that was reported to enhance E-cadherin expression in certain cell lines [26], was identified negative in the screening. We then tested these results in a more detailed manner. MEF cells transduced with four factors were cultured under the indicated concentration of chemicals in DMEM medium for 3 days and the protein levels of E-cadherin were tested by Western blotting. Notably, Apigenin significantly increased E-cadherin expression to approximately 1.4-fold relative to control under optimal concentration of 10 μM; Luteolin also significantly increased E-cadherin expression to approximately 1.6-fold under optimal concentration of 7.5 μM. SAC, however, showed no effect under the concentrations tested and was used as a negative control here (Fig. 1A). Apigenin and Luteolin could also increase E-cadherin expression in MEF cells in the absence of reprogramming factors after 3 days treatment with each compound, as determined by RT-PCR analysis (Supporting Information Fig. S2B). The chemicals were then added to the mES medium after the transduced MEF cells were reseeded upon feeder cells on day 4 to monitor their ability to affect the reprogramming efficiency (Fig. 1B). Remarkably, addition of Apigenin or Luteolin on their optimized concentrations caused a nearly fourfold increase in the number of AP+ colonies, whereas addition of SAC showed no obvious effect (Fig. 1C and Supporting Information Fig. S1A). A similar fourfold increase was also seen when number of Oct4::GFP positive colonies were counted (Supporting Information Fig. S1C). The increased iPS cell generation correlated well with the upregulation of E-cadherin protein level (Fig. 1C, 1D). We also found that Apigenin promotes iPS cell formation in a dose-dependent manner (Fig. 1E and Supporting Information Fig. S1B), which correlates with its ability to upregulate endogenous E-cadherin expression (Fig. 1F). The similar dose-dependent increase was also obvious when number of Oct4::GFP positive colonies were used to compare the reprogramming efficiency (Supporting Information Fig. S1D). The result of MTT assay revealed that both Apigenin and Luteolin inhibit cell proliferation, indicating that the increased iPS generation is not achieved by facilitating cell proliferation (Supporting Information Fig. S2A). It has been reported that Apigenin has direct interference with Smad2, a signal transducer of Tgf-β pathway known to regulate E-cadherin expression [27, 28]. We thus explored whether these two chemicals increase E-cadherin expression through regulating Tgf-β pathway. We found that treatment with Apigenin (10 μM) or Luteolin (7.5 μM) directly reduced Smad2 expression (Supporting Information Fig. S2C), indicating attenuated Tgf-β signaling in MEF cells may contribute to the upregulation of E-cadherin expression. These observations demonstrate that E-cadherin promoting chemicals have positive effect on the iPS cell generation.

Figure 1.

Chemicals that upregulate E-cadherin can enhance reprogramming efficiency. (A): Immunoblotting of E-cadherin expression in four-factor transduced MEF cells under small-molecule treatment, with the concentrations listed. Equal volume of dimethyl sulfoxide (DMSO) was used as control. Quantification by densitometric analysis is shown. (B): Schematic representation of induced pluripotent stem (iPS) cell generation with chemical treatment. (C): Quantification of the efficiency of four-factor induced iPS cell generation with small-molecule treatment. The error bars denote the standard error derived from quantification of three separate wells of cells. **p < .01, versus control. (D): Fold change in the E-cadherin protein level under chemical treatment. The error bars denote the standard error derived from densitometric analysis of three independent experiments. *p < .05, versus control. (E): Quantification of the efficiency of four-factor induced iPS cell generation under Apegenin (5, 10, and 15 μM) treatment. **p < .01, versus control. (F): Fold change in the E-cadherin protein level under Apigenin (5, 10, and 15 μM) treatment. The error bars denote the standard error derived from densitometric analysis of three independent experiments. *p < .05, versus control. Abbreviations: AP+, alkaline phosphatase-positive; Ctr, control;; ES, embryonic stem; MEF, mouse embryonic fibroblast; SAC, S-allyl-L-cysteine.

E-Cadherin Is Upregulated During the Early Stage of Reprogramming

E-cadherin has been previously shown to express at a high level in ES cells and in a variety of adult stem cells in Drosophila and mice [20, 29, 30]. To test the function of E-cadherin in iPS cells, we derived stable 4F (four factors; Oct4, Sox2, Klf4, and c-Myc) iPS cells and 3F (three factors; Oct4, Sox2 and Klf4) iPS cells from Oct4::GFP MEF cells using the standard protocol for retroviral transduction. Single iPS colonies were picked and expanded into stable cell lines (named 4F-iPS and 3F-iPS throughout the article). Both 4F-iPS and 3F-iPS cells were morphologically indistinguishable from ES cells and expressed the Oct4::GFP at a high level (Fig. 2A). Activation of Nanog as well as endogenous Oct4 and Sox2 was thought to be a property of fully pluripotent iPS cells [2]. We selected three clones of 4F-iPS cells and two clones of 3F-iPS cells for analysis. These iPS cells showed activation of endogenous Oct4 and Sox2, as tested by primers specific for endogenous transcripts (Fig. 2B), and expressed the pluripotency marker Nanog (Fig. 2B, 2C). One of these iPS cell lines, 4F-iPS-1, was injected to blastocyst and transplanted into pseudopregnant ICR female mice to test its ability for chimeric formation. As expected, live chimeras were obtained, proving the full pluripotency of this iPS cell line (Supporting Information Fig. S3A). We found that E-cadherin protein was expressed at a similar level in these stabilized iPS cells as in ES cells but was absent in MEF cells (Fig. 2B, 2C). Notably, N-cadherin, another classic cadherin whose tissue distribution is distinct from that of E-cadherin, was expressed in MEF cells but had been shut down in pluripotent cells (Fig. 2B, 2C), indicating that an N-cadherin to E-cadherin switch occurred during the reprogramming process.

Figure 2.

E-cadherin is upregulated during the early stage of reprogramming. (A): Morphology of 4F-iPS cells and 3F-iPS cells on MEF feeder cells. Scale bars represent 100 μm. (B): RT-PCR analysis of pluripotency markers, E-cadherin (E-cad) and N-cadherin (N-cad) expression in 4F-iPS cells, 3F-iPS cells, and mES cells. Primer sets used for Oct4 and Sox2 were specific for the endogenous transcripts. β-actin was used as a loading control [31]. (C): Western blot analysis of E-cadherin, N-cadherin, and Nanog expression on 4F-iPS cells, 3F-iPS cells, and mES cells. (D): Immunofluorescence staining of MEF cells, mES cell, and 4F-iPS cells for pluripotency and cell–cell contact markers. DAPI staining indicates the nuclear. Scale bars represent 50 μm. (E): Real-time PCR analysis of Oct4, Nanog, E-cadherin, and N-cadherin on the indicated days in four-factor transduced MEF cells during the reprogramming process. β-actin was used as an internal control. The error bars denote the standard error derived from quantification of three independent experiments. (F): Western blot analysis of E-cadherin, N-cadherin, and Nanog expression on four-factor transduced MEF cells during the reprogramming process. (G): Immunofluorescence staining of four-factor transduced MEF cells for E-cadherin and N-cadherin on day 6 post-virus infection. Scale bars represent 50 μm. Abbreviations: iPS, induced pluripotent stem; MEF, mouse embryonic fibroblast; mES, moue embryonic stem; SSEA1, stage-specific embryonic antigen-1; ZO-1, zona occludens 1.

To further clarify the function of E-cadherin in iPS cells, subcellular localization of E-cadherin as well as several other cell adhesion regulators was compared by immunofluorescence study in MEF cells and iPS cells. Consistent with their spindle-shaped morphology, MEF cells did not express E-cadherin and the tight junction protein ZO-1. Instead, these cells showed high levels of N-cadherin which mainly accumulated in the cytosol. The cadherin binding partner β-catenin also showed a cytosolic distribution in MEF cells. However, iPS cells derived from MEF cells formed compatible colonies as those of mouse ES cells, accompanied by membrane localization of E-cadherin, β-catenin, and ZO-1, and virtually complete loss of N-cadherin (Fig. 2D). These alterations indicated that iPS cells have undergone a transition in which four-factor transduced MEF cells established and maintained E-cadherin-mediated cell–cell contact, a typical feature of ES cells. This observation raised the possibility that establishment of cell–cell contact has a correlation with reprogramming and E-cadherin may be a key component that contributes to this correlation.

The sequential activations of endogenous pluripotency markers, such as SSEA1, Oct4, and Nanog, have been proposed to reflect the progress of reprogramming [32, 33]. To further investigate the nature of E-cadherin in iPS cell generation, we then explored the dynamic changes of E-cadherin expression during the reprogramming process and compared it with those of SSEA1 and Nanog. Surprisingly, E-cadherin was activated as early as 2 days post-virus infection, a time point when the four reprogramming factors started to express, and further upregulated for several days upon viral transduction, then reached a steady level (Fig. 2E, 2F and Supporting Information Fig. S3B). However, as a late cornerstone in the reprogramming process, Nanog only became detectable on day 8 post-viral transduction (Fig. 2E, 2F). During the whole reprogramming period, upregulation of SSEA1 has been defined as an early event [32, 33]. Immunofluorescent analysis of E- cadherin and SSEA1 expression in four-factor transduced MEF cells demonstrated that emergence of E-cadherin protein was clearly detected at day 4, whereas SSEA1 expression could be detected until day 6 (Supporting Information Fig. S3C). These data indicated that activation of endogenous E-cadherin marks an early event in the reprogramming process. Interestingly, the expression of N-cadherin downregulated at day 3 and kept a reduced expression level afterward in the entire population, showing a contrary manner with E-cadherin, as in many other biological processes such as epithelial-mesenchymal transition (EMT) (Fig. 2E, 2F). Immunofluorescent study further confirmed that E-cadherin and N-cadherin were exclusively regulated in the reprogrammed MEF cells. In the compatible cells that underwent morphological changes, E-cadherin was expressed and localized to the cell membrane, indicating its functional roles here; meanwhile, N-cadherin was downregulated in these cells, in contrast to the surrounding cells which still showed typical fibroblast-like morphology, resulting in an explicit boundary between these two groups of cells (Fig. 2G). These data demonstrate that an N-cadherin to E-cadherin switch occurs during the early stage of reprogramming.

Ectopic Expression of E-Cadherin Enhances iPS Cell Generation

We next examined whether ectopic expression of E-cadherin is sufficient to facilitate the reprogramming process and thus promote the iPS cell generation. Four or three transcription factors accompanied with human E-cadherin were introduced into MEF cells with retroviral vectors. In four-factor induced reprogramming, we typically got 100 AP+ colonies per 20,000 starting cells. The overall efficiency was calculated approximately 0.5%, similar as previously reported by other groups [34]. When human E-cadherin was introduced together, the AP+ colony number boosted to 400 per 20,000 cells (approximately 2% of efficiency, Fig. 3A). We reproduced the experiment and consistently got a fourfold increase in the reprogramming efficiency (approximately 2% of efficiency in the presence of E-cadherin vs. 0.5% in the absence of it), through counting either the number of AP+ or GFP+ colonies (Fig. 3B, 3C). In the three-factor induced reprogramming, the typical efficiency we got was approximately 30 colonies per 20,000 starting cells (Fig. 3A). When E-cadherin was added to the recipe, we consistently got a threefold to fourfold increase in the reprogramming efficiency, by counting either the number of AP+ or GFP+ colonies (Fig. 3A–3C). These data suggest that overexpression of E-cadherin by viral transduction is sufficient to enhance the iPS cell generation.

Figure 3.

Ectopic expression of E-cadherin enhances reprogramming efficiency. (A): Representative images of AP+ colonies. MEF cells were transduced with retrovirus containing four factors (Oct3/4, Sox2, Klf4, and c-Myc) plus E-cadherin or three factors (Oct3/4, Sox2, and Klf4) plus E-cadherin and cultured in the standard iPS induction conditions. Cultures were fixed and stained for alkaline phosphatase activity (Upper panel). Parallel-transduced MEF cells were tested for E-cadherin overexpression by Western blot using antibody specific for E-cadherin (Lower panel). (B): Quantification of AP+ colony numbers after E-cadherin overexpression. Bars represent numbers that normalized as relative colony numbers per 20,000 starting cells. The error bars denote the standard error derived from quantification of three independent experiments. (C): Quantification of Oct4::GFP positive (GFP+) colony numbers after E-cadherin overexpression. As a more stringent quantification for reprogramming efficiency, GFP+ colony numbers were counted under a fluorescence microscope. Bars represent numbers that normalized as relative colony numbers per 20,000 starting cells. The error bars denote the standard error derived from quantification of three independent experiments. (D): Morphology of iPS cells derived by the four factors plus E-cadherin method (E4F-iPS cells). Phase contrast photo shows an ES-like morphology of E4F-iPS cells. Antibody staining shows that E4F-iPS cells express pluripotency markers SSEA1 and Nanog. Represent pictures of iPS cell line E4F-iPS-1 are shown. The scale bars represent 100 μm. (E): Morphology of E4F-iPS cells maintained in medium with and without LIF. E4F-iPS cells were passaged on MEF feeder cells and maintained in + LIF medium or – LIF medium for 3 days. Morphology (phase) and Oct4::GFP expression (Oct4::GFP) were scanned and photographed. The scale bars represent 100 μm. (F): Real-time PCR analysis of the markers for pluripotency and all three germ layers in embryoid bodies derived from E4F-iPS cells. Results from cell line E4F-iPS-1 are shown. Bars represent expression levels relative to that of ES cells. β-actin was used as an internal control. (G): Teratomas derived from E4F-iPS cells. Shown are representatives HE staining pictures for ectoderm (neural tissue), mesoderm (cartilage) and endoderm (gut). Scale Bars represent 50 μm. (H): Two-week-old chimeric mice derived from E4F-iPS cells. Abbreviations: AP+, alkaline phosphatase-positive; Bry, brachyury; ES, embryonic stem; ES-EB, embryonic stem cells derived embryoid bodies; GFP, green fluorescent protein; FGF5, fibroblast growth factor 5; 3F, three factors; 4F, four factors; SSEA1, stage-specific embryonic antigen-1.

To address whether enhanced reprogramming efficiency by E-cadherin overexpression could finally result in bona fide iPS cells, we then generated stable iPS cell lines using the four factors plus E-cadherin method. These iPS cells were referred to E4F-iPS cells throughout the article. E4F-iPS cells maintain an ES-like morphology and show significant high expression of the pluripotency markers AP, SSEA1, and Nanog (Fig. 3D). The overexpression of E-cadherin was confirmed by genomic PCR genotyping (Supporting Information Fig. S4). In vitro differentiation assays further confirmed the pluripotency of the E4F-iPS cells. In the presence of LIF in ES medium, E4F-iPS cells expressed high level of Oct4::GFP. In contrast, in the absence of LIF in the medium, E4F-iPS cells began to differentiate, losing the colony morphology as well as the Oct4::GFP expression (Fig. 3E). When cultured in suspension, E4F-iPS cells formed embryoid bodies just like ES cells. Pluripotency markers Oct4 and Nanog were largely silenced in day 6 EBs of E4F-iPS cells, whereas markers of differentiated cell lineages were activated to a similar level as that in ES cells derived EBs (ES-EBs) (Fig. 3F). Teratomas that contain all three germ layers were formed when E4F-iPS cells were injected into NOD-SCID mice, further confirming their differentiation ability (Fig. 3G). Finally, we confirmed whether E4F-iPS cells are fully reprogrammed iPS cells by testing their ability to form chimeric mice following blastocyst injection. Live chimeras with high contribution from E4F-iPS cells were obtained, proved the in vivo differentiation potential of these cells (Fig. 3H). Our results thus show that iPS cells generated with E-cadherin overexpression are indeed fully pluripotent iPS cells.

Ablation of Endogenous E-Cadherin Reduces the Reprograming Efficiency

To address the role of E-cadherin in iPS cell generation, we performed comparison experiments using the standard four-factor induction protocol. Four factors were cointroduced into Oct4::GFP MEF cells together with either E-cadherin shRNA or control shRNA and cultured under standard conditions. shRNA hairpins were tested to functionally downregulate E-cadherin expression in mouse hepatocytes, which express endogenous E-cadherin (Supporting Information Fig. S5A). Although E-cadherin expression was unaffected by the control hairpin, E-cadherin shRNA demonstrated 80%–90% downregulation of E-cadherin expression, as tested by Western blot (Supporting Information Fig. S5A). Next we explored the functional effect of knocking down E-cadherin on iPS cell generation. For quantification, AP+ colonies were photographed and counted on day 10 post-viral transduction (Fig. 4A). When MEF cells were transduced with four factors plus control shRNA, we consistently got comparable iPS cell generation efficiency as reported, approximately 150 colonies from 20,000 starting cells [34]. However, only approximately 30 colonies (approximately 80% decrease) were available in a paralleled experiment when E-cadherin expression was disrupted using E-cadherin shRNA (Fig. 4A, 4B). As a more stringent quantification standard, Oct4::GFP+ colonies were photographed and counted on day 16 post-viral transduction (Fig. 4C). Here, we typically got 45 GFP+ colonies from 20,000 starting cells. However, 12 GFP+ colonies from 20,000 starting cells (approximately a 75% decrease) were generated in the E-cadherin shRNA experiment (Fig. 4C and Supporting Information Fig. S5B). Notably, the decreased iPS cell generation was coincidental with E-cadherin protein level (Fig. 4A). E-cadherin has been reported to arrest cell growth in certain cancer cell lines [35, 36]; however, no difference in proliferation was observed between the control and E-cadherin knockdown samples in our experiment as determined by MTT assay, indicating that the reduction in reprogramming efficiency we observed was independent of cell proliferation rates (Supporting Information Fig. S5C). To further address that the impaired efficiency was specifically resulted from E-cadherin knockdown, we examined whether E-cadherin reintroduction was sufficient to rescue iPS cell generation. As expected, reprogramming efficiency was fully restored to normal levels when human E-cadherin, which was not targeted by mouse E-cadherin shRNA, was expressed (Fig. 4B, 4C). These data indicate that activation of endogenous E-cadherin may be essential for normal iPS cell generation.

Figure 4.

Knockdown of endogenous E-cadherin reduces the reprograming efficiency. (A): Representative images of AP+ colonies. MEF cells were infected with retrovirus containing four factors in combination with mock, E-cadherin (E-cad) shRNA, E-cad shRNA control, and human E-cadherin (hE-cad) as indicated. After 10 days, cells were fixed and stained for AP activity. (B): Quantification of AP+ colonies as indicated. Expression of E-cadherin was examined by Western blot (bottom). β-actin was used as a loading control. The error bars indicate the standard error derived from quantification of three independent experiments. ***p < .001, versus control. (C): Quantification of Oct4::GFP positive (GFP+) colonies on day 16 after retroviral infection. The error bars indicate the standard error derived from quantification of three independent experiments. **p < .01, versus control. Abbreviations: AP+, alkaline phosphatase-positive; GFP, green fluorescent protein.

Adhesive Binding Activity of E-Cadherin Is Required for Reprograming

Our data have suggested that E-cadherin plays an important role in the reprogramming process. To further investigate the functional role of E-cadherin, we used the adhesion inhibitory peptide and antibody to specifically abrogate E-cadherin-mediated cell–cell contact throughout the reprogramming process. HAV containing peptide (termed SHAVSA) treatment of mES cells has been reported to result in loss of cell–cell contact by inhibiting cadherin engagement [19, 23]. Indeed, when cultured under 10 mM SHAVSA for 24 hours, mES cells lost their typical colony morphology and developed a fibroblast-like phenotype, showing the cell–cell contact had been disrupted. In contrast, SHGVSA, the malfunctioned counterpart of SHAVSA, did not have this effect (Fig. 5A). The inhibiting peptide was then added to iPS cell culture medium to see if it affected the iPS cell generation process. When the concentration of SHAVSA reached 5 mM, a significant reduction in reprogramming efficiency was observed. When the concentration of SHAVSA reached 10 mM, the reprogramming efficiency declined to approximately 60%. The control peptide SHGVSA still exhibited no obvious effect (Fig. 5B and Supporting Information Fig. S6A).

Figure 5.

Abrogation of cell–cell contact compromises iPS cell generation. (A): Phenotype of mES cells treated under control peptide (SHAVSA) or with inhibiting peptide (SHGVSA) for 24 hours. Immunofluorescence analysis shows Oct4 and Nanog protein expression in peptide-treated mES cells. Scale bars represent 100 μm. (B): Quantification of AP+ colonies on four-factor transduced MEF cells treated with SHAVSA and SHGVSA. Four factors transduced MEF cells were cultured in standard ES medium or ES medium containing indicated concentrations of SHAVSA and SHGVSA peptide. The error bars denote the standard error derived from quantification of three independent experiments. *p < .05, **p < .01, versus control. (C): Representative pictures of mES cells after treatment with the antibody DECMA-1 or the denatured DECMA-1 for 24 hours. The concentration of DECMA-1 used here is 16 μg/ml. (D): Quantification of AP+ colonies on four-factor transduced MEF cells treated with DECMA-1 (4, 8, and 16 μg/ml). Heat denatured DECMA-1 (16 μg/ml) was used as control. The error bars denote the standard error derived from quantification of three independent experiments. **p < .01, versus control. (E): Quantification of the efficiency of four-factor induced iPS cell generation after E-cadherin knockdown and rescue. MEF cells were infected by four factors in combination with mock, E-cadherin shRNA (E-cad shRNA) and full-length (FL) E-cadherin or mutant E-cadherin (EΔ71, EΔC, and W2A). The number of colonies staining positive for AP (AP+) were counted on day 10 after infection. The error bars denote the standard error derived from quantification of three independent experiments. *p < .05, **p < .01, versus control. Expression of the proteins was examined by Western blot (bottom). (F): Quantification of the efficiency of four-factor induced iPS cell generation with overexpression of FL E-cadherin or mutant E-cadherin (EΔ71, EΔC, and W2A). The number of AP+ colonies was counted. The error bars denote the standard error derived from quantification of three independent experiments. *p < .05, **p < .01, versus control. Expression of the proteins was examined by Western blot (bottom). Abbreviations: AP+, alkaline phosphatase-positive; DAPI, 40,6-diamidino-2-phenylindole; DIC, differential interference contrast; FL, full length.

Inhibition of E-cadherin hemophilic binding by DECMA-1 has been shown to decrease cell–cell contact in mouse ES cells by specific blockade of the extracellular domain of E-cadherin [29]. As reported, we observed loss of cell–cell contact in mES cells after addition of DECMA-1 to the growth medium for 24 hours, which was similar to the effect of HAV-containing peptide SHAVSA (Fig. 5C). Therefore, we used DECMA-1 to further substantiate our conclusion. MEF cells were virally transduced with four factors and cultured under DECMA-1 containing medium. Heat-denatured DECMA-1 was used as control. As expected, DECMA-1 decreased the number of AP+ colonies in a dosage-dependent manner. A significant decrease was observed when 16 μg/ml of DECMA-1 was used (Fig. 5D and Supporting Information Fig. S6B). These data further reinforce the conclusion that E-cadherin-mediated cell–cell contact is required for efficient reprogramming.

E-cadherin itself could not only mediate cell–cell contact but also could provide downstream signals [37–39]. In the reprogramming process, it is rather interesting to clarify whether this protein is only implicated in intercellular adhesion and therefore keeps cells together, or if it is also engaged in intracellular signaling. To address this issue, we next investigated which region of E-cadherin protein was required for iPS cell generation. EΔ71, the E-cadherin mutant lacking the β-catenin binding domain, EΔC, the E-cadherin mutant lacking the entire intracellular domain, and, W2A, and E-cadherin point mutant bearing a Trp-2 to alanine mutation, which was defective in hemophilic binding, were constructed and applied in the iPS generation recipe (Supporting Information Fig. S6C). The properties of these mutants, including the hemophilic binding activity, membrane localization, and stability have been well-investigated previously [18, 40, 41]. Expression of the E-cadherin mutants was verified by Western blotting using antibodies that recognize the intracellular domain and extracellular domain, respectively (Supporting Information Fig. S6D). As quality control, immnofluorescence and fluorescent flow cytometry analysis demonstrated normal cell surface expression of full length E-cadherin as well as its mutated forms EΔ71, EΔC, and W2A (Supporting Information Fig. S6E). We first investigated which construct(s) could rescue the knock down effect of E-cadherin in four-factor transduced MEF cells. Surprisingly, EΔ71, the β-catenin binding domain lacking protein, could fully rescue the reprogramming efficiency compared with full-length E-cadherin, whereas EΔC, the protein lacking nearly the entire intracellular domain could also partially rescue the knockdown effect (Fig. 5E). However, the W2A mutant exhibited no rescuing effect (Fig. 5E). We then addressed whether these mutants would facilitate iPS cell generation when overexpressed along with the four factors. When EΔ71 and EΔC (the two intracelluar domain damaging proteins) were transduced together with the four reprogramming factors, we observed a significant increase in iPS generation efficiency, the same as their full length counterpart, whereas the W2A mutation, which shows a dominant negative effect on adhesive interactions, exhibited significant inhibition on the reprogramming efficiency (Fig. 5F). We compared the effect on cell proliferation of these mutants with full length E-cadherin by MTT, and no obvious difference was observed (Supporting Information Fig. S6F). Together, these data suggest that adhesive activity of E-cadherin is critical for iPS cell generation.

DISCUSSION

In this study, we have identified a new route to improve the reprogramming efficiency that involves E-cadherin-mediated cell–cell contact. We confirm that current reprogramming method can be optimized either by application of chemicals that upregulate E-cadherin expression or by directly overexpression of E-cadherin. These observations may lead to the application of new chemicals or new factors that are involved in cell–cell contact to the current reprogramming protocol. Notably, E-cadherin improves the reprogramming efficiency at comparable extents between four factors and three factors induced reprogramming, indicating its c-Myc independent mechanisms, which is similar to many previously reported chemicals [7, 8, 10]. We noticed that Apigenin and Luteolin have also been shown to possess antioxidant properties [42], which may contribute to improve the reprogramming efficiency here, as hypoxia has been reported to enhance the generation of both mouse and human iPS cells [43].

It has been reported that the major function of E-cadherin in ES cells is to maintain the colony morphology [21, 29]. However, our observations reveal that E-cadherin is directly involved in the restoration of pluripotency. Interestingly, the reprogramming efficiency is enhanced by not only the full length E-cadherin protein but also its C-terminal truncated forms, demonstrating a new concept that certain domains of a protein or factor is enough to facilitate reprogramming. It seems reasonable to assume that transformed MEF cells require E-cadherin-mediated cell–cell contact to sense adhesion dependent signals from neighboring cells, for sustaining reprogramming process [44, 45]. Candidates of the downstream signals initiated by E-cadherin upregulation include notch signals and gap junction dependent pathways [46–48]. Further clarifying the responsible signals may contribute to more efficient reprogramming. Maintaining cell–cell contact is also expected to enable cells more resistant to external and internal perturbations [49]. This aspect may be particularly important for iPS cell generation because the reprogramming process is so complicated that any obstacle encountered by the intermediate cells will drive them away from the reprogramming process. The hypothesis is also supported by the recent derivation of FAB stem cells, which are believed to be trapped in a state of partial pluripotency because lacking of appropriate E-cadherin expression [21]. Following reprogramming, MEF cells are believed to quickly get into the process of mesenthymal to epithelia transition (MET) upon the expression of reprogramming factors [50] and upregulation of E-cadherin is one of the important features of MET [51]. Insteadly, the early upregulation of E-cadherin during reprogramming may be a direct consequence of the expression of reprogramming factors because previously we have found that the promoter region of E-cadherin contains Oct3/4 binding sites ([22] and unpublished data). It is reported that enhanced MET may accelerate the reprogramming process [50]. Thus, upregulation of E-cadherin may also serve to enhance the MET transition, resulting in improved reprogramming efficiency. Our findings therefore extended the role of E-cadherin as well as MET transition during fibroblast to iPS cell reprogramming.

The reprogramming process is believed to be a stochastic process amenable to acceleration through two distinct mechanisms, namely, cell-division-rate-dependent ways versus cell-division-rate-independent ways [52]. The mechanism by which E-cadherin improves the reprogramming efficiency can be categorized to the later because E-cadherin shows no effect on cell proliferation rate when introduced together with reprogramming factors. Therefore, we propose that E-cadherin-mediated cell–cell contact may be an intrinsic property that defines the basis of stem-cell specific identity. It is of great interest to identify the downstream signals that contribute to this property and potentially influence the reprogramming efficiency. It is also an opened question whether other factors and signals that function in the intercellular communication system of the pluripotent stem cells may be applied to further optimize the reprogramming technology. Finally, it is reported that E-cadherin also involved in the regulation of human ES cell pluripotency [53, 54]. Thus, our findings may also be applicable to optimize the generation of human iPS cells and possibly other mammalian iPS cells.

CONCLUSION

We have demonstrated that E-cadherin-mediated cell-cell contact is critical for the generation of induced pluripotent stem cells. Knockdown of E-cadherin results in reduced reprogramming efficiency due to the loss of cell-cell contact. Further more, upregualtion of E-cadherin by chemicals or by direct overexpression significantly facilitates reprogramming. Our results provide a novel function of E-cadherin and give new insights into the mechanisms of reprogramming.

Acknowledgements

We thank Dr. Duanqing Pei for continued help and instructions on the iPS generation technology. We thank Dr. Junlin Guan, Dr. Jian Zhao, and Dr. Xin Xie for helpful discussion and Jing Liao and Lei Qian for technical assistance. We appreciate Shunmei Xin, Shan Chen, and Xianglu Zeng for their technical assistance. We thank all members of the lab for sharing reagents and advice. We thank Chinese National Compound Resource Center for providing us the chemical libraries for initial screenings. We thank FulenGen for providing us with some of the plasmids. This research was supported by grants from the Ministry of Science and Technology (2007CB947904, 2007CB947100, 2007CB948000, 2009CB9 41100, 2009CB940903, 2010CB944900), National Natural Science Foundation of China (30625014, 30623003, 30871285, 90713047, 90919028, 30871430), Shanghai Municipal Commission for Science and Technology (07DZ22919, 08dj1400500, 08dj1400502, 09PJ1410900), Natural Science Foundation of Shanghai (10ZR1435300), and Chinese Academy of Sciences (SIBS2008001, KSCX2-YW-R-110, KSCX2-YW-R-229).

DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST

The authors indicate no potential conflicts of interest.

Ancillary