Modulating Glypican4 Suppresses Tumorigenicity of Embryonic Stem Cells While Preserving Self-Renewal and Pluripotency

Authors

  • Annalisa Fico,

    1. Institut de Biologie de Développement de Marseille-Luminy (IBDML), CNRS UMR 7288, Case 907, Parc Scientifique de Luminy, Aix-Marseille Université, Marseille Cedex 09, France
    Search for more papers by this author
  • Antoine De Chevigny,

    1. Institut de Biologie de Développement de Marseille-Luminy (IBDML), CNRS UMR 7288, Case 907, Parc Scientifique de Luminy, Aix-Marseille Université, Marseille Cedex 09, France
    Search for more papers by this author
    • Author contributions: A.F.: conception and design, collection, assembly of data, data analysis and interpretation, and final approval of the manuscript; A.D.C.: contribution to design and performed and interpreted the shRNA experiments in postnatal brains; J.E. and M.R.B.: contribution to design and performed and interpreted the chimeric mouse experiments; H.C.: contribution to conception and design of the shRNA experiments in postnatal brains and financial support; F.M.: contribution to experimental design, data interpretation, and financial support; R.D.: contribution to conception and design, data analysis and interpretation, provision of study materials, financial support, and manuscript writing.

  • Joaquim Egea,

    1. IRBLLEIDA, Campus de Ciències de la Salut, Universitat de Lleida, C. de Montserrat Roig 2, Lleida, Spain
    Search for more papers by this author
    • Author contributions: A.F.: conception and design, collection, assembly of data, data analysis and interpretation, and final approval of the manuscript; A.D.C.: contribution to design and performed and interpreted the shRNA experiments in postnatal brains; J.E. and M.R.B.: contribution to design and performed and interpreted the chimeric mouse experiments; H.C.: contribution to conception and design of the shRNA experiments in postnatal brains and financial support; F.M.: contribution to experimental design, data interpretation, and financial support; R.D.: contribution to conception and design, data analysis and interpretation, provision of study materials, financial support, and manuscript writing.

  • Michael R. Bösl,

    1. Max-Planck-Institute of Neurobiology, Am Klopferspitz 18, Martinsried, Germany
    Search for more papers by this author
    • Author contributions: A.F.: conception and design, collection, assembly of data, data analysis and interpretation, and final approval of the manuscript; A.D.C.: contribution to design and performed and interpreted the shRNA experiments in postnatal brains; J.E. and M.R.B.: contribution to design and performed and interpreted the chimeric mouse experiments; H.C.: contribution to conception and design of the shRNA experiments in postnatal brains and financial support; F.M.: contribution to experimental design, data interpretation, and financial support; R.D.: contribution to conception and design, data analysis and interpretation, provision of study materials, financial support, and manuscript writing.

  • Harold Cremer,

    1. Institut de Biologie de Développement de Marseille-Luminy (IBDML), CNRS UMR 7288, Case 907, Parc Scientifique de Luminy, Aix-Marseille Université, Marseille Cedex 09, France
    Search for more papers by this author
    • Author contributions: A.F.: conception and design, collection, assembly of data, data analysis and interpretation, and final approval of the manuscript; A.D.C.: contribution to design and performed and interpreted the shRNA experiments in postnatal brains; J.E. and M.R.B.: contribution to design and performed and interpreted the chimeric mouse experiments; H.C.: contribution to conception and design of the shRNA experiments in postnatal brains and financial support; F.M.: contribution to experimental design, data interpretation, and financial support; R.D.: contribution to conception and design, data analysis and interpretation, provision of study materials, financial support, and manuscript writing.

  • Flavio Maina,

    1. Institut de Biologie de Développement de Marseille-Luminy (IBDML), CNRS UMR 7288, Case 907, Parc Scientifique de Luminy, Aix-Marseille Université, Marseille Cedex 09, France
    Search for more papers by this author
    • Author contributions: A.F.: conception and design, collection, assembly of data, data analysis and interpretation, and final approval of the manuscript; A.D.C.: contribution to design and performed and interpreted the shRNA experiments in postnatal brains; J.E. and M.R.B.: contribution to design and performed and interpreted the chimeric mouse experiments; H.C.: contribution to conception and design of the shRNA experiments in postnatal brains and financial support; F.M.: contribution to experimental design, data interpretation, and financial support; R.D.: contribution to conception and design, data analysis and interpretation, provision of study materials, financial support, and manuscript writing.

  • Rosanna Dono

    Corresponding author
    1. Institut de Biologie de Développement de Marseille-Luminy (IBDML), CNRS UMR 7288, Case 907, Parc Scientifique de Luminy, Aix-Marseille Université, Marseille Cedex 09, France
    • IBDML, CNRS UMR 7288, Case 907, Parc Scientifique de Luminy, Aix-Marseille Université, 13288 Marseille Cedex 09, France
    Search for more papers by this author
    • Author contributions: A.F.: conception and design, collection, assembly of data, data analysis and interpretation, and final approval of the manuscript; A.D.C.: contribution to design and performed and interpreted the shRNA experiments in postnatal brains; J.E. and M.R.B.: contribution to design and performed and interpreted the chimeric mouse experiments; H.C.: contribution to conception and design of the shRNA experiments in postnatal brains and financial support; F.M.: contribution to experimental design, data interpretation, and financial support; R.D.: contribution to conception and design, data analysis and interpretation, provision of study materials, financial support, and manuscript writing.

    • Telephone: +33-4-91269266; Fax: +33-4-91820682


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

Abstract

Self-renewal and differentiation of stem cell depend on a dynamic interplay of cell-extrinsic and -intrinsic regulators. However, how stem cells perceive the right amount of signal and at the right time to undergo a precise developmental program remains poorly understood. The cell surface proteins Glypicans act as gatekeepers of environmental signals to modulate their perception by target cells. Here, we show that one of these, Glypican4 (Gpc4), is specifically required to maintain the self-renewal potential of mouse embryonic stem cells (ESCs) and to fine tune cell lineage commitment. Notably, Gpc4-mutant ESCs contribute to all embryonic cell lineages when injected in blastocyts but lose their intrinsic tumorigenic properties after implantation into nude mice. Therefore, our molecular and functional studies reveal that Gpc4 maintains distinct stemness features. Moreover, we provide evidence that self-renewal and lineage commitment of different stem cell types is fine tuned by Gpc4 activity by showing that Gpc4 is required for the maintenance of adult neural stem cell fate in vivo. Mechanistically, Gpc4 regulates self-renewal of ESCs by modulating Wnt/β-catenin signaling activities. Thus, our findings establish that Gpc4 acts at the interface of extrinsic and intrinsic signal regulation to fine tune stem cell fate. Moreover, the ability to uncouple pluripotent stem cell differentiation from tumorigenic potential makes Gpc4 as a promising target for cell-based regenerative therapies. Stem Cells2012;30:1863–1874

INTRODUCTION

Over the past years, stem cells (SC) have become an attractive system for research and cell-based regenerative therapies due to their unique ability for self-renewal while maintaining multilineage differentiation potential [1–3]. However, concerns have arisen on whether our current knowledge permits an efficient and safe SC use in tissue engineering strategies. Uncontrolled SC expansion is a potential cause of tissue degeneration and cancer [4]. Conversely, premature SC loss reduces the number of differentiated progeny, thus disrupting tissue structure and function [5]. The identification of new regulators of SC fate and insights into their mechanism of action is expected to improve SC-based therapeutic approaches.

In vivo, SCs reside in specialized microenvironments, or niches, where they are exposed to a variety of signals acting in a spatially and temporally restricted manner. Such signals, which include soluble cytokines, growth factors, and extracellular matrix proteins, converge on the network of transcription factors determining the alternative SC fate choices of self-renewal and differentiation at each cell division [6–8]. Moreover, a continuous crosstalk between cell-extrinsic and -intrinsic regulators results in a robust molecular system that remodels SC fate in response to the environmental changes [6–8]. However, how extracellular signaling cues are properly perceived by SCs to ensure reproducibility of their developmental programs remains poorly understood.

Embryonic stem cells (ESCs) provide a cellular model to investigate the molecular strategies underlying SC fate decisions. For example, studies on mouse ESCs have first revealed that propagation of the self-renewing pluripotent state of ESCs is stimulated by leukemia inhibitory factor (LIF) and bone morphogenetic proteins (Bmps) acting through signal transducer and activator of transcription 3 (Stat3)/Phosphatidylinositol 3-kinases and Smad1/5/8, respectively [9, 10]. LIF and Bmps have subsequently been shown to hold lineage commitment in check by either preserving the transcriptional regulatory network of ESCs [9, 11, 12] or by inducing the expression of inhibitors of differentiation such as the Id proteins [10]. Recent data have also revealed that ESCs may be maintained by suppressing the prodifferentiation mitogen-activated protein kinase (Mapk)/extracellular signal-regulated kinases (Erks) pathway [7]. Mapk/Erks activity is driven mainly by autocrine fibroblast growth factor4 (Fgf4) signaling [7], indicating that maintenance of the ESC state requires shielding them from distinct inductive signals in the environment. Interestingly, Bmps also cooperate with LIF/Stat3 to shield ESCs by the Fgf4/Erks action, thus preventing lineage commitment [7]. Together, these findings indicate that ESC fate can be experimentally determined by modulating levels of opposing self-renewal and differentiation inducing signals.

The canonical Wnt/β-catenin pathway is an intriguing regulator of SC fate for its capability to act at multiple levels and elicit different responses. Canonical Wnt signaling directs mesodermal and endodermal differentiation of ESCs [13] by rendering cells competent to respond to differentiation pathways [14]. On the other hand, Wnt signaling also impacts on pluripotency and self-renewal of mouse and human ESCs [15, 16]. In agreement with this, pharmacological inhibition of glycogen synthase kinase3 (GSK3), a known intracellular regulator of Wnt signaling, promotes ESC self-renewal and inhibits differentiation [17]. It has been recently shown that GSK3 inhibition consolidates ESC pluripotency in part by stabilizing intracellular β-catenin [18]. This is consistent with results showing that downregulating GSK3 or adenomatous polyposis coli (APC) triggers high β-catenin levels and impairs ESC differentiation [19, 20]. However, it is still unclear whether this Wnt signaling role reflects instructive or permissive functions as Wnts synergize with LIF and FGFs in mouse and human ESCs, respectively, to maintain stemness [21, 22].

The overall picture emerging from these studies is that ESC fate choice relies upon distinct signaling networks and their molecular interactions, which establish the signaling thresholds underlying distinct fates and changes. Threshold signaling is a fundamental strategy to promote cell fate specification and tissue patterning during development [23]. In embryos, distinct cell surface proteins such as members of the heparan sulfate proteoglycan (HSPG) superfamily [24] mediate the interaction of cells with extracellular signal peptides. This property allows them to impact on mechanisms such as receptor–ligand interactions, the temporal and quantitative supply of active signals to cells, and the maintenance of morphogen gradients [23–26]. So far, the functional relevance of HSPGs in mouse ESCs has been investigated in a broad manner by targeting enzymes such as N-deacetylase/N-sulfotransferase 1 and 2 (Ndst1/2) and exostosin protein 1 (Ext1), which are required for heparan sulfate chain modifications [27, 28]. These mutant ESC lines fail to develop endothelial and neuronal cell types, respectively [29, 30]. Other studies revealed that mutant cells are incapable of exiting self-renewal [27, 28], thus indicating a more general role for HSPGs in ESC differentiation [31]. However, it remains to be determined whether HSPGs have an impact on ESC self-renewal/pluripotency and which family member is functionally implicated.

Recent studies have revealed that Glypican4 (Gpc4) is among the most abundant HSPG expressed by human and mouse ESCs [32, 33]. In mice, Gpc4 transcripts are abundant in both embryonic and adult neural precursors [34–36]. Reduction of Gpc4 activity in embryos disrupts cell movements during gastrulation [37], dorsal forebrain patterning [38], cartilage and bone morphogenesis [39]. Biochemical and genetic studies have shown that the main function of glypican family members is to regulate signaling of Wnts, sonic hedgehog (Shh), Fgf, and Bmps with remarkable tissue and stage specificity. Depending on the cellular context, glypicans exert either stimulatory or inhibitory activities. For example, Gpc4 enhances Wnt11 signaling during gastrulation [37, 40] and positively modulates Fgfs during dorsal forebrain patterning [38]. The Drosophila Gpc4, Dally-like, shows biphasic activities: as a repressor of short-range Wingless signaling and as an activator of long-range responsiveness [25, 41]. The pleiotropic action of glypicans on different signaling pathways thus becomes restricted as they are used for distinct developmental events [25, 41].

In this study, we have investigated the role of Gpc4 in genetically modified mouse ESCs. We show that Gpc4 is required for the maintenance of self-renewing ESCs in vitro by modulating Wnt/β-catenin signaling. Remarkably, impairment of Gpc4 function in ESCs disrupts teratoma formation even though cells remain pluripotent in chimeric mouse embryos as highlighted by blastocyst injection. We demonstrate that Gpc4 also regulates self-renewal and maintenance of neural SCs in postnatal brains, revealing that Gpc4-mediated stemness maintenance is not restricted to ESCs. Thus, our results uncover new molecular strategies as to how SC numbers and fate can be regulated during developmental and tissue homeostasis. They also provide insights into how the fine tuning of fate decisions in pluripotent SCs can minimize the risk of teratoma formation.

MATERIALS AND METHODS

ESC Cultures and Differentiation

The ESC lines, PST132, Ex194, and XST060, were a gift of Dr. William Skarnes. E14Tg2a have been described previously [42]. All ESC lines were maintained on a monolayer of mitomycin-C-inactivated mouse fibroblasts in presence of Dulbecco's modified Eagle's medium supplemented with 15% fetal bovine serum (FBS), 100 μM 2-mercaptoethanol, minimum essential medium (MEM) nonessential amino acids, 2 mM L-glutamine, 1 mM sodium pyruvate, and 100 units/ml LIF ESGRO (Chemicon, Temecula, CA, http://www.chemicon.com). Neuronal ESC differentiation was done by using a serum-free neural induction protocol [43]. For differentiation into embryoid bodies (EBs), the hanging drop method was used [44]. Differentiating EBs were exposed to recombinant Wnt3A (R&D Systems Inc., Minneapolis, MN, http://www.rndsystems.com), 6-bromoindirubin-30-oxime (BIO; Calbiochem, San Diego, CA, http://www.emdbiosciences.com), CHIR 99021 (Selleck, Houston, TX, http://www.selleckchem.com/), AR A014418 (Sigma, St. Louis, MO, www.sigma-aldrich.com), or LiCl from day0 to day5 at the indicated concentrations.

Teratoma Assays

ESCs were trypsinized into single-cell suspensions and resuspended in phosphate buffered saline (PBS) at 2.5 × 107cells per milliliter concentration. Cells were injected subcutaneously into the hind limbs of severe combined immunodeficiency mice (SCID) using a 25-gauge needle (200 μl). Teratomas were collected after 4 weeks, fixed, embedded, sectioned, and stained as described [45].

Statistical Analysis

Each experiment presented consists of a minimum of two biological replicates unless stated otherwise. Statistical significance was determined using Student's t test. Mann–Whitney U test was used to assess differences between data groups using Instat software (Graph Pad Software). Statistical significance was defined as ns, p > .05; *, p < .05; **, p < .01; ***, p < .001. Error bars represent SEM.

RESULTS

Gpc4 in Mouse ESCs

We studied the functional relevance of Gpc4 in ESCs using a series of genetically modified cells. As Gpc4 is X-linked in men and mice, this analysis was performed in hemizygous male mouse ESCs. By screening the database of the Sanger Institute Gene Trap Resource (SIGTR; [42]), we identified two mouse ESC lines, called PST132 and Ex194, carrying insertion mutations in the Gpc4 gene. Molecular analysis revealed that these two ESC lines harbor single insertion sites within the first Gpc4 gene intron at two different positions (Fig. 1A; Supporting Information Fig. S1). Integration of the trap vector is expected to generate Gpc4 fusion proteins where the first 308 and 286 amino acids of Gpc4 (for PST132 and Ex194, respectively) are fused in-frame to the betageo reporter, thus removing most of the Gpc4 functional motifs (Supporting Information Fig. S1C; sequence tags identified by 5′rapid amplification of cDNA ends (RACE) available from the SIGTR; [25, 41]). Thus, the PST132 and Ex194 ESC lines, referred to as Gpc4gt-1 and Gpc4gt-2 in this study, carry two distinct Gpc4-mutant alleles. This genetic approach was further complemented by the generation of a Gpc4-mutant ESC line where exon 3 was deleted by Cre–Lox-mediated recombination (Fig. 1B; Gpc4flox). Excision of this exon results in the generation of a nonfunctional gene product (Fig. 1B; Gpc4Δex3).

Figure 1.

Mouse embryonic stem cells (ESCs) carrying Gpc4 mutant alleles retain self-renewal properties in vitro. (A): Schematic representation of the Gpc4 alleles with the gene trap insertion. p1, p2, and p3 indicate the position of oligos used for semiquantitative reverse transcriptase polymerase chain reaction (RT-PCR). (B): Schematic representation of conditional Gpc4 mutant allele generated by using the LoxP-Cre system. p4 and p5 indicate oligos used for semiquantitative RT-PCR. (C): Semiquantitative RT-PCRs showing Gpc4 transcript levels differentially downregulated in Gpc4gt-1, Gpc4gt-2, and Gpc4Δex3 mutant versus control (wt or Gpc4flox) ESCs. Gpc4gt-1 line carries a stronger hypomorphic allele than Gpc4gt-2. lacZ gene expression is only detected in Gpc4gt-1 and Gpc4gt-2 ESCs (fourth panel). Gapdh expression (bottom panel) was used as internal control in all RT-PCR studies reported except as otherwise indicated. Primers used are indicated. (D): Western blot analysis showing the full length Gpc4 core protein in Gpc4-mutant and control ESCs. Actin protein levels were used as loading controls in all biochemical studies reported. (E): β-Gal staining of wt, Gpc4gt-1, and Gpc4gt-2 ESCs. Scale bars = 50 μm. (F): Quantitative RT-PCR analysis of total RNA levels of Oct4, Nanog, and Sox2 from ESCs grown in optimal cultured conditions. (G): Scatter plot analysis of genome-wide mRNA expression profile of Gpc4Δex3 versus control (Gpc4flox) ESCs. The solid line across the plot is the reference line for identity. Note that mutant cells show an expression profile resembling that of control ESCs (R2>0.99). Abbreviations: β-Gal, β-galactosidase; ctl, control cells; Gapdh, glyceraldehyde 3-phosphate dehydrogenase; wt, wild type.

We analyzed the levels of wild-type (wt) and Gpc4-mutant transcripts by semiquantitative reverse transcriptase polymerase chain reaction (RT-PCR). As shown in Figure 1C, the Gpc4 gene is expressed in mouse ESCs and mRNA levels are markedly reduced in all Gpc4-mutant ESC lines. In particular, Gpc4Δex3 ESCs only expressed low levels of a truncated gpc4-transcript lacking exon3 (Fig. 1C; compare RT-PCRs with p1+p2, p1+p5, and p4+p5 oligos). Quantitative analysis revealed that the full-length Gpc4 transcript was 70% ± 1% and 50% ± 0.5% reduced in Gpc4gt-1 and Gpc4gt-2 cells, respectively (p < .001). Moreover, Gpc4gt-1 and Gpc4gt-2 ESCs expressed high levels of Gpc4 exon1-LacZ fusion transcript, thus showing that the trap insertion efficiently interrupts proper splicing (Fig. 1C; compare RT-PCR with p1+p3 oligos). Western blot analysis confirmed the reduction of Gpc4 protein levels in Gpc4-mutant ESC lines in comparison to either wt or Gpc4flox (Fig. 1D; Supporting Information Fig. S2A). Moreover, a functional β-galactosidase (β-Gal) fusion protein was expressed in Gpc4gt-1 and Gpc4gt-2 ESC lines (Fig. 1E; Supporting Information Fig. S2A).

In the mouse, the glypican gene family comprises five additional members [25, 41]. We analyzed whether loss of Gpc4 levels altered gene expression levels of the other glypicans. No significant compensatory changes were detected in the transcript levels of Gpc1 to Gpc6, indicating that the mutations described above selectively affect Gpc4 (Supporting Information Fig. S2B). We also found that Gpc4flox ESCs behaved indistinguishable from wt ESCs in all experiments performed. Therefore, we refer them as control cells, although mutant cells were systematically compared with their corresponding controls in all studies.

Gpc4 Loss-of-Function Mutations Accelerate the Differentiation of Mouse ESCs

Gpc4 expression levels in ESCs rapidly decrease during the onset of neural and EB-mediated mesodermal–endodermal differentiation and are differentially upregulated as differentiation progresses (Supporting Information Fig. S2C, S2D). This dynamic expression profile suggests that Gpc4 might participate in specific stages during ESC fate determination. In the present studies, we focused on the functional requirement of Gpc4 in self-renewal and lineage commitment of ESC.

In order to address the function of Gpc4 in stemness, we cultured control and Gpc4-mutant ESCs under growth conditions optimal for self-renewal (with LIF and embryonic fibroblast feeder layers). Gpc4gt-1, Gpc4gt-2, and Gpc4Δex3 cells exhibited morphological and growth properties similar to controls (data not shown). Mutant cells expressed pluripotency markers such as Oct4, Nanog, Sox2, and Rex1 at similar levels (Fig. 1F; Supporting Information Fig. S2E) and lacked differentiation markers (Figs. 2B, 2D, 3A). Moreover, genome-wide mRNA expression profiles demonstrated that Gpc4-mutant ESCs display a gene expression signature highly similar to control cells (Fig. 1G). For example, genes such as Oct4, Nanog, Sox2, Rex1, Fgf5, and Klf4 were expressed at similar levels in Gpc4-mutant and control ESCs (Fig. 1G; Supporting Information Table 2). Thus, Gpc4-mutant cells retain molecular and cellular features of ESCs when grown in self-renewal conditions. Intriguingly, Gpc4gt-1 ESCs grown under suboptimal conditions (in the absence of feeder layer) derepressed early markers of ectoderm, mesoderm, and endoderm differentiation and downregulated Sox2 expression (Supporting Information Fig. S3A).

Figure 2.

Onset and efficiency of EB differentiation depends on GPC4 expression levels. (A): EB differentiation was monitored as numbers of beating foci from day0 to day15 (mean ± SEM; n = 3). (B): Semiquantitative reverse transcriptase polymerase chain reaction (RT-PCR) analysis of cellular markers from total RNA extracted at different time points during EB differentiation of control and Gpc4gt-1 ESCs. (C): Enhanced onset of EB differentiation in Gpc4Δex3 cells versus control (mean ± SEM; n = 2). (D): Semiquantitative RT-PCR analysis of cellular markers similar to panel (B) performed on Gpc4Δex3 and control cells. (E): Time course analysis of beating foci in differentiating EBs from control, Gpc4gt-1, and Gpc4gt-1 mutant cells with restored Gpc4-transcript levels (Gpc4over1 and Gpc4over2; mean ± SEM; n = 3). In all RT-PCR studies, Gapdh expression (bottom panel) was used to normalize total RNA levels between the different samples of each experiment. (F): Semiquantitative RT-PCR analysis of total RNA levels from EBs at day7 in control, Gpc4gt-1, Gpc4over1, and Gpc4over2 cells showing expression levels of Oct4, α-Mhc, and Gpc4. Note that Gpc4over1 and Gpc4over2 cells show beating foci profiles and expression marker levels similar to control. Abbreviations: ctl: control cells; EBs, embryoid bodies; ESC, embryonic stem cells.

Figure 3.

Gpc4 levels regulate neural lineage entry. (A): Immunofluorescence showing the distribution of Oct4 and Nestin-positive cells in neural differentiating control and Gpc4gt-1 embryonic stem cells (ESCs) at different time points (×40 magnification. Scale bars = 100 μm). (B): Quantitative analysis of Oct4 and Nestin-positive cells in neural differentiating control and Gpcgt-1 cells (mean ± SEM; n = 2; ***, p < .001; **, p < .01). (C): Semiquantitative reverse transcriptase polymerase chain reaction (RT-PCR) analysis of total RNA extracted from neural differentiating control and Gpc4gt-1 cells at different time points. (D): Immunofluorescence of Tuj1 (×40 magnification; scale bars = 100 μm), GABA, and Calretinin-positive cells (×10 magnification; scale bars = 200 μm) in control and Gpc4gt-1 cells. (E): Quantitative analysis of phospho-histone H3-positive cells in neural differentiating control and Gpc4gt-1 cells at different time points (mean ± SEM; n = 2; ***, p < .001; **, p < .01). (F): Quantitative analysis of cleaved-caspase3-positive cells in neural differentiating wt and Gpc4gt-1 cells at different time points (mean ± SEM; n = 2; ***, p < .001; **, p < .01). Abbreviations: ctl, control cells; DAPI, 4′,6-diamidino-2-phenylindole; ESC, embryonic stem cell; GABA, γ-aminobutyric acid (GABA).

As Gpc4-mutant cells appear sensitive to changes in self-renewal promoting signals, we assessed the role of Gpc4 during in vitro ESC differentiation. We first triggered the fate of Gpc4-mutant and control ESCs toward mesoderm–endoderm using EBs. EBs differentiate into different specialized cell types including cardiomyocytes, identifiable by the presence of contracting foci [44]. The size of Gpc4-mutant and control EBs was similar after 5 days in culture (Gpc4-mutant: 14,625 ± 625 cells; control: 12,825 ± 325 cells; p > .05; Supporting Information Fig. S3B). Moreover, no major changes were observed in the spatial and temporal distribution of the basement membrane protein, Laminin1 (Supporting Information Fig. S3B). In contrast, the differentiation kinetics of Gpc4-mutant EBs was drastically different. Whereas beating EBs first appeared in control cultures at day7 (3% ± 0.3%) and numbers increased up to 65% ± 1% at day15 (Fig. 2A, 2C), a prominent incidence of Gpc4gt-1 and Gpc4Δex3 beating EBs was detected at day6 (Gpc4gt-1: 80% ± 1%; Gpc4Δex3: 65% ± 1.8%) and reached 90% by day7 (Fig. 2A, 2C). These findings correlated with extended areas of Gpc4-mutant EBs expressing the sarcomeric protein α-Actinin (Supporting Information Fig. S3C). Interestingly, the number of beating EBs in Gpc4gt-2 cells was intermediate to controls and Gpc4gt-1 or Gpc4Δex3 cells (Fig. 2A, 2C), suggesting that Gpc4 might regulate ESC fate in a dose-dependent manner.

This EB functional analysis was corroborated by molecular marker studies. While the expression of pluripotency markers as Oct4 and Cripto was reduced in Gpc4-mutant EBs at day5, these changes were only apparent at later time points in control cells (Fig. 2B, 2D). Consistently, markers for the early mesoderm (Brachyury), cadiomyocytes (Nkx2.5 and α-Mhc), and endothelial cells (CD31) were prematurely upregulated in mutant cells (Fig. 2B, 2D; Supporting Information Fig. S3C). Endodermal differentiation (e.g., Gata4) also occurred in control and mutant EBs, but no drastic expression profile changes were observed in differentiated cells such as hepatocytes (Tdo marker; Fig. 2B, 2D). Moreover, restoring Gpc4 expression in Gpc4gt-1 ESCs to levels similar to those in control cells (Gpc4over1 and Gpc4over2 cells) rescued Gpc4gt-1 cell behavior and molecular phenotypes (Fig. 2E, 2F). Thus, rescue experiments together with the use of distinct alleles (either resulting from gene trap or conditional mutagenesis) reveal that the loss-of-function phenotypes observed in Gpc4-mutant ESCs are not caused by clonal variation or genetic background. Together, these findings show that Gpc4 is required to regulate cell lineage entry.

Gpc4 Regulates the Onset of ESC Differentiation Toward Neural Lineages

To understand whether Gpc4 acts in a more general fashion during lineage commitment, we examined neural cell lineage differentiation of Gpc4-mutant ESCs using established protocols [43]. Differentiating Gpc4-mutant cells revealed a 39% reduction of Oct4-positive cells at day4 in line with EB results (Fig. 3A, 3B, 3C; Supporting Information Fig. S3D). This correlated with a 31% increase of nestin-positive neural precursors by day4 in contrast to controls (Fig. 3A, 3B, 3C; Supporting Information Fig. S3D). As evidenced by following the neuronal markers Tuj1, Map-2, γ-aminobutyric acid (GABA), and Calretinin, Gpc4-mutant ESCs also developed a large number of differentiated neurons at early time points when compared to controls (Fig. 3C, 3D; Supporting Information Fig. S3D, S3E).

The increased numbers of neural derivatives of Gpc4-mutant ESCs could arise either from their precocious differentiation or from an increased proliferation and/or survival. Quantitative analysis of phopho-histone H3 and activated caspase-3 positive cells did not reveal significant changes when compared to controls (Fig. 3E, 3F). Taken together, these results indicate that Gpc4 controls the balance of ESC self-renewal and differentiation during the onset of lineage entry, independently of the differentiation pathway.

Pluripotency Is Maintained by Gpc4-Mutant ESCs in Developing Embryos

The studies above report that the enhanced differentiation potential of Gpc4-mutant ESCs in vitro does not interfere with their ability to differentiate into cell types belonging to all three germ layers. To assess whether these cells also maintain pluripotency features in vivo, we generated chimeric mouse embryos by injecting them into wt blastocysts [1]. Embryos were analyzed at different developmental stages as inhibition of Gpc4 functions alters gastrulation and brain patterning in Xenopus embryos [38]. For these studies, we took advantage of Gpc4gt-1 cells expressing a functional β-Gal transgene. This allows visualization of the distribution of Gpc4 expressing mutant cells in developing embryos, although with a pattern restricted to those positive for Gpc4 expression. The percentage of β-Gal positive embryos was 85.7% at embryonic days E6.75 and E7.5 (12/14 embryos), 72% at E9.5 (13/18 embryos), and 62.5% at E13.5 (10/16 embryos). Whole mount β-Gal staining and tissue sections revealed a significant contribution of Gpc4-mutant ESCs to tissues derived from all three germ layers (Fig. 4). This included the neuroepithelium, head mesenchyme, developing heart, muscle, lung, and gut (Fig. 4D–4G). The degree of chimerism, as estimated by whole mount β-Gal expression, decreased at later embryonic stages (approximately 50% reduction between E6.75/E7.5 and E9.5). Moreover, a fraction of embryos with high levels of chimerism showed gross morphological defects (Fig. 4C). Consistently, the β-Gal distribution recapitulates the expression profile of Gpc4 mRNA at the corresponding developmental stages (Fig. 4A–4C; Supporting Information Fig. S4, [36]). Together, these results show that Gpc4-mutant ESCs are pluripotent when exposed to a blastocyst environment.

Figure 4.

Gpc4-mutant cells retain the ability to contribute to the development of embryonic tissues. (A): β-Gal staining of mouse chimeric embryos obtained by injecting Gpc4gt-1 embryonic stem cells (ESCs) into blastocysts. β-Gal distribution in embryos at E7.5. Scale bars = 200 μm. (B, C): β-Gal distribution in embryos at E9.5. Panel (C) shows a mouse embryo with tail bud defects (*). Scale bars = 500 μm. (D): Transverse section of the embryo in (B) at the level of the forebrain showing β-Gal activity in the developing neuroepithelium (white arrows) and head mesenchyme (arrowheads). Scale bars = 200 μm. (E): Transverse section of the embryo in (B) at the level of the heart region. Arrows point to cardiac muscle. (F, G): Transverse sections of a chimeric embryo at E13.5 showing β-Gal activity in liver (F) and gut (G). Scale bars = 200 μm. Abbreviations: a: atrial chamber; β-Gal, β-galactosidase; fb, forebrain; g, gut; hb, hindbrain; l, liver; mb, midbrain; ov: otic vesicle; s, somites; sc: spinal cord; v: ventricle.

Lack of Gpc4 Function Discharges ESCs from Developing Teratomas

The pluripotency properties of Gpc4-mutant ESCs together with their intriguing feature to rapidly lose stemness under differentiation conditions suggest that they are sensitive to the environmental stimuli. To evaluate this possibility, we transplanted Gpc4-mutant ESCs into SCID mice and assessed their ability to develop teratoma [1]. As expected, mice injected with control ESCs developed large teratomas by 4 weeks (Fig. 5A, 5C). The average weight of teratomas was 4.5 ± 1.9 g (Fig. 5B), and they were composed of tissues derived from all three germ layers (Fig. 5D, 5E). Strikingly, injection of Gpc4gt-1 ESCs led to the formation of only small tissue masses (Fig. 5A, 5F) with an average weight of 0.8 ± 0.4 g (Fig. 5B). Histological analysis revealed the presence of β-Gal expressing Gpc4gt-1 cells but the absence of germ layer-derived structures (Fig. 5G, 5H). Consistently, injection of Gpc4Δex3 cells did not lead to teratoma formation, and only tissue masses were found (Supporting Information Fig. S5A). To exclude the possibility of off-target effects, we performed rescue experiments by injecting Gpc4over1 cells (used for studies shown in Fig. 2E, 2F) in into SCID mice. After 3 weeks of injection, Gpc4over1 cells mice developed teratoma tissues significantly similar to those obtained with control ESCs as estimated from average tissue weight (control: 1 ± 0.32 g; Gpc4over1: 1.1 ± 0.36 g) and histological analysis (Supporting Information Fig. S5B). Thus, genetic downregulation of Gpc4 functions in ESCs uncouples distinct stemness properties: while preserving self-renewal and pluripotency of ESC, tumorigenesis is suppressed.

Figure 5.

Reduced Gpc4 levels in mouse embryonic stem cells (ESCs) prevent teratoma formation. (A): A severe combined immunodeficiency mouse subcutaneously injected with 5 × 106 controls and Gpc4gt-1 ESCs. Arrows point to tissue masses developed after 4 weeks. (B): Quantitative analysis of tissue weights after dissection of ectopic masses developed in mice injected as in (A). Results represent two independent experiments (mean ± SEM; n = 10; ***, p < .001). (C): Representative image of dissected ectopic tissues derived from injected control ESCs. (D, E): Hematoxylin/eosin staining of paraffin sections (×10; scale bars = 200 μm) from tissue in (C) showing structures as mucus-producing epithelium [endoderm; white arrowheads in (D)], neuroepithelium [ectoderm; white arrows in (D)], and cartilage/bone [mesoderm; white arrows in (E)]. (F): Representative image of dissected ectopic tissues derived from injected Gpc4gt-1 ESCs. (G, H): Hematoxylin/eosin staining of paraffin sections from tissue in (F) showing that Gpc4-mutant cells do not develop teratomas (×10). Arrows point to distribution of injected mutant cells as revealed by β-Gal staining. Abbreviation: ctl: control cells.

Gpc4 Regulates the Response of Mouse ESCs to Wnt Signaling

The fate choice between self-renewal and lineage commitment of ESC is regulated by the intercalation of distinct autocrine and paracrine signals. LIF, Bmp4, and Wnt3a maintain ESC self-renewal whereas differentiation is triggered by autoinductive stimulation of Erks pathway by Fgf4 [6–8]. As Glypicans function as modulators of these signals, genetically reducing Gpc4 might result in: (a) impaired sensitivity to signals controlling self-renewal or (b) enhanced response to differentiation stimuli. To discriminate between these two possibilities, we investigated the activation of distinct intercellular signaling pathways in Gpc4 loss-of-function background.

We first evaluated Wnt/β-catenin signaling activity by transfecting cells with the T-cell factor (TCF)-β-catenin reporter construct TOPFlash, containing TCF binding sites [46] and by using luciferase activity as readout of Wnt/β-catenin activation following Wnt3a stimulation. Gpc4-mutant ESCs displayed reduced activation of the TOPFlash reporter even after treatment with high Wnt3a levels (Fig. 6A). Similar expression levels of Wnt receptors, Frizzleds, were found in control and mutant ESCs (Fig. 6B), indicating that ESCs require Gpc4 to perceive properly the incoming Wnt signal. In contrast, stimulation of control and Gpc4-mutant ESCs with either LIF or Bmp4 resulted in similar increases of Stat3 and Smad1/5/8 phosphorylation, respectively (Fig. 6C). Moreover, neural differentiation of both control and Gpc4-mutant ESCs was inhibited when Bmp4 was added to the neural differentiation media (Supporting Information Fig. S5C; [13]). Finally, Erks were phosphorylated in both controls and Gpc4-mutant ESCs upon Fgf stimulation, although to a slightly reduced extent in mutant cells (Supporting Information Fig. S5D). Therefore, Gpc4 appears to modulate the ESC sensitivity to signals controlling self-renewal (Wnt, but not Bmp and LIF), rather than tuning down differentiation commitment stimuli (such as Fgfs).

Figure 6.

Gpc4 regulates embryonic stem cell (ESC) response to Wnt/β-catenin signaling. (A): TOPflash reporter activation in control and Gpc4-mutant ESCs following Wnt3a stimulation at indicated concentrations. The lack of FOPFlash activation, containing mutated T-cell factor (TCF)-binding sites, demonstrated that the activity was specific for TCF-β-catenin complexes (mean ± SEM; n = 2). (B): Semiquantitative reverse transcriptase polymerase chain reaction (RT-PCR) analysis showing no major changes in the expression levels of Frizzled (Fzd) family members in total RNA from Gpc4gt-1, Gpc4gt-2, and Gpc4Δex3 mutant versus control (wt or Gpc4flox) ESCs. (C): Western blot analysis of p-Stat3 and p-Smad1/5/8 protein levels in control and Gpc4gt-1 cells stimulated with increasing concentration of LIF and Bmp4 (indicated). (D): Quantitative analysis of beating EBs from control, Gpc4gt-1, and Gpc4gt-1 cells exposed at the GSK3 inhibitor, BIO, from day0 to day5. BIO concentrations are indicated (mean ± SEM; n = 3). (E): Semiquantitative RT-PCR analysis of total RNA extracted from differentiating EBs at day10 following BIO treatment at the indicated concentrations. Abbreviations: BIO, 6-bromoindirubin-30-oxime; ctl: control cells; EB, embryoid body; FOPFlash, mutant TCF reporter plasmid; fzd, Frizzled; LIF, leukemia inhibitory factor; TOPflash, TCF reporter plasmid; wt, wild type.

In line with results showing impaired Wnt signal response, treating differentiating EBs for 5 days with increased Wnt3a doses did not rescue the differentiation phenotype of Gpc4-mutant ESCs, while maintaining control EBs in an undifferentiated state (Supporting Information Fig. S6A, S6B, S6G). These results prompted us to examine whether intracellular activation of the canonical Wnt/β-catenin pathway was sufficient to rescue Gpc4-mutant ESCs. Among other functions, GSK3 acts as an inhibitor of Wnt signal transduction. Therefore, we analyzed the functional consequence of blocking GSK3 during EB differentiation and quantified the numbers of beating foci. GSK3 was transiently blocked by exposing developing EBs to BIO for 5 days [17]. All experiments were performed using doses shown to be effective and specific for GSK3 inhibition [47]. In the presence of BIO, Gpc4-mutant ESCs acquired differentiation features of control ESCs from approximately day8 onward (Fig. 6D). Notably, the observed phenotypic rescue of EBs exposed to BIO correlates with an increased activation of the TOPFlash reporter in both control and Gpc4-mutant ESCs (Supporting Information Fig. S6C). Therefore, the functional rescue mediated by GSK3 inhibition can be in part explained by reactivation of the intracellular Wnt/β-catenin pathway. Similarly, inhibition of GSK3 by LiCl [48], by CHIR 99021 [7], or by AR A014418 (a GSK3 inhibitor, which lacks side effects as impairment of cyclin functions [49]), restored the differentiation potential of Gpc4-mutant ESCs to a similar extent (Supporting Information Fig. S6D, S6E, S6F). Consistently, molecular analysis of Oct4 and α-Mhc expression in day10 Gpc4-mutant EBs exposed to GSK3 inhibitors showed that a dose-dependent increase of Oct4 transcript levels correlates with a decrease in α-Mhc expression levels (Fig. 6E; Supporting Information Fig. S6H, S6I). In conclusion, GSK3 inhibition leads, among other possible effects such as those on cell metabolism, to the intracellular activation of Wnt/β-catenin signaling [15, 16], which alleviates the premature exit from self-renewal in Gpc4-mutant ESCs.

Gpc4 Controls Maintenance of Adult Neural Stem Cells

The functional implication of Gpc4 in maintenance of ESCs prompted us to assess whether Gpc4 is also implicated in controlling fate of other SC types. We found that Gpc4 is expressed in the neurogenic areas of newborn and adult mouse brains where the neural SCs reside (Supporting Information Fig. S7A; [49, 50]). Postnatal olfactory bulb (OB) neurons originate from radial glial neural SCs (type B cells) located in the subventricular zone (SVZ; [51]). Radial glial neural SCs divide slowly to generate fast-replicating transit-amplifying precursors (type C cells: positive for Mash1), which in turn give rise to neuroblasts expressing neuronal markers as Doublecortin (DCX). Neuroblasts migrate through the rostral migratory stream (RMS) to the OB, where they differentiate into local interneurons [49]. We have previously described a RNA interference strategy based on in vivo electroporation to analyze gene functions during OB neurogenesis in postnatal brains [52, 53]. This approach was used to study the potential role of Gpc4 in neural SCs. We first validated the activity of different Gpc4-specific short hairpin RNA (shRNA) plasmids by coexpressing them with a mouse Myc-epitope tagged Gpc4 expression constructs in NIH3T3 cells (Supporting Information Fig. S7B, S7C).

The Gpc4 shRNA or control vectors were coelectroporated with a green fluorescent protein (GFP) expression plasmid, which allowed detection of transfected cells. Electroporation was done into the lateral ventricular wall of 1-day-old pups (P1), and effects were analyzed at P7. We examined the global distribution of transfected cells in the ventricular zone (VZ), SVZ, and RMS assessing GFP expression. At P7 many GFP-positive cells had entered the RMS in both control and Gpc4 shRNA expressing pups (Fig. 7). In the SVZ, however, there was a striking difference in the proportion of the different GFP-positive cell types. In particular, the fraction of GFP-positive neural SCs (identified by their apical and basal processes contacting the ventricle and pial surface, respectively) was drastically reduced in Gpc4 shRNA pups (control: 26.5% ± 3.9%, Gpc4 shRNA: 6% ± 0.5%; percentage over total GFP-positive cells; Fig. 7B, 7C, 7E; Supporting Information Fig. S7D). This neural SCs reduction was directly proportional to the extent of Gpc4 downregulation observed with different shRNA constructs (Supporting Information Fig. S7C, S7D). These findings indicate that Gpc4 maintains the pool of neural SCs in vivo.

Figure 7.

Gpc4 is crucial for maintenance of neural stem cells in vivo. (A): GFP-positive cells in the SVZ of mouse brain at P7 coelectroporated at P1 with short hairpin RNA (shRNA) control and a GFP expression plasmid. Squares indicate enlargements shown in (B) and (C). Scale bars = 200 μm. (B, C): Enlargements of the SVZ in (A) showing neuroblasts (arrow in B) and radial glia cells (asterisks in C). Scale bars = 50 μm. (D): GFP-positive cells in the SVZ of mouse brain at P7 coelectroporated at P1 with shRNA Gpc4 and a GFP expression plasmid. (E): Enlargements of (D) showing radial glia cells (asterisk), C cells (arrowheads), and neuroblasts (arrows). (F, G): Neuroblast distribution in the RMS of P7 mice electroporated with the shRNA control (F) and shRNA Gpc4 (G). Scale bars = 100 μm. (H, I, J): Immunofluorescence analysis of Mash1-positive cells in the SVZ of a mouse brain at P7 coelectroporated as in (A). Scale bars = 20 μm. (H): GFP expression in radial glia cells (asterisks), C cells (arrowheads), and neuroblasts (arrows). (I): Mash1 expression in C cells (arrowheads) but not in neuroblasts (arrows) and radial glia cells (asterisks). (J): Merged image. (L, M, N): Immunofluorescence analysis of Mash1-positive cells in the SVZ of a mouse brain at P7 coelectroporated as in (D). Scale bars = 20 μm. (L): GFP expression in neuroblasts (arrows) and C cells (arrowheads) forming often clusters. (M): Mash1 expression in C cells (arrowheads) but not in neuroblasts (arrows) and radial glia cells (asterisks). (N): Merged image. (O): Quantitative analysis of Mash1-negative/DCX-negative, Mash1-positive/DCX-negative, and Mash1-negative/DCX-positive cells in electroporated brains. Cells were counted and reported as percentage of Mash1-positive and DCX-positive cells over the total number of GFP expressing cells. Abbreviations: Ctl, control; DCX, Doublecortin; GFP, green fluorescent protein; RMS, rostral migratory stream; SVZ, subventricular zone.

To follow the fate of neural SCs with impaired Gpc4 activity, we determined the proportion of Mash1-positive transit-amplifying precursors and DCX-positive migratory neuroblasts in the SVZ. As shown in Figure 7O, the fraction of GFP-positive cells expressing DCX was similar in controls and Gpc4 shRNA expressing brains (control: 64.7% ± 2.3%; shRNA: 65.6% ± 2.2%). In contrast, Mash1-positive cells were significantly increased in Gpc4 shRNA brains (control: 8.8% ± 1.6%, shRNA: 28.4% ± 2.2%; Fig. 7O). Moreover, Mash1-positive cells formed clusters and expressed higher GFP levels in brains electroporated with Gpc4 shRNA (Fig. 7H-7N). Taken together, the downregulation of Gpc4 in neural SCs shifts the balance from maintenance of SC pool toward intermediate daughter cells in vivo. Gpc4 is also expressed in other regions where SCs reside such as in the hippocampus and in the small crypts of the intestine [34, 54]. Therefore, it is likely that modulation of Gpc4 activity fine tunes self-renewal and lineage commitment of different SC types in vivo.

DISCUSSION

Finely tuned environmental signals coordinate the alternative fates of self-renewal and differentiation in SCs. Understanding the mechanisms involved in precise orchestration of signals regulating SC fate has a broad impact in biomedical science ranging from elucidating the causes of diseases to the use of SCs in regenerative medicine. Our study identifies the cell surface protein, Gpc4, as a new component of the signaling machinery regulating ESC maintenance. Our findings also highlight three key features of Gpc4-mediated SC regulation. First, Gpc4 modulates the response of ESCs to Wnt ligands, resulting in the activation of TCF/β-catenin signaling. Second, Gpc4-mutant cells retain distinct ESC properties in vivo: they are competent to integrate into all three germ layers when implanted in mouse embryos but are dispensed from developing teratomas when injected in nude mice. Third, Gpc4-mediated maintenance of stemness is not restricted to pluripotent SC types as it is also involved in maintaining multipotent neural SCs in postnatal brains. Thus ESCs and neural SCs both require Gpc4 to avoid early depletion of progenitors despite differences in self-renewal timing mechanisms.

The main function of glypican family members is to regulate the action of signals, such as Wnts, Shh, Fgfs, and Bmps. Despite this apparent promiscuous interaction, signaling modulation occurs with surprising tissue and stage specificity. Our previous studies established that Gpc4 enhances Fgf signaling during early brain development [38]. Here, we show that in ESCs Gpc4-mediated cell fate is predominantly restricted to the positive modulation of canonical Wnt signaling. In particular, Gpc4 ensures full activation of Wnt pathway, which in turn enables maintenance of the SC state. Thus, our findings demonstrate that Gpc4 acts as an extracellular sensor to modulate the sensitivity of SCs to incoming signals, thereby defining the signaling thresholds underlying distinct fate changes.

Previous functional analysis established that Wnt signaling acts on ESCs at least at two levels: (a) it maintains stemness under permissive conditions [15–20] and (b) it promotes mesoderm and cardiomyocyte lineage commitment while preventing neuronal differentiation [13, 14]. Our studies show that regulation of canonical Wnt by Gpc4 does not alter the cell lineage commitment as Gpc4-mutant cells retain the ability to contribute to all three germ layers. Instead, they rapidly lose stemness properties when depleted of self-renewal maintaining signals and exposed to differentiation promoting conditions. The premature onset of lineage commitment and differentiation in Gpc4-mutant ESCs arises most likely as a direct consequence of reduced Wnt signal reception. It has been shown that activation of Fgf4/Erks signaling in ESCs is required to trigger lineage commitment [7]. Moreover, ESC lines deficient in Ndst1/2 and Ext1 enzymes, which are required for heparan sulfate chain modifications, fail to initiate differentiation in vitro because of an attenuated response to Fgfs [27, 28]. In contrast, Gpc4-mutant ESCs retain basal Erk levels similar to controls, although respond to Fgf signaling with slightly reduced sensitivity. These findings indicate that loss-of-Gpc4 activity does not act on Fgf to prime ESCs toward differentiation.

Although we cannot exclude that Gpc4 also modulates other signaling pathways, intracellular enhancement of Wnt signal transduction by inhibiting GSK3 alleviates the premature in vitro differentiation phenotype of mutant cells. Even though the canonical Wnt pathway maintains pluripotency, this property appears to require the synergistic action of other signaling pathways such as LIF [21, 55]. Such combinatorial action of different signaling networks provides an additional level to safeguard stemness [56]. In agreement, Gpc4-mutant ESCs with impaired Wnt signal transduction retain self-renewal potential when cultured in the presence of LIF and feeder layers and when injected into mouse blastocysts. In contrast, removal of self-renewal signals renders mutant cells more prone to differentiation. Therefore, our findings support the possibility that the role of Wnt signaling in self-renewal is permissive rather than instructive. As Wnt signaling is also required for self-renewal and maintenance of pluripotency in human ESCs [17, 57], it is likely that also in this context these properties are ensured by Gpc4-mediated regulation.

Recent breakthroughs in SC biology, especially the establishment of induced pluripotent stem (iPS) cells, have generated great enthusiasm for their therapeutic potential in regenerative medicine [58]. However, the risk of teratoma development after transplantation is a serious concern and may hamper clinical applications [59]. Modulating the activity of genes such as Cripto, known to influence Nodal pathway in ESCs, interferes with teratoma development [60]. Moreover, genetic manipulation of APC and GSK3, which deregulate Wnt/β-catenin signal by imposing a constitutive high levels of β-catenin, also impairs the teratoma potential of ESCs [19, 20]. However, acting on these three genes causes either upregulation of stemness markers or impaired ESCs differentiation into all three embryonic cell lineages. We show that attenuation of Wnt signal transduction by genetic reduction of Gpc4 disrupts the intrinsic potential of teratoma development by ESCs without interfering with cell pluripotency. Indeed, preliminary studies show the ability of Gpc4-mutant ESCs to generate neuronal cell types when implanted in adult rat brains, without developing teratoma or tissue masses (unpublished results). Thus, our studies suggest that fine tuning Wnt activity through Gpc4 may provide a tool to uncouple ESC tumorigenic features from their developmental potential in vivo. As Gpc4 is a cell surface protein rather than an intracellular component, it may become an attractive target to design “drugable” strategies that enable the manipulation of SC fate and the impairment of teratoma potential of iPS and ESCs.

Acknowledgements

We thank E. Arenas, K. Dudley, P. Durbec, R. Kelly, G. Minchiotti, and R. Zeller for critically reading the manuscript; S. DeFalco, A. Moumen, and lab members for helpful discussions; G. Luxardi for his contribution at the initial stage of the project; the IBDML imaging service, R. Taté and S. Arbucci of the IGB-CNR for technical support; W.C. Skarnes for providing the Gpc4-mutant ESC lines; Institut Clinique de la Souris (Strasburg, France) for generating Gpc4flox ESCs. This research was supported by the Marie Curie Host Grant for TOK (MTKD-CT-2004-509804) by the FRM (no. DLC20060206414) and by AFM (no. 13683) to R.D. and F.M. H.C. was supported by grants from the ANR (MNP FORDOPA) and the FRM (Label Equipe FRM). A.F. was supported by FIRB-Mertit-N-RBNE08LN4P-002. A.F. is currently affiliated with Institute of Genetics and Biophysics “Adriano Buzzati Traverso,” CNR, 80131 Naples, Italy.

Disclosure of Potential Conflicts of Interest

The authors indicate no potential conflicts of interest.

Ancillary