Advertisement

Expression of Nodal, Lefty-A, and Lefty-B in Undifferentiated Human Embryonic Stem Cells Requires Activation of Smad2/3*

  • Daniel Besser
    Correspondence
    To whom correspondence should be addressed: Max-Delbrück-Center, Robert-Rössle-Straβe 10, D-13125, Berlin, Germany. Tel.: +49-30-9406-3776; Fax: +49-30-9406-2656;
    Affiliations
    Laboratory of Molecular Vertebrate Embryology, The Rockefeller University, New York, New York 10021
    Search for articles by this author
  • Author Footnotes
    * This work was supported by a grant from The Rockefeller University. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
    The on-line version of this article (available at http://www.jbc.org) contains Supplementary Data.
Open AccessPublished:August 11, 2004DOI:https://doi.org/10.1074/jbc.M404979200
      Human embryonic stem cells will remain undifferentiated or undergo differentiation when grown in conditioned or non-conditioned medium, respectively. The factors and signaling events that control the maintenance of the undifferentiated state are not well characterized and their identification is of major importance. Based on the data from global expression analyses, we set out to identify genes and the signaling pathways controlling them that are regulated in the early phase of the differentiation process. This study shows that nodal and the inhibitors of Nodal signaling, lefty-A and lefty-B, are down-regulated very early upon differentiation. High expression of these genes in undifferentiated cells is maintained by activation of the transcription factor Smad2/3, downstream of the activin-linked kinases (ALK) 4/5/7. Treatment of differentiating cells with Activin A leads to activation of Smad2/3 and expression of nodal, lefty-A and lefty-B, while inhibition of ALK4/5/7 by the kinase inhibitor SB-431542 blocks activation of Smad2/3 and expression of these genes in the undifferentiated state. In addition, when cells are maintained undifferentiated by treatment with the GSK3-inhibitor, BIO, high expression of nodal, lefty-A, and lefty-B also requires activation of ALK4/5/7. Conversely, BMP signaling leading to Smad1/5/8 activation via ALK2/3/6 is blocked in undifferentiated cells and becomes activated upon differentiation. Taken together, these observations establish that Smad2/3 is activated in undifferentiated hESCs and required for the expression of genes controlling Nodal signaling. Moreover, there appears to be cross-talk between inhibition of GSK3, a hallmark of Wnt signaling and the Activin/Nodal pathway.
      Embryonic stem cells (ESCs)
      The abbreviations used are: ESCs, embryonic stem cells; TGFβ, transforming growth factor β; ALK, activin-linked kinase; BMP, bone morphogenic protein; GDF, growth differentiation factor; TDGF, teratoma-derived growth factor; LIF, leukemia inhibitory factor; SB, SB-431542; BIO, 6-bromoindirubin-3′oxime; CM, conditioned medium; non-CM, non-conditioned medium; MEFs, mouse embryo fibroblasts; mESCs, mouse ESCs; hESCs, human ESCs; HSP, homology sequence pair; RT, reverse transcriptase; ca, constitutively activated.
      1The abbreviations used are: ESCs, embryonic stem cells; TGFβ, transforming growth factor β; ALK, activin-linked kinase; BMP, bone morphogenic protein; GDF, growth differentiation factor; TDGF, teratoma-derived growth factor; LIF, leukemia inhibitory factor; SB, SB-431542; BIO, 6-bromoindirubin-3′oxime; CM, conditioned medium; non-CM, non-conditioned medium; MEFs, mouse embryo fibroblasts; mESCs, mouse ESCs; hESCs, human ESCs; HSP, homology sequence pair; RT, reverse transcriptase; ca, constitutively activated.
      are pluripotent cells, which can proliferate indefinitely and contribute to the formation of basically all cell types in vitro and in vivo. Over the last decade the study of mammalian embryonic stem cells, especially mouse ESCs, has provided valuable insights into early embryogenesis in mammals. More recently, researchers have started to analyze human ESCs (hESCs) allowing a glimpse into early human embryogenesis and the development of tools for regenerative medicine. While our understanding of the basic biology of mouse ESCs (mESCs) and their in vivo properties is quite advanced, we know very little about the biological properties of hESC. Studying human cells is of special importance since recent molecular studies have shown that their properties differ from mouse cells (
      • Sato N.
      • Meijer L.
      • Skaltsounis L.
      • Greengard P.
      • Brivanlou A.H.
      ,
      • Reubinoff B.E.
      • Pera M.F.
      • Fong C.Y.
      • Trounson A.
      • Bongso A.
      ,
      • Smith A.G.
      ). These findings are not unexpected given the differences in early embryonic development between mice and humans (
      • Ko M.S.
      ).
      One of the foremost questions in the study of hESCs is the identification of factors required for maintenance of the undifferentiated state. hESCs, unlike mESCs (reviewed in Ref.
      • Smith A.G.
      ), cannot be maintained undifferentiated with leukemia inhibitory factor (LIF, Refs.
      • Sato N.
      • Meijer L.
      • Skaltsounis L.
      • Greengard P.
      • Brivanlou A.H.
      and
      • Reubinoff B.E.
      • Pera M.F.
      • Fong C.Y.
      • Trounson A.
      • Bongso A.
      ). However, both human and mouse ESCs stay undifferentiated on feeder layers of mouse embryo fibroblasts (MEFs) or in the presence of MEF-conditioned medium (CM), suggesting that MEFs produce soluble factors that are sufficient to promote the undifferentiated state. Embryonic stem cells have the potential to differentiate into cell types of all different lineages, i.e. ecto-, endo-, and mesodermal and extra-embryonic and germ cell lineages (reviewed in Ref.
      • Smith A.G.
      ). Complex signaling networks control these differentiation processes in vivo (reviewed in Ref.
      • Loebel D.A.
      • Watson C.M.
      • De Young R.A.
      • Tam P.P.
      ).
      A recent study reported global transcriptional profiling in which the entire repertoire of transcripts in mouse (
      • Ramalho-Santos M.
      • Yoon S.
      • Matsuzaki Y.
      • Mulligan R.C.
      • Melton D.A.
      ) and hESCs, grown both under pluripotent and differentiating conditions were compared (
      • Sato N.
      • Sanjuan I.M.
      • Heke M.
      • Uchida M.
      • Naef F.
      • Brivanlou A.H.
      ). Interestingly, 24% overlap was observed between the two sets. The genes that are down-regulated in both mouse and human ESCs, include members of the transforming growth factor β (TGFβ) signaling pathway, such as the teratoma-derived growth factor-1 (TDGF-1/Cripto), which is an EGF-CFC co-receptor for Nodal signaling, as well as lefty-A (EBAF, lefty2 in the mouse) and lefty-B (lefty1 in the mouse), which are inhibitors of Nodal signaling. It has also been shown that nodal itself is highly expressed in undifferentiated hESCs and mESCs (
      • Ramalho-Santos M.
      • Yoon S.
      • Matsuzaki Y.
      • Mulligan R.C.
      • Melton D.A.
      ,
      • Rosler E.S.
      • Fisk G.J.
      • Ares X.
      • Irving J.
      • Miura T.
      • Rao M.S.
      • Carpenter M.K.
      ,
      • Zeng X.
      • Miura T.
      • Luo Y.
      • Bhattacharya B.
      • Condie B.
      • Chen J.
      • Ginis I.
      • Lyons I.
      • Mejido J.
      • Puri R.K.
      • Rao M.S.
      • Freed W.J.
      ).
      TGFβ signaling is involved in a variety of cell fate decisions in early embryogenesis (reviewed in Ref.
      • Harland R.
      • Gerhart J.
      ). The TGFβ superfamily of ligands signals through two main branches: the BMP branch activates Activin-linked kinases (ALK) 2/3/6 (TGFβ type I receptors), leading to the phosphorylation and activation of the transcription factors Smad1/5/8. The Activin/Nodal branch involves the activation of the type I receptors ALK4/5/7 and subsequent phosphorylation and activation of the effectors Smad2/3. Smad7 inhibits both branches of TGFβ signaling and provides a repressive input on these pathways (reviewed in Ref.
      • Shi Y.
      • Massague J.
      ).
      Activation of Smad2 downstream of Nodal signaling is involved in several processes in the development of the embryo. Nodal induces mesoderm and endoderm, patterns the nervous system, and determines left-right asymmetry in vertebrates (reviewed in Ref.
      • Schier A.F.
      ). Lefty1 blocks Nodal signaling by a dual mechanism, it binds Nodal directly and also binds EGF-CFC co-receptors, such as TDGF-1/Cripto, thus preventing the assembly of an active Nodal/Activin receptor complex (
      • Chen C.
      • Shen M.M.
      ). A similar mechanism has been proposed for the highly homologous factor Lefty2. Loss of function analysis of Nodal leads to a phenotype where no mesoderm is formed (
      • Conlon F.L.
      • Lyons K.M.
      • Takaesu N.
      • Barth K.S.
      • Kispert A.
      • Herrmann B.
      • Robertson E.J.
      ) while Lefty2-deficient mice show excessive mesoderm formation and expansion of the primitive streak as a consequence of overactivation of Nodal signaling (
      • Meno C.
      • Gritsman K.
      • Ohishi S.
      • Ohfuji Y.
      • Heckscher E.
      • Mochida K.
      • Shimono A.
      • Kondoh H.
      • Talbot W.S.
      • Robertson E.J.
      • Schier A.F.
      • Hamada H.
      ). No obvious defects in the stem cell compartment, the inner cell mass, were observed in Lefty2-deficient animals. However, redundancy between Lefty1 and Lefty2 might explain the lack of phenotypic consequences. The Lefty1-deficient mouse shows a phenotype in which the left-right axis is disturbed (
      • Meno C.
      • Shimono A.
      • Saijoh Y.
      • Yashiro K.
      • Mochida K.
      • Ohishi S.
      • Noji S.
      • Kondoh H.
      • Hamada H.
      ). There are currently no double knockouts reported. Nodal and both Lefties are under the control of Activin/Nodal and Smad2/3 signaling in the embryonic node (reviewed in Ref.
      • Hamada H.
      • Meno C.
      • Watanabe D.
      • Saijoh Y.
      ).
      This study demonstrates that nodal, lefty-A, and lefty-B, are the earliest genes showing decreased expression upon differentiation in non-conditioned medium (non-CM). It is shown that Activin/Nodal signaling in hESCs regulates all three genes. Activin A treatment induces expression of these genes, whereas treatment with the ALK4/5/7 kinase inhibitor SB-431542 (
      • Inman G.J.
      • Nicolas F.J.
      • Callahan J.F.
      • Harling J.D.
      • Gaster L.M.
      • Reith A.D.
      • Laping N.J.
      • Hill C.S.
      ) inhibits their expression in undifferentiated cells. In hESCs grown under undifferentiated conditions in CM, Smad2/3 is phosphorylated, reflecting its activation, but its phosphorylation decreases when hESCs undergo differentiation in non-CM. When the regulatory sequence of lefty-A, which has multiple binding sites for the forkhead transcription factor FoxH1 (Fast1), a coactivator for Smad2/3, and sites for Smad2/3 itself, was fused to a luciferase reporter, comparable regulation during both pluripotency and differentiation was observed compared with the endogenous lefty-A expression. Luciferase expression increased when constitutively active ALK4 was co-expressed while co-expression of the repressive factor Smad7 inhibited luciferase expression. In addition, it has recently been reported that inhibition of GSK3, a hallmark of Wnt-signaling, using the kinase inhibitor 6-bromoindirubin-3′oxime (BIO,
      • Sato N.
      • Meijer L.
      • Skaltsounis L.
      • Greengard P.
      • Brivanlou A.H.
      ,
      • Meijer L.
      • Skaltsounis A.L.
      • Magiatis P.
      • Polychronopoulos P.
      • Knockaert M.
      • Leost M.
      • Ryan X.P.
      • Vonica C.A.
      • Brivanlou A.
      • Dajani R.
      • Crovace C.
      • Tarricone C.
      • Musacchio A.
      • Roe S.M.
      • Pearl L.
      • Greengard P.
      ), is sufficient to block the decision to undergo differentiation in both mouse and human ESCs (
      • Sato N.
      • Meijer L.
      • Skaltsounis L.
      • Greengard P.
      • Brivanlou A.H.
      ). Here it is shown that undifferentiated hESCs grown in BIO without CM, maintain high mRNA levels for nodal, lefty-A, and lefty-B as well as phosphorylation of Smad2/3 utilizing an active ALK4/5/7 receptor complex. These results suggest an interaction between inhibition of GSK-3 and Activin/Nodal signaling.

      EXPERIMENTAL PROCEDURES

      Reagents and Antibodies—Human Activin A, and BMP4 were purchased from R&D. The TGFβ inhibitor SB-431542 (
      • Inman G.J.
      • Nicolas F.J.
      • Callahan J.F.
      • Harling J.D.
      • Gaster L.M.
      • Reith A.D.
      • Laping N.J.
      • Hill C.S.
      ) was purchased from Tocris. The GSK-3 inhibitor, 6-bromoindirubin-3′oxime (BIO, Refs.
      • Sato N.
      • Meijer L.
      • Skaltsounis L.
      • Greengard P.
      • Brivanlou A.H.
      and
      • Meijer L.
      • Skaltsounis A.L.
      • Magiatis P.
      • Polychronopoulos P.
      • Knockaert M.
      • Leost M.
      • Ryan X.P.
      • Vonica C.A.
      • Brivanlou A.
      • Dajani R.
      • Crovace C.
      • Tarricone C.
      • Musacchio A.
      • Roe S.M.
      • Pearl L.
      • Greengard P.
      ) was generously provided by A. H. Brivanlou (The Rockefeller University, New York). All other reagents were purchased from Sigma.
      Antibodies used were rabbit polyclonal anti-phospho-Smad1/5/8 and Smad2/3 (Cell Signaling); rabbit polyclonal Smad1/5/8 and Smad2/3 (Upstate Biotechnology) and mouse monoclonal α-tubulin (Santa Cruz Biotechnology), and secondary anti-mouse and anti-rabbit antibodies linked to horseradish peroxidase (Jackson Immunoresearch Laboratories).
      Cell Culture—HESCs lines H1 (WiCell Research Institute) and BGN2 (BresaGen) were cultured on matrigel (BD Biosciences)-coated tissue culture plates in MEF-conditioned F12 medium supplemented with 20% knock-out serum replacement (KSR), non-essential minimal amino acids, penicillin/streptomycin, l-glutamine, β-mercaptoethanol, and 4 ng/ml FGF-2. The medium was conditioned by incubating 12 ml of medium for 24 h on hygromycin-resistent MEFs culture (5 × 106 cells/100 mm-diameter dish). MEFs (Specialty Media) were rendered growth-incompetent by treatment with mitomycin C by the manufacturer. MEFs were plated in Dulbecco's modified Eagle's medium supplemented with 10% calf serum and penicillin/streptomycin for the first 24 h until used for conditioning stem cell medium. All tissue-culture reagents were purchased from Invitrogen and used in the recommended concentration unless noted otherwise.
      Plasmid and Probes—The leftyA-luc construct was generated by PCR from bacterial artificial chromosome CITBI-E1–2519G10 (Research Genetics). The lefty-A promoter region (+42 bp to -4805 bp in regard to the transcription start site) was amplified using a primer pair from Integrated DNA Technologies: forward primer: 5′-ttccctgttcttcaaacaccgtcc-3′, reverse primer: 5′-tcctctagggaggttgaaggagg-3′. The amplified PCR fragment was purified with the PCR purification kit (Qiagen) and subcloned into pGL2-basic (Promega). pRL-null (Renilla) luciferase (Promega) was used as a control for transfection efficiency. Cells were co-transfected with components of the TGFβ signaling pathway: constitutive active ALK3 and ALK4 (kindly provided by R. Harland, University of California, Berkeley) and Smad7 in pcDNA3, (kindly provided by P. Ten Dijke, Ludwig Institute, Amsterdam, Netherlands, Ref.
      • Nakao A.
      • Afrakhte M.
      • Moren A.
      • Nakayama T.
      • Christian J.L.
      • Heuchel R.
      • Itoh S.
      • Kawabata M.
      • Heldin N.E.
      • Heldin C.H.
      • ten Dijke P.
      ).
      RT-PCR Assays—Total RNA was isolated from BGN2 stem cells using TRIzol (Invitrogen) and reverse-transcribed using MMLV reverse transcriptase (Invitrogen). PCR amplification was performed on a GeneAmp PCR system 9600 (PerkinElmer Life Sciences) using TaqDNA polymerase (Promega). The PCR reaction for all primer pairs consisted of 25 cycles, apart from β-actin (21 cycles) and lck (28 cycles). Oligonucleotides used for RT-PCR were custom-made by Integrated DNA Technologies (for primer sequences see Supplemental Data).
      Cell Transfection, Western Blotting, and Luciferase Assays—H1 and BGN2 cells were transfected using LipofectAMINE plus reagent (Invitrogen) as described by the manufacturer. Cells were either harvested for preparation of whole cell extracts (extraction buffer: 50 mm HEPES-NaOH, pH 7.4, 150 mm NaCl, 1% Triton X-100, 10% glycerol, 10 mm EDTA, 1 mm Na3VO4, and complete miniprotease inhibitor mixture (Roche Applied Science)) or luciferase extracts following the protocol of the manufacturer of the dual-luciferase assay (Promega). Cells for luciferase experiments were harvested 36 h after transfection. Firefly and Renilla luciferase activity was determined as recommended by Promega in a dual-luciferase luminometer (Lumat LB-9507, Berthold Technologies) and Western blotting of the respective whole cell extract (25 μg) was carried out following standard methods (
      • Sambrook J.
      • Fritsch E.F.
      • Maniatis T.
      ) and instructions by the manufacturer of the polyvinylidene difluoride blotting membranes (Millipore) and the immunodetection system (Amersham Biosciences).

      RESULTS

      Decrease of nodal, lefty-A, and lefty-B Expression Is the Earliest Indication of Differentiation in hESCs—A previous report has identified 918 genes in H1 hESCs as the signature of the undifferentiated pluripotent state (
      • Sato N.
      • Sanjuan I.M.
      • Heke M.
      • Uchida M.
      • Naef F.
      • Brivanlou A.H.
      ). 24% of this set, 204 genes, was also present in mESCs analyzed by similar means (
      • Ramalho-Santos M.
      • Yoon S.
      • Matsuzaki Y.
      • Mulligan R.C.
      • Melton D.A.
      ). In order to identify genes and/or signaling pathways that are involved in maintaining the undifferentiated state within this pool, 64 genes were selected based on their potential involvement in signal transduction and/or transcriptional regulation. The expression of these genes was assessed by comparative RT-PCR in H1 cells in a time course after withdrawal of MEF-conditioned medium (CM) (see Supplemental Data). The time-dependent regulation of 6 of the 64 genes is shown in Fig. 1A. Among the different genes tested in hESCs, nodal, lefty-A and lefty-B showed the earliest down-regulation after withdrawal of CM (Supplemental Data and Fig. 1). The level of nodal, lefty-A, and lefty-B transcripts drop to about 3% of their level in undifferentiated H1 cells within 24 h of CM withdrawal while the levels for all other 64 genes was still at 20% or higher of their initial value (Fig. 1B and Supplemental Data). These genes are also regulated in a similar way in another hESC line, BGN2 (Fig. 1C). Nodal, lefty-A, and lefty-B mRNA levels decreased the most rapidly upon differentiation compared with all other genes tested. For this reason and in consideration of the fact that Nodal and Lefty-A are involved in early cell fate decisions in the induction of endo- and mesoderm, we decided to analyze their regulation in further detail.
      Figure thumbnail gr1
      Fig. 1Down-regulation of nodal, lefty-A, and lefty-B is the earliest indication of differentiation in hESCs. A, RT-PCR analysis of the mRNA levels of nodal, lefty-A, and lefty-B, compared with other markers of pluripotency such as the genes for transcription factor oct3/4 (
      • Nichols J.
      • Zevnik B.
      • Anastassiadis K.
      • Niwa H.
      • Klewe-Nebenius D.
      • Chambers I.
      • Scholer H.
      • Smith A.
      ), and nanog (
      • Chambers I.
      • Colby D.
      • Robertson M.
      • Nichols J.
      • Lee S.
      • Tweedie S.
      • Smith A.
      ,
      • Mitsui K.
      • Tokuzawa Y.
      • Itoh H.
      • Segawa K.
      • Murakami M.
      • Takahashi K.
      • Maruyama M.
      • Maeda M.
      • Yamanaka S.
      ); the surface proteins tdgf-1 (cripto) (
      • Baldassarre G.
      • Romano A.
      • Armenante F.
      • Rambaldi M.
      • Paoletti I.
      • Sandomenico C.
      • Pepe S.
      • Staibano S.
      • Salvatore G.
      • De Rosa G.
      • Persico M.G.
      • Viglietto G.
      ), and thy-1 (
      • Henderson J.K.
      • Draper J.S.
      • Baillie H.S.
      • Fishel S.
      • Thomson J.A.
      • Moore H.
      • Andrews P.W.
      ); and the Src family kinase lck (
      • Sato N.
      • Sanjuan I.M.
      • Heke M.
      • Uchida M.
      • Naef F.
      • Brivanlou A.H.
      ). These marker molecules were analyzed in H1 hESCs over a time course of 0, 6, 12, 24, 78 h and 1 (1w) and 3 (3w) weeks of differentiation. β-actin was used as control for the RT-PCR reaction and gel loading, and as negative control for omission of reverse transcriptase (β-actin -RT). B, phosphorimager-based quantification of the experiment in A is shown in dual-logarithmic scale. C, same experiment as in A with the hESCs line BGN2 using two early time points (24 and 72 h) for differentiation.
      Activin/Nodal Signaling Is Required for the Regulation of nodal, lefty-A, and lefty-B—As mentioned above, nodal, lefty-A, and lefty-B are regulated in response to Smad2 activation in the embryonic node (reviewed in Ref.
      • Hamada H.
      • Meno C.
      • Watanabe D.
      • Saijoh Y.
      ). Therefore, we analyzed their regulation by modulation of the Activin/Nodal signaling pathway, which regulates the activation of Smad2 (Fig. 2A). Activation of the pathway was achieved by treatment of BGN2 cells with Activin A whereas inhibition was carried out by treatment of the cells with the ALK4/5/7 kinase inhibitor SB-431542 (
      • Inman G.J.
      • Nicolas F.J.
      • Callahan J.F.
      • Harling J.D.
      • Gaster L.M.
      • Reith A.D.
      • Laping N.J.
      • Hill C.S.
      ). Expression of all three genes decreased in cells grown in non-CM compared with cells grown in CM. Activin A led to increased expression of lefty-A in cells grown in non-CM while SB-431542 decreased its expression in cells grown in CM, both in a dose-dependent manner. When the cells were treated with Activin A and SB-431542 together, the levels of the genes were low, consistent with the fact that Activin A binds to ALK4 to activate Smad2/3. These results, taken together, demonstrate that nodal, lefty-A, and lefty-B are selectively responsive to Activin/Nodal signaling in hESCs.
      Figure thumbnail gr2
      Fig. 2Activin/Nodal signaling regulates nodal, lefty-A, and lefty-B expression and expression of components of the TGFβ signaling pathway in hESCs and MEFs. A, RT-PCR analysis for the expression levels for nodal, lefty-A, and lefty-B mRNA (as indicated) in BGN2 cells. Cells were grown in CM (lane 1), differentiated in non-CM over time (24 and 72 h, lanes 2 and 3), in non-CM with increasing amounts (10, 30, and 100 ng/ml) of recombinant Activin A (lanes 4–6), in CM with increasing amounts of SB-431542 (1, 3, 10 μm) (lanes 7–9), in non-CM with BIO (2 μm) alone (lane 11) or in non-CM in the presence of SB-431542 (10 μm) with Activin A (30 ng/ml, lane 10) or BIO (2 μm, lane 12). The treatment with Activin-A, SB-431542, BIO, or combination of these agents was done for 72 h in CM or non-CM as indicated. β-Actin expression was used as loading control and as negative control for omission of reverse transcriptase (β-actin -RT). B, RT-PCR analysis for mRNA expression levels of tgfβ-1,2,3, inhibin α, inhibin βA, inhibin βB, and gdf1 in MEFs. hprt expression was used as positive control and as negative control for omission of reverse transcriptase (hprt-RT). Expression of inhibin βC, inhibin βE, and nodal was not detected. C, RT-PCR analysis of activin type I receptors alk1–7, as described in , for undifferentiated (CM) and differentiated for 72 h (72) H1 and BGN2 cells.
      BIO Requires Activation of Smad2/3 for the Regulation of nodal, lefty-A, and lefty-B—It has been shown that inhibition of GSK3 by a small molecular weight compound, called BIO (
      • Sato N.
      • Meijer L.
      • Skaltsounis L.
      • Greengard P.
      • Brivanlou A.H.
      ,
      • Meijer L.
      • Skaltsounis A.L.
      • Magiatis P.
      • Polychronopoulos P.
      • Knockaert M.
      • Leost M.
      • Ryan X.P.
      • Vonica C.A.
      • Brivanlou A.
      • Dajani R.
      • Crovace C.
      • Tarricone C.
      • Musacchio A.
      • Roe S.M.
      • Pearl L.
      • Greengard P.
      ), is sufficient to maintain the undifferentiated state of hESCs and high expression of marker genes such as oct3/4, nanog, and rex-1 (
      • Sato N.
      • Meijer L.
      • Skaltsounis L.
      • Greengard P.
      • Brivanlou A.H.
      ). We tested the effect of BIO on the expression levels of nodal, lefty-A, and lefty-B with or without SB-431542. Fig. 2A shows that BIO in the absence of CM maintained the endogenous expression of all three genes to levels comparable with the expression in CM, while SB-431542 blocked induction of the genes induced by BIO. These results suggest that the induction of the genes whether it derives from conditioned medium or BIO treatment, requires a functional ALK4/5/7 pathway.
      Expression Status of Components of the TGFβ Signaling Pathway—A list of TGFβ pathway components present in the H1 line is shown in Table I. The values were determined by global gene expression analysis (
      • Sato N.
      • Sanjuan I.M.
      • Heke M.
      • Uchida M.
      • Naef F.
      • Brivanlou A.H.
      ) of H1 cells maintained undifferentiated or induced to differentiate. The list comprises 44 different molecules involved in TGFβ signaling either as ligands, receptors, Smads, or modulators of the pathway. The table compares the expression values for the molecules in undifferentiated (CM) and differentiated (non-CM) conditions and gives the relative indices for the expression (Fold CM/n-CM). While most genes are not regulated (fold value between 0.5 and 2) several genes show a significantly higher expression in undifferentiated cells, including lefty-A, lefty-B, and tdgf-1 (cripto), as stated above, as well as cerberus, gdf3, foxH1, and oaz. Other genes, such as TGFβ2, follistatin splice variant 344 (FST sv344), TGFβR-II, and ltbp-1, are up-regulated upon differentiation.
      Table IComponents of the TGFβ signaling pathway in H1 hESCs Table I shows the expression levels of several TGFβ pathway components as observed in a global expression profile using microarrays on the Affymetrix U133A Chip (
      • Sato N.
      • Sanjuan I.M.
      • Heke M.
      • Uchida M.
      • Naef F.
      • Brivanlou A.H.
      ). The second column gives the relative expression value for the respective gene in the presence of CM. The third column gives the relative expression value in the absence of CM (non-CM) for three weeks. The fourth column gives the index for the expression between CM and non-CM, thus a measure whether the gene decreases (>1), increases (<1) or remains unchanged upon differentiation (=1). A value between 0.5 and 2 (not bolded) does not allow a conclusive statement on the gene regulation since this value falls into the range of uncertainty (2-fold).
      CMnon-CMFold (CM/non-CM)
      Ligands
      tgfβ2417740.05
      inhibin βC4712571.8
      inhibin βE1201131.1
      bmp-21501461
      bmp-71561051.5
      gdf33721362.7
      nodal
      There were relevant genes, e.g. nodal, that are not represented on the U133A chip. However, nodal was found to be expressed in hESCs in other studies (8, 9)
      Present
      Extracellular inhibitors
      inhibin α671290.5
      lefty-A10441616.5
      lefty-B406010041
      decorin9057811.2
      tmeff110183532.9
      cerberus9372194.3
      follistatin (sv317)1371760.8
      follistatin (sv344)2968880.3
      follistatin-like 1240945430.5
      chordin-like145861.7
      drm/gremlin59660.9
      Type I receptors
      Note that genes such as ALK4/5/7, which were detected by RT-PCR (Fig. 2C), were not detected in the microarray analysis
      bmpRIA/ALK37744031.9
      actRI/ALK23163920.8
      Type II receptors
      tβR-II1043500.3
      actRII2302700.9
      actRIIB3833161.2
      Co-receptors
      tdgf1/cripto40635267.7
      Miscellaneous regulators
      ltbp-11485230.3
      fkbp12126217250.7
      arip-1/magi1131031.1
      tak13303510.9
      Smads
      smad25466880.8
      smad3771590.5
      smad44804681
      smad51561700.9
      smad72823740.8
      Smad-Adapter
      sara1592070.8
      Ubiquitin E3-Ligase
      smurf-12293660.6
      smurf-24389640.5
      Nuclear transport
      nup2142682311.2
      nup15310968881.2
      Co-activator
      foxH1/fast-13431272.7
      oaz6201773.5
      Transcription repressors
      drap-12264250.5
      ecsit1591411.1
      yy1217316361.3
      ski120681.8
      ski-1541360.4
      a There were relevant genes, e.g. nodal, that are not represented on the U133A chip. However, nodal was found to be expressed in hESCs in other studies (
      • Rosler E.S.
      • Fisk G.J.
      • Ares X.
      • Irving J.
      • Miura T.
      • Rao M.S.
      • Carpenter M.K.
      ,
      • Zeng X.
      • Miura T.
      • Luo Y.
      • Bhattacharya B.
      • Condie B.
      • Chen J.
      • Ginis I.
      • Lyons I.
      • Mejido J.
      • Puri R.K.
      • Rao M.S.
      • Freed W.J.
      )
      b Note that genes such as ALK4/5/7, which were detected by RT-PCR (Fig. 2C), were not detected in the microarray analysis
      Next, the expression of TGFβ ligands and receptors in MEFs and hESCs, respectively, was examined. MEFs secrete factor/s into the conditioned medium, which are able to maintain the hESCs undifferentiated. Transcripts for several TGFβ ligands were present in MEFs (Fig. 2B) including tgfβ-1,2,3, activin A,B and A/B (homo-(A or B) or heterodimer (A/B) of two inhibin β chains) and gdf1.
      RT-PCR analysis showed that all TGFβ type I receptors except ALK1 are present in H1 and BGN2 hESCs, making them competent to respond to Activin/Nodal/TGFβ signaling leading to Smad2/3 activation downstream of ALK4/5/7 (Fig. 2C).
      Regulation of Smad2/3 and Smad1/5/8 Activation in hESCs—The regulation of nodal, lefty-A, and lefty-B expression by Activin/Nodal signaling suggested that phosphorylation and activation of Smad2/3 is required for this process, since these transcription factors are the major downstream components of this signaling pathway. To address this question, Smad2/3 activation in hESCs grown either in CM or non-CM was determined.
      Activation of the effector Smads, including Smad2/3 and Smad1/5/8 can be measured by their phosphorylation with phosphospecific antibodies. Fig. 3 shows the results of a Western blot analysis using antibodies specific for the phosphorylated forms of Smad2/3 in H1 (Fig. 3A) and BGN2 cells (Fig. 3B). Significant levels of phosphorylated Smad2/3 were detected in cells when cultured under pluripotency conditions (CM). This experiment does not allow for the determination of which Smad molecule, Smad2 and/or Smad3 is involved in the activation of nodal, lefty-A, and lefty-B, because both are expressed in H1 cells (Table I), and the antibodies recognize both molecules. When the cells were induced to differentiate by growing them in non-CM, the level of phospho-Smad2/3 dropped dramatically in both cell lines. Treatment of hESCs in differentiating conditions with Activin A, led to phosphorylation of Smad2/3, while challenging BGN2 cells with the inhibitor SB-431542 in CM or in the presence of Activin A, strongly reduced the phosphorylation of Smad2 (Fig. 3B). If BGN2 cells are treated with BIO a significant induction of Smad2/3 phosphorylation can be observed, although not to the same level as observed in CM or after Activin A treatment. Also this activation is sensitive to SB-431542 suggesting that BIO treatment leads to the regulation of an extracellular event to activate Smad2/3. Treatment of the cells with BMP4, an activator of the Smad1/5/8 branch, in the presence of CM did not show a significant effect on the activation of Smad2/3, as expected.
      Figure thumbnail gr3
      Fig. 3Phosphorylation of Smad2/3 and Smad1/5/8 in hESCs. A, H1 human embryonic stem cells were grown in the presence of CM or in non-CM for 24 or 72 h. Cells were also treated with 30 ng/ml Activin A (Act), or 25 ng/ml BMP4 (BMP) for 72 h as indicated. The phosphorylated form of Smad1/5/8 (Smad1/5/8P, upper panel) and Smad2/3 (Smad2/3P, third panel from the top) were detected on Western blots (25 μg of total protein of whole cell extracts) using phosphospecific antibodies. The membranes were stripped and probed with antibodies against Smad1/5/8 (second panel from the top), Smad2/3 (fourth panel from the top), and α-tubulin (bottom panel), as indicated, to confirm that similar amounts of whole cell extracts were used. B, as in A for BGN2 cells, these cells were also treated with 10 μm SB-431542 (SB) and 2 μm BIO and with 10 μm SB-431542 in combination with 30 ng/ml Activin A or 2 μm BIO (as indicated). The treatment with Activin-A, SB-431542, BIO or combination of these agents was done for 72 h in CM or non-CM as indicated.
      To assess whether the Smad1/5/8 branch of TGFβ signaling was also regulated during differentiation, an antibody against the phosphorylated form of Smad1/5/8 was used. Interestingly, Smad1/5/8 signaling showed the opposite pattern of Smad2/3 activation. While phosphorylation levels were negligible in CM, they increased dramatically upon differentiation in H1 (Fig. 3A) and BGN2 (Fig. 3B). While treatment with Activin A and BIO in the absence of CM led to the decrease of Smad1/5/8 phosphorylation in both cell lines, although not completely to the same low level as observed in CM, BMP4 increases the level of phospho-Smad1/5/8 very strongly. Treatment of BGN2 cells with Activin A in non-CM reduced the level of phospho-Smad1/5/8 but not to the same level as in CM suggesting that inhibition of Smad1/5/8 is at least in part dependent on Smad2/3 activation. On the other hand, treatment with SB-431542 in CM, increased the phosphorylation compared with CM but not to same extent as in non-CM suggesting that factor/s in the CM are also involved in the inhibition of Smad1/5/8. Taken together, it can be proposed that factors in the conditioned medium, which are not affected by SB-431542 treatment, and factors from the hESCs themselves, which are regulated by the Smad2/3 pathway inhibit the Smad1/5/8 branch of TGFβ signaling in the undifferentiated state.
      These results suggest that activation of Smad2/3 and Smad1/5/8 in hESCs are oppositely regulated upon differentiation modulating expression of nodal, lefty-A, and lefty-B by Smad2/3, but not by Smad1/5/8.
      Characterization of lefty-A Regulatory Sequences Maintaining Expression in the Undifferentiated State—A 4.8-kb promoter region upstream of the lefty-A gene was cloned by PCR, using a BAC construct (CITBI-E1–2519G10) as template. This BAC contains around 200 kb sequence of the human chromosomal region 1q42 comprising the lefty-A and lefty-B locus (Fig. 4A). The regulatory regions of human lefty-A and mouse lefty2 contain seven homology sequence pairs (HSP) indicative of conserved regulatory sequences (Fig. 4A). As previously reported, two FoxH1 (Fast1) binding sites and five potential binding sites for the Smad2/4 heterodimer are present in the most upstream HSP 1 (
      • Saijoh Y.
      • Adachi H.
      • Sakuma R.
      • Yeo C.Y.
      • Yashiro K.
      • Watanabe M.
      • Hashiguchi H.
      • Mochida K.
      • Ohishi S.
      • Kawabata M.
      • Miyazono K.
      • Whitman M.
      • Hamada H.
      ). It has been shown that these binding sites are recognized by their respective transcription factors and that these binding sites are required to allow the expression of Lefty-A in vivo in the embryonic node (
      • Saijoh Y.
      • Adachi H.
      • Sakuma R.
      • Yeo C.Y.
      • Yashiro K.
      • Watanabe M.
      • Hashiguchi H.
      • Mochida K.
      • Ohishi S.
      • Kawabata M.
      • Miyazono K.
      • Whitman M.
      • Hamada H.
      ).
      Figure thumbnail gr4
      Fig. 4Characterization of lefty-A responsive regulatory sequences. The promoter of lefty-A is down-regulated in differentiated H1 and BGN2 cells compared with undifferentiated cells, and is responsive to activation of Smad2/3. A, organization of the human genomic region 1q42. Both lefty genes (red) and the promoter region for lefty-A (green) are indicated. The lefty gene region (middle panel) shows the organization of the lefty-A locus (from -5142kb to +4618 with regard to the transcription start site). Seven homology sequence pairs (HSP, black lines labeled 1–7) between the human lefty-A and the mouse lefty2 promoter are indicated including two FoxH1 (Fast-1) binding sites in the first most upstream HSP1 (red dots). The lefty-A gene (transcribed region) is indicated in red showing the four exons, the translation start site (ATG) and stop site underneath. A 4.8-kb PCR product (blue line) including the primer locations (arrowheads) comprising the upstream promoter region (+42 to -4805 in relation to the transcription start site) is indicated. The lower panel shows the sequence information for HSP1 (-4684 to -4377 bp upstream of the transcription start site) including the two conserved FoxH1 (Fast1) binding sites (bold red letters), the potential Smad(2/3)/4 heterodimer binding sites (green letters; one of these sites is conserved between human and mouse (bold green letters). B, H1 and BGN2 cells were transfected with 500 ng of a firefly luciferase reporter construct harboring 4.8 kb of the upstream promoter region of human lefty-A (lefty-A-luc). 50 ng of pRL-null were included to normalize for transfection efficiency. The cells were grown in the presence of conditioned medium (CM), non-conditioned medium (non-CM), and BIO for 36 h after transfection. Luciferase activity in extracts (10 μg of total protein) was determined as described under “Experimental Procedures.” The ratio between firefly and Renilla luciferase was determined and normalized with respect to that of the lefty-A-luc value in cells in conditioned medium. Each column represents the mean value of duplicate samples with the positive original value indicated as error bar; the value were normalized with respect to that of leftyA-luc with cells in CM. Experiments were performed twice with similar results.
      The 4.8-kb PCR product (+42bp to -4805 bp with regard to the transcription start site) of the promoter region was introduced upstream of a luciferase gene in pGL2-basic construct to produce leftyA-luc. In order to determine whether the promoter of the lefty-A gene is regulated in hESCs in the same way as the endogenous gene, the leftyA-luc construct was transiently transfected into H1 and BGN2 cells and assayed for luciferase expression. The cells were grown in CM or non-CM without or with BIO for 36 h after transfection. Fig. 4B shows that luciferase expression was significantly higher in cells grown in the presence of conditioned medium and BIO, than in cells that engage in differentiation in non-CM alone. This indicates that the 4.8-kb upstream region contains the regulatory elements that are sufficient to mimic the responsiveness of endogenous lefty-A gene (Figs. 1 and 2).
      Activation of Smad2/3 Selectively Regulates the lefty-A Promoter—To test whether this promoter construct displayed the same regulation as the endogenous gene, the activation and inhibition of the Smad2/3 pathway in cells cultured in CM and non-CM was performed. Smad2/3 activation was induced by two independent methods, either by the addition of Activin A or by coexpression of the constitutively activated (ca) type I receptor ca-ALK4. Activation of Smad1/5/8 by coexpression of ca-ALK3 was used as a negative control.
      Fig. 5A shows that in non-CM, activation of Smad2/3 by Activin A treatment was sufficient to induce the activation of the lefty-A promoter to the same or higher levels than observed in CM in H1 and BGN2 cells. Similarly, activation of the Smad2/3 branch by coexpression of ca-ALK4 in BGN2 cells maintained high expression of leftyA-luc in non-CM (Fig. 5B). ca-ALK3 did not show any influence on luciferase expression.
      Figure thumbnail gr5
      Fig. 5Activin/Nodal signaling regulates the lefty-A promoter. A, H1 and BGN2 cells were transfected with 500 ng lefty-A-luc and 50 ng pRL-null grown in CM or non-CM and presented with 30 ng/ml Activin A (Act), 10 μm SB-431532 (SB), and 2 μm BIO (BIO) alone or in combination for 36 h after transfection. Luciferase activity was determined and displayed as described in . B, BGN2 cells were transfected with 500 ng lefty-A-luc and 50 ng pRL-null and co-transfected with 500 ng pDH105 (CM and non-CM) or expression vectors for ca-ALK3, ca-ALK4, and the inhibitory Smad7. Luciferase activity was determined 36 h after transfection and displayed as described in .
      Inhibition of Smad2/3 signaling was achieved using either the ALK4/5/7 inhibitor SB-431542 or a cell autonomous inhibitory Smad, Smad7. It is noteworthy that Smad7 inhibits both Smad1/5/8 and Smad2/3 (
      • Shi Y.
      • Massague J.
      ). SB-431542 abrogated the activity of the lefty-A promoter in CM or in Activin A-treated BGN2 cells (Fig. 5A). Similarly, co-expression of Smad7 inhibited the induction of lefty-A in cells grown in CM (Fig. 5B).
      As shown above, inhibition of GSK3 by BIO is sufficient to induce the expression of nodal, lefty-A, and lefty-B in cells grown in the absence of CM dependent on ALK4/5/7 activation (Fig. 2A) and increased phosphorylation of Smad2/3 (Fig. 3B). To test whether the leftyA-luc construct is regulated in a similar way, its behavior was examined in cells treated with BIO with or without SB-431542 (Fig. 5A). As expected, the induction of the lefty-A promoter by BIO was suppressed when the drug SB-431542 was co-presented with BIO in BGN2 cells in a similar manner as described for the induction in CM and Activin A (Fig. 5A), suggesting that activation of the lefty-A promoter as well as the regulation of endogenous nodal, lefty-A, and lefty-B and the activation of Smad2/3 in BIO-treated cells requires active signaling through ALK4/5/7.

      DISCUSSION

      The identification of factors and signaling pathways involved in differentiation of ESCs is of importance for the field of stem cell research and subject of intense investigation. We show here that the genes for the ligand Nodal and its inhibitors Lefty-A and Lefty-B constitute a group of genes that respond immediately after differentiation is induced by withdrawal of CM. The rapid down-regulation of these genes is dependent on the loss of Activin/Nodal signaling and subsequent Smad2/3 activation. Direct activation of the pathway by Activin A and indirect activation by inhibition of GSK3 using BIO rescues the expression of the genes in the absence of CM. Conversely, an increase in BMP signaling and Smad1/5/8 activation is observed in cells undergoing differentiation. Altogether the findings presented here suggest that this group of genes, nodal and the lefties, and the signaling events they are regulating are not only important for early cell fate decisions in vivo but are also essential for the undifferentiated state of hESCs.
      TGFβ Signaling Components in Human Embryonic Stem Cells—The data presented here demonstrate that nodal, lefty-A, and lefty-B are the earliest genes detected whose expression decreases upon differentiation of hESCs and that their regulation is directly dependent on Smad2/3 activation. Interestingly, similar findings have been described in the embryonic node in the mouse (reviewed in Refs.
      • Hamada H.
      • Meno C.
      • Watanabe D.
      • Saijoh Y.
      and
      • Saijoh Y.
      • Adachi H.
      • Sakuma R.
      • Yeo C.Y.
      • Yashiro K.
      • Watanabe M.
      • Hashiguchi H.
      • Mochida K.
      • Ohishi S.
      • Kawabata M.
      • Miyazono K.
      • Whitman M.
      • Hamada H.
      ), suggesting that the regulation of these genes by Activin/Nodal signaling is a common mechanism during different stages of early development. In the case of the embryonic node, the relevant ligand activating Smad2/3 appears to be Nodal. This might suggest that Nodal itself is also involved in regulating Smad2/3 in the inner cell mass, i.e. the stem cell compartment, in vivo and raises the question whether the regulation of Smad2/3 also occurs in mouse embryonic stem cells (mESCs). Microarray analysis using mESCs grown on feeder cells, showed that mouse nodal, lefty1, and lefty2 were strongly down-regulated upon differentiation (
      • Ramalho-Santos M.
      • Yoon S.
      • Matsuzaki Y.
      • Mulligan R.C.
      • Melton D.A.
      ). However, when mESCs were grown in the presence of LIF no regulation of lefty1 or Smad2/3 was observed (data not shown). It appears possible that different culture conditions between hESCs (in CM) and mESCs (in LIF) affect the activation of Smad2/3 and regulation of the lefty genes in ESCs. These differences need to be further investigated.
      The inhibition of BMP signaling in hESCs seems to be dependent on multiple factors secreted by the MEFs into CM and by the hESCs themselves. Several BMP inhibitors produced by both cell types were detected in global gene expression profiles for both cell types, e.g. cerberus in hESCs (
      • Sato N.
      • Sanjuan I.M.
      • Heke M.
      • Uchida M.
      • Naef F.
      • Brivanlou A.H.
      ) and gremlin in MEFs.
      D. Besser, unpublished observation.
      The finding that activation of Smad1/5/8 is low in the undifferentiated state is in agreement with a recent report showing that active BMP signaling promotes differentiation of hESCs into trophoblast cells (
      • Xu R.H.
      • Chen X.
      • Li D.S.
      • Li R.
      • Addicks G.C.
      • Glennon C.
      • Zwaka T.P.
      • Thomson J.A.
      ). It cannot be excluded that the strong increase of Smad1/5/8 activation during the differentiation may have a direct effect on the down-regulation of the expression levels of nodal, lefty-A, and lefty-B, either by competing with the active Smad2/3 to complex with Smad4, which dimerizes with Smad2/3 and Smad1/5/8 to activate transcription (reviewed in Ref.
      • Shi Y.
      • Massague J.
      ) or by direct binding of Smad1/5/8 as a transcriptional repressor to the promoter of nodal, lefty-A, and lefty-B.
      Although none of the mice deficient for Nodal, Lefty1, or Lefty2 has obvious defects in the inner cell mass (
      • Conlon F.L.
      • Lyons K.M.
      • Takaesu N.
      • Barth K.S.
      • Kispert A.
      • Herrmann B.
      • Robertson E.J.
      ), it might be speculated, that the observed regulation of Nodal, Lefty-A, and Lefty-B in the hESCs establishes a signaling network that allows the temporal and spatial control of Smad2/3 activation in the stem cell compartment to establish the correct induction of endoderm and especially of mesoderm. This mechanism allows a positive feedback loop in which expression of Nodal via the activation of Smad2/3 induces its own expression and a negative feedback loop in which both Lefties inhibit the function of Nodal. Thus, the question arises which factors are supplied in the conditioned medium from MEFs to maintain the stem cells undifferentiated and expression of nodal, lefty-A, and lefty-B high. It is very likely, that the MEFs might provide other Smad2/3 activating factors such as Activin A, A/B, and B and TGFβ1–3 (Fig. 2B) that are required to establish an input into Smad2/3 signaling that cannot be inhibited by the Lefties. Another possibility is that MEFs provide other molecules activating Nodal signaling such as ligands of the Wnt family, which induce Nodal-related factors in lower vertebrates, ligands of the Notch family, which have been shown to activate Nodal signaling in the Node, or proteases such as the convertases Spc1 and Spc4 (also know as Furin and Pace4, respectively), which are required for the maturation of the Nodal protein (reviewed in Ref.
      • Schier A.F.
      ). In addition, the MEF-conditioned medium contains gremlin, which appears to be responsible for the inhibition of BMP-Smad1/5/8 pathway in the undifferentiated state. Further experiments regarding the factor/s produced by the MEFs, which establish the maintenance of the undifferentiated state and the high expression of nodal, lefty-A, and lefty-B are required.
      Synergism between the Wnt Pathway and Smad2/3 Activation in Embryonic Systems—Activation of Wnt signaling has been implicated in the establishment of stemness in different adult stem cells, e.g. in skin (
      • Gat U.
      • DasGupta R.
      • Degenstein L.
      • Fuchs E.
      ), hematopoetic (
      • Reya T.
      • Duncan A.W.
      • Ailles L.
      • Domen J.
      • Scherer D.C.
      • Willert K.
      • Hintz L.
      • Nusse R.
      • Weissman I.L.
      ), and embryonic stem cells (
      • Sato N.
      • Meijer L.
      • Skaltsounis L.
      • Greengard P.
      • Brivanlou A.H.
      ). Our data suggest that a cross-talk between inhibition of GSK3 by BIO, which has been implicated in Wnt-signaling (
      • Sato N.
      • Meijer L.
      • Skaltsounis L.
      • Greengard P.
      • Brivanlou A.H.
      ,
      • Meijer L.
      • Skaltsounis A.L.
      • Magiatis P.
      • Polychronopoulos P.
      • Knockaert M.
      • Leost M.
      • Ryan X.P.
      • Vonica C.A.
      • Brivanlou A.
      • Dajani R.
      • Crovace C.
      • Tarricone C.
      • Musacchio A.
      • Roe S.M.
      • Pearl L.
      • Greengard P.
      ) and Activin/Nodal signaling occurs in the regulation of nodal, lefty-A, and lefty-B expression in undifferentiated stem cells. The interaction between these two pathways has been documented in early embryonic development (reviewed in Ref.
      • Harland R.
      • Gerhart J.
      ). For instance, ectodermal explants of UV-irradiated Xenopus embryos differentiate to dorsal mesoderm when treated with a combination of Wnt8 and Activin, but not when exposed to either factor alone (
      • Sokol S.Y.
      • Melton D.A.
      ). Activation of Wnt signaling in Xenopus embryos by lithium chloride ultimately leads to increased phosphorylation of Smad2/3 (
      • Schohl A.
      • Fagotto F.
      ). The mechanism of Wnt/TGFβ cooperation may also be direct, since the downstream effector molecules of these pathways, Lef1 and Smad2/3 have been reported to form complexes and act synergistically to activate the Xtwn gene in Xenopus (
      • Labbe E.
      • Letamendia A.
      • Attisano L.
      ).
      Results from studies of targeted deletions in mouse provide further evidence of extensive cross-talk between Wnt signaling and Smad2/3 activation. Primitive streak formation is absent or impaired in Wnt3-/- mice (
      • Liu P.
      • Wakamiya M.
      • Shea M.J.
      • Albrecht U.
      • Behringer R.R.
      • Bradley A.
      ), and a similar phenotype is also observed in mice deficient for Nodal (
      • Conlon F.L.
      • Lyons K.M.
      • Takaesu N.
      • Barth K.S.
      • Kispert A.
      • Herrmann B.
      • Robertson E.J.
      ), ALK4 (
      • Gu Z.
      • Nomura M.
      • Simpson B.B.
      • Lei H.
      • Feijen A.
      • van den Eijnden-van Raaij J.
      • Donahoe P.K.
      • Li E.
      ), Smad2 (
      • Weinstein M.
      • Yang X.
      • Li C.
      • Xu X.
      • Gotay J.
      • Deng C.X.
      ,
      • Nomura M.
      • Li E.
      ,
      • Waldrip W.R.
      • Bikoff E.K.
      • Hoodless P.A.
      • Wrana J.L.
      • Robertson E.J.
      ), FoxH1 (Fast1) (
      • Hoodless P.A.
      • Pye M.
      • Chazaud C.
      • Labbe E.
      • Attisano L.
      • Rossant J.
      • Wrana J.L.
      ), and Smad4 (
      • Sirard C.
      • de la Pompa J.L.
      • Elia A.
      • Itie A.
      • Mirtsos C.
      • Cheung A.
      • Hahn S.
      • Wakeham A.
      • Schwartz L.
      • Kern S.E.
      • Rossant J.
      • Mak T.W.
      ,
      • Yang X.
      • Li C.
      • Xu X.
      • Deng C.
      ). These common phenotypes are indicative of the functional relevance of these molecules and their cross-talk in early mammalian development might be reflected in the regulation of these pathways in hESCs shown here.

      CONCLUSIONS

      Here, it is shown that activation of Smad2/3 is necessary for the expression of nodal, lefty-A, and lefty-B. A model (Fig. 6) is proposed in which Smad2/3 is directly regulated by ligands in the MEF-CM and expression of Nodal protein may establish a positive forward loop. In addition, inhibition of GSK3 by BIO, which has been implicated in Wnt-signaling, regulates these genes depending on ALK4/5/7. Inhibition of GSK3 establishes crosstalk to the Activin/Nodal signaling pathway via a yet unknown factor/s to induce increased levels of expression of nodal, lefty-A, and lefty-B.
      Figure thumbnail gr6
      Fig. 6Proposed model of the cross-talk between Wnt and TGFβ signaling in hESCs leading to regulation of Nodal, Lefty-A, and Lefty-B in the undifferentiated state. A previous study has shown that BIO (light red), by inhibiting GSK-3, can maintain hESCs undifferentiated (
      • Sato N.
      • Meijer L.
      • Skaltsounis L.
      • Greengard P.
      • Brivanlou A.H.
      ). hESCs are also undifferentiated when grown in MEFs-conditioned medium (MEF-CM). Here it is shown that activation of Smad2/3 downstream of ALK4/5/7 leads directly to high expression of nodal, lefty-A, and lefty-B in the undifferentiated state or when the cells are treated with Activin A. Inhibition of ALK4/5/7 by SB-431542 (light red) can completely abolish this activation. Furthermore, cross-talk between inhibition of GSK3, a hallmark of Wnt-signaling (light green) and activation of Smad2/3 (green) is required for expression of all three genes in cells that are treated with BIO. The results suggest that this cross-talk is established by an extracellular ligand leading to Smad2/3 activation. In addition, factors (orange) produced by the MEFs (e.g. Gremlin) and by the hESCs themselves (e.g. Cerberus) in response to Smad2/3 activation, inhibit the Smad 1/5/8 branch of TGFβ signaling (red) that is activated when cell undergo differentiation.
      Interestingly, in the undifferentiated state of hESCs, there are two mechanisms that account for the inhibition of ALK2/3/6 signaling, one is dependent on the Smad2/3 activation and appears to involve factors from the hESCs, possibly Cerberus (Table I), the other one appears to involve factors secreted by the MEFs, such as Gremlin.
      Further experiments are needed to assess the effect of Smad2/3 regulation on the behavior of other markers for the pluripotent state and stemness itself. Furthermore, the role of Smad2/3 activation may also have a role in the maintenance of other progenitor cell types that require Wnt signaling. In the absence of stringent in vivo assays in humans, it is difficult to address the in vivo relevance of these findings. Assays to address the formation of teratomas and embryoid bodies in the presence or absence of Smad2/3 activation in mESCs and hESCs in different settings, including BIO and MEF-CM are required and should address the in vivo relevance of these findings.
      The data reported here on the regulation of Nodal, Lefty-A and Lefty-B and the opposing regulation of the two TGFβ signaling branches offer new insights into the signaling events controlling the maintenance of pluripotency.

      Acknowledgments

      I thank Ali H. Brivanlou for his generous support during this study in his laboratory. I would also like to thank Ariel Levine for her contribution to Fig. 2C, Noboru Sato for H1 cells for the RT-PCR experiment in the Supplemental Data, Makiko Uchida for excellent technical support, and Daylon James, Ariel Levine, Jacqueline Bromberg, Alin Vonica, Scott Noggle, Noboru Sato, Patric Turowski, and Inés Ibañez-Tallon for discussion and critical reading of the manuscript. I would like to thank WiCell (Wisconsin) for providing the H1 cell line, BresaGen, Inc. for the BGN2 cell line, Ali H. Brivanlou (The Rockefeller University, New York) for BIO, Richard Harland (University of California, Berkeley) for the ca-ALK3 and -4 expression plasmids and Peter Ten Dijke (Ludwig Insitute, Amsterdam, Netherlands) for the Smad7 expression construct.

      References

        • Sato N.
        • Meijer L.
        • Skaltsounis L.
        • Greengard P.
        • Brivanlou A.H.
        Nat. Med. 2004; 10: 55-63
        • Reubinoff B.E.
        • Pera M.F.
        • Fong C.Y.
        • Trounson A.
        • Bongso A.
        Nat. Biotechnol. 2000; 18: 399-404
        • Smith A.G.
        Annu. Rev. Cell Dev. Biol. 2001; 17: 435-462
        • Ko M.S.
        Trends Biotechnol. 2001; 19: 511-518
        • Loebel D.A.
        • Watson C.M.
        • De Young R.A.
        • Tam P.P.
        Dev. Biol. 2003; 264: 1-14
        • Ramalho-Santos M.
        • Yoon S.
        • Matsuzaki Y.
        • Mulligan R.C.
        • Melton D.A.
        Science. 2002; 298: 597-600
        • Sato N.
        • Sanjuan I.M.
        • Heke M.
        • Uchida M.
        • Naef F.
        • Brivanlou A.H.
        Dev. Biol. 2003; 260: 404-413
        • Rosler E.S.
        • Fisk G.J.
        • Ares X.
        • Irving J.
        • Miura T.
        • Rao M.S.
        • Carpenter M.K.
        Dev. Dyn. 2004; 229: 259-274
        • Zeng X.
        • Miura T.
        • Luo Y.
        • Bhattacharya B.
        • Condie B.
        • Chen J.
        • Ginis I.
        • Lyons I.
        • Mejido J.
        • Puri R.K.
        • Rao M.S.
        • Freed W.J.
        Stem. Cells. 2004; 22: 292-312
        • Harland R.
        • Gerhart J.
        Annu. Rev. Cell Dev. Biol. 1997; 13: 611-667
        • Shi Y.
        • Massague J.
        Cell. 2003; 113: 685-700
        • Schier A.F.
        Annu. Rev. Cell Dev. Biol. 2003; 19: 589-621
        • Chen C.
        • Shen M.M.
        Curr. Biol. 2004; 14: 618-624
        • Conlon F.L.
        • Lyons K.M.
        • Takaesu N.
        • Barth K.S.
        • Kispert A.
        • Herrmann B.
        • Robertson E.J.
        Development. 1994; 120: 1919-1928
        • Meno C.
        • Gritsman K.
        • Ohishi S.
        • Ohfuji Y.
        • Heckscher E.
        • Mochida K.
        • Shimono A.
        • Kondoh H.
        • Talbot W.S.
        • Robertson E.J.
        • Schier A.F.
        • Hamada H.
        Mol. Cell. 1999; 4: 287-298
        • Meno C.
        • Shimono A.
        • Saijoh Y.
        • Yashiro K.
        • Mochida K.
        • Ohishi S.
        • Noji S.
        • Kondoh H.
        • Hamada H.
        Cell. 1998; 94: 287-297
        • Hamada H.
        • Meno C.
        • Watanabe D.
        • Saijoh Y.
        Nat. Rev. Genet. 2002; 3: 103-113
        • Inman G.J.
        • Nicolas F.J.
        • Callahan J.F.
        • Harling J.D.
        • Gaster L.M.
        • Reith A.D.
        • Laping N.J.
        • Hill C.S.
        Mol. Pharmacol. 2002; 62: 65-74
        • Meijer L.
        • Skaltsounis A.L.
        • Magiatis P.
        • Polychronopoulos P.
        • Knockaert M.
        • Leost M.
        • Ryan X.P.
        • Vonica C.A.
        • Brivanlou A.
        • Dajani R.
        • Crovace C.
        • Tarricone C.
        • Musacchio A.
        • Roe S.M.
        • Pearl L.
        • Greengard P.
        Chem. Biol. 2003; 10: 1255-1266
        • Nakao A.
        • Afrakhte M.
        • Moren A.
        • Nakayama T.
        • Christian J.L.
        • Heuchel R.
        • Itoh S.
        • Kawabata M.
        • Heldin N.E.
        • Heldin C.H.
        • ten Dijke P.
        Nature. 1997; 389: 631-635
        • Sambrook J.
        • Fritsch E.F.
        • Maniatis T.
        Molecular Cloning: A Laboratory Manual. 2nd Ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY1989
        • Saijoh Y.
        • Adachi H.
        • Sakuma R.
        • Yeo C.Y.
        • Yashiro K.
        • Watanabe M.
        • Hashiguchi H.
        • Mochida K.
        • Ohishi S.
        • Kawabata M.
        • Miyazono K.
        • Whitman M.
        • Hamada H.
        Mol. Cell. 2000; 5: 35-47
        • Xu R.H.
        • Chen X.
        • Li D.S.
        • Li R.
        • Addicks G.C.
        • Glennon C.
        • Zwaka T.P.
        • Thomson J.A.
        Nat. Biotechnol. 2002; 20: 1261-1264
        • Gat U.
        • DasGupta R.
        • Degenstein L.
        • Fuchs E.
        Cell. 1998; 95: 605-614
        • Reya T.
        • Duncan A.W.
        • Ailles L.
        • Domen J.
        • Scherer D.C.
        • Willert K.
        • Hintz L.
        • Nusse R.
        • Weissman I.L.
        Nature. 2003; 423: 409-414
        • Sokol S.Y.
        • Melton D.A.
        Dev. Biol. 1992; 154: 348-355
        • Schohl A.
        • Fagotto F.
        Development. 2002; 129: 37-52
        • Labbe E.
        • Letamendia A.
        • Attisano L.
        Proc. Natl. Acad. Sci. U. S. A. 2000; 97: 8358-8363
        • Liu P.
        • Wakamiya M.
        • Shea M.J.
        • Albrecht U.
        • Behringer R.R.
        • Bradley A.
        Nat. Genet. 1999; 22: 361-365
        • Gu Z.
        • Nomura M.
        • Simpson B.B.
        • Lei H.
        • Feijen A.
        • van den Eijnden-van Raaij J.
        • Donahoe P.K.
        • Li E.
        Genes Dev. 1998; 12: 844-857
        • Weinstein M.
        • Yang X.
        • Li C.
        • Xu X.
        • Gotay J.
        • Deng C.X.
        Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 9378-9383
        • Nomura M.
        • Li E.
        Nature. 1998; 393: 786-790
        • Waldrip W.R.
        • Bikoff E.K.
        • Hoodless P.A.
        • Wrana J.L.
        • Robertson E.J.
        Cell. 1998; 92: 797-808
        • Hoodless P.A.
        • Pye M.
        • Chazaud C.
        • Labbe E.
        • Attisano L.
        • Rossant J.
        • Wrana J.L.
        Genes Dev. 2001; 15: 1257-1271
        • Sirard C.
        • de la Pompa J.L.
        • Elia A.
        • Itie A.
        • Mirtsos C.
        • Cheung A.
        • Hahn S.
        • Wakeham A.
        • Schwartz L.
        • Kern S.E.
        • Rossant J.
        • Mak T.W.
        Genes Dev. 1998; 12: 107-119
        • Yang X.
        • Li C.
        • Xu X.
        • Deng C.
        Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 3667-3672
        • Nichols J.
        • Zevnik B.
        • Anastassiadis K.
        • Niwa H.
        • Klewe-Nebenius D.
        • Chambers I.
        • Scholer H.
        • Smith A.
        Cell. 1998; 95: 379-391
        • Chambers I.
        • Colby D.
        • Robertson M.
        • Nichols J.
        • Lee S.
        • Tweedie S.
        • Smith A.
        Cell. 2003; 113: 643-655
        • Mitsui K.
        • Tokuzawa Y.
        • Itoh H.
        • Segawa K.
        • Murakami M.
        • Takahashi K.
        • Maruyama M.
        • Maeda M.
        • Yamanaka S.
        Cell. 2003; 113: 631-642
        • Baldassarre G.
        • Romano A.
        • Armenante F.
        • Rambaldi M.
        • Paoletti I.
        • Sandomenico C.
        • Pepe S.
        • Staibano S.
        • Salvatore G.
        • De Rosa G.
        • Persico M.G.
        • Viglietto G.
        Oncogene. 1997; 15: 927-936
        • Henderson J.K.
        • Draper J.S.
        • Baillie H.S.
        • Fishel S.
        • Thomson J.A.
        • Moore H.
        • Andrews P.W.
        Stem. Cells. 2002; 20: 329-337