Activation of a Nodal-independent signaling pathway by Cripto-1 mutants with impaired activation of a Nodal-dependent signaling pathway

Abstract

Cripto-1, a co-receptor for Nodal, can activate Nodal- dependent and Nodal-independent signaling pathways. In this study we have investigated whether Cripto-1 mutants, that fail to activate a Nodal-dependent signaling pathway, are capable to activate a Nodal-independent signaling pathway in mammary epithelial cells. Cripto-1 mutants expressed in EpH4 mouse mammary epithelial cells are fully functional in regard to activation of a Nodal-independent signaling pathway, leading to phosphorylation of mitogen-activated protein kinase (MAPK) and Akt and to enhanced proliferation and motility of these cells, suggesting that Cripto-1 mutants with impaired Nodal signaling are still active in a Nodal-independent signaling pathway.

1. Introduction

Human and mouse Cripto-1 (CR-1/Cr-1) are members of the epidermal growth factor (EGF)–CFC (Cripto-1 in humans, FRL1 in Xenopus and Cryptic in mice) family of proteins [1]. EGF–CFC proteins contain multiple domains including a modified EGF-like domain and a cystein-rich CFC motif [1]. During embryogenesis, Cripto-1 can function as a co-receptor for the transforming growth factor β (TGFβ) family member Nodal regulating the formation of the primitive streak and specification of the mesoderm and endoderm [1]. Cripto-1-dependent Nodal signaling depends upon the Activin type II (Act RII) and type I (Alk4) serine/threonine kinase receptors that activate the Smad-2/Smad-3 intracellular signaling pathway [2]. Evidence suggests that Cripto-1 recruits Nodal to the Act RII/Alk4 receptor complex by interacting with Nodal through the EGF-like domain and with Alk4 through the CFC domain [3]. Cripto-1 is expressed at high levels in several different types of human tumors, including breast and colon cancer [1]. In fact, Cripto-1 overexpression in human cancers has been associated with numerous aspects of tumor initiation and progression, including enhanced cellular proliferation, epithelial-to-mesenchimal transition and tumor angiogenesis [4–6]. Regulation of cell proliferation, motility and survival by Cripto-1 is dependent upon activation of the ras/raf/mitogen-activated protein kinase (MAPK) and phopshatidylinositol 3′ kinase (PI3K)/Akt signaling pathways [2,7]. Activation of these two intracellular signaling pathways is independent of Nodal and is mediated by binding of Cripto-1 to the heparan sulfate proteoglycan Glypican- 1, which can then activate the tyrosine kinase c-Src triggering activation of MAPK and Akt [7]. O-fucosylation is a rare form of glycosylation in which fucose is transferred to Threonine (Thr) or Serine (Ser) residues within an EGF-like module of several proteins [8]. Human and mouse Cripto-1 proteins are fucosylated at a conserved Thr residue within the EGF-like domain (Thr⁸⁸ for human CR-1 and Thr⁷² for mouse Cr-1) [9–11]. A single point mutation in the fucosylation consensus sequence of human CR-1 (Thr⁸⁸ to alanine [Ala]) or of mouse Cr-1 (Thr⁷² to Ala) results in loss of Cripto-1-dependent Nodal signaling in different cell-based assays [9,11]. Interestingly, a recent study has shown that O-fucosylation is not required by Cripto-1 to function in Nodal signaling pathway and that, in contrast, Thr⁷², to which O-fucose is attached, is absolutely required by Cripto- 1 to bind to Nodal [10]. However, none of these studies have investigated whether O-fucose and/or Thr^88/72 are required for the activation of a Nodal-independent signaling pathway. In the present study we have assessed the biological activity of fucosylation impaired and/or Thr Cripto-1 mutants in Nodal-independent signaling assays in mousemammary epithelial cells.

2. Materials and methods

2.1. Cell culture

EpH4, COS7 and 293T cells were grown as previously described [2]. EpH4 cells were transfected with wild-type, T72A and T72S Cripto-1 plasmids (kindly provided by Pamela Stanley, Albert Einstein College of Medicine, New York) using Fugene 6 (Roche, Indianapolis, IN) and cell lines were selected with G418 (400 μg/ml) (Invitrogen, Carlsbad, CA) for 2 weeks [7].

2.2. Cell proliferation, migration and invasion assays

Proliferation assay was performed as previously described [6]. Migration and invasion assays were performed in fibronectin-coated or Matrigel-coated Boyden chambers (Chemicon, Temecula, CA) [6]. Briefly, EpH4 WT, Cr-1, T72A or T72S cells were seeded in 12-well plates at 2 × 10⁵ cells per well and incubated overnight at 37 °C. The following day, the cells that had migrated or invaded the Matrigel through the filter were stained with a crystal violet solution, the stain solution was eluted and the absorbance was read at 595 nm.

2.3. Western blot analysis

EpH4 WT, Cr-1, T72A and T72S cells were seeded in 60-mm plates (5 × 10⁵ cells/plate) and serum starved for 24 h. Western blot analysis for phospho- and total-MAPK (1:1000 dilution, Cell Signaling, Beverly, MA) and phospho- and total-Akt (1:1000 dilution, Cell Signaling) was performed as previously described [2]. For Smad-2 activation, EpH4 WT, Cr-1, T72A and T72S were transiently transfected with a Nodal expression vector [12] using Fugene 6 and serum-starved for 24 h. Western blot for phospho- and total Smad-2 was performed using anti-phospho- and total-Smad-2 antibodies (1:1000, Cell Signaling). Western blots for Cr-1 and β-actin were performed using an antimouse Cr-1 goat polyclonal antibody (1:1000, Santa Cruz Biotechnology, Santa Cruz, CA) and an anti-β-actin mouse monoclonal antibody (1:10 000, Sigma, St. Louis, MO).

2.4. Coimmunoprecipitation assay

COS7 cells (1 × 10⁶ cells/60-mm plates) were transiently transfected with Glypican-1-Fc and Cr-1, T72A or T72S expression vectors using Fugene 6. 48 h after transfection cells were lysed and Glypican-1-Fc was immunoprecipitated as previously described [7]. Western blots were performed using anti-mouse Cr-1 goat polyclonal antibody (Santa Cruz) and anti-human Fc antibody (1:5000, Amersham, Piscataway, NJ) for Glypican-1Fc.

2.5. Transfection of EpH4 cells with siRNA

EpH4 WT, Cr-1, T72A and T72S cells were transfected with 3 μg of mouse Glypican-1 siRNA (Santa Cruz Biotechnology) or control siRNA (Santa Cruz Biotechnology) using Amaxa technology (Amaxa Biosystems, Gaithersburg, MD). After nucleofection cells were plated at 1 × 10⁶ cells in 100-mm plates. The following day, plates for Western blot analysis were washed with PBS and incubated in serum-free medium for additional 24 h. 48 h after nucleofection, proteins were extracted from plates in serum-free medium and RNA was isolated from plates of cells growing in complete medium. Western blot analysis was then performed on serum-starved cells using phospho-and total-MAPK or phospho- and total-Akt antibodies, as described above.

2.6. Real-time PCR

RNA was isolated using RNeasy mini kit (Qiagen, Valencia, CA) and cDNA synthesis was performed using 1 μg of total RNA with the RETROscript kit (Ambion Applied Biosystems, Foster City, CA) following the manufacturer’s protocol. Quantitative Real time PCR was performed on Startagene MX300P using Brilliant II SYBR Green PCR master mix (Stratagene, Cedar Creek, TX). Primers for mouse Glypican-1 were purchased from Santa Cruz Biotechnology. Glypican-1 expression levels were normalized to mouse 28S RNA expression (forward primer 5′-TTGAAAATCCGGGGGAGAG-3′, reverse primer 5′-ACATTGTTCCAACATGCCAG-3′).

2.7. Luciferase assay

Nodal-responsive reporter assay was performed as previously described [13]. Briefly, 293T cells plated in 24-well plates were transfected with an optimized amount of expression vectors ((n2)7-Lux, 50 ng/well; TK-renilla (Promega, Madison, WI), 5 ng/well; FAST-1, 25 ng/well; Nodal, 100 ng/well; Alk4, 50 ng/well; Cr-1, T72A or T72S, 100 ng/well). 24 h after transfection, dual-luciferase assays were performed using a dual-luciferase kit (Promega).

3. Results

3.1. Generation of EpH4 mouse mammary epithelial cells stably expressing Cr-1 mutants

We have previously demonstrated that overexpression of mouse Cr-1 can induce an increase in cell proliferation, anchorage-independent growth, migration and invasion of mouse mammary epithelial cells [4]. We therefore investigated the biological effects of the Cr-1 T72A mutant, which is not fucosylated, and the Cr-1 T72S mutant, which is fucosylated on the Ser residue, in EpH4 mouse mammary epithelial cells, which are negative for Cr-1 expression [4]. EpH4 wild-type (WT) control cells were transfected with expression vectors encoding for wild-type Cr-1 or T72A and T72S Cr-1 mutants. A specific band for Cr-1 could be detected by Western blot analysis in EpH4 Cr-1, T72A and T72S cells using an anti- Cr-1 polyclonal antibody (Fig. 1).

Fig. 1.

Expression of wild-type, T72A and T72S Cr-1 plasmids in EpH4 cells. Western blot analysis for Cr-1 and β-actin in EpH4 cells expressing Cr-1, T72A and T72S plasmids.

3.2. Expression of wild-type, T72A or T72S Cr-1 enhance cell proliferation, migration and invasion of EpH4 mammary epithelial cells

To ascertain if overexpression of wild-type or mutant Cr-1 proteins might alter cell proliferation, EpH4 WT, Cr-1, T72A and T72S cells were grown in serum-free conditions for 7 days. In agreement with previous results [4], Cr-1 transfected cells exhibited an approximately three-fold enhanced growth rate after 6 days under serum-free conditions compared to EpH4 WT control cells (Fig. 2A). EpH4 T72A or T72S cells showed a proliferation rate similar to EpH4 Cr-1 cells (Fig. 2A). We next determined the effect of T72A and T72S Cr-1 mutants on EpH4 cell motility. As shown in Fig. 2B and C, EpH4 Cr-1 expressing cells exhibited a statistically significant increase in their migratory (three-fold increase) (P < 0.01) and invasive (two-fold increase) (P < 0.01) behavior when compared to EpH4 WT cells. Likewise, EpH4 T72A and EpH4 T72S cells also showed enhanced motility with levels of migration and invasion comparable to EpH4 Cr-1 cells.

Fig. 2.

Biological activity of T72A and T72S Cr-1 mutants and activation of MAPK and Akt signaling pathways in EpH4 cells. (A) Proliferation assay in serum-free conditions of Cr-1, T72A and T72S EpH4 expressing cells compared to wild-type control cells. Data are representative of three experiments with triplicate samples. (B) Migration and (C) invasion assay of EpH4 WT, Cr-1, T72A and T72S cells. OD: optical density. The absorbance is reflective of the number of migrating or invading cells. ^*P < 0.01. Data are representative of four experiments with duplicate samples. Western blot analysis for phospho- and total-MAPK (D) and phospho- and total-Akt (E) in serum-starved EpH4 cells.

3.3. T72A and T72S mutants activate Nodal-independent signaling pathways in EpH4 cells

We therefore investigated whether T72A or T72S Cr-1 mutants were capable of activating MAPK and Akt signaling pathways in EpH4 mammary epithelial cells. When cultured in serum-free conditions, EpH4 Cr-1, T72A or T72S showed an approximately equal two-fold increase in the phosphorylation of MAPK or four-fold increase in the phosphorylation of Akt, after normalizing to total MAPK and Akt levels of expression (Fig. 2D and E) as compared to EpH4 WT cells. Similar results, with respect to cell proliferation, migration, invasion and activation of intracellular signaling molecules, were obtained in HC11 mouse mammary epithelial cells expressing Cr-1, T72A or T72S plasmids (data not shown).

3.4. T72A and T72S Cr-1 mutants fail to activate a Nodal-dependent signaling pathway in EpH4 and 293T cells

To confirm that the T72A and T72S Cr-1 mutants are indeed defective in their ability to activate a Nodal-dependent signaling pathway, we transfected 293T cells with a Nodal-responsive luciferase reporter construct (n2)⁷-Lux together with Cr-1 wild-type, T72A or T72S and Nodal in the presence of Alk4 and FAST [13]. A 10-fold increase in luciferase activity was observed in 293T cells co-transfected with wild-type Cr-1 and Nodal compared to control cells (P < 0.01) (Fig. 3A). In contrast, the activity of both T72A and T72S Cr-1 mutants was significantly reduced compared to Cr-1 wild-type expressing cells (P < 0.01), confirming that Thr⁷² in mouse Cr-1 is required for Cr-1-dependent Nodal signaling (Fig. 3A). We also tested whether T72A and T72S Cr-1 mutants were able to induce together with Nodal activation of Smad-2 in EpH4 mammary epithelial. A Nodal expression vector was transiently transfected into EpH4 cells, which lack Nodal expression [4], and the levels of phosphorylated Smad-2 were examined by Western blot analysis. The phospho-Smad-2 signal was enhanced by approximately 2.5-fold in EpH4 Cr-1 cells that were transiently expressing Nodal, as compared to EpH4 WT cells transfected with Nodal alone (Fig. 3B and C). In contrast, Smad-2 phosphorylation was strongly reduced in EpH4 T72A or T72S cells that were expressing Nodal (Fig. 3B and C).

Fig. 3.

T72A and T72S Cr-1 mutants fail to activate a Nodal-dependent signaling pathway in 293T and EpH4 cells. (A) Luciferase reporter assay in 293T cells transiently transfected with (n2)₇-Lux reporter, FAST-1, Nodal, Alk4 expression vectors and Cr-1, T72A or T72S plasmids. ^*P < 0.01, compared with control; ^**P < 0.01, compared to Cr-1. Data are representative of three experiments with triplicate samples. RLU: relative luciferase units. (B) Western blot analysis for phospho- and total-Smad-2 in EpH4 WT, Cr-1, T72A and T72S cells transiently transfected with Nodal. (C) Fold difference in phospho-Smad-2 expression normalized to total-Smad-2 content by densitometric analysis. Densitometric analysis is representative of three Western blot experiments. ^*P < 0.05, compared with control.

3.5. T72A and T72S Cr-1 mutants bind to Glypican-1 in COS7 cells

Because we have previously demonstrated that Cripto-1 activates a Nodal and Alk4-independent signaling pathway through Glypican-1 [7], we evaluated whether the T72A and T72S Cr-1 mutants interact with Glypican-1. Wild-type, T72A or T72S Cr-1 proteins were all found to interact to a similar extent with Glypican-1 in a coimmunoprecipitation assay in COS7 cells (Fig. 4A). Western blot analysis of the cell lysates for Cr-1 or Glypican-1 confirmed expression of the transfected plasmids in COS7 cells (Fig. 4B). To ascertain if Glypican-1 is required by Cripto-1 to induce activation of intracellular signaling molecules in mouse mammary epithelial cells, EpH4 WT, Cr-1, T72A or T72S cells were transfected with a specific Glypican-1 siRNA or with a control non-silencing siRNA and analyzed by Western blot for activation of MAPK and Akt signaling pathways. As shown in Fig. 4C, transfection of Glypican-1 siRNA strongly reduced endogenous Glypican-1 mRNA expression in EpH4 WT, Cr-1, T72A or T72S cells as compared to cells transfected with a control siRNA, as assessed by real-time PCR. Furthermore, Western blot analysis showed a strong reduction in the phosphorylaion of MAPK (Fig. 4D) and Akt (Fig. 4F) in serum-starved EpH4 Cr-1, T72A or T72S cells transfected with Glypican- 1 siRNA as compared to cells transfected with a control siRNA (Fig. 4E and G).

Fig. 4.

Glypican-1 is required by wild-type, T72A or T72S Cr-1 to activate MAPK and Akt signaling molecules in EpH4 cells. (A) Immunoprecipitation for Glypican-1 in COS7 cells transiently transfected with Glypican-1-Fc alone or in combination with Cr-1, T72A or T72S and Western blot analysis with an anti-Cr-1 antibody or anti-human Fc antibody. (B) Western blot analysis for Cr-1 and Glypican-1 on cell lysates of COS7 transfected cells. (C) Real-time PCR for Glypican-1 in EpH4 WT, Cr-1, T72A or T72S cells transfected with control or Glypican-1 siRNAs. Data were normalized to 28S mRNA expression. Data are representative of two experiments with duplicate samples. ^*P < 0.05. Western blot analysis for phospho- and total-MAPK in serum-starved EpH4 cells transfected with Glypican-1 siRNA (D) or control siRNA (E). Western blot analysis for phospho- and total-Akt in serum-starved EpH4 cells transfected with Glypican-1 siRNA (F) or control siRNA (G).

4. Discussion

Several potential N-linked and O-linked glycosylation sites have been identified within the protein structure of Cripto-1 [1]. O-fucosylation is a rare type of glycosylation that occurs on Ser or Thr residues in the context of EGF repeats [8]. Cripto- 1 is fucosylated at a conserved Thr residue in the EGF-like domain and a single point mutation in the fucosylation consensus sequence results in loss of fucosylation and inability to facilitate Cripto-1-dependent Nodal signaling [9]. However, subsequent studies revealed that fucosylation is not required by Cripto-1 to function in a Nodal signaling assay [10]. For example, embryos that are null for GDP-fucose protein O-fucosyltransferase- 1 (Pofut1) expression, which is the enzyme responsible for the addition of O-fucose to EGF repeats, die at day 9.5, two days later than Cr-1 null embryos [14]. In addition, Cr-1, although not fucosylated in Pofut1^−/− embryonic stem (ES) cells, can still activate a Nodal-dependent signaling pathway [10]. Furthermore, Pofut1^−/− or T72A mutant ES cells can also differentiate into beating cardiomyocytes, a process that requires a functional Cripto-1/Nodal signaling pathway [10,15]. In contrast, the Thr residue to which O-fucose is attached is absolutely required for Cr-1 function in a Nodal-dependent signaling assay. In fact, site directed mutagenesis of Thr⁷² with eight other amino acids or substitution with a Ser residue, that can bind O-fucose, give rise to Cr-1 mutants with a strongly impaired activity in Nodal assays. However, all these studies lack any description of activity of Cripto-1 fucosylation or Thr mutants in a Nodal-independent signaling pathway(s). Cripto-1 can function as an oncogene promoting cellular proliferation and transformation in vitro and in vivo mainly through activation of the ras/raf/MAPK and PI3K/Akt signaling pathways [1,2,7]. In fact, treatment of mammary epithelial or endothelial cells with specific MAPK or PI3K inhibitors strongly interferes with the ability of Cr-1 to enhance proliferation, migration and invasion of these cells [4,6,7]. On the contrary, an Alk4 inhibitor has very little effect on the migratory behavior of endothelial cells, suggesting that these biological effects of Cr-1 are mainly dependent on activation of a Nodal-independent signaling pathway [6]. In the present study we demonstrate for the first time that mutation of Thr⁷² in mouse Cr-1 give rise to a fully functional protein with respect to activation of a Nodal-independent signaling pathway, whereas Thr⁷² is still required for the Cripto-1 Nodal-dependent pathway. Because the EGF-like domain of Cr-1 directly interacts with Nodal, Thr⁷² is required for binding of Cr- 1 to Nodal [11]. Interaction of Cr-1 with Glypican-1 might be also mediated by the EGF-like domain, since a peptide corresponding to the EGF-like domain of Cr-1 results in activation of MAPK signaling pathway downstream of Glypican-1 [1]. However, Thr⁷² mutation does not appear to perturb binding of Cr-1 to Glypican-1 (Fig. 4A) and the subsequent activation of MAPK and Akt signaling pathways in EpH4 cells (Fig. 2D and E). Probably, different amino acid residues within the EGF-like domain mediate interaction of Cr-1 with Glypican- 1. Moreover, we demonstrate that Glypican-1 is absolutely required by both wild-type and T72A or T72S mutant Cr-1 proteins to fully induce activation of MAPK and Akt signaling pathways in EpH4 cells (Fig. 4D and F). Parisi and coauthors have shown that glycine 71 and phenylalanine 78 in the EGF-like domain and thryptophan 107 in the CFC domain of Cr-1 are required for cardiac differentiation of ES cells through a Nodal-dependent signaling pathway [15]. It would be informative to assess if these Cr-1 mutants are also essential for interaction with Glypican-1 and activation of a Nodal-independent signaling pathway. In this regard, a recent report by D’Andrea et al. has demonstrated that Cr-1 mouse embryos carrying the amino acid substitution phenylalanine 78 to alanine (F78A), which fail to stimulate a canonical Nodal-dependent pathway, are able to establish an A/P axis, initiate germ layer formation and gastrulation movements, unlike Cr-1 null embryos, probably through activation of a Nodal-independent signaling pathway [16]. In conclusion, identification of amino acid residues within Cripto-1 protein with different binding affinities for signaling partners might be useful in defining the mechanism by which Cripto-1 can activate in a selective manner a Nodal-dependent versus a Nodal-independent pathway leading to a variety of biological responses.