If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
To whom correspondence should be addressed: Division of Molecular and Cellular Biology Research, Research Bldg., S-218, 2075 Bayview Ave., Toronto, Ontario M4N 3M5, Canada. Tel.: 416-480-6100 (ext. 3350);
* This work was by the Canadian Institute of Health Research. 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.
Loss-of-function mutations of the GPC3 gene are the cause of the human Simpson-Golabi-Behmel syndrome. Based on the overgrowth phenotype of the Simpson-Golabi-Behmel syndrome patients and the key role played by the insulin-like growth factor (IGF) signaling system in regulating embryonic growth, it was speculated that GPC3 regulates IGF signaling. In order to test the validity of this hypothesis, we mated GPC3 knockout mice with insulin receptor substrate-1 (IRS-1) nullizygous mice. We found that GPC3 regulates organism growth independent of IRS-1, suggesting that GPC3 does not modulate IGF signaling. Instead, we found that GPC3 knockout mice exhibit alterations in the Wnt signaling pathway, which is also associated with the regulation of cell proliferation. In particular, the loss of GPC3 led to the inhibition of the non-canonical Wnt/JNK signaling pathway, while concomitantly causing the activation of canonical Wnt/β-catenin signaling. These in vivo findings were confirmed in vitro upon the ectopic overexpression of GPC3 in mesothelioma cells. In these cells, the GPC3-induced increase in JNK activity was associated with an enhanced response to Wnt5a. Most interestingly, the heparan sulfate chains of GPC3 were not required for its stimulatory activity on Wnt5a signaling and for the formation of GPC3-Wnt5a complexes. We propose that at least in some cell types GPC3 serves as a selective regulator of Wnt signaling, by potentiating non-canonical Wnt signaling, while inhibiting the canonical Wnt signaling pathway.
Glypicans are cell-surface heparan sulfate proteoglycans that are bound to the exoplasmic surface by a glycosylphosphatidylinositol anchor (
). All glypicans have similar structural domains, including a large, N-terminal globular cysteine-rich domain and a cluster of C-terminal heparan sulfate glycanation sites, near the point of membrane attachment (
). The involvement of glypicans in Wnt signaling has also been recently corroborated in Zebrafish and Xenopus. Specifically, glypicans appear necessary for cellular movements during gastrulation, which are known to be regulated by non-canonical Wnt signaling (
The exact mechanism by which glypicans regulate Wnt activity is still under study, but there is experimental evidence to suggest that glypicans work at the level of signal reception by facilitating the interaction between Wnts and their signaling receptors Frizzleds (
). This was largely based on the observation that loss-of-function mutations of GPC3 can lead to enhanced organism growth in utero. In particular, the loss of GPC3 has been causally associated with the Simpson-Golabi-Behmel syndrome, a pediatric condition characterized by somatic overgrowth (
Although the evidence to date suggests that GPC3 does not directly regulate IGF signal transduction at the level of ligand-receptor interactions, it still remains possible that GPC3 could stimulate a pathway that interacts with components of the IGF signaling system downstream of the receptors. In particular, the insulin receptor substrate-1 (IRS-1), a key component of the IGF signaling pathway, can be activated by molecules that are not considered intrinsic components of such a signaling pathway (
Here we provide genetic and molecular data that demonstrate that GPC3 regulates organism size by a mechanism that is independent of IRS-1. Instead, we report that GPC3 regulates Wnt signaling by potentiating non-canonical Wnt signaling, both in vivo and in vitro, while concomitantly inhibiting canonical Wnt signaling.
Mouse Strains, Genetic Crosses, and Tissue Collection—The generation of mice carrying targeted deletions in either GPC3 or IRS-1 was described previously (
). All mice were maintained in a C57Bl/6 background from at least the 5th backcross. To assess the effect of the lack of GPC3 on the body size of IRS-1-null mice, at least five newborn mice of each genotype were compared. For tissue collection, whole embryos were decapitated prior to snap-freezing on dry ice. All tissues were then stored at -80 °C until use. Newborn pups were collected prior to noon of the day of parturition and weighed. Unless otherwise indicated, at least two embryos with the same genotype were pooled for Western blot analysis or for the measurement of enzymatic activity. Experiments were performed at least twice, and a representative blot is shown.
Expression Vectors—The pEF-BOSGPC3 and pEF-BOSΔGAG constructs were generated from a GPC3 cDNA, as described previously (
). The pGPC3IRES2EGFP and pΔGAGIRES2EGFP constructs were subsequently created by subcloning the EcoRI-AccI fragment containing the human GPC3 cDNA from pEFGPC3 and the EcoRI-SacII fragment containing the GPC3ΔGAG cDNA from pEFΔGAG into the EcoRI-AccI and EcoRI-SacII sites of pIRES2EGFP (Clontech), respectively. The placZIRES2EGFP plasmid was created by sub-cloning the galactosidase cDNA from pINDlacz (Invitrogen) into the EcoRI site of pIRES2EGFP. The pLNCXWnt5a(HA) plasmid containing an hemagglutinin A (HA)-tagged Wnt5a cDNA was a kind gift from Dr. J. Kitajewski. The pRC/CMVTGFα plasmid was generated by subcloning the HindIII-XbaI TGFα cDNA from pRC/CMVTGFα(as) (kindly provided by Dr. B. Ziober) into the HindIII site of pRC/CMV (Invitrogen) by using HindIII linkers.
Cell Culture—II14 cells and II14 cells stably expressing rat GPC3 (
) were cultured in Dulbecco's modified Eagle's medium/F-12 supplemented with 10% fetal calf serum, 1% glutamine/penicillin/streptomycin, and hydrocortisone, insulin, holotransferrin, selenium (HITS), whereas 293T cells, L cells overexpressing Wnt3a and Wnt5a (ATCC), and Rat-2 cells overexpressing Wnt-1 (kind gift from Dr. M. Semënov) were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 1% glutamine/penicillin/streptomycin. All cells were grown in a 37 °C incubator in the presence of 5% CO2.
The pooled population of II14 cells stably expressing LacZ, GPC3, and ΔGAG were generated by transfecting II14 cells with placZIRES2EGFP, pGPC3IRES2EGFP, and pΔGAGIRES2EGFP respectively, using Lipofectin, as described previously (
). After 3 weeks of selection with 600 μg/ml of geneticin (Invitrogen), colonies were pooled together and subjected to two rounds of fluorescence-activated cell sorting based on green fluorescent protein (GFP) fluorescence. Expression of GPC3 was confirmed by Western blotting with the 1G12 antibody (
Wnt-containing conditioned media from L and Rat-1 cells were collected from cultures grown to confluence, centrifuged at 1,000 rpm for 10 min, snap-frozen on dry ice, and stored at -80 °C until use. To stimulate cells with Wnt-containing conditioned media, cells were incubated in serum-free media for 1 h prior to the addition of the media for up to 3 h at 37 °C. Stimulated cells were quickly washed with ice-cold phosphate-buffered saline (PBS) and lysed for β-catenin and phosphoprotein analyses, as described below.
Analysis of Cytoplasmic β-Catenin Levels—E11.5 decapitated embryos were suspended in a hypotonic buffer (20 mm Hepes, pH 7.6, 1.5 mm MgCl, 10 mm KCl, 5 mm sodium orthovanadate, 25 mm tetrasodium pyrophosphate, 50 mm sodium fluoride, 1 mm phenylmethanesulfonyl fluoride, 50 mm okadaic acid, 10 μg/ml leupeptin, 10 μg/ml aprotinin, 1 μg/ml pepstatin, 2 mm benzamidine hydrochloride) for 10 min. Embryos were then briefly centrifuged, resuspended in 800 μl of homogenizing buffer (20 mm Hepes, pH 7.6, 1.5 mm MgCl, 10 mm KCl, protease inhibitors) per embryo, and homogenized with 40 strokes of a Dounce homogenizer. A fraction of the crude homogenate was quickly resuspended in an equal volume of 2× detergent lysis buffer (2% Triton X-100, 0.2% SDS, 200 mm NaCl, 2 mm EDTA, protease inhibitors, phosphatase inhibitors) and incubated on ice for 30 min. Any insoluble material was removed by centrifugation at 15,000 rpm for 15 min at 4 °C. The supernatant was collected (total fraction) and protein concentration determined. The remainder of the crude homogenate was centrifuged at 2,000 rpm for 10 min at 4 °C. Cytoplasmic fractions were obtained by ultracentrifugation of the cleared homogenates at 28,000 rpm for 1 h at 4 °C. The resulting supernatant was then collected and protein concentration determined.
In experiments with tissue culture cell lines, cytoplasmic extracts were obtained using saponin, as described previously (
). Briefly, cells were lysed on the dish with 250 μl of saponin lysis buffer (25 mm Hepes, 75 mm potassium acetate, 0.1% saponin, phosphatase inhibitors, and protease inhibitors). Cells were extracted twice, pooled, and then centrifuged at 15,000 rpm for 15 min at 4 °C to remove insoluble materials. The resulting supernatants were collected, and their protein concentration was determined.
All samples were resuspended with an equal volume of 2× SDS-PAGE loading buffer. Proteins were resolved by SDS-PAGE under reducing conditions, transferred, and then analyzed by Western blotting with an anti-β-catenin antibody (Signal Transduction Laboratories), followed by an anti-mouse horseradish peroxidase-conjugated antibody (StressGen). Protein bands were detected using the ECL Reagent (Amersham Biosciences). Protein loading was determined by probing membranes with either an anti-actin (Sigma) or an anti-protein kinase B (PKB) antibody (Cell Signaling).
Analysis of Phosphoproteins—Due to the limited amount of tissue available for analysis, embryonic extracts were made using the Trizol reagent, as per the manufacturer's instructions. Briefly, an E11.5 embryo was homogenized with 20 strokes of a tight-fitting pestle in 1 ml of Trizol (Invitrogen), and incubated on ice for 15 min. Following the addition of 200 μl of chloroform and the separation of the aqueous phase by centrifugation, the organic phase was extracted with 300 μl of ethanol and centrifuged at 3,000 rpm for 5 min at 4 °C. Proteins were then precipitated from the supernatant with 3 volumes of pre-cooled acetone (-20 °C) and stored at -20 °C for 15 min. Following centrifugation at 15,000 rpm for 5 min at 4 °C, the protein pellet was washed twice with an 80% ethanol/water mixture at room temperature for 10 min. Pellets were air-dried at room temperature and redissolved in a resuspension buffer (1% SDS, 1 mm dithiothreitol) prior to protein quantification.
Protein extracts from tissue culture cell lines were obtained by lysing cells directly on dishes with 1× detergent lysis buffer (1% Triton X-100, 0.1% SDS, 199 mm NaCl, 1 mm EDTA, proteinase inhibitors, phosphatase inhibitors) for 30 min on ice. Crude lysates were collected by scraping and centrifuged at 15,000 rpm for 15 min to remove insoluble material. The supernatants were then collected, and protein concentration was determined.
All protein samples were resolved by SDS-PAGE, transferred to a membrane, and then subjected to Western blotting with either an anti-phospho-Jun kinase (JNK) antibody (Cell Signaling), an antibody against phospho-protein kinase C (PKC) (Cell Signaling), or an antibody directed against phospho-c-Jun (Santa Cruz Biotechnology). Protein bands were detected using a secondary horseradish peroxidase-conjugated antibody (StressGen) and the ECL Reagent (Amersham Biosciences). Total JNK, total PKC, and c-Jun levels were determined by stripping and reprobing the membranes with anti-JNK (Cell Signaling), anti-pan-PKC (Cell Signaling), or anti-c-Jun (Santa Cruz Biotechnology) antibodies.
Immunostaining—II14 cells were grown on coverslips for 2 days. Coverslips were then washed with PBS and fixed in 4% paraformaldehyde/PBS at room temperature. After rinsing the coverslips with PBS and permeabilizing the cells with 0.2% Triton X-100 in PBS plus 100 mm glycine for 15 min at room temperature, nonspecific sites were blocked with 0.5% bovine serum albumin in PBS for an additional 15 min at room temperature. The actin cytoskeleton was then stained with phalloidin-rhodamine. For Disheveled (Dsh) staining, cells were fixed for 10 min at room temperature with a 1:1 methanol/acetone mixture. After extensive washings in PBS, nonspecific sites were blocked using a 5% normal rabbit serum solution in PBS. The coverslips were then incubated for 2 h with 4 μg/ml of a goat anti-Dsh antibody (C-19, Santa Cruz Biotechnology), followed by a 1-h incubation with a fluorescein isothiocyanate-conjugated anti-goat antibody (Jackson ImmunoResearch). Nuclei were counterstained with 300 nm 4,6-diamidino-2-phenylindole for 2 min prior to mounting with anti-fade (Dako).
Analysis of Wnt Expression in II14 Cells—RNA was extracted using Trizol, and cDNA was generated using reverse transcriptase. For assessment of Wnt5a expression, the following primers were employed: 5′-CTTCCGCAAGGTGGGCGATGC-3′ and 5′-TTGCACAGGCGTCCCTGCGTG-3′. PCRs were completed in the presence of 150 nm MgCl and 200 nm dNTP using the following conditions: denaturation at 94 °C for 5 min; 30 cycles of 94 °C for 1 min, 55 °C for 1 min, 72 °C for 1 min; and elongation at 72 °C for 10 min. The expected size of the Wnt5a PCR product was 204 bp.
Co-immunoprecipitation—1 million 293T cells were plated on 35-mm dishes. The next day cells were transfected with 2 μg of each of the indicated expression vectors using Lipofectamine 2000. 48 h following transfection, cells were lysed in 1% Triton X-100, 0.5% deoxycholate, phenylmethanesulfonyl fluoride, aprotinin. For immunoprecipitation studies, 500 μg of protein lysates were diluted to 1 μg/μl protein and pre-cleaned by incubating with protein G beads for 1 h at 4 °C. Protein complexes were immunoprecipitated from cleared lysates with either 2.5 μg/ml of anti-GPC3 1G12 antibody or 10 μg/ml of the anti-carcinoembryonic antigen (CEA) J22.4.4 antibody (kind gift from Dr. C. Stanners) overnight at 4 °C, followed by a 1-h incubation at 4 °C with protein G. Immune complexes were then collected and washed three times with lysis buffer, prior to resuspension in 1× SDS-PAGE sample buffer. Proteins from immunoprecipitated samples or 30 μg of crude cell lysates were separated by SDS-PAGE, transferred to a membrane, and subjected to Western blotting using antibodies against either GPC3 (1G12), HA (3F10, Roche Applied Science), CEA (J22.4.4), or transforming growth factor-α (TGFα) (Ab-1, Oncogene Research). All protein bands were detected using an horseradish peroxidase-conjugated secondary antibody and chemiluminescence reagents.
GPC3 Regulates Organism Size by a Mechanism That Is Independent of IRS-1—To determine whether GPC3 regulates embryonic growth through IRS-1, we first investigated whether the level of IRS-1 phosphorylation is altered in GPC3-null mice. Analysis of embryos at day 12.5 post-coitum, the earliest developmental stage at which overgrowth can be detected (
), showed no change in IRS-1 phosphorylation in the mutant mice compared with wild-type littermates (Fig. 1A). As expected from the IRS-1 results, the level of tyrosine phosphorylation of the IGF receptor-1 was not altered in the GPC3-null mice (Fig. 1A). As a complement to the molecular analyses of the GPC3 knockout mice, we also performed genetic studies in which GPC3 knockout mice were mated with IRS-1 knockout mice. We speculated that if GPC3 regulates organism size through the IGF signaling pathway, then mice deficient for both GPC3 and IRS-1 should exhibit less overgrowth than that observed in GPC3 knockout mice. By measuring newborn weights, however, we found that the loss of GPC3 results in similar overgrowth irrespective of IRS-1 levels, because GPC3-null mice displayed a 31% overgrowth over wild-type mice and a 38% overgrowth in an IRS-1-null background. Similar results were obtained when E12.5 embryos were weighed (data not shown). These results indicate that GPC3 regulates organism growth independent of IGF signaling through IRS-1.
PKB and GSK-3 Activities Are Not Altered in GPC3-null Mice—Given that GPC3 appears to regulate organism size independent of IRS-1, we decided to investigate whether GPC3 acts through molecules that are further downstream of this signaling molecule, particularly along the phosphatidylinositol 3-kinase/PKB pathway. This signaling pathway has been implicated in organism growth regulation (
). In order to assess the degree of activation of this pathway in GPC3 knockout mice, we first determined the levels of PKB phosphorylation. As shown in Fig. 1B, GPC3 knockout mice do not appear to exhibit altered levels of PKB phosphorylation relative to wild-type littermates. An alternative method for assessing the activity of PKB is to determine the levels of GSK-3 phosphorylation, one of the known targets of PKB (
). Similar to what it was found with PKB, the GPC3-null mice did not display any obvious changes in the levels of GSK-3 phosphorylation (Fig. 1B). This result was confirmed by direct measurement of GSK-3 activity, which was similar in both the GPC3 knockouts and the normal littermates (data not shown).
GPC3 Knockout Mice Exhibit Alterations in Wnt Signaling— Cyclin D1 is a critical component of the cell cycle machinery necessary for the transition from the G1 phase to the S phase. Consequently, mice deficient for cyclin D1 are smaller than their wild-type littermates (
). We therefore speculated that GPC3 knockout mice could express higher levels of cyclin D1. Indeed, as shown in Fig. 1C, the levels of cyclin D1 in the GPC3 knockout mice are about 50% higher than those of wild-type littermates. Given that this increase in cyclin D1 levels is not associated with alterations in either PKB or GSK-3 activities, and in light of the experimental evidence indicating that glypicans can regulate Wnt signaling (
), it seemed plausible that the up-regulation of cyclin D1 in the GPC3 knockout mice could be due to alterations in the canonical Wnt signaling pathway, which is known to regulate cyclin D1 expression (
). To address this question, we homogenized GPC3 knockout and wild-type embryos under detergent-free conditions, and we assessed the levels of cytoplasmic β-catenin. We predicted that if the dysregulation in Wnt signaling was the cause for the increase in cyclin D1 levels, then GPC3 knockout mice would exhibit an increase in canonical Wnt signaling, which would be reflected by an increase in cytoplasmic β-catenin levels. Indeed, we found that GPC3 knockout mice embryos contain higher levels of cytoplasmic β-catenin than embryos from wild-type mice, whereas the levels of total β-catenin remained relatively unchanged (Fig. 2A). In order to ascertain if the up-regulation of cyclin D1 in GPC3-null mice could be specifically associated with elevated levels of canonical Wnt signaling, we also determined the level of extracellular signal-regulated kinase (ERK) phosphorylation, because the activation of the ERK pathway can also increase cyclin D1 expression (
). By assessing the levels of phosphorylated ERK, we found that there are no changes in the level of ERK activity in the GPC3 knockout mice relative to wild-type littermates (data not shown). These findings suggest that the increase in cyclin D1 levels observed in GPC3 knockout mice is specifically associated with an increase in canonical Wnt signaling.
Given that glypicans have been shown to regulate both canonical or non-canonical signaling (
), it is possible that the increase in canonical Wnt signaling observed in the GPC3-null embryos could be the result of either the direct stimulation of the canonical Wnt pathway or an indirect consequence of the inhibition of a non-canonical Wnt pathway. We therefore decided to investigate whether there are alterations in non-canonical Wnt signaling in the mice lacking GPC3. In particular, we measured protein PKC and JNK activation, because these molecules are known targets of the Ca2+ and planar cell polarity (PCP) non-canonical pathways, respectively (
). Fig. 2, B and C, shows that whereas the loss of GPC3 was not associated with changes in the level of PKC phosphorylation, the GPC3 knockout embryos exhibited almost a 60% reduction in the levels of JNK phosphorylation relative to wild-type littermates. This change in phosphorylation levels was not due to changes in the amount of total JNK protein, which were not altered in the knockout embryos. These findings indicate that embryonic non-canonical Wnt signaling is inhibited in the absence of GPC3 and suggest that GPC3 is required for optimal PCP/non-canonical Wnt signaling.
GPC3 Regulates Wnt Signaling in Mesothelioma Cells—In order to better characterize the role of GPC3 in Wnt signaling, we investigated if ectopic GPC3 could lead to changes in Wnt activity in cultured cells. Based on the knockout mice results, we expected that the overexpression of GPC3 would lead to the activation of the PCP pathway, while inhibiting the canonical Wnt signaling. For these experiments we made use of the II14 mesothelioma cell line, and derivative clones stably expressing GPC3 (
). To characterize further the effect of ectopic GPC3, we compared the proliferation rate of GPC3-transfected and vector control cells in the presence of serum. We found that the GPC3-expressing clones proliferate at a slower rate, particularly at higher cell densities (Fig. 3A). This reduction in proliferation rate did not appear to correlate with contact inhibition, because the GPC3-expressing clones eventually went on to reach cell densities similar to vector control cells and even formed foci (data not shown). We also noticed that the GPC3-expressing clones have markedly different cellular morphology than the vector control cell line, as seen on phase contrast (Fig. 3B). The vector control cells appeared more spindle shaped and fibroblast-like, whereas GPC3-expressing cells had a much flatter morphology. Most intriguingly, the morphological changes induced by ectopic GPC3 were associated with changes in the organization of their actin filaments (F-actin). In particular, rhodamine-phalloidin staining revealed that the GPC3-expressing clones have a marked reduction in the number of stress fibers, and most of the F-actin appeared localized around the nuclear and plasma membranes (Fig. 3B). The F-actin arrangement at the plasma membrane was characteristically very densely arranged parallel to the cell surface. These changes in actin cytoskeleton could result from the stimulation of the non-canonical PCP pathway acting through the Rho family of small GTPases (
). However, attempts at assessing the effect of ectopic GPC3 on Rho and Rac activities by using an in vitro kinase assay have so far been unable to detect any changes (data not shown). It may be possible that the PCP pathway regulates cytoskeletal changes through other GTPases (
). As an alternative method to assessing whether the PCP pathway is stimulated by GPC3, we determined the level of JNK activation. As shown in Fig. 3C, we found that GPC3-expressing clones exhibit increased levels of JNK phosphorylation compared with vector control cells. This was further confirmed by the assessment of the phosphorylation levels of c-Jun, a downstream target of the kinase activity of JNK. Consistent with the changes observed in JNK phosphorylation, cells expressing GPC3 exhibited higher levels of c-Jun phosphorylation (Fig. 3C).
Given that GPC3 knockout mice also showed alterations in the canonical β-catenin pathway, we assessed whether ectopic GPC3 could affect canonical Wnt signaling in the mesothelioma cells. We found that GPC3-expressing clones display reduced levels of cytoplasmic β-catenin compared with the vector control cells (Fig. 3D), indicating that the canonical pathway was inhibited by GPC3 expression. Put together, the in vitro and in vivo findings suggest that GPC3 can stimulate the non-canonical PCP pathway, while concomitantly inhibiting canonical Wnt signaling.
Because changes in Dsh localization have been associated with the activation of Wnt signaling, we immunostained the GPC3-transfected and vector control II14 cells with an anti-Dsh antibody. In vector control cells Dsh was predominantly localized to the cytoplasm, with punctate vesicular-like staining (Fig. 4A), as reported previously (
). Comparatively, clones expressing GPC3 exhibited prominent perinuclear Dsh staining, with relatively little cytoplasmic localization. These changes in Dsh localization were not associated with changes in total Dsh protein levels (data not shown), suggesting that the altered localization is a reflection of changes in Dsh activity. Increased perinuclear staining of Dsh has been reported previously (
GPC3 Potentiates Wnt5a Signal Transduction—Our findings suggest that GPC3 regulates Wnt-dependent autocrine activation of non-canonical and canonical pathways. In order to determine which Wnts may be regulated by GPC3, we assessed the types of Wnts that are expressed endogenously by II14 cells using RT-PCR. We found that II14 cells produce Wnt5a, Wnt11, and Wnt2b (Fig. 5A, and data not shown). Given that Wnt5a is known to stimulate the non-canonical pathways (
), we speculated that GPC3 stimulates Wnt5a-activated non-canonical signaling. We therefore assessed the effect of ectopic GPC3 on such signaling by incubating the various II14 clones with conditioned media from Wnt5a-expressing cells. We found that in vector control clones the Wnt5a conditioned media increase JNK phosphorylation by ∼2-fold, whereas in GPC3-expressing cells Wnt5a-dependent activation of JNK was ∼11-fold above basal levels, which is almost 5-fold higher than what it was observed in control cells (Fig. 5B). These findings suggest then that GPC3 could stimulate non-canonical Wnt signaling by increasing autocrine Wnt5a activity in II14 mesothelioma cells.
Given that Wnt5a-dependent activation of the non-canonical pathway is potentially capable of inhibiting the canonical β-catenin pathway (
), we speculated that the ability of GPC3 to stimulate Wnt5a non-canonical signaling may be the underlying mechanism by which GPC3 exerts inhibitory effects on the canonical pathway. In order to address this question, we assessed the ability of the Wnt5a conditioned media to inhibit the canonical Wnt pathway in the GPC3-expressing II14 clones. We found that the Wnt5a stimulation of either vector control or GPC3-transfected clones was not able to enhance the degradation of cytoplasmic β-catenin (Fig. 5C). So far all the evidence supporting the view that non-canonical Wnt pathways are able to suppress canonical signaling has been generated under chronic Wnt stimulation through the transfection of cells, as opposed to transient stimulation with conditioned media. Therefore, as an alternative approach to answering whether the GPC3-mediated stimulation of the non-canonical Wnt pathway is sufficient to induce the inhibition of the canonical pathway in mesothelioma cells, we assessed the ability of ectopic GPC3 to regulate the response of II14 cells to canonical Wnts. We hypothesized that if GPC3 was indeed inhibiting canonical signaling by suppressing the activity of the canonical pathway through cross-talk from the non-canonical pathway, then cells stably expressing GPC3, and thus chronically activating the non-canonical pathway, should be able to suppress the stimulation of the Wnt canonical pathway by exogenous Wnt. When vector control cells were stimulated with conditioned media from Wnt3a-expressing cells, we were able to observe an approximate 5-fold induction of β-catenin stabilization (Fig. 5D). Wnt3a stimulation of GPC3-expressing clones generated a 7-fold increase in the level of cytoplasmic β-catenin, suggesting that there is no general suppression of the canonical pathway. Although it could be argued that excessive levels of exogenous Wnt3a abrogated the inhibition of the canonical pathway generated by the GPC3-induced activation of the non-canonical pathway, we observed the same phenomenon when cells were stimulated with Wnt-1-conditioned media, which was only able to induce approximately a 2-fold increase in β-catenin stability (data not shown). These results would suggest then that, at least in the II14 mesothelioma cells, GPC3 can stimulate Wnt5a-induced activation of the non-canonical PCP pathway, but that the GPC3-induced inhibition of the canonical pathway may be independent of the effect of GPC3 on non-canonical Wnt signaling.
GPC3 Stimulation of the Wnt5a Signaling Does Not Require the HS Chains—Given that glypicans are cell-surface heparan sulfate proteoglycans and that Wnts are heparin-binding peptides, it has been proposed that the HS chains are required for the glypican-induced regulation of Wnt activity (
). We decided therefore to investigate whether the GPC3-mediated stimulation of Wnt5a signaling was mediated by the HS chains. In order to tackle this question, we generated a pooled population of II14 cells stably expressing wild-type GPC3 and a GPC3 mutant that lacks HS chains (
) (Fig. 6A). These cells were then stimulated with Wnt5a conditioned media, and the level of JNK activation was assessed. Based on the degree of JNK phosphorylation, it would appear that the HS chains are not required to stimulate Wnt5a signaling, although the stimulation of such signaling may be increased when the HS are present (Fig. 6B).
Because it has been shown that glypicans can form complexes with Wnts (
), the results showing that the HS chains are not required for the GPC3-induced stimulation of Wnt5a signaling suggest that the HS chains are not necessary for the formation of a GPC3-Wnt5a complex. To test this hypothesis, we transiently transfected 293 cells with expression vectors encoding Wnt5a and GPC3 or the GPC3 mutant that lacks HS chains, and we investigated whether Wnt5a is able to co-immunoprecipitate with wild-type and mutant GPC3. Fig. 7C shows that this is indeed the case. In order to address issues of specificity, we also determined if CEA, an unrelated glycosylphosphatidylinositol-anchored cell-surface molecule, could also co-immunoprecipitate Wnt5a. Fig. 7A shows that CEA was unable to pull down any detectable amount of Wnt5a. We also assessed the ability of GPC3 to co-immunoprecipitate TGFα, a growth factor that is not related to Wnt5a and whose signaling activity, as of yet, has not been shown to be regulated by glypicans. Fig. 7B shows that TGFα was unable to co-immunoprecipitate with GPC3. Together, all this evidence suggests that GPC3 can specifically exist in a complex with Wn5a and that the HS chains are not required for the formation of such a complex.
To date, studies from invertebrates and vertebrates alike have implicated the IGF signaling system as the main embryonic growth regulatory pathway (
). This may be due in part to the fact that alterations in the IGF signaling pathway generally lead to organisms that are proportionally smaller, with limited dysmorphic features and consequently with limited impact on embryo survival (
). However, the evidence presented here showing that the overgrowth of GPC3-null mice does not require IRS-1, together with previous molecular and genetic studies, strongly suggests that GPC3 exerts its growth regulatory effects in embryos by a mechanism that is completely independent of the IGF signaling pathway (
). Moreover, our molecular analyses of GPC3 null mice indicates that there are no alterations at the level of downstream targets of IRS, notably PKB and GSK-3.
By having ruled out a role for the IGF signaling system in the overgrowth of GPC3-null mice, evidence presented in this study led us to speculate that GPC3 determines organism size through the Wnt signaling pathway. This is based on the observation that GPC3 selectively potentiates non-canonical PCP signaling while inhibiting canonical Wnt signaling in the embryo. Although a role of the Wnt signaling pathway in the regulation of body size remains to be proven, there is experimental evidence suggesting that such a pathway can regulate growth in utero (
) brings up the intriguing possibility that the stimulation of the non-canonical and/or the inhibition of the canonical Wnt signaling pathways can lead to decreases in organ and organism size. Our findings that GPC3 can stimulate non-canonical Wnt5a signaling in vitro, while inhibiting canonical Wnt signaling, would be consistent with this hypothesis. The loss of GPC3 consequently would lead to a reduction in non-canonical Wnt signaling and stimulation of the canonical signaling pathway, and hence increased organ and overall organism size.
It remains to be seen whether reductions in Wnt5a signaling underlie the overgrowth exhibited by GPC3 knockout mice. Although Wnt5a knockout mice do not exhibit reductions in organism size, it would appear that Wnt5a is important for the outgrowth of many structures (
), which may mask its effect on overall size. Moreover, given the complexity of the Wnt signaling pathway, there may be many overlapping or redundant functions in vivo. In this context, it may be possible that GPC3 may regulate non-canonical signaling by acting on various Wnts in vivo.
Given the numerous reports showing that activation of the non-canonical pathways can lead to the inhibition of the canonical pathway (
), the GPC3-dependent activation of the non-canonical pathway may lead to the inhibition of canonical Wnt signaling in the embryo. Although we cannot formally rule out such a cross-talk mechanism, the inability of GPC3 to suppress the response of II14 mesothelioma cells to exogenous canonical Wnts is inconsistent with this model. On the other hand it may be possible that our inability to observe an increase in Wnt5a-induced inhibition of canonical signaling in GPC3-transfected mesothelioma cells is due to the particular configuration of the Wnt signaling system in such cells.
The exact mechanism by which GPC3 stimulates non-canonical Wnt activity still remains to be elucidated, but we provide some evidence to suggest that GPC3 may form a complex with Wnt5a that enhances Wnt5a-dependent non-canonical signaling. Experimental evidence supporting such mechanism of action for glypicans has also been provided by others (
One important finding of this study is that GPC3 does not require the HS chains to bind to Wnt5a and to stimulate its signaling strength, although they may be required for optimal signaling. Given that Wnts are heparin-binding molecules, the traditional model has been that the HS chains of glypicans are required for Wnt signaling (
). The reason for this disparity is not clear at this time. Whether this is due to cell type-specific processing of the GPC3 core protein or the presence of other cellular proteins remains to be seen. In the latter case, it is of interest to mention that glypicans have been suggested to form a complex with the Frizzled receptors (