GPC3 Decreases IGF-1R–Grb10 Interaction and Promotes Cell Invasion in Hepatocellular Carcinoma in a Gender-Dependent Manner


 Background: Glypican-3 (GPC3) mRNA was more frequently overexpressed in women and patients with invasive HCC. We explore possible molecular mechanisms of the effect of GPC3 on growth factor receptor-bound protein 10 (Grb10) and insulin-like growth factor 1 receptor (IGF-1R) interaction of tumor invasion in women. Methods: For in vitro experiments, GPC3 and pertinent mutants were transfected, and Western blotting (HEK293T cells), confocal microscopy (HeLa and PLC-PRF-5 cells), luciferase assays for AP-1 reporter activities (NIH3T3 and HuH-7 cells), gelatin zymography (PLC-PRF-5 cells) and cell culture in 3D collagen I gels (NIH3T3 and R- cells) were performed. For in vivo experiments, GPC3 and IGF-1R coexpression was evaluated in hepatocellular carcinoma clinical samples. Results: We found that interaction of IGF-1R with Grb10 was hindered by GPC3, and GPC3 causes IGF-1R colocalization with Grb10 to a lesser extent after IGF-1 stimulation; moreover, it promoted IGF-1-stimulated AP-1 activation and matrix metalloproteinase -2 and 9 (MMP-2 and MMP-9) secretion in vitro, which seemingly play a role in tumor invasion or recurrence. Further, gender differences existed among patients with hepatocellular carcinoma in terms of GPC3 and IGF-1R coexpression in vivo.Conclusions: We believe that a more intensive surveillance of GPC3 expression in female patients with hepatocellular carcinoma should contribute to the prediction of recurrence, and this may highlight new strategies for treating hepatocellular carcinoma in women.


Introduction
Hepatocellular carcinoma (HCC) is the second leading cause of death due to malignancy worldwide (1).
Gender differences in the incidence of HCC are evident, with the male:female ratio ranging from 2:1 to 4:1 (2,3). Gender differences have also been reported in HCC recurrence among recipients of a liver transplant (LT), with the incidence being nearly three times higher in women than in men among those with high α-fetoprotein (AFP) levels at the time of transplantation (4). We previously reported that glypican-3 (GPC3) mRNA was more frequently overexpressed in women and patients with invasive HCC (5), demonstrating that higher levels of insulin-like growth factor 1 (IGF-1) receptor (IGF-1R) were found in GPC3-expressing cells because GPC3 decreased IGF-1-induced IGF-1R ubiquitination and degradation (6).
Glypicans are a family of heparan sulfate proteoglycans linked to the exocytoplasmic surface of the plasma membrane by a glycosylphosphatidylinositol anchor (7). Glypicans act as coreceptors, facilitating the formation of ligand-receptor complexes and effectively lowering the required concentration of ligands (8). Loss-of-function mutations in GPC3 cause Simpson-Golabi-Behmel syndrome, a condition that is characterized by prenatal and postnatal overgrowth and an increased risk of embryonal tumor development (9). Studies have also demonstrated that GPC3 plays a key role in cancer development (7,10,11). GPC3 regulates the signaling activity of various morphogens, including Wnts, Hedgehogs, bone morphogenetic proteins and broblast growth factors (12)(13)(14). GPC3 reportedly promotes the in vitro and in vivo growth of HCC cells by interacting with the Wnt ligand to facilitate Wnt/Frizzled binding (10). Furthermore, GPC3 seems to be both a serum marker (15) and a therapeutic target for HCC (16)(17)(18).
We previously demonstrated that GPC3 binds to IGF-2 and IGF-1R through its N-terminal proline-rich domain, causes the phosphorylation of IGF-1R and extracellular signal-regulated kinase (ERK) and induces oncogenicity (11). Further, GPC3 enhances IGF-1R signaling, leading to ERK phosphorylation, c-Myc expression and increased oncogenicity (6). GPC3 binds to and potentially sequesters growth factor receptor-bound protein 10 (Grb10), thereby blocking IGF-1R ubiquitination and degradation in the proteasome. Given that GPC3 is located in chromosome Xq26 and is more frequently overexpressed in women and patients with invasive HCC (5), we herein further investigated whether IGF-1R is associated with the mechanism of invasion by GPC3 and whether gender differences exist in GPC3 and IGF-1R coexpression in HCC.

Tissues and cells
Surgically resected liver tissue specimens were obtained from the Department of Pathology, National Taiwan University Hospital and Keelung Hospital, Ministry of Health and Welfare. These specimens were used in accordance with appropriate regulations and with approval by the Institutional Review Board of the Ethics Committees of the NTUH (200701095R) and KLH (TYGH100038). The HCC cell lines used were PLC-PRF-5 (CRL-8024) cells, which were purchased from ATCC (VA, USA), and HuH-7 cells (JCRB-0403) (19), which were from JCRB (Osaka, Japan). HEK293 (CRL-1573) and HEK293T cells were obtained from ATCC (CRL-3216). Prior to use, reauthentication of these cells was performed by Short Tandem Repeat (STR) DNA pro ling analysis. NIH3T3 cells (BCRC-60008) were purchased from BCRC in 2003 (Hsinchu, Taiwan); NIH3T3 is a nonhuman cell line, and the morphology and growth speed of these cells have not changed since we obtained them (data not shown). HEK293T and NIH3T3 cells were used for transient transfection and stable clone selection. HeLa cells (BCRC-60005), which were authenticated through STR DNA pro ling analysis, were purchased from BCRC in 2014. They were used for immuno uorescence staining within 6 months of cell resuscitation. Cells were cultured in Dulbecco's modi ed Eagle's medium supplemented with 10% fetal calf serum. G-418 (Promega, Fitchburg, WI, USA) was used for the selection of stable clones.

Confocal microscopy
After transfection, cells were cultured in a serum-free medium, stained with IGF-1R (CD221, ThermoFisher, USA) antibody and stimulated with 50 ng/ml of IGF-1 for 10 min. They were then treated with 0.5% Triton X, xed in 4% paraformaldehyde and stained with anti-Grb10 (Santa Cruz Biotechnology). The secondary antibodies were conjugated with Alexa 488 (green) for IGF-1R and Alexa 594 (red) for Grb10, and images were captured using a confocal laser scanning microscope.
Cell culture in 3D collagen I gels We added the desired amount of collagen I (Corning, 354249) to obtain a nal collagen I concentration of 2.3 mg/ml in a solution containing 10% 10 × minimal essential medium, 1% fetal calf serum (Sigma-Aldrich), and distilled water. The pH was adjusted to 7.4 with 1 N NaOH. Cells (2.5 × 10 5 cells/ml) were added to the aforementioned solution before it solidi ed. The mixture was then immediately transferred to a 96-well plate and allowed to solidify at 37 °C. After 30 min, 100 µl of normal complete culture medium was added on top of the cells and the collagen I gel mixture. Cells were cultured in the 3D collagen I gel for 3 d.
Luciferase assays for AP-1 reporter activities Cells were transiently transfected with a combination of plasmids, as described in the legend of Fig. 4. After overnight transfection, they were treated with serum-free media for 48 h. Fire y luciferase (Promega) and Renilla luciferase (Applied Biosystems) assays were performed according to manufacturer instructions. For each experiment, transfections were performed in triplicate. Values represent means ± SD of at least three independent experiments. All values were normalized for transfection e ciency (Renilla luciferase activity).
Gelatin zymography PLC-PRF-5 cells were cultured in serum-free medium for 3 d. The conditioned media with 8 µg of proteinafter spin concentration-was mixed with nonreducing sodium dodecyl sulfate gel sample buffer and applied without boiling to a 10% polyacrylamide gel containing 0.1% sodium dodecyl sulfate and 1 mg/ml gelatin solution. After electrophoresis, the gels were washed with 50 mmol/l Tris-HCl (pH 7.5) containing 0.15 mol/l NaCl, 5 mmol/l CaC1 2 , 5 µmol/l ZnCl, 0.02% NaN 3 and 0.25% Triton X-100 (three changes) at room temperature and then incubated in the same buffer without Triton X-100 (two changes) at 37 °C for 20 h. Proteins were stained with Coomassie Brilliant Blue R-250.

Results
Gender differences in the coexpression of GPC3 and IGF-1R in HCC We previously found that GPC3 mRNA was more frequently overexpressed in female than in male patients with HCC (5) and later con rmed this nding through an immunohistochemistry study (6), wherein we observed that GPC3 expression was highly correlated with IGF-1R expression in HCC specimens (6). Herein, we studied 126 patients with HCC (98 men and 28 women) to determine if gender differences are evident in GPC3 and IGF-1R coexpression. GPC3 expression was observed more frequently in women than in men ( Fig. 1A and B, 92.8% vs. 52%), and GPC3 and IGF-1R coexpression was found to be correlated with gender (Table 1 and Fig. 1C, p = 0.0001 by Fisher's exact test). Interestingly, GPC3 and IGF-1R were more often coexpressed in women (100%) than in men (57.1%), with AFP levels > 320 ng/ml (5). Less interaction between Nedd4, Grb10 and IGF-1R in the presence of GPC3 We previously demonstrated that GPC3 coimmunoprecipitated with Grb10 but not with Nedd4 (6). Herein we stimulated HCC cells with IGF-1 and found that GPC3 decreased the ability of Grb10 to pull down IGF-1R by approximately 0.5-fold (Fig. 2, lane 4). Nedd4 also weakly bound to Grb10 in GPC3-expressing cells stimulated with IGF-1 (Fig. 2, lane 4). It is therefore likely that GPC3 bound to and sequestered Grb10, thereby preventing its binding to IGF-1R. In HeLa and PLC-PRF-5 cells, IGF-1R was internalized to a lesser extent and was colocalized with Grb10 after stimulating GPC3-expressing cells with IGF-1 ( Fig. 3A and   3B). As control, IGF-1R was colocalized with Grb10 in the presence of the GPC3 mutant P26-30A (the GPC3 mutant cannot bind to IGF-1R or be internalized and colocalized with Grb10 after IGF-1 stimulation; Fig. 3A) or vector control ( Fig. 3A and 3B) upon stimulation with IGF-1.
Using an AP-1 luciferase reporter, we demonstrated that GPC3 overexpression increased luciferase activity in NIH3T3 cells by up to 2.5-fold in serum-free conditions. Moreover, IGF-1 treatment further increased the luciferase activity in GPC3-overexpressing NIH3T3 cells, suggesting that the effects of GPC3 and IGF-1 are additive (Fig. 4A). HuH-7 cells expressed high levels of endogenous GPC3, and therefore, the effect of GPC3 overexpression on AP-1 activity was minimal (Fig. 4B). By inducing the expression of ∆GPC3 (the deleted GPC3 mutant cannot protect IGF-1R from degradation) in HuH-7 cells, we demonstrated that ∆GPC3 not only decreased the basal levels of AP-1 activity in the cells but also inhibited the effects of IGF-1 stimulation (Fig. 4B).

GPC3 increased MMP-2 and MMP-9 expression
We previously found GPC3 mRNA was more frequently overexpressed in invasive HCC (5). The AP-1 binding site is reportedly a direct regulator of MMP activity (22),, which is critical for the invasive potential of tumors and is a predictor of tumor recurrence and survival in patients with HCC after surgical resection (23). Increased production of MMP-2 has been observed in cells overexpressing IGF-1R (24) and via the phosphoinositide 3-kinase (PI3-K) and mitogen-activated protein kinase (MAPK) pathways (25).
In the case of PLC-PRF-5 cells, more MMP-2 and MMP-9 were found in the conditioned media of GPC3expressing cells than in the conditioned media of the vector control or GPC3 mutant cells (Fig. 5A). In addition, more activated ERK was found in PLC-PRF-5 cells expressing GPC3 than in the vector control or GPC3 mutant cells (Fig. 5B). Similarly, increased p-ERK expression was observed in GPC3-expressing NIH3T3 cells (Fig. 6A), and 3D collagen I gels exhibited more invasive outgrowth with dramatic invadopodia (Fig. 6C). By contrast, GPC3 did not activate ERK (Fig. 6B) or promote the outgrowth of invadopodia in R-cells (IGF-1R knockout mouse embryonal broblasts; Fig. 6D). These observations suggest that GPC3 plays a role in the promotion of invasive properties through IGF-1R.

Discussion
In this study, we demonstrated that GPC3 causes IGF-1R colocalization with Grb10 to a lesser extent after IGF-1 stimulation and it promotes more IGF-1-stimulated activation of AP-1 and more secretion of both MMP-2 and MMP-9, all of which may play a role in tumor invasion or recurrence. In addition, gender differences were evident in GPC3 and IGF-1R coexpression in HCC clinical samples.
GPC3 was also found to decrease the ability of Grb10 to pull down IGF-1R and Nedd4 (an E3 ubiquitin ligase). Similar ndings have been reported for another E3 ligase, namely MDM2, which was observed to bind to isoforms of p73 (TAp73α and ΔNp73α) and to a much lesser extent to isoforms of p63 (TAp63α and ΔNp63α). Further, MDM2 was noted to inhibit the transcriptional activity of the isoforms of p73 but to have almost no effect on those of p63 (26). Therefore, less IGF-1R interaction with Grb10 and Nedd4 caused by GPC3 seems to protect IGF-1R from degradation and enhance IGF-1R signaling.
The extracellular MMP-9 and MMP-2 are critical for the invasive potential of tumors and for the prediction of tumor recurrence and survival in patients with HCC after surgical resection (23). In the prostate cancer cell line DU145, IGF-1 has been shown to regulate MMP-2 and MMP-9 activity and expression via the PI3-K and MAPK pathways, and the invasive capacity of DU145 cells was inhibited by blocking IGF-1R (25).
In GPC3-expressing cells, increased levels of IGF-1R, p-ERK, MMP-2 and MMP-9 and higher invasive capacity were found. By contrast, the invasive capacity or ERK activation of GPC3 was eliminated in IGF-1R knockout R-cells. Therefore, IGF-1R is clearly essential for GPC3-induced invasion.
In the present study as well as in previous studies, GPC3 and IGF-1R were found to be more frequently coexpressed in female patients with HCC; moreover, gender differences regarding their role in HCC were statistically signi cant. GPC3 was more frequently overexpressed in women and closely correlated with elevation of serum AFP levels (5). IGF-1R mediates cancer invasion (27), and risk of tumor recurrence (28). HCC recurrence in LT recipients was reported to be nearly three times higher in women than in men with high AFP levels (> 100 ng/dl) at LT (4), and this correlates the nding of gender differences in the coexpression of GPC3 and IGF-1R in HCC samples with high AFP levels in this study.
In conclusion, we found evidence of gender differences in the coexpression of GPC3 and IGF-1R in HCC and that GPC3-expressing HCC cells exhibit higher MMP-2 and MMP-9 secretion. Considering that such substantial gender differences exist in HCC in terms of GPC3 expression and its effects, a subset of female patients with HCC may bene t from more intensive surveillance of GPC3 expression. Different strategies to target GPC3 for the effective treatment of HCC in women are at present in development. The human tissue involved in the study were de-linked and approved by Institutional Review Board of the Ethics Committees of the NTUH (200701095R) and KLH (TYGH100038). There was no animal study. All methods were carried out in accordance with relevant guidelines and regulations.

Consent for publication
Not applicable.

Availability of data and material
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Competing Interests
None declared.