Published OnlineFirst April 13, 2010; DOI: 10.1158/1078-0432.CCR-10-0018

Clinical Human Cancer Biology Cancer Research Differential Functions of Growth Factor Receptor–Bound 7 (GRB7) and Its Variant GRB7v in Ovarian Carcinogenesis

Yajun Wang, David W. Chan, Vincent W.S. Liu, PM Chiu, and Hextan Y.S. Ngan

Abstract Purpose: Aberrant overexpression of growth factor receptor–bound protein 7 (GRB7) and its variant GRB7v has been found in numerous human cancers. The goal of this study was to characterize the func- tions of GRB7 and GRB7v in the ovarian carcinogenesis and to investigate the differential roles of GRB7 and GRB7v in the modulation of signaling pathways. Experimental Design: Quantitative reverse transcription–PCR, Western blot, and immunohistochem- ical analyses were used to evaluate the levels of GRB7 and GRB7v. The cellular localization, functions, and signaling pathways regulated by GRB7 and GRB7v were investigated by enforced expression of GRB7 and GRB7v. Results: Quantitative reverse transcription–PCR and Western blot analyses showed that GRB7 and GRB7v were frequently upregulated in ovarian cancer samples. The overexpressed GRB7 (P =0.009) and GRB7v (P = 0.017) were significantly correlated with high-grade ovarian cancer. Immunohistochem- ical analysis on ovarian cancer tissue array confirmed that the upregulated GRB7 was significantly corre- lated with high-grade ovarian cancer (P = 0.001). Confocal microscopy analysis showed that GRB7 and GRB7v predominately localized in cytoplasm of ovarian cancer cells, consistent with their roles as signal- ing adaptors. Enforced expression of GRB7 promoted cell proliferation, migration, and invasion, whereas GRB7v only increased cell proliferation and anchorage-independent growth ability. With the treatment of specific kinase inhibitors, we showed that both GRB7 and GRB7v promoted cell proliferation through activating extracellular signal-regulated kinase signaling, whereas GRB7 enhanced cell migration/invasion

by activating c-Jun NH2 terminal kinase signaling. Conclusions: Our studies implicate that the overexpressed GRB7 and GRB7v are associated with high- grade tumors and exert distinct tumorigenic functions through regulating different signaling pathways in ovarian cancer cells. Clin Cancer Res; 16(9); 2529–39. ©2010 AACR.

Ovarian cancer is one of the most deadly malignancies molecular pathogenesis of ovarian cancer may assist the in women (1). Approximately 70% of ovarian cancer cases development of novel therapeutic interventions for this are diagnosed at the advanced stage, and the mortality rate disease. has remained high for decades (2). It is widely accepted signaling pathways are the major regu- that genetic variations act in concert with other factors, lators of cell growth, cell division, and motility. The such as hormonal and environmental factors, to potentiate aberrant activation of these pathways attributes to the de- the risk of ovarian cancer. However, the knowledge on the velopment of many human cancers (4, 5). The human molecular basis of ovarian cancer development remains growth factor receptor–bound protein 7 (GRB7) is a mem- poor (3). Undoubtedly, a better understanding of the ber of GRB7 adaptor protein family, which also includes GRB10 and GRB14, and acts as an adaptor bridging the growth factor receptor kinase and their down- Authors' Affiliation: Department of Obstetrics and Gynecology, LKS stream signaling molecules (6–9). It is composed of sever- Faculty of Medicine, The University of Hong Kong, Hong Kong SAR, al protein binding domains, such as an amino-terminal People's Republic of China proline-rich region, followed by a RA (ras-associating) Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/). domain, a PH (pleckstrin homology) domain, a BPS (between PH and SH2 domains) motif, and a carboxyl- Corresponding Author: Hextan Y.S. Ngan, Department of Obstetrics and Gynecology, 6th Floor, Professorial Block, Queen Mary Hospital, terminal SH2 domain (7). Its variant GRB7v lacks an Pokfulam, Hong Kong SAR, People's Republic of China. Phone: 852- 88-bp exon in the COOH terminus due to alternative spic- 2255-4260; Fax: 852-2255-0947; E-mail: [email protected]. ing, resulting in the substitution of the SH2 domain by a doi: 10.1158/1078-0432.CCR-10-0018 hydrophobic sequence (10). Recent findings have reported ©2010 American Association for Cancer Research. that aberrant upregulation of GRB7 is frequently found in

www.aacrjournals.org 2529

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2010 American Association for Cancer Research. Published OnlineFirst April 13, 2010; DOI: 10.1158/1078-0432.CCR-10-0018

Wang et al.

proved by the local institutional ethics committee. Two Translational Relevance immortalized human ovarian surface epithelial (HOSE) cells, HOSE12-11 and HOSE17-1 (gifts from Prof. G.S.W. – Growth factor receptor bound protein 7 (GRB7) is Tsao, Department of Anatomy, University of Hong Kong), an adaptor molecule of signal transduction. It is fre- and ovarian cancer cell lines A2780CP, A2780S, OV2008, quently upregulated in a variety of human cancers and C13 (gifts from Prof. B.K. Tsang, Department of Ob- and plays an essential role in tumor progression. Our stetrics and Gynecology, University of Ottawa) and study here reports that not only GRB7 but also its SKOV3, OVCAR3, OVCA420, OVCA429, and OVCA433 isoform GRB7v was upregulated and correlated to (American Type Culture Collection) were used in this high-grade ovarian cancer. Intriguingly, GRB7 and study. All were cultured at 37°C in 5% CO2 in MEM or GRB7v showed different tumorigenic functions DMEM supplemented with 10% fetal bovine serum. through regulating differential signaling pathways in Plasmid and cell transfection. The pEGFP/GRB7 was con- ovarian cancer cells. These results suggest that the over- structed by subcloning the GRB7 full-length cDNA frag- expressed GRB7 and GRB7v play different roles in the ment from PKH3/GRB7 plasmid (gift from Prof. J.L. tumorigenicity of high-grade ovarian cancer. The un- Guan, Department of Cell and Developmental Biology, derstanding of their distinct functional roles and mo- University of Michigan) into pEGFP-C3 vector (Invitrogen lecular mechanisms of GRB7 and GRB7v may facilitate Life Technology). The pEGFP/GRB7v was constructed by the exploration of these proteins as novel therapeutic PCR amplification on cDNAs of normal ovarian tissues targets in ovarian cancers. using the primer pair (GRB7s, 5′-GCCTGGAGGAAGAAGA- CAAA-3′ and GRB7a, 5′-AAGTACAGGCGACCCTCCTC- 3′) and subcloned into pEGFP-C3 vector (Invitrogen). The sequences of the pEGFP/GRB7 and pEGFP/GRB7v various human cancers, such as breast cancer (11, 12), were verified by sequencing analysis. Lipofectamine 2000 esophageal cancer (13, 14), and gastric cancer (15, 16). (Invitrogen) was used for cell transfection according to the Its variant GRB7v has also been found to be overexpressed manufacturer's instructions. Stable cells overexpressing in esophageal cancer (10). Clinicopathologic studies have GFP/GRB7 and GFP/GRB7v were harvested after 14 days revealed that the overexpressed GRB7 significantly associ- of G418 selection and verified by Western blot analysis. – ates with metastatic tumor phenotype (7). Indeed, GRB7 RNA extraction and reverse transcription PCR analysis. has been shown to enhance cell migration through associ- Total RNA from the tumor tissues and the cultured cells ation with focal adhesion kinase (17, 18). These findings was prepared using the TRIzol reagent (Invitrogen). The suggest that GRB7 is involved in the progression of can- cDNA was prepared using Taqman reverse transcription cers. However, the tumorigenic roles of the overexpressed kit according to the manufacturer's instructions (Applied GRB7 and GRB7v in human cancers, such as ovarian can- Biosystems). The real-time quantitative reverse transcrip- – cer, remain unknown. tion PCR (Q-PCR) was performed using the Taqman gene In this study, we report that both GRB7 and GRB7v are expression assay kit in an ABI 7700 system (Applied Biosys- upregulated in ovarian cancers and the overexpressed tems). The primers and probes for GRB7 (Assay ID GRB7- GRB7 and GRB7v are significantly correlated with high- 1213) and GRB7v (Assay ID GRB7v-1113) were purchased grade ovarian cancer. We further provide evidence that from Applied Biosystems. Each sample was performed in the exogenous overexpression of GRB7 promotes cell pro- triplicate, and the relative expression level of each gene liferation, migration, and invasion through extracellular was normalized with glyceraldehyde-3-phosphate dehydroge- nase (Assay ID Hs99999905_m1; Applied Biosystem). signal-regulated kinase (ERK) and c-Jun NH2 terminal kinase (JNK) signaling pathways, whereas the enforced Immunohistochemical and Western blot analyses. Immu- expression of GRB7v increases cell proliferation and nohistochemical staining for GRB7 and GRB7v was per- anchorage-independent growth ability via ERK signaling. formed on an ovarian cancer tissue array (OVC1021; Our findings confer that GRB7 and GRB7v function differ- Pantomics, Inc.). The section was immunostained with ently through regulating differential signaling pathways in primary polyclonal anti-GRB7 antibody (H70, Santa Cruz ovarian cancer cells. Biotechnology, Inc.) in 1:50 dilution. For negative con- trols, the primary antibody was replaced with TBS. The percentages of positive-stained cells in tumors and normal Materials and Methods epithelia were assessed. The proportions of positive cells were ranged from 10% to 100%, whereas the intensity Clinical samples and cell lines. Atotalof76tumor of staining was scored as 0 (negative), 1 (faint), 2 (mod- samples surgically resected from primary ovarian cancer erate), 3 (strong), and 4 (marked) in the most strongly patients and 53 normal ovary samples from benign dis- stained tumor area. The immunoreactivity score for each eases were randomly selected for this study. The histologic case was taken as percentage of positive cells multiplied subtypes and disease stages of the tumors were classified by the intensity of staining. according to International Federation of Gynecology and For Western blot analysis, cells were lysed with cell lysis Obstetrics criteria. The use of clinical specimens was ap- buffer (Cell Signaling Technology) containing protease

2530 Clin Cancer Res; 16(9) May 1, 2010 Clinical Cancer Research

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2010 American Association for Cancer Research. Published OnlineFirst April 13, 2010; DOI: 10.1158/1078-0432.CCR-10-0018

GRB7 and GRB7v in Ovarian Cancer

Fig. 1. Upregulation of GRB7 and GRB7v in ovarian cancer. A, real-time Q-PCR analysis showed that GRB7 and GRB7v were upregulated in ovarian cancer cell lines comparing with immortalized normal HOSEs. B, Western blotting revealed the upregulation of GRB7/GRB7v protein in the ovarian cancer cell lines. β-Actin was used as loading control. C, real-time Q-PCR analysis showed that the expressions of GRB7 and GRB7v were significantly higher in clinical ovarian cancer samples compared with normal ovaries; P = 0.016. D, immunohistochemical analysis of GRB7/GRB7v in ovarian cancer tissue array samples. Representative figures showing GRB7/GRB7v staining: a, benign (arrow, the epithelial cell layer); b, serous, Grade 1; c, serous, Grade 2; d, serous, Grade 3; e, cytoplasmic localization (arrow, cytoplasmic localization of GRB7/GRB7v). Magnification, ×200 (a–d) and ×400 (e).

inhibitor cocktail (Roche) and phenylmethylsulfonyl fluo- and AKT (Cell Signaling). Blots were then incubated with ride (Sigma Chemical Co.). The samples were resolved by goat anti-rabbit or anti-mouse secondary antibody conju- SDS-PAGE and electroblotted onto Immobilon-P Transfer gate with horseradish peroxidase (Amersham Pharmacia Membrane (Millipore Corporation). Blots were blocked Biotech) and visualized by enhanced chemiluminescence with 5% skim milk, followed by incubation with GRB7 (Amersham). (H70, Santa Cruz), GFP (Abcam, Inc.), β-actin (Sigma Confocal microscopy. Cells were cultured on coverslips Chemical Co.), phosphorylated ERK, ERK, phosphorylated and transiently transfected with either pEGFP/GRB7 or P38, P38, phosphorylated JNK, JNK, phosphorylated AKT, pEGFP/GRB7v expressing plasmids. After 24 hours, the

www.aacrjournals.org Clin Cancer Res; 16(9) May 1, 2010 2531

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2010 American Association for Cancer Research. Published OnlineFirst April 13, 2010; DOI: 10.1158/1078-0432.CCR-10-0018

Wang et al.

Table 1. GRB7 expression and clinicopathologic factors in ovarian cancer

Characteristics Total Expression P ≤3-fold >3-fold

Age (y) <55 49 25 (51.0%) 24 (49.0%) 0.705 ≥55 27 15 (55.6%) 12 (44.4%) Grade 1 and 2 27 19 (70.4%) 8 (29.6%) 0.009 3 35 13 (37.1%) 22 (62.9%) Stage Early 23 11 (47.8%) 8 (52.2%) 0.525 Late 52 29 (55.8%) 23 (44.2%) Cell types Clear cell 16 5 (31.3%) 11 (68.7%) 0.054 Others 60 35 (58.3%) 25 (41.7%) Recurrence + 37 16 (43.2%) 21 (56.8%) 0.155 − 35 21 (60.0%) 14 (40.0%)

transfected cells were fixed with 4% paraformaldehyde Anchorage-independent growth assay in soft agar. A total and treated with 0.1% triton. After the cells were counter- of 5 × 104 OVCA433 cells were suspended in 2 mL of full stained with 4′,6-diamidino-2-phenylindole (Sigma), the medium containing 0.3% agar. The mixtures were plated fluorescent signals were examined and photographed by on top of solidified bottom layer with 1% agar in full me- confocal microscopy. dium. Viable colonies with 20 cells or more were scored Cell viability analysis. Cell viability was analyzed with and photographed after 21 days. The experiment was per- cell proliferation kit II {2,3-bis[2-methoxy-4-nitro-5- formed in triplicate and was repeated thrice. sulfophenyl]-2H-tetrazolium-5-carboxanilide inner salt Wound healing assay. Equal numbers of cells were cul- (XTT); Roche} according to the manufacturer's instruc- tured in a six-well plate until confluenced monolayer cells tions. The assay was performed in triplicate for each were obtained. Cells were pretreated with mitomycin C time point and was repeated in at least three indepen- (10 μg/mL, 3 hours; Sigma), washed with PBS, and then dent experiments. wounded with a sterilized micropipette tip. The width of

Table 2. GRB7v expression and clinicopathologic factors in ovarian cancer

Characteristics Total Expression P ≤3-fold >3-fold

Age (y) <55 49 13 (26.5%) 36 (73.5%) 0.432 ≥55 27 5 (18.5%) 22 (81.5%) Grade 1 and 2 27 10 (37.0%) 17 (63.0%) 0.017 3 35 4 (11.4%) 31 (88.6%) Stage Early 23 4 (17.4%) 19 (82.6%) 0.373 Late 52 14 (26.9%) 38 (73.1%) Cell types Clear cell 16 2 (12.5%) 14 (87.5%) 0.054 Others 60 16 (26.7%) 44 (73.3%) Recurrence + 37 7 (18.9%) 30 (81.1%) 0.335 − 35 10 (28.6%) 25 (71.4%)

2532 Clin Cancer Res; 16(9) May 1, 2010 Clinical Cancer Research

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2010 American Association for Cancer Research. Published OnlineFirst April 13, 2010; DOI: 10.1158/1078-0432.CCR-10-0018

GRB7 and GRB7v in Ovarian Cancer

the gaps was photographed at different time points. The Statistical analysis. Data were expressed as mean ± SD or width of wound was measured, and the relative velocity mean ± SEM. Student's t test was used to compare the va- of cell migration was calculated as width/time. All the lues between subgroups. Crosstabs and Pearson χ2 test experiments were done in triplicate and were repeated were used to test the correlation between the expression for at least thrice. levels of GRB7 and GRB7v and clinical parameters. The sta- Transwell cell migration and cell invasion analyses. tistical analyses were performed using the SPSS 13.0 soft- Quantification of cell migration and invasion was per- ware (SPSS). A P value of <0.05 was considered significant. formed by the Transwell cell migration and cell invasion assay kits (Chemicon International, Inc.) according to the Results manufacturer's instructions. In brief, 1.5 × 105 cells were resuspended in serum-free culture medium (for phosphor- Overexpression of GRB7 and GRB7v in ovarian cancer. ylated JNK inhibition test, 10 μmol/L SP600125 was Our preliminary study using cDNA microarray analysis re- included) and seeded on the upper chamber. The full me- vealed that the expression levels of GRB7 in the ovarian dium was placed in the lower chamber as chemoattractant. cancer cell lines were severalfolds higher than that in The cells were allowed to pass through the pores to the HOSE cells (data not shown). To confirm the upregulation lower surface of the membrane (12 hours for cell migra- of GRB7 in ovarian cancer cell lines, we evaluated the ex- tion and 48 hours for cell invasion assay) and were stained pression levels of GRB7 and GRB7v by Q-PCR analysis. In- with the staining buffer. Three different fields of the terestingly, we found that not only the expressions of stained cells were photographed and counted for each GRB7 were upregulated 9-fold to 40-fold in most of ovar- Transwell filter. The experiments were repeated thrice. ian cancer cell lines (such as SKOV3, OVCAR3, OVCA420,

Fig. 2. The effects of GRB7 and GRB7v on cell migration/invasion of ovarian cancer cells. A, Western blot analysis verified the expressions of GFP/GRB7 and GFP/GRB7v fusion protein in three OVCA433 clones of each gene. B, wound healing assay showed GRB7-enhanced, but not GRB7v-enhanced, cell migration. Complete wound closure was observed in GFP/GRB7 stable clone (C11) after 11 hours, but not in the vector control (V) and GRB7v stable clone (C5). The gap width was illustrated by arrows. The relative migration velocities of three independent assays were calculated and plotted (right). C, the Transwell cell migration assay showed that more cells of GFP/GRB7 stable clones (C11 and C19) transmigrated to the detection fields compared with the vector control (V) and GFP/GRB7v stable clones (C1 and C5); *, P < 0.05. The cell numbers in the random chosen fields were counted for three independent assays, and the results were plotted (right). D, the Transwell cell invasion assay showed that more cells of GFP/GRB7 stable clones (C11 and C19) invaded to the detection fields compared with the vector control (V) and GFP/GRB7v stable clones (C1 and C5); *, P < 0.05. The cell numbers in the random chosen fields were counted for three independent assays, and the results were plotted (right).

www.aacrjournals.org Clin Cancer Res; 16(9) May 1, 2010 2533

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2010 American Association for Cancer Research. Published OnlineFirst April 13, 2010; DOI: 10.1158/1078-0432.CCR-10-0018

Wang et al.

OVCA429, and OVCA433) but also the expressions of (P = 0.016; Fig. 1C). Clinicopathologic correlation showed GRB7v were upregulated 5-fold to 30-fold in the above cell that overexpressions of GRB7 and GRB7v were significantly lines compared with average expressions in HOSE cells associated with high-grade ovarian cancers (GRB7, P = (Fig. 1A). We further examined the protein levels of 0.009; GRB7v, P = 0.017; Tables 1 and 2). In addition, there GRB7 in the ovarian cancer cell lines by Western blotting was a trend showing that the overexpressed GRB7 and with specific anti-GRB7 antibody. Consistent with the re- GRB7v correlated with clear cell subtype of ovarian cancer sults of Q-PCR analysis, the expressions of GRB7 (and (Tables 1 and 2). These findings suggest that both GRB7 GRB7v) were remarkably higher in the above ovarian can- and GRB7v are overexpressed in ovarian cancer and signifi- cer cell lines (Fig. 1B). However, the positive band could cantly associated with high-grade ovarian cancers. not distinguish the expression patterns of GRB7 and Immunohistochemical analysis of GRB7 and GRB7v GRB7v because of their small size difference. expression in ovarian cancer tissue array. To further evaluate We further evaluated the expression status of GRB7 and the protein expression levels of GRB7 and GRB7v in the GRB7v in ovarian cancer tissues (n = 76) and normal ovarian clinical samples, immunohistochemical analysis was per- tissues (n = 53) by Q-PCR analysis. Based on the compara- formed on an ovarian cancer tissue array. The results tive computed tomography method using glyceraldehyde-3- showed that the cancer cells showed strong to marked phosphate dehydrogenase as the endogenous control, very staining of GRB7 and GRB7v (the antibody recognized low levels of GRB7 and GRB7v were detected in normal both isoforms simultaneously) in 65 of 97 tumor tissues ovaries but were significantly higher in ovarian cancers (Fig.1D).Butonlyweaktomoderatestainingwas

Fig. 3. The effects of GRB7/GRB7v on cell proliferation, anchorage-independent growth ability, and signaling activities of ovarian cancer cells. A, XTT assay showed enforced expression of GRB7 and GRB7v promoted cell proliferation rates. As compared with vector control (V), GFP/GRB7 stable clones showed a mild increase of cell proliferation (60% increase for C19; *, P = 0.004 and 20% increase for C7 and C11; **, P < 0.05), whereas GFP/GRB7v showed a higher increase of cell proliferation rate (∼110% increase for C1 and C5; *, P < 0.01 and 60% increase for C9; **, P = 0.013). B, the soft agar assay showed that GRB7v, but not GRB7, enhanced the cell anchorage-independent growth ability in OVCA433 cells. Bigger size of colonies (left) and more number of colonies (right) were observed in GFP/GRB7v stable clones (C1, C5, and C9) compared with vector control (V) and GFP/GRB7 stable clones (C11 and C19).

2534 Clin Cancer Res; 16(9) May 1, 2010 Clinical Cancer Research

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2010 American Association for Cancer Research. Published OnlineFirst April 13, 2010; DOI: 10.1158/1078-0432.CCR-10-0018

GRB7 and GRB7v in Ovarian Cancer

observed in the five cases of normal or benign tumor ovarian cancer cells, stable expression of GRB7 or GRB7v tissues (Fig. 1D). In addition, the staining signals were were established in OVCA433 cells (GFP/GRB7 expres- predominantly found in the cytoplasm of tumor cells sion, C7, C11, and C19; GFP/GRB7v expression, C1, C5 (Fig. 1D). The clinicopathologic analysis showed that and C9; Fig. 2A). Recent studies have shown that GRB7 high-score staining (ranking >75% percentile of all cases) is able to promote cell migration (18, 19). Thus, we firstly was significantly correlated with high-grade tumors (P = examined the effects of GRB7 and GRB7v on cell migra- 0.001; Fig. 1D; Supplementary Table S1). Nevertheless, tion by wound healing assay. Consistent to previous find- other clinical factors, such as age, histologic subtypes, ings, OVCA433 cells with overexpression of GRB7 and stages, showed no significant correlation with (C7, C11, and C19) showed significant faster wound GRB7/GRB7v expressions (Supplementary Table S1). closure compared with the vector control (Fig. 2B). These results support that the protein levels are consistent Surprisingly, enforced expression of GRB7v in OVCA433 with the mRNA levels of GRB7 and GRB7v in ovarian can- cells (C1, C5, and C9) could not show obvious difference cer and their upregulated expressions significantly associ- in wound closure compared with the vector control ated with high-grade ovarian cancers. (Fig. 2B). To better quantify the effect on cell migration, Both GRB7 and GRB7v localize in cytoplasm. Accumulat- the Transwell cell migration kit was used to examine ing evidence that GRB7 functions as a signaling adaptor GRB7 and GRB7v stably expressing clones. Consistent with has been reported (6–9). Thus, it is of interest to examine wound healing analysis, GRB7 stably expressing clones the subcellular localization of GRB7 and GRB7v to sup- (C11 and C19) showed 2-fold to 3-fold of more cells mi- port their roles as signaling adaptors. By confocal micros- grating through the membrane compared with vector con- copy analysis, we found that both GFP/GRB7 and GFP/ trol and GRB7v stably expressing clones (C1 and C5; GRB7v fusion proteins were predominately localized in Fig. 2C). Moreover, we also examined the cell invasion the cytoplasm of HEK293 cells and OVCA433 ovarian can- ability using Transwell cell invasion kit. We found that cer cells (Supplementary Fig. S1). This finding was consis- enforced expression of GRB7 in OVCA433 cells (C11 and tent with immunohistochemical data (Fig. 1D), indicating C19) showed ∼2-fold more invaded cells through the both GRB7 and GRB7v may function as signaling adaptors Matrigel compared with vector control and GRB7v stably in cytoplasm of ovarian cancer cells. expressing clones (C1 and C5; Fig. 2D). These results show GRB7, but not GRB7v, enhances cell migration and cell that GRB7, but not GRB7v, significantly enhances cell invasion. To study the function of GRB7 and GRB7v in migration and cell invasion in ovarian cancer cells.

Fig. 3. Continued. C, GRB7 and GRB7v regulate differential signaling targets. Western blot analysis revealed the upregulation of candidate signaling targets (phosphorylated ERK, phosphorylated JNK, and phosphorylated AKT) in the GFP/GRB7 and GFP/GRB7v OVCA433 stable clones (left) and transient transfection lysates in HEK293T cells (right).

www.aacrjournals.org Clin Cancer Res; 16(9) May 1, 2010 2535

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2010 American Association for Cancer Research. Published OnlineFirst April 13, 2010; DOI: 10.1158/1078-0432.CCR-10-0018

Wang et al.

Fig. 4. Inhibition of specific kinase activity abrogates GRB7 and GRB7v functions. A, the cells treated by the specific kinase inhibitors U0126, SP600125, and Wortmannin showed significant reduction of phosphorylated ERK, phosphorylated JNK, and phosphorylated AKT levels, respectively. B, the wound healing assay showed that SP600125, but not Wortmannin, could inhibit the cell migration rate of OVCA433 cells compared with the control using DMSO. The relative migration velocities of three independent assays were plotted. C, SP600125 neutralized the function of GRB7 on cell migration. Left, the wound healing analysis showed the vector control (V) and GFP/GRB7 stable clones (C11 and C19) closed the wounds at similar rates under treatment. The relative migration velocities were plotted. Right, Transwell cell migration analysis showed that the migrated cells of GFP/GRB7 stable clone (C11) significantly reduced compared with DMSO control (*, P = 0.026). D, XTT cell proliferation assay showed that, upon U0126 treatment, the increased cell proliferation rate found in GFP/GRB7 (C11 and C19) and GFP/GRB7v (C5 and C9) stable clones was totally abrogated.

Differential effects of GRB7 and GRB7v on cell prolifera- important tumorigenic behaviors found in high-grade tu- tion and anchorage-independent growth ability. Apart from mors (20). We firstly evaluated the effects of GRB7 and cell migration and invasion, higher cell growth rate and gain GRB7v on cell proliferation by XTT cell viability assay (21). of anchorage-independent growth ability are another two OVCA433 cells with enforced expression of GRB7v (C1, C5,

2536 Clin Cancer Res; 16(9) May 1, 2010 Clinical Cancer Research

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2010 American Association for Cancer Research. Published OnlineFirst April 13, 2010; DOI: 10.1158/1078-0432.CCR-10-0018

GRB7 and GRB7v in Ovarian Cancer

and C9) exhibited 60% to 110% higher cell proliferation rate ment of SP600125, we found that the increased cell migra- than the vector control (Fig. 3A). On the other hand, overex- tion ability from enforced expression of GRB7 was pression of GRB7 in OVCA433 (C7, C11, and C19) could significantly abrogated in GRB7 stably expressing clones only increase 20% to 60% cell proliferation rate compared (C11 and C19; Fig. 4C). Moreover, by Transwell cell mi- with vector control (Fig. 3A). We next used soft agar assay gration assay, we further confirmed that the enhanced cell to study the effects of enforced expression of GRB7 and migration ability in GRB7 stable clone (C11) was signifi- GRB7v in cell anchorage–independent growth ability. In- cantly reduced upon treatment of the JNK inhibitor terestingly, we found that only the enforced expression SP600125 (Fig. 4C). On the other hand, under continuous of GRB7v was able to increase number (2-fold to 3-fold; treatment of the ERK inhibitor U0126, we found that the P < 0.01) and size of colonies compared with the vector con- increased cell proliferation rates in GRB7 (C11 and C19) trol (Fig. 3B). These data indicate that GRB7v has higher on- and GRB7v (C5 and C9) stable clones were remarkably re- cogenic effect than GRB7 in promoting cell proliferation and duced (Fig. 4D). Taken together, these results suggest that cell anchorage–independent growth ability in ovarian the ERK signaling is important for both GRB7-mediated cancer cells. and GRB7v-mediated cell proliferation enhancement, GRB7 and GRB7v regulate differential signaling targets. whereas the JNK signaling is the crucial signaling pathway Recent report has documented that GRB7 is able to upregu- regulated by GRB7 in promoting cell migration. late the levels of phosphorylated ERK and phosphorylated AKT (19), suggesting that GRB7 is a crucial regulator in Discussion AKT/mitogen-activated protein kinase (MAPK) signaling axis. In this study, we aimed to investigate whether GRB7 has been known as an adaptor molecule mediat- the different functions of GRB7 and GRB7v in affecting ing signal transduction from cell surface receptors to vari- cell migration/invasion, cell proliferation, and anchor- ous downstream signaling pathways through interaction age-independent growth ability are due to their differen- with receptor tyrosine kinases, nonreceptor tyrosine tial effects on different signaling pathways in ovarian kinases, and many adaptor molecules (6–9). Emerging cancer cells. By Western blot analysis, the levels of phos- evidence has documented that GRB7 is frequently upregu- phorylated ERK, phosphorylated JNK, and phosphoryla- lated in a variety of human cancers and plays an essential ted AKT were found to be elevated in either GRB7 stably role in tumor progression (11–16). Consistent with these expressing clones of OVCA433 cells (C7, C11, and C19) studies, we have also found that GRB7 was overexpressed or HEK293T cells transient transfected with GRB7 in ovarian cancer cell lines and clinical samples. Intrigu- (Fig. 3C). In contrast, only the levels of phosphorylated ingly, its isoform GRB7v also showed an accordant expres- ERK were elevated in both transient transfection and stable sing pattern. More importantly, both overexpressed GRB7 expression of GRB7v (C1, C5, and C9; Fig. 3C). In addi- and GRB7v were significantly correlated with high-grade tion, there was no change of phosphorylated P38 expres- ovarian cancers. This indicates that GRB7 and GRB7v sion levels in OVCA433 cells with enforced expression of may have tumorigenic roles in ovarian cancer progression. either GRB7 or GRB7v (data not shown). These data confer The high-grade ovarian cancers are undifferentiated and that GRB7 and GRB7v regulate different downstream tar- poor in prognosis. These cancers have high-grade nuclei gets in AKT/MAPK signaling axis. and numerous mitotic figures (22, 23). The progression GRB7 and GRB7v exhibit different tumorigenic effect on to invasive carcinoma is a fast and stepwise process (22, ovarian cancer cells via AKT/MAPK signaling pathway. To 23). This indicates that these high-grade ovarian cancers further investigate the differential roles of GRB7 and tend to have higher cell proliferation and metastatic beha- GRB7v in regulating downstream targets of AKT/MAPK viors. The upregulation of GRB7 and GRB7v may thus in- signaling pathways related to the tumorigenic behaviors, volve these tumorigenic properties. Indeed, our functional we used specific kinase inhibitors targeting ERK studies using ovarian cancer cell models have shown that (U0126), JNK (SP600125), and AKT (Wortmannin). We the upregulation of GRB7 mainly promotes cell migration/ firstly evaluated the optimal dose of each protein kinase invasion and slightly in cell proliferation whereas GRB7v inhibitor in suppressing its target. We found that U0126 significantly increases cell proliferation and anchorage- (10 μmol/L for 4 hours) was able to reduce the level of independent growth ability. It is not surprising that GRB7 phosphorylated ERK at ∼50%, SP600125 (10 μmol/L for affects cell migration/invasion because GRB7 is frequently 4 hours) was able to reduce the level of phosphorylated overexpressed in invasive and metastatic cell lines and can- JNK at ∼60%, and Wortmannin (0.1 μmol/L for 4 hours) cer types (10, 14, 19, 24, 25). It has been shown that the was able to reduce the level of phosphorylated AKT at interaction of the SH2 domain of GRB7 to phosphotyrosine ∼80% in GRB7 stable clones (C11) of OVCA433 cells 14 (pY14) on caveolin-1 or EphB1 promotes cell migration (Fig. 4A). Moreover, inhibition of phosphorylated JNK (26, 27). Conversely, the expression of GRB7 with mutated and phosphorylated AKT by SP600125 and Wortmannin, SH2 domain or a cyclic peptide inhibitor to SH2 domain respectively, reduced the cell motility rates of OVCA433 remarkably reduces cell migration (18, 28), suggesting cells (Fig. 4B). Between these two protein kinase inhibi- that the presence of SH2 domain in GRB7 is crucial for tors, only SP600125 exhibited significant inhibitory effect its biological functions, e.g., cell migration regulation. on cell motility (Fig. 4B). Interestingly, under the treat- On the other hand, the absence of COOH terminal SH2

www.aacrjournals.org Clin Cancer Res; 16(9) May 1, 2010 2537

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2010 American Association for Cancer Research. Published OnlineFirst April 13, 2010; DOI: 10.1158/1078-0432.CCR-10-0018

Wang et al.

domain in GRB7v indicates it is incapable of binding growth factor-1, RasGTPase effectors, and specific phos- and phosphorylation by upstream kinase protein to ex- phoinositides (27, 34, 35). All of these possibilities may hibit any biological functions similar to GRB7. Indeed, explain that GRB7v, although in the absence of SH2 do- our results showed that GRB7v promotes cell prolifera- main, may still regulate downstream targets different from tion and anchorage-independent growth ability instead GRB7. However, further study for the identification of of cell migration/invasion, implying that GRB7v may GRB7v interaction partners and crucial phosphorylation be activated and regulated by distinct phosphorylation sites in each domain is needed. sites and subsequently regulating downstream targets In conclusion, our study implicates that the overex- different from GRB7. pressed GRB7 and GRB7v both associate with high-grade A recent study has proposed that ERK and AKT are the tumors but exert distinct tumorigenic functions through putative downstream signaling targets because their phos- regulating different signaling pathways in ovarian cancer phorylation levels are elevated, accompanied with GRB7 cells. Further understanding of their distinct functional overexpression (19). We thus examined the expressions roles and molecular mechanisms of GRB7 and GRB7v of the main signaling molecules in MAPK and AKT path- may facilitate the exploration of these proteins as novel ways. In this study, we have found that the levels of phos- therapeutic targets in ovarian cancers. phorylated JNK, phosphorylated ERK, and phosphorylated AKT were elevated by GRB7 enforced expression. On the Disclosure of Potential Conflicts of Interest other hand, overexpression of GRB7v just increased the level of phosphorylated ERK. To couple these increased No potential conflicts of interest were disclosed. signaling activities and biological functions, we applied specific kinase inhibitors to the GRB7-or GRB7v-overex- Acknowledgments pressing cell lines. As expected, the increased cell prolifer- ation due to the upregulation of either GRB7 or GRB7v We thank Prof. G.S.W. Tsao (Department of Anatomy, University of was totally inhibited by the ERK inhibitor U0126. In ad- Hong Kong) for providing immortalized HOSE cells, Prof. B.K. Tsang dition, the enhanced cell migration/invasion ability due to (Department of Obstetrics and Gynecology, University of Ottawa) for providing ovarian cancer cell lines, and Prof. J.L. Guan (Department of GRB7 expression was also abrogated by the JNK inhibitor Cell and Developmental Biology, University of Michigan) for providing SP600125. Our findings are consistent with previous stud- the GRB7 full-length cDNA. ies that the increased phosphorylated ERK promotes cell proliferation and the elevated phosphorylated JNK en- Grant Support hances cell migration (29–33). These results further confer the molecular basis of differential functions between HKU Seed Funding Programme for Basic Research grant 200811159046 GRB7 and GRB7v. and Wong Check She Charitable Foundation. The costs of publication of this article were defrayed in part by the Although there is a compelling evidence that the inter- payment of page charges. This article must therefore be hereby marked action with upstream tyrosine kinases is mediated by the advertisement in accordance with 18 U.S.C. Section 1734 solely to SH2 domain to relay the signal to downstream targets (18, indicate this fact. 28). The adjacent BPS, RA, and PH domains have also Received 01/06/2010; revised 02/25/2010; accepted 03/01/2010; been shown to interact with the insulin and insulin-like published OnlineFirst 04/13/2010.

References 1. Landen CN, Jr., Birrer MJ, Sood AK. Early events in the pathogenesis associated with invasive esophageal carcinoma. J Clin Invest 1998; of epithelial ovarian cancer. J Clin Oncol 2008;26:995–1005. 102:821–7. 2. Hoskins WJ. Prospective on ovarian cancer: why prevent? J Cell 11. Kao J, Pollack JR. RNA interference-based functional dissection of Biochem Suppl 1995;23:189–99. the 17q12 amplicon in breast cancer reveals contribution of 3. Salehi F, Dunfield L, Phillips KP, Krewski D, Vanderhyden BC. Risk coamplified genes. Genes Chromosomes Cancer 2006;45:761–9. factors for ovarian cancer: an overview with emphasis on hormonal 12. Stein D, Wu J, Fuqua SA, et al. The SH2 domain protein GRB-7 is co- factors. J Toxicol Environ Health B Crit Rev 2008;11:301–21. amplified, overexpressed and in a tight complex with HER2 in breast 4. Pawson T. Specificity in signal transduction: from phosphotyrosine- cancer. EMBO J 1994;13:1331–40. SH2 domain interactions to complex cellular systems. Cell 2004;116: 13. Tanaka S, Mori M, Akiyoshi T, et al. Coexpression of Grb7 with 191–203. epidermal growth factor receptor or Her2/erbB2 in human advanced 5. Bell HS, Ryan KM. Intracellular signalling and cancer: complex esophageal carcinoma. Cancer Res 1997;57:28–31. pathways lead to multiple targets. Eur J Cancer 2005;41:206–15. 14. Tanaka S, Sugimachi K, Kawaguchi H, Saeki H, Ohno S, Wands JR. 6. Han DC, Shen TL, Guan JL. The Grb7 family proteins: structure, Grb7 signal transduction protein mediates metastatic progression of interactions with other signaling molecules and potential cellular esophageal carcinoma. J Cell Physiol 2000;183:411–5. functions. Oncogene 2001;20:6315–21. 15. Kishi T, Sasaki H, Akiyama N, et al. Molecular cloning of human 7. Shen TL, Guan JL. Grb7 in intracellular signaling and its role in cell GRB-7 co-amplified with CAB1 and c-ERBB-2 in primary gastric regulation. Front Biosci 2004;9:192–200. cancer. Biochem Biophys Res Commun 1997;232:5–9. 8. Daly RJ. The Grb7 family of signalling proteins. Cell Signal 1998;10: 16. Varis A, Wolf M, Monni O, et al. Targets of gene amplification and over- 613–8. expression at 17q in gastric cancer. Cancer Res 2002;62:2625–9. 9. Holt LJ, Daly RJ. Adapter protein connections: the MRL and Grb7 17. Han DC, Guan JL. Association of focal adhesion kinase with Grb7 protein families. Growth Factors 2005;23:193–201. and its role in cell migration. J Biol Chem 1999;274:24425–30. 10. Tanaka S, Mori M, Akiyoshi T, et al. A novel variant of human Grb7 is 18. Han DC, Shen TL, Guan JL. Role of Grb7 targeting to focal contacts

2538 Clin Cancer Res; 16(9) May 1, 2010 Clinical Cancer Research

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2010 American Association for Cancer Research. Published OnlineFirst April 13, 2010; DOI: 10.1158/1078-0432.CCR-10-0018

GRB7 and GRB7v in Ovarian Cancer

and its phosphorylation by focal adhesion kinase in regulation of cell 28. Porter CJ, Matthews JM, Mackay JP, et al. Grb7 SH2 domain migration. J Biol Chem 2000;275:28911–7. structure and interactions with a cyclic peptide inhibitor of cancer 19. Chu PY, Huang LY, Hsu CH, et al. Tyrosine phosphorylation of cell migration and proliferation. BMC Struct Biol 2007;7:58. growth factor receptor-bound protein-7 by focal adhesion kinase in 29. Bessard A, Fremin C, Ezan F, Fautrel A, Gailhouste L, Baffet G. the regulation of cell migration, proliferation, and tumorigenesis. RNAi-mediated ERK2 knockdown inhibits growth of tumor cells J Biol Chem 2009;284:20215–26. in vitro and in vivo. Oncogene 2008;27:5315–25. 20. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100: 30. Brand S, Olszak T, Beigel F, et al. Cell differentiation dependent 57–70. expressed CCR6 mediates ERK-1/2, SAPK/JNK, Akt signaling 21. Chan DW, Liu VW, Tsao GS, et al. Loss of MKP3 mediated by resulting in proliferation and migration of colorectal cancer cells. oxidative stress enhances tumorigenicity and chemoresistance of J Cell Biochem 2006;97:709–23. ovarian cancer cells. Carcinogenesis 2008;29:1742–50. 31. Byun HJ, Hong IK, Kim E, et al. A splice variant of CD99 increases 22. Vang R, Shih Ie M, Kurman RJ. Ovarian low-grade and high-grade motility and MMP-9 expression of human breast cancer cells through serous carcinoma: pathogenesis, clinicopathologic and molecular the AKT-, ERK-, and JNK-dependent AP-1 activation signaling biologic features, and diagnostic problems. Adv Anat Pathol 2009; pathways. J Biol Chem 2006;281:34833–47. 16:267–82. 32. Caceres M, Tobar N, Guerrero J, Smith PC, Martinez J. c-jun- 23. Cho KR, Shih Ie M. Ovarian cancer. Annu Rev Pathol 2009;4:287–313. NH2JNK mediates invasive potential and EGFR activation by 24. Haran M, Chebatco S, Flaishon L, et al. Grb7 expression and cellular regulating the expression of HB-EGF in a urokinase-stimulated migration in chronic lymphocytic leukemia: a comparative study of pathway. J Cell Biochem 2008;103:986–93. early and advanced stage disease. Leukemia 2004;18:1948–50. 33. Kwon GT, Cho HJ, Chung WY, Park KK, Moon A, Park JH. Isoliquir- 25. Itoh S, Taketomi A, Tanaka S, et al. Role of growth factor receptor itigenin inhibits migration and invasion of prostate cancer cells: bound protein 7 in hepatocellular carcinoma. Mol Cancer Res 2007; possible mediation by decreased JNK/AP-1 signaling. J Nutr 5:667–73. Biochem 2009;20:663–76. 26. Han DC, Shen TL, Miao H, Wang B, Guan JL. EphB1 associates with 34. Stein EG, Gustafson TA, Hubbard SR. The BPS domain of Grb10 Grb7 and regulates cell migration. J Biol Chem 2002;277:45655–61. inhibits the catalytic activity of the insulin and IGF1 receptors. FEBS 27. Shen TL, Han DC, Guan JL. Association of Grb7 with phosphoinosi- Lett 2001;493:106–11. tides and its role in the regulation of cell migration. J Biol Chem 2002; 35. Ponting CP, Benjamin DR. A novel family of Ras-binding domains. 277:29069–77. Trends Biochem Sci 1996;21:422–5.

www.aacrjournals.org Clin Cancer Res; 16(9) May 1, 2010 2539

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2010 American Association for Cancer Research. Correction Clinical Cancer Correction: Differential Functions of Growth Research Factor Receptor–Bound Protein 7 (GRB7) and Its Variant GRB7v in Ovarian Carcinogenesis

In this article (Clin Cancer Res 2010;16:2529–39), which was published in the May 1, 2010 issue of Clinical Cancer Research (1), a reader pointed out potential image reuse. The authors admitted that due to an oversight that occurred during figure preparation, identical b-actin blots were erroneously used in Fig. 2A (GRB7v; OVCA433) and Fig. 3C (transient transfection GFP/GRB7). The authors provided the original blots, which showed the correct b-actin blots for each figure. The corrected Fig. 3C appears below. The results and conclusions put forth in this article remain unchanged. The authors regret these errors.

Stable expression Transient transfection OVCA433 HEK293T

GRB7 GRB7v GFP/GRB7 (mg) GFP/GRB7v (mg)

V C7 C11 C19 C1 C5 C9 V 0.5 1.0 3.0 V 0.5 1.0 3.0

GFP

Phospho-ERK

ERK

Phospho-JNK

JNK

Phospho-AKT

AKT

b-Actin

Figure 3C.

Reference 1. Wang Y, Chan DW, Liu VWS, Chiu PM, Ngan HYS. Differential functions of growth factor receptor- bound protein 7 (GRB7) and its variant GRB7v in ovarian carcinogenesis. Clin Cancer Res 2010;16: 2529–39.

Published online March 1, 2016. doi: 10.1158/1078-0432.CCR-15-2863 2016 American Association for Cancer Research.

1296 Clin Cancer Res; 22(5) March 1, 2016 Published OnlineFirst April 13, 2010; DOI: 10.1158/1078-0432.CCR-10-0018

Differential Functions of Growth Factor Receptor−Bound Protein 7 (GRB7) and Its Variant GRB7v in Ovarian Carcinogenesis

Yajun Wang, David W. Chan, Vincent W.S. Liu, et al.

Clin Cancer Res 2010;16:2529-2539. Published OnlineFirst April 13, 2010.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-10-0018

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2010/04/12/1078-0432.CCR-10-0018.DC1

Cited articles This article cites 35 articles, 10 of which you can access for free at: http://clincancerres.aacrjournals.org/content/16/9/2529.full#ref-list-1

Citing articles This article has been cited by 1 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/16/9/2529.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/16/9/2529. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 2010 American Association for Cancer Research.