SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Eph receptor 2 (EphA2) overexpression is frequently accompanied by the loss of its cognate ligand during tumor progression. However, the molecular mechanism of this ligand-independent promotion of tumor by EphA2 remains unclear in highly malignant and fatal cholangiocarcinoma (CC). We examined the biological role of EphA2 in tumor growth and metastasis in CC tissues and cells according to the degree of differentiation and we explored the downstream signaling pathways of EphA2. Growth factor-mediated EphA2 overexpression itself leads to the activation of the mammalian target of rapamycin complex 1 (mTORC1) and extracellular signal-regulated kinase (ERK) pathways through ligand-independent activation of EphA2 (phosphorylation of S897). An in vitro soft agar assay and in vivo orthotopic or subcutaneous tumor model showed that EphA2 enhanced colony formation and accelerated tumor growth, and which seemed to be mainly associated with Akt (T308)/mTORC1 activation. Aberrant expression and activation of EphA2 was also associated with poorer differentiation and higher metastatic ability. Enhanced metastatic ability was also observed in an orthotopic tumor model or lung metastasis model, correlating with Pyk2(Y402)/c-Src/ERK activation in addition to activation of the canonical Raf/MEK/ERK pathway. The mTORC1 and Raf/Pyk2 pathways also appeared to affect each other. These results suggest that growth factor-mediated EphA2 might be involved in tumor growth and metastasis through activation of the mTORC1 and Raf/Pyk2 pathways. Therapeutic strategies that target EphA2 and its downstream effectors may be useful to control CC. (HEPATOLOGY 2013;57:2248–2260)

Recently, members of the Eph receptor tyrosine kinase (RTK) family, including Eph receptor 2 (EphA2), have been linked to tumor progression and neovascularization.1 EphA2 is a transmembrane RTK that is found at low levels in adult epithelial cells2; however it is frequently overexpressed and functionally altered in many types of carcinomas.3-6 EphA-Fc receptor proteins that disrupt endogenous receptor activation significantly inhibit growth and neovascularization of tumors in vivo.7-9 These data suggest that EphA2 functions as a pro-oncogene. In contrast, EphA2 has also been reported to have a tumor suppressor role. EphA2−/− gene-trap mice display increased susceptibility to chemical carcinogen-induced skin cancer compared to control littermates, along with increased tumor cell proliferation and phosphorylation of extracellular signal-regulated kinase (ERK).9 Overexpression of EphA2 in lung and breast cancer cell lines has been found to negatively regulate proliferation and induce apoptosis.10, 11 Furthermore, ligand stimulation of EphA2 inhibits integrin signaling and the Ras/ERK pathway, which is correlated with inhibition of cell proliferation and migration.12 Treatment of human breast cancer cell lines with soluble ephrin-A1-Fc ligand attenuates epidermal growth factor (EGF)-mediated phosphorylation of ERK and inhibits transformation of NIH3T3 cells expressing v-ErbB2.13 In addition, EphA2 is a gene targeted by the p53 family and causes apoptosis when overexpressed.11, 14, 15 The exact role of EphA2 in tumor development and progression has remained controversial. EphA2 overexpression is frequently accompanied by the loss of its cognate ligands,16, 17 but it can promote tumor progression in a ligand-independent manner.18

Human cholangiocarcinoma (CC) is a highly malignant, generally fatal neoplasm originating in the bile duct epithelial cells or cholangiocytes of the intra- and extrahepatic biliary system. Mortality rates have risen sharply in recent years19 and because CC has no proven adjuvant therapy, only surgical resection with tumor-free margins is associated with improvement in survival, with 5-year survival rates in the range of 20%-40%.20 However, resection is precluded by distant metastases, extensive regional lymph node metastasis, and vascular encasement or invasion. Abnormal EGF and EGF receptor (EGFR) signaling can also participate in the genesis and progression of CC.21 There is accumulating evidence implicating EGF and EGFR signaling pathways that regulate or mediate the epithelial-to-mesenchymal transition (EMT) in the development of CC.22, 23 However, reduced EGFR expression and activation was shown to correlate with a lack of differentiation in CC, even in CC with an altered EMT.24 The molecular mechanism underlying tumor progression, by which tumor cells acquire the phenotypes of invasiveness and metastasis, is therefore still unclear in CC.

We therefore investigated whether this ligand-independent, growth factor-mediated overexpression of EphA2 contributes to tumor progression and metastasis and explored the pro-oncogenic downstream pathways.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Cell Lines and Tissues.

CC cell lines, including Choi-CK, Cho-CK, JCK, and SCK cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS; Invitrogen, Carlsbad, CA), 1% penicillin and streptomycin (Sigma Chemical, St. Louis, MO), 3 mmol/L taurine, and 25 mmol/L HEPES (Invitrogen, Carlsbad, CA). All cell lines were grown at 37°C in the presence of 5% CO2, as described.25 The human cholangiocyte cell line (H69) immortalized with simian virus 40 (SV40) large T antigen was cultured in hormonally supplemented medium containing 1.64 ± 10−6 mol/L EGF (R&D Systems, Minneapolis, MN), as described.26 CC tissues were obtained from the archives of the Department of Pathology, Chonbuk National University Hospital (Jeonju, South Korea). To quantify the levels of phospho-EphA2 or EphA2 in the immunostained tissue samples, we used an immunohistochemistry (IHC) scoring system. Briefly, the mean percentage of phospho-EphA2 or EphA2-positive tumor cells in each tissue section was determined by analysis of at least five random fields at ×400 magnification. The intensity of the phospho-EphA2 or EphA2 immunoreactive signals was scored as: 1+, weak; 2+, moderate; 3+, intense. The percentage of positive tumor cells and the staining intensity of each tissue section were then multiplied to produce the phospho-EphA2 or EphA2 IHC score. These judgments were made by two independent pathologists, neither of whom had any knowledge or information pertaining to the patients' clinical status. This study was approved by the Ethics Committee of Chonbuk National University Hospital.

Luciferase Reporter Gene Assay.

Transcriptional activity assays were performed using the Luciferase Assay System (Promega) according to the manufacturer's instructions. Cho-CK cells were simultaneously transfected with EphA2-Luc reporter plasmids10 and p53mt-choi, p53mt-cho, p53mt-jck, or p53mt-sck expression plasmids.25 Luciferase activity was measured using a Dual-Luciferase Reporter (DLR) Assay Kit (Promega) as instructed by the supplier. Firefly and Renilla luciferase activities were measured for normalization using a luminometer (Lumat LB9507, Berthold, Bad Wildbad, Germany).

Phospho-Proteome Profiling.

Cells were rinsed with cold phosphate-buffered saline (PBS) and immediately solubilized in NP-40 lysis buffer (1% NP-40, 20 mM Tris-HCl [pH 8.0], 137 mM NaCl, 10% glycerol, 2 mM EDTA, 1 mM sodium orthovanadate, 10 μg/mL aprotinin, 10 μg/mL leupeptin) by rocking the lysates gently at 4°C for 30 minutes. Following microcentrifugation at 14,000g for 5 minutes, supernatants were transferred into a clean test tube and sample protein concentrations were determined using the Pierce Protein Assay Kit (Rockford, IL). Lysates (500 μg) were diluted and incubated with Human Phospho Kinase Arrays (Proteome Profiler; R&D Systems) per the manufacturer's instructions. Array data were developed on x-ray film following exposure to chemiluminescent reagents.

Soft Agar Colony Formation Assay.

Cells were plated at a density of 5 × 104 cells per well in 60 mm plates in growth medium containing 0.7% agar (5 mL per well) on top of a layer of growth medium containing 1.4% agar (3 mL per well). Growth medium (500 μL) with 10% FBS was added on top of the agar. The cell suspension was plated and cultured in a 37°C incubator for 2 weeks. After 2 weeks, viable colony formation was observed using an optical microscope.

Mouse Tumorigenicity Assay.

Four-week-old female athymic nude mice (BALB/cAnN/CrljOri-nu, Orient, Seongnam, South Korea) were used in all experiments. The animals were maintained in a specific pathogen-free environment. The animal room was kept at 20-22°C under a 12-hour light/dark cycle. Cho-CK cells were stably transfected with the EphA2 expression plasmid (EphA2-20 and EphA2-22), and 5 × 106 cells in 60 μL DMEM/20 μL Matrigel were then injected into the left lobe of the mouse liver, subcutaneously 5 × 106 cells in 200 μL PBS into both flanks of the nude mice. Growth curves were plotted based on mean tumor volume within each experimental group at the indicated timepoints. Tumors were monitored for changes in length and width. Tumor volume was calculated according to the following equation: V (mm3) = width2 (mm2) × length (mm)/2. Tumor growth was observed for 5 or 8 weeks. In vivo tumorigenic experiments were performed with seven mice in each treatment group. This animal study was conducted according to a protocol approved by the institutional Animal and Care Committee at Chonbuk National University Hospital.

In Vivo Metastasis Assay.

A tail vein injection assay was used to assess the effect of EphA2 on tumor metastasis. Cho-CK cells (1 × 106 cells in 0.01 mL PBS per mouse) previously transfected with either recombinant vector containing full-length EphA2 or empty vector were injected into the tail veins of athymic nude mice. JCK-luc cells with knockdown of EphA2 by lentiviral delivery of short hairpin RNA (shRNA) or control nontarget shRNA were injected into the tail veins of athymic nude mice. Mice were assessed for long-distance lung metastasis at 12 weeks (all, six mice per group). The numbers of lung metastasis nodules were counted to analyze the effects of EphA2 on spontaneous tumor metastasis.

Bioluminescence Imaging and Analysis.

An orthotopic nude mouse lung metastasis tumor model with JCK-luc cells with knockdown of EphA2 by lentiviral delivery of shRNA or control nontarget shRNA was established. Briefly, 5 × 105 cells in 0.08 mL of culture medium were injected into the tail veins of Balb/c nude mice (4 weeks old, n = 6). Tumor growth was monitored once a week for 4 weeks using a Xenogen IVIS 100 imaging system (Alameda, CA; 1 minute, Level B/FOV15). Mice were anesthetized with 3% isoflurane after administration of 150 mg/kg body weight firefly D-luciferin (Xenogen) by intraperitoneal injection for imaging.

Statistical Analysis.

Statistical analyses were carried out using SPSS 16.0. Correlations were calculated using Spearman rank-order coefficients. The data are expressed as averages ± standard error of the mean (SEM). The differences were analyzed by dependent or independent t tests, and P values ≤ 0.05 were considered significant.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

EphA2 Expression According to the Level of Differentiation in CC.

First, we tested the specificity of the rabbit polyclonal EphA2 antibody used in our study. The EphA2 antibody specifically detected myc- or GFP-tagged EphA2 fusion protein (Supporting Fig. S1) in transient transfection experiments. We previously described four CC cells with varying levels of differentiation, including well-differentiated Choi-CK, moderately differentiated Cho-CK, poorly differentiated JCK, and sarcomatoid SCK cells.25 EphA2 expression was more prominent in less-differentiated CC cells, such as JCK and SCK cells (Fig. 1A). However, both of these cell types showed lower expression levels of the known EphA2 ligand, EFNA1, as well as EGFR, compared to the better-differentiated Choi-CK and Cho-CK cells. Accordingly, EphA2 transcript levels were elevated in poorly differentiated JCK and SCK cells. IHC examination of EphA2 expression in CC tissues revealed that normal bile duct epithelial cells were negative or barely positive for EphA2 immunoreactivity. As expected, the poorly differentiated CC tissues were more strongly positive for EphA2 than the better-differentiated CC tissues (Fig. 1B). These results suggest that EphA2 expression correlates with poorer differentiation and more progression in CCs. Among the various growth factors tested, EGF induced EphA2 expression and reciprocal inhibition of EFNA1 in CC cells (Fig. 1C, upper panels). Concomitantly, EGF down-regulated EFNA1 promoter activity (Supporting Fig. S2). Therefore, EGF seems to transcriptionally regulate EphA2 and EFNA1. Inhibition assays showed that RTK inhibitors such as AG1478 and gefitinib efficiently inhibited EGF-mediated EphA2 expression and concurrently enhanced EFNA1 expression in the differentiated CC cells (Fig. 1C, lower panels). MEK and ERK inhibitors (U0126 and PD98059) also inhibited EphA2 expression, but less efficiently. Inhibitors of phosphatidylinositol 3-kinase (wortmannin and LY294002) had no effect on the EGF-induced expression of EphA2, but inhibitors of c-Src (PP2 and lavendustin) inhibited the EGF-induced expression of EphA2 (Supporting Fig. S3). The EGF-mediated morphological changes were consistent with EMT changes (Fig. 1D), as we previously described. Among the cells treated with various growth factors, the promoter activity of EphA2 was highest in the EGF-treated cells (Fig. 1E). EGF dose-dependently induced EphA2 overexpression and phosphorylation on S897, but not on Y594 (Fig. 1F). These results suggest that aberrant expression of EphA2 resulting from abnormal growth factor signaling is crucial for tumor promotion and metastasis and this abnormal growth factor signaling contributes to the down-regulation of EFNA1 and EGFR.

thumbnail image

Figure 1. EphA2 expression in cholangiocarcinoma cells. (A) EphA2, Ephrin A1, and EGFR expression according to the level of differentiation in CC cells. A semiquantitative RT polymerase chain reaction assay was performed to determine the EphA2 mRNA level in the CC cells. (B) Immunohistochemical detection of EphA2 in CCs according to differentiation (Mod, moderate; Sar, sarcomatoid). The results shown are representative of three independent experiments (n = 3). (C) Regulation of EFNA1 and EphA2 by growth factors such as EGF (10 or 40 ng/mL), TGF-α (10 ng/mL), TGF-β (10 ng/mL), and FGF (20 ng/mL) (upper panels). Blockage of EGF-mediated EphA2 induction by receptor tyrosine kinase inhibitors (5 μM AG1478 and 2 μM gefitinib) and MEK/ERK inhibitors (10 μM U01226 and 50 μM PD98059) in Cho-CK cells treated with the drugs for 48 hours (lower panels). The results shown are representative of three independent experiments. (D) EMT-like morphological changes of CC cells treated with various growth factors. (E) Increased promoter activity of EphA2 in Cho-CK cells induced by various growth factors. Each bar represents the mean ± SEM. *P < 0.05; **P < 0.01. (F) EGF-mediated EphA2 expression and phosphorylation at S897 in Cho-CK and JCK cells. The results shown are representative of three independent experiments.

Download figure to PowerPoint

Morphological Changes Caused by EphA2.

Overexpression of EphA2 is correlated with disease progression and highly malignant cellular phenotypes.7, 8 We therefore investigated the morphological changes induced by EphA2 overexpression in Cho-CK cells. EphA2 expression was examined in stably transfected Cho-CK cells by immunoblot analysis (Fig. 2A). The stable transfectants expressing EphA2 (A2-20 and A2-22) showed a more migratory and less adhesive phenotype compared to the vector control cells (VC4 and VC10) (Fig. 2B). We carried out an immunofluorescence assay in order to define the subcellular localization of EphA2 after transient transfection with the GFP-tagged EphA2 expression plasmid in H69 cells (Fig. 2C) and Cho-CK cells (Supporting Fig. S4A). Endogenous EphA2 was mainly expressed along the cytoplasmic membrane and less in the cytoplasm (white arrows). Ectopically overexpressed EphA2 was similarly localized in the cytoplasmic membrane and cytoplasm. In contrast to secretory EFNA1, EphA2 was not a secretory protein in the culture supernatants of CC cells (Supporting Fig. S4B). To assess the ligand-mediated EphA2 expression, we treated the SCK cells with recombinant mouse ephrin-A1 Fc ligand and found that it efficiently down-regulated EphA2 expression in a dose-dependent manner compared to treatment with recombinant human IgG1-Fc (Supporting Fig. S5). EGF treatment resulted in an increase in EphA2 expression and phosphorylation at S897, but not Y594. In contrast, the ephrin-A1 Fc ligand decreased EphA2 expression, but increased phosphorylation of EphA2 at Y594 (Fig. 2D). The cell morphology changed from fibroblastoid type to epitheloid type, i.e., reverse EMT (MET) (Fig. 2E). Alternatively, we examined the F-actin staining of stress fibers with fluorescent phalloidin in the EphA2-expressing cells treated with either recombinant IgG1-Fc or recombinant ephrin-A1 Fc ligand. Treatment with recombinant ephrin-A1 Fc ligand resulted in repressed expression of EphA2 and the formation of F-actin stress fibers (Fig. 2F).

thumbnail image

Figure 2. Morphological changes according to EphA2 expression and subcellular localization. (A) Western blot analysis of EphA2 in Cho-CK cells stably expressing the EphA2 protein. (B) Morphological changes in stable transfectants (A2-20 and A2-22) were compared with those of the vector control cells (VC-4 and VC-10). (C) Anti-EphA2 Ab immunoreactivity (red) was assayed by immunofluorescence; Cho-CK cells were transfected with the GFP-tagged EphA2 expression vector (GFP-EphA2) or an empty vector control (GFP). The cells were processed by staining with 1 μg/mL Hoechst 33258 to visualize the nuclei (blue) and for indirect immunofluorescence by staining for EphA2 (red). Trans, transmission. Scale bar = 20 μm. (D) Regulation and phosphorylation of EphA2 in Cho-CK cells treated with EGF (40 ng/mL) or EFNA1-Fc (1 μg/mL) for three days compared with the Fc-control cells (1 μg/mL). The results shown are representative of three independent experiments. VC, vehicle control. (E) Morphological reversion of fibroblastoid SCK cells to epitheloid cells induced by treatment with EFNA1-Fc (μg/mL). (F) SCK cells stably expressing EphA2 were processed for immunofluorescence staining for EphA2 (green) and F-actin (red). EphA2 was suppressed in association with F-actin by EFNA1-Fc in SCK cells. Scale bar = 20 μm.

Download figure to PowerPoint

Enhanced Tumorigenicity of EphA2.

In order to determine if human EphA2 has pro-oncogenic properties, the capacity of EphA2 to anchor independent cell growth was compared to vector control cells by determining the level of colony growth on soft agar. The stable EphA2 transfectants formed a greater number of larger colonies on soft agar (Fig. 3A). Furthermore, tumor growth was orthotopically or subcutaneously faster in the mouse tumor model with the stable transfectants than with vector control cells (Fig. 3B; Supporting Fig. S6). These results suggest that EphA2 plays a crucial role in tumor growth in vitro and in vivo. In addition, intrahepatic tumor metastases were more enhanced in orthotopic xenoplants from the stable EphA2 transfectants than from the vector control cells. We next determined whether ERK1/2 activation is also involved in EphA2-mediated tumor growth. EphA2 expression was linked to weak ERK1/2 activation in EphA2 overexpressed cells, but its activation was strongly correlated with ERK1/2 activation in invasive CC cells (Fig. 3C). Previously, EphA2 was reported to be downstream of tumor suppressor p53, which is related to cell growth inhibition and apoptosis.10 However, Cho-CK and other CC cells contain mutant p53.25 Those p53 mutant CC cells showed decreased expression of EphA2 promoter activity rather than up-regulation (Fig. 3D), implying that EphA2 expression is independent of p53 function in CC cells. In addition, the transformation ability of EphA2 was further examined by determining the level of cell growth as colonies in soft agar (Supporting Fig. S7). Intriguingly, it was found that EphA2 overexpression in H69 cells dramatically enhanced the anchorage-independent cell growth in soft agar, implying that EphA2 seems to be potentially oncogenic in these cells.

thumbnail image

Figure 3. Enhanced tumorigenicity of Cho-CK cells induced by EphA2. (A) Cho-CK cells stably expressing EphA2 generated more and much larger colonies on soft agar than the vector control cells. The graph shows quantification of colony numbers (lower panel). Each bar represents the mean ± SEM. (B) Growth of the tumor masses and their intrahepatic metastases from EphA2-expressing cells (A2-20 and A2-22) and vector control cells (VC-4 and VC-10) that were injected into the liver of nude mice (upper). The main tumor volume and the number of tumor nodules were quantified 5 weeks after orthotopic transplantation. Each value represents the mean ± SEM (n = 7). (C) Aberrant expression of EphA2 is associated with phosphorylation of ERK in Cho-CK cells transiently transfected with EphA2 expression plasmids and CC cells according to differentiation. The results shown are representative of three independent experiments. (D) EphA2 up-regulation is independent of p53 in CC cells. The transcriptional regulation of EphA2 by wildtype p53 or mutant p53 was assessed in Cho-CK cells using a luciferase reporter with the EphA2 promoter. Each bar represents the mean ± SEM. The results shown are representative of three independent experiments.

Download figure to PowerPoint

Acquisition of EMT Characteristics by EphA2.

The EMT of epithelial cancers is correlated with aggressive tumors that are believed to play a critical role in metastatic cancer.27 We therefore examined whether EphA2 overexpression contributed to tumor invasiveness and metastasis by affecting the EMT phenomena in CC cells. We found that adhesion molecules such as E-cadherin and desmoplakin I/II were down-regulated, and locomotion molecules, including fibronectin (FN), vimentin (VIM), N-cadherin, and alpha smooth muscle actin (α-SMA), were up-regulated in the cells stably expressing EphA2 (Fig. 4A). Wound repair assays were performed on Cho-CK cells stably expressing EphA2, and the stable transfectants demonstrated a visibly enhanced capacity to migrate into the wounded area compared to the vector control cells (Fig. 4B). In a modified Boyden chamber assay, the stable transfectants readily penetrated the matrix and colonized the bottom surface of the Matrigel-coated membrane, while the vector control cells did not (Fig. 4C). Therefore, EphA2 expression appeared to enhance the invasiveness of these cells. The role of EphA2 expression in the metastatic phenotype of the Cho-CK cells was examined by first injecting stable transfectants (100 μL of 5 × 105 cells/mL) into the tail vein of nude mice and subsequently monitoring multiple metastatic nodules in the lungs. The stable EphA2 transfectants first colonized and then continued growing into the lungs with many more metastatic nodules (black arrowhead) than were achieved with the vector control cells (Fig. 4D). A histocytological examination confirmed these results (Fig. 4E). Many metastatic nodules strongly expressed EphA2 in the lungs.

thumbnail image

Figure 4. Migratory and metastatic abilities induced by EphA2. (A) Western blot analyses showed acquisition of mesenchymal and loss of epithelial marker proteins in the stable EphA2 transfectants, compared with vector control cells. FN, fibronectin; VIM, vimentin; N-cad, N-cadherin; αSMA, alpha smooth muscle actin; E-cad, E-cadherin; CK, cytokeratin; Des, desmoplakin. The results shown are representative of three independent experiments. (B) Migratory ability was determined in the Cho-CK cells stably expressing EphA2 by way of the wound healing assay. Phase contrast images were taken 36 hours later to assess cell migration. The results shown are representative of three independent experiments. (C) Photomicrographs of a modified Boyden chamber assay show numerous stable EphA2 transfectants while only a few vector control cells had traversed the Matrigel-coated membrane after 48 hours. The cells that invaded were normalized to the viable cell mass on both sides of the membrane (right panel). Each bar represents the mean ± SEM. **P < 0.01. (D) Mice were inoculated with 5 × 105 Cho-CK cells, transfected as indicated, by way of the tail vein. Forty-five days after tumor implantation, animals were euthanized and the lungs were evaluated for tumor nodules. The graph shows results from two independent experiments conducted in four groups (n = 5). Black arrowheads indicate metastatic nodules. Each bar represents the mean ± SEM. ***P < 0.001. (E) Immunohistochemistry for EphA2 in the numerous metastatic nodules observed in the lungs of the mice. Scale bars = 200 μm.

Download figure to PowerPoint

Akt and Pyk2 Activation by EphA2.

Next, we sought to identify the phosphorylation of downstream molecules using Human Phospho-Kinase Arrays (Fig. 5A). EphA2 overexpression led to phosphorylation of Akt and Pyk2, as confirmed by immunoblot and immunoprecipitation, respectively (Fig. 5B). Akt phosphorylation occurred at T308 and Pyk2 at Y402 in both JCK and SCK cells. However, other phosphorylation sites, S473 of Akt and Y881 of Pyk2, were not affected by EphA2. Mammalian target of rapamycin (mTOR) is known to function downstream of Akt as part of the mTOR complex 1 (mTORC1) protein complex.28, 29 We therefore investigated the mTORC1 pathway in stable EphA2 transfectants of the four CC cell lines. EphA2 overexpression resulted in phosphorylation of mTORC1, 4EBP1, and p70S6K in JCK and SCK cells (Fig. 5C). EphA2 overexpression also led to c-Src phosphorylation (Fig. 5D). Next, we knocked down EphA2 expression using lentiviral delivery of shRNA to determine whether downstream Akt/mTOR and Pyk2/c-Src activation had changed. EphA2 suppression substantially decreased mTOR, 4EBP1, and p70S6K phosphorylation in JCK cells (Supporting Fig. S8A). As expected, the Akt inhibitor similarly inhibited the mTOR pathway (Supporting Fig. S8B). Although EphA2 suppression resulted in the inhibition of Pyk2 and c-Src activation (Supporting Fig. S9A), Akt inhibition had almost no effect on the activation of Pyk2 and c-Src (Supporting Fig. S9B). These results suggest that the Akt/mTORC1 pathway is independent of the Pyk2/c-Src pathway. The knockdown experiments revealed that EphA2 knockdown itself resulted in EFNA1 induction and was linked to the inhibition of canonical upstream activators of ERK signaling, including b/c-Raf and MEK (Fig. 5E), but not Ras activation (Supporting Fig. S10). EphA2 itself did not transcriptionally down-regulate EFNA1 (Supporting Fig. S11). The inhibition assay using c-Src inhibitor PP2 demonstrated that c-Src activation was associated with neither the Pyk2 nor the Raf/MEK1/2 activation. Rapamycin was used as a negative control (Fig. 5F). Similar results were obtained by a transient transfection assay with active c-Src or a dominant negative c-Src (DN-c-Src) (Supporting Fig. S12). These findings suggest that the canonical MEK or c-Src pathway independently regulates ERK1/2 activation.

thumbnail image

Figure 5. Downstream signaling of EphA2. (A) Dot blot screens using the Phospho-Kinase Array Kit based on spotted pairs (vertical doublets) of antibodies for the signaling proteins indicated in the upper and lower boxes (Akt and Pyk2). The phosphorylation state was determined using a mixture of phosphopeptide-specific antibodies. The two doublets in the corners are internal positive controls. The results shown are representative of three independent experiments. (B) Phosphorylation of Akt and Pyk2 by immunoblot and immunoprecipitation assays in the stable EphA2 transfectants compared to the vector control cells (left); phosphoactivation of Akt and Pyk2 in the four CC cell lines (right). The results shown are representative of three independent experiments. (C) Activation of Akt and mTORC1 in the stable EphA2 transfectants compared to the vector control cells (left); activation of Akt and mTORC1 in the four CC cell lines (right). The results shown are representative of three independent experiments. (D) Activation of Pyk2 and c-Src in the stable EphA2 transfectants compared to the vector control cells (left); activation of Pyk2 and c-Src in the four CC cell lines (right). The results shown are representative of three independent experiments. (E) Inhibition of the c-Raf/MEK/ERK pathway by EphA2 knockdown with lentiviral delivery of shRNA. The results shown are representative of three independent experiments. (F) Regulation of Pyk2/c-Src and c-Raf/MEK/ERK pathways by c-Src inhibitors (10 μM PP2 and 400 nM Laven). VC, vehicle control; Laven, lavendustin; Rapa, rapamycin. The results shown are representative of three independent experiments.

Download figure to PowerPoint

Proliferation and Metastatic Ability of Akt/mTORC1 and c-Src Activation.

To determine the major pathway linked to tumor cell proliferation and metastatic ability, we measured cell proliferation in EphA2 transfectants treated with Akt inhibitor (5 μM), mTORC1 inhibitor (rapamycin 20 ng/mL), or c-Src inhibitor (PP2 20 μM). Both the Akt inhibitor and rapamycin effectively inhibited EphA2-mediated tumor cell proliferation. PP2 also substantially inhibited EphA2-mediated tumor cell proliferation, but not completely (Fig. 6A). In the wound healing assay, PP2 completely inhibited EphA2-mediated wound closure. Rapamycin, however, caused only a partial inhibition of wound repair (Fig. 6B). The Matrigel-coated transwell invasion assay similarly showed that PP2 completely inhibited invasion of Matrigel pores, while rapamycin provided only partial inhibition (Fig. 6C). These results suggest that the Akt/mTORC1 pathway plays a major role in regulating cell proliferation while the Pyk2/c-Src pathway generally acts to modulate metastatic ability. However, the Akt/mTOR and Pyk2/c-Src pathways also appeared to substantially influence each other. We further examined the morphological changes occurring in the stable transfectants expressing EphA2 after rapamycin or PP2 treatment. Rapamycin partially inhibited the migratory characteristics and stress fiber formation, while PP2 completely suppressed these phenomena (Fig. 6D).

thumbnail image

Figure 6. Inhibition of tumor cell proliferation and metastatic ability. (A) Proliferation inhibition of EphA2 transfectants treated with Akt inhibitor (10 μM), rapamycin (Rapa, 20 ng/mL), or PP2 (10 μM). Each value represents the mean ± SEM. (B) The wound healing assay demonstrated complete and partial inhibition of EphA2-mediated wound closure by PP2 and rapamycin, respectively. Quantitative measurements are shown. Each bar represents the mean ± SEM (n = 3). *P < 0.05 and ***P < 0.001 (n = 3). (C) The Matrigel assay revealed a significant number of EphA2 transfectants invading through pores, which was significantly inhibited by PP2, but less inhibited by rapamycin. Quantitative measurements are shown. Each bar represents the mean ± SEM (n = 3). **P < 0.01 and ***P < 0.001. (D) Morphological changes in stable EphA2 transfectants after treatment with rapamycin (Rapa, 20 ng/mL) or PP2 (10 μM). Immunofluorescent staining for EphA2 (FITC, green) and F-actin (phalloidin-TRITC, red). Scale bar = 20 μm.

Download figure to PowerPoint

Inhibition of Metastasis by EphA2 Knockdown.

To exclude the cell-specific function of EphA2 we used metastasis-prone CC cells, JCK, to determine whether EphA2 expression is critical for metastasis of CC cells because these cells express high levels of EphA2. EphA2 knockdown cells were established by lentiviral delivery of EphA2 shRNA with stable expression of luciferase (JCK-luc) (Fig. 7A). The metastatic phenotype of the JCK-luc cells was examined by injection into the tail veins (100 μL of 5 × 105 cells/mL) of nude mice. Sufficient bioluminescence data were collected at 4 weeks postinjection (Fig. 7B). The parent and non-target cells first colonized and then continued growing into the lungs with many metastatic nodules (Fig. 7C). A histological examination demonstrated that injection of the parent and nontarget cells resulted in numerous EphA2-positive metastatic lesions, and the target cell injection resulted in EphA2-negative and fewer metastatic nodules. These results suggest that EphA2 expression plays a major role in tumor metastasis and invasion in CC cells.

thumbnail image

Figure 7. Inhibition of the metastatic ability of JCK cells by EphA2 knockdown. (A) Knockdown of EphA2 expression in JCK cells by the lentiviral delivery of nontarget (shNT) or target shRNA (sh413s and sh3352s). (B) Bioluminescent images demonstrate that parent and nontarget tumor cells spread to the lungs (n = 6 mice/group, representative anterior-posterior images, 1 minute exposure time). Quantitative measurement of photon flux (right). Each bar represents the mean ± SEM. ***P < 0.001. (C) Immunohistochemical staining of EphA2 in metastatic nodules. Scale bars = 500 μm. (D) Schematic model of EGF/EphA2-mediated tumor growth and metastasis through the Akt/mTORC1 and Pyk2/c-Src pathways. p, phosphorylation.

Download figure to PowerPoint

Metastatic CCs showed stronger immunoreactivity for EphA2 and phosphorylated EphA2 (S897) than CCs at the original site (Supporting Fig. S13A). Furthermore, the EphA2 immunoreactivity of CCs with N1 (positive adjacent lymph node) was stronger than the CC tissues with N0 (negative lymph node), according to tumor differentiation (Supporting Fig. S13B). A schematic model of the signaling pathways downstream of EphA2 activation (phosphorylation of S897) is depicted in Fig. 7D. In addition to the canonical Raf/MEK activation, overexpression of EphA leads to activation of Akt and Pyk2 by phosphorylation. In turn, activated Akt and Pyk2 phosphorylate mTORC1 and c-Src, respectively. The activation of molecules downstream of mTORC1 and ERK contribute to tumor growth and/or metastasis in CCs.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Ligand-dependent inhibition of cell proliferation and migration by EphA2 may help maintain homeostasis in normal tissues and suppress tumor development at early stages of tumorigenesis.30 In contrast, ligand-independent stimulation of cell migration by EphA2 promotes tumor progression and provides an oncogenic function.31 Our data revealed that EphA2 overexpression resulted in an increase in tumor growth and metastasis, and was even associated with oncogenesis. EphA2 has previously been described as a substrate for Akt serine/threonine kinase in vitro and in vivo. The phosphorylation of EphA2 by Akt occurs predominantly at a single residue, S897. It is rapidly induced upon stimulation with serum and all growth factors. Therefore, EphA2 serves as a common downstream effector molecule for growth factor signaling. Previously, adhesion-induced EphA2 expression was shown to be dependent on the activation of MEK and Src.32 In contrast, EGF-induced activation of the MAPK pathway was shown to control the expression of the EphA2 receptor and its ligand.13 Accordingly, our experiments showed that EGF-mediated EphA2 induction was partially blocked only by MEK/ERK inhibitors. Inhibitors of phosphatidylinositol 3-kinase (wortmannin and LY294002), p38 (SB203580), JNK, and c-Src (PP2 and lavendustine) had no effect on EGF-induced EphA2 expression.33 In contrast, the ligand-mediated stimulation of EphA2 causes rapid dephosphorylation of phospho-S897 concomitant with dephosphorylation and inactivation of Akt kinase.18 Our studies revealed that ligand-dependent activation of EphA2 resulted in phosphorylation of EphA2 at Y594, which did not seem to be relevant to growth factor signaling. The ligand-independent activation and phosphorylation of EphA2 at S897 has been suggested to be associated with aberrant stimulation of the growth factor(s), which enhanced tumorigenicity and metastasis in vitro and in vivo. This autocrine or paracrine signaling of growth factor(s) seems to be responsible for the reciprocal down-regulation of EFNA1 and EGFR.

The Akt-mTORC1 pathway is often activated in cancer cells due to loss of the tumor suppressor, PTEN, a lipid phosphatase that dephosphorylates PI(3,4,5)P3 to PI(4,5)P2.34 Alternatively, activating mutations in PI3 kinase or Akt and deregulation of growth factor receptors can also result in activation of the Akt-mTORC1 pathway in cancer cells.35 This pathway can promote cancer cell growth as well as migration and invasiveness.36 In this study, we first proposed that EphA2 expression and activation itself can activate the Akt-mTORC1 pathway, which primarily contributes to tumor cell proliferation and additionally to tumor metastasis. EphA2 expression resulted in phosphorylation of Akt at T308, but not at S473, an arrangement that seemed to be critical for ligand-independent activation of EphA2 linked to the activation of the mTORC1 pathway. In contrast, ephrine-A1 ligand-dependent Akt activation is related to the dephosphorylation of Akt at both T308 and S473.37

There is increasing evidence of an interaction between Pyk2 and c-Src in CC cell proliferation and invasiveness. In addition, the role of the ERK/MAPK-signaling pathway in the progression of cancer is related to Pyk2 activation.38 Pyk2 is phosphorylated at the major autophosphorylation site (Tyr-402) and the potential Grb2-binding site (Tyr-881) during EMT. In contrast, phosphorylation of Fak at the corresponding autophosphorylation site (Tyr-397) occurs even in sedentary epithelial cells, whereas phosphorylation at Tyr-407 and Tyr-861 is induced during EMT.39 In our study, EphA2 expression induced a change in EMT that was correlated with phosphorylation of Pyk2 at T402, but not at Tyr-881, which is linked to c-Src activation. In the inhibition assay, tumor cell proliferation appeared to depend primarily on activation of the Akt/mTORC1 pathway, whereas tumor invasion and metastasis seemed to result from Pyk2/c-Src activation. Furthermore, both knockdown of EphA2 and Akt inhibition suppressed the activation of the mTORC1 pathway. In contrast, knockdown of EphA2 led to suppression of the Pyk2/c-Src pathway, while Akt inhibition did not. Thus, EphA2 appears to be a frequent upstream participant in both pathways while Akt is only an upstream target of mTORC1.

Next, we determined whether Pyk2/c-Src directly activates ERK or if it activates ERK through the Raf/MEK pathway. Inhibition assays using c-Src inhibitor PP2 or dn-c-Src revealed that PykA2/c-Src activates ERK in a way that is independent of Raf/MEK activation. In addition, activation of the EphA2 receptor has been reported to stimulate the MAPK pathway.40 Our data also showed that activation of the EphA2 receptor is linked to the activation of the Raf/MEK/ERK signaling cascade. Therefore, the separate pathways of Pyk2/c-Src and Raf/MEK converge with ERK activation.

In conclusion, our studies clarified the molecular pathway of EphA2 signaling and provide insight into the mechanisms that control cancer progression and metastatic spread inCC.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The authors thank Dr. Chen Jin (Vanderbilt University School of Medicine, Nashville, TN) for the generous gift of the EphA2 plasmid, Dr. S. Lemay (McGill University, Quebec, Canada) for kindly providing the EphA2-luc (-1881∼+137) reporter plasmid.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
HEP_26253_sm_SuppFig1.eps363KSupporting Information Figure 1.
HEP_26253_sm_SuppFig2.eps965KSupporting Information Figure 2.
HEP_26253_sm_SuppFig3.eps588KSupporting Information Figure 3.
HEP_26253_sm_SuppFig4.eps679KSupporting Information Figure 4.
HEP_26253_sm_SuppFig5.eps1074KSupporting Information Figure 5.
HEP_26253_sm_SuppFig6.eps557KSupporting Information Figure 6.
HEP_26253_sm_SuppFig7.eps719KSupporting Information Figure 7.
HEP_26253_sm_SuppFig8.eps1252KSupporting Information Figure 8.
HEP_26253_sm_SuppFig9.eps963KSupporting Information Figure 9.
HEP_26253_sm_SuppFig10.eps805KSupporting Information Figure 10.
HEP_26253_sm_SuppFig11.eps663KSupporting Information Figure 11.
HEP_26253_sm_SuppFig12.eps577KSupporting Information Figure 12.
HEP_26253_sm_SuppFig13.eps830KSupporting Information Figure 13.
HEP_26253_sm_SuppFig14.eps2644KSupporting Information Figure 14.
HEP_26253_sm_SuppInfo.doc36KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.