- Email: [email protected]
Akt pathway in colorectal carcinoma
FEBS Letters 588 (2014) 1921–1929
journal homepage: www.FEBSLetters.org
Phosphorylation and changes in the distribution of nucleolin promote tumor metastasis via the PI3K/Akt pathway in colorectal carcinoma Dong-ming Wu a,1, Peng Zhang b,1, Ru-yan Liu c, Ya-xiong Sang a, Cong Zhou a, Guang-chao Xu a, Jin-liang Yang a, Ai-ping Tong a,⇑, Chun-ting Wang a,⇑ a b c
State Key Laboratory of Biotherapy and Cancer Center, West China Hospital, West China Medical School, Sichuan University, Chengdu, PR China Department of Radiation Oncology, Sichuan Cancer Hospital, Chengdu, Sichuan, PR China Graduate School, Guangxi Medical University, Nanning, Guangxi, PR China
a r t i c l e
i n f o
Article history: Received 12 November 2013 Revised 22 March 2014 Accepted 24 March 2014 Available online 5 April 2014 Edited by Zhijie Chang Keywords: Nucleolin Metastatic VEGF Distribution PI3K/Akt Colorectal carcinoma
a b s t r a c t Here, we investigated the molecular mechanism underlying the changes in the distribution of nucleolin. Our study identified PI3K/Akt signaling as an essential pathway regulating the distribution of nucleolin. Furthermore, nucleolin can interact with phospho-PI3K-p55, and changes in the distribution of nucleolin were related to its phosphorylation. Subsequently, we analyzed the correlation of VEGF and nucleolin, and found that distribution of nucleolin related to metastatic potential. Finally, blocking cell surface nucleolin influences the process of epithelial–mesenchymal transitions. This indicates that nucleolin may be a novel cancer therapy target and a predictive marker for tumor migration in colorectal carcinoma. Ó 2014 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
1. Introduction Worldwide, colorectal cancer is the fourth most common tumor in men and the third most common in women [11, 23]. Tumor metastasis remains the principal cause of high mortality rate and poor prognosis in patients with colorectal cancer; colorectal carcinoma that is localized within the colon or has only spread to the lymph nodes is curable by surgery with or without chemotherapy, and has a 5-year survival rate of 70% and 90%, respectively. However, cancer that has metastasized to distant sites is generally incurable and has a 5-year survival rate of 12% [30]. Hence, many studies aimed at identifying a novel predictor or therapy target for colorectal carcinoma have been reported [17,31]. Nucleolin is a multifunctional protein with well-characterized roles in the organization of nucleolar chromatin, packaging of pre-rRNA, transcription of rDNA, and ribosome assembly [1]. The importance of cell surface nucleolin has been suggested by studies ⇑ Corresponding authors. Address: State Key Laboratory of Biotherapy, West China Hospital, West China Medical School, Sichuan University, Gaopeng Street, Keyuan Road 4, Chengdu 610041, PR China. Fax: +86 028 85164060. E-mail addresses: [email protected] (A.-p. Tong), [email protected]. cn (C.-t. Wang). 1 These authors contributed equally to this work.
showing that functional blockade or down-regulation of cell surface nucleolin in endothelial cells inhibits migration and capillary-tubule formation and causes endothelial cell apoptosis [37].Targeting nucleolin on the plasma membrane of cancer cells seems to be an effective approach to inhibit cancer cell growth and angiogenesis in various in vitro and in vivo experimental models [26]. In recent years, studies have suggested that high levels of nucleolar expression of nucleolin are associated with a better prognosis in pancreatic ductal adenocarcinoma patients [25], and that nucleolin on the cell surface may be a novel cancer therapy target (Abdelmohsen and Gorospe, 2012), [38,39]. Previous studies have indicated that VEGF can increase the cell surface nucleolin expression in human microvascular endothelial cells [16]. VEGF plays a major role in tumor angiogenesis and migration; its expression is inversely correlated with survival of patients in many human cancers, including colorectal carcinomas. Both cell surface nucleolin and VEGF are upregulated in different types of cancer and have been implicated in migration progression [3,4]. Here, we showed that VEGF, through the VEGFR, leads to phosphorylation of nucleolin and that this then results in nucleolin localization to the membrane. We also found that distribution of nucleolin is related to the phosphorylation states of nucleolin.
http://dx.doi.org/10.1016/j.febslet.2014.03.047 0014-5793/Ó 2014 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
1922
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
Our findings suggested that increased expression of cell-surface and cytosolic nucleolin were related to tumor metastasis. Thus, nucleolin may be a novel theraputic target for colorectal cancer, while the distribution of nucleolin may be a predictive indicator in clinical diagnosis. 2. Materials and methods
ylmethanesulfonyl fluoride) and incubated for 30 min on ice; lysates were then centrifuged at 18 000g for 30 min at 4 °C and the supernatants (500 mg) incubated for 2 h at 4 °C with 2 mg of anti-nucleolin antibody. Concurrently, 2 mg of rabbit IgG was used as the control. Immunoprecipitated samples were washed three times with RIPA buffer; these samples as well as pre-immunoprecipitated samples (50 mg) were subjected to Western blotting analysis using an anti-nucleolin and an anti-PI3K antibody.
2.1. Materials Recombinant human VEGF165 was purchased from PeproTech (Suzhou, China). The anti-nucleolin, anti-PI3K, and anti-Akt polyclonal antibodies were obtained from ProteinTech Group, Inc. (Wuhan, China). The phospho-Akt (Thr308) and phospho-PI3 kinase P85 (Tyr458)/P55 (Tyr199) antibodies were purchased from Cell Signaling Technology, Inc. (Danvers, MA). The phosphoserine/ threonine/tyrosine antibody was purchased from GeneTex International Corporation (CA, USA). Anti-Thr76/Thr84-phosphorylated nucleolin antibody was purchased from BioLegend (San Diego, USA). The secondary antibody (goat anti-rabbit IgG) was purchased from Santa Cruz Biotechnology (Heidelberg, Germany). Tumor microarrays were purchased from Biomax US (MD, USA). HB19 (Cell surface nucleolin inhibitor) and 9 Args were synthetized by Shanghai Science Peptide (Shanghai, China). All other kits or reagents were purchased from the Beyotime Institute of Biotechnology (Shanghai, China). 2.2. Immunofluorescence Cells were fixed with 4% paraformaldehyde, and then blocked with PBS containing 10% normal goat serum. Cells were then stained with anti-nucleolin polyclonal antibody for 30 min at 37 °C, and then stained with Cy3-conjugated secondary antibody for 30 min at 37 °C. Then those samples were re-stained by Hochest 33258 for 15 min at 37 °C. All immunofluorescence images were obtained using an Olympus BX51 microscope equipped with either a 20 or a 40 objective lens (Olympus) and a DP 50 camera (Olympus). Images were processed using DPC controller software (Olympus). 2.3. Fluorescence-activated cell-sorting (FACS) analysis After treatment, cells were trypsinized with 0.02% EDTA, and resuspended in PBS supplemented with 0.1% bovine serum albumin. Cells were incubated with anti-nucleolin antibody for 30 min on ice, and then incubated on ice with anti-rabbit secondary antibody conjugated to Cy3. Then, cells were directly analyzed using a FACScan (Becton Dickinson, San Jose, CA).
2.6. Immunohistochemical staining and assessment of VEGF and nucleolin expression Tissues were dewaxed and antigens retrieved using high pressure for 3 min. Then, the activity of endogenous peroxidases was blocked using 3% hydrogen peroxide for 10 min at room temperature. After immersion in normal goat serum for 30 min at 37 °C, tissues were subjected to primary antibody incubation at 4 °C overnight. Subsequently, sections were washed with PBS for 15 min at room temperature and incubated with secondary antibody that had been labeled with biotin for 30 min at 37 °C. After washing with PBS for 15 min at room temperature, the sections were incubated with horseradish peroxidase (HRP) complex for 30 min at 37 °C and visualized using diaminobenzidine (DAB). All immunohistochemical images were obtained using an Olympus BX51 microscope equipped with either a 20, a 40, or 100 objective lens (Olympus) and a DP 50 camera (Olympus). Images were processed using DPC controller software (Olympus). The immunohistochemical staining was evaluated using a semiquantitative scoring method. The scores were decided by two pathologists independently. The total nucleolin and VEGF staining was scored as per staining intensity: no staining (0), light positive staining (1), medium positive staining (2), and strong positive staining (3). The positively stained area was scored as: <5% (0), 5–25% (1), 26–50% (2), 51–75% (3), and >75% (4). A final score was calculated by multiplying the intensity score by the expression score and the rating (from 0 to 12) was determined for each case. 2.7. Statistical analysis Each experiment was performed at least three times independently. The associations of VEGF and nucleolin with clinicopathological parameters were generated using a paired t-test or one-way ANOVA analyses implemented in GraphPad Prism 5 (GraphPad, San Diego, CA). Statistical significance was defined as P < 0.05. 3. Results 3.1. VEGF changes the distribution of nucleolin in colon cancer cells
2.4. Subcellular fractionation Subcellular fractions of cells comprising cytosolic, nuclear, and cell membrane extracts were prepared as follows. Cell monolayers in 100-mm plates were washed extensively with PBS before being scraped and pelleted. Washed cells (2 106) were then used for extraction of cytosolic, nuclear, and cell membrane proteins using a Membrane and Cytosol Protein Extraction Kit and a Nuclear and Cytoplasmic Protein Extraction Kit according to the manufacturer’s recommendations. These samples were then analyzed by Western blot or co-immunoprecipitation analysis. 2.5. Co-immunoprecipitation assay Cells were resuspended in ice-cold RIPA (Radio-Immunoprecipitation Assay) cell lysis buffer containing 1 mM PMSF (Phen-
In order to investigate the function of VEGF in the distribution of nucleolin expression in colon cancer, HCT116 cells and DLD-1 cells treated with 0, 5, 10, or 20 ng/ml VEGF for 24 h were harvested for FACS analysis. We found that VEGF increased nucleolin expression on the cell surface of HCT116 and DLD-1 cells, in a concentration-dependent manner (Fig. 1A). To further assess the regulation of nucleolin distribution in colon cells by VEGF, HCT116 and DLD-1 cells were stained with an anti-nucleolin antibody for immunofluorescence analysis. As shown in Fig. 1B, we found that cell surface nucleolin was increased after VEGF treatment, in a dose-dependent manner, and there was no change in the mRNA level of nucelolin (Supplement Fig. 2). To further confirm that VEGF can change the distribution of nucleolin in colon cancer cells, we treated HCT116 and DLD-1 cells with different doses of VEGF; after 24 h, we separated the
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
1923
Fig. 1. VEGF changes the distribution of nucleolin in colon cancer cells. (A) Flow cytometry showed that VEGF treatment of cells increased cell surface nucleolin expression in HCT116 cells (Left) and DLD-1 cells (Right). (B) Immunofluorescence analysis cell surface nucleolin after cells were treated with VEGF in HCT116 cells (Up) and DLD-1 cells (Down). (C) The differences in the distribution of nucleolin expression were analyzed after treatment of cells with the indicated dose of VEGF in HCT116 cells (Up) and DLD-1 cells (Down).
1924
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
subcellular fractions of these cells, and analyzed the expression of nucleolin in the cytosolic, nuclear, and cell membrane extracts by Western blotting. As shown in Supplement Fig. 1, the separation efficiency of those kits was detected. We found that VEGF clearly increased nucleolin levels in the cell membrane and cytoplasmic fractions, while nucleolin levels in the nucleus was decreased, and the total nucleolin remained unchanged in both HCT116 and DLD-1 cells (Fig. 1C). 3.2. VEGF changes the distribution of nucleolin via the PI3K/Akt pathway It has been suggested that the migration of nucleolin in the cytosol and nucleus to the cell membrane occurs via the PI3K/ Akt pathway [20]. We tested this suggestion: HCT116 cells were
cultured with 50 lM LY294002 (PI3K inhibitor) for 5 h, and then treated with VEGF; 24 h later, cells were harvested for FACS analysis. As shown in Fig. 2A, the cell surface nucleolin levels were decreased when cells were pretreated with LY294002, as compared to cells not treated with LY294002, and there was no change in the mRNA level of nucelolin (Supplement Fig. 3). This indicated that blocking the PI3K/Akt pathway can reduce the cell surface nucleolin level effectively in VEGF-treated cells. Anti-nucleolin staining by immunofluorescence analysis in HCT116 cells further supported this finding: we found that LY294002 clearly blocked the VEGF-mediated distribution of nucleolin at the cell surface (Fig. 2B). We treated HCT116 cells with VEGF and/or LY294002; after 24 h, we separated the subcellular fraction of HCT116 cells, and analyzed the expression of nucleolin in the cytosolic, nuclear,
Fig. 2. VEGF changes the distribution of nucleolin via the PI3K/Akt pathway. (A) Treatment of cells with LY294002 reduced the VEGF-mediated increase in cell surface nucleolin. (B) Expression of cell surface nucleolin was determined by immunofluorescence analysis after cells were treated with VEGF, following pre-treatment with LY294002. (C) The distribution of nucleolin and cell membrane extracts as well as total nucleolin levels were analyzed using Western blotting. (D) VEGF treatment caused an increase in the phosphorylation of Akt. (E) LY294002 treatment reduced the VEGF-mediated phosphorylation of Akt.
Fig. 3. Distribution of nucleolin is related to the phosphorylation states of nucleolin. (A) VEGF increased PI3K-p55 phosphorylation. (B) Nucleolin interacts with phosphoPI3K-p55. Normal IgG was as the control. (C) VEGF increase nucleolin phosphorylation. Normal IgG was as the control. (D) VEGF increase cell surface nucleolin phosphorylation. Normal IgG was as the control. (E) Nucleolin phosphorylation is related to tumor stage. Lymph node metastases have higher phosphorylation level than colorectal carcinoma and normal lymph node. Normal IgG was as the control.
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
1925
Fig. 4. Distribution of nucleolin is associated with tumor metastasis. (A) VEGF and nucleolin expression in normal colon, an adenocarcinoma group, and a metastases group. (B) Immunohistochemical staining of VEGF (top) and nucleolin (bottom) in normal colon (left), adenocarcinoma (middle), and metastases (right). (C) VEGF (left) and nucleolin (right) expression was investigated in clinical colorectal cancer samples. Immunohistochemical staining of VEGF and nucleolin in normal colon (top), adenocarcinoma (middle), and lymph node metastases and liver metastases (bottom) indicated that distribution of nucleolin correlates with tumor metastasis.
and cell membrane extracts, as well as the total nucleolin in cells, by Western blotting. We found that the nucleolin levels in the cytosolic and cell membrane fractions decreased, while nuclear nucleolin levels increased, without changes in total nucleolin levels after LY294002 treatment. This indicated that LY294002 blocked the distribution of nucleolin from the nucleus (Fig. 2C).
We next evaluated the phosphorylation states of Akt in HCT116 cells. As shown in Fig. 2D, the phosphorylation levels of Akt increased with VEGF treatment in a concentration-dependent manner, while total Akt levels remained unchanged. After treatment with LY294002, the phosphorylation of Akt was markedly decreased (Fig. 2E).
1926
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
Table 1 Clinicopathological association of nucelolin with colorectal carcinoma. Factors
Number of patients
Average score
P value
Group Adenocarcinoma Metastatic adenocarcinoma
30 30
3.43 ± 0.55 4.03 ± 0.65
0.49
Gender Male Female
22 8
3.68 ± 0.69 2.75 ± 0.86
0.47
Age <60 years P60 years
21 9
2.91 ± 0.63 4.67 ± 1.05
0.15
Differentiation Well Moderate Poor
12 27 15
3.83 ± 0.90 4.74 ± 0.70 3.00 ± 0.64
0.25
Clinical stage I II III IV
5 16 6 3
2.60 ± 1.03 4.17 ± 0.79 4.67 ± 1.50 2.25 ± 1.32
0.52
Clinical and TNM stagings were according to American Joint Committee on Cancer: AJCC Cancer Staging Manual, 6th edition.
Table 2 Clinicopathological association of VEGF with colorectal carcinoma. Factors
Number of patients
Average score
P value
Group Adenocarcinoma Metastatic adenocarcinoma
30 30
4.93 ± 0.40 6.80 ± 0.70
0.02
Gender Male Female
22 8
5.18 ± 0.48 4.75 ± 0.94
0.66
Age <60 years P60 years
21 9
4.52 ± 0.49 5.78 ± 0.68
0.16
Differentiation Well Moderate Poor
12 27 15
4.58 ± 0.73 5.29 ± 0.53 6.00 ± 0.84
0.46
Clinical stage I II III IV
5 16 6 3
3.40 ± 0.87 5.44 ± 0.60 5.00 ± 1.03 5.00 ± 2.00
0.47
Clinical and TNM stagings were according to American Joint Committee on Cancer: AJCC Cancer Staging Manual, 6th edition.
3.3. Distribution of nucleolin is related to the phosphorylation states of nucleolin It has been demonstrated that nucleolin physically interacts with PI3K [18]; therefore, we tested whether VEGF increases PI3K phosphorylation levels in HCT116 cells. As shown in Fig. 3A, VEGF promoted PI3K-P55 phosphorylation, while total PI3K and nucleolin levels remained unchanged in HCT116 cells. Given that nucleolin interacts with PI3K, we tested its phosphorylation subunits, and found that nucleolin interacts with the phospho-PI3K-p55 subunit (Fig. 3B). This indicated that nucleolin interacts with phosphorylated p55 in the cell nucleus and then translocates from the nucleus to the cytosol and cell membrane. We next tested the phosphorylation state of nucleolin in HCT116 cells, by immunoprecipitating total nucleolin, and then analyzing the phosphorylation state of nucleolin using a phosphoserine/threonine/tyrosine antibody. After treatment of cells with VEGF, we found that the phosphorylation of nucleolin increased.
We next tested whether LY294002 could block the nucleolin phosphorylation; we found that LY294002 could reduce the VEGF-mediated increase in nucleolin phosphorylation (Fig. 3C). Subsequently, we analyzed the correlation between different phosphorylation states of nucleolin and differences in the distribution of nucleolin were analyzed. We found that distribution of nucleolin was related to the phosphorylation state of nucleolin in HCT116 cells; phosphorylation of nucleolin was only observed in the cytosolic and cell membrane extracts, but not in nuclear nucleolin. Furthermore, cytosolic and cell membrane extracts demonstrated greater levels of nucleolin phosphorylation after VEGF treatment, but this phosphorylation was blocked when LY294002 was added (Fig. 3D). In order to test whether different stages of colorectal carcinoma show different phosphorylation levels of nucleolin, we analyzed the nucleolin phosphorylation level in colorectal carcinoma samples and compared these levels with those in lymph node metastases. We found that lymph node metastases had higher nucleolin phosphorylation levels compared to colorectal carcinoma. Then, the phosphorylation state of nucleolin in different distributions was determined by separating colorectal carcinoma and lymph node metastases tissue. We found that the cytosolic and cell membrane extracts demonstrated higher phosphorylation levels of nucleolin in the lymph node metastases than those in colorectal carcinoma (Fig. 3F). It has been proved that Thr76/Thr84-phosphorylated nucleolin play a role in process of tumors, so we tested whether the expression of phospho-Thr76/Thr84 nucleolin changed with VEGF stimulation. As shown in Supplement Fig. 4, VEGF increased phospho-Thr76/Thr84 of nucleolin, while treatment with LY294002 inhibited the effects of VEGF. 3.4. Distribution of nucleolin is association with tumor metastasis To study the role of VEGF and nucleolin in colorectal carcinoma and metastases, we assessed the levels of VEGF and nucleolin expression in colorectal carcinoma using a tissue microarray. As shown in Fig. 4A and B, there was no difference between normal colon tissue, colon adenocarcinoma tissue, and colon metastases in terms of immunohistochemical staining of total nucleolin. No significant difference was observed between well, moderately, and poorly differentiated carcinoma tissues; different clinical stages showed no significant differences in total nucleolin expression (Table 1). Statistical analysis indicated that VEGF expression was significantly lower in colon adenocarcinoma than in metastases (P = 0.0241; Table 2). We demonstrated that total nucleolin expression levels were not statistically significantly different between the adenocarcinoma group and the metastases group, which suggested that the expression of total nucleolin does not change with tumor metastasis. To further explore the roles of nucleolin distribution in colorectal carcinoma metastasis, clinicopathological association analyses were specifically performed for colorectal carcinoma and metastases. Along with the increase in metastatic capacity from colorectal carcinoma, through lymph node metastases to liver metastases, the metastatic colorectal carcinoma showed more cytosolic staining and less nuclear staining for nucleolin (Fig. 4C). 3.5. Inhibition of cell surface nucleolin by HB19 inhibits EMT in colon cancer cells Thus far, we have shown that inhibition of the migration of nucleolin to the cell surface is associated with a greater risk of tumor metastasis. Then, we investigated whether inhibition of cell surface nucleolin would affect the process of EMT in colon cancer cells. As shown in Fig. 5A and Supplement Fig. 5, we confirmed that
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
1927
Fig. 5. Inhibition of cell surface nucleolin by HB19 inhibits the process of epithelial–mesenchymal transitions (EMT) in colon cancer cells. (A) Inhibition of cell surface nucleolin by HB19. 9R was as the control. (B) Morphological changes after HCT116 cells were treated with HB19 and/or VEGF. (C) Representative immunofluorescence images of staining for E-cadherin (above) and FSP1 (below) in serum-starved HCT116 cells treated with HB19 and/or VEGF. (D) Biomarkers for EMT: vimentin, FSP1, and E-cadherin were analyzed by Western blotting. (E) Quantification of cell migration by the transwell assay; data from three independent experiments are shown (P < 0.05).
the synthesis of peptide HB19 [8] can block cell surface nucleolin expression in HCT116 cells. 9R (9 Args) was as the control of HB19. Treatment with HB19 promoted morphological changes in these cells that were characterized by loss of the ‘‘fibroblastoidlike’’ phenotype and acquisition of an epithelial-like compact morphology (Fig. 5B). As shown in Fig. 5C, immunofluorescence staining revealed that HB19 upregulated expression of the epithelial
maker E-cadherin and decreased expression of the mesenchymal maker fibroblast-specific protein 1 (FSP1) in HCT116 cells. We next analyzed cells treated with VEGF and/or HB19 with biomarkers for EMT: vimentin, FSP1and E-cadherin (Fig. 5D). HB19 markedly reduced vimentin and FSP1 expression, and increased E-cadherin expression in HCT116 cells, and also inhibited the changes caused by VEGF treatment. Furthermore, we
1928
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
Fig. 6. Schematic representation of the proposed mechanism leading to nucleolin distribution and phosphorylation upon VEGF stimulation. Binding of exogenous VEGF to its receptor, VEGFR, leads to PI3K/Akt pathway activation and PI3K-p55 phosphorylation. Activated PI3K-p55 interacts with nucleolin, either directly or indirectly, causing phosphorylation of nucleolin and a change in nucleolin distribution.
proved that LY294002 could block the process of EMT actived by VEGF (Supplement Fig. 6). Then, metastatic capacity of cells treated with VEGF and/or HB19 were analyzed by the transwell assays. As shown in Fig. 5E, VEGF significantly enhanced the metastatic capacity of HCT116 cells, while HB19 treatment decreased the metastatic capacity of these cells (P < 0.001). It was also showned that HB19 treatment was able to abolish the effects of VEGF on the metastatic capacity of HCT116 cells. 4. Discussion Colorectal carcinoma is one of the most common cancers worldwide [41], hence, a number of prognostic and predictive indicators are used to guide therapy in patients with colorectal carcinoma. Nucleolin is a protein that is overexpressed on the surface of tumor and endothelial cells. Recent studies have underlined the involvement of cell surface nucleolin in tumor growth and angiogenesis [14]. Cell surface nucleolin interacts with receptors associated with malignancies, such as ErbB1, facilitating their activation and leading to enhanced cell growth [9]. Moreover, it binds a variety of ligands that play critical roles in tumorigenesis and angiogenesis, such as hepatocyte growth factor [21], endostatin [33], tumor-homing peptide F3 [5], laminin [35], P-selectin [28], and midkine [20]. The effect of these signals, however, on the migration of nucleolin from the nucleus to the cell surface remains unclear, and not all tumor cells express cell surface nucleolin [34]. VEGF can increase tumour metastatic capacity and its receptor, VEGFR, is a target for therapy. Moreover, overexpression of the VEGFR is a marker for poor prognosis although paradoxically neither it or VEGF are useful markers for response to anti-VEGF therapy. Moreover, recent studies have revealed that cell surface nucleolin regulates human endothelial and glioma cell migration [20]; hence, we attempted to determine whether VEGF and nucleolin distribution in tumors are related. In this study, we proved that VEGF regulates the distribution of nucleolin in colon cancer cells. It has been suggested that both the regulation of the nucleuscell surface shuttling of nucleolin by phosphorylation and the reported related physical interaction between nucleolin and PI3K, indicates a possible direct regulation of nucleolin by PI3K [18,28]. Previous studies have proven that the distribution of nucleolin at the cell surface is caused by phosphorylation of nucleolin [6]; we proposed that VEGF would phosphorylate nucleolin and cause it to move to the cell surface. VEGF can activate expression of many downstream genes, for example, those encoding protein kinases
(e.g., PKC, Akt) [12], and thus could regulate the distribution of nucleolin by the PI3K/Akt pathway and lead to phosphorylation of nucleolin by such protein kinases. PI3K-p55 is one of the regulatory subunits of PI3K, which plays an important role in the regulation of PI3K activity. Previous studies have shown that PI3K-p55 is mainly located in the nucleus and can be regulated via the PI3K/ Akt pathway [15]. In this paper, we demonstrated that nucleolin interacts with the phospho-PI3K-p55 subunit, and that the distribution of nucleolin is related to phosphorylation of nucleolin. Nucleolin and its phosphorylation have been proved to be involved in PRL-3-mediated cancer progression/metastasis signaling [32]. In the present study, we found that nucleolin phosphorylation and re-location were involved in tumor metastasis. These findings may provide some clues for tumor diagnosis and the treatment of tumor metastasis. Recently, studies have shown that cytoplasmic staining and nucleolar staining of nucleolin correlated with patients’ prognosis in gastric cancer [27]. In this paper, we demonstrated that lymph node metastases samples show more cell cytoplasmic staining for nucleolin than those with adenocarcinoma in situ. This indicated that nucleolin distribution may be related to the process of metastasis in colorectal carcinoma. Other studies have demonstrated that high levels of nucleolar nucleolin is associated with a better prognosis in pancreatic ductal adenocarcinoma patients [25]; this further supporting the possibility that nucleolin has a biologically functional role in tumours that progress to a more malignant phenotype. EMT is an indispensable process that is associated with normal tissue development and organogenesis, as well as with tissue remodeling and wound healing [19]. The involvement of EMT in cancer metastasis is based on the observation that acquisition of mesenchymal markers, such as vimentin or FSP1 by epithelial carcinoma cells is associated with increased metastatic potential, and loss of epithelial cell adhesion molecules, such as E-cadherin [2,24]. In this paper, we found that blocking cell surface distribution of nucleolin by treatment with specific inhibitor HB19 inhibited the process of EMT, indicating that nucleolin located in cell surface plays an important role in EMT. Thus inhibition of cell surface nucleolin may be a promising theraputic approach for preventing tumor metastasis. In this paper, we showed that VEGF expression were corelated with nucleolin distribution in clinical samples of colorectal carcinoma. VEGF treatment promoted phosphorylation and re-location of nucleolin through PI3K/Akt pathway in colorectal tumor cell lines. Further the data showed that the distribution of nucleolin was related to tumor metastasis both in clinical tumor samples and in cellular models in vitro (Fig. 6). Our findings indicate that
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
nucleolin may serve as a promising treatment target for colorectal carcinoma.
[20]
Funding source This study was funded by the National Key Basic Research Program of China (2010 CB 529900) and the National Natural Science Foundation of China (81102062 and 81071818). Conflict of interests
[21]
[23] [24]
The authors have no conflict of interests to declare. Acknowledgements We thank Dr. Chun-laiNie (State Key Laboratory of Biotherapy and Cancer Center, West China Hospital, West China Medical School, Sichuan University) for providing HCT116 and DLD-1 cells.
[25]
[26]
[27]
Appendix A. Supplementary data [28]
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.febslet.201 4.03.047.
[30] [31]
References [32] [1] Bremer, S., Klein, K., Sedlmaier, A., Abouzied, M., Gieselmann, V. and Franken, S. (2013) Hepatoma-derived growth factor and nucleolin exist in the same ribonucleoprotein complex. BMC Biochem. 14, 2. [2] Cannito, S., Novo, E., Compagnone, A., Valfrè di Bonzo, L., Busletta, C., Zamara, E., et al. (2008) Redox mechanisms switch on hypoxia-dependent epithelialmesenchymal transition in cancer cells. Carcinogenesis 29, 2267–2278. [3] Cascio, S., Ferla, R., D’Andrea, A., Gerbino, A., Bazan, V., Surmacz, E., et al. (2009) Expression of angiogenic regulators, VEGF and leptin, is regulated by the EGF/ PI3K/STAT3 pathway in colorectal cancer cells. J. Cell. Physiol. 221, 189–194. [4] Chekhonin, V.P., Shein, S.A. and Korchagina, A.A. (2013) GurinaOI. VEGF in tumor progression and targeted therapy. Curr. Cancer Drug Targets 13, 423– 443. [5] Christian, S., Pilch, J., Akerman, M.E., Porkka, K., Laakkonen, P. and Ruoslahti, E. (2003) Nucleolin expressed at the cell surface is a marker of endothelial cells in angiogenic blood vessels. J. Cell Biol. 4, 871–878. [6] Ding, Y., Song, N., Liu, C., He, T., Zhuo, W., He, X., et al. (2012) Heat shock cognate 70 regulates the translocation and angiogenic function of nucleolin. Arterioscler. Thromb. Vasc. Biol. 32, e126–e134. [8] El Khoury, D., Destouches, D., Lengagne, R., Krust, B., Hamma-Kourbali, Y., Garcette, M., et al. (2010) Targeting surface nucleolin with a multivalent pseudopeptide delays development of spontaneous melanoma in RET transgenic mice. BMC Cancer 10, 325–337. [9] Farin, K., Schokoroy, S., Haklai, R., Cohen-Or, I., Elad-Sfadia, G., Reyes-Reyes, M.E., et al. (2011) Oncogenic synergism between ErbB1, nucleolin, and mutant Ras. Cancer Res. 6, 2040–2051. [11] J. Ferlay, I. Soerjomataram, M. Ervik, R. Dikshit, S. Eser, C. Mathers, et al. GLOBOCAN 2012 v1.0, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 11 [Internet]. [12] Gliki, G., Wheeler-Jones, C. and Zachary, I. (2002) Vascular endothelial growth factor induces protein kinase C (PKC)-dependent Akt/PKB activation and phosphatidylinositol 30 -kinase-mediates PKC delta phosphorylation: role of PKC in angiogenesis. Cell Biol. Int. 26, 751–759. [14] Hovanessian, A., Soundaramourty, C., El Khoury, D., Nondier, I., Svab, J. and Krust, B. (2010) Surface expressed nucleolin is constantly induced in tumor cells to mediate calcium-dependent ligand internalization. PLoS ONE 5 (12), e15787. [15] Hu, J., Xia, X., Cheng, A., Wang, G., Luo, X., Reed, M.F., et al. (2008) A peptide inhibitor derived from p55PIK phosphatidylinositol 3-kinase regulatory subunit: a novel cancer therapy. Mol. Cancer Ther. 7, 3719–3728. [16] Huang, Y., Shi, H., Zhou, H., Song, X., Yuan, S. and Luo, Y. (2006) The angiogenic function of nucleolin is mediated by vascular endothelial growth factor and nonmuscle myosin. Blood 107, 3564–3571. [17] Huang, C.J., Teng, H.W., Chien, C.C., Lin, J.K. and Yang, S.H. (2013) Prognostic significance of C – reactive protein polymorphism and KRAS/BRAF in synchronous liver metastasis from colorectal cancer. PLoS ONE 8, e65117. [18] Huddleson, J.P., Ahmad, N. and Lingrel, J.B. (2006) Up-regulation of the KLF2 transcription factor by fluid shear stress requires nucleolin. J. Biol. Chem. 281, 15121–15128. [19] Kiefel, H., Bondong, S., Pfeifer, M., Schirmer, U., Erbe-Hoffmann, N., Schäfer, H., et al. (2012) EMT-associated up-regulation of L1CAM provides insights into
[33]
[34] [35]
[37]
[38]
[39] [41]
1929
L1CAM-mediated integrin signaling and NF-kappaB activation. Carcinogenesis 33, 1919–1929. Koutsioumpa, M., Polytarchou, C., Courty, J., Zhang, Y., Kieffer, N., Mikelis, C., et al. (2013) Interplay between alphavbeta3 integrin and nucleolin regulates human endothelial and glioma cell migration. J. Biol. Chem. 288, 343–354. Lee, Y., Marquez, A., Mungunsukh, O. and Day, R. (2010) Hepatocyte growth factor inhibits apoptosis by the profibrotic factor angiotensin II via extracellular signal-regulated kinase 1/2 in endothelial cells and tissue explants. Mol. Biol. Cell 23, 4240–4250. Lyon, France. International Agency for Research on Cancer 2013. Available from: (accessed day/month/year.). Nam, E.H., Lee, Y., Park, Y.K., Lee, J.W. and Kim, S. (2012) ZEB2 upregulates integrin alpha5 expression through cooperation with Sp1 to induce invasion during epithelial-mesenchymal transition of human cancer cells. Carcinogenesis 33, 563–571. Peng, L., Liang, J., Wang, H., Song, X., Rashid, A., Gomez, H.F., et al. (2010) High levels of nucleolar expression of nucleolin are associated with better prognosis in patients with stage II pancreatic ductal adenocarcinoma. Clin. Cancer Res. 16, 3734–3742. Pichiorri, F., Palmieri, D., De Luca, L., Consiglio, J., You, J., Rocci, A., et al. (2013) In vivo NCL targeting affects breast cancer aggressiveness through miRNA regulation. J. Exp. Med. 210, 951–968. Qiu, W., Zhou, F., Zhang, Q., Sun, X., Shi, X., Liang, Y., et al. (2013) Overexpression of nucleolin and different expression sites both related to the prognosis of gastric cancer. APMIS, http://dx.doi.org/10.1111/apm.12131. Reyes-Reyes, E.M. and Akiyama, S.K. (2008) Cell-surface nucleolin is a signal transducing P-selectin binding protein for human colon carcinoma cells. Exp. Cell Res. 314, 2212–2223. Siegel, Rebecca, Naishadham, Deepa and Jemal, Ahmedin (2013) Cancer Statistics. CA Cancer J. Clin. 2013 (63), 11–30. Saigusa, S., Inoue, Y., Tanaka, K., Toiyama, Y., Okugawa, Y., Shimura, T., et al. (2013) Decreased expression of DUSP4 is associated with liver and lung metastases in colorectal cancer. Med. Oncol. 30, 620. Semba, S., Mizuuchi, E. and Yokozaki, H. (2010) Requirement of phosphatase of regenerating liver-3 for the nucleolar localization of nucleolin during the progression of colorectal carcinoma. Cancer Sci. 10, 2254–2261. Song, N., Ding, Y., Zhuo, W., He, T., Fu, Z., Chen, Y., et al. (2012) The nuclear translocation of endostatin is mediated by its receptor nucleolin in endothelial cells. Angiogenesis 4, 697–711. Storck, S., Shukla, M., Dimitrov, S. and Bouvet, P. (2007) Functions of the histone chaperone nucleolin in diseases. Subcell. Biochem. 41, 125–144. Turck, N., Lefebvre, O., Gross, I., Gendry, P., Kedinger, M. and Simon-Assmann, P. (2006) Effect of laminin-1 on intestinal cell differentiation involves inhibition of nuclear nucleolin. J. Cell. Physiol. 2, 545–555. Wise, J.F., Berkova, Z., Mathur, R., Zhu, H., Braun, F.K., Tao, R.H., et al. (2013) Nucleolin inhibits Fas ligand binding and suppresses Fas-mediated apoptosis in vivo via a surface nucleolin-Fas complex. Blood 121, 4729–4739. Wu, C.D., Chou, H.W., Kuo, Y.S., Lu, R.M., Hwang, Y.C., Wu, H.C., et al. (2012) Nucleolin antisense oligodeoxynucleotides induce apoptosis and may be used as a potential drug for nasopharyngeal carcinoma therapy. Oncol. Rep. 27, 94– 100. Abdelmohsen, K. and Gorospe, M. (2012) RNA-binding protein nucleolin in disease. RNA Biol. 9 (6), 799–808. Zinser-Sierra, J.W., Rodríguez-Ramírez, S., Villalobos-Valencia, R. and RamírezMárquez, M. (2011) Use of bevacizumab in metastatic colorectal cancer: report from the Mexican opinion and analysis forum on colorectal cancer treatment with bevacizumab (September 2009). Drugs R D 11, 101–111.
Further reading [7] Dumler, I., Stepanova, V., Jerke, U., Mayboroda, O.A., Vogel, F., Bouvet, P., et al. (1999) Urokinase-induced mitogenesis is mediated by casein kinase 2 and nucleolin. Curr. Biol. 9, 1468–1476. [10] Fahling, M., Steege, A., Perlewitz, A., Nafz, B., Mrowka, R., Persson, P.B., et al. (2005) Role of nucleolin in posttranscriptional control of MMP-9 expression. Biochim. Biophys. Acta 1731, 32–40. [13] Hegde, P.S., Jubb, A.M., Chen, D., Li, N.F., Meng, Y.G., Bernaards, C., et al. (2013) Predictive impact of circulating vascular endothelial growth factor in four phase III trials evaluating bevacizumab. Clin. Cancer Res. 19, 929–937. [22] Litchfield, L.M., Riggs, K.A., Hockenberry, A.M., Oliver, L.D., Barnhart, K.G., Cai, J., et al. (2012) Identification and characterization of nucleolin as a COUP-TFII coactivator of retinoic acid receptor beta transcription in breast cancer cells. PLoS ONE 7, e38278. [29] Reyes-Reyes, E.M., Teng, Y. and Bates, P.J. (2010) A new paradigm for aptamer therapeutic AS1411 action: uptake by macropinocytosis and its stimulation by a nucleolin-dependent mechanism. Cancer Res. 70, 8617–8629. [36] Weickhardt, A.J., Tebbutt, N.C. and Mariadason, J.M. (2010) Strategies for overcoming inherent and acquired resistance to EGFR inhibitors by targeting downstream effectors in the RAS/PI3K pathway. Curr. Cancer Drug Targets 10, 824–833. [40] Xu, Q., Sun, Q., Zhang, J., Yu, J., Chen, W. and Zhang, Z. (2013) Downregulation of miR-153 contributes to epithelial-mesenchymal transition and tumor metastasis in human epithelial cancer. Carcinogenesis 34, 539–549.
journal homepage: www.FEBSLetters.org
Phosphorylation and changes in the distribution of nucleolin promote tumor metastasis via the PI3K/Akt pathway in colorectal carcinoma Dong-ming Wu a,1, Peng Zhang b,1, Ru-yan Liu c, Ya-xiong Sang a, Cong Zhou a, Guang-chao Xu a, Jin-liang Yang a, Ai-ping Tong a,⇑, Chun-ting Wang a,⇑ a b c
State Key Laboratory of Biotherapy and Cancer Center, West China Hospital, West China Medical School, Sichuan University, Chengdu, PR China Department of Radiation Oncology, Sichuan Cancer Hospital, Chengdu, Sichuan, PR China Graduate School, Guangxi Medical University, Nanning, Guangxi, PR China
a r t i c l e
i n f o
Article history: Received 12 November 2013 Revised 22 March 2014 Accepted 24 March 2014 Available online 5 April 2014 Edited by Zhijie Chang Keywords: Nucleolin Metastatic VEGF Distribution PI3K/Akt Colorectal carcinoma
a b s t r a c t Here, we investigated the molecular mechanism underlying the changes in the distribution of nucleolin. Our study identified PI3K/Akt signaling as an essential pathway regulating the distribution of nucleolin. Furthermore, nucleolin can interact with phospho-PI3K-p55, and changes in the distribution of nucleolin were related to its phosphorylation. Subsequently, we analyzed the correlation of VEGF and nucleolin, and found that distribution of nucleolin related to metastatic potential. Finally, blocking cell surface nucleolin influences the process of epithelial–mesenchymal transitions. This indicates that nucleolin may be a novel cancer therapy target and a predictive marker for tumor migration in colorectal carcinoma. Ó 2014 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
1. Introduction Worldwide, colorectal cancer is the fourth most common tumor in men and the third most common in women [11, 23]. Tumor metastasis remains the principal cause of high mortality rate and poor prognosis in patients with colorectal cancer; colorectal carcinoma that is localized within the colon or has only spread to the lymph nodes is curable by surgery with or without chemotherapy, and has a 5-year survival rate of 70% and 90%, respectively. However, cancer that has metastasized to distant sites is generally incurable and has a 5-year survival rate of 12% [30]. Hence, many studies aimed at identifying a novel predictor or therapy target for colorectal carcinoma have been reported [17,31]. Nucleolin is a multifunctional protein with well-characterized roles in the organization of nucleolar chromatin, packaging of pre-rRNA, transcription of rDNA, and ribosome assembly [1]. The importance of cell surface nucleolin has been suggested by studies ⇑ Corresponding authors. Address: State Key Laboratory of Biotherapy, West China Hospital, West China Medical School, Sichuan University, Gaopeng Street, Keyuan Road 4, Chengdu 610041, PR China. Fax: +86 028 85164060. E-mail addresses: [email protected] (A.-p. Tong), [email protected]. cn (C.-t. Wang). 1 These authors contributed equally to this work.
showing that functional blockade or down-regulation of cell surface nucleolin in endothelial cells inhibits migration and capillary-tubule formation and causes endothelial cell apoptosis [37].Targeting nucleolin on the plasma membrane of cancer cells seems to be an effective approach to inhibit cancer cell growth and angiogenesis in various in vitro and in vivo experimental models [26]. In recent years, studies have suggested that high levels of nucleolar expression of nucleolin are associated with a better prognosis in pancreatic ductal adenocarcinoma patients [25], and that nucleolin on the cell surface may be a novel cancer therapy target (Abdelmohsen and Gorospe, 2012), [38,39]. Previous studies have indicated that VEGF can increase the cell surface nucleolin expression in human microvascular endothelial cells [16]. VEGF plays a major role in tumor angiogenesis and migration; its expression is inversely correlated with survival of patients in many human cancers, including colorectal carcinomas. Both cell surface nucleolin and VEGF are upregulated in different types of cancer and have been implicated in migration progression [3,4]. Here, we showed that VEGF, through the VEGFR, leads to phosphorylation of nucleolin and that this then results in nucleolin localization to the membrane. We also found that distribution of nucleolin is related to the phosphorylation states of nucleolin.
http://dx.doi.org/10.1016/j.febslet.2014.03.047 0014-5793/Ó 2014 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.
1922
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
Our findings suggested that increased expression of cell-surface and cytosolic nucleolin were related to tumor metastasis. Thus, nucleolin may be a novel theraputic target for colorectal cancer, while the distribution of nucleolin may be a predictive indicator in clinical diagnosis. 2. Materials and methods
ylmethanesulfonyl fluoride) and incubated for 30 min on ice; lysates were then centrifuged at 18 000g for 30 min at 4 °C and the supernatants (500 mg) incubated for 2 h at 4 °C with 2 mg of anti-nucleolin antibody. Concurrently, 2 mg of rabbit IgG was used as the control. Immunoprecipitated samples were washed three times with RIPA buffer; these samples as well as pre-immunoprecipitated samples (50 mg) were subjected to Western blotting analysis using an anti-nucleolin and an anti-PI3K antibody.
2.1. Materials Recombinant human VEGF165 was purchased from PeproTech (Suzhou, China). The anti-nucleolin, anti-PI3K, and anti-Akt polyclonal antibodies were obtained from ProteinTech Group, Inc. (Wuhan, China). The phospho-Akt (Thr308) and phospho-PI3 kinase P85 (Tyr458)/P55 (Tyr199) antibodies were purchased from Cell Signaling Technology, Inc. (Danvers, MA). The phosphoserine/ threonine/tyrosine antibody was purchased from GeneTex International Corporation (CA, USA). Anti-Thr76/Thr84-phosphorylated nucleolin antibody was purchased from BioLegend (San Diego, USA). The secondary antibody (goat anti-rabbit IgG) was purchased from Santa Cruz Biotechnology (Heidelberg, Germany). Tumor microarrays were purchased from Biomax US (MD, USA). HB19 (Cell surface nucleolin inhibitor) and 9 Args were synthetized by Shanghai Science Peptide (Shanghai, China). All other kits or reagents were purchased from the Beyotime Institute of Biotechnology (Shanghai, China). 2.2. Immunofluorescence Cells were fixed with 4% paraformaldehyde, and then blocked with PBS containing 10% normal goat serum. Cells were then stained with anti-nucleolin polyclonal antibody for 30 min at 37 °C, and then stained with Cy3-conjugated secondary antibody for 30 min at 37 °C. Then those samples were re-stained by Hochest 33258 for 15 min at 37 °C. All immunofluorescence images were obtained using an Olympus BX51 microscope equipped with either a 20 or a 40 objective lens (Olympus) and a DP 50 camera (Olympus). Images were processed using DPC controller software (Olympus). 2.3. Fluorescence-activated cell-sorting (FACS) analysis After treatment, cells were trypsinized with 0.02% EDTA, and resuspended in PBS supplemented with 0.1% bovine serum albumin. Cells were incubated with anti-nucleolin antibody for 30 min on ice, and then incubated on ice with anti-rabbit secondary antibody conjugated to Cy3. Then, cells were directly analyzed using a FACScan (Becton Dickinson, San Jose, CA).
2.6. Immunohistochemical staining and assessment of VEGF and nucleolin expression Tissues were dewaxed and antigens retrieved using high pressure for 3 min. Then, the activity of endogenous peroxidases was blocked using 3% hydrogen peroxide for 10 min at room temperature. After immersion in normal goat serum for 30 min at 37 °C, tissues were subjected to primary antibody incubation at 4 °C overnight. Subsequently, sections were washed with PBS for 15 min at room temperature and incubated with secondary antibody that had been labeled with biotin for 30 min at 37 °C. After washing with PBS for 15 min at room temperature, the sections were incubated with horseradish peroxidase (HRP) complex for 30 min at 37 °C and visualized using diaminobenzidine (DAB). All immunohistochemical images were obtained using an Olympus BX51 microscope equipped with either a 20, a 40, or 100 objective lens (Olympus) and a DP 50 camera (Olympus). Images were processed using DPC controller software (Olympus). The immunohistochemical staining was evaluated using a semiquantitative scoring method. The scores were decided by two pathologists independently. The total nucleolin and VEGF staining was scored as per staining intensity: no staining (0), light positive staining (1), medium positive staining (2), and strong positive staining (3). The positively stained area was scored as: <5% (0), 5–25% (1), 26–50% (2), 51–75% (3), and >75% (4). A final score was calculated by multiplying the intensity score by the expression score and the rating (from 0 to 12) was determined for each case. 2.7. Statistical analysis Each experiment was performed at least three times independently. The associations of VEGF and nucleolin with clinicopathological parameters were generated using a paired t-test or one-way ANOVA analyses implemented in GraphPad Prism 5 (GraphPad, San Diego, CA). Statistical significance was defined as P < 0.05. 3. Results 3.1. VEGF changes the distribution of nucleolin in colon cancer cells
2.4. Subcellular fractionation Subcellular fractions of cells comprising cytosolic, nuclear, and cell membrane extracts were prepared as follows. Cell monolayers in 100-mm plates were washed extensively with PBS before being scraped and pelleted. Washed cells (2 106) were then used for extraction of cytosolic, nuclear, and cell membrane proteins using a Membrane and Cytosol Protein Extraction Kit and a Nuclear and Cytoplasmic Protein Extraction Kit according to the manufacturer’s recommendations. These samples were then analyzed by Western blot or co-immunoprecipitation analysis. 2.5. Co-immunoprecipitation assay Cells were resuspended in ice-cold RIPA (Radio-Immunoprecipitation Assay) cell lysis buffer containing 1 mM PMSF (Phen-
In order to investigate the function of VEGF in the distribution of nucleolin expression in colon cancer, HCT116 cells and DLD-1 cells treated with 0, 5, 10, or 20 ng/ml VEGF for 24 h were harvested for FACS analysis. We found that VEGF increased nucleolin expression on the cell surface of HCT116 and DLD-1 cells, in a concentration-dependent manner (Fig. 1A). To further assess the regulation of nucleolin distribution in colon cells by VEGF, HCT116 and DLD-1 cells were stained with an anti-nucleolin antibody for immunofluorescence analysis. As shown in Fig. 1B, we found that cell surface nucleolin was increased after VEGF treatment, in a dose-dependent manner, and there was no change in the mRNA level of nucelolin (Supplement Fig. 2). To further confirm that VEGF can change the distribution of nucleolin in colon cancer cells, we treated HCT116 and DLD-1 cells with different doses of VEGF; after 24 h, we separated the
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
1923
Fig. 1. VEGF changes the distribution of nucleolin in colon cancer cells. (A) Flow cytometry showed that VEGF treatment of cells increased cell surface nucleolin expression in HCT116 cells (Left) and DLD-1 cells (Right). (B) Immunofluorescence analysis cell surface nucleolin after cells were treated with VEGF in HCT116 cells (Up) and DLD-1 cells (Down). (C) The differences in the distribution of nucleolin expression were analyzed after treatment of cells with the indicated dose of VEGF in HCT116 cells (Up) and DLD-1 cells (Down).
1924
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
subcellular fractions of these cells, and analyzed the expression of nucleolin in the cytosolic, nuclear, and cell membrane extracts by Western blotting. As shown in Supplement Fig. 1, the separation efficiency of those kits was detected. We found that VEGF clearly increased nucleolin levels in the cell membrane and cytoplasmic fractions, while nucleolin levels in the nucleus was decreased, and the total nucleolin remained unchanged in both HCT116 and DLD-1 cells (Fig. 1C). 3.2. VEGF changes the distribution of nucleolin via the PI3K/Akt pathway It has been suggested that the migration of nucleolin in the cytosol and nucleus to the cell membrane occurs via the PI3K/ Akt pathway [20]. We tested this suggestion: HCT116 cells were
cultured with 50 lM LY294002 (PI3K inhibitor) for 5 h, and then treated with VEGF; 24 h later, cells were harvested for FACS analysis. As shown in Fig. 2A, the cell surface nucleolin levels were decreased when cells were pretreated with LY294002, as compared to cells not treated with LY294002, and there was no change in the mRNA level of nucelolin (Supplement Fig. 3). This indicated that blocking the PI3K/Akt pathway can reduce the cell surface nucleolin level effectively in VEGF-treated cells. Anti-nucleolin staining by immunofluorescence analysis in HCT116 cells further supported this finding: we found that LY294002 clearly blocked the VEGF-mediated distribution of nucleolin at the cell surface (Fig. 2B). We treated HCT116 cells with VEGF and/or LY294002; after 24 h, we separated the subcellular fraction of HCT116 cells, and analyzed the expression of nucleolin in the cytosolic, nuclear,
Fig. 2. VEGF changes the distribution of nucleolin via the PI3K/Akt pathway. (A) Treatment of cells with LY294002 reduced the VEGF-mediated increase in cell surface nucleolin. (B) Expression of cell surface nucleolin was determined by immunofluorescence analysis after cells were treated with VEGF, following pre-treatment with LY294002. (C) The distribution of nucleolin and cell membrane extracts as well as total nucleolin levels were analyzed using Western blotting. (D) VEGF treatment caused an increase in the phosphorylation of Akt. (E) LY294002 treatment reduced the VEGF-mediated phosphorylation of Akt.
Fig. 3. Distribution of nucleolin is related to the phosphorylation states of nucleolin. (A) VEGF increased PI3K-p55 phosphorylation. (B) Nucleolin interacts with phosphoPI3K-p55. Normal IgG was as the control. (C) VEGF increase nucleolin phosphorylation. Normal IgG was as the control. (D) VEGF increase cell surface nucleolin phosphorylation. Normal IgG was as the control. (E) Nucleolin phosphorylation is related to tumor stage. Lymph node metastases have higher phosphorylation level than colorectal carcinoma and normal lymph node. Normal IgG was as the control.
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
1925
Fig. 4. Distribution of nucleolin is associated with tumor metastasis. (A) VEGF and nucleolin expression in normal colon, an adenocarcinoma group, and a metastases group. (B) Immunohistochemical staining of VEGF (top) and nucleolin (bottom) in normal colon (left), adenocarcinoma (middle), and metastases (right). (C) VEGF (left) and nucleolin (right) expression was investigated in clinical colorectal cancer samples. Immunohistochemical staining of VEGF and nucleolin in normal colon (top), adenocarcinoma (middle), and lymph node metastases and liver metastases (bottom) indicated that distribution of nucleolin correlates with tumor metastasis.
and cell membrane extracts, as well as the total nucleolin in cells, by Western blotting. We found that the nucleolin levels in the cytosolic and cell membrane fractions decreased, while nuclear nucleolin levels increased, without changes in total nucleolin levels after LY294002 treatment. This indicated that LY294002 blocked the distribution of nucleolin from the nucleus (Fig. 2C).
We next evaluated the phosphorylation states of Akt in HCT116 cells. As shown in Fig. 2D, the phosphorylation levels of Akt increased with VEGF treatment in a concentration-dependent manner, while total Akt levels remained unchanged. After treatment with LY294002, the phosphorylation of Akt was markedly decreased (Fig. 2E).
1926
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
Table 1 Clinicopathological association of nucelolin with colorectal carcinoma. Factors
Number of patients
Average score
P value
Group Adenocarcinoma Metastatic adenocarcinoma
30 30
3.43 ± 0.55 4.03 ± 0.65
0.49
Gender Male Female
22 8
3.68 ± 0.69 2.75 ± 0.86
0.47
Age <60 years P60 years
21 9
2.91 ± 0.63 4.67 ± 1.05
0.15
Differentiation Well Moderate Poor
12 27 15
3.83 ± 0.90 4.74 ± 0.70 3.00 ± 0.64
0.25
Clinical stage I II III IV
5 16 6 3
2.60 ± 1.03 4.17 ± 0.79 4.67 ± 1.50 2.25 ± 1.32
0.52
Clinical and TNM stagings were according to American Joint Committee on Cancer: AJCC Cancer Staging Manual, 6th edition.
Table 2 Clinicopathological association of VEGF with colorectal carcinoma. Factors
Number of patients
Average score
P value
Group Adenocarcinoma Metastatic adenocarcinoma
30 30
4.93 ± 0.40 6.80 ± 0.70
0.02
Gender Male Female
22 8
5.18 ± 0.48 4.75 ± 0.94
0.66
Age <60 years P60 years
21 9
4.52 ± 0.49 5.78 ± 0.68
0.16
Differentiation Well Moderate Poor
12 27 15
4.58 ± 0.73 5.29 ± 0.53 6.00 ± 0.84
0.46
Clinical stage I II III IV
5 16 6 3
3.40 ± 0.87 5.44 ± 0.60 5.00 ± 1.03 5.00 ± 2.00
0.47
Clinical and TNM stagings were according to American Joint Committee on Cancer: AJCC Cancer Staging Manual, 6th edition.
3.3. Distribution of nucleolin is related to the phosphorylation states of nucleolin It has been demonstrated that nucleolin physically interacts with PI3K [18]; therefore, we tested whether VEGF increases PI3K phosphorylation levels in HCT116 cells. As shown in Fig. 3A, VEGF promoted PI3K-P55 phosphorylation, while total PI3K and nucleolin levels remained unchanged in HCT116 cells. Given that nucleolin interacts with PI3K, we tested its phosphorylation subunits, and found that nucleolin interacts with the phospho-PI3K-p55 subunit (Fig. 3B). This indicated that nucleolin interacts with phosphorylated p55 in the cell nucleus and then translocates from the nucleus to the cytosol and cell membrane. We next tested the phosphorylation state of nucleolin in HCT116 cells, by immunoprecipitating total nucleolin, and then analyzing the phosphorylation state of nucleolin using a phosphoserine/threonine/tyrosine antibody. After treatment of cells with VEGF, we found that the phosphorylation of nucleolin increased.
We next tested whether LY294002 could block the nucleolin phosphorylation; we found that LY294002 could reduce the VEGF-mediated increase in nucleolin phosphorylation (Fig. 3C). Subsequently, we analyzed the correlation between different phosphorylation states of nucleolin and differences in the distribution of nucleolin were analyzed. We found that distribution of nucleolin was related to the phosphorylation state of nucleolin in HCT116 cells; phosphorylation of nucleolin was only observed in the cytosolic and cell membrane extracts, but not in nuclear nucleolin. Furthermore, cytosolic and cell membrane extracts demonstrated greater levels of nucleolin phosphorylation after VEGF treatment, but this phosphorylation was blocked when LY294002 was added (Fig. 3D). In order to test whether different stages of colorectal carcinoma show different phosphorylation levels of nucleolin, we analyzed the nucleolin phosphorylation level in colorectal carcinoma samples and compared these levels with those in lymph node metastases. We found that lymph node metastases had higher nucleolin phosphorylation levels compared to colorectal carcinoma. Then, the phosphorylation state of nucleolin in different distributions was determined by separating colorectal carcinoma and lymph node metastases tissue. We found that the cytosolic and cell membrane extracts demonstrated higher phosphorylation levels of nucleolin in the lymph node metastases than those in colorectal carcinoma (Fig. 3F). It has been proved that Thr76/Thr84-phosphorylated nucleolin play a role in process of tumors, so we tested whether the expression of phospho-Thr76/Thr84 nucleolin changed with VEGF stimulation. As shown in Supplement Fig. 4, VEGF increased phospho-Thr76/Thr84 of nucleolin, while treatment with LY294002 inhibited the effects of VEGF. 3.4. Distribution of nucleolin is association with tumor metastasis To study the role of VEGF and nucleolin in colorectal carcinoma and metastases, we assessed the levels of VEGF and nucleolin expression in colorectal carcinoma using a tissue microarray. As shown in Fig. 4A and B, there was no difference between normal colon tissue, colon adenocarcinoma tissue, and colon metastases in terms of immunohistochemical staining of total nucleolin. No significant difference was observed between well, moderately, and poorly differentiated carcinoma tissues; different clinical stages showed no significant differences in total nucleolin expression (Table 1). Statistical analysis indicated that VEGF expression was significantly lower in colon adenocarcinoma than in metastases (P = 0.0241; Table 2). We demonstrated that total nucleolin expression levels were not statistically significantly different between the adenocarcinoma group and the metastases group, which suggested that the expression of total nucleolin does not change with tumor metastasis. To further explore the roles of nucleolin distribution in colorectal carcinoma metastasis, clinicopathological association analyses were specifically performed for colorectal carcinoma and metastases. Along with the increase in metastatic capacity from colorectal carcinoma, through lymph node metastases to liver metastases, the metastatic colorectal carcinoma showed more cytosolic staining and less nuclear staining for nucleolin (Fig. 4C). 3.5. Inhibition of cell surface nucleolin by HB19 inhibits EMT in colon cancer cells Thus far, we have shown that inhibition of the migration of nucleolin to the cell surface is associated with a greater risk of tumor metastasis. Then, we investigated whether inhibition of cell surface nucleolin would affect the process of EMT in colon cancer cells. As shown in Fig. 5A and Supplement Fig. 5, we confirmed that
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
1927
Fig. 5. Inhibition of cell surface nucleolin by HB19 inhibits the process of epithelial–mesenchymal transitions (EMT) in colon cancer cells. (A) Inhibition of cell surface nucleolin by HB19. 9R was as the control. (B) Morphological changes after HCT116 cells were treated with HB19 and/or VEGF. (C) Representative immunofluorescence images of staining for E-cadherin (above) and FSP1 (below) in serum-starved HCT116 cells treated with HB19 and/or VEGF. (D) Biomarkers for EMT: vimentin, FSP1, and E-cadherin were analyzed by Western blotting. (E) Quantification of cell migration by the transwell assay; data from three independent experiments are shown (P < 0.05).
the synthesis of peptide HB19 [8] can block cell surface nucleolin expression in HCT116 cells. 9R (9 Args) was as the control of HB19. Treatment with HB19 promoted morphological changes in these cells that were characterized by loss of the ‘‘fibroblastoidlike’’ phenotype and acquisition of an epithelial-like compact morphology (Fig. 5B). As shown in Fig. 5C, immunofluorescence staining revealed that HB19 upregulated expression of the epithelial
maker E-cadherin and decreased expression of the mesenchymal maker fibroblast-specific protein 1 (FSP1) in HCT116 cells. We next analyzed cells treated with VEGF and/or HB19 with biomarkers for EMT: vimentin, FSP1and E-cadherin (Fig. 5D). HB19 markedly reduced vimentin and FSP1 expression, and increased E-cadherin expression in HCT116 cells, and also inhibited the changes caused by VEGF treatment. Furthermore, we
1928
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
Fig. 6. Schematic representation of the proposed mechanism leading to nucleolin distribution and phosphorylation upon VEGF stimulation. Binding of exogenous VEGF to its receptor, VEGFR, leads to PI3K/Akt pathway activation and PI3K-p55 phosphorylation. Activated PI3K-p55 interacts with nucleolin, either directly or indirectly, causing phosphorylation of nucleolin and a change in nucleolin distribution.
proved that LY294002 could block the process of EMT actived by VEGF (Supplement Fig. 6). Then, metastatic capacity of cells treated with VEGF and/or HB19 were analyzed by the transwell assays. As shown in Fig. 5E, VEGF significantly enhanced the metastatic capacity of HCT116 cells, while HB19 treatment decreased the metastatic capacity of these cells (P < 0.001). It was also showned that HB19 treatment was able to abolish the effects of VEGF on the metastatic capacity of HCT116 cells. 4. Discussion Colorectal carcinoma is one of the most common cancers worldwide [41], hence, a number of prognostic and predictive indicators are used to guide therapy in patients with colorectal carcinoma. Nucleolin is a protein that is overexpressed on the surface of tumor and endothelial cells. Recent studies have underlined the involvement of cell surface nucleolin in tumor growth and angiogenesis [14]. Cell surface nucleolin interacts with receptors associated with malignancies, such as ErbB1, facilitating their activation and leading to enhanced cell growth [9]. Moreover, it binds a variety of ligands that play critical roles in tumorigenesis and angiogenesis, such as hepatocyte growth factor [21], endostatin [33], tumor-homing peptide F3 [5], laminin [35], P-selectin [28], and midkine [20]. The effect of these signals, however, on the migration of nucleolin from the nucleus to the cell surface remains unclear, and not all tumor cells express cell surface nucleolin [34]. VEGF can increase tumour metastatic capacity and its receptor, VEGFR, is a target for therapy. Moreover, overexpression of the VEGFR is a marker for poor prognosis although paradoxically neither it or VEGF are useful markers for response to anti-VEGF therapy. Moreover, recent studies have revealed that cell surface nucleolin regulates human endothelial and glioma cell migration [20]; hence, we attempted to determine whether VEGF and nucleolin distribution in tumors are related. In this study, we proved that VEGF regulates the distribution of nucleolin in colon cancer cells. It has been suggested that both the regulation of the nucleuscell surface shuttling of nucleolin by phosphorylation and the reported related physical interaction between nucleolin and PI3K, indicates a possible direct regulation of nucleolin by PI3K [18,28]. Previous studies have proven that the distribution of nucleolin at the cell surface is caused by phosphorylation of nucleolin [6]; we proposed that VEGF would phosphorylate nucleolin and cause it to move to the cell surface. VEGF can activate expression of many downstream genes, for example, those encoding protein kinases
(e.g., PKC, Akt) [12], and thus could regulate the distribution of nucleolin by the PI3K/Akt pathway and lead to phosphorylation of nucleolin by such protein kinases. PI3K-p55 is one of the regulatory subunits of PI3K, which plays an important role in the regulation of PI3K activity. Previous studies have shown that PI3K-p55 is mainly located in the nucleus and can be regulated via the PI3K/ Akt pathway [15]. In this paper, we demonstrated that nucleolin interacts with the phospho-PI3K-p55 subunit, and that the distribution of nucleolin is related to phosphorylation of nucleolin. Nucleolin and its phosphorylation have been proved to be involved in PRL-3-mediated cancer progression/metastasis signaling [32]. In the present study, we found that nucleolin phosphorylation and re-location were involved in tumor metastasis. These findings may provide some clues for tumor diagnosis and the treatment of tumor metastasis. Recently, studies have shown that cytoplasmic staining and nucleolar staining of nucleolin correlated with patients’ prognosis in gastric cancer [27]. In this paper, we demonstrated that lymph node metastases samples show more cell cytoplasmic staining for nucleolin than those with adenocarcinoma in situ. This indicated that nucleolin distribution may be related to the process of metastasis in colorectal carcinoma. Other studies have demonstrated that high levels of nucleolar nucleolin is associated with a better prognosis in pancreatic ductal adenocarcinoma patients [25]; this further supporting the possibility that nucleolin has a biologically functional role in tumours that progress to a more malignant phenotype. EMT is an indispensable process that is associated with normal tissue development and organogenesis, as well as with tissue remodeling and wound healing [19]. The involvement of EMT in cancer metastasis is based on the observation that acquisition of mesenchymal markers, such as vimentin or FSP1 by epithelial carcinoma cells is associated with increased metastatic potential, and loss of epithelial cell adhesion molecules, such as E-cadherin [2,24]. In this paper, we found that blocking cell surface distribution of nucleolin by treatment with specific inhibitor HB19 inhibited the process of EMT, indicating that nucleolin located in cell surface plays an important role in EMT. Thus inhibition of cell surface nucleolin may be a promising theraputic approach for preventing tumor metastasis. In this paper, we showed that VEGF expression were corelated with nucleolin distribution in clinical samples of colorectal carcinoma. VEGF treatment promoted phosphorylation and re-location of nucleolin through PI3K/Akt pathway in colorectal tumor cell lines. Further the data showed that the distribution of nucleolin was related to tumor metastasis both in clinical tumor samples and in cellular models in vitro (Fig. 6). Our findings indicate that
D.-m. Wu et al. / FEBS Letters 588 (2014) 1921–1929
nucleolin may serve as a promising treatment target for colorectal carcinoma.
[20]
Funding source This study was funded by the National Key Basic Research Program of China (2010 CB 529900) and the National Natural Science Foundation of China (81102062 and 81071818). Conflict of interests
[21]
[23] [24]
The authors have no conflict of interests to declare. Acknowledgements We thank Dr. Chun-laiNie (State Key Laboratory of Biotherapy and Cancer Center, West China Hospital, West China Medical School, Sichuan University) for providing HCT116 and DLD-1 cells.
[25]
[26]
[27]
Appendix A. Supplementary data [28]
Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.febslet.201 4.03.047.
[30] [31]
References [32] [1] Bremer, S., Klein, K., Sedlmaier, A., Abouzied, M., Gieselmann, V. and Franken, S. (2013) Hepatoma-derived growth factor and nucleolin exist in the same ribonucleoprotein complex. BMC Biochem. 14, 2. [2] Cannito, S., Novo, E., Compagnone, A., Valfrè di Bonzo, L., Busletta, C., Zamara, E., et al. (2008) Redox mechanisms switch on hypoxia-dependent epithelialmesenchymal transition in cancer cells. Carcinogenesis 29, 2267–2278. [3] Cascio, S., Ferla, R., D’Andrea, A., Gerbino, A., Bazan, V., Surmacz, E., et al. (2009) Expression of angiogenic regulators, VEGF and leptin, is regulated by the EGF/ PI3K/STAT3 pathway in colorectal cancer cells. J. Cell. Physiol. 221, 189–194. [4] Chekhonin, V.P., Shein, S.A. and Korchagina, A.A. (2013) GurinaOI. VEGF in tumor progression and targeted therapy. Curr. Cancer Drug Targets 13, 423– 443. [5] Christian, S., Pilch, J., Akerman, M.E., Porkka, K., Laakkonen, P. and Ruoslahti, E. (2003) Nucleolin expressed at the cell surface is a marker of endothelial cells in angiogenic blood vessels. J. Cell Biol. 4, 871–878. [6] Ding, Y., Song, N., Liu, C., He, T., Zhuo, W., He, X., et al. (2012) Heat shock cognate 70 regulates the translocation and angiogenic function of nucleolin. Arterioscler. Thromb. Vasc. Biol. 32, e126–e134. [8] El Khoury, D., Destouches, D., Lengagne, R., Krust, B., Hamma-Kourbali, Y., Garcette, M., et al. (2010) Targeting surface nucleolin with a multivalent pseudopeptide delays development of spontaneous melanoma in RET transgenic mice. BMC Cancer 10, 325–337. [9] Farin, K., Schokoroy, S., Haklai, R., Cohen-Or, I., Elad-Sfadia, G., Reyes-Reyes, M.E., et al. (2011) Oncogenic synergism between ErbB1, nucleolin, and mutant Ras. Cancer Res. 6, 2040–2051. [11] J. Ferlay, I. Soerjomataram, M. Ervik, R. Dikshit, S. Eser, C. Mathers, et al. GLOBOCAN 2012 v1.0, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 11 [Internet]. [12] Gliki, G., Wheeler-Jones, C. and Zachary, I. (2002) Vascular endothelial growth factor induces protein kinase C (PKC)-dependent Akt/PKB activation and phosphatidylinositol 30 -kinase-mediates PKC delta phosphorylation: role of PKC in angiogenesis. Cell Biol. Int. 26, 751–759. [14] Hovanessian, A., Soundaramourty, C., El Khoury, D., Nondier, I., Svab, J. and Krust, B. (2010) Surface expressed nucleolin is constantly induced in tumor cells to mediate calcium-dependent ligand internalization. PLoS ONE 5 (12), e15787. [15] Hu, J., Xia, X., Cheng, A., Wang, G., Luo, X., Reed, M.F., et al. (2008) A peptide inhibitor derived from p55PIK phosphatidylinositol 3-kinase regulatory subunit: a novel cancer therapy. Mol. Cancer Ther. 7, 3719–3728. [16] Huang, Y., Shi, H., Zhou, H., Song, X., Yuan, S. and Luo, Y. (2006) The angiogenic function of nucleolin is mediated by vascular endothelial growth factor and nonmuscle myosin. Blood 107, 3564–3571. [17] Huang, C.J., Teng, H.W., Chien, C.C., Lin, J.K. and Yang, S.H. (2013) Prognostic significance of C – reactive protein polymorphism and KRAS/BRAF in synchronous liver metastasis from colorectal cancer. PLoS ONE 8, e65117. [18] Huddleson, J.P., Ahmad, N. and Lingrel, J.B. (2006) Up-regulation of the KLF2 transcription factor by fluid shear stress requires nucleolin. J. Biol. Chem. 281, 15121–15128. [19] Kiefel, H., Bondong, S., Pfeifer, M., Schirmer, U., Erbe-Hoffmann, N., Schäfer, H., et al. (2012) EMT-associated up-regulation of L1CAM provides insights into
[33]
[34] [35]
[37]
[38]
[39] [41]
1929
L1CAM-mediated integrin signaling and NF-kappaB activation. Carcinogenesis 33, 1919–1929. Koutsioumpa, M., Polytarchou, C., Courty, J., Zhang, Y., Kieffer, N., Mikelis, C., et al. (2013) Interplay between alphavbeta3 integrin and nucleolin regulates human endothelial and glioma cell migration. J. Biol. Chem. 288, 343–354. Lee, Y., Marquez, A., Mungunsukh, O. and Day, R. (2010) Hepatocyte growth factor inhibits apoptosis by the profibrotic factor angiotensin II via extracellular signal-regulated kinase 1/2 in endothelial cells and tissue explants. Mol. Biol. Cell 23, 4240–4250. Lyon, France. International Agency for Research on Cancer 2013. Available from:
Further reading [7] Dumler, I., Stepanova, V., Jerke, U., Mayboroda, O.A., Vogel, F., Bouvet, P., et al. (1999) Urokinase-induced mitogenesis is mediated by casein kinase 2 and nucleolin. Curr. Biol. 9, 1468–1476. [10] Fahling, M., Steege, A., Perlewitz, A., Nafz, B., Mrowka, R., Persson, P.B., et al. (2005) Role of nucleolin in posttranscriptional control of MMP-9 expression. Biochim. Biophys. Acta 1731, 32–40. [13] Hegde, P.S., Jubb, A.M., Chen, D., Li, N.F., Meng, Y.G., Bernaards, C., et al. (2013) Predictive impact of circulating vascular endothelial growth factor in four phase III trials evaluating bevacizumab. Clin. Cancer Res. 19, 929–937. [22] Litchfield, L.M., Riggs, K.A., Hockenberry, A.M., Oliver, L.D., Barnhart, K.G., Cai, J., et al. (2012) Identification and characterization of nucleolin as a COUP-TFII coactivator of retinoic acid receptor beta transcription in breast cancer cells. PLoS ONE 7, e38278. [29] Reyes-Reyes, E.M., Teng, Y. and Bates, P.J. (2010) A new paradigm for aptamer therapeutic AS1411 action: uptake by macropinocytosis and its stimulation by a nucleolin-dependent mechanism. Cancer Res. 70, 8617–8629. [36] Weickhardt, A.J., Tebbutt, N.C. and Mariadason, J.M. (2010) Strategies for overcoming inherent and acquired resistance to EGFR inhibitors by targeting downstream effectors in the RAS/PI3K pathway. Curr. Cancer Drug Targets 10, 824–833. [40] Xu, Q., Sun, Q., Zhang, J., Yu, J., Chen, W. and Zhang, Z. (2013) Downregulation of miR-153 contributes to epithelial-mesenchymal transition and tumor metastasis in human epithelial cancer. Carcinogenesis 34, 539–549.