ANTICANCER RESEARCH 28: 327-334 (2008)

Overexpression of Vimentin Contributes to Prostate Cancer Invasion and Metastasis via Src Regulation

JUNCHENG WEI*, GANG XU*, MINGFU WU, YONGTAO ZHANG, QIONG LI, PING LIU, TAO ZHU, ANPING SONG, LIANGPIN ZHAO, ZHIQIANG HAN, GANG CHEN, SHIXUAN WANG, LI MENG, JIANFENG ZHOU, YUNPING LU, SHIXUAN WANG and DING MA

Cancer Biology Research Center, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei 430030, P.R. China

Abstract. A significant proportion of prostate cancer patients Prostate carcinoma is the second leading cause of cancer- treated with curative intent develop advanced disease. At a related death in the United States and Europe (1), mainly fundamental biological level, very little is known about what due to its high potential for bone metastasis. However, the makes the disease aggressive and metastatic. Observational molecular mechanisms of prostate cancer metastasis are not pathology reports and experimental data suggest that an well understood. Among the currently available techniques, epithelial-mesenchymal transition (EMT) is involved in prostate cancer proteomics permits the analysis of thousands of cancer invasiveness. The mechanism by which vimentin modified or unmodified proteins simultaneously and has promotes prostate cancer cell invasion and metastasis was become increasingly popular for identifying biomarkers for examined. The highly metastatic human prostate epithelial cell the early detection, classification and prognosis of tumors, line PC-3M-1E8 (1E8-H) and the low metastatic line PC-3M- as well as for pinpointing molecular targets for improving 2B4 (2B4-L) were used for comparative proteomic analysis by treatment outcomes (2). two-dimensional gel electrophoresis, followed by matrix-assisted It is known that tumor progression to malignancy requires laser desorption/time of flight mass spectrometry (MALDI- a change from an epithelial phenotype to a fibroblast or TOF-MS). A transwell assay was performed to test cell mesenchymal phenotype, a re-programming known as an migration and invasion and immunoblotting was used to epithelial-mesenchymal transition (EMT) (3-6), which is analyze the relative expression of proteins. High vimentin characterized by the acquisition of an invasive, mesenchymal expression was detected in 1E8-H compared to 2B4-L cells. phenotype by epithelial cells that enables their migration into Down-regulation of vimentin in 1E8-H by antisense-vimentin a new microenvironment. E-cadherin is a prototypic type I transfection led to a significant inhibition of invasiveness, and cadherin that forms homophilic interactions through its selective stimulation of vimentin activity in 2B4-L by delivery extracellular immunoglobulin (Ig) domain, which connects to of recombinant vimentin promoted cell invasiveness. Vimentin actin filaments indirectly via ·-catenin and ‚-catenin through activity was associated with C-src, ‚-catenin and E-cadherin their cytoplasmic domains (7-9). The appropriate expression expression. PP2, a specific inhibitor of src family kinases, of E-cadherin on the plasma membrane is essential for cells reduced phospho-‚-catenin expression and induce E-cadherin to retain an epithelial morphology. Evidence from previous expression. Vimentin promotes tumor cell invasiveness and the studies has shown a correlation between loss of cell–cell targeting of vimentin/C-src may be a promising strategy for adhesion during EMT and decreased E-cadherin function, preventing or blocking prostate cancer metastasis. resulting in the aggressiveness, de-differentiation and metastasis of many carcinomas (10-17). The transgenic mouse model generated by Perl et al. has demonstrated that the loss of E-cadherin-mediated intercellular adhesion is a rate- *Both authors contributed equally to this manuscript. limiting step in the progression from adenoma to invasive carcinoma in vivo (18). The intermediate filament protein Correspondence to: Dr. Ding Ma, Cancer Biology Research Center, (IFP) vimentin, expressed in mesenchymal cells, is a well- Tongji Hospital, Tongji Medical College, Huazhong University of known marker for EMT (3). Vimentin expression and Science and Technology, Wuhan, Hubei 430030, P.R. China. Fax: 86-27-83662779 e-mail: [email protected] perturbation of E-cadherin-mediated cell adhesion are therefore both regarded as hallmarks of EMT-associated Key Words: Vimentin, prostate cancer, invasion, C-src, E- events. The heterogeneity and reversibility of E-cadherin cadherin/‚-catenin. protein production also remains a subject of controversy (19).

0250-7005/2008 $2.00+.40 327 ANTICANCER RESEARCH 28: 327-334 (2008)

Src is a signal-transducing protein kinase that plays 2-DE and image analysis. The 2-DE was performed as previously central roles in the control of cell growth and described (25) using the PROTEAN IEF and PROTEAN Plus differentiation. Overexpression and aberrant activation of Dodeca cell systems (Bio-Rad). For preparative 2-DE, the total proteins (260 Ìg/185 Ìl) premixed with re-hydration buffer were src family kinases have been identified in various human run in an isoelectric focusing (IEF) system using a pre-cast 11 cm cancer cells (20, 21). It has been demonstrated that C-src immobilized pH gradient (IPG) strip with pI ranges of 4-7 (Bio- kinases are enriched at the cell-cell adherent junctions of Rad). The total Vh was 52,000-69,000. SDS-PAGE gels were run various types of cells, including hepatocytes (22). at constant voltages of 150 V for 45 minutes and 200 V for 6 h until Transfection of a V-src mutant into Madin-Darby canine the bromophenol blue dye reached the bottom of the gels. After kidney cells led to increased association of E-cadherin with electrophoresis in the second dimension, the gels were stained with ‚-catenin, which caused a disruption of intercellular “Blue Silver”, a sensitive colloidal Coomassie G-250 stain compatible with mass spectrometry (MS) analysis (26). The gels adhesion and an increase in vulnerability to invasion (23). were scanned using a GS710 imaging densitometer (Bio-Rad) and In the present study, the human prostate epithelial analyzed by PDQuest software (Bio-Rad). cancer cell line PC-3M and its clonal cells, PC-3M-1E8 (1E8-H) cells and PC-3M-2B4 (2B4-L) cells, which possess Protein identification by matrix-assisted laser desorption/time of flight opposite metastatic potentials but similar genetic mass spectrometry (MALDI-TOF-MS). Protein identification was backgrounds (24), were selected as model systems for the performed as described previously (27), with some modifications. study of the molecular events involved in prostate Briefly, spots were cut out from the gels, destained for 20 minutes in 30 mM KCN/100 mM Na S O 1:1 (v/v) and washed with MilliQ metastasis. Using a transfection approach the proteins 2 2 3 water until the gels became clear. Background gel was cut out and associated with metastasis were screened and the potential used as a negative control. Isolated gel fragments were kept in 0.2 candidates involved in regulating tumor migration and M NH4HCO3 for 20 minutes, lyophilized and digested overnight in invasion were investigated. Preliminary data showed that 12.5 ng/l trypsin containing 0.1 M NH4HCO3. The peptides were vimentin is one of the predominant overexpressed proteins extracted three times with 50% acetonitrile (CAN) containing 0.1% in the highly metastatic cell line 1E8-H compared to the trifluoroacetic acid (TFA), mixed with the hydroxycinnamic (HCCA) low metastatic potential 2B4-L cells. Our studies therefore matrix (Sigma), applied to the target and analyzed using a Bruker REFLEX III MALDI-TOF mass spectrometer (BrukerFranzen focused on the role of vimentin in mediating tumor Analytik, Bremen, Germany). Protein identification using peptide metastasis and the possibility of cross-linking between mass fingerprinting was performed by Mascot search vimentin and C-src, as well as sustained E-cadherin/‚- (http://www.matrixscience.com, MatrixScience Ltd, London, UK) catenin complex conformation. against the National Center for Biotechnology Information (NCBI) non-redundant protein database. The errors in peptide masses Materials and Methods were within the range of 0.01%-0.1%. One missed tryptic cleavage site per peptide was allowed during the search. Cysteines were Cell culture and reagents. The human prostate epithelial cancer cell carboamidomethylated and methionines were oxidated. Proteins lines 1E8-H and 2B4-L were obtained from the Peking University matching more than four peptides and with a Mascot score higher Health Science Center (Beijing, China) and were cultured in RPMI than 63 were considered significant (p<0.05). growth media supplemented with 10% fetal bovine serum (FBS), 100 Ìg/ml streptomycin and 100 U/ml penicillin. The cells were Immunofluorescence microscopy. The cells were seeded onto sterile kept at 37ÆC in a humidified atmosphere with 5% CO2. PP2 (src glass coverslips at 50% confluency. After 24 h, they were fixed at kinase inhibitor) was purchased from Calbiochem (La Jolla, CA, room temperature in 3.7% formaldehyde in PBS for 10 min and USA). A stock solution of PP2 was prepared in Me2SO (Sigma, St. then washed three times in PBS. The cells were permeabilized for Louis, MO, USA) and stored at –70ÆC. 10 min in 0.2% Triton X-100 and blocked in 1% bovine serum albumin (BSA). After a series of washes, the coverslips were Protein preparation for two-dimensional gel electrophoresis (2-DE). incubated with mouse anti-human ‚-catenin (1:100 dilution, Santa The cells were cultured in 75 mm tissue culture dishes until 90% Cruz Biotechnology, Santa Cruz, CA, USA) or E-cadherin (1:100 confluent and harvested with 0.25% trypsin. The cell pellets were dilution, Santa Cruz Biotechnology) for 1 h, and then washed in rinsed with cold PBS before lysis in 500 Ìl buffer (8 M urea, 4% PBS. This was followed by a 1 h incubation with a FITC-conjugated 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate anti-mouse IgG secondary antibody (1:100 dilution, Santa Cruz (CHAPS), 65 mM DL-dithiothreitol (DTT), 40 mM Tris-HCl, Biotechnology). Rhodamine phalloidin used to detect filamentous 4 mM EDTA, 0.2% Bio-lyte ampholytes, protease inhibitors and actin (F-actin) was obtained from Molecular Probes (Eugene, endonuclease). The cell pellets were frozen and thawed three Oregon, USA). The negative control was performed without times with liquid nitrogen, sheared with a 27 gauge needle to adding the primary antibody. The slides were sealed in glycerin and remove DNA and centrifuged at 25,000 xg for 45 minutes at 4ÆC. observed by fluorescent microscopy (Zeiss, Germany) at a The supernatant was purified and concentrated using a wavelength of 450 nm. ReadyPrep Cleanup kit (Bio-Rad, Hercules, CA, USA). Protein lysates were stored at –80ÆC until recovered for use. The protein Immunoblotting and co-immunoprecipitation. The cells were concentration was determined by using the Bio-Rad RC DC treated with lysis buffer containing 15% Nonidet P40 (NP40), 5 M protein assay (Bio-Rad). NaCl, 2 M Tris base (pH 7.4), 0.5 M EDTA (pH 8.0), 2 mM

328 Wei et al: Vimentin Promotes Prostate Cancer Invasion phenylmethane-sulfonyl fluoride, 10 Ìg/ml leupeptin, 10 Ìg/ml pcDNAVIMs, while the 1E8-H cells were transfected with 3 Ìg aprotinin, 0.5 mM benzamidine, 1.5 mM sodium fluoride, 300 ÌM pcDNA3. 1 (as a blank control) or 3 Ìg pcDNAVIMas according to sodium vanadate and 10 mM sodium pyrophosphate. Protein the protocol supplied with the LipofectAMINE transfection concentrations were determined using a Bio-Rad protein assay. reagent (Invitrogen Life Technologies, Inc, CA, USA). Briefly, Equal amounts of protein denatured in SDS sample buffer (2% 2x105 cells were plated in 6-well dishes and incubated with the SDS, 62.5 mM Tris·base (pH 6.8), 10% glycerol, 5% appropriate DNA and LipofectAMINE in serum-free media for 5 ‚-mercaptoethanol and 0.005% bromphenol blue) were loaded h, and then equal volumes of media containing 20% FBS were onto 10% SDS-PAGE and the gels were transferred onto added. After 24 h, the media were replaced with media containing nitrocellulose membranes. The membranes were blocked overnight 1 mg/ml Geneticin (G418). The surviving colonies were selected in tris buffered saline (TBS) containing 5% (w/v) powder skim after 2 weeks and then maintained in 300 Ìg/ml G418. The positive milk, and after a series of washes, blots were stained with the cell clones were chosen and amplified, and neo gene expression was recommended dilution of primary antibodies against vimentin detected by RT-PCR using primers 5'-AGAGGCTATTCTGCT (1:500), anti-C-src (Santa Cruz Biotechnology), anti-‚-catenin, anti- ATGAC-3' (sense) and 5'-GCTTCAGTGACAAC GTCGAG-3' E-cadherin, anti-phospho-C-src (Santa Cruz Biotechnology), anti- (antisense). The blank pcDNA3.1 vector-transfected clones were phospho-‚-catenin (abcam, Cambridge, UK) or ‚-actin (C2; Santa established by ring cloning. Changes in vimentin levels were Cruz Biotechnology). After further washing, the blots were assayed by Western blot. incubated with a 1:1000 dilution of goat-anti-mouse IgG antibody conjugated to horseradish peroxidase (Santa Cruz Biotechnology) Motility and invasion assays. To evaluate the effect of vimentin on and then developed using the enhanced chemiluminescence motility and invasion activity in the 1E8-H and 2B4-L cells, assays Western blot detection kit (Pierce Biotechnology, Inc, Rockford, were performed in vitro by testing the ability of the cells to invade IL, USA). The co-immunoprecipitation was carried out with a from the top well to the lower chamber. Cell motility was assessed ProFoundì co-immnoprecipitation kit (Pierce) according to the using 24 transwell units with polycarbonate filters (Costar Corp, manufacturer's instructions. Cambridge, MA, USA) containing 8 Ìm pores. The cells were plated at 1.0x105 cells/well in 0.5 ml serum-free media. The outer Expression vector construction. The total RNA was isolated from the chambers were filled with 0.5 ml media containing 10% FBS. After 1E8-H cells using TRIZolì Reagent (Gibco BRL, Gaithersburg, 24 h, cells migrating to the undersurface of the filters were MD, USA) according to the manufacturer's instructions. The RNA counted. The top surface of the membrane was gently scrubbed (2 Ìg) was used for cDNA synthesis by reverse transcription. The with a cotton swab, and cells on the undersurface were fixed in RNA samples were incubated at 70ÆC for 5 min with 0.5 Ìg oligo methanol and stained with haematoxylin before undergoing a series deoxythymidine primers in a final volume of 10 Ìl and then at 37ÆC of washes. The same five microscopic fields were used to count the for 60 min in a 25 Ìl reaction buffer containing 125 mmol/l number of cells passing to the undersurface of each filter. For the deoxynucleotide triphosphate, 200 U Muloney murine leukemia invasion assays, the insert chambers were replaced with Matrigel virus reverse transcriptase (RT) and Muloney murine leukemia (Sigma) reconstituted basement membrane layer. virus RT buffer (Promega, Madison, USA). The obtained cDNAs were amplified by using cloning primers as follows: vimentin: 5'- Inhibition of Src. PP2, a specific inhibitor of src family kinases was CGGGATCCCGCCCTCGTTCGCCTCTTCTC-3' (sense) and 5'- used to treat the 1E8-H and 2B4-LVIMs cells. The PP2 was CGGAATTCCGATATCGCCTGCCACTGAGTG- 3' (antisense); applied to 90% confluent 1E8-H cells at doses of 1x10–5, 1x10–4 antisense-vimentin: 5'-CTGCTCGAGCGCCCTC GTTCGCCT and 1x10–3 M for 24, 48 and 72 hrs. Subsequently both 1E8-H and CTTCTC- 3' (sense) and 5'-CGGAATTCCGATATC GCCTGCC 2B4-LVIMs cells were treated with 1x10–4 M PP2 for 48 h. ACTGAGTG-3' (antisense). The enzyme sites for BamHπ, EcoRπ and Xhoπ are underlined. The PCR profile was 95ÆC for 1 minute, Statistical analysis. All the experiments were repeated at least three 94ÆC for 30 sec, 56ÆC for 1 minute and 72ÆC for 1 minute for 30 times. The Student's t-test was used to evaluate differences cycles, followed by extension for 7 minutes at 72ÆC. The amplified between experimental and control groups. Significance was defined products with different restriction sites, including BamHI and at the level of p<0.05. EcoRI or EcoRI and XhoI, were purified using a PCR kit (New England Biolabs, Herts, UK) and ligated into the pcDNA3.1 vector Results (Promega) by following the instruction manual. The ligation product was transformed into E. coli DH5· competent cells. The recombinant plasmids were screened by digestion and sequenced Proteomic analysis of 1E8-H and 2B4-L cells. The average to confirm the insert sequences of vimentin. The pcDNA3.1 number of protein spots detected were 873±73 in the 1E8- plasmid with full-length vimentin was designated as pcDNAVIMs, H cells and 857±67 in the 2B4-L cells. The scatter plots while the pcDNA3.1 plasmid with full-length antisense-vimentin showed good correlation between these two cell types with a was designated as pcDNAVIMas. correlation coefficient of 0.86, implying a similarity in the expression of most proteins, with differences in only a small DNA transfection and clone selection. To determine the effects of number. A representative gel image was selected from the vimentin on invasion and metastasis in prostate cancer cells, one clone of 2B4-L cells was selected for transfection with the sense- 1E8-H cells and the 2B4-L cells (Figure 1A). In total, 14 vimentin vector and one clone of 1E8-H cells was picked up for spots (including both housekeeping and differentially transfection with the antisense-vimentin vector. The 2B4-L cells expressed proteins) were identified from preparative “blue were transfected with 3 Ìg pcDNA3 .1 (as a blank control) or 3 Ìg silver” stained gels using MALDI-TOF-MS analysis. The

329 ANTICANCER RESEARCH 28: 327-334 (2008)

Figure 2. Cross-expression of appropriate proteins involved in invasion. 1E8-H cell lysates were immunoprecipitated with vimentin (VIM) antibodies (A) or C-src antibodies (B). C-src, ‚-catenin (‚-cat) and E- cadherin (E-cad) were detected in the anti-vimentin immunoprecipitate. Vimentin, ‚-catenin and E-cadherin were detected in the anti-C-src immunoprecipitate. Preimmune mouse IgG was used as a control (Cont).

identity of spot 2307R. To confirm the proteomic result above, Western blot and 2D Western blot analyses were used to enhance sensitivity and to effectively detect protein isoforms generated through post-translational modification. Increased expression of vimentin was detected in the 1E8-H cells compared to that in the 2B4-L cells (Figure 1C and D).

Expression of vimentin in relation to C-src, ‚-catenin and E- cadherin proteins. As shown in Figure 2, the blots immunoprecipitated with vimentin antibody were stained with C-src, ‚-catenin or E-cadherin antibodies and showed appropriate positive staining (Figure 2A). Meanwhile, the blots immunoprecipitated with C-src antibody were stained with vimentin, ‚-catenin or E-cadherin antibodies and showed appropriate positive staining (Figure 2B). These cross-reactions verified that vimentin was associated with C- src, ‚-catenin and E-cadherin.

Vimentin transfection effect on invasion and motile activity in Figure 1. Proteomic analysis of 1E8-H and 2B4-L cells. (A) A gel image was selected from 1E8-H and 2B4-L cells; the intensity of spot 2307R in tumor cells. The invasion of the 1E8-HVIMas cells was 1E8-H cells was significantly stronger compared to that in 2B4-L cells. (B) much lower than that of the 1E8-H cells (Figure 3A and 3C, The MALDI-TOF mass spectrum of peptides derived from 2307R. 2D p<0.01), whereas the invasion ability of the 2B4-LVIMs Western blot (C) and Western blot analysis (D) confirmed that a high cells was noticeably higher than that of the 2B4-L cells expression of vimentin was detected in 1E8-H cells compared to that in (Figure 3A and 3C, p<0.01). Figure 3B shows the motility 2B4-L cells. ‚-Actin was used to demonstrate equal loading. of the tumor cells which was represented as the mean numbers of cells counted on the undersurface of each filter per observed field (Figure 3D). The results of the motility assay were consistent with the invasion assays in the same intensity of spot 2307R in the 1E8-H cells was significantly cell lines (2B4-L vs. 2B4-LVIMs and 1E8-H vs. 1E8- stronger compared to that in the 2B4-L cells (Figure 1A) HVIMas p<0.01, respectively). and the MALDI-TOF mass spectrum of peptides derived from 2307R is shown in Figure 1B. A Mascot search using The effects of vimentin transfection on the expression of proteins the peptide mass fingerprint (PMF) data indicated that 19 involved in invasion. Western blot assays showed significantly peptides were matched with peptides from vimentin, giving decreased vimentin expression in the antisense-vimentin- sequence coverage of 38% and a summary score of 171. transfected 1E8-H cells, designated as 1E8-HVIMas, These results strongly suggested that vimentin was the compared to the non-transfected or pcDNA3.1 vector-

330 Wei et al: Vimentin Promotes Prostate Cancer Invasion

Figure 3. The effects of vimentin transfection on invasion and motility. 1.0x105 2B4-L/2B4-LVIMs or 1E8-H/1E8-HVIMas cells were seeded per filter and incubated for 24 h. Both invasion (A) and motility (B) assays were carried out on three filters for each cell line and cells were counted from the same five fields of each filter under 200x magnification. The results of the invasion assay are expressed as the number of cells passing through Matrigel-coated filters (C). The results of the motility assay are expressed as the number of cells counted per field passing through the filter (D). Values and error bars shown in this graph represent the averages and standard deviations, respectively, of three independent experiments. transfected cells. Meanwhile, an induced vimentin level was cell contact in the 1E8-HVIMas and 2B4-L cells, forming detected in the sense-vimentin-transfected 2B4-L cells, classic zipper-like structures (Figure 5A and B, arrows), named 2B4-LVIMs, which was much higher than in the non- whereas in the 1E8-H and 2B4-LVIMs cells, E-cadherin and transfected or blank vector-transfected cells (Figure 4A). ‚-catenin were either not visible or presented a weak The expression levels of the relevant proteins, including staining at boundaries between opposing cells, indicating C-src, phospho-C-src, ‚-catenin, phospho-‚-catenin and E- that the 1E8-H and 2B4-LVIMs cells were easy to detach. cadherin were tested, in both the 1E8-H and 2B4-L cells and in the 1E8-HVIMas and 2B4-LVIMs cells. As shown in Inhibition of C-src and the expression level of corresponding Figure 4B, the 1E8-HVIMas cells displayed decreased proteins involved in invasion. As shown in Figure 6A, B and expression of C-src, phospho-C-src and phospho-‚-catenin C, PP2 caused dose-dependent and time-dependent compared to non-transfected 1E8-H cells, while E-cadherin inhibition of the expression of C-src in the 1E8-H cells. The protein levels increased and the expression of ‚-catenin lowest effective dose of PP2, 1x10–4 M, and an appropriate protein showed no change. Induced expression of C-src, time point, 48 h, were used to test its effect on ‚-catenin phospho-C-src and phospho-‚-catenin and a reduced E- and E-cadherin expression in the 1E8-H and 2B4-LVIMs cadherin level were detected in the transfected 2B4-LVIMs cells. As shown in Figure 6D, a reduced expression of cells, compared with the non-transfected 2B4-L cells. The phospho-‚-catenin and a remarkably increased E-cadherin ‚-catenin level remained stable after vimentin transfection expression were detected in the PP2-treated cells compared (Figure 4B). Immunofluorescence assays showed that strong to the non-treated cells. The ‚-catenin level remained stable E-cadherin and ‚-catenin staining existed at points of cell- after PP2 administration (Figure 6D).

331 ANTICANCER RESEARCH 28: 327-334 (2008)

Figure 4. Western blot analysis of cells following transfection with pcDNAVIMas or with pcDNAVIMs. 1E8-H cells were transfected with 3 Ìg pcDNAVIMas and 2B4-L cells were transfected with 3 Ìg pcDNAVIMs. Cells transfected with blank pcDNA3.1 vector were used as controls. (A) After transfection, protein extracts from the cells were subjected to Western blot analysis with an antibody against vimentin (VIM). (B) After transfection, protein extracts from the cells were subjected to Western blot analysis with antibodies to C-src, phospho-C-src (P-C-src), ‚-catenin (‚-cat), phospho-‚-catenin (P-‚-cat) and E-cadherin (E-cad). ‚-Actin was used to demonstrate equal loading.

Discussion

Significant expression of vimentin was detected in 1E8-H cells (high metastatic potential) compared with 2B4-L cells Figure 5. E-cadherin and ‚-catenin staining by immunofluorescence (low metastatic potential). Western blot and 2D Western assay. (A) and (B) cells were first stained with rhodamine phalloidin and blot analyses further verified this result in the present study. then stained with E-cadherin or ‚-catenin antibodies, followed by the Some investigators have also demonstrated that vimentin is appropriate FITC-conjugated secondary antibody. Positive staining was highly expressed in invasive cancer cells, as determined by visualized by confocal microscopy. E-cadherin and ‚-catenin ‘zippers' were clearly seen at cell boundaries (arrows), where actin and E-cadherin or proteomic analysis or microarrays (28-31). Previous studies actin and ‚-catenin overlapped. Scale bar=20 Ìm. have demonstrated that vimentin is expressed in epithelial cells that undergo tumor invasion (32, 33). Lang et al. have also demonstrated that vimentin is expressed by motile prostate cell lines and positively correlates with poorly determined by Matrigel invasion assays. This was consistent differentiated carcinomas and bone metastases (34). The with results from overexpression of vimentin in human important contribution of vimentin to the invasive breast cancer cells resulting in phenotypic interconversion phenotype of prostate cancer cells was demonstrated by the and invasive behavior conversion (33). present transfection results. The overexpression of vimentin The immunoblot and immunofluorescence analysis in the increased invasiveness in the stable, vimentin-transfected present study demonstrated that loss of E-cadherin activation 2B4-LVIMs cells, while in contrast, the decreased vimentin correlated inversely with increasing vimentin expression in expression in antisense-vimentin-transfected 1E8-HVIMas the stable, vimentin-transfected 2B4-LVIMs cells. Therefore, cells led to a significant decrease in invasiveness, as both vimentin activity and E-cadherin-mediated cell adhesion

332 Wei et al: Vimentin Promotes Prostate Cancer Invasion

linked and functionally coupled in the human prostate cancer cells, indicating the possibility that vimentin affected the E- cadherin/‚-catenin complex via C-src. Early work suggested that the E-cadherin/‚-catenin complex was further stabilized by serine phosphorylation of E-cadherin at residues 684, 686 and 692. Structural analysis revealed that these residues only interact with ‚-catenin when phosphorylated (35-37). Phosphorylation of ‚-catenin tyrosine residue 654 by C-src causes an approximately six-fold reduction in the affinity of ‚-catenin for E-cadherin (38). In our study, phospho-‚-catenin expression increased in the 2B4- LVIMs cells and decreased in the 1E8-HVIMas cells, while the expression of ‚-catenin protein in the transfected cells was the same as in the non-transfected cells. Furthermore, src inhibition reduced phospho-‚-catenin expression and induced E-cadherin expression in highly metastatic cells, and could probably inhibit detachment of these cells. There is some evidence that phosphorylation of two of the critical tyrosines of ‚-catenin that disrupt the core E-cadherin complex, 142 and 654, has the potential to regulate the exchange of ‚-catenin from its association with E-cadherin to its role in nuclear localization and transcription (39). The phosphorylation of tyrosine 654 increased the ability of ‚- catenin to bind the basal transcription factor TATA-binding protein and enhanced transcriptional activity of ‚-catenin. Thus, ‚-catenin levels remained stable following vimentin transfection or PP2 administration in the prostate cancer cells. In conclusion, vimentin can promote tumor cell invasiveness by regulating the E-cadherin/‚-catenin complex via C-src in prostate cancer cells. Therefore, vimentin might be a target for therapeutic agents aiming to block the metastasis of prostate cancer, either through antisense- vimentin transfection or through the use of specific inhibitors, such as C-src inhibitors.

Acknowledgements

Grants support: National Science Foundation of China (30571950; 36072227) and the “973” Program of China (No. 2002CB513100).

References

Figure 6. The effects of PP2 on expression of C-src and related proteins. 1 Anazawa Y, Nakagawa H, Furihara M et al: PCOTH, a novel 1E8-H cells treated with PP2 at doses of 1.0x10–5 M, 1.0x10–4 M and gene overexpressed in prostate cancers, promotes prostate 1.0x10–3 M, for 24 h (A) 48 h (B) or 72 h (C). 1E8-H and 2B4-LVIMs cancer cell growth through phosphorylation of oncoprotein cells treated with 1.0x10–4 M PP2 for 48 h (D). Protein extracts from the TAF-Ibeta/SET. Cancer Res 65(11): 4578-4586, 2005. cells were subjected to Western blot analysis with antibodies to C-src, ‚- 2 Simpson RJ and Dorow DS: Cancer proteomics: from signaling catenin (‚-cat), phospho-‚-catenin (P-‚-cat) and E-cadherin (E-cad). ‚- networks to tumor markers. Trends Biotechnol 19(Suppl 10): Actin was used to demonstrate equal loading. S40-48, 2001. 3 Gilles C and Thompson E: The epithelial to mesenchymal transition and metastatic progression in carcinoma. Breast J 2(2): 83-96, 1996. appear to be hallmarks of EMT-associated events implicated 4 Jechlinger M, Grunert S, Tamir IH et al: Expression profiling in cell migration and invasion. Furthermore, the data also of epithelial plasticity in tumor progression. Oncogene 22(46): demonstrated that vimentin and C-src kinase were physically 7155-7169, 2003

333 ANTICANCER RESEARCH 28: 327-334 (2008)

5 Thiery JP: Epithelial-mesenchymal transitions in tumour 24 Liu Y, Zheng J, Fang W et al: Isolation and characterization of progression. Nat Rev Cancer 2(6): 442-454, 2002. human prostate cancer cell subclones with different metastatic 6 Boyer B, Valles AM and Edme N: Induction and regulation of potential. [Zhonghua Bing Li Xue Za Zhi] 28(5): 361-364, 1999. epithelial-mesenchymal transitions. Biochem Pharmacol 60(8): 25 Gorg A, Obermaier C, Boguth G et al: The current state of two- 1091-1099, 2000. dimensional electrophoresis with immobilized pH gradients. 7 Takeichi M: Morphogenetic roles of classic cadherins. Curr Electrophoresis 21(6): 1037-1053, 2000. Opin Cell Biol 7(5): 619-627, 1995. 26 Candiano G, Bruschi M, Musante L et al: Blue silver: a very 8 Kemler R: From cadherins to catenins: cytoplasmic protein sensitive colloidal Coomassie G-250 staining for proteome interactions and regulation of cell adhesion. Trends Genet 9(9): analysis. Electrophoresis 25(9): 1327-1333, 2004 317-321, 1993. 27 Gerner C, Frohwein U, Gotzmann J et al: The Fas-induced 9 Tepass U, Truong K, Godt D, Ikura M and Peifer M: Cadherins apoptosis analyzed by high throughput proteome analysis. J Biol in embryonic and neural morphogenesis. Nat Rev Mol Cell Biol Chem 275(50): 39018-39026, 2000. 1(2): 91-100, 2000. 28 Liu X, Wu Y, Zehner ZE, Jackson-Cook C and Ware JL: 10 Beavon IR: The E-cadherin-catenin complex in tumour Proteomic analysis of the tumorigenic human prostate cell line metastasis: structure, function and regulation. Eur J Cancer M12 after microcell-mediated transfer of chromosome 19 36(13 Spec No): 1607-1620, 2000. demonstrates reduction of vimentin. Electrophoresis 24(19-20): 11 Christofori G and Semb H: The role of the cell-adhesion 3445-3453, 2003. molecule E-cadherin as a tumour-suppressor gene. Trends 29 Singh S, Sadacharan S, Su S, Belldegrun A, Persad S and Singh Biochem Sci 24(2): 73-76, 1999. G: Overexpression of vimentin: role in the invasive phenotype 12 Birchmeier W and Behrens J: Cadherin expression in carcinomas: in an androgen-independent model of prostate cancer. Cancer role in the formation of cell junctions and the prevention of Res 63(9): 2306-2311, 2003. invasiveness. Biochim Biophys Acta 1198(1): 11-26, 1994. 30 Vasko V, Espinosa AV, Scouten W et al: Gene expression and 13 Mareel M, Bracke M and Van Roy F: Invasion promoter versus functional evidence of epithelial-to-mesenchymal transition in invasion suppressor molecules: the paradigm of E-cadherin. papillary thyroid carcinoma invasion. Proc Natl Acad Sci USA Mol Biol Rep 19(1): 45-67, 1994. 104(8): 2803-2808, 2007. 14 Takeichi M: Cadherins in cancer: implications for invasion and 31 Craven RA, Stanley AJ, Hanrahan S et al: Proteomic analysis metastasis. Curr Opin Cell Biol 5(5): 806-811, 1993. of primary cell lines identifies protein changes present in renal 15 Shino Y, Watanabe A, Yamada Y et al: Clinicopathologic cell carcinoma. Proteomics 6(9): 2853-2864, 2006. evaluation of immunohistochemical E-cadherin expression in 32 Gilles C, Polette M, Zahm JM et al: Vimentin contributes to human gastric carcinomas. Cancer 76(11): 2193-2201, 1995. human mammary epithelial cell migration. J Cell Sci 112(Pt 24): 16 Tamura S, Shiozaki H, Miyata M et al: Decreased E-cadherin 4615-4625, 1999. expression is associated with haematogenous recurrence and 33 Hendrix MJ, Seftor EA, Seftor RE and Trevor KT: Experimental poor prognosis in patients with squamous cell carcinoma of the co-expression of vimentin and keratin intermediate filaments in oesophagus. Br J Surg 83(11): 1608-1614, 1996. human breast cancer cells results in phenotypic interconversion and 17 Hunt NC, Douglas-Jones AG, Jasani B, Morgan JM and increased invasive behavior. Am J Pathol 150(2): 483-495, 1997. Pignatelli M: Loss of E-cadherin expression associated with 34 Lang SH, Hyde C, Reid IN et al: Enhanced expression of lymph node metastases in small breast carcinomas. Virchows vimentin in motile prostate cell lines and in poorly Arch 430(4): 285-289, 1997. differentiated and metastatic prostate carcinoma. Prostate 18 Perl AK, Wilgenbus P, Dahl U, Semb H and Christofori G: A 52(4): 253-263, 2002. causal role for E-cadherin in the transition from adenoma to 35 Stappert J and Kemler R: A short core region of E-cadherin is carcinoma. Nature 392(6672): 190-193, 1998. essential for catenin binding and is highly phosphorylated. Cell 19 Hirohashi S: Inactivation of the E-cadherin-mediated cell Adhes Commun 2(4): 319-327, 1994. adhesion system in human cancers. Am J Pathol 153(2): 333- 36 Lickert H, Bauer A, Kemler R and Stappert J: Casein kinase II 339, 1998. phosphorylation of E-cadherin increases E-cadherin/beta- 20 Irby RB, and Yeatman TJ: Role of Src expression and catenin interaction and strengthens cell-cell adhesion. J Biol activation in human cancer. Oncogene 19(49): 5636-5642, 2000. Chem 275(7): 5090-5095, 2000. 21 Coluccia AM, Benati D, Dekhil H et al: SKI-606 decreases 37 Simcha I, Kirkpatrick C, Sadot E et al: Cadherin sequences that growth and motility of colorectal cancer cells by preventing inhibit beta-catenin signaling: a study in yeast and mammalian pp60(c-Src)-dependent tyrosine phosphorylation of beta-catenin cells. Mol Biol Cell 12(4): 1177-1188, 2001. and its nuclear signaling. Cancer Res 66(4): 2279-2286, 2006. 38 Roura S, Miravet S, Piedra J, Garcia de Herreros A and Dunach 22 Tsukita S, Oishi K, Akiyama T, Yamanashi Y, Yamamoto T M: Regulation of E-cadherin/Catenin association by tyrosine and Tsukita S: Specific proto-oncogenic tyrosine kinases of src phosphorylation. J Biol Chem 274(51): 36734-36740, 1999. family are enriched in cell-to-cell adherens junctions where the 39 Piedra J, Martinez D, Castano J, Miravet S, Dunach M and de level of tyrosine phosphorylation is elevated. J Cell Biol 113(4): Herreros AG: Regulation of beta-catenin structure and activity by 867-879, 1991. tyrosine phosphorylation. J Biol Chem 276(23): 20436-20443, 2001. 23 Behrens J, Vakaet L, Friis R et al: Loss of epithelial differentiation and gain of invasiveness correlates with tyrosine phosphorylation of the E-cadherin/beta-catenin complex in cells Received August 12, 2007 transformed with a temperature-sensitive v-SRC gene. J Cell Revised November 15, 2007 Biol 120(3): 757-766, 1993. Accepted December 11, 2007

334