Ovarian cancer has one of the highest mortalities in malignancies in women, but little is known of its tumour progression properties and there is still no effective molecule that can monitor its growth or therapeutic responses. MSLN (mesothelin), a secreted protein that is overexpressed in ovarian cancer tissues with a poor clinical outcome, has been previously identified to activate PI3K (phosphoinositide 3-kinase)/Akt signalling and inhibit paclitaxel-induced apoptosis. The present study investigates the correlation between MSLN and MMP (matrix metalloproteinase)-7 in the progression of ovarian cancer, and the mechanism of MSLN in enhancing ovarian cancer invasion. The expression of MSLN correlated well with MMP-7 expression in human ovarian cancer tissues. Overexpressing MSLN or ovarian cancer cells treated with MSLN showed enhanced migration and invasion of cancer cells through the induction of MMP-7. MSLN regulated the expression of MMP-7 through the ERK (extracellular-signal-regulated kinase) 1/2, Akt and JNK (c-Jun N-terminal kinase) pathways. The expression of MMP-7 and the migrating ability of MSLN-treated ovarian cancer cells were suppressed by ERK1/2- or JNK-specific inhibitors, or a decoy AP-1 (activator protein 1) oligonucleotide in in vitro experiments, whereas in vivo animal experiments also demonstrated that mice treated with MAPK (mitogen-activated protein kinase)/ERK- or JNK-specific inhibitors could decrease intratumour MMP-7 expression, delay tumour growth and extend the survival of the mice. In conclusion, MSLN enhances ovarian cancer invasion by MMP-7 expression through the MAPK/ERK and JNK signal transduction pathways. Blocking the MSLN-related pathway could be a potential strategy for inhibiting the growth of ovarian cancer.

INTRODUCTION

The invasion and metastasis of cancer cells are complex and complicated processes that involve several classes of proteins, including calcium-dependent cadherins, integrins, extracellular proteases, angiogenetic factors and lymphangiogenic factors. The MMPs (matrix metalloproteinases) are conventionally divided into collagenases, gelatinases, stromelysins and matrilysins on the basis of their specificity for ECM (extracellular matrix) components [1]. Previous studies have demonstrated that MMPs are a growing group of proteolytic enzymes involved in ECM remodelling. Moreover, MMP expression and activity are increased in almost every type of human cancer. Thus MMPs are associated with advanced tumour stage and poor survival of cancer patients [2,3].

MMP-7 has a minimal domain organization and is one of the few MMPs that is overexpressed by carcinoma cells rather than stromal cells [4]. It plays roles not only in tumour progression, but also in early cancer development [5]. Overexpression of the Mmp7 gene in transgenic mice has led to enhanced tumorigenesis in a breast cancer model [6] and intestinal tumorigenesis [7]. MMP-7 is also implicated in cancer invasion/metastasis and in initiation or growth in gastrointestinal and other cancers [8]. However, its role in tumour invasion and progression in ovarian cancer has not been investigated.

MSLN (mesothelin), originally found in mesothelial cells, is a secreted protein anchored at the cell membrane by a glycosylphosphatidylinositol linkage [9]. It has been identified as a potential new tumour antigen for mesotheliomas and ovarian cancers [10,11]. MSLN has also been found to be overexpressed in cancers of the pancreas [12], stomach [13] and endometrium [14]. Our previous study [15] has shown that MSLN can inhibit paclitaxel-induced apoptosis through a PI3K (phosphoinositide 3-kinase)/Akt-dependent pathway, suggesting that MSLN can influence chemotherapeutic sensitivity [15]. Some papers have also reported that MSLN can be a potential tumour marker in monitoring cancer progression [16,17]. However, the biological effects of the MSLN influence on tumour progression or invasion remains unclear.

The WF-0 cell is a peritoneal-based tumour model created from C57BL/6 mice by introducing various oncogenes [18]. The continuous isolation and passage of early-stage tumour cells (WF-0) from the ascites of the tumour-challenged C57BL/6 mice resulted in an aggressive tumour cell line named WF-3. One of the genes that is highly expressed in WF-3, but not in WF-0, cells is MSLN. WF-3 cells can be an ascitogenic malignant tumour model which demonstrates morphological features of intraperitoneal tumorigenesis. WF-3 cells also demonstrate relatively highly proliferative and migrating rates compared with WF-0 tumour cells. Furthermore, the WF-3 tumour cells can express high levels of MSLN, commonly observed in intraperitoneal tumors such as malignant mesotheliomas and ovarian cancers. Therefore the WF-3 tumour model serves as an excellent pre-clinical tumour model for ovarian cancer to study clinically relevant issues, such as screening and evaluating the effectiveness of therapeutic intervention by monitoring MSLN levels.

The present study aimed to clarify the molecular mechanism of the invasive and metastatic abilities of ovarian cancer induced by MSLN. The hypothesis was that MMP-7 acts as a mediator of MSLN-induced tumour invasiveness and metastasis of ovarian cancer.

MATERIALS AND METHODS

Cell culture and transfection

The WF-0, MSLN-transfectants (WF-0/M1, WF-0/M2 and WF-0/M3) and WF-3 cancer cell lines were generated and cultured as described previously [15,18]. The OVCAR-3 ovarian cancer cell line was obtained from the A.T.C.C. (Manassas, VA, U.S.A.).

siRNA (short interfering RNA) for MSLN was designed and synthesized using the Silencer™ siRNA construction kit (Ambion). The sequences of the double-stranded RNA used to block MSLN and MMP7 expression were 5′-AGAAGAUGAGGCUCUCGUCUAUCUC-3′ (sense) and 5′-GAGAUAGACGAGAGCCUCAUCUUCU-3′ (antisense) [15], and 5′-UAGUCUAGUGGACAACCUCUU-3′ (sense) and 5′-AAGAGGUUGUCCACUAGACUA-3′ (antisense) respectively. Briefly, OVCAR-3 cells were transfected with 50 pM MSLN siRNA using 12 μg of Lipofectamine™ 2000 transfection reagent in a total of 2 ml of serum-free RPMI 1640 medium. After incubation at 37°C, 5% CO2 for 6 h, 2 ml of culture medium was added.

In the MMP-7 blockage experiments, WF-0 cells were transfected with 50 pM MMP7 siRNA using 12 μg of Lipofectamine™ 2000 transfection reagent in a total of 2 ml of serum-free RPMI 1640 medium. At 12 h later, recombinant MSLN was added to the medium. Cell invasion and migration assays were performed.

RNA isolation, microarray and data analysis

Total RNA from the parental WF-0 and various MSLN-overexpressing WF-0 cell lines was isolated using the TRIzol® RNA isolation kit (Invitrogen). Reverse transcription of isolated RNA was performed in a final reaction volume of 20 μl containing 5 μg of total RNA in MMLV (Moloney murine leukaemia virus) RT (reverse transcriptase) buffer (Promega), consisting of 10 mM DTT (dithiothreitol), all four dNTPs (each at 2.5 mM), 1 μg of random hexamer and 200 units of MMLV RT. The reaction mixture was incubated at 37°C for 2 h and the reaction was terminated by heating at 70°C for 10 min.

To perform the microarray experiments, RNA isolation and the RT reaction from parental WF-0 cells and various MSLN-transfected WF-0 cell lines were performed as described above. Alterations in gene expression were evaluated by reverse transcription of poly(A)+ RNAs in the presence of Cy3 (indocarbocyanine) or Cy5 (indodicarbocyanine) fluorescent labelling dyes followed by hybridization to a mouse GEM2 microarray chip (UniGene 1). Gene subsets were selected for further study on the basis of differential Cy3/Cy5 expression ratios that were ≥2.0 in response to antibody cross-linking treatment. Differential expression of representative selected genes was confirmed by RT–PCR and/or Northern blot hybridization.

Patients and specimens

Eighty patients with advanced-staged ovarian carcinoma who underwent surgery and adjuvant chemotherapy were enrolled in the study. The Institutional Review Board approved the experimental protocols and all patients provided written informed consent. Ovarian cancer specimens were obtained intra-operatively and were frozen immediately at −70°C until analysis.

Generation of murine MSLN protein

Recombinant murine MSLN protein was purchased from Abnova, as described previously [15].

RT–PCR

RT–PCR was used to assess the target gene expression in different experiments. Briefly, 1 μl of the reaction mixture was amplified by PCR using the following primers: human MSLN, 5′-CAAGAAGTGGGAGCTGGAAG-3′ (sense) and 5′-GTCTCCAGGGACGTCACATT-3′ (antisense); human MMP7, 5′-GAGTGCCAGATGTTGCAGAA-3′ (sense) and 5′-AAATGCAGGGGGATCTCTTT-3′ (antisense); mouse MSLN, 5′-TTGTGCCCACTTCTTCTCCCTCA-3′ (sense) and 5′-CTCATCCAACACTGCTACCAAGC-3′ (antisense); MMP2, 5′-ACCAGAACACCATCGAGACC-3′ (sense) and 5′-CCATCAGCGTTCCCATACTT-3′ (antisense); MMP7 5′-CCCGGTACTGTGATGTACCC-3′ (sense) and 5′-AATGGAGGACCCAGTGAGTG-3′ (antisense); MMP9, 5′-CGTCGTGATCCCCACTTACT-3′ (sense) and 5′-AGAGTACTGCTTGCCCAGGA-3′ (antisense); MMP10, 5′-AAGTTCCTCGGGTTGGAGAT-3′ (sense) and 5′-GGGTGCAAGTGTCCATTTCT-3′ (antisense); MMP13, 5′-AGTTGACAGGCTCCGAGAAA-3′ (sense) and 5′-TCCTTGGAGTGATCCAGACC-3′ (antisense); and GAPDH (glyceraldehyde-3-phosphate dehydrogenase), 5′-ACCCAGAAGACTGTGGATGG-3′ (sense) and 5′-TGCTGTAGCCAAATTCGTTG3′ (antisense).

Briefly, total RNAs from the ovarian normal tissues, cancerous tissues, cancer cell lines WF-0, WF-3 or ectopic MSLN-expressing cells (WF-0 MSLN-tranfectants) were isolated using the TRIzol® RNA isolation kit following the manufacturer's instructions (Invitrogen). Reverse-transcribed cDNA products were amplified by PCR with primers specific for the target gene as described above. The amplification products were separated by 1% agarose gel electrophoresis and visualized after staining with ethidium bromide.

Immunoblotting

Frozen ovarian normal and cancerous tissues were homogenized with lysis buffer [137 mM NaCl, 2.7 mM KCl, 1 mM MgCl2, 1 mM CaCl2, 1% Nonidet P40, 10% glycerol, 1 mg/ml BSA, 20 mM Tris/HCl (pH 8.0) and 2 mM orthovanadate] [15]. Tumour cells including WF-0, WF-3 or ectopic MSLN-expressing WF-0 transfectants were seeded and serum-free medium was replaced overnight, followed by treatment with various concentrations of MSLN for the indicated time. The cells were lysed in the immunoprecipitation assay buffer and analysed as described previously [15]. The protein extracts were quantified using the BCA (bicinchoninic acid) protein assay kit (Pierce). Then 50 μg of each cell lysate was resolved by SDS/PAGE (12% gel), transferred on to a PVDF/nylon membrane (Millipore), and probed with antibodies specific to MMP-7, β-actin, ERK (extracellular-signal-regulated kinase), phospho-ERK, Akt, JNK (c-Jun N-terminal kinase), phospho-JNK, p38, phospho-p38 (Upstate Biotechnology) or phospho-Akt (Ser473) (Chemicon International). The membrane was then probed with either HRP (horseradish peroxidase)-conjugated goat anti-mouse or goat anti-rabbit antibody. The specific bands were visualized by an ECL (enhanced chemiluminescence) Western blotting system (GE Healthcare).

In the signal transduction pathways blockage experiments, WF-0 cells were treated for 30 min with different concentrations of inhibitors, including PD98059, LY294002, SP600125 or SB203580 (Sigma) followed by 10 nM MSLN for 24 h. Cells were extracted, transferred on to membranes, and then probed with the anti-MMP-7 or anti-β-actin antibody.

Immunohistochemistry

Tissue sections for immunostaining were obtained from formalin-fixed paraffin-embedded primary ovarian cancers. Formalin-fixed tissue sections were deparaffinized and incubated with 0.01 M citrate buffer followed by microwave treatment (95°C for 10 min). The sections were immersed in 0.3% H2O2 in absolute methanol and treated with 5% FBS (fetal bovine serum) at room temperature (25±2°C) for 30 min. Overnight incubation at 4°C with the anti-MSLN antibody (clone 5B2, Thermo Scientific) or the anti-MMP-7 antibody (clone ID2, Thermo Scientific) was followed by incubation with biotinylated anti-mouse IgG and biotin/streptavidin peroxidase (Super Sensitive Multi-link HRP Detection System, BioGenex). A DAB (diaminobenzidine) solution was used for colour development, and haematoxylin solution was used for counterstaining. Samples without a primary antibody added were used as a negative control. Intensity of immunohistochemical staining ≥10% of cancerous tissues were grouped as high expression, and <10% of cancerous tissues were assigned to the low expression group.

Real-time PCR

MSLN and MMP7 RNA were reverse-transcribed to cDNA. Real-time PCR was then carried out using the LightCycler Real-Time detection system (Roche Diagnostics) [15]. 5′-CTATTCCTCAACCCAGATGCGT-3′ and 3′-GCACATCAGCCTCGCTCA-5′ were the primers used to detect MSLN, and 5′-TGTATGGGGAACTGCTGACATC-3′ and 3′-AGTTGCAGCATACAGGAAGTTAATC-5′ were used to detect MMP7. Amplification was performed in 50 cycles with 10 s at 95°C, 10 s at 60°C, and 10 s at 72°C. Detection of GAPDH was carried out using the LightCycler h-G6PDH housekeeping gene set (Roche Applied Science) for 50 cycles of 10 s at 95°C, 15 s at 55°C, and 15 s at 72°C. Calibration of MMP7 and MSLN expression by real-time PCR was carried out as described previously [15].

Tumour cell invasion assays

Tumour cell invasion assays were performed using Boyden chambers with filter inserts (pore size, 8 μm) coated with Matrigel™ (40 μg; Collaborative Biomedical, Becton Dickinson) in 24-well dishes (Nucleopore) as described previously [20]. Briefly, 2×105 cells of WF-0, WF-3 or ectopic MSLN-expressing WF-0/M1, WF-0/M2 or WF-0/M3 cells with or without MSLN stimulation were seeded in the upper chamber, whereas the same medium was placed in the lower chamber. The plates were incubated for 24 h. Then the cells were fixed in methanol and stained with 0.05% Crystal Violet in PBS for 1 h at room temperature. Cells on the upper side of the filters were removed by cotton-tipped swabs, and the filters were washed with PBS. The cells on the lower side of the filters were defined as invasive cells and counted at ×200 magnification in ten different fields of each filter.

For MSLN blockage experiments, OVCAR-3 tumour cells were treated with siMSLN and seeded in the upper chamber and incubated. The filters were treated and invasive cells were counted as described above.

For the signal transduction pathways blockage experiments, WF-0 cells were first treated with PD098059, LY294002, SP600125 or SB203580 for 30 min and then treated with PBS or 10 nM MSLN for 24 h. The filters were treated and the invasive cells were counted as described above.

Cell migration assays

Cell migration was determined using wound-healing assays with some modifications [21]. WF-0 cells were seeded into six-well plates overnight. A sterile 200 μl pipette tip was used to scratch the cells to form a wound. The cells were then cultured with different concentrations of MSLN for 24 h. Cells which migrated to the wound were defined as migrating cells and were visualized with an inverted Olympus phase-contrast microscope. Four different areas in each assay were counted to calculate the number of migrating cells to the origin of the wound. The healing rate was quantified by measuring the numbers of migrating cells.

Synthesis of AP-1 (activator protein 1) decoy oligonucleotides

Phosphorothioate double-stranded decoy oligonucleotide (a cis-element decoy) carrying the AP-1 (AP-1 decoy oligonucleotide) consensus sequence was purchased from Invitrogen. The AP-1 decoy oligonucleotide sequence was 5′-TGTCTGACTCATGTC-3′ (sense) and 3′- ACAGACTGAGTACAG-5′ (antisense) [22]. The blank control decoy oligonucleotide sequence was 5′-TTGCCGTACCTGACTTAGCC-3′ (sense) and 3′-AACGGCATGGACTGAATCGG-5′ (antisense) [22].

Luciferase activity assays

To evaluate the transcriptional activity of Mmp-7, a 0.7-kb segment at the 5′-flanking region of the mouse Mmp7 promoter region was used [20]. WF-0 cells (5×105) were transiently transfected with 5 μg of pGL3/MMP-7 or pGL3-basic promoter plasmid mixed with 15 μg of Lipofectamine™ 2000 transfection reagent (3:1). After transfection for 24 h, the cells were serum-starved overnight and then treated with MSLN for a further 12 h. Luciferase and β-galactosidase activities were quantified using the Dual-Luciferase® Reporter Assay System (Promega). Co-transfection with pSV-β-galactosidase and the conducted data were normalized following all transient transfections.

For the signal transduction pathways blockage experiments, the Mmp-7 promoter-transfected WF-0 cells were first treated with the respective signal transduction inhibitor, followed by culturing with MSLN. For the AP-1 decoy oligonucleotide experiments, decoy oligonucleotide (5 μM) and Mmp-7 promoter plasmid were mixed (5 μg) and then transfected into WF-0 cells with TransFast™ (Promega) as described previously [21]. Cells were then incubated with MSLN for a further 12 h. Cell lysates were isolated to determine luciferase activities.

In vivo tumour growth experiments

Female C57BL/6J mice (6–8 weeks old) were purchased and kept in the animal facility of the School of Medicine, National Taiwan University. All animal procedures were performed according to approved protocols and in accordance with recommendations for the proper use and care of laboratory animals. C57BL/6 mice (five per group) were inoculated with 2×105 WF-3 tumour cells subcutaneously into both hind limbs. When the tumour masses were palpable (approximately 2–3 mm in diameter) at day 7 of tumour inoculation, the mice were given an intra- or peri-tumour injection with either DMSO only or a signal transduction pathway-specific inhibitor mixed with DMSO to evaluate whether blocking the signalling pathway of MSLN could inhibit tumour growth [22]. The anti-tumour effect was determined by measuring tumour volume calculated using the following formula:

 
formula

In the intraperitoneal injection model, C57BL/6 mice (five per group) were inoculated with 2×105 WF-3 tumour cells by intraperitoneal injection. At 3 days after WF-3 tumour cell implantation, the mice were injected intraperitoneally with either DMSO only or with a signal transduction pathway-specific inhibitor mixed with DMSO to evaluate whether blocking the signalling pathway of MSLN could inhibit or delay the intraperitoneal tumour growth [22]. The anti-tumour effect was determined by measuring the percentages of tumour-free or live mice.

Statistical analyses

Statistical analyses were done using the Statistical Package of Social Studies 6.0 (SPSS). The results were confirmed by conducting at least three independent experiments in all of the in vitro and in vivo experiments in the present study. One-way ANOVA [23] and the Mann–Whitney U test were used. The event/time distributions for different mice in the tumour-free and survival experiments were compared using log-rank analysis. P<0.05 was considered statistically significant.

RESULTS

WF-0/MSLN-overexpressing tumour cells expressed higher levels of MMP-7

To evaluate the gene expression profiles between the parental WF-0 cell line and various MSLN-transfected WF-0 cell lines, microarray analysis using gene chips was performed. Table 1 summarizes the genes that were expressed more than 2-fold higher in the MSLN-transfected WF-0 tumour cells than those in parental WF-0 tumour cells. Genes including those for emerin (GenBank® accession number AA414844, 4.7-fold), lipocalin 2 (GenBank® accession number AA087193, 4.6-fold), fibroblast growth factor receptor 3 (GenBank® accession number AI425744, 3.8-fold), MAPK (mitogen-activated protein kinase)–activated protein kinase 2 (GenBank® accession number W45833, 3.3-fold), protein kinase Cβ (GenBank® accession number AA289586, 3.1-fold), MMP7 (GenBank® accession number AA689037, 2.9-fold), MSLN (GenBank® accession number AA673869, 2.6-fold), peroxisome-proliferator-activated receptor-binding protein (GenBank® accession number AA415615, 2.6-fold), interleukin 2 (GenBank® accession number AA290106, 2.6-fold) and Src-like adapter protein (GenBank® accession number AA867150, 2.5-fold) were expressed more than 2.5-fold higher in the MSLN-transfected WF-0 tumour cells than in parental WF-0 tumour cells (Supplementary Table S1 at http://www.BiochemJ.org/bj/442/bj4420293add.htm).

Table 1
The representative 18 genes whose expression are elevated at least 2-fold in ectopic MSLN-expressing tumour cells compared with the parental WF-0 tumour cells
Gene name Systematic name Ectopic MSLN-expressing tumour cells/parental WF0 cells 
Emerin AA414844 4.7 
Lipocalin 2 AA087193 4.6 
Fibroblast growth factor receptor 3 AI425744 3.8 
MAPK-activated protein kinase 2 W45833 3.3 
Protein kinase Cβ AA289586 3.1 
MMP-7 AA689037 2.9 
MSLN AA673869 2.6 
Peroxisome-proliferator-activated receptor-binding protein AA415615 2.6 
Interleukin-2 AA290106 2.6 
Src-like adapter protein AA867150 2.6 
Vascular endothelial growth factor C AA726944 2.4 
Epidermal growth factor receptor AI156625 2.2 
Metallothionein 1 AA638765 2.2 
Doublecortin AA271274 2.2 
Thrombospondin 3 AA791128 2.1 
Calsenilin W83313 2.1 
Caveolin 3 AA002425 2.1 
Calbindin 2 AA244813 2.1 
Somatostatin AA051655 2.1 
Gene name Systematic name Ectopic MSLN-expressing tumour cells/parental WF0 cells 
Emerin AA414844 4.7 
Lipocalin 2 AA087193 4.6 
Fibroblast growth factor receptor 3 AI425744 3.8 
MAPK-activated protein kinase 2 W45833 3.3 
Protein kinase Cβ AA289586 3.1 
MMP-7 AA689037 2.9 
MSLN AA673869 2.6 
Peroxisome-proliferator-activated receptor-binding protein AA415615 2.6 
Interleukin-2 AA290106 2.6 
Src-like adapter protein AA867150 2.6 
Vascular endothelial growth factor C AA726944 2.4 
Epidermal growth factor receptor AI156625 2.2 
Metallothionein 1 AA638765 2.2 
Doublecortin AA271274 2.2 
Thrombospondin 3 AA791128 2.1 
Calsenilin W83313 2.1 
Caveolin 3 AA002425 2.1 
Calbindin 2 AA244813 2.1 
Somatostatin AA051655 2.1 

MSLN correlated well with MMP-7 in ovarian cancer, regardless of RNA and protein levels

To evaluate the correlations between MSLN and MMP-7, RT–PCR and immunoblotting analyses were performed. Representative images of RT–PCR and immunoblotting are shown in Figures 1(A) and 1(B). The expression of MSLN and MMP-7 was higher in the cancerous tissues than in normal tissues. Moreover, ovarian cancerous tissues that expressed higher MSLN RNA also expressed higher MMP7 RNA (Figure 1A). Ovarian cancerous tissues that contained higher MSLN protein also contained higher MMP-7 protein (Figure 1B).

MSLN expression correlated with that of MMP-7 in human ovarian cancer

Figure 1
MSLN expression correlated with that of MMP-7 in human ovarian cancer

(A) RT–PCR analyses and (B) Western blot analyses of MSLN and MMP-7 expression from ovarian cancer tissues. Neither MSLN nor MMP-7 was expressed in normal ovarian tissues. The expression level of MSLN correlated well with that of MMP-7 in various ovarian cancer tissues in both analyses. (C) Immunohistochemical examination of ovarian cancer tissues with low expression of MSLN (top left-hand panel) and MMP-7 (top right-hand panel), and high expression of MSLN (bottom left-hand panel) and MMP-7 (bottom right-hand panel). Magnification, ×100. (D) There was a good correlation of mRNA expression between MSLN and MMP7 by real-time PCR (r=0.694, P<0.001, Spearman's correlation). Ct, threshold cycle value.

Figure 1
MSLN expression correlated with that of MMP-7 in human ovarian cancer

(A) RT–PCR analyses and (B) Western blot analyses of MSLN and MMP-7 expression from ovarian cancer tissues. Neither MSLN nor MMP-7 was expressed in normal ovarian tissues. The expression level of MSLN correlated well with that of MMP-7 in various ovarian cancer tissues in both analyses. (C) Immunohistochemical examination of ovarian cancer tissues with low expression of MSLN (top left-hand panel) and MMP-7 (top right-hand panel), and high expression of MSLN (bottom left-hand panel) and MMP-7 (bottom right-hand panel). Magnification, ×100. (D) There was a good correlation of mRNA expression between MSLN and MMP7 by real-time PCR (r=0.694, P<0.001, Spearman's correlation). Ct, threshold cycle value.

To further evaluate the exact cells that secrete MSLN and MMP-7, immunohistochemical staining was performed. Both MSLN and MMP-7 were secreted by ovarian cancer cells (Figure 1C). Ovarian cancer cells that expressed more MSLN also expressed more MMP-7. In addition, cancer cells that secreted lower amounts of MSLN also secreted less MMP-7.

We further quantified the correlation between the expression levels of MSLN and MMP-7 in respective ovarian cancerous tissues by real-time PCR. The correlation between MSLN and MMP-7 expression levels in ovarian cancer tissues evaluated by real-time PCR analysis also showed a positive correlation (r=0.694, P<0.001, Spearman's correlation; Figure 1D).

MSLN enhanced MMP-7 expression in ovarian cancer cells

To elucidate the positive correlation of MSLN and MMP-7 expression, two ascitogenic murine cancer cell lines, WF-3 and WF-0, were used. WF-3 tumour cells have stronger abilities of tumour invasion and migration than WF-0 tumour cells [18]. The WF-3 cells expressed higher levels of MSLN and MMP-7 than WF-0 cells as measured by both RT–PCR (Figure 2A) and immunoblotting analyses (Figure 2B).

Expression of MMP-7 in WF-0, WF-3 and various MSLN-transfectants of WF-0 cells

Figure 2
Expression of MMP-7 in WF-0, WF-3 and various MSLN-transfectants of WF-0 cells

(A) RT–PCR analyses and (B) Western blot analyses of MMP-7 expression in low MSLN-expressing WF-0 and high MSLN-expressing WF-3 cancer cell lines. The expression of MMP-7 in WF-3 cells were significantly higher than that in WF-0 cells at both the mRNA and protein levels. (C) RT–PCR analyses and (D) Western blot analyses of MMP-7 expression in various MSLN-transfected WF-0 cell lines. MSLN-overexpressing WF-0/M2 and WF-0/M3 cells expressed higher levels of MMP-7 than MSLN low-expressing WF-0/M1 and parental WF-0 cells at both the mRNA and protein levels.

Figure 2
Expression of MMP-7 in WF-0, WF-3 and various MSLN-transfectants of WF-0 cells

(A) RT–PCR analyses and (B) Western blot analyses of MMP-7 expression in low MSLN-expressing WF-0 and high MSLN-expressing WF-3 cancer cell lines. The expression of MMP-7 in WF-3 cells were significantly higher than that in WF-0 cells at both the mRNA and protein levels. (C) RT–PCR analyses and (D) Western blot analyses of MMP-7 expression in various MSLN-transfected WF-0 cell lines. MSLN-overexpressing WF-0/M2 and WF-0/M3 cells expressed higher levels of MMP-7 than MSLN low-expressing WF-0/M1 and parental WF-0 cells at both the mRNA and protein levels.

To further characterize the relationship between MSLN and MMP-7, several endogenous MSLN transfectants, such as WF-0/M1, WF-0/M2 and WF-0/M3 were used. The highly MSLN-expressing clones WF-0/M3 and WF-0/M2 expressed higher MMP-7 RNA and protein than parental WF-0 cells and the low MSLN-expressing clone WF-0/M1 (Figures 2C and 2D).

MSLN enhanced the invasive capability of ovarian cancer cells

To investigate whether MSLN plays a role in the invasive capability of cancer cells, invasive experiments were performed using Boyden chambers. Representative images of Boyden chamber experiments are shown in Figure 3(A). The highly MSLN-expressing clone WF-0/M3 had the strongest invasive abilities (Figure 3B). The clone revealed an ~8-fold increased invasive ability (865±124%) compared with parental WF-0 cells (Figure 3B). Furthermore, the endogenous MSLN-expressing WF-3 tumour cells also revealed a 2-fold increased invasive ability (225±35%) compared with WF-0 parental cells (Figure 3B).

MSLN enhanced the invasive and migrating capabilities of ovarian cancer cells

Figure 3
MSLN enhanced the invasive and migrating capabilities of ovarian cancer cells

(A) Representative Figures of cell invasion using Boyden chamber experiments. A1, WF-0 cells; A2, WF-3 cells; A3, WF-0/mock cells; and A4, WF-0/M3 cells. (B) Quantitative results of cell invasion by various MSLN-expressing cancer cell lines using Boyden chamber experiments. Results are presented as the fold change (tested group/parental WF-0 group) from three independent experiments and are means±S.E.M. * indicates P<0.05. MSLN-overexpressing tumour cells had a higher invasive ability than tumour cells with low expression of MSLN. (C) Quantitative results of WF-0 cells treated with various concentrations of MSLN in cell invasion experiments. Results are presented as the fold change (MSLN-treated WF-0 group/parental WF-0 group) from three independent experiments and are means±S.E.M. * indicates P<0.05. Exogenous MSLN enhanced the invasive ability of WF-0 cells. (D) Quantitative results of cell invasion assessed by Boyden chamber experiments in various MSLN-knockdown OVCAR-3 cell lines. Results are presented as the fold change (MSLN down-regulated group/parental OVCAR-3 group) from three independent experiments and are means±S.E.M. * indicates P<0.05. The invasive ability of OVCAR-3 cells was decreased by down-regulating MSLN. (E) Representative images of migration assays in WF-0 cells treated with recombinant MSLN. E1, WF-0 treated with PBS; E2, WF-0 cells treated with 1 nM MSLN; E3, WF-0 cells treated with 3 nM MSLN; and E4, WF-0 cells treated with 10 nM MSLN. (F) Quantitative results of WF-0 cells treated with recombinant MSLN in migration experiments. Results are presented as the fold change (MSLN-treated group/parental WF-0 group) and are means±S.E.M. * indicates P<0.05. The numbers of migrating WF-3 cells increased in a dose-dependent manner by treating with MSLN. MSLN at a concentration of 10 nM significantly enhanced cell migrating activity.

Figure 3
MSLN enhanced the invasive and migrating capabilities of ovarian cancer cells

(A) Representative Figures of cell invasion using Boyden chamber experiments. A1, WF-0 cells; A2, WF-3 cells; A3, WF-0/mock cells; and A4, WF-0/M3 cells. (B) Quantitative results of cell invasion by various MSLN-expressing cancer cell lines using Boyden chamber experiments. Results are presented as the fold change (tested group/parental WF-0 group) from three independent experiments and are means±S.E.M. * indicates P<0.05. MSLN-overexpressing tumour cells had a higher invasive ability than tumour cells with low expression of MSLN. (C) Quantitative results of WF-0 cells treated with various concentrations of MSLN in cell invasion experiments. Results are presented as the fold change (MSLN-treated WF-0 group/parental WF-0 group) from three independent experiments and are means±S.E.M. * indicates P<0.05. Exogenous MSLN enhanced the invasive ability of WF-0 cells. (D) Quantitative results of cell invasion assessed by Boyden chamber experiments in various MSLN-knockdown OVCAR-3 cell lines. Results are presented as the fold change (MSLN down-regulated group/parental OVCAR-3 group) from three independent experiments and are means±S.E.M. * indicates P<0.05. The invasive ability of OVCAR-3 cells was decreased by down-regulating MSLN. (E) Representative images of migration assays in WF-0 cells treated with recombinant MSLN. E1, WF-0 treated with PBS; E2, WF-0 cells treated with 1 nM MSLN; E3, WF-0 cells treated with 3 nM MSLN; and E4, WF-0 cells treated with 10 nM MSLN. (F) Quantitative results of WF-0 cells treated with recombinant MSLN in migration experiments. Results are presented as the fold change (MSLN-treated group/parental WF-0 group) and are means±S.E.M. * indicates P<0.05. The numbers of migrating WF-3 cells increased in a dose-dependent manner by treating with MSLN. MSLN at a concentration of 10 nM significantly enhanced cell migrating activity.

Exogenous MSLN could significantly enhance the invasive activities of WF-0 parental cells in a dose-dependent manner (106.8±13.5% for 1 nM MSLN, 125.7±20.4% for 3 nM MSLN and 237.1±36.9% for 10 nM MSLN, when taking the invasive activity of WF-0 cells treated with PBS only as 100%) (Figure 3C).

Suppression of MSLN could reduce the invasive capabilities of cancer cells, and the numbers of invasive cells in siRNA/MSLN transfectants were significantly fewer than those in parental OVCAR-3 or OVCAR-3/mock group (105.6±3.7% for OVCAR-3/mock, 20.48±3.6% for OVCAR-3/siMSLN c12, 31.5±5.9% for OVCAR-3/siMSLN c13 and 13.8±4.6% for OVCAR-3/siMSLN c14, when taking the invasive activity of parental OVCAR-3 cells as 100%) (Figure 3D).

MSLN enhanced the migrating ability of ovarian cancer cells

To evaluate whether MSLN could also enhance the migrating ability of ovarian cancer cells, cell migration assays were performed. Representative images of cell migration assays are shown in Figure 3(E). The numbers of migrating cells in MSLN-treated WF-0 cells increased when the concentration of MSLN increased (51.4±10.6 for PBS, 87.1±23.2 for 1 nM MSLN, 122.8±29.5 for 3 nM MSLN and 657.0±125.4 for 10 nM MSLN, P<0.001, one-way ANOVA) (Figure 3F).

MSLN regulated MMP-7 to effect the invasive and migrating abilities of ovarian cancer cells

Members of the MMP family, such as MMP-2, -9, -10 or -13, have been reported to be involved in the progression of ovarian cancer. The expression of MMP family members is critical for tumour progression of ovarian cancer [24]. In light of this, we further analysed whether MMPs were involved in MSLN-induced invasive behaviours of ovarian cancer. RT–PCR analysis showed that only MMP7, not MMP2, MMP9, MMP10 or MMP13, was induced by 6 h after MSLN stimulation. The expression of MMP7 persisted even 48 h after MSLN stimulation (Figure 4A). The enhancement of MMP-7 protein in WF-0 cells also showed a similar phenomenon after MSLN stimulation, as assessed by immunoblot analysis (Figure 4B).

MSLN enhanced MMP-7 expression in ovarian cancer cells

Figure 4
MSLN enhanced MMP-7 expression in ovarian cancer cells

(A) RT–PCR analyses of MMP2, 7, 9, 10 and 13 in WF-0 cells treated with recombinant MSLN. Only MMP7 expression could be regulated by MSLN. (B) Western blot analyses of MMP-7 in WF-0 cells treated with recombinant MSLN. MSLN enhanced mRNA and protein levels of MMP-7 at 6 h after MSLN treatment, and MMP-7 expression persisted even 48 h after MSLN treatment. (C) Western blot analyses of MSLN (top panel) and MMP-7 (middle panel) expression levels in various siMSLN-transfected OVCAR-3 derivatives. MMP-7 expression was decreased in a similar manner to MSLN when MSLN was blocked by siMSLN. (D) Luciferase activities of MMP7 promoter-transfected WF-0 cells treated with recombinant MSLN. The activity of MMP7 promoter-driven luciferase was enhanced by MSLN in a dose- and time-dependent manner. Results are means±S.E.M. (E) Representative images of migration assays in WF-0 cells treated with siMMP-7 followed by recombinant MSLN. E1, WF-0 cells treated with MSLN for 12 h; E2, WF-0 cells treated with MSLN for 24 h; E3, WF-0 cells treated with MSLN for 48 h; E4, WF-0/siMMP-7 cells treated with MSLN for 12 h; E5, WF-0/siMMP-7 cells treated with MSLN for 24 h; and E6, WF-0/siMMP-7 cells treated with MSLN for 48 h. (F) Quantitative cell migration results of WF-0 cells treated with siMMP-7 followed by recombinant MSLN. Results are presented as the number of migrating cells and are means±S.E.M. * indicates P<0.05. The number of migrating WF-0 cells was enhanced by MSLN in a time-dependent manner. A concentration of 10 nM MSLN significantly enhanced cell migrating activity. Blockage of MMP-7 expression in the WF-0/siMMP-7 cells could inhibit the cell migration activity enhanced by MSLN.

Figure 4
MSLN enhanced MMP-7 expression in ovarian cancer cells

(A) RT–PCR analyses of MMP2, 7, 9, 10 and 13 in WF-0 cells treated with recombinant MSLN. Only MMP7 expression could be regulated by MSLN. (B) Western blot analyses of MMP-7 in WF-0 cells treated with recombinant MSLN. MSLN enhanced mRNA and protein levels of MMP-7 at 6 h after MSLN treatment, and MMP-7 expression persisted even 48 h after MSLN treatment. (C) Western blot analyses of MSLN (top panel) and MMP-7 (middle panel) expression levels in various siMSLN-transfected OVCAR-3 derivatives. MMP-7 expression was decreased in a similar manner to MSLN when MSLN was blocked by siMSLN. (D) Luciferase activities of MMP7 promoter-transfected WF-0 cells treated with recombinant MSLN. The activity of MMP7 promoter-driven luciferase was enhanced by MSLN in a dose- and time-dependent manner. Results are means±S.E.M. (E) Representative images of migration assays in WF-0 cells treated with siMMP-7 followed by recombinant MSLN. E1, WF-0 cells treated with MSLN for 12 h; E2, WF-0 cells treated with MSLN for 24 h; E3, WF-0 cells treated with MSLN for 48 h; E4, WF-0/siMMP-7 cells treated with MSLN for 12 h; E5, WF-0/siMMP-7 cells treated with MSLN for 24 h; and E6, WF-0/siMMP-7 cells treated with MSLN for 48 h. (F) Quantitative cell migration results of WF-0 cells treated with siMMP-7 followed by recombinant MSLN. Results are presented as the number of migrating cells and are means±S.E.M. * indicates P<0.05. The number of migrating WF-0 cells was enhanced by MSLN in a time-dependent manner. A concentration of 10 nM MSLN significantly enhanced cell migrating activity. Blockage of MMP-7 expression in the WF-0/siMMP-7 cells could inhibit the cell migration activity enhanced by MSLN.

To test whether down-regulation of MSLN could reduce the expression of MMP-7, the expression of MSLN in the siRNA/MSLN OVCAR-3 transfectants was lowered compared with that in parental OVCAR-3 and OVCAR-3/mock groups (Figure 4C). The expression of MMP-7 was also decreased in these siRNA/MSLN OVCAR-3 transfectants as compared with that in the parental OVCAR-3 and OVCAR-3/mock groups (Figure 4C).

MSLN induced the transcriptional activity of MMP-7

To further examine whether MSLN could regulate MMP-7 expression at the transcriptional level, luciferase activity assays were performed. Luciferase activities of MMP-7 were significantly increased by MSLN in a dose-dependent manner (1.24±0.16 fold for 1 nM MSLN, 1.57±0.24 for 3 nM MSLN and 3.69±0.29 for 10 nM MSLN, when taking the luciferase activity by PBS as 100%) (Figure 4D). However, luciferase activities of the mock control (PGL3-Basic) did not show any change after treatment with MSLN.

To address whether MSLN-mediated tumour invasion is dependent on MMP-7 activity, we used siRNA experiments to knockdown MMP7 expression in vitro. As shown in Figure 4(E), cell migration activity was significant reduced when the MMP-7 activity was being blocked by siRNA, even in cells treated with MSLN (78.0±35.6 for WF-0 cells treated with MSLN for 12 h, 109.7±41.7 for WF-0 cells treated with MSLN for 24 h, 442.7±81.2 for WF-0 cells treated with MSLN for 48 h, 58.3±16.9 for WF-0/siMMP7 cells treated with MSLN for 12 h, 82.7±26.6 for WF-0/siMMP7 cells treated with MSLN for 24 h and 138.3±61.2 for WF-0/siMMP7 cells treated with MSLN for 48 h respectively) (Figure 4F). Invasion experiments using a Boyden chamber also demonstrated similar results (results not shown).

These results suggest that MMP-7 plays an important role in MSLN-mediated cell invasive and migrating behaviours.

MSLN directed MMP-7 by activating MAPK/ERK and JNK signal transduction pathways

To investigate the possible signalling pathways for MSLN-induced migration and invasion of ovarian cancer cells, the phosphorylation of various molecules in MSLN-treated WF-0 cells was evaluated by immunoblot analyses. The phosphorylation of ERK1/2, Akt and JNK was quickly up-regulated when treated with MSLN for only 30 min (Figure 5A). However, the phosphorylation of p38 (MAPK) did not change when stimulated with MSLN (Figure 5A).

Signal-transduction mechanisms involved in MSLN-mediated MMP-7 up-regulation in ovarian cancer cells

Figure 5
Signal-transduction mechanisms involved in MSLN-mediated MMP-7 up-regulation in ovarian cancer cells

(A) Western blotting analyses of phosphorylation of ERK1/2, Akt, JNK and p38 MAPK molecules in WF-0 cells treated with recombinant MSLN. ERK1/2, Akt and JNK phosphorylation were up-regulated after treatment with MSLN for 30 min. (B) Migration assays of WF-0 cells treated with various inhibitors of signal transduction molecules. Results are presented as the fold change (tested group/parental WF-0 group) from three independent experiments and are means±S.E.M. * indicates P<0.05. The migrating ability of MSLN-directed WF-0 cells was abolished by ERK1/2 (PD098059)- or JNK (SP600125)-specific inhibitors, but not by PI3K/Akt (LY294002)- or p38 (SB203580)-specific inhibitors. (C) Western blotting analyses of MMP-7 expression in cells treated with recombinant MSLN and/or various inhibitors of signal transduction molecules. MMP-7 expression induced by MSLN was blocked by the ERK1/2 (PD098059)- or JNK (SP600125)-specific inhibitors, but not by the PI3K/Akt (LY294002)- or p38 (SB203580)-specific inhibitors. (D) Luciferase activities of MMP7 promoter-transfected WF-0 cells treated with recombinant MSLN and/or various inhibitors of signal transduction molecules. Results are presented as the fold change (experimental group/WF-0 cell group with PBS only) and are means±S.E.M. from three independent experiments. The MSLN-stimulated luciferase activity (3.29±0.3-fold) could be significantly inhibited by the ERK1/2 inhibitor (PD098059 25 μM, 1.34±0.21-fold) and the JNK inhibitor (SP600125 10 μM, 1.52±0.23-fold) in WF-0 cells transfected with the MMP7 promoter (P<0.01, one-way ANOVA). (E) Luciferase activities of MMP7 promoter-transfected WF-0 cells treated with MSLN and/or AP-1 decoy oligonucleotide. Results are presented as the fold change (experimental group/WF-0 cell group with PBS only) and are means±S.E.M. from three independent experiments. Luciferase activity significantly decreased when pre-treated with AP-1 decoy oligonucleotide (1.31±0.20-fold for AP-1 agonist, 3.41±0.41-fold for control, P<0.01, one-way ANOVA).

Figure 5
Signal-transduction mechanisms involved in MSLN-mediated MMP-7 up-regulation in ovarian cancer cells

(A) Western blotting analyses of phosphorylation of ERK1/2, Akt, JNK and p38 MAPK molecules in WF-0 cells treated with recombinant MSLN. ERK1/2, Akt and JNK phosphorylation were up-regulated after treatment with MSLN for 30 min. (B) Migration assays of WF-0 cells treated with various inhibitors of signal transduction molecules. Results are presented as the fold change (tested group/parental WF-0 group) from three independent experiments and are means±S.E.M. * indicates P<0.05. The migrating ability of MSLN-directed WF-0 cells was abolished by ERK1/2 (PD098059)- or JNK (SP600125)-specific inhibitors, but not by PI3K/Akt (LY294002)- or p38 (SB203580)-specific inhibitors. (C) Western blotting analyses of MMP-7 expression in cells treated with recombinant MSLN and/or various inhibitors of signal transduction molecules. MMP-7 expression induced by MSLN was blocked by the ERK1/2 (PD098059)- or JNK (SP600125)-specific inhibitors, but not by the PI3K/Akt (LY294002)- or p38 (SB203580)-specific inhibitors. (D) Luciferase activities of MMP7 promoter-transfected WF-0 cells treated with recombinant MSLN and/or various inhibitors of signal transduction molecules. Results are presented as the fold change (experimental group/WF-0 cell group with PBS only) and are means±S.E.M. from three independent experiments. The MSLN-stimulated luciferase activity (3.29±0.3-fold) could be significantly inhibited by the ERK1/2 inhibitor (PD098059 25 μM, 1.34±0.21-fold) and the JNK inhibitor (SP600125 10 μM, 1.52±0.23-fold) in WF-0 cells transfected with the MMP7 promoter (P<0.01, one-way ANOVA). (E) Luciferase activities of MMP7 promoter-transfected WF-0 cells treated with MSLN and/or AP-1 decoy oligonucleotide. Results are presented as the fold change (experimental group/WF-0 cell group with PBS only) and are means±S.E.M. from three independent experiments. Luciferase activity significantly decreased when pre-treated with AP-1 decoy oligonucleotide (1.31±0.20-fold for AP-1 agonist, 3.41±0.41-fold for control, P<0.01, one-way ANOVA).

The migrating ability of WF-0 cells enhanced by MSLN were abolished by ERK1/2 (PD098059)- or JNK (SP600125)-specific inhibitors, but not by PI3K/Akt- (LY294002) or p38 (SB203580)-specific inhibitors (270.5±12.4 for MSLN alone, 113.8±19.3 for PD098059 plus MSLN, 220.5±28.6 for LY294002 plus MSLN, 126.9±15.5 for SP600125 plus MSLN and 300.2±21.5 for SB203580 plus MSLN; P<0.01, one-way ANOVA) (Figure 5B).

Next we evaluated whether MSLN-induced MMP-7 expression could be suppressed by various inhibitors of transduction molecules. The protein expression of MMP-7 enhanced by MSLN was decreased when WF-0 cells were pre-treated with ERK1/2 (PD098059)- or JNK (SP600125)-specific inhibitors, but not when pre-treated with PI3K/Akt (LY294002)- or p38 (SB203580)-specific inhibitors (Figure 5C).

To further elucidate whether the MSLN-induced transduction molecules enhanced MMP-7 expression at the transcriptional level, luciferase activity assays were performed. When taking the luciferase activity in the control group as 100%, luciferase activities enhanced by MSLN (3.29±0.30-fold) could be suppressed by the ERK1/2 inhibitor (PD098059 25 μM, 1.34±0.21-fold) and the JNK inhibitor (SP600125 10 μM, 1.52±0.23-fold), but not by the PI3K/Akt inhibitor (LY294002 25 μM 2.94±0.19-fold) or p38 inhibitor (SB203580 10 μM, 3.10±0.15-fold) in WF-0 cells transfected with the Mmp7 promoter (P<0.01, one-way ANOVA) (Figure 5D).

MAPK/ERK and JNK signal transduction-mediated AP-1 activation was involved in MSLN-induced MMP-7 expression

To elucidate whether AP-1, the key transcription factor that cross-links to ERK and JNK, was involved in the MSLN-regulated MMP7 gene transcription, luciferase activity assays were performed. The luciferase activity enhanced by MSLN (3.41±0.41-fold) was significantly decreased when pre-treated with AP-1 decoy oligonucleotide (1.31±0.20-fold, P<0.01, one-way ANOVA) (Figure 5E).

MSLN-mediated tumour growth was suppressed by a MAPK/ERK or JNK inhibitor through down-regulation of MMP-7 expression

To examine whether in vitro experiments could be further translated in vivo, in vivo tumour growth experiments were performed. Representative images of tumour sizes are shown in Figure 6(A). The PD098059 (MAPK/ERK inhibitor)- or SP600125 (JNK inhibitor)-treated groups had smaller tumour volumes than DMSO-treated groups (Figure 6B). The decrease in tumour volume compared with the MSLN-overexpressing WF-3 cell line was ~65% for PD098059 treatment and ~56% for SP600125 treatment. However, tumour volumes did not decrease in the LY294002-treated group (results not shown). Intratumour Mmp7 expression in various groups was further measured. There was a significantly lower Mmp7 expression in tumour tissues of the PD098059- or SP600125-treated groups compared with that in the DMSO-treated group (Figure 6C).

Signal-transduction inhibitors suppressed MSLN-mediated tumour growth in ovarian cancer-bearing mice

Figure 6
Signal-transduction inhibitors suppressed MSLN-mediated tumour growth in ovarian cancer-bearing mice

(A) Representative images of tumour growth. WF-3 cells were subcutaneously inoculated into both hind limbs and treated with either DMSO or signal transduction pathway-specific inhibitors. A1, 42 days after DMSO or PD098059 injection; and A2, 42 days after DMSO or SP600125 injection. (B) The tumour volumes in mice receiving either PD098059 or SP600125. The tumour volumes in the PD098059 and SP600125 groups were significantly smaller than those in the respective DMSO-only group (P<0.01, one-way ANOVA). (C) RT–PCR analyses of intratumor Mmp7 expression in various groups. Mmp7 expression levels in the PD098059 and SP600125 groups were lower than those in the DMSO-only group. (D) In vivo tumour therapeutic experiments. WF3 cells were inoculated with 2×105 WF-3 tumour cells intraperitoneally. DMSO or a signal transduction pathway-specific inhibitor was injected intraperitoneally 3 days after tumour challenge. Results are presented as the percentage of tumour-free mice. PD098059- and SP600125-treated groups could prolong the percentage of tumor-free mice compared with DMSO-treated groups (P=0.008, log-rank test). (E) Percentage of live mice. Mice were inoculated with 2×105 WF-3 cells, and DMSO or a signal transduction pathway-specific inhibitor was injected as described above. Results are presented as the percentage of live mice. More than 50% of mice treated with either PD098059 or SP600125 were still alive even 50 days after the WF-3 challenge as compared with the untreated group (P=0.007, log-rank test). In contrast, all of the mice in the untreated group died within 50 days after tumour challenge.

Figure 6
Signal-transduction inhibitors suppressed MSLN-mediated tumour growth in ovarian cancer-bearing mice

(A) Representative images of tumour growth. WF-3 cells were subcutaneously inoculated into both hind limbs and treated with either DMSO or signal transduction pathway-specific inhibitors. A1, 42 days after DMSO or PD098059 injection; and A2, 42 days after DMSO or SP600125 injection. (B) The tumour volumes in mice receiving either PD098059 or SP600125. The tumour volumes in the PD098059 and SP600125 groups were significantly smaller than those in the respective DMSO-only group (P<0.01, one-way ANOVA). (C) RT–PCR analyses of intratumor Mmp7 expression in various groups. Mmp7 expression levels in the PD098059 and SP600125 groups were lower than those in the DMSO-only group. (D) In vivo tumour therapeutic experiments. WF3 cells were inoculated with 2×105 WF-3 tumour cells intraperitoneally. DMSO or a signal transduction pathway-specific inhibitor was injected intraperitoneally 3 days after tumour challenge. Results are presented as the percentage of tumour-free mice. PD098059- and SP600125-treated groups could prolong the percentage of tumor-free mice compared with DMSO-treated groups (P=0.008, log-rank test). (E) Percentage of live mice. Mice were inoculated with 2×105 WF-3 cells, and DMSO or a signal transduction pathway-specific inhibitor was injected as described above. Results are presented as the percentage of live mice. More than 50% of mice treated with either PD098059 or SP600125 were still alive even 50 days after the WF-3 challenge as compared with the untreated group (P=0.007, log-rank test). In contrast, all of the mice in the untreated group died within 50 days after tumour challenge.

To more closely model the ovarian tumour microenvironment, an intraperitoneal injection was used. As shown in Figure 6(D), PD098059- and SP600125-treated groups had lower percentages of tumour-bearing mice than the DMSO-treated group 10 days after tumour challenge (P=0.008, log-rank test). The survival experiments also indicated that more than 50% of mice receiving either PD098059 or SP600125 were still alive even 50 days after WF-3 challenge (P=0.007, log-rank test). In contrast, all of the mice in the untreated group developed tumours and died within 20 and 50 days after tumour challenge (Figures 6D and 6E).

Our results indicated that the signal transduction inhibitors PD098059 or SP600125 could delay the in vivo growth of MSLN-expressing WF-3 tumour cells and extend the survival of mice challenged with WF-3 tumour cells.

DISCUSSION

Although MSLN can be used as a new marker for detecting ovarian epithelial cancers [17], its exact function in tumour progression remains unclear. The elevation of MSLN is also found in other types of malignancies, such as pancreatic cancer [25], biliary carcinomas [26] and mesotheliomas [27]. A previous study has confirmed that MSLN can play an important role in chemotherapy-induced apoptosis in ovarian cancer [15]. Identifying the function of MSLN will enhance its clinical application in ovarian cancer, including early detection, chemo-response, prognosis and therapeutic targeting. The MSLN receptor has not yet been identified, and clarifying it is also an important issue to address the biological activity of MSLN-mediated signalling networks. The MSLN receptor may even be a potential therapeutic target of malignancies.

MSLN can activate the ERK signal pathway to execute its biological function. The aggressiveness of invasion and migration of tumour cells determines the metastatic potential that remains a significantly poor prognostic feature of ovarian cancer [28]. In the human breast cancer model, MSLN has been found to prevent anoikis and promote cell growth by suppressing Bim induction, despite being dependent on the activation of the ERK signalling cascades [29]. However, MSLN may have other molecular mechanisms not yet identified that mediate its function. The present study shows that MSLN can activate the ERK signal pathway to promote the migrating and invasive capabilities of ovarian cancer cells. ERK is a critical mediator of cell proliferation [30]. This is consistent with previous reports that the inhibition of the ERK pathway can eliminate the growth of many types of cancer cells in in vitro studies [31]. Blocking the MAPK/ERK cascade can inhibit MMP-7 expression and suppress tumour growth (Figures 5 and 6).

ERK and JNK are two known so-called mitogenic pathways in mammalian cells that can induce MMP-9 up-regulation [32]. Simon et al. [33] have shown that inhibition of p38 leads to reduced phorbol ester-induced MMP-9 expression and invasion by tumour cells. However, the inhibition of p38 does not reduce MSLN-enhancing MMP-7 expression and the invasion of ovarian cancer cells in the present study (Figures 5A and 5B). Activation of JNK correlates with the migration and invasion of tumour cells, and this is well-established in previous studies [34,35]. Such evidence suggests that JNK plays an important role in tumour invasion and metastasis. MSLN, in the present study, has the function of enhancing both ERK and JNK activation, inducing MMP-7 expression, and promoting cancer invasion activity. However, Kim et al. [36] reported that MMP-7 expression can be induced by JNK1/2 activation, but not in the case of ERK1/2 or p38 in human colorectal cancer. The explanation is that different signal pathways are involved in MMP-7 expression in different histological types of cancers.

MSLN activates MMP-7 to enhance tumour progression in ovarian cancer. MMP-7 has the minimal domain organization required for secretion, latency and activity [4]. Its correlation with tumour progression in ovarian cancer has been identified previously [37,38]. The CoexpressDB database has reported that MMP-7 may co-express with MSLN (http://coxpresdb.jp/cgi-bin/coex_list.cgi?gene=4316&sp=Hsa). However, the regulation of MMP-7 in ovarian cancer has not yet been identified. In the present study, we also observed that the expression of MMP-7 correlates well with MSLN in ovarian cancerous tissue and cancer cell lines (Figure 1). MMP-7 expression is decreased by down-regulating MSLN expression (Figure 4C). In addition, MSLN enhances the expression of MMP7 at the transcriptional level (Figures 4D, 5D and 5E). MSLN has the function of inducing MMP-7 secretion to enhance the migration and invasion of ovarian cancer cells.

The AP-1 transcription factor can form either a homo- or hetero-complex that can directly bind as c-fos/c-fos, c-jun/c-jun or c-fos/c-jun to activate endo-or exo-genous stimulation [36]. AP-1 is also a key transcriptional regulator in both ERK and JNK signal transductions [39]. Previous studies demonstrate that AP-1 can be an important transcription factor to mediate MMP7 expression [40]. The activation of MMP7 gene transcription induced by MSLN in ovarian cancer can be mediated by AP-1 in the present study (Figure 5E). Previous studies also demonstrate that MMP7 can be up-regulated by Tcf or Lef-1 [41]. It will be interesting to investigate whether MSLN can regulate MMP7 gene transcription by these factors.

MSLN could be an emerging marker for the diagnosis and target-based therapy of ovarian epithelial cancers. Previous studies have shown that MSLN is expressed in ovarian cancer tissues, and ovarian cancer tissues overexpressing MSLN have a higher possibility of chemoresistance [17,19,42], suggesting that MSLN could be an ideal marker for cancer diagnosis and target-based therapy. The molecular markers involved in the activity of chemotherapeutic agents could shed light on the successes and failures of ovarian cancer treatment and could provide a basis for individualized therapy. Clinical trial studies reveal that an anti-MSLN monoclonal antibody, MORAb-009, can inhibit the MSLN–CA-125 interaction and may be a useful strategy for the prevention of tumour metastasis in mesotheliomas and ovarian cancers [16,43].

Targeted therapy can be an effective cancer therapeutic strategy in the ERK or JNK signalling pathways. These two pathways are activated by a wide variety of mitogenic stimuli that interact with distinct receptors and turn on biological behaviours, including mitosis [44], differentiation [45] and tumour progression [46]. The present study has identified that the activation of ERK and JNK are crucial in MSLN-induced MMP-7 for the migration and invasion of ovarian cancer. Therefore specifically blocking both the ERK and JNK pathways can abolish tumorigenesis (Figures 5 and 6). As such, the ERK and JNK pathways represent attractive therapeutic targets for ovarian cancer. With the universal role of ERK and JNK pathways in regulating normal cellular functions, sensitive and specific inhibitors for cancer cells are important to minimize the undesirable effects on normal cells.

Abbreviations

     
  • AP-1

    activator protein 1

  •  
  • Cy3

    indocarbocyanine

  •  
  • Cy5

    indodicarbocyanine

  •  
  • ECM

    extracellular matrix

  •  
  • ERK

    extracellular-signal-regulated kinase

  •  
  • GAPDH

    glyceraldehyde-3-phosphate dehydrogenase

  •  
  • HRP

    horseradish peroxidase

  •  
  • JNK

    c-Jun N-terminal kinase

  •  
  • MAPK

    mitogen-activated protein kinase

  •  
  • MMLV

    Moloney murine leukaemia virus

  •  
  • MMP

    matrix metalloproteinase

  •  
  • MSLN

    mesothelin

  •  
  • PI3K

    phosphoinositide 3-kinase

  •  
  • RT

    reverse transcriptase

  •  
  • siRNA

    small interfering RNA

AUTHOR CONTIBUTION

Wen-Fang Cheng was the key principal investigator. He designed the experiments and wrote the paper with input from all of the other authors. Chi-An Chen and Wen-Fang Cheng provided clinical specimens and correlated the clinical parameters. Ming-Cheng Chang participated in all of the experimental procedures, including the design and performance of all of the experiments. Chi-An Chen and Pao-Jen Chen participated in the interpretation of experimental results and discussion. Ying-Cheng Chiang also advised on and evaluated the animal experiments. Yu-Li Chen offered technical support for the experiments, and helped with the interpretation and discussion of the results. Han-Wei Lin and Wen-Hsien Lin Chiang were the research assistants for the project, and assisted in the execution of all of the experiments.

We thank Dr. L.M. Matrisian (Vanderbilt University School of Medicine, Nashville, TN, U.S.A.) for providing a 0.7-kb segment of the 5′-flanking region of the mouse Mmp7 promoter region.

FUNDING

This work was supported, in part, by the 2nd and 7th core laboratory facility of the Department of Medical Research of National Taiwan University Hospital and granted by the National Science Committee of Taiwan [grant number 97-2314-B-002-064-MY3], National Taiwan University Hospital [grant number 97-000837] and Min-Sheng General Hospital, Taoyuan, Taiwan-National Taiwan University Hospital Joint Research Program [grant number 97MSN05].

References

References
1
Egeblad
M.
Werb
Z.
New functions for the matrix metalloproteinases in cancer progression
Nat. Rev. Cancer
2002
, vol. 
2
 (pg. 
161
-
174
)
2
Sillanpaa
S.
Anttila
M.
Suhonen
K.
Hamalainen
K.
Turpeenniemi-Hujanen
T.
Puistola
U.
Tammi
M.
Sironen
R.
Saarikoski
S.
Kosma
V. M.
Prognostic significance of extracellular matrix metalloproteinase inducer and matrix metalloproteinase 2 in epithelial ovarian cancer
Tumour Biol.
2007
, vol. 
28
 (pg. 
280
-
289
)
3
Noriyuki
M.
Sumi
T.
Zhi
X.
Misugi
F.
Nobeyama
H.
Yoshida
H.
Matsumoto
Y.
Yasui
T.
Honda
K.
Ishiko
O.
Vascular endothelial growth factor, matrix metalloproteinases, and cyclooxygenase-2 influence prognosis of uterine cervical cancer in young women
Int. J. Oncol.
2007
, vol. 
31
 (pg. 
531
-
536
)
4
Gaire
M.
Magbanua
Z.
McDonnell
S.
McNeil
L.
Lovett
D. H.
Matrisian
L. M.
Structure and expression of the human gene for the matrix metalloproteinase matrilysin
J. Biol. Chem.
1994
, vol. 
269
 (pg. 
2032
-
2040
)
5
Chambers
A. F.
Matrisian
L. M.
Changing views of the role of matrix metalloproteinases in metastasis
J. Natl. Cancer Inst.
1997
, vol. 
89
 (pg. 
1260
-
1270
)
6
Rudolph-Owen
L. A.
Chan
R.
Muller
W. J.
Matrisian
L. M.
The matrix metalloproteinase matrilysin influences early-stage mammary tumorigenesis
Cancer Res.
1998
, vol. 
58
 (pg. 
5500
-
5506
)
7
Wilson
C. L.
Heppner
K. J.
Labosky
P. A.
Hogan
B. L.
Matrisian
L. M.
Intestinal tumorigenesis is suppressed in mice lacking the metalloproteinase matrilysin
Proc. Natl. Acad. Sci. U.S.A.
1997
, vol. 
94
 (pg. 
1402
-
1407
)
8
Yamamoto
H.
Itoh
F.
Adachi
Y.
Sakamoto
H.
Adachi
M.
Hinoda
Y.
Imai
K.
Relation of enhanced secretion of active matrix metalloproteinases with tumor spread in human hepatocellular carcinoma
Gastroenterology
1997
, vol. 
112
 (pg. 
1290
-
1296
)
9
Bast
R. C.
Jr
Klug
T. L.
St John
E.
Jenison
E.
Niloff
J. M.
Lazarus
H.
Berkowitz
R. S.
Leavitt
T.
Griffiths
C. T.
Parker
L.
, et al. 
A radioimmunoassay using a monoclonal antibody to monitor the course of epithelial ovarian cancer
N. Engl. J. Med.
1983
, vol. 
309
 (pg. 
883
-
887
)
10
Chang
K.
Pastan
I.
Willingham
M. C.
Isolation and characterization of a monoclonal antibody, K1, reactive with ovarian cancers and normal mesothelium
Int. J. Cancer
1992
, vol. 
50
 (pg. 
373
-
381
)
11
Chang
K.
Pastan
I.
Molecular cloning of MSLN, a differentiation antigen present on mesothelium, mesotheliomas, and ovarian cancers
Proc. Natl. Acad. Sci. U.S.A.
1996
, vol. 
93
 (pg. 
136
-
140
)
12
Argani
P.
Iacobuzio-Donahue
C.
Ryu
B.
Rosty
C.
Goggins
M.
Wilentz
R. E.
Murugesan
S. R.
Leach
S. D.
Jaffee
E.
Yeo
C. J.
, et al. 
Mesothelin is overexpressed in the vast majority of ductal adenocarcinomas of the pancreas: identification of a new pancreatic cancer marker by serial analysis of gene expression (SAGE)
Clin. Cancer Res.
2001
, vol. 
7
 (pg. 
3862
-
3868
)
13
Hippo
Y.
Yashiro
M.
Ishii
M.
Taniguchi
H.
Tsutsumi
S.
Hirakawa
K.
Kodama
T.
Aburatani
H.
Differential gene expression profiles of scirrhous gastric cancer cells with high metastatic potential to peritoneum or lymph nodes
Cancer Res.
2001
, vol. 
61
 (pg. 
889
-
895
)
14
Frierson
H. F.
Jr
Moskaluk
C. A.
Powell
S. M.
Zhang
H.
Cerilli
L. A.
Stoler
M. H.
Cathro
H.
Hampton
G. M.
Large-scale molecular and tissue microarray analysis of mesothelin expression in common human carcinomas
Hum. Pathol.
2003
, vol. 
34
 (pg. 
605
-
609
)
15
Chang
M. C.
Chen
C. A.
Hsieh
C. Y.
Lee
C. N.
Su
Y. N.
Hu
Y. H.
Cheng
W. F.
Mesothelin inhibits paclitaxel-induced apoptosis through the PI3K pathway
Biochem. J.
2009
, vol. 
424
 (pg. 
449
-
458
)
16
Hassan
R.
Bera
T.
Pastan
I.
Mesothelin: a new target for immunotherapy
Clin. Cancer Res.
2004
, vol. 
10
 (pg. 
3937
-
3942
)
17
Huang
C. Y.
Cheng
W. F.
Lee
C. N.
Su
Y. N.
Chien
S. C.
Tzeng
Y. L.
Hsieh
C. Y.
Chen
C. A.
Serum mesothelin in epithelial ovarian carcinoma: a new screening marker and prognostic factor
Anticancer Res.
2006
, vol. 
26
 (pg. 
4721
-
4728
)
18
Cheng
W. F.
Hung
C. F.
Chai
C. Y.
Chen
C. A.
Lee
C. N.
Su
Y. N.
Tseng
W. Y.
Hsieh
C. Y.
Shih Ie
M.
Wang
T. L.
Wu
T. C.
Generation and characterization of an ascitogenic mesothelin-expressing tumor model
Cancer
2007
, vol. 
110
 (pg. 
420
-
431
)
19
Reference deleted
20
Fingleton
B. M.
Heppner Goss
K. J.
Crawford
H. C.
Matrisian
L. M.
Matrilysin in early stage intestinal tumorigenesis
APMIS
1999
, vol. 
107
 (pg. 
102
-
110
)
21
Chu
C. Y.
Cha
S. T.
Chang
C. C.
Hsiao
C. H.
Tan
C. T.
Lu
Y. C.
Jee
S. H.
Kuo
M. L.
Involvement of matrix metalloproteinase-13 in stromal-cell-derived factor 1 α-directed invasion of human basal cell carcinoma cells
Oncogene
2007
, vol. 
26
 (pg. 
2491
-
2501
)
22
Chen
Y. C.
Su
Y. N.
Chou
P. C.
Chiang
W. C.
Chang
M. C.
Wang
L. S.
Teng
S. C.
Wu
K. J.
Overexpression of NBS1 contributes to transformation through the activation of phosphatidylinositol 3-kinase/Akt
J. Biol. Chem.
2005
, vol. 
280
 (pg. 
32505
-
32511
)
23
Ganesan
K.
Ivanova
T.
Wu
Y.
Rajasegaran
V.
Wu
J.
Lee
M. H.
Yu
K.
Rha
S. Y.
Chung
H. C.
Ylstra
B.
, et al. 
Inhibition of gastric cancer invasion and metastasis by PLA2G2A, a novel β-catenin/TCF target gene
Cancer Res.
2008
, vol. 
68
 (pg. 
4277
-
4286
)
24
Davidson
B.
Goldberg
I.
Gotlieb
W. H.
Kopolovic
J.
Risberg
B.
Ben-Baruch
G.
Reich
R.
Coordinated expression of integrin subunits, matrix metalloproteinases (MMP), angiogenic genes and Ets transcription factors in advanced-stage ovarian carcinoma: a possible activation pathway?
Cancer Metastasis Rev.
2003
, vol. 
22
 (pg. 
103
-
115
)
25
Jhala
N.
Jhala
D.
Vickers
S. M.
Eltoum
I.
Batra
S. K.
Manne
U.
Eloubeidi
M.
Jones
J. J.
Grizzle
W. E.
Biomarkers in diagnosis of pancreatic carcinoma in fine-needle aspirates
Am. J. Clin. Pathol.
2006
, vol. 
126
 (pg. 
572
-
579
)
26
Swierczynski
S. L.
Maitra
A.
Abraham
S. C.
Iacobuzio-Donahue
C. A.
Ashfaq
R.
Cameron
J. L.
Schulick
R. D.
Yeo
C. J.
Rahman
A.
Hinkle
D. A.
, et al. 
Analysis of novel tumor markers in pancreatic and biliary carcinomas using tissue microarrays
Hum. Pathol.
2004
, vol. 
35
 (pg. 
357
-
366
)
27
Creaney
J.
Christansen
H.
Lake
R.
Musk
A. B.
de Klerk
N.
Robinson
B. W.
Soluble mesothelin related protein in mesothelioma
J. Thorac. Oncol.
2006
, vol. 
1
 (pg. 
172
-
174
)
28
Siu
M. K.
Chan
H. Y.
Kong
D. S.
Wong
E. S.
Wong
O. G.
Ngan
H. Y.
Tam
K. F.
Zhang
H.
Li
Z.
Chan
Q. K.
, et al. 
p21-Activated kinase 4 regulates ovarian cancer cell proliferation, migration, and invasion and contributes to poor prognosis in patients
Proc. Natl. Acad. Sci. U.S.A.
2010
, vol. 
107
 (pg. 
18622
-
18627
)
29
Uehara
N.
Matsuoka
Y.
Tsubura
A.
Mesothelin promotes anchorageindependent growth and prevents anoikis via extracellular signal-regulated kinase signaling pathway in human breast cancer cells
Mol. Cancer Res.
2008
, vol. 
6
 (pg. 
186
-
193
)
30
Bermudez
Y.
Yang
H.
Cheng
J. Q.
Kruk
P. A.
Pyk2/ERK1/2 mediate Sp1- and c-Myc-dependent induction of telomerase activity by epidermal growth factor
Growth Factors
2008
, vol. 
26
 (pg. 
1
-
11
)
31
Kohno
M.
Pouyssegur
J.
Pharmacological inhibitors of the ERK signaling pathway: application as anticancer drugs
Prog. Cell Cycle Res.
2003
, vol. 
5
 (pg. 
219
-
224
)
32
Gum
R.
Wang
H.
Lengyel
E.
Juarez
J.
Boyd
D.
Regulation of 92 kDa type IV collagenase expression by the Jun amino terminal kinase- and the extracellular signal-regulated kinase-dependent signaling cascades
Oncogene
1997
, vol. 
14
 (pg. 
1481
-
1493
)
33
Simon
C.
Goepfert
H.
Boyd
D.
Inhibition of the p38 mitogen-activated protein kinase by SB 203580 blocks PMA-induced Mr 92000 type IV collagenase secretion and in vitro invasion
Cancer Res.
1998
, vol. 
58
 (pg. 
1135
-
1139
)
34
Whitmarsh
A. J.
Davis
R. J.
Role of mitogen-activated protein kinase kinase 4 in cancer
Oncogene
2007
, vol. 
26
 (pg. 
3172
-
3184
)
35
Lizundia
R.
Chaussepied
M.
Huerre
M.
Werling
D.
Di Santo
J. P.
Langsley
G.
c-Jun NH2-terminal kinase/c-Jun signaling promotes survival and metastasis of B lymphocytes transformed by Theileria
Cancer Res.
2006
, vol. 
66
 (pg. 
6105
-
6110
)
36
Kim
M.
Murakami
A.
Kawabata
K.
Ohigashi
H.
(−)-Epigallocatechin-3-gallate promotes pro-matrix metalloproteinase-7 production via activation of the JNK1/2 pathway in HT-29 human colorectal cancer cells
Carcinogenesis
2005
, vol. 
26
 (pg. 
1553
-
1562
)
37
Furuya
M.
Ishikura
H.
Ogawa
Y.
Kawarada
Y.
Shibata
M.
Fujimoto
S.
Yoshiki
T.
Analyses of matrix metalloproteinases and their inhibitors in cyst fluid of serous ovarian tumors
Pathobiology
2000
, vol. 
68
 (pg. 
239
-
244
)
38
Shigemasa
K.
Tanimoto
H.
Sakata
K.
Nagai
N.
Parmley
T. H.
Ohama
K.
O'Brien
T. J.
Induction of matrix metalloprotease-7 is common in mucinous ovarian tumors including early stage disease
Med. Oncol.
2000
, vol. 
17
 (pg. 
52
-
58
)
39
Wen-Sheng
W.
ERK signaling pathway is involved in p15INK4b/p16INK4a expression and HepG2 growth inhibition triggered by TPA and Saikosaponin a
Oncogene
2003
, vol. 
22
 (pg. 
955
-
963
)
40
Yamamoto
H.
Itoh
F.
Senota
A.
Adachi
Y.
Yoshimoto
M.
Endoh
T.
Hinoda
Y.
Yachi
A.
Imai
K.
Expression of matrix metalloproteinase matrilysin (MMP-7) was induced by activated Ki-ras via AP-1 activation in SW1417 colon cancer cells
J. Clin. Lab. Anal.
1995
, vol. 
9
 (pg. 
297
-
301
)
41
Gustavson
M. D.
Crawford
H. C.
Fingleton
B.
Matrisian
L. M.
Tcf binding sequence and position determines β-catenin and Lef-1 responsiveness of MMP-7 promoters
Mol. Carcinog.
2004
, vol. 
41
 (pg. 
125
-
139
)
42
Cheng
W. F.
Huang
C. Y.
Chang
M. C.
Hu
Y. H.
Chiang
Y. C.
Chen
Y. L.
Hsieh
C. Y.
Chen
C. A.
High mesothelin correlates with chemoresistance and poor survival in epithelial ovarian carcinoma
Br. J. Cancer
2009
, vol. 
100
 (pg. 
1144
-
1153
)
43
Hassan
R.
Cohen
S. J.
Phillips
M.
Pastan
I.
Sharon
E.
Kelly
R. J.
Schweizer
C.
Weil
S. C.
Laheru
D.
Phase I clinical trial of the chimeric anti-mesothelin monoclonal antibody MORAb-009 in patients with mesothelin expressing cancers
Clin. Cancer Res.
2010
, vol. 
16
 (pg. 
6132
-
6138
)
44
Bogre
L.
Meskiene
I.
Heberle-Bors
E.
Hirt
H.
Stressing the role of MAP kinases in mitogenic stimulation
Plant Mol. Biol.
2000
, vol. 
43
 (pg. 
705
-
718
)
45
Choi
B. H.
Kim
C. G.
Bae
Y. S.
Lim
Y.
Lee
Y. H.
Shin
S. Y.
p21 Waf1/Cip1 expression by curcumin in U-87MG human glioma cells: role of early growth response-1 expression
Cancer Res.
2008
, vol. 
68
 (pg. 
1369
-
1377
)
46
Uzgare
A. R.
Isaacs
J. T.
Enhanced redundancy in Akt and mitogen-activated protein kinase-induced survival of malignant versus normal prostate epithelial cells
Cancer Res.
2004
, vol. 
64
 (pg. 
6190
-
6199
)

Supplementary data