PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of transoncolGuide for AuthorsAbout this journalExplore this journalTranslational Oncology
 
Transl Oncol. 2010 August; 3(4): 230–238.
Published online 2010 August 1.
PMCID: PMC2915414

Effect of Ovarian Cancer Ascites on Cell Migration and Gene Expression in an Epithelial Ovarian Cancer In Vitro Model1

Abstract

A third of patients with epithelial ovarian cancer (EOC) present ascites. The cellular fraction of ascites often consists of EOC cells, lymphocytes, and mesothelial cells, whereas the acellular fraction contains cytokines and angiogenic factors. Clinically, the presence of ascites correlates with intraperitoneal and retroperitoneal tumor spread. We have used OV-90, a tumorigenic EOC cell line derived from the malignant ascites of a chemonaive ovarian cancer patient, as a model to assess the effect of ascites on migration potential using an in vitro wound-healing assay. A recent report of an invasion assay described the effect of ascites on the invasion potential of the OV-90 cell line. Ascites sampled from 31 ovarian cancer patients were tested and compared with either 5% fetal bovine serum or no serum for their nonstimulatory or stimulatory effect on the migration potential of the OV-90 cell line. A supervised analysis of data generated by the Affymetrix HG-U133A GeneChip identified differentially expressed genes from OV-90 cells exposed to ascites that had either a nonstimulatory or a stimulatory effect on migration. Ten genes (IRS2, CTSD, NRAS, MLXIP, HMGCR, LAMP1, ETS2, NID1, SMARCD1, and CD44) were upregulated in OV-90 cells exposed to ascites, allowing a nonstimulatory effect on cell migration. These findings were validated by quantitative polymerase chain reaction. In addition, the gene expression of IRS2 and MLXIP each correlated with prognosis when their expression was assessed in an independent set of primary cultures established from ovarian ascites. This study revealed novel candidates that may play a role in ovarian cancer cell migration.

Introduction

Ovarian cancer is the fifth cause of cancer-related deaths in woman and the most lethal of all gynecological cancers. Largely asymptomatic, more than 70% of patients with ovarian cancer have already reached an advanced stage of disease at initial diagnosis [1,2], and the overall 5-year survival rate for these patients is less than 30% [3]. Intraperitoneal dissemination is common, and malignant cells can implant anywhere in the peritoneal cavity but are more likely to implant in sites of stasis along the peritoneal fluid circulation [4]. Ascites, a voluminous exudative fluid with a cellular fraction consisting of ovarian cancer cells, lymphocytes, and mesothelial cells, is present in more than one-third of ovarian cancer patients. The acellular fraction is known to harbor cytokines and angiogenic factors [5–9]. Clinically, the presence of ascites correlates with intraperitoneal and retroperitoneal tumor spread, suggesting that it may facilitate metastasis [10].

Previously, we used an in vitro invasion assay to monitor the effect of ascites on the potential of chemonaive epithelial ovarian cancer (EOC) cell lines to degrade and migrate across a Matrigel-based barrier [11]. We used the OV-90 cell line to perform our assays. This cell line is derived from an ovarian malignant ascites, is able to form tumors in a xenograft model, and has been extensively characterized at both the cellular and the molecular levels [11–13]. Ascites from more than 50 patients were tested and compared with either 5% fetal bovine serum (FBS) or no serum in an invasion assay. The ascites were classified into one of two categories for their effect on invasion on the basis of comparison to FBS-treated cells: stimulatory (≥5% FBS activity) or nonstimulatory (<20% of FBS activity). We focused on gene expression profiles generated from the OV-90 cell line [12] treated with ascites possessing either stimulatory or nonstimulatory invasive potential. A supervised analysis of gene expression microarray data sets identified differentially expressed genes, which were validated by quantitative polymerase chain reaction (Q-PCR) assays on the basis of OV-90 cells exposed to a large number of ascites from different patients. In a previous study [11], the proliferation rates and the capacity to form three-dimensional spheroids in hanging drop cultures of the OV-90 cell line treated with different ascites were also described. The results from this previous study strongly supported the notion that ascites affect the cellular and molecular behaviors of ovarian cancer cells.

In the present study, we have further assessed the effect of the same ascites samples on the migration potential of OV-90 cells using an in vitro wound-healing assay and extended our analysis to include a larger number of ascites samples derived from ovarian cancer patients. We also assessed differential gene expression using OV-90 cells treated with no serum, with 5% of ascites, or with 5% FBS that correlated with the cellular migration potential of this cell line. Some candidate genes identified in our analysis were further validated using Q-PCR, and their association with survival was tested in an independent set of primary cultures derived from ovarian cancer patients with ascites.

Materials and Methods

Cell Culture Conditions, Material, and Patients

OV-90 cells were maintained in ovarian surface epithelium (OSE) complete medium (cat. no. 316-030-CL; Wisent, Quebec, Canada) supplemented with 10% FBS, 2.5 µg/ml of amphotericin B (Wisent), and 50 µg/ml of gentamicin (Invitrogen, Ontario, Canada) at 37°C [14]. Ascites were collected at the time of clinical intervention from ovarian cancer patients at the Centre hospitalier de l'Université de Montréal (Montreal, Quebec, Canada). Informed consent for this study was obtained from all of the patients. After centrifuging at 1250g for 5 minutes, the acellular fractions of ascites were stored at -20°C and tested within 6 months of collection. Histopathological diagnosis, grade, and stage of ovarian tumor samples were assigned according to the criteria of the International Federation of Gynecology and Obstetrics. Of the 31 ascites samples, one third were from patients diagnosed with papillary serous adenocarcinoma and most presented as stage IIIC grade 3 tumor diseases (Table 1). A small proportion (4/31) of patients had already received chemotherapy before ascites collection. The cohort of patients with ovarian cancer and the accompanying ascites used for this research and included in the survival analysis have been described previously [11].

Table 1
Compiled Clinical Characteristics of Patients from Which Ascites Were Obtained.

In Vitro Migration Assays

Cellular migration was assayed by determining the ability of cells to migrate in a culture plate using a wound-healing assay. To evaluate migration assays, OV-90 cells were plated in 12-well dishes and were grown at 37°C until confluent. Cell monolayers were scraped using a sterile 200-µl yellow plastic tip to produce small wounds of similar size. Wounded monolayers were then washed with phosphate-buffered saline to remove cell debris, and OSE medium was added with no serum or with either 5% FBS or 5% ascites. For inactivation assays, ascites were heated for 10 minutes at 100°C to denature the proteins. Cells were incubated at 37°C under 5% CO2 for different lengths of time to evaluate their migration (0, 6, 24, 30, 48, and 54 hours after scratch formation). At the different time points, cells were methanol-fixed and treated with Giemsa stain (Sigma-Aldrich, Inc, St Louis, MO). Digital images were obtained at each time point of the experiment. Images were analyzed, and wound closures were quantified using Image Pro Plus software (Version 5.1; MediaCybernetics, Bethesda, MD) and Microsoft Excel. All experiments were performed twice using triplicate samples and were normalized to FBS-treated cultures.

RNA Preparation

Total RNA was extracted using TRIzol reagent (Gibco/BRL, Life Technologies, Inc, Grand Island, NY) according to the manufacturer's protocol. RNA was extracted from OV-90 cells grown to 80% confluence in 100-mm Petri dishes. The quality of RNA was assessed using a 2100 Bioanalyzer with the RNA 6000 Nano LabChip kit (Agilent Technologies, Mississauga, Ontario, Canada) according to the manufacturer's protocol.

Microarray hybridization experiments were performed at McGill University and the Genome Quebec Innovation Center (Montreal, Quebec, Canada) using the HG-U133A GeneChip arrays. This chip allows the analysis of approximately 18,400 transcripts and variants, including 14,500 well-characterized human genes, composed of more than 22,000 probe sets. Protocols are available at the Affymetrix Web site (www.affymetrix.com; Affymetrix, Santa Clara, CA). Methods for labeling and hybridization of RNA were previously described [11].

Gene Expression Statistical Analysis

Gene expression profiles were analyzed using R (version 2.4.0; www.r-project.org), a statistical programming language, Bioconductor [15], and an open-source software library for the analyses of genomic data, which is based on R. Background subtraction, normalization (quantile normalization), and expression value calculations were performed using the justGCrma function available as part of the Bioconductor's gcrma package. Bioconductor's gene filter package was used to filter genes with insufficient variation in expression across all samples tested. Expression values retained after this filtering process had intensities greater than 100 units in at least two samples and a log base 2 scale of at least 0.2 for the interquartile range across all tested samples. Differentially expressed genes were identified using Bioconductor's limma package that implements moderate t tests by fitting a linear model for each group of samples and using empirical Bayes method to moderate SEs of the estimated log-fold changes between the predefined groups.

Kaplan-Meier survival plots, univariate Cox proportional hazard regressions, as well as log-rank tests were performed to determine the significance of using gene expression levels to predict survival of EOC patients as described earlier [11]. The expression threshold cutoff was determined by survival tree using the recursive partitioning and regression tree (RPART) method [16]. The survival analysis and tree building were performed using R's survival and RPART packages, respectively. The Pearson correlation coefficient test (two-tailed) was used to calculate the correlation between gene expression and migration rate and was performed with SPSS software 11.0 (SPSS, Inc, Chicago, IL).

Quantitative Reverse Transcription-PCR Validation

The complementary DNA synthesis was prepared using the QuantiTech Reverse Transcription for two-step reverse transcription-PCR (Qiagen, Inc, Mississauga, Ontario, Canada) according to the manufacturer's instruction. First-strand synthesis for reverse transcription-PCR was performed with 1 µg of RNA and a mix of Oligo dT and random hexamers. Samples were diluted 1:10 in water before Q-PCR.

Q-PCR was performed using Rotor-gene 3000 (Corbett Research, Montreal Biotech, Inc, Montreal, Quebec, Canada). The Quantitect SYBR Green PCR (Qiagen, Inc) reaction mixture was used to label 5 µl of sample complementary DNA and 10 pg of the different primers in a final volume of 25 µl. Reactions were performed at least twice according to the manufacturer's instructions. Serial dilutions were performed to generate a standard curve for each gene tested to define the efficiency of the Q-PCR, and a melt curve was constructed to confirm reaction specificity. The analytical method of Pfaffl [17] was used to measure the relative quantity of gene expression. The first sample (with 5% FBS) served as the reference sample in each experiment. Gene expression was evaluated in the OV-90 cell line with no serum or with either 5% FBS or 5% ascites, under the same conditions used to evaluate the migration potential of the OV-90 cell line. β-Actin was used as reference gene on the basis of its stable expression in all samples by microarray analysis. For every marker, a Pearson correlation was calculated between the scored migration result (1 < 100% and 2 ≥ 100% of migration) and the scored gene expression (1 < median and 2 ≥ median). All experiments were performed in duplicate.

Results

Effect of Ascites on OV-90 Migration Potential

Ascites from 31 EOC patients were studied to determine their effect on OV-90 cell migration in an in vitro scratch assay (Table 1 and Figure 1). Cells were grown in a monolayer until confluence in a 12-well plate. After scratching the monolayer with a pipet tip, cells were maintained in medium with or without 5% FBS. To test the effect of ascites, FBS was replaced by 5% ascites (acellular fraction) from EOC patients and sampled after 0, 6, 24, 30, 48, and 54 hours of incubation. OV-90 cells are able to seal the wound in 48 hours, similar to the average seen in comparable EOC cell lines [18]. The influence of each ascites sample on cell migration was scored in comparison to medium supplemented with 5% FBS (Figure 1). In OV-90 cells, a difference in cell migration was observed between cells in contact with medium alone and cells in medium supplemented with 5% FBS. When cells were incubated with the medium without FBS (Figure 1B), cell migration was reduced by 40% in comparison to 5% FBS. In contrast, a large number of ascites samples (n = 23) stimulated OV-90 cell migration at levels similar to 5% FBS (Figure 1B). A smaller number of ascites samples (n = 8) did not stimulate cell migration when compared with migration in the presence of FBS (Figure 1B).

Figure 1
Effect of ascites on OV-90 cell migration. Migration was assessed by determining the ability of cells to migrate in a culture plate using a wound-healing assay in the presence of ascites compared with 5% FBS (% migration) after 54 hours of incubation. ...

To determine whether the migration effect of ascites is protein component-dependant, two stimulatory and two nonstimulatory ascites were selected and heated for 10 minutes at 100°C before adding to the OV-90 cell cultures (Figure 2). The results suggest that protein inactivation abolished the stimulatory effect (FBS, A2647 and A2839). For the two nonstimulatory ascites, samples A2295(2) and A2090, we also observed a decreased effect on cell migration from ascites that had been heat-treated. The pH of FBS and the four ascites was not altered by heating (data not shown).

Figure 2
Effect of heat-treated ascites in OV-90 cell migration assays in vitro. Both stimulatory (A2647 and A2839) and nonstimulatory (A2295(2) and A2090) ascites were heated at 100°C for 10 minutes to inactivate the proteins before adding to OV-90 cell ...

Effect of Ascites on Gene Expression in OV-90 Cells

To identify potential molecular players in migration regulated by ascites, gene profiling analysis was performed. Total RNA was extracted from OV-90 cells after a 24-hour exposure to no serum, 5% FBS, or 5% of one of eight ascites sampled from ovarian cancer patients as previously described (Table 1) [11]. The RNA samples were each hybridized on Affymetrix HG-U133A GeneChip arrays, and gene expression profiles were analyzed using 6489 probe sets (see Materials and Methods). A supervised analysis was performed using expression data sets representing the following groups. The GSTIMUL group contained the ascites (A1946, A1835, A2085, A1337, and A1592) that stimulated cell migration and included the 5% FBS control. The GnSTIMUL group contained ascites (A1322, A1317, and A2090) that demonstrates less cell migration and included the no-FBS sample. This supervised analysis identified 129 genes that were differentially expressed (P < .05, t test) between the GSTIMUL and GnSTIMUL groups (Table 2).

Table 2
List of Differentially Expressed Genes among the GSTIMUL and GnSTIMUL Groups That Were Tested in Q-PCR.

Differential Expression Validation of Selected Candidates by Q-PCR

Forty gene candidates involved in the migration potential of OV-90 cells were selected for further validation by Q-PCR on the basis of their P values obtained in the previous microarray analysis and also on their gene functions. The candidates tested by Q-PCRare described in Table 2. Of the 40 selected genes, 21 were downregulated in the GnSTIMUL group (HIST1H2AC, F1P1L1, RBM10, H2BS, BTG3, CNTNAP2, IGF1R, DKC1, HIST1H2BD, CDCA4, MDC1, SMARCD3, TSTA3, SMARCD2, PDPK1, ADAMTS1, ASH2L, SMARCA4/BRG1, CMKOR1, ANAPC5, and GPR125) and 19 were upregulated in the GnSTIMUL group (HSPA1B, CALM3, IRS2, CBFB, CEBPA, LIPG, CRLF1, MDK, CTSD, NRAS, MLXIP, HMGCR, CALM1, LAMP1, MKRN1, ETS2, NID1, SMARCD1, and CD44). Q-PCR was performed on RNA derived from OV-90 cells exposed individually to the entire panel of 31 ascites (Table 1). The relative expression ratio of each candidate, based on the Pfaffl method, was quantified, and for each experiment, themedian ratiowas calculated, and the result was categorized as above or below the median expression. Pearson correlations were calculated for the migratory effect (stimulatory or nonstimulatory) of each ascites with a gene expression score, as shown in Table 3. Of the 40 candidates, 10 (IRS2, CTSD, NRAS, MLXIP, HMGCR, LAMP1, ETS2, NID1, SMARCD1, and CD44) tested by Q-PCR correlated significantly with the ascites migration effect. Three candidates (MDC1, SMARCA4, and GPR125) correlated significantly, but their downregulated expression pattern by Q-PCR was in the opposite direction expected from the microarray expression analysis.

Table 3
Correlation between Migration and Gene Expression.

Survival

In a previous study [11], we showed that the genes exhibiting a significant correlation with ascites invasion effect could be good prognosis predictors. The prognostic potential of the candidate genes that Q-PCR expression correlated significantly with the ascites migration effect was evaluated using microarray expression in 28 primary cultures derived from the cellular fraction of ascites of ovarian cancer patients. A survival tree was used to determine the expression cutoff that could lead to the identification of prognostic groups among patients. In the case of IRS2 and MLXIP, Kaplan-Meier curves coupled with a log-rank test identified a patient group with an overall good survival rate associated with a high expression (Figure 3). A trend toward significance was observed for the HMGCR candidate, although its association between gene expression and patient survival was not significant (P = .061; Figure 3).

Figure 3
Relationship between IRS2 and MLXIP expression and cumulative survival of patients with ovarian cancer in the context of concomitant ascites. The cutoffs were determined by survival tree. Kaplan-Meier graphical representation of the survival curves illustrates ...

Discussion

EOC is the second most common gynecological cancer and accounts for more than half of the deaths associated with gynecological pelvic malignancies [19,20]. This gynecological malignancy is associated with vague symptoms, which results in a diagnosis at a late stage [2]. The need for reliable biomarkers in ovarian cancer detection is increasing because EOC mortality has not significantly decreased during the past years because of a poor understanding of the biology [4]. Although ascites are commonly found in patients with EOC, its association with a poor prognostic factor is not universally accepted, and mechanisms that lead to ascites formation are not well characterized in EOC [10,21,22]. The presence of ascites correlates with intraperitoneal and retroperitoneal tumor spread, which supports a role in metastasis [2,10]. It is also known that ascites contain factors that increase vascular permeability [5]. In this study, we assessed the effects of ascites on OV-90 cell migration and correlated this effect with the alteration of gene expression that occurred in the same cell line as a consequence of exposure to several heterologous ascites obtained in different clinical settings (Table 1) [12].

In control tests, the presence of FBS in culture medium stimulated the cellular migration of the OV-90 cell line. An analysis of the migration behavior of OV-90 with 31 different ascites showed that more than two-thirds of the ascites samples stimulated cell migration in a fashion similar to the serum control. In contrast, one-third of the ascites did not stimulate migration in a fashion similar to the non-FBS control. Of the ascites samples collected from the same patient (A2295 and A2295(2), respectively) but at different times during the course of her treatment, it is interesting that the prechemotherapy ascites stimulated cell migration, whereas the postchemotherapy ascites inhibited cell migration. This observation raises the intriguing possibility that chemotherapy treatment is associated with a diminution of tumor cell migratory potential but requires validation with a larger sample set.

In a previous study, we determined how ascites affected the invasive capacity of OV-90 cells [11]. Because several of the ascites tested in that study were also tested in the present study, we looked for correlations between the migration and invasion results. Although some ascites were able to increase both the migratory and the invasive properties of OV-90 cells, no correlation was observed between ascites stimulating migration and invasion (data not shown), suggesting that these two events are activated by different stimuli. Both the stimulatory and nonstimulatory effects of selected ascites and FBS were lost when the samples were heated. With this treatment, the results were similar to those observed with the medium lacking FBS, suggesting that the effects of FBS and both stimulatory and nonstimulatory ascites were due to the presence of a protein or a protein component rather than other soluble factors.

Gene expression was assessed to link molecular events with the migration effect of ascites. For this purpose, RNA from OV-90 cells exposed to five stimulatory ascites or 5% FBS (GSTIMUL) were compared with RNA from cells in the absence of serum or treated with three nonstimulatory ascites (GnSTIMUL). Microarray analysis identified 129 genes differentially expressed between the group of ascites that stimulated cell migration and the group that did not stimulate cell migration. Among those genes, 40 were tested in Q-PCR. Pearson correlations were calculated for the migratory effect (stimulatory or nonstimulatory) of each ascites with a gene expression score, as shown in Table 3. Among theses 40 genes, differential expression of 13 candidates was confirmed by Q-PCR in a larger set of 31 ascites samples (Table 3). Three candidates correlated significantly, but the expression pattern was opposite to that expected by microarray analysis. Of 27 additional candidates tested by Q-PCR, no significant correlation between gene expression and OV-90 cell migration could be established. The fact that the differential expression of these genes was identified by microarray analysis on a limited number of ascites samples (only eight), although the validation was tested on 31 different ascites, could explain this poor validation rate. These results also suggest that an extensive validation in a supervised analysis by Q-PCR of gene candidates is essential to uncover the richness of genes implicated in migration process.

Many genes differentially expressed in cells stimulated by the two groups of ascites are closely related to the mitogen-activated protein kinase (MAPK) pathway and apoptosis. A similar association between MAPK-related genes and invasion was also observed in our recent study [11], although different gene candidates were identified. In the present study, the MAPK pathway-related genes include NRAS, ETS2, Cathepsin D (CSTD), and HMGCR. Interestingly, recent studies suggest that Cathepsin D could be responsible for the positive regulation of proliferation, survival, motility, and invasion of fibroblasts by triggering activation of ras/MAPK/Rds [23]. The Ras proteins were some of the first proteins identified involving the regulation of cell growth [24]. Nearly 30% of human cancers are associated with mutations in the Ras genes [25]. In ovarian cancer research, Ahmed et al. [26] showed that ascites enhanced the activation of Ras by increasing Ras-GTP levels in the study of four ovarian cancer cell lines. This study also presented evidence that activation of Ras and downstream Erk pathway is involved in maintaining growth, adhesion, and invasiveness of cancer cells. It was also recently shown that a high expression of HMG-CoA reductase (HMGCR) induced acceleration of the cholesterol synthesis pathway in cancer cells, and this may have promoted Ras isoprenylation, a posttranslational modification activating Ras [27]. In addition, the Ras-MAP kinase signaling pathway leads to the phosphorylation of ETS transcription factors, including Ets-2, which plays an important role in the regulation of growth and cell cycle-related genes [28] and protects cells from apoptosis [29].

The other candidate genes identified in this study have been linked to tumor growth and apoptosis. For example, CD44, a gene upregulated in the group of nonstimulatory ascites, mediates the interaction between ovarian carcinoma cells and the mesothelial cells lining abdominal organs [30]. Through strong affinity binding to the extracellular matrix in ovarian carcinoma cells, CD44 has been shown to affect cell adhesion [30] and migration [31] as well as to increase tumor growth [32]. In addition, increased CD44 expression is associated with an increased expression of Bcl-2, an antiapoptotic factor [33]. Mammary tumor cells that are deficient in Irs-2, another candidate, have been shown to be significantly more sensitive to apoptotic stimuli such as serum deprivation [29]. Conversely, BRG1/SMARCA4, a member of the SWI/SNF complex, which is downregulated in OV-90 cells treated with stimulatory ascites, is required for the growth arrest induction and cell senescence induced by p21 [34,35]. These results show that gene candidates that decrease the migratory potential of OV-90 cells are also involved in growth promotion and apoptotic protection of tumor cells, whereas genes upregulated by migratory ascites are involved in cell growth arrest. Taken together, this suggests that cell migration and growth may be sequential events during tumor progression where cells need to stop growth to be able to start migrating. Further experiments will be needed to confirm this hypothesis. In line with this is the fact that most gene candidates alone are not associated with survival of patients, which also means that the migratory potential of tumor cell is necessary but insufficient per se to affect the aggressive behavior of tumor cells.

In summary, this study provides evidence for some novel gene candidates and molecular pathways that may play an important role in ovarian cancer cell migration. Combined with our previous research [11], this work continues to define the importance of studying the effect of both ascites and the tumor environment on ovarian cancer cells. Our data also suggest that ascites may contain either positive or negative regulators of tumor behaviors and can play a role in future clinical outcomes. Future studies assessing the relative expression of these candidates in clinical specimens will no doubt help to better refine those that are important in ovarian cancer progression.

Acknowledgments

The authors thank Louise Champoux, Manon de Ladurantaye, and Lise Portelance for technical assistance; PierreDrouin, Philippe Sauthier, and Philippe Gauthier for providing specimens; and Luke Masson for reviewing the manuscript.

Footnotes

1This research was supported by The Cancer Research Society, Inc, Strategic Grant Program in Genomics and Proteomics of Metastatic Cancer award to A.-M.M.-M., P.N.T., M.C. and D.M.P. The ovarian tissue bank was supported by the Banque de tissus et de données of the Réseau de recherche sur le cancer of the Fonds de la Recherche en Santé du Québec, affiliated with the Canadian Tumor Repository Network. The authors have no financial disclosures or conflict of interest to declare.

References

1. Naora H, Montell D. Ovarian cancer metastasis: integrating insights from disparate model organisms. Nat Rev Cancer. 2005;5:355–366. [PubMed]
2. Brown MR, Blanchette JO, Kohn EC. Angiogenesis in ovarian cancer. Baillieres Best Pract Res Clin Obstet Gynaecol. 2000;14:301–918. [PubMed]
3. Cannistra SA. Cancer of the ovary. N Engl J Med. 2004;351:2519–2529. [PubMed]
4. Cvetkovic D. Early events in ovarian oncogenesis. Reprod Biol Endocrinol. 2003;1:68. [PMC free article] [PubMed]
5. Zebrowski B, Liu W, Ramirez K, Akagi Y, Mills G, Ellis L. Markedly elevated levels of vascular endothelial growth factor in malignant ascites. Ann Surg Oncol. 1999;6:373–378. [PubMed]
6. Miyamoto S, Hirata M, Yamazaki A, Kageyama T, Hasuwa H, Mizushima H, Tanaka Y, Yagi H, Sonoda K, Kai M. Heparin-binding EGF-like growth factor is a promising target for ovarian cancer therapy. Cancer Res. 2004;64:5720–5727. [PubMed]
7. Mills G, Eder A, Fang X, Hasegawa Y, Mao M, Lu Y, Tanyi J, Tabassam F, Wiener J, Lapushin R, et al. Critical role of lysophospholipids in the pathophysiology, diagnosis, and management of ovarian cancer. Cancer Treat Res. 2002;107:259–283. [PubMed]
8. Saltzman A, Hartenbach E, Carter J, Contreras D, Twiggs L, Carson L, Ramakrishnan S. Transforming growth factor-alpha levels in the serum and ascites of patients with advanced epithelial ovarian cancer. Gynecol Obstet Invest. 1999;47:200–204. [PubMed]
9. Abendstein B, Stadlmann S, Knabbe C, Buck M, Muller-Holzner E, Zeimet A, Marth C, Obrist P, Krugmann J, Offner F. Regulation of transforming growth factor-beta secretion by human peritoneal mesothelial and ovarian carcinoma cells. Cytokine. 2000;12:1115–1119. [PubMed]
10. Ayhan A, Gultekin M, Taskiran C, Dursun P, Firat P, Bozdag G, Celik NY, Yuce K. Ascites and epithelial ovarian cancers: a reappraisal with respect to different aspects. Int J Gynecol Cancer. 2007;17:68–75. [PubMed]
11. Puiffe M-L, Le Page C, Filali-Mouhim A, Zietarska M, Ouellet V, Tonin N, Chevrette M, Provencher D, Mes-Masson A. Characterization of ovarian cancer ascites on cell invasion, proliferation, spheroid formation, and gene expression in an in vitro model of epithelial ovarian cancer. Neoplasia. 2007;9:820–829. [PMC free article] [PubMed]
12. Provencher D, Lounis H, Champoux L, Tetrault M, Manderson E, Wang J, Eydoux P, Savoie R, Tonin P, Mes-Masson A. Characterization of four novel epithelial ovarian cancer cell lines. In Vitro Cell Dev Biol Anim. 2000;36:357–361. [PubMed]
13. Tonin N, Hudson T, Rodier F, Bossolasco M, Lee PD, Novak J, Manderson EN, Provencher D, Mes-Masson A. Microarray analysis of gene expression mirrors the biology of an ovarian cancer model. Oncogene. 2001;20:6617–6626. [PubMed]
14. Kruk P, Maines-Bandiera S, Auersperg N. A simplified method to culture human ovarian surface epithelium. Lab Invest. 1990;63:132–136. [PubMed]
15. Dudoit S, Gentleman R, Quackenbush J. Open source software for the analysis of microarray data. Biotechniques. 2003;(Suppl):45–51. [PubMed]
16. Therneau T. An Introduction to Recursive Partitioning Using the RPART Routine. Rochester, NY: Section of Biostatistics, Mayo Clinic; 1997. Technical Report.
17. Pfaffl M. A new mathematical model for relative quantification in realtime RT-PCR. Nucleic Acids Res. 2001;29:e45. [PMC free article] [PubMed]
18. Ouellet V, Zietarska M, Portelance L, Lafontaine J, Madore J, Puiffe M, Arcand S, Shen Z, Hébert J, Tonin P, et al. Characterization of three new serous epithelial ovarian cancer cell lines. BMC Cancer. 2008;8:152. [PMC free article] [PubMed]
19. Parkin DM, Pisani P, Ferlay J. Global cancer statistics. CA Cancer J Clin. 1999;49:33–64. [PubMed]
20. Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun MJ. Cancer statistics. CA Cancer J Clin. 2003;53:5–26. [PubMed]
21. Jandu N, Richardson M, Singh G, Hirte H, Hatton MWC. Human ovarian cancer ascites fluid contains a mixture of incompletely degraded soluble products of fibrin that collectively possess an antiangiogenic property. Int J Gynecol Cancer. 2006;16:1536–1544. [PubMed]
22. Shen-Gunther J, Mannel RS. Ascites as a predictor of ovarian malignancy. Gynecol Oncol. 2002;87:77–83. [PubMed]
23. Laurent-Matha V, Maruani-Herrmann S, Prébois C, Beaujouin M, Glondu M, Noël A, Alvarez-Gonzalez ML, Blacher S, Coopman P, Baghdiguian S, et al. Catalytically inactive human cathepsin D triggers fibroblast invasive growth. J Cell Biol. 2005;168:489–499. [PMC free article] [PubMed]
24. Downward J. Targeting Ras signalling pathways in cancer therapy. Nat Rev Cancer. 2002;3:11–22. [PubMed]
25. Parikh C, Subrahmanyam R, Ren R. Oncogenic NRAS, KRAS, and HRAS exhibit different leukemogenic potentials in mice. Cancer Res. 2007;67:7139–7146. [PMC free article] [PubMed]
26. Ahmed N, Riley C, Oliva K, Rice G, Quinn M. Ascites induces modulation of α6β1 integrin and urokinase plasminogen activator receptor expression and associated functions in ovarian carcinoma. Br J Cancer. 2005;92:1475–1485. [PMC free article] [PubMed]
27. Kaneko R, Tsuji N, Asanuma K, Tanabe H, Kobayashi D, Watanabe N. Survivin down-regulation plays a crucial role in 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitor-induced apoptosis in cancer. J Biol Chem. 2007;282:19273–19281. [PubMed]
28. Oikawa T. ETS transcription factors: possible targets for cancer therapy. Cancer Sci. 2004;95:626–633. [PubMed]
29. Nagle J, Ma Z, Byrne M, White M, Shaw L. Involvement of insulin receptor substrate 2 in mammary tumor metastasis. Mol Cell Biol. 2004;24:9726–9735. [PMC free article] [PubMed]
30. Lessan K, Aguiar DJ, Oegema T, Siebenson L, Skubitz A. CD44 and β1 integrin mediate ovarian carcinoma cell adhesion to peritoneal mesothelial cells. Am J Pathol. 1999;154:1525–1537. [PubMed]
31. Casey RC, Skubitz A. CD44 and β1 integrins mediate ovarian carcinoma cell migration toward extracellular matrix proteins. Clin Exp Metast. 2000;18:67–75. [PubMed]
32. Strobel T, Swanson L, Cannistra SA. In vivo inhibition of CD44 limits intra-abdominal spread of a human ovarian cancer xenograft in nude mice: a novel role for CD44 in the process of peritoneal implantation. Cancer Res. 1997;57:1228–1232. [PubMed]
33. Klingbeil P, Marhaba R, Jung T, Kirme R, Ludwig T, Zöller M. CD44 variant isoforms promote metastasis formation by a tumor cell-matrix cross-talk that supports adhesion and apoptosis resistance. Mol Cancer Res. 2009;7:168–179. [PubMed]
34. Hendricks K, Shanahan F, Lees E. Role for BRG1 in cell cycle control and tumor suppression. Mol Cell Biol. 2004;24:362–376. [PMC free article] [PubMed]
35. Kang H, Cui K, Zhao K. BRG1 controls the activity of the retinoblastoma protein via regulation of p21 (CIP1/WAF1/SDI) Mol Cell Biol. 2004;24:1188–1199. [PMC free article] [PubMed]

Articles from Translational Oncology are provided here courtesy of Neoplasia Press