Influence of monolayer, spheroid, and tumor growth conditions on chromosome 3 gene expression in tumorigenic epithelial ovarian cancer cell lines
- Neal AL Cody†1,
- Magdalena Zietarska†2,
- Ali Filali-Mouhim2,
- Diane M Provencher2, 3,
- Anne-Marie Mes-Masson2, 4 and
- Patricia N Tonin1, 5Email author
© Cody et al; licensee BioMed Central Ltd. 2008
Received: 19 December 2007
Accepted: 07 August 2008
Published: 07 August 2008
Expression microarray analyses of epithelial ovarian cancer (EOC) cell lines may be exploited to elucidate genetic and epigenetic events important in this disease. A possible variable is the influence of growth conditions on discerning candidates. The present study examined the influence of growth conditions on the expression of chromosome 3 genes in the tumorigenic EOC cell lines, OV-90, TOV-21G and TOV-112D using Affymetrix GeneChip® HG-U133A expression microarray analysis.
Chromosome 3 gene expression profiles (n = 1147 probe sets, representing 735 genes) were extracted from U133A expression microarray analyses of the EOC cell lines OV-90, TOV-21G and TOV-112D that were grown as monolayers, spheroids or nude mouse xenografts and monolayers derived from these tumors. Hierarchical cluster analysis was performed to compare chromosome 3 transcriptome patterns of each growth condition. Differentially expressed genes were identified and characterized by two-way comparative analyses of fold-differences in gene expression between monolayer cultures and each of the other growth conditions, and between the maximum and minimum values of expression of all growth conditions for each EOC cell line.
An overall high degree of similarity (> 90%) in gene expression was observed when expression values of alternative growth conditions were compared within each EOC cell line group. Two-way comparative analysis of each EOC cell line grown in an alternative condition relative to the monolayer culture showed that overall less than 15% of probe sets exhibited at least a 3-fold difference in expression profile. Less than 23% of probe sets exhibited greater than 3-fold differences in gene expression in comparisons of the maximum and minimum value of expression of all growth conditions within each EOC cell line group. The majority of these differences were less than 5-fold. There were 17 genes in common which were differentially expressed in all EOC cell lines. However, the patterns of expression of these genes were not necessarily the same for each growth condition when one cell line was compared with another.
The various alternative in vivo and in vitro growth conditions of tumorigenic EOC cell lines appeared to modestly influence the global chromosome 3 transcriptome supporting the notion that the in vitro cell line models are a viable option for testing gene candidates.
The molecular genetic analysis of ovarian cancer has been facilitated by the establishment and characterization of spontaneously immortalized epithelial ovarian cancer (EOC) cell lines that have been derived from malignant cells by long-term growth in cell culture . In our laboratory, we have studied the properties of three EOC cell lines derived from malignant ovarian tumors (TOV-21G and TOV112D) and ascites (OV-90) [2, 3]. These EOC cell lines were derived from patient samples prior to chemotherapy. They have been extensively characterized and shown to exhibit many of molecular genetic features, cytogenetic anomalies, and somatic mutations in tumor suppressor genes frequently associated with malignant ovarian cancers [2–4]. An attractive feature of these EOC cell lines is that they develop tumors at subcutaneous and intraperitoneal sites in nude mouse xenograft models . The phenotypes of the EOC cell lines are also reflected in global analyses of gene expression using large-scale gene expression microarrays analyses where the differentially expressed genes have been shown to overlap with those observed independently in the molecular analyses of ovarian cancers [5–11]. The application of various growth conditions to capture the full spectrum of the disease along with large-scale gene expression analyses could be important in our understanding of the biological and genetic factors that influence the phenotypic characteristics of the disease [1, 12].
A possible variable in the application of EOC cell line models is the influence of growth conditions on discerning and then characterizing gene candidates which initially exhibit differential gene expression in in vitro EOC cell line models. Recently, our group has reported on global differences in gene expression between EOC cell lines that were cultured as monolayers, spheroids, or nude mouse xenografts suggesting that microenviroment could impact the transcriptome . To further assess the variability of gene expression of EOC cell lines propagated in different contexts, we have extracted chromosome 3 gene expression profiles from the Affymetrix expression microarray data from three tumorigenic EOC cell lines, TOV-21G, TOV-112D and OV-90, that have been propagated as monolayers, spheroids or nude mouse xenografts, and monolayers derived from these xenografts . We have focused our analysis on the chromosome 3 gene expression because of our interest in elucidating genes located on this chromosome in ovarian cancer and the use of these well established EOC cell lines as models to both identify and characterize chromosome 3 gene candidates potentially important in this disease [7, 8, 14, 15]. These EOC cell lines were derived from ovarian cancer samples from chemotherapy naïve patients and have been shown to exhibit unique karyotypic abnormalities . Both OV-90 and TOV-112D exhibit complex karyotypic anomalies consistent with those typically seen in the majority of EOCs, whereas TOV-21G exhibited an atypical diploid karyotype with trisomy 10 as the only gross abnormality [2, 16]. Karyotype analysis has demonstrated evidence of an unique chromosome 3 abnormality in OV-90 comprised of a chromosome 22 derived homogeneously staining region replacing the 3p arm but not affecting the 3q arm [2, 5]. In particular, OV-90 has emerged as an interesting in vitro model with the potential for identifying and testing chromosome 3 tumor suppressor genes because of extensive loss of heterozygosity of the 3p arm , and the recent demonstration of suppression of tumorigenicity in chromosome 3 fragment transfer experiments attributable to functional complementation of 3p genes . The EOC cell line TOV-21G has shown no evidence of chromosome 3 karyotypic abnormalities  but has demonstrated evidence of microsatellite instability consistent with mismatch repair anomalies .
The present study was focused on addressing the magnitude and extent of transcriptome modifications for different EOC cell line models that may be influenced by tumor microenvironment. As each cell line exhibits unique molecular genetic characteristics comparisons of chromosome 3 transcriptome modifications were made with respect to gene expression profiles generated with each growth condition within each experimental cell line model.
EOC cell lines
The EOC cell lines were derived from a stage III/grade 3 clear cell carcinoma (TOV-21G), a stage IIIc/grade 3 endometrioid carcinoma (TOV-112D), and from the ascites fluid of a stage IIIc/grade 3 adenocarcinoma (OV-90), all from chemotherapy naïve patients, as described . Cells were cultured in OSE medium consisting of 50:50 medium 199:105 (Sigma), with 2.5 μg/mL amphotericin B and 50 μg/mL gentamicin . Culture media was supplemented with 10% FBS.
Source of chromosome 3 expression profiles
Chromosome 3 gene expression profiles were extracted from normalized Affymetrix GeneChip® HG-U133A microarray analyses of the OV-90, TOV-21G and TOV-112D EOC cell lines that were each grown under different conditions as described previously , and will be made available at Gene Expression Omnibus http://www.ncbi.nlm.nih.gov/geo/. These conditions include monolayer cultures (L), spheroid cultures (S), nude mouse xenografts at subcutaneous (TSC) or intraperitoneal (TIP) sites, and monolayer cultures of subcutaneous (LSC) and intraperitoneal (LIP) xenografts, as described previously . Data normalization, which is intended to eliminate systematic biases when comparing expression values from independently derived GeneChip® experiments, was achieved from the raw expression data using the MAS5 software (Affymetrix Microarray Suite®) by multiplying the value for an individual probe set by 100 and dividing by the mean of the raw expression values for the given sample data set as described previously [5, 10, 17]. The software also generates a reliability score for each probe set, which reflects the level of non-specific binding. A high reliability score of Present (P call) represents minimal hybridization to the mismatch probe set and consistent hybridization across all matched probes, in contrast to a borderline score of Marginal (M call) or a score of Ambiguous (A call).
The normalized data set was then used to extract the expression profiles associated with probe sets representing chromosome 3 genes. Probe sets corresponding to chromosome 3 genes were identified using the Affymetrix NetAffx Batch Query tool http://affymetrix.com/analysis/index.affx and the UniGene Homo sapiens database, based on UniGene Build 198 http://www.ncbi.nlm.nih.gov/UniGene/. Additional mapping information was obtained from the University of California Santa Cruz (UCSC) Human Genome Browser database, March 2006 (NCBI Build 36.1) hg 18 assembly http://genome.ucsc.edu. Based on these databases, 1147 probe sets were identified that mapped to chromosome 3, representing 735 genes and ESTs. Chromosome 3 alignment of probe sets (represented genes) was determined based on the UCSC Human Genome Browser database, where 535 probe sets mapped to genes on the chromosome 3p arm and 612 probe sets mapped to genes on the chromosome 3q arm.
The normalized expression data sets were also rescaled to eliminate systematic biases due to low expression values. Low values with A-calls are considered to be technical noise, which may influence fold-difference comparisons and overestimate expression differences that result from high variability of low expression values. Probe sets containing A-calls may also reflect either absent expression or poorly designed probe sets . To reduce this technical noise, values below 15 were reassigned a threshold value of 15, based on the mean expression value of data with low reliability scores of the chromosome 3 extracted probe set data.
Hierarchical Cluster Analysis
Hierarchical cluster analysis was performed on normalized and rescaled gene expression data analyzed using Bioconductor, an open-source software library for the analyses of genomic data  based on R, a language and environment for statistical computing and graphics http://www.r-project.org. In order to determine the significance of the differential expression, modified t-tests were performed with Bioconductor's limma package, where p-values from the resulting comparison were adjusted for multiple testing as described . This method controls for the false discovery rate, which was set to 0.05. Bioconductor's genefilter package was used to filter out probe sets with insufficient variation in gene expression across all tested samples for the analysis of each EOC cell line data set. In the remaining expression values, a log base 2 scale of at least 0.5 for the interquartile range was required across all tested samples for each EOC cell line group as described. Hierarchical clustering analysis was performed with R's cluster package, using the Pearson correlation distance.
Two-way comparative analyses
Two-way comparative analyses based on fold differences of expression values were performed on normalized and rescaled gene expression data derived from each EOC cell line. The expression values with at least one high-reliability score or P call for each EOC cell line sample set (data containing expression values generated from each growth condition) were evaluated in two-way comparative analyses. Differentially expressed genes were defined as those which exhibited at least a 3-fold difference in two-way comparative analyses with expression value for monolayer culture and each growth condition, or between the maximum and minimum value of expression observed within a set of data for each EOC cell line.
Hierarchical cluster analysis
Two-way comparisons relative to the reference monolayer culture
Two-way comparisons of gene expression values of any alternative growth condition compared with the monolayer cultures.
EOC cell line
Chromosomal location of probe sets
Number of probe sets analyzed
Number (%) of probe sets exhibiting > 3-fold differences in gene expression values in two-way comparisons
Two-way comparative analysis of the range of gene expression
Two-way comparisons of the maximum and minimum value of expression exhibited by a probe set of any alternative growth condition
Number (%) of probe sets exhibiting > 3-fold differences in gene expression values in two-way comparisons
EOC cell line
Number of probe sets analyzed
There were 17 genes which were found differentially expressed greater than 3-fold in all EOC cell lines (Additional file 1). These genes may represent those that could be affected by growth condition or tumor microenvironment . Notable is that the patterns of expression of these 17 genes were not necessarily the same for each growth condition when one cell line is compared with another. For example in OV-90, the maximum value of expression of RIS1 was found with the monolayer culture (L) and both subcutaneous (TSC) and intraperitoneal (TIP) xenografts exhibited the lowest values of expression of this gene (Figure 5), whereas the highest level of expression of RIS1 in TOV-112D was found with the subcutaneous (TSC) xenograft sample and the lowest value was observed with intraperitoneal (TIP) xenograft (Figure 9).
In this study, we described chromosome 3 transcriptome changes for three well characterized EOC cell lines (OV-90, TOV-112D, and TOV-21G) that each responded differently in relation to various growth conditions such as in three dimensional spheroid culture and nude mouse xenograft models, relative to the conventional monolayer culture. However, the alternative in vitro and in vivo growth conditions of tumorigenic EOC cell lines appeared to have modestly influenced the expression of chromosome 3 genes. This was reflected in the hierarchical cluster analysis where there was an overall high degree of correlation (> 90%) in gene expression in each EOC cell line group tested irrespective of growth condition. It was also reflected in the two-way comparative analyses where a 3-fold cut-off was applied. Although we have previously shown that replicates of Affymetrix GeneChip® expression data derived from the EOC cell lines grown as monolayer cultures were highly reproducible [5, 6], a lower cut-off (such as a 2-fold cut-off) would have captured differences in gene expression attributable to experimental variability [5, 17]. Unlike earlier studies using Affymetrix GeneChip® expression microarrays of the EOC cell lines, we have used a lower threshold level of 15 rather than 50 or 100 depending on the GeneChip® used [5, 6, 8–10]. A lower threshold value would increase the number of differentially expressed genes occurring in the low range of gene expression and this perhaps explains the large number of differentially expressed genes with values falling below 150 for all growth conditions (see Figures 5 to 10). The two-way comparison analyses were consistent with hierarchical cluster analyses which indicated a high correlation in gene expression patterns in the EOC cell line regardless of growth condition. Our results with chromosome 3 genes were consistent with whole genome transcriptome analyses of the EOC cell lines which also showed a high correlation (> 85%) of gene expression regardless of growth condition suggesting that microenvironment modestly influenced gene expression .
The EOC cell line lines exhibited unique patterns of gene expression as shown by the hierarchical cluster analysis. These unique differences are also reflected in a previous global analyses of gene expression from the entire Affymetrix U133A microarray . Thus while gene expression profiles of OV-90 cell line grown as tumors or spheroid clustered together, which may indicate that gene expression patterns could be associated with growth as 3D structures, this was not the case in chromosome 3 transcriptome profiles for TOV-21G and TOV-112D. The differences in the clustering patterns and differentially expressed genes observed in the three EOC cell lines was not surprising. The EOC cell lines were derived from long-term passages of tumor tissues representing different histopathological subtypes of ovarian cancer . These EOC cell lines also differ in their molecular genetic characteristics, in that OV-90 and TOV-112D harbor somatic mutations in TP53, whereas TOV-21G harbors a somatic mutation in KRAS and exhibits microsatellite instability. OV-90 also is monoallelic for the 3p arm, however this gross genomic anomaly did not significantly impact on global patterns of gene expression of the chromosome 3p arm as assayed by Affymetrix expression analyses of this cell line and the other EOC cell line used in the present study [3, 8, 15]. The EOC cell lines also differ in their response to ionizing radiation and chemotherapeutic agents . Thus the unique patterns of gene expression as shown in Figures 5 to 10 could in part reflect molecular genetic differences of the these cell lines.
Given the molecular genetic differences in the EOC cell lines, it is not surprising that there were few similarly differentially expressed genes found in common with all of EOC cell lines. Indeed there were only 17 genes in common in all three EOC cell lines which exhibited differential expression greater than 3-fold in all comparative analyses (Additional file 1). A review of gene ontology suggests that some examples of the differentially expressed genes have been associated with cellular shape (ARPC4 and NCK1), cell growth and division (CDC25A and SKIL), and extracellular signaling/cell-cell junctions (ROBO1, SKIL, TM4SF1 and WNT5A) (Additional file 1). Some of these genes have recently been identified as differentially expressed in ovarian cancer samples relative to normal tissue. For example, TMEM158, HEG1, PLOD2 and ATP13A3, were recently found differentially expressed greater than 3-fold in a comparative analysis of primary cultures of normal ovarian surface epithelial cells and malignant serous ovarian tumors . However, the 17 differentially expressed genes observed in common with all three EOC cell lines do not necessarily exhibit that the same differences in gene expression patterns relative to monolayer cultures, suggesting that they each may have responded differently to alternative growth conditions. Further analysis is required to determine if these 17 genes are indeed responding to differences in microenvironment as consequence of growth alternative growth conditions.
Future experiments are required to determine if the differences observed in the EOC cell lines grown in alternative conditions are biologically relevant or a reflection of experimental design. The magnitude of the differences in gene expression observed in the EOC cell lines grown under the various in vitro and in vivo growth conditions may all still be significantly different when each is compared with normal cells . The EOC cell lines, with their capacity to grow in different contexts, provides an opportunity to further examine the biological relevance of transcriptional differences that may be influenced by the microenvironment wherein which they are propagated. Recently our group has applied such a strategy to specifically identify genes transcriptionally modified based on microenvironment, and one such gene, S100A6, was found differentially expressed relative to culture conditions and further validated by RT-PCR and immunohistochemistry . While this finding may be disconcerting and discourage the use of in vitro model systems for studying gene candidates, our results in the present study show that a high correlation of gene expression in the transcriptomes generated from ovarian cancer cell lines propagated in different contexts. Overall these results attests to the validity of the EOC cell lines as an in vitro model for studying gene candidates but point out that some genes may be influenced by microenvironment, a factor that should be taken into consideration when investigating the molecular biology of specific genes. As our EOC cell lines are amenable to propagation in alternative growth conditions one could assay and further investigate the magnitude of transcriptional effects for specific candidate genes of interest and their consequences at the protein level to further understand the biological relevance gene expression differences associated with microenvironment.
The ability to culture tumorigenic EOC cells under different in vivo and in vitro growth conditions affords the opportunity to study gene expression of candidates in contexts that more closely mimic tumor growth in vitro. However, the analyses of chromosome 3 transcriptomes are highly comparable within each EOC cell line context. These observations would argue that gene expression studies using monolayer cultures of ovarian cancer lines is still a viable option for initial studies involving the characterization of gene expression pattern of candidate genes.
NAL Cody is a recipient of a studentship awards from the Research Institute of McGill University Health Centre and the Department of Human Genetics at McGill University. M Zietarska is a recipient of studentship awards from the Fondation Marc Bourgie de l'Institute du cancer de Montréal and the Faculté des études supérieures de l'Universitié de Montreal. This research was supported by grants from the Canadian Institutes of Health Research to PN Tonin, A-M Mes-Masson and D Provencher, and from the Réseau Cancer: Axe banque de tissus et de données du Fonds de recherche en Santé du Québec (FRSQ) and Valorisation Recheche Québec to A-M Mes-Masson, D Provencher and PN Tonin. PN Tonin is a medical scientist at the Research Institute of the McGill University Health Centre, which receives support from the FRSQ.
- Garson K, Shaw TJ, Clark KV, Yao DS, Vanderhyden BC: Models of ovarian cancer – are we there yet?. Mol Cell Endocrinol. 2005, 239 (1–2): 15-26. 10.1016/j.mce.2005.03.019.View ArticlePubMedGoogle Scholar
- Provencher DM, Lounis H, Champoux L, Tetrault M, Manderson EN, Wang JC, Eydoux P, Savoie R, Tonin PN, Mes-Masson AM: Characterization of four novel epithelial ovarian cancer cell lines. In Vitro Cell Dev Biol Anim. 2000, 36 (6): 357-361. 10.1290/1071-2690(2000)036<0357:COFNEO>2.0.CO;2.View ArticlePubMedGoogle Scholar
- Samouelian V, Maugard CM, Jolicoeur M, Bertrand R, Arcand SL, Tonin PN, Provencher DM, Mes-Masson AM: Chemosensitivity and radiosensitivity profiles of four new human epithelial ovarian cancer cell lines exhibiting genetic alterations in BRCA2, TGFbeta-RII, KRAS2, TP53 and/or CDNK2A. Cancer Chemother Pharmacol. 2004, 54 (6): 497-504. 10.1007/s00280-004-0843-9.View ArticlePubMedGoogle Scholar
- Manderson EN, Mes-Masson AM, Provencher D, Tonin PN: Mutations in a 10-bp polyadenine repeat of transforming growth factor beta-receptor type II gene is an infrequent event in human epithelial ovarian cancer. Clin Genet. 2000, 57 (2): 151-153. 10.1034/j.1399-0004.2000.570211.x.View ArticlePubMedGoogle Scholar
- Arcand SL, Mes-Masson AM, Provencher D, Hudson TJ, Tonin PN: Gene expression microarray analysis and genome databases facilitate the characterization of a chromosome 22 derived homogeneously staining region. Mol Carcinog. 2004, 41 (1): 17-38. 10.1002/mc.20038.View ArticlePubMedGoogle Scholar
- Benoit MH, Hudson TJ, Maire G, Squire JA, Arcand SL, Provencher D, Mes-Masson AM, Tonin PN: Global analysis of chromosome X gene expression in primary cultures of normal ovarian surface epithelial cells and epithelial ovarian cancer cell lines. Int J Oncol. 2007, 30 (1): 5-17.PubMedGoogle Scholar
- Birch AH, Quinn MC, Filali-Mouhim A, Provencher DM, Mes-Masson AM, Tonin PN: Transcriptome analysis of serous ovarian cancers identifies differentially expressed chromosome 3 genes. Mol Carcinog. 2008, 47 (1): 56-65. 10.1002/mc.20361.View ArticlePubMedGoogle Scholar
- Manderson EN, Mes-Masson AM, Novak J, Lee PD, Provencher D, Hudson TJ, Tonin PN: Expression profiles of 290 ESTs mapped to chromosome 3 in human epithelial ovarian cancer cell lines using DNA expression oligonucleotide microarrays. Genome Res. 2002, 12 (1): 112-121. 10.1101/gr.174202.View ArticlePubMedPubMed CentralGoogle Scholar
- Presneau N, Dewar K, Forgetta V, Provencher D, Mes-Masson AM, Tonin PN: Loss of heterozygosity and transcriptome analyses of a 1.2 Mb candidate ovarian cancer tumor suppressor locus region at 17q25.1-q25.2. Mol Carcinog. 2005, 43 (3): 141-154. 10.1002/mc.20096.View ArticlePubMedGoogle Scholar
- Presneau N, Mes-Masson AM, Ge B, Provencher D, Hudson TJ, Tonin PN: Patterns of expression of chromosome 17 genes in primary cultures of normal ovarian surface epithelia and epithelial ovarian cancer cell lines. Oncogene. 2003, 22 (10): 1568-1579. 10.1038/sj.onc.1206219.View ArticlePubMedGoogle Scholar
- Tonin PN, Hudson TJ, Rodier F, Bossolasco M, Lee PD, Novak J, Manderson EN, Provencher D, Mes-Masson AM: Microarray analysis of gene expression mirrors the biology of an ovarian cancer model. Oncogene. 2001, 20 (45): 6617-6626. 10.1038/sj.onc.1204804.View ArticlePubMedGoogle Scholar
- Le Page C, Provencher D, Maugard CM, Ouellet V, Mes-Masson AM: Signature of a silent killer: expression profiling in epithelial ovarian cancer. Expert Rev Mol Diagn. 2004, 4 (2): 157-167. 10.1586/1473718.104.22.168.View ArticlePubMedGoogle Scholar
- Zietarska M, Maugard CM, Filali-Mouhim A, Alam-Fahmy M, Tonin PN, Provencher DM, Mes-Masson AM: Molecular description of a 3D in vitro model for the study of epithelial ovarian cancer (EOC). Mol Carcinog. 2007, 46 (10): 872-885. 10.1002/mc.20315.View ArticlePubMedGoogle Scholar
- Cody NA, Ouellet V, Manderson EN, Quinn MC, Filali-Mouhim A, Tellis P, Zietarska M, Provencher DM, Mes-Masson AM, Chevrette M, et al: Transfer of chromosome 3 fragments suppresses tumorigenicity of an ovarian cancer cell line monoallelic for chromosome 3p. Oncogene. 2007, 26 (4): 618-632. 10.1038/sj.onc.1209821.View ArticlePubMedGoogle Scholar
- Lounis H, Mes-Masson AM, Dion F, Bradley WE, Seymour RJ, Provencher D, Tonin PN: Mapping of chromosome 3p deletions in human epithelial ovarian tumors. Oncogene. 1998, 17 (18): 2359-2365. 10.1038/sj.onc.1202152.View ArticlePubMedGoogle Scholar
- Wang J, Mes-Masson AM, Tonin PN, Provencher D, Eydoux P: Trisomy of chromosome 10 in two cases of ovarian carcinoma. Cancer Genet Cytogenet. 2000, 118 (1): 65-68. 10.1016/S0165-4608(99)00170-3.View ArticlePubMedGoogle Scholar
- Novak JP, Sladek R, Hudson TJ: Characterization of variability in large-scale gene expression data: implications for study design. Genomics. 2002, 79 (1): 104-113. 10.1006/geno.2001.6675.View ArticlePubMedGoogle Scholar
- Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Dudoit S, Ellis B, Gautier L, Ge Y, Gentry J, et al: Bioconductor: open software development for computational biology and bioinformatics. Genome Biol. 2004, 5 (10): R80-10.1186/gb-2004-5-10-r80.View ArticlePubMedPubMed CentralGoogle Scholar
- Benjamini Y, Yekutieli D: Quantitative trait Loci analysis using the false discovery rate. Genetics. 2005, 171 (2): 783-790. 10.1534/genetics.104.036699.View ArticlePubMedPubMed CentralGoogle Scholar
- Kashuba VI, Li J, Wang F, Senchenko VN, Protopopov A, Malyukova A, Kutsenko AS, Kadyrova E, Zabarovska VI, Muravenko OV, et al: RBSP3 (HYA22) is a tumor suppressor gene implicated in major epithelial malignancies. Proc Natl Acad Sci USA. 2004, 101 (14): 4906-4911. 10.1073/pnas.0401238101.View ArticlePubMedPubMed CentralGoogle Scholar
- Shu J, Jelinek J, Chang H, Shen L, Qin T, Chung W, Oki Y, Issa JP: Silencing of bidirectional promoters by DNA methylation in tumorigenesis. Cancer Res. 2006, 66 (10): 5077-5084. 10.1158/0008-5472.CAN-05-2629.View ArticlePubMedGoogle Scholar
- Kristiansen G, Pilarsky C, Wissmann C, Stephan C, Weissbach L, Loy V, Loening S, Dietel M, Rosenthal A: ALCAM/CD166 is up-regulated in low-grade prostate cancer and progressively lost in high-grade lesions. Prostate. 2003, 54 (1): 34-43. 10.1002/pros.10161.View ArticlePubMedGoogle Scholar
- van Kempen LC, Oord van den JJ, van Muijen GN, Weidle UH, Bloemers HP, Swart GW: Activated leukocyte cell adhesion molecule/CD166, a marker of tumor progression in primary malignant melanoma of the skin. Am J Pathol. 2000, 156 (3): 769-774.View ArticlePubMedPubMed CentralGoogle Scholar
- Oh JJ, Grosshans DR, Wong SG, Slamon DJ: Identification of differentially expressed genes associated with HER-2/neu overexpression in human breast cancer cells. Nucleic Acids Res. 1999, 27 (20): 4008-4017. 10.1093/nar/27.20.4008.View ArticlePubMedPubMed CentralGoogle Scholar
- Jia HL, Ye QH, Qin LX, Budhu A, Forgues M, Chen Y, Liu YK, Sun HC, Wang L, Lu HZ, et al: Gene expression profiling reveals potential biomarkers of human hepatocellular carcinoma. Clin Cancer Res. 2007, 13 (4): 1133-1139. 10.1158/1078-0432.CCR-06-1025.View ArticlePubMedGoogle Scholar
- Cavatorta AL, Fumero G, Chouhy D, Aguirre R, Nocito AL, Giri AA, Banks L, Gardiol D: Differential expression of the human homologue of drosophila discs large oncosuppressor in histologic samples from human papillomavirus-associated lesions as a marker for progression to malignancy. Int J Cancer. 2004, 111 (3): 373-380. 10.1002/ijc.20275.View ArticlePubMedGoogle Scholar
- Fuja TJ, Lin F, Osann KE, Bryant PJ: Somatic mutations and altered expression of the candidate tumor suppressors CSNK1 epsilon, DLG1, and EDD/hHYD in mammary ductal carcinoma. Cancer Res. 2004, 64 (3): 942-951. 10.1158/0008-5472.CAN-03-2100.View ArticlePubMedGoogle Scholar
- Gardiol D, Zacchi A, Petrera F, Stanta G, Banks L: Human discs large and scrib are localized at the same regions in colon mucosa and changes in their expression patterns are correlated with loss of tissue architecture during malignant progression. Int J Cancer. 2006, 119 (6): 1285-1290. 10.1002/ijc.21982.View ArticlePubMedGoogle Scholar
- Watson RA, Rollason TP, Reynolds GM, Murray PG, Banks L, Roberts S: Changes in expression of the human homologue of the Drosophila discs large tumour suppressor protein in high-grade premalignant cervical neoplasias. Carcinogenesis. 2002, 23 (11): 1791-1796. 10.1093/carcin/23.11.1791.View ArticlePubMedGoogle Scholar
- The pre-publication history for this paper can be accessed here:http://www.biomedcentral.com/1755-8794/1/34/prepub
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.