Identification of novel variant, 1484delG in the 3'UTR of H3F3B, a member of the histone 3B replacement family, in ovarian tumors

Previous studies have implicated the chromosomal region at 17q25 as harboring tumor suppressor genes based on the frequent loss of heterozygosity (LOH) observed in epithelial ovarian cancers (EOC). RT-PCR validation of Affymetrix GeneChip expression of H3F3B, a member of the 3B histone family that maps to 17q25.1, revealed a doublet band in cDNA from one of four EOC cell lines, OV90. In contrast to three other EOC cell lines (TOV81D, TOV112D and TOV21G) and primary cultures derived from normal ovarian surface epithelial cells (NOSE), sequence analysis of the cDNA revealed a deletion of G at position 1484 of the transcribed sequence which is located within the 3'UTR of H3F3B. OV90 was derived from ascites fluid of an undifferentiated adenocarcinoma of ovarian origin. The variant allele was identified in 1 of 65 (2%) healthy women with no prior history of cancer and in 5 participants with ovarian tumors comprising of 4 of 79 (5%) malignant EOC, none of 10 low malignancy potential tumors, and 1 of 8 (13%) benign tumors. All carriers of the variant alleles were heterozygous and tumor samples did not exhibit preferential LOH of the normal allele. The variant allele was identified in EOC samples of clear cell (1 of 20), mucinous (1 of 8), mixed cell (1 of 3) and undifferentiated (1 of 2) histopathological subtypes but none of 34 serous or 12 endometrioid subtype tumors. One of 3 mucinous benign tumors also harbored the variant allele. The functional significance of the variant is unknown, however its presence in rare subtypes of ovarian epithelial tumors warrants further investigation.     

Loss of the inactive X chromosome and replication of the active X in BRCA1-defective and wild-type breast cancer cells

In females, X chromosome inactivation (XCI) begins with the expression of the XIST gene from the X chromosome destined to be inactivated (Xi) and the coating of XIST RNA in cis. It has recently been reported that this process is supported by the product of the BRCA1 tumor suppressor gene and that BRCA1-/- cancers show Xi chromatin structure defects, thus suggesting a role of XCI perturbation in BRCA1-mediated tumorigenesis. Using a combined genetic and epigenetic approach, we verified the occurrence of XCI in BRCA1-/- and BRCA1wt breast cancer cell lines. It was ascertained that the Xi was lost in all cancer cell lines, irrespective of the BRCA1 status and that more than one active X (Xa) was present. In addition, no epigenetic silencing of genes normally subjected to XCI was observed. We also evaluated XIST expression and found that XIST may be occasionally transcribed also from Xa. Moreover, in one of the BRCA1wt cell line the restoring of XIST expression using a histone deacetylase inhibitor, did not lead to XCI. To verify these findings in primary tumors, chromosome X behavior was investigated in a few BRCA1-associated and BRCA1-not associated primary noncultured breast carcinomas and the results mirrored those obtained in cancer cell lines. Our findings indicate that the lack of XCI may be a frequent phenomenon in breast tumorigenesis, which occurs independently of BRCA1 status and XIST expression and is due to the loss of Xi and replication of Xa and not to the reactivation of the native Xi.     

Patterns of expression of chromosome 17 genes in primary cultures of normal ovarian surface epithelia and epithelial ovarian cancer cell lines

Oligonucleotide microarray analysis was applied to assess the expression profile of 332 probe sets representing 308 genes or expressed sequence tags (ESTs) that map to chromosome 17 in order to address epigenetic events that result in alterations in gene expression in epithelial ovarian cancer (EOC). Expression profiles were generated from 12 primary cultures derived from normal ovarian surface epithelium (NOSE) and four long-term cultures (TOV-81D, TOV-112D, TOV-21G and OV-90) derived from EOCs that have been previously characterized and shown to mimic features of the tumoral cells from which they were derived. The expression values of all 332 probe sets is highly correlated across the 12 NOSEs (89% correlation coefficients >0.90). In two-way comparisons, differential patterns of gene expression were observed for 157 probe sets for which the expression value of at least one EOC cell line fell outside the limits of the range of expression of the 12 NOSEs. When compared to NOSEs, four genes displayed similar differential patterns of gene expression across all four EOC cell lines, and 26 genes displayed similar differential patterns of gene expression across the three EOC cell lines (TOV-112D, TOV-21G and OV-90) representing tumoral cells derived from the most aggressive disease. A total of 17 genes displayed differential patterns of gene expression greater than threefold in at least one EOC cell line in comparison to NOSE, and three genes were differentially expressed greater than threefold across all aggressive cell lines. The analysis of a selected number of genes by RT-PCR revealed patterns of gene expression comparable to those observed by microarray analysis in the majority of samples tested. Comparison of expression profiles of differentially expressed genes identified by microarray analysis in two-way comparisons of the EOC cell lines and the NOSEs with published reports revealed 10 genes previously implicated in ovarian tumorigenesis and 18 in tumorigenesis of other types of cancer. The differential pattern of gene expression of genes that map to both the p and q arm of chromosome 17 is consistent with the hypothesis that a number of genes that map to this chromosome are implicated in the etiology of ovarian cancer.     

Chromosome 17q25 genes, RHBDF2 and CYGB, in ovarian cancer

It has been proposed that the frequent loss of heterozygosity (LOH) of an entire chromosome 17 contig in epithelial ovarian cancers (EOC) is the consequence of the inactivation of multiple tumour suppressor genes on this chromosome. We report the characterization of a 453 Kb 17q25 locus shown previously to exhibit a high frequency of LOH in EOC samples. LOH analysis further defined the minimal region of deletion to a 65 Kb interval flanked by D17S2239 and D17S2244, which contains RHBDF2, CYGB and PRCD as tumour suppressor gene candidates. Tissue specific expression excluded PRCD as a candidate. RHBDF2 was expressed at low levels in the majority of benign and low malignant potential (LMP) tumours, and in a subset of malignant ovarian tumour samples, as compared with primary cultures of normal ovarian surface epithelial cell (NOSE) samples. CYGB was expressed at low levels in the majority of LMP and malignant samples compared with benign and NOSE samples. In contrast to CYGB expression, RHBDF2 was expressed at low or undetectable levels in EOC cell lines exhibiting tumourigenic characteristics and up-regulated in a genetically modified EOC cell line rendered non-tumourigenic. DNA sequence analysis identified variants but no apparent deleterious mutations in either gene. Methylation-specific PCR analysis suggested that promoter methylation of CYGB but not RHBDF2 occurred in 6 of 31 malignant samples. The results combined suggest that RHBDF2 and CYGB may play distinctive roles in ovarian cancer and could be added to the growing roster of chromosome 17 genes implicated in this disease.     

Microarray analysis of gene expression mirrors the biology of an ovarian cancer model

We have previously described an ovarian cancer model based on four independent spontaneously immortalized epithelial ovarian cancer cell lines (TOV-21G, TOV-81D, TOV-112D and OV-90) from patients who were never exposed to chemotherapy or radiation therapy. These cell lines are particularly interesting since they retain characteristics of the original epithelial ovarian cancers (EOC) from which they were derived. Here we report the characterization of this model system using high-density DNA microarrays in order to assess gene expression. Expression profiles were generated from total RNAs extracted from the four EOC cell lines. For comparison, expression profiling is also provided for a primary culture of normal ovarian surface epithelium (NOV-31) and a fresh EOC sample (TOV-578G). Comparison of expression profiles revealed patterns of expression that distinguish NOV-31 from that of all tumor derived samples. The expression pattern of TOV-81D, an EOC cell line that was derived from a patient with indolent disease, most closely resembles NOV-31 while profiles of samples derived from patients with more aggressive disease (TOV-21G, OV-90, TOV-112D and TOV-578G) showed more divergent patterns of expression. The microarray analysis (http://genome.mcgill.ca) results confirm the usefulness of an ovarian cancer model based on the characterization of these EOC cell lines.     

A statistical analysis of chromosome 11 and 17 loss of heterozygosity in epithelial ovarian cancer

Many regions of the genome exhibit loss of heterozygosity (LOH) in epithelial ovarian cancer (EOC) suggesting sites of recessive genetic elements such as tumor suppressor genes. We performed detailed LOH studies of chromosomes 17 and 11 using 24 microsatellite repeat markers in a population of 47 patients with EOC. Univariate statistical analysis revealed that significant co-losses of chromosomal loci occurred between 17p and 17q whole arms (p=0.0003), NME1 (17q21) with D11S922 (11p15.5) (p=0.0067) and D11S912 (11q24) with D11S935 (11p13) (p=0.0073). Statistical analysis of the relationship between LOH on particular chromosomal arms and clinicopathological factors revealed a significant association between serous histological subtype of ovarian adenocarcinoma and chromosome 17p (p=0.0052) and telomeric 17q (p=0.0007) LOH. An analysis of specific polymorphic chromosomal loci demonstrated that adverse survival was significantly associated with LOH at 11q24 (p=0.0067) and 17q21 (p=0.0076). There were nonsignificant trends suggesting a relationship between chromosome 17p LOH and poorly differentiated (p=0.025) and advanced FIGO stage (p=0.031) tumours. Considering these statistical associations, a preliminary multistep model for involvement of chromosomes 11 and 17 in ovarian neoplasia can be constructed.     

Chromosomes 6 and 18 induce neoplastic suppression in epithelial ovarian cancer cells

Metaphase comparative genomic hybridisation (CGH) studies indicate that chromosomes 4, 5, 6, 13, 14, 15 and 18 are frequently deleted in primary ovarian cancers (OCs). Therefore we used microcell‐mediated chromosome transfer (MMCT) to establish the functional effects of transferring normal copies of these chromosomes into 2 epithelial OC cell lines (TOV112D and TOV21G). The in vitro neoplastic phenotype (measured as anchorage dependent and independent growth and invasion) was compared between recipient OC cell lines and multiple MMCT hybrids. Chromosomes 6 and 18 showed strong evidence of functional, neoplastic suppression for multiple hybrids in both cell lines. We also found evidence in 1 cancer cell line suggesting that chromosomes 4, 13 and 14 may also cause functional suppression. Array CGH and microsatellite analyses were used to characterise the extent of genomic transfer in chromosome 6 and 18 hybrids. A 36 MB deletion on chromosome 6 in 2 hybrids from 1 cell line mapped the candidate region proximal to 6q15 and distal to 6q22.2; and an ～10 MB candidate region spanning the centromere on chromosome 18 was identified in 2 hybrids from the other cell line. These data support reported functional effects of chromosome 6 in OC cell lines; but to our knowledge, this is the first time that functional suppression for chromosome 18 has been reported. This suggests that these chromosomes may harbour tumour suppressor‐“like” genes. The future identification of these genes may have a significant impact on the understanding and treatment of the disease and the identification of novel therapeutic targets. © 2008 Wiley‐Liss, Inc.     

Molecular description of a 3D in vitro model for the study of epithelial ovarian cancer (EOC)

Epithelial ovarian cancer (EOC) cell lines are useful tools for the molecular and biological characterization of ovarian cancer. The use of an in vitro multidimensional (3-D) culture model recapitulates some of the growth conditions encountered by tumor cells in vivo. Here we describe a molecular comparison of spheroid based 3D EOC models versus monolayer cultures and xenografts using cell lines from malignant ovarian tumors (TOV-21G and TOV-112D) and ascites (OV-90) previously established and characterized in our laboratory. Gene expression analyses of the three models were performed using the Affymetrix HG-U133A high density DNA array. Cluster analysis identified a set of genes that stratified expression profiles from the EOC cell lines grown as spheroids and xenografts from that of monolayer cultures. The gene expression analysis results were validated by Q-PCR analyses on an independent set of RNAs. Differential expression observed for the S100A6 gene between the monolayer, spheroid cultures and xenografts was confirmed at the protein level by immunohistochemistry. The analysis was extended to various ovarian tumor tissues using an EOC tissue array. This result represents an example of a gene that, if studied in vitro, is more representative of the in vivo disease in a 3D model rather than the monolayer culture. Identification of genes in spheroid models that mimic the in vivo tumor gene expression patterns may allow a better understanding of the community effect observed in human disease that is determined by direct or indirect interactions of cells with their environment or other surrounding cells.     

Gene expression microarray analysis and genome databases facilitate the characterization of a chromosome 22 derived homogeneously staining region

Karyotype and fluorescence in situ hybridization (FISH) analyses previously identified a homogeneously staining region (HSR) derived from chromosome 22 in OV90, an epithelial ovarian cancer (EOC) cell line. Affymetrix expression microarrays in combination with the UniGene and Human Genome Browser databases were used to identify the candidate genes comprising the amplicon of the HSR, based on comparison of expression profiles of OV90, EOC cell lines lacking HSRs and primary cultures of normal ovarian surface epithelial (NOSE) cells. A group of probe sets displaying a minimum 3-fold overexpression with a high reliability score (P-call) in OV90 were identified which represented genes that mapped within a 1-2 Mb interval on chromosome 22. A large number of probe sets, some of which represent the same genes, displayed no evidence of overexpression and/or low reliability scores (A-call). An investigation of the probe set sequences with the Affymetrix and Sanger Institute Chromosome 22 Group databases revealed that some of the probe sets displaying discordant results for the same gene were complementary to intronic sequences and/or the antisense strand. Microarray results were validated by RT-PCR. Genomic analysis suggests that the HSR was derived from the amplification of a 1.1 Mb interval defined by the chromosomal map positions of ZNF74 and Hs.372662, at 22q11.21. The deduced amplicon is derived from a complex region of chromosome 22 that harbors low-copy repeats (LCRs). The amplicon contains 18 genes as likely targets for gene amplification. This study illustrates that large-scale expression microarray analysis in combination with genome databases is sufficient for deducing target genes associated with amplicons and stresses the importance of investigating probe set design before engaging in validation studies.     

Cyclin D1, p53, and p21Waf1/Cip1 expression is predictive of poor clinical outcome in serous epithelial ovarian cancer.

PURPOSE: Dysregulation of cell cycle control, in particular G(1)-S-phase transition, is implicated in the pathogenesis of most human cancers, including epithelial ovarian cancer (EOC). However, the prognostic significance of aberrant cell cycle gene expression in EOC remains unclear. EXPERIMENTAL DESIGN: The expression of selected genes from the pRb pathway that regulates G(1)-S-phase progression, including cyclin D1, p16(Ink4a), cyclin E, p27(Kip1), p21(Waf1/Cip1), and p53, was examined in a consecutive series of 134 serous EOC using immunohistochemistry and the results correlated to disease outcome. RESULTS: Molecular markers predictive of reduced overall survival in univariate analysis were overexpression of cyclin D1 (P = 0.03) and p53 (P = 0.03) and reduced expression of p27(Kip1) (P = 0.05) and p21(Waf1/Cip1) (P = 0.02), with the latter three also being prognostic for a shorter progression-free interval. In addition, patients displaying overexpression of p53 with concurrent loss of p21(Waf1/Cip1) had a significantly shorter overall (P = 0.0008) and progression-free survival (P = 0.0001). On multivariate analysis, overexpression of cyclin D1 and combined loss of p21(Waf1/Cip1) in the presence of p53 overexpression were independent predictors of overall survival. Similarly, the combination of p21(Waf1/Cip1) loss and p53 overexpression was independently predictive of a shorter progression-free interval. Overexpression of p53 and cyclin E and reduced expression of p27(Kip1) and p21(Waf1/Cip1) were significantly associated with increasing tumor grade. CONCLUSIONS: This study confirms that dysregulation of cell cycle genes is common in EOC, and that aberrant expression of critical cell cycle regulatory proteins can predict patient outcome in serous EOC.     

The role of S100A14 in epithelial ovarian tumors

S100A14 is an EF-hand calcium-binding protein that has been reported to be involved in the progression of many malignancies. However, its role in ovarian cancer has not yet been clarified. In this study, we investigated the significance of S100A14 expression in epithelial ovarian cancers (EOCs) as well as it''s mechanism of action. On both RNA and protein levels, S100A14 was overexpressed in transformed cells. Immunohistochemical staining demonstrated that S100A14 expression was associated with advanced stage (P < 0.001) and poor tumor grade (P < 0.001). Moreover, S100A14 overexpression was an independent prognostic factor for overall survival (HR = 4.53, P = 0.029). We also investigated S100A14''s functional role by employing lentiviral-mediated overexpression and knockdown in EOC cells. S100A14 overexpression promoted cell proliferation, tumorigenesis, migration, and invasion, whereas S100A14 knockdown inhibited these properties. TOV112D cells that overexpressed S100A14 also exhibited greater tumor growth potential in xenografted mice. S100A14 promoted such a malignant phenotype in EOC cells through the PI3K/Akt pathway. Taken together, our data indicate that S100A14 has a crucial role in EOC progression, and its overexpression is associated with poor prognosis. Further study of S100A14''s molecular mechanisms may lead to the development of a novel therapeutic target for ovarian cancer.     

正常卵巢上皮细胞和卵巢上皮癌细胞基因表达差异的研究

目的：探讨卵巢上皮癌细胞相关基因的差异表达。方法：采用舍384条肿瘤相关基因的cDNA阵谱检测了卵巢上皮癌细胞株SKOV-3及正常卵巢上皮细胞的基因谱，分析卵巢上皮癌细胞相关基因的差异表达。结果：在384条候选基因中，与卵巢癌相关的差异表达基因33条，其中22条表达上调，11条表达下调。结论：cDNA阵谱技术是筛查卵巢癌相关基因的有效方法。     

The disappearing Barr body in breast and ovarian cancers

Interest has recently reawakened in whether loss of the heterochromatic X chromosome (Barr body) is prevalent in certain breast and ovarian cancers, and new insights into the mechanisms involved have emerged. Mitotic segregation errors commonly explain the loss of the inactive X chromosome (Xi), but compromise of Xi heterochromatin in some cancers may signal broader deficits of nuclear heterochromatin. The debated link between BRCA1 and Xi might reflect a general relationship between BRCA1 and heterochromatin, which could connect BRCA1 to both epigenetic and genetic instability. We suggest that heterochromatic instability is a common but largely unexplored mechanism, leading to widespread genomic misregulation and the evolution of some cancers.     

Hedgehog signaling pathway regulates the growth of ovarian cancer spheroid forming cells

The hedgehog (Hh) pathway has been shown to be activated in numerous malignancies as well as in cancer stem cells. We sought to determine the importance of the Hh pathway in regulating growth and development of ovarian cancer spheroid-forming cells (SFCs). Ovarian cancer cell lines (ES2, TOV112D, OV90, and SKOV3) as well as a normal ovarian epithelial cell line (IOSE80) were grown in non-adherent growth conditions to form SFCs. Western blot analysis was used to determine the expression of Hh pathway proteins SMOH, PTCH, GLI1. SFCs were treated with Hh agonists (SHH and IHH) as well as an Hh inhibitor (cyclopamine) to determine changes in spheroid growth and survival. All ovarian cancer cell lines readily formed spheroids in non-adherent growth conditions while IOSE80 failed to form SFCs. Compared to IOSE80, ovarian cancer cell lines demonstrated significant activation of the Hh pathway as defined by increased expression of intranuclear GLI1. Both Hh agonists demonstrated significant increases in spheroid volume of at least 42-fold for SHH-treated cells and 46-fold for IHH-treated cells. With regard to survival, SFCs were 30-50% more resistant to cyclopamine than their corresponding monolayer cells. Despite this resistance, inhibition of the Hh pathway with cyclopamine prevented further growth of SFCs with a 10-, 5-, and 4-fold restriction in growth for ES2, SKOV3, and TOV112D, respectively. The hedgehog pathway appears to be important in regulating growth of ovarian cancer spheroid-forming cells. The activation and inhibition of this pathway demonstrates significant correlation to enhanced growth and growth restriction, respectively.     

Monoclonal antibody targeting MUC1 and increasing sensitivity to docetaxel as a novel strategy in treating human epithelial ovarian cancer

The purpose of this study was to investigate the in vitro effect of anti-MUC1 monoclonal antibody (MAb) C595 alone and in combination with docetaxel, on the growth and survival of different epithelial ovarian cancer (EOC) cell lines. MUC1 expression was assessed on EOC cell lines (OVCAR-3, IGROV-1, A2780, CAOV-3, TOV-21G, TOV-112D, SKOV-3 and OV-90) using immunofluorescence labeling and flow cytometry. The effect of MAb C595 alone or in combination with docetaxel on the cell lines was studied by proliferation, colony and TUNEL assays. Our results indicate that all primary and metastatic EOC cell lines tested were positive to MAb C595 (MUC1); MAb C595 inhibited EOC cell proliferation in a MUC1- and dose-dependent manner; low-dose MAb C595 (1/2 of IC₅₀) combined with docetaxel greatly improved efficiency of cell killing in EOC cells and induced apoptosis; the additive effect of MAb C595 was further confirmed in colony forming assays; and cell death following single or combined treatments was associated with the release of cytochrome c and increased caspase-3 activity. These results suggest that MAb C595 used either alone, or combined with docetaxel, is an attractive strategy for targeting human EOC.