Gene expression profile analysis identifies metastasis and chemoresistance-associated genes in epithelial ovarian carcinoma cells

Abstract

The purpose of this study was to identify genes that associated with higher ability of metastasis and chemotherapic resistance in epithelial ovarian carcinoma (EOC) cells. An oligonucleotide microarray with probe sets complementary to 41,000(+) unique human genes and transcripts was used to determine whether gene expression profile may differentiate three epithelial ovarian cell lines (RMG-I-C, COC1 and HO8910) from their sub-lines (RMG-I-H, COCI/DDP and HO8910/PM) with higher ability of metastasis and chemotherapic resistance. Quantitative real-time PCR and immunohistochemical staining validated the microarray results. Hierarchic cluster analysis of gene expression identified 49 genes that exhibited ≥2.0-fold change and P value ≤0.05. Highly differential expression of GCET2, NLRP4, FOXP1 and SNX29 genes was validated by quantitative PCR in all cell line samples. Finally, FOXP1 was validated at the protein level by immunohistochemistry in paraffin embedded ovarian tissues (i.e., for metastasis, 15 primary EOC and 10 omental metastasis [OM]; for chemoresistance, 13 sensitive and 13 resistant EOC). The identification of higher ability of metastasis and chemotherapic resistance-associated genes may provide a foundation for the development of new type-specific diagnostic strategies and treatment for metastasis and chemotherapic resistance in epithelial ovarian cancer.

Fig. 1

Representative scatter plot of changes in gene expression levels. Scatter plot is a visualization that is useful for assessing the variation (or reproducibility) between chips. All detected probe point values on the chip were plotted. The central diagonal lines were used to classify gene expression levels into three groups: group I, >twofold change increase in gene expression; group II, gene expression levels within a twofold change; and group III, >twofold change decrease in gene expression

Fig. 2

Hierarchical clustering map of DEGs. The result of hierarchical clustering on conditions shows a distinguishable gene expression profiling among samples

Fig. 3

Volcano plot of DEGs. The vertical lines correspond to twofold up and down, respectively, and the horizontal line represents a P value of 0.05. So the red point in the plot represents the differentially expressed genes with statistical significance

Fig. 4

Quantitative real-time PCR validation for 4 selected genes. Quantitative real-time PCR for selected genes (GCET2, NLRP4, FOXP1 and SNX29) found to be differentially expressed in gene microarrays. The relative expression of GCET2 and CFTR was significantly higher in RMG-I-H, COC1/DDP, HO8910/PM than RMG-I-C, COC1, HO8910, respectively. The relative expression of FOXP1 and GARS was significantly lower in RMG-I-H, COC1/DDP, HO8910/PM than RMG-I-C, COC1, HO8910, respectively. (P < 0.05, one-way ANOVA)

Fig. 5

Immunohistochemical staining for FOXP1. Representative immunohistochemical staining for FOXP1. Left panel chemotherapeutic sensitive sample shows a positive nuclear staining for FOXP1. Right panel chemotherapeutic resistant sample displays a negative nuclear staining for FOXP1

Fig. 6

Kaplan–Meier survival analysis of chemotherapic ovarian cancer patients. Kaplan–Meier survival analysis shows that the positive nuclear staining of FOXP1 is an independent risk factor in ovarian cancer patients and strongly correlates with good prognosis

Fig. 7

Interaction network of the differentially expressed gene. Genes with more links are shown in bigger size. Proteins shown in red are encoded by up-regulated genes, while those in green are encoded by down-regulated genes, the gray represents the predicted genes. Arrow line represents definite control relationship, dotted line represents predicted control relationship, solid line represents inhibition