Transcriptome profile of OVCAR3 cisplatin-resistant ovarian cancer cell line

Background

The NIH:OVCAR-3 is a cisplatin refractory cell line established from malignant ascites of a patient with progressive adenocarcinoma of the ovary after combination chemotherapy with cyclophosphamide, Adriamycin, and cisplatin [1]. Thus, OVCAR3 serves as a model cell line for drug resistance in ovarian cancer. Here, we perform a comparative transcriptome analysis from the US National Cancer Institute human tumor cell line anticancer drug screen (NCI60) dataset [2]. Our results indicate a specific gene transcription profile of OVCAR3 genes relative to non-cancerous Human Ovarian Surface Epithelial cells (HOSE) and drug sensitive Serous Ovarian Cancer Epithelial Samples (CEPI) and SKOV3 cell lines. Pathway enrichment analysis from OVCAR3 unique transcripts was conducted using KEGG; Disease and Drug term enrichment used the PharmGKB [3] databases.

Materials and methods

Datasets from the NCI60 were obtained from the Gene Expression Omnibus (GEO) of NCBI [4] (OVCAR3 and SKOV3 from series GSE2003, OSE and CEPI from series GDS3592 and GSE14407). Transcriptome data analysis was conducted with Partek Genomics Suite version 6.6. The WEB-based GEne SeT AnaLysis Toolkit (WebGestalt) was used to perform enrichment analysis [5]. Genes present in KEGG pathway, PharmGKB Disease, and Drug terms enrichment sets were connected and expanded to one degree of biological interaction using the Michigan Molecular Interactions databases plugin [6] and visualized using Cytoscape version 2.8.3 [7].

Results

Transcriptome analysis of OVCAR3 specific gene expression changes resulted in 160 significant transcripts with a fold change > ±2 and an ANOVA derived Benjamini Hochberg adjusted p-value < 0.001. Enrichment analysis using a Hypergeometric test identified 189 PharmGKB Disease terms, 90 Drug terms and 31 KEGG pathways associated with these genes. A union of the disease, drug and KEGG pathway gene lists yielded 14 common genes for the dataset which were unique to OVCAR3 cells versus SKOV3 and CEPI (Table 1).

Table 1 Transcripts differentially expressed in OVCAR3 versus SKOV3 and CEPI having significant enrichment scores in KEGG Pathway, PharmGKB Drug and Disease databases.

Conclusions

This list of OVCAR3 unique genes, and the resulting interactions graph (Figure 1) represent potential pathways of drug resistance associated genes in ovarian cancer. Notably, ERBB2 (HER2) and FYN are the hub genes of the interaction network specific for OVCAR3 cell line. Thus, they may provide valuable insights into the drug resistance etiology of ovarian cancer. ERBB2 (HER2) has previously been reported to interact with Estrogen Receptor (ESR2) in Breast Cancer [8] and FYN has been implicated in Glioblastoma and T-cell Lymphomas [9, 10]; however their detailed roles in Ovarian Cancer have only recently been studied, warranting further investigation [11, 12].

Figure 1

Interaction network of OVCAR3 unique genes. Diamond and rectangle nodes are seed nodes of 14 OVCAR3 unique genes; circular nodes are 1 degree of biological interactions; rounded rectangular nodes are the highly connected hubs in the network (FYN and ERBB2).