Objective: The purpose of this study was to identify genes that are highly differentially expressed in metastatic serous papillary ovarian tumors (MET) when compared with primary ovarian serous carcinomas (OSPC).
Study design: An oligonucleotide microarray with probe sets complementary to >14,500 human genes was used to determine whether patterns of gene expression may differentiate OSPC from MET in 31 snap-frozen serous papillary ovarian carcinomas (ie, 14 primary OSPC and 17 omental metastasis [MET]).
Results: Hierarchic cluster analysis of gene expression in OSPC and MET identified 156 genes that exhibited > 2-fold differences (P < .05) and that distinguished OSPC from MET. A number of invasion and metastasis predictive genes (including plasminogen activator; matrix metalloproteinase; matrix structural constituent genes encoding products with collagen, heparin, and hyaluronic acid binding activity; genes encoding receptors for insulin-like growth factors; vascular endothelial growth factor; endothelin type A; fibroblast growth factor; thrombospondin 1 and 2; type A and B integrins, and chemokines [stromal cell-derived factor 1 (CXCL12)]) were found among the 120 genes that were highly differentially overexpressed in MET, when compared with OSPC. Down-regulated genes in MET compared with OSPC included hepsin and testisin. Overexpression of CXCL12, matrix metalloproteinase, plasminogen activator, type A and B integrins, and hepsin genes was validated by quantitative real-time polymerase chain reaction in all samples. Finally, overexpression of CXCL12 in MET, when compared with OSPC, was validated at the protein level by immunohistochemistry.
Conclusion: Gene expression profiling may differentiate metastatic ovarian cancer from primary OSPC. The identification of metastasis-associated genes may provide a foundation for the development of new type-specific diagnostic strategies and treatment for metastatic ovarian cancer.