::  A N C H A   L A B  ::
Expression profile of ovarian tumors: distinct signature of Sertoli-Leydig cell tumor
A. Baranova, S.Gowder, S.Naouar, S.King, K.Schlauch, M.Jarrar, A.Van Meter, Y.Ding, F. Gorreta, Dr. Cook, V.Chandhoke, A.Christensen

This study was published in: Int J Gynecol Cancer. 2006 Nov-Dec;16(6):1963-72.

Motivation: Epithelial ovarian cancer comprises the majority of malignant ovarian tumors in adult women and its high mortality is attributed mostly to late diagnosis.

Methods: In this study we reveal distinct expression signatures of previously uncharacterized ovarian carcinoma subtypes, including endometrioid component of mixed ovarian tumor and Sertoli-Leydig tumor. Both subtypes were compared to the most common and well characterized ovarian epithelial carcinoma of the serous type. These comparisons were performed by cDNA microarrays allowing high fidelity measurements of the expression levels of 39,360 human individual cDNA species representing both known and unknown human genes.

Results: A subset of the 146 mRNAs that corresponds to 85 functional genes associated with UniGene descriptions is listed in Table 1. We have estimated the level of congruency between expression levels of the genes represented in Table 1 in the three types of ovarian tumors. Pearson's correlation coefficients computed for gene expression ratios in the log scale also demonstrates much higher congruency for serous and endometrioid ovarian tumors (0.843), than for serous and Sertoli-Leydig (0.711) or endometrioid and Sertoli-Leydig (0.633). In addition, we retrieved all cDNA probes differentially expressed between Sertoli-Leydig ovarian tumors and normal ovaries, and combined the two lists of differentially expressed genes to yield a gene list representing only probes independently revealed by both comparisons. Correlation coefficients computed for gene expression ratios in the log scale showed exceptional congruency between these gene lists (0.95), thus supporting the validity of the chosen approach. The resulting list of 907 genes representing Sertoli-Leydig ovarian tumor signature is given in Supplementary Table 1.

Validation: To validate our microarray technology, we obtained quantitative real-time PCR measurements of the mRNA level of 8 genes: namely CIAO1, FBLN5, KRT8, MET, PLXNB1, POFUT2, SBNO1 and SHC1. When normal ovaries were compared with the Sertoli-Leydig sample, in all but one case (FBLN5), the fold differences measured by real-time PCR were in correspondence with the microarray fold difference values. In the case of FBLN5, contradictory results were obtained.

Support: This research was supported by grants ‘‘Cancer Genomics and Development of Diagnostic Tools and Therapies'' (Commonwealth Technology Research Fund). A.B. was partially covered by NIH 1R15CA113331-01.

Contact: abaranov@gmu.edu