Novel Computational Biology Technologies to Identify Ovarian Cancer Biomarkers that Predict Survival
Posted October 1, 2007
Igor Jurisica, Ph.D., University Health Network, Toronto, Ontario

Dr. Igor Jurisica of the University Health Network, Toronto, developed novel computational biology tools. I2D (Interologous Interaction Database; is a repository-based system for protein-protein interactions, which is essential during the analysis of existing and emerging data sets on altered gene and protein expression patterns in ovarian and other cancers. OPHID incorporates data from several sources: existing databases, high-throughput interaction screens, manual and text mining extraction of interaction from over 17 million published articles in PubMed, and computational predictions across organisms using orthologous mapping. I2D covers six main organisms (human, rat, mouse, fly, worm, and yeast), comprising over half a million interactions, making it significantly larger than any other single existing database. A companion system, NAViGaTor (Network Analysis, Visualization and Graphing system; was developed by these researchers to create visual renderings of the interactions. These tools are freely available to the academic community.

After developing these tools, Dr. Jurisica in collaboration with Dr. Kevin Brown from Samuel Lunenfeld Research Institute used them to compare gene expression patterns in normal and cancerous ovarian cell cultures. Cell cultures were derived from tissue samples from patients undergoing surgery for prophylactic oophorectomy, uterine tumors, or ovarian tumors. The researchers found 17 genes that were differentially regulated by androgens (hormones associated with prostate cancer) in ovarian cancer cells as compared to controls. Two of these genes, Basic leucine zipper transcription factor 2 (BACH2) and acetylcholinesterase (ACHE), were upregulated by androgen in ovarian surface epithelial cells derived from women with germline BRCA1 or BRCA2 mutations who underwent prophylactic oophorectomy. Dr. Jurisica's group further investigated BACH2 and ACHE using an ovarian cancer tissue microarray from a separate set of 149 clinical samples. Both cytoplasmic ACHE and BACH2 levels were significantly increased in ovarian cancer relative to benign cases. High levels of cytoplasmic ACHE staining correlated with reduced survival, whereas nuclear localization of BACH2 correlated with shorter times to disease recurrence. These results suggest that ACHE may be a new biomarker for predicting survival, while BACH2 may predict risk of disease recurrence.


Brown K and Jurisica I. Unequal evolutionary conservation of human protein interactions in interologous networks. Genome Biology 8(5), 2007.

Motamed-Khorasani A, Jurisica I, Letarte M, Shaw PA, Parkes RK, Zhang X, Evangelou A, Rosen B, Murphy KJ, and Brown TJ. 2007. Differentially androgen-modulated genes in ovarian epithelial cells from BRCA mutation carriers and control patients predict ovarian cancer survival and disease progression. Oncogene 26(2):198-214.

Kotlyar M and Jurisica I. 2006. Predicting protein-protein interactions by association mining. Information Systems Frontiers 8:37-47.

Otasek D, Brown K, and Jurisica I. 2006. Confirming protein-protein interactions by text mining. SIAM Conference on Text Mining, Bethesda, Maryland.

Jurisica I, Wigle DA, and Wong B. 2007. Cancer informatics in the post-genomics era. In: Implementing Information-Based Medicine, Series: Cancer Treatment and Research, Volume 137. Springer Verlag.


Public and Technical Abstracts: Integrated Computational Biology Approach to Marker Selection for Early Detection and Treatment of Epithelial Ovarian Cancer

Top of Page

New Animal Model to Study Ovarian Cancer
Posted September 21, 2007
Rong Wu, M.D. and Kathleen Cho, M.D., University of Michigan, Ann Arbor, Michigan

Dr. Rong Wu and Dr. Kathleen Cho of the University of Michigan are studying ovarian endometrioid adenocarcinoma (OEA), the second most common of the four major types of epithelial ovarian cancer, which include serous, clear cell, and mucinous. The researchers hypothesized that OEA has unique genetic alterations not typically observed in the other ovarian cancer subtypes. With funding from a Fiscal Year 2003 New Investigator Award, Dr. Wu, Dr. Cho, and colleagues showed that mutations predicted to deregulate the canonical Wnt signaling pathway (via mutant CTNNB1/ß-catenin) and the PI3K/Pten/Akt signaling pathway (via mutant PIK3CA and/or PTEN) often occurred together in human OEAs, suggesting that these pathways cooperate during OEA pathogenesis. OEAs with these gene mutations were usually low-grade, low-stage tumors. TP53 and K-RAS did not show cooperativity with the Wnt and PI3K/Pten/Akt pathways.

Based on these findings, the researchers generated a unique mouse model of OEA derived from deregulation of Wnt and PI3K/Pten/Akt signaling specifically in the murine ovarian surface epithelium. The mice developed tumors with similar morphology, biological behavior, and gene expression patterns to human OEAs. This model system could be highly useful for preclinical testing of new chemopreventive and chemotherapeutic agents that target these signaling pathways in ovarian cancer cells. Such therapies may eventually be particularly beneficial for women with recurrent disease.


Wu R, Hendrix-Lucas N, Kuick R, Zhai Y, Schwartz DR, Akyol A, Hanash S, Misek DE, Katabuchi H, Williams BO, Fearon ER, and Cho KR. 2007. Mouse model of human ovarian endometrioid adenocarcinoma based on somatic defects in the Wnt/ß-catenin and PI3K/Pten signaling pathways. Cancer Cell 11:321-333.


Public and Technical Abstracts: Development and Characterization of a Murine Model of Ovarian Endometrioid Adenocarcinoma Induced by the Tissue Specific Expression of Oncogenic beta-Catenin

Top of Page

Ovarian Granulosa Cells Show a New Mechanism for BRCA1-Induced Ovarian Tumorigenesis
Posted September 6, 2007
Louis Dubeau, M.D., Ph.D., University of Southern California, Los Angeles, California

Women with germline mutations of BRCA1 have a 40% risk of developing ovarian cancer by age 70 and are predisposed to breast and fallopian tube tumors. The conventional hypothesis is that BRCA1 functions as a classical tumor suppressor gene, whereby both alleles become inactivated in ovarian epithelial cells. Dr. Louis Dubeau proposed a new hypothesis that granulosa cells in the ovary are the critical cells affected by BRCA1 inactivation and that these cells regulate the growth of ovarian epithelial tumors. Granulosa cells normally produce sex hormones that regulate the ovulatory cycle and influence cell growth. With funding from a Department of Defense Ovarian Cancer Research Program Fiscal Year 2003 Idea Development Award, Dr. Dubeau and his team tested the hypothesis that reduced levels of BRCA1 in granulosa cells induce ovarian cancer through a cell-nonautonomous mechanism. Dr. Dubeau and his colleagues inactivated the BRCA1 gene specifically in ovarian granulosa cells in a mouse model. Two-thirds of the animals developed serous adenomas in the ovaries and uterine horns. However, the tumor cells had epithelial morphology and normal copies of the BRCA1 gene, suggesting that they were not derived from granulosa cells. Instead, these findings suggest that ovarian lesions in this animal model originated from cells derived from the mullerian tract. Furthermore, tumor development occurred through a cell-nonautonomous mechanism through an effector secreted by granulosa cells (see figure). Identifying this secreted factor could lead to new, more effective screening and prevention strategies based on blood-based detection of this factor. The effector may also provide a therapeutic target for treating ovarian cancer.


Dubeau L. 2007. BRCA1 induced ovarian oncogenesis. Published in Proceedings of the 1st International Conference on Ovarian Cancer: State of the Art and Future Directions. Springler Science, New York (in press).

Chodankar R, Kwang S, Sangiorgi F, Hong H, Yen H-Y, Deng C, Pike MC, Shuler CF, Maxson R, and Dubeau L. 2005. Inactivation of Brca1 in mouse ovarian granulosa cells causes serous epithelial cystadenomas carrying functional Brca1 alleles in the ovary and uterus. Current Biology 15:561-565.

Dell H. 2005. Remote control (News & Views article featuring Dr. Dubeau's findings). Nature 434:839.


Public and Technical Abstracts: Mechanism of Ovarian Epithelial Tumor Predisposition in Individuals Carrying Germline BRCA1 Mutations

Top of Page

Ovarian Cancer Detection by Elevated Urinary Levels of Bcl2
Posted March 28, 2007
Patricia A. Kruk, Ph.D., University of South Florida, Tampa, Florida

Ovarian cancer has the highest mortality rate among gynecological cancers, with approximately 12,000 deaths annually in the United States and 114,000 deaths worldwide. Because of a lack of early symptoms diagnosis occurs later in the disease when it has spread beyond the ovary; consequently, the prognosis is poor. No reliable screening test to aid in early diagnosis is currently available. Standard diagnostic tools for the detection of ovarian cancer include physical pelvic examination, ultrasound, or measuring blood levels for CA125. None of these methods provides a reliable and accurate means to detect ovarian cancer. Dr. Patricia Kruk of the University of South Florida received a Fiscal Year 2001 Department of Defense Ovarian Cancer Research Program Idea Development Award to investigate the anti-apoptotic role of telomerase in ovarian cancer. With this award, Dr. Kruk studied the role of telomerase in the cellular resistance to apoptosis in ovarian cancer cells and the clinical relationship between telomerase activity and apoptotic events in clinical specimens of normal and malignant epithelium. Pilot studies on urinary samples from healthy volunteers, patients with benign gynecologic disease, and patients in various stages of ovarian cancer revealed that the levels of BCL-2, a protein with anti-apoptotic activity, were 10 times higher in samples from patients with ovarian cancer than from the healthy volunteers or patients with benign gynecologic disease. Analysis of data from additional studies demonstrated that elevated urinary BCL-2 levels were associated with 92 % of ovarian cancers whereas elevated blood CA125 levels were associated with only 68% of ovarian cancers, suggesting that BCL-2 might be a useful biomarker for the detection of ovarian cancer. Dr. Kruk's study of urinary levels of BCL-2 holds a promising outlook for the future development of a reliable, simple, safe diagnostic tool for ovarian cancer that could lead to improvements in earlier diagnosis and reduce the mortality rates of this insidious disease.


Public and Technical Abstracts: Detection of Ovarian Cancer by Urinary Levels of Bcl-2

Top of Page