, , , ,

RamaDevinRama Raghavan, M.S., is a bioinformatics software developer at the University of Kansas School of Medicine and has extensive experience in computer programming, exposure to statistical analysis of large genomic datasets, and performed analysis of NGS data using various standard pipelines. Her fundamental interest is to use computer science in solving computationally intensive problems.
Devin Koestler, Ph.D., is an assistant professor in the Department of Biostatistics at the University of Kansas School of Medicine. His primary research involves the development and application of statistical methodologies for high-throughput ‘omics’ data. In addition to methodological interests in statistical genomics, multivariate statistics, mixture models, and mixed effects models, he also has a deeply rooted interest in epigenetics and molecular epidemiology.

Epithelial ovarian cancer (EOC) remains a leading cause of women’s gynecologic cancer death in developed countries, with late detection and platinum-resistant relapses contributing to unfavorable prognoses. Because only 10% to 15% of patients who present with advanced disease experience long-term remission, most patients are subject to repetitive and frequent treatment cycles, disease recurrence, and/or unchecked disease progression. Along with the physical burdens suffered by affected patients, the costs to the health care system are significant, with recent estimates suggesting that EOC accounts for $5.1 billion annually. The enormous physical, societal, and economic burdens associated with EOC underscore the pressing need for continued and future research aimed at elucidating factors that underlie the development and progression of this disease and, importantly, the discovery of novel therapeutics for its treatment.

From 2007 to 2009, 30% to 40% of drugs that were launched in the United States were repositioned, reformulated, or new combinations of existing drugs. Drug repositioning reduces the risk of failure due to adverse toxicology and bypasses much of the early research and development costs.  According to a recent survey of 30 pharmaceutical and biotechnology forms, the cost to relaunch a repositioned drug averages $8.4 million, compared to $41.3 million for launching a new formulation of a drug in its original indication. In addition, whereas 10% and 50% of new molecular compounds make it to the market from phase 2 and 3 clinical trials, respectively, rates for repurposed compounds are significantly higher (25% and 65%, respectively).  One of the most well known and successful examples of repositioning is thalidomide, which was initially launched in 1957 as a sedative but was later found to be an effective treatment of pain relief in leprosy and Kaposi sarcoma, generating over $500 million in annual revenue.

So how can we go about determining which (if any) existing drugs or compounds can be repositioned for the treatment of ovarian cancer?  One approach involves the identification of gene signatures that are associated with clinically relevant outcomes of ovarian cancer (e.g., survival, time to disease recurrence, etc.), followed by the identification of existing drugs or compounds that target those same genes. Specifically, by making use of publicly available data deposited in The Cancer Genome Atlas (TCGA) consisting of the gene expression signatures of patients with high-grade serous ovarian cancer (the most common subtype of ovarian cancer), our group has focused on identifying specific genes whose expression associates with downstream clinical outcomes of the patients. Genes identified from this initial analysis are then explored using novel bioinformatics tools, such as the Connectivity Map (CMAP), enabling the identification of existing drugs that share a similar expression signature. Drugs or compounds discovered from this approach therefore represent potentially new therapeutic agents for the treatment of high-grade serous ovarian cancer.

The nuts and bolts of this approach is the CMAP algorithm, developed by Broad Institute at MIT. This algorithm represents an innovative way to connect diseases with a specific gene expression signature and potential drug targets through the common vocabulary of genome-wide expression profiling. CMAP references contain 1,309 FDA-approved small molecules and over 7,000 expression profiles. The approach starts with a “query signature” (e.g., drug-resistant versus drug-sensitive leukemia) and assesses its similarity to each of the reference expression profiles in the dataset, yielding a connectivity score. Some recent findings of CMAP include: ursolic acid for reducing fasting-induced muscle atrophy, new analgesic and antinociceptive properties of phenoxybenzamine, and phenothiazine antipsychotic drugs to inhibit proliferation of tamoxifen-resistant breast cancer cells.

Unraveling the complexity of ovarian cancer and the multipurpose potential of existing drugs and compounds make efficient use of existing resources and have tremendous implications for the future treatment of this disease. Importantly, in the advent of the genomic era, existing publicly available data from repositories such as the TCGA and innovative software tools like CMAP represent an incredibly valuable resource and hold great promise to accelerate ovarian cancer drug discovery.