Many cancers rely on what scientists call oncogene addiction, which refers to malignant processes that are driven by a single alteration, such as mutations in the EGFR gene in lung cancer or over-expression of HER2 in breast cancer. Therapies have been successfully developed targeting those key alterations, turning the presence of the abnormality into a hallmark of activity, or a predictive biomarker, Dr. Jimeno said.
But the Hopkins researchers, led by Manuel Hidalgo, M.D., Ph.D., suspected that pancreatic cancer − a disease in which EGFR-targeted therapies have proved to be of use, but where amplifications or mutations in the EGFR gene are rare − is an example of a cancer driven by pathway addiction. The problem, however, is that no one knows what the determinant of Tarceva activity is in pancreatic cancer, Jimeno says.
The team used a tool known as gene set enrichment analysis (GSEA), which classifies 25,000 genes assayed in a chip into 198 predetermined pathways, be they metabolic, structural, or signal transduction. This analysis was conducted by Aik Choon Tan, Ph.D., from the Institute of Computational Medicine at Johns Hopkins. The researchers looked at messenger RNA levels to determine which pathways were more actively expressed in patients whose disease responded to Tarceva. Two of the most frequently expressed pathways in Tarceva-sensitive cases were the closely related EGFR and MAPK pathways, they found.
What we did that was different is that rather than examining gene expression datasets looking for outliers or individual genes, we employed GSEA to detect modest but coordinated changes in expression of genes involved in a common pathway or biological function, Dr. Jimeno said.
They then validated the prediction in a prospective cohort of patient-derived tumors xenografted on mice that were treated with Tarceva, and found that the EGFR pathway was again high
|Contact: Greg Lester|
American Association for Cancer Research