Predicting Growth In Cancer Stem Cells

Cancer is a frightening and thorny adversary. Therapies to battle human cancer are frequently harsh with debilitating side effects. But often, these initial therapies are successful in destroying the bulk of tumor cells and driving cancer into remission.

However, in many cases, later, sometime years later, cancer can return aggressively. Frequently when it returns, it is resistant to the therapies that originally drove it into remission. Why is this scenario so common?

One idea is the cancer stem cell theory. Although there are many complexities to cancer stem cell biology, the basic model suggests that many cancers are composed of different types of cells arranged in a hierarchy, much like a normal tissue. At the top of the hierarchy are the cancer stem cells. These are a minority population with an unlimited capacity to divide though they may divide infrequently. The bulk of the tumor is thought to be composed of partially differentiated descendants of the cancer stem cells that divide rapidly but for a limited time before they cease to be proliferative.

Many therapies target and destroy this rapidly dividing population, but the cancer stem cells persist and give rise to new rapidly dividing descendants that may harbor resistance to therapy.  One solution to this challenge is to target cancer stem cells, and preclinical and clinical trials attempting to do so are underway. Unfortunately, there are obstacles. As mentioned, cancer stem cells may grow slowly and make up a minority population in tumors, making measurement of a therapy’s effectiveness difficult.

To identify new potential targets for cancer stem cells, a group of cell biologists collaborated with bioinformaticists at the Oklahoma Medical Research Foundation. The bioinformaticists used a computer program one of them had developed called GAMMA (Global Microarray Meta-Analysis). Simply put, GAMMA scans a large portion of the thousands of published experimental gene expression studies, not just those in the cancer field. GAMMA then applies a “Guilt by Association” algorithm to rank proteins’ likelihood of participation in many biological processes.

The cell biologists took these predictions and tested 50 genes anticipated to be involved in cancer stem cell division. They used a cell culture model of breast cancer stem cells developed in Robert Weinberg’s laboratory at MIT. Three related breast cell lines, which lack cancer stem cell characteristics, served for comparison. The researchers used RNA interference to deplete cells of each of the proteins coded by the 50 selected genes.

They found that 21 of the target genes showed preferential growth inhibition of the breast cancer stem cells versus the other three breast cell lines. These 21 genes participate in a variety of biological pathways, and some have been little studied. The researchers examined 6 of the genes in more detail finding that 4 were implicated in the process of chromosome segregation during cell division. Determining the exact biochemical pathways in which the 21 genes function and testing whether these pathways can be targeted for therapeutic inhibition of cancer stem cells will require further study.

These studies are described in the article entitled Predictive bioinformatics identifies novel regulators of proliferation in a cancer stem cell model, recently published in the journal Stem Cell Research. The studies were a collaboration between Constantin Georgescu and Jonathan D. Wren from the Arthritis and Clinical Immunology Research Program and Evan Fields, John R. Daum and Gary J. Gorbsky in the Cell Cycle and Cancer Biology Research Program at the Oklahoma Medical Research Foundation in Oklahoma City.