Research in Cancer Genetics: Toward Improved Cancer Prevention and Treatment

Some gene mutations are carried by the germ cells, sperm or egg, from parent to child; these can result in inherited illnesses. Mutations, both spontaneous or caused by environmental exposure, also occur in the body's other cells during the course of our everyday existence. These single mutations in our non-germ cells are often harmless and the cell simply dies. Mutations in growth regulatory genes, however, can start a cell down the path to cancer.
Fortunately, the body still has many lines of defense left, and most clinically apparent cancers can only be established when the original cell continues to divide and picks up four to six additional mutations over the course of 10 to 20 years. Molecular approaches have allowed identification of some key mutations that initiate cancer progression; however, sequencing of the human genome is a watershed event that will lead eventually to a fuller understanding of the relation between genes and cancer. Here, I briefly relay two new lines of cancer genetics research that are proceeding at UNC-Chapel Hill, one defining cancer susceptibility and one attempting to improve cancer treatment. These are large projects undertaken by teams of investigators.

Cancer Prevention: Can We Define Families or Populations with Heightened Susceptibility?

We know that certain families exhibit a high risk of cancer because they carry a mutated gene and pass it from generation to generation. While these familial cancer syndromes constitute only between 5 percent and 10 percent of cancers, identification of the involved genes has been enormously instructive. Most of these familial cancer genes function in control of DNA replication or proofreading and repair of damaged DNA. These are the cell's fail-safe mechanisms that prevent mutations and allow perfection in the passage of our DNA to daughter cells.
When the capacity to police DNA imperfections is compromised, many more mutations occur, making cancer-causing mutations all the more likely. One hypothesis of the UNC-Chapel Hill cancer genetics effort is that the other 90 percent of cancers do not arise spontaneously in an equally susceptible population. Rather, we believe that individuals or families vary in cancer susceptibility because they inherit a complex set of gene alterations ­ each one of which, by itself, is insufficient to increase susceptibility, but the constellation, when inherited together, increases cancer risk. The risk in individual or families with these complex patterns of inheritance may be particularly high in conjunction with exposure to specific environmental, hormonal, dietary or infectious agents.
As a team at Lineberger, molecular epidemiologists, basic scientists studying DNA repair processes, medical and molecular geneticists, statisticians and clinicians are charting this major UNC-Chapel Hill initiative. We will use population-based and clinical studies, high throughput genotyping, and quantitative trait and statistical analysis to create a body of knowledge about the complex inheritance of subtle cancer susceptibility. Our ultimate aim is to understand who is susceptible and then target prevention to the individuals and families who can benefit the most.

Gene Expression Analysis: Prediction of Cancer Outcome and Response to Therapy

Our current classification of tumors relies heavily on the light microscope coupled with the recent addition of several individual predictive molecular markers. Unfortunately, these techniques often fall short of perfection. For example, two women with microscopically identical breast tumors, or two men with similar prostate cancers, can have widely divergent outcomes.
A second Lineberger team is striving to create a molecular definition of tumor types by analyzing the expression of all genes in a patient's tumors. We will go beyond the assessment of 2-10 gene products by current techniques and use a new technology designed to measure the expression of 10,000 genes in a single tumor sample.
This cDNA microarray technology uses robots to spot 10,000 circles of individual genes' DNA on a microscope slide. Comparison of genes expressed in a common reference pool to those from an individual's cancer, both before and after cancer treatment, can be accomplished by sophisticated scanners that detect fluorescence intensity in each of the slide's 10,000 spots.
The challenge is to convert this mountain of data into computerized information capable of being grasped by the human mind. This is the task of our bioinformatics group, who will draw on ways to cluster gigabytes of information in displays that can be visualized and then analyzed. In our examples we will demonstrate that molecular profiling of gene expression is becoming a reality and may enable prediction of a patient's response to a specific therapy.