A team of center researchers led by Dr. Lenora Loo, a postdoctoral researcher in Dr. Peggy Porter's lab, has used a sensitive genetic technique to distinguish between different forms of breast cancer. The approach, which identifies a full spectrum of chromosome abnormalities in breast-cancer cells, could form the basis for a test to help doctors predict the development of invasive cancer or predict the outcome in patients who have cancer.
The analysis makes use of a technique called array comparative genomic hybridization (CGH), which reveals the segments of chromosomes that are lost or duplicated in cancer cells. Such genetic rearrangements provide signposts to genes which, when present in abnormal numbers of copies in the cell, may contribute to cancer.
The approach could help scientists learn the order in which chromosome defects are acquired as a tumor becomes more advanced as well as define which sets of defects occur in different types of cancer. Both types of information are important for detecting early changes that precede invasive cancer or customizing therapy to an individual patient's disease.
The study, published in the December 1 issue of Cancer Research, includes work from colleagues in the Human Biology and Public Health Sciences divisions as well as researchers in the Genomics and Flow Cytometry shared resources.
The researchers analyzed 44 breast-tumor samples that had been identified based on a previous study — led by PHS investigators Drs. Christopher Li, Janet Daling and Kathi Malone — which identified groups of women who had either the ductal or lobular form of the disease. Although ductal breast cancer is more common, Li and colleagues have found that the occurrence of lobular cancer has increased during the last decade. This rise is associated with the use of combination hormone-replacement therapy. Lobular cancer is highly responsive to therapy but more difficult for doctor's to detect.
Using tissue-analysis methods known as immunohistochemistry, pathologists Drs. Ming Gang Lin and Xiaopu Yuan in Porter's lab further classified the tumor types based on whether they were estrogen-receptor (ER) positive or negative, a characteristic doctors routinely assess to evaluate prognosis and therapy. ER-positive cancers are typically more responsive to therapy and are sensitive to cancer-prevention agents like tamoxifen.
Porter's lab next applied a powerful enrichment step that separated normal cells away from the cancer cells using a method called flow cytometry. This elimination of normal cells was critical to the investigators' ability to detect chromosome losses.
Loo then used array comparative genomic hybridization to determine whether chromosome differences could be detected among the different cancer subtypes. The technique involves affixing bits of the human genome onto a glass slide. The slide contains more than 4,000 segments of the genome, representing much — although not all — of the genome. The slides were created through collaboration between Dr. Barbara Trask's laboratory and Dr. Jeff Delrow and colleagues in the Genomics shared resource.
Genetic material is then extracted from normal and tumor cells, treated with substances that allow them to be distinguished from each other and applied to the slides. This step reveals which parts of the genome are present and in what amounts in the tumor cells, allowing the researchers to evaluate which regions of the genome are missing or duplicated in the tumor cells. By looking for overlapping patterns from tumor samples, researchers can identify regions of the genome that are consistently altered in different cancer subtypes. These regions likely contain genes that contribute to the particular form of the disease. Several of the segments identified are known to contain genes that have previously been implicated in many types of cancer.
Although other groups have used this technique to analyze chromosomes from breast and other tumors, the Fred Hutchinson study is the first to systematically compare chromosome abnormalities among different subtypes of cancer, Loo said.
Working with statisticians Drs. Li Hsu and Doug Grove, Loo found a consistent association between each form of cancer and particular additions or losses of parts of different chromosomes. Several of the segments identified are known to contain genes that have previously been implicated in many types of cancer.
"Even though this was a small study, this is good indication that there are genomic profiles that could potentially differentiate between cancers based on their histological (ductal or lobular) type and their estrogen receptor status," Loo said. "In future studies, we may be able to identify other subgroupings. Ultimately, we may be able to correlate these chromosome alterations with a patient's prognosis."
The Porter lab is using array CGH and another array system that identifies the loss of one of two copies of individual genes through the genome to help define the molecular changes associated with important subtypes of breast cancer, such as estrogen receptor negative tumors and ductal and lobular forms or the disease. The genomic differences detected will help identify markers that can be used to predict behavior in the specific subtypes.
Dr. Barbara Trask, one of the study co-authors, noted that many other researchers at the center are using the arrays developed in the Genomics Resource to analyze other types of human tumors and mouse tumors. A growing group of investigators meet twice each month to discuss their research using array CGH.