|
||||||||
|
PATHOLOGY Surprising Likeness Found in Genetic Profiles of Invasive and Noninvasive Breast TumorsCancer Cause May Include Interplay of Factors Beyond Gene Expression Even the most virulent breast cancer, one that has broken through the milk-producing ducts to invade the surrounding fatty tissue, starts out as a tidy self-contained cluster that if removed would cause little trouble. So different in appearance are the ragged-edged invasive tumors from the neater in situ variety that scientists have assumed to get from one to the other must require the turning on or off of genes.
"Quite frankly, I think this study is just the beginning," said Dennis Sgroi. "It probably raises more questions than it answers. But some of the answers we have gotten from a global perspective are consistent with what others are seeing." (Photo by Graham Ramsay) Yet from a molecular point of view, HMS researchers have found little to distinguish an invasive breast tumor from its more isolated predecessor. Taking a kind of molecular bird's-eye view, Dennis Sgroi and his colleagues analyzed the gene expression patterns of invasive ductal carcinomas (IDC) and ductal carcinomas in situ (DCIS) taken from 36 women. Though each differed genetically from nearby swatches of normal breast tissue--1,940 genes were either turned on or off--invasive and preinvasive cells were almost identical in their gene expression patterns. The findings appear online in Proceedings of the National Academy of Sciences Early Edition for the week of April 21.
On the face of it, the new findings pose fresh dilemmas for researchers and patients. Cancer cells were thought to leap from one stage to another by accumulating new mutations or by activating or inactivating preexisting genes. Sgroi's study suggests a more complex and, in some ways, richer story. Rather than a linear narrative featuring the turning on and off of individual tumor suppressors or oncogenes--the heroes and villains of the piece--cancer progression may depend on contingencies such as levels and timing of gene expression. To be fully told, the story may require weaving together sequences of events happening at different levels--gene, protein, tissue--that do not bear predictable relationships to one another. "Gene expression is only one aspect of it all. There is also the protein level," said Sgroi. "So we are only looking at one part of the problem. You have to be very cautious about how you expand upon the interpretation of this data. Maybe RNA message levels have not changed but the protein levels have. So we could be missing very subtle things." Typing TumorsThe new findings might appear to put those looking for solutions to the problem of breast cancer behind a kind of molecular eightball. Yet Sgroi's new study does offer hope to researchers, and potentially to clinicians and their patients. To begin, it provides fresh clues about why some patients' cancers are more aggressive than others. For years, clinicians observed that the more undifferentiated, bloblike, and quick to reproduce the cells are in a preinvasive cluster, the more likely the cluster is to become malignant. On this basis, they classified breast cancers into three grades--I, II, and III--the latter being the most aggressive. When Sgroi and his colleagues separated the patient tumor samples according to their grade, they observed telling differences. For example, a subset of the genes associated with breast cancer was more highly expressed in grade III than grade I tumors. Since tumors at both grades can become invasive, there may be more than one genetic pathway by which breast cancers arise, said Sgroi. Exploring the genetic differences between grades I and III could lead to new hints about why some tumors progress so rapidly, and how they might be stopped.In addition, the discovery that, in general, breast tumors differ from normal tissue by the expression of 1,940 genes could provide hints about why clusters of cells become cancerous in the first place. "We now have data that we never had," said Sgroi. "This is a very important database that will be important to basic researchers. If they want to know if a gene is perhaps involved in breast cancer, they can access this database." The information they retrieve could lead to not just a better understanding, but also new means for detecting and treating breast cancer. "We may be able to come up with markers that distinguish normal from abnormal tissue," Sgroi said. "Picking up breast cancer at an early stage, we have a greater impact on the outcome. If we can pick it up early, we may be able to cure the patient." Harvesting mRNAIt was curiosity about which genes were nudging normal tissue into the earliest stages of breast cancer that first drew Sgroi's interest, but he faced a problem. Many breast tumors are a motley mix of cells at various stages. Cells from the earliest stages, ADH and DCIS, are a minority in many biopsies. What is more, cancer cells are so tiny that it can be difficult to harvest enough messenger RNA to perform microarray analyses. Bedeviled by these twin problems of heterogeneity and size, researchers had yet to come up with a clear molecular characterization of the three stages of breast cancer progression.As it turned out, researchers at the National Cancer Institute and Arcturus Engineering had developed techniques that got around the problems. The first of these, laser capture microdissection, essentially trains laser-heated plastic on specific cells, which are then netted and plucked out by the melted plastic. Using this method, Sgroi and colleagues were able to isolate cells from the three stages of cancer progression, along with adjacent normal cells, in the archived tissue specimens of 38 MGH patients. At least two of the three stages were represented synchronously in 31 patient specimens. Using a method for amplifying RNA that they developed, Arcturus's Mark Erlander and colleagues gleaned enough reagent to perform genetic expression profiles on the captured samples--a laborious undertaking since two samples for each tumor stage represented in each biopsy were subjected to microarray analysis. "Some people have said that is excessive," Sgroi said. "But we wanted to show how consistent and reproducible this is. We wanted to be very, very careful in our analysis." Though surprising, the lack of qualitative differences between breast cancer stages had been hinted at by other researchers. In small-scale studies using different methods, Dana-Farber Cancer Institute's Kornelia Polyak, an HMS assistant professor of medicine, had observed gene expression patterns to be more similar than previously believed. "So our findings are not completely off the wall. They seem to be consistent," Sgroi said. Furthermore, the pairing of laser capture microdissection and DNA microarrays could yield results outside the breast cancer and general oncology communities. "I believe we can apply this approach to any tumor," Sgroi said. "In my opinion, we should be able to apply this to just about any problem--not only breast cancer but any disease process." --Misia Landau |
