February 17, 2010 Bar Harbor, Maine -- Researchers at The Jackson Laboratory have discovered telltale variations in mRNA processing -- the cell’s protein-building function -- that correspond to cancer. The team showed that they could distinguish among similar tumor subtypes with at least 74 percent accuracy; the current standard in molecular diagnostics is about 10 percent.
Jackson Associate Professor Joel Graber, Ph.D., a physicist-turned-computational scientist who collaborated with cancer researcher Assistant Professor Kevin Mills and other Jackson scientists on the project, notes that while molecular cancer diagnostics are an important clinical advance in cancer management, “new methods are still needed to improve accuracy.”
In molecular diagnostics, DNA is extracted from a patient’s tissue sample (such as a tumor) and analyzed using gene expression microarrays to find activity patterns associated with various diseases. “Microarray diagnostics have been with us for 15 years, but so far haven’t lived up to their promise,” Dr. Mills says. “The profiles that come off microarrays don’t accurately detect most cancer cases, and even when they do, they don’t categorize them correctly.” The Jackson team found that the molecular profiles, which basically indicated which genes were “on” or “off,” can miss an important factor: subtle differences in genes that orchestrate the processes in other genes.
“Almost every gene in the human genome can come in more than one form, or isoform,” Dr. Graber explains. “For things to work right, not only do the right genes have to turn on, but they also have to be processed the right way, by the protein products of other genes. Disruption of mRNA processing genes can potentially alter lots of things that are going on in the cell, possibly leading to cancer.”
Dr. Mills studies genetic mechanisms, such as DNA repair, that can cause cancer when they fail. Using mouse models of B-cell lymphoma, a kind of leukemia, he was looking at three different kinds of tumors in two mouse varieties. Even though the tumors were indistinguishable from one another by standard laboratory methods of histology and cell sorting, the mice had three distinct survival rates.
He shared his findings with Dr. Graber, who was interested in expanding the gene expression analysis, specifically to include systematic changes in the balance between isoforms of the genes measured with the microarrays. “When you use that analysis,” Mills says, “you get this huge quantum leap in your prediction of the subtypes. Joel was getting almost 90 percent accuracy. I almost fell off my chair the first time I saw the data.”
While truncation of a genetic region known as 3'-UTR was the most commonly observed pattern, they also detected genes with elongated transcripts. An analysis of cellular samples drawn from similar tumors that respond differently to chemotherapy also revealed potentially diagnostic changes in processing, the researchers note in the paper.
They confirmed their findings by analyzing microarray data from human primary tumor samples.
Dr. Mills suggests that the patterns detected by the team may indicate cellular processes that directly influence cancer development. “Because it works so well,” he says, “we theorize it’s pinpointing the pathways that are really relevant. That accuracy has to be telling us something functionally important as well as diagnostically accurate.”
Drs. Graber and Mills are continuing their collaboration, characterizing other tumor types and analyzing cancer patient data using the new method.
Dr. Mills comments, “One thing that motivates us is that we’re working in childhood cancers. These really early-onset leukemias tend to be worse than adult cancers, in that they’re more aggressive and therapy-resistant. Couple that with the fact that this is a population that is least able to tolerate the chemotherapy.”
He adds, “We want to find targeted ways to treat each patient’s cancer, based on that child’s own genetic makeup. We want to have a better idea of which patients can respond to and tolerate a treatment.”
The Jackson Laboratory is an independent, nonprofit biomedical research institution based in Bar Harbor, Maine, with a facility in Sacramento, Calif. Its mission is to discover the genetic basis for preventing, treating and curing human diseases, and to enable research and education for the global biomedical community.
Help employers find you! Check out all the jobs and post your resume.
