What makes a cancer come back? No one really knows, but for a long time researchers have thought that certain DNA changes in cancer cells may protect them from chemotherapy and allow them to stage a resurgence. Examining one of the most studied cancers, a childhood leukemia, a team in Tennessee has now cataloged exactly what happens to these leukemia cells in youngsters who relapse. What they found, although not necessarily true for other cancers, could be a big step toward new ways to prevent relapse and identify children at risk.
St. Jude Children's Research Hospital in Memphis holds one of the country's best tissue libraries of acute lymphoblastic leukemia (ALL), the most common childhood cancer. Although cure rates for ALL have soared from single digits to more than 80% today, some children still relapse. For them, the chance of survival drops dramatically, to about 30%, says James Downing, St. Jude's scientific director. These children remain one of the greatest challenges for ALL specialists.
Wondering which changes in DNA might be responsible for these relapses, Downing, hematologist and pathologist Charles Mullighan, and their St. Jude colleagues focused on one particular type of DNA pattern, called copy number. Increasingly, researchers are finding that extra stretches of DNA, or DNA that's deleted in some people--so-called copy number alterations (CNAs)--seem to play a role in disease, particularly in neuropsychiatric conditions such as autism or mental retardation (ScienceNOW, 14 November).
From their tissue bank, the team culled samples from 61 ALL patients who had relapsed and compared samples taken at the time of diagnosis with those taken when the cancer came back. Tomorrow in Science, Downing and colleagues report that 92% of leukemia cells from relapsed patients harbored changes in CNAs; most often, they found a loss of DNA that caused certain genes to shut down. In all, 15 genes had lost DNA at relapse, and two genes had gained DNA. Many of the genes control cell division or immune cell development, which goes awry in some leukemias. To Downing's surprise, CNAs didn't appear to be hitting genes involved in drug resistance. Some doctors have assumed that resistance was important because patients who relapse with ALL often respond poorly to chemotherapy.
Where did the CNAs come from? In many cases, CNAs present at diagnosis appeared in far more cells at relapse, suggesting that they managed to elude treatment and proliferate while other leukemia cells died. Another possibility is that DNA changes appeared as a result of chemotherapy. Several specialists dismissed this theory, however, because the changes showed up in genes associated with leukemia, not randomly across the genome.
In terms of treatment, hematologist-oncologist Matthew Walter of Washington University in St. Louis, Missouri, suggests focusing on CNAs that may mark an "ancestor" preleukemia cell in the bone marrow; Downing's group found sets of CNAs in relapse and diagnosis samples from the same child that were overlapping but not the same. Some CNAs were lost at relapse and others were gained. This happened, they deduced, because all of the CNAs probably were part of ancestor cells that helped spawn the leukemia and, later, the relapse. "Maybe those are the mutations that our therapies need to be addressing," he says.
The work "is fascinating from a lot of viewpoints" and "raises perhaps more questions than answers," says oncologist Bert Vogelstein of Johns Hopkins University in Baltimore, Maryland. Although it's been known for years that ALL and other cancers that relapse typically acquire new genetic changes, "one would not have expected there to be so many" as seen here, says Vogelstein.
Puzzles remain. Not all of the findings implicated CNAs, for example. Seven children actually lost all of the CNAs present before relapse, and five children had no change in CNAs between diagnosis and relapse. In addition, far fewer CNAs show up in a more aggressive and deadly childhood leukemia, acute myelogenous leukemia. Downing is now studying whether the ALL findings can be used at diagnosis to identify patients who have a high or low risk of relapse.