Photo by Clay Eals
Highly sensitive tests developed at the Hutch may help doctors predict the success or failure of treatment for acute myelogenous leukemia (AML), a disease that occurs in 6,000 to 10,000 people a year in the United States.
The tests, including one that can detect one mutant cell among 10,000 to 1 million, identify genetic abnormalities that indicate an aggressive form of the leukemia. The tests may let physicians tailor therapies for patients and provide early clues about the likelihood of relapse.
The Southwest Oncology Group (SWOG), whose statistical center is based in the Public Health Sciences Division at the Hutch, will begin a clinical trial using this method early next year for patients at the Seattle Cancer Care Alliance and other SWOG sites around the country.
With conventional chemotherapy, about 25 to 45 percent of AML patients can be cured. Patients who relapse may have benefited most from a bone-marrow or stem-cell transplant, a procedure which carries its own degree of risk. Doctors have been at a loss to predict which individuals will fail standard chemotherapy and which will be cured.
Because no outwardly visible clues can answer this question, scientists now turn to DNA-based methods that can detect subtle genetic abnormalities that correlate with disease severity and response to treatment.
Mutations in a gene called FLT3 are found in 30 to 40 percent of patients with AML. In pediatric and adult leukemia patients, these mutations spell a poor prognosis, said Dr. Derek Stirewalt, a medical-oncology research associate in the Clinical Research Division and lead author of the study.
Analysis prior to treatment
"With these tests for the FLT3 abnormalities, patients with AML could have genetic analysis for mutations in the gene prior to treatment," he said. "If we detect a mutation, we could use that information to stratify patients into risk groups and select the most appropriate treatment for individual patients.
"We also have developed tests for FLT3 mutations that will allow us to follow patients prospectively throughout treatment. These tests for minimal residual disease after treatment may be able to predict the likelihood of relapse after treatment."
Stirewalt, a member of the lab of Dr. Jerald Radich in the Clinical Research Division, in collaboration with Dr. Cheryl Willman at the University of New Mexico Health Sciences Center, published these techniques in this month's issue of Leukemia Research. Stirewalt's research is funded by a training grant from the American Cancer Society.
Trace cells with cancer-causing mutations that persist or recur after treatment are characteristic of what is known as minimal residual disease, a condition studied extensively in Radich's lab. Such patients have a high probability of relapse.
Radich and colleagues have shown that polymerase chain reaction, a technique that enables minute quantities of DNA to be expanded, detects minimal residual disease in patients with chronic myelogenous leukemia (CML). These patients can then be treated aggressively to minimize the chance of relapse.
Stirewalt used a similar approach to develop a test that can be used both as a prognostic indicator for AML prior to treatment and to monitor minimal residual disease. Unlike CML, which always is caused by the same genetic defect, AML is believed to arise from multiple DNA abnormalities that vary from patient to patient.
One common mutation, found in about 30 percent of AML patients, is a rearrangement in the FLT3 gene that results in a tandem duplication of a segment of the gene. In addition, another 5 to 10 percent of AML patients have different mutation in FLT3. Together, these two genetic alterations make mutations in the FLT3 gene the most common genetic mutation found in AML patients.
FLT3 specifies production of a protein known as a tyrosine kinase and is one of a family of proteins involved in normal cell growth and development, Stirewalt said.
"FLT3 is important in normal regulation of early bone-marrow progenitor (ancestor) cells," he said. "Normally, it responds to a growth factor that modulates cell division. But in AML, the mutated FLT3 protein appears to be inappropriately activated without the growth factor, and this may drive cells to proliferate and prevent normal differentiation."
Exquisitely sensitive test
His test for FLT3 is exquisitely sensitive, detecting as few as one mutant cell in a background of 10,000 to 1 million normal cells.
Early next year, SWOG will begin a clinical trial for patients at the Alliance and all other SWOG sites that use Stirewalt's method.
"We will test the status of FLT3 in AML patients as part of their initial workup," he said. "Patients will be prospectively followed to determine if these tests can accurately predict who will relapse."
Also, a trial a similar trial, headed by Dr. Soheil Meshinchi at the Alliance, is beginning to accrue pediatric AML patients.
Stirewalt said use of the FLT3 test will impact other research on AML treatment.
"There is a lot of excitement about new drugs that inhibit the FLT3 protein in mice," he said. "We and others are aggressively pursuing ways to block the activity of the mutated FLT3 gene. If these targeted therapies prove useful in humans, we could use our molecular test to identify those patients who would respond to the drug."
Application to ALL
Recently, the Radich lab has shown that molecular detection of minimal residual disease can be applied successfully to another blood cancer, acute lymphoblastic leukemia (ALL). Cells from some 5 to 20 percent of patients with ALL contain a genetic abnormality called the Philadelphia chromosome, the same mutation that is found in all patients with CML.
At the American Society for Hematology meeting earlier this month, Stirewalt and Hutch colleagues presented results of a study that found that their PCR test for minimal residual disease accurately predicts which patients with ALL are likely to relapse after transplantation.
"We're really excited about these results," Stirewalt said. "We can use this technique to identify those patients who are at high risk for relapse before there are any observable signs of disease. This demonstrates that we have developed a sensitive prognostic tool."