It looked like a scientific dead end – a clinical trial that found no benefit to a potential drug for a form of leukemia. But, like police detectives working a cold case, Dana-Farber scientists gathered hundreds of tissue samples that had been collected for the study – most of them languishing in laboratory cabinets, destined for disposal – and analyzed their molecular makeup.
The result was discovery of a distinct genetic subtype of acute myeloid leukemia (AML) that often has a worse prognosis than other forms but is rich in targets for new “smart” drugs. The research, published in the journal Blood, is testimony to the wealth of information that can be extracted from tumor tissue in the genomic age – if scientists have the foresight to recognize its value.
The research trail began in 2008 with the opening of an international phase 3 clinical trial of a novel drug for patients with secondary AML. Diagnosed in several thousand people a year in the United States, the disease can arise after a previous malignancy of the bone marrow or after treatment for another cancer. The trial, co-led by Dana-Farber’s Richard Stone, MD, examined whether a combination of a new drug, called amonafide L-malate, and the chemotherapy agent cytarabine worked better than the standard therapy of cytarabine and daunorubicin (another chemotherapy drug). Nearly 450 patients at 150 treatment centers around the world were randomly assigned to receive either the new or standard combination.
The results were unambiguous: the new combination was no better (or no worse) than standard therapy at producing complete remissions in patients.
The lackluster showing brought an end to clinical testing of amonafide L-malate, and the company that produced it began to shut down operations, says Stone. But instead of allowing this particular line of research to go fallow, Stone and Dana-Farber’s Donna Neuberg, ScD, began discussing what amounted to a scientific salvage operation.
“As part of the trial, the participating research centers had collected tumor tissue samples from hundreds of patients, often at different stages of their treatment,” Stone remarks. “We thought that by collecting this material and analyzing it for genomic abnormalities, we might get insights into why some patients responded better to the treatment than others.”
Stone and Neuberg asked their Dana-Farber colleague R. Coleman Lindsley, MD, PhD – a physician-scientist with a research interest in myelodysplastic syndromes – if he thought the plan was feasible. He said he’d investigate.
The tissue specimens arrived at Dana-Farber in January 2012 in 20 storage crates that, Lindsley says, were “not well organized, but well labeled,” allowing him to track the samples to individual patients. The tissue had been affixed to glass slides for viewing under a microscope. Lindsley used a razor blade to scrape the tissue off and collected each sample in a separate vial. He separated out the DNA from the tissue samples and analyzed it for mutations.
The laboratory work was performed under the supervision of Lindsley’s mentor, Benjamin Ebert, MD, PhD, and was funded by a grant from the Friends of Dana-Farber Cancer Institute and by gifts to Dana-Farber’s Adult Leukemia Program, including gifts from Team Flames of the Pan-Mass Challenge and the Ted Rubin Memorial Golf Tournament.
“The tumor tissue samples were an extraordinary resource,” Lindsley comments. “Coming from hundreds of patients on five continents, they represented one of the best and most broadly representative sample sets from people with this disease.”
The researchers analyzed tumor samples from nearly 200 study participants, divided into two groups: secondary AML (s-AML), which arises after a previous bone marrow cancer, and therapy-related AML (t-AML), which occurs after treatment for another cancer. By comparing the genetic landscape of these two rare subtypes with that of “de novo” AML – a more common variety that arises in the absence of previous cancer – investigators identified a specific molecular signature for s-AML.
Because de novo AML and s-AML are often treated differently, the researchers wondered whether the newly defined genetic signature provides a more precise way of identifying these subtypes than the standard approach. They analyzed tissue from 105 leukemia patients treated at Dana-Farber whose disease had been categorized by conventional means. They found that one third of older patients with de novo AML carried the s-AML signature in their tumor cells, and that these patients’ disease tended to behave much like s-AML, with a poor response to chemotherapy and a tendency to relapse.
“The findings provide an objective way of identifying patients with this subtype of AML, and predicting which patients are most likely to benefit from standard chemotherapy,” Lindsley remarks. “It can also help direct our efforts to develop better treatment regimens for patients whose tumor tissue carries the s-AML signature.”
Adds Neuberg, “the decision to obtain and analyze tissue from an unsuccessful clinical trial shows where an academic approach to science can complement a commercial approach. The patients enrolled in the clinical trial did so with the understanding that their participation could help improve treatment. Our research has allowed that to happen.”