NEW YORK (GenomeWeb News) – Using whole-genome sequencing and targeted deep sequencing on tumor and normal samples from several individuals with acute myeloid leukemia, a Washington University team has uncovered some of the patterns contributing to AML relapse after treatment.
"It's a way to really write the dictionary for what went wrong with the human genome when these cancer cells evolved and progressed and relapsed over time," senior author John DiPersio, chief of Washington University's oncology division, told GenomeWeb Daily News.
As they reported today in Nature, the researchers sequenced primary tumor and relapse tumor samples from eight individuals with AML, along with matched normal samples from each. Through comparisons of these genomes, coupled with deep sequencing of candidate mutations, they not only found genes that are altered in relapse tumor genomes, but also traced the progression of cell populations within primary tumors that survived treatment and eventually evolved into relapse tumors.
Their findings suggest that relapse tumors can arise when either a dominant founder clone or one of several clones in the original tumor survives treatment and goes on to further mutate and expand. Along with mutations shared between these persistent primary clones and the relapse cell populations, the team noted, relapse tumors typically contain additional mutations, including changes that appear to be related to DNA damage caused by compounds used to treat the primary tumors.
"AML relapse is associated with the addition of new mutations and clonal evolution, which is shaped, in part, by the chemotherapy that the patients receive to establish and maintain remissions," the study's authors noted.
The report represents the latest in a series of cancer sequencing studies by Washington University researchers.
Elaine Mardis, co-director of the Genome Institute at Washington University, outlined findings from one of the AML relapse genomes at a meeting in early 2010. That relapse tumor, which came from the first woman to have her primary tumor genome sequenced back in 2008, is included in the new analysis.
Using Illumina paired-end sequencing, the researchers sequenced primary tumor, relapse tumor, and matched normal skin samples from eight individuals with AML, generating at least 25 times haploid coverage and more than 97 percent diploid coverage for each genome.
The patients had all gotten essentially the same treatment for their primary AML tumors: an induction therapy involving treatment with anthracycline and cytarabine, also known as Ara-C, followed by consolidation chemotherapy.
After identifying potential somatic mutations and structural variants in the tumor genomes, the team used targeted sequence capture and deep sequencing of skin, primary tumor, and relapse tumor samples to look more closely at the frequency of candidate mutations in each tumor, offering a peek at the clonal cell populations present in the original tumor genome and in the relapse tumor.
By generating sequence covering these regions at a median depth of 590x, researchers could define these clonal populations and see how clones from primary tumors fared following treatment.
The analysis suggests there are a few potential routes that could lead to relapse, DiPersio explained. Some patients had a dominant founder clone in their primary tumor that persisted despite treatment and then came back, he noted, while other patients had several clonal populations in their original tumors — some that were susceptible to treatment and at least one that was not.
The clonal populations that did remain after chemotherapy continued evolving and acquiring mutations, he added, including some mutations that appear to have arisen as a consequence of the treatment itself.
"Chemotherapy is effective at eliminating some of these founder clones and is contributing to the accumulation of changes in the persisting clones," DiPersio said.
But, he emphasized, there is no evidence so far to suggest that chemotherapy causes the mutations that make the clonal populations more aggressive down the road.
"Whether those chemotherapy-associated alterations in the remaining founder clone contribute to its resistance and its biological behavior is not clear," he said. "It's associated with treatment, but it's not known if those changes could make the disease worse or recurrent."
Among the specific genes that were mutated in the tumor genomes, the team found some of the same genes that have been linked to AML in the past, such as DNMT3A, FLT3, IDH1, and IDH2. Not surprisingly, some mutations were shared between primary and relapse tumors from each individual, while others turned up in just one of the tumors.
Within the relapse genomes, though, the researchers saw a range of different mutations with relatively little overlap from one individual to the next.
"There's quite a bit of variation in what we found from the relapse samples patient-to-patient. Much like newly diagnosed AML, there's going to be a number of very diverse mutations associated with relapse," DiPersio said. "They're not largely shared by each patient that relapses. They're unique mutations to that particular patient, though some are shared at a low frequency."
The relapse tumors did contain some recurrent mutations in genes not reported in AML in the past, including WAC, SMC3, DIS3, DDX41, and DAXX. Even so, DiPersio said no pathways have popped out yet as being frequently affected during relapse.
And he predicted that the group will have to sequence around 50 to 100 additional primary-relapse tumor sets to get an idea of the frequency of most of the relapse-related mutations. That work is underway and expected to be completed within a few years.
"As we look at more and more of these mutations, it's more likely that we'll get a handle on what pathways are more frequently involved with relapse than other pathways," DiPersio said. "And that might provide us with some insights into therapeutic advances."