AML Genome Reveals Complexities, Potential Driver Mutations

May 4, 2013
Anna Azvolinsky

Researchers have characterized acute myeloid leukemia, providing new genetic driver leads to help classify the disease and even stratify AML patients by risk.

Researchers have characterized both the genetic and epigenetic aberrations found in acute myeloid leukemia (AML). The results are published in the New England Journal of Medicine with data sets publically available through the Cancer Genome Atlas Research Network.

The results show that the genetics of the cancer are complex but do provide new genetic driver leads that will help to better classify the disease and to stratify AML patients by risk. “This study was designed to identify potential new drivers,” said Timothy J. Ley, MD, of the Genome Institute at Washington University and Siteman Cancer Center in St. Louis, and lead author of the study.  While somatic mutations have been known to drive AML, as many as half of AML patients do not have detectable mutations in known genetic genes that drive AML. “By identifying both rarely mutated genes that cluster together in pathways or gene sets, as well as recurrently mutated genes, we are narrowing down the potential driver list for functional studies,” Ley added.

The putative driver mutations from this research can now be tested for their role in AML development, a process that will take many more years.

Ley and colleagues analyzed 200 samples from patients with de novo AML, representing the major cytogenetic and morphological subtypes of the disease. Whole-genome sequencing was performed on the primary tumor samples and matched normal skin samples from 50 of the patients, and exome sequencing was performed on the remaining 150 patient samples and matching normal skin samples.

“This is a very different [genetic] landscape from that of epithelial tumors,” Ley told Cancer Network. “[There are] fewer mutations in the genome compared to that of most solid tumors and less evidence of genomic instability.” 

The study identified at least one potential driver mutation in each of the 200 patient samples and 23 genes that were frequently mutated, including known drivers of AML such as DNMT3A, FLT3, NPM1, IDH1, IDH2, and CEBPA. Twenty-eight percent (56 samples) had a mutation the FLT3 gene, which is a target of several drugs in development for treatment of AML.

While 10% of the AML samples had a mutation in TP53, which was associated with a more complex cytogenetics profile, the overall mutation burden in these TP53-mutated leukemias was not greatly different compared to patients with normal karyotypes, said Ley.

The researchers are now in the process of sequencing thousands of archived samples from around the world to understand the frequency of the recurrently mutated genes found in the current study. Ley believes that this will lead to a set of rules for patient risk-assessment. While there are generic biomarker approaches used to evaluate an AML patient’s risk profile, none have proven completely accurate suggesting that further understanding of the genetic aberrations in AML (which this study partly provides) are needed. Three types of DNA fusions are associated with a favorable-risk profile-these patients tend to respond to current chemotherapy treatment regimens. Those with a poor-risk profile who tend to have complex genetic alterations are more likely to relapse after initial treatment and require an allogeneic transplant.

The main question is how to classify those patients who fall in between a favorable- and high-risk profile. Many of these normal karyotype patients do well on treatment but some have poor outcomes. Even before using newly validated driver mutations to develop new medications, associating patterns of mutations with risk could facilitate treatment decisions-whether a patient should receive a transplant up front or is a candidate for chemotherapy first is necessary, said Ley.

The current study also revealed strong links between mutations, messenger RNA, and micro RNA expression patterns and epigenetic patterns. These patterns were not previously known and researchers do not yet understand their meaning. “These results suggest regulatory relationships and important biology that will need to be clarified experimentally in the future,” said Ley. “There is lots to do with this data.”