Three recent epigenetic studies of patients with acute myelogenous leukemia (AML) reveal new disease subtypes and outcome predictors based on methylation patterns of abnormal blast cells. AML, known to vary widely by biology, clinical prognosis, and outcome, is characterized by relatively few genetic mutations that do not explain the clinical and molecular diversity.
Because epigenetic changes occur at relatively high frequency, in contrast to DNA mutations, the hope, researchers say, is that epigenetic research can help better define disease characteristics, determine prognosis, and measure drug efficacy in clinical trials. Epigenetic changes can also be pharmacologically modified without changing genetic sequence.
In cancer, measuring gene expression and detecting genetic mutations is not enough, said Ari M. Melnick, MD, Director of the Raymond and Beverly Sackler Center for Biomedical and Physical Sciences as well as the Epigenomics Core Facility at Weill Cornell Medical College in New York City, lead author of one of the studies (Figueroa ME et al: Cancer Cell 2010;17:13–27). “Cancer is an epigenetic disease, not only a genetic disorder.”
Epigenetic marks including DNA methylation, histone modifications, and microRNAs control the timing, levels, and even the splicing of genes. Applications of genetic research in cancer have been limited until now, but the three studies represent progress in moving basic epigenetics research to the clinic.
The first author of the second study (Blood 2010;115:636-6), Lars Bullinger, MD, a hematology-oncology fellow at the University of Ulm in Germany, explained that in AML, as in other cancers, DNA methylation plays a critical role, with implications for diagnosis and treatment. Epigenetic research in leukemia is a bit ahead of work in most solid tumors except for colorectal cancer, in part because of the ease of obtaining cells in blood vs obtaining tumor samples, which also have greater heterogeneity.
“Leukemias are characterized by fusion proteins and genetic translocations, many of which also have epigenetic causes,” said Manel Esteller, MD, PhD, Director of the Center for Cancer Epigenomics and Biology Program at the University of Barcelona, coauthor of the third study, a poster presented by Sara Alvarez, MD, PhD, at the most recent American Society of Hematology Annual Meeting (Abstract 2394).
Taken together, the new data show that understanding epigenetic changes in AML can give clinicians a clearer, more complete picture of how the disease develops, can help determine which patients have more resistant or less resistant disease, and can point to better ways of treating both types of patients and gauge the efficacy of treatments.
DNA Methylation Profiling
While the researchers each used a different method of methylation profiling, all found new clinical biomarkers from epigenetic characteristics of AML patients that had not been apparent from genetic analyses.
Dr. Melnick's team measured the DNA methylation of 14,000 genes in 344 AML patients. The results showed that patients were clearly grouped into 16 different subtypes based on their epigenetic profiles, each with a unique and extensive aberrant DNA methylation signature. Five of the groups represent previously unknown, new AML subtypes, three clusters were identical to WHO-classified AML subtypes, and eight were associated with specific genetic or epigenetic lesions.
Each AML subtype displayed unique epigenetic signatures when compared with normal bone marrow CD34+ cells. In most cases, the difference between normal CD34+ cells and AML cells consisted mainly of hypermethylation, but unexpectedly, a few subgroups were predominantly hypomethylated as compared with normal.
“This was a surprise, since up to now it was thought that tumors are practically always hypermethylated as gene promoters,” Dr. Melnick said. “One of the implications of this finding is that for certain subgroups of patients, epigenetic drugs that target DNA hypermethylation may not be indicated.”
Deregulating Signal Pathways
Not surprisingly, in each of the 16 subgroups, aberrant DNA methylation signatures affected important cellular signaling pathways, including the p53 tumor suppressor, cytotoxic T-cell-mediated apoptosis, damage repair, and immunodeficiency signaling.
Another important finding was the discovery of a set of 45 genes that were consistently deregulated in almost all AML patients regardless of the subgroup, the majority of which were hypermethylated. This set could be used as a diagnostic panel to confirm AML.
Unlike genetic mutations which are seen in relatively few AML patients, these epigenetic lesions, or “epi-mutations,” occurred almost universally. Dr. Melnick hypothesizes that deregulation of these 45 genes may be necessary for normal hematopoietic cells to become leukemia cells.
In addition to the 45 deregulated genes, 15 genes were identified that accurately predicted overall patient survival independently of other known biomarkers. The next step will be testing patients prospectively to stratify them according to epigenetic signatures in trials, possibly including trials with methyltransferase inhibitors, he said.
Dr. Bullinger and his colleagues discovered new AML subgroups by examining methylation patterns associated with outcome and developed a methylation-based outcome model. This study assessed a much smaller number of gene promoter regions than Dr. Melnick's, and correlated them with patient prognoses.
Using mass spectrometry, Bullinger et al screened 92 genomic regions of 182 patient samples for promoter methylation, and compared them with 74 AML controls. Two main groups of genes were identified, the larger characterized by low methylation levels, and the smaller, with higher levels.
The researchers also found that a number of genetic abnormalities have an impact on epigenetic translocation profiles. “Our results provide a proof of principle that epigenetic markers can be used as surrogates for patient outcomes in AML,” Dr. Bullinger said.
Dr. Esteller and colleagues used high-throughput methylation array profiling to find epigenomic variations in 116 patients whose AML also displayed biologic and clinical diversity. Two major methylation patterns: one with minimal DNA methylation changes and another with major aberrant DNA methylation.
Interestingly, the first group corresponded to the poor-prognosis group, as defined by traditional cytogenetics, and the second one included 80% of the good-prognostic group.
“Why this is so is not yet understood,” Dr. Esteller said. “One explanation is that DNA methylation can affect other parts of the genome, such as inactivation of microRNAs, which may be important and which these arrays don't detect.”
One unanswered question, he added, is whether those leukemias with more hypermethylation alterations are those that will respond better to treatment with DNA demethylation agents.
Not the Whole Picture
Writing in an editorial accompanying Dr. Melnick's study, Dr. Bullinger noted that measuring methylation alone is not enough to determine the entire picture of epigenetic changes and to know which changes are relevant to clinical outcome. “Aberrant gene expression in leukemia is influenced by many factors in addition to DNA methylation,” he and Scott A. Armstrong, MD, PhD, wrote.
Dr. Esteller agreed: “Current epigenetic studies examine only about 10% of the genome, but in the future, using the new technique of “deep”- or “ultra”-sequencing, we will be able to study the remaining 90%, such as noncoding microcoding RNAs.”
“The field of epigenetics is still in its infancy, Dr. Melnick observed. “We don't even fully understand yet how DNA encodes. The first step to fully understanding how epigenetic changes work in AML is to study normal lymphocytes.”
Such an effort is now under way, Dr. Esteller said, begun by the recently established International Human Epigenomic Consortium.