The Identification of Epigenetic Biomarkers for Myeloid Malignancies

Abstract

Myeloid malignancies comprise a heterogeneous group of blood cancers that arise from abnormal stem or progenitor cells of the bone marrow and often portend a poor prognosis. The major disorders can be divided into four subsets based on clinical and genetic analyses: myeloproliferative neoplasms (MPN), myelodysplastic syndromes (MDS), mixed MPN/MDS, and acute myeloid leukemias (AML). Assigning accurate diagnosis can be quite challenging, and few reliable biomarkers are available for diagnostic, prognostic, and therapeutic purposes. As such, the purpose of this study was to elucidate useful biomarkers in various subsets of myeloid malignancies. It is now well established that epigenetic changes, such as aberrant DNA methylation, underlie all stages of cancer evolution. Epigenetic changes are reversible, so, in theory, cancer cells may be “reset” by targeted treatment with drugs that modulate the epigenetic system. Furthermore, clinical trials using demethylating agents such as azacytidine and decitabine have yielded promising results in hematopoietic malignancies. Intriguingly, acquired chromosomal lesions, primarily large genomic deletions have major clinical prognostic value at diagnosis, and appear to predict response to epigenetic therapies, but the mechanisms are completely unknown. Using microsatellites and SNP arrays, we identified common regions of loss, gain, uniparental disomy (UPD), and loss of heterozygosity (LOH) in subsets of myeloid malignancy patient samples. Common regions of LOH were found on chromosome 7 in AML and MDS and common regions of loss were found on chromosomes 13 and 20 in the classical myeloproliferative disorder myelofibrosis (MF). As shown in other studies, we also found common UPD on chromosome 9, including the JAK2 locus in MPDs. We have interrogated these global regions of change, analyzing the common genes through rigorous bioinformatic analyses to find candidate tumor suppressor genes. We focused on functional pathways that were connected to genes in the common genomic regions, including Wnt, Ras, Jak/STAT, and DNA repair. We reasoned that one gene abnormality may only define a small subset of patients; however, alterations in a common pathway may accurately define the majority of a disease subset. This represents the hypothesis generation component of our analysis. We present early validation studies of our candidate genes. We have preliminary evidence that SOCS family members are methylated in MPNs, offering an explanation for cases with no JAK2 mutation. Additional results investigating signaling, DNA repair, and cell cycle regulatory pathways will be presented across multiple myeloid malignancy samples. Further studies and collaborations are required to validate these targets and pathways in large numbers of patient samples. Common pathway analysis of epigenetic alterations, in concert with genetic alterations, will undoubtedly lead to a better understanding of the pathogenesis of cancers and provide useful biomarkers for diagnosis, prognostic stratification, and drug studies, with the ultimate goal of improving patient survival.