Bioinformatic Analyses of Broad H3K79me2 Domains in Different Leukemia Cell Line Data Sets

Abstract

A subset of expressed genes is associated with a broad H3K4me3 (histone H3 trimethylated at lysine 4) domain that extends throughout the gene body. Genes marked in this way in normal cells are involved in cell-identity and tumor-suppressor activities, whereas in cancer cells, genes driving the cancer phenotype (oncogenes) have this feature. Other histone modifications associated with expressed genes that display a broad domain have been less studied. Here, we identified genes with the broadest H3K79me2 (histone H3 dimethylated at lysine 79) domain in human leukemic cell lines representing different forms of leukemia. Taking a bioinformatic approach, we provide evidence that genes with the broadest H3K79me2 domain have known roles in leukemia (e.g., JMJD1C). In the mixed-lineage leukemia cell line MOLM-13, the HOXA9 gene is in a 100 kb broad H3K79me2 domain with other HOXA protein-coding and oncogenic long non-coding RNA genes. The genes in this domain contribute to leukemia. This broad H3K79me2 domain has an unstable chromatin structure, as was evident by enhanced chromatin accessibility throughout. Together, we provide evidence that identification of genes with the broadest H3K79me2 domain will aid in generating a panel of genes in the diagnosis and therapeutic treatment of leukemia in the future.