Epigenetic and Transcriptomic Profiling of MLL Rearranged Infant Acute Lymphoblastic Leukemia

2017 
Abstract [Background] Infant acute lymphoblastic leukemia (ALL) accounts for less than 5% of childhood leukemia, and is characterized by Mixed lineage leukemia (MLL also called KMT2A ) rearrangements found in approximately 80% of patients. Also found in 5-10% of ALL in older children and adults, MLL -rearrangement positivity is highly associated with poor outcome in ALL; there is an urgent need for molecular findings of potential therapeutic targets and/or risk stratification markers, which can be translated into the clinical practice. Recently, the mechanisms that MLL fusion proteins recruit epigenetic regulators, including histone H3 lysine 79 methyltransferase DOT1L and BET family histone readers, to activate the transcription of target genes have been reported, suggesting epigenetic dysregulation should play a central role in the leukemogenesis. However, DNA-level methylation profiles of MLL -rearranged leukemia and their associations with gene expression profiles are not fully characterized. [Materials and Methods] Our cohort included samples from the Infant Leukemia Sub-committee, Japan Children's Cancer Group. Both methylation array and RNA sequencing were performed on the diagnostic leukemia samples from 38 infants with ALL, consisting of 30 cases with MLL -rearrangements identified by fluorescence in situ hybridization or Southern blotting, and 8 MLL -germline cases. Methylation beta values were normalized with beta-mixture quantile normalization method. For RNA sequencing data analysis, we used publicly available Genomon Pipeline. [Results] RNA sequencing confirmed MLL -rearrangements in 30 cases, whose partner genes comprised AFF1 (n=15), MLLT1 (n=6), MLLT3 (n=3), MLLT10 (n=2), EPS15 (n=1), AFF2 (n=1), ABI1 (n=1), and SARNP (n=1). As expected, DNA methylation-based hierarchical clustering divided the cases into 2 clusters with highly distinctive methylation profiles: MLL -rearranged cluster (MLL -R cluster) and MLL -germline cluster (MLL -G cluster). Intriguingly, 3 out of 30 (10%) MLL -rearranged cases were classified into MLL -G cluster; these cases revealed uncommon partner genes, such as ABI1 , SARNP , and MLLT10, respectively. ABI1, an Eps8 SH3 domain-binding protein, and SARNP, a SAP domain containing ribonucleoprotein, are rare fusion partners of MLL. The MLL -germline leukemia-like methylation profiles suggested different leukemogenic mechanisms from common fusion partners, such as AFF1 and MLLT1, which directly or indirectly interact with DOT1L and positive transcription elongation factor b. Because DOT1L-binding domain of MLLT10 was retained in both of the 2 cases with MLL-MLLT10 in our cohort, additional mechanisms causing MLL-germline leukemia-like methylation in a single case may be present. As a result of inter-cluster comparison, MLL -R clusterexhibited global hypermethylation compared to MLL -G cluster (P -value [Conclusion] Corresponding to known aberrant histone modifications in MLL -rearranged leukemia, largely different DNA methylation profiles between MLL -R and MLL -G clusters have been demonstrated. Because DNA-level methylation on CpG dinucleotides interactively affects histone modifications, these results implicated aberrant DNA methylation in the pathogenesis of MLL -rearranged leukemia as a part of epigenetic dysregulation. We also identified a subset of MLL -rearranged cases with MLL -germline leukemia-like methylation/expression profiles, indicating that not only MLL -rearrangement positivity, but also the types of MLL -rearrangement partners may affect a biological feature of leukemic cells. These cases may possibly be therapeutically stratified as relatively lower risk group compared to other MLL -rearranged cases. Further detailed analyses using clinical information are planned. Disclosures No relevant conflicts of interest to declare.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    0
    Citations
    NaN
    KQI
    []