Background
Acute myeloid leukaemia (AML) is the third most common form of leukaemia in children, typically characterised by the rapid proliferation of primitive haematopoietic myeloid progenitor cells[
1]. Paediatric AML is a highly heterogeneous disease, which presents a major barrier towards the development of accurate disease classification, risk stratification and targeted therapies within the clinic. The French-American-British (FAB)[
2] and more recently World Health Organisation (WHO)[
3] classifications of leukaemia take into account cell morphology, cytogenetic aberrations and common genetic lesions. However, not all patients fall into these well-defined categories. Additionally, the recurrent chromosomal and genetic lesions frequently found in AML fail to induce leukaemogenesis and do not explain the recognised clinical heterogeneity[
4,
5].
One of the hallmarks of nearly all human cancers is the disruption of the epigenetic profile, including gross aberrations in DNA methylation. Increasing evidence in adult AML has indicated that epigenetic events play critical roles in the onset, progression, and outcome of AML[
6] and may help tailor disease treatment. However, the need for similar elucidations in childhood disease is paramount. Aberrant methylation of cytosine residues at palindromic CpG sites (often clustered in dense CpG ‘islands’) near gene promoter regions is widely studied in carcinogenesis and haematological malignancies[
6,
7]. It is now well established that elevated DNA methylation is an important mechanism of gene transcriptional inactivation[
8,
9] and genes such as
ESR1,
IGSF4 and
CDKN2B/p15 are epigenetically silenced in adult leukaemia[
6]. Previous studies have subdivided adult AML into 16 epigenetic sub-groups based on DNA methylation signatures, correlating with patient clinical outcome and distinct from both normal haematopoietic cells and normal stages of myeloid differentiation[
4]. Despite such emerging findings in an adult context, the utility of individual DNA methylation disruptions in paediatric AML has yet to be fully evaluated[
6].
MicroRNA (miRNA) represent an alternative epigenetic regulator, having been implicated in the regulation of critical gene expression networks in plants and animals. The role of miRNA in haematopoiesis, cancer and disease is also beginning to be appreciated[
10,
11]. The global influence of individual miRNA on the genome is difficult to dissect, as miRNA can modulate the expression of hundreds of genes, and each gene can harbour binding sites for several miRNA[
12]. Human miRNA are initially transcribed (pri-miRNA), and processed by several complexes to form a 70 bp hairpin-loop (pre-miRNA)[
13]. After successive enzymatic steps, a miRNA:miRNA* complementary duplex is formed where the ‘functional’ strand is combined with RISC (RNA Induced Silencing Complex) and Argonaute proteins to guide, and inhibit, specific target messenger RNA (mRNA) through base pair recognition[
14,
15]. However, the miRNA transcriptome is becoming increasingly complex, emphasised by Next Generation Sequencing (NGS) technologies. NGS has highlighted that alternate miRNA* transcripts, as well as miRNA sequence variants (isomiRs[
16]) may play a biological role, similar to their canonical miRNA relatives[
17,
18].
Links between miRNA deregulation and cancer diagnosis were first identified in adult Chronic Lymphoblastic Leukaemia (CLL), where the loss or down-regulation of tumour-suppressing miRNA cluster
miR-15a/16-1 directly caused leukaemic transformation[
19,
20]. At present, no such association has been identified for childhood leukaemia. The expression of paediatric disease-associated miRNA has to date only identified a distinction between leukaemia of different lineages and the differentiation of rearranged AMLs within a limited number of cytogenetic subtypes[
21,
22]. Paediatric
MLL can be distinguished from others by differentially expressed miR-126, miR-146a, miR-181a/b/d, miR-100, miR-21, miR-196a/b, miR-29 and miR-125b[
21]. However concordance among studies is often low and the mechanism of deregulation is often unknown[
22,
23].
Genes encoding miRNA can be regulated epigenetically in a similar manner to protein coding genes[
22]. Studies have demonstrated epigenetically regulated miRNA in adult AML, including hypermethylation and down-regulation of miR-124a and associated deregulation of target mRNA
EVI1, CEBPA and
CDK6 independent of diagnostic cytogenetic subtype (reviewed in[
22]). Additionally, miR-193a targeting
KIT, and miR-14b targeting
CREB have been identified in adult investigations as specifically controlled by DNA methylation (reviewed in[
22]). However, the identification of DNA methylation and miRNA expression connections in paediatric leukaemia is lacking.
Paediatric AML has distinct cytogenetic and clinical features relative to their adult counterparts[
5,
21,
24‐
26]. Therefore, there is a critical need to improve our understanding of the biology of childhood leukaemia as separate entities, distinct from adult disease. Cognisant of this, we aimed to identify differential DNA methylation within paediatric AML on a genome-scale using defined clinical subtypes and age-matched controls. We identified a number of significantly altered DNA methylation loci, with associated gene and miRNA expression change, between paediatric AML and non-leukaemic counterparts. Specifically we describe here the epigenetic deregulation of
DLEU2, which has associated alterations in downstream
miR-15a/16-1 miRNA cluster expression.
Acknowledgements
We thank Dr Elizabeth Algar and Priscilla Siswara for access to flash frozen bone marrow and gene abnormality identification, and Dr. Pei Tian for facilitating sample procurement. Dr. Jovana Maksimovic we thank for HM450 data analysis.
Financial support
This work was supported by the Victorian Cancer Agency Grant and National Health & Medical Research Council (NHMRC), Australia [Project Grant number 607306] (N.W.); MCRI is supported by the Victorian Government’s Operational Infrastructure Support Program; the National Health & Medical Research Council Dora Lush Postgraduate Scholarship to L.M.; NHMRC Senior Research Fellowship to R.S.; The Leukaemia Foundation of Australia Phillip Desbrow Postdoctoral Fellowship (N.W.) and Project Grant in Aid( N.W. & R.S.); N.W. was also supported by My Room and the Children’s Cancer Centre Foundation.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
MH and MP-B carried out cell sorting, extractions and the accumulation of samples. JN assisted in accumulation of samples, extractions, conversions and assay running, as well as preliminary data analysis. FM and NG facilitated sample procurement and conceived the study. ZC participated in statistical and data analysis, study concepts and helped to draft the manuscript. RS and NW conceived the study, acquired the samples, interpreted the data, participated in drafting the manuscript and provided critical revisions for the approved final version. LM designed the study, acquired the samples, organised sample processing, participated in sample processing, undertook all data analysis and statistical evaluations and drafted the manuscript. All authors read and approved the final manuscript.