Mir-10a Regulates Autophagy and Cell Cycle in Normal Karyotype Acute Myeloid Leukaemia

Aim: Normal karyotype (NK) Acute Myeloid Leukemia (AML) accounts for approximately half of all AML cases. Almost two-thirds of these carry the Nucleophosmin 1(NPM1) gene mutation (NPM1^c+). MicroRNAs (miRNA/miRs), short RNA molecules that act as post-transcriptional regulators of gene expression, have multiple roles in regulating key gene signalling networks in cancer, including in NK-AML. miR-10a is significantly overexpressed in NK-AML compared to normal bone marrow, particularly in those carrying the NPM1^c+mutation in which it is expressed at levels ~20-fold higher than in NPM1 wildtype (NPM1^wt) NK-AML. We have previously shown that repression of miR-10a in the NPM1^c+ AML cell line OCI-AML3 leads to inhibited proliferation and initiation of apoptosis. Here we aimed to further delineate the role of miR-10a in NK-AML.

Results: miR-10a was shown to directly bind to and downregulate key members of the p53-mediated tumour suppressor gene network, including the CDKN1A (p21) inhibitor Transcription Factor AP2-gamma (TFAP2C) and Retinoblastoma (RB1) regulator Rb1-Inducible Coiled-Coil 1 (RB1C11). Inhibition of miR-10a resulted in significant repression of the cell cycle inhibitor p21, upregulation of RB1CC1, decrease in the expression of autophagy marker LC3-II, and the induction of apoptosis in serum-deprived cells. The individual modulation of these miR-10a target genes via siRNA (CDKN1A) and overexpression (RB1CC1) had similar effects, triggering apoptosis and also sensitizing OCI-AML3 cells to chemotherapy agents cytarabine and daunorubicin in transfected cells.

Due to miR-10a's potential role in regulating the central p53-p21-Rb1 tumour suppressor signalling axis in AML, miR-10a levels were quantified by QPCR in primary NK-AML patient samples and correlated to clinical outcome. Analysis of 61 patients from the German SAL AML2003 randomized clinical trial who received post-induction chemotherapy + allogeneic bone marrow transplantation (BMT) in a risk-adapted manner demonstrated that high miR-10a expression (> median expression) was associated with significantly poorer relapse-free survival (hazard ratio (HR) = 5.6, p < 0.001) compared those patient samples with low miR-10a expression. This is consistent with these patients being more resistant to post-induction chemotherapy and/or upfront BMT, and suggests that these patients may benefit from alternate therapy.

To investigate the mechanism of miR-10a dysregulation in NPM1^c+ AML cells, we assayed methylation of the miR-10a promoter in NPM1^c+ (OCI-AML3) versus NPM1^wt(HL60, MV4) cell lines. Consistent with the observed upregulation of miR-10a in NPM1^c+ AML cells, hypomethylation of the miR-10a promoter was observed in OCI-AML3 cells (<5% CpG methylation) compared to the HL60 and MV4 (60% and 92% CpG methylation, respectively; p < 0.01), which was mirrored in NPM1^c+ and NPM1^WT primary patient samples. Exposure of the NPM1^WT AML cells to hypomethylating agents decitabine and azacitidine resulted in reduced DNA methylation of the miR-10a promoter and a corresponding 20-100-fold upregulation of miR-10a (p < 0.001 for both lines). There was no significant association between miR-10a promoter hypomethylation with mutations of epigenetic modifiers such as DNA Methytransferanse 3A (DNMT3A; which are highly enriched in NPM1^c+ AML) in primary patient specimens.

Conclusion: Overexpression of the pro-survival miRNA miR-10a in NPM1^c+ AML is epigenetically regulated, and results in modulation of the p53-p21-Rb1 tumour suppressor signalling axis in NK-AML. This may have clinical significance and identifies miR-10a as a potential novel biomarker and drug target for AML.