MicroRNA (miRNA)-29b Targets DNMT3A and B and Induces Re-Expression of the Hypermethylated ESR1 and p15 Genes in Acute Myeloid Leukemia (AML).

Aberrant DNA methylation has been shown to play an important role in AML by silencing structurally normal genes important for hematopoiesis. MiRNAs are small non-coding RNAs that down-regulate gene expression. Although aberrant methylation of miRNAs has been reported, it is not known whether miRNAs themselves can modulate DNA methylation. To address this question, we tested the functional role of miRNAs that are predicted to target both DNA methyltransferase 3A and 3B (DNMT3A and B), in AML cells. We cloned the 3′UTR of both DNMT3s into a luciferase reporter and co-transfected K562 cells with scrambled or sense sequences of each of the miRNAs predicted to interact with DNMT3A and B. We found that miR-29b significantly repressed luciferase activity of both DNMT3A and B constructs (42% and 67%, respectively) compared with the controls. Transfection of the pre-miR-29b, but not of the scrambled, resulted in a marked reduction of the DNMT3A and B mRNA levels measured by qRT-PCR in MV4-11 cells (80% and 90%, respectively) and in K562 cells (23% and 67% respectively).We also observed a significant down-regulation of DNMT3s protein by immunoblotting in K562 cells transfected with miR-29b relative to the controls. These results were then validated in AML patients. Both miR-29b and DNMT3B expression levels were measured by qRT-PCR in bone marrow samples obtained from 12 AML patients with primary disease. We found that miR-29b expression was inversely correlated to DNMT3B mRNA levels (Pearson coefficient R=−0.66; P=0.03). Functionally, we showed that the mRNA expression of p15 and ESR1, that are silenced through promoter hypermethylation in AML, were increased in K562 and MV4-11 cells transfected with miR-29b compared with the controls. ESR1 was 1.9 and 1000 fold upregulated in miR-29b-tranfected K562 and MV4-11cells, respectively and p15 was 11.6 fold upregulated in miR-29b-tranfected MV 4–11 cells. Interestingly, we observed that the expression of miR-29b-1, which is silenced in K562 cell lines, was 6.5 fold up-regulated following treatment with 2.5 uM decitabine, a DNA hypomethylating nucleoside compared with untreated controls. Consistent with these results, we also showed 5.6 fold increase in miR-29b-1 expression in post-treatment CD34+ selected bone marrow blasts from AML patients (n=3) who received decitabine (20 mg/m2/day x 10 days) on a clinical protocol (OSU 0336) compared with pretreatment baselines. These results were consistent with the methylation analysis of CpGs up to 30kb upstream from the 5′ encoding regions of the miR-29b-1 precursors in the same patients. Using MassArray (sequenom) technology, we showed that all the CpGs were highly methylated (95 to 100%) in pre-treatment baseline samples and these methylation levels decreased by 20% after decitabine treatment. These preliminary data provide a functional link between miRNAs and aberrant epigenetics by suggesting that:

• miR-29b targets DNMT3A and B;

• miR-29b is silenced in AML cell lines and patient primary blasts and its expression is restored in vitro and in vivo by ypomethylating agents;

3) increase of miR-29b levels results in DNMTs down-regulation and re-expression of otherwise hypermethylated genes. Altogether, there data support a rationale for developing novel miRNA-based therapeutic strategies that alone and in combination with other hypomethylating agents may target aberrant epigenetics in AML.