Altered microRNA (miRNA) expression is frequently observed in acute myelogenous leukemia (AML) and has been implicated in leukemic transformation., somatic mutation of miRNA genes appears to be rare in AML and cancer in general. We recently reported whole genome or exome sequencing of 200 cases of de novo AML (The Cancer Genome Atlas, NEJM 2013). Recurring point mutations of only one miRNA gene were identified. Specifically, heterozygous point mutations of MIR142 were identified in 3 cases and bi-allelic mutations were identified in 1 case (total incidence of 2%). The miRNA “seed” sequence, located at positions 2-8 from the 5’ end of the mature miRNA, is critical in mediating the specificity of miRNA-mRNA target interaction. Of import, all AML associated MIR142 point mutations localized to this critical seed region in the mature miRNA-142-3p sequence leading to the prediction that these mutations alter normal miRNA-142-3p function. Surprisingly, when we transiently over-expressed MIR142 mini-genes containing these mutations, we observed decreased expression of miR-142-5p, suggesting that the mutations in the seed sequence of miR-142-3p are affecting miRNA processing of the hairpin. To explore this possibility, we sequenced the small RNA transcriptome of 28 total cases of de novo AML, including the four cases harboring MIR142 point mutations. As a control, we also sequenced CD34+ progenitors from healthy donors (n = 4). In general, miRNA-142-5p is expressed at higher levels than 3p (average ratio of miRNA-142-5p/3p: 1.8 ± 0.23). However, in the three cases with heterozygous MIR142 mutations, a significantly decreased expression of miRNA-142-5p relative to 3p was observed (0.41± 0.13). Most strikingly, in the AML case with bi-allelic MIR142 mutations, miRNA-142-5p levels were markedly reduced. Of note, bioinformatic target prediction programs show the loss of many predicted miRNA-142-3p targets upon mutation. However, target prediction programs did not show any commonly gained mRNA targets between the three identified AML associated miRNA-142-3p mutants. Collectively, these data suggest that these MIR142 mutations are unlikely to be gain-of-function but probably lead to a complete loss of normal miRNA-142-3p function in addition to decreased processing or stability of miRNA-142-5p.
Our small RNA sequencing data also revealed marked over-expression of miRNA-142 in our AML patient cohort. Compared with control CD34+ cells, miRNA-142-5p and 3p levels were increased 16.34 ± 3.6 and 30.92.34 ± 7.0 fold, respectively. Thus, we next examined the effects of enforced expression of wild-type or mutant MIR142 on hematopoiesis. C-kit+ murine hematopoietic progenitors transduced with lentivirus expressing wild-type or mutant MIR142 were transplanted into lethally irradiated recipients. Surprisingly, over-expression of wild type miRNA-142 resulted in a complete loss of repopulating activity as early as 6 wks post transplant. In contrast, hematopoietic progenitors transduced with mutant MIR142 were able to generate multi-lineage engraftment for at least 7 months post transplant. These data suggest that miRNA-142 over-expression in AML may be compensatory to transformation and serve to inhibit leukemic cell growth. It follows that the loss-of-function MIR142 observed in AML may disrupt this negative feedback loop, thereby promoting AML expansion. Studies to characterize the miRNA-142 target repertoire within the hematopoietic system are currently underway and should provide insights on how disruption of miRNA-142 function may contribute to the pathogenesis of AML.
No relevant conflicts of interest to declare.
