LAUSR

MicroRNAs (miRNAs) are a class of small noncoding RNAs that have critical functions in gene silencing by binding to complementary mRNAs to induce their degradation or translational repression [1]. Interestingly, only a small fraction of miRNAs can exert their function in the form of miRNA-induced silencing complex (miRISC) [2]. Unfortunately, although dysregulated expression or processing of cancer-related miRNAs has been demonstrated to play a crucial role in oncogenesis, the contribution of miRNA dysfunction is still poorly understood. MLL-rearranged leukemias generally have a poor prognosis and account for 10% of overall acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML) cases [3]. MLL translocations encoding chimeric fusion proteins comprising the N-terminus of MLL in frame with various fusion partner proteins are characteristically found in MLL leukemias [3]. MLL is a histone H3 Lysine4 (H3K4) methyltransferase and is proteolytically cleaved into two distinct subunits, MLL and MLL, which noncovalently interact to form an intramolecular complex involved in epigenetic transcriptional regulation. Interestingly enough, we recently uncovered an unexpected role for MLL protein in miRNA-mediated translational repression [4]. Our studies showed that the MLL subunit alone could colocalize with miRISC components in cytoplasmic processing bodies (P-bodies), and affect the function of a subset of miRNAs such as the let-7 family [4]. Multiple miRNAs have been found to be dysregulated in MLL leukemias [5, 6]. However, since only a small fraction of miRNAs are functional, it is not clear whether the expression levels of miRNAs reflect their actual contribution to the pathogenesis of MLL leukemia. Considering that the abundance of wild-type MLL protein was reduced in MLL leukemic cells [7, 8], we reasoned that the function of a subset of miRNAs would be compromised and may play a critical role in the pathogenesis of MLL leukemia. We first addressed how P-body formation and miRNAmediated gene silencing were affected in MLL leukemic cells. Immunofluorescence results showed that the number of DDX6and DCP1A-marked P-bodies in MLL leukemic cells, including RS4;11, SEM, KOPN8, and THP-1 cells, was significantly fewer than in non-MLL leukemic lines such as JM1, REH, and U937 cells (Fig. 1a (i), and Supplementary Fig. S1a–c). Furthermore, the capacity of ectopically expressed let-7a or CXCR4 to silence their bulged miRNA reporters, but not perfect siRNA reporters, was markedly reduced in MLL leukemic cells (Fig. 1a (ii) and Supplementary Fig. S2a, b). These results were consistent with our previous findings that MLL was required for miRNA-mediated translational repression of partially matched mRNAs, but not for cleavage of perfectly matched mRNAs [4]. These defects in the MLL leukemic cells were associated with a reduced level of MLL but not P-body proteins (Supplementary Fig. S2c). Further introduction of MLL restored miRNA-mediated gene silencing in MLL leukemic cells (Supplementary Fig. S2d), indicating that the impairment in miRNA-mediated gene silencing in MLL leukemic cells was caused by downregulating wild-type MLL, especially the MLL subunit. These authors contributed equally: Shouhai Zhu, Xiaoyan Cheng, Ruiheng Wang, Yuting Tan