The Role of microRNA-155 in Mouse Models of MLL-AML

Background: Genetic rearrangements that fuse the mixed lineage leukemia (MLL) gene, now termed KMT2A, to one of a variety of partners are seen in 5% – 20% of acute myeloid leukemia (AML). MLL-fusions are especially common in childhood AML and many of them are associated with poor prognosis. The MLL-fusions perturb transcription through different mechanisms and they are often associated with high expression of the transcription factors HOXA9 and MEIS1. Based on a micro-RNA screen in an AML mouse model mimicking the step-wise development of aggressive AML we have found that concurrent Hoxa9 and Meis1 overexpression is associated with upregulation of micro-RNA-155 (miR-155). Expression of miR-155 was also found to be higher in bone marrow samples from patients with MLL-AML compared with bone marrow from healthy donors (p<0.001), as were the expression of HOXA9 and MEIS1 (p<0.05).

In lymphomas, miR-155 plays a pivotal role as an oncogene. It is frequently upregulated in samples from lymphoma patients and a mouse model of lymphoma showed a certain degree of miR-155-addiction which could be targeted by miR-155 inhibitors. Despite the differences in the pathobiology of AML and lymphoma, the upregulation of miR-155 in AML with high HOXA9 and MEIS1 expression may indicate miR-155 as a relevant therapeutic target also in MLL-AML.

Methods: To test the importance of miR-155 and its potency as a drug target in MLL-AML we used a miR-155 knock-out mouse model (miR-155^–/–) (Thai et al, Science, 2007). MLL-fusion genes of varying leukemic potential; MLL-AF5 (KMT2A-AFF4), MLL-ENL (KMT2A-MLLT1), MLL-AF9 (KMT2A-MLLT3) were retrovirally expressed in miR-155^–/– mouse bone marrow (mbm) cells and in wild-type mbm cells (miR-155^+/+).

Results: In concordance with the previous findings in human AML patient samples, miR-155^+/+ cells expressing MLL-AF5, MLL-ENL, or MLL-AF9 showed upregulation of miR-155 (p < 0.05). Also, Hoxa9 and Meis1 transcripts were increased (p<0.05). Interestingly, the magnitude of upregulation of both miR-155 and Meis1 correlated with the degree of aggressiveness based on disease latency and survival observed in these leukemia models with highest upregulation in MLL-ENL and MLL-AF9 and lowest in MLL-AF5 (p<0.05). Expression of the MLL-fusion genes in miR-155^–/– mbm cells resulted in similar induction of Hoxa9 and Meis1 expression as in miR-155^+/+ mbm cells, indicating that miR-155 is downstream of the Hoxa9/Meis1 axis.

To determine the leukemic potential in vivo, we transplanted recipient mice with miR-155^+/+ mbm cells and miR-155^–/– mbm cells expressing MLL-ENL or MLL-AF9. Engraftment of leukemic cells, based on peripheral blood analysis, did not differ between mice transplanted with miR-155^+/+ mbm cells and miR-155^-/- mbm cells expressing MLL-fusions. Also, disease development induced by MLL-AF9 and MLL-ENL (4–8 weeks and 10–32 weeks, respectively) was similar in mice transplanted with miR-155^–/– mbm cells and  mice transplanted with miR-155^+/+ mbm cells. In accordance with the in vivo results, functional studies in vitro showed that the proliferative capacity and colony forming ability of MLL-fusion expressing cells were similar in miR-155^+/+ mbm cells and miR-155^–/– mbm cells, indicating that miR-155 is not essential for MLL-ENL- or MLL-AF9-induced leukemic transformation.

Conclusions: In summary, miR-155 is upregulated in MLL-AML in both mice and man, seemingly through an MLL>HOXA9/MEIS1>miR-155 axis. Since absence of miR-155 does not alter the leukemic potential induced by MLL-AF9 or MLL-ENL, miR-155 may contribute to, but is not pivotal for MLL leukemogenesis. We therefore conclude that miR-155 is not a therapeutic target in MLL-AML.