Essential role of PR-domain protein MDS1-EVI1 in MLL-AF9 leukemia

Acute myeloid leukemias bearing mixed-lineage leukemia (MLL) fusion proteins (MFPs) have a poor prognosis; chemotherapy is inadequate, indicating the need for more effective therapies.^1-3  Insights into the molecular pathogenesis will greatly facilitate the development of targeted agents.

In a subset of MFP AMLs, specifically those lacking monocytic features,⁴  MFPs can bind to and activate transcription of MECOM,^4-6  a highly conserved protooncogene encoding MDS1-EVI1 (ME), and EVI1 isoforms via distinct transcription sites (see Figure 1A). Relative to EVI1, ME possesses a PRDI-BF1-RIZ1 (PR) domain with histone methyltransferase (HMT) activity.⁷  In this study, using mouse alleles where ME is constitutively (ME^m1) or conditionally (ME^fl4) lost,⁸  we reveal the PR domain as being essential in MFP transformation in mice. These results strongly suggest that ME is a novel target for therapeutic intervention.

ME^m1 and ME^fl4 (Evi1^fl3 previously) alleles⁸  and Esr-Cre⁹  have been described. All procedures were approved by the animal care and use committees of the involved institutions.

The MLL-AF9 expression construct¹⁰  was moved to mCherry.¹¹  pMIGR1-ME (p1183) was made by inserting MDS1⁸  into BglII-cut p854.¹²  pMIGR1-MDS (p1211) was made by polymerase chain reaction (5′atgaattcgcatgagatccaaaggcagggc3′/5′atgaattcttcacctggtctcccatccatagctg3′) and cloning.

Lineage-negative/Sca-1⁺/c-kit⁺ (LSK) cells were isolated (t = 0), infected⁸  with retrovirus (t = 15 hours), sorted (t = 63 hours), and plated in M3434 (StemCell). For Figure 1F, add-back infections were performed following MLL-AF9 sort (t = 64 hours), sorted for green fluorescent protein (GFP) (t = 90 hours), and plated in M3434.¹³ 

Primary leukemias were harvested and explanted to culture, ^+/− 4-OH tamoxifen (TAM); 10⁶ spleen cells per mouse were transplanted into sublethally irradiated secondary recipient mice via tail vein injection. Leukemia development in transplanted mice was monitored by assessing complete blood counts (CBCs) every week.

To test the role of ME in MLL-AF9 leukemogenesis, we infected bone marrow cells from ME^m1/m1 mice,⁸  which bear a lacZ insertion in exon 1 of Mds1 (Figure 1A) and lack MDS1 and ME but express normal levels of EVI1, and performed a serial replating assay.^13-16  Although MLL-AF9 induced wild-type (WT) cells to form colonies at each cycle, it was unable to transform ME^m1/m1 cells (Figure 1B). These findings indicate a dependency of MLL-AF9 transformation on functional Mds1 and/or ME.

We then tested if the transformation block was specific for MLL-AF9: ME^m1/m1 cells were transduced with MLL-AF9, Nup98-HoxA9, MLL-ENL, E2A-Hlf, or Meis1-HoxA9 and were assayed for transformation. This revealed that cells lacking ME were resistant to transformation only by MLL-AF9 or MLL-ENL (Figure 1C). Thus, the block to transformation is oncogene specific. To further test this finding, ME^fl4/m1/Esr-Cre (Figure 1D) and ME^fl4/+/Esr-Cre (Figure 1E) LSK cells were transduced with the same oncogenes, split into 2 treatment groups (vehicle or 4-OH TAM), and assayed for transformation. All vehicle-treated cell samples, regardless of genotype, were fully capable of being transformed with all oncogenes (Figure 1D-E), whereas 4-OH TAM–treated ME^fl4/m1/Esr-Cre cells displayed the same selective resistance to transformation; 4-OH TAM–treated ME^fl4/+/Esr-Cre cells were susceptible to transformation by all oncogenes. Together, these data indicate a selective requirement for a functional ME allele for transformation by MFP oncogenes.

To test which isoform of Mecom is required for MLL-AF9 transformation, we retrovirally transduced LSK cells from ME^m1/m1 mice with MIGR1, or with constructs for MDS1, EVI1, or ME, and then infected them with MLL-AF9 retrovirus; transduced cells were assayed for transformation. Although neither MDS1 nor EVI1 was able to rescue the deficiency, ME was able to (Figure 1F), confirming that in the context of the ME^m1/m1 genotype, ME is the essential isoform that is lacking. Because the major difference between ME and EVI1 is the PR domain, it is clear that this domain is critical for ME activity in the setting of MLL-AF9–induced leukemogenesis.

To test if ME is required for transformed MLL-AF9 AML cells to survive in vivo, we established transplantable leukemias and then deleted the gene and assayed for leukemogenesis by transplantation (Figure 2A). ME^fl4/m1/Esr-Cre or ME^fl4/+/Esr-Cre LSK cells were transduced with MLL-AF9 and transplanted into irradiated recipients. After 3 months, primary AMLs developed, and leukemic spleen cells were harvested, induced with 4-OH TAM to delete ME (yielding genotypes ME^ko4/m1 and ME^ko4/+, respectively), and injected into irradiated recipients. Weekly CBCs noted an increasing leukocyte count and persistently low hematocrit and platelet count in mice receiving ME^fl4/m1/Esr-Cre cells not treated with 4-OH TAM (Figure 2B-D), indicating infiltration of leukemic cells into the blood and bone marrow; in mice receiving 4-OH TAM–treated ME^fl4/m1/Esr-Cre cells, a normal leukocyte level was maintained, and platelets and hematocrit recovered to normal levels. The control group mice receiving ME^fl4/+/Esr-Cre cells with and without 4-OH TAM treatment all develop leukemia as expected (not shown).

Four weeks posttransplant, mice receiving ME^fl4/+/Esr-Cre cells with and without 4-OH TAM treatment, and mice receiving ME^fl4/m1/Esr-Cre cells untreated with 4-OH TAM became moribund; in contrast, the cohort receiving ME^fl4/m1/Esr-Cre cells treated with 4-OH TAM remained healthy. Necropsy confirmed that the moribund mice had widely disseminated disease, with enlarged spleens (Figure 2E-F), and, on histopathology, leukemic cell infiltration of liver (Figure 2G), bone marrow (Figure 2H), and spleen (not shown). In contrast, mice receiving 4-OH TAM–treated ME^fl4/m1/Esr-Cre leukemic cells remained healthy and had normal-sized spleens at necropsy (Figure 2E-F), with no or minimal AML infiltration into organs, as assessed by histopathology (Figure 2G-H).

Bindels et al showed overexpression of EVI1 in a subset of MFP leukemias; those expressing EVI1 were phenotypically distinct in that they rarely showed monoblastic phenotype.⁴  Thus, it appears that not all cases of MFP-induced leukemias express the MECOM locus and that there is a distinct phenotype (monoblastic) when MECOM is not activated. It is likely that MECOM nonexpressing MFP-induced leukemias arise via transformation of a cell that is beyond the hematopoietic stem cell/common myeloid progenitor (HSC/CMP) stage, at which the MECOM locus is normally silenced.¹⁷  Arai et al established that MFPs can upregulate transcription of both EVI1 and ME.⁵  Although our results reveal an essential role for ME, they do not preclude that EVI1 isoforms are also essential. Nor do they exclude the possibility that it is the ratio of ME to EVI1 that is critical.

Currently, there are no clinically available targeted therapies for MFP leukemias. Published reports suggest several targeted therapies for these leukemias: inhibitors of glycerol synthase kinase,¹⁸  DOT1L,¹⁹  and the MFP–multiple endocrine neoplasia 1 interaction.²⁰  However, none of these has yet made it to the clinic, and furthermore, each is likely toxic, based on interpolation from genetic studies.^21-23  Thus, despite these reports, there is still a need for additional avenues of therapeutic intervention. This study reveals the novel discovery we have made that ME is necessary for MFP transformation.

The fact that ME but not EVI1 can rescue the mutant phenotype centers attention on the PR domain as being essential for ME function in the setting of MFP leukemogenesis. Recent studies have shown that this domain harbors H3K9 monomethyltransferase activity,⁷  which was expected based on the finding of HMT activity in closely related proteins.²⁴  This may open an avenue for novel therapeutic intervention in the treatment of MFP leukemias. The fact that ME is a nonessential gene for cellular and organismal survival⁸  suggests that therapies that inhibit its function are likely to be nontoxic and well tolerated. Future studies will be focused on further characterization of the PR domain and inhibition of its HMT activity.

The authors thank Johan Jansson and Katie Brock for technical help, and Hongbo Yu for histology.

This work was supported by grants from the National Institutes of Health, National Cancer Institute (R01CA120313) (A.S.P.) and New York State Stem Cell Science (C026423) (Y.Z.) (C026406) (A.S.P).

Contribution: Y.Z., F.C., and A.S.P. designed the study; Y.Z., K.O., K.K., L.H., and F.C. performed the study; and Y.Z., C.H.G., and A.S.P. wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Archibald S. Perkins, Department of Pathology and Laboratory Medicine, University of Rochester, Box 626, 601 Elmwood Ave, Rochester, NY 1464; e-mail: archibald_perkins@URMC.rochester.edu; and Yi Zhang, Department of Pathology and Laboratory Medicine, University of Rochester, Box 626, 601 Elmwood Ave, Rochester, NY 14642; e-mail: yi_zhang@urmc.rochester.edu.