Mapping of MN1 Sequences Necessary for Myeloid Transformation

The MN1 oncogene is deregulated in human acute myeloid leukemia and its overexpression induces proliferation and represses myeloid differentiation of primitive human and mouse hematopoietic cells, leading to myeloid leukemia in mouse models. To delineate the sequences within MN1 necessary for MN1-induced leukemia, we tested the transforming capacity of in-frame deletion mutants, using retroviral transduction of mouse bone marrow. We found that integrity of the regions between amino acids 12 to 458 and 1119 to 1273 are required for MN1’s in vivo transforming activity, generating myeloid leukemia with some mutants also producing T-cell lympho-leukemia and megakaryocytic leukemia. Although both full length MN1 and a mutant that lacks the residues between 12–228 (Δ12–228 mutant) repressed myeloid differentiation and increased myeloproliferative activity in vitro, the mutant lost its transforming activity in vivo. Both MN1 and Δ12–228 increased the frequency of common myeloid progentiors (CMP) in vitro and microarray comparisons of purified MN1-CMP and Δ12–228-CMP cells showed many differentially expressed genes including Hoxa9, Meis1, Myb, Runx2, Cebpa, Cebpb and Cebpd. This collection of immediate MN1-responsive candidate genes distinguishes the leukemic activity from the in vitro myeloproliferative capacity of this oncoprotein.


Introduction
Acute myeloid leukemia (AML) in adults is a leading cause of leukemia-related deaths, and is characterized by uncontrolled proliferation and impaired differentiation of hematopoietic cells that results in accumulated myeloid blasts in the bone marrow and periphery. [1] Tight control of the balance between proliferation and differentiation is essential for the maintenance of normal hematopoeisis. In various AML subtypes, deregulation of different or sometimes overlapping genes disrupts this balance and causes AML. [1,2] These genes mostly control the survival/proliferation/differentiation programs of the hematopoietic stem/progenitor cells (HSPC). [1,2].
MN1 (Meningioma 1) is located on human chromosome 22 and encodes a 1319 amino acid (aa) long protein, which is unique as it does not show homology to any known proteins. [3] MN1 is involved in AML either as a partner of the t(12;22)(p12;q12), creating an MN1-TEL fusion protein, [4] or as an overexpressed gene. [5][6][7] About half of patients with AML carry leukemic cells with a normal karyotype [8] in which elevated MN1 expression correlates with poor prognosis. [9] In addition, increased expression of MN1 cooperates with CBFb-SMMHC [7], NUP98-HOXD13 [10] and MLL-ENL [11] fusion proteins to induce leukemia, suggesting that deregulation of MN1 expression contributes to leukemogenesis. Indeed, others and we have shown that ectopic expression of MN1 in mouse HSPC (Hematopoietic Stem-Progenitor Cells) causes myeloid leukemia [7,12] and MN1 induces proliferation and inhibits myeloid differentiation of both mouse and human HSPC. [13] The differentiation inhibitory and proliferative effects of MN1 can be prevented by re-introduction of CEBPA. [13].
Although the transforming capacity of MN1 is well established, the molecular mechanisms and pathways that regulate its leukemogenic activity remain elusive. We hypothesized that identification of the domains within MN1 contributing to its leukemic activity, and determination of the gene expression profiles of cells that ectopically express MN1 or a mutant lacking leukemic activity, could provide more in-depth information about the genetic programs involved in MN1-induced myeloid leukemia. Here, we mapped the regions of MN1 that confer myeloproliferative, differentiation-inhibitory and leukemogenic activity on mouse HSPC. In addition, we could distinguish expression profiles associated with MN1's myeloproliferative and differentiation inhibitory effects from its myelo-transforming activity. This was accomplished by comparing the transcriptome of highly purified common myeloid progenitors (CMP) overexpressing MN1 or a MN1 deletion mutant, which induced myeloprolifera- Immuno-cytochemical analysis of the above cells using the same MN1-antibodies (red). Blue shows nuclear staining with 4,6-diamidino-2-phenylindole (DAPI). Images were captured with an Olympus BX-50 microscope (equipped with UPlanFL 406/0.75 numerical apertures with a SPOT camera and SPOT Advanced Imaging software of Diagnostic Instruments. The original magnification was x400. (D, E) FACS analysis showing CD11b expression of indicated U937 stable cell lines 3 days after treatment with vehicle, vitamin D 3 (Vit-D3) or ATRA (mean 6 SEM of duplicates; *P,.05, ***P,.001, # not significant). doi:10.1371/journal.pone.0061706.g001 tion and prevented myeloid differentiation in vitro, but did not cause leukemia in mice.

Ethics Statement
This study was carried out in accordance with the recommendations in the Guide for Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Institutional Animal Care and Users Committee of St Jude Children's Research Hospital (Approved protocol Number: 209-100171-04/12). All efforts were made to minimize suffering.

Generation of MN1 and MN1-mutant Cell Lines
The U937 cell line [15] was maintained and transduced with the corresponding MIG viruses for two days as described (2.5610 5 /well, 12 well plate). [13] GFP + cells were sorted using FACS and differentiation was induced with vitamin-D 3 (100 nM; D1530, Sigma-Aldrich, St. Louis, MO, USA), all-trans retinoic acid (ATRA, 1 uM) (R2625, Sigma-Aldrich) or the same amount of vehicle (100% Ethanol) for 3 days. Differentiation was assessed using FACS analysis of CD11b expression.

Western Blot and Immunocytochemistry
Total cell lysates were prepared in 2% SDS containing protease inhibitors (Sigma) followed by three 15-second pulses of sonication. Lysates (from 4610 5 cells) were separated on 4-20% stain free Tris-HCL polyacrylamide gel (PAGE), (Biorad, Hercules, CA, USA) and transferred to polyvinylidene diflouride membranes (Millipore, Bedford, MA, USA). Membranes were incubated with anti-human MN1 antibodies recognizing either the N-terminus [7] or C-terminus of the protein (sc-27349; Santa Cruz, Santa Cruz, CA, USA) followed by incubation with horseradish peroxidase (HRP)-conjugated secondary antibodies. Protein bands were visualized with chemiluminescence (G&E Healthcare Biosciences Pittsburgh, PA, USA). A stain free gel imaging system (Biorad) or Anti-human ACTIN antibody (4968; Cell Signaling, Danvers, MA, USA) were used as a control for equal gel loading.
Immunocytochemistry of the cells using anti-MN1 antibody was performed as described. [7].
For differentiation assays, cells were expanded for 96 hours after the last retroviral transduction and subjected to a second lineage depletion using the mouse hematopoietic progenitor enrichment kit (Catalog #19756; Stemcell Technologies, Vancouver, BC, Canada) following the manufacturer's instructions. Cells were maintained in growth medium and marker analysis was performed at Day-0 and Day-2 after the second lineage depletion using FACS.
In a parallel experiment, cells were transduced and maintained in liquid culture in growth medium and seeded at the same density 24 hours after the last transduction (2610 6 cells/well in 4 ml in a 6 well plate). Cells were split (2610 6 cells/well in 4 ml) 3 times a week and the growth potential of GFP + cells over that of the GFP 2 cells in the same culture was measured by performing a periodic GFP analysis using FACS at the indicated time points. Methylcellulose assays with GFP + sorted cells were performed as described. [13].

Microarray Analysis
Two days after the first transduction (after total 5 days in culture) with the indicated retroviruses, mouse CMP/GFP + cells were collected into TRIzol using FACS and RNA was isolated according to the manufacturers' protocol. RNA was isolated from bulk bone marrow cells of 2 leukemic MN1-mice using TRIzol and used in microarray experiments. RNA quality was confirmed by analysis on the Agilient 2100 Bioanalyzer. Total RNA (100 ng) was processed in the St Jude microarray core according to the Affymetrix GeneChip eukaryote two-cycle target labeling protocol (http://media. affymetrix.com/support/downloads/manuals/expression_analysis_ technical_manual.pdf). Biotin-labeled cRNA (6 mg) was added to a hybridization cocktail and then processed automatically on an HT MG-430 PM array plate using the Affymetrix GeneTitan system. The HT MG-430 PM array contains 45,037 probesets that interrogate more than 39,000 mouse transcripts. Normalized transcript measures were generated from scan intensity files using the RMA algorithm. [17] A two-fold threshold was applied to identify differentially expressed transcripts among cell populations. Probeset annotations were obtained from the Affymetrix NetAffx website (http://www.affymetrix.com/analysis/index.affx). Gene Ontology and pathway enrichment analysis was performed using the DAVID bioinformatics resources [18] (http://david.abcc.ncifcrf. gov/). Gene Set Enrichment Analysis (GSEA) [19] was also performed using canonical pathways obtained from the Broad Institute (C2), Ingenuity Pathways and GeneGo Pathways.  The microarray data of this study have been deposited in GEO under accession no GSE38767.

Bone Marrow Transplantation and Analysis of Mouse Tissues
Wild type C57Bl6 mice (8-12 weeks old) were used as donors and recipients. Lineage depleted bone marrow cells were isolated, cultured, and transduced as described above. At day 5 of culture, which included the retroviral transductions, cells were harvested, washed once with PBS and diluted in PBS containing 100 U/ml heparin (APP Pharmaceuticals, Schaumburg, IL USA). Lethally irradiated mice (950 rad at 60 rad/min with a cesium irradiator) received transplants 24 hours after irradiation of 5610 5 unsorted cells per mouse (tail vein injections). Mice were maintained on Baytril-water for 21 days, starting immediately after irradiation. Complete blood counts were performed with a Hemavet 3700 (Drew Scientific). Peripheral blood samples were analyzed for the presence of GFP using FACS at the indicated times. Tissues were collected, processed, and stained with antibodies against GFP (A11122; Invitrogen) and MPO (A0398; DAKO Carpinteria, CA, USA), and analyzed as described. [20] Cells were also analyzed for expression of CD3, GATA1, and B220 (data not shown).

Statistical Analysis
Two-way or one-way ANOVA analysis, Kaplan-Meier survival plots and comparison of Kaplan-Meier plots (log-rank (Mantel-Cox) test) were performed using the GraphPad Prism, version 4.0 c for Mac (GraphPad Software; www.graphpad.com).

Identification of MN1 Sequences Inhibiting Myeloid Differentiation of U937 Cells
MN1 causes acute myeloid leukemia when overexpressed in mouse HSPC. [7,12] Its overexpression also caused impaired myeloid differentiation, increased proliferation, and increased selfrenewal of primary mouse or human CD34 + -HSPC in vitro. [13] Based on data of MN1 sequences essential for the transforming activity of the MN1-TEL fusion protein [21], we first tested the oncogenic activity of MN1 deletion mutants D12-228, D18-458 and D570-1273 ( Figure 1A). To ensure that mutants would localize to the nucleus [14], we added a SV40-NLS to all MN1 constructs. Addition of the NLS did not affect MN1 function as overexpression of full length MN1 or MN1-NLS similarly impaired vitamin-D 3 or ATRA induced monocytic or granulocytic differentiation, as judged by expression of the differentiation marker CD11b ( Figure 1D). We confirmed the ectopic expression and nuclear localization of each MN1 mutant in FACS-sorted GFP + -U937 cell lines using western blot ( Figure 1B) and immunocytochemistry (IC) ( Figure 1C). The three deletion mutants showed a differentiation inhibitory effect with the D12-228 mutant showing the strongest inhibition (1.2 and 1.3-fold better differentiation than MN1-cells, in response to vitamin D 3 and ATRA, respectively) and the D18-458 mutant the least inhibition (2.2 fold improvement in vitamin D 3 response and complete recovery of the ATRA response)( Figure 1E). This indicated that aa residues between 12-228 contributed to but were not essential for the differentiation block observed in U937 cells.

MN1 Increases the Frequency of CMP in vitro
Mouse HSPC transduced with retroviruses encoding full length or the three MN1 deletion mutants were analyzed for their in vitro differentiation potential in response to cytokines. FACS analysis performed 2 days after the first transduction showed that, relative to mock transduced cells, the percentage of CMP cells in the GFP + /Lin 2 /c-Kit + /Sca1 2 fraction in both MN1 and D12-228expressing cells was increased, albeit more modestly in the latter (Figure 2A). This increase slightly enlarged the downstream MEP population, but decreased the respective downstream GMP populations by 10.4% and 4.2%, respectively (Figure 2A). In contrast, cells expressing the D18-458 mutant showed the same CMP, MEP and GMP distribution as mock transduced cells (Figure 2A).
To test the differentiation potential of transduced cells beyond the GMP stage, we repeated the same experiment and included the D570-1273 mutant. Following transduction and a 96 hours expansion of the cells, a second lineage depletion was performed (referred to as Day-0) to purge cells that had differentiated during the transduction and expansion process. FACS analysis of Day-0 cells showed similar low levels of Mac-1 and Gr-1 expression in all samples, indicating efficient depletion of maturing myeloid cells from each culture ( Figure 2B). Differentiation of the cells was evaluated 48 hours later (Day-2). FACS results of Day-2 MN1cells or MN1-mutant-cells compared with mock-cells did not show any increase in the Sca-1 + /c-Kit + population, which is enriched in hematopoietic stem cells ( Figure 2C). As expected, 50% of mocktransduced cells expressed myeloid differentiation markers (Mac-1, Gr-1 or Mac-1/Gr-1) at Day-2, which was 2.7 fold higher than at Day-0 ( Figure 2B). In contrast, only 12% of MN1-transduced cells and 26% of D12-228-cells expressed these markers at Day-2. The D18-458 and D570-1273 cells showed fully restored differentiation with 47% and 51% of cells expressing myeloid differentiation markers, respectively ( Figure 2B). The GFP-negative population in each culture served as an internal control and expressed equal levels (47%-57%) of myeloid markers in all samples at Day-2 ( Figure 2B). Collectively, D12-228 cells displayed a less pronounced inhibition of myeloid differentiation than MN1 cells, while D18-458 or D570-1273 cells showed no inhibition of differentiation in this assay. Mapping of Sequences Essential for MN1-mediated Myeloproliferation and Self-renewal in vitro Given that sequences in the N-terminal (aa 18-458) and Cterminal (aa 570-1273) regions of MN1 were required to impair the myeloid differentiation of primary mouse HSPC, we tested if these same regions were important for MN1-mediated myeloproliferative activity. Liquid culture of MN1 and D12-228-transduced mouse HSPC for 14 days showed increased proliferation, outgrowing the co-cultured uninfected cells as shown by sequential FACS analyses over time. The percentage of GFP + cells increased by 3.2-fold in the MN1-transduced cultures (from 28% at Day-0 to 91% at Day-14), by 2.7-fold in D12-228-transduced cultures (35.8% at Day-0 and 97.2% at Day-14) but did not increase in the D18-458, D570-1273, or mock-transduced cultures ( Figure 3A). Previously we showed that forced MN1 expression increases proliferation of mouse [7] and human HSPC in vitro. [13] Our current results suggest that sequences within aa 18-458 and aa 570-1273 are essential for this MN1-induced effect.
In addition to abnormal growth and impaired differentiation, acquisition of increased self-renewal capacity by HSPC is essential for the development of leukemia. Therefore, we evaluated the in vitro self-renewal potential of the MN1-deletion mutants using a methylcellulose (MC) serial-replating assay. Consistent with the in vitro proliferation assays, D12-228-cells recapitulated the effect of full length MN1 and showed increased replating and proliferation capacity, as was judged by both the increased number and size of the colonies, whereas the replating activity of the cells transduced with vector, D18-458 or D570-1273 mutants was exhausted after the initial MC culture ( Figure 3B). Giemsa staining and differential counts of cytospin preparations from the MC-1 plates demonstrated a 5.6-fold and 4.2-fold increase in the number of promyelocytes in MN1 and D12-228-cultures, respectively ( Figure 3C and 3D). Similar to the liquid culture results ( Figure 2B), MN1 cells did not mature into neutrophils (1% in MN1 vs. 45% in mock; Figure 3D). The D12-228 mutant partially reproduced this effect, which, compared with mock-cells, showed a 2-fold inhibition of maturation into neutrophils ( Figure 3D). In contrast, D18-458 and D570-1273 cells showed a moderate increase in the number of mature neutrophils compared to mock-transduced cells ( Figure 3D). Therefore, both MN1 and D12-228 increased the pool of proliferating promyelocytes and myelocytes at the expense of the pool of mature metamyelocytes and neutrophils ( Figure 3E). In conclusion, deletion of aa 12-228 of MN1 moderately decreased its ability to block myeloid differentiation in methylcellulose assays whereas the self-renewal (replating capacity of colonies) and proliferative activity of MN1 were maintained. Thus, sequences comprised within aa 18-458 and aa 570-1273 of MN1 are necessary to induce aberrant self-renewal, increased proliferation, and impaired differentiation of primary mouse HSPC in vitro.

Sequences in both the N-terminal and C-terminal Regions of MN1 are Required for its Leukemogenic Activity
The MN1-TEL1 fusion protein only contains MN1 sequences encoded by its first exon but not the 60 aa encoded by its second exon [14]. Therefore we tested the oncogenic activity of the first exon of MN1 (D1260-1319) versus that of full-length MN1 in our in vivo leukemogenesis assay. Mice receiving transplants with MN1 or D1260-1319 transduced bone marrow both developed AML with or without maturation between 35-52 and 30-62 days post transplant, respectively ( Figure 4A). Next, we tested the in vivo transforming capacity of the D12-228, D18-458 and D570-1273 MN1-mutants. Periodic FACS analysis of the GFP content of peripheral blood samples showed successful engraftment in each group ( Figure 4B and Table 1). Mice receiving transplants of MN1-transduced bone marrow developed myeloid leukemia within 48-88 days (median survival 71 days) (P,0.0001, log-rank test; Figure 4C). However, mice receiving mock and D18-458 transduced bone marrow remained disease free till the end of the observation period (6 months), while 3 of 7 mice receiving D570-1273 bone marrow developed GFP 2 T cell lymphoma ( Figure 4C, see below). Despite its increased self-renewal and myeloproliferative activity in vitro, mice receiving transplants of D12-228expressing bone marrow (transduction rate of the transplanted cells were similar to that of full length MN1, Table 1) did not develop leukemia in two independent experiments ( Figure 4C). One of 8 mice in D12-228-group was found dead at 131 days after transplantation but the cause of death could not be determined due to autolysis of the tissues, and was therefore omitted from the study.
To further delineate the domains in MN1 essential for leukemogenesis we generated additional deletions in the 18-570 and 570-1273 regions. Given the results with the D12-228 mutant we resided to testing the leukemogenic potential of the additional mutants in vivo only. To further dissect the 570-1273 region, we generated the D570-950 ( Figures 5A, 5B), D570-1010 ( Figures 5A,  5B), D570-1080, D570-1109, D570-1119, D570-1175 and D570-1209 deletion mutants ( Figure 5A) and tested their leukemogenic activity in mice receiving transplants of bone marrow transduced with each mutant. With the exception of the latter two, all mutants caused myeloid leukemia between 30 and 130 days post transplant. Increasing the size of the deletion did not affect the type of AML, which in all cases was with or without maturation. Seven mice expressing mutant MN1 developed T cell lymphoma (one D578, one D570-1119, two D570-1175 and three D570-1273 mice) while one (D570-1080) developed both T-cell lymphoma and myeloid leukemia. Although the D570-1175 and D570-1273 mutants developed T cell lymphoma only, it is currently unclear if there is a direct correlation with mutant MN1 expression because four lymphomas were GFP + (one D578, one D570-1119, one D570-1175 and the one D578 mixed leukemia/ lymphoma) while the others were GFP 2 (data not shown). Immunohistochemical analysis of selected leukemic bone marrow samples marking the transplanted cells (GFP staining) and myeloid cells (Myeloperoxidase (MPO)) are shown in the Figure 5C.

Expression Profiling of Purified CMP Populations
We next sought to distinguish the genetic program mediating MN1's role in increased self-renewal/proliferation and inhibition of differentiation in vitro from that mediating MN1's transforming activity. Given that ectopic expression of MN1 increased the frequency of CMP in vitro, we performed a microarray analysis using FACS-sorted CMP + /GFP + cells expressing vector, MN1, the D12-228, or the D18-458-deletion mutant. We used DAVID bioinformatics database (http://david.abcc.ncifcrf.gov/) to func- tionally categorize differentially expressed genes. Pair-wise comparisons for each of the three samples using vector-CMP cells as baseline ( Figure 6A, 6B and 6C) with a threshold of .2-fold difference was used to identify differentially expressed transcripts. As shown in the Venn diagram in Figure 6D, 2038 transcripts were differentially expressed in MN1-CMP only and 1265 transcripts were differentially expressed in both MN1 and D12-228-CMP, while 1110 of these were regulated in the same direction. Therefore, these 2038 MN1-specific transcripts are potentially involved in MN1-induced myeloid leukemia, whereas the 1110 overlapping transcripts between MN1-cells and D12-228-cells may play a role in MN1-induced enhanced myeloproliferation and impaired differentiation of murine bone marrow cells. To identify similarities between the immediate MN1-induced expression profiles and those in leukemic MN1-mice, we determined the differentially expressed transcripts in two bulk leukemic MN1-bone marrow samples using the mock-CMP profile as baseline. Further comparison of CMP-MN1-specific transcripts with the leukemic bone marrow transcripts showed that 502 of 2038 MN1-specific transcripts were differentially regulated in the same direction in the leukemic cells ( Figure 6E). Table 2 lists selected transcripts representing potentially important genes. We have also used GSEA to analyze our microarray data. Using canonical pathways (C2 from Broad, Ingenuity, and GeneGo) we did not obtain significant differences (FDR,0.05) between cells transduced with MN1 or the MN1 deletion mutants. Comparison of MN1 vs. GFP comparison indicated that only the BMP2 pathway was significantly activated in MN1 cells ( Figure S1). The microarray data of this study have been deposited in GEO under accession no GSE38767.

Discussion
In this study we mapped the minimal regions of MN1 necessary for its leukemogenic activity as well as the regions of MN1 that distinguish its leukemic in vivo activity from its in vitro myeloproliferative/self-renewal and differentiation inhibitory activities. Furthermore, we identified early changes in the expression profiles of MN1-CMP cells that presumably contribute to a transformed stage, by performing microarray analysis of highly enriched FACS-sorted-CMP cells. We found that 24.6% of differentially expressed genes in MN1-CMP cells were also regulated in the same direction in leukemic bone marrow cells of MN1-mice.
Our study indicates that integrity of two regions of MN1, comprised within residues 18-458 and 1119-1273, are required for its myeloid leukemic activity, and presence of only one of these regions in the absence of the other is insufficient for in vivo transformation. Via which mechanism these regions contribute to the transforming activity of MN1 will be a question that we aim to answer in future studies. We favor the explanation that the deletions remove domains of MN1, which interact with partner proteins that are necessary for its leukemogenic activity, but we cannot exclude other possibilities such as the deletions altering the conformation of the protein thereby inactivating it. As mentioned above, it is currently unclear if some of the deletion mutants of MN1 directly generate T-cell lymphoma, given that 4 out of 8 cases were GFP + and the remaining 4 cases GFP 2 . Because we were interested in MN1's function as a myeloid disease gene, we did not investigate this issue further and will address this observation in a future study. It is intriguing that the D458-560+D570-1119 mutant was unique in that it also caused GFP + megakaryocytic leukemia, indicating a correlation between the presence of aa 458-560 and a switch from myeloid to megakaryocytic leukemia.
We hypothesized that deletion of residues 12-228 of MN1 would affect the transcriptional activation of targets that modulate MN1-induced leukemogenesis, given that this mutant does not exhibit in vivo leukemic activity. We reasoned that comparison of the transcriptomes of cells expressing full length MN1 or a nonleukemogenic MN1-mutant could provide a more refined insight into the genetic program causing MN1-induced myeloid leukemia. Given that MN1 overexpression expanded the CMP population in vitro (Figure 2A), and transforms CMP but not GMP cells in vivo [22], we used microarray analysis of highly purified CMP cells overexpressing MN1, D12-228 or D18-458 to identify crucial transcriptional targets of MN1.
Microarray expression profiling identified 2038 transcripts that discriminate MN1-CMP cells from two of the mutants (D12-228 or D18-458) that do not exhibit in vivo transforming activity. Hoxa9, Meis1, Mll, Runx2, and Myb were among these differentially upregulated genes ( Table 2). All these genes play a well-documented role in the transformation of myeloid cells and are frequently involved in myeloid leukemia. [23][24][25][26] Interestingly, the activity of Meis and Hox genes renders mouse CMP cells susceptible to MN1 overexpression [22], while MN1 (ref [7]) or Runx2 (ref [26]) overexpression cooperates with Cbfb-SMMHCinduced mouse AML. Moreover, important regulators of myeloid differentiation including Cebp-a, Cebp-b, and the M-CSF and GM-CSF receptors were down regulated in both MN1-CMP and D12-228-CMP populations ( Table 2). Disruption of Cebp-a in mice blocks the transition from CMP to GMP, which not only impairs granulopoiesis, but also enhances the self-renewal activity of hematopoietic stem cells. This generates increased numbers of myeloblasts in the bone marrow but does not cause leukemia [27] alone, but does collaborate with other events in the leukemogenic process. [28] Nonetheless, Cebp-a deficient cells can generate granulocytes in response to IL3 and/or GM-CSF in a Cebp-bdependent manner. [29] Therefore, besides the many genes that are deregulated in mouse MN1-CMP-cells immediately after transduction, it is tempting to speculate that combined altered expression of Hoxa9/Meis1/c-myb and Cebps are at least partially responsible for the MN1-induced leukemic transformation. Insufficient expression of Cebp's and other myeloid differentiation factors (Table 2) in D12-228-cells may impair differentiation and enhance proliferation in vitro. However, compared with MN1-cells D12-228-cells show reduced Hoxa9/ Meis1/c-myb expression, which despite its induced reduction in differentiation factor expression, may render this mutant inadequate to induce leukemia in mice ( Figure 4C). The higher expression of Hoxa9/Meis/Myb in MN1-cells most likely also contributes to the more pronounced differentiation defect of these cells (Figures 2 and 3). [25] Downregulation of CEBP-a expression is an important aspect of MN1-imposed changes in CD34 + human HSPC, given that forced Cebp-a expression (both U937-MN1 and HSPC-MN1) reversed the MN1-induced defects in myeloid differentiation and proliferation in vitro. [13].
In our transplantation experiments we used unfractionated and unsorted mouse HSPC that were transduced with MN1-retrovirus. Analysis of the bone marrow samples of leukemic MN1-mice showed a dramatic increase in lin 2 (8% vs. 50% and 62% in MN1) and c-Kit + /Sca1 2 cells (13% vs. 38% and 70% in MN1) ( Figure 7A) when compared to that in wild type mice, which resulted in an increase in absolute numbers of MEP, CMP and GMP cells. Distribution of these three populations in the bone marrow of MN1-mice was skewed towards the GMP ( Figure 7A). This suggests that MN1 more severely impairs terminal differentiation downstream of the GMP in vivo, giving rise to accumulation of cells that are arrested at that stage, which is in agreement with our in vitro findings ( Figure 2B).
Future studies that will identify the proteins interacting with MN1, the domains within MN1 required for their interaction, and validation of the role of these potential players in MN1-induced leukemia will further increase our understanding of the precise role of MN1 overexpression in AML. These studies will also form the basis for the design of targeted therapeutic approaches to better treat patients overexpressing MN1 in their leukemic blasts. Figure S1 (TIF)