Mixed-lineage-leukemia (MLL) fusion oncogenes are closely involved in infant acute leukemia, which is frequently accompanied by mutations or overexpression of FMS-like receptor tyrosine kinase 3 (FLT3). Earlier studies have shown that MLL fusion proteins induced acute leukemia together with another mutation, such as an FLT3 mutant, in mouse models. However, little has hitherto been elucidated regarding the molecular mechanism of the cooperativity in leukemogenesis. Using murine model systems of the MLL-fusion-mediated leukemogenesis leading to oncogenic transformation in vitro and acute leukemia in vivo, this study characterized the molecular network in the cooperative leukemogenesis. This research revealed that MLL fusion proteins cooperated with activation of Ras in vivo, which was substitutable for Raf in vitro, synergistically, but not with activation of signal transducer and activator of transcription 5 (STAT5), to induce acute leukemia in vivo as well as oncogenic transformation in vitro. Furthermore, Hoxa9, one of the MLL-targeted critical molecules, and activation of Ras in vivo, which was replaceable with Raf in vitro, were identified as fundamental components sufficient for mimicking MLL-fusion-mediated leukemogenesis. These findings suggest that the molecular crosstalk between aberrant expression of Hox molecule(s) and activated Raf may have a key role in the MLL-fusion-mediated-leukemogenesis, and may thus help develop the novel molecularly targeted therapy against MLL-related leukemia.
Multistep oncogenesis has been suggested in malignancy by the observation of more than two heterogeneous genetic and/or epigenetic lesions.1 In leukemogenesis, recurring chromosomal translocations are frequently found in hematological malignancies, which sometimes coincide with subtle but critical genetic mutations leading to functional aberration.2, 3, 4 Earlier studies showed that many of the translocation target genes are transcription factors involved in hematopoietic differentiation and/or self-renewal, whereas coincident mutations often occur on the genes involved in cell proliferation.4 These results lead to a hypothetical model of leukemogenesis in which these two kinds of genetic alterations may cooperate to induce acute leukemia. This concept has been recently exemplified in experimental models using combinations of fusion genes, including mixed-lineage leukemia (MLL, also called ALL1 or HRX) or AML1 fusion genes, and other coincident genetic mutations.5, 6, 7, 8, 9
MLL is a proto-oncogene that is rearranged in human acute leukemia with chromosome 11 band q23 (11q23) translocation,10, 11 encoding a histone methyltransferase that assembles in a chromatin-modifying supercomplex.12 Meanwhile, MLL fusion gene leads to leukemogenesis through several HOX genes directly transactivated by MLL fusion protein itself.4, 11, 13, 14 It is noteworthy that most of the genetically engineered mice carrying the MLL fusion developed hematological malignancy after a long latency, suggesting that secondary genotoxic stress is required to develop overt acute leukemia.15, 16, 17, 18 An earlier study presented direct evidence that MLL fusion proteins induced myeloproliferative disease (MPD) with a long latency, and caused acute leukemia with a short latency together with a coincident mutation of FMS-like tyrosine kinase 3 (FLT3).6
Recent studies revealed that genetic alterations, including FLT3, NRAS (neuroblastoma RAS viral (v-ras) oncogene homolog) and KRAS (v-Ki-ras2 Kirsten rat sarcoma viral oncogene homolog), are frequently accompanied by 11q23 translocation.19, 20 FLT3 is a receptor tyrosine kinase involved in leukemogenesis and normal hematopoiesis.21 The mutations of FLT3 are mainly classified into length mutations such as internal tandem duplication (ITD) of the juxtamembrane domain, and point mutations within the activation loop of the second tyrosine kinase domain (TKD).21 Interestingly, FLT3-TKD, as well as overexpression of the wild type of FLT3, is found to be frequently associated with infant acute lymphoid leukemia (ALL), with rearrangements of MLL.19, 22 Both types of FLT3 mutations result in a constitutive activation of FLT3 kinase activity, followed by activation of signaling pathways, including signal transducer and activator of transcription 5 (STAT5) and Ras/Raf/mitogen-activated protein (MAP) kinase.23, 24 Both STAT5 and Ras/Raf/MAP kinase (MAPK) are involved in cellular proliferation, survival and differentiation.25, 26 Constitutively active mutants of Ras induce oncogenic transformation through activation of the MAPK cascade.26 However, little has so far been elucidated regarding the molecular mechanism of collaboration in leukemogenesis.
To further clarify the molecular mechanism of MLL-fusion-mediated leukemogenesis, we focused on signal transduction associated with malignant transformation that collaborates with MLL fusion protein in vitro, and highlighted the contrastive roles of STAT5 and MAPK in leukemogenesis. Interestingly, comparative analyses suggested synergistic collaboration with activated Ras in MLL-fusion-mediated leukemogenesis, and also activation of Raf in malignant transformation in vitro, but not with STAT5 activation in vivo and in vitro. Thus, the activation of Ras/Raf/MAPK cascade may have an important role in multistep leukemogenesis with 11q23 translocations.
Materials and methods
Construction of the plasmids and retrovirus production
Fragments of murine constitutively active mutants of STAT5A (#227 and 1*628) fused with a FLAG tag at the C-terminus, a coding region of human NRASG12V and MLL-eleven nineteen leukemia (ENL) short form6 were inserted upstream of the internal ribosomal entry site (IRES)-enhanced green fluorescent protein (EGFP) cassette of pMYs-IRES-EGFP.29 Fragments of coding regions of a wild type of NRAS and NRASG12V were inserted into pMXs-puro.29 A fragment of murine Hoxa930 (a kind gift from Dr G Sauvageau) was inserted into pMXs-IRES-EGFP.29 A fragment of a dominant negative mutant (dn) of STAT5A23 was inserted upstream of the IRES-Kusabira-Orange (KO)31 cassette of pMXs-IRES-KO, in which the EGFP cassette in pMXs-IRES-EGFP29 was replaced with the KO cassette of phKO1-S1 (MBL, Nagoya, Japan). pMXs-neo-MLL-SEPT6,6 pMY-FLT3-ITD-IRES-EGFP,6 pMY-FLT3D835V-IRES-EGFP6 and pBabe-puro-ΔRaf-estrogen receptor (ER)28 were described earlier. Retroviruses were harvested 48 h after transfection with each retroviral construct into PlatE cells29 in which appropriate expression of the transgenes was confirmed by western blot analysis, as described earlier.6
An MLL-SEPT6-immortalized murine myelomonocytic cell line, HF6, was established through colony-replating assays using retroviral transduction with pMXs-neo-MLL-SEPT6 as described earlier.6 A Hoxa9-immortalized murine myelomonocytic cell line, A9G, was established through infection with retroviruses harboring Hoxa9 in pMXs/IRES-EGFP29 as reported earlier.32 The HF6,6 A9G and murine pro-B Ba/F328 cells were cultured in the presence of interleukin-3 (IL-3) (R&D Systems, Minneapolis, MN, USA). HF6 cells transduced with FLT3 mutants were cultured in the same medium, except for the absence of IL-3. The expression levels of FLT3 in these cells were evaluated using a phycoerythrin (PE)-conjugated anti-CD135 antibody, or an anti-mouse immunoglobulin G1,κ, as the isotype-matched control (BD Biosciences, San Diego, CA, USA) using fluorescence-activated cell sorting (FACS) Calibur (BD Biosciences) as described earlier.33
Immunoprecipitation and western blot analysis
Fifty million parental and additionally transduced HF6 cells, or 10 million parental and transduced Ba/F3 cells, were harvested in the lysis buffer, and the lysates were either suspended with 1 × sodium dodecyl sulfate sample buffer after immunoprecipitation using polyclonal anti-STAT5A antibody (L-20) (Santa Cruz Biotechnology, Santa Cruz, CA, USA) or directly mixed with an equal volume of 2 × sodium dodecyl sulfate sample buffer and then boiled, as described earlier.25 In some experiments, the parental HF6 cells had been deprived of IL-3 8 h before harvest. Western blot analysis of each sample was performed using the polyclonal anti-STAT5A (L-20), monoclonal anti-phosphotyrosine (4G10) (Upstate Biotechnology, Lake Placid, NY, USA), polyclonal anti-extracellular signal-related kinase (ERK)1/2, monoclonal anti-phospho-ERK1/2 (E10) (Cell Signaling Technology, Danvers, MA, USA), monoclonal anti-α-tubulin (Sigma-Aldrich, St Louis, MO, USA), monoclonal anti-FLAG (M2), polyclonal anti-ERα (MC-20) and monoclonal anti-N-Ras (F155) (Santa Cruz Biotechnology) antibodies to probe membranes, as described earlier.25
Evaluation of cellular effects by inhibition of signal transduction in vitro
The response to the drug was evaluated as described earlier.23 In brief, HF6 cells expressing the FLT3 mutants (3 × 105) were infected with retroviruses harboring or not harboring the dnSTAT5A in pMXs-IRES-KO in the presence of polybrene, as described earlier.6 Viable cell numbers were counted with standard Trypan blue staining, and the expression of the dnSTAT5A was monitored by assessment of KO positivity using the FL2 channel on the FACS Calibur, daily after infection. At 48 h after infection, to evaluate the status of phosphorylated STAT5, half a million of these cells were fixed with fixation buffer, permeablized with Perm Buffer III and analyzed with an Alexa Fluor 647-conjugated anti-phospho-STAT5 (Y694) (all from BD Biosciences) antibody, or the anti-mouse immunoglobulin G1,κ, as the isotype-matched control antibody, using the FL4 channel on the FACS Calibur, according to the manufacturer's recommendation. As controls, the parental HF6 cells with and without IL-3 stimulation after deprivation of IL-3 for 8 h were used. Meanwhile, these HF6 cells (1 × 104) were cultured for 72 h in 24-well plates in the presence of various concentrations of a MAPK kinase (MEK) inhibitor, U0126, or a PI3 kinase inhibitor, LY294002 (Calbiochem-Novabiochem, San Diego, CA, USA) and each vehicle control (ethanol for U0126 and dimethyl sulfoxide for LY294002). Viable cell numbers were counted with standard Trypan blue staining after each treatment, followed by calculation of the 50% inhibitory concentration (IC50) of each drug using a logistic regression model. To evaluate the inhibitory effect of U0126 on ERK1/2, five million of the cells were treated for 2 h, harvested and analyzed with the anti-ERK1/2 or the anti phospho-ERK1/2 antibody after western blotting.
Myeloid transformation assays in vitro
In a series of transformation assays, the acquisition of IL-3-independent proliferation was examined in IL-3-dependent cells. HF6 and Ba/F3 cells were infected with retroviruses harboring NRAS, NRASG12V or mock in pMXs-puro; ΔRaf-ER or mock in pBabe puro; and STAT5A1*6, STAT5A#2 or none (only GFP) in pMYs-IRES-EGFP, respectively, in the presence of polybrene, as described earlier.6 A9G cells were also retrovirally transduced with NRAS, NRASG12V ΔRaf-ER or each mock in the same way. For puromycin selection, the transduced cells were cultured with 1 μg/ml of puromycin 24–96 h after infection, followed by propagation for 5 days in the absence of puromycin. Next, 1 × 105 puromycin-resistant cells transduced with NRAS, NRASG12V or mock were cultured in 24-well plates in the absence of IL-3, whereas those transduced with ΔRaf-ER or mock were cultured under the same condition, except for the presence of 1 μM of 4-hydroxy-tamoxifen or a vehicle control (ethanol). The cells transduced with STAT5A1*6, STAT5#2 or none were purified on the basis of the expression of GFP using a FACS Aria (BD Biosciences) 36 h after infection. Immediately, these purified cells (1 × 104) were cultured in 96-well plates in the absence of IL-3, to avoid excessive signals caused by STAT5A#2 or 1*6 in the presence of IL-3, which led to cell death as described earlier.25 Viable cell numbers were counted periodically after standard Trypan blue staining.
Leukemogenesis assays in vivo
Leukemogenesis assays in vivo using C57BL/6 mice produced by a combination of two kinds of transgenes were performed with lethal conditioning using lethally (9.5 Gy) irradiated recipients, or with sublethal conditioning using sublethally (5.25 Gy) irradiated recipients receiving no radioprotective bone marrow (BM) cells, as described earlier6 (Supplementary Figure 1). In brief, hematopoietic progenitors were harvested from 6- to 10-week-old Ly-5.1 C57BL/6 mice 4 days after intraperitoneal administration of 150 mg/kg 5-fluorouracil, and cultured overnight in alpha minimal essential medium supplemented with 20% fetal calf serum and 50 ng/ml each of mouse stem cell factor, human IL-6, human FLT3-ligand (R&D Systems) and human thrombopoietin (Kirin Brewery, Takasaki, Japan). The prestimulated cells were infected with several combinations of the retroviruses for 60 h in the α minimal essential medium supplemented with the same fetal calf serum and cytokines using RetroNectin (Takara Bio Inc., Otsu, Japan) according to the manufacturer’s recommendations, followed by intravenous injection of 105 of the cells into Ly-5.2 mice together with either a radioprotective dose (2 × 105) of Ly-5.2 cells under lethal conditioning or none under sublethal conditioning. Morbid mice and their tissue samples were analyzed, and immunophenotyping of BM, splenic and thymic cells was performed using the FACS Calibur, as described earlier.33 The hematopoietic neoplasms were diagnosed mainly on the basis of morphology as described earlier.6 The probabilities of murine overall survival were estimated using Kaplan–Meier method and compared using the log-rank test. All animal studies were performed in accordance with the guidelines of the Animal Care Committees of the Institute of Medical Science, the University of Tokyo and the Mie University.
Southern blot analysis
Genomic DNA was extracted from spleens, digested with NheI or BamHI for detecting proviral integration and clonality, respectively, and analyzed with the Neo or puro probe (Supplementary Figure 1) as described earlier.34
Reverse transcriptase-polymerase chain reaction (PCR)
Total RNA was extracted from cell lines, spleen or BM, and reverse transcribed to complementary DNA as described earlier.6 The conditions, reagents for reverse transcriptase-PCR and the primers specific for β2microglobulin (β2MG), Hoxa9 and MLL-SEPT6 have been described earlier,6 except that PCR amplification for MLL-SEPT6 transcripts was sometimes run for 35 cycles. To detect the transcript of NRASG12V, PCR amplification was run for 21 cycles using the following primers: NRAS-S, 5′-IndexTermGTGGTTATAGATGGTGAAACCTGTT-3′ and NRAS-AS, 5′-IndexTermGACCATAGGTACATCTTCAGAGTCCT-3′.
MLL-SEPT6 cooperates with both types of FLT3 mutations through different modes of signal transduction
To clarify the molecular mechanism of cooperation between MLL fusion proteins and FLT3 mutants, signaling pathways of FLT3-ITD and FLT3-TKD that cooperate with MLL-SEPT6 were examined using the IL-3-dependent MLL-SEPT6-immortalized cell line, HF6.6 Earlier, STAT5 and MAPK ERK1/2 had been found to be activated downstream of FLT3 mutants in factor-dependent cell lines.23, 24 Therefore, the activation of these molecules was first examined using parental HF6 and transformed HF6 cells expressing FLT3-ITD (HF6ITD) or FLT3D835V (HF6D835V) described earlier.6 Nearly equal levels of expression of the FLT3 mutants in the transformed HF6 cells were confirmed (Figure 1a). A western blot analysis after immunoprecipitation of the lysates from these cells revealed constitutive phosphorylation of STAT5A in HF6 cells expressing the FLT3 mutants in the absence of IL-3, but little in the parental HF6 cells that had been deprived of IL-3 (Figure 1b). In addition, a western blot analysis of the same lysates also revealed constitutive phosphorylation of ERK1/2 in those cells expressing the FLT3 mutants, but little in the parental HF6 cells that had been deprived of IL-3 (Figure 1b).
Next, to determine whether STAT5 and/or MAPK were important in the transformation of HF6 cells expressing FLT3 mutants, each signaling pathway was inhibited using dnSTAT5A or MEK inhibitor U0126. After retroviral transduction with the dnSTAT5A, the proliferation of HF6ITD cells expressing dnSTAT5A was suppressed more efficiently than that of HF6D835V cells expressing dnSTAT5A (Figure 2a). KO-positive cells expressing dnSTAT5A showed higher levels of phosphorylated STAT5 than KO-negative cells (Figure 2b). This finding is consistent with the earlier report showing that the dnSTAT5A exerts its effect on endogenous STAT5A and 5B with persistent phosphorylation of the dnSTAT5A itself.35 In contrast, U0126 retarded the proliferation of the HF6D835V cells more effectively than the HF6ITD cells (Figure 2c, each IC50 is 0.67±0.35 μM for HF6D835V and 6.09±0.90 μM for HF6ITD in the absence of IL-3). Indeed, U0126 inhibited phosphorylation of ERK1/2 in the HF6ITD and HF6D835V cells in a semidose-dependent manner (Figure 2d). In addition, another important signaling pathway downstream of FLT3, through PI3 kinase, was inhibited using LY294002. LY294002 also retarded the growth of the HF6D835V and HF6ITD cells in a dose-dependent manner, but there was no remarkable difference between both types of HF6 cells (Supplementary Figure 2, each IC50 is 4.18±0.55 μM for HF6D835V and 8.12±1.54 μM for HF6ITD in the absence of IL-3).
Taken together, these results in vitro suggested that the activation of MAPK was more critical for transformation by FLT3-TKD than by FLT3-ITD in HF6 cells, whereas activation of STAT5 was more critical for transformation by FLT3-ITD than by FLT3-TKD.
Activation of Ras-MAPK cascade enables HF6 cells to grow without IL-3 through cooperation between Hoxa9 and Raf
We further examined whether direct activation of either STAT5 or MAPK cascade is sufficient to confer factor-independent growth on HF6 cells. Although the constitutively active mutants of STAT5A, the relatively stronger mutant STAT5A1*6 and weaker mutant STAT5A#2, enabled Ba/F3 cells (Ba/F31*6, Ba/F3#) to grow without IL-3 as reported earlier,25, 27, 28 both failed to confer factor-independent growth on HF6 cells with limited elongation of survival time without IL-3 (Figures 3a and c). In contrast, the oncogenic NRAS mutant, NRASG12V, which had been detected in a case of AML with MLL-SEPT6,20 enabled HF6 cells (HF6G12V) to grow without IL-3, while it conferred no factor-independent growth on Ba/F3 with limited elongation of survival time without IL-3 (Figures 3b and c). In addition, Raf-1, a signal molecule downstream of Ras in Ras-MAPK cascades associated with malignant transformation, was tested with an activation-inducible system using ΔRaf-ER, consisting of the catalytic domain of human RAF-1 (ΔRaf) and the hormone-binding domain of the ER (Figure 3d), as described earlier.28 Unlike transduced Ba/F3 (Ba/F3ΔRaf-ER) cells, transduced HF6 (HF6ΔRaf-ER) cells grew without IL-3 only in the presence of 4-hydroxy-tamoxifen (Figure 3e). In these HF6ΔRaf−ER cells treated with 4-hydroxy-tamoxifen, STAT5A was not found to be secondarily activated by induction of activation of Raf/MAPK cascade in the absence of IL-3, whereas it was found to be weakly activated by stimulation with IL-3 for 15 min (data not shown).
Furthermore, we examined whether Hoxa9, which is one of the well-known target genes of MLL fusion proteins,10, 11, 13, 14 is involved in cooperation between MLL fusion protein and Ras/Raf/MAPK cascade. In the myeloid transformation assays, the murine BM progenitors immortalized by Hoxa9 in the presence of IL-3 (named A9G) proliferated without IL-3 after retroviral transduction of NRASG12V (Figure 3b). In the inducible transformation system using ΔRaf-ER, transduced A9G (A9GΔRaf−ER) cells grew without IL-3 only in the presence of 4-hydroxy-tamoxifen (Figure 3e). Expression level of Hoxa9 in A9G cells was shown in comparison with those in Ba/F3 and HF6 (negative and positive controls, respectively) cells by reverse transcriptase-PCR (Figure 3f).
Taken together, these results in vitro suggested the essential role of activation of the Ras/Raf/MAPK cascade together with Hoxa9 upregulated by MLL fusion proteins in the transformation of the cells expressing MLL fusion protein.
MLL fusion proteins and oncogenic NRAS cooperate to induce acute leukemia, at least partly through aberrant expression of Hoxa9
The findings on the transformation of HF6 cells in vitro led to the hypothesis that MLL fusion proteins might cooperate with activation of Ras to induce AML in vivo. To test this hypothesis, the oncogenic potential of NRASG12V (G12V) or STAT5A#2 (#2) to cooperate with MLL-SEPT6 (MS6) or MLL-ENL short form was examined in the leukemogenesis assays in vivo (Supplementary Figure 1). STAT5A1*6 was not used owing to its too strong oncogenic potential in vivo as reported earlier.36 The transduction efficiencies of NRASG12V, STAT5A#2 and MLL-ENL were 30–50, 20–40 and 5–10%, respectively, as determined by GFP expression (data not shown).
The mice receiving the BM cells transduced with MLL-SEPT6 and NRASG12V (MS6/G12V) died with significantly shorter latencies (26±2.4 days; P<0.05, log-rank test) than the MS6/GFP mice that died of MPD (137±9.0 days) as described earlier,6 but, unexpectedly, the neo/G12V mice died as early as the MS6/G12V mice (31±1.4 days) (Figure 4a, Table 1, and data not shown). The MS6/#2 mice died with significantly shorter latencies (82±11 days; P<0.05, log-rank test) than the MS6/GFP mice, but as early as the neo/#2 mice (80±8.0 days) (Figure 4a and Table 1). Notably, the phenotypes of the MS6/G12V mice were very different from those of the neo/G12V mice and from MPD in the MS6/GFP mice, whereas those of the MS6/#2 mice were rather similar to MPD in the MS6/GFP mice than those of the neo/#2 mice.
The morbid MS6/G12V mice showed hepatosplenomegaly with various ranges of leukocytosis, anemia and thrombocytopenia, whereas the morbid neo/G12V mice showed no hepatomegaly but mild splenomegaly, and severe pancytopenia (Figure 4b and Table 1). Histopathological analyses of the morbid MS6/G12V mice showed that immature myelomonocytic blasts accounted for more than 30% of BM cells, and severely infiltrated the spleen and the liver (Figures 4c and d, and data not shown). Immunophenotyping analyses of the BM cells also revealed that a majority of these cells expressed GFP, which indicated expression of NRASG12V, with high level of CD11b, intermediate level of Gr-1 (a myeloid differentiation marker also known as Ly-6G) and low level of c-Kit (CD117, the receptor of stem cell factor) (Figure 4e). In addition, Southern blot analysis of genomic DNAs derived from the spleens of the MS6/G12V mice showed oligoclonal bands of proviral integration (Figure 4g). These results indicated that the MS6/G12V mice developed AML similar to the mice receiving BM cells transduced with MLL-SEPT6 and FLT3-ITD, as described earlier.6 In contrast, the morbid neo/G12V mice showed extremely hypocellular marrows and extramedullary hematopoiesis in the spleen, where a majority of the cells did not express Ly5.1 (Figure 4f), with little expression of Hoxa9 in comparison with the morbid MS6/G12V mice (Supplementary Figure 3a). Thus, this finding suggested that, in our leukemogenesis assays under lethal conditioning, NRAS might develop BM aplasia presumably due to engraftment failure. Meanwhile, the MS6/#2 mice died of MPD, showing myeloid hyperplasia consisting predominantly of mature granulocytic elements in the BM cells, where a very small population (1.0%) expressed STAT5A#2, with splenomegaly similar to the MS6/GFP mice (Figures 4b–d, and Table 1). The neo/#2 mice showed neither hepatosplenomegaly nor hematological abnormalities in the peripheral blood, but relative myeloid hyperplasia in the BM, where only a small population (9.4%) expressed STAT5A#2 (Figures 4b and f, data not shown and Table 1), thus implying that STAT5A#2 might induce lethal BM abnormality owing to paracrine expression of some cytokines as in the earlier report using STAT5A1*6.36
To generalize leukemogenic cooperation between MLL fusion proteins and oncogenic NRAS and avoid the early death caused by transduction of NRASG12V, the BM cells transduced with MLL-ENL and/or oncogenic NRAS were also transplanted into recipient mice under sublethal conditioning. The MLL-ENL short form was used for leukemogenesis assays under sublethal conditioning with oncogenic NRAS (NRASG12V), in which retroviral vectors were exchanged, so that the expression of GFP indicated that of MLL-ENL (Supplementary Figure 1). These leukemogenesis assays under sublethal conditioning confirmed that the combination of MLL-ENL and NRASG12V reproduced AML, and that MLL-ENL (and puro) induced the phenotype of MPD (Figures 5a, b and d, and Table 1). Meanwhile, NRASG12V (and GFP) led to thymoma, sometimes together with leukocytosis, with a long latency (Figures 5a, c and d, and Table 1). In addition, to examine the possibility that the phenotypes associated with STAT5A#2 might change, similar to oncogenic NRAS, the BM cells transduced with STAT5A#2 (in pMYs-IRES-EGFP) and/or MLL-SEPT6 (in pMXs-neo) were again transplanted into recipient mice under sublethal conditioning. Within an observation period of 160 days, two of three neo/#2 mice under sublethal conditioning died with longer latencies (134 and 139 days) and showed the same phenotype of myeloid hyperplasia in the BM, where a small population (15%) expressed STAT5A#2, although these had different phenotypes of pancytopenia and splenomegaly (Supplementary Figure 3b and data not shown). In contrast, two of three MS6/#2 mice and all of the three MS6/GFP mice survived and showed no hematological abnormalities in the peripheral blood, whereas one of the MS6/#2 mice died (125 day) but could not be analyzed because of post-mortem change, within the observation period. Histopathological analysis of one MS6/#2 mouse, which was killed 150 days after the transplantation, showed no significant hepatosplenomegaly but mild myeloid hyperplasia in the BM (data not shown). Only 30% of the BM cells were positive for donor-derived Ly-5.1, and 7% of the BM cells were positive for GFP, indicating expression of STAT5A#2 (Supplementary Figure 3c), whereas reverse transcriptase-PCR analysis of the BM cells gave very weak signals of MLL-SEPT6 after 30 cycles (data not shown), but clearly visible signals after 35 cycles (Supplementary Figure 3c). Therefore, sublethal conditioning seemed to be inappropriate for leukemogenesis assays using oncogenes, such as MLL-SEPT6 and STAT5A#2, which had relatively weak oncogenic potential in comparison with MLL-ENL and NRASG12V.
Finally, we examined whether Hoxa9 may be involved in cooperation between the MLL fusion protein and oncogenic NRAS in vivo, such as in transformation assays in vitro. The leukemogenesis assays using the BM cells transduced with Hoxa9 and oncogenic NRAS were carried out under lethal conditioning, because preliminary leukemogenesis assays under sublethal conditioning were unsuccessful probably because of engraftment failure (data not shown). The combination of Hoxa9 and NRASG12V (A9/G12V) led to death with short latencies (28±7.5 days) (Figure 6a and Table 1), whereas Hoxa9 (and GFP) per se induced no lethal disease within 120 days, as reported earlier.37 The A9/G12V mice showed remarkable hepatosplenomegaly and had a tendency toward leukocytosis, anemia and thrombocytopenia (Table 1). Histopathological and immunophenotyping analyses of the BM cells revealed that the A9/G12V mice had a few, but prominent, myelomonocytic blasts (Figure 6b), with high expression of CD11b and Gr-1, and low level of c-Kit (Figure 6c). A Southern blot analysis of genomic DNAs derived from the spleens of the A9/G12V mice gave oligoclonal bands (data not shown). These results indicated that Hoxa9 cooperated with oncogenic NRAS to rapidly induce lethal myeloid malignancy that was not identical but similar to the acute leukemia induced by MLL fusion proteins and oncogenic NRAS.
Taken together, these results in vivo suggested that MLL fusion proteins rapidly induce acute leukemia together with activated NRAS, at least in part through aberrant expression of Hoxa9.
The present study provides several evidences that MLL-fusion-mediated leukemogenesis cooperated synergistically with Ras activation, but not with STAT5 activation. Although all known MLL fusion proteins were not tested in this study, we showed that this synergistic cooperation was not limited to the specific MLL fusion, using two different well-characterized types of MLL fusion proteins. In the light of the role of FLT3 mutations in MLL-fusion-mediated leukemogenesis described earlier,6 signaling pathways downstream of FLT3 mutations were analyzed in the transfectants of HF6, a cell line expressing MLL-SEPT6. The immortalized cells, such as HF6 and A9G, used in this study might have acquired additional mutations. However, the phenotypes including IL-3 dependency, expression patterns of lineage markers and growth rates were not changed since their establishment (data not shown), thus suggesting that at least no mutations leading to critical transformation had occurred in these cell lines. Although recent studies have disclosed the differences in activation of signal molecules, including MAPK and STAT5, between FLT3-TKD and FLT3-ITD,24, 38 our experiments using transduction with FLT3 mutants and inhibition of the signal molecules first showed a crucial role of activation of MAPK rather than STAT5 in the factor-independent survival and proliferation of HF6 cells. Next, the myeloid transformation assays in vitro revealed that the activation of Raf-1, as well as oncogenic NRAS, transformed HF6 cells, but that constitutively active mutants (1*6 and #2) of STAT5A did not. The leukemogenesis assays in vivo also showed that oncogenic NRAS rapidly induced acute leukemia together with MLL fusion proteins, which differed from the original phenotype induced by each molecule. In contrast, the active STAT5A mutant did not confer obvious synergistic effects on the MLL-fusion-mediated leukemogenesis. Thus, these results in vitro and in vivo suggested that activation of the Ras/Raf/MAPK pathway may be sufficient for the transformation of HF6 cells and development of MLL-fusion-mediated leukemia.
Oncogenic NRAS induced thymoma in the leukemogenesis assays under sublethal conditioning, which is consistent with the development of T-lymphoma by FLT3-TKD in our experimental system (Ono et al., unpublished data), whereas it led to the development of BM aplasia in our leukemogenesis assays under lethal conditioning. This difference in the disease phenotypes implies that forced expression of oncogenic NRAS in BM progenitors might be involved in its inhibitory effects on the engraftment of radioprotective cells as well as the antiproliferative effect of oncogenic NRAS in the early phase of the transplantation.39 These disease phenotypes were also different from the development of MPD in the earlier reports.39, 40 This discrepancy might be due to the differences in the experimental systems, such as the retroviral transduction and mice strains. Meanwhile, the BM progenitors transduced with Hoxa9 and NRASG12V seemed to result in engraftment failure under sublethal conditioning, but these rapidly developed myeloid malignancy under lethal conditioning. A recent study using BM transplantation showed the possibility of drastic fluctuation in the engraftment of donor cells receiving pathological modification under sublethal conditioning;41 hence, our unsuccessful results under sublethal conditioning might be associated with some instability of the transplantation.
Our leukemogenesis assays showed a definitively synergistic cooperation between MLL fusion proteins and oncogenic NRAS in the acceleration of disease onset and change of the phenotypes. Interestingly, the synergistic cooperation between MLL fusion proteins and Ras/Raf/MAPK activation closely correlated with recent clinical studies reporting the frequent coincidence of MLL fusion genes and mutations of RAS20 or RAF.42 It was reported that the additional expression of oncogenic KRAS induced an acute promyelocytic leukemia-like disease in transgenic mice expressing promyelocytic leukemia/retinoic acid receptor-α with an increased penetrance and decreased latency, although neither the penetrance nor the latency was significantly different from those in mice that died of MPD by expression of oncogenic KRAS alone.43 Other groups recently reported that the combination of oncogenic NRAS and MLL-AF944 or MLL-ENL45 is capable of developing AML, and that induced repression of oncogenic NRAS on the combination reverted AML to MPD by the MLL fusion gene (MLL-AF9) alone.44 Although our findings that MLL fusion proteins and oncogenic NRAS cooperate to induce AML confirmed these notions, the present study further analyzed the involvement of Hoxa9 and Raf, downstream of the cooperation between MLL fusion proteins and oncogenic NRAS. The myeloid transformation assays in vitro showed that the activation of Raf-1, as well as oncogenic NRAS, transformed A9G, a cell line expressing Hoxa9. The leukemogenesis assays in vivo also showed that Hoxa9 and oncogenic NRAS rapidly developed myeloid malignancy. These results in vitro and in vivo suggested that, as downstream molecules, Hoxa9 and Raf may have important roles in the synergistic leukemogenesis by MLL fusion proteins and oncogenic NRAS.
Our findings suggest a possible model of MLL-fusion-mediated leukemogenesis that was essentially recapitulated by Hoxa9 expression and Ras/Raf/MAPK activation (Figure 7). In the context of secondary genetic alterations, such as FLT3 mutations, this model explains the clinical features of acute leukemia with 11q23 translocations. First, overexpression, as well as TKD mutations, of FLT3 frequently detected in the MLL-rearranged infant acute leukemia may be involved in the leukemogenesis mainly through activation of Ras/Raf/MAPK, because several studies reported that the signaling pathway of wild-type FLT3 is similar to FLT3-TKD rather than FLT3-ITD.24, 38 Second, besides FLT3, other unknown molecular pathways that lead to the activation of Ras/Raf/MAPK might also be involved in the MLL-rearranged leukemia carrying no known genetic alterations, as FLT3 alterations are not found very frequently in most MLL-rearranged leukemia except in infants.46, 47 Meanwhile, in the context of MLL fusion proteins, we analyzed the role of the Hoxa9-mediated pathway leading to leukemogenesis. Recent studies revealed that one of the Hox-cofactor molecules, Meis1, is an essential molecule involved in normal hematopoiesis48 as well as Hoxa9-mediated leukemogenesis.49 However, our experimental system6 using BM cells transduced with MLL fusion proteins did not detect any significant upregulation of Meis1 in comparison with the mock transduction as reported earlier,50 in contrast with the findings by other groups.14 Therefore, we focused on Hoxa9, one of the key molecules directly upregulated by MLL fusion proteins. Interestingly, a recent study showed that the combination of Hoxa9 and Meis1 cooperated with Trib1, which enhanced the phosphorylation of ERK, to induce acute leukemia in the BM transplantation assays.51 Their study is not inconsistent with our findings; thus, the HOX and Ras/Raf/MAPK axes may have central roles in the molecular network of MLL-mediated leukemogenesis, which might be additively affected by other pathways, such as activation of STAT5 (Figure 7). In addition, at least, endogenous expression of Meis1 in A9G cells is also considered to be important in this network, but further analysis will be required to clarify the role of Meis1 in the collaboration between HOX and MAPK axes.
This study suggests that MLL fusion proteins synergistically cooperate with Ras/Raf/MAPK activation in leukemogenesis, at least partly through the upregulation of Hoxa9. Future studies analyzing the molecular crosstalk between Hoxa9 and the Ras/Raf/MAPK cascade are expected to provide novel insights into the molecular mechanism of MLL-fusion-mediated leukemogenesis.
Conflict of interest
The authors declare no conflict of interest.
Vogelstein B, Kinzler KW . Cancer genes and the pathways they control. Nat Med 2004; 10: 789–799.
Look AT . Oncogenic transcription factors in the human acute leukemias. Science 1997; 278: 1059–1064.
Rowley JD . The critical role of chromosome translocations in human leukemias. Annu Rev Genet 1998; 32: 495–519.
Gilliland DG, Tallman MS . Focus on acute leukemias. Cancer Cell 2002; 1: 417–420.
Kelly LM, Kutok JL, Williams IR, Boulton CL, Amaral SM, Curley DP et al. PML/RARalpha and FLT3-ITD induce an APL-like disease in a mouse model. Proc Natl Acad Sci USA 2002; 99: 8283–8288.
Ono R, Nakajima H, Ozaki K, Kumagai H, Kawashima T, Taki T et al. Dimerization of MLL fusion proteins and FLT3 activation synergize to induce multiple-lineage leukemogenesis. J Clin Invest 2005; 115: 919–929.
Schessl C, Rawat VP, Cusan M, Deshpande A, Kohl TM, Rosten PM et al. The AML1-ETO fusion gene and the FLT3 length mutation collaborate in inducing acute leukemia in mice. J Clin Invest 2005; 115: 2159–2168.
Stubbs MC, Kim YM, Krivtsov AV, Wright RD, Feng Z, Agarwal J et al. MLL-AF9 and FLT3 cooperation in acute myelogenous leukemia: development of a model for rapid therapeutic assessment. Leukemia 2008; 22: 66–77.
Watanabe-Okochi N, Kitaura J, Ono R, Harada H, Harada Y, Komeno Y et al. AML1 mutations induced MDS and MDS/AML in a mouse BMT model. Blood 2008; 111: 4297–4308.
Ayton PM, Cleary ML . Molecular mechanisms of leukemogenesis mediated by MLL fusion proteins. Oncogene 2001; 20: 5695–5707.
Meyer C, Kowarz E, Hofmann J, Renneville A, Zuna J, Trka J et al. New insights to the MLL recombinome of acute leukemias. Leukemia 2009; 23: 1490–1499.
Yokoyama A, Somervaille TC, Smith KS, Rozenblatt-Rosen O, Meyerson M, Cleary ML . The menin tumor suppressor protein is an essential oncogenic cofactor for MLL-associated leukemogenesis. Cell 2005; 123: 207–218.
Daser A, Rabbitts TH . Extending the repertoire of the mixed-lineage leukemia gene MLL in leukemogenesis. Genes Dev 2004; 18: 965–974.
Hess JL . MLL: a histone methyltransferase disrupted in leukemia. Trends Mol Med 2004; 10: 500–507.
Corral J, Lavenir I, Impey H, Warren AJ, Forster A, Larson TA et al. An MLL-AF9 fusion gene made by homologous recombination causes acute leukemia in chimeric mice: a method to create fusion oncogenes. Cell 1996; 85: 853–861.
Drynan LF, Pannell R, Forster A, Chan NM, Cano F, Daser A et al. MLL fusions generated by Cre-loxP-mediated de novo translocations can induce lineage reassignment in tumorigenesis. EMBO J 2005; 24: 3136–3146.
Wang J, Iwasaki H, Krivtsov A, Febbo PG, Thorner AR, Ernst P et al. Conditional MLL-CBP targets GMP and models therapy-related myeloproliferative disease. EMBO J 2005; 24: 368–381.
Chen W, Li Q, Hudson WA, Kumar A, Kirchhof N, Kersey JH . A murine MLL-AF4 knock-in model results in lymphoid and myeloid deregulation and hematologic malignancy. Blood 2006; 108: 669–677.
Taketani T, Taki T, Sugita K, Furuichi Y, Ishii E, Hanada R et al. FLT3 mutations in the activation loop of tyrosine kinase domain are frequently found in infant ALL with MLL rearrangements and pediatric ALL with hyperdiploidy. Blood 2004; 103: 1085–1088.
Liang DC, Shih LY, Fu JF, Li HY, Wang HI, Hung IJ et al. K-Ras mutations and N-Ras mutations in childhood acute leukemias with or without mixed-lineage leukemia gene rearrangements. Cancer 2006; 106: 950–956.
Gilliland DG, Griffin JD . The roles of FLT3 in hematopoiesis and leukemia. Blood 2002; 100: 1532–1542.
Armstrong SA, Staunton JE, Silverman LB, Pieters R, den Boer ML, Minden MD et al. MLL translocations specify a distinct gene expression profile that distinguishes a unique leukemia. Nat Genet 2002; 30: 41–47.
Murata K, Kumagai H, Kawashima T, Tamitsu K, Irie M, Nakajima H et al. Selective cytotoxic mechanism of GTP-14564, a novel tyrosine kinase inhibitor in leukemia cells expressing a constitutively active Fms-like tyrosine kinase 3 (FLT3). J Biol Chem 2003; 278: 32892–32898.
Choudhary C, Schwable J, Brandts C, Tickenbrock L, Sargin B, Kindler T et al. AML-associated Flt3 kinase domain mutations show signal transduction differences compared with Flt3 ITD mutations. Blood 2005; 106: 265–273.
Nosaka T, Kawashima T, Misawa K, Ikuta K, Mui AL, Kitamura T . STAT5 as a molecular regulator of proliferation, differentiation and apoptosis in hematopoietic cells. EMBO J 1999; 18: 4754–4765.
Schubbert S, Shannon K, Bollag G . Hyperactive Ras in developmental disorders and cancer. Nat Rev Cancer 2007; 7: 295–308.
Ariyoshi K, Nosaka T, Yamada K, Onishi M, Oka Y, Miyajima A et al. Constitutive activation of STAT5 by a point mutation in the SH2 domain. J Biol Chem 2000; 275: 24407–24413.
Onishi M, Nosaka T, Misawa K, Mui AL, Gorman D, McMahon M et al. Identification and characterization of a constitutively active STAT5 mutant that promotes cell proliferation. Mol Cell Biol 1998; 18: 3871–3879.
Kitamura T, Koshino Y, Shibata F, Oki T, Nakajima H, Nosaka T et al. Retrovirus-mediated gene transfer and expression cloning: powerful tools in functional genomics. Exp Hematol 2003; 31: 1007–1014.
Kroon E, Krosl J, Thorsteinsdottir U, Baban S, Buchberg AM, Sauvageau G . Hoxa9 transforms primary bone marrow cells through specific collaboration with Meis1a but not Pbx1b. EMBO J 1998; 17: 3714–3725.
Sakaue-Sawano A, Kurokawa H, Morimura T, Hanyu A, Hama H, Osawa H et al. Visualizing spatiotemporal dynamics of multicellular cell-cycle progression. Cell 2008; 132: 487–498.
Calvo KR, Sykes DB, Pasillas M, Kamps MP . Hoxa9 immortalizes a granulocyte-macrophage colony-stimulating factor-dependent promyelocyte capable of biphenotypic differentiation to neutrophils or macrophages, independent of enforced meis expression. Mol Cell Biol 2000; 20: 3274–3285.
Ono R, Ihara M, Nakajima H, Ozaki K, Kataoka-Fujiwara Y, Taki T et al. Disruption of Sept6, a fusion partner gene of MLL, does not affect ontogeny, leukemogenesis induced by MLL-SEPT6, or phenotype induced by the loss of Sept4. Mol Cell Biol 2005; 25: 10965–10978.
Nosaka T, van Deursen JM, Tripp RA, Thierfelder WE, Witthuhn BA, McMickle AP et al. Defective lymphoid development in mice lacking Jak3. Science 1995; 270: 800–802.
Moriggl R, Gouilleux-Gruart V, Jähne R, Berchtold S, Gartmann C, Liu X et al. Deletion of the carboxyl-terminal transactivation domain of MGF-STAT5 results in sustained DNA binding and a dominant negative phenotype. Mol Cell Biol 1996; 16: 5691–5700.
Schwaller J, Parganas E, Wang D, Cain D, Aster JC, Williams IR et al. STAT5 is essential for the myelo- and lymphoproliferative disease induced by TEL/JAK2. Mol Cell 2000; 6: 693–704.
Nakamura T, Largaespada DA, Shaughnessy Jr JD, Jenkins NA, Copeland NG . Cooperative activation of Hoxa and Pbx1-related genes in murine myeloid leukaemias. Nat Genet 1996; 12: 149–153.
Grundler R, Miething C, Thiede C, Peschel C, Duyster J . FLT3-ITD and tyrosine kinase domain mutants induce 2 distinct phenotypes in a murine bone marrow transplantation model. Blood 2005; 105: 4792–4799.
MacKenzie KL, Dolnikov A, Millington M, Shounan Y, Symonds G . Mutant N-ras induces myeloproliferative disorders and apoptosis in bone marrow repopulated mice. Blood 1999; 93: 2043–2056.
Parikh C, Subrahmanyam R, Ren R . Oncogenic NRAS rapidly and efficiently induces CMML- and AML-like diseases in mice. Blood 2006; 108: 2349–2357.
Santaguida M, Schepers K, King B, Sabnis AJ, Forsberg EC, Attema JL et al. JunB protects against myeloid malignancies by limiting hematopoietic stem cell proliferation and differentiation without affecting self-renewal. Cancer Cell 2009; 15: 341–352.
Christiansen DH, Andersen MK, Desta F, Pedersen-Bjergaard J . Mutations of genes in the receptor tyrosine kinase (RTK)/RAS-BRAF signal transduction pathway in therapy-related myelodysplasia and acute myeloid leukemia. Leukemia 2005; 19: 2232–2240.
Chan IT, Kutok JL, Williams IR, Cohen S, Moore S, Shigematsu H et al. Oncogenic K-ras cooperates with PML-RAR alpha to induce an acute promyelocytic leukemia-like disease. Blood 2006; 108: 1708–1715.
Kim WI, Matise I, Diers MD, Largaespada DA . RAS oncogene suppression induces apoptosis followed by more differentiated and less myelosuppressive disease upon relapse of acute myeloid leukemia. Blood 2009; 113: 1086–1096.
Zuber J, Radtke I, Pardee TS, Zhao Z, Rappaport AR, Luo W et al. Mouse models of human AML accurately predict chemotherapy response. Genes Dev 2009; 23: 877–889.
Chillon MC, Fernandez C, Garcia-Sanz R, Balanzategui A, Ramos F, Fernandez-Calvo J et al. FLT3-activating mutations are associated with poor prognostic features in AML at diagnosis but they are not an independent prognostic factor. Hematol J 2004; 5: 239–246.
Bacher U, Haferlach C, Kern W, Haferlach T, Schnittger S . Prognostic relevance of FLT3-TKD mutations in AML: the combination matters--an analysis of 3082 patients. Blood 2008; 111: 2527–2537.
Hisa T, Spence SE, Rachel RA, Fujita M, Nakamura T, Ward JM et al. Hematopoietic, angiogenic and eye defects in Meis1 mutant animals. EMBO J 2004; 23: 450–459.
Wong P, Iwasaki M, Somervaille TC, So CW, Cleary ML . Meis1 is an essential and rate-limiting regulator of MLL leukemia stem cell potential. Genes Dev 2007; 21: 2762–2774.
Horton SJ, Grier DG, McGonigle GJ, Thompson A, Morrow M, De Silva I et al. Continuous MLL-ENL expression is necessary to establish a ‘Hox Code’ and maintain immortalization of hematopoietic progenitor cells. Cancer Res 2005; 65: 9245–9252.
Jin G, Yamazaki Y, Takuwa M, Takahara T, Kaneko K, Kuwata T et al. Trib1 and Evi1 cooperate with Hoxa and Meis1 in myeloid leukemogenesis. Blood 2007; 109: 3998–4005.
We thank Dr Guy Sauvageau (Laboratory of Molecular Genetics of Stem Cells, Institute for Research in Immunology and Cancer, Canada) for the plasmid harboring a fragment of Hoxa9, and Dr Yusuke Satoh (Hematology and Oncology, Osaka University Graduate School of Medicine, Osaka, Japan) for technical advice. We are also grateful to R&D Systems for providing cytokines, and Brian Quinn for language assistance. This work was supported in part by Chugai Pharmaceutical Company Ltd, Grants-in-Aid from the Ministry of Education, Culture, Sports, Science, and Technology in Japan, the Novartis Foundation (Japan) for the Promotion of Science and the Japan Leukaemia Research Fund.
About this article
International Journal of Molecular Sciences (2019)
The Impact of PI3-kinase/RAS Pathway Cooperating Mutations in the Evolution of KMT2A-rearranged Leukemia
De novo activating mutations drive clonal evolution and enhance clonal fitness in KMT2A-rearranged leukemia
Nature Communications (2018)
Cooperation ofMLL/AF10(OM-LZ)withPTPN11activating mutation induced monocytic leukemia with a shorter latency in a mouse bone marrow transplantation model
International Journal of Cancer (2017)
Biochimica et Biophysica Acta (BBA) - Reviews on Cancer (2017)