Functional contribution of EEN to leukemogenic transformation by MLL-EEN fusion protein

Article metrics


The EEN (extra eleven nineteen) gene was originally cloned from a case of acute myeloid leukemia M5 subtype with translocation t (11; 19)(q23; p13), in which EEN was fused with MLL. To explore the involvement of EEN in leukemogenesis caused by MLL-EEN, we studied the transformation potential of the MLL-EEN fusion protein. MLL-EEN had oncogenic features, while, as a control, MLLΔ, the truncated form of MLL lacking the EEN moiety, did not show any oncogenic potential. MLL-EEN exerted a dominant-negative effect over wild-type EEN in terms of subcellular localization. Normally, EEN was found in the cytoplasm, but the MLL-EEN fusion protein was located in the nucleus, and EEN could be delocalized by MLL-EEN. This interaction is via a coiled-coil dimerization domain of EEN, which is reserved in the fusion protein. In addition, MLL-EEN might act as a potential transcriptional factor with the MLL part providing the DNA-binding domain and the EEN part providing the transcription activation domain, though EEN seems to have no direct role in transcriptional regulation. As an aberrant transcriptional factor, MLL-EEN could transactivate the promoter of HoxA7, a potential target gene of MLL.


Chromosomal translocations involving 11q23 are recurrent cytogenetic abnormalities in hematological malignancies, especially in acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL). The human myeloid-lymphoid leukemia gene (MLL/ALL-1/HRX) has been previously identified at the breakpoint of 11q23 in a number of different kinds of leukemia (Rubnitz et al., 1996; Rowley, 1998; Ayton and Cleary, 2001; Collins and Rabbitts, 2002; Ernst et al., 2002). Up to date, at least 54 chromosomal regions have been reported to be translocated to 11q23 and at least 31 genes have been cloned as fusion partners (Huret et al., 2001). The most frequent translocations are t(9;11)(p22;q23), t(4;11)(q21;q23) and t(11;19)(q23;p13), with the fusion partners AF9, AF4 and ENL, respectively. MLL-associated chromosomal rearrangements usually occur in a lineage-specific manner, suggesting a crucial role for the MLL fusion partner in determining disease phenotype.

EEN (named from extra eleven nineteen) is the third partner gene of MLL been identified on 19p13 (So et al., 1997). EEN belongs to a new SH3-domain family, which includes at least three members, EEN/SH3GL1/SH3p8/endophilin II, EEN-B1/SH3GL2/SH3p4/endophilin I and EEN-B2/SH3GL3/SH3p13/endophilin III (Sparks et al., 1996; Giachino et al., 1997). Upon binding by proline-rich ligands, these SH3 domains containing proteins play critical roles in a wide variety of biological processes (Mayer, 2001). For example, endophilin I has been found to be involved in synaptic vesicle endocytosis (Ringstad et al., 1997; McPherson, 1999; Reutens and Begley, 2002). Due to the high-sequence homology with endophilin I, EEN might play similar roles in the process of endocytosis.

Although the leukemogenic effect of MLL fusion protein has been well established in a numbers of settings, the lack of functional information about the MLL partners has made it difficult to address their contribution to the oncogenic potential of MLL fusion proteins. It is still unclear whether MLL fusion proteins act through a loss-of-function or gain-of-function mechanism in transforming cells and there are data supporting both scenarios (Ayton and Cleary, 2001). Several observations, however, indicated that fusion partners play essential roles in determining the oncogenic capacity of the MLL fusion proteins. First, all partner genes are fused in-frame to MLL, generating a full-length fusion protein, whereas terminal deletions of MLL have not been identified in leukemic cells. Second, the DNA-binding motifs of MLL, as well as the transcriptional transactivation activity of ENL, are required for in vitro immortalization of murine myeloid cells (Slany et al., 1998). Most convincingly, knockin mice expressing an MLL-AF9 fusion gene under the control of the natural MLL promoter developed AML (Corral et al., 1996). To explore the mechanism of leukemogenesis caused by MLL-EEN, especially the functional role of the EEN part of the fusion protein, we analysed the phenotypic differences between the cells bearing MLL-EEN and the cells bearing the truncated form of MLL.


Characterization of oncogenic features of MLL-EEN fusion gene

NIH3T3 cells have been utilized to demonstrate the transforming potential of several leukemia-specific fusion proteins (Kamps et al., 1991; Hajra et al., 1995; Frank et al., 1999). In our study, we first tested the transforming potential of MLL-EEN fusion protein in NIH3T3 cells. NIH3T3 cells transfected with pCI-neo, pCI-MLLΔ and pCI-MLL-EEN (see Materials and methods) were designated NIH3T3-Control, NIH3T3-MLLΔ and NIH3T3-MLL-EEN. The structures of MLL, EEN, MLL-EEN and MLLΔ are shown in Figure 1a. To check for the synthesis of the corresponding MLL-EEN fusion protein in the NIH3T3-MLL-EEN cells, cell lysates of transfected cells were immunoblotted with an anti-EEN antibody (Figure 1b). Only in the NIH3T3-MLL-EEN cells, there was an EEN-reactive protein at the expected size (about 200 kDa) corresponding to MLL-EEN protein. The extra bands around 46 kDa are likely to be endogenous EEN of NIH3T3 cells. The presence of DNA sequences and transcripts in the corresponding cells was further verified by genomic DNA PCR and RT–PCR experiments (Figure 1c). These G418-resistant cells were plated for growth in soft agar. No colonies were detected with either empty vector or MLLΔ expression vector, and soft agar colonies were obtained with MLL-EEN at a frequency of 1.1±0.6 % (Table 1). Typical morphologies of colonies were shown in Figure 1d. To determine the in vivo oncogenicity of MLL-EEN, NIH3T3 cells were injected subcutaneously into the flanks of nude mice. Six out of six sites injected with NIH3T3-MLL-EEN cells developed tumors, with a latency period of 2 weeks; these tumors grew to >1 cm in diameter by 6 weeks of observation (Table 1). Sites injected with NIH3T3-Control and NIH3T3-MLLΔ cells did not develop tumors by 8 weeks of observation (Table 1). These results demonstrate that expression of MLL-EEN in NIH3T3 cells leads to in vitro and in vivo transformation and provide direct evidence that the MLL-EEN fusion protein, but not the truncated form of MLL, is oncogenic.

Figure 1

NIH3T3 transformation assay – colonies in soft agar. (a) The structures of MLL, EEN, MLL-EEN and truncated MLLΔ proteins. Numbers represent amino-acid (aa) positions. MLL-EEN has a fusion site at aa number 1365 of MLL with aa number 16 of EEN giving rise to a 1718-aa fusion protein. Various putative domains are labeled as follows: AT, AT hook DNA-binding motifs; MT, cysteine-rich motif homologous to DNA methyltransferase; PHD, PHD zinc-fingers; TA, transactivation domain; SET, SET domain; CC, coiled-coil domain; SH3, SH3 domain. (b) Western blot analysis of expressing MLL-EEN fusion protein in stably transfected NIH3T3 cells. (c) Genomic DNA PCR and RT–PCR analysis of neo, MLLΔ and MLL-EEN in the stable transfected NIH3T3 cells. For RT–PCR analysis, an RT–PCR performed on samples in the absence of reverse transcriptase (RT−) served as a control for lack of DNA contamination. (d) Representative phase-contrast microscopy images (× 40 magnification) of colonies formed by NIH3T3 cells

Table 1 Summary of the transforming properties of MLL-EEN

To investigate the effects of MLL-EEN on hematopoietic cells, HL60 cells were transfected with pCI-MLL-EEN (designated HL60-MLL-EEN). HL60 cells, derived from human acute myeloblastic leukemia with promyelocytic features, have been used extensively as a model of leukemic myeloid cell differentiation and they can be induced to differentiate to the macrophage phenotype using TPA, to the monocytic phenotype using 1,25-Vitamin D3, or to granulocyte-like cells by retinoic acid (Collins, 1987; Mollinedo et al., 1998). HL60 cells express bcl-2 and bax, overexpress c-myc, but do not express p53, and N-ras is mutated (Mollinedo et al., 1998). Since additional events, including mutations in p53 and N-ras, and deregulated bcl-2 function, have been documented to accompany MLL rearrangements in some cases (Naoe et al., 1993; Pocock et al., 1995; Lanza et al., 1996; Mahgoub et al., 1998), it is reasonable to speculate that the oncogenic potential of MLL fusion protein will be potentiated in the cells with such mutations. We thus took advantage of HL60 cells to study the oncogenic properties of MLL-EEN. Previous knockin experiments in mice (Corral et al., 1996) and myeloid transformation assay in primary murine bone marrow progenitor cells (Lavau et al., 1997), as well as our data in NIH3T3 cells (see above), indicate that truncated form of MLL alone lacks the ability of transformation. Therefore, HL60 cells transfected with pCI-MLLΔ (designated HL60-MLLΔ) and HL60 cells transfected with pCI-neo (designated HL60-Control) were taken as controls. Validation of the presence and transcription of neo, MLLΔ and MLL-EEN genes in the stable transfected HL60 cells by PCR was shown in Figure 2a. The synthesis of MLL-EEN fusion protein was also testified by Western blotting using anti-EEN antibody (Figure 2b). The bands corresponding to the endogenous EEN of HL60 cells are also indicated.

Figure 2

The oncogenic feature of HL60 cells bearing MLL-EEN. (a) Genomic DNA PCR and RT–PCR analysis of neo, MLLΔ and MLL-EEN in the stable transfected HL60 cells. An RT–PCR performed on samples in the absence of reverse transcriptase (RT−) served as a control for lack of DNA contamination. (b) Western blot analysis of expressing wild-type EEN and MLL-EEN fusion protein in stably transfected HL60 cells. (c) The cell cycle distribution of stably transfected HL60 cells. The result of only one representative experiment is shown. (d) The cell proliferation curves of HL60 cells. Differentiation assay of the transfected HL60 cells by testing the cell surface antigen CD11b (e) and CD14 (f). (g) Apoptosis assay of the transfected HL60 cells. (h) Survival curve of HL60 cells under the treatment of different concentrations of DNR. Values in (d–f) and (h) are means±s.d.'s of at least three individual experiments

HL60-MLL-EEN cells showed increased proliferation as compared with HL60-Control and HL60-MLLΔ cells as revealed by the cell cycle distribution (Figure 2c). The percentage of HL60-MLL-EEN cells in proliferation phases (S+G2/M phases) of the cell cycle (61.4%) was higher than the ones of HL60-Control (51.8%) and HL60-MLLΔ cells (51.5%). Figure 2d shows the mean proliferation curve of these cells. After 6 days, the difference of the cell numbers between HL60-MLL-EEN and HL60-Control, as well as that between HL60-MLL-EEN and HL60-MLLΔ, was statistically significant as determined by Student's t-test (P<0.05). It implies that the HL60-MLL-EEN cells grow faster than HL60-MLLΔ cells and HL60-Control cells.

Since the MLL-EEN fusion protein was related to a phenotype of acute monocytic leukemia (So et al., 1997), we studied differentiation along the monocytic lineage of these stably transfected HL60 cells. It was well documented that HL60 cells could be induced to differentiate into monocytes by 1,25-Vitamin D3 (Collins, 1987). Previously described changes in antigen expression by differentiating cells include the induction of CD11b and CD14 by Vitamin D3 (Collins, 1987; Trayner et al., 1998). Being treated with Vitamin D3, the percentage CD11b-positive cells of HL60-Control, HL60-MLLΔ and HL60-MLL-EEN cells were 60.8±4.7, 58.3±2.6 and 28.0±0.2%, respectively (Figure 2e); the percentage CD14-positive cells of HL60-Control, HL60-MLLΔ and HL60-MLL-EEN cells were 87.5±1.5, 86.2±0.3 and 77.0±0.4%, respectively (Figure 2f). The difference between HL60-Control and HL60-MLL-EEN cells, as well as the difference between HL60-MLLΔ and HL60-MLL-EEN cells, were statistically significant as determined by Student's t-test (P<0.05). These data suggest that HL60-MLL-EEN cells are less sensitive to induction of monocytic lineage differentiation by Vitamin D3. Supporting our results, SN-1 cells with t(11;16) (MLL-CBP) were also found to be relatively insensitive to Vitamin D3 (Hayashi et al., 2000), and downregulation of MLL-CBP fusion gene expression is associated with differentiation of SN-1 cells (Niitsu et al., 2001).

Patients who have 11q23 translocations often have aggressive clinical features (Rubnitz et al., 1996). There is a clear correlation with the poor prognosis and resistance against chemotherapy. Several MLL fusion proteins, including MLL-AF4, MLL-ENL and MLL-AF9, were found to prevent apoptosis of leukemic cells (Akao et al., 1998; Kersey et al., 1998; Dorrie et al., 1999; Kawagoe et al., 2001). Since apoptosis is often deregulated in leukemogenesis (Kitada et al., 2002), we tried to study the effect of MLL-EEN on apoptosis in HL60 cells. Our data showed that the HL60-MLL-EEN cells were more resistant to cell death induced by serum deprivation (Figure 2g). In addition, we also observed higher resistance of HL60-MLL-EEN cells to apoptosis induced by daunorubicin (DNR) (Figure 2g), which is known to induce apoptosis of HL60 cells by activation of caspase-3 (Kwon et al., 2002). In both cases, the percentage of PI/Annexin V cells in HL60-MLL-EEN cells was higher than the ones in HL60-Control and HL60-MLLΔ cells, with statistically significant difference as determined by Student's t-test (P<0.05). We further studied the survival of HL60-Control, HL60-MLLΔ and HL60-MLL-EEN cells in different concentrations of DNR (Figure 2h). IC50 of DNR to HL60-Control, HL60-MLLΔ and HL60-MLL-EEN were 0.138±0.010, 0.105±0.009 and 0.265±0.001 μg/ml, respectively. The IC50 value of HL60-MLL-EEN cells is higher than those of HL60-Control and HL60-MLLΔ cells, with statistically significant difference as determined by Student's t-test (P<0.05). These results indicate that the expression of MLL-EEN make the HL60 cells more resistant to serum deprivation and DNR treatment.

Delocalization of EEN by MLL-EEN

To understand the mechanism of MLL-EEN in leukemogenesis, its subcellular localization must be identified first. Sequence analysis indicates that a potential nuclear localization signal (NLS) KKLEGRRLDFDYKKKRQG was present from aa159 to aa176 in EEN, but most SH3-domain proteins are cytoplasmic (Mayer, 2001). As shown in Figure 3a, the wild-type EEN protein indicated by red fluorescence was diffusely distributed in the cytoplasm of the transfected NIH3T3 cells, while the MLL-EEN fusion protein and the truncated MLLΔ protein indicated by green fluorescence showed nuclear localization (Figure 3b and c). The subcellular localization difference between EEN and MLL-EEN implied that MLL-EEN and EEN exert their functions in two different ways. When the wild-type EEN was cotransfected with MLLΔ into NIH3T3 cells, the EEN protein remained in the cytoplasm and the MLLΔ in the nucleus (Figure 3d–f). Interesting enough, while EEN was cotransfected with MLL-EEN into NIH3T3 cells, EEN could be found both in the nucleus and the cytoplasm while MLL-EEN was only in the nucleus. Furthermore, while keeping the total amount of plasmids unchanged but increasing the ratio of MLL-EEN over EEN used in transfection, more cells with nuclear location of EEN were observed. The typical results were showed in Figure 3g–i (the ratio of MLL-EEN over EEN was 1 : 2) and Figure 3j–l (the ratio of MLL-EEN over EEN was 2 : 1). This indicates that MLL-EEN protein directs EEN to the nucleus in a competitive manner.

Figure 3

Subcellular localization studies. Subcellular localization of EEN (a), MLL-EEN (b), MLLΔ (c), EEN together with MLLΔ (d-f) and EEN together with MLL-EEN (g-l). EEN was indicated by red fluorescent and MLL-EEN and MLLΔ were indicated by green fluorescent. Representative results are shown

Identification a coiled-coil dimerization domain of EEN

To further explore possible direct interaction between EEN and MLL-EEN, and to identify the regions involved in this interaction, the MLL-EEN and several EEN deletion mutants were expressed as GST fusion proteins and applied to a pull-down assay. The structures of various truncated versions of EEN are shown in Figure 4a. Through sequence analysis, we found there are a coiled-coil domain (119–227 aa) and an SH3 domain (306–368 aa) in EEN protein, both of which are implicated in protein–protein interactions. The GST pull-down assay showed that MLL-EEN could bind with EEN in vitro. Analysis of the various deletion constructs indicated that all polypeptides containing 119–227 aa of EEN, including EEN1, EEN2, EEN3 and EEN5, could bind to EEN (Figure 4b). Thus, the coiled-coil domain of EEN is responsible for dimerization.

Figure 4

Dimerization of EEN. (a) Schematic depictions of various truncated versions of EEN. Numbers in brackets represent corresponding amino acids. (b) GST pull-down assay. MLL-EEN GST fusion protein and various EEN mutants fused to GST proteins immobilized on glutathione agarose beads were incubated with 35S-labelled wild-type EEN (bottom panel). Top panel: the input GST proteins for each construct. (c) Coimmunoprecipitation assay. Lanes 1 and 3, COS7 cells cotransfected with EEN-Myc and GFP; lanes 2 and 4, COS7 cells cotransfected with EEN-Myc and EEN-GFP. WB: Western blotting of whole-cell lysates; IP: Western blotting of immunoprecipitation

The pull-down results also pointed out that EEN could form homodimer in vitro. To determine whether EEN can form homodimer in vivo, we performed an immunoprecipitation assay. When GFP-tagged EEN and Myc-tagged EEN were cotransfected into COS7 cells, GFP-tagged EEN (about 75 kDa) could be coimmunoprecipitated with Myc-tagged EEN from detergent extracts of these cells using an anti-GFP antibody (Figure 4c, lane 4). While as a negative control, GFP (about 29 kDa) could not be coimmunoprecipitated with Myc-tagged EEN (Figure 4c, lane 3), indicating that this interaction takes place between EEN and EEN itself. These data suggest that EEN might form homodimer in vivo. Consistent with our finding, Ringstad et al. also observed self-association of endophilins I and II and this interaction requires the central coiled-coil domain (Ringstad et al., 2001).

Transactivity feature of EEN and deregulation of HoxA7 by MLL-EEN

Since many of the MLL fusion partners function as transcriptional activators in reporter gene assays, it has been thought that the transcriptional effector functions of MLL fusion partners should be essential to leukemogenesis (Ayton and Cleary, 2001). Although wild-type EEN is located in cytoplasm in physiological context and seems to have no relation with transcription, its relation with transcriptional regulation is not known. Moreover, MLL-EEN is localized in the nucleus. MLL-EEN thus may aberrantly affect the transcriptional regulation of the target genes of MLL. To test whether EEN has latent transcription activity, a reporter gene assay was performed. As shown in Figure 5a, the middle coiled-coil domain (119–227 aa) and the C-terminal SH3 domain (306–368 aa) both had transactivation property. Interestingly, whereas the full-length EEN (EEN1) failed to activate transcription, EEN8, which lacks the N-terminal 15 amino acids and is identical to the reserved part of EEN in MLL-EEN fusion protein, did have transactivation activity. Results from yeast one-hybrid also confirmed this finding (Figure 5b). We speculate that the EEN in MLL-EEN fusion protein might work as a potential activator of transcription. Thus, with the MLL part providing the DNA-binding domain (Zeleznik et al., 1994; Birke et al., 2002) and the EEN part providing the transcriptional activation domain, MLL-EEN might function as a novel transcriptional factor.

Figure 5

(a) Transactivity assay of EEN in COS7 cells. The indicated EEN mutants were fused to the GAL4 DNA-binding domain in the pBIND vector. Numbers in brackets indicate the amino-acid positions of the respective mutants. To prove the functionality of the GAL4 fusion constructs, the respective plasmid was cotransfected into COS7 cells together with the pG5-luc reporter plasmid. Firefly luciferase activities were normalized for Renilla luciferase and presented as relative luciferase units. The mean of relative luciferase units of the pBIND control was assigned a value of 1.0. Means and s.d.'s of triplicates are given. (b) Transactivity assay of EEN in yeast. (c) MLL-EEN increases HoxA7 promoter activity in HL60 cells. Data are presented as the fold activation of firefly luciferase activity relative to the empty pCI-neo vector and are normalized for Renilla luciferase activity. The average firefly luciferase/Renilla luciferase activity on pGL3-HoxA7 of the pCI-neo control was assigned a value of 1.0. Results are expressed as means of three independent experiments±s.d.'s

To investigate how will MLL-EEN affect its target genes, taken that MLL-EEN might serve as a potential transcriptional factor, we tested the effect of MLL-EEN on HoxA7 promoter. HoxA7 was identified as one of the targets of MLL (Yu et al., 1995; Ernst et al., 2002), and other MLL fusion protein, such as MLL-ENL and MLL-ELL, strongly transactivated the promoter HoxA7 (Schreiner et al., 1999; Dimartino et al., 2000). Interestingly, HoxA7 is involved in leukemogenesis. The murine HoxA7 gene is constitutionally activated by a retroviral insertion in myeloid leukemia that outgrew in BXH-2 mouse (Nakamura et al., 1996). Moreover, a genetic approach has demonstrated that HoxA7 is required for the full oncogenic activity of an MLL fusion protein (Ayton and Cleary, 2003). Since the EEN part in MLL-EEN fusion protein could serve as a potential activator of transcription, it is possible that the EEN part could activate transcription unspecifically. To exclude the possibility that MLL-EEN could activate the transcription of HoxA7 through its EEN part in an indirect way, we introduced another control pCI-GAL4-EEN, which fused the GAL4 DNA-binding domain with 16–368 aa of EEN. To further exclude the possibility that MLL-EEN could be a global activator of transcription, we performed luciferase assays using a luciferase reporter plasmid with a TATA-like promoter region from the herpes simplex virus thymidine kinase promoter (pTAL-Luc). The data showed in Figure 5c indicated that MLL-EEN was a strong transcriptional activator of HoxA7 and accomplished a 17-fold acivation. In contrast, neither MLLΔ nor GAL4-EEN altered the promoter activity of pGL3-HoxA7 plasmid, indicating that the alteration of promoter activity by MLL-EEN depended on both of the MLL part and the EEN part. In addition, none of these constructs could activate luciferase activities exhibited by pTAL-luc. Therefore, MLL-EEN seems to be a specific activator of HoxA7 transcription, rather than a global activator of transcription. It means MLL-EEN might play a role in leukemogenesis through deregulating the target genes of wild-type MLL.


In our study, we demonstrated that MLL-EEN was an oncogene. Using hematopoietic progenitor transformation assay, other groups also demonstrated that MLL-EEN fusion gene was capable of immortalizing hematopoietic progenitor cells (Chan and Yam, 2002; Yam et al., 2003). This result will further support our finding that MLL-EEN fusion gene could cause oncogenic transformation. We also found that the expression of MLL-EEN could render the HL60 cells proliferation advantage, apoptosis-resistance and refractory toward differentiation stimuli. This implies that the pathogenesis caused by MLL-EEN might be associated with deregulation of proliferation, apoptosis and differentiation. These pathways might collaborate and finally lead to the development of leukemia.

Our work revealed interesting features of protein localization of wild-type EEN and the MLL-EEN fusion protein. The chimeric MLL-EEN protein showed nuclear localization pattern, similar to other MLL chimera proteins (Joh et al., 1996; Butler et al., 1997; Rogaia et al., 1997; Yano et al., 1997). It was reported that the motifs within the N-terminal one-third of the MLL protein, including the AT hooks and methyltransferase homologous domain, were responsible for nuclear localization (Yano et al., 1997). Interestingly, when MLL-EEN was cotransfected with wild-type EEN into NIH3T3 cells, the fusion protein directed wide-type EEN protein into the nucleus. Since EEN shares high homology with endophilin I, which plays a role in the clathrin-coated vesicle endocytotic pathway regulating synaptic vesicle formation (Reutens and Begley, 2002), wild-type EEN might be also involved in endocytosis in a similar way. Recently, Chen et al. (2003) demonstrated that the interaction of EEN and Ca2+ channel is essential for clathrin-mediated synaptic vesicle endocytosis. Receptor-mediated endocytosis has been traditionally considered an effective mechanism to attenuate ligand-activated responses, and EEN/endophilin family members are also involved in this process. For example, endophilin-CIN85-Cbl complex can mediate ligand-induced downregulation of EGF receptors and ligand-dependent downregulation of c-Met (Petrelli et al., 2002; Soubeyran et al., 2002). It was proposed that impaired endocytosis might enhance cell replication through prolonged signaling by growth-factor receptors (Floyd and De Camilli, 1998). In the cells bearing MLL-EEN, wild-type EEN might be partially redistributed into the nucleus by MLL-EEN, and there might be insufficient dose of EEN localized in cytoplasm. As a result, endocytosis-mediated regulation of signaling might be attenuated, leading to prolonged activation of signaling pathways. Thus, cells with less EEN proteins might display an enhanced proliferative response to growth factors. On the other hand, mistargeted EEN might also exert unfavorable effects to disrupt normal cell functions. Besides EEN, MLL-EEN fusion protein was also able to delocalize EBP into the nucleus through direct interaction between EEN and EBP, and interfere with the Ras-suppressing activities of EBP (Yam et al., 2003). Interestingly, EAF1, an ELL-associated factor, was also delocalized by expression of the MLL-ELL fusion protein (Simone et al., 2001; Polak et al., 2003). These results, in company with our finding, suggest that delocalization of the target proteins through the fusion partners might represent a novel pathway for oncogenesis by MLL fusion proteins.

Previous studies strongly suggest that the transcriptional effector functions of MLL fusion partners are essential for leukemogenesis (Ayton and Cleary, 2001; Zeisig et al., 2003). The ability of MLL-ENL, MLL-ELL, MLL-CBP, MLL-AFX and MLL-FKHRL1 to immortalize myeloid progenitors correlates well with their ability to transactivate reporter genes in transient transcriptional assays (Slany et al., 1998; Dimartino et al., 2000; Lavau et al., 2000; So and Cleary, 2002; So and Cleary, 2003). Although EEN appears to be cytoplasmic protein with no known function in transcription (So et al., 1997), according to our results, EEN might also be involved in transcriptional activation indirectly. MLL-EEN fusion protein, which is localized in nucleus, might recognize the MLL target genes by the DNA-binding domain provided by MLL part (Zeleznik et al., 1994; Birke et al., 2002) and influence the transcription of these genes by the transcription activation domain provided by EEN part. As we showed here, MLL-EEN did activate the promoter of HoxA7 gene.

Protein–protein interactions, including dimerization, may also contribute to the leukemogenic potential of the MLL fusion proteins. Two fusion partner proteins of MLL, AF10 and AF17, contain the leucine-zipper dimerization motifs and the leucine-zipper of AF10 is required for the leukemic transformation by MLL-AF10 (Dimartino et al., 2002). When MLL was fused with lacZ, which assemble into a tetrameric form, is sufficient to cause acute leukemia in chimeric mice (Dobson et al., 2000). Partial duplication of MLL, which functionally could be equivalent to a dimer of the N-terminal portion of the MLL, was also found in acute leukemia (Schichman et al., 1994). Moreover, several recent studies demonstrated that dimerization also plays an essential role in the leukemogenesis caused by MLL fusion proteins (Martin et al., 2003; So et al., 2003), indicating a novel mechanism for the oncogenic activity of MLL chimeric proteins (Hsu and Look, 2003). Since the dimerization domain of EEN is retained in MLL-EEN fusion protein, it is possible that EEN contributes to the leukemogenic feature of MLL-EEN by providing such dimerization interface. In addition, we also demonstrated that the dimerization domain of EEN might also be involved in transcriptional activation. It is possible that EEN might obtain transcriptional activity by interacting with other coiled-coil domain containing transcriptional factors through this coiled-coil domain.

In summary, by comparing the phenotypic difference between cells bearing MLL-EEN and MLLΔ, we show here that the EEN part plays an essential role in leukemogenesis caused by MLL-EEN. Through the EEN part, MLL-EEN could delocalize other proteins and deregulate the transcription of the target genes of wild-type MLL. The concurrence of these pathways may confer the fully oncogenic feature on the MLL-EEN fusion protein.

Materials and methods


Full-length EEN was inserted into pCI-neo (Promega, Madison, WI, USA; for use in in vitro transcription/translation reactions) (plasmid pCI-EEN). Full-length EEN was also subcloned into pCMV-Myc (Clontech, Palo Alto, CA, USA; for use in immunofluorescent microscopy and coimmunoprecipitation) (plasmid pCMV-Myc-EEN). EEN, MLL-EEN and truncated form of MLL (designated MLLΔ, 1–1360 aa according to full-length MLL) (Figure 1a) were cloned into pEGFP-C1 (Clontech, Palo Alto, CA, USA; for use in coimmunoprecipitation or fluorescent microscopy) (plasmid pEGFP-EEN, pEGFP-MLL-EEN and pEGFP-MLLΔ). Fragments encoding various regions of EEN (Figure 1b) were generated by PCR and subcloned into pGEX-5X1 (Amersham Pharmacia Biotech, Little Chalfont, England; for use in GST-fusion protein expression), pBIND (Promega, Madison, WI, USA; for use in mammalian transactivation assay), and pGBKT7 (Clontech, Palo Alto, CA, USA; for use in yeast one-hybrid assay and two-hybrid screening). The fragment of GAL4-EEN from pBIND-EEN8 was cloned into pCI-neo to generate pCI-GAL4-EEN. All constructs were confirmed by sequencing. Cloning details are available upon request.

Cell culture, transfection and electroporation

HL60 cells were cultured at density below 1 × 106/ml in RPMI-1640. NIH3T3 and COS7 cells were cultured in DMEM. The media were supplemented with 10% fetal calf serum (FCS, Gibco BRL Life Technologies, Gaithersburg, MD, USA), 2 mM L-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. All cell cultures were incubated at 37°C in humidified air with 5% CO2. Transfections of NIH3T3 and COS7 cells were carried out with SuperFect (QIAGEN GmbH, Hilden, Germany) according to the manufacturer's instructions. Transfection of HL60 cells was performed by electroporation according to the Gene Pulser Electroprotocol (Bio-Rad Laboratories, Hercules, CA, USA). The stable cell lines were obtained after being cultured in the medium plus 1 mg/ml G418 for about 3 weeks, and were maintained in the medium plus 0.2 mg/ml G418. For all studies, pooled populations of G418-resistant cells were used to eliminate clonal variation.

Genomic DNA PCR, RT–PCR and Western blot

Genomic DNA was purified according to standard methods. Total RNA was isolated using Trizol LS Reagent (Invitrogen, Carlsbad, CA, USA) and treated with DNase I to remove contaminating genomic DNA. A measure of 1 μg RNA was converted to random-primed cDNA using SuperScript first-strand synthesis system (Invitrogen) according to the manufacturer's instructions. Genomic DNA and cDNA were subjected to PCR analysis. The primers for amplification of MLL-EEN were designed to span the breakpoint between the fusion partners: forward primer: ccatcagcaagagaggatcctg; reverse primer: gccggcatccagcaatgcgtc. The primers for amplification of MLLΔ were: forward primer: ccatcagcaagagaggatcctg; reverse primer, which was corresponding to the pCI-neo vector sequence downstream to the insert: aattaaccctcactaaagggaa. Primers for amplification of neo: gattgcacgcaggttctcc and gtagccaacgctatgtcctg. The expected sizes for the amplification products were 660 bp for neo, 360 bp for MLLΔ and 640 bp for MLL-EEN.

Stably transfected NIH3T3 cells and HL60 cells were lysed in SDS buffer. After separation on standard SDS–PAGE gels (8%) the proteins were blotted on to nitrocellulose. MLL-EEN proteins were detected with polyclonal antibody against EEN (Santa Cruz Biotech, Santa Cruz, CA, USA).

Colony growth in soft agar

Essentially, in six-well plates prepared with a lower layer of 0.7% agar solution in DMEM with 10% fetal calf serum (Gibco BRL Life Technologies, Gaithersburg, MD, USA) containing 1 mg/ml of G418 sulfate (Promega, Madison, WI, USA), 2 × 103 stably transfected NIH3T3 cells were suspended in a 0.35% agar solution in growth medium. Colonies were scored 21 days after being plated (colonies larger than 0.1 mm in diameter were scored as positive).

Tumor growth in nude mice

A total of 1 × 106 stably transfected NIH3T3 cells in 200 μl PBS were injected into the flank of 4-week-old athymic, Balb-c/nu/nu mice. The mice were inspected weekly for the development of tumors and observed for up to 8 weeks.

Analysis of cell cycle

Cells were collected, washed, and fixed overnight in 70% cold ethanol. After being washed with PBS, cells were incubated in PBS supplemented with 50 μg/ml RNase A. Cell cycle distribution was then analysed by staining the cells with 50 μg/ml propidium iodine (PI, SantaCruz Biotech, Santa Cruz, CA, USA) and evaluated in EPICS XL flow cytometer (Beckman Coulter, Hialeah, FL, USA). Gating was used to remove debris and doublets before collection. Data were quantitated using MultiCycle software (Phoenix Flow Systems, San Diego, CA, USA).

Characterization of cell differentiation

After being incubated in media with or without Vitamin D3 (1 × 10−7M) for 72 h, aliquots of 0.5 × 106 cells were subjected to single label staining. Cell surface myeloid-specific antigen CD11b or CD14 expression was analysed by direct staining with either fluorescein isothiocyanate (FITC)- or phycoerythrin (PE)-conjugated monoclonal antibody (Becton Dickinson, San Jose, CA, USA). Nonspecific fluorescence was assessed using isotype-matched control. Analysis was performed by EPICS XL flow cytometer (Beckman Coulter, Hialeah, FL, USA).

Cell survival and apoptosis assays

Viability and apoptosis analysis of the cells submitted to cellular stress either in the absence of serum or in the media containing 1 μg/ml DNR for 60 h were determined by double staining with Annexin V (SantaCruz Biotech, Santa Cruz, CA, USA) and PI. Analysis was performed by EPICS XL flow cytometer (Beckman Coulter, Hialeah, FL, USA). PI/Annexin V cells were scored as normal cells, PI/Annexin V+ cells were scored as early apoptotic cells and PI+ cells were scored as late apoptotic cells undergoing secondary necrosis or dead cells (Ormerod, 1998).

Drug sensitivities were determined by the MTT colorimetric method. Cells were plated onto 96-well plates at a density of 2000 cells per well in 100 μl of RPMI-1640 plus 10% fetal calf serum. After cells were incubated for 1 h, varying concentrations of DNR were added, and the cells were maintained for 72 h. Cell aliquots treated without any drug were used as controls. After this period of drug exposure, 10 μl MTT (5 mg/ml in RPMI-1640) were added to each well and incubated at 37°C for 4 h. Formazan was dissolved in 100 μl of DMSO and the absorbance was measured at 570 nm wavelength on a microplate reader (Titertek, Thermo Labsystems, Finland). The IC50 (drug concentration resulting in 50% inhibition of MTT dye formation, compared to controls) was determined directly from semilogarithmic dose–response curves. Values represent the mean±s.d. of three experiments.

Immunofluorescent and fluorescent microscopy

Cells were washed in PBS, fixed in −20°C methanol for 30 min, permeabilized in PBS with 0.25% Triton X-100, and blocked with 3% BSA in PBS. Primary anti-Myc antibody (Clontech, Palo Alto, CA, USA) was prepared in 3% BSA-PBS (1 : 200) and applied for 1 h, 25°C. Mounts were washed with PBS with 1% NP-40, and incubated with rhodamine-conjugated rabbit anti-mouse antibody (1 : 200; Calbiochem, La Jolla, CA, USA) for 1 h, 25°C, then washed extensively and placed on coverslips. The fluorescent signals were observed with confocal microscope (BioRad Radiance 2000 Confocal).

GST pull-down in vitro protein interaction assay

MLL-EEN and various EEN fragments (Figure 4a) were expressed as GST fusion proteins in the host strain BL21 induced by IPTG. EEN subcloned into pCI-neo was translated into protein in vitro using the TNT coupled reticulocyte lysate system (Promega, Madison, WI, USA) according to the manufacturer's instruction. GST fusion proteins were preincubated with glutathione sepharose beads (Amersham Pharmacia Biotech, Little Chalfont, England) in PBS (plus 1% Triton X-100, 5 mM DDT, 100 μg/ml PMSF) for 40 min at 4°C and washed using PBS twice. 35S-labelled in vitro synthesized EEN protein was then mixed with the GST sepharose beads and incubated in NETN buffer (0.5% NP-40, 1 mM EDTA, 20 mM Tris pH 8.0, 100 mM NaCl) at 4°C for 3 h. The beads were washed five times in buffer H (20 mM HEPES pH 7.7, 50 mM KCl, 20% glycerol, 0.1% NP-40 and 0.007% β-mercaptoethanol). Bound proteins were resolved by SDS–PAGE and visualized by autoradiography.


pEGFP-EEN and pCMV-Myc-EEN were cotransfected into COS7 cells, empty pEGFP-C1 vector was also cotransfected with pCMV-Myc-EEN as negative control. At 48 h after transfection, cells were lysed with CytoBuster protein extraction reagent (Novagen) plus EDTA-free protease inhibitor cocktail (Calbiochem, La Jolla, CA, USA). The whole cell lysates were precleared with protein A/G agarose beads (SantaCruz Biotech, Santa Cruz, CA, USA) with rotation at 4°C for 2.5 h. Anti-Myc antibodies (Clontech, Palo Alto, CA, USA) were then added to the precleared whole-cell lysates for 2 h of rotation at 4°C. After this, protein A/G agarose beads were added to the lysates for another 2 h at 4°C. Immunoprecipitates were washed four times with PBS plus 0.1% Triton X-100 and analysed by Western blotting using anti-GFP antibody (SantaCruz Biotech, Santa Cruz, CA, USA).

Transactivity assays

Transcriptional activities of different fragments of EEN protein were first tested with a luciferase reporter assay. pBIND vectors containing different fragments of EEN fused with a GAL4 DNA-binding domain were cotransfected with pG5-luc reporter vector (Promega, Madison, WI, USA), which contains five GAL4-binding sites upstream of a minimal TATA box that is upstream of the firefly luciferase gene. Empty pBIND vector was taken as control. The cells were harvested 48 h after transfection and assayed for firefly and Renilla luciferase activity using the dual-luciferase reporter assay system according to the manufacturer's instructions (Promega).

We also verified these transcriptional activities by yeast one-hybrid assay. Yeast reporter strain AH109 (Clontech, Palo Alto, CA, USA) was transformed with pGBKT7 plasmids containing different EEN domains fused to GAL4 DNA-binding domain. Positive clones on SD/−Trp plates were restreak onto SD/−Trp/−His/−Ade plates. Only those yeasts with transactivation domain containing constructs could grow up on SD/−Trp/−His/−Ade plates.

Reporter luciferase assay

HL60 cells were used to test the HoxA7 promoter activity. Cells were cotransfected by mammalian expression vectors pCI-MLL-EEN and pGL3-HoxA7 (Schreiner et al., 1999) (from Dr Robert Karl Slany, University of Erlangen, Germany) reporter vector. Empty pCI-neo, pCI-GAL4-EEN and pCI-MLLΔ were used as controls. The pRL-TK (Promega, Madison, WI, USA) vector, which leads to constitutive expression of Renilla luciferase, was also cotransfected as an internal control. After incubation for 36 h, cells were harvested and subjected to the luciferase assay using the dual-luciferase reporter assay system (Promega, Madison, WI, USA) on Lumat LB 9507 luminometer (EG&G Berthold, Bad Wildbad, Germany). As a control experiment, reporter luciferase assay was also performed with the pTAL-luc reporter plasmid (Clontech, Palo Alto, CA, USA). The relative luciferase units were calculated according to the firefly luciferase activities normalized by the activities of Renilla luciferase. Each set of experiments was repeated at least three times.


  1. Akao Y, Mizoguchi H, Misiura K, Stec WJ, Seto M, Ohishi N and Yagi K . (1998). Cancer Res., 58, 3773–3776.

  2. Ayton PM and Cleary ML . (2001). Oncogene, 20, 5695–5707.

  3. Ayton PM and Cleary ML . (2003). Genes Dev., 17, 2298–2307.

  4. Birke M, Schreiner S, Garcia-Cuellar MP, Mahr K, Titgemeyer F and Slany RK . (2002). Nucleic Acids Res., 30, 958–965.

  5. Butler LH, Slany R, Cui X, Cleary ML and Mason DY . (1997). Blood, 89, 3361–3370.

  6. Chan LC and Yam JW . (2002). Blood, 100 (Suppl), 199b.

  7. Chen Y, Deng L, Maeno-Hikichi Y, Lai M, Chang S, Chen G and Zhang JF . (2003). Cell, 115, 37–48.

  8. Collins EC and Rabbitts TH . (2002). Trends Mol. Med., 8, 436–442.

  9. Collins SJ . (1987). Blood, 70, 1233–1244.

  10. Corral J, Lavenir I, Impey H, Warren AJ, Forster A, Larson TA, Bell S, McKenzie AN, King G and Rabbitts TH . (1996). Cell, 85, 853–861.

  11. Dimartino JF, Ayton PM, Chen EH, Naftzger CC, Young BD and Cleary ML . (2002). Blood, 99, 3780–3785.

  12. Dimartino JF, Miller T, Ayton PM, Landewe T, Hess JL, Cleary ML and Shilatifard A . (2000). Blood, 96, 3887–3893.

  13. Dobson CL, Warren AJ, Pannell R, Forster A and Rabbitts TH . (2000). EMBO J., 19, 843–851.

  14. Dorrie J, Schuh W, Keil A, Bongards E, Greil J, Fey GH and Zunino SJ . (1999). Leukemia, 13, 1539–1547.

  15. Ernst P, Wang J and Korsmeyer SJ . (2002). Curr. Opin. Hematol., 9, 282–287.

  16. Floyd S and De Camilli P . (1998). Trends Cell Biol., 8, 299–301.

  17. Frank RC, Sun X, Berguido FJ, Jakubowiak A and Nimer SD . (1999). Oncogene, 18, 1701–1710.

  18. Giachino C, Lantelme E, Lanzetti L, Saccone S, Bella VG and Migone N . (1997). Genomics, 41, 427–434.

  19. Hajra A, Liu PP, Wang Q, Kelley CA, Stacy T, Adelstein RS, Speck NA and Collins FS . (1995). Proc. Natl. Acad. Sci. USA, 92, 1926–1930.

  20. Hayashi Y, Honma Y, Niitsu N, Taki T, Bessho F, Sako M, Mori T, Yanagisawa M, Tsuji K and Nakahata T . (2000). Cancer Res., 60, 1139–1145.

  21. Hsu K and Look AT . (2003). Cancer Cell, 4, 81–83.

  22. Huret JL, Dessen P and Bernheim A . (2001). Leukemia, 15, 987–989.

  23. Joh T, Kagami Y, Yamamoto K, Segawa T, Takizawa J, Takahashi T, Ueda R and Seto M . (1996). Oncogene, 13, 1945–1953.

  24. Kamps MP, Look AT and Baltimore D . (1991). Genes Dev., 5, 358–368.

  25. Kawagoe H, Kawagoe R and Sano K . (2001). Leukemia, 15, 1743–1749.

  26. Kersey JH, Wang D and Oberto M . (1998). Leukemia, 12, 1561–1564.

  27. Kitada S, Pedersen IM, Schimmer AD and Reed JC . (2002). Oncogene, 21, 3459–3474.

  28. Kwon KB, Park EK, Ryu DG and Park BH . (2002). Exp. Mol. Med., 34, 32–37.

  29. Lanza C, Gaidano G, Cimino G, Pastore C, Nomdedeu J, Volpe G, Vivenza C, Parvis G, Mazza U, Basso G, Madon E, Lo CF and Saglio G . (1996). Genes Chromosomes Cancer, 15, 48–53.

  30. Lavau C, Du C, Thirman M and Zeleznik-Le N . (2000). EMBO J., 19, 4655–4664.

  31. Lavau C, Szilvassy SJ, Slany R and Cleary ML . (1997). EMBO J., 16, 4226–4237.

  32. Mahgoub N, Parker RI, Hosler MR, Close P, Winick NJ, Masterson M, Shannon KM and Felix CA . (1998). Genes Chromosomes Cancer, 21, 270–275.

  33. Martin ME, Milne TA, Bloyer S, Galoian K, Shen W, Gibbs D, Brock HW, Slany R and Hess JL . (2003). Cancer Cell, 4, 197–207.

  34. Mayer BJ . (2001). J. Cell Sci., 114, 1253–1263.

  35. McPherson PS . (1999). Cell Signal., 11, 229–238.

  36. Mollinedo F, Santos-Beneit AM and Gajate C . (1998). Animal Cell Culture Techniques. Clynes M (ed). Springer-Verlag: New York, pp. 264–297.

  37. Nakamura T, Largaespada DA, Shaughnessy Jr JD, Jenkins NA and Copeland NG . (1996). Nat. Genet., 12, 149–153.

  38. Naoe T, Kubo K, Kiyoi H, Ohno R, Akao Y, Yoshida J, Kato K, Kojima S and Matsuyama T . (1993). Blood, 82, 2260–2261.

  39. Niitsu N, Hayashi Y and Honma Y . (2001). Oncogene, 20, 375–384.

  40. Ormerod MG . (1998). Leukemia, 12, 1013–1025.

  41. Petrelli A, Gilestro GF, Lanzardo S, Comoglio PM, Migone N and Giordano S . (2002). Nature, 416, 187–190.

  42. Pocock CF, Malone M, Booth M, Evans M, Morgan G, Greil J and Cotter FE . (1995). Br. J. Haematol., 90, 855–867.

  43. Polak PE, Simone F, Kaberlein JJ, Luo RT and Thirman MJ . (2003). Mol. Biol. Cell, 14, 1517–1528.

  44. Reutens AT and Begley CG . (2002). Int. J. Biochem. Cell Biol., 34, 1173–1177.

  45. Ringstad N, Nemoto Y and De Camilli P . (1997). Proc. Natl. Acad. Sci. USA, 94, 8569–8574.

  46. Ringstad N, Nemoto Y and De Camilli P . (2001). J. Biol. Chem., 276, 40424–40430.

  47. Rogaia D, Grignani F, Carbone R, Riganelli D, LoCoco F, Nakamura T, Croce CM, Di Fiore PP and Pelicci PG . (1997). Cancer Res., 57, 799–802.

  48. Rowley JD . (1998). Annu. Rev. Genet., 32, 495–519.

  49. Rubnitz JE, Behm FG and Downing JR . (1996). Leukemia, 10, 74–82.

  50. Schichman SA, Caligiuri MA, Gu Y, Strout MP, Canaani E, Bloomfield CD and Croce CM . (1994). Proc. Natl. Acad. Sci. USA, 91, 6236–6239.

  51. Schreiner SA, Garcia-Cuellar MP, Fey GH and Slany RK . (1999). Leukemia, 13, 1525–1533.

  52. Simone F, Polak PE, Kaberlein JJ, Luo RT, Levitan DA and Thirman MJ . (2001). Blood, 98, 201–209.

  53. Slany RK, Lavau C and Cleary ML . (1998). Mol. Cell Biol., 18, 122–129.

  54. So CW, Caldas C, Liu MM, Chen SJ, Huang QH, Gu LJ, Sham MH, Wiedemann LM and Chan LC . (1997). Proc. Natl. Acad. Sci. USA, 94, 2563–2568.

  55. So CW and Cleary ML . (2002). Mol. Cell Biol., 22, 6542–6552.

  56. So CW and Cleary ML . (2003). Blood, 101, 633–639.

  57. So CW, Lin M, Ayton PM, Chen EH and Cleary ML . (2003). Cancer Cell, 4, 99–110.

  58. Soubeyran P, Kowanetz K, Szymkiewicz I, Langdon WY and Dikic I . (2002). Nature, 416, 183–187.

  59. Sparks AB, Hoffman NG, McConnell SJ, Fowlkes DM and Kay BK . (1996). Nat. Biotechnol., 14, 741–744.

  60. Trayner ID, Bustorff T, Etches AE, Mufti GJ, Foss Y and Farzaneh F . (1998). Leuk. Res., 22, 537–547.

  61. Yam JW, Jin DY, So CW and Chan LC . (2003). Blood, DOI:10.1182/blood-2003-07-2452.

  62. Yano T, Nakamura T, Blechman J, Sorio C, Dang CV, Geiger B and Canaani E . (1997). Proc. Natl. Acad. Sci. USA, 94, 7286–7291.

  63. Yu BD, Hess JL, Horning SE, Brown GA and Korsmeyer SJ . (1995). Nature, 378, 505–508.

  64. Zeisig BB, Schreiner S, Garcia-Cuellar MP and Slany RK . (2003). Leukemia, 17, 359–365.

  65. Zeleznik L, Harden AM and Rowley JD . (1994). Proc. Natl. Acad. Sci. USA, 91, 10610–10614.

Download references


We are grateful to Robert Karl Slany for the gift of the pGL3-HoxA7 plasmid. We thank Jian-Xiang Liu for critical reading of this manuscript. We greatly appreciate the technical assistance of Ting Wang and the members in Shanghai Institute of Hematology. This work was supported in part by grants from the Chinese National Key Program for Basic Research (973), the National Natural Science Foundation of China, the Shanghai Commission for Science and Technology, the Shanghai Commission for Education and the Samuel Waxman Cancer Research Foundation.

Author information

Correspondence to Sai-Juan Chen.

Rights and permissions

Reprints and Permissions

About this article


  • transcriptional factor
  • HoxA7
  • leukemogenesis

Further reading