E2A-HLF Usurps Control of Evolutionarily Conserved Survival Pathways

E2A-HLF usurps control of evolutionarily conserved survival pathways

Markus G Seidel1 and A Thomas Look*,1

1Pediatric Oncology Department, Dana-Farber Cancer Institute, 44 Binney Street, M-630, Boston, Massachusetts, MA 02115, USA

E2A-HLF, the chimeric fusion protein resulting from the partner, extradenticle (Kamps et al., 1990; Nourse et al., leukemogenic translocation t(17;19), appears to employ 1990; Izraeli et al., 1992; Numata et al., 1993); for a review evolutionarily conserved signaling cascades for its of E2A-PBX1, see C Murre, in this issue of Oncogene transforming and antiapoptotic functions. These arise Reviews). A third, only recently described E2A chimeric from both impairment of normal E2A function and gene, results from cryptic rearrangements of chromo- activation of a survival pathway triggered through the some 19, apparently fusing 5 aminoterminal sequences of HLF bZip DNA binding and dimerization domain. Recent E2A with various fragments of a previously unknown reports identify wild-type E2A as a tumor suppressor in T gene located at the long arm of chromosome 19 (q13.4), lymphocytes. Moreover, E2A-HLF has been shown to called FB1 (Brambillasca et al., 1999). This region has activate SLUG, a mammalian homologue of the cell death been implicated in frequent cancer-associated loss of speci®cation protein CES-1 in Caenorhabditis elegans, heterozygosity and is also thought to harbor tumor which appears to regulate an evolutionarily conserved cell suppressor functions (Takeuchi et al., 1995; Maintz et al., survival program. Recently, several key mouse models 1997; Bicher et al., 1997; McDonald et al., 1998; Mehenni have been generated, enabling further elucidation of these et al., 1997). The resultant E2A-FB1 fusion, like E2A- pathways on a molecular genetic level in vivo. In this PBX1, gives rise to pre-B-ALL in children; however, the review, we discuss the characteristics of both components oncogenic mechanisms and prevalence of this newly of the fusion protein with regard to their contribution to identi®ed E2A-fusion gene remain to be elucidated. the regulation of cell fate and the oncogenic potential of The E2A-HLF fusion protein contains the N- E2A-HLF. Oncogene (2001) 20, 5718 ± 5725. terminal transactivation domains of E2A joined to the basic region and leucine zipper domain of hepatic Keywords: apoptosis; acute leukemia; chromosomal leukemia factor (HLF) (Inaba et al., 1992; Hunger et translocation; circadian rhythm; hepatic leukemia al., 1992). Two dierent types of genomic rearrange- factor; developmental cell fate ments leading to E2A-HLF fusion have been described (for reviews, see Look, 1997a; Hunger, 1996); both result in a functional protein that retains the DNA- Introduction binding capacity and transcriptional activation domains (Hunger et al., 1994). While E2A plays a major The E2A-HLF fusion protein gives rise to a distinct role in B-cell development and is usually expressed in but fortunately rare form of high risk pro-B-cell acute lymphocytes, the expression of HLF is aberrantly lymphoblastic leukemia (ALL) in adolescents. The driven by the E2A promoter when the fusion gene is unique clinical features of this disease, which often expressed in pro-B lymphoblasts. Therefore, interfer- proves refractory to intensive chemotherapy and is ence with normal E2A function as well as inappropri- frequently associated with disseminated intravascular ate HLF expression, and thereby altered transcriptional coagulation and hypercalcemia at diagnosis, together activity from the fusion transcription factor, are both with the apoptosis-linked oncogenic mechanisms of the likely to be important leukemogenic mechanisms fusion protein, have attracted considerable scienti®c employed by E2A-HLF. Finally, homozygous p16INK4A attention (for previous reviews see Hunger, 1996; and p15INK4B deletions were reported in two E2A-HLF- Look, 1997a,b; Ferrando and Look, 2000). positive cell lines, and hypermethylation of the p15 In addition to the translocation t(17;19)(q22;p13), promoter was seen in one of two primary (t(17;19)+ which generates E2A-HLF, two other E2A transloca- leukemia specimens, providing clues to the other tions are involved in leukemias arising from early B- genetic alterations that may contribute to E2A-HLF- lineage progenitors. Most prevalent in childhood B mediated leukemogenesis (Maloney et al., 1998). lineage ALL is the t(1;19)(q23;p13), resulting in a fusion of E2A amino-terminal sequences, similar to those represented in E2A-HLF, to PBX1, the mammalian E2A as a tumor suppressor homologue of the major homeotic protein heterometric B-cell maturation

*Correspondence: A Thomas Look; The family of basic region/helix ± loop ± helix (bHLH) E-mail: [email protected] transcription factors is characterized by a typical E2A-HLF usurps control of evolutionarily conserved survival pathways MG Seidel and AT Look 5719 protein-nucleic acid interaction motif allowing for T cell maturation and cell fate homo- or heterodimerization and sequence-speci®c DNA binding, and contains a variety of members B-cells are unique in that homodimers of E2A proteins that are implicated in lymphocytic development as (also referred to as class A bHLH proteins) are well as leukemogenesis (E2A, Tal1/Sc1, Tal2, Lyl1, required for the proper E2A function (Bain et al., BHLHB1 (Murre et al., 1989a,b; Begley and Green, 1993; Murre et al., 1991; Shen and Kadesch, 1995), 1999; Baer, 1993; Wang et al., 2000). The E2A locus whereas in other parts of the body, tissue-speci®c encodes three proteins, E12, E47 and E2-5, that are bHLH proteins (class B, such as MyoD (muscle), Tal1 generated by alternative splicing and bind to speci®c and Lyl1 (T cells), and NeuroD (neurons)) hetero- E-box elements that are relevant in immunoglobulin dimerize with E2A to cooperatively mediate bHLH and T-cell receptor gene rearrangements during functions. E2A is also highly expressed in thymocytes development (Murre et al., 1989a,b; Sun and (Roberts et al., 1993; Sawada and Littman, 1993; Baltimore, 1991; Henthorn et al., 1990a,b). Studies Takeda et al., 1990), and T-cell-speci®c genes (eg. the in genetically modi®ed mice have increased the TCRb enhancer, CD4 enhancer and CD4 silencer) have knowledge about E2A substantially. Two dierent been shown to contain E-box sites (Duncan et al., knockout strategies yielded E2A null mice (Table 1; 1996; Sawada and Littman, 1993; Takeda et al., 1990), Bain et al., 1994; Zhuang et al., 1994) with a suggesting that E2A might also be important in T cells. phenotype that included impaired immunoglobulin A more recent ®nding, that a de®ciency of E2A in E2A DJ and V(D)J gene re-arrangements, but normal T- null mice or E2A antagonism in Id1-transgenic mice cell receptor gene rearrangements and normal also leads to partial T-cell maturation arrest and the myeloid cell development. These ®ndings demon- development of T-cell lymphomas, has led to the strated the necessity of E2A proteins in B-cell hypothesis that the T-cell leukemia oncoproteins Tal1 formulation. The E2A7/7 mouse models, as well as and Lyl1 act as dominant inhibitors of E2A (Table 1; a third strategy based on overexpression of a Bain et al., 1997; Yan et al., 1997; Kim et al., 1999; bHLH-sequestering antagonist (the Id1 gene) in a Begley and Green, 1999). In particular, the down- transgenic approach (Table 1; Sun, 1994), showed B- regulation of E47 activity has been shown to be critical cell maturation arrest at the earliest stage of for maturation and enhanced cell death of immature T- committed B-lineage development, right after the cells, and for positive selection of thymocytes (Bain et cells have become common lymphoid progenitors. In al., 1999). This study points out the similarity of the T- addition, it has been shown that E2A is speci®cally lineage phenotypes of E477/7 and Bcl-2 transgenic mice induced during B-cell activation, and although not (such as alteration of the CD4/CD8 ratio), and required for normal proliferation of mature B-cells, emphasizes a possible role of cell fate regulation by E2A is essential for immunoglobulin isotype class E2A proteins. switching during B-cell activation (Quong et al., In fact, other work from C Murre's laboratory 1999). (Engel and Murre, 1999) as well as a report from the

Table 1 Mouse models relevant to the roles of E2A and E2A-HLF Symptoms, malignancy, Genetic manipulation Hematopoietic phenotype life expectancy

E2A7/7 mice Deletion of bHLH domains of E12 B- and T-cell developmental block, T-cell lymphoma associated with or of E12 and E47 by deficient DJ and V(D)J rearrangements c-myc overexpression homologous recombination and B220 expression, and reduced Reduced life-span (Bain et al., 1994; Zhuang et al., 1994) RAG-1 expression Myeloid lineages normal Id1 tg mice Transgene under lck promoter Thymic hypoplasia T-cell lymphoma at (Kim et al., 1999) T-cell maturation arrest 2.5 ± 4 months (mostly DN cells) Reduced life-span Increased apoptosis of differentiating T-cells E2A-HLF tg1 mice Transgene under Ig enhancer Thymic hypoplasia Infections (mainly and promoter Increased T-cell apoptosis bacterial) (Honda et al., 1999) B-cell maturation block T-cell ALL at 3 ± 6 months (reduced B220 expression) Reduced life span E2A-HLF tg2 mice Transgene under Em enhancer Thymic hypoplasia T-cell (and rarely B cell) and SV40 early promoter T-cell maturation arrest and lymphoma after 10 months (Smith et al., 1999) increased apoptosis Reduced life-span Progressive expansion of the primitive DN T cell pool Slug7/7 mice Targeted deletion by homologous recombination Clonogenic progenitor pools enlarged Postnatal growth delay (Inoue et al., 1999) Normal peripheral blood cell counts Chronic eyelid inflammation Hypersensitive to ionizing irradiation Normal life expectancy

Abbreviations: tg, transgenic; DN, CD4 and CD8 double negative; ALL, acute lymphoblastic leukemia

Oncogene E2A-HLF usurps control of evolutionarily conserved survival pathways MG Seidel and AT Look 5720 laboratory of XH Sun (Park et al., 1999) con®rmed, factor. Thus it is likely that the transforming properties that apoptosis can be promoted by either ectopic of E2A-HLF do not depend on perturbation of normal expression of E47 or E12 in E2A-de®cient lymphomas E2A function, but also on the induction of novel, or by restoration of E47 activity in Jurkat cells, a Tal1- potentially cancer-linked target genes of the chimeric expressing T-cell line. These ®ndings support the idea fusion protein (Figure 1). In the fusion protein, the that the regulation of cell fate decisions, or facilitation dimerization and DNA-binding domains of E2A are of pro-apoptotic events, may be one of the functions of replaced by the corresponding domains of HLF. E2A. Thus, loss of normal E2A function may lead to Together with DBP (albumin promoter D-box binding aberrant survival, as suggested by the anti-apoptotic protein; Mueller et al., 1990), and TEF (thyrotroph activity of amino-terminal fragments of E2A-HLF embryonic factor; (Drolet et al., 1991), HLF belongs to (Inukai et al., 1998). Thus, one mechanism of the PAR (proline and acidic amino acid-rich) subfamily oncogenicity used by E2A-HLF might be impairment of bZIP transcription factors (Hunger et al., 1992; of normal E2A function. It is intriguing, that mice, Inaba et al., 1992). By primary amino acid sequence engineered to constitutively overexpress E2A-HLF, similarity, it is closely related to the Drosophila gene develop T-cell lymphomas and B- and T-lineage vrille, a developmental transcription factor that plays a maturation defects similar to those seen in E2A null major role in circadian rhythm control. Vrille is or Id1 transgenic mice (Table 1; Honda et al., 1999; regulated by the bHLH transcription factors Smith et al., 1999; Kim et al., 1999). Collectively, these (dCLOCK and CYCLE in an oscillating manner via ®ndings demonstrate a role of the E2A component of E-box motifs and apparently acts by negatively the fusion protein in the regulation of cell fate, regulating other clock genes, such as period and suggesting that compromised E2A function may timeless (Blau and Young, 1999). Likewise, murine contribute to early B lineage leukemias in progenitors DBP mRNA was shown to oscillate not only in the harboring the t(17;19). liver and in other organs, but also in the suprachiasmatic nuclei of the hypothalamus, a region thought to comprise the central mammalian pacemaker (Wuarin et Functions in cell cycle control al., 1992; Fonjallaz et al., 1996; Lavery et al., 1996; In addition to the function of E2A in cell fate Lopez-Molina et al., 1997). Recent publications suggest decisions, bHLH proteins such as E2A and MyoD that DBP functions in part similarly to VRILLE in the are involved in the induction of G1 cell cycle arrest mammalian clock, by virtue of its ability to regulate (Peverali et al., 1994; Sorrentino et al., 1990). mPer1 and to be rhythmically induced by CLOCK Correspondingly, the bHLH-sequestering Id proteins (Yamaguchi et al., 2000; Ripperger et al., 2000). have been shown to promote G1-S transition (Barone Although cycling expression levels of HLF and TEF et al., 1994; Hara et al., 1994). One molecule acting at have also been observed in various organs (Falvey et that transition during cell cycle, p21CIP1/WAF1,an al., 1995; Fonjallaz et al., 1996), HLF did not appear inhibitor of cyclin-dependent kinases (Cdk), was to be highly expressed in the suprachiasmatic nuclei recently shown to be regulated by E2A and Id proteins (Hitzler et al., 1999). Thus it is likely that other (Prabhu et al., 1997). Other laboratories have demon- members of this protein family, more similar by amino strated that Cdk-2-dependent phosphorylation of Id2 acid sequence to the Drosophila protein VRILLE than and Id3 prevents these proteins from antagonizing HLF, e.g. DBP or NFIL3, may represent better E2A-mediated cell cycle arrest (Hara et al., 1997; Deed candidates for VRILLE homologues in the mammalian et al., 1997). In addition, Chu and Kohtz (2001) have circadian rhythm regulation pathways. recently identi®ed G1 Cdk sites in the amino-terminal Recently, an in vitro study suggested a role of HLF domain of E12/E47 suggesting that E2A itself is a alone or in combination with DBP in promoting the target of Cdk phosphorylation ± mediated regulation. expression of Factor VIII and Factor IX genes in the These studies implicate E2A as an integrating molecule liver, a function that could also be demonstrated for in signals regulating G1-S phase cell cycle checkpoints. E2A-HLF (Begbie et al., 1999). It remains to be seen In summary, E2A serves as a tumor suppressor par whether this ®nding is related to the clinical predis- excellence, mediating essential functions in the regula- position of t(17;19)+ ALL patients to present with tion of dierentiation, maturation, cell fate, and the disseminated intravascular coagulation. cell cycle, so that modi®cations of its expression level likely contribute in important ways to leukemogenesis. Cell fate regulation by HLF-related genes HLF expression in the brain and other organs during mouse development was studied extensively, because Both components of E2A-HLF contribute to another closely related bZIP transcription factor with leukemogenic pathways the same DNA-binding consensus site has been discovered to play a role in neuronal apoptosis, namely Hepatic leukemia factor (HLF) and related genes the Caenorhabditis elegans cell death speci®cation The DNA-binding speci®city of E2A-HLF is distinct protein CES-2, which is required for the regulated cell from that of E2A, due to in-frame fusion to the basic death of two sister cells of serotoninergic neurosecre- region/leucine zipper domains of hepatic leukemia tory motor neurons during the development of the

Oncogene E2A-HLF usurps control of evolutionarily conserved survival pathways MG Seidel and AT Look 5721

Figure 1 Cell fate pathways regulated by E2A, HLF and related bZip proteins, SLUG and related zinc ®nger transcription factors, and the E2A-HLF chimera worm (Ellis and Horvitz, 1991; Metzstein et al., 1996). fact, the CES-1-related zinc ®nger transcription factor The CES-2 homologue E2A-HLF was shown to SLUG is an upregulated target of E2A-HLF and mediate survival of leukemia cells and to inhibit promotes survival of hematopoietic cells if over- apoptosis after growth factor withdrawal from cell expressed in vitro in an analogous fashion to E2A- lines (Inaba et al., 1996), suggesting involvement of HLF (Inukai et al., 1999). HLF in neuronal apoptosis. However, Hitzler et al. (1999) could not identify a role for HLF in Functional domains of E2A-HLF developmentally regulated neuronal cell death. NFIL3/E4BP4, encoded by an IL-3-regulated de- The ability of E2A-HLF to cause malignant transfor- layed early gene, is another HLF-related but PAR- mation was demonstrated by induction of anchorage- de®cient bZIP transcription factor that has been independent growth of NIH3T3 ®broblasts in soft agar implicated in IL-3-mediated survival of pro-B lympho- and subsequent tumor formation when these cells were cytes (Cowell et al., 1992; Cowell and Hurst, 1994; transplanted in nude mice (Yoshihara et al., 1995). Zhang et al., 1995; Ikushima et al., 1997; Kuribara et This oncogenic activity has been shown to require at al., 1999). This putative role becomes very important least one of the two trans-activation domains of E2A, with regard to the oncogenic transcription factor E2A- as well as the leucine zipper dimerization domain of HLF, because the latter carries the same capacity to HLF (Yoshihara et al., 1995). In addition, the promote survival of murine hematopoietic cells if transforming ability also depends on homodimerization overexpressed in vitro. Thus, survival of native of E2A-HLF (Inukai et al., 1997). However, the t(17;19)-harboring cells could be ascribed to constitu- antiapoptotic function of E2A-HLF in vitro requires tive activation of a survival pathway physiologically neither the basic (DNA-binding) region nor the leucine activated by hematopoietic growth factors via NFIL3 zipper domain needed for protein ± protein dimeriza- (Inaba et al., 1996; Ikushima et al., 1997; Kuribara et tion (Inukai et al., 1998). Furthermore, in contrast to al., 1999). These results support the contention that all other cell lines tested, the HAL-01 patient-derived E2A-HLF, through the aberrant localization of two t(17;19)-positive leukemia cell line, which harbors a trans-activating domains of E2A at pathological HLF- point mutation within the leucine zipper region of speci®c consensus sites, might actively employ a HLF, does not express the E2A-HLF target gene survival signaling pathway to perform its oncogenic SLUG on a level detectable by RT ± PCR (Inukai et al., functions. These eects apparently occur as a main 1999). Consistent with these ®ndings, deletion of the oncogenic event in addition to the perturbation of entire homeodomain of PBX1 did not impair the normal E2A function in this leukemia (Figure 1). In potential of the E2A-PBX1 fusion protein to transform

Oncogene E2A-HLF usurps control of evolutionarily conserved survival pathways MG Seidel and AT Look 5722 ®broblasts as well as to induce thymoma in mice (Gilks et al., 1993; Grimes et al., 1996). GFI-1 shares (Kamps et al., 1996; Monica et al., 1994). Moreover, the SNAG (Snail and G®-1) transcriptional repressor the entire PBX1-derived portion of E2A-PBX1, which domain with SLUG, and was shown to inhibit BAX in is comparable to the carboxy-terminal HLF part in vitro and in primary cells of GFI-1-induced thymomas, E2A-HLF, can be replaced by DNA-binding/dimeriza- supporting the hypothesis that SLUG/Snail family tion or even only dimerization motifs from other proteins, like CES-1, can regulate cell fate (Figure 1; proteins, without losing the in vitro transforming Grimes et al., 1996). However, GFI-1 appears to be a potential of the intact oncoprotein. This hallmark of rather weak oncogenic and requires cooperation with oncogenicity was completely retained even in the c-myc and pim in oncogenicity (Schmidt et al., 1998). absence of PBX1 and depended only on dimerization PAG-3, a GFI-1 homologue in C. elegans, is a zinc- of the E2A fragment of E2A-PBX1 (Bayly and ®nger transcription factor involved in neuronal devel- LeBrun, 2000). Therefore, in both of the fusion opment and is thought to share functional properties proteins, E2A-HLF and E2A-PBX1, the E2A compo- with GFI-1 (Jia et al., 1997). A clear prediction is that nent appears to be sucient to trigger oncogenic PAG-3 may also play a role in cell fat decisions during transformation if misexpressed and dimerized due to C. elegans development. the fusion partner. The striking hematopoietic similarity between two Target genes of E2A-HLF and an evolutionarily independently derived E2A-HLF transgenic mouse conserved apoptosis program models, E2A-PBX1 transgenic mice, and E2A-de®cient mice ± in particular with regard to the emergence of Inukai et al. (1999) identi®ed human SLUG as a target increasingly immature lymphoid cell populations, gene of E2A-HLF by representational dierence analysis lymphocytic maturation arrest and development of T of RNA pools obtained from a t(17;19)+ leukemia cell lymphomas ± lends further support to the conclu- patient cell line, which was stably transfected with an sion that impairment of normal E2A function provides inducible dominant negative form of E2A-HLF (Inukai a central mechanism of oncogenesis (Table 1) (Smith et et al., 1999). Other genes identi®ed in this screen as being al., 1999; Honda et al., 1999; Dedera et al., 1993; Bain upregulated in an E2A-HLF-dependent manner also et al., 1994; Zhuang et al., 1994). However, Slug- serve as potential downstream targets of E2A-HLF, de®cient mice also show a phenotype of perturbed namely, ANNEXIN VIII and SRPUL (sushi repeat hematopoiesis and appear to be sensitive to DNA protein upregulated in leukemia) (Kurosawa et al., damage induced apoptosis within the hematopoietic 1999). However, the latter two genes could not prevent compartment (Table 1; Inoue et al., 1999; Inoue et al., apoptosis when overexpressed in murine IL-3-dependent manuscript submitted). These ®ndings indicate the leukemia cell lines, suggesting that they are not causally operation of an exclusively E2A-HLF-mediated me- involved in the antiapoptotic functions of E2A-HLF chanism to promote the survival of lymphoblasts, one (Kurosawa et al., 1999). As pointed out previously, the that is unconnected to impaired E2A function. E2A-HLF oncogene is closely related to ces-2, a cell In summary, two characteristic features of E2A- death speci®cation gene in Caenorhabditis elegans HLF, loss of one E2A allele and abnormal expression (Metzstein et al., 1996; Inaba et al., 1996). By amino of the HLF DNA-binding domain, likely contribute acid sequence homology and functional properties as a both eciently and synergistically to the oncogenic target of E2A-HLF, SLUG represents a mammalian properties of E2A-HLF. homologue of CES-1, which acts downstream of CES-2 in C. elegans (Figure 1; Metzstein and Horvitz, 1999; Inukai et al., 1999). Ces-2 and ces-1 control programmed cell death in sister cells of neurosecretory motor neurons SLUG and cell fate regulation by E2A-HLF during development in C. elegans (Ellis and Horvitz, 1991; Metzstein et al., 1996, 1998; Metzstein and Slug and snail transcription factors in cancer Horvitz, 1999). The resemblance of cell death pathways Slug belongs to the Slug/Snail family of evolutionarily including CES-2/CES-1 in the worm and E2A-HLF/ conserved zinc ®nger transcriptional repressors (Boulay SLUG in human pro-B leukemia suggested that SLUG et al., 1987; Hemavathy et al., 2000b). Besides many might have an important regulatory role in the survival functions in embryonic development, Snail-related of lymphoid or other types of hematopoietic cells. In the proteins are also involved in carcinogenesis through worm, downregulation of ces-1 to abolish the CES-1- aberrant regulation of apoptosis (Metzstein and mediated repression of EGL-1 promotes the develop- Horvitz, 1999; Inukai et al., 1999; Grimes et al., mental cell death of two neuronal sister cells of 1996) and possibly an invasiveness/metastasis by neurosecretory motor neurons. EGL-1, the C. elegans regulating the expression of E-cadherins (Figure 1; homologue of pro-apoptotic Bcl-2 family members, Oda et al., 1998; Batlle et al., 2000; Cano et al., 2000; contains a BH3 (Bcl-2 homology) domain and is required for review, see Hemavathy et al., 2000a). GFI-1, a to sequester CED-9 from CED-4 leading to activation of more distant relative of SLUG in the SLUG/Snail CED-3, which are the homologues of the Bcl-2, Apaf-1 family, is an oncoprotein that can render IL-2- and caspase-9 mammalian mediators of the apoptosis dependent T cells growth factor independent and was cascade (Metzstein and Horvitz, 1999). According to this shown to cause T cell lymphomas in transgenic mice model, SLUG may inhibit mitochondrial dysfunction

Oncogene E2A-HLF usurps control of evolutionarily conserved survival pathways MG Seidel and AT Look 5723

Figure 2 Models for the roles of CES-1 and SLUG in apoptosis. (Right) Programmed cell death of NSM sister neurons during C. elegans development involves downregulation of CES-1, which in turn directly or indirectly leads to increased activity of the BH3-only cell death protein, EGL-1. (Left) In pro-B leukemia cells, the CES-1 related protein, SLUG, is postulated to act through the downregulation of mammalian BH3-containing proteins to promote the aberrant survival of cells with defective immunoglobulin genes that ordinarily would be targeted for apoptotic cell death (see Inukai et al., 1998) and the cytochrome c/Apaf-1/caspase-9 apoptotic path- cells. These results suggest that hematopoietic cells in way by transcriptionally downregulating one of the these mice might be prone to apoptosis (Inoue et al., mammalian EGL-1 homologs, e.g., BID/BAX/BAK, manuscript submitted). The study by Inoue et al. whose interaction plays a major role in caspase-9 implicates SLUG as a survival factor in normal activation (Figure 2; Inukai et al., 1999; Lindsten et al., hematopoietic cell development and supports the 2000). model of cooperative oncogenic signaling triggered by Jiang et al. (1998) reported that mice lacking Slug the two components of the E2A-HLF fusion protein. are fertile but show postnatal growth delay and eyelid Future research will focus on the identi®cation of abnormalities; their mesoderm and neural crest devel- downstream targets of the E2A-HLF/SLUG pathway opment appeared normal. Inoue et al. (1999) indepen- in promoting the aberrant survival of lymphoid dently derived a Slug-de®cient mouse model to leukemia cells. The death speci®cation pathways investigate the role of Slug in hematopoiesis and normally regulated by Slug will also be important apoptosis. As summarized in Table 1, Inoue and areas for study, given that Slug expression appears to colleagues observed increased numbers of clonogenic be aberrantly activated by E2A-HLF and not normally hematopoietic progenitors in the bone marrow and expressed by early B cell progenitors. Genetically spleen of Slug7/7 knockout mice, although the engineered animal models lacking components of peripheral blood cell counts were normal, possibly evolutionarily conserved apoptosis signaling cascades re¯ecting a feedback loop to compensate for loss of represent a powerful tool for identifying these cell dierentiating hematopoietic cells (Inoue et al., 1999; death speci®cation pathways and de®ning their roles in Inoue et al., manuscript submitted). Slug-de®cient mice normal hematopoiesis. also showed enhanced hematopoietic cell destruction when challenged with total-body g-irradiation, given as a DNA-damaging agent. This eect could be prevented Acknowledgments by a single-dose treatment with a truncated pegylated The authors thank John Gilbert for editorial review and form of megakaryocyte growth and development factor Stephen P Hunger and David W Sternberg Jr for reading (MGDF), which acts as a survival factor for myeloid the manuscript and their critical comments.

References

Baer R. (1993). Sem. Cancer Biol., 4, 341 ± 347. Bain G, Robanus-Maandag EC, Izon DJ, Amsen D, Bain G, Engel I, Maandag EC, te Riele HPJ, Voland JR, Kruisbeek AM, Weintraub BC, Krop I, Schlissel MS, Sharp LL, Chun J, Huey B, Pinkel D and Murre C. (1997). Feeney AJ, van Roon M, van der Valk M, te Riele HPJ, Mol. Cell. Biol., 17, 4782 ± 4891. Berns A and Murre C. (1994). Cell, 79, 885 ± 892. Bain G, Gruenwald S and Murre C. (1993). Mol. Cell. Biol., Barone VM, Pepperkok R, Peverali FA and Philipson L. 13, 3522 ± 3529. (1994). Proc. Natl. Acad. Sci. USA, 91, 4985 ± 4988. Bain G, Quong MW, Solo RS, Hedrick SM and Murre C. (1999). J. Exp. Med., 190, 1605 ± 1616.

Oncogene E2A-HLF usurps control of evolutionarily conserved survival pathways MG Seidel and AT Look 5724 Batlle E, Sancho E, Franci C, Dominguez D, Monfar M, Honda H, Inaba T, Suzuki T, Oda H, Ebihara Y, Tsuiji K, Baulida J and Garcia DH. (2000). Nat. Cell Biol., 2, 84 ± Nakahata T, Ishikawa T, Yazaki T and Hirai H. (1999). 89. Blood, 93, 2780 ± 2790. Bayly R and LeBrun DP. (2000). Mol. Cell Biol., 20, 5789 ± Hunger SP. (1996). Blood, 87, 1211 ± 1224. 5796. Hunger SP, Brown R and Cleary ML. (1994). Mol. Cell. Begbie M, Mueller C and Lillicrap D. (1999). DNA Cell Biol., Biol., 14, 5986 ± 5996. 18, 165 ± 173. Hunger SP, Ohyashiki K, Toyama K and Cleary ML. (1992). Begley CG and Green AR. (1999). Blood, 93, 2760 ± 2770. Genes Dev., 6, 1608 ± 1620. Bicher A, Ault K, Kimmelman A, Gershenson D, Reed E Ikushima S, Inukai T, Inaba T, Nimer SD, Cleveland JL and and Liang B. (1997). Gynecol. Oncol., 66, 36 ± 40. Look AT. (1997). Proc. Natl. Acad. Sci. USA, 94, 2609 ± Blau J and Young MW. (1999). Cell, 99, 661 ± 671. 2614. Boulay JL, Dennefeld C and Alberga A. (1987). Nature, 330, Inaba T, Inukai T, Yoshihara T, Seyschab H, Ashmun RA, 395 ± 398. Canman CE, Laken SJ, Kastan MB and Look AT. (1996). Brambillasca F, Mosna G, Colombo M, Rivolta A, Caslini Nature, 382, 541 ± 544. C,MinuzzoM,GiudiciG,MizziL,BiondiAandPrivitera InabaT,RobertsWM,ShapiroLH,JollyKW,Raimondi E. (1999). Leukemia, 13, 369 ± 375. SC, Smith SD and Look AT. (1992). Science, 257, 531 ± Cano A, Perez-Moreno MA, Rodrigo I, Locascio A, Blanco 534. MJ, del Barrio MG, Portillo F and Nieto MA. (2000). Nat. InoueA,MorimotoA,HitzlerJK,PestinaTI,JacksonCW, Cell Biol., 2, 76 ± 83. Tanaka R, Bronson RT, Chong M, McKinnon PJ, Inukai Chu C and Kohtz DS. (2001). J. Biol. Chem., 276, 8524 ± T, Ashmun RA, Grosveld GC and Look AT. (1999). 8534. Blood, 94, 3040 ± 3040. Cowell IG and Hurst HC. (1994). Nucleic Acids Res., 22, 59 ± Inukai T, Inaba T, Ikushima S and Look AT. (1998). Mol. 65. Cell. Biol., 18, 6035 ± 6043. Cowell IG, Skinner A and Hurst HC. (1992). Mol. Cell Biol., Inukai T, Inaba T, Yoshihara T and Look AT. (1997). Mol. 12, 3070 ± 3077. Cell. Biol., 17, 1417 ± 1424. Dedera DA, Waller EK, LeBrun DP, Sen-Majumdar A, Inukai T, Inoue A, Kurosawa H, Goi K, Shinjyo T, Ozawa Stevens ME, Barsh GS and Cleary ML. (1993). Cell, 74, K, Mao M, Inaba T and Look AT. (1999). Mol. Cell, 4, 833 ± 843. 343 ± 352. Deed RW, Hara E, Atherton GT, Peters G and Norton JD. Izraeli S, Kovar H, Gadner H and Lion T. (1992). Blood, 80, (1997). Mol. Cell. Biol., 17, 6815 ± 6821. 1413 ± 1417. Drolet DW, Scully KM, Simmons DM, Wegner M, Chu KT, Jia Y, Xie G, McDermott JB and Aamodt E. (1997). Swanson LW and Rosenfeld MG. (1991). Genes Dev., 5, Development, 124, 2063 ± 2073. 1739 ± 1753. Jiang R, Lan Y, Norton CR, Sundberg JP and Gridley T. Duncan DD, Adlam M and Siu G. (1996). Immunity, 4, 301 ± (1998). Dev. Biol., 198, 277 ± 285. 311. Kamps MP, Murre C, Sun XH and Baltimore D. (1990). Ellis RE and Horvitz HR. (1991). Development, 112, 591 ± Cell, 60, 547 ± 555. 603. Kamps MP, Wright DD and Lu Q. (1996). Oncogene, 12, Engel I and Murre C. (1999). Proc. Natl. Acad. Sci. USA, 96, 19 ± 30. 996 ± 1001. Kim D, Peng XC and Sun XH. (1999). Mol. Cell. Biol., 19, Falvey E, Fleury-Olela F and Schibler U. (1995). EMBO J., 8240 ± 8253. 14, 4307 ± 4317. Kuribara R, Kinoshita T, Miyajima A, Shijyo T, Yoshihara Ferrando AA and Look AT. (2000). Semin. Hematol., 37, T, Inukai T, Ozawa K, Look AT and Inaba T. (1999). Mol. 381 ± 395. Cell. Biol., 19, 2754 ± 2762. Fonjallaz P, Ossipow V, Wanner G and Schibler U. (1996). Kurosawa H, Goi K, Inukai T, Inaba T, Chang KS, Shinjyo EMBO J., 15, 351 ± 356. T, Rakestraw K, Naeve CW and Look AT. (1999). Blood, Gilks CB, Bear SE, Grimes HL and Tsichlis PN. (1993). Mol. 93, 321 ± 332. Cell Biol., 13, 1759 ± 1768. Lavery DJ, Schmidt EE and Schibler U. (1996). The PAR Grimes HL, Gilks CB, Chan TO, Porter S and Tsichlis PN. transcription factor family and circadian gene expression: (1996). Proc. Natl. Acad. Sci. USA, 93, 14569 ± 14573. Greppin H, Degli Agosti R and Bonzon M (eds) Geneva- Hara E, Hall M and Peters G. (1997). EMBO J., 16, 332 ± 135 ± 145. 342. LindstenT,RossAJ,KingA,ZongW,RathmellJC,Shiels HaraE,YamaguchiT,NojimaH,IdeT,CampisiJ, HA, Ulrich E, Waymire KG, Mahar P, Frauwirth K, Chen Okayama H and Oda K. (1994). J. Biol. Chem., 269, Y,WeiM,EngVM,AdelmanDM,SimonMC,MaA, 2139 ± 2145. Golden JA, Evan G, Korsmeyer SJ, MacGregor GR and Hemavathy K, Ashraf SI and Ip YT. (2000a). Gene, 257, 1± Thompson CB. (2000). Mol. Cell., 6, 1389 ± 1399. 12. Look AT. (1997a). E2A-HLF chimeric transcription factors Hemavathy K, Guru S, Harris J, Chen JD and Ip YT. in pro-B cell acute lymphoblastic leukemia. Chromosomal (2000b). Mol. Cell. Biol., 26, 5087 ± 5095. Translocations and Oncogenic Transcription Factors. Henthorn P, Kiledjian M and Kadesch T. (1990a). Science, Rauscher F, III and Vogt PK (eds) Berlin: Springer- 247, 467 ± 470. Verlag, pp. 45 ± 53. Henthorn P, McCarrick-Walmsley R and Kadesch T. Look AT. (1997b). Science, 278, 1059 ± 1064. (1990b). Nucleic Acids Res., 18, 677. Lopez-Molina L, Conquet F, Dubois-Dauphin M and Hitzler JK, Soares HD, Drolet DW, Inaba T, O'Connell S, Schibler U. (1997). EMBO J., 16, 6762 ± 6771. Rosenfeld MG, Morgan JI and Look AT. (1999). Brain Res., 820, 1±11.

Oncogene E2A-HLF usurps control of evolutionarily conserved survival pathways MG Seidel and AT Look 5725 Maintz D, Fiedler K, Koopman J, Rollbrocker B, Nechev S, Quong MW, Harris DP, Swain SL and Murre C. (1999). Lenartz D, Stangl AP, Louis DN, Schramm J, Wiestler EMBO J., 18, 6307 ± 6318. OD and Von Deimling A. (1997). J. Neuropathol. Exp. Ripperger JA, Shearman LP, Reppert SM and Schibler U. Neurol., 56, 1098 ± 1104. (2000). Genes Dev., 14, 679 ± 689. Maloney KW, McGavran L, Odom LF and Hunger SP. Roberts VJ, Steenbergen R and Murre C. (1993). Proc. Natl. (1998). Leukemia, 12, 1417 ± 1421. Acad.Sci.USA,90, 7583 ± 7587. McDonald JM, Douglass EC, Fisher R, Geiser CF, Krill CE, Sawada S and Littman DR. (1993). Mol. Cell. Biol., 13, Strong LC, Virshup D and Hu V. (1998). Cancer Res., 58, 5620 ± 5628. 1387 ± 1390. Schmidt T, Karsunky H, Gau E, Zevnik B, Elsasser HP and Mehenni H, Aalto-Setala K, Blouin J-L, Radhakrishna U, Moroy T. (1998). Oncogene, 17, 2661 ± 2667. BhardwajSS,BhardwajK,DixitVB,RichardsKF, Shen CP and Kadesch T. (1995). Mol. Cell. Biol., 15, 4518 ± Bermejo-Fenoll A, Leal AS, Raval RC and Antonarakis 4524. SE. (1997). Am.J.Hum.Genet.,61, 1327 ± 1334. Smith KS, Rhee JW, Naumovski L and Cleary ML. (1999). Metzstein M and Horvitz HR. (1999). Mol. Cell., 4, 309 ± Mol. Cell. Biol., 19, 4443 ± 4451. 319. Sorrentino V, Pepperkok R, Davis RL, Ansorge W and Metzstein MM, Hengartner MO, Tsung N, Ellis RE and Philipson L. (1990). Nature, 345, 813 ± 815. Horvitz HR. (1996). Nature, 382, 545 ± 547. Sun XH. (1994). Cell, 79, 893 ± 900. Metzstein MM, Stan®eld GM and Horvitz HR. (1998). TIG, Sun XH and Baltimore D. (1991). Cell, 64, 459 ± 470. 14, 410 ± 416. Takeda J, Cheng A, Mauxion F, Nelson CA, Newberry RD, Monica K, LeBrun DP, Dedera DA, Brown R and Cleary Sha WC, Sen R and Loh DY. (1990). Mol. Cell. Biol., 10, ML. (1994). Mol. Cell. Biol., 14, 8304 ± 8314. 5027 ± 5035. Mueller CR, Maire P and Schibler U. (1990). Cell, 61, 279 ± Takeuchi S, Bartram CR, Wada M, Reiter A, Hatta Y, Seriu 291. T, Lee E, Miller CW, Miyoshi I and Koeer HP. (1995). Murre C, McCaw PS and Baltimore D. (1989b). Cell, 56, Cancer Res., 55, 5377 ± 5382. 777 ± 783. Wang J, Jani-Sait SN, Escalon EA, Carroll AJ, de Jong PJ, Murre C, McCaw PS, Vaessin H, Caudy M, Jan LY, Cabrera Kirsch IR and Aplan PD. (2000). Proc. Natl. Acad. Sci. CV, Buskin JN, Hauschka SD, Lassar AB, Weintraub H USA, 97, 3497 ± 3502. and Baltimore D. (1989a). Cell, 58, 537 ± 544. Wuarin J, Falvey E, Lavery D, Talbot D, Schmidt E, Murre C, Voronova A and Baltimore D. (1991). Mol. Cell Ossipow V, Fonjallaz P and Schibler U. (1992). J. Cell Biol., 11, 1156 ± 1160. Sci., 16, 123 ± 127. Nourse J, Mellentin JD, Galili N, Wilkinson J, Stanbridge E, Yamaguchi S, Mitsui S, Yan L, Yagita K, Miyake S and Smith SD and Cleary ML. (1990). Cell, 60, 535 ± 545. Okamura H. (2000). Mol. Cell. Biol., 20, 4773 ± 4781. Numata S, Kato K and Horibe K. (1993). Leukemia, 7, Yan W, Young AZ, Soares VC, Kelley R, Benezra R and 1441 ± 1444. Zhuang Y. (1997). Mol. Cell. Biol., 17, 7317 ± 7327. Oda H, Tsukita S and Takeichi M. (1998). Dev. Biol., 203, Yoshihara T, Inaba T, Shapiro LH, Kato J and Look AT. 435 ± 450. (1995). Mol. Cell. Biol., 15, 3247 ± 3255. Park ST, Nolan GP and Sun XH. (1999). J. Exp. Med., 189, Zhang W, Zhang J, Kornuc M, Kwan K, Frank R and Nimer 501 ± 508. SD. (1995). Mol. Cell. Biol., 15, 6055 ± 6063. Peverali FA, Ramqvist T, Sarich R, Pepperkok R, Barone Zhuang Y, Soriano P and Weintraub H. (1994). Cell, 79, MV and Philipson L. (1994). EMBO J., 13, 4291 ± 4301. 875 ± 884. Prabhu S, Ignatova A, Park ST and Sun X-H. (1997). Mol. Cell. Biol., 17, 5888 ± 5896.

Oncogene