Leukemia (2003) 17, 2454–2459 & 2003 Nature Publishing Group All rights reserved 0887-6924/03 $25.00 www.nature.com/leu B-cell development in the presence of the MLL/AF4 oncoprotein proceeds in the absence of HOX A7 and HOX A9 expression FE Bertrand1, JD Spengeman1, N Shah2 and TW LeBien2

1Department of Microbiology and Immunology, Brody School of Medicine at East Carolina University, USA; and 2University of Minnesota Cancer Center, University of Minnesota, USA

Infant acute lymphoblastic leukemia (ALL) is frequently char- stages.1,11,12 Infant ALL is a distinct subset characterized by acterized by the t(4;11)(q21;q23) cytogenetic abnormality the t(4;11)(q21;q23) cytogenetic abnormality harboring the encoding the MLL/AF4 oncogene, increased HOX expres- sion and a pro-B/monocytoid phenotype. We have previously MLL/AF4 [mixed-lineage leukemia (MLL)/ALL-1 fused gene from established a novel MLL/AF4-positive cell line, B-lineage 3 4 (AF4)] fusion oncogene. MLL/AF4-positive (BLIN-3), which retains several features of normal B-lineage leukemic blasts generally express a ‘mixed-lineage’ (pro-B/ development (functional Ig gene rearrangement and apoptotic monocyte) phenotype.11,12 While some MLL/AF4-expressing sensitivity to stromal cell withdrawal) not generally observed in cell lines have exhibited limited acquisition of myeloid lineage infant ALL. We now use microarray analysis to identify patterns characteristics upon supraphysiological treatment with phorbol of in BLIN-3 that may modulate MLL/AF4 13 oncogenesis and contribute to the retention of normal B- esters, loss of B-lineage developmental potential is a common lineage developmental characteristics. Comparison of 6815 feature. This loss of B-lineage developmental potential in MLL/ expressed in BLIN-3 with published microarray data on AF4-positive leukemias is evident not only in failure to leukemic blasts from t(4;11) patients indicated that BLIN-3 was differentiate, but also by loss of the normal requirement for unique in lacking the expression of certain HOX-A cluster BM stromal cells/growth factors for regulated cell growth and genes. These results were validated by RT-PCR showing no loss of apoptotic sensitivity due to growth factor deprivation. expression of HOX A7 or HOX A9 in BLIN-3. A HOX C8 promoter reporter was active in BLIN-3, indicating that lack of HOX gene A role for MLL (also called ALL-1 and Hrx) and AF4 in expression in BLIN-3 was not due to a nonfunctional MLL/AF4. hematopoiesis is not well defined. MLL is a 430 kDa transcrip- Our results suggest that B-lineage development can proceed in tion factor highly homologous to the Drosophila trithorax gene t(4;11) leukemic blasts in the absence of HOX-A gene expres- (trx).12 It has a complex structure that includes an AT-hook sion. domain for DNA binding, a methyltransferase region and a SET Leukemia (2003) 17, 2454–2459. doi:10.1038/sj.leu.2403178 domain that bears the most similarity to trx. MLL null mice have Published online 16 October 2003 Keywords: B-cell development; MLL; AF4; t(4;11); HOX; defects in HOX gene expression and homeotic defects, microarray; leukemia suggesting a role for MLL in the maintenance of HOX gene expression.14,15 A recent report demonstrated that MLL can mediate high levels of HOX C8, HOX A7 and HOX A9 expression through the binding of the MLL SET domain to HOX Introduction 16 promoters. However, as truncated MLL does not contain the SET domain, the mechanism(s) of elevated HOX gene expres- B cells differentiate from hematopoietic stem cells in a stepwise sion as a result of MLL fusions remain elusive. Recently programmed fashion through a series of developmental check- 1 published work by Cleary’s group suggests that dimerization of points (reviewed by LeBien ). The pro-B cell is the first readily 17 MLL fusions may partly explain this apparent paradox. AF4 identifiable B-lineage committed precursor. Pro-B cells exhibit contains nuclear localization signals and GTP-binding domains, surface expression of the ‘stem cell antigen’ CD34 and the pan 12 but its function is unknown. AF4-deficient mice exhibit B-cell marker CD19.2,3 Most pro-B cells bear DJ rearrange- inhibition of T-cell development and a modest alteration in B- ments, although VDJ rearrangements and m-heavy-chain ex- 18 cell development. pression can be found in a minority of the CD34 þ CD19 þ The MLL/AF4 fusion preserves the AT-hook and compartment.2,4 Pre-B cells are defined by cytoplasmic expres- 5 þ methyltransferase domains of MLL and the GTP binding, and sion of m-heavy chain. These cells are CD19 and have lost 12 nuclear localization regions of AF4. Although the contribution surface expression of CD34. A subset of pre-B cells express m- of MLL/AF4 to leukemogenesis is not precisely defined, MLL- heavy on the cell surface in association with Vpre and l5.6 lacZ fusions are sufficient to confer oncogenesis in mice, Together with the signaling accessory molecules, Iga and Igb, 7 supporting a gain of function role for MLL in neoplastic these form the pre-BCR (see, Burrows et al for review). Survival, 19,20 transformation. Downstream targets of MLL/AF4 are just proliferation and screening of m-heavy chains have been beginning to be identified, but HOX A7 and HOX A9 have been attributed to pre-BCR signaling, contributing to the hypothesis implicated as targets for dysregulation in MLL-mediated that it functions to select pre-B cells for subsequent develop- 21–22 oncogenesis. ment. These features of B-cell development are highly con- HOX genes play a pivotal role in hematopoietic lineage served in mouse and human, and only rarely have these steps commitment and differentiation. HOX A4, HOX A5, HOX A7, been recapitulated in transformed human B-lineage cells.8–10 þ and HOX A9 are often considered to be pivitol HOX genes B-lineage ALL is a neoplastic expansion of CD19 precursors 12,14–16 required for MLL trasformation. Furthermore, enforced representing one of the several distinct developmental expression of individual HOX genes can lead to leukemogen- esis.23,24 In human, there are at least 39 HOX genes organized Correspondence: Dr F Bertrand, Department of Microbiology and into four loci located on different spanning 100 kb Immunology, Brody School of Medicine at East Carolina University, 25 Greenville, NC 27858, USA; Fax: þ 1 252 744 3104 each. Based on position and sequence similarity, HOX genes Received 7 August 2003; accepted 1 September 2003; Published can be classified into paralog groups that span HOX gene online 16 October 2003 clusters. For example, HOX A1, HOX B1 and HOX D1 form Absence of HOX gene expression in t(4;11) B-cell development FE Bertrand et al 2455 paralog group 1, and HOXA1 shares more Affymetrix (Affymetrix, Inc., Santa Clara, CA, USA) recom- with these genes than with other HOX A genes such as HOX mended protocols. Each BLIN-3-labeled RNA sample was A2.26 During development, HOX gene expression occurs in a probed with a U95A Total Named Gene Microarray (Affymetrix, temporal and spatial fashion, with 30 HOX genes (paralog Inc.). Microarrays were hybridized and scanned at the Uni- groups 1–4) being expressed earlier in anterior regions followed versity of Minnesota Biomedical Genomics Center. Data were by more 50 HOX genes in posterior regions.27 Interestingly, there analyzed using MicroArray Suite 5.0 and DataMining Tool is evidence for a similar 30–50 expression pattern in hemato- (Affymetrix, Inc.). In order to permit relative comparisons across poietic cell differentiation.23 HOX A and HOX B cluster gene multiple chips, gene expression values for each chip data set expression is found in CD34 þ subsets and HOX B genes are (n.CHP files) were scaled to the same expression value range. expressed in peripheral B and T cells.27 There is also evidence Genes with an expression P-value o0.05 were collected from for lineage-specific HOX gene expression, such as the restriction each of four BLIN-3 arrays and each of eight t(4;11) patient of HOX A10 to myeloid cells.28 The exact molecular pathways arrays21 (www.genome.wi.mit.edu/cgi-bin/cancer/datasets.cgi). regulated by HOX gene expression in hematopoietic cell growth The Mann–Whitney test was performed to determine if a given and differentiation are not fully characterized, but a variety of gene changed in expression between BLIN-3 samples and studies indicate that altered HOX expression can influence t(4;11) patient samples. A Mann–Whitney P-value o0.05 differentiation and growth factor responses (Magli et al23 and indicated the expression change. The average signal value for references therein). The expression of HOX B8 in a mouse each gene was calculated for the BLIN-3 samples and the t(4;11) 29 myeloid cell line blocks IL-6-induced differentiation. HOX B7 patient samples. For each gene, log2(fold change) was calcu- is expressed in human bone marrow upon GM-CSF stimulation, lated as log2((average BLIN-3 signal)/(average t(4;11) signal)). and reduced HOX B7 expression inhibits the formation of GM- Data were sorted by log2(fold change) and by Mann–Whitney CSF-induced colony formation.30 HOX A9À/À mice exhibit a change direction (ie up or down). This BLIN-3 data set and the reduction in peripheral lymphocytes, myeloid progenitors and unprocessed BLIN-3 array data are available at www.ecu.edu/ pre-B cells.31 These mice also have a profound reduction in microbiology/BLIN3_array.htm. GM-CSF responses. The overexpression of HOX A9 murine RT-PCR and Southern blotting was performed as described bone marrow stem cells results in stem cell expansion and a previously.9 PCR primer pairs for HOX A9, HOX A7 and Meis1 reduced in vitro IL-7 proliferative response of B-cell precur- genes have been published.34,35 sors.32 Interestingly, these mice did not exhibit lymphoid lineage transformation, and B-lineage bone marrow subcompartments were relatively normal. Luciferase assays We described a novel t(4;11) ALL cell line termed B-lineage 3 (BLIN-3), which expresses the MLL/AF4 oncoprotein and A total of 1 Â 107 cells were electroporated in a BioRad Gene exhibits a normal pro-B cell to pre-B cell developmental Pulser II (Hercules, CA, USA) at 250 V, 950 mF with 20 pmol of program, stromal cell dependency, an IL-7 response and an either HOX C8-luc, pGL-Control, or pGL-Basic and 10 mgof intact apoptotic program upon growth factor withdrawal.9 These CMV-b gal plasmids. HOX C8-luc contains the HOX C8 biological characteristics of BLIN-3 suggested that the expres- promoter fused to the luciferase gene (kind gift of Jay Hess, sion of MLL/AF4 alone does not perturb the normal develop- University of Pennsylvania16). pGL-control bears the luciferase mental program of human B-lineage cells. gene under control of the SV40 promoter and enhancer and was In order to identify genes that may cooperate with MLL/AF4 to used as a positive control. pGL-Basic contains only the modulate B-lineage leukemogenesis, we used microarray luciferase gene with no promoter or enhancer, and was used analysis to compare patterns of gene expression in BLIN-3 to as a background control. CMV-b gal has the b-galactosidase freshly isolated t(4;11) leukemic blasts.21 Interestingly, BLIN-3 (b-gal) gene driven by the CMV promoter and was used as a completely lacked the expression of HOX A9, HOX A7, HOX control for transfection efficiency. Luciferase and b-gal assays A5 and HOX A4, suggesting that the expression of these genes were performed using the Luciferase Assay System and b- may be required to abrogate B-cell development in the presence Galactosidase Assay System from Promega (Madison, WI, USA) of the MLL/AF4 oncoprotein. as per the manufacturer’s instruction. Luciferase assays were read in a Turner Systems Luminometer (Sunnyvale, CA, USA), as per the manufacturer’s instruction. Luciferase values were Materials and methods normalized to b-gal for each experiment, and are reported as fold expression over that of the pGL-Basic reporter for each cell Cells lines and tissue culture: BLIN-3 cells (t(4;11) infant- line tested. ALL) were cultured on stromal cell layers in XVIVO-10 medium supplemented with IL-7.9 RS4;11 (t(4;11) infant-ALL),13 SEM K2 (t(4;11) infant ALL),33 U937 (AML), NALM-6 (pre-B ALL) and Results BLIN-1 (pre-B ALL)8 were maintained in RPMI-1640 supple- mented with 10% heat-inactivated fetal calf serum. An original Microarray gene chip analysis cyropreservation of BLIN-3 (OR BLIN-3) that had been cultured for 2 weeks, but never passaged, was thawed rapidly in a 371C Four independently harvested cultures of m þ BLIN-3 cells were water bath, washed twice with media and then used for analyzed on Affymetrix U95A chips. These data were compared subsequent RT-PCR analysis. with published array data on eight t(4;11) patient samples.21 While such a comparison between a cell line and patient specimens may introduce unexpected variability due to Microarray gene chip analysis differences in sample selection or processing, the salient features of this analysis were consistent with known biological and Total RNA, double-stranded cDNA and labeled RNA were developmental features of BLIN-3 (see below). There were 6815 prepared from four independent cultures of m þ BLIN-3, using genes expressed in BLIN-3, 5029 of which were expressed at

Leukemia Absence of HOX gene expression in t(4;11) B-cell development FE Bertrand et al 2456 levels similar to that in t(4;11) patient samples. A total of 706 difference between Flt3 expression was observed between genes showed a two-fold or greater increase and 312 genes BLIN-3 samples and t(4;11) patient samples. showed a two-fold or greater decrease in BLIN-3 compared to t(4;11) patient samples. Of the 6815 BLIN-3 expressed genes, 768 genes were expressed in at least one of the four BLIN-3 BLIN-3 cells lack HOX A9, HOX A7, HOX A5 and HOX replicates but not in any t(4;11) sample. Of these 768, 112 genes A4 gene expression were expressed in all four BLIN-3 replicates, but not in any of the t(4;11) patient samples. MLL-fusions are associated with an increase in HOX gene 17 genes (not among the 6815 BLIN-3 expressed genes) were expression.21,22 HOX A9, HOX A7, HOX A5 and HOX A4 were expressed by all of the t(4;11) patient samples and none of the not expressed in BLIN-3, but were expressed in t(4;11) patient BLIN-3 replicates. Eight were cDNAs of unknown identity. The samples (Table 1). HOX A7 was expressed in eight of eight remaining nine were CD44 (L05424), AHNAK (M80899), t(4;11) patient samples, HOX A9 in six of eight, HOX A5 in three lipocortin (X05908), TGF-b (M38449), NF-IL6-b (M83667), of eight and HOX A4 in one of eight. HOX A1, which is not HOX A7 (AJ005814), EDN (X55988), motilin (X15393) and targeted by MLL,16,36 exhibited no difference in expression BST-1 (D21878). between BLIN-3 and t(4;11) samples. HOX C8 expression was not detected in BLIN-3 or t(4;11) samples. RT-PCR confirmed that HOX A7 and HOX A9 were not BLIN-3 cells express B-lineage-associated genes detected in BLIN-3 (Figure 1, left panel), but were readily detected in the well-characterized t(4;11) cell line RS4;11 and BLIN-3 cells exhibited increased or unique expression of several the U937 monocytoid cell line. Importantly, HOX A9 and HOX genes associated with B-lineage development, consistent with A7 were not detected in mÀ BLIN-3 cells (pro-B) or in m þ BLIN-3 our previous studies demonstrating that BLIN-3 is capable of cells (pre-B), indicating that HOX A9 and HOX A7 expression 9 undergoing pro-B to pre-B cell development. Notable was an was not being lost as a function of B-cell differentiation. eight-fold increase in 14.1 (l5) expression and a four-fold Moreover, HOX A7 and HOX A9 were not expressed in OR increase in m-heavy-chain transcripts (Table 1). The tyrosine BLIN-3, an original cryopreservation of the BLIN-3 sample kinase BTK was also increased in BLIN-3. The expression of (Figure 1, right panel). Thus, it is unlikely that HOX gene myeloid lineage-associated genes CD44, Mir-10, MNDA as expression was lost as a result of prolonged culture with IL-7 and well as TGF-b was found only in t(4;11) patient samples. No stromal cells. The OR BLIN-3 cells expressed the MLL/AF4 fusion gene (Figure 1). HOX A7 and HOX A9 were also not detected in BLIN-1, a pre-B ALL cell line lacking the t(4;11) abnormality that undergoes pre-B to immature B-cell differ- 8 Table 1 Gene expression in BLIN-3 cells as compared with t(4;11) entiation. patient samples Transformation involving HOX A9 is enhanced by coopera- tion with Meis1- and Pbx-family members.36–40 BLIN-3 samples B-lineage related exhibited unique or increased expression of Pbx2 and Pbx3, Expressed only in BLIN-3 respectively (Table 1). The expression of Meis1 was similar l gene locus DNA D87002 between BLIN-3 and t(4;11) samples, and the expression was Increased in BLIN-3 confirmed via RT-PCR (Figure 1). 14.1 Ig-related M27749 (+8.3) Rearranged mRNA for m-HC enhancer constant region X58529 (+4.4) CD24 L33930 (+4.7) BTK U78027 (+3.2) Oct-binding factor Z49194 (+3.8) Decreased in BLIN-3 IL-7R M29696 (À3.0)

Myeloid lineage related Expressed only in t(4;11) samples CD44 L05424 MIR-10 AF004231 MNDA myeloid cell nuclear differentiation antigen M81750 TGF-b M38449

Hox Expressed only in BLIN-3 PBX2 X59842 Figure 1 BLIN-3 t(4;11) cells lack HOX A7 and HOX A9 Hox7 M97676 À þ Increased expression in BLIN-3 expression. (Left panel) m and m BLIN-3 cells do not express PBX3 X59841 (+4.0) detectable Hox A7 or Hox A9, following 30 cycles of PCR Expressed only in t(4;11) samples amplification and Southern blotting. BLIN-1 is a B-lineage ALL cell HoxA4 M74297 line that does not bear the t(4;11) abnormality. U937 is a cell line of HoxA5 M26679 the monocyte lineage. RS4;11 is a well-characterized t(4;11) cell line. HoxA7 AJ005814 RT means no reverse transcriptase. (Right panel) An original HoxA9 U41813 cryopreservation of BLIN-3 (OR BLIN-3) that had been cultured for 2 weeks, but never passaged, was thawed and analyzed by RT-PCR for Lineage-associated and HOX gene expression in BLIN-3 as deter- HOX A7, HOX A9 and MLL/AF4 gene expression. The faster migrating mined from U95A Affymetrix arrays. Accession number and fold band in SEMK2 is due to a difference in the t(4;11) break point.9,33 RT change (parentheses) as compared with the expression in t(4;11) means no reverse transcriptase. Left and right panels are independent patient samples21 are indicated. experiments.

Leukemia Absence of HOX gene expression in t(4;11) B-cell development FE Bertrand et al 2457 BLIN-3 MLL/AF4 oncoprotein can activate the HOX C8 expression to previously published array data21 on eight promoter t(4;11) patient samples. Our array study revealed that four independent analyses of BLIN-3 lacked the expression of HOX MLL fusions have been shown to activate HOX C8 and HOX A9 A4, HOX A5, HOX A7 and HOX A9 (Table 1). Lack of HOX A7 promoters.16 To test if a HOX promoter would be transcription- and HOX A9 expression in BLIN-3 was verified by RT-PCR ally active in BLIN-3, we transfected m þ BLIN-3 cells with a (Figure 1). In contrast, these genes were expressed in other HOX C8 promoter luciferase (HOX C8-Luc) reporter,16 and t(4;11) cell lines (Figure 1). compared HOX C8 promoter activity in BLIN-3 to that in the MLL is a member of the trx family of genes that in Drosophila t(4;11) cell lines SEMK2 and RS4;11. In these experiments, HOX are positive regulators of HOX gene expression.12 In the mouse, C8-Luc is a surrogate for MLL/AF4 oncoprotein activity. As MLL has been shown to augment the expression of HOX A7, shown in Figure 2, the HOX C8 reporter was highly active in HOX A9 and HOX C8, and MLL is necessary to maintain HOX each of the t(4;11) cell lines including BLIN-3, but was gene expression.14,15,37 Increased expression of HOX A5, HOX minimally active in NALM-6 or U937 that do not bear the A7 and HOX A9 has been documented in MLL-associated t(4;11) abnormality. These data provide evidence that the BLIN- leukemias,21,22,41 and enforced expression of individual HOX 3 MLL/AF4 oncoprotein is functional in terms of transcriptional genes has been shown to transform primary hematopoietic stem activation of HOX promoters. cells.23,24 Although it has been suggested through microarray analysis of MLL-positive T- and B-lineage leukemic patient samples that overexpression of HOX is a key step in oncogen- 41 Discussion esis, a direct requirement for HOX gene expression in MLL- mediated transformation or an altered leukemic phenotype in We recently described the characterization of a t(4;11) cell line, MLL-positive leukemia in the absence of HOX gene expression BLIN-3, which has retained stromal cell/growth factor depen- has yet to be demonstrated. In this study, we present evidence dency and apoptotic responsiveness similar to that of normal that lack of HOX gene expression may influence the biological properties of an MLL-positive leukemia. pro-B and pre-B cells, despite the expression of the MLL/AF4 À 9 Our PCR analysis did not detect HOX A7 or HOX A9 in m oncoprotein. Moreover, we demonstrated that BLIN-3 would þ phenotypically and functionally differentiate from the pro-B cell BLIN-3 or m BLIN-3, confirming the array data and suggesting to the pre-B cell stage as accessed by: stochastic DJ to V(D)J that HOX gene expression was not being lost as a function of B- rearrangement, subsequent surface expression of the pre-BCR cell development (Figure 1). Moreover, we found that HOX A7 (m-cLC), and a proliferative response to pre-BCR crosslinking.9 and HOX A9 were not expressed in an original sample of BLIN- Importantly, both mÀ BLIN-3 (pro-B) and m þ BLIN-3 (pre-B) 3 (OR BLIN-3) that had never been passaged, although this expressed MLL/AF4 message and protein similar to that of other sample expressed the MLL/AF4 fusion gene (Figure 1). Thus, well-characterized t(4;11) cell lines that exhibit a more classical HOX gene expression was not lost in BLIN-3 as a result of t(4;11) phenotype.9 prolonged culture. Although other unidentified factors may The aforementioned data suggested that MLL/AF4 expression contribute to the unique biological properties of BLIN-3, the alone may not be sufficient to confer apoptotic resistance to absence of HOX gene expression is unusual and novel for MLL/ growth factor withdrawal and to arrest B-lineage development. AF4-expressing cells, as HOX A4, HOX A5, HOX A7 and HOX In this study, we used microarray technology to screen gene A9 upregulation is well-documented leukemic t(4;11) cells. expression in m þ BLIN-3 cells and compared that gene Importantly, we demonstrated that the BLIN-3 MLL/AF4 oncoprotein was functionally capable of transactivating the HOX C8 promoter (Figure 2) similar to other t(4;11) cell lines, such as RS4;11 and SEMK2 that exhibit arrested B-cell development, apoptotic resistance and express HOX A7 and HOX A9 (Figure 1 and 2). It is not clear why HOX genes that are normally overexpressed in the presence of truncated MLL are not expressed in BLIN-3, although it is possible that BLIN-3 may bear cytogenetically silent mutation(s) in the HOX A cluster that prevent expression of these genes. HOX A genes are not normally expressed in B-lineage committed cells,23,24 thus an alternative explanation may be that BLIN-3 has not retained enough of a monocytoid genetic program to permit the expression of the endogenous HOX gene locus. Comparison of HOX C8-Luc activity in BLIN-3 to that of the HOX A9- expressing t(4;11) cell lines RS4;11 and SEMK2 provides an indirect measure of the functional ability of the BLIN-3 MLL/AF4 fusion to activate a HOX promoter known to be upregulated in MLL-positive leukemia. BLIN-3 expresses a functional MLL/AF4 oncoprotein, and its lack of HOX gene expression provides evidence that MLL/AF4 dysregulated HOX gene expression is a necessary step in altering B-cell development in t(4;11) leukemogenesis. Figure 2 The HOX C8 promoter is activated in the presence of the Our data suggest a model (Figure 3) of leukemogenesis in BLIN-3 MLL/AF4 fusion gene. Luciferase assays were performed as which HOX A9 and/or HOX A7 gene expression enhance MLL/ described in Materials and methods. Data shown are normalized to a b-gal reporter and presented as fold expression over activity of a AF4 oncogenesis. In conventional t(4;11), MLL/AF4 leads to promoter-less luciferase vector. Data are representative of two transformation, consistent with studies demonstrating that independent experiments. truncated MLL is sufficient for transformation.19,20 The presence

Leukemia Absence of HOX gene expression in t(4;11) B-cell development FE Bertrand et al 2458 4 Bertrand FE, Billips LG, Burrows PD, Gartland GL, Kubagawa H, Schroeder Jr HW. Ig D(H) gene segment transcription and rearrangement before surface expression of the pan-B-cell marker CD19 in normal human bone marrow. Blood 1997; 90: 736–744. 5 Gathings WE, Lawton AR, Cooper MD. Immunofluorescent studies of the development of pre-B cells, B lymphocytes and immuno- globulin isotype diversity in humans. Eur J Immunol 1977; 7: 804– 810. 6 Lassoued K, Nunez CA, Billips L, Kubagawa H, Monteiro RC, LeBien TW et al. Expression of surrogate light chain receptors is restricted to a late stage in pre-B cell differentiation. Cell 1993; 73: 73–86. 7 Burrows PD, Stephan RP, Wang YH, Lassoued K, Zhang Z, Cooper MD. The transient expression of pre-B cell receptors governs B cell development. Semin Immunol 2002; 14: 343–349. 8 Wormann B, Anderson JM, Liberty JA, Gajl-Peczalska K, Brunning RD, Silberman TL et al. Establishment of a leukemic cell model for studying human pre B to B cell differentiation. J Immunol 1989; 142: 110–117. 9 Bertrand FE, Vogtenhuber C, Shah N, LeBien TW. Pro-B cell to pre- Figure 3 Model illustrating putative role for HOX A7 and HOX A9 B cell development in B-lineage acute lymphoblastic leukemia. in modulating MLL/AF4 oncogenesis during B-cell development. Blood 2001; 98: 3398–3405. 10 Wang YH, Zhang Z, Burrows PD, Kubagawa H, Bridges Jr SL, of MLL/AF4 also promotes aberrant expression of HOX A7 and Findley HW et al. V(D)J recombinatorial repertoire diversification HOX A9. These interact with Meis1 and Pbx37,40 to augment during intraclonal pro-B to B-cell differentiation. Blood 2003; 101: MLL oncogenesis, by potentially enhancing apoptotic resis- 1030–1037. 11 Greaves M. Molecular genetics, natural history and the demise of tance, growth factor independent growth and loss of develop- childhood leukaemia. Eur J Cancer 1999; 35: 173–185. mental potential. In the BLIN-3 t(4;11) cell line, truncated MLL 12 DiMarino JF, Cleary ML. MLL rearrangements in haematological leads to transformation/immortalization. However, lack of HOX malignancies: lessons from clinical and biological studies. Br J A7 and HOX A9 expression blocks Hox/Meis/Pbx-mediated Haematol 1999; 106: 614–626. enhancement of MLL oncogenesis, resulting in the retention of 13 Stong RC, Korsmeyer SJ, Parkin JL, Arthur DC, Kersey JH. Human normal B-developmental features. Interestingly, HOX A5 and acute leukemia cell line with the t(4;11) chromosomal rearrange- HOX A10 have been shown to interact with FKHR regulating ment exhibits B lineage and monocytic characteristics. Blood 42,43 1985; 65: 21–31. IGFBP-1 expression, providing a potential pathway leading 14 Yu BD, Hess JL, Horning SE, Brown GA, Korsmeyer SJ. Altered from altered HOX gene expression to loss of stromal cell/growth Hox expression and segmental identity in Mll-mutant mice. Nature factor dependency in MLL leukemia. As recent studies have 1995; 378: 505–508. indicated that mechanisms of HOX gene regulation differ 15 Yu BD, Hanson RD, Hess JL, Horning SE, Korsmeyer SJ. MLL, a between wild-type MLL and truncated MLL,16 BLIN-3 is a useful mammalian trithorax group gene functions as a transcriptional model for examining the regulation and potential requirement maintenance factor in morphogenesis. Proc Natl Acad Sci USA 1998; 95: 10632–10636. for HOX gene expression in MLL-mediated oncogenesis during 16 Milne TA, Briggs SD, Brock HW, Martin ME, Gibbs D, Allis CD human B-cell development. et al. MLL targets SET domain methyltransferase activity to Hox gene promoters. Mol Cell 2002; 10: 1107–1117. 17 So CW, Lin M, Ayton PM, Chen EH, Cleary ML. Dimerization Acknowledgements contributes to oncogenic activation of MLL chimeras in actue leukemias. Cancer Cell 2003; 4: 99–110. We thank John Kersey, Paula Croonquist, Michael Linden, Brian 18 Isnard P, Core N, Naquet P, Djabali M. Altered lymphoid Van Ness, Kevin Roberg-Perez (University of Minnesota Cancer development in mice deficient for the mAF4 proto-oncogene. Center) and Paul Doran (Affymetrix) for helpful discussions. We Blood 2000; 96: 705–710. 19 Dobson CL, Warren AJ, Pannel R, Forster A, Rabbitts TH. also thank Jay Hess (University of Pennsylvania) for kindly Tumorigenesis in mice with a fusion of the leukemia oncogene providing the Hox C8-Luc vector. This work was supported by MLL and the bacterial lacZ gene. EMBO J 2000; 19: 843–851. National Institutes of Health Grant RO1 CA31685 (TWL) and the 20 Collins EC, Rabbitts TH. The promiscuous MLL gene links Leukemia Research Fund (TWL), University of Minnesota. FEB chromosomal translocations to cellular differentiation and tumor was supported by the Chris P Tkalcevic Foundation for Leukemia tropism. Trends Mol Med 2002; 8: 436–442. Research as a Special Fellow of the Leukemia and Lymphoma 21 Armstrong SA, Staunton JE, Silverman LB, Pieters R, den Boer ML, Minden MD et al. MLL translocations specify a distinct gene Society. TWL holds the Apogee Enterprises Chair in Cancer expression profile that distinguishes a unique leukemia. Nat Genet Research. 2002; 1: 41–47. 22 Yeoh EJ, Ross ME, Shurtleff SA, Williams WK, Patel D, Mahfouz R et al. Classification, subtype discovery, and prediction of outcome in pediatric acute lymphoblastic leukemia by gene expression References profiling. Cancer Cell 2002; 1: 133–143. 23 Magli MC, Largman C, Lawrence HJ. Effects of HOX 1 LeBien TW. Fates of human B-cell precursors. Blood 2000; 96: genes in blood cell differentiation. J Cell Phys 1997; 173: 168–177. 9–23. 24 Cillo C, Cantile M, Faiella A, Boncinelli E. Homeobox genes in 2 Wolf ML, Weng WK, Stieglbauer KT, Shah N, LeBien TW. normal and malignant cells. J Cell Phys 2001; 188: 161–169. Functional effect of IL-7-enhanced CD19 expression on human B 25 Apiou F, Flagiello D, Cillo C, Malfoy B, Poupon MF, Dutrillaux B. cell precursors. J Immunol 1993; 151: 138–148. Fine mapping of human HOX gene clusters. Cytogenet Cell Genet 3 Billips LG, Nunez CA, Bertrand FE, Stankovic AK, Gartland GL, 1996; 73: 114–115. Burrows PD et al. Immunoglobulin recombinase gene activity is 26 Acampora D, D’sposito M, Faiella A, Pannese M, Migliaccio E, modulated reciprocally by interleukin 7 and CD19 in B cell Morelli F et al. The human HOX gene family. Nucleic Acids Res progenitors. J Exp Med 1995; 182: 973–982. 1989; 17: 10385–10402.

Leukemia Absence of HOX gene expression in t(4;11) B-cell development FE Bertrand et al 2459 27 Sauvageau G, Lansdorp PM, Eaves CJ, Hogge DE, Dragowska WH, 36 Hanson RD, Hess JL, Yu BD, Ernst P, van Lohuizen M, Berns A Reid DS et al. Differential expression of homeobox genes in et al. Mammalian trithorax polycomb-group homologues are functionally distinct CD34+ subpopulations of human bone antagonistic regulators of homeotic development. Proc Natl Acad marrow cells. Proc Natl Acad Sci USA 1994; 91: 12223–12227. Sci USA 1999; 96: 14372–14377. 28 Lawrence HJ, Sauvageau G, Ahmadi N, Lopez AR, LeBeau MM, 37 Kroon E, Krosl J, Thorsteinsdottir U, Baban S, Buchberg AM, Link M et al. Stage- and lineage-specific expression of the Sauvageau G. Hoxa9 transforms primary bone marrow cells HOXA10 homeobox gene in normal and leukemic hematopoietic through specific collaboration with Meis1a but not Pbx1b. EMBO cells. Exp Hematol 1995; 23: 1160–1166. J 1998; 17: 3714–3725. 29 Blatt C, Loten J, Sachs L. Inhibition of specific pathways of myeloid 38 Shen W-F, Rozenfeld S, Kwong A, Komuves LG, Lawrence HJ, cell differentiation by an activated HOX 2.4 homeobox gene. Cell Largman C. HOXA9 forms triple complexes with PBX2 and MEIS1 Growth Differ 1992; 3: 671–676. in myeloid cells. Mol Cell Biol 1999; 19: 3051–3061. 30 Lill MC, Fuller JF, Herzig R, Crooks GM, Gasson JC. The role of the 39 Thorsteinsdottir U, Kroon E, Jerome L, Blasi F, Sauvageau G. homeobox gene, HOX B7, in human myelomonocytic differentia- Defining roles for HOX and MEIS1 genes in induction of acute tion. Blood 1995; 85: 692–697. myeloid leukemia. Mol Cell Biol 2001; 21: 224–234. 31 Lawrence HJ, Helgason CD, Sauvageau G, Fong S, Izon DJ, 40 Calvo KR, Knoepfler PS, Sykes DB, Pasillas MP, Kamps MP. Humphries RK et al. Mice bearing a targeted interruption of the Meis1a suppresses differentiation by G-CSF and promotes pro- homeobox gene HOXA9 have defects in myeloid, erythroid, and liferation by SCF: potential mechanisms of cooperativity with lymphoid hematopoiesis. Blood 1997; 89: 1922–1930. Hoxa9 in myeloid leukemia. Proc Natl Acad Sci USA 2001; 98: 32 Thorsteinsdottir U, Mamo A, Kroon E, Jerome L, Bijl J, Lawrence J 13120–13125. et al. Overexpression of the myeloid leukemia-associated Hoxa9 41 Ferrando AA, Armstrong SA, Neuberg DS, Sallan SE, Silverman LB, gene in bone marrow cells induces stem cell expansion. Blood Korsmeyer SJ et al. Gene expression signatures in MLL-rearranged 2002; 99: 121–129. T-lineage and B-precursor acute leukemias: dominance of HOX 33 Pocock CF, Malone M, Booth M, Evans M, Morgan G, Greil J et al. dysregulation. Blood 2003; 102: 262–268. BCL-2 expression by leukaemic blasts in a SCID mouse model of 42 Foucher I, Volovitch M, Frain M, Kim JJ, Souberbielle JC, Gan L biphenotypic leukaemia associated with the t(4;11)(q21;q23) et al. Hoxa5 overexpression correlates with IGFBP1 upregulation translocation. Br J Haematol 1995; 90: 855–867. and postnatal dwarfism: evidence for an interaction between 34 Afonja O, Smith Jr JE, Cheng DM, Goldenberg AS, Amorosi E, Hoxa5 and Forkhead box transcription factors. Development Shimamoto T et al. MEIS1 and HOXA7 genes in human acute 2002; 129: 4065–4074. myeloid leukemia. Leukemia 2000; 24: 849–855. 43 Kim JJ, Taylor HS, Akbas GE, Foucher I, Trembleau A, Jaffe RC et al. 35 Drabkin HA, Parsy C, Ferguson K, Guilhot F, Lacotte L, Roy L et al. Regulation of insulin-like growth factor binding protein-1 promoter Quantitative HOX expression in chromosomally defined subsets of activity by FKHR and HOXA10 in primate endometrial cells. Biol acute myelogenous leukemia. Leukemia 2002; 16: 186–195. Reprod 2003; 68: 24–30.

Leukemia