BP1, a New Homeobox Gene, Is Frequently Expressed in Acute

Home , DLX gene family, Homeobox, Hox gene

Leukemia (2000) 14, 1867–1875  2000 Macmillan Publishers Ltd All rights reserved 0887-6924/00 $15.00 www.nature.com/leu BP1, a new homeobox gene, is frequently expressed in acute leukemias SB Haga1,SFu2, JE Karp3, DD Ross3,4, DM Williams5,6, WD Hankins5, F Behm7, FW Ruscetti8, M Chang9, BD Smith10, D Becton11, SC Raimondi7 and PE Berg2

1Division of Human Genetics, University of Maryland School of Medicine, Baltimore, MD; 2Department of Biochemistry and Molecular Biology, The George Washington University Medical Center, Washington, DC; 3University of Maryland Greenebaum Cancer Center, University of Maryland School of Medicine, Baltimore, MD; 4Baltimore Veterans Medical Center, Department of Veterans Affairs, Baltimore, MD; 5Department of Medicine, Johns Hopkins University, Baltimore, MD; 6ProED, Inc., Bethesda, MD; 7Department of Pathology and Laboratory Medicine, St Jude’s Children’s Research Hospital, Memphis, TN; 8Laboratory of Leukocyte Biology, NCI-FRDC, Frederick, MD; 9Statistical Ofﬁce, University of Florida, Gainesville, FL; 10Johns Hopkins Oncology Center, Johns Hopkins University, Baltimore, MD; and 11University of Arkansas for Medical Sciences, Department of Pediatrics, Little Rock, AR, USA

Aberrant expression of homeobox genes has been described shown to be leukemogenic in mice, directly implicating this in primary leukemia blasts. We recently cloned a new cDNA, class of genes in leukemia.3,4,10,11 BP1, which is a member of the homeobox gene family. BP1 expression was investigated in bone marrow samples from The aim of this study was to discover whether BP1 is acti- acute myeloid leukemia (AML), acute T cell lymphocytic leuke- vated in acute leukemias. In addition to BP1, analyses of two mia (ALL) and pre-B cell ALL. Expression levels of two appar- other cDNAs are included in these experiments, DLX7 and ent isoforms of BP1, DLX7 and DLX4, were measured in the DLX4.12–14 BP1, DLX7 and DLX4 appear to represent isoforms same samples. They are weakly if at all detectable in normal (alternative splice variants) of a common gene. Evidence for bone marrow, PHA-stimulated T cells or B cells. BP1 RNA was this includes our observations that the three homeoboxes are highly expressed in 63% of AML cases, including 81% of the identical and the cDNAs share sequence identity in other pediatric and 47% of the adult cases, and in 32% of T-ALL 15 cases, but was not found in any of the pre-B ALL cases. Co- regions. In addition, BP1 maps to chromosome 17q21, the expression of BP1, DLX7 and DLX4 occurred in a significant site of both DLX7 and DLX4.12,13 The transcriptional targets number of leukemias. Our data, including co-expression of BP1 of DLX7 and DLX4 are unknown. with c-myb and GATA-1, markers of early progenitors, suggest BP1 expression was measured in leukemia cell lines and in that BP1 expression occurs in primitive cells in AML. Analysis + − AML and ALL. AML is the most frequent adult leukemia and of CD34 and CD34 normal bone marrow cells revealed BP1 is 16,17 expressed in CD34− cells and virtually extinguished in CD34+ the second most frequent pediatric leukemia. Survival is cells. Ectopic expression of BP1 in the leukemia cell line K562 poor in AML, with only 10–35% survival in adults and 30– increased clonogenicity, consistent with a role for BP1 in leuke- 40% survival in children. On the other hand, ALL is the most mogenesis. The presence of BP1 RNA in leukemic blasts may frequent leukemia in children, but is less common in adults.18 therefore be a molecular marker for primitive cells and/or may Although aberrant expression of several molecular markers indicate that BP1 is an important upstream factor in an onco- has been associated with AML and ALL,3–5,17,19 there is clearly genic pathway. Leukemia (2000) 14, 1867–1875. Keywords: homeobox; BP1; leukemia; clonogenicity; CD34 a need to identify additional markers which could aid in diag- nosis, detection of residual disease and/or determination of appropriate therapy, as well as potentially serve as thera- peutic targets. Introduction Previously, we found that BP1 was highly expressed in erythroid and monocytic cell leukemia lines, with weak We have cloned a novel cDNA which is a member of the expression in megakaryocytic and monocytic/granulocytic Distal-less (DLX) family of homeobox (HB) genes. This cDNA, leukemia cell lines.1 Here, we extended that analysis to show called BP1, encodes a repressor of two silencers of the human that BP1 RNA was present in nine of the 14 lymphoid leuke- β-globin gene (GenBank Accession No. AF254115).1 HB mia cell lines examined. However, BP1 expression was very genes are characterized by a conserved 180 bp DNA low or undetectable in normal bone marrow, T lymphocytes sequence (the homeobox) encoding a DNA binding domain.2 or B lymphocytes, suggesting aberrant BP1 expression might The largest family of HB genes, consisting of 39 members in be occurring in the BP1-positive cell lines. humans, is called HOX. Recent studies have demonstrated Analysis of BP1 in both adult and pediatric AML was perfor- that HB genes are expressed during hematopoiesis, where at med, revealing that BP1 is present in an average of 63% of least one of them has a role in differentiation.3,4 cases, with more frequent expression in pediatric than adult Altered expression of HB genes has been implicated in AML. Expression of two other transcription factors was also acute leukemias (reviewed in Look5). In pre-B cell acute lym- assessed in each AML bone marrow sample, GATA-1 and c- phocytic leukemias (ALL), for example, the HB gene Pbx1 is myb, which are markers of early progenitors.20–24 To compare fused to E2A6 and, in acute myeloid leukemia (AML), HOXA9 the expression of BP1 in myeloid and lymphoid leukemias, or PMX1 (another HB gene) is fused to NUP98.7,8 In other expression analysis was performed in pediatric T cell ALL and cases of leukemia not involving obvious cytogenetic changes, in pediatric pre-B ALL. Interestingly, although BP1 RNA was a HOX gene which is normally silent is either activated or found in 32% of T-ALL cases, it was not detectable in patients perhaps not down-regulated in an early hematopoietic pro- 9 with pre-B ALL. To look at the distribution of BP1 in normal genitor. Ectopic expression of several HOX genes has been + − bone marrow, we analyzed expression in CD34 and CD34 progenitor cells. BP1 mRNA is enriched in CD34− cells and barely detectable in CD34+ cells. Correspondence: P Berg, George Washington University Medical The clonogenicity of BP1 was measured by testing the Center, Dept of Biochemistry and Molecular Biology, Ross Building, Room 533, 2300 Eye Street NW, Washington, DC 20037, USA; Fax: ability of K562 cells ectopically expressing cloned BP1 cDNA (202)294–8974 to grow in soft agar. The number of colonies was found to be Received 10 February 2000; accepted 20 June 2000 significantly increased in the overexpressing clones compared BP1 expression in acute leukemia SB Haga et al 1868 with controls, indicating high level BP1 expression confers an Expression was scored as negative (−), positive (+)or increased oncogenic potential. ambiguous (+/−) by normalizing against β-actin: a ratio was calculated by dividing the c.p.m. in the band representing BP1, DLX7 or DLX4 by the c.p.m. in the β-actin band from the Materials and methods same sample. One normal BM and one remission BM were included in each AML experiment; their ratios (6 repeats of Culture of leukemic cell lines each) averaged 0.01 for BP1, DLX7 and DLX4, and were the guide against which we scored a sample as −(0.0–0.10), T and B-ALL cell lines were grown in RPMI 1640 (Gibco-BRL, +/−(0.11–0.15), +(.0.15) or ++(.0.45). Two phytohemagglut- Gaithersburg, MD, USA) supplemented with 2 mM glutamine, inin (PHA)-stimulated normal T cell cultures were the control 100 U/ml penicillin, 100 µg/ml streptomycin, 0.1 mM non- for T cell ALL; their average ratios (10 repeats) were 0.05; the essential amino acids, 1 mM sodium pyruvate and 10% fetal same criteria were used for scoring BP1, DLX7 and DLX4 as bovine serum. Erythromyeloid cell lines were cultured in above. c-myb and GATA-1 are expressed in normal BM (our RPMI 1640 supplemented with 2 mM glutamine, 100 U/ml observations and Refs 20, 21). The average ratios in BM for penicillin and 100 µg/ml streptomycin, 10 mM Hepes and c-myb and GATA-1 were 0.57 and 2.1, respectively. Samples 10% fetal bovine serum. with a ratio between 0.10 and 1.14 (c-myb) or 0.10 and 4.2 (GATA-1) were scored as (+), ie within the range of normal BM, and ratios greater than 1.14 (c-myb) or 4.2 (GATA-1) Clinical samples were scored as (++). Each RT-PCR and hybridization was performed twice, independently. Normal bone marrows (BM) were obtained from the Brain Tissue Bank for Developmental Disorders at the University of Maryland. Adult AML bone marrow samples were obtained Isolation of CD34 cells from the University of Maryland Greenebaum Cancer Center and The Johns Hopkins Cancer Center Cytogenetics Core Lab- Bone marrow cells were aspirated from the posterior iliac crest oratory. Bone marrow samples from pediatric T and B-ALL of consenting healthy adult donors following guidelines patients were obtained from the Pediatric Oncology Group approved by the Institutional Review Board for NIH. Cells (POG). Informed consent was obtained from each patient or were first prepared with Histopaque-1077 density gradient guardian at their local institutions. Leukemia samples con- centrifugation. Mononuclear cells were then incubated with tained at least 85% blasts. T and B lymphocytes from periph- CD34 (QBEnd10)-conjugated magnetic microbeads eral blood of normal donors were obtained by countercurrent (AmCellCorp, Sunnyvale, CA, USA) and processed through a elutriation followed by lineage specific separation using MACS magnetic separation column (Miltenyi Biotec, Bergisch magnetic beads (Miltenyi Biotec, Auburn, CA, USA). Gladbach, Germany) to obtain purified CD34+ cells. For higher purity of CD34+ cells, a second column run was used. Purity of isolated CD34+ cells was generally greater than 90%, RT-PCR and cell viability as evaluated by propidium iodide exclusion was always higher than 95%. Purity of the remaining CD34− Mononuclear cells were isolated for RNA analysis by Ficoll– cells was greater than 95%. Paque. Total RNA was extracted from all samples using TRIzol Reagent (Gibco-BRL, Gaithersburg, MD, USA). For the reverse transcription (RT) reaction, 1 µg of DNase I-treated RNA was Construction of cell lines overexpressing BP1 added to 2.5 µM oligo-d(T) primers (PE Biosystems, Foster City, CA, USA), 10 U RNase inhibitor (Gibco-BRL), 0.5 mM A BP1 cDNA fragment of 1013 bp containing the complete dNTPs, and 100 U MMLV-reverse transcriptase (Promega, open reading frame was amplified by RT-PCR from K562 cell Madison, WI, USA). The RT reaction was performed in a ther- RNA using primers designed from the original cDNA. This mal cycler at 25°C for 10 min, 4 min ramp, 42°C for 50 min, fragment was cloned into pGEM7, sequenced, then sub- and 99°C for 5 min. One to 2 µl of this reaction was used in cloned into pRC/RSV (Invitrogen, Carlsbad, CA, USA) using the PCR reaction. Primers specific for BP1, DLX4 and DLX7 HindIII and XbaI. were designed, and the PCR products were verified by restric- tion enzyme analysis. Linearity assays for each primer set have been performed and cycling conditions adjusted accordingly Statistical analysis (data not shown). Primer sequences and PCR cycling conditions are shown in Table 1. The significance of correlation between expression of BP1, DLX4, and DLX7 was assessed by Fisher’s exact test. All P values reported were two-sided. Semi-quantitative RT-PCR

An oligonucleotide specific for each product was end-labelled Results with γ-32P-dATP and added to 10% of the PCR product (2.5 µl). Hybridization was in the thermal cycler (94°C for 15 s; Expression of BP1 in normal cells 42°C for 1 min). The hybridized product was electrophoresed on a 5% polyacrylamide gel, the gel was dried, and then RNA levels for BP1 were determined in four normal bone mar- exposed to film. The autoradiograph was aligned with the gel row samples and one bone marrow from an AML patient in and bands were excised and quantitated by scintillation remission, and in five PHA-stimulated T cell and five B cell counting. preparations from normal bone marrow. Results obtained after

Leukemia BP1 expression in acute leukemia SB Haga et al 1869 Table 1 PCR primers

Primers Sequence PCR conditions Product Ref.

BP1 Upper: 59CACCTCCTGTCTTACCCCTACACC39 94°C/1 min; 62°C/1 min; 581 bp Chase et al1 Lower: 59GCCCTTCCCCAGATTCACATCATC39 72°C/1.5 min for 30 cycles DLX7 Upper: 59CCTACACCGTGTTGTGCTGC39 94°C/1 min; 60°C/1 min; 406 bp This paper Lower: 59CTGTTGCCATAGCCACTG39 72°C/1.5 min for 30 cycles DLX4 Upper: 59CACGGTGTGGCGGGGGAGACAT39 94°C/1 min; 60°C/1 min; 350 bp This paper Lower: 59CTGCGGTGGGAGGTCGGAGTTC39 72°C/1.5 min for 30 cycles β-actin Upper: 59GGATCTTCATGAGGTAGTCAGTC39 94°C/1 min; 60°C/1 min; 626 bp Raff et al25 Lower: 59CCTCGCCTTTGCCGATCC39 72°C/1.5 min for 20 cycles c-myb Upper: 59ATTAGGTAATGAATTGTAGCCAG39 94°C/30 s; 55°C/30 s; 228 bp Majello et al26 Lower: 59ACTTAGAGTAATGCTTTTACTGA39 72°C/1 min for 28 cycles Shimamoto et al27 GATA-1 Upper: 59CCATTGCTCAACTGTATGGAGGG39 94°C/30 s; 58°C/30 s; 249 bp Tsai et al28 Lower: 59ACTATTGGGGACAGGGAGTGATG39 72°C/1 min for 28 cycles Shimamoto et al27 SCL Upper: 59CAATCGAGTGAAGAGGAGACCTCC39 94°C/45 s; 55°C/45 s; 144 bp Chen et al29 Lower: 59TTGCGGAGCTCGGCAAAGGC39 72°C/1.5 min for 30 cycles

semi-quantitative RT-PCR analysis of six representative phoma cell lines (Table 2). DLX7 and DLX4 RNA levels were samples are shown in Figure 1. BP1 RNA was barely visible also determined in the T cell ALL cell lines (Figure 2), with in any of these control samples; results were similar for DLX4 co-expression of the three isoforms in the same cell lines. BP1 and DLX7. After a longer exposure of the autoradiograph, a mRNA was also readily detectable in four of six erythromye- faint band was seen in all six samples for several of the loid cell lines (K562, HEL, THP-1 and U937), with less isoforms (data not shown). expression in MEG-01 and HL60 cells.1 Analysis of DLX7 and DLX4 in those cell lines showed that BP1, DLX7 and DLX4 were frequently co-expressed, with greatest expression of all Expression of BP1 in cell lines three in K562 and U937 cells (data not shown). The observation that there was little or no expression of any of the iso- Expression of BP1 was examined by RT-PCR in a number of forms in normal bone marrow (described above), compared leukemia cell lines. Fourteen cell lines of lymphoid origin with the expression found in diverse myeloid and lymphoid were analyzed for BP1 expression. BP1 RNA was present in leukemia cell lines, led us to examine the relative expression four of eight T cell ALL cell lines and in three of four lym- of BP1, DLX7 and DLX4 in the bone marrow of acute leukemia patients using semi-quantitative RT-PCR.

Expression of BP1 in acute myeloid leukemias

Expression of BP1 was examined in both pediatric and adult AML patients. A total of 39 AML patients were studied, of whom 18 were under the age of 18 (pediatric; cases 8–27) and 21 were 18 years of age or older (adult; cases 29–49). Table 3 summarizes the analysis of the AML samples and the clinical data available for each patient, including cytogenetics, expression of surface markers and initial response to treatment; outcome data were not available for all patients and thus were excluded. Data are grouped according to the French–American–British (FAB) criteria.30 Assignment of expression levels was made as described in Materials and methods. For this analysis, normal bone marrow was used as a negative baseline against which the clinical samples were compared. Samples classiﬁed as +/− were excluded from statistical analysis. BP1 was overexpressed in a deﬁnitive manner in 81% Figure 1 Expression of BP1, DLX7 and DLX4 in normal bone mar- (13/16) of the pediatric bone marrow samples and in 47% row (BM), B lymphocytes (B) and T lymphocytes (T). Semi-quantitative (9/19) of the adult cases. DLX7 was overexpressed in 59% RT-PCR was used to measure expression, as described in Materials and methods. One normal BM (lane 1), one remission BM (lane 2), (10/17) of pediatric and 38% (6/16) of adult cases, while DLX4 two normal B lymphocyte samples (labeled B) and two PHA-stimu- was overexpressed in 79% (11/14) of pediatric and 79% lated T lymphocyte samples (labeled T) are shown. (15/19) of adult cases. An example of RT-PCR analysis show-

Leukemia BP1 expression in acute leukemia SB Haga et al 1870 Table 2 Expression of BP1 in hematopoietic cell lines ing expression in selected samples is seen in Figure 3. Here, patient 15 showed the greatest BP1 and DLX7 levels, with less BP1 expression DLX4 expression. The correlation between BP1 and DLX7 (P = 0.0002), BP1 and DLX4 (P = 0.0016), and DLX7 and DLX4 Lymphoid (P = 0.023) were statistically significant by Fisher’s exact test. B cell lineage Analysis of GATA-1 and c-myb, markers of primitive cells, Normal +/− ALL was performed on the AML samples to indicate the differen- REH + tiation state of cells expressing BP1. GATA-1 is believed to be RS4;11 + involved in positive regulation of myeloid development and Burkitt’s lymphoma thought to be expressed in early progenitors, then downregul- Raji − ated in myeloid (but not erythroid) differentiation, making its + Daudi expression in the myeloid pathway a sign of immaturity.22 It T cell lineage is detected in acute leukemias characterized by expansion of Normal +/− early progenitors.21 c-myb is expressed in immature hemato- ALL poietic cells, and its expression decreases as cells differen- Jurkat + 20,23,24 + tiate. Since c-myb and GATA-1 are expressed in normal MOLT-3 20,21 MOLT-4 + bone marrow (data not shown), a classification of positive MOLT-13 + was given to samples expressing at the level of normal cells CCRF-CE − (see Materials and methods for details). We found that every HSB2 − AML sample exhibited c-myb expression at least equal to that MOLT-16 − in normal bone marrow, and 42% (13/31) exhibited greater − RPMI 840 expression. A higher proportion of pediatric than adult cases Lymphoma HUT78 + showed high-level c-myb expression, 59% in children com- Sup-T1 + pared with 21% in adults. GATA-1 was present in 77% (24/31) of AML cases, with 72% expression in children and Myeloid 85% in adults. Erythroid Co-expression of DLX7 and DLX4 with BP1 was striking: + K562 84% of BP1+ samples were DLX7+ and 100% were DLX4+. HEL + Monocytic GATA-1 was co-expressed in 74% of the samples, and all THP-1 + were c-myb positive, with 45% exhibiting high levels of c- U937 + myb. Another parameter in this study was analysis of surface Megakaryocytic markers. Interestingly, 64% of the BP1-positive samples were MEG-01 +/− CD34 negative, while 73% were CD13 positive and 100% Monocytic/granulocytic were CD33 positive. HL60 +/− Among the 39 AML patients, two had an abnormal chromosome 17q arm, the locus of BP1.15 Patient 41, with acute pro- myelocytic leukemia, exhibited a t(15;17) translocation pre- sumably involving the retinoic acid receptor on chromosome 17.5 Patient 8 exhibited a t(11;17)(q23;q25) translocation. This translocation has been identified as fusing the MLL gene on chromosome 11q23 to either the AF17q25 gene or to MSF, both located on 17q25 and differing by a single base.31–33 Since the MLL gene is known to activate several HOX genes,34 it is tempting to speculate that the fusion protein may activate BP1 in this case.

Expression of BP1 in acute lymphoid leukemias

For comparison with the myeloid lineage leukemias, we examined 19 cases of pediatric T cell ALL (Table 4). Here, 32% (6/19) were BP1 positive. Outcome data were not available for these patients. Analysis of DLX7 and DLX4 showed that 40% (6/15) of the cases were DLX7 positive, and 40% (6/15) were DLX4 positive. A comparison of expression of the three isoforms in AML and T cell ALL is seen in Figure 4. We next analyzed 19 pediatric patients with pre B-ALL. No detectable BP1 was observed in any of these cases, although the β-actin controls were normally expressed (data not shown). Figure 2 Expression of BP1, DLX7 and DLX4 in T cell lineage lymphoid cell lines. Samples were analyzed using RT-PCR. BP1 expression in CD34+ and CD34− cells

To examine more precisely the reason for very low BP1 expression in normal bone marrow and to determine whether

Leukemia BP1 expression in acute leukemia SB Haga et al 1871 Table 3 Expression of BP1, DLX7 and DLX4 and clinico-biological features in patients with AML

FAB Patient BP1 DLX7 DLX4 c-myb GATA-1 CD34 CD33 CD13 Karotype Response

Normal BM −−−++ M0 12a ++ + +/−+++−++ND PR 37 +/−−+++++++46,XY,t(1:15),trisomy 1q NR 47 −−−ND ND +++47,XY,+13 NR M1 15a +++++++++++46,XY,t(11;9)(q23;p13.1) CR 17a +/−− +/−++++++46,Y,t(X;7)(q13;p15)dup CR (12)(p11p13) 19a −−−++++++46,XX CR 22a ++ ++ ++ +++ + + + + 47,XX,+21 ND 26a +++/−+++++46,XX CR 27a −−++++++46,XY CR 29 −−+++++++46,XY,del(5)(q31q34) NR 40 ++ ++ ++ ++ +/−− + + 46,XY,t(1;12) CTC 46 +++ND ND −++46,XY,del(1)(q32),del(13) CR (q12–22) M2 18a −−−+/−++++47,XX,+8CR 20a +−++−−++46,XY CR 44 − ND +/− ND ND +++46,XY CR 45 −−−ND ND ++−45,X,-Y CR (,6 months) 48 ++/− ND ND ND +++47,XY,+8NR M3 41 ++++++−++46, XX, t(15;17)(q21;q12-21) CR M4 23a +/−−−+−+++46,XX,inv(16)(p13q22) CR 31 −−−++−++45,XY,−7NR 32 −−−++−++45,XX,−16 NR 35 ++/−+++−++46,XX NR 49 −−+ND ND +++46,XY,inv(16)(p13q22) CR M5 8a +++++−−+−46,XX,t(11;17)(q23;q25) CR 9a ++/−+/−+−−+−ND CR 11a +++++−−++46,XY NR 13a ++++++−+−46,XY,der(1)t(1;6) CR (q32;p21.1),add(11)(q23),der(22) t(1;22)(q23;q11.2) 14a +++++−++46,XX CR 16a +−+++−++46,XX NA 30 −+/−++−−++47,XY,+3NA 33 −−++++++46,XY NR 34 +++++++−48,XX,add(4)(p15.1),del(5) NR (q21q33),t(7;12)(p10;q10), add(11)(q23),+i(12)(p10), add(16)(q11.2),−17,+18,−20, +der(21)t(17;21)(q10;q10),+mar 36 +−+++++−46,XY,del(7)(q22), NR +8(q24.3) 38 +/−−++++++46,XX CR 39 −+/−++++−++48,XY,t(1;16),+7,+19 CTC 42 ++++−13% + 62% 47,XY,t(6;9)(q23;q34),+8RD 43 +++ND ND −++46,XY,t(3;5)(q25;q34) NR M7 21a + ++ ++ +++ +++ + + − 46,XY CR 24a +++++++++45,XY,t(3;3)(q21;q26),−7NR aAge less than 18. CR, complete response; NR, no response; PR, partial response; CTC, complete tumor clearance; RD, residual disease; ND, not determined; NA, data not available.

BP1 is expressed in early progenitors, BP1 expression was Clonogenicity of K562 cells overexpressing BP1 measured in CD34+ and CD34− cells. There was clear expression in two independent isolates of CD34− cells (Figure Since the function of BP1 is not known in the context of leuke- 5, lanes 2 and 3). In contrast, there was barely detectable BP1 mia, we tested the effect of enforced BP1 expression on the mRNA in three independent samples of CD34+ bone marrow oncogenic potential of the erythroleukemia cell line K562. (lanes 4, 5 and 6). Expression in K562 cells is shown for com- Four K562 cell lines stably overexpressing BP1 were con- parison in lane 1, and β-actin was measured as a loading structed, as well as controls containing an empty vector (see control. Materials and methods). The BP1 mRNA levels in the stable

Leukemia BP1 expression in acute leukemia SB Haga et al 1872 8a, 8c and 8d, exhibited signiﬁcantly increased numbers of colonies able to grow in 0.5% soft agar per 15000 cells plated compared with controls; there was a 45-fold increase in the number of colonies for 8c compared with 7a or 7b. These results indicate BP1 overexpression is associated with increased clonogenicity in K562 cells, predictive of an increase in oncogenicity.

Discussion

We have detected significant RNA expression of BP1, a new homeobox gene, in the bone marrow of 81% of pediatric AML patients, compared with 47% of adult patients. In contrast, expression of BP1 was not reproducibly found in normal bone marrow. The highest percentage of BP1 positives occurred in the FAB classification M5 (monocytic), in which 77% of AML cases were BP1 positive; bone marrow cells from 100% of the children in this category were BP1 positive. Two splice variants, DLX7 and DLX4, were co-expressed in 48% and 79% of AML patients, respectively. BP1, DLX7 and DLX4 levels were also assessed in 19 cases of pediatric T cell ALL. Although the frequency of expression was less in comparison with AML, BP1 was overexpressed in 32%, DLX7 in 50% and DLX4 in 40% of T cell ALLs, compared with weak or no expression in normal PHA-stimulated T lymphocytes. In sharp contrast, no BP1 expression was detected in pre-B ALL. Figure 3 Expression of BP1, DLX7 and DLX4 in acute myeloid leu- − Although the reason for expression of BP1 in T cell ALL but kemia. Samples from six patients are shown. ( ) indicates expression not in pre B cell ALL is not obvious, this difference may equal to or less than normal controls; (+) indicates expression greater than controls; (+) indicates an expression ratio (defined in Materials provide a useful diagnostic distinction. and methods) at least three times greater than that for (+). Patients 8, c-myb was expressed in all AML samples, either at a level 11, 15, and 19 are less than 18 years of age, while patients 27 and comparable to normal bone marrow or at a higher level. 34 are over 18 years old. Expression of c-myb is associated with immaturity,20,23,24 so those cases within the normal range may be arrested at an early progenitor stage. We speculate that higher c-myb Table 4 Expression of BP1, DLX7 and DLX4 in children with T expression may be part of the leukemogenic process since cell ALL activation of c-myb causes leukemia in mice.35 In this regard, of 11 samples which highly expressed c-myb and could be Patient BP1 DLX7 DLX4 evaluated for BP1 expression, nine were BP1 positive. There is a substantial body of data on expression of HOX 1 −−− 2 −+−genes in malignant hematopoietic cell lines. The HOX genes 3 −+−are clustered on four chromosomes, and the DLX genes are 4 +−−located in pairs on the same chromosomes.3,4,12,36 BP1 is situ- 5 +++ated at the 39 end of the HOXB cluster on chromosome 17.15 6 −−−HOXB genes are preferentially expressed in erythroid cells, −++ 7 including K562 and HEL cell lines.3,37–39 It is believed that 8 −−− 9 −−−genes in the HOX clusters are switched off or on in blocks in 38,40 10 +++myeloid cells. Our data suggest that BP1 may be part of 11 − ND − this coordinate regulation since its pattern of expression in 12 +−−erythromyeloid cell lines is similar to that of the adjacent 13 −−+HOXB genes. Transcripts of HOX genes have also been found 14 +++ − in AML, T-ALL and pre B-ALL but, unlike BP1, they are readily 15 ND ND 9,41–44 16 +++detectable in normal bone marrow. Notably, the 17 − ND ND expression of HOX genes is downregulated during normal 18 − ND ND hematopoietic differentiation but not in AML.43 19 − ND ND BP1 RNA expression in acute leukemias may represent a marker for the differentiation stage of the leukemic blasts ND, not determined. and/or may be directly involved in leukemogenesis. Our data point to the possibility that BP1 expression in AML occurs in early progenitors: (1) As described above, all of the BP1-posi- cell lines (8a, 8c, 8d and 8e) relative to the empty vector con- tive cells are also c-myb positive and 74% are GATA-1 posi- trols (7a and 7b) varied from five- to 21-fold, determined by tive, two indicators of early progenitors.20,22–24 (2) The barely RT-PCR (Table 5). Stable cell lines were tested for their clon- detectable expression of BP1 seen in normal bone marrow is ogenicity (ability to grow in soft agar), an indicator of oncog- compatible with expression in primitive cells, which comprise enic potential. Three of the four cell lines overexpressing BP1, a very small sub-population of normal bone marrow. (3) Over-

Leukemia BP1 expression in acute leukemia SB Haga et al 1873

Figure 4 Comparison of expresssion of BP1, DLX7 and DLX4 in AML and T cell ALL. Black bars represent adult AML, gray bars pediatric AML, and white bars pediatric T cell ALL.

CD34+ stem cells express several HOX genes, and this expression is down-regulated in CD34− cells.4,48 In contrast, BP1 is expressed in CD34− cells and down-regulated in CD34+ cells. These results are in agreement with the data in AML samples (point 5), in which BP1 was primarily found in CD34− cells. The very low expression in CD34+ cells could represent either expression in a few CD34+ cells or contami- nation of the CD34+ cells with CD34− cells. Recent papers support the existence of a sub-population of primitive CD34− lin− stem cells with repopulating ability in both mice and humans.49–51 In mice, CD34− lin− stem cells can convert to CD34+ stem cells upon activation; this has not been investigated in humans.50,51 Since our CD34− cells contain both lin+ and lin− subpopulations, we do not know whether BP1 is expressed in stem cells. However, it is clear that BP1 is acti- + − Figure 5 Expression of BP1 in CD34 and CD34 cells. Cells were vated early in hematopoesis. We hypothesize that BP1 sorted as described (Materials and methods), and BP1 mRNA was expression is then repressed during differentiation. This idea measured by RT-PCR. is strengthened by our observation that BP1 is down-regulated during erythroid differentiation of the cell line MB-02.1 Table 5 Clonogenicity of K562 cells overexpressing BP1 In support of a possible oncogenic role for BP1, we observed that stable cell lines overexpressing BP1 exhibited Cell line BP1 expression Colonies up to a 45-fold increase in clonogenicity. Moreover, its high frequency of expression in AML may indicate BP1 is an 7a (control) 1 23 ± 15 upstream factor in an oncogenic pathway. Further experi- 7b (control) 1 21 ± 7 ments are needed to delineate the roles of BP1 in normal hem- 8a 21 517 ± 108 atopoiesis, to directly determine if it plays a role in neoplastic 8c 12 924 ± 199 transformation, and to examine the clinical signiﬁcance of its ± 8d 5 223 34 expression in acute leukemias. 8e 7 57 ± 18

Acknowledgements expression of BP1 in both myeloid and lymphoid leukemia We thank Dr Valerie Hu for helpful comments on the manu- argues that leukemogenesis may occur in a stem cell or multi- script. This work was supported in part by grants from the Elsa potent hematopoietic progenitor. (4) Further supporting this U Pardee Foundation (PEB), National Institutes of Health idea is the observation that 59% of BP1 positive blasts are Grant R01 DK53533 (PEB), NIH CA-21765 (FB and SCR), and found in FAB classes considered to be primitive and associa- by the American Lebanese Syrian Associated Charities ted with stem cell leukemias, ie M0 (minimally differentiated), (ALSAC) (FB and SCR). Procurement and processing of some M5 (monocytic) or M7 (megakaryocytic).45–47 (5) 64% of BP1- leukemia samples was supported by National Institutes of positive cases are CD34 negative. Health Grant 5P30CA06973 (BDS).

Leukemia BP1 expression in acute leukemia SB Haga et al 1874 References erythroid transcriptional factor in GM- and G-CSF-dependent myeloid cell lines. Nucleic Acids Res 1990; 18: 6863–6869. 23 Gonda T, Metcalf D. Expression of myb, myc and fos proto-onco- 1 Chase MB, Haga S, Fu S, Davenport G, Morgan D, Mah A, Berg genes during the differentiation of a murine myeloid leukemia. PE. BP1, a novel homeobox gene with repressor activity, down- Nature 1984; 310: 249–251. regulates erythroid differentiation (submitted). 24 Luscher B, Eisenman RN. New light on Myc and Myb. Part II. 2 Levine M, Hoey T. Homeobox proteins as sequence-specific tran- Myb. Genes Dev 1990; 4: 2235–2241. scription factors. Cell 1988; 55: 537–540. 25 Raff T, van der Giet M, Endemann D, Wiederholt T, Paul M. 3 Lawrence HJ, Sauvageau G, Humphries RK, Largman C. The role Design and testing of β-actin primers for RT-PCR that do not co- of HOX homeobox genes in normal and leukemic hematopoiesis. amplify processed pseudogenes. BioTechniques 1997; 23: 456– Stem Cells 1996; 14: 281–291. 460. 4 van Oostveen JW, Biji JJ, Raaphorst FM, Walbooners JJM, Meijer 26 Majello B, Kenyon LC, Dalla-Favera R. Human c-myb protoonco- CJLM. The role of homeobox genes in normal hematopoiesis and gene: nucleotide sequence of cDNA and organization of the gen- hematological malignancies. Leukemia 1999; 13: 1675–1690. omic locus. Proc Natl Acad Sci USA 1986; 83: 9636–9640. 5 Look AT. Oncogenic transcription factors in the human acute leu- 27 Shimamoto T, Nakamura S, Bollekens J, Ruddle FH, Takeshita K. kemias. Science 1997; 278: 1059–1064. Inhibition of Dlx-7 homeobox gene causes decreased expression 6 Lu Q, Wright DD, Kamps MP. Fusion with E2A converts the Pbx1 of GATA-1 and c-myc genes and apoptosis. Proc Natl Acad Sci homeodomain protein into a constitutive transcriptional activator USA 1997; 94: 3245–3249. in human leukemias carrying the t(1;19) translocation. Mol Cell 28 Tsai SF, Martin DIK, Zon LI, D’Andrea AD, Wong GG, Orkin SH. Biol 1994; 14: 3938–3948. Cloning of the cDNA for the major DNA-binding protein of the 7 Borrow J, Shearman AM, Stanton VP Jr, Becher R, Collins T, Willi- erythroid lineage through expression in mammalian cells. Nature ams AJ, Dube I, Katz F, Kwong YL, Morris C, Ohyashiki K, Toyama 1989; 339: 446–451. K, Rowley J, Housman DE. The t(7;11)(p15;p15) translocation in 29 Chen Q, Cheng J-T, Tsai L-H, Schneider N, Buchanan G, Carroll acute myeloid leukaemia fuses the genes for nucleoporin NUP98 A, Crist W, Ozanne B, Siciliano MJ, Baer R. The tal gene and class I homeoprotein HOXA9. Nat Genet 1996; 12: 159–167. undergoes chromosome translocation in T cell leukemia and 8 Nakamura T, Yamazaki Y, Hatano Y, Miura I. NUP98 is fused to potentially encodes a helix–loop–helix protein. EMBO J 1990; 9: PMX1 homeobox gene in human acute myelogenous leukemia 415–424. with chromosome translocation t(1;11)(q23;p15). Blood 1999; 94: 30 Bennett JM, Catovsky D, Daniel MT, Flandrin G, Galton DAG, 741–747. Gralnick HR, Sultan C. Proposed revised criteria for the classi- 9 Petrini M, Quaranta MT, Testa U, Samoggia P, Tritarelli E, Care A, fication of the French–American–British Cooperative Group. Ann Cianetti L, Valtieri M, Barletta C, Peschle C. Expression of selected Intern Med 1985; 103: 620–625. human HOX genes in B/T acute lymphoid leukemia and interleu- 31 Baer MR, Stewart CC, Lawrence D, Arthur DC, Mrozek K, Strout kin-2/interleukin-1 b-stimulated natural killer lymphocytes. Blood MP, Davey FR, Schiffer CA, Bloomfield CD. Acute myeloid leuke- 1992; 80: 185–193. mia with 11q23 translocations: myelomonocytic immunopheno- 10 Hawley RG, Fong AZC, Reis MD Zhang N, Lu M, Hawley TS. type by multiparameter flow cytometry. Leukemia 1998; 12: Transforming function of the HOX11/TCL3 homeobox gene. Can- 317–325. cer Res 1997; 57: 337–345. 11 Thorsteinsdottir U, Sauvageau G, Hough MR, Dragowska W, 32 Taki T, Ohnishi H, Shinohara K, Sako M, Bessho F, Yanagisawa Lansdorp PM, Lawrence HJ, Largman C, Humphries RK. Over- M, Hayashi Y. AF17q25, a putative septin family gene, fuses the expression of HOXA10 in murine hematopoietic cells perturbs MLL gene in acute myeloid leukemia with t(11;17)(q23;q25). Can- both myeloid and lymphoid diffferentiation and leads to acute cer Res 1999; 59: 4261–4265. myeloid leukemia. Mol Cell Biol 1997; 17: 495–505. 33 Osaka M, Rowley JD, Zeleznik-Le NJ. MSF (MLL septin-like 12 Nakamura S, Stock DW, Wydner KL, Bollekens JA, Takeshita K, fusion), a fusion partner gene of MLL, in a therapy-related acute Nagai BM, Chiba , Kitamura T, Freeland TM, Zhao Z, Minowada myeloid leukemia with a t(11;17)(q23;q25). Proc Natl Acad Sci J, Lawrence JB, Weiss KB, Ruddle FH. Genomic analysis of a new USA 1999; 96: 6428–6433. mammalian Distal-less gene: Dlx-7. Genomics 1996; 38: 314– 34 Yu BD, Hess JL, Horning SE, Brown GAJ, Korsmeyer S. Altered 324. Hox expression and segmental identity in Mll-mutant mice. Nat- 13 Quinn LM, Johnson BV, Nicholl J, Sutherland GR, Kalionis B. Iso- ure 1995; 378: 505–508. lation and identification of homeobox genes from human placenta 35 Wolff L, Koller R, Bies J, Nazarov V, Hoffman B, Amanullah A, including a novel member of the Distal-less family, DLX4. Gene Krall M, Mock B. Retroviral insertional mutagenesis in murine pro- 1997; 187: 55–61. monocytic leukemias: c-myb and Mml1. Curr Topics Micro 14 Price JA, Bowden DW, Wright JT, Pettenati MJ, Hart TC. Identifi- Immuno 1996; 211: 191–199. cation of a mutation in DLX3 associated with tricho-dento-osseous 36 Simeone A, Acampora D, Pannese M, D’Esposito M, Stornaiuolo (TDO). Hum Mol Gen 1998; 7: 563–569. A, Gulisano M, Mallamaci A, Kastury K, Druck T, Huebner K. 15 Fu S, Strovel JW, Haga SB, Stamberg J, Berg PE. Mapping of a new Cloning and characterization of two members of the vertebrate homeobox gene, BP1, near its isoform DLX7 and characterization Dlx gene family. Proc Natl Acad Sci USA 1994; 91: 2250–2254. of their roles in repression of the beta globin gene. Am J Hum 37 Shen W-F, Largman C, Lowney P, Corral JC, Detmer K, Hauser Gen 1998; 63: A181. CA, Simonitch TA, Hack FM, Lawrence HJ. Lineage-restricted 16 Pui C-H. Childhood leukemias. New Engl J Med 1995; 332: expression of homeobox-containing genes in human hematopo- 1618–1630. etic cell lines. Proc Natl Acad Sci USA 1989; 86: 8536–8540. 17 Karp JE. Acute leukemia: mechanisms of cell survival as targets 38 Magli CM, Barba, P, Celetti A, De Vita G, Cillo, C, Boncinelli E. for therapy. Int J Oncol 1997; 11: 657–674. Coordinate regulation of HOX genes in human hematopoietic 18 Copelan EA, McGuire EA. The biology and treatment of acute lym- cells. Proc Natl Acad Sci USA 1991; 88: 6348–6352. phoblastic leukemia in adults. Blood 1995; 85: 1151–1168. 39 Mathews CHE, Detmer K, Boncinelli E, Lawrence HJ, Largman C. 19 Tenen DG, Hromas R, Licht JD, Zhang D-E. Transcription factors, Erythroid-restricted expression of homeobox genes of the human normal myeloid development and leukemia. Blood 1997; 90: HOX2 locus. Blood 1991; 78: 2248–2252. 489–491. 40 Celetti A, Barba P, Cillo C, Rotoli B, Boncinelli E, Magli MC. 20 Gewirtz AM, Calabretta B. A c-myb antisense oligodeoxynucleo- Characteristic patterns of HOX gene expression in different types tide inhibits normal human hematopoiesis in vitro. Science 1988; of human leukemia. Int J Cancer 1993; 53: 237–244. 242: 1303–1306. 41 Lawrence HJ, Sauvageau G, Ahmadi N, Lopez AR, LeBeau MM, 21 Guerrasio A, Saglio G, Rosso C, Alfarano A, Camaschella C, Lo Link M, Humphries K, Largman C. Stage- and lineage-specific Coco F, Biondi A, Ranbaldi A, Nicolis S, Ottolenghi S. Expression expression of the HOXA10 homeobox gene in normal and leu- of GATA-1 mRNA in human myeloid leukemic cells. Leukemia kemic hematopoietic cells. Exp Hematol 1995; 23: 1160–1166. 1994; 6: 1034–1038. 42 Biji JJ, van Oostveen JW, Walboomers JMM, Brink ATP, Vos W, 22 Crotta S, Nicolis S, Ronchi A, Ottolenghi S, Ruzzi L, Shimada Y, Ossenkoppele GJ, Meijer CJLM. Differentiation and cell-type- Migliaccio AR. Progressive inactivation of the expression of an restricted expression of HOXC4, HOXC5 and HOXC6 in myeloid

Leukemia BP1 expression in acute leukemia SB Haga et al 1875 leukemias and normal myeloid cells. Leukemia 1998; 12: R, Bruno A, Aronica G, Maffei L, Suppo G, Simone MD, Forte L, 1724–1732. Cordero V, Postorino M, Tuﬁlli V, Isacchi G, Masi M, Papa G, 43 Kawagoe H, Humphries RK, Blair A, Sutherland HJ, Hogge DE. Amadori S. Minimally differentiated acute myeloid leukemia Expression of HOX genes, HOX cofactors, and MLL in pheno- (AML-M0): comparison of 25 cases with other French–American– typically and functionally deﬁned subpopulations of leukemic and British subtypes. Blood 1997; 89: 621–629. normal human hematopoietic cells. Leukemia 1999; 13: 687–698. 48 Sauvageau G, Lansdorp PM, Eaves CJ, Hogge DE, Dragowska WH, 44 Salvati PD, Ranford PR, Ford J, Kees UR. HOX11 expression in Reid DS, Largman C, Lawreence J, Humphries RK. Differential pediatric acute lymphoblastic leukemia is associated with T cell expression of homeobox genes in functionally distinct CD34+ sub- phenotype. Oncogene 1995; 11: 1333–1338. populations of human bone marrow cells. Proc Natl Acad Sci USA 45 Cuneo A, Mecucci C, Kerim S, Vandenberghe E, Dal Cin P, Van 1994; 91: 12223–12227. Orshoven A, Rodhain J, Bosly A, Michaux JL, Martiat P, Boogaerts 49 Bhatia M, Bonnet D, Murdoch B, Gan OI, Dick JE. A newly disco- M, Carli MG, Castoldi G, Van Den Berghe H. Multipotent stem vered class of human hematopoietic cells with SCID-repopulating cell involvement in megakaryoblastic leukemia: cytologic and activity. Nature Med 1998; 4: 1038–1045. cytogenetic evidence in 15 patients. Blood 1989; 74: 1781–1790. 50 Goodell MA. CD34+ or CD34−: does it really matter? Blood 1999; 46 Bonnet D, Dick JE. Human acute leukemia is organized as a hier- 94: 2545–2547. archy that originates from a primitive hematopoietic cell. Nature 51 Sato T, Laver JH, Ogawa M. Reversible expression of CD34 by Med 1997; 3: 730–737. murine hematopoietic stem cells. Blood 1999; 94: 2548–2554. 47 Venditti A, Del Poeta G, Buccisano F, Tamburini A, Cox MC, Stasi

Leukemia