Role of an Inverted CCAAT Element in Human Topoisomerase IIα Gene Expression in ICRF-187-Sensitive and -Resistant CEM Leukemic Cells

Materials and Methods

Cell Lines.

Human leukemic CEM cells and the ICRF-187-resistant subline CEM/ICRF-8 were cultured as described (Morgan et al., 2000). All cell lines were incubated at 37°C in a humidified chamber containing 5% CO₂/95% air.

Western Blot Analysis.

Nuclear extracts were prepared from logarithmically growing cells (5 × 10⁵cells/ml) as described previously (Mo and Beck, 1997). For analysis of NF-YA, NF-YB, and YB-1 transcription factors, proteins (100 μg/well) were separated on a 10% SDS-polyacrylamide gel, electrophoretically transferred onto nitrocellulose, and incubated with either purified mouse anti-NF-YA monoclonal antibody (PharMingen, SanDiego, CA), rabbit anti-NF-YB polyclonal antibody (Biodesign International, Kennebunk, ME), or rabbit anti-YB-1 antibody (generously provided by Dr. Manfred Dietel, Humbolt University, Berlin, Germany). Bound antibody was detected using the enhanced chemiluminescence (ECL) detection method (Amersham Pharmacia Biotech, Arlington Heights, IL) according to the manufacturer's instructions. Autoradiographic signals were quantified by densitometric scanning using a GS-700 Imaging Densitometer and Molecular Analyst Software (Bio-Rad, Hercules, CA). Equal loading of total cellular protein was determined by stripping and reblotting the membrane with anti-β-actin antibody (Oncogene Science, Cambridge, MA).

Construction of TopoIIα Promoter-Luciferase Plasmids for Transient Expression.

The recombinant plasmid p557 that contains the full-length topoIIα promoter (nt −557 to +90) and recombinant plasmids containing serial deletions of the full-length promoter: p382 (nt −382 to +90), p252 (nt −252 to +90), p182 (nt −182 to +90), p162 (nt −162 to +90), and p90 (nt −90 to +90) were provided by Dr. Q. Wang (Fibrogen, Inc., San Francisco, CA; Wang et al., 1997b). Bases are numbered with respect to the major transcription start site (designated +1) (Hochhauser et al., 1992).

The topoIIα promoter constructs with mutations in specific ICE sites were generated using PCR overlap extension. PCR amplification of each half of the mutated promoter constructs was performed separately, followed by joining of the halves in a third PCR. Conditions for PCR were as described previously (Wang et al., 1997b). The topoIIα promoter with a mutation in ICE5 was generated using the following primers: the first half of the mutant ICE5 promoter was PCR-amplified with the forward primer 5′-GGATCGGTACCGGGGTTGAGGCAGATGCCAG-3′ (f-557) (nt −561 to −542), which contains aKpnI restriction site, and the reverse primer 5′-CCAGGAACTGTCCAGCTATT-3′ (r-mICE5) (nt −398 to −379), which alters the ICE sequence from CCAAT to TCCAG (underlined). The second half was generated with the forward primer 5′-AATAGCTGGACAGTTCCTGG-3′ (f-mICE5) (nt −398 to −379), and the reverse primer 5′-GATCAGATCTGGTGACGGTCGTGAAGGGGC-3′ (r-557)(nt +71 to +90), which contains a BglII restriction site. Primer r-mICE5 contains nt complementary to the f-mICE5 primer. Each half of the mutant ICE4 topoIIα promoter construct was generated using the following primers: the first half was PCR-amplified with the forward primer 5′-GGATCGGTACCGTTCCTGGAGAATAAACATC-3′ (f-382)(nt −386 to −367), which contains a KpnI restriction site, and the reverse primer 5′-AGGGAATCTGGACTCTGAGA-3′ (r-mICE4) (nt −263 to −244), which alters the ICE sequence from CCAAT to ACTCT (underlined). The second half of the mutant ICE4 topoIIα promoter was generated using the forward primer 5′-TCTCAGAGTCCAGATTCCCT-3′ (f-mICE4) (nt −263 to −244), and the reverse primer r-557. Primer r-mICE4 contains nt complementary to the f-mICE4 primer. Each half of the mutant ICE3 topoIIα promoter construct was generated using the following primers: the first half was PCR-amplified with the forward primerf-557, and the reverse primer 5′-GTTTGAATAAACTACTCAGG-3′ (r-mICE3) (nt −179 to −160), which alters the ICE sequence from CCAAT to CTACT (underlined). The second half was generated with the forward primer 5′-CCTGAGTAGTTTATTCAAAC-3′ (f-mICE3) (nt −179 to −160), and the reverse primer r-557. Primer r-mICE3 contains nt complementary to the f-mICE3 primer.

Each half of the mutated ICE3, ICE4, or ICE5 PCR-amplified DNA fragments containing the overlapping, complementary sequences were gel purified (Geneclean III kit; Bio 101, Vista, CA), mixed together with its complementary half (1 μg each), and subjected to a second PCR amplification using a pair of the corresponding external primers as described above. Primers f-557, r-mICE5 and f-mICE5, r-557 were used to construct a topoIIα promoter containing a mutation in ICE5 (named p557-mt ICE5). Primers f-382, r-mICE4 and f-mICE4, r-557 were used to construct a topoIIα promoter containing a mutation in ICE4 (p382-mt ICE4). Primers f-557, r-mICE3 and f-mICE3, r-557 were used to construct a topoIIα promoter containing a mutation in ICE3 (p557-mt ICE3). All constructs were subcloned into the pGL2-Basic vector upstream of the luciferase reporter gene (PROMEGA, Madison, WI) and all mutations were verified by DNA sequencing.

Transient Transfections and Luciferase Reporter Assays.

DNA was introduced into the cells by electroporation using the Gene Pulser II apparatus with an extender (Bio-Rad), according to the manufacturer's instructions. Each luciferase plasmid was cotransfected with pSV-β-galactosidase control vector (Promega) into CEM and CEM/ICRF-8 cells and cell extracts were prepared after electroporation as described previously (Morgan et al., 2000). Luciferase activity was measured by a luminometer with an auto-injector (Model TD-20/20; Turner Designs, Sunnyvale, CA). Luciferase activities were normalized to β-galactosidase activities.

Preparation of Nuclear Extracts.

Nuclear extracts were prepared from logarithmically growing cells (5 × 10⁵ cells/ml). Cells were resuspended in ice-cold buffer A (10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, and 6 mM MgCl₂), incubated on ice for 10 min and centrifuged (1,200g, 5 min at 4°C). The supernatant was removed and the pellet was incubated in buffer A containing 0.5% Nonidet P-40. After 5 min on ice, the nuclei were sedimented (3,300g, 15 min at 4°C) and the nuclear pellets were lysed by the addition of ice-cold buffer B (20 mM HEPES, pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 10 μg/ml aprotinin, 2.5 μg/ml pepstatin, 10 μg/ml trypsin inhibitor, and 1 mM benzamidine). The nuclear extracts were incubated on ice for 30 min with occasional vortexing, centrifuged (15,000g, 15 min at 4°C), and the cleared supernatant was transferred to a new tube.

Electrophoretic Mobility Shift Assays (EMSAs).

Nuclear extracts (15 μg of protein/assay) were incubated for 30 min at room temperature in a 20-μl reaction mixture containing 10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 1 mM MgCl₂, 0.5 mM dithiothreitol, 0.5 mM EDTA, 8% glycerol, 1 μg poly(dI-dC), 20 μg bovine serum albumin, and ³²P-labeled double-stranded oligonucleotide probe (0.2 pmol). The oligonucleotides were labeled with ³²P using Klenow polymerase. Competition for protein binding activity was carried out in the presence of a 50-fold molar excess of unlabeled wild-type or mutant oligonucleotide probe. DNA-protein complexes and free DNA were applied to 4% nondenaturing polyacrylamide gels and separated by electrophoresis at 100 V for 2.5 h in a buffer containing 25 mM Tris, 190 mM glycine, and 1 mM EDTA. The gels were dried and exposed to X-ray film with intensifying screens. For supershift assays, approximately 2 μg of a rabbit anti-NF-YB antibody (Biodesign International) were preincubated with nuclear extract for 1 h at room temperature before the addition of³²P-labeled oligonucleotide probe. The following DNA oligonucleotides, along with their complementary strands were synthesized (Life Technologies, Gaithersburg, MD) and used for EMSA analysis: ICE3 (5′-CCTCCCTAACCTGATTGGTTTATTCAAACAAACC-3′, nt −189 to −156 of topoIIα promoter); mutant ICE3 (5′-CCTCCCTAACCTGAgTaGTTTATTCA AACA AACC-3′, nt −189 to −156); ICE4 (5′-GGTGAGCCCTTCTCATTGG CCAGATTCCC-3′, nt −273 to–245); and ICE5 (5′- GGGATCTTAAATAGATTGGCAGTTCCTGGAG-3′, nt −407 to −377). The criteria for assigning bands as nonspecific versus specific ICE-protein complexes is based on two controls. First, a DNA-protein complex is specific if it is eliminated by the addition of a molar excess of unlabeled competitor DNA of the same sequence but not by the addition of a molar excess of the same sequence containing mutation(s). Second, the specificity of the band shift is further confirmed when no DNA-complex is formed when the mutated³²P-labeled oligonucleotide alone is used in the EMSA.

Results

TopoIIα Promoter Activity Is Increased in the ICRF-187-Resistant CEM/ICRF-8 Cells.

We have recently demonstrated that the levels of topoIIα protein in the CEM/ICRF-8 cells are increased ∼ 5-fold compared with the parental CEM cells (Morgan et al., 2000). Similarly, we have also demonstrated by Northern blot analysis that CEM/ICRF-8 cells express greater than 2-fold increase in topoIIα mRNA (Morgan et al., 2000). To determine whether the increased expression levels of topoIIα are caused by transcriptional up-regulation, basal transcriptional activity of the full-length topoIIα promoter was measured by transiently transfecting into drug-sensitive and -resistant cells a construct containing the wild-type topoIIα promoter (p557) fused to the luciferase reporter gene. These experiments revealed that CEM/ICRF-8 expressed approximately 2.3-fold more promoter activity than the parental CEM cells (Table 1, plasmid p557). These results are consistent with our previous data (Morgan et al., 2000) and suggest that the up-regulation of topoIIα in the drug-resistant cells occurs in part at the transcriptional level. Importantly, because sequence analysis of the full-length topoIIα promoter from CEM/ICRF-8 compared with CEM revealed no mutations (data not shown), we suggest that the observed transcriptional up-regulation of the topoIIα gene in the drug-resistant cells may be caused by alterations in trans-acting factors of the transcriptional machinery.

Determination of the Promoter Sequences Responsible for Mediating TopoIIα Up-Regulation in ICRF-187-Resistant Cells.

To identify the promoter sequences responsible for enhanced topoIIα transcriptional activity in the drug-resistant cells, we measured basal luciferase activity in CEM and CEM/ICRF-8 cells transiently transfected with a series of topoIIα promoter-luciferase deletion constructs (Fig. 1). Sequence analysis of the topoIIα promoter has revealed the presence of five ICEs (Hochhauser et al., 1992), numbered 1 through 5. The CCAAT sequence is a commoncis-regulatory element; mutation of this motif has been shown to alter the transcriptional activity of eukaryotic genes (Santoro et al., 1988; Hochhauser et al., 1992). Stepwise deletion of the 5′ topoIIα promoter sequences, which includes stepwise deletion of each ICE (Fig. 1), resulted in a concomitant decrease in luciferase activity in both the CEM and CEM/ICRF-8 cells (Table 1). This overall pattern of basal promoter activity in Table 1 is consistent with previous reports (Hochhauser et al., 1992; Wang et al., 1997b) and confirms that ICEs play an important role in the transcriptional regulation of topoIIα expression.

One exception, however, was the p182 construct, in which promoter activity in CEM and CEM/ICRF-8 transfected with p182 was increased approximately 2- and 2.4-fold, respectively, compared with luciferase activity derived from p252 (Table 1). This observation suggests that the 70 base pairs between nt −252 and −182 may be important for binding of transcription factors and topoIIα expression in our leukemic cell lines. There were statistically significant differences in luciferase activity in CEM and CEM/ICRF-8 transfected with p557, p382, p252, and p182 but not with p162 or p90, which is of importance to our study. The low level of promoter activity associated with the p162 and p90 constructs is because our data were expressed as a percentage of the corrected luciferase activity of control CEM cells (Table 1). For example, with respect to the p162 construct, the absolute luciferase activities were actually 509 units in CEM/ICRF-8 and 283 units in CEM, which are then normalized to control CEM and expressed as a percentage. Data from Table 1 suggest that promoter sequences between nt −557 and −162, which include ICE3, ICE4, and ICE5, seem to be involved in the up-regulation of topoIIα promoter activity in the drug-resistant CEM/ICRF-8 cells. Notably, luciferase activity of the p182 promoter construct was significantly increased (3.5-fold) in CEM/ICRF-8 relative to CEM; this was the greatest difference compared with any of the other topoIIα promoter-luciferase plasmids (Table 1). Therefore, one of the criticalcis-acting promoter elements involved in topoIIα transcriptional up-regulation in the ICRF-187-resistant cells seems to include a 20-base-pair region (nt −182 to −162) that contains ICE3.

An Inverted CCAAT Element in the TopoIIα Promoter Plays a Role in the Down-Regulation of Promoter Activity in CEM and CEM/ICRF-8 Cells.

To better determine the roles of ICE3, ICE4, and ICE5 in the regulation of topoIIα expression, CEM and CEM/ICRF-8 cells were transiently transfected separately. Each of the topoIIα promoter-luciferase constructs contained a mutated ICE3, ICE4, or ICE5. As shown in Fig. 2, point mutation of ICE3 in the context of the full-length topoIIα promoter resulted in an increase in topoIIα transcriptional activity (1.33-fold) in the drug-sensitive CEM cells. Although this increase seems small, it is statistically significant as determined by the Student's ttest, suggesting that ICE3 may have an important role in topoIIα expression by negatively regulating basal promoter activity. In the case of the drug-resistant cells, CEM/ICRF-8 exhibited ∼2-fold more wild-type topoIIα promoter activity compared with CEM. Importantly, this activity did not significantly change when ICE3 was mutated (Fig.2). From these data, we conclude that (a) ICE3 may serve as a negativecis-acting element, and (b) because enhanced transcriptional activity in CEM/ICRF-8 is independent of a wild-type ICE3 sequence, then there may exist alterations in transcription factor binding to ICE3 in the drug-resistant cells.

Luciferase activity of the topoIIα promoter plasmids containing wild type ICE4 or ICE5 was significantly increased in CEM/ICRF-8 compared with CEM cells (Fig. 3A and B). Mutation of ICE4 or ICE5 contributed to a decrease in luciferase activity in both the CEM and CEM/ICRF-8 cells, although the decrease was not significant, suggesting that neither ICE4 nor ICE5 alone seems to contribute to topoIIα promoter activity in these cells.

Differential Binding of Protein Complexes to ICE3 in the TopoIIα Promoter from CEM and CEM/ICRF-8 Cell Lines.

To further explore the role of ICE3 in the regulation of topoIIα gene expression, protein factor binding to this element was analyzed in CEM and CEM/ICRF-8 by EMSA using a radiolabeled double-stranded DNA oligonucleotide spanning ICE3 (nt −187 to −155). Nuclear extracts derived from CEM and CEM/ICRF-8 cells contained a factor that specifically bound to the ICE3 probe (Fig.4, lanes 1 and 6). This DNA-protein complex was eliminated by the addition of 50-fold molar excess of unlabeled competitor DNA of identical sequence (Fig. 4, lanes 2 and 7) but not by the addition of 50-fold molar excess mutated ICE3 probe (Fig. 4, lanes 3 and 8). No DNA-protein complex was formed when the mutated ICE3 oligonucleotide alone was used in the EMSA (Fig. 4, lanes 4 and 9), confirming the specificity of this band shift. The addition of an antibody against NF-YB, a specific ICE-binding factor (Hooft van Huijsduijnen et al., 1990; Maity et al., 1992), resulted in both inhibition of complex formation and a supershift (Fig. 4, lanes 5 and 10). At a longer gel exposure time, the antibody supershift became more apparent in extracts derived from CEM/ICRF-8 (data not shown). The major nonspecific band seems to decrease in intensity with NF-YB antibody coincubation. Given that we are using a polyclonal antibody, it is possible that the anti-NFYB antibody cross-reacts nonspecifically with other protein in the prepared cell extracts, thus resulting in the inhibition of nonspecific DNA-protein complex formation. As a control, the addition of a nonspecific rabbit polyclonal antibody of the same isotype as anti-NF-YB gave no supershift (data not shown). These results suggest that NF-YB makes up at least one component of the ICE3 protein-binding complex. Our gel shift profiles showing ICE-protein complexes and NF-YB supershifts are in agreement with others (Isaacs et al., 1996; Herzog and Zwelling, 1997; Furukawa et al., 1998). Interestingly, although the result is qualitatively similar to that obtained using extracts from the parental CEM cells, the level of DNA-protein complexes formed was lower in the ICRF-187-resistant cell line (Fig. 4, compare lane 1 to lane 6). These quantitative differences are not caused by differences in cell growth in culture or in extract preparation because only cells growing logarithmically were used for each independent experiment, and all have yielded the same results.

Nonetheless, to confirm these quantitative differences in protein-ICE3 complex formation between the drug-sensitive and -resistant cells, we incubated increasing pmol amounts of ICE3 probe with a constant amount (in micrograms) of nuclear extract derived from CEM and CEM/ICRF-8 cells and analyzed these ICE3-protein interactions by EMSA. Incubating increasing picomolar amounts of ICE3 probe with a constant microgram amount of nuclear extract derived from CEM resulted in a general increase in the formation of ICE3-protein complexes (Fig.5A). Using ICE3 probe concentrations greater than 1.0 pmol interferes with DNA-protein complex formation. By contrast, incubating the same increasing amounts of ICE3 probe with a constant microgram amount of nuclear extract derived from the ICRF-187-resistant cells resulted in little or no increase in ICE3 binding (Fig. 5A). Importantly, although there is little or no increase in binding for the drug-resistant cells, the amount of nonspecific binding and free probe increased in both the drug-resistant and -sensitive cells alike. These results, quantified in Fig. 5B, demonstrate that at 1.0 pmol ICE3 substrate, up to 3-fold more DNA-protein complexes were formed in CEM compared with CEM/ICRF-8 cells (Fig. 5B).

Data from Figs. 4 and 5 indicate that ICE3 in the topoIIα promoter is recognized by a protein complex that contains NF-YB and that this complex is present at a reduced level in the drug-resistant CEM/ICRF-8 cells. The observation that enhanced topoIIα promoter activity in the drug-resistant cells (Table 1) correlates with decreased protein complex binding to ICE3 suggests that the protein factors that specifically bind to ICE3 may negatively regulate basal topoIIα promoter activity.

Analysis of Protein Expression Levels of the CCAAT Box Binding Factors NF-YA, NF-YB, and YB-1 in CEM and CEM/ICRF-8 Cells.

By Western blot analysis, the protein levels of the CCAAT box binding factors, NF-YB and NF-YA (Li et al., 1992; Mantovani et al., 1992), were similarily expressed in CEM and CEM/ICRF-8 cells (Fig.6). The apparent sizes of the isoforms for NF-YB (∼32 and 34 kDa) and for NF-YA (∼42 and 44 kDa) in Fig. 6are in agreement with previously published results (Li et al., 1992;Mantovani et al., 1992). Another known CCAAT binding factor includes a 35-kDa YB-1 protein (Didier et al., 1988), which was also present in equal amounts in the CEM and CEM/ICRF-8 cells (Fig. 6). Equal loading of nuclear protein was confirmed by blotting for β-actin (data not shown). Thus, our observations suggest that the transcriptional up-regulation of topoIIα in the drug-resistant cells is not caused by altered expression of these trans-acting factors, suggesting that such alterations as mutations in or post-translational modifications of NF-Y or other binding proteins in the complex may exist.

Discussion

We have shown recently that topoIIα expression is up-regulated at both the protein and mRNA levels in a leukemic CEM cell line (CEM/ICRF-8) selected for resistance to the topoII catalytic inhibitor, ICRF-187 (Morgan et al., 2000). Here, we demonstrate that this up-regulation of topoIIα occurs at the transcriptional level because topoIIα promoter activity was significantly increased in CEM/ICRF-8 compared with CEM. We have localized the region of the promoter (nt −557 to −162) that seems to be responsible for topoIIα transcriptional up-regulation in the drug-resistant cells. An ICE located between nt −182 and −162 (ICE3) may play a role as a negativecis-acting element in our leukemic cell lines. We base this conclusion on the following: 1) Using EMSA, ICE3 was recognized by a NF-YB-containing protein complex and this DNA-protein complex was present at a reduced level in the topoIIα-overexpressing ICRF-187-resistant cells compared with the drug-sensitive cells, and (2) point mutation of ICE3 resulted in a significant increase in topoIIα transcriptional activity in the drug-sensitive CEM cells.

ICEs are known to be involved in either up- or down-regulating topoIIα gene expression. For example, wild-type p53 represses topoIIα gene expression through functional interaction with ICE1 (Wang et al., 1997b). Moreover, ICE1 was involved in topoIIα transcriptional repression in serum-starved cells (Falck et al., 1999), and heat-shock activation of topoIIα required a decrease in nuclear binding activity to ICE1 (Furukawa et al., 1998). In contrast, ICE2 was identified as a positive regulatory element for topoIIα expression in proliferating cells (Isaacs et al., 1996).

Results here suggest that the up-regulation of topoIIα transcriptional activity in the ICR F-187-resistant cells may be caused in part by differential binding of the NF-Y family of transcription factors to a region of the topoIIα promoter that contains the third inverted CCAAT cis-acting element. The presence of ICE3 binding activity has been reported (Herzog and Zwelling, 1997; Wang et al., 1997a), but ours is the first study to demonstrate differential ICE3 binding activity as a function of drug-resistance. Accordingly, we offer the suggestion that transcription factor binding to ICE3 negatively regulates topoIIα expression in CEM cells, possibly by displacing positive regulatory factors flanking this region of the topoIIα promoter. In this regard, one study has shown that NF-Y DNA binding activity seems to induce bends in the DNA and the extent of bending depends upon the promoter sequences flanking the CCAAT box (Ronchi et al., 1995). The putative negative role of an ICE is further supported by the observation made by Takano et al. (1999) that binding of a nuclear protein factor to ICE1 down-regulates topoIIα gene expression in an etoposide-resistant cell line.

In terms of our ICRF-187-resistant cells, regulation of topoIIα expression seems to be attenuated by dissociation of these NF-Y-containing nuclear proteins from the ICE3 promoter region. Altered binding is not caused by mutations in the topoIIα promoter, because sequence analysis revealed a wild-type promoter in the ICRF-187-resistant cells, just as in the drug-sensitive CEM cells. Furthermore, by Western blot analysis, there were no apparent differences in the levels of the CCAAT box binding proteins, NF-YB, NF-YA, or YB-1 in drug-sensitive and -resistant cells that could account for changes seen in nuclear protein-ICE3 complex formation. These observations suggest that in CEM/ICRF-8, such alterations as mutations in or post-translational modifications of NF-Y or other binding proteins in the complex may exist. If this were the case, then protein-protein interactions and protein-DNA binding activity may all be affected. In the case of interactions between NF-Y and histones, a recent study has suggested a model in which NF-Y interacts/recruits histone acetyltransferases to the promoter, thereby stimulating histone acetylation and activating G₂/M-dependent transcription of the topoIIα gene (Adachi et al., 2000). Studies are currently underway to determine whether differential phosphorylation and/or gene mutation in CCAAT box binding proteins, and differential histone acetylation, would alter transcription factor binding to the topoIIα promoter in our ICRF-187-resistant cells.

Although our data presented herein suggest that ICE3 is a negative regulator of topoIIα expression, it does not seem to be the sole contributor to the altered regulation of topoIIα expression in the ICRF-187-resistant cells, as suggested by our mutational analysis in Fig. 2. The topoIIα promoter fragment (nt −557 to −162) found responsible for topoIIα transcriptional up-regulation in the drug-resistant cells contains ICE3, ICE4, and ICE5. Although there were no differences in the amount of ICE4 or ICE5 protein-complexes formed between the drug-sensitive and -resistant cells as reported by EMSA (data not shown), the possibility exists that promoter elements further upstream or downstream of these ICEs may cooperate with thesecis-acting elements to enhance topoIIα expression in the drug-resistant cells. Additional changes in chromatin structure and/or methylation status within specific regions of the topoIIα promoter may also contribute to the altered topoIIα expression observed here; studies examining some of these possibilities are currently underway.

We have also shown here that ICE1 and ICE2 are neither necessary nor sufficient for topoIIα up-regulation in CEM/ICRF-8 cells. It is not known what distinguishes one ICE from another, but the specific sequences and regulatory promoter elements flanking each ICE that contribute to the overall regulation of inverted CCAAT boxes may be involved. Indeed, by EMSA analysis, the ICE5 oligonucleotide seems to bind to a different nucleoprotein complex in cell extracts compared with the ICE3 and ICE4 oligonucleotides (Herzog and Zwelling, 1997; S.E.M. and W.T.B., unpublished observations). In addition to ICEs, other cis-acting promoter elements have been known to affect topoIIα expression (Fraser et al., 1995; Brandt et al., 1997). Larger DNA fragments of ∼200 bp each, spanning the entire topoIIα promoter, were used in EMSAs to determine whether cis-acting promoter elements, independent of ICEs, play a role in topoIIα up-regulation in CEM/ICRF-8, but we have obtained no clear evidence of differential DNA-protein interactions outside of ICE3 in CEM or CEM/ICRF-8 cells (data not shown).

In the context of drug resistance, very little is known about the factors that regulate topoIIα expression at the transcriptional level. Our study is the first to show that topoIIα transcriptional up-regulation in ICRF-187-resistant cells is mediated in part by altered regulation of the third inverted CCAAT box in the topoIIα promoter. Furthermore, we have demonstrated that ICE3 seems to play a repressive role in topoIIα expression. Our present studies involve characterization of additional ICE3-binding proteins andcis-acting elements and delineation of differences in the overall chromatin structure of the topoIIα promoter that may account for topoIIα over-expression in our ICRF-187-resistant cells.