Daniel M. Sullivan H. Lee Moffitt Cancer Center and Research Institute, 12902 Magnolia Drive, Tampa, FL 33612, USA.
The resistance of several leukaemic and myeloma cell lines (CCRF, L1210, HL-60, KG-1a and RPMI 8226) to VP-16 was found to increase with cell density and to be maximal (3.5- to 39-fold) in plateau phase cell cultures, as measured by clonogenic and MTT assays. Non-transformed confluent Flow 2000 human fibroblasts and Chinese hamster ovary (CHO) cells were also five- and 15-fold resistant to VP-16 respectively. The transition from log to plateau phase was accompanied by a drastic decrease in topoisomerase (topo) IIα content in CHO cells and human fibroblasts, while the leukaemic cells maintained constant cellular levels of topo IIα and topo IIβ. However, the nuclear topo IIα content was found to decrease as a result of translocation of the enzyme to the cytoplasmic compartment in the leukaemic cells. This was confirmed by subcellular fractionation experiments, Western blotting analyses and immunocytochemistry studies. The quantity of topo IIα in plateau phase cytoplasmic fractions ranged from 18% in L1210 cells to 50% in HL-60 and 8226 cells, as measured by both immunoblotting and quantification of the label in immunofluorescent images. The cytoplasmic fraction from plateau phase cells retained topo II catalytic activity, as measured by the decatenation of kinetoplast DNA. The nuclear–cytoplasmic ratio of topo IIα may be critical in determining the sensitivity of leukaemic cells to topo II inhibitors. Cytoplasmic trafficking of topo IIα was observed in plasma cells obtained from patients with multiple myeloma, and perhaps contributes to drug resistance in this disease.
DNA topoisomerase function is required for replication, transcription, recombination, DNA repair, chromatin and chromosome condensation and decondensation, nucleolar formation and sister chromatid segregation ( Wang, 1996). Topoisomerase (topo) I activity results in single-strand breaks in DNA, and this enzyme is involved in changing the superhelical density of DNA by virtue of its relaxation/supercoiling activities. Topo II reactions are ATP dependent and are accompanied by the formation of a covalent complex between dimeric topo II and the 5′ phosphoryl end of each cleaved strand. Topo I reactions do not require ATP, and the formation of the transient covalent complex is at the 3′ phosphoryl end of cleaved DNA. Several clinically important anti-tumour drugs (VP-16, doxorubicin and mitoxantrone) inhibit topo II religation activity and lead to an increase in DNA scission ( Liu, 1989). These DNA breaks probably lead to cell death by interfering with the progress of the replication fork ( Qiu et al, 1996 ). This process results in high-molecular-weight DNA fragmentation and further endonuclease activation, which trigger apoptotic events, such as the appearance of low-molecular-weight DNA fragments and chromatin compaction ( Sugimoto et al, 1998 ), protease activation, lamin disruption and structural disorganization of cellular compartments.
In culture, most cell lines display a differential sensitivity to topo II inhibitors, which depends on cell density ( Sullivan et al, 1986 , 1987; Chow & Ross, 1987; Markovits et al, 1987 ). When plateau cell density (or confluency) is reached, cells become extremely drug resistant. The cellular content of topo II and topoisomerase activity are regulated in a cell cycle-dependent and proliferation-dependent manner ( Tricoli et al, 1985 ; Estey et al, 1987 ; Hsiang et al, 1988 ; Robbie et al, 1988 ; Schneider et al, 1988 ; Chapuis et al, 1992 ). Drug resistance to topo II poisons may be achieved by different mechanisms, which may involve quantitative or qualitative changes. The transition from log to plateau density may lead to an attenuation of cellular topo IIα amount and activity, with concomitant drug resistance, in non-transformed cell lines ( Sullivan et al, 1987 ). This transition does not necessarily lead to a decrease in cellular topo II content in tumour cell lines ( Sullivan et al, 1986 , 1987), yet the cells become resistant. This discrepancy led us to investigate the overall cellular quantity of topo IIα and topo IIβ, drug sensitivity as a function of cell density, and the subcellular distribution of both isoforms of topo II.
MATERIALS AND METHODS
Etoposide (VP-16), phenylmethylsulphonyl fluoride (PMSF), ATP, dimethyl sulphoxide (DMSO), EDTA, Nonidet P-40, dithiothreitol (DTT), MTT, cis-platinum, vincristine and standard chemicals were obtained from Sigma Chemical (St Louis, MO, USA). Topotecan and mitoxantrone were generously provided by SmithKline Beecham Pharmaceuticals (Philadelphia, PA, USA) and Lederle Laboratories (Pearl River, NY) respectively. [3H]-VP-16 (7.4 GBq/mmol) was supplied by Moravek Biochemicals (Brea, CA, USA). DNA polymerase I (Klenow fragment), anti-topoisomerase IIα (Ki-S1) and anti-histone monoclonal antibodies were obtained from Roche Molecular Biochemicals (Indianapolis, IN, USA). The FPLC and prepacked Mono Q column were obtained from Pharmacia (Piscataway, NJ, USA).
HL-60, RPMI-8226 and CEM/CCRF cells were grown in RPMI-1640 medium containing 10% fetal calf serum (FCS) and an additional 2 mmol/l L-glutamine, whereas L1210 cells were grown in RPMI-1640 medium with 20% FCS. KG-1a cells were grown in suspension culture in Iscove's media with 20% FCS. All media were supplemented with 100 U/ml penicillin and 100 μg/ml streptomycin from Gibco BRL (Gaithersburg, MD, USA). The log density of all five cell lines was defined as 2–4 × 105 cells/ml. The plateau density was different for each cell line: 9 × 105 cells/ml for RPMI-8226, 1.6 × 106 cells/ml for HL-60, 2 × 106 cells/ml for CEM/CCRF and L1210, and 8 × 106 cells/ml for KG-1a cells. Plateau phase is defined as a cell viability ≥ 95% (by trypan blue dye exclusion) and no net gain in cell number. The non-transformed cell lines were grown in monolayer culture in α-MEM media supplemented with 5% FCS for Chinese hamster ovary (CHO) cells and 10% FCS for Flow fibroblasts. Log cell densities were 2 × 104 cells/ml (for 10 ml of media in a 25-cm2 flask). Plateau phase was reached at 8 × 105 cells/ml for CHO cells and at 2 × 105/ml for Flow fibroblasts.
Cytotoxicity assays were performed by treating 1 × 105 cells from either log or plateau densities (plateau density cells were adjusted to 1 × 105 cells/ml with fresh media and incubated for 30 min before drug treatment). Drug treatment was for 1 h with different concentrations of VP-16, mitoxantrone, topotecan, cis-platinum and vincristine (diluents were DMSO, water, methanol, DMSO and water, respectively, and controls had the same concentration of diluent). For the MTT assay, cells were plated at 3 × 103 cells/well (in 200 μl) for 7–10 days. Formazan formation as an index of viability was read at 540 nm with Dynex Revelation (Chantilly, VA, USA) after treatment for 4 h with 0.4 mg/ml MTT. The clonogenicity assay was performed by plating 2500 drug-treated cells in 0.3% soft agar in 15% FCS in RPMI-1640 media for 10–14 days at 37°C in the presence of 5% CO2. Colonies were counted manually.
Drug uptake and thymidine incorporation
Equilibrium concentrations of [3H]-VP-16 were measured as described previously ( Sullivan et al, 1986 , 1987). The incorporation of [3H]-thymidine by log and plateau phase cells was assayed as described by Robbie et al (1988 ).
Equal numbers of log and plateau phase 8226, L1210, CCRF, Kg-1a, HL-60, CHO and Flow fibroblast cells were fixed with 70% ethanol at 1 × 106 cells/ml in a 200 μl volume. After RNase digestion for 1 h, propidium iodide was added at 0.1 mg/ml immediately before reading, and DNA histograms were determined by FACS-Scan (Becton Dickinson, San Jose, CA, USA).
Whole-cell lysates were prepared by lysing cell suspensions in 1 × sample buffer (2% SDS, 0.1 mol/l DTT, 10% glycerol and 0.025% bromophenol blue) and disrupting genomic DNA with 10 bursts from a Branson sonifier ( Heck & Earnshaw, 1986). The fractions from the subcellular separation (see below) were mixed with 4 × sample buffer and boiled for 3 min. Proteins were separated on 7.5% SDS–PAGE gels and transferred overnight onto polyvinylidene difluoride (PVDF) membranes (Bio-Rad Laboratories, Hercules, CA, USA). Membranes were blocked for 1 h with 0.5% Tween-20 in TBS (0.5 mol/l NaCl, 20 mmol/l Tris, pH 7.5) and treated with either polyclonal antibody 454 against topo IIα or polyclonal antibody JB1 against topo IIβ at dilutions of 1:5000 in TBS buffer supplemented with 0.05% Tween-20. The polyclonal antibody JB1 was produced by immunizing a rabbit with a total of 5 mg of FPLC-purified recombinant topo IIβ expressed in a yeast system ( Austin et al, 1995 ). This system was generously provided by Dr Carolyn Austin (University of Newcastle, Newcastle upon Tyne, UK). The JB1 antiserum specifically recognizes topo IIβ by immunoblotting, immunocytochemistry and immunoprecipitates only this topo II isoform. The secondary donkey anti-rabbit antibody was conjugated to horseradish peroxidase and diluted in 0.05% Tween-20 in TBS. The signal was detected by ECL (Amersham, Arlington Heights, IL, USA). Topo I was detected with the C-21 murine monoclonal IgM antibody directed against an epitope in the C-terminal 67 kDa of topo I (a generous gift from Dr Y.-C. Cheng, Yale University Medical School, New Haven, CT, USA) ( Beidler & Cheng, 1995). For topo I, the membrane was blocked with 5% non-fat milk and 0.15% Tween-20 in PBS. ECL detection of the signal was with a secondary sheep anti-mouse antibody conjugated to horseradish peroxidase (Amersham, Arlington Heights, IL, USA).
Preparative subcellular fractionation into cytoplasmic and nuclear compartments
For the separation of nuclear and cytoplasmic fractions, a protocol involving the use of EDTA was utilized, as it ensures an efficient disruption of the tight perinuclear shell of intermediate filaments. Generally, 5 × 106 log or plateau phase cells were washed with PBS and the pellet resuspended in 150 μl of 0.15 mol/l sucrose, 2.5 mmol/l EDTA, 1 mmol/l PMSF, 0.25% NP-40, pH 8.0, and layered over 50 μl of 0.25 mol/l sucrose, 0.1 mmol/l EDTA, 2 mmol/l PMSF, pH 6.8. Tubes were centrifuged for 1 min at 14 000 × g, and the resulting supernatant contained the cytoplasmic contents, distinct from the transparent nuclear pellet. The fractions were then analysed by immunoblotting.
Cell suspensions from log and plateau densities were diluted to 5 × 104 cells/ml with PBS and cytocentrifuged for 3 min at 500 r.p.m. onto double cytoslides (Shandon, Pittsburgh, PA, USA). The slides were fixed with 4% paraformaldehyde for 10 min at room temperature and permeabilized with 0.5% Triton X-100, 1% glycine, PBS for 1 h. Rabbit polyclonal antibody 454 against topo IIα ( Sullivan et al, 1989 ) was used simultaneously with a mouse monoclonal antibody against histone (Roche Molecular Biochemicals). Both were diluted 1:100 with 0.1% NP-40, 1% BSA in PBS, and the incubation was for 1 h at room temperature. After several washes with PBS for 2 h, slides were incubated with a mixture of goat anti-rabbit IgG-TRITC-labelled antibody (Sigma) diluted 1:80 and goat anti-mouse IgG-FITC-labelled antibody (Sigma) diluted 1:125 in 0.1% NP-40, 1% BSA in PBS for 35 min in the dark at room temperature. When labelling topo I with the C-21 murine monoclonal IgM antibody and topo IIβ with the JB1 polyclonal antiserum, both were used at a dilution of 1:100 and the incubation, washes and secondary antibodies were the same as above. For concurrent identification of the light-chain-producing plasma cells and topo IIα, a combination of mouse monoclonal antibody Ki-S1 (Roche Molecular Biochemicals), specific for the C-terminal domain of topo IIα ( Boege et al, 1995 ), and either anti-κ or anti-λ FITC-labelled Fab fragments of rabbit anti-human antibodies (Dako, Glostrup, Denmark) were used. Ki-S1 antibody was used at a dilution of 1:100, while anti-κ and anti-λ antibodies were diluted 1:50.
After several washes in PBS, all slides were covered with coverslips in mounting media of antifade/DAPI (Vector, Burlingame, CA, USA). Immunofluorescence was observed with a Leitz Orthoplan 2 microscope, and images were captured by a CCD camera with the Smart Capture program (Vysis, Downers Grove, IL, USA). Scanning confocal microscopy was performed with a Zeiss Axiovert 100M laser confocal microscope, and images were captured with LSM 510 (Gottingen, Germany).
Quantitative measurement of the immunofluorescence of topo IIα
Quantification of the topo IIα present in the nuclear and cytoplasmic compartments of 200 cells (50 cells/experiment performed four times) from log and plateau phase cells was performed using the Smart Capture program, with measurements confirmed by Adobe Photoshop 5.0 (Adobe Systems, San Jose, CA, USA). Each compartment was quantified separately in pixels, and nuclear/cytoplasmic ratios were calculated individually for each cell. The ratios were plotted on scattergrams in which 50 values/column were presented by means of Graph Pad Prism 2.0 (GraphPad Software, San Diego, CA, USA). The mean values and standard deviations were also calculated with GraphPad software.
Measurement of topoisomerase II enzyme activity
The enzyme activity of topo II was measured by two different methods: catalytic activity by the decatenation of kDaNA; and VP-16-induced cleavage activity by the selective K/SDS precipitation assay ( Liu et al, 1983 ; Nelson et al, 1984 ). With kDNA as a substrate, it is imperative to minimize any nuclease activity that could mimic topo II activity (especially in cytoplasmic fractions in which lysosomal nucleases are abundant). A partial purification of topo II was performed to eliminate the interfering nuclease activity.
To measure the topo II enzyme activity present in the nuclear and cytoplasmic compartments, an initial subcellular fractionation was performed followed by salt extraction of the enzyme and partial purification by FPLC on a Mono Q column (Pharmacia). All procedures were carried out at 4°C and were accomplished within 3 h to minimize proteolysis, although 2 mmol/l PMSF and a 1 × protease inhibitors cocktail (consisting of 20 μg/ml each of antipain, aprotinin, chymostatin, leupeptin, soybean trypsin inhibitor, benzamidine and pepstatin A) were present.
Nuclear and cytoplasmic compartments from log and plateau phase human 8226 myeloma cells were separated as described above on two different occasions. The starting cell numbers were identical for log and plateau phase (3 × 108 cells). The cytoplasmic compartment (around 2.4 ml total volume) was adjusted to 0.5 mol/l KCl, 4 mmol/l DTT and 50 mmol/l Tris, pH 7.0, and extracted on ice for 20 min. After 10 min of centrifugation in a microfuge (14 000 × g) at 4°C, the supernatant was centrifuged further for 45 min in a Beckman 100S airfuge (35 000 × g) at 4°C. The nuclear pellet was extracted with 0.5 mol/l KCl, 4 mmol/l DTT and 50 mmol/l Tris, pH 7.0, for 45 min at 4°C. The nuclear extract was centrifuged at 100 000 × g for 45 min at 4°C to remove genomic DNA. After centrifugation, the protein concentrations of both supernatants were determined, and the nuclear and cytoplasmic fractions were loaded separately on a Mono Q column after diluting the salt concentration to 0.1 mol/l KCl with a buffer containing 20 mmol/l Tris, pH 7.0, 4 mmol/l DTT, 10% glycerol, 1 mmol/l Na2EDTA, 1 mmol/l PMSF. A step gradient of KCl (0 to 0.2 mol/l to 0.4 mol/l) was used, and 0.5 ml fractions were collected ( Sullivan et al, 1989 ). Topo II activity consistently eluted with 0.4 mol/l KCl in three fractions from both nuclear and cytoplasmic extracts, and the protein was detected on Western blots in only these three fractions. Each fraction was titrated for decatenation activity, in which 1 unit of activity is defined as the amount of protein needed to completely decatenate 1 μg of kDNA in 30 min at 30°C.
Processing of bone marrow aspirates from patients with multiple myeloma
Approximately 5 ml of bone marrow aspirate was collected in EDTA tubes and diluted to 30 ml with PBS. This was then layered over 15 ml of Ficoll-Paque Plus (Pharmacia Biotech) and centrifuged at 400 × g in a swinging bucket rotor at 4°C for 30 min. The mononuclear cell interface was collected and washed twice with cold PBS. The cell number was determined, and 105 cells were cytospun onto double cytoslides. The cells were fixed with 4% paraformaldehyde and stored at −85°C. Aliquots of cells were preserved in freezing media consisting of 40% RPMI-1640, 50% FCS and 10% DMSO in liquid nitrogen. The samples were obtained from patients with multiple myeloma on a high-dose chemotherapy study, approved by the Institutional Review Board.
Table 2. Table II. Cell cycle distribution and [3H]-thymidine incorporation of log and plateau phase non-transformed and tumour cell lines*.
Drug uptake of [3H]-VP-16
To examine the possibility that plateau phase drug resistance was caused by an alteration in VP-16 transport, the 1 h equilibrium concentration of 25 μmol/l [3H]-VP-16 was determined for log and plateau phase HL-60 and 8226 cells. We have demonstrated previously that the plateau phase VP-16 resistance of CHO, CCRF and L1210 cells is not the result of a decrease in equilibrium concentrations of [3H]-VP-16 ( Sullivan et al, 1986 , 1987). The [3H]-VP-16 uptake by log and plateau phase HL-60 cells was 22 155 c.p.m./mg dry weight and 19 561 c.p.m./mg respectively. For 8226 log and plateau phase cells, the values were 17 089 c.p.m./mg and 17 990 c.p.m./mg. We conclude that an alteration in the drug transport of VP-16 was unlikely to account for the observed drug resistance.
Cellular topo IIα and β content by immunoblotting
The overall quantity of topo II was determined separately for the topo II isoforms at log and plateau densities in the leukaemic, myeloma and non-transformed cell lines. A remarkable difference between malignant and non-transformed cell lines was the near disappearance of topo IIα from CHO and fibroblast cells at a time when they reach confluency and lose their proliferative potential (Fig 1). In contrast, all tumour cell lines preserved their total cellular topo IIα content, independent of cell density. Topo IIβ did not show any drastic quantitative changes in the lines studied and was evidently not downregulated in either normal or neoplastic cells. The content of topo I was not found to be significantly changed by the log to plateau phase transition in fibroblasts, 8226, HL-60, KG-1a, CCRF or L1210 cells. Thus, the relatively high level of resistance of plateau phase L1210 cells to topotecan (see 1 Table I) does not appear to result from a downregulation of cellular topo I.
Subcellular distribution of topo II by immunocytochemistry
The above results were confirmed by immunofluorescent microscopy, in which the subcellular distribution of histones and topo IIα were also examined. At log density, both antigens were strictly nuclear, while at plateau cell density, the tumour cell lines demonstrated a translocation of topo IIα from the nucleus to the cytoplasm (Fig 2). In the case of CHO cells and Flow fibroblasts, there was a translocation of topo IIα to the cytoplasm. However, considering the major degradation of topo IIα that occurs at confluency (Fig 1), the contribution of cytoplasmic trafficking of topo IIα to drug resistance in these cells was probably minimal. In order to evaluate the amount of topo IIα at the single cell level statistically, we measured the pixel intensity of nuclear and cytoplasmic topo IIα in 200 individual cells (50 cells/experiment four times) at log, mid-log and plateau densities. These measurements are presented as a ratio of nuclear/cytoplasmic topo IIα from 50 cells on scattergrams (Fig 3). CHO cells had a proportional decrease in the nuclear/cytoplasmic ratio, such that, at plateau phase, the ratio was < 1. This suggested that, for the remaining undegraded topo IIα, there was relatively more in the cytoplasm than in the nucleus.
The relative subcellular distribution of topo IIα may be critical in determining an individual cell's drug sensitivity, but the ratio does not take into account the absolute amount of topo IIα in the compartments. As the leukaemic cells progressed from early log phase to plateau phase, there was a significant decrease in the topo IIα nuclear/cytoplasmic ratio for all cell lines except KG-1a (see Fig 3 for mean nuclear/cytoplasmic ratios). Thus, at confluent densities, all cell lines (except KG-1a) had significantly more topo IIα in the cytoplasmic compartment. The immunolabelling of three cell lines (CCRF, L1210 and HL-60) for topo I and topo IIβ demonstrated no significant trafficking of these topoisomerases to the cytoplasm during the transition from log to plateau phase (data not shown).
Separation of nuclear and cytoplasmic topo IIα in tumour cell lines
We performed subcellular fractionation experiments in order to separate nuclear and cytoplasmic fractions from cells at log and plateau cell densities. The quantity of topo IIα in these two compartments in the leukaemic and myeloma cell lines is shown in Fig 4. This biochemical analysis confirmed the immunofluorescence studies above. The relative amount of topo IIα in the cytoplasm from the four cell lines was determined by densitometric analysis of the immunoblots: 29% for 8226 cells; 70% for HL-60 cells; 27% for CCRF cells; and 16% for L1210 cells.
We next examined the topo II enzyme activity present in the nuclear and cytoplasmic fractions from log and plateau density 8226 cells. To eliminate the interference of nucleases in the kDNA decatenation assay, the nuclear and cytoplasmic fractions were partially purified by FPLC on two separate occasions, with similar results. Fig 5 shows the VP-16-induced cleavage activity (K/SDS assay) present in the crude nuclear fractions from log and plateau phase 8226 cells (Fig 5A), as well as the catalytic activity (decatenation) present in partially purified fractions from log phase and confluent cells (Fig 5B and C). Although there was no significant difference in VP-16-induced covalent topo II * DNA complex formation between nuclear compartments, there was a significant difference in catalytic activity present in the cytoplasmic fractions. Ten per cent of the total decatenation activity was present in the cytoplasm of log phase cells, whereas 70% of the total activity was in the cytoplasm of confluent cells. The lack of a difference in VP-16-induced cleavage in the nuclear fraction may reflect the activity of the topo IIβ isoform, whose activity was also measured in this assay.
To investigate the mechanism underlying the trafficking of topo IIα into the cytoplasmic compartment further, we performed experiments that bring log phase HL-60 and 8226 cells to a plateau density artificially. Immunofluorescent microscopy demonstrated that, after a short period of time (1–5 h), a redistribution of topo IIα occurred in 8226 cells, which was similar to that shown above (which required several days for cells to achieve plateau-phase) (Fig 6). Cells began to translocate topo IIα into the cytoplasm after 2 h and, after 4 h, 8226 cells were found to be fivefold resistant to VP-16 by colony-forming cytotoxicity assays. Experiments with HL-60 cells gave similar results, in that 1 h ‘artificial’ plateau HL-60 cells did not demonstrate drug resistance, while VP-16 resistance was twofold after 5 h.
Subcellular distribution of topoisomerases in clinical multiple myeloma samples
We next studied the distribution of topo I, topo IIα and topo IIβ in clinical samples: bone marrow aspirates from myeloma patients. Representative aspirates from two out of 10 patients examined are shown in Fig 7. These aspirates (both with > 90% plasma cells) were obtained before high-dose chemotherapy but after standard-dose chemotherapy. In all patients, there was a considerable proportion of topo IIα in the cytoplasmic compartment of the plasma cells (Fig 7C and O). The percentage of plasma cells that was brightly fluorescent and displayed a nuclear distribution of topo IIα was always in the range 10–15% and reflected cells that were probably in the S- or G2/M-phase of the cell cycle (Fig 7A and M). Topo I was found to be strictly nuclear or nucleolar except for one patient (90% of cases) (Fig 7E and Q). In the one patient who had relapsed after high-dose chemotherapy, topo I was found to be exclusively cytoplasmic. The topo IIβ subcellular distribution was mostly a mixed nuclear and cytoplasmic pattern and, in ≈ 50% of the patients, showed nucleolar staining as well. Generally, topo IIβ labelling demonstrated a low expression by immunofluorescence and was nearly undetectable on Western blots. To confirm the plasma cell identity, double staining for κ- and λ-light chains was combined with staining by the monoclonal antibody for topo IIα, Ki-S1 (Fig 7G, H, J and K).
When the nuclear/cytoplasmic ratios of topoisomerases were examined in the malignant plasma cells from 10 myeloma patients by confocal laser scanning microscopy, only topo IIα was present with ratios below 1 in the majority of cells (cytoplasmic-predominant ratio). The relationship of the overall protein content of topo IIα in malignant plasma cells and their nuclear/cytoplasmic ratios clearly showed that only cells with a considerable absolute amount of the enzyme had high ratios, while cells with cytoplasmic-predominant ratios (below 1) generally contained 15–20 times less topo IIα (Fig 8A and B). This led to a misrepresentation on Western blots for topo IIα separated into nuclear and cytoplasmic compartments (Fig 8C), in the sense that the contribution of cytoplasmic topo IIα was underestimated. A cytoplasmic distribution of topo IIα was the dominant type of localization for nearly 85% of the malignant plasma cells, whereas the nuclear distribution was representative of the proliferative fraction of the malignancy.
It is a well-established phenomenon that, when cell lines become confluent and maintain extensive cell–cell interactions, they become intrinsically drug resistant ( Croix & Kerbel, 1997). How cell adhesion controls cell growth, apoptotic signals, cell cycle progression ( Sechler & Schwarzbauer, 1998) and drug resistance remains to be elucidated. As many of the anti-tumour drugs examined are topo II poisons, the distribution of topo II (the nuclear target) is of clinical importance. There is controversy between investigators regarding the location ( Hirano & Mitchison, 1993) of topo II isoforms, as this depends on technical details and possible labelling artifacts ( Sumner, 1996). Taken together, the cellular adherence status could well affect cell cycle progression, the redistribution of cell cycle-dependent pro-teins ( Taagepera et al, 1998 ), as well as the nuclear import, export ( Nigg, 1997) and shuttling of cell cycle-controlling proteins. This may involve critical targets of anti-tumour drugs, including topo I and topo IIα.
One mechanism of drug resistance to topo II inhibitors that has been demonstrated in vitro is an attenuation of topo IIα content ( Valkov & Sullivan, 1997), caused either by decreased transcript levels ( Isaacs et al, 1996 ; Furukawa et al, 1998 ) or by enhanced degradation ( Nakajima et al, 1996 ). This may be a common in vivo event. Genetic changes, such as allele inactivation, chromosome deletions, amplification of genetic suppressor elements ( Gudkov et al, 1993 ) and point mutations, have been described for cell lines, but are rarely detected in patient samples. One interesting point mutation has been described in the topo IIα gene that leads to the formation of a stop codon and truncation of the C-terminal part of the enzyme, thus eliminating one critical nuclear localization signal and relegating topo IIα to the cytoplasm in the HL-60 cell line ( Harker et al, 1995 ). Other epigenetic mechanisms that may be involved in drug resistance to topo II inhibitors are the phosphorylation status of topo II ( Dang et al, 1994 ; Ritke et al, 1995 ) and modulating interactions with p53 ( Hochhauser et al, 1999 ).
In the present study, the expression and distribution of topo IIα and β in non-transformed cell lines and in leukaemic cell lines was examined, as was the corresponding drug sensitivity to topo II inhibitors as a function of cell density. We realize that the plateau density of cell cultures is far below the cell density existing in the tissues and biological fluids of patients. However, we consider that the changes that occur in the transition from the proliferative state in log culture to quiescent plateau densities are a reasonable model system to study the molecular mechanisms of acquiring drug resistance without drug selection pressure. Topo IIα was detected by Western blot analyses in the cytoplasmic fraction of plateau density leukaemic cell lines. We do not consider this finding an exception or a phenomenon attributable to tumour cells, as the same observation has been made regarding topo II location in Drosophila living embryos, suggesting that distinct subpopulations of the enzyme ( Swedlow et al, 1993 ) exist in the cytoplasm.
The present study confirms and explains the possible mechanism of cytoplasmic localization of topo IIα observed previously in tissue sections of clinical material ( Turley et al, 1997 ). We have obtained several bone marrow samples from pretreated myeloma patients and frequently found topo IIα in the cytoplasm of the malignant plasma cells, in contrast to the normal myeloid precursors. The fraction of proliferative plasma cells (presumably in S- or G2/M-phase) was ≈ 10–15% of all the malignant cells, and these displayed a prominent nuclear topo IIα staining. The amount of topo I in these clinical samples was studied by Western blot and found to be greater than the amount of topo IIα, relative to CCRF cell controls. In addition, topo I was localized to the nucleus in 9/10 patients examined. Topo IIβ was distributed evenly between the cytoplasm and the nucleus, but its total amount was low relative to topo I and topo IIα. These preliminary observations are in agreement with the acquired drug resistance of myeloma patients to chemotherapy, and suggest that the trafficking of topo I and topo IIα may be a relevant clinical issue.
The cytoplasmic localization of topo IIα has thus far been attributed only to C-terminally truncated proteins that have lost a critical nuclear localization signal either through deletion or through alternative splicing ( Wessel et al, 1997 ). We have now described a ‘natural’ mechanism of trafficking of topo IIα that is dependent on cell density and is probably triggered by increased cell–cell contact. It is difficult to speculate on the biological reason or cellular benefit in placing in an ‘unusual’ compartment a typical nuclear protein intimately involved in DNA metabolism ( Smalheiser, 1996). In this respect, the leukaemic cell lines were similar and obviously differed from the non-transformed lines, which downregulated and degraded topo IIα. The existence of a cytoplasmic fraction of topo IIα may also lead to drug resistance by binding topo II-directed drugs to the target protein in a compartment in which no cytocidal effect could occur because of a lack of genomic DNA. The formation of etoposide * topoisomerase II competent complexes in the absence of DNA has been described already ( Burden et al, 1996 ).
The authors thank Walter Trudeau and Dr Mayer Fishman for critical discussion and technical support. This work was supported in part by National Institutes of Health Grant CA59747, CA82050 and Moffitt Research Funds to D.M.S.