Induction of MDR1 gene expression by anthracycline analogues in a human drug resistant leukaemia cell line

The effects of 4-demethoxydaunorubicin (idarubicin, IDA) and MX2, a new morpholino-anthracycline, on up-regulation of the MDR1 gene in the low-level multidrug resistant (MDR) cell line CEM/A7R were compared at similar concentrations (IC10, IC50and IC90) over a short time exposure (4 and 24 h). The chemosensitivity of each drug was determined by a 3-day cell growth inhibition assay. Compared with epirubicin (EPI), IDA and MX2 were 17- and eightfold more effective in the CEM/A7R line respectively. No cross-resistance to 5-FU was seen in the CEM/A7R line. Verapamil (5 μM) and PSC 833 (1 μM), which dramatically reversed resistance to EPI in the CEM/A7R line, had no sensitizing effect on the resistance of this line to MX2, but slightly decreased resistance to IDA. The sensitivity to 5-FU was unchanged by these modulators. The induction of MDR1 mRNA expression by IDA, MX2 and 5-FU was analysed by Northern blotting and semiquantitatively assessed by scanning Northern blots on a phosphorimager. The relative level of MDR1 expression was expressed as a ratio of MDR1 mRNA to the internal RNA control glyceraldehyde-3-phosphate dehydrogenase (GAPDH). IDA, MX2 and 5-FU differentially up-regulated MDR1 mRNA in the CEM/A7R line in a dose-dependent manner. Both IDA and MX2 induced MDR1 expression within 4 h. 5-FU up-regulated MDR1 expression only when drug exposure was prolonged to 24 h. Based on MRK 16 binding, flow cytometric analysis of P-glycoprotein (Pgp) expression paralleled the increase in MDR1 mRNA levels. For the three anthracyclines, the increase in MDR1 expression was stable in cells grown in the absence of drug for more than 3 weeks after drug treatment. The induction of MDR1 expression by 5-FU was transient, associated with a rapid decrease in the increased Pgp levels which returned to baseline 72 h after the removal of 5-FU. This study demonstrates that MDR1 expression can be induced by analogues of anthracyclies not pumped by Pgp, and that this induction appears to be stable despite a 3-week drug-free period. © 1999 Cancer Research Campaign

Drug resistance is a common problem in acute leukaemia (Mickenna et al, 1997). Patients often relapse with unresponsive disease after an initial response to treatment with cytotoxic drugs (Rothenberg et al, 1989). The expression of P-glycoprotein (Pgp) encoded by the MDR1 gene (Deuchars and Ling, 1989), is a wellknown mechanism of multidrug resistance in relapsed acute leukaemia (Shustik et al, 1995;Bosch and Croop, 1996). The level of MDR1 gene expression in acute leukaemia is commonly increased after chemotherapy (Grogan et al, 1993;Nooter and Sonneveld, 1994). In addition, there appears to be a direct correlation between expression of the MDR1 gene de novo and outcome (survival) in this disease (Poeta et al, 1996).
Although the mechanisms underlying the effect of cytotoxic drugs on MDR1 gene expression and in turn the multidrug resistant (MDR) phenotype remain poorly understood, the acquisition of Pgp-mediated drug resistance during chemotherapy is usually thought to be due to the selection of drug-resistant cells (Chaudhary and Roninson, 1993;Gekeler et al, 1994;Manzano et al, 1996). Sikic and colleagues have clearly demonstrated this in a cell line model (Beketic-Oreskovic et al, 1994;Chen et al, 1994;Dumontet et al, 1996), under conditions in which up-regulation of Pgp was prevented. Although several attempts have been made to demonstrate that Pgp expression can be up-regulated, these studies have generally required long exposure times (72 h or more). Consequently, the selection of Pgp-expressing cells could not be ruled out from these studies (Manzano et al, 1996).
However, we have demonstrated that a rapid induction of the MDR1 gene can occur after the exposure of cells to anthracyclines in a human MDR cell line (CEM/A7R) known to express lowlevel MDR1 mRNA and P-glycoprotein (Hu et al, 1995). The induction of MDR1 was dose and time related and correlated with an increase in Pgp expression and drug resistance. The rapid increase in MDR1 gene expression after exposure to daunorubicin or epirubicin strongly supported the previous findings that the MDR1 gene promoter was activated by anti-cancer agents (Kohno et al, 1989). Similar effects have been reported in rodent cell lines (Chin et al, 1990;Fardel et al, 1997). This induction of MDR1 gene expression may have an important role in our understanding and treatment of drug-resistant tumours, as it can be prevented by cyclosporin A (CyA) and its analogue PSC 833 (Hu et al, 1996).
Alternative approaches designed to overcome the MDR phenotype include the use of new lipid-soluble compounds that are not substrates for Pgp (De Vries et al, 1990;Michieli et al, 1993). However, the effect of these agents on MDR1 gene expression is unknown. The present study was designed to investigate whether two new lipid-soluble anthracyclines, 4-demethoxydaunorubicin (idarubicin, IDA) and MX2, a new morpholino-anthracycline, Induction of MDR1 gene expression by anthracycline analogues in a human drug resistant leukaemia cell line XF Hu 1 , A Slater 1 , D Rischin 1 , P Kantharidis 1 , JD Parkin 2 and J Zalcberg 1 were able to up-regulate MDR1 gene expression compared with the classic anthracycline epirubicin (EPI).

Materials
EPI and IDA and verapamil were obtained commercially from Farmitalia (Melbourne, Australia). MX2 was a gift from Kirin Brewery Company (Japan). Verapamil was dissolved in 0.9% saline solution. PSC 833 was obtained from Sandoz Pharma (Basel, Swizerland) and initially dissolved in absolute alcohol before being diluted in RPMI-1640 to give a stock solution of 0.5 mg ml -1 (the final ethanol concentration was 35%). RPMI-1640 was purchased as a powder (Gibco) and supplemented with 10% fetal calf serum (FCS) (Trace Biosciences, Melbourne, Australia), gentamicin (80 µg ml -1 ), minocycline (1 µg ml -1 ), Hepes (20 mM), sodium bicarbonate (0.21%) and glutamine (0.8 mM). Monoclonal antibody (mAb) MRK 16 to Pgp was generously provided by Dr Takashi Tsuruo (Division of Experimental Chemotherapy, Japanese Foundation for Cancer Research). A fluorescein-labelled F(ab) 2 fragment of sheep anti-mouse IgG was purchased from Silenus Laboratories (Melbourne, Australia). The cDNA probe pHDR5A was a gift from Dr M Gottesman and Dr Ira Pastan (Laboratory of Molecular Biology, National Institutes of Health, Bethesda, MD, USA). The cDNA glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was a gift from Dr Mark Ross (Ludwig Institute for Cancer Research, Melbourne, Australia). Propidium iodide was purchased from Sigma Chemical (St Louis, MO, USA).

Cell lines and drug treatment
The studies were carried out in a variant human T-cell leukaemia MDR cell line, CEM/A7R. This line was derived from a classic MDR cell line CEM/A7, selected for low-level doxorubicin (DOX) resistance by stepwise selection of the parental line, CCRF-CEM, cultured in increasing concentrations of DOX (Zalcberg et al, 1994). The resistant line CEM/A7 was maintained in conditioned medium containing 0.07 µg ml -1 of DOX. The variant line (now stable for over 3 years) was established by culturing the CEM/A7 cells in the absence of DOX before being subcloned and designated as the CEM/A7R line. This line was not exposed to DOX or other Pgp substrates except in the specific experiments detailed below. At the time of these experiments, all lines were mycoplasma-free based on the mycoplasma TC Rapid kit (Gen-Probe, San Diego, CA, USA).
The CEM/A7R cells in exponential growth phase were collected 2 days after subculture. The cells were washed, counted and resuspended in 20 ml of fresh medium to a total cell number of 5 × 10 6 -1 × 10 7 . The cells were treated with various concentrations of EPI, IDA and MX2 for 4-24 h, then harvested and washed three times with cold phosphate-buffered saline (PBS) for the analysis of MDR1 gene expression. Cell viability was determined (after staining by trypan blue) using phase-contrast microscopy to detect cells of abnormal size or granularity. Non-viable cells were excluded from flow cytometric analysis by propidium iodide staining.

Growth assays
The sensitivity of each of the cell lines to a variety of chemotherapeutic drugs was determined by a standard growth inhibition assay (Tsuruo et al, 1981). Briefly, after determining cell viability, 2 × 10 5 of the tested cells were exposed to varying concentrations of each tested drug in the presence or absence of verapamil (Vp) or PSC 833 in 12-well plates. The cells were incubated at 37°C in a humidified chamber containing 5% carbon dioxide in air for 3 days and counted using an automated coulter counter (Hu et al, 1990a). Results are expressed as the increase in cell number of drug-exposed cells as a percentage of the increase in untreated control cells. The IC 10 , IC 50 and IC 90 for each drug were determined by calculating the drug concentration required to inhibit cell growth by 10%, 50% and 90%. Relative resistance represents the ratio of the IC 50 of the resistant CEM/A7R cell line compared with the parental CCRF-CEM cell line.

RNA extraction and Northern blot analysis
RNA was isolated by the guanidinium thiocyanate method described by Chomczynski and Sacchi (1987). Twenty micrograms of total cellular RNA was size fractionated on a 1.5% agarose gel containing 2.2 M formaldehyde and transferred onto nylon filters (Hybond-N, Amersham, UK) for MDR1 hybridization. The filters were probed with the plasmid pHDR5A containing a 1.4-kb MDR1 cDNA (Ueda et al, 1987) and then reprobed with a 32 P-labelled GAPDH cDNA for normalization. The pHDR5A probe predominantly recognizes the MDR1 gene under the high stringency conditions used in this study. The filters were prehybridized overnight at 42°C in hybridization buffer containing 50% formamide, 5 × SSPE (1 × SSPE containing 0.15 M sodium chloride, 0.001 M sodium dihydrogen phosphate and 0.001 M EDTA), 5 × Denhardt's solution, 0.5% sodium dodecyl sulphate (SDS) and 1% skimmed milk powder. Hybridization was carried out in the same buffer. The pHDR5A and GAPDH cDNAs were randomly primed with [α-32 P]-dCTP. Labelled cDNA (10 6 c.p.m.) was added to each millilitre of hybridization buffer. The filters were washed sequentially in 2 × SSPE with 0.1% SDS at 42°C for 15 min, 1 × SSPE with 0.1% SDS at 65°C for 30 min and finally 0.1 × SSPE with 0.1% SDS at room temperature for 15 min. The filters were then exposed to radiographic film at -70°C using intensifying screens, or radioactive signals were quantitated by scanning on a phosphorimager using Image Quant software (Molecular Dynamics, Melbourne, Australia).

Pgp expression
Flow cytometry was used to measure Pgp expression. Cells were collected and washed three times in medium containing 10% FCS. MRK16, a mAb to an external epitope of Pgp (final concentration 10 µg ml -1 ), was added to cells at room temperature (RT) for 20 min. A non-specific murine mAb (IgG 2a , Becton Dickinson, Sydney, Australia) was used as the isotype control. After a further three washes, cell pellets were resuspended in the same volume of phosphate-buffered saline (PBS) containing 10 µl of a 1:10 dilution of a fluorescein-conjugated F(ab′) 2 fragment of sheep antimouse IgG antibody (Silenus Laboratories) for 20 min at RT in the dark. Cells were washed once again (×3) and then analysed in a FACScan flow cytometer (Becton Dickinson). Mean fluorescence intensity was recorded for each tested population (after correcting for non-specific binding) to provide an estimate of relative MRK-16 binding.   (Hu et al, 1990b). Cells were washed three times with PBS and resuspended in fresh medium for 1 h at 37°C in a humidifier before performing drug accumulation studies. The cells were adjusted to a concentration of 5 × 10 6 ml -1 and viability assessed with trypan blue. Cells were added to 96-well plates to give a final number of 5 × 10 5 cells per well and incubated at 37°C with tracer amounts of [ 3 H]daunomycin (final concentration 1.85 × 10 4 Bq ml -1 , 0.05 µg ml -1 ). The cells were harvested onto glass-fibre filters at designated times with an automated cell harvester (Cambridge Technology). The filter papers were dried and dissolved in 5 ml of a liquid scintillation cocktail (Ultima Cold, Packard) before radioactivity was measured. All assays were performed in triplicate.

Statistics
Analysis of variance was used to compare intracellular levels of [ 3 H]daunomycin after exposure of cells to various experimental conditions.

Cytotoxicity of IDA, MX2 and 5-FU
The relative cytotoxicity of the two lipid-soluble anthracyclines IDA and MX2 was compared with the classic anthracycline epirubicin (EPI), as well as the unrelated drug 5-FU -a drug not pumped by Pgp. The IC 10 , IC 50 and IC 90 for each drug were determined in a 3-day growth inhibition assay (Table 1) in the drug-resistant CEM/A7R and the parental drug-sensitive line CCRF-CEM. On a weight basis, IDA and MX2 were 17-and eightfold more active than EPI in the drug-resistant line CEM/A7R and 11-and fivefold more active than EPI in the parental line CCRF-CEM, respectively, when comparing doses that inhibited cell growth by 50% (IC 50 ). Relative to the drug-sensitive line, there was a fivefold increase in resistance to EPI and a threefold increase in the resistance to IDA and MX2 in the drug-resistant line. As expected, no increase in the resistance to 5-FU was observed in the CEM/A7R line compared with the parental line (Table 1).
PSC 833 (0.1 µM) dramatically reversed the resistance of the CEM/A7R line to EPI ( Figure 1A). However, the addition of 5 µM verapamil (data not shown) or 1 µM PSC 833 to the culture medium only slightly increased the sensitivity of this line to IDA ( Figure 1A), and had no detectable effect on the sensitivity of the line to MX2 ( Figure 1B). Neither modulator had any effect on the sensitivity of the MDR line to 5-FU ( Figure 1B).

Gene expression
The effects of EPI, IDA, MX2 or 5-FU on the induction of the MDR1 gene in the CEM/A7R line were compared at their respective IC 10 , IC 50 and IC 90 concentrations over a 4-or 24-h period of drug exposure. All samples collected for the analysis of MDR1 gene expression were harvested immediately after drug treatment. The up-regulation of MDR1 gene expression by each drug was semiquantitatively assessed by scanning Northern blots on a phosphorimager.
We had previously demonstrated that 1.5 µg ml -1 of EPI (IC 90 concentration) induced the expression of MDR1 mRNA as early as 4 h after such exposure. In the present study, the effect of EPI on the induction of the MDR1 gene was examined at the IC 10 and IC 90 concentrations over a short time period (4-6 h). Up-regulation of MDR1 mRNA was observed over 4-6 h with a two-to fivefold increase in MDR1 mRNA noted at 0.1 (IC 10 ) and 1.0 (IC 90 ) µg ml -1 of EPI respectively (Figure 2A and B).
The impact of IDA and MX2 on the induction of MDR1 gene expression was examined at equieffective concentrations over 4 h. Both drugs induced a similar increase in MDR1 mRNA levels in a dose-dependent manner (Figure 3). A two-to threefold increase in MDR1 mRNA was observed at the IC 10 (0.002, 0.02 µg ml -1 ) and IC 50 (0.02, 0.05 µg ml -1 ) concentrations of IDA and MX2, and a fivefold increase at the IC 90 concentrations (0.08, 0.1 µg ml -1 ) of

Figure 3
Effect of IDA or MX2 on induction of MDR1 gene expression in the CEM/A7R cells after a 4-h exposure. CEM/A7R cells in exponential growth phase were treated with and without IDA at the IC 10 (0.002 µg ml -1 ), IC 50 (0.02 µg ml -1 ), and IC 90 (0.08 µg ml -1 ) levels or MX2 at IC 10 (0.02 µg ml -1 ) IC 50 (0.05 µg ml -1 ), IC 90 (0.1 µg ml -1 ) for 4 h. RNA was then extracted and subjected to Northern blot analysis and scanned on a phosphorimager as described in the Materials and methods section. Results are expressed as the ratio of MDR1 to GAPDH RNA after signals were normalized with respect to the CCRF-CEM signal. Columns, means of triplicate analyses of Northern blots; bars, s.d. These experiments were repeated on three separate occasions with similar findings. The data presented are representative of one such experiment Figure 4 The effects of 5-FU on the induction of the MDR1 gene in the CEM/A7R line over a 24-h exposure. Cells in exponential growth phase were treated with (or without) 5-FU at the IC 10 (0.2 µg ml -1 ), IC 50 (0.5 µg ml -1 ), IC 90 (2 µg ml -1 ) or IDA at the IC 50 level (0.02 µg ml -1 ) as a comparison. The 32 Plabelled MDR1 and GAPDH bands were scanned on a phosphorimager using Image Quant Software (Molecular Dynamics) and the results are expressed as the ratio of MDR1 to GAPDH RNA after signals were normalized with respect to the CCRF-CEM signal. Columns, means of triplicate analyses of Northern blots; bars, s.d. Similar results were seen in repeated experiment IDA and MX2 respectively (Figure 3). When the exposure time was prolonged up to 24 h, a fivefold increase of MDR1 mRNA was observed at the IC 50 concentrations of IDA (0.02 µg ml -1 ) or MX2 (0.05 µg ml -1 ) (data not shown). The potential induction of the MDR1 gene by 5-FU was examined at 0.2, 0.5 and 2.0 µg ml -1 (IC 10 , IC 50 and IC 90 ) levels respectively. Although no obvious increase in MDR1 expression was observed at any concentration of 5-FU after 4 h (data not shown), Northern analysis revealed induction had occurred in a dosedependent manner after 24 h (Figure 4). A threefold induction in MDR1 expression was seen at the IC 90 level (2 µg ml -1 ) compared with the fivefold induction of MDR1 at the IC 50 level (0.02 µg ml -1 ) of IDA (Figure 4).
For each of the four drugs used in these experiments, treated cells were stained with trypan blue to assess cells for visible cell damage. Careful examination using phase-contrast microscopy for change in cell morphology and/or increased granularity revealed no visible evidence of cell damage within 4-24 h at any of the drug concentrations used.

Pgp expression
To determine whether the increased levels of MDR1 mRNA expression induced by IDA, MX2 and 5-FU were accompanied by corresponding changes in Pgp levels, antigen density (mean channel fluorescence or mcf) was measured by flow cytometry. The CEM/A7R cells were exposed for 24 h to approximately the IC 90 levels of IDA, MX2 (0.1 µg ml -1 ), 5-FU (2.0 µg ml -1 ) or EPI (0.5 µg ml -1 ) before being subjected to Pgp analysis ( Figure 5). In parallel to the increase in the MDR1 mRNA levels (Figure 3), Pgp expression was increased five-to sevenfold with EPI, MX2 and IDA (mcf 36.65, 25.98 and 22.20 respectively) and threefold with 5-FU (mcf 18.74). In comparison, the mcf of untreated CEM/A7R cells was 4.2.
To examine whether the induction in MDR1 gene expression was stable, CEM/A7R cells were treated overnight with 0.1 µg ml -1 MX2 (IC 90 ) or 2.0 µg ml -1 5-FU (IC 90 ). The drug treated cells were then washed three times and allowed to grow in drug-free medium for either 72 h or 3 weeks before being subjected to a further analysis of MDR1 mRNA levels (data not shown) or Pgp expression. In the MX2-and IDA-treated CEM/A7R cells, Pgp expression remained two-to threefold higher (mcf 13.50 and 11.09) than the untreated CEM/A7R cells (mcf 4.66) 3 weeks after drug exposure. In contrast, the increased Pgp levels in cells treated with 5-FU returned to base line (mcf 4.69) within 72 h of stopping drug treatment (data not shown).

Figure 7
The chemosensitivity of untreated (q q-q q) or treated CEM/A7R cells, 3 weeks after an overnight exposure to 0.5 µg ml -1 of EPI (v-v), 0.1 µg ml -1 of IDA (q-q) and MX2 (s s-s s), and 2 µg ml -1 of 5-FU (v v-v v) (0.1 µg ml -1 ) or 5-FU (2.0 µg ml -1 ), the cells were maintained in drug-free medium for 3 weeks before drug accumulation was quantified. Compared with the untreated CEM/A7R cells, the intracellular levels of [ 3 H]daunomycin were significantly lower in the EPI-, IDA-and MX2-treated CEM/A7R cells, but were unchanged in the 5-FU-treated cells ( Figure 6). Growth inhibition assays also showed a two-to threefold increase in resistance to EPI in MX2-, IDA-and EPI-treated cells after 3 weeks in drugfree medium. In contrast, the drug sensitivity of 5-FU-treated cells was the same as that of untreated CEM/A7R cells (Figure 7).

DISCUSSION
In this report, we investigated whether the two lipid-soluble anthracycline analogues IDA and MX2 could up-regulate MDR1 gene expression in a human MDR, T-cell leukaemia line, CEM/A7R, and its parental line, CCRF-CEM. Both drugs are thought to be poor substrates for Pgp-mediated transport (de Vries and Zijlstra, 1990;Ross et al, 1995). Initially, we compared the chemosensitivity (in the presence or absence of Vp or PSC 833) of each drug in the cell lines. Verapamil (5 µM) and PSC 833 (1 µM) dramatically decreased the resistance of the CEM/A7R line to EPI, but had no effect on the sensitivity of this drug-resistant line to MX2 and only slightly increased the sensitivity to IDA (Figure 1). These findings are consistent with the observation that both anthracycline analogues are highly lipophilic and diffuse rapidly through the cell membrane to bind to DNA, thereby interfering with the capacity of cells expressing Pgp to pump these drugs (Horichi et al, 1990;Watanabe et al, 1991;Berman and McBride, 1992) and thus overcoming drug resistance. Although IDA and MX2 are less susceptible to this mechanism of resistance compared with the drug-sensitive line, detectable levels of resistance to IDA and MX2 were still observed in the resistant CEM/A7R cells (Table 1).
The effects of IDA, MX2 and 5-FU on the induction of MDR1 gene expression were investigated at equieffective concentrations with respect to the inhibition of cell growth in a 72-h cell growth inhibition assay. When MDR1 mRNA expression was analysed by Northern blotting, a two-to fivefold increase was observed in the CEM/A7R cells 4 h after treatment with IDA, MX2 or EPI at concentrations ranging from the IC 10 to IC 90 level. In contrast, 5-FU failed to up-regulate MDR1 expression after a 4-h exposure, although induction of MDR1 gene expression was seen after a 24h exposure. The effects of these drugs on MDR1 gene expression was confirmed by flow analysis, demonstrating a concomitant increase in Pgp expression. Up-regulation of Pgp expression by IDA and MX2 was stable, as demonstrated by the increase in Pgp levels, drug resistance and decrease in drug accumulation in the IDA-and MX2-treated CEM/A7R cells 3 weeks after the removal of these agents from the culture medium (Figures 5-7).
Other investigators have reported the up-regulation of MDR1 promoter activity in reporter gene assays by cytotoxic drugs (Kohno et al, 1989) as well as many other factors (Rohlff and Glazer, 1995). These findings have been controversial with respect to their relevance in the clinical context, as the endogenous promoter does not always behave in an analogous manner to the transfected promoter. For example, Tanimura et al (1992) and Ferrandis and Benard (1993) have demonstrated that a transfected promoter is active in drug-sensitive cells in which the endogenous promoter is not active.
Increased MDR1 mRNA and Pgp expression have also been reported in human cell lines after their exposure to cytotoxic drugs (Chaudhary and Roninson, 1993;Gekeler et al, 1994). However, the possibility that selection had occurred could not be excluded (reviewed in Manzano et al, 1996), as drug-induced MDR1 expression usually occurred after a 72-h exposure to sublethal concentrations of cytotoxics and was associated with the appearance of visible morphological cell damage.
The present study was designed to examine the role of the two anthracycline analogues IDA and MX2 on the up-regulation of MDR1 gene expression for a range of concentrations (IC 10 , IC 50 and IC 90 levels) within 4 and 24 h. In our study, the increase in MDR1 expression was not related to the appearance of cytological damage and was seen at concentrations as low as the IC 10 level for each drug.
Up-regulation of MDR1 gene expression has mainly been observed for drugs which are substrates for Pgp, although one study demonstrated that a non-Pgp substrate, cytarabine, was able to induce MDR1 expression (after a 72-h exposure) which was stable after withdrawal of the drug for more than 6 weeks (Chaudhary and Roninson, 1993). In this report, we observed that 5-FU, also not a substrate for Pgp, could up-regulate MDR1 mRNA levels. However, this appeared to be a transient increase because the level of Pgp returned to baseline within 72 h of removal of the drug. These findings suggest that the mechanism by which 5-FU acts probably differs from that used by other anthracyclines which result in a sustained increase in Pgp expression.
It is likely that the cellular response to cytotoxic stress is to activate a number of defence mechanisms via stress response pathways (Osborn and Chambers, 1996). The up-regulation of Pgp expression may represent only one of the results of this response to these stress mechanisms. However, this finding would not be expected to confer resistance to IDA or other drugs that are not pumped by Pgp (Consul et al, 1996;Hargrave et al, 1995). Thus, the up-regulation of Pgp expression in response to such drugs does not necessarily confer a survival advantage. In fact, it is generally accepted that the treatment of relapsed refractory leukaemia with idarubicin/cytarabine is associated with higher rates of complete remission and longer survival compared with daunorubicin/cytarabine (Carella et al, 1993;Berman, 1993).
The mechanism by which these lipid-soluble anthracyclines, or indeed the classic Pgp substrate EPI, up-regulate Pgp expression is not known. The fact that induction can occur rapidly (in 4 h) strongly suggests that up-regulation involves transcription of the MDR1 gene, perhaps involving stress response pathways. Osborn and Chambers (1996) demonstrated that a c-jun NH 2 -terminal protein kinase (JNK), a member of the mitogen-activated protein kinase family, is activated by a variety of stressful stimuli, including exposure to cytotoxic agents. In their system, exposure to doxorubicin resulted in increased JNK activity and increased MDR1 mRNA levels. Rohlff and Glazer (1995) suggested a number of signalling pathways may act on MDR1 gene expression. Which, if any, of these pathways are involved in the induction response observed in our line is the focus of further work in our laboratory.