Growth inhibition, cytokinesis failure and apoptosis of ... - Nature

Leukemia (1999) 13, 768–778  1999 Stockton Press All rights reserved 0887-6924/99 $12.00 http://www.stockton-press.co.uk/leu

Growth inhibition, cytokinesis failure and apoptosis of multidrug-resistant leukemia cells after treatment with P-glycoprotein inhibitory agents G Lehne1,2, P De Angelis3, M den Boer4 and HE Rugstad1 1

Department of Clinical Pharmacology, 2Institute for Surgical Research, and 3Institute for Pathology, The National Hospital, Rikshospitalet, University of Oslo, Norway; and 4Department of Pediatrics, Free University Hospital, Amsterdam, The Netherlands

The multidrug transporter P-glycoprotein (Pgp), which is frequently overexpressed in multidrug resistant leukemia, has many proposed physiological functions including involvement in transmembraneous transport of certain growth-regulating cytokines. Therefore, we studied cell growth of three pairs of drug resistant and sensitive leukemia cell lines (KG1a, K562 and HL60) exposed to three different inhibitors of Pgp. The resistant KG1a and K562 sublines, which expressed high levels of Pgp, responded to low doses of the cyclosporin SDZ PSC 833, the cyclopeptolide SDZ 280–446, and the cyclopropyldibenzosuberane LY335979 with a dose-dependent growth inhibition. In the resistant variants of KG1a and K562 cells the mean half-maximal growth inhibitory doses (GI50) of SDZ PSC 833 were 312 (SE 41) and 414 (SE 50) nM, those of SDZ 280– 446 were 685 (SE 51) and 578 (SE 54) nM, and those of LY335979 were 66 (SE 1) and 48 (SE 8) nM, respectively. Exposure to 1 ␮M SDZ PSC 833 resulted in tetraploidization, cytokinesis failure and apoptosis of the KG1a and K562 MDR variants. Conversely, parental cells with no or low levels of Pgp and the non-Pgp resistant variant of HL60 cells were not receptive to these cytotoxic effects. We conclude that inhibition of Pgp may exercise selective cytotoxicity in Pgp-rich leukemia cells indicating a possible therapeutic target in multiresistant leukemia. Keywords: P-glycoprotein; apoptosis; leukemia; resistance modulator

Introduction Failure to achieve complete and durable responses from cancer chemotherapy is a common clinical problem that limits the curative potential of anti-cancer drugs in clinical oncology. Multidrug resistance (MDR) is believed to be a major cause for treatment failure, and is frequently associated with overexpression of the multidrug transporter, P-glycoprotein (Pgp), which is an integral plasma membrane protein capable of drug expulsion and maintenance of tolerable intracellular levels of certain cytotoxic drugs.1 Several clinically important anti-cancer drugs may be removed from neoplastic cells by Pgp-mediated transport, despite the diversity in chemical structures and mechanisms of action.2 This ability is apparently reflected in a negative correlation between Pgp expression and chemosensitivity or survival in leukemias,3,4 lymphomas5 and some solid tumors.6–8 Pgp has a broad specificity for substrates, and several noncytotoxic drugs may competitively inhibit efflux of cytotoxic drugs by Pgp and thereby downmodulate MDR.9 The calcium channel blocker verapamil was the first agent that was shown to modify MDR in vivo and in vitro,10 but unfortunately the MDR modulating activity required concentrations that are associated with severe cardiac toxicity in patients.11 Therefore, a new generation of highly specific and potent Pgp Correspondence: G Lehne, Department of Clinical Pharmacology, Rikshospitalet, The National Hospital of Norway, N-0027 Oslo; Fax: 47 22 11 19 87 Received 28 September 1998; accepted 28 December 1998

inhibitors with less toxic potential have been developed including the non-immunosuppressive cyclosporin SDZ PSC 833,12 the cyclopeptolide SDZ 280–446,13 and the cyclopropyldibenzosuberane LY 335979.14 These agents exercise high affinity for Pgp and interfere with binding of known substrates such as azidopine and calcein.15–17 SDZ PSC 833 which is the most extensively studied of these, has proven to be well tolerated in phase I trials,18–20 and is currently being investigated as an adjunct to cancer chemotherapy in phase II and III trials. Although Pgp contributes to renal and hepatic elimination of toxic compounds,21 adrenal secretion of steroid hormones22 and tissue protection at the blood–brain barrier,23 the physiological role of Pgp is not fully understood. Two recent reports have shown that Pgp may also take part in the transport of growth stimulating cytokines IL-2 and IL-4.24,25 Thus, the participation of Pgp in growth regulation could represent a new target for growth control in neoplastic cells expressing Pgp. Recently, we discovered a selective cytotoxic effect of SDZ PSC 833 which engaged apoptosis (programmed cell death) in multidrug resistant leukemia cells.26 As this effect appeared to involve Pgp, the present paper addresses the effect of other modulators on both Pgp- and non-Pgp-expressing leukemia cells, and further focuses on the induction of apoptosis in relation to cell cycle progression. The cells were characterized with respect to MDR phenotype and expression of proteins involved in the promotion (p53, Bax, Apol/Fas) and suppression (Bcl-2, Bcl-xL) of apoptosis. Pgp overexpression was the only common denominator which was exclusively associated with a dose-dependent growth inhibitory response to the modulators SDZ PSC 833, SDZ 280–446, and LY335979. Furthermore, the KG1a and K562 MDR variants, which shared these characteristics, also demonstrated attributes of cytokinesis failure and apoptosis after exposure to the modulators. Materials and methods

Chemicals The human leukemia cell lines were propagated in RPMI 1640 medium with l-glutamine (BioWhittaker, Walkersville, MA, USA) supplemented with 10% fetal calf serum, streptomycin (100 ␮g/ml), penicillin (100 U/ml), nystatin (40 U/ml) and HEPES (only K562 cell lines). MRK16 was kindly provided by Professor T Tsuruo (Institute of Molecular and Cellular Biosciences, The University of Tokyo, Japan), and UIC2 was provided by Immunotech (Marseille, France). MRPr1 was a gift from Dr MJ Flens and Professor RJ Scheper (Department of Pathology, University Hospital Vrije Universiteit, Amsterdam, The Netherlands). The IgG2a isotypic control antibody MOUSE IgG2a was purchased from Monosan (Uden, The Netherlands). SDZ PSC 833 and SDZ 280–446 were supplied by Novartis (Basel, Switzerland). LY335979 and vincristine

Selective cytotoxicity of P-glycoprotein inhibitory agents G Lehne et al

were supplied by Eli Lilly & Co. (Indianapolis, IN, USA). Daunorubicin was provided by Rhoˆne Poulenc Rorer (Vitry, France).

Cells and culture conditions Multidrug-resistant human acute myelogenous leukemia cell lines were obtained by subculturing American Type Culture Collection (ATCC) lines KG1a27 and HL6028 in stepwise increased concentrations of daunorubicin and vincristine from 10 to 100 ng/ml (KG1a/200) and from 2 to 65 ng/ml (HL60/130). The parental cells were designated KG1 a/0 and HL60/0, respectively. The human chronic myelogenous K562 (ATCC) cell line29 designated K562/0, and a multistep selected MDR subline (K562/150) adapted to 150 nm vincristine30 were provided by Dr Astrid Gruber (Karolinska Hospital, Stockholm, Sweden). The cells were maintained and propagated as previously described.31 The number of doublings (n) were calculated according to the equation: n = 3.32 · log (X/X0) where X0 is the cell count at seeding time point and X is the cell count on day x of the culture.

Cell growth inhibition assay Approximately 5 × 104 cells were plated in 16 mm-diameter wells (Costar Corporation, Cambridge, MA, USA) and grown in 1 ml of drug-free medium for the first 24 h. The wells were then supplemented with the drug or combination of drugs in a certain dose range. At least three replicate cultures were made for each of the dose levels and for untreated controls. The KG1a cells were cultured for 96 h, and the K562 and HL60 cells were cultured for 72 h ensuing two to three doublings. The actual mean numbers of doublings (standard error of the mean in parenthesis) for KG1a/0 were 3.1 (SE 0.2), that for KG1a/200 were 2.7 (SE 0.2), that for K562/0 were 3.3 (SE 0.2), that for K562/150 were 3.1 (SE 0.4), that for HL60/0 were 3.3 (SE 0.1), and that for HL60/130 were 2.0 (SE 0.04). Harvested cells were counted in a Coulter Counter ZM (Coulter Electronics, Luton, UK). Cell numbers were assessed electronically with a lower threshold set at 7.5 ␮m particle diameter to discard cell debris and shrunken (apoptotic) cells. The dose level required for 50% inhibition of cell growth (GI50) was calculated from linear plots of dose vs cell number. The resistance factor (RF) was defined as the ratio between the GI50 values obtained in the resistant and sensitive cells, respectively. By co-incubation with either SDZ PSC 833, SDZ 280–446 or LY335979 modulation of growth inhibition was assessed. The modulating factor (MF) was defined as the ratio between the GI50 values of multidrug resistant cells with and without the modulating agent, respectively. Each experiment was performed at least three times.

described.31 Determination of MRP required fixation in 0.74% v/v formaldehyde in acetone for 10 s at room temperature to allow staining of a cytoplasmic epitope with MRPr1 (1.7 ␮g/ml) antibody as previously described.32 Immunofluorescence distributions were generated using a FACScan flow cytometer (Becton Dickinson, San Jose´, CA, USA) tuned to 488 nm laser excitation wavelength. FITC fluorescence of gated populations was collected through a bandpass filter (FL1, bandwidth 515–545 nm). Calculations of logarithmically amplified fluorescence values were performed in arithmetic mode using the CellQuest computer program (Becton Dickinson). Reproducibility was ensured by at least two replicate experiments.

Intracellular accumulation of daunorubicin: The cells were grown in drug-free medium for 24 h prior to flow cytometric analysis. A FACScan flow cytometer was used to generate daunorubicin fluorescence which provides excitation maximum at 490 nm and emission maximum at 592 nm.33 The fluorescence was transmitted through a bandpass filter of 564–606 nm (FL2) and logarithmically amplified. Correlated forward angle (a relative measure of cell size) and right angle (a measure of cell granularity) light scatter measurements were generated to exclude dead cells and debris from analyses. Acquired data from 10 000 events were analyzed using Cell Quest. By virtue of daunorubicin fluorescence intracellular accumulation of the drug was measured as fluorescence distributions of cell samples after 1 h incubation at 37°C with 4.4 ␮m daunorubicin alone or in the presence of SDZ PSC 833, SDZ 280–446, or LY335979 at 1 ␮m concentrations. Each experiment was performed three times.

TUNEL assay for detection of apoptotic cells: Terminal deoxynucleotidyl transferase(TdT)-mediated d-uridine-5⬘triphosphate(dUTP)-biotin nick end-labeling (TUNEL) utilizes the ability of the enzyme TdT to catalyze binding of deoxynucleotides to 3⬘-OH ends of DNA strand breaks. The method used in this study is based on the method of Gorczyca et al34 and modified as described.26 Briefly, the cells were cultured for 48 h in the presence of 1 ␮m SDZ PSC 833, SDZ 280–446 or LY335979, and then 106 cells were fixed in 1% ice-cold paraformaldehyde and subsequently in methanol at −20°C before 30 min incubation at 37°C in a TdT solution (Boehringer Mannheim, Mannheim, Germany) which included biotin-labeled dUTP. Subsequently, the cells were incubated on ice with streptavidin-fluorescein isothiocyanate (FITC) for 30 min, and finally counterstained with 5 ␮g/ml propidium iodide (PI) to measure DNA content. Analyses were performed using FACScan and FACSVantage (Becton Dickinson) flow cytometers. FITC green fluorescence represented DNA strand breaks as in apoptosis, and PI red fluorescence represented DNA content. Doublet discrimination and exclusion of debris and clumps were achieved using red fluorescence pulse-width analyses. Each experiment was performed three times.

Flow cytometry Determination of MDR phenotype: Pgp immunofluorescence was obtained in viable cells by a three-layer staining technique using MRK16 (20 ␮g/ml) and UIC2 (20 ␮g/ml) antibodies to reveal surface epitopes of Pgp as previously

APO-1/Fas expression: The IgG2a antibody BMS138 and the IgG2b antibody BMS140 (both Bender MedSystems, Vienna, Austria) were selected as primary APO-1/Fas antibodies. Cell suspensions were washed with PBS/BSA and

769


770

incubated on ice for 60 min with BMS138 (10 ␮g/ml), BMS140 (10 ␮g/ml), MOUSE IgG2a (25 ␮g/ml), or MOUSE IgG2b (25 ␮g/ml) in PBS/BSA. The second and third layer staining protocols were carried out as described for Pgp detection and followed by flow cytometry as described in ‘Determination of MDR phenotype’. Reproducibility was ensured by two replicate experiments.

Cytospin preparations Cytospin preparations were made by centrifugation of aliquots of approximately 105 cells in 300 ␮l RPMI and 300 ␮l fixative (95% methanol and 5% carbowax 1540) at 1000 r.p.m. for 5 min using Shandon Cytospin 2 (Shandon Scientific, Cheshire, UK). Automated Papanicolaou staining was performed in a Jung Autostainer XL (Leica, Germany) as previously described.26

Immunoblotting Cell lines which were to be analyzed for expression of several proteins involved in the regulation of apoptosis were prepared for immunoblotting as follows: 2–3 × 106 cells from each cell line, washed once in PBS, were boiled for 5 min in sodium dodecyl sulphate (SDS) sample buffer containing 0.1 mm phenylmethylsulfonylfluoride (PMSF). These were then cooled on ice and frozen at −20°C until use. The proteins and Kaleidoscope molecular weight standards (BioRad, Hercules, CA, USA) were first separated by electrophoresis in 8.0% or 12.0% SDS-polyacrylamide gels.35 Immunoblotting was done as described previously36 using the following antibodies: antiBax P-19 (Santa Cruz Biotechnology, Santa Cruz, CA, USA) at final concentration 2 ␮g/ml, anti-Bcl-xL S-18 (Santa Cruz Biotechnology) at 2 ␮g/ml, anti-Bcl-2 clone 124 (Dako, Oslo, Norway) at 5 ␮g/ml, anti-p53 Pab 1801 (Calbiochem, Cambridge, MA, USA) at 2.5 ␮g/ml and anti-p-83 34C1 (kindly provided by Dr T Stokke, The Norwegian Radium Hospital, Oslo, Norway) at 1:30 dilution of stock. The amount of p83 was determined on the same blots as a control for gel loading and cell concentration as described previously.36 An amplified biotin-streptavidin alkaline phosphatase staining procedure (BioRad) was used to detect primary antibodies. Reproducibility was ensured by two replicate experiments.

Statistics Comparisons of mean values were performed by two-tailed unpaired t tests using Prism software (GraphPad, San Diego, CA, USA). P values ⬍0.05 were regarded as statistically significant. Summary statistics were reported as means provided with standard errors of the mean (SE) or 95% confidence intervals (CI). Histogram comparisons were performed by Kolmogorov–Smirnov statistics37 to generate D values ranging between 0 and 1.0 according to the difference between two distributions (higher values indicate greater difference). D values ⭓0.15 were regarded as significant differences between two histograms.38

Results

MDR phenotype The MDR phenotypes were determined by flow cytometric immunofluorescence detection using anti-Pgp (MRK16, UIC2) and anti-MRP (MRPr1) antibodies. The degree of protein expression was determined by the expression factor (EF), which was defined as the ratio between mean fluorescence values for relevant (numerator) and irrelevant (denominator) antibody. To distinguish between wild-type and MDR phenotype a discrimination factor (DF) was defined as the ratio between the EF values for resistant (numerator) and parental (denominator) cells. Using MRK16 the KG1a/200 and K562/150 cells were highly positive for Pgp with EF values of 65.5 and 71.9, respectively. The parental KG1a/0 cells were slightly Pgp-positive (EF = 8.8) and the parental K562/0 cells were completely negative (EF = 0.7). Both the HL60/0 and HL60/130 cells expressed negligible levels of Pgp (EF = 2.9 and 1.6, respectively). In comparison with parental cells the resistant KG1a and K562 variants overexpressed Pgp multifold (DF = 7 and 103, respectively), whereas Pgp expression was not increased in the resistant HL60 variant (DF = 0.6). These results were reproduced using UIC2 which resulted in corresponding DF values (5, 69 and 0.5, respectively). Overexpression of MRP occurred only in the resistant HL60 subline (DF = 5) using MRPr1. Only trace amounts of MRP were detected in the other leukemia cell lines of this study.

Pgp function Cytotoxic daunorubicin is a known substrate of both Pgp39 and MRP.40 Therefore, we assessed Pgp or MRP function in terms of cellular accumulation of daunorubicin using flow cytometry and fluorescence detection in cells that were incubated for 60 min at 37°C with 4.4 ␮m of the drug. Accumulation of daunorubicin was restricted in all the resistant sublines compared to the corresponding wild types (P ⭐ 0.02). By combining data from triplicate experiments with concurrent exposure to 1 ␮m of SDZ PSC 833, SDZ 280–446 or LY335979, none of which were fluorescent, Pgp function and modulator activity were demonstrated (P ⬍ 0.0001). The combined fluorescence values (mean channel fluorescence) of KG1a/200 cells increased from 10.5 (SE 0.03) to 38.5 (SE 1.1), that of KG562/150 cells from 18.2 (SE 0.5) to 37.7 (SE 0.4), and that of parental KG1a/0 cells from 17.4 (SE 0.9) to 27.9 (SE 0.7). The minor modulator activity in the KG1a wild type was due to a low constitutive Pgp expression as previously demonstrated. SDZ PSC 833 and SDZ 280–446 contributed equally to enhance the accumulation of daunorubicin in MRP expressing HL60/130 cells as treatment with these modulators increased the combined fluorescence values from 38.8 (SE 1.9) to 64.1 (SE 3.5) (P = 0.002). Conversely, LY 335979 did not contribute to any change in daunorubicin content of cells with no or very low levels of Pgp as demonstrated by stable fluorescence values of 33.9 (SE 0.8) and 34.4 (SE 1.1) in K562/0 cells (NS), of 85.4 (SE 5.2) and 79.4 (SE 6.8) in HL60/0 cells (NS), and of 38.8 (SE 1.9) and 32.9 (SE 3.4) in HL60/130 cells (NS). These findings indicate LY335979 exclusively blocks the function of Pgp, whereas SDZ PSC 833 and SDZ 280–446 block the function of both Pgp and MRP.


Cytotoxic effect of daunorubicin A measure of cytotoxicity was provided by determination of the drug dose that reduced the cell numbers in continuously growing cultures by 50% (GI50). There was a 3-, 13- and 17fold relative resistance to daunorubicin in the resistant sublines of KG1a, K562 and HL60 cells, respectively. Low-dose treatment with SDZ PSC 833 (120 and 210 nm), SDZ 280–446 (420 and 340 nm), or LY 335979 (40 nm) sensitized the MDR variants KG1a/200 and K562/150 6.5-fold (SE 0.3) and 3.6fold (SE 0.5), respectively, to the cytotoxic effect of daunorubicin. This substantial circumvention of MDR was in agreement with the highly significant increment of daunorubicin accumulation in these cells following combined treatment with anthracycline and modulator. In the MRP positive HL60/130 cells SDZ PSC 833 and SDZ 280–446 reduced the mean GI50 value for untreated cells from 445 (SE 19) nm to 220 (SE 41) nm, but the effect was obtained within a higher dose range (412–1235 nm). Apparently, LY335979 (1 ␮m) did not exercise any effect on the resistance of the HL60/130.

observed following treatment with SDZ 280–446 and LY335979 as well (Figure 4), and by using these agents the G2/M populations increased to 65% and 60%, respectively. These populations were shown to consist of true tetraploid cells and not doublets or clumps, by the use of pulse-processed red fluorescence signals, which allowed for doublet discrimination. Correspondingly, microscopic images demonstrated several binucleated cells as well as nuclear fragmentation (Figure 5). The wild-type KG1a cells were not receptive to these changes in cell cycle distribution. The cell cycle distribution of the resistant K562 cells underwent similar changes as the resistant KG1 a cells after exposure to the three modulators. In particular, the changes induced by LY335979 were pronounced with reduction of the G1 phase population from 49% to 4%, reduction of the S phase population from 19% to 6%, increment of the G2/M phase population from 26% to 42% and emergence of large populations with subdiploid (23%) and polyploid (25%) cells. The cell cycle distribution of the wild-type K562 cells remained unchanged after equivalent treatment.

Cytotoxic effect of the Pgp inhibitors

Expression of proteins involved in apoptosis regulation

To study the effect of Pgp inhibition on cell proliferation we performed four replicate experiments measuring cell growth in the three pairs of leukemia cell lines during single-drug treatment with increasing doses of the resistance modulating agents. The growth of the KG1a/200 and K562/150 cells was inhibited in a dose-dependent manner by treatment with nanomolar concentrations of either SDZ PSC 833, SDZ 280–446 or LY335979 (Figure 1), whereas the growth of the corresponding parental lines and both sensitive and resistant HL60 lines remained unaffected by equivalent treatment (Figure 2). The half-maximal growth inhibitory doses for the responsive cells are listed in Table 1.

Expression of several proteins involved in the regulation of apoptosis, p53, Bcl-xL, Bcl-2, and Bax was assessed using immunoblotting (Figure 6). None of the cell lines expressed the apoptosis transcription factor p53, which promotes apoptosis subsequent to DNA damage.41 On the other hand, expression of the pro-apoptosis Bax protein was moderate in all cell lines. Both the wild type and the MDR variant of KG1a cells expressed high levels of the anti-apoptotic proteins Bcl2 and Bcl-xL. Both K562 and both HL60 cell types expressed high levels of Bcl-xL. Conversely, Bcl-2 was present at a low level in the K562 cells and basically absent in the HL60 cells. Interestingly, there was no difference in the pattern of expression within each pair of MDR and wild-type cells. Thus, selection for drug resistance did not change the levels of the apoptotic regulators studied in the KG1a, K562 and HL60 cells. Treatment with 1 ␮m SDZ PSC 833 for 48 h did not affect the expression level of these proteins either. The surface membrane protein APO1/Fas, which is an important transducer of certain apoptotic signals,42 was measured by flow cytometry and immunofluorescence detection. From these analyses it was evident that APO1/Fas was present in both HL60 cell types, and absent in both K562 cell types, whereas differential expression of APO1/Fas was demonstrated for the wild type and the MDR variant of KG1a cells (Figure 7). Only the MDR variant was APO1/Fas positive, which might favor a cytotoxic effect of the resistance modulators in these cells.

Pgp inhibition is associated with apoptosis induction Previously, we demonstrated by microscopy that exposure of drug resistant KG1a cells to SDZ PSC833 resulted in chromatin condensation and nuclear fragmentation,26 which are characteristic morphological features of apoptosis. In the present study we measured DNA fragments resulting from apoptosis in cells exposed for 48 h to 1 ␮m of either SDZ PSC 833, SDZ 280–446 or LY335979 using TUNEL assay and flow cytometry. Cellular DNA content was measured simultaneously. The results showed that the KG1a/200 and K562/150 MDR variants were clearly apoptotic as demonstrated by at least a two-fold increase in the FITC-fluorescence signal from dUTP-labeled DNA fragments compared to a neglible increase in the parental cells (Figure 3). Furthermore, striking changes in the DNA distribution of KG1a/200 cells occurred after treatment for 48 h with the modulators. In the control cultures the percentages of cells in each cell cycle compartment were as follows: G1 phase 35%, S phase 35% and G2/M phase 30%. In the SDZ PSC 833treated cultures the G1 phase population was reduced to 4%, the S phase population to 6% and the G2/M phase population increased to 62%, which indicates G2/M arrest. In addition, cells of both higher (18%) and lower (10%) ploidy emerged, the latter representing advanced stages of apoptosis with only remnants of the nuclei left in the cells. Similar results were

Discussion Pgp belongs to the ATP binding cassette (ABC) family of transporter molecules,43 which also include the MDR-associated protein (MRP).44 These proteins contribute to MDR by their functional activity as efflux pumps for a wide range of cytotoxic agents. Furthermore, Pgp contributes to transmembrane efflux of certain cytokines24,25 which may promote growth of various leukemia cells45–47 and provide protection against apoptosis in vitro.48 Therefore, one possible consequence of Pgp inhibition could be reduction of growth stimuli available

771


772

Figure 1 Growth inhibition curves for wild-type leukemia cells (KG1a/0, K562/0) and their resistant sublines (KG1a/200, K562/150) after exposure to increasing doses of SDZ PSC 833, SDZ 280–446 or LY335979. Both MDR variants responded to the three agents with significant growth inhibition. Each point represent the mean of triplicates. The error bars represent 95% confidence intervals.

to proliferating leukemia cells followed by apoptosis. In the present study we have demonstrated that Pgp overexpression is associated with a growth inhibitory response to the Pgp inhibitors SDZ PSC 833, SDZ 280–446 and LY335979 in the MDR variants of KG1a and K562 leukemia cells. On the other hand, the non-Pgp MDR variant of HL60 cells, which overexpresses MRP, was not receptive to similar growth inhibitory effects. Thus, the cytotoxic effect of the modulators is associated with Pgp in particular and not with MDR in general. The KG1a and K562 leukemia cells are poorly differentiated and share many characteristics with primitive hematopoietic blast cells.49–50 Chaudhary and Roninson51 have proposed that Pgp in hematopoietic stem cells may be involved in the export of a growth regulatory molecule which was hitherto not ident-

ified. The putative role for Pgp in the transmembrane transport of cytokines coincides with an autocrine growth regulation involving a number of cytokines in cell lines derived from patients with AML. It has been demonstrated that leukemic myeloid blast cells generally express high levels of IL1 ␣, IL3, and IL6 receptors,52 and that blast cells from AML patients both express and produce a number of cytokines including GM-CSF,53 IL-1,54 IL-6,55 IL-956 and IL-11.57 Furthermore, cell proliferation may occur in autocrine loops based on endogenously produced and extracellularly released growth factors.55,57 The mechanism behind this kind of cellular signal transduction is unknown, but it is likely that any kind of interference with the processing of vital growth factors, such as transport inhibition by modulating agents, will have a detri-


Table 1 The half-maximal growth inhibitory concentrations of SDZ PSC 833, SDZ 280–446 and LY335979 in resistant KG1a/200 and K562/150 leukemia cells

nM GI50 for

KG1a/200 K562/150

SDZ PSC 833

SDZ 280–446

LY335979

312 (SE 41) 414 (SE 50)

685 (SE 51) 578 (SE 54)

66 (SE 1) 48 (SE 8)

The cells were cultured for 72 (K562) and 96 (KG1a) h with increasing concentrations of each modulator to obtain the concentration required for 50% inhibition of cell growth (GI50). The values are means of four replicate experiments. Standard errors of the mean (SE) are given in parenthesis.

Figure 2 Growth inhibition curves for sensitive HL60/0 and resistant HL60/130 cells after exposure to increasing doses of SDZ PSC 833, SDZ 280–446, or LY335979. Growth inhibition did not occur in either of the cell types at the applied dosages. Each point represent the mean of triplicates. The error bars represent the 95% confidence intervals.

mental effect on cell survival. Correspondingly, withdrawal of the supporting growth factor commonly leads to cell death by apoptosis within a few days in most factor-dependent cell lines, and the apoptotic process can be suppressed by new addition of growth factors.58 In the present study the TUNEL assay, the DNA distribution analyses and the microscopic images provided evidence that apoptosis contributes substantially to the selective cytotoxicity of the modulating agents studied. The MDR variant of KG1a/200 cells expressed the APO-1/Fas apoptosis transducer, in contrast to the wild type, which suggests that this resistant line is more receptive to apoptosis induction than its parent line. On the other hand, absence of APO-1/Fas does not prevent resistant K562/150 cells from undergoing apoptosis, which indicates that apoptosis induced by SDZ PSC

Figure 3 Overlay histograms showing the flow cytometric distribution of FITC-dUTP-labeled DNA nick ends in resistant KG1a/200 (a) and K562/150 (b) cells after treatment with 1 ␮m of SDZ PSC 833, SDZ 280–446 or LY335979 for 48 h in comparison with the corresponding wild-type cells. The column to the right lists the values of relative fluorescence (rF) which is the ratio between mean fluorescence values of treated (numerator) and untreated (denominator) populations. The Y-axis represents the number of cells. Kolmogorov– Smirnov statistics revealed that dUTP fluorescence increased significantly compared to non-treated cells for KG1a/200 (D = 0.74, D = 0.58 and D = 0.70 for SDZ PSC 833, SDZ 280–446 and LY335979, respectively) and K562/150 (D = 0.33, D = 0.33 and D = 0.54) cells. The histograms presented are representative for three replicate experiments.

773


774

Figure 4 Bivariate scatterplots presenting cell cycle distributions for apoptotic and viable KG1a/200 (a) and KG1a/0 (b) cells in 48 h regular cultures or cultures supplemented with 1 ␮m SDZ PSC 833, SDZ 280–446 or LY335979. Cellular DNA content is expressed as DNA-PI fluorescence (FL3) along the X-axis, and DNA strand breaks as dUTP-FITC labeling of DNA nick ends (FL1) along the Y-axis. Correspondingly, the horizontal histogram reveals the DNA ploidy (2N, 4N, 8N), and the vertical histogram depicts dUTP-FITC incorporation. The arrows demarcated as APO + and the horizontal cut-off lines in the dUTP-FITC histograms indicate apoptotic regions. Treatment of KG1a/200 cells with a modulator results in G2/M arrest, cytokinesis failure, polyploidization and apoptosis, whereas the parental KG1a/0 cells were unaffected. These data were representative for three replicate experiments.

833, SDZ 280–446 and LY335979 is independent of the APO1/Fas pathway. The BCR-ABL anti-apoptosis gene is a cytogenetic hallmark of chronic myelogenous leukemia (CML),59 and is constitutively expressed in the CML derived K562 cells, which accordingly, are highly resistant to apoptosis induction by a variety of stimuli.60,61 Therefore, it is quite remarkable that the MDR variant of K562 cells also undergoes apoptosis after exposure to nanomolar concentrations of the modulators studied. In addition, the anti-apoptotic regulators Bcl-2 and Bcl-xL contribute to drug resistance of the MDR variants of both KG1a and K562 cells. Both Bcl-2 and Bcl-xL are frequently coexpressed with Pgp in acute myeloblastic leukemia cells and associated with autonomous growth in vitro and poor prognosis of leukemia.62,63 The apoptotic promoter p5364 was not expressed in any of the cell lines studied, which is consistent with the total absence of p53 mRNA and protein expression previously reported for the wild types of KG1a,65 HL6066 and K56267 cells. Although loss of wild-type p53 function is usually associated with resistance to apoptosis,68,69 the Pgp-rich variants of KG1a and K562 were receptive to apoptosis induction in

the present study. Inhibition of Pgp in these cells was also associated with cytokinesis failure as shown by the accumulation of large tetraploid populations and the emergence of binucleated cells using flow cytometry and microscopy, respectively. Thus, our study strongly indicates that overexpression of Pgp is involved in acquisition of cytokinesis failure and apoptosis following exposure to Pgp inhibitors. Cells with a low constitutive level of Pgp appear to be at less risk of these events. Further studies are needed to identify the trigger mechanism(s) and apoptotic pathway(s). Currently, investigations of these are being done using time course flow cytometric DNA analyses and cell sorting of cultures treated with Pgp inhibitory agents. Development of hypersensitivity to cytotoxic drugs, local anesthetics, narcotic analgesics and calcium antagonists during in vitro selection for drug resistance have previously been reported as collateral sensitivity.70–72 Most of these compounds are not substrates of Pgp, and the phenomenon has been associated with alteration of membrane biophysical properties rather than Pgp inhibition.70–73 Collateral sensitivity has been demonstrated for drug concentrations in the


775

Figure 4

Continued

micromolar range.72,74 Conversely, the selective cytotoxicity of the modulators in this study appeared at considerably lower concentrations, which facilitates possible therapeutic exploitation, and resulted specifically in cytokinesis failure and apoptosis induction. The maximally tolerable plasma concentration of PSC 833 in cancer patients is 2.5–3 ␮m,19 which demonstrates that the concentrations applied in our study are clinically relevant. The present paper discloses selective cytotoxicity of low doses of the Pgp inhibitory agents SDZ PSC 833, SDZ 280– 446 and LY335979. Previously, we have reported a similar effect of high-dose verapamil (GI50 11.6 ␮m) and a 60-fold lower dose of SDZ PSC 833, which was associated with the MDR phenotype in leukemia cells but not in hepatoma cells.26 In the present paper we demonstrate that increased MRP expression is not a predisposing factor due to unrestrained growth of MRP expressing HL60 leukemia cells in the presence of the modulators, and efficient growth inhibition of the MDR variants of KGla and K562 leukemia cell lines induced by LY335979 which does not interact with MRP. SDZ PSC 833 and SDZ 280–446 appear to interact with both Pgp and MRP, but the contribution of other transporters cannot be excluded from these experiments. Another MDR-associated ABC transporter, the multispecific organic anion transporter (cMOAT), is only inefficiently inhibited by compounds known

to block MRP,75 and thus represents a less likely target for the modulators studied. For several reasons it appears that functional restriction of the Pgp efflux pump might be responsible for the particular phenomenon disclosed in this paper. First, the cytotoxicity was conferred by drugs specifically targeted at Pgp interaction. Second, the results were consistent irrespective of structural diversity of the Pgp inhibitory agents. Third, these agents were shown to induce functional inhibition of Pgp at comparable concentrations. Fourth, susceptibility to these agents required Pgp overexpression. Finally, expression of a specific apoptotic regulatory protein did not seem to be the differential factor between receptive and non-receptive cell lines. To further prove the relationship between Pgp and the said phenomenon experiments using homologue cell lines with different Pgp expression levels, cell lines transfected with the MDR1 gene, or cell lines depleted for Pgp by MDR1 antisense treatment are warranted. In conclusion, our results demonstrate that low-dose treatment of Pgp-rich leukemia cells with different Pgp inhibitory agents results in growth inhibition and cell death by apoptosis. This finding represents a potential therapeutic option that is currently being studied in vivo, and should be considered during the ongoing development of strategies to overcome multidrug resistance.


776

Figure 5 The microphotograph of resistant KG1a/200 leukemia cells treated with PSC 833 and stained with Papanicolau dyes demonstrates morphological features of cytokinesis failure and apoptosis. Examples of binucleated cells (A, B), cells with fragmented nuclei (C) and cells with condensed nuclei (D) are indexed.

Figure 6 Immunoblots demonstrating the distribution of apoptotic proteins in human leukemia wild-type cells and their resistant sublines (KG1a/200, K562/150, HL60/130). Lane A represents a positive control for bax and bcl-2 (normal colonic mucosa). Lane B represents a positive control for bcl-2, bcl-xL and p53 (HTC116 human colon cancer cell line treated with 5FU). Each cell type was examined untreated and after treatment with 1 ␮m SDZ PSC 833 for 48 h. The protein expression appeared to be unaffected by this treatment. The blots were representative for two replicate experiments.

Acknowledgements This study was supported by grants from Medinnova SF, The Norwegian Cancer Society and The Research Council of Norway. The authors are grateful to Karen Johanne Beckstrøm, Reidun Hauge, May Ellen Lauritsen at the National Hospital in Oslo and Karen Kazemier at the Free University of Amsterdam for excellent technical assistance.

Figure 7 Overlay histograms showing the flow cytometric distribution of fluorescence-labeled APO-1/Fas in wild types (KG1a/0, K562/0, HL60/0) and corresponding MDR variants (KG1a/200, K562/150, HL60/130) of human leukemia cells. The primary antibody was either DMS138 (a) or DMS140 (b). The column to the right lists the expression factor (EF) for APO-1/Fas which is the ratio between mean fluorescence values of populations stained with relevant (numerator) and irrelevant (denominator) MAbs. The Y-axis represents the number of cells. Histogram comparisons using Kolmogorov–Smirnov statistics revealed that APO-1/Fas fluorescence differs significantly from the isotypic controls for KG1a/200 (D = 0.60 and 0.72 for DMS138 and DMS140, respectively), HL60/0 (D = 0.91 and 0.92) and HL60/130 (D = 0.93 and 0.92) cells. The two figures provide mutual confirmation by means of different anti-Apo-1/Fas Mabs.

References 1 Juliano RL, Ling V. A surface glycoprotein modulating drug permeability in Chinese hamster ovary cell mutants. Biochim Biophys Acta 1976; 455: 152–162. 2 Pastan I, Gottesman M. Multiple-drug resistance in human cancer. New Engl J Med 1987; 316: 1388–1393. 3 Marie JP, Zittoun R, Sikic BI. Multidrug resistance (mdr1) gene expression in adult acute leukemias: correlations with treatment outcome and in vitro drug sensitivity. Blood 1991; 78: 586–592. 4 Campos L, Guyotat D, Archimbaud E, Calmard-Oriol P, Tsuruo T, Troncy J, Treille D, Fiere D. Clinical significance of multidrug resistance P-glycoprotein expression on acute nonlymphoblastic leukemia cells at diagnosis. Blood 1992; 79: 473–476. 5 Yuen AR, Sikic BI. Multidrug resistance in lymphomas. J Clin Oncol 1994; 12: 2453–2459. 6 Chan HS, Thorner PS, Haddad G, Ling V. Immunohistochemical detection of P-glycoprotein: prognostic correlation in soft tissue sarcoma of childhood. J Clin Oncol 1990; 8: 689–704.


7 Chan HS, Haddad G, Thorner PS, DeBoer G, Lin YP, Ondrusek N, Yeger H, Ling V. P-glycoprotein expression as a predictor of the outcome of therapy for neuroblastoma. New Engl J Med 1991; 325: 1608–1614. 8 Baldini N, Scotlandi K, Barbanti-Brodano G, Manara MC, Maurici D, Bacci G, Bertoni F, Picci P, Sottili S, Campanacci M et al. Expression of P-glycoprotein in high-grade osteosarcomas in relation to clinical outcome. New Engl J Med 1995; 333: 1380– 1385. 9 Ford JM. Modulators of multidrug resistance. Preclinical studies. Hematol Oncol Clin North Am 1995; 9: 337–361. 10 Tsuruo T, Iida H, Tsukagoshi S, Sakurai Y. Overcoming of vincristine resistance in P388 leukemia in vivo and in vitro through enhanced cytotoxicity of vincristine and vinblastine by verapamil. Cancer Res 1981; 41: 1967–1972. 11 Candell J, Valle V, Soler M, Rius J. Acute intoxication with verapamil. Chest 1979; 75: 200–201. 12 Twentyman PR, Bleehen NM. Resistance modification by PSC833, a novel non-immunosuppressive cyclosporin. Eur J Cancer 1991; 27: 1639–1642. 13 Loor F, Boesch D, Gaveriaux C, Jachez B, Pourtier-Manzanedo A, Emmer G. SDZ 280–446, a novel semi-synthetic cyclopeptolide: in vitro and in vivo circumvention of the P-glycoprotein-mediated tumour cell multidrug resistance. Br J Cancer 1992; 65: 11–18. 14 Starling JJ, Shepard RL, Cao J, Law KL, Norman BH, Kroin JS, Ehlhardt WJ, Baughman TM, Winter MA, Bell MG, Shih C, Gruber J, Elmquist WF, Dantzig AH. Pharmacological characterization of LY335979: a potent cyclopropyldibenzosuberane modulator of Pglycoprotein. Adv Enzyme Regul 1997; 37: 335–347. 15 Dantzig AH, Shepard RL, Cao J, Law KL, Ehlhardt WJ, Baughman TM, Bumol TF, Starling JJ. Reversal of P-glycoprotein-mediated multidrug resistance by a potent cyclopropyldibenzosuberane modulator, LY335979. Cancer Res 1996; 56: 4171–4179. 16 Naito M, Watanabe T, Tsuge H, Koyama T, Oh-hara T, Tsuruo T. Potentiation of the reversal activity of SDZ PSC833 on multi-drug resistance by an anti-P-glycoprotein monoclonal antibody MRK16. Int J Cancer 1996; 67: 435–440. 17 Tiberghien F, Loor F. Ranking of P-glycoprotein substrates and inhibitors by a calcein-AM fluorometry screening assay. Anticancer Drugs 1996; 7: 568–578. 18 Sonneveld P, Marie JP, Huisman C, Vekhoff A, Schoester M, Faussat AM, van Kapel J, Groenewegen A, Charnick S, Zittoun R, Lowenberg B. Reversal of multidrug resistance by SDZ PSC 833, combined with VAD (vincristine, doxorubicin, dexamethasone) in refractory multiple myeloma. A phase I study. Leukemia 1996; 10: 1741–1750. 19 Boote DJ, Dennis IF, Twentyman PR, Osborne RJ, Laburte C, Hensel S, Smyth JF, Brampton MH, Bleehen NM. Phase I study of etoposide with SDZ PSC 833 as a modulator of multidrug resistance in patients with cancer. J Clin Oncol 1996; 14: 610–618. 20 Kornblau SM, Estey E, Madden T, Tran HT, Zhao S, Consoli U, Snell V, Sanchez-Williams G, Kantarjian H, Keating M, Newman RA, Andreeff M. Phase I study of mitoxantrone plus etoposide with multidrug blockade by SDZ PSC-833 in relapsed or refractory acute myelogenous leukemia. J Clin Oncol 1997; 15: 1796–1802. 21 Endicott JA, Ling V. The biochemistry of P-glycoprotein-mediated multidrug resistance. Annu Rev Biochem 1989; 58: 137–171. 22 Ueda K, Okamura N, Hirai M, Tanigawara Y, Saeki T, Kioka N, Komano T, Hori R. Human P-glycoprotein transports cortisol, aldosterone, and dexamethasone, but not progesterone. J Biol Chem 1992; 267: 24248–2452. 23 Tatsuta T, Naito M, Oh-hara T, Sugawara I, Tsuruo T. Functional involvement of P-glycoprotein in blood–brain barrier. J Biol Chem 1992; 267: 20383–20391. 24 Drach J, Gsur A, Hamilton G, Zhao S, Angerler J, Fiegl M, Zojer N, Raderer M, Haberl I, Andreeff M, Huber H. Involvement of Pglycoprotein in the transmembrane transport of interleukin-2 (IL2), IL-4, and interferon-gamma in normal human T lymphocytes. Blood 1996; 88: 1747–1754. 25 Raghu G, Park SW, Roninson IB, Mechetner EB. Monoclonal antibodies against P-glycoprotein, an MDR1 gene product, inhibit interleukin-2 release from PHA-activated lymphocytes. Exp Hematol 1996; 24: 1258–1264. 26 Lehne G, Rugstad HE. Cytotoxic effect of the cyclosporin PSC 833

27 28 29 30

31

32

33

34

35 36

37 38

39 40

41 42 43 44

45

46

47 48

by inhibition of P-glycoprotein in multidrug resistant leukemia cells. Br J Cancer 1998; 78: 593–600. Knowles BB, Howe CC, Aden DP. Human hepatocellular carcinoma cell lines secrete the major plasma proteins and hepatitis B surface antigen. Science 1980; 209: 497–499. Collins SJ, Gallo RC, Gallagher RE. Continuous growth and differentiation of human myeloid leukaemic cells in suspension culture. Nature 1977; 270: 347–349. Lozzio CB, Lozzio BB. Human chronic myelogenous leukemia cell-line with positive Philadelphia chromosome. Blood 1975; 45: 321–334. Gruber A, Briese B, Arestrom I, Vitols S, Bjorkholm M, Peterson C. Effect of verapamil on daunorubicin accumulation in human leukemic cells with different levels of MDR1 gene expression. Leukemia Res 1993; 17: 353–358. Lehne G, De Angelis P, Clausen OP, Egeland T, Tsuruo T, Rugstad HE. Binding diversity of antibodies against external and internal epitopes of the multidrug resistance gene product P-glycoprotein. Cytometry 1995; 20: 228–237. Den Boer ML, Zwaan CM, Pieters R, Kazemier KM, Rottier MM, Flens MJ, Scheper RJ, Veerman AJ. Optimal immunocytochemical and flow cytometric detection of P-gp, MRP and LRP in childhood acute lymphoblastic leukemia. Leukemia 1997; 11: 1078–1085. Lehne G, De Angelis P, Clausen OPF, Rugstad HE. Human hepatoma cells rich in P-glycoprotein are sensitive to aclarubicin and resistant to three other anthracyclines. Br J Cancer 1996; 74: 1719–1729. Gorczyca W, Tuziak T, Kram A, Melamed MR, Darzynkiewicz Z. Detection of apoptosis-associated DNA strand breaks in fineneedle aspiration biopsies by in situ end labeling of fragmented DNA. Cytometry 1994; 15: 169–175. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970; 227: 680–685. De Angelis P, Stokke T, Smedshammer L, Lothe RA, Lehne G, Chen Y, Clausen OP. P-glycoprotein is not expressed in a majority of colorectal carcinomas and is not regulated by mutant p53 in vivo. Br J Cancer 1995; 72: 307–311. Young IT. Proof without prejudice: use of the Kolmogorov–Smirnov test for the analysis of histograms from flow systems and other sources. J Histochem Cytochem 1977; 25: 935–941. Leith CP, Chen IM, Kopecky KJ, Appelbaum FR, Head DR, Godwin JE, Weick JK, Willman CL. Correlation of multidrug resistance (MDR1) protein expression with functional dye/drug efflux in acute myeloid leukemia by multiparameter flow cytometry: identification of discordant MDR−/efflux+ and MDR1+/efflux− cases. Blood 1995; 86: 2329–2342. Scambia G, Panici PB, Contu G, De Vincenzo R, Ferrandina G, Isola G, Maccio A, Mancuso S. Mechanisms and modulation of resistance to anthracyclines. Int J Oncol 1994; 4: 951–959. Stride BD, Grant CE, Loe DW, Hipfner DR, Cole SPC, Deeley RG. Pharmacological characterization of the murine and human orthologs of multidrug-resistance protein in transfected human embryonic kidney cells. Mol Pharmacol 1997; 52: 344–353. Velculescu VE, El-Deiry WS. Biological and clinical importance of the p53 tumor suppressor gene. Clin Chem 1996; 42: 858–868. Nagata S, Golstein P. The Fas death factor. Science 1995; 267: 1449–1456. Germann UA, Pastan I, Gottesman MM. P-glycoproteins: mediators of multidrug resistance. Semin Cell Biol 1993; 4: 63–76. Krishnamachary N, Center MS. The MRP gene associated with a non-P-glycoprotein multidrug resistance encodes a 190-kDa membrane bound glycoprotein. Cancer Res 1993; 53: 3658– 3661. Sahraoui Y, Allouche M, Ammar A, Spanakis E, Clemenceau C, Jasmin C, Perraki M, Varela-Millot C, Georgoulias V. Interleukin 2 production and interleukin 2 receptor expression by human immature leukemic T cells. Leukemia 1992; 6: 1025–1035. Ghaderi AA, Richardson P, Cardona C, Millsum MJ, Ling N, Gillis S, Ledbetter J, Gordon J. Stimulation of B-chronic lymphocytic leukemia populations by recombinant interleukin-4 and other defined growth-promoting agents. Leukemia 1988; 2: 165–170. Drexler HG, Zaborski M, Quentmeier H. Cytokine response profiles of human myeloid factor-dependent leukemia cell lines. Leukemia 1997; 11: 701–708. Dancescu M, Rubio-Trujillo M, Biron G, Bron D, Delespesse G,

777


778

49 50 51 52

53 54

55

56

57 58 59

60

61

Sarfati M. Interleukin 4 protects chronic lymphocytic leukemic B cells from death by apoptosis and upregulates Bcl-2 expression. J Exp Med 1992; 176: 1319–1326. Koeffler HP, Billing R, Lusis AJ, Sparkes R, Golde DW. An undifferentiated variant derived from the human acute myelogenous leukemia cell line (KG-1). Blood 1980; 56: 265–273. Lozzio BB, Lozzio CB, Bamberger EG, Feliu AS. A multipotential leukemia cell line (K-562) of human origin. Proc Soc Exp Biol Med 1981; 166: 546–550. Chaudhary PM, Roninson IB. Expression and activity of P-glycoprotein, a multidrug efflux pump, in human hematopoietic stem cells. Cell 1991; 66: 85–94. Ferrari S, Grande A, Manfredini R, Tagliafico E, Zucchini P, Torelli G, Torelli U. Expression of interleukins 1, 3, 6, stem cell factor and their receptors in acute leukemia blast cells and in normal peripheral lymphocytes and monocytes. Eur J Haematol 1993; 50: 141–148. Young DC, Wagner K, Griffin JD. Constitutive expression of the granulocyte–macrophage colony-stimulating factor gene in acute myeloblastic leukemia. J Clin Invest 1987; 79: 100–106. Bradbury D, Rogers S, Reilly IA, Kozlowski R, Russell NH. Role of autocrine and paracrine production of granulocyte–macrophage colony-stimulating factor and interleukin-1 beta in the autonomous growth of acute myeloblastic leukaemia cells – studies using purified CD34-positive cells. Leukemia 1992; 6: 562–566. Zhu YM, Bradbury DA, Keith FJ, Russell N. Absence of retinoblastoma protein expression results in autocrine production of interleukin-6 and promotes the autonomous growth of acute myeloid leukemia blast cells. Leukemia 1994; 8: 1982–1988. Lemoli RM, Fortuna A, Tafuri A, Fogli M, Amabile M, Grande A, Ricciardi MR, Petrucci MT, Bonsi L, Bagnara G, Visani G, Martinelli G, Ferrari S, Tura S. Interleukin-9 stimulates the proliferation of human myeloid leukemic cells. Blood 1996; 87: 3852–3859. Kobayashi S, Teramura M, Sugawara I, Oshimi K, Mizoguchi H. Interleukin-11 acts as an autocrine growth factor for human megakaryoblastic cell lines. Blood 1993; 81: 889–893. Williams GT, Smith CA, Spooncer E, Dexter TM, Taylor DR. Haemopoietic colony stimulating factors promote cell survival by suppressing apoptosis. Nature 1990; 343: 76–79. Stam K, Heisterkamp N, Grosveld G, de Klein A, Verma RS, Coleman M, Dosik H, Groffen J. Evidence of a new chimeric ber/cabl mRNA in patients with chronic myelocytic leukemia and the Philadelphia chromosome. New Engl J Med 1985; 313: 1429– 1433. McGahon A, Bissonnette R, Schmitt M, Cotter KM, Green DR, Cotter TG. BCR-ABL maintains resistance of chronic myelogenous leukemia cells to apoptotic cell death. Blood 1994; 83: 1179– 1187. Bedi A, Barber JP, Bedi GC, el-Deiry WS, Sidransky D, Vala MS, Akhtar AJ, Hilton J, Jones RJ. BCR-ABL-mediated inhibition of apoptosis with delay of G2/M transition after DNA damage: a

62

63

64 65 66 67 68

69

70 71 72 73

74

75

mechanism of resistance to multiple anticancer agents. Blood 1995; 86: 1148–1158. Pallis M, Zhu YM, Russell NH. Bcl-x(L) is heterogenously expressed by acute myeloblastic leukaemia cells and is associated with autonomous growth in vitro and with P-glycoprotein expression. Leukemia 1997; 11: 945–949. Bosanquet AG, Bell PB, Burlton AR, Amos TA. Correlation of bcl2 with P-glycoprotein expression in chronic lymphocytic leukaemia and other haematological neoplasms but of neither marker with ex vivo chemosensitivity or patient survival. Leuk Lymphoma 1996; 24: 141–147. Canman C, Kastan M. Role of p53 in apoptosis. Adv Pharmacol 1997; 41: 429–460. Clave E, Carosella ED, Gluckman E, Dubray B, Socie G. Ionizing radiation effects on the KG1 a primitive hematopoietic cell line. Int J Radiat Oncol Biol Phys 1996; 35: 709–719. Wolf D, Rotter V. Major deletions in the gene encoding the p53 tumor antigen cause lack of p53 expression in HL-60 cells. Proc Natl Acad Sci USA 1985; 82: 790–794. Law JC, Ritke MK, Yalowich JC, Leder GH, Ferrell RE. Mutational inactivation of the p53 gene in the human erythroid leukemic K562 cell line. Leukemia Res 1993; 17: 1045–1050. Wang Y, Ramqvist T, Szekely L, Axelson H, Klein G, Wiman KG. Reconstitution of wild-type p53 expression triggers apoptosis in a p53-negative v-myc retrovirus-induced T-cell lymphoma line. Cell Growth Differ 1993; 4: 467–473. Aas T, Borresen AL, Geisler S, Smith-Sorensen B, Johnsen H, Varhaug JE, Akslen LA, Lonning PE. Specific P53 mutations are associated with de novo resistance to doxorubicin in breast cancer patients. Nature Med 1996; 2: 811–814. Bech-Hansen NT, Till JE, Ling V. Pleiotropic phenotype of colchicine-resistant CHO cells: cross-resistance and collateral sensitivity. J Cell Physiol 1976; 88: 23–31. Hill BT. Modulation of antitumour drug resistance: experimental laboratory data and results of clinical evaluation. Cancer Treat Rev 1990; 17: 197–202. Callaghan R, Riordan JR. Collateral sensitivity of multidrug resistant cells to narcotic analgesics is due to effects on the plasma membrane. Biochim Biophys Acta 1995; 1236: 155–162. Stow MW, Warr JR. Reduced influx is a factor in accounting for reduced vincristine accumulation in certain verapamil-hypersensitive multidrug-resistant CHO cell lines. FEBS Lett 1993; 320: 87–91. Quesada AR, Barbacid MM, Mira E, Aracil M, Marquez G. Chemosensitization and drug accumulation assays as complementary methods for the screening of multidrug resistance reversal agents. Cancer Lett 1996; 99: 109–114. Evers R, Kook M, van Deemter L, Janssen H, Calafat J, Oomen LC, Paulusma CC, Oude Elferink RP, Baas F, Schinkel AH, Borst P. Drug export activity of the human canalicular multispecific organic anion transporter in polarized kidney MDCK cells expressing cMOAT (MRP2) cDNA. J Clin Invest 1998; 101: 1310–1319.