Cellular Pharmacology of Mitoxantrone in P

Home , Mitoxantrone

Cellular pharmacology of mitoxantrone in p-glycoprotein-positive and -negative human myeloid leukemic cell lines U Consoli, NT Van, N Neamati, R Mahadevia, M Beran, S Zhao and M Andreeff

Department of Hematology, Section of Experimental Hematology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA

Previous reports suggest that resistance to mitoxantrone in dif- protein is particularly important because it confers cross-resist- ferent tumor cell lines is unrelated to the overexpression of p- ance to several structurally unrelated drugs.9 The mechanisms glycoprotein. In order to determine the role of p-glycoprotein in the cellular pharmacology of mitoxantrone flow cytometry of action of mitoxantrone resistance are not well known. and confocal microscopy were used to study two human Although mitoxantrone seems to select for a unique drug- myeloid leukemia cell lines selected for resistance to mitoxan- resistance phenotype, some cross-resistance with anthracy- trone (HL-60MX2) and doxorubicin (HL-60DOX). To optimize the clines has been shown in vitro.10,11 Human colon carci- detection of intracellular mitoxantrone, we determined the noma,12 leukemia13 and gastric carcinoma14 cell lines selec- maximum excitation (607 nm) and emission (684 nm) wavelength by fluorescence spectroscopy. The modified flow cyto- ted by continuous exposure to increasing concentrations of metric conditions using 568.2 nm laser emission for excitation mitoxantrone had no p-glycoprotein overexpression. Clinical and a 620 nm long pass filter for fluorescence collection cross-over studies have not demonstrated complete cross- resulted in a 1-log increase in sensitivity, compared with stan- resistance between mitoxantrone and doxorubicin,15,16 pro- dard 488-nm laser excitation. Uptake and retention of mitoxan- viding rationale for the use of mitoxantrone in disease- trone in the presence of verapamil, a calcium channel blocker known to inhibit p-glycoprotein, were analyzed. Our results relapsed patients following anthracycline treatment. In con- showed no change in uptake and retention of the drug in p- trast to these reports, mitoxantrone like the anthracyclines, has glycoprotein-negative mitoxantrone-resistant HL-60MX2 cells also been shown to select for cells exhibiting the MDR pheno- and in its sensitive parental line, HL-60s. In contrast, 3.1- and type mediated by p-glycoprotein overexpression.17 Further- 2.4-fold increases were found in uptake and retention of mitox- more, a resistance mechanism associated with enhanced antrone in p-glycoprotein-positive cells (HL-60DOX) incubated with verapamil. Confocal microscopy of intracellular drug dis- efflux involving a pathway distinct from the MDR-1-encoded 18 tribution demonstrated reduced nuclear uptake, which could be p-glycoprotein has also been reported. reversed by verapamil, in HL-60DOX. A characteristic punctate Flow cytometry and confocal laser microscopy have been pattern was observed for the intracytoplasmic drug distribution successfully used to analyze cellular pharmacology and the in HL-60DOX and HL-60MX2 cells and was partially modified by intracellular distribution of anthracyclines based on their flu- the presence of verapamil in HL-60DOX cells. Verapamil increased cytotoxicity of mitoxantrone two-fold in HL-60DOX orescence emission after laser excitation. This method pro- cells, 1.4-fold in HL-60MX2, and had no effect in HL-60s. Our vides a powerful tool to evaluate the role of p-glycoprotein in study demonstrates that the cellular pharmacology of mitoxan- the cellular pharmacology of anthracylines and in the devel- trone is affected by p-glycoprotein and can be reversed at least opment of MDR in various systems in vitro19,20 and in in part by verapamil. Other mechanisms of resistance however, vivo.21,22 Although mitoxantrone shares only part of its chemi- seem to play a determinant role in the modulation of mitoxantrone cytotoxicity. cal structure with other anthracyclines, like doxorubicin, it has Keywords: mitoxantrone; p-glycoprotein; multidrug resistance a planar electron-rich chromophore of a polycyclic dye, which is responsible for its fluorescent properties. In this study, we used a fluorimetric method to evaluate the Introduction intracellular content of mitoxantrone in order to clarify the influence of p-glycoprotein on the cellular pharmacology of Mitoxantrone is an anticancer anthraquinone drug1,2 that has mitoxantrone and in the development of MDR. Uptake and shown therapeutic efficacy comparable with that of standard retention of mitoxantrone were evaluated in two human cell induction and salvage treatment regimens in advanced breast lines selected for resistance to mitoxantrone (HL-60MX2, p- cancer,3 acute leukemia,4,5 non-Hodgkin’s lymphoma,6 and glycoprotein negative) and doxorubicin (HL-60DOX, p-glyco- chronic myeloid leukemia in blastic transformation.7 Although protein positive) and their drug-sensitive parent cell line (HL- mitoxantrone is structurally similar to anthracycline doxorub- 60s, p-glycoprotein negative) in the presence of verapamil, a icin, it has an improved tolerability profile,8 and the mech- calcium channel blocker known to inhibit MDR possibly by anism of action appears to be quite different. Like other interfering with p-glycoprotein function.23 Cytotoxicity assays anthracyclines, mitoxantrone intercalates into DNA and in the presence and absence of verapamil were also perfor- causes DNA-protein cross-links and protein-associated dou- med to analyze the role of p-glycoprotein in protecting cells ble-stranded and single-stranded breaks in the DNA. In from mitoxantrone cytotoxicity. Lastly, to evaluate the cellular addition, this drug is able to stabilize the DNA topoisomerase pharmacology of mitoxantrone better, we studied the different II cleavable complex and to prevent DNA synthesis.8 patterns of intracellular distribution using confocal laser scan- The development of cellular resistance to anticancer drugs ning microscopy with optical sectioning capability. In this greatly limits the activity of cancer chemotherapy against study, we provide evidence that mitoxantrone is a substrate malignant tumors. In this regard, overexpression of p-glyco- for p-glycoprotein and that higher intracellular concentrations of mitoxantrone in p-glycoprotein-positive cells incubated in the presence of verapamil are associated with increased cyto- Correspondence: U Consoli, Institute of Hematology, Ospedale Ferra- toxicity, circumventing, at least in part, p-glycoprotein- rotto, Via Citelli 6, Catania 95124, Italy mediated MDR. Received 23 November 1996; accepted 8 September 1997 P-glycoprotein and mitoxantrone U Consoli et al 2067 Materials and methods Cell uptake and retention studies

Drugs Cells were incubated at the concentration of 2 × 106/ml in medium plus FCS 5% with or without verapamil (10 ␮M) for Mitoxantrone and verapamil were obtained as pharmaceutical 2 h at 37°C with mitoxantrone at the concentration of 5 ␮g/ml. preparations from Lederle Parenterals (Carolina, Puerto Rico) For uptake studies, aliquots of 5 × 105 cells were taken after and American Regent Laboratory (Shirley, NY, USA). 15, 30, 60, 90 and 120 min of incubation. Cells were washed in ice-cold PBS without Ca2+ or Mg2+, resuspended in 0.5 ml ice-cold PBS, and mitoxantrone fluorescence was measured within 5 min. For retention studies, at the end of a 2 h Cell lines exposure, every sample was washed twice in ice-cold PBS and resuspended for 2 h in fresh medium, with or without 10 ␮M The human acute myeloid leukemia HL-60s and its mitoxan- verapamil. Cells were incubated at 37°C, and aliquots of cells trone-selected HL-60MX2 cell lines were generously provided were analyzed at 15, 30, 60, 90 and 120 min in the same by Dr Graydon Harker (VA Medical Center, Salt Lake City, conditions used for the uptake studies. Experiments were per- UT, USA). The multiple myeloma parent cell line 8226/S and formed in triplicate, and a representative experiment is shown. its doxorubicin-resistant mutants 8226/DOX6 and The uptake and retention were calculated by dividing the 8226/DOX40 were kindly provided by Dr WS Dalton median level fluorescence channel, after 2 h of incubation (University of Arizona, Tucson, AZ, USA). Those cell lines with verapamil, by the fluorescence signal of the cells incu- were grown in RPMI 1640 supplemented with 10% FCS, 1% bated with the drug alone, measured at the same time point. penicillin, 1% streptomycin and 1% L-glutamine (all from The level of drug fluorescence increase corrected for autoflu- GIBCO-BRL, Gaithersburg, MD, USA). Doxorubicin-resistant orescence was expressed as the mean of three experiments. selected HL-60DOX cells were grown in IMDM supplemented with 10% FCS, 1% penicillin, 1% streptomycin and 1% L- glutamine. Cell cultures were maintained at 37°C in a humidi- fied 5% CO2 atmosphere. Cytotoxicity (MTT assay)

Drug cytotoxicity in vitro was assessed for each cell line by RNA extraction and reverse transcriptase PCR using the MTT dye reduction assay (Sigma Chemical Com- (RT-PCR) pany, St Louis, MO, USA). This technique relies on converting MTT (3,(4,-dimethylthiazol-2-yl)-2-5 diphenyl tetrazolium RNA was isolated according to the acid-guanidium-phenol- bromide) to a red formazan product by reductive enzymes chloroform method.24 RT-PCR was performed as previously present only in metabolically viable cells. Hence, the quantit- reported.25 ation of insoluble formazan was possible by dissolving it in a lysing buffer and measuring the optical density of each sample with a microplate spectrophotometer (Molecular Devices USA, Menlo Park, CA, UA) at a 570 nm wavelength. Cells Antibody staining (135 ␮l) were plated in 96-well microassay culture plates × 4 ° (2 10 cells/well) and grown for 24 h at 37 CinaCO2 incu- For the study of p-glycoprotein MoAb 4E3 (kindly provided bator. Drugs were then added to the wells (15 ␮l) to achieve by Dr R Arceci, Dana-Farber Cancer Center, Boston, MA, a final drug concentration of 0.025 to 5 ␮g/ml (eight wells USA) was used as previously described.26 were used for each different concentration). The plates were ° incubated at 37 C in a 5% CO2 incubator for 1 h. Cells were then resuspended in fresh medium and incubated for 72 h. Flow cytometry Wells containing culture medium in the presence or absence of verapamil were used as controls. When incubation was complete, 25 ␮l of the stock solution of MTT dye in a 0.9% Flow cytometric measurements were performed with a dual- aqueous NaCl solution was added to each well to achieve a laser FACS Vantage (Becton Dickinson, Mountain View, CA, final concentration of 0.5 ␮g/ml. The plates were incubated USA). The krypton/argon spectrum laser (Coherent, Mountain ° ␮ at 37 CinaCO2 incubator for 4 h. Subsequently, 85 lof View, CA, USA) emits several lines in the orange and red areas lysing buffer (20% solution of SDS in DMF and water (1:1, of the light spectrum, all suitable for mitoxantrone binding pH 4.7)) was added, and the cells were incubated for an studies, but the orange line at 568.2 nm (150 mW) was chosen ° additional 24 h at 37 C in a 5% CO2 incubator. The optical as the best combination of high efficiency of excitation (% of density of each well was then measured with a microplate maximum), wide range of collectable drug fluorescence emis- spectrophotometer (Molecular Devices USA) at a wavelength sion, and availability of filters to allow the measurement of of 570 nm. The percentage of cell viability was calculated drug emission free of interference from incident light. For the using the following equation: 4E3 MoAb, an argon ion laser (Spectra Physics, Mountain View, CA, USA) operating at 488 nm and 15 mW was used, % cell viability = (mean optical density of treated wells/ and detection light was collected through a 530/30 nm band mean optical density of control wells) × 100 pass filter. Data acquisition and analysis were performed with the Lysys II software (Becton Dickinson) and results were The viability values were plotted against the drug concen- expressed as D values according to the Kolmogorov–Smir- tration used, and the ID50 was calculated from the curve. nov statistic. Experiments were repeated at least three times. P-glycoprotein and mitoxantrone U Consoli et al 2068 Laser confocal microscopy

An upright version of a Zeiss (New York, NY, USA) confocal laser scanning microscope equipped with a He/Ne laser (543- nm emission, power 5 mW) and a LSMv2.0 software, was used. Signals were collected by a photomultiplier through a 590-nm long pass filter. Digitized images were transmitted to a MAC-based image analysis system through a GPIB interface using BDS-LSM software (Biological Detection Systems, Pitts- burgh, PA, USA). IPLab software was used for image pro- cessing (Signal Analytics, Vienna, VA, USA). Teflon-coated glass microscopy slides (Cel-Line Associates, Newfield, NJ, USA) with 10 wells, each 7 mm in diameter and containing 10 ␮l, were used. Cells adhering to the microscope slide were analyzed for mitoxantrone distribution under oil immersion microscopy. HL-60s, HL-60DOX and HL-60MX2 were incubated in the presence of mitoxantrone either with or without verapamil as described above for uptake study. Drug distribution was recorded within 5 min after 30 and 90 min of incubation with the drugs.

Results

Mitoxantrone ﬂuorescence properties

The excitation spectrum of mitoxantrone in aqueous solution (Figure 1a) showed peak excitation at 607 nm and demonstrated that excitation at 488 nm was only 6% of the maximum. This greatly limits the usefulness of the 488 nm line of the Argon laser for pharmacological distribution studies of mitoxantrone, specially at low drug concentrations. The emission spectrum of mitoxantrone obtained with a SPF-500TM spectrofluorimeter (SLM Instrument, Urbana, IL, USA) showed a relative fluorescence emission at 568.2 nm, which is 20-fold higher than the emission obtained at 488 nm (Figure 1b). This maximum was also confirmed by absorption spectrophotome- try (data not shown). The 568.2 nm wavelength used for drug binding studies by flow cytometry resulted in a 1-log improve- ment in sensitivity, compared with the 488 nm wavelength, which was also confirmed by fluorescence spectroscopy showing a 10-fold higher emission level when mitoxantrone was excited at 567 nm rather than at 488 nm.

Analysis of MDR-1 mRNA by PCR ampliﬁcation using MDR-1-speciﬁc primers

To ensure that our cell lines did not change MDR-1 expression Figure 1 Excitation spectra of mitoxantrone in aqueous solution. status on prolonged culture, we directly measured MDR-1 by Sensitivity at 568.2 nm was six-fold higher than at 488 nm (a). Emis- RT-PCR. Results are shown in Figure 2. A sensitive (8226/s) sion spectra of mitoxantrone excited at 568.2 nm (———) and and two MDR (8226/DOX6 and 8226/DOX40) cell lines were 488 nm (------) (b). Fluorescence maximum was 20-fold higher after included as controls. MDR-1 expression was undetectable in excitation at 568.2 nm as compared with 488 nm. 8226/s and elevated in 8226/DOX6 and 8226/DOX40 cell lines. A strong band was detected for MDR-1 mRNA in HL- of p-glycoprotein. No p-glycoprotein-related ﬂuorescence was 60DOX cells, and no band was present in HL-60s and HL- detected in HL-60MX2 or HL-60s cell lines (D value of 0.02 60MX2 cells. and 0.03, respectively). Background ﬂuorescence was determined by incubation with an isotype control.

Flow cytometric analysis of p-glycoprotein expression Effect of verapamil on mitoxantrone uptake and To relate protein expression to MDR-1 mRNA, we measured retention in HL-60/DOX cells p-glycoprotein of these cell lines by ﬂow cytometry. Figure 3 shows that HL-60DOX cells incubated with MoAb 4E3 were We studied the functional integrity of the drug pump by strongly ﬂuorescent (D value of 0.98), indicating the presence adding verapamil to our drug incubation mixtures. Figure 4 P-glycoprotein and mitoxantrone U Consoli et al 2069 the intracellular content of mitoxantrone was observed in the presence of verapamil during uptake and retention studies for HL-60MX2 (Figure 5) and HL-60s (Figure 6).

Cytotoxicity studies

The MTT dye reduction assay was used to determine the cytotoxicity of mitoxantrone in p-glycoprotein-positive and p-glycoprotein-negative cell lines. Resistant index (RI) is expressed

as a ratio of the ID50 for HL-60MX2 or HL-60DOX cells vs HL-60s cells (Table 1). HL-60MX2 and HL-60DOX were more resistant than HL-60s cells (RI 28.9 and 15.9, respectively). In HL-60s, mitoxantrone cytotoxicity was not modified in the ± presence of verapamil (ID50 0.49 0.03 without verapamil vs ± Figure 2 Analysis of MDR-1 in HL-60DOX, HL-60MX2 and HL- ID50 0.50 0.02 with verapamil) and was only partially 60s cell lines. Sensitive (8226/S) and multidrug-resistant cell lines ± reduced for HL-60MX2 (1.4-fold) (ID50 14 0.7 without vera- (8226/DOX6, 8226/DOX40) were used as controls. ± pamil vs ID50 10 1.0 with verapamil). Mitoxantrone cytotoxicity in p-glycoprotein-positive HL-60DOX cells increased approximately two-fold by concomitant incubation with vera- shows the results of uptake and retention studies in HL- ± ± 60DOX cells after exposure of 5 ␮g/ml of mitoxantrone alone pamil (ID50 7.8 0.4 without verapamil vs ID50 4.2 0.2 with verapamil). Verapamil alone showed no toxicity in any cell or in combination with verapamil 10 ␮M. Cells incubated in drug-free medium with or without verapamil, were used as lines tested (data not shown). controls. In HL-60DOX cells, incubation with mitoxantrone and verapamil resulted in a 3.1-fold increase in mitoxantrone uptake when compared with the same cells incubated in the Confocal microscopy absence of verapamil. During efflux studies, the presence of verapamil resulted in a 2.4-fold increase in mitoxantrone- Optical section of interphase HL-60s cells (Figure 7a and b) related fluorescence retention maintaining the same level revealed a predominantly cytoplasmic and perinuclear depo- reached during uptake in the presence of mitoxantrone. sition of mitoxantrone after 30 min of incubation (Figure 7a) with an increased concentration within the nuclear structures after 90 min (Figure 7b). Cytoplasmic mitoxantrone fluor- Effect of verapamil on mitoxantrone uptake and escence was predominantly detected in a diffuse distribution retention in HL-60MX2 and HL-60s cells pattern with partial accumulation in the perinuclear zone and the Golgi area. Intense staining was noticed in the nuclear The cellular pharmacology of mitoxantrone in HL-60MX2 and membrane and in nucleolar-like structures. This intracellular HL-60s cells was studied as described above. No increase in distribution was not modified by verapamil (data not shown).

Figure 3 P-glycoprotein expression in HL-60s (B), HL-60MX2 (D) and HL-60DOX (F) cells using ﬂow cytometry analysis with the MoAb 4E3 and a secondary goat-anti-mouse FITC-conjugated antibody. An isotype IgG2a was used as control (A, C, E). The x-axis had a range of 4 log. P-glycoprotein and mitoxantrone U Consoli et al 2070

Figure 4 In vitro cellular uptake and efﬂux in human acute leukemia cell line HL-60s. 106 cells/ml were incubated for 120 min with 5 ␮g/ml mitoxantrone with (——G——) and without (-----Ì-----) verapamil. After 120 min of uptake, every sample was reincubated in drug-free medium with (——G——, ——I——) and without (-----H-----, -----Ì-----) verapamil for 120 min. Control experiments using medium alone and verapamil alone showed curve overlap.

Figure 5 In vitro cellular uptake and efﬂux in human acute leukemia cell line HL-60MX2. 106 cells/ml were incubated for 120 min with 5 ␮g/ml mitoxantrone with (——G——) and without (-----Ì-----) verapamil for 120 min of uptake. Every sample was reincubated in drug-free medium with (——G——, ——I——) and without (-----H-----, -----Ì-----) verapamil for 120 min. Control experiments using medium alone and verapamil alone showed curve overlap.

HL-60MX2 cells (Figure 7c and d) showed very low nuclear (Figure 7e). Incubation with verapamil resulted in a consistent fluorescence, and no nuclear membrane staining was increase in mitoxantrone-related fluorescence with distri- observed, as compared with HL-60s. We observed reduced bution of the drug within the nuclear structure and the cyto- cytoplasmic fluorescence with condensation of mitoxantrone plasm. Although mitoxantrone distribution in the presence of in highly fluorescent cytoplasmic inclusions (Figure 7c). The verapamil was similar to that obtained for HL-60s cells, few addition of verapamil did not result in modification of intra- highly fluorescent cytoplasmic inclusions were maintained cellular drug distribution in these cells (Figure 7d). HL-60DOX (Figure 7f). cells (Figure 7e and f) after 90 min showed minimal mitoxantrone-related fluorescence in all cell compartments, with only a few highly fluorescent inclusions present in the cytoplasm P-glycoprotein and mitoxantrone U Consoli et al 2071

Figure 6 In vitro cellular uptake and efﬂux in human acute leukemia cell line HL-60DOX. 106 cells/ml were incubated for 120 min with 5 ␮g/ml mitoxantrone with (——G——) and without (-----Ì-----) verapamil. After 120 min of uptake, every sample was reincubated in drug- free medium with (——G——, ——I——) and without (-----H-----, -----Ì-----) verapamil for 120 min. Control experiments using medium alone and verapamil alone showed curve overlap.

Table 1 In vitro cytotoxicity of mitoxantrone with and without ver- lines can be cross-resistant to mitoxantrone,30 and agents that apamil in HL-60S, HL-60DOX and HL-60MX2 cells circumvent p-glycoprotein have been used to increase mitoxantrone cellular concentration in murine Friend leukemia31 ␮ b c + Cell lines ID50 ( g/ml) RI RI VER and cytotoxicity in p-glycoprotein-positive K562 cells.32 These findings strongly suggest an involvement of p-glycoprotein in + MITOX MITOX VER mechanisms of resistance to mitoxantrone, but the role of p- glycoprotein in cellular pharmacology and cytotoxicity of ± ± HL-60DOX 7.8 0.4 4.2 0.2 15.9 8.4 mitoxantrone has not been well defined. Several reports have HL-60MX2 14 ± 0.7 10 ± 1.0 28.5 20.0 HL-60S 0.49 ± 0.03 0.50 ± 0.02 been published on the cellular pharmacology of mitoxantrone, but few have investigated the distribution of drug mol- 33 34 aCytotoxicity was determined after 1 h exposure of cells to different ecules within intact, living cells. Smith et al showed the concentrations of drugs and 72 h incubation in drug-free media, feasibility of using confocal microscopy to analyze the subcel- using MTT assay as described in Materials and methods. All data lular distribution of mitoxantrone, but they obtained only low are mean ± s.d. of three independent experiments. levels of mitoxantrone fluorescence in neutral aqueous sol- b Values for ID50 were determined from the plot of cell survival frac- ution with an excitation at 514 nm, as we also found (Figure tions vs drug concentrations. c 1). Smith et al attempted to correct this problem by absorbing RI (resistance index) represent the ratio between the ID50 of the drug-resistant and parental cell line. the drug on to a silica matrix and by using high intensity laser power with a high photomultiplier gain setting. In this study, we optimized the spectral conditions under which mitoxan- Discussion trone uptake, retention and distribution were measured by using the 568.2 nm laser line for excitation and 620 nm long Development of cellular resistance to anticancer drugs limits pass filter for fluorescence collection. These conditions did the activity of chemotherapy and is a pivotal clinical problem. not require excessive laser power or a high voltage photomul- The limited data available on mitoxantrone resistance suggest tiplier and resulted in a 1-log enhancement in sensitivity when that this drug selects an atypical resistance profile and displays compared with drug binding studies using 514 nm or lower only partial cross-resistance with other intercalating agents. In wavelength excitation. Under these conditions, we were able tumor cell lines selected for mitoxantrone resistance, various to apply to mitoxantrone a reliable flow cytometric method mechanisms including changes in topoisomerase II,27 different that has been used extensively in cellular pharmacology of drug compartmentalization,28 differences in glutathione S- other anthracyclines, for which the 488 nm line laser is a suit- transferase,29 and increased p-glycoprotein17 and non p-glyco- able source for excitation.19,26,35 Combining these methods protein-mediated drug efflux14 have been demonstrated either with cytotoxicity studies, we demonstrated that verapamil separately or in combination and are postulated to have a role could increase mitoxantrone uptake and retention in HL- in cross-resistance with other structurally unrelated drugs. 60DOX cells that displayed the MDR phenotype related to the Unlike the anthracyclines, vinca alkaloids and natural pro- presence of p-glycoprotein and that an increase in mitoxan- ducts that usually select for the MDR-1 phenotype, tumor cells trone content was associated with a two-fold enhanced cyto- selected for resistance to mitoxantrone do not always over- toxic effect. We also studied HL-60MX2 cells to analyze the express p-glycoprotein. However, p-glycoprotein-positive cell cellular pharmacology in a mitoxantrone-resistant and p-gly- P-glycoprotein and mitoxantrone U Consoli et al 2072 microscopic images demonstrated the different patterns of mitoxantrone distribution in HL-60s cells (diffuse) and HL- 60MX2 cells (punctate), confirming the involvement of a mechanism of resistance that increased trapping of the drug in intracellular organelles, primarily in the Golgi apparatus and in lysosomes.34 Interestingly, in HL-60DOX cells, more than one mechanism of resistance seemed to be present as we observed both sequestration in highly fluorescent intracytoplasmic vesicles and restoration of a diffuse pattern of distribution when p-glycoprotein was blocked by verapamil. Both mechanisms of intracytoplasmic sequestration and p-glycoprotein probably facilitate cell survival by decreasing drug concentration in the cytosol and nucleoplasm, but only p-glycoprotein is able to reduce the total amount of drug-related fluorescence within the cell. Compared with the other two cell lines, HL-60DOX showed a consistent decrease of the mitoxantrone-related fluorescence. This decrease was observed in both flow cytometric and confocal microscopy studies and in both cases was reverted by simultaneous incubation with a p-glycoprotein blocker. However, increased levels in intracytoplasmic mitoxantrone were not accompanied by a comparable increase in its cytotoxicity. The lack of linear correlation between increase in drug-related fluorescence and cytotoxicity in the presence of a p-gp blocker, observed also for doxorubicin and idarubicin,26 can be explained by the presence of additional mechanisms of resistance and especially by mechanisms that increase trapping of the drug in intracellular organelles as observed by confocal microscope. We also analyzed our cell lines for the presence of the multidrug resistance-associated protein MRP,36 and LRP-56,37 which are not reversed by verapamil38,39 but no differences were observed between sensitive and resistant cell lines (data not shown). Recently, we also suggested that expression of genes related to the bcl-2 family in HL-60 is altered during development of anthracycline resistance,40 which further Figure 7 Confocal microscope analysis of mitoxantrone fluor- implicates multiple distinct mechanisms acting at various escence distribution in optical section of interphase HL-60s cells. Mitoxantrone-related fluorescence was visualized live after 30 (a) and points in conferring resistance to chemotherapy. Confocal 90 (b) min. Mitoxantrone after 30 min was first detected in the cyto- microscopy confirmed the increased intracellular mitoxan- plasmic area with a diffuse pattern of distribution and a distribution trone-related fluorescence in HL-60DOX cells when incu- in the nuclear membrane, and nucleolar-like structures were noticed bated in the presence of verapamil. A different pattern of intra- after 90 min. In HL-60MX2 cells, mitoxantrone was present in highly cellular mitoxantrone distribution was observed in sensitive fluorescent cytoplasmic inclusions, and a reduced distribution in the nuclear structures was also observed (c). This distribution was not and resistant cells, with cytoplasmic punctate distribution modified by the presence of verapamil (d). Intracellular drug distri- observed in HL-60MX2 and HL-60DOX cells. Despite this bution after 90 min in HL-60DOX cells was minimal and mostly lim- finding, blocking drug efflux alone was followed by only a ited to the cytoplasm, with a punctate pattern localization (e). Incu- partial sensitization of p-glycoprotein-positive cells to chemo- bation of the cells in the presence of verapamil determined a therapy, indicating that p-glycoprotein represents only one of consistent increase in mitoxantrone accumulation and a diffuse pat- the mechanisms responsible for mitoxantrone resistance. tern of cytosol distribution (f). Although this distribution was similar to that of HL-60s cells, highly fluorescent cytoplasmic inclusions still Although we demonstrated a role of p-glycoprotein in the persist, indicating the possible coexistence of different mechanisms cellular pharmacology of mitoxantrone, different mechanisms of resistance. including drug compartmentalization, changes in topoisomerase II, and drug detoxification may primarily contribute to the coprotein-negative cell line that was selected for resistance to clinical onset of resistance to mitoxantrone. mitoxantrone and is cross-resistant to other drugs that inhibit We report herein a reliable method to study the cellular topoisomerase II.27 In agreement with the results of Harker et pharmacology and distribution of mitoxantrone within intact, al13 we found no p-glycoprotein overexpression by PCR and living cells. Our results demonstrate that the blockade of p- MoAb in this cell line. As expected, uptake, retention and glycoprotein-related drug efflux increased the concentrations cytotoxicity of mitoxantrone in HL-60MX2 cells were mini- of mitoxantrone in p-glycoprotein-positive cells suggesting mally modified by verapamil. that mitoxantrone is a substrate for p-glycoprotein. However, Although this cytofluorometric approach reflects the intra- cytotoxicity was only partially improved by a p-gp blocker cellular mitoxantrone content in single cells, no information and other mechanisms of resistance were reported. These can be obtained by flow cytometry regarding the subcellular observations suggest the presence of a complex cellular mech- distribution of the drug. Therefore, confocal microscopy under anism responsible for mitoxantrone resistance which include optical conditions similar to those used in flow cytometry was intracellular uptake and efflux, formation of membrane ves- used to study the intracellular drug distribution. Confocal icles and drug compartmentalization. P-glycoprotein and mitoxantrone U Consoli et al 2073 Acknowledgements C, Cowan KH. Reduced intracellular drug accumulation in the absence of P-glycoprotein (mdr1) overexpression in mitoxantrone- resistant human MCF-7 breast cancer cells. Cancer Res 1992; 52: We wish to thank Dan Stowt, for his assistance in the 6175–6181. measurements of mitoxantrone fluorescence spectra. This 19 Berman E, Adams M, Duigou-Osterndorf R, Godfrey L, Clarkson work was supported by grants No. CA 57639 and CA 16672 B, Andreeff M. Effect of tamoxifen on cell lines displaying the mul- from the National Institutes of Health. UC is in part supported tidrug-resistant phenotype. Blood 1991; 4: 818–825. by a scholarship from the Fondazione Catanese per lo Studio 20 Priebe W, Van TN, Burke TG, Perez-Soler R. Removal of the basic e la Cura delle Malattie Neoplastiche del Sangue (Catania, center from doxorubicin partially overcomes multidrug resistance Italy). and decreases cardiotoxicity. Anti-Cancer Drug 1993; 4: 37–48. 21 Marie JP, Faussat-Suberville AM, Zhou D, Zittoun R. Daunorub- icin uptake by leukemic cells: correlation with treatment outcome and mdr-1 expression. Leukemia 1993; 7: 825–831. References 22 Ross DD, Wooten JP, Sridhara R, Ordonez JV, Lee ED, Schiffer CA. Enhancement of daunorubicin accumulation, retention, and cytotoxicity by verapamil or cyclosporin A in blast cells from 1 van der Graaf WTA, de Vries EGE. Mitoxantrone: bluebeard for patients with previously untreated acute myeloid leukemia. Blood malignancies. Anti-Cancer Drugs 1990; 1: 109–125. 1993; 82: 1288–1299. 2 Albert DS, Peng Y, Bowden GT, Dalton WS, Makel C. Pharma- cology of mitoxantrone: mode of action and pharmacokinetics. 23 Tsuruo T, Kawabata H, Nagumo N, Iida H, Kitatani Y. Potentiation Invest New Drugs 1985; 3: 101–107. of antitumor agents by calcium channel blockers with special ref- 3 Henderson IC, Allegra JC, Woodcock T, Wolff S, Bryan S, Cart- erence to cross-resistance patterns. Cancer Chemother Pharmacol wright K, Dukart G, Henry D. Randomized clinical trial comparing 1985; 15: 16–19. mitoxantrone with doxorubicin in previously treated patients with 24 Chomcyznski P, Sacchi N. Single-step method of RNA isolation by metastatic breast cancer. J Clin Oncol 1989; 7: 560–571. acid guandium isothiocyanate-phenol-chloroform extraction. Anal 4 Bezwoda WR, Bernasconi C, Hutchinson RM, Winfield DA, Biochem 1987; 102: 156–159. deBock R, Mandelli F. Mitoxantrone for refractory and relapsed 25 Drach D, Zhao S, Drach J, Mahadevia R, Gattringer C, Huber H, acute leukemia. Cancer 1990; 66: 418–422. Andreeff M. Subpopulation of normal peripheral blood and bone 5 Arlin ZA, Case DC Jr, Moore J, Wiernik P, Feldman E, Saletan S, marrow cells express a functional multidrug resistance phenotype. Desai P, Sia L, Cartwright K. Randomized multicenter trial of cyto- Blood 1992; 80: 2729–2734. sine arabinoside with mitoxantrone or doxorubicin in previously 26 Consoli U, Priebe W, Ling YH, Mahadevia R, Perez-Soler R, And- untreated adult patients with acute nonlymphocytic leukemia reeff M. Annamycin, a novel anthracycline, is not affected by p- (ANLL). Leukemia 1990; 4: 177–183. glycoprotein related MDR: comparison with doxorubicin and ida- 6 Bajetta E, Buzzoni R, Valagussa R, Bonadonna G. Mitoxantrone: rubicin. Blood 1996; 88: 633–644. an active agent in refractory non-Hodgkin’s lymphomas. Am J Clin 27 Harker WG, Slade DL, Drake FH, Parr RL. Mitoxantrone resistance Oncol Cancer Clin Trials 1988; 11: 100–103. in HL-60 leukemia cell: reduced nuclear topoisomerase II catalytic 7 Arlin ZA, Berman E, Jhanwar S, Gams R, Schoch I, Dukart G. activity and drug-induced DNA cleavage in association with Treatment of blastic transformation of chronic myelogenous leuke- reduced expression of the topoisomerase II beta isoform. Bio- mia with mitoxantrone. Cancer Chemother Pharmacol 1986; 17: chemistry 1991; 30: 9953–9961. 189–190. 28 Reymann A, Arps H, Dietel M, Schaefer B, Woermann C. Reduced 8 Smith PJ, Morgan SA, Fox ME, Watson JV. Mitoxantrone-DNA mitoxantrone contents in a mitoxantrone-resistant gastric carci- binding and the induction of topoisomerase II associated DNA noma cell line unrelated to the multidrug-resistance phenotype. damage in multi-drug resistant small cell lung cancer cells. Naunyn-Schmiedeberg’s Arch Pharmacol 1990; 341: 41 (Abstr.). Biochem Pharmacol 1990; 40: 2069–2078. 29 Peters WH, Roelofs HMJ. Biochemical characterization of resist- 9 Harris AL, Hochhauser D. Mechanisms of multidrug resistance in ance to mitoxantrone and adriamycin in Caco-2 human colon cancer treatment. Acta Oncol 1992; 31: 205–213. adenocarcinoma cells: a possible role for glutathione s-transferase. 10 Capolongo L, Belvedere G, D’Incalci M. DNA damage and cyto- Cancer Res 1992; 52: 1886–1890. toxicity of mitoxantrone and doxorubicin in doxorubicin-sensitive 30 Dalton WS, Durie BGM, Albert DS, Gerlach JH, Cress AE. Charac- and -resistant human colon carcinoma cells. Cancer Chemother terization of a new drug-resistant human myeloma cell line that Pharmacol 1990; 25: 430–434. expresses p-glycoprotein. Cancer Res 1986; 46: 5125–5130. 11 Jensen PB, Vindelov L, Roed H, Demant EJ, Sehested M, 31 Reymann A, Looft G, Woermann C, Dietel M, Erttmann R. Rever- Skovsgaard T, Hansen HH. In vitro evaluation of the potential of sal of multidrug resistance in Friend leukemia cells by dexniguldi- aclarubicin in the treatment of small cell carcinoma of the lung pine-HCl. Cancer Chemother Pharmacol 1993; 32: 25–30. (SCCL). Br J Cancer 1989; 60: 838–844. 32 Genne P, Duchamp O, Solary E, Pinard D, Belon JP, Dimanche- 12 Dalton WS, Cress AE, Alberts DS, Trent JM. Cytogenetic and Boitrel MT, Chauffert B. Comparative effect of quinine and cincho- phenotypic analysis of a human colon carcinoma cell line resistant nine in reversing multidrug resistance on human leukemic cell to mitoxantrone. Cancer Res 1988; 48: 1882–1888. line K562/ADM. Leukemia 1994; 8: 160–164. 13 Harker WG, Slade DL, Dalton WS, Meltzer PS, Trent JM. Multi- 33 Freedman J, Brandwein J, Kardish M, Ellis J. Detection of multidrug drug resistance in mitoxantrone-selected HL-60 leukemia cells in resistance due to mitoxantrone by flow cytometry. Blood 1992; the absence of P-glycoprotein overexpression. Cancer Res 1989; 80: 203a (Abstr.). 49: 4542–4549. 34 Smith PJ, Sykes HR, Fox ME, Furlong IJ. Subcellular distribution 14 Dietel M, Arps H, Lage H, Niendorf A. Membrane vesicle formation due to acquired mitoxantrone resistance in human gastric car- of the anticancer drug mitoxantrone in human and drug-resistant cinoma cell line EPG85-257. Cancer Res 1990; 50: 6100–6106. murine cells analyzed by flow cytometry and confocal microscopy 15 Henderson IC, Allegra JC, Woodcock T, Wolff S, Bryan S, Cart- and its relationship to the induction of DNA damage. Cancer Res wright K, Dukart G, Henry D. Randomized clinical trial comparing 1992; 52: 4000–4008. mitoxantrone with doxorubicin in previously treated patients with 35 Berman E, McBride M. Comparative cellular pharmacology of metastatic breast cancer. J Clin Oncol 1989; 7: 560–571. daunorubicin and idarubicin in human multidrug-resistant leuke- 16 Neidhart JA, Gochnour D, Roach R, Hoth D, Young D. A compari- mia cells. Blood 1992; 12: 3267–3273. son of mitoxantrone and doxorubicin in breast cancer. J Clin 36 Cole SPC, Bhardwaj G, Gerlach JH, Mackie JE, Grant CE, Almquist Oncol 1986; 4: 672–677. KC, Stewart AJ, Kurz EU, Duncan AMV, Deeley RG. Over- 17 Muller C, Laval F, Soues S, Birck C, Charcosset JY. High cell den- expression of a transporter gene in a multidrug-resistant human sity-dependent resistance and P-glycoprotein-mediated multidrug lung cancer cell line. Science 1992; 258: 1650–1654. resistance in mitoxantrone-selected Chinese hamster cells. 37 Scheper RJ, Broxterman HJ, Scheffer GL, Kaaijk P, Dalton WS, van Biochem Pharmacol 1992; 43: 2091–2102. Heijningen THM, van Kalken CK, Slovak ML, de Vries EGE, van

18 Nakagawa M, Schneider E, Dixon KH, Horton J, Kelley K, Morrow der Valk P, Meijer CJLM, Pinedo HM. Overexpression of a Mr P-glycoprotein and mitoxantrone U Consoli et al 2074 110 000 vesicular protein in non-p-glycoprotein-mediated multid- modulation in acute myeloid leukemia. Leukemia 1996; 10: rug resistance. Cancer Res 1993; 53: 1475–1479. 36–38. 38 Slapak CA, Mizunuma N, Kufe DW. Expression of the multidrug 40 Zhao S, El-Tounsi I, Lin Y, Andreeff M. Increased expression of resistance associated protein and p-glycoprotein in doxorubicin- bcl-2 family gene in HL-60 doxorubicin resistant cell line and selected human myeloid leukemia cells. Blood 1994; 84: 3113– acute myeloid leukemia cells treated with chemotherapy. Blood 3121. 1995; 86: 2063a (Abstr.). 39 List AF. The role of multidrug resistance and its pharmacological