In Resistance to Adriamycin and Taxol by the K562 Leukemia Cell Line

[CANCER RESEARCH56. 3461-3467, August 1, 19961 Proton Nuclear Magnetic Resonance Spectroscopy Reveals Cellular Lipids Involved in Resistance to Adriamycin and Taxol by the K562 Leukemia Cell Line

Laurence Le Moyec,' Roger Tatoud, Armelle Degeorges, Cynthia Calabresse, Guy Bauza, Michel Eugene, and Fabien Calvo

Laboratoire ik RMN et Exploration Fonctionnelle IL L M., G. B., M. E.] and Laboratoire de Pharmacologie ExpÃ©rimentale [R. T., A. D., F. C.], HÃ´pital Saint Louis. 27 rue Juliette Dodu, and Laboratoire de Biologic Cellulaire HÃ©matopolÃ©tique.Institut Universitaire d'HÃ©matologie, I avenue Claude Vellefaux (C. C.]. 75010 Paris, France

ABSTRACT modification of the plasma membrane structure (1). In the case of Taxol, which normally inhibits microtubule disaggregation, abnormal Proton nuclear magnetic resonance spectroscopy was performed on tubulin requiring Taxol to polymerize has been evidenced (4) in whole cells to study lipids and metabolites in Adriamycin- and Taxol Taxol-resistant cells. resistant K562 cells expressing multidrug resistance (MDR) and their sensitive counterparts. With one-dimensional spectra, both resistant cell The plasma membrane plays a key role in the control of intracel lines showed lower fatty acid methylene:methyl ratios and higher choline: lular concentration, and its lipid composition may alter drug efflux methyl ratios than sensitive cells. Using two-dimensional COSY spectra, a and influx. Membrane lipid composition of sensitive and resistant decrease In the glutamine content was evidenced in resistant cells. When tumor cells has been studied, in particular, with anthracycline antibi these cells were maintained in culture mediwn without the drug, the fatty otics (Adriamycin and daunorubicin). These DNA-targeting drugs acid signals were partially recovered. Adriamycm-resistant K562 cells also have membrane effects and, when linked to polymers carriers, were also treated for 4 days with a high dose of verapamil, a MDR can be active without entering cells. They bind electrostatically to reversing agent. The nuclear magnetic resonance spectra of vernpamil membrane lipids and to negatively charged phospholipids via weaker treated cells also showed partial recovery of fatty acid signals. These hydrophobic interactions. Adriamycin interacts with membranes by results could be paralleled with the reversion of the resistant phenotype, as evidenced by measuring the inhibiting concentration of Adriamycin increasing the turnover of phosphatidylinositol (5). Thus, modifica and vinblastine in K562adr cells cultured without the drug or after tions in membrane composition have consequences on cell surface short-term exposure to verapamil. Conversely, P-glycoprotein and mRNA charges and on membrane fluidity, altering drug binding and drug expression and DNA amplification of the mdr gene were not modified uptake. The nature of these lipid modifications has been studied in when compared to resistant cells, suggesting that the MDR phenotype several works, but a general lipid profile for resistant cells has not yet could be partially reversed independently of the mdr gene amplification come forth. For example, P388 daunorubicin-resistant cells bind less and expression. These results demonstrate the role of lipids in the resist drug than their sensitive counterparts (6). This result was related to a ance phenomenon. lower phosphatidylserine content and a higher cholesterol content in drug-resistant cells. In human leukemic lymphoblasts, the develop INTRODUCTION ment of vinblastine resistance was associated with an increase in ether lipids, cholesterol, and phospholipids (7). Another way to investigate The term MDR2 phenomenon describes cellular expression of the relationship between lipids and resistance is to modify the lipid resistance to drugs without necessarily related origin or chemical membrane composition. In human ovarian cancer cells, Adriamycin similarities (for review, see Refs. 1â€”3).TheMDR is related to a lower resistance can be increased by lowering the membrane cholesterol intracellular concentration of drugs. The most widely evoked and content when cells are cultured in medium without lipoproteins, and studied mechanism is the presence, in the membrane of MDR cells, of this effect is reversed by vp (8). Polyunsaturated fatty acids were a P-gp (or gp-170), working as an AlP-dependent pump to increase shown to decrease the vincristine resistance of human neuroblastoma the drug efflux. This P-gp is encoded by a family of mdr genes cells by changing the phospholipid membrane composition (9) and overexpressed in resistant cell lines. that of human cervical carcinoma (10) by increasing membrane flu However, other mechanisms are also responsible for the decrease in idity, which was also related to increased sensitivity of resistant drug accumulation. Drug uptake may be reduced, intracellular degra Ll2lO cells (1 1). The mechanism may not simply be a matter of drug dation of drugs may be increased, and metabolism into active corn diffusion since the use of known membrane fluidizers decreases drug pounds may be altered (1). Besides, drug effects may be altered by uptake by P-gp in rat liver canalicular membrane vesicles (12). In modification of cellular targets (enzyme, protein, or DNA) and by Ehrlich ascites tumor cells, the level of resistance was reported to be improvement of DNA repair. These mechanisms may coexist with dependent on the level ofphospholipids of the plasma membrane (13). P-gp and with each other, depending on the type of drug used to Proton NMR spectroscopy is able to detect the mobile fraction of induce resistance. In the case of Adriamycin resistance, several mech lipids contained in the cell, and thus is sensitive to membrane fluidity anisms were described: increased enzyme activities such as that of the modifications related to lipid composition changes (14â€”16).Since the enzyme responsible for detoxification by glutathione, decreased ac chemical analysis of membrane lipids reflects the quantity of different tivity of the nuclear enzyme topoisomerase II, decreased transforma lipids, NMR spectroscopy can provide an assessment of the physical tion of the drug into semiquinone, which can react with oxygen to state of these lipids and detect changes in membrane fluidity in form active oxygen radicals, and reduced intracellular diffusion by relation to drug resistance. With this aim, we studied the human leukemia K562 cell line using Received 3/4/96; accepted 5/28/96. The costs of publication of this article were defrayed in part by the payment of page proton NMR spectroscopy and compared it to Adriamycin- (K562adr) charges. This article must therefore be hereby marked advertisement in accordance with and Taxol- (K562tax) resistant cell lines, both overexpressing the mdr 18 U.S.C. Section 1734 solely to indicate this fact. gene. Because the lipids detected using NMR were affected in resist I To whom requests for reprints should be addressed. @ 2 abbreviations used are: MDR, multidrug resistance; P-gp, P-glycoprotein; NMR, ant cells, we altered resistance by maintaining the cells in medium nuclear magnetic resonance; Glu, glutamine-glutamate; Ct, creatine-creatinine; Chol, without the drug and measured the consequences on NMR spectra. On choline and derivative compounds; lao, inositol; Eth, ethanolamine; vp. verapamil; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; IC,0, 10% inhibitory concentra K562adr, the consequences of resistance reversion using this method tion. and short-term exposure to high doses of vp were investigated with 3461

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 1996 American Association for Cancer Research. PROTON NMR OF MDR K562 CELLS proton spectroscopy, measurement of drug sensitivity, and expression denaturing gel containing 2% formaldehyde. RNAs were transferred to an of the mdr gene (mRNA, P-gp, and DNA). Amersham Hybond C membrane. Hybridization was performed at 42Â°Cfor 12 h in buffer containing 50% (v/v) formamide, lx Denhardt's solution, 5X SSC,2.5 mMsodiumphosphate,0.25mg/mIsingle-strandherringspermDNA, MATERIALS AND METHODS and 10% dextran sulfate, with the random priming 32P-labeled probe (Multi prime DNA labeling system; Arnersham). After removing unbound probe, the Cell Culture. The K562wtcell line is a humanleukemiacell line (17). The filters were exposed at â€”80Â°ConKodak X-OMAT AR film with intensifying Adriamycin-resistant derived cell line K562adr was a gift from Dr. T. Tsuruo screens. (Japanese Foundation for Cancer Research, Tokyo, Japan) and the Taxol Probes used were: mdrl, 600 bp of pCHP1 genomic fragment (provided by resistant derived cell line K562tax was a gift from C. TempÃ©te(Institut de Dr. V. Ling, Ontario Cancer Institute, Toronto, Ontario, Canada); p53, 1.3-kb Chimie des Substances Naturelles, Gif sur Yvette, France). PvuH cDNA fragment from the p53-H-8 clone (20); and GAPDH, 1.2-kb Pst! K562 cells were cultured in RPMI 1640 (Sigma, St. Quentin Fallavier, fragment from rat GAPDH (21). France) supplemented with 10% FCS (ATGC, Noisy le Grand, France) and 4 Cytofluorimetry. The monoclonal antibody MRK16 is a murine IgG2a mM glutamine (Sigma). The resistant cells were cultured in the same medium developed and kindly offered by Dr. H. Hamada and Dr. T. Tsuruo (22). This containing 0.5 ,.@g/mlAdriamycin(Farmitalia Carlo Erba, Rueil Malmaison, antibody reacts with an extracellular determinant ofthe P-gp that is specifically France) for K562adr and 0.8 @g/mlTaxol (a gift from Dr. F. Picot, Institut de expressed on the plasma cell membrane of multidrug-resistant cells. Cells were Chimie des Substances Naturelles) for K562tax. incubated for 1 h at 4Â°Cwith MRK16 (10 @&g/ml)orwith an !gG2a control (10 For resistance reversion experiments, K562adr and K562tax cells were @g/ml).Theywere washed with cold PBS and incubated for an additional 30 cultured in medium without the drug and respectively labeled K562ads and mm at 4Â°Cin darkness with fluorescein-conjugated goat anti-mouse IgG2a. K562tas along with the number of passages without the drug (K562ads-pn or After two washes with cold PBS, cells were resuspended in cold PBS and K562tas-pn). vp experiments were performed by addition of vp (Sigma) to analyzed with an Epics Profile Analyzer (Coulter Corporation). Results were K562adr cells (K562adr vp) in the culture medium without Adriamycin. The expressed as a percentage of fluorescent cells compared to their control. The toxicity of vp toward K562wt and K562adr was determined with the method analysis was repeated three times. described below on one experiment performed in triplicate; IC10 and IC50 were Cellular Resistance to Adriamycin and Vinblastine. A total of 8 X i0@ used for experiments. cells were plated in 96-well microtiter culture plates for 96 h at 37Â°Cin 7% Cells were mass cultured under the same conditions after seeding l0@ CO2.The quantificationof survivingcells with increasingconcentrationsof cells/ml in 15 ml containing flasks at day 0 and were used at day 3 (experi Adriamycin or vinblastine and without drug was determined by measurement ments with Taxol) or day 4 (experiments with Adriamycin) for NMR exper of the endogenous enzyme hexosaminidase (18). The experiments were per iments; 50â€”90x 106 cells were collected and washed three times with 2 ml formed in triplicate, and the points obtained on six to eight experiments for the PBS (Sigma)indeuteriumoxide(D20; Sigma).Thefinalpelletwas suspended same drug and same cell line were pooled to fit the theoretical equation of a in PBS/D20 to obtain a final volume of 0.5 ml and placed in a 5-mm NMR single action site of a drug: y = 100 X IC@J(IC50 + x), where y is the tube. percentage of surviving cells and x is the drug concentration. The values NMR Spectroscopy. Spectra were recorded on a Bruker (Wissembourg, obtained on K562wt, K562ads, and K562adr vp were compared using variance France) AM400WB spectrometer at 400 MHz. The signal from residual water analysis, and P < 0.05 was considered as significant. was suppressed by the presaturation technique with an irradiation of 0.08 W for 2 s. Resonances chemical shifts were expressed in ppm, in reference to TSP RESULTS (sodium trimethylsilylpropionate, external) assigned to 0 ppm. One-dimensional experiments were performed at 20Â°Cwith a 60Â°flip angle, NMR Spectra. The one-dimensional spectra obtained with and 64 transients were accumulated. Acquisition time was 0.68 s on 8000 data K562wt, K562adr, and K562ads-p84 for an equivalent number of points corresponding to a sweep width of 6000 Hz. The Fourier transformation cells (Fig. 1) demonstrated differences in lipid intensity between was performed after a zero filling to 16,000 data points and exponential multiplication corresponding to 1-Hz line broadening. Two-dimensional COSY experiments were performed with 1000 data points in the F2 direction and 256 data points in the Fl direction. The sweep width was reduced to 2700 Hz, eliminatingthe aromaticregionof the spectrumwhereno resonanceswere detected in one-dimensional acquisitions. Thus, the experiments could be Chol performed within 3 h. The two-dimensional Fourier transformation was ap Glu CH3 plied after zero filling to 512 data points in the Fl direction and a sine-bell function in both directions. Each experiment consisted of a one-dimensional acquisition, a two-dimen sional COSY spectrum, and a control one-dimensional spectrum. Peak assign ments were made based on data from the literature and using spectra obtained on standards. The peak areas were measured using manual integration in WINNMR software (Broker). The groups of resonances measured included: methyl (CH3) resonances from fatty acids integrated between 0.75 and 1.10 ppm; methylene resonances (CH2) from fatty acids between 1.15 and 1.5 ppm; Glu between 2.5 and 1.8 ppm; Ct between 2.9 and 3. 10 ppm; N-trimethyl of Chol between 3.3 and 3.1 ppm; the Ino between 3.6 and 3.5 ppm; Eth between 3.8 and 3.7 ppm. The lipid resonance areas were reported to CH3and the metabolite areas to Ct. The peak area ratios were compared by variance analysis, and the Student Newman-Keuls test for multiple comparison differences were considered as significant for P < 0.05. Southern and Northern Blots. High molecular weight DNA was isolated from cells as described (19). Ten @xgDNA digested with EcoR! (Boehringer 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 Mannheim, Meylan, France) were electrophoresed in 0.8% agarose gel and transferred to a Hybond C membrane (Amersham, Les Ulis, France). Total (ppm) cellular RNA was isolated using the guanidine-cesium-chloride method (19). Fig. 1. Aliphatic region of one-dimensional spectra obtained at day 4 on K562wt (top), Twenty @xgoftotal cellular RNA were electrophoresed on a 1% agarose K652adr (middle), and K562ads (bottom). 3462

SD)CellTable 1Peak intensity ratios obtainedfromone-dimensional spectraobtained at day 4 aftertransplantation of IU@cells/mi (mean Â± lineNo. of experiments K652wt 5 3.50 Â±0.56 0.66 Â±0.23 0.43 Â±0.08 8.02 Â±1.17 0.30 Â±0.06 K562adr 8 1.63 Â±0.16a 1.07 Â±0.21â€• l@4 Â±o.23'@ 6.14 Â±0.83Â° 0.30 Â±0.07 K562ads (passages 50â€”120) 7CH@/CH3 2.25 Â±0.l6'@Chol/CH3 0.87 Â±0.24mo/Ct 1.29 Â±0.2l@@cGlu/Ct 6.43 Â±0.49'Eth/Ct 0.31 Â±0.05 aK562wtvs.K562adrsignificantlydifferent. b K562adr vs. K562ads significantly different. C K562wt vs. KS62ads significantly different. sensitive and resistant cells. The integration results are presented in possible to obtain a better assignment of the resonance groups as Table 1. defined in one-dimensional spectra. The resonances located between When compared to K562wt, K562adr cells showed lower CH2/CH3 1.8 and 2.5 ppm arose from glutamate, glutamine, glutathione, and and Glu:Ct ratios, higher Chol:CH3 and Ino:Ct ratios, and an identical methionine in K562wt cells, as seen in Fig. 3. From the two-dimen Eth:Ctratio. sional spectra, the decrease in the Glu:Ct ratio could be attributed to Experiments were performed on K562ads (passages > 50) and a decrease in the glutamine level, since the cross-peaks of glutamate pooled in the same group as the CH2:CH3 ratio after passage 50 was and glutathione-methionine were saved in K562adr and K562tax outside the 95% confidence interval of the K562adr CH2:CH3 ratio. two-dimensional spectra while a glutamine cross-peak was absent. After 50 passages (up to 120) in medium without Adriamycin, The inositol detected in NMR spectra was mainly myo-inositol, K562ads, when compared to K562adr, showed an increase in the since the concentrations of inositol phosphates are below detectable CH2:CH3 and Ino:Ct ratios, whereas other values remained identical levels. In Fig. 3, cross-peaks of inositol were clearly detected and to K562adr. Nevertheless, the CH2:CH3 and Ino:Ct ratios of K562ads were increased in K562adr and K562tax two-dimensional spectra. were still significantly different from those of K562wt. The choline group around 3.2 ppm could include the N-trimethyl Fig. 2 shows the spectra obtained with K562wt, K562tax, and resonances from choline, phosphorylcholine, and glycerophosphoryl K562tas-p63. In this case also, the lipid intensity differences are choline metabolites of the membrane phosphatidylcholine. In two clearly seen on the spectra acquired on an equivalent number of cells. dimensional COSY spectra, these three molecules gave rise to three Table 2 reports the integration results of the major resonances. different cross-peaks between the two CH2 of the choline molecule. K652tax cells showed a significant decrease in the CH2:CH3, The only cross-peak detectable in our spectra corresponded to the Glu:Ct, and Eth:Ct ratios, a significant increase in the Chol:CH3 ratio, chemical shifts in phosphorylcholine. and no change in the Ino:CH3 ratio when compared to K652wt. K562wt and K562adr cells were treated with vp for 4 days, and the Experiments on K562tas after passage 10 were pooled in the same IC10 and IC50 values to vp were determined. For K562wt, IC10 was 35 group as the CH2:CH3 ratio after passage 10 was outside the 95% .LMand IC50 was 65 @LM.For K562adr, IC30 was 2 p.M and IC50 was confidence interval of the K562tax CH2:CH3 ratio. The CH2:CH3 40 @LM.TheseIC10 values were used for 8 days and the IC50 values for ratio of these K562tas cells was significantly higher than that for the 4 days before NMR analysis; spectra integration results are shown in K562tax cells and also significantly lower than that for the K562wt Table 3. For K562wt, IC10 VP did not modify the spectra, but IC50 cells. The other peak ratios measured were not significantly modified increased the CH2:CH3 and Glu:Ct ratios. For K562adr, the IC10 vp from KS62tax cells. decreased the Ino:Ct ratio without modification of other ratios. With Fig. 3 represents the two-dimensional COSY spectrum obtained IC50 the CH2:CH3 ratio was increased, but other ratios were unmod with K562wt. In this spectrum, fatty acids are responsible for the ified. presence of cross-peaks (a-d) corresponding to functions inside the Cellular Drug Resistance: Measurement of IC@. Cellular resist fatty acid chains, whereas in spectra of both Adriamycin- and Taxol ance to Adriamycin and vinblastine was assayed on K562wt, resistant cells, the cross-peak a was the only correlation detected. K652adr, K562ads-p84, and K562adr treated for 4 days with 40 p.M The two-dimensional COSY spectra improved resolution, making it VP(K562adr vp), and IC50 values are reported in Table 4. IC50 values obtained with K562adr vp were measured after treatment with Vp, but in the absence of vp. The resistance of the K562adr cell line was increased by a factor of 100 when compared to K562wt for both drugs. K562ads cells showed increased sensitivity to both Adriamycin and vinblastine, although !C50s were not decreased to the level observed in K562wt cells. KS62adr vp cells were again more sensitive to Adriamycin and vinblastine, but slightly more resistant to Adriamycin than K562ads. mdr mRNA Expression and DNA Amplification. mdrl gene amplification and P-gp RNA expression were investigated in K562wt, K562adr, K562ads-p84, and K562adr with IC50 VP for 1, 5, and 10 passages. Amplification of the P-gp gene sequences was only detected in restriction digested genomic DNA extracted from the multidrug resistant cell lines (Fig. 4A, Lanes 1â€”4and 6). All variations in the P-gp gene signal for K562ads and K562adr treated with vp oc curred in parallel with a variation in the intensity of the p53 4.0 3.5 3.0 2.5 2.0 1.5 1.0 hybridization signal, reflecting unevenness in DNA loading. (ppm) Southern blot analysis failed to demonstrate a significant variation Fig. 2. Aliphatic region of one-dimensional spectra obtained at day 3 on K562wt (top), in the P-gp gene amplification between K562adr, K562ads, and K652tax (middle) and K562tas (bottom). K562adr vp. 3463

SD)CellTable 2 Peakintensity ratios obtainedfromone-dimensional spectraobtained at day 3 aftertransplantation of 10@cells/in! (mean Â± lineNo. of experiments K562wt 4 2.14 Â±0.09 0.97 Â±0.08 0.48 Â±0.08 8.53 Â±0.63 0.37 Â±0.05 K562tax 3 1.23 Â±o.05a 1.85 Â±0.15a 0.62 Â±0.05 6.96 Â±0.53â€• 0.15 Â±o.o1@z K562tas (passages 10-63) 5CH@/CH3 1.53 Â±021b.cChol/CH3 2.02 Â±0.35ctao/Ct 0.67 Â±0.11Glu/Ct 737 Â±0.49cEth/Ct 0.17 Â±O.05c

a K562wt vs. K562tax significantly different. b KS62tax vs. K562tas significantly different.

C K562wt vs. K562tas significantly different.

Fig. 4B shows the 4.8-kb mdrl gene product in the K562 cell lines. concentration partially recovered the sensitive phenotype and lipid Expression of the mdrl gene could not be detected in the K652wt cell signals. In these cells (K562ads and K652adr vp), mdr gene amplifi line, but could be demonstrated in K562adr, K562ads, and K562adr cation and expression (mRNA and protein) were not decreased. Thus, VP cell lines. Expression of P-gp mRNA was not modified after long the reversed resistance may involve poorer efficiency of the P-gp exposure to vp (Fig. 4B, Lanes 2, 3, and 4) or culture in medium pump or other mechanisms, suggesting that partial reversion of re without Adriamycin (Fig. 4B, Lane 1). sistance was not only related to P-gp, but also to a modification in P-gp Expression. The expression of P-gp was assessed with cellular lipids. MRK16 monoclonal antibody in K562wt, K562adr, K562ads-p84, The membrane origin of lipid signals detected on whole cells has and K562adr with IC50 vp for one passage (K562adr vp). A mouse been widely discussed in the literature. Membrane phospholipids monoclonal antibody was used as a negative control. The three cx presenting restricted molecular motion should produce extremely periments performed showed reproducible results. As shown in Fig. 5, broad resonances. Nevertheless, it has been shown that several ma less than 5% of the K562wt cells were positive for P-gp when lignant cell lines, tissues of tumoral origin, and activated immune cells compared to the control. For the three other cell lines, 75â€”85%of the can produce spectra with elevated fatty acid methyl and methylene cells were positive for P-gp expression with similar levels, demon signals (15, 23, 24). These resonances arise from molecules presenting strating no significant difference among these cell lines. possible isotropic diffusion; thus, they should exist in membrane domains where lipids are relatively free to move (16). It has been DISCUSSION shown that this type of lipid signal can exist without being associated with the presence of intracellular lipid droplets (25), but the presence Proton NMR spectra of Adriamycin- and Taxol-resistant K562 cells of such droplets may also affect the proton spectra. Membrane lipids presented a lower fatty acid CH2:CH3 ratio and a higher Chol:CH3 exist in different cellular sites (mitochondria, nucleus, and plasma ratio than sensitive K562 cells. Both resistant cell lines also showed a membrane), but the plasma membrane seems to be the most probable decreased glutamine:creatine ratio. Parallel to these common differ origin of these signals (24). On mammary adenocarcinoma cell lines, ences, Adriamycin-resistant K562 cells displayed an increased inosi we have shown that the membrane-enriched fractions produced in tol content when compared to creatine, whereas Taxol-resistant cells tense fatty acid resonances for cell lines with the higher lipid signals showed a decrease in ethanolamine. When resistant K562 cells were detected on whole cells (15). maintained in culture medium without a drug, fatty acid signals were Chemical methods have been used to analyze membrane lipids in partially recovered. resistant cells and show modified composition with respect to their K562adr cells cultured without Adriamycin or exposed to a high vp sensitive counterparts (8â€”13).The lipid modifications described may therefore affect the proton NMR spectra by changing the mobility of molecules inside the plasma membrane. In this study, the proton spectra performed on whole K562-sensitive and -resistant cells dem onstrated that the NMR visible fatty acid fraction was lower in Adriamycin- and Taxol-resistant cells than in sensitive cells, whereas choline resonance was increased. The fatty acid signal increase can be related to a change in the quantity of detectable lipids and in the mobility of these lipids. Phospholipids have been widely described in tumors in vitro and in vivo using 31P-NMR spectroscopy (26); phos phorylcholine, which seems to be the major compound contributing to this signal, could contribute to cell proliferation (27). In addition to lipid signal modification, NMR spectra demonstrated that other metabolites were affected along with drug resistance, such as the glutamine and inositol content. Glutamine, glutamate, and glutathione are metabolites used by the cell for detoxification via glutathione conjugation. Glutamine is provided in the culture medium. Its decrease in resistant cells may be due to a higher turnover of these metabolites in the metabolic pathway of glutamine to glutamate to glutathione, to increase glutathione conjugation as already described, by the increase in enzyme activities such as glutathione S-transferase (28). This mechanism contributes to the resistance by decreasing the free drug intracellular content. In Taxol-resistant cells, glutathione 4.0 3.0 2.0 1.0 depletion alters the microtubule synthesis and may account for Taxol resistance (29). Fig. 3. Aliphatic region of two-dimensional COSY spectrum obtained at day 4 on K562wtcells.Glu.glutamate;GIn,glutamine;GSH,glutathione;Met,methionine;Lys, Inositol is also known to participate in detoxification mechanisms, lysine; Taur, Taurine; Cys, cysteine; PChol, phosphorylcholine. a-d, cross-peaks from since it can produce glucuronic acid used for conjugation (30). The fatty acid chains. 3464

Table 3 Peak intensity ratios obtainedfrom one-dimensional spectra obtained at day 4 after transplantation of 1O@cells/miâ€• experimentsCH2/CH3Cbol/CH3Ino/CiGIu/CtEth/CtK652wt53.50Cell lineNo. of 0.06K562wt, Â±0.560.66 Â±0.230.43 Â±0.098.02 Â±1.170.30 Â± â€˜pIC10 Â±0.37 Â±0.09 Â±0.03 Â±0.72 Â±0.07 K562wt,0.14K562adr81.63 VpIC503 33.40 4.64 Â±023b0.66 0.44 Â±0.010.42 0.35 Â±0.049.21 13.55 Â±0.83'@0.24 0.22 Â± 0.07K562adr, Â±0.161.07 Â±0.211.04 Â±0.236.14 Â±0.830.30 Â± VpIC,0 Â±0.33 Â±0.20 Â±0.06c Â±0.75 Â±0.002 K562adr, Vp IC503 31.65 2.54 Â±0.19â€•1.29 1.19 Â±0.060.63 0.95 Â±0.135.56 6.97 Â±0.120.22 0.20 Â±0.01 aIntheabsenceorpresenceofvpfortwopassagesatIC10andonepassageatICso(meanÂ±SD). @ vs. K562wt vp lCso significantly different.

C K562adr vs. K562adr vp IC10 significantly different. @ d vs. K562adr vp IC50 significantly different.

non, K562adr cells were used. K562ads and K562adr vp cells exhib Table 4 ICso values obtained in Adriamycin and vinblastine Â±SDof the curve fitting iting partial recovery oflipid NMR signals were compared to K562wt perfonned on n points and K562adr cells for cellular resistance, mdr gene amplification, and

(,sM)K562wt0.Cell lineAdriamycin (tiM)Vinblastine expression. 17 Â±0.03 Â±0.0001 There is converging evidence that the function of P-gp is linked (n74)K562adr21.5 (n=80)0.0039 to its lipidic environment. It has been reported that oVerexpression Â±2.8 Â±0.04 of P-gp in the plasma membrane is accompanied by modifications (n74)K562ads8.0 (n=80)0.36 Â±0.9 Â±0.02 in the membrane lipid content (5â€”8),and changes in the membrane (n74)K562adr, (n68)0.185 lipid composition of resistant cells can modulate resistance (9â€”13). â€˜pICso14.6 0.032(nÂ±2.40.15 Â± Moreover, it has been demonstrated that retention of P-gp ATPase 56)(n M) activity requires the presence of lipids during its purification (31), increase in inositol compared to creatine was found in K562adr cells and this enzyme is preferentially activated by specific types of but not in K562tax cells, suggesting that a detoxification mechanism lipids (32). Detergents can also affect this ATPase activity (32), via glucuroconjugation may be involved in Adriamycin resistance but and have been reported to reverse the MDR phenotype by altering not in Taxolresistance. P-gp drug binding (33). When MDR is reversed by surfactants, a When resistant cell lines were cultured for two passages without decrease in lipid membrane fluidity has been shown using the drugs, the spectra were not modified, suggesting that lipid signal fluorescence polarization technique (34). On a whole, these studies changes were not dependent on the presence of the drug in the media show that there is an evident link between P-gp activity and the or interaction with membrane lipids. In contrast, the long-term ab quantitative and qualitative nature of the membrane lipids on sence of the drug produced partial recovery of fatty acid signals, but which MDR modulators, including vp, indirectly interact (35). The the other ratios measured did not differ from those of the resistant degree of fluidity of these lipids required for activation or inhibi cells. This increase in lipid signals was measured after 50 passages for tion of P-gp remains unclear, since contradictory hypotheses and K562adr and 10 passages for K562tax. results are reported (9â€”13, 34), probably because the notion of Adriamycin- and Taxol-resistant cells showed common lipid mod membrane fluidity is not dependent on a single parameter but is the ifications, but at the same time, metabolic modifications seemed to be result of several types of interactions (36) and the coexistence of dependent on the drug inducing the resistance. The partial recovery of lipid and protein domains within the membrane (37). lipid signals was also common to both resistant cell lines cultured VP 15a Ca2@ channel blocker agent shown to modulate the MDR without drugs. To suggest the relationship between the lipid content, phenotype (38, 39). In different experimental models, vp was reported as measured using NMR spectroscopy and the resistance phenome to either decrease (40â€”42)or increase P-gp expression (43). A mech

A B 123456 123456 @ 23.1 MDR

Fig. 4. mdrl gene amplification and gene cx pression analysis. A. Southern blot hybridization of DNA from 1(562cell lines with mdrl probe; 20 @sg &oRI digested DNA were loaded in each lane. MDR 4.4 â€”0. 28s â€”0 Size markers (A DNA/HindU!) are indicated in kb units. Blot was rehybridized with a p53 probe as 3.5 â€”0 quantitative control. B, Northern blot hybridization of total RNA from K562 cell lines with mdrl 18sâ€”Ã¸@ probe; 20 ,@gof total RNA were loaded in each lane, and the blot was rehybridized with a GAPDH probe as a quantitative control. Lane 1, K562ads p84; Lane 2. K562adr vp p1; Lane 3, K562adr vp p5; Lane 4, K562adr vp p10; Lane 5, K562wt; and Lane 6, K562adr. p53 GAPDH

3465

K562wt KS62adr CellCount Cell Count

Fig. 5. Flow cytometric analysis of P-gp expression in the K562 cell lines. Cells were incubated with MRK16 anti-P-gp monoclonal anti body (â€”) or with an IgG2a as control (same isotype as MRK16, ). p-n' was detected in K562adr, KS62ads, and K562adr vp. The same scale was used for the four cell lines (fluorescence intensity K562ads K562adrVp and cell count). Cell Count Cell Count

FluorescenceIntensity(a.u.) FluorescenceIntensity(a.u.)

anism evoked for vp (and other reversing molecules) as a drug ACKNOWLEDGMENTS modulator concerns the effects on lipid physical properties; indeed, We acknowledge Drs. L. Ma, C. Chomienne, C. GauvillÃ©,andJ. Soudon for it may be concentrated in lysosomes where it inhibits lipid metab their help in carrying out this work and E. Quiniou and J. Borredon for olism (44). In this case, the resulting accumulation of polar lipids providing two-dimensional software. would interfere with the membrane lipids. Thus, lipid signal mod ifications might originate from the membrane or the intracellular lipids since vp may interfere with both. However, the change in REFERENCES lipid content detected using proton spectroscopy concerns fatty I. Vendrik, C. P. J., Bregers, J. J., Dc Jong, W. H., and Steerenberg, P. A. Resistance acid signals, without changes in the phospholipids. Thus, even to cytostatic drugs at cellular level. Cancer Chemother. Pharmacol., 29: 413â€”429, 1992. though lysosome accumulation cannot be excluded from the con 2. Simon, S. M., and Schindier, M. Cell biochemical mechanisms of multidrug resist tribution of the lipid NMR signal, the nature of the lipid accumu ance in tumors. Proc. Natl. Acad. Sci. USA, 91: 3497â€”3504,1994. 3. Nielsen, D., and Skovsgaard, T. P-glycoprotein as a multidrug transporter: a critical lated is not the one usually described with the lysosome accumu review of current multidrug resistant cell lines. Biochim. Biophys. Acta, 1139: lation. This interference of vp with lipids was also found in 169â€”183,1992. K562wt cells, since the CH2:CH3 ratio was also increased in these 4. Cabral, F. R., Wible, L., Brenner, S., and Brinidey, B. R. Taxol-requiring mutant of Chinese hamster ovary cells with impaired mitotic spindle assembly. J. Cell Biol., 97: cells treated with vp. 30â€”39,1983. Interestingly, when Adriamycin was absent from the culture 5. Grunicke, H., and Hofman, J. Cytotoxic and cytostatic effects of antitumor agents induced at the plasma membrane level. Pharmacol. Ther., 55: 1â€”30,1992. medium of resistant cells, such an effect on lipids could not be 6. Escriba, P. V., Fen'er-Montiel,A. V., Ferragut,J. A., and Gonzalez-Ros,J. M. Role of evoked. Nevertheless, the same results were obtained: an increase membranelipids in the interactionof daunomycinwith plasma membranesfrom tumor cells: Implications in drug-resistance phenomena. Biochemistry, 29: 7275-7282, 1990. in fatty acid NMR signals, decreased resistance to Adriamycin and 7. May, G. L, Wright, L. C., Dyne, M., Mackinnon, W. B., Fox, R. M., and Mountford, Vinbiastine, unchanged P-gp and mdr mRNA expression, and mdr C. E. Plasma lipid composition of vinblastine sensitive and resistant human leukaemic gene amplification. In this case, a mechanism other than the lymphoblasts. lot. J. Cancer, 42: 728â€”733,1988. 8. Mazzoni, A., and Trave, F. Cytoplasmic membrane cholesterol and doxorubicin interaction with lipids is responsible for the modification in cellu cytotoxicity in drug-sensitive and multidrug-resistant human ovarian cancer cells. lar lipid content with resistance reversion. But in both cases, the OncoL Res., 5: 75â€”82,1993. 9. Ikushima, S., Fujiwara, F., Todo, S., and Imashuku, S. Effects of polyunsaturated lipid content modification, as detected using proton NMR spec fatty acids on vincristine-resistance in human neuroblastoma cells. Anticancer Res., troscopy, is related to a decrease in resistance without involvement 11: 1215â€”1220,1991. of the MDR system. 10. Madhavi, N., and Das, U. N. Effects of n-6 and n-3 fauy acids on the survival of vincristine sensitive and resistant human cervical carcinoma cells in vitro. Cancer Reversion of the MDR phenotype could be obtained in this cell Left., 84: 31â€”41,1994. system, and was not related to mdr gene expression and amplification. 11. Guffy, M. M., North, J. A., and Burns, C. P. Effects of cellular fatty acid alteration on Adriamycin sensitivity in cultured L1210 murine leukemia cells. Cancer Res., 39: This reversion was associated with modifications in the lipids detected 1863â€”1866,1984. using proton NMR spectroscopy. The increased fatty acid signals 12. Sinicrope, F. A., Dudeja, P. K., Bissonnene, B. M., Safa, A. R., and Brasitus, T. A. Modulation of P-glycoprotein-mediated drug transport by alterations in lipid fluidity observed may be related to higher membrane fluidity or an increase in of rat liver canalicular membrane vesicles. J. Biol. Chem.. 267: 24995â€”25002,1992. the mobile lipid regions within the membrane. Such lipid composition 13. Litman, T., Nielsen, D., Skovsgaard, T., and Bukhave K. Lipid composition of sensitive and multidrug resistant ascites tumour cells. Cell. Pharmacol., 2: 9â€”14, modifications could improve intracellular drug uptake or alter the 1995. P-gp pump function and thus increase the sensitivity to drugs. 14. Mackinnon, W. B., May, G. L, and Mountford, C. E. Esterified cholesterol and 3466

triglyceride are present in plasma membranes of Chinese ovary cells. Eur. J. Bio 29. Liebmann, J. E., Hahn, S. M., Cook, J. A., Lipdchultz, C., Mitchell, i. B., and chem.,205:827â€”839,1992. Kaufman, D. C. Glutathione depletion by i-buthionine sulfoximine antagonizes Taxol 15. Le Moyec, L, Tatoud, R., Eugene, M., GauvillÃ©,C.,Primot, I., Charlemagne, D., and cytotoxicity. Cancer Rca., 53: 2066â€”2070, 1993. Calvo, F. Cell and membrane lipid analysis by proton magnetic resonance spectros 30. Ross, B. D. Biochemical considerations in 1H spectroscopy. Glutamate and gluta copyin fivebreastcancercelllines.Br.J. Cancer,66:626â€”628,1992. mine; myo-inositol and related metabolites. NMR Biomed., 4: 59â€”63,1991. 16. Dingley, A. J., King, N. J. C., and King, 0. F. An NMR investigation of changes in 31. Urbatsch, I. L, and Senior, A. E. Effects of lipids on ATPase activity of purified plasma membrane triglyceride and phospholipid precursors during the activation of Chinese hamster P-glycoprotein. Arch. Biochem. Biophys., 316: 135â€”140,1995. T-lymphocytes. Biochemistry. 31: 9098-9106, 1992. 32. Doige, C. A., Yu, X., and Sharom, F. J. The effects of lipids and detergents on 17. Jeannesson, P., Trentesaux, C., Gerard, B., Jardiller, I. C., Ross, K. L, and TdkÃ¨s, ATPase-active P-glycoprotein. Biochim. Biophys. Acts, 1146: 65â€”72,1993. z. A. Induction of erythroid differentiation in human leukemic K562 cells by 33. Zordan-Nudo, T., Ling, V., and Georges E. Effects of nonionic detergents on membrane-directed action of Adriamycin co-valently bound to microspheres. Cancer P-glycoprotein drug binding and reversal of multidrug resistance. Cancer Res., 53: Rca.,50:1231â€”1236,1990. 5994â€”6000,1993. 18. Landegren, U. Measurement of cell numbers by means of the endogenous enzyme 34. Dudeja, P. K., Anderson, K. M., Harris, J. S., Buckingham, L, and Coon, J. S. hexosaminidase. Applications to detection of lymphokines and cell surface antigens. Reversal of multidrug phenotype by surfactants: relationship to membrane lipid J. Immunol. Methods, 67: 379â€”388,1984. fluidity. Arch. Biochem. Biophys., 319: 309â€”315,1995. 19. Davis, L 0., Dibner, M. D., and Battey, J. F. Basic Methods in Molecular Biology. 35. Wadkins, R. M., and Houghton, P. J. The role of drug-lipid interactions in the New York: Elsevier Science Publishing Co., 1986. biological activity of modulators of multi-drug resistance. Biochim. Biophys. Acts, 20. Harris, N., Brill, E., Shohat, 0., Prokocimer, M., Wolf, D., Arai, N., and Rotter, V. 1153: 225â€”236,1993. Molecular basis for heterogeneity of the human p53 protein. Mol. Cell. Biol., 6: 36. Cribier, S., Morrot, G., and Zachowski, A. Dynamics of the membrane lipid phase. 4650â€”4656, 1986. Prostaglandins Leukotrienes Essent. Fatty Acids, 48: 27-32, 1993. 21. Fort, P., Marty,L, Piechaczyk,M.,El Sabrouty,S., Dani,C., Jeanteur,P., and 37. Glaser, M. Lipid domains in biological membranes. Curr. Opin. Struct. Biol., 3: Blanchard,J.M.VariousratadulttissuesexpressonlyonemajormRNAspeciesfrom 475â€”481,1993. the glyceraldehyde-3-phosphate-dehydrogenase multigenic family. Nucleic Acids 38. Tsuruo, T., lids, H., Tsukagoshi, S., and Sakurai, Y. Increased accumulation of Res., 13: 1431â€”1442,1985. vincristine and Adriamycin in drug-resistant P388 tumor cells following incubation 22. Hamada, H., and Tsuruo, T. Functional role for the 170- to 180-kDa glycoprotein with calcium antagonists and calmodulin inhibitors. Cancer Res., 42: 4730â€”4733, specific to drug-resistant tumor cells as revealed by monoclonal antibodies. Proc. 1982. Natl. Acad. Sci. USA, 83: 7785â€”7789,1986. 39. Pereira, E., Borrel, M. N., Fiallo, M., and Garnier-Sullerot, A. Non-competitive 23. Lean, C. L, Mackinnon, W. B, Delikatny, E. J., Whitehead, R. H., and Mountlord, inhibition of THP-Adriamycin by verapamil in living K562 leukemia cells. Biochim. C.E.Cell-surfacefucosylationandmagneticresonancespectroscopycharacterization Biophys. Acts, 1225: 209â€”216,1994. of human malignant colorectal cells. Biochemistry, 31: 11095â€”11105, 1992. 40. Muller, C., Bailly, J. D., Goubin, F., Laredo, J., JaffrÃ©zou,J.P., Bordier, C., and 24. Mountford, C. E., and Mackinnon, W. B. Proton magnetic resonance spectroscopy of Laurent, G. Verapamil decreases P-glycoprotein expression in multidrug-resistant lymphocytes: an historical perspective. Immunomethods, 4: 98â€”112,1994. human leukemic cell lines. tat. J. Cancer, 56: 749â€”754,1994. 25. Mackinnon, W. B., Dyne, M., Holmes, K. T., Mountford, C. E., and Gupta, R. S. 41. Koizumi, S., Konishi, M., Ichihara, T., Wads, H., Matsukawa, H., Goi, K., and Further evidence that the narrow IH magnetic resonance signals from malignant cells Mizutani, S. Flow cytometric functional analysis of multidrug resistance by fluo-3: a do not arise from lipid droplets. NMR Biomed., 2: 161â€”164,1989. comparison with rhodamine-123. Eur. J. Cancer, 31A: 1682â€”1688,1995. 26. de Certaines, J. D., Larsen, V. A., Podo, F., Carpinelli, G., Briot, 0., and Henricksen, 42. Bidder, J. L., and Spengler, B. A. Reverse transformation ofmultidrug-resistant cells. 0. In vivo 3lP MRS of experimental tumours. NMR Biomed., 6: 345-365, 1993. Cancer Metastasis Rev., 13: 191-207, 1994. 27. Cuadrado, A., Camero, A., Dolfi, F., Jimenez, B., and Lseal, J. C. Phosphoryicholine: 43. Herzog C. E., Tsokos, M., Bates, S. E., and Fojo, A. Increased mdr-l/P-glycoprotein a novelsecondmessengerformitogenicactivityof growthfactors.Oncogene,8: expression after treatment of human colon carcinoma with P-glycoprotein antago 2959â€”2968, 1993. nists. J. Biol. Chem., 268: 2946â€”2952, 1993. 28. Saint-Ruf, C., Malfoy, B,. Scholl, S., Zafrani, B., and Dutrillaux, B. GSTp gene is 44. JaffrÃ©zou,J. P., and Laurent, G. The intriguing link between modulation of both frequently coamplifled with Th1T2and HSTF1 proto-oncogenes in human breast multidrug resistance and ligand-toxin conjugate cytotoxicity. FEBS Lea., 323: 191â€” cancers. Oncogene, 6: 403â€”406, 1991. 197, 1993.

3467

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 1996 American Association for Cancer Research. Proton Nuclear Magnetic Resonance Spectroscopy Reveals Cellular Lipids Involved in Resistance to Adriamycin and Taxol by the K562 Leukemia Cell Line

Laurence Le Moyec, Roger Tatoud, Armelle Degeorges, et al.

Cancer Res 1996;56:3461-3467.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/56/15/3461

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/56/15/3461. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.