Interplay between MRP Inhibition and Metabolism of MRP Inhibitors

Nov 18, 2003 - Biopharmaceutics & Drug Disposition 2017 38 (1), 20-32 .... M. Wortelboer , Michiel G.J. Balvers , Mustafa Usta , Peter J. van Bladeren...
0 downloads 4 Views 159KB Size
Download PDF
1642

Chem. Res. Toxicol. 2003, 16, 1642-1651

Interplay between MRP Inhibition and Metabolism of MRP Inhibitors: The Case of Curcumin Heleen M. Wortelboer,*,† Mustafa Usta,† Astrid E. van der Velde,† Marelle G. Boersma,‡ Bert Spenkelink,‡ Jelmer J. van Zanden,‡,§ Ivonne M. C. M. Rietjens,‡,§ Peter J. van Bladeren,‡,§,| and Nicole H. P. Cnubben†,§ TNO Nutrition and Food Research, P.O. Box 360, 3700 AJ Zeist, The Netherlands, Division of Toxicology, Wageningen University, P.O. Box 8000, 6700 EA Wageningen, The Netherlands, Wageningen University/TNO Centre for Food Toxicology, P.O. Box 8000, 6700 EA Wageningen, The Netherlands, and Nestle´ Research Centre, P.O. Box 44, CH-1000 Lausanne 26, Switzerland Received May 26, 2003

The multidrug resistance proteins MRP1 and MRP2 are efflux transporters with broad substrate specificity, including glutathione, glucuronide, and sulfate conjugates. In the present study, the interaction of the dietary polyphenol curcumin with MRP1 and MRP2 and the interplay between curcumin-dependent MRP inhibition and its glutathione-dependent metabolism were investigated using two transport model systems. In isolated membrane vesicles of MRP1- and MRP2-expressing Sf9 cells, curcumin clearly inhibited both MRP1- and MRP2mediated transport with IC50 values of 15 and 5 µM, respectively. In intact monolayers of MRP1 overexpressing Madin-Darby canine kidney (MDCKII-MRP1) cells, curcumin also inhibited MRP1-mediated activity, although with a 3-fold higher IC50 value than the one observed in the vesicle model. Interestingly, MRP2-mediated activity was hardly inhibited in intact monolayers of MRP2-overexpressing MDCKII (MDCKII-MRP2) cells upon exposure to curcumin, whereas the IC50 value in the vesicle incubations was 5 µM. The difference in extent of inhibition of the MRPs by curcumin in isolated vesicles as compared to intact cells, observed especially for MRP2, was shown to be due to a swift metabolism of curcumin to two glutathione conjugates in the MDCKII cells. Formation of both glutathione conjugates was about six times higher in the MDCKII-MRP2 cells as compared with the MDCKII-MRP1 cells, a phenomenon that could be ascribed to the significantly lower glutathione levels in the cell line. The efflux of both conjugates, identified in the present study as monoglutathionyl curcumin conjugates, was demonstrated to be mediated by both MRP1 and MRP2. From dose-response curves with Sf9 membrane vesicles, glutathionylcurcumin conjugates appeared to be less potent inhibitors of MRP1 and MRP2 than their parent compound curcumin. In conclusion, curcumin clearly inhibits both MRP1- and MRP2-mediated transport, but the glutathione-dependent metabolism of curcumin plays a crucial role in the ultimate level of inhibition of MRP-mediated transport that can be achieved in a cellular system. This complex interplay between MRP inhibition and metabolism of MRP inhibitors, the latter affecting the ultimate potential of a compound for cellular MRP inhibition, may exist not only for a compound like curcumin but also for many other MRP inhibitors presently or previously developed on the basis of vesicle studies.

Introduction Multidrug resistance is a major problem in chemotherapy of human cancer. One of the processes that contributes to multidrug resistance is the active removal of drugs, both by increased detoxification of the carcinostatic drug and by enhanced efflux by transmembrane transporters, thereby decreasing the intracellular concentration of the therapeutic compound (1). The membrane transport proteins belonging to the ATP-binding cassette family such as Pgp1 (MDR1) and members of the multidrug resistance associated protein family (especially * To whom correspondence should be addressed. Tel: +31-306944484. Fax: +31-30-6960264. E-mail: [email protected]. † TNO Nutrition and Food Research. ‡ Division of Toxicology, Wageningen University. § Wageningen University/TNO Centre for Food Toxicology. | Nestle ´ Research Centre.

MRP1, MRP2, and MRP3) are directly involved in this process (1-3). Pgp transport activity is independent of intracellular glutathione levels, whereas the activity of both MRP1 and MRP2 on glutathione is not (3). MRP1 and MRP2 transport a broad range of anionic substrates such as sulfate, glucuronide, and glutathione conjugates, and also, unmodified anticancer drugs such as vincristine and daunorubicin are transported by MRPs whereby GSH efflux seems to be crucial (3, 4). In addition, in many tumors and anticancer drug resistant cell lines, an 1 Abbreviations: calcein-AM, calcein acetoxymethylester; CDNB, 1-chloro-2,4-dinitrobenzene; CUR, curcumin; CURSG, glutathionyl curcumin conjugate; DNPSG, S-(2,4-dinitrophenyl)glutathione; EASG, glutathionyl ethacrynic acid conjugate; GST, glutathione-S-transferase; HBSS, Hanks’ balanced salt solution; MDCK cells, Madin-Darby canine kidney cells; MDR1, multidrug resistance protein 1; Pgp, P-glycoprotein; MRP1, human multidrug resistance-associated protein 1; MRP2, human multidrug resistance protein 2; Sf9, Spodoptera frugiperda insect cells.

10.1021/tx034101x CCC: $25.00 © 2003 American Chemical Society Published on Web 11/18/2003

Curcumin and Multidrug Resistance Proteins

Figure 1. Structure of curcumin (CUR).

overexpression is observed of not only MRP transporter proteins but also of several GSH-associated enzymes (58). On the basis of these observations, inhibition of GSTrelated detoxification and/or MRP-mediated efflux in tumor cells can be expected to enhance the pharmacological effect of anticancer drugs. On the other hand, inhibition of these processes in nontumor cells can enhance the toxicity of electrophiles known to be metabolized and excreted by GSH-dependent pathways. Natural polyphenols have been reported to interact with both GSH-associated biotransformation enzymes as well as membrane transport proteins (9-13). In addition, it has been shown that several plant polyphenols can sensitize cancer cells to anticancer agents, even at low concentrations (14-16). In this respect, the dietary polyphenol CUR has been shown to modulate the glutathione-related biotransformation system (17). CUR [1,7-bis(4-hydroxy-3-methoxyphenyl)-1,6-heptadiene-3,5dione] (Figure 1) is the major yellow pigment extracted from turmuric, a commonly used spice and coloring agent, derived from the rhizome of the herb Curcuma longa. CUR possesses many beneficial biological activities, such as anticarcinogenic, antiviral, antiinflammatory, and antioxidant activity (18-21) and seems to be nontoxic even at high doses in humans (22). Today, there is much attention for CUR in its clinical use, e.g., as an antiinflammatory compound for humans (21, 22), but its metabolism and pharmacokinetics are only partly understood (23). In a study with human melanoma cells (17), CUR was shown to affect the excretion of DNPSG at various levels of interaction, not only by inhibition of GST activity but also by depletion of GSH, and a direct inhibition of DNPSG efflux. Although the effects of CUR on DNPSG transport proteins were not studied in detail, these data point to an effect of CUR and/or its metabolites on transporter proteins. Interaction of CUR and/or its metabolites with transporters such as MRP1 and MRP2 may result in potentially harmful (or helpful?) interactions with various drugs or disease states. Therefore, the objective of the present study was to investigate the interaction of CUR with MRP1- and MRP2-mediated activity. In isolated membrane vesicles of either MRP1 or MRP2 overexpressing Sf9 insect cells, in which CUR metabolism is not an issue, the effect of the parent compound CUR and its glutathione conjugates was studied. In addition, the interplay between the glutathione-dependent metabolism of CUR and MRP inhibition was investigated in intact MDCKII cells overexpressing either MRP1 or MRP2.

Experimental Procedures Materials. CUR (>99%) and monoclonal antibodies to human MRP1 (MRPr1) and human MRP2 (M2III-6) were obtained from Alexis Biochemicals (Kordia, Leiden, The Netherlands). Sf9 cells were obtained from Invitrogen (Groningen, The Netherlands). Recombinant baculoviruses containing either the human MRP1 cDNA or the human MRP2 cDNA were a kind gift from Prof. Dr. B. Sarkadi, National Institute of Haematology and Immunology, Research Group of the Hungarian Academy of Sciences, Budapest, Hungary. The MDCKII cell lines, stably

Chem. Res. Toxicol., Vol. 16, No. 12, 2003 1643 expressing either human MRP1 cDNA (hereafter called MRP1 cells) or MRP2 cDNA (hereafter called MRP2 cells) or transfected with an empty vector (hereafter called control cells) were kindly provided by Prof. P. Borst (NKI, Amsterdam). Dulbecco’s minimum essential medium (DMEM) with GlutaMax, Grace’s insect medium, fetal calf serum, penicillin/streptomycin, and gentamicin were all from Gibco (Paisley, Scotland). [14C]Inulin (37 Bq/mg) was purchased from Sigma (St. Louis, MO). [14C]Ethacrynic acid ([2,3-dichloro-4-(2-methylene-1-oxobutyl)phenoxy]acetic acid) with a specific activity of 555 MBq/ mmol was purchased from Amersham (Buckinghamshire, U.K.). The radioactive glutathione conjugate of ethacrynic acid was synthesized according to Ploemen et al. (24). DNPSG was synthesized analogously to Sokolovsky et al. (25). The MRP1 inhibitor MK 571 was obtained from BioMol (Plymouth Meeting, PA), and the Pgp inhibitor PSC833 was a kind gift from Novartis Pharma AG (Basel, Switzerland). The MRP2 inhibitor cyclosporin A was from Fluka (Zwijndrecht, The Netherlands). Calcein acetoxymethyl ester was obtained from Molecular Probes (Eugene, OR). Glutathione reductase and creatine kinase were from Boehringer (Mannheim, Germany). HPLC grade trifluoroacetic acid was obtained from Baker (Deventer, The Netherlands). HPLC grade methanol was from Merck (Darmstadt, Germany). All other chemicals were from Sigma Chemical Co., unless stated otherwise. Expression of MRP1 and MRP2 in Insect Cells. Sf9 insect cells were infected with recombinant baculoviruses containing either MRP1 cDNA or MRP2 cDNA. Briefly, cells were cultured in spinner flasks in Grace’s insect medium with 10% fetal calf serum and 10 µg/mL gentamicin at 27 °C. For infection, cells were cultured on 145 cm2 culture disks and infected with a baculovirus (multiplicity of infection of five) for 3 days essentially as described by Zaman et al. (26). Virusinfected Sf9 cells were harvested and frozen at -80 °C until membrane preparation. Membrane Preparation and Immunoblotting. Membranes from infected Sf9 cells were isolated as described by van Aubel et al. (27). Membrane protein concentrations were determined according to Bradford (28) adapted for 96 well measurements on a BioRad 3550 microplate reader. Vesicles were prepared by passing the suspension 30 times through a 26-gauge needle with a syringe. Aliquots of 25 µL membrane vesicles containing 1 mg protein/mL were quickly frozen in liquid nitrogen and stored at -80 °C until used in transport assays. The expression of MRP1 and MRP2 in Sf9 membranes as well as the expression of MRP1 and MRP2 in isolated membranes of MDCKII cells was assessed using immunoblotting with monoclonal antibodies MRPr1 and M2III-6, raised against human MRP1 and human MRP2, respectively. The results obtained indicate a Mr of 190 kDa for the MRP proteins in MDCKII cells and an apparent Mr of 150 kDa in the Sf9transfected cells. This is in line with literature data describing that human MRPs are produced in an underglycosylated form in Sf9 cells, which has been demonstrated not to affect their transport functions (29, 30). In the wild-type MDCKII cells, a very light band was detected when stained with monoclonal antibody against human MRP2 as noted earlier by Evers et al. (31). In these wild-type kidney cells, canine multidrug resistance proteins, which are possibly not specifically recognized by the specific monoclonal antibodies raised against human MRPs, could still be present but to a much lower level than the human MRPs in the transfected cells. [14C]EASG Uptake in Sf9 Membrane Vesicles. The uptake of [14C]EASG in isolated Sf9 cell membrane vesicles was determined as described earlier (26). In brief, vesicles were rapidly thawed and preincubated for 1 min at 37 °C in Trissucrose buffer (TS buffer: 10 mM Tris, 250 mM sucrose, pH 7.4) containing 4 mM ATP, 10 mM MgCl2, 10 mM creatine phosphate, and 100 mg/mL creatine kinase. The reaction was started by addition of [14C]EASG to a final concentration of 4 µM (555 MBq/mmol). Samples were taken at 1, 2.5, and 5 min, diluted with 1 mL of ice-cold TS buffer, and rapidly filtered

1644 Chem. Res. Toxicol., Vol. 16, No. 12, 2003 through presoaked nitrocellulose filters (0.45 µm pore size) (Schleicher and Schuell, Dassel, Germany) in a 1225 sampling Manifold filtration unit (Millipore, Ettenleur, The Netherlands). Filters were rinsed with 5 mL of TS buffer, and the radioactivity associated with the filters was measured by liquid scintillation counting. In control experiments, ATP was replaced by 4 mM AMP-PCP (R,β-methylene adenosine 5′-triphosphate). ATPdependent [14C]EASG uptake was calculated by subtracting the values obtained in the presence of AMP-PCP from those in the presence of ATP. The concentration at which 50% transport inhibition was obtained (IC50) was determined after an incubation of 1 min in the presence of various amounts (0-50 µM of inhibitor; final concentration DMSO, 0.5%). MDCKII Cell Culture. The Madin-Darby canine kidney cell lines (control, MRP1, or MRP2 transfected) were cultured in DMEM with GlutaMax (4.5 g of glucose per liter), 10% fetal calf serum, and 1% penicillin/streptomycin (each 10 000 units/ mL) and were grown in a humidified atmosphere of 5% CO2 at 37 °C, as described before (31, 32). Cell lysates were prepared by scraping the cells in ice-cold PBS and sonicating for 10 s on ice. GST activities were determined according to the method of Habig et al. (33), and GSH was measured according to the method of Anderson (34). For transport experiments, 105 cells/cm2 were seeded and grown to confluency on microporous polycarbonate filters (0.4 µm pore size, 24.5 mm, Costar Corp., Cambridge, MA). It was shown earlier (31, 32) that in these polarized cell lines MRP1 routes to the basolateral plasma membrane, whereas MRP2 localizes to the apical plasma membrane. Confluency of the monolayers was checked by determination of the paracellular flux of inulin[14C]carboxylic acid (185 kBq/mol, 4.2 µM) from the apical to the basolateral side in three separate monolayers of cells, and radioactivity was analyzed using a Wallac 1414 Liquid Scintillation Counter. The percentage of leakage was less than 2% in all experiments. In addition, the transepithelial electric resistance (TEER) value of each monolayer was measured using a Millicell-ERS epithelial volt/ohmmeter (Millipore, Bedford). The TEER value of a confluent monolayer of MDCKII cells ranged between 200 and 250 Ω cm2 as reported before (35). The leukotriene D4 receptor antagonist MK-571 was used as a MRP1 inhibitor (1), and cyclosporin A was used as MRP2 inhibitor (36). The cytotoxicity of CUR, MK-571, and cyclosporin A was checked using a neutral red test (37). No cytotoxicity of CUR, MK571, or cyclosporin A was observed up to 50 µM during the course of the experiments. Efflux of DNPSG and GSH from MDCKII Cells. The efflux of DNPSG was determined by incubating control, MRP1, and MRP2 cells in 24.5 mm diameter transwells (Costar Corp.) with CDNB essentially as described by Evers et al. (32, 38) with the modification that DNPSG was determined using HPLC. Confluent monolayers of cells were incubated in HBSS containing 0.5 mM acivicin (to prevent degradation of DNPSG by γ-glutamyltranspeptidase) and either CUR or MRP inhibitor. DMSO was used as a vehicle in a final concentration of 0.5% (v/v). After the cells were preincubated for 10 min, a 100 times concentrated stock solution of 0.25 mM CDNB in HBSS was added to both the apical and the basolateral compartment of the monolayer, yielding a final CDNB concentration of 2.5 µM. The resulting hydrophilic glutathione conjugate DNPSG can only leave the cell by active transport processes and is known to be a good substrate for both MRP1 and MRP2 (32, 38). Aliquots of medium (200 µL) from both sides were taken at three time points (7, 14, and 20 min), directly mixed with 5 µL of 0.04 M N-acetyl L-cysteine to trap unconjugated CDNB, frozen on dry ice, and stored at -20 °C until DNPSG analysis. DNPSG was analyzed by reversed phase HPLC (Agilent Technologies system) on an Inertsil ODS-3 column (4.6 mm × 250 mm, GLSciences, Japan), eluted at a flow rate of 0.8 mL/min with 0.1% (v/v) trifluoroacetic acid in water (solvent A) and 0.1% (v/v) trifluoroacetic acid in methanol (solvent B), 5 min isocratically at 30% B, followed by a linear gradient of 30-90% B over 20 min, followed by a linear gradient to 100% in 2.5 min, and

Wortelboer et al. equilibrated with an isocratic elution at 30% B for 10 min. Quantification of DNPSG was performed by peak area integration at 340 nm, using concentration/absorbance curves of DNPSG standards. The efflux of GSH from MDCKII cells was determined in parallel cultures exposed to DMSO (v/v 0.5%) only. Aliquots of media (100 mL) were used to determine the GSH concentration using the method of Anderson (34). Intracellular GSH Concentration and DNPSG Level. Intracellular DNPSG and GSH concentrations were determined at the end of the efflux experiments. Thus, after 20 min of incubation with CDNB, cells were washed twice with ice-cold PBS. Filters with cells were cut out and collected in 1.4 mL of HBSS with 0.5 mM acivicin and 35 µL of 0.04 M N-acetyl L-cysteine. Cells were disrupted by sonication on ice (10 s), and the cell lysate was centrifuged at 2800g for 5 min, and stored frozen at -20 °C until further analysis. Aliquots of supernatant were used to determine the GSH concentration using the method of Anderson (34) and the intracellular DNPSG levels using HPLC as described above. Efflux of Calcein in MDCKII Cells. The efflux of calcein, which is a good substrate for MRP1 and MRP2 (39), was determined using the confluent monolayer of the control, MRP1, and MRP2 cells cultured on 12 mm diameter transwells (Costar Corp.). First, cells were loaded with calcein-AM (1 µM) in DMEM without phenol red for 2 h. Once inside the cells, cleavage of this nonfluorescent calcein-AM ester by intracellular esterases leads to formation of the fluorescent derivative calcein. The nonfluorescent calcein-AM, however, is a good substrate for both Pgp and MRP1 (39), and in order to diminish the MRPdependent efflux of calcein-AMsand because it was preferred to use no MRP inhibitors during loading timescells were loaded with calcein-AM at a temperature of 7 °C. In addition, PSC833 (0.1 µM) was added as a Pgp inhibitor. Calcein-AM uptake and intracellular conversion to calcein were checked for each cell line as follows. After several time intervals of exposure to calcein-AM, cells were washed with ice-cold PBS; filters with cells cut out and collected in 0.5 mL of PBS. The filters with cells were sonicated for 10 s, and fluorescence of intracellular calcein was determined using a Cytofluor 2300 (Millipore) with excitation at 485 nm and emission at 530 nm. Efflux of calcein was determined as follows. Loaded cells were washed with DMEM without phenol red and exposed to fresh medium (37 °C) containing either 2-50 µM CUR, 50 µM MK571 (as a typical MRP1 inhibitor), or 30 µM cyclosporin A (as a typical MRP2 inhibitor). Control cells received vehicle only (0.5% DMSO v/v). Efflux of calcein was measured in media samples from both the apical and the basolateral compartment after 25 and 45 min. Fluorescence of calcein was determined as described above. Cellular Formation of Glutathione Metabolites. To investigate the biotransformation of CUR in control, MRP1, and MRP2 cells, cells were plated on 75 mm diameter transwells (Costar Corp.). At confluence, cells were exposed to HBSS containing 0.5 mM acivicin and 40 µM CUR. After 4 h, media samples were collected and analyzed for the presence of CUR and its metabolites by reversed phase HPLC using an Alltima C18 column (4.6 mm × 150 mm, 5 µm, Alltech, IL), eluted at a flow rate of 1.0 mL/min with 0.1% (v/v) trifluoroacetic acid in water (solvent A) and acetonitril (solvent B). The solvent program started isocratically with 0% B over 5 min, followed by a linear gradient to 29% B over 6 min and, thereafter, an isocratic elution at 29% over 20 min, followed by a linear gradient to 95% B over 7 min, and equilibrated with an isocratic elution at 0% over 12 min. Detection was performed using a Waters 996 photodiode array detector (200-500 nm). Synthesis and Purification of CURSGs. To elucidate the structure of the presumed glutathione conjugates of CUR, both conjugates were synthesized as follows. GSH (10 mM) was dissolved in demineralized water adjusted to pH 8.5. CUR (50 µL of a 10 mM stock solution in DMSO) was added every half hour (for five times), and the mixture was incubated at 37 °C

Curcumin and Multidrug Resistance Proteins in the dark. The reaction mixture was monitored by reversed phase HPLC (Agilent Technologies System) as described above. For LC/MS analysis, separation of peaks was performed on an Alltima C18 column (2.1 mm × 150 mm, 5 µm, Alltech) using an elution flow of 0.2 mL/min and similar elution conditions as described above. Mass spectrometric analysis (Finnigan MAT 95, San Jose, CA) was performed in the positive mode using a spray voltage of 4.5 kV and a capillary temperature of 180 °C with nitrogen as the sheath and auxiliary gas. To isolate the CUR glutathionyl conjugates (as a mixture), preparative HPLC was performed using a Zorbax ODS column (21.2 mm × 250 mm), and the metabolites eluted at a flow rate of 4.0 mL/min with 0.1% (v/v) trifluoroacetic acid in water (solvent A) and acetonitrile (solvent B). The solvent program was linear to 29% B over 5 min, followed by an isocratic elution at 29% over 23 min, and finally a linear gradient to 95% B over 7 min. Detection was performed at 360 nm. The two peaks eluting from 25 to 33 min, representing the two CUR-derived metabolites, were collected, and the eluent evaporated. The dry mixture of the two CURSGs was stored at -20 °C. MRP-Mediated Efflux of CUR Metabolites in MDCKII Cells. The efflux of the CUR metabolites from control, MRP1, and MRP2 cells, respectively, was followed in time. In brief, cells were cultured to confluence on 75 mm diameter transwells (Costar Corp.) and incubated with HBSS containing 0.5 mM acivicin for 10 min, added at both sides of the filter. Hereafter, CUR (final concentration 40 µM) was added at both the apical and the basolateral side. At 10, 20, 40, 60, and 90 min thereafter, 150 µL of medium (from both the apical and the basolateral side) was sampled, immediately frozen on dry ice, and stored at -20 °C until analysis by reversed HPLC (see above). The glutathione conjugates of CUR were stable at -20 °C for at least 2 weeks. MRP-Mediated Efflux of CUR in MDCKII Cells. The efflux of CUR from control, MRP1, and MRP2 cells, respectively, was followed in time. First, cells were loaded with CUR (40 µM) in HBSS containing 0.5 mM acivicin and 0.1 µM PSC833, added at both sides of the filter, for 15 min. Hereafter, cells were washed with HBSS and fresh HBSS medium (1 mL) containing 0.5 mM acivicin and 0.1 µM PSC833 was added to both sides of the cells. Aliquots of medium (150 µL) from both sides were taken at three time points (30, 60, and 90 min), directly frozen on dry ice, and stored at -20 °C until analysis by reversed HPLC (see above). Statistical Analysis. Data are presented as means ( SD where appropriate. Statistical differences between group means were determined with one way ANOVA, followed by Dunnetts’ multiple comparison test. A probability level of p < 0.05 was considered significant.

Results Transport Studies in Isolated Membrane Vesicles of Either MRP1- or MRP2-Transfected Sf9 Cells. To validate the transport model with inside-out membrane vesicles of MRP1- or MRP2-transfected Sf9 cells, the uptake of the radiolabelled glutathione conjugate of ethacrynic acid ([14C]EASG) was determined. In the membrane vesicles of both the MRP1- and the MRP2transfected Sf9 cells, the uptake of [14C]EASG was shown to be time- and ATP-dependent, indicating active transport by MRP (Figure 2). As a control, the uptake of [14C]EASG was determined in membrane vesicles of the mocktransfected Sf9 cells, which appeared to be 0.09 nmol/ min/mg protein with ATP and 0.08 nmol/min/mg protein without ATP. CUR (up to 50 µM) clearly inhibited the ATP-dependent [14C]EASG uptake in both MRP1- and MRP2-containing vesicles (Figure 3). The dose-response curve observed for MRP1 reveals an IC50 of 15 µM CUR. For MRP2, a biphasic concentration-dependent inhibition of [14C]EASG uptake by CUR was observed. This obser-

Chem. Res. Toxicol., Vol. 16, No. 12, 2003 1645

Figure 2. Transport of [14C]EASG in Sf9 cell membrane vesicles, expressing either MRP1 (A) or MRP2 (B), in the presence of 4 mM ATP (closed symbols) or AMP-PCP (open symbols). Each point represents the average ( SD of an incubation in triplicate.

Figure 3. Dose-dependent effect of CUR on the ATP-dependent [14C]EASG uptake in Sf9 cell membrane vesicles, containing MRP1 (A) or MRP2 (B). Each value is the average ( SD of an experiment performed in triplicate, expressed as the percentage of transport in the absence of CUR.

Figure 4. Typical time-dependent efflux pattern of DNPSG in MDCKII-MRP1 (A) and MDCKII-MRP2 (B) cells. Monolayers of MDCKII cells were treated with 2.5 µM CDNB to study DNPSG efflux to the basolateral compartment (open symbols) and to the apical compartment (closed symbols). Each point represents the average ( SD of an incubation done in triplicate.

vation might indicate different inhibition sites for CUR in the MRP2 protein. The dose-response curve presented reveals an IC50 of the first phase of about 5 µM CUR for MRP2 inhibition in these membrane vesicles of MRP2transfected Sf9 cells. DNPSG Efflux, GSH, and GST Levels in Control, MRP1-, and MRP2-Transfected MDCKII Cells. The efflux of DNPSG from all three MDCKII cell lines was determined after exposure to 2.5 µM CDNB (added to both sides). As shown in Figure 4, a time-dependent increase of DNPSG level in the apical and basolateral media was observed. As expected, the efflux of DNPSG from MRP1 cells is mainly to the basolateral side (Figure 4A), whereas in MRP2 cells DNPSG is predominantly excreted to the apical side (Figure 4B), in line with data already described by others (32, 38). In Table 1, total, intra-, and extracellular levels of DNPSG are shown, as

1646 Chem. Res. Toxicol., Vol. 16, No. 12, 2003

Wortelboer et al.

Table 1. Distribution of DNPSG upon Exposure to CDNB, GST Activities, and GSH Levels in the MDCKII (Control), MDCKII-MRP1, and MDCKII-MRP2 Cellsa MDCKII (control) total intracellular excreted to apical side excreted to basolateral side GST (cell lysate) (nmol/min/mg protein) GSH (cell lysate) (nmol/mg protein) excreted to the apical side excreted to basolateral side

MDCKII-MRP1

MDCKII-MRP2

DNP-SG (nmol/monolayer 4.1 ( 0.15 (100%) 4.4 ( 0.09 (100%) 0.7 ( 0.02 (17%) 0.5 ( 0.01 (12%) 1.3 ( 0.09 (32%) 0.5 ( 0.00* (11%) 2.1 ( 0.05 (51%) 3.4 ( 0.10* (77%) 44.3 ( 4.2 32.5 ( 3.1*

5.4 ( 0.58* (100%) 0.1 ( 0.01* (1%) 4.4 ( 0.62* (82%) 0.9 ( 0.03* (17%) 22.9 ( 2.5*

29.1 ( 1.3

34.4 ( 0.5*

2.5 ( 0.01*

GSH efflux (pmol/min/monolayer 3.4 ( 0.6 2.3 ( 0.5 7.6 ( 1.2 18.8 ( 3.1*

18.8 ( 3.1* 4.4 ( 1.2

a For details, see the Experimental Procedures. Data are means ( SD of three incubations of a typical experiment performed in quadruplicate. Those marked with asterisks differ significantly (ANOVA + Dunnetts’ test) from the corresponding value in MDCKII-MII (control) cells (*P < 0.05).

measured at the end of the transport experiments (i.e., after exposure to CDNB for 20 min). It appeared that the total amount of DNPSG formed after 20 min of incubation with CDNB is comparable in all three cell lines (4.0-5.5 nmol DNPSG/monolayer). As expected, as a result of the enhanced efflux of DNPSG in MRP1 and MRP2 cells, the intracellular DNPSG levels were higher in the control cells as compared to the MRP1 and MRP2 cells. Table 1 also presents the GST activities and GSH levels as measured in cell lysates of the three different cell lines. The results reveal that GST activities in all three cell lines do not vary by more than a factor of two, whereas the GSH levels vary significantly. In the MDCKII-MRP1 cells, the GSH concentration appeared to be 2.5 nmol/mg cell lysate, which amounts to only 10% of the GSH levels observed in the other two cell lines. Important to note is that even this low cellular GSH concentration detected in MRP1 cells did not result in lower DNPSG formation in these MRP1 cells (Table 1). The low intracellular GSH levels in the MRP1 cells as compared to the MRP2 cells could not be explained by enhanced GSH efflux, as the total GSH efflux in the MRP1 and MRP2 cells appeared to be identical (Table 1). Effect of CUR on DNPSG Efflux in Control, MRP1-, and MRP2-Transfected MDCKII Cells. The effect of CUR (0, 20, and 50 µM) on the efflux of DNPSG was determined in the controls, MDCKII-MRP1, and MDCKII-MRP2 cells. Cells were preincubated for 10 min with either CUR or a typical MRP1 or MRP2 inhibitor, before addition of CDNB. The extracellular DNPSG levels in the apical and basolateral compartment, the intracellular DNPSG levels, and the total amount of formed DNPSG as measured in all three MDCKII cell lines are shown in Figure 5. In the control MDCKII cells (Figure 5A), CUR inhibited the efflux of DNPSG to both the apical and the basolateral side dose-dependently, with a concomitant increase of intracellular DNPSG. In these control cells, MK571 (10 µM) inhibited the basolateral DNPSG efflux, whereas cyclosporin A (20 µM) inhibited the apical DNPSG efflux, and both transporter inhibitors increased the intracellular DNPSG levels. In MRP1 cells (Figure 5B), CUR also decreased the efflux of DNPSG, to both the basolateral as well as the apical compartment, and this decreased DNPSG efflux was again accompanied by a concomitant increase of intracellular DNPSG in cells exposed to the highest

Figure 5. Effect of CUR on the formation and excretion of DNPSG by MDCKII-MII (A), MDCKII-MRP1 (B), and MDCKIIMRP2 (C) cells cultured on transwells. Results present DNPSG concentrations (nmol/monolayer) in the apical compartment (first bar), basolateral compartment (second bar), and intracellular compartment (third bar). Additionally, the resulting total formation of DNPSG is presented (fourth bar). For comparison, the effect on DNPSG excretion of the model inhibitors MK571 (10 µM in MII cells, 50 µM in MRP1 cells) and cyclosporine (20 µM in MII and MRP2 cells) is given. Each bar represents means ( SD of incubations performed in triplicate. Those marked with asterisks differ significantly (ANOVA + Dunnetts’ test) from the corresponding value in DMSO-treated cells (P < 0.05).

concentration of CUR (50 µM). An almost identical distribution pattern of DNPSG to that obtained for MRP1

Curcumin and Multidrug Resistance Proteins

Chem. Res. Toxicol., Vol. 16, No. 12, 2003 1647

Figure 6. Effect of CUR on calcein excretion from MDCKII-MRP1 cells (A) and MDCKII-MRP2 cells (B). Concentrations of CUR used were 0 (closed square), 20 (open square), 30 (closed diamond), 40 (open diamond), or 50 µM (closed circle). In MRP1 cells, 50 µM MK571 was used as MRP1 inhibitor (A, open circle). In MRP2 cells, 40 µM cyclosporine A was used to inhibit MRP2 (B, open circle). Panels C and D show dose-response curves of CUR on both calcein efflux (closed circles) and DNPSG efflux (open circles) for MDCKIIMRP1 cells (C, basolateral efflux data) and MDCKII-MRP2 cells (D, apical efflux data). Each point represents average ( SD of experiments performed in triplicate.

cells treated with 50 µM CUR was observed when MRP1 cells were treated with 50 µM MK571 (a typical inhibitor for MRP1, Figure 5B). Treatment of MRP2 cells with CUR (Figure 5C) did not affect the efflux of DNPSG to the apical side, where the MRP2 protein is localized. In contrast, the relatively low efflux of DNPSG to the basolateral side was affected upon exposure to CUR, which resulted in a slightly enhanced intracellular DNPSG level. Cyclosporine A reduced the efflux of DNPSG to the apical side to 60% of the efflux in DMSO-treated MRP2 cells, whereas the efflux to the basolateral side was enhanced to a level of 145% of that in the DMSO-treated cells, a phenomenon that was observed before (40). Cyclosporine A enhanced the intracellular DNPSG level in the MRP2-transfected cells by a factor of 4. As shown in Figure 5, the total formation of DNPSG in MRP1 cells, however, appeared to be significantly reduced upon exposure to CUR, as compared to the control and MRP2 cells. These results indicate a possible interference of CUR with the formation of DNPSG in MRP1 cells, in which process the low GSH levels as observed in these MRP1 cells (see Table 1) could play a role. Calcein Efflux in Control, MRP1-, and MRP2Transfected MDCKII Cells. To study the MRP1mediated transport in a model system where the efflux could not be hampered by low GSH levels, transport experiments were performed with calcein, an organic anion substrate, which is transported by MRP1 and MRP2 without being conjugated to GSH. Incubation of MDCKII cells with the nonfluorescent ester of calcein,

i.e., calcein-AM for 90 min at low temperature, resulted in an intracellular calcein level (due to intracellular esterases, which converts calcein-AM in calcein) comparable for all three cell lines (359 ( 41 fluorescence units/ monolayer). A typical “polarized” efflux pattern of calcein in time, comparable to that of DNPSG (see Figure 4), was observed in the MRP1 and MRP2 cells (data not shown). In MRP1 cells, calcein is predominantly excreted to the basolateral side (seven times higher than apical efflux) whereas in MRP2 cells the efflux of calcein is predominantly to the apical side (10 times higher than basolateral efflux). In this assay, an IC50 value of 5 µM MK571 for MRP1 basolateral calcein efflux was determined, whereas an IC50 value of 10 µM cyclosporine A for MRP2 apical calcein efflux was determined (results not shown). Effect of CUR on Calcein Efflux in Control, MRP1-, and MRP2-Transfected MDCKII Cells. CUR clearly inhibited the efflux of calcein from MRP1 cells to the basolateral side (Figure 6A), whereas in the MRP2 cells, CUR only slightly inhibited the apical efflux of calcein (Figure 6B). Because of the very limited efflux of calcein to the apical side in MRP1 cells and to the basolateral side in MRP2 cells, inhibitory data of CUR are only given for the major efflux side of calcein. The typical MRP1 inhibitor MK571 inhibited the calcein efflux from MRP1 cells to the basolateral compartment almost completely (Figure 6A), whereas in the MRP2 cells, the MRP2 inhibitor cyclosporine A inhibited the apical calcein efflux, although not completely (Figure 6B). Figure 6C,D shows the dose-dependent effect of CUR on calcein efflux to the basolateral side in MRP1 cells and the apical side in MRP2 cells, respectively. For

1648 Chem. Res. Toxicol., Vol. 16, No. 12, 2003

Wortelboer et al.

Figure 7. Representative HPLC chromatogram of the glutathionyl adducts of CUR (peaks 1 and 2) and the parent compound CUR (peak 3) present in the apical medium of MDCKII-MRP2 cells treated with 40 µM CUR for 90 min. Inserts represent LC-MS spectra, revealing an m/z of 675.9 and 675.8 for the monoglutathionyl CUR adducts.

Figure 8. Time-dependent efflux of CUR metabolites, i.e., peak 1 monoglutathionyl CUR (open symbols) and peak 2 monoglutathionyl CUR (closed symbols) in control (A), MRP1 cells (B), and MRP2 cells (C) both to the basolateral compartment (circles) and to the apical compartment (triangles). Cells were exposed to 40 µM CUR.

comparison, similar data on DNPSG efflux are included in Figure 6C,D. For the CUR-mediated inhibition of both the calcein and the DNPSG efflux in MRP1 cells, an IC50 value of 45 µM CUR was revealed (Figure 6C). Interestingly, in MRP2 cells, the calcein and DNPSG efflux was only inhibited to a limited extent, even by the maximal noncytotoxic concentration of CUR (50 µM) (Figure 6D). Comparison of these results with the inhibition curves obtained for the MRP1 and MRP2 containing Sf9 vesicles (Figure 3) reveals that higher levels of CUR are needed in MDCKII cells, as compared to the membrane vesicles to inhibit MRP1- and especially MRP2-mediated transport. Formation of Glutathione Conjugates of CUR in MDCKII Cells. To find a rationale for the less efficient CUR inhibition in cells as compared to vesicle studies, it was investigated whether biotransformation of CUR in the MDCKII-MRP1 and the MDCKII-MRP2 cells could be a factor causing a reduced concentration of the free inhibitor, CUR, resulting in less efficient inhibition.

Because biotransformation of CUR can be expected to proceed by GSH conjugation (41, 42), the glutathionedependent metabolism of CUR was investigated in control, MRP1-, and MRP2-transfected MDCKII cells, upon exposure to 40 µM CUR. HPLC diode array and LC/MS analysis of the culture media reveal the formation of two CUR derivatives (Figure 7), with retention times, UV spectra, and LC/MS characteristics identical to the chemically synthesized glutathione CUR adducts (i.e., tR ) 22.7 and 23.7 min). Both glutathione metabolites appeared to have identical absorption spectra with a λmax of 380 nm. LC/MS analysis revealed an m/z of 675.8 and 675.9 for the subsequent isolated metabolites, which is identical to the mass expected for a protonated monoglutathionyl CUR adduct. Figure 8 presents the efflux of the mono-CURSGs to the apical and basolateral side of control, and MRP1- or MRP2-transfected MDCKII cells. The amount of excreted monoglutathionyl CURs was the highest in MRP2 cells, followed by the control cells and MRP1 cells. The low

Curcumin and Multidrug Resistance Proteins

Figure 9. Dose-dependent effect of the mixture of monoglutathionyl CUR (55% CURSG1 and 45% CURSG2) on the ATPdependent [14C]EASG uptake in Sf9 cell membrane vesicles, containing MRP1 (A) or MRP2 (B). Data are given as percentage of control uptake in the absence of CUR. Each point is the average ( SD of triplicate incubations.

efflux of mono-CURSGs from MRP1 cells, as compared to the other two cell lines, is most likely the result of the low intracellular GSH levels in these cells. A typical polarized efflux pattern of these CURSGs was observed similar to the calcein and DNPSG efflux, indicating that these glutathionyl conjugates of CUR are also substrates of both MRP1 and MRP2. Curcumin Efflux in MRP1 and MRP2 Cells. To study whether curcumin itself is a substrate for MRP1 or MRP2, both control and MRP1- and MRP2-transfected MDCKII cells were loaded with curcumin (80 nmol/ monolayer), followed by studying the efflux of curcumin into fresh medium. The excretion of curcumin in all three MDCKII cell lines was identical, i.e., 12.3 ( 2.6 nmol/ monolayer for the first 30 min (of which 24% was in the apical compartment and 76% in the basolateral compartment). A typical “polarized” efflux pattern of the glutathionyl curcumin conjugates was observed (Figure 8), and no such efflux pattern of the parent compound curcumin was observed, which indicates that curcumin is not a substrate for either MRP1 or MRP2. Inhibition of MRP1- and MRP2-Mediated [14C]EASG Transport in Sf9 Membrane Vesicles by the Mixture of CURSGs. To study whether the CURSGs interact with MRP1 and/or MRP2 transport activity, a mixture of the monoCURSGs (of which 55% CURSG1 and 45% CURSG2) was used to determine the effect on the ATP-dependent uptake of [14C]EASG uptake in Sf9 vesicles. The dose-response curves (Figure 9) reveal an IC50 value of about 20 and 30 µM for inhibition of MRP1 and MRP2, respectively, by the mixture of CURSGs.

Discussion In the present study, the interaction of the dietary polyphenol CUR with human multidrug resistance proteins MRP1 and MRP2 and the interplay between CURdependent MRP inhibition and its glutathione-dependent metabolism were investigated. Using isolated membrane vesicles of MRP1- and MRP2expressing Sf9 cells, it was demonstrated that CUR inhibits both MRP1- and MRP2-mediated transport with IC50 values of 15 and 5 µM, respectively. The MRP1 and MRP2 inhibitory potential of CUR is comparable to that of other MRP inhibitors tested in similar model systems using membrane vesicles of MRP-expressing cells, such as specific tyrosine kinase inhibitors (43), MK-571 (36), indomethacin analogues (44), sulfinpyrazone (26), vera-

Chem. Res. Toxicol., Vol. 16, No. 12, 2003 1649

pamil analogues (45), probenecid (26, 46), and prostaglandins (47), all showing IC50 values in the micromolar range. It must be stressed, however, that outcomes of these different studies should be compared with care since various types and concentrations of substrates were used to monitor MRP-mediated transport and various types of MRP-expressing cells were used as source for membrane vesicles. The IC50 values observed indicate that in vivo inhibition of MRP1 and MRP2 by CUR may occur, because upon use of commercially available CUR supplements that contain up to 2-4 g of CUR per day or upon a recommended therapeutic oral dose of CUR of 6-8 g per day, suggested to be used in future phase II studies (22), local concentrations of CUR in the gastrointestinal tract are expected to reach the micromolar range. Upon oral intake of 8 g per day, a serum concentration of about 2 µM was detected (22), while concentrations observed in the liver and colon mucosa were even 75-150 times higher than those observed in plasma (22, 48). Thus, CUR intake can be expected to result in concentrations that are able to modulate uptake and kinetics of other food ingredients or even drugs through this inhibitory effect on both MRP1 and MRP2. The mechanism of inhibition of MRP1 and MRP2 transport activity by CUR is unknown. For the ABC transporter Pgp, it has been demonstrated recently that CUR inhibits both verapamil-stimulated ATPase activity and the photo affinity labeling of Pgp with a prazosin analogue, indicating that CUR interacts directly with the transporter, at either the substrate or the ATPase site (10). Because CUR contains two R,β-unsaturated ketone moieties, this compound can conjugate with glutathione through Michael addition nonenzymatically or through a GST-catalyzed reaction to form glutathione conjugates (42). Besides CUR, its glutathione conjugates may also interact with transmembrane transporters of the MRP family. Especially in a cellular system in which glutathione and GSTs are present, glutathione conjugation of CUR may occur and may affect the level of CURmediated inhibition of MRP1 and MRP2. However, because CUR, but also its glutathione conjugates, were able to inhibit MRP1 and MRP2, the concentration of CUR and its glutathione conjugates at the target site of MRP as well as their intrinsic potency to inhibit MRPmediated transport activity can be expected to determine the ultimate efficacy of CUR-mediated MRP inhibition. Results of the present study reveal that this glutathionedependent metabolism of CUR indeed interferes with the efficiency of its MRP1 and MRP2 inhibition. Namely, in intact polarized monolayers of MDCKII cells overexpressing MRP1, CUR inhibited MRP1-mediated activity with a 3-fold higher IC50 value than the one observed in the vesicle model. Interestingly, MRP2-mediated activity was hardly inhibited in intact MDCKII cells overexpressing MRP2 upon exposure to CUR, whereas the IC50 value for MRP2 inhibition in the vesicle incubations was 5 µM. The difference in extent of inhibition of the MRPs by CUR in isolated vesicles as compared to intact cells, observed especially for MRP2, was shown to be due to a swift metabolism of CUR to two glutathione conjugates. Formation of both glutathione conjugates was about six times higher in the MDCKII-MRP2 cells as compared to the MDCKII-MRP1 cells, a phenomenon that can be ascribed to the significantly lower glutathione levels observed in the latter cell line.

1650 Chem. Res. Toxicol., Vol. 16, No. 12, 2003

Until present, the formation of glutathionylcurcumin conjugates was never detected in vivo or in cellular systems (49). Recently, Awasthi et al. (42) showed the synthesis of several products in an incubation of glutathione with CUR, although no exact positive identification of the products could be made. In the present study, both glutathione conjugates of CUR for the first time could be identified by HPLC diode array and LC/MS analysis as monoglutathionylcurcumin conjugates. The exact structure of the monoglutathionyl conjugates of CUR awaits their characterization by NMR, which is hampered by the unstable nature of the isolated conjugates. Both monoglutathionylcurcumin conjugates, in contrast to the parent compound CUR, appear to be transported by MRP1 and MRP2, respectively, because cellular efflux of the mono-CURSGs was observed predominantly at the basolateral side in the MDCKII cells overexpressing MRP1 and predominantly at the apical side in the MDCKII cells overexpressing MRP2. The monoglutathionyl conjugates of CUR, however, are less potent inhibitors of MRP1 and MRP2 as compared to the parent compound CUR as is concluded from their higher IC50 values for inhibition of MRP-mediated transport activity in vesicles as compared to the IC50values of CUR. A less potent inhibitory potential of the glutathione conjugates of CUR as compared to CUR is in accordance with the results obtained for MRP inhibition in the MDCKII-MRP1 and MDCKII-MRP2 cells, where enhanced formation of glutathione conjugates of CUR results in less inhibition of MRP. For other compounds with an R,β-unsaturated moiety in their structure like PGA1, PGA2, and ethacrynic acid, it was demonstrated that their GS metabolites are high affinity substrates for MRP1, whereas the parent compounds PGA1, PGA2, and ethacrynic acid are not substrates or inhibitors of MRP1 (26, 47). Apparently, the activity profile of compounds containing an R,β-unsaturated moiety for their ability to inhibit MRP1- and MRP2mediated transport processes varies with the compound, suggesting that some of these compounds interact with different sites of the MRP molecules. In conclusion, the data of the present study clearly indicate that CUR inhibits both MRP1- and MRP2mediated transport and that the glutathione-dependent metabolism of CUR plays a crucial role in the ultimate level of inhibition of MRP-mediated transport that can be achieved in a cellular system. The results of the present study also point at possible CUR-drug or CURxenobiotic interactions, especially when the pharmacoor toxicokinetics of the drug or xenobiotic is dependent on MRP1- and/or MRP2-mediated transport processes. Further research is needed to study whether glutathione conjugation of CUR also occurs in human (intestinal and liver) cells. The complex interplay between MRP inhibition and metabolism of MRP-inhibitors, the latter affecting the ultimate potential of a compound for cellular MRP inhibition may exist not only for a compound like CUR but also for many other MRP inhibitors presently or previously developed on the basis of vesicle studies.

Acknowledgment. We thank Prof. P. Borst from the National Cancer Institute (Amsterdam, The Netherlands) who kindly provided the transfected MDCKII cell lines. The baculovirus encoding MRP1 or encoding MRP2 were kind gifts from Prof. Dr. B. Sarkadi from the National Institute of Haematology and Immunology, Research

Wortelboer et al.

Group of the Hungarian Acadamy of Sciences, Budapest, Hungary. This work was supported by Grant TNOV20002169 from the Dutch Cancer Society.

References (1) Keppler, D., Leier, I., Jedlitschky, G., and Koning, J. (1998) ATPdependent transport of glutathione S-conjugates by the multidrug resistance associated protein MRP1 and its apical isoform MRP2. Chem.-Biol. Interact. 111-112, 153-161. (2) Ko¨nig, J., Nies, A. T., Cui, Y., Leir, I., and Keppler, D. (1999) Conjugate export pumps of the multidrug resistance protein (MRP) family: localization, substrate specificity and MRP2mediated drug resistance. Biochim. Biophys. Acta 1461, 377-394. (3) Borst, P., Evers, R., Kool, M., and Wijnholds, J. (2000) A family of drug transporters; the Multidrug resistance-associated proteins. J. Natl. Cancer Inst. 92, 1295-1302. (4) Renes, J., de Vries, E. G. E., Jansen, P. L. M., and Mu¨ller, M. (2000) The (patho)physiological functions of the MRP family. Drug Resist. Updates 3, 289-302. (5) McLellan, L. I., and Wolf, C. R. (1999) Glutathione and glutathione dependent enzymes in cancer drug resistance. Drug Resist. Updates 2, 153-164. (6) Tew, K. D., O’Brien, M., Laing, N. M., and Shen, H. (1998) Coordinate changes in expression of protective genes in drugresistant cell lines. Chem.-Biol. Interact. 111-112, 199-211. (7) Cnubben, N. H. P., Rietjens, I. M. C. M., Wortelboer, H., van Zanden, J., and van Bladeren, P. J. (2001) The interplay of glutathione-related processes in antioxidant defence. Environ. Toxicol. Pharmacol. 10, 141-152. (8) Locigno, R., and Castranovo, V. (2001) Reduced glutathione system: role in cancer development, prevention and treatment (review). Int. J. Oncol. 19, 221-236. (9) Versantvoort, C. H. M., Rhodes, T., and Twentyman, P. R. (1996) Acceleration of MRP-associated efflux of rhodamine 123 by genistein and related compounds. Br. J. Cancer 74, 1949-1954. (10) Anuchapreeda, S., Leechanachai, P., Smith, M. M., Ambudkar, S. V., and Limtrakul, P. (2002) Modulation of P-glycoprotein expression and function by curcumin in multidrug-resistant human KB cells. Biochem. Pharmacol. 64, 573-582. (11) Romiti, N., Tongiani, R., Cervelli, F., and Chieli, E. (1998) Effects of curcumin on P-glycoprotein in primary cultures of rat hepatocytes. Life Sci. 62, 2349-2358. (12) Leslie, E. M., Mao, Q., Curtis, J., Oleschuk, C. J., Deeley, R. G., and Cole, S. P. C. (2001) Modulation of multidrug resistance protein 1 (MRP1/ABCC1) transport and ATPase activities by interaction with dietary flavonoids. Mol. Pharmacol. 59, 11711180. (13) van Zanden, J. J., Haman. O. B., van Iersel, M. L. P. S., Boeren, S., Cnubben, N. H. P., Lo Bello, M., Vervoort, J., van Bladeren, P. J., and Rietjens, I. M. C. M. (2003) Inhibition of human glutathione S-transferase P1-1 by the flavonoid quercetin. Chem.Biol. Interact. 145, 139-148. (14) Zhang, K., and Wong, K. P. (1997) Glutathione conjugation of chlorambucil: measurement and modulation by plant polyphenols. Biochem. J. 325, 417-422. (15) Zhang, K., Mack, P., and Wong, K. P. (1998) Glutathione-related mechanisms in cellular resistance to anticancer drugs. Int. J. Oncol. 12, 871-882. (16) Harbottle, A., Daly, A. K., Atherton, K., and Campbell, C. (2001) Role of glutathione S-transferase P1, P-glycoprotein and multidrug resistance associated protein 1 in acquired doxorubicin resistance. Int. J. Cancer 92, 777-783. (17) Iersel, M. L. P. S., van Ploemen, J. P. H. T. M., Struik, I., van Amersfoort, C., Keyzer, A. E., Schefferlie, J. G., and van Bladeren, P. J. (1996) Inhibition of glutathione S-transferase activity in human melanoma cells by R,β-unsaturated carbonyl derivatives. Effects of acrolein, cinnamaldehyde, citral, crotonaldehyde, curcumin, ethacrynic acid, and trans-2-hexenal. Chem.-Biol. Interact. 102, 117-132. (18) Osawa, T., Sugiyama, Y., Inayoshi, M., and Kawakashi, S. (1995) Antioxidative activity of tetrahydrocurcuminoids. Biosci., Biotechnol., Biochem. 59, 1609-1612. (19) Stoner, G. D., and Mukhtar, H. (1995) Polyphenols as cancer chemopreventive agents. J. Cell. Biochem. 22, 169-180. (20) Commandeur, J. N. M., and Vermeulen, N. P. E. (1996) Cytotoxicity and cytoprotective activities of natural compounds. The case of curcumin. Xenobiotica 26, 667-680. (21) Levi, M. S., Borne, R. F., and Williamson, J. S. (2001) A review of cancer chemopreventive agents. Curr. Med. Chem. 8, 13491362.

Curcumin and Multidrug Resistance Proteins (22) Cheng, A.-L., Hsu, C.-H., Lin, J.-K., Hsu, M.-W., Ho, Y.-F., Shen, T.-S., Ko, J.-Y., Lin, J.-T., Lin, B.-R., Wu, M.-S., Yu, H.-S., Jee, S.-H., Chen, G.-S., Chen, T.-M., Chen, C.-A., Lai, M.-k., Pu, Y.S., Pan, M.-H., Wang, Y.-J., Tsai, C.-C., and Hsieh, C.-Y. (2001) Phase I Clinical Trial of curcumin, a chemopreventive agent, in patients with high risk or pre-malignant lesions. Anticancer Res. 21, 2895-2900. (23) Ireson, C. R., Jones, D. J. L., Orr, S., Coughtrie, M. W. H., Boocock, D. J., Williams, M. L., Farmer, P. B., Steward, W. P., and Gesher, A. J. (2002) Metabolism of the cancer preventive agent curcumin in human and rat intestine. Cancer Epidemiol. Biomarkers Prev. 11, 105-111. (24) Ploemen, J. H. T. M., Van Schanke, A., van Ommen, B., and van Bladeren, P. J. (1994) Reversible conjugation of ethacrynic acid with glutathione and human glutathione S-transferase P1-1. Cancer Res. 54, 915-919. (25) Sokolovsky, M., Sadeh, T., and Patchornik, A. (1964) Specific chemical cleavage of cysteinyl peptides. J. Am. Chem. Soc. 86, 1212-1217. (26) Zaman, G. J. R., Cnubben, N. H. P., van Bladeren, P. J., Evers, R., and Borst, P. (1996) Transport of the glutathione conjugate of ethacrynic acid by the human multidrug resistance protein MRP. FEBS Lett. 391, 126-130. (27) van Aubel, R. A. M. H., Masereeuw, R., and Russel, F. G. M. (2000) Molecular pharmacology of renal organic anion transporters. Am. J. Physiol. Renal Physiol. 279, F216-F232. (28) Bradford, M. (1976) A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248-254. (29) Bakos, E., Hegedus, T., Hollo, Z., Welker, E., Tusnady, G. E., Zaman, G. J. R., Flens, M. J., Varadi, A., and Sarkadi, B. (1996) Membrane topology and glycosylation of the human multidrug resistance-associated protein. J. Biol. Chem. 271, 1232212326. (30) Gao, M., Loe, D. W., Grant, C. E., Cole, S. P. C., and Deeley, R. G. (1996) Reconstruction of ATP-dependent leukotriene C4 transport by coexpression of both half-molecules of human multidrug resistance protein in insect cells. J. Biol. Chem. 271, 2778287. (31) Evers, R., Kool, M., van Deemter, L., Janssen, H., Calafat, J., Oomen, L. C. J. M., Paulusma, C. C., Oude Elferink, R. P. J., Baas, F., Schinkel, A. H., and Borst, P. (1998) Drug export activity of the human canalicular multispecific organic anion transporter in polarized kidney MDCK cells expressing cMOAT (MRP2) cDNA. J. Clin. Invest. 101, 1310-1319. (32) Evers, R., de Haas, M., Sparidans, R., Beijnen, J., Wielinga, P. R., Lankelma, J., and Borst, P. (2000) Vinblastine and sulfinpyrazone export by the multidrug resistance protein MRP2 is associated with glutathione export. Br. J. Cancer 83, 375-383. (33) Habig, W. H., Pabst, M. J., and Jakoby, W. B. (1974) Glutathione S-transferases, the first step in mercapturic acid formation. J. Biol. Chem. 249, 7130-7139. (34) Anderson, M. E. (1985) Determination of glutathione and glutathione disulfide in biological samples. Methods Enzymol. 113, 548-555. (35) Irvine, J. D., Takahashi, L., Lockhart, K., Cheong, J., Tolan, J. W., Selick, H. E., and Grove, J. R. (1998) MDCK (Madin Darby Canine Kidney) Cells: A tool for membrane permeability screening. J. Pharm. Sci. 88, 28-33. (36) Chen, Z.-S., Kawabe, T., Ono, M., Aoki, S., Sumizawa, T., Furukawa, T, Uchiumi, T., Wada, M., Kuwano, M., and Akiyama, S.-I. (1999) Effect of multidrug resistance-reversing agents on

Chem. Res. Toxicol., Vol. 16, No. 12, 2003 1651

(37) (38)

(39)

(40)

(41) (42)

(43)

(44)

(45)

(46)

(47)

(48)

(49)

transporting activity of human canalicular multispecific organic anion transporter. Mol. Pharmacol. 56, 1219-1228. Borenfreund, E., and Peurner, J. A. (1985) Toxicity determined in vitro by morphological alterations and neutral red absorption. Toxicol. Lett. 24, 119-124. Evers, R., Zaman, G. J. R., van Deemter, L., Jansen, H., Calafat, J., Oomen, L. C. J. M., Oude Elferink, R. P. J., Borst, P., and Schinkel, A. H. (1996) Basolateral localization and export activity of the human multidrug resistance-associated protein in polarized pig kidney cells. J. Clin. Invest. 97, 1211-1218. Hollo, Z., Homoloya, L., Hegedus, T., Muller, M., Szakacs, G., Jakab, K., Antal, F., and Sarkadi, B. (1998) Parallel functional and immunological detection of human multidrug resistance proteins, P-glycoprotein and MRP1. Anticancer Res. 18, 29812987. Evers, R., Kool, M., Smith, A. J., van Deemter, L., de Haas, M., and Borst, P. (2000) Inhibitory effect of the reversal agents V-104, GF120918 and Pluronic L61 on MDR1 Pgp-, MRP1- and MRP2mediated transport. Br. J. Cancer 83, 366-374. Mathews, S., and Rao, M. N. A. (1991) Interaction of curcumin with glutathione. Int. J. Pharm. 76, 257-259. Awasthi, S., Pandya, U., Singhal, S. S., Lin, J. T., Thiviyananathan, V., Seifert, W. E., Jr., Awasthi, Y. C., and Ansari, G. A. S. (2000) Curcumin-glutathione interactions and the role of human glutathione S-transferase P1-1. Chem.-Biol. Interact. 128, 1938. Hegedus, T., Orfi, L., Sepofdi, A., Varadi, A., Sarkadi, B., and Keri, G. (2002) Interaction of tyrosinase kinase inhibitors with the human multidrug transporter proteins MDR1 and MRP1. Biochim. Biophys. Acta 1587, 318-325. Touhey, S., O’Connor, R., Plunkett, S., Maguire, A., and Clynes, M. (2002) Structure-activity relationship of indomethacin analogues for MRP1, COX-1, COX-2 inhibition: identification of novel chemotherapeutic drug resistance modulators. Eur. J. Cancer 38, 1661-1670. Loe, D. W., Oleschuk, C. J., Deeley, R. G., and Cole, S. P. (2000) Structure-activity studies of verapamil analogues that modulate transport of leukotrieneC4 and reduced glutathione by multidrug resistance protein MRP1. Biochem. Biophys. Res. Commun. 275, 795-803. Bakos, E., Evers, R., Sinko, E., Varadi, A., Borst, P., and Sarkadi B. (2000) Interactions of the human multidrug resistance proteins MRP1 and MRP2 with organic anions. Mol. Pharmacol. 57, 760768. Evers, R., Cnubben, N. H. P., Wijnholds, J., Van Deemter, L., Van Bladeren, P. J., and Borst, P. (1997) Transport of glutathione prostaglandin A conjugates by the multidrug resistance protein 1. FEBS Lett. 419, 112-116. Sharma, R. A., Ireson, C. R., Verschoyle, R. D., Hill, K. A., Williams, M. L., Leuratti, C., Manson, M. M., Marnett, L. J., Steward, W. P., and Gescher A. (2001) Effect of dietary curcumin on glutathione S-transferases and malondialdehyde-DNA adducts in rat and colon mucosa: relationship with drug levels. Clin. Cancer Res. 7, 1452-1458. Ireson, C., Orr, S., Jones, D. J. L., Verschoyle, R., Lim, C.-K., Luo, J.-L., Howells, L., Plummer, S., Jukes, R., Williams, M., Steward, W. P., and Gesher, A. (2001) Characterization of metabolites of the chemopreventive agent curcumin in human and rat hepatocytes and in the rat in vivo, and evaluation of their ability to inhibit phorbol ester-induced prostaglandin E2 production. Cancer Res. 61, 1058-1064.

TX034101X