Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

GSH and its associated phase II enzymes, the glutathione S-transferases (GSTs), represent a major detoxification route in the metabolism of harmful xenobiotics and endogenous products of oxidative stress (O'Brien and Tew, 1996). They have been extensively studied for their role in metabolizing anticancer agents. GSH has been shown to form thioether conjugates with several important chemotherapeutic drugs, most notably alkylating agents or their metabolites (Hayes and Pulford, 1995). Overexpression of GSTs has often been associated with resistance to such drugs and has therefore become an obstacle in the effective treatment of clinical disease. Overexpression of GSTs is usually isozyme specific. The literature is replete with examples of increased expression of GSTP1-1 in tumors or cell lines resistant to the anthracycline, Vinca alkaloid, and epipodophyllotoxin groups of anticancer agents, despite the fact that GSH conjugates to these drugs have never been reported (Tew, 1994). Several chemical inhibitors of GSTs have been studied for their role in potentiating anticancer drugs, however, they lack strong isozyme specificity (Tew et al., 1988; Hall et al., 1989; Smith et al., 1989; Ford et al., 1991; Hansson et al., 1991). Furthermore, these drugs have additional pharmacological properties independent of their GST inhibitory function, which make them less suitable for the clinical modulation of drug resistance (Bach et al., 1987; Kunigi et al., 1991; Xu et al., 1992).

Efforts have been made over the last decade to synthesize inhibitors of specific isoforms of GST. Earlier work provided the basis for synthesis of structure-based analogs of GSH to be tested as isozyme-specific inhibitors of the GSTs. Substituting the C-terminal glycine of GSH (-glu-cys-gly) altered the V_max and K_m for each isozyme in the conjugation of GSH to 1-chloro-2,4-dinitrobenzene (Adang et al., 1990). Askelof et al. (1975) showed that functionalizing the sulfur of the cysteine residue of GSH with various n-alkyl groups affected the potency and specificity of the resultant GSH analogs as inhibitors of rat GSTs. More recently, Lyttle et al. (1994) used this information in the systematic synthesis of isozyme-specific inhibitors of human GSTs. One compound, -glutamyl-S-(benzyl)cysteinyl-R-()-phenyl glycine (TER117) (Fig. 1), was produced by substituting glycine with phenyl glycine, and functionalizing the thiol of cysteine with a benzyl group to yield a compound specific for the inhibition of GSTP1-1 (Flatgaard et al., 1993). The modifications of GSH resulting in TER117 allow for its binding at the G site of the enzyme with an affinity ~1000-fold greater than that for GSH. Ethyl esterification of the two free carboxyl groups of the peptidomimetic resulted in TER199, and thus membrane-permeating potential. Following cellular uptake, the esters are cleaved and TER117 is generated. TER199 was found to inhibit GSTP1-1 and to potentiate the toxicity of chlorambucil in several human derived tumor cell lines and to enhance the efficacy of melphalan in a xenograft of human colonic tumor cells that overexpress GSTP1-1 (Morgan et al., 1996).

View larger version (8K): [in this window] [in a new window]   Fig. 1.   Structure of TER117. The structure of TER117 resembles that of GSH, with the addition of a benzyl group to the sulfur atom and the substitution of glycine with phenylglycine.

Because TER117 and TER199 resemble a GSH conjugate, we sought to investigate if this drug could modulate multidrug resistance by a mechanism in addition to GSTP1-1 inhibition. Multidrug resistance-associated protein1 (MRP1) is an ATP-dependent transporter that has been implicated in multidrug resistance to a broad category of anticancer drugs (Cole et al., 1994; Breuninger et al., 1995). With inside-out vesicles prepared from MRP1-overexpressing cells, it has recently been demonstrated that MRP1 transports GSH conjugates and oxidized glutathione disulfide (Jedlitschky et al., 1994, 1996; Muller et al., 1994; Leier et al., 1996; Loe et al., 1996; Zaman et al., 1996). This technique also has been used to determine whether MRP1 transports the anthracycline, epipodophyllotoxin, and Vinca alkaloid drugs. Despite the fact that cells overexpressing MRP1 are resistant to these drugs, direct transport has generally not been shown with the aforementioned technique. Vincristine and daunorubicin have been shown to be transported in the presence of GSH (Loe et al., 1996; Renes et al., 1999). Other methods, however, have shown these three drug classes to be substrates for MRP1. These include the stimulation of ATPase activity of purified MRP1 by anthracyclines and Vinca alkaloids (Chang et al., 1997), the ATP-dependent vanadate trapping of MRP1 in the presence of epipodophyllotoxins and Vinca alkaloids (Taguchi et al., 1997), as well as a fluorescence technique used by Marbeuf-Gueye et al. (1998) to describe the kinetics of anthracycline efflux from MRP1-expressing cells. If TER117 were a substrate of MRP1, it could be used as an inhibitor of MRP1-mediated transport for the various chemotherapeutic agents, or their metabolites, to which it confers multidrug resistance. The combined modulation of drug resistance at both the level of GST-mediated conjugation and of MRP1 transport inhibition represents a potentially useful means of using one compound to circumvent a multifaceted drug-resistance phenotype.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials and Cell Lines. [³H]Dinitrophenylglutathione ([³H]GS-DNP; 1.45Ci/mmol) was synthesized with the following procedure. All steps before QAE-Sephadex chromatography were conducted under nitrogen or with nitrogen-saturated solutions. [³H]GSH (250 µCi) (43.8 Ci/mmol; NEN Life Science Products, Boston, MA) was removed from dithiothreitol by ethyl acetate extraction as previously described (Butler et al., 1996). The dithiothreitol-free [³H]GSH was diluted with unlabeled GSH to 4 µmol and was mixed with 8 µmol of 1-chloro-2,4-dinitrobenzene. The reaction was started by addition of 10 µg of rat liver glutathione transferase in a final volume of 1 ml containing 10 mM Tris-HCl, pH 7.4. After 2 h of incubation at 37°C, the sample was pipetted onto a QAE-Sephadex column (0.4-ml bed volume) pre-equilibrated with 10 mM Tris-HCl, pH 7.4, and washed with 5 volumes of the same buffer. The bound [³H]GS-DNP was eluted with 0.5 N formic acid and lyophilized. The resulting purified [³H]GS-DNP was resuspended in 0.5 ml of distilled H₂0 and stored at 80°C. Doxorubicin, daunorubicin, etoposide, vincristine, 5-fluorouracil (5-FU), mitoxantrone, mitomycin C, cisplatin, melphalan, paclitaxel, ATP, and creatine phosphate were obtained from Sigma Chemical Co. (St. Louis, MO). [³H]Daunorubicin was purchased from NEN Life Science Products. Creatine kinase was purchased from Boehringer Mannheim (Indianapolis, IN). NIH3T3/MRP1 (transfected with the pSRtkneo plasmid containing MRP1) and NIH3T3/MSV (transfected with the pSRtkneo plasmid) cell lines were kindly provided by Dr. Gary Kruh (Fox Chase Cancer Center). The HL-60 cell line was purchased from American Type Culture Collection (Rockville, MD).

Analysis of Cellular Drug Accumulation. NIH3T3 transfectants were seeded at a density of 7 × 10⁵ cells/flask in 25-cm² culture dishes. After overnight growth in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum, the media was replaced with serum-free media containing 1 µM daunorubicin at 0.5 µCi of [³H]daunorubicin and 50 µM TER199, and allowed to incubate for the various time points indicated. The cells were then washed with PBS, lysed with 0.5% SDS in PBS, and counted in a liquid scintillation counter.

Analysis of Intracellular Daunorubicin Distribution. NIH3T3/MRP1 cells were seeded at a density of 1 × 10⁵ on sterilized glass coverslips in 35-mm culture dishes. After growth for 24 h, the cells were incubated in serum-free media containing 20 µM daunorubicin with or without 50 µM TER199 for 2 h. The cells were then washed and analyzed immediately after incubation in DMEM containing 50 µM TER199 or its solvent, 0.05% dimethyl sulfoxide (DMSO), for various time points. Daunorubicin fluorescence was detected by a cooled charge-coupled device camera with a 555 excitation filter and a multiband pass emission filter on a Nikon eclipse TE300 inverted microscope. All images were taken with a 60× Plan Fluor long working-distance objective. The images accumulated by a Photometrics Quantix camera were acquired and analyzed with ISEE software (Inovision, Durham, NC).

Immunofluorescence Analysis. Cellular immunofluorescent studies of MRP in NIH3T3 cells and NIH3T3/MRP1 cells was accomplished with standard protocols. Cells were grown on glass coverslips to a density of 1 ×10⁵cells/ml. The cells were fixed with 4% paraformaldehyde and 0.1% Triton X-100, and blocked with 10% fetal calf serum for 30 min. Monoclonal anti-MRP antibody raised in mouse (Kamiya Biomedical Company, Seattle, WA) was added to the cells to a final concentration of 4 µg/coverslip. After a 1-h incubation at 37°C, cells were treated with a secondary goat anti-mouse antibody (Vector Laboratories, Inc. Burlingame, CA), tagged with biotin as per manufacturer instructions for 1 h at 37°C. Then an antibiotin tertiary antibody raised in mouse (Sigma Chemical Co.) tagged with fluorescein isothiocyanate was applied to the cells for 1 h at 37°C in the dark. Between each step, multiple washes with PBS, 0.1% Triton X-100, and 1% BSA were performed to remove excess reagents. The prepared coverslips were then mounted with a Miowol antifade solution and observed by laser scanning confocal microscopy on a Bio-Rad MRC 600 laser scanning confocal microscope with the 60× objective of a Nikon Optiphot2.

Preparation of Plasma Membrane Vesicles. Plasma membrane vesicles were isolated according to Leier et al. (1994). Membrane preparations were frozen and stored in liquid nitrogen.

[³H]GS-DNP Transport Studies. Transport of [³H]GS-DNP into plasma membrane vesicles and its inhibition by unlabeled TER117 were measured by rapid filtration through nitrocellulose filters (Leier et al., 1994). The stock solution of TER117 in DMSO was made fresh on the day of experiment. The final DMSO concentration was 0.9%. In the control experiments without inhibitors, an appropriate dilution of DMSO was added to the incubation solution to control for any effects of the solvent. After preincubation at 37°C of 50 µl of the reaction mixture (250 mM sucrose, 10 mM Tris-HCl, pH 7.4, 4 mM ATP/4 mM 5'-AMP, 10 mM MgCl₂, 10 mM creatine phosphate, and 100 µM creatine kinase) containing various concentrations of [³H]GS-DNP, the incubation was started by addition of 60 µl of membrane vesicles (30 µg of protein). Twenty-microliter aliquots were taken at indicated time points and diluted in 1 ml of ice-cold incubation buffer. The diluted samples were filtered through presoaked nitrocellulose filters and rinsed three times with 3 ml of incubation buffer. Filters were dissolved and counted in a liquid scintillation counter. ATP-dependent transport rate was calculated by subtracting the corresponding values in the presence of 5'-AMP from those in the presence of ATP.

Establishment of TER199-Resistant Cells. HL-60 human myeloid leukemia cells were chronically exposed to TER199 and resistant cells were selected. Resistance was initiated with 2.5 µM TER199 and the drug concentration was subsequently increased in increments of 5 µM after three to four passages.

Analysis of Transcript Levels by Reverse Transcription-Polymerase Chain Reaction (RT-PCR). Total RNA was isolated from logarithmically growing cells with the Quiagen Rneasy total RNA kit. First-strand cDNA synthesis was performed in a 20-µl reaction volume containing the following: 1 µg of total RNA; 2.5 µM random hexamers (Promega Biotec, Madison, WI); 10 mM dithiothreitol; 0.2 M dATP, dCTP, dGTP, and dGTP; 50 mM Tris-HCl; 75 mM KCl; 3 mM MgCl₂; and 200 U of Superscript II reverse transcriptase (Gibco BRL, Paisley, Scotland). The PCR was performed in a 100 µl of reaction volume containing the following: 20 mM Tris-HCl, 50 mM KCl, 1.0 mM MgCl₂, 0.5 µM sense and antisense primers, 5 U of Taq DNA polymerase, and 50 ng of cDNA. The following primer sequences were used: ATC-AGT-GGG-GAC-AGG-TAA-AAC and GCT-CCT-AAG-TCA-GTT-AAG-AAC-C for -glutamyl cysteine synthetase (-GCS) light; TGA-AAC-TCT-GCA-AGA-GAA-GGG-G and GCT-TCA-TCT-GGA-AAG-AAG-AGG-G for -GCS heavy; GGC-TCA-CTA-AAA-GCC-TCC-TG and AGT-GCC-TTC-ACA-TAG-TCA-TC for GSTP1-1; CCA-GGA-GAC-TGC-TAT-CAT-GG and ACC-AGA-TGA-ATG-TGA-ATG-TCA-GCC-CG for GST; GGA-AAA-GAA-GTA-CAC-GAT-GGG-G and TTC-ACG-AAG-GAT-AGT-GGG-TAG-C for GSTµ; CTG-CTT-TCC-TGC-TGA-TCT-ATG-C and TGA-AGA-TTT-CAT-AAG-CTG-CTC-C for multidrug resistance1 gene (MDR1); CTG-GGC-TTA-TTT-CGG-ATC-AAC-G and TGA-CAC-GAA-GTC-CTT-CAG-GTG-G for MRP1; TGG-GAC-CCT-CAG-TCT-TAG-CAG-G and AGC-CGC-TCT-GTG-GAA-ACA-CTG-G for canalicular multispecific organic anion transporter (cMOAT); GAT-ACG-CTC-GCC-ACA-GTC-C and CAG-TTG-GCC-GTG-ATG-TGG-CTG for MRP3; and GGA-CGA-CAT-GGA-GAA-AAT-GTG-G and GGA-TAG-CAA-CGT-ACA-TGG-CTG-G for actin. PCR products were run on a 2% Metaphore (FMC Corp., Rockland, ME) agarose gel and stained with ethidium bromide to visualize DNA. Positive-negative photos were taken, and these were analyzed for relative band intensity with NIH Image software (version 1.57).

Western Blot Analysis. Protein concentrations were assayed with the Bradford reagent (Bio-Rad Laboratories, Hercules, CA). When immunoblotting for GCS heavy and light subunits, 200 µg of whole-cell extract was loaded onto a 10% SDS-polyacrylamide gel electrophoresis and run at 50 V for 16 h. The gel was blotted on polyvinylidene difluoride membrane (NEN Life Science Products) and transferred in transfer buffer containing 25 mM Tris base, 190 mM glycine, 20% methanol, and 0.1% SDS for 6 h at 50 V. For MRP1 immunoblotting 60 µg of the S-100 fraction of whole-cell lysate was run on 10% SDS-polyacrylamide gel electrophoresis (Bio-Rad Laboratories). The blots were blocked with 10% nonfat dry milk in Tris-buffered saline-T (50 mM Tris, pH 7.5, 2.3% NaCl, 0.2% Tween 20) for 2 h and incubated with rabbit polyclonal antibodies for -GCS heavy and -GCS light and with mouse monoclonal for MRP1. Antisera for -GCS heavy and light were generated against the following peptide epitopes: GLLSQGSPLSWEETK for -GCS heavy and LLTHNDPKELLSEAS for -GCS light. The monoclonal antisera against MRP1, MRPmAb-1, was kindly donated by the laboratory of Dr. Gary Kruh (Breuninger et al., 1995). The blots were washed in TBS-T three times for 20 min each and then incubated in secondary mouse or rabbit antibody conjugated to horseradish peroxidase and the proteins detected with a chemiluminescence detection kit (NEN Life Science Products).

Glutathione Measurements. Total intracellular glutathione levels were measured with a modified method of Tietze (1969). Briefly, an aliquot of cells was counted with a Coulter counter, and the remaining cells pelleted and resuspended in 400 µl of 3% 5-sulfosalicylic acid. They were sonicated for 30 s and placed at 4°C for 2 h before centrifugation at 10,000g. One hundred microliters of supernatant was then used in the spectrophotometric 1-chloro-2,4-dinitrobenzene assay as described by Tietze (1969).

Cytotoxicity Assays. Drug sensitivities were determined by the Cell Titer 96 aqueous nonradioactive cell proliferation assay (Promega Biotec). Cells (1000 cells/well for NIH3T3 or 10,000 cells/well for HL-60) were seeded in 100 µl of DMEM-10% fetal calf serum in 96-well plates. HL-60 and HL-60/TER199 cells were exposed to drug at varying drug concentrations for 72 h. After incubation with drug, 20 µl of 2 mg/ml 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt) and 45.9 µg/ml phenazine methosulfate) were added to each well and incubated at 37°C for ~40 min. Production of formazan was detected at 490 nm on a 7520 microplate reader (Cambridge Technology, Inc., Cambridge, MA). NIH3T3/MRP1 and NIH3T3/MSV cells were coincubated with drug and 50 µM TER199 for 24 h. The media were removed, and the cells were allowed to grow in normal growth media for 24 h before the addition of 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium and phenazine methosulfate.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Maintenance of Intracellular Daunorubicin Concentrations in TER199-Treated Cells. CCD analysis was used to examine the level and overall distribution pattern of daunorubicin in cells that were coincubated in the presence or absence of TER199. Cells were exposed to saturating levels of daunorubicin for 2 h before replacement of the media with drug-free media containing TER199 or solvent. Images taken immediately, or at 30 min and 2 h after removal of daunorubicin, show that the drug is gradually cleared from vehicle-treated cells (Fig. 2, D-F), but remains higher in the TER199-treated cells (Fig. 2, A-C). For example, NIH3T3/MSV cells treated with vehicle (Fig. 2, D-F) show a net decrease in fluorescence intensity from 815 to 488 (gray level scale, arbitrary units) or 40%. NIH3T3/MSV cells treated with TER199 (Fig. 2, A-C) decrease from 965 to 818, a net change of 15%. NIH3T3/MRP1 cells treated with vehicle (Fig. 2, J-L) decrease from1088 to 665, a net change of 39%. NIH3T3/MRP1 cells treated with TER199 (Fig. 2, G-I) decrease from 533 to 441, a net change of 17%. Analysis of a large number of cells confirmed these differences. At the 30-min and 2-h time points, the intracellular distribution pattern of daunorubicin in solvent-treated cells is distinct from that of TER199-treated cells. In addition to a decrease in drug fluorescence in the solvent-treated cells, the drug is packaged into vesicles, whereas in the TER199-treated cells, there remains a more dispersed fluorescence. Immunofluorescent staining with confocal imaging reveals that MRP1 is localized to vesicle structures in NIH3T3/MRP1 cells and is present, albeit at lower levels, in NIH3T3/MSV cells (Fig. 3).

View larger version (130K): [in this window] [in a new window]   Fig. 2.   Cooled charge-coupled device analysis of daunorubicin fluorescence in NIH3T3/MSV and NIH3T3/MRP1 cells. Cells were treated with saturating doses of daunorubicin before drug removal and replacement with DMEM containing vehicle or TER199. NIH3T3/MSV cells (A-C), NIH3T3/MRP1 cells (G-I), 50 µM TER199 + NIH3T3/MSV cells (D-F), TER199 + NIH3T3/MRP1 cells (J-L). Images were captured immediately following daunorubicin removal (A, D, G, J) or at time periods of 30 (B, E, H, K), or 120 min after daunorubicin removal (C, F, I, L). The cells were maintained on an incubated stage at 37°C during image acquisition. All images were 60× magnification.

View larger version (59K): [in this window] [in a new window]   Fig. 3.   Confocal images (60×) showing immunofluorescence of MRP in NIH3T3/MSV cells (A) and NIH3T3/MRP1 cells (B). A monoclonal anti-MRP antibody was obtained from Kamiya Biomedical Company; secondary antibody was antimouse tagged with biotin; tertiary antibody, antibiotin tagged with fluorescein isothiocyanate.

TER199 Enhances Daunorubicin Accumulation in MRP1-Transfected Cells. Kinetic studies of [³H]daunorubicin accumulation in NIH3T3/MRP1 were done in both the presence and absence of 50 µM TER199. As seen in Fig. 4, the rate of drug accumulation was significantly higher in TER199-treated cells compared with vehicle alone. The levels of daunorubicin accumulation are similar to NIH3T3/MSV, which do not overexpress MRP1.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Accumulation of [3H]daunorubicin in NIH3T3/MSV and NIH3T3/MRP1 cells. Cells were treated with [3H]daunorubicin (1 µM) for the indicated time points in the presence () or absence of 50 µM TER199 (, ) before lysis and scintillation counting to determine intracellular levels of [3H]daunorubicin. Sham-transfected NIH3T3/MSV cells are indicated by  and MRP1-transfected NIH3T3/MRP1 cells by  and . Data are represented as means ± S.D. of three independent experiments.

TER117 Inhibits ATP-Dependent Transport of GS-DNP in MRP1-Containing Membrane Vesicles. To analyze whether the interaction of TER117 with MRP1 is as a transported substrate or as an allosteric effector, we initiated competition studies of an established glutathione conjugate, GS-DNP. Inside-out membrane vesicles were made from NIH3T3/MSV and NIH3T3/MRP1 cells. ATP-dependent uptake of [³H]GS-DNP was measured as a function of time (Fig. 5A). We found [³H]GS-DNP to be transported with a K_m value of 8.08 µM and a V_max value of 3.1 nmol/mg/min in these vesicles. There was little uptake into vesicles made from sham-transfected NIH3T3/MSV (data not shown). Competition studies were performed with varying concentrations of [³H]GS-DNP and 1 mM. Figure 5B shows Lineweaver-Burk analysis that the interaction of TER117 with GS-DNP is reversible with a K_i value of 752 µM.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   ATP-dependent transport of GS-DNP into NIH3T3/MRP1 membrane vesicles in the presence or absence of TER117. A, vesicle solutions (30 µg) were incubated with [3H]GS-DNP (50 nM) at 37°C in the presence of 4 mM ATP (, ) and an ATP-regenerating system. In control experiments, ATP was replaced by 5'-AMP (, ; top panel, bottom line of graph). ATP-dependent transport was calculated by subtraction of the 5'-AMP values from ATP values (, ). Data are represented as means ± S.D. of three independent experiments. Closed symbols represent transport in the presence of 1 mM and open in its absence. B, double reciprocal plots according to Lineweaver and Burk  is absence of inhibitor and  is presence of 1 mM.

TER199 Reversal of Drug Resistance in NIH3T3/MRP1 Cells. To assess whether the interaction of TER117 with MRP1 could translate into an apparent cellular reversal of drug resistance to anticancer agents, a series of cytotoxicity assays was performed in the presence or absence of the esterified TER199. In every experiment, NIH3T3 and NIH3T3/MRP1 cells treated with only the indicated concentration of TER199 were analyzed to be sure that there was no toxicity compared with cells maintained in normal growth media. We found that resistance to several agents in NIH3T3/MRP1 cells could be partially reverted to the NIH3T3/MSV phenotype by coincubation with TER199 (Table 1) and that resistance to vincristine was completely reversed. Resistance to doxorubicin, daunorubicin, etoposide, and mitoxantrone also was reduced by 54, 51, 77, and 26%, respectively.

Drug Resistance of HL-60/TER199 Cells. HL-60 cells were made resistant to TER199 via stepwise selection in progressively increasing doses of TER199. HL-60/TER199 cells were examined for their cross-resistance to various chemotherapeutic agents. As shown in Table 2, HL-60/TER199 cells displayed 8-fold resistance to vincristine. More moderate levels of resistance were observed for paclitaxel, mitoxantrone, doxorubicin, daunorubicin, etoposide, and 5-FU. HL-60/TER199 cells were equally sensitive as parental HL-60 cells to the alkylating agents cisplatin, melphalan, and mitomycin C. 

Induction of Genes Involved in GSH Metabolism and Transport in TER199-Resistant HL-60 Cells. Quantitative RT-PCR analysis was undertaken to examine HL-60/TER199 cells for the possible increased expression of MRP1, and of other genes involved in the metabolism and disposition of GSH. Figure 6A shows the levels of these transcripts with respect to their expression in the parental HL-60 line. Compared with parental HL-60, MRP1 transcript levels were increased by 52 ± 16-fold in HL-60/TER199. Western analysis confirms that MRP1 is overexpressed in HL-60/TER199 cells at ~10 times the level expressed in HL-60 cells (Fig. 6B). The resistant cells did not express increased levels of MDR1, cMOAT, or MRP3. However, an increase in a novel transporter ABC2 (Laing et al., 1998) was apparent (data not shown). The functional significance of this transporter is not yet known.

View larger version (109K): [in this window] [in a new window]   Fig. 6.   Altered mRNA transcript and protein levels in chronically exposed HL-60/TER199 cells. Quantitative RT-PCR (A) and Western blot analysis (B) of GSH-associated and transporter genes associated with multidrug resistance. The molecular weight of MRP1 is 190 kDa; -GCS heavy 73 kDa; -GCS light 30 kDa; and actin 42 kDa. Lanes 1 are parental HL-60 cells, and lanes 2 are HL-60/TER199 cells. The images shown are representative of three independent experiments.

View larger version (43K): [in this window] [in a new window]   Fig. 7.   GSH levels in HL-60/TER199 cells. GSH concentrations were measured according to the method of Tietze. Data are represented as means ± S.D. of three independent experiments.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

TER199 has been shown to potentiate the cytotoxicity of several anticancer agents, earlier shown to be a result of its inhibition of GSTP1-1 activity (Morgan et al., 1996). The present study provides evidence for a further mechanism of modulation. The benzyl moiety pendent to the sulfur atom of TER199 presumably causes the cell to recognize the drug as a GSH conjugate, and process it accordingly. Where TER199 serves as an antagonist for MRP1-mediated transport of anticancer drugs, it becomes a candidate for combination with those agents to enhance their overall cellular concentrations and attenuate their cytotoxic properties.

Like many of the ABC transporters, MRP1 has a promiscuous substrate specificity permitting it to efflux a broad range of drugs and their conjugates. The highest affinity substrate for MRP1 has been shown to be LTC₄, a GSH conjugate (Jedlitschky et al., 1994). Our in vitro vesicle assays yield a K_m for GS-DNP transport of 8.08 µM and a K_i for inhibition of such transport of 752 µM. The reversible nature of this inhibition suggests that TER117 is a substrate for MRP1. Although the K_i is modest in the vesicle uptake studies, the inhibitory effects of TER199 appear to be meaningful in a biological system. The biological importance is most apparent by the increased accumulation and retention of daunorubicin in TER199-treated NIH3T3/MRP1 cells (Figs. 2 and 4) and may be due to a concentration of TER117 intracellularly, a phenomenon known to occur with other drugs (Jordan et al., 1991). Immunofluorescent data confirmed the increase in MRP1 in the transfected cells and suggest a vesicular subcellular localization. This distribution is consistent with the punctate daunorubicin and MRP patterns shown in Figs. 2 and 3, suggesting that the drug may be sequestered into these vesicles as an intermediate step in removal from the cell.

Furthermore, cytotoxicity assays indicate that TER199 sensitizes multidrug-resistant NIH3T3/MRP1 cells to various agents. NIH3T3/MRP1 cells are resistant to anthracyclines, epipodophyllotoxins, and Vinca alkaloids. The most dramatic shift in resistance occurred for vincristine with drug sensitivity completely restored. Resistance to doxorubicin, daunorubicin, etoposide, and mitoxantrone was reduced considerably. Although reversal of drug resistance was not complete, a partial reversal of resistance of this order of magnitude is significant, and may ultimately result in therapeutic advantage in modulation of resistance. It is unclear why the reversal of resistance was not complete, but it is most likely a result of TER199's weak inhibition of MRP1-mediated transport of drugs, drug metabolites, or any downstream toxic species produced as a result of drug treatment.

The mechanism by which resistance to these drugs was reversed was probably a result of inhibition of MRP1 activity and not a result of GSTP1-1 inhibition. Although TER117 is a reversible inhibitor of GSTP1-1, having a K_i value of 420 nM (Flatgaard et al., 1993), none of the drugs affected in this study have been shown to form GSH conjugates even though there frequently is resistance to these drugs in cells/tumors having an increased expression of GSTP1-1 (Tew, 1994). In addition, NIH3T3/MSV and NIH3T3/MRP1 cells have very little detectable GSTP1-1 (M.L.O., G. Kruh, K.D.T.). Furthermore, vanadate trapping assays, such as those used for the elucidation of substrates for P-glycoprotein or MRP1 (Urbatsch et al., 1995a,b; Taguchi et al., 1997) confirmed that TER199 could increase 8-azido-ATP binding to MRP1 (data not shown). These results support the kinetic data for transport inhibition.

One of the more useful aspects of drug resistance models selected through chronic exposure is the high probability that cellular adaptations will accurately reflect information pertinent to the drug's mechanism of action. Thus, HL-60/TER199 cells have been maintained at various drug concentrations ranging from 50 µM TER199 to 75 µM TER199. One finding was the lower levels of total cellular GSH relative to the parental HL-60 line. This is consistent with the previous finding of lower GSH in MRP1-overexpressing cells (Cole et al., 1990). We also examined the mRNA transcript and protein expression of several GSH-associated enzyme systems, including GST, GST, GSTµ, -GCS heavy and light subunits, and the ABC transporters MRP1, MDR1, cMOAT, and MRP3. We found no change in transcript expression for any of the GSTs and no change in protein expression of GST. The transcript expression of -GCS increased by a factor of 2 for both the heavy and light subunits of the enzyme, however, the increase in protein levels was only 50%. This result was curious in light of the decreased GSH levels present in these cells. One explanation is that the higher levels of MRP1 in HL-60/TER199 cells more effectively export GSH from the cell, either in an unconjugated or conjugated form. This would serve to lower the GSH pools and maintain a GSH homeostasis distinct from the parent cell line. Furthermore, MRP1-mediated transport of vincristine and daunorubicin has been shown to be dependent on the presence of GSH (Loe et al., 1996, 1998; Renes et al., 1999). Although the precise reason for the importance of GSH has yet to be determined, coefflux of the drug and GSH may occur. Thus, additional expression of MRP1 may be a contributory factor in diminished cellular GSH levels. No cMOAT, MRP3, or MDR1 transcript expression was detected in these cells.

Consistent with the observation that HL-60/TER199 cells have increased expression of MRP1, these cells displayed a cross-resistance profile similar to that described for other MRP1-overexpressing cell lines (Cole et al., 1994; Breuninger et al., 1995). This includes resistance to vincristine, daunorubicin, doxorubicin, and etoposide. It is notable that HL-60/TER199 cells also displayed resistance to mitoxantrone and 5-FU, two agents which are not typically included in the standard MRP1 phenotype. With MRP1-transfected cells, Breuninger et al. (1995) described a modest degree of resistance to mitoxantrone that may result from the planar structural similarity it has to anthracyclines. At this time, there is no evidence linking 5-FU efflux with MRP1. We are unsure whether the approximate 2-fold resistance to 5-FU is mediated through MRP1 or another as yet uncharacterized mechanism in the HL-60/TER199 cells.

In summary, the mechanism of action of the GSH peptidomimetic drug TER199 appears to be a composite of effects on at least two cellular targets. In addition to the inhibitory properties against GSTP1-1, the de-esterified molecule interferes with the transport function of MRP1. This characteristic has broadened the utility of the agent as a modulator of other cancer drugs.