Coordinated induction of MRP/GS-X pump and gamma-glutamylcysteine synthetase by heavy metals in human leukemia cells.

We recently reported that GS-X pump activity, as assessed by ATP-dependent transport of the glutathione-platinum complex and leukotriene C4, and intracellular glutathione (GSH) levels were remarkably enhanced in cis-diamminedichloroplatinum(II) (cisplatin)-resistant human leukemia HL-60 cells (Ishikawa, T., Wright, C. D., and Ishizuka, H. (1994) J. Biol. Chem. 269, 29085-29093). Now, using Northern hybridization and RNase protection assay, we provide evidence that the multidrug resistance-associated protein (MRP) gene, which encodes a human GS-X pump, is expressed at higher levels in cisplatin-resistant (HL-60/R-CP) cells than in sensitive cells, whereas amplification of the MRP gene is not detected by Southern hybridization. Culturing HL-60/R-CP cells in cisplatin-free medium resulted in reduced MRP mRNA levels, but these levels could be induced to rise within 30 h by cisplatin and heavy metals such as arsenite, cadmium, and zinc. The increased levels of MRP mRNA were closely related with enhanced activities of ATP-dependent transport of leukotriene C4 (LTC4) in plasma membrane vesicles. The glutathione-platinum (GS-Pt) complex, but not cisplatin, inhibited ATP-dependent LTC4 transport, suggesting that the MRP/GS-X pump transports both LTC4 and the GS-Pt complex. Expression of gamma-glutamylcysteine synthetase in the cisplatin-resistant cells was also co-induced within 24 h in response to cisplatin exposure, resulting in a significant increase in cellular GSH level. The resistant cells exposed to cisplatin were cross-resistant to melphalan, chlorambucil, arsenite, and cadmium. These observations suggest that elevated expression of the MRP/GS-X pump and increased GSH biosynthesis together may be important factors in the cellular metabolism and disposition of cisplatin, alkylating agents, and heavy metals.

We recently reported that GS-X pump activity, as assessed by ATP-dependent transport of the glutathioneplatinum complex and leukotriene C 4 , and intracellular glutathione (GSH) levels were remarkably enhanced in cis-diamminedichloroplatinum(II) (cisplatin)-resistant human leukemia HL-60 cells (Ishikawa, T., Wright, C. D., and Ishizuka, H. (1994) J. Biol. Chem. 269, 29085-29093). Now, using Northern hybridization and RNase protection assay, we provide evidence that the multidrug resistance-associated protein (MRP) gene, which encodes a human GS-X pump, is expressed at higher levels in cisplatin-resistant (HL-60/R-CP) cells than in sensitive cells, whereas amplification of the MRP gene is not detected by Southern hybridization. Culturing HL-60/ R-CP cells in cisplatin-free medium resulted in reduced MRP mRNA levels, but these levels could be induced to rise within 30 h by cisplatin and heavy metals such as arsenite, cadmium, and zinc. The increased levels of MRP mRNA were closely related with enhanced activities of ATP-dependent transport of leukotriene C 4 (LTC 4 ) in plasma membrane vesicles. The glutathioneplatinum (GS-Pt) complex, but not cisplatin, inhibited ATP-dependent LTC 4 transport, suggesting that the MRP/GS-X pump transports both LTC 4 and the GS-Pt complex. Expression of ␥-glutamylcysteine synthetase in the cisplatin-resistant cells was also co-induced within 24 h in response to cisplatin exposure, resulting in a significant increase in cellular GSH level. The resistant cells exposed to cisplatin were cross-resistant to melphalan, chlorambucil, arsenite, and cadmium. These observations suggest that elevated expression of the MRP/GS-X pump and increased GSH biosynthesis together may be important factors in the cellular metabolism and disposition of cisplatin, alkylating agents, and heavy metals.
Development of drug resistance in tumor cells is a significant obstacle to long-term, sustained patient response to chemother-apy. There is accumulating evidence that active export of anticancer drugs from cells is one of the major mechanisms of drug resistance. The GS-X pump 1,2 has been shown to eliminate a potentially cytotoxic glutathione-platinum (GS-Pt) complex from tumor cells, thereby modulating glutathione (GSH)associated resistance to cisplatin (1). The GS-X pump is functionally overexpressed in cisplatin-resistant human leukemia HL-60 cells (HL-60/R-CP), in which the cellular GSH level is substantially enhanced (2).
The GS-X pump is an ATP-dependent export pump for organic anions such as cysteinyl leukotrienes, glutathione disulfide (GSSG), glutathione S-conjugates, and glucuronide conjugates, and certain organic anions such as methotrexate. It plays a physiologically important role in inflammation, oxidative stress, xenobiotic metabolism, and tumor drug resistance (3)(4)(5). Although the kinetic properties and substrate specificity of this novel transporter have been studied, only recently has its molecular nature been identified. Studies by Mü ller et al. (6) and Leier et al. (7) have provided important evidence that overexpression of the multidrug resistance-associated protein (MRP) gene in human cancer cells results in increased ATPdependent GS-conjugate transport, thus demonstrating that the MRP gene product is a human GS-X pump (8). Moreover, the yeast cadmium factor (YCF1) gene from Saccharomyces cerevisiae has been identified on the basis of its ability to confer cadmium resistance (9). The YCF1 gene encodes an ATP-binding cassette protein with extensive sequence homology to human MRP (9). More importantly to our present study, however, this gene product was recently found to be a vacuolar GS-X pump in yeast cells. 3 Thus, MRP and YCF1 gene products are members of the GS-X pump family occurring in both the animal and plant kingdoms. Based on those recent findings, we have examined in this study the expression of the MRP gene in HL-60/R-CP cells as well as its potential role in cellular resistance to heavy metals.
A recent publication by de Vries et al. (10) pointed out that the overexpression of the MRP/GS-X pump per se does not necessarily result in resistance to anticancer drugs (10). If so, then the ultimate effect of the export pump would most likely require other fundamentally important factors, one being the biosynthesis of GSH. Through the conjugation reaction with cellular GSH, electrophilic organic compounds as well as heavy metals are converted to organic anions; thus, the multivalent negative charge is apparently crucial for the recognition of substrates by the MRP/GS-X pump (3,11). Cellular GSH is synthesized by ␥-glutamylcysteine synthetase (␥-GCS) and GSH synthetase. The first reaction, catalyzed by ␥-GCS, is a rate-limiting step in overall GSH biosynthesis, and the cellular GSH level is substantially regulated by ␥-GCS (12). Furthermore, correlation between cisplatin resistance and expression of ␥-GCS has been demonstrated in human ovarian cancer cells (13). In HL-60/R-CP cells, cisplatin-resistant leukemia cells established in our laboratory, the cellular GSH level was 7-to 8-fold higher than in sensitive cells. To gain more insight into the mechanism responsible for this, we examined the expression of ␥-GCS in the cisplatin-resistant and -sensitive cells. We present strong evidence that the ␥-GCS expression was remarkably induced by cisplatin within 24 h and that the cellular GSH level was concomitantly increased in the cisplatin-resistant but not the sensitive cells.
The present study demonstrates for the first time that both the MRP/GS-X pump and ␥-GCS are induced by cisplatin and heavy metals. Their coordinate induction is likely to be an important determinant for the acquired, cellular resistance to cisplatin, heavy metals, and alkylating agents.

MATERIALS AND METHODS
Biochemicals and Chemicals-GSH, GSSG, ATP, creatine phosphate, creatine kinase, phenylmethylsulfonyl fluoride, glutathione reductase, and a random-primed labeling kit were purchased from Boehringer Mannheim (Mannheim, Germany Cell Culture-Human myelocytic leukemia HL-60 cells (ATCC No. CCL240), obtained from the American Type Culture Collection (Rockville, MD), were maintained in a humidified chamber (37°C, 5% CO 2 ) in RPMI 1640 medium supplemented with glutamine (2 mM), 10% (v/v) heat-inactivated fetal calf serum, and gentamicin (50 g/ml). The cell numbers were determined by trypan blue dye exclusion and counting with a hemacytometer and were kept at 1.5 ϫ 10 5 cells/ml by passaging every 5 days. A cisplatin-resistant subline, named HL-60/R-CP, was established by maintaining HL-60 cells in the presence of cisplatin (Bristol-Myers Squibb, Seattle, WA), as described previously (2). For the induction experiments, HL-60/R-CP cells were maintained in the cisplatin-free medium for 1 month and subsequently treated with cisplatin or heavy metals for various lengths of time.
Determination of Cell Sensitivity to Anticancer Drugs and Heavy Metals-Cells (1.5 ϫ 10 5 cells/ml) were incubated in 100 l of culture medium containing melphalan, chlorambucil, cadmium chloride, or sodium arsenite at different concentrations in 96-well plates in a humidified tissue culture chamber (37°C, 5% CO 2 ). After 72 h, the number of surviving cells was determined.
Determination of GS-X Pump Activity in Plasma Membrane Vesicles-Plasma membrane vesicles were prepared from HL-60 and HL-60/R-CP cells as described previously (2) and stored at Ϫ85°C until use. Frozen stocked membrane vesicles were thawed quickly at 37°C and stored on ice until used. Of the total membrane vesicles, 42 to 46% were inside-out as assessed by sialidase accessibility assay (14). The standard incubation medium for the assay of GS-X pump activity in the preparation contained plasma membrane vesicles (50 g of protein), 10 nM [ 3 H]LTC 4 , 0.25 M sucrose, 10 mM Tris/HCl, pH 7.4, 10 mM MgCl 2 , 1 mM ATP, 10 mM creatine phosphate, and 100 g/ml creatine kinase in a final volume of 110 l. The reaction was started by adding [ 3 H]LTC 4 to the incubation medium. The reaction was carried out at 37°C, and the amount of [ 3 H]LTC 4 incorporated into the vesicles was measured by a rapid filtration technique as described previously (14).
cDNA Probes for MRP and ␥-GCS-An MRP-cDNA probe (2.86 kb) was prepared from total RNA of human mononuclear cells by RT-PCR using MRP-specific primers: forward primer (225-245), 5Ј-TGCCTTGG-GATTTTTGCTGTG; backward primer (3085-3066), 5Ј-CGATCCCTT-GTGAAATGCCC. The PCR reaction consisted of 34 cycles of 94°C for 30 s, 55°C for 60 s, and 72°C for 160 s. The resulting RT-PCR product was ligated with the pCR TM II vector (Invitrogen) and amplified using One Shot TM INVaFЈ competent cells (Invitrogen). The DNA sequence of the insert was identical with the partial sequence (225-3085) of the MRP cDNA (15). The 2.86-kb insert was excised by EcoRI digestion, purified, and used for Northern and Southern hybridizations.
A cDNA probe for the catalytic subunit of ␥-GCS was prepared from total RNA of HL-60 cells by RT-PCR using forward primer 5Ј-GCTG-CATCTCCCTTTTACCGAG and backward primer 5Ј-TGGCAACTGT-CATTAGTTCTCCAG. The 0.88-kb PCR product had a sequence identical with the partial cDNA sequence (841-1723) of ␥-GCS (16). The PCR product was ligated with the pCR TM II vector and amplified using One Shot TM INVaFЈ competent cells. The 0.88-kb insert was excised by EcoRI digestion, purified, and used for Northern hybridization.
Sequence Analysis-DNA sequences of the PCR products ligated with pCR TM II vector were determined using a DNA sequence analyzer (Applied Biosystems Model 373A). Sequencing reactions utilized Sequenase (U. S. Biochemical Corp.) and synthetic oligonucleotide primers corresponding to T7, SP6, and internal sequences of MRP or ␥-GCS.
Northern Hybridization-Total cellular RNA was prepared by acid guanidinium thiocyanate-phenol-chloroform extraction from samples of 1 ϫ 10 7 cells as described by Chomczynski and Sacchi (17). The RNA (10 g/lane as determined by absorbance at 260 nm) was fractionated by electrophoresis in 1.0% (w/v) agarose gels containing formaldehyde (18) and transferred to Nytran membranes (Schleicher & Schuell). The membranes were then baked at 80°C for 2 h. 32 P-Labeled DNA probes were prepared by a random-primer labeling method. Hybridization with the DNA probe (1 ϫ 10 6 cpm/ml) was performed at 42°C for 24 h in a mixture containing 5 ϫ SSPE (750 mM sodium chloride, 5 mM EDTA, and 50 mM sodium phosphate, pH 7.4), 50% formamide, 5 ϫ Denhardt's solution, 0.1% SDS, 10% dextran sulfate, and 200 g/ml denatured salmon sperm DNA. After hybridization, the membrane was washed in 0.1% SDS-2 ϫ SSC (300 mM sodium chloride and 30 mM sodium citrate, pH 7.0) at room temperature for 20 min and subsequently in 0.1% SDS-0.1 ϫ SSC (15 mM sodium chloride and 1.5 mM sodium citrate, pH 7.0) at 55°C for 30 min. The membranes were then exposed to Kodak X-Omat AR films at Ϫ85°C using intensifying screens.
Southern Hybridization-Ten g each of genomic DNA from HL-60 and HL-60/R-CP cells were digested with restriction endonuclease EcoRI, separated by 1% agarose gel electrophoresis, and transferred to a nitrocellulose membrane. Hybridization was carried out according to procedures described previously (18) using the 32 P-labeled 2.86-kb MRP cDNA as a probe. Hybridization signals were detected by a Phosphor-Imager (Molecular Dynamics, Sunnyvale, CA).
RNase Protection Assay-A 293-nucletotide fragment of MRP cDNA spanning nucleotides 44 to 336 from the translation start site was synthesized by RT-PCR using the following primers: forward primer 5Ј-GGGAATTCTGGGACTGGAATGTCACG (EcoRI site is underlined) and backward primer 5Ј-CGGGATCCAGGAATATGCCCCGACTTC (BamHI site). The PCR product was digested with EcoRI and BamHI and cloned into EcoRI/BamHI sites of the pCR TM II vector. The resultant plasmid DNA was linearized with XbaI, and the antisense RNA probe was synthesized using SP6 RNA polymerase. To prepare the probe for detecting ␥-GCS mRNA level, pCR 0.88 ␥-GCS cDNA was linearized with PstI, and the antisense probe was synthesized using SP6 RNA polymerase. Either 20 g (for MRP and ␥-GCS probes) or 1 g (for 18 S rRNA probe) of total RNA from HL-60/R-CP cells was hybridized with 32 P-labeled antisense RNA probes (2 ϫ 10 5 cpm) and subjected to the RNase protection assay as described previously (19).
Determination of Cellular Level of Total Glutathione (GSH ϩ GSSG)-A cell suspension (a total of 1 ϫ 10 7 cells) was withdrawn from the cell culture and centrifuged at 40 ϫ g for 5 min at 4°C. The precipitated cells were resuspended in 10 ml of ice-cold phosphatebuffered saline (0.9% (w/v) NaCl containing 10 mM potassium phosphate, pH 7.4) and again centrifuged at 400 ϫ g for 5 min. The resulting cell pellet was resuspended in 750 l of phosphate-buffered saline. From the cell suspension, a 500-l aliquot was taken, mixed with 300 l of 20% perchloric acid, and homogenized at 4°C with an ultrasonicator. After centrifugation at 16,000 ϫ g for 5 min, a 200-l aliquot of the resulting supernatant was withdrawn and neutralized by K 2 HCO 3 . The concentration of total glutathione (GSH ϩ GSSG) in the neutralized sample was determined according to the method of Tietze (20) with a modification described in Ref. 21.

Expression of the MRP/GS-X Pump in HL-60/R-CP Cells-
We previously reported that the GS-X pump is responsible for transporting the GS-Pt complex in human HL-60 cells and that GS-X pump activity is significantly enhanced (4-to 5-fold) in cisplatin-resistant cells (2). Our recent RT-PCR study has shown that the MRP gene encoding a human GS-X pump is expressed at high levels in HL-60/R-CP cells (5). To confirm the finding, we performed Northern hybridization and RNase pro-tection assays using an MRP-specific probe (provided by Drs. Cole and Deeley; see Ref. 15) that encodes the 3Ј region (1-kb) sequence of MRP cDNA. MRP mRNA of about 8.0 kb was observed in both HL-60/R-CP and HL-60 cells; however, the level was remarkably (about 5-fold) higher in the resistant cells than in the sensitive cells (Fig. 1A). RNase protection assay also showed increased expression of MRP in HL-60/R-CP cells (data not shown). On the other hand, the Southern hybridization of the EcoRI-digested genomic DNA from HL-60 and HL-60/R-CP cells with the MRP-specific probe (2.86-kb) (see "Materials and Methods") exhibited four hybridized bands at 3.7, 4.1, 9.0, and 13.0 kb, and no amplification of the MRP gene was observed in HL-60/R-CP cells (Fig. 1B). These data suggest transcriptional up-regulation of the MRP gene and/or enhanced stability of its transcript in the cisplatin-resistant cells.
Induction of MRP/GS-X Pump by Cisplatin and Heavy Metals-Expression of the MRP gene in HL-60/R-CP cells was reversible and dependent on cisplatin. MRP mRNA level remarkably decreased when the cells were maintained without cisplatin for 1 month. GS-X pump activity, assessed by ATPdependent transport of LTC 4 in plasma membrane vesicles, decreased from 2.51 Ϯ 0.21 pmol/mg of protein/10 min to 0.30 Ϯ 0.04 pmol/mg of protein/10 min after the cisplatin-free incubation. On the other hand, when the cells were re-exposed to 20 M cisplatin, the level of MRP mRNA (8.0 kb in size) increased significantly (7-fold) (Fig. 1A). Fig. 2B shows the time course of MRP mRNA levels during cisplatin incubation. Interestingly, MRP was also induced by arsenite, cadmium, and zinc, when the cells were incubated with these metals at noncytotoxic concentrations for 48 h. Fig. 3 demonstrates the increased levels of MRP mRNA detected by Northern hybridization (A) and by RNase protection assay (B). increased from 0.30 Ϯ 0.04 pmol/mg of protein/10 min to 1.95 Ϯ 0.11 pmol/mg of protein/10 min (n ϭ 3), which is consistent with the elevated levels of MRP mRNA (Fig. 5A).
It would be of importance to examine whether the zincinduced MRP/GS-X pump had an affinity for the GS-Pt complex. We previously reported that transport of the GS-Pt complex is an ATP-dependent process and inhibited by LTC 4 , GSSG, and S- (2,4-dinitrophenyl)glutathione, suggesting that the GS-X pump exports the GS-Pt complex from cells (2). Fig.  5C demonstrates that LTC 4 transport was dose dependently inhibited by the GS-Pt complex but not by cisplatin. The GS-Pt complex inhibited GS-X pump activity 50% at a concentration of about 150 M, which is similar to the K m value (130 M) for the complex (2). Thus, these results strongly suggest that the heavy metal-inducible MRP/GS-X pump transports both LTC 4 and the GS-Pt complex.
Induction of ␥-GCS by Cisplatin in HL-60/R-CP Cells-Biosynthesis of cellular GSH is a key factor in the function of the MRP/GS-X pump as well as in the overall metabolism and disposition of cisplatin in cancer cells. In HL-60/R-CP cells maintained with cisplatin, the intracellular GSH level was 7.52 Ϯ 0.74 nmol/10 6 cells, about 7-fold higher than in HL-60 cells (1.15 Ϯ 0.12 nmol/10 6 cells) (2). After the incubation medium was replaced with cisplatin-free medium, the intracellular GSH level in the resistant cells decreased to 2.01 nmol/10 6 cells, with a half-time of 5 days. This suggests a reversible regulation of intracellular GSH levels.
As Fig. 5 illustrates, ␥-GCS, which is a rate-limiting enzyme of cellular GSH biosynthesis, was induced by cisplatin in HL-60/R-CP cells. Prior to this experiment, cells had been maintained in the cisplatin-free medium for 1 month. The cisplatin- were withdrawn from the incubation medium, and the levels of 4.0-kb ␥-GCS mRNA were determined. Prior to the experiment, HL-60/R-CP cells had been maintained in cisplatin-free medium for 1 month, as described under "Materials and Methods." The relative level of ␥-GCS mRNA was determined using a scanning densitometer and normalized to the level of HL-60 cells at 0 h. resistant and -sensitive cells were then exposed to 20 M cisplatin for 24 h (Fig. 5A). It is important to note that the ␥-GCS mRNA level in HL-60/R-CP cells significantly increased during the incubation with cisplatin (Fig. 5). By contrast, ␥-GCS mRNA levels in HL-60 cells remained almost constant throughout the incubation (Fig. 5A). No amplification of the ␥-GCS gene was detected in both HL-60 and HL-60/R-CP cells (Fig. 1B). Fig. 6A shows time courses of intracellular GSH levels in the cisplatin-resistant and -sensitive cells during the incubation with cisplatin. Intracellular GSH level in HL-60/R-CP cells increased up to 5.8 nmol/10 6 cells, corresponding to about 80% of the GSH level in cells continuously exposed to cisplatin. On the other hand, the cellular GSH level in the sensitive cells was virtually unchanged during the first 18-h incubation, and it began to decrease after 24 h owing to cell damage. The increase in cellular GSH level in the cisplatin-resistant cells is intimately linked with the increase of ␥-GCS mRNA level (Fig. 5). As shown in Fig. 6B, the increase of intracellular GSH level in HL-60/R-CP cells after a 12-h exposure to cisplatin was directly proportional to cisplatin concentration up to 40 M. Table I demonstrates that cisplatin and heavy metals, i.e. arsenite, cadmium, and zinc, significantly enhanced cellular GSH levels and GS-X pump activity in HL-60/R-CP cells. The results also suggest that induction of both ␥-GCS and MRP/GS-X pump occurs in concert in the cisplatinresistant cells.

Effect of Cisplatin and Heavy Metals on Cellular GSH and GS-X Pump Activity-
Cross-resistance of HL-60/R-CP Cells to Melphalan, Chlorambucil, Cadmium, and Arsenite- Fig. 7 shows the sensitivity of HL-60 and HL-60/R-CP cells to melphalan, chloram-bucil, cadmium, and arsenite and clearly demonstrates the cross-resistance of HL-60/R-CP cells to these alkylating agents and heavy metals. As assessed on the basis of IC 50 value, the extents of resistance of HL-60/R-CP cells were 3.2-, 2.6-, 2.9-, and 3.0-fold for melphalan, chlorambucil, cadmium, and arsenite, respectively. As previously reported (2), HL-60/R-CP cells were about 10-fold more resistant to cisplatin than HL-60 cells. It is remarkable that cisplatin-resistant HL-60/R-CP cells grew normally in the presence of cadmium at concentrations of up to 100 M, whereas the growth and viability of cisplatin-sensitive HL-60 cells were greatly affected in that same concentration range (Fig. 7).

Induction of the MRP/GS-X Pump by Cisplatin and Heavy
Metals-We have previously established HL-60/R-CP cells and demonstrated an elevated GS-X pump activity in this cisplatinresistant cell line. We also demonstrated that the ATP-dependent transport of the GS-Pt complex and LTC 4 measured with plasma membrane vesicles was about 4-to 5-fold greater in HL-60/R-CP cells than in HL-60 cells. The K m value for the GS-Pt complex was estimated to be 130 M (2), similar to that for GSSG, one of the endogenous substrates for the GS-X pump (21). Since our studies were published, several laboratories have reported that the human GS-X pump is encoded by the MRP gene (6 -8). The present study demonstrates that the human MRP gene is expressed at higher levels in HL-60/R-CP cells. Furthermore, we showed that LTC 4 and the GS-Pt complex mutually inhibit their ATP-dependent transport in plasma membrane vesicles prepared from HL-60/R-CP cells (Fig. 4C  and Ref. 2). These results strongly suggest that MRP and GS-X pump have similar, if not identical, substrate specificity. Recent studies by Fujii et al. (22) and Goto et al. (23) also showed that GS-X pump activity was significantly enhanced in cisplatin-resistant human epidermoid carcinoma KB, colonic cancer HCT8, and ovarian cancer A2780 cells. Collectively, these results are consistent with the idea that the overexpressed GS-X pump is responsible for eliminating the GS-Pt complex as well as GS-conjugates from these cisplatin-resistant cells.
Our present results also demonstrate that HL-60/R-CP cells are cross-resistant to heavy metals (cadmium and arsenite) in addition to alkylating agents, melphane and chlorambucil (Fig.  7). Like cisplatin, these heavy metals react with GSH to form GSH chelate complexes. Thus, it is reasonable to speculate that the overexpressed MRP/GS-X pump in these cells is responsible for detoxifying these heavy metals. Several recent findings support this contention. (i) The primary structure of the mammalian MRP gene product deduced from its cDNA sequence is remarkably similar to that of YCF1, which is responsible for cadmium resistance in S. cerevisiae (9). Deletion of the YCF1 gene results in defective transport of the glutathione-bimane conjugate into vacuoles of yeast cells, 3 suggesting that the YCF1 gene encodes a vacuolar GS-X pump in yeast cells and that its MRP homolog functions as a transporter for GSHmetal complexes as well. (ii) The primary sequences of MRP and YCF1 genes are both remarkably similar to that of the ltpgpA gene (24), which is a putative pump for As 3ϩ -GSH complex (25) and confers arsenite resistance in Leishmania (26,27). (iii) Cole et al. (28) have suggested a potential role for the MRP gene in cellular resistance to some heavy metal anions, including arsenite and antimony. Furthermore, Zaman et al. (29) have recently reported that arsenite efflux from MRPtransfected cells was accompanied by a significant increase of GSH efflux, suggesting that arsenite may be transported as a GSH-chelate complex by the same export pump. These results, taken together, strongly suggest that the mechanisms of eliminating cytotoxic heavy metals via GSH conjugation are conserved throughout evolution.
We would like to stress that overexpression of MRP/GS-X is not the only mechanism that is involved in cisplatin resistance in cancer cells. A number of cisplatin-resistant cell lines have failed to display increased MRP/GS-X pump (30). 4 Other mechanisms, such as removal of platinum-DNA adducts, may be involved in these non-MRP/GS-X pump-overexpressing cells (Ref. 31 and references therein). Likewise, overexpression of the MRP/GS-X pump does not necessarily have to result in cisplatin resistance (28,32,33). Since formation of the GS-Pt complex is a prerequisite for the function of GS-X pump, one can envision that intracellular levels of GSH would also play an important role in the overall drug resistance mechanism (see below).
Induction of ␥-GCS by Cisplatin-Glutathionation allows substrate recognition by the MRP/GS-X pump. While GSH transferases catalyze the conjugation reactions with electrophilic organic compounds, GSH spontaneously reacts with heavy metals to form GSH-metal chelate complexes (1,34,35). Increased cellular GSH propels the reaction of cytotoxic heavy metals with GSH to generate GSH-metal complexes, thereby providing an important defense mechanism. Previous study demonstrated that cisplatin resistance of cancer cells was often observed with increased levels of intracellular GSH and high levels of ␥-GCS expression (13). Transcriptional up-regulation of ␥-GCS gene expression was also reported for melphalanresistant human prostate carcinoma cells (36). The present study demonstrates that ␥-GCS was induced by cisplatin in HL-60/R-CP cells and that cellular GSH levels were concomitantly increased in the cisplatin-resistant cells (Figs. 5 and 6). These results strongly suggest that replenishment of intracellular GSH is required for the elevated GS-X pump activity in these drug-resistant variants. Furthermore, our results demonstrate that the cellular GSH status is dynamic and can be modulated by extracellular stimuli. Rapid up-regulation of ␥-GCS expression must therefore be a critical determinant for cellular tolerance to heavy metals and electrophilic compounds.
␥-GCS exists as a holoenzyme in vivo, consisting of dissociable heavy and light subunits (37,38). The heavy subunit expresses the catalytic function whereas the light subunit, regulatory function. Recent transfection experiments using expression vectors containing cDNA encoding these subunits demonstrated that, while co-transfecting both heavy and light subunits yielded high levels of intracellular GSH, transfecting either subunit alone could produce moderate increases of GSH (39). While the levels of the regulatory subunit expression in HL-60/R-CP cells remain to be determined, our present results apparently are consistent with the notion that overexpression of the heavy subunit is sufficient to increase GSH levels.
Our HL-60/R-CP cells exhibit cross-resistance to other heavy metals, e.g. cadmium, arsenite (Fig. 4). A recent report also showed that a cisplatin-resistant human ovarian carcinoma cell line exhibited cross-resistance to antimony and cadmium (40). However, these resistance profiles are slightly different from those reported for the MRP-transfected cells (28,29). These reports demonstrated that cells transfected with the MRP gene displayed resistance to doxorubicin, vincristine, arsenite, and antimony, but not to cisplatin and cadmium, whereas our HL-60/R-CP cells show no cross-resistance to doxorubicin and vincristine (2). These results, taken together, suggest that overexpression of the MRP/GS-X pump per se is necessary but not sufficient to generate a full spectrum of drug resistance. Multiple biochemical events such as reduction, deglycosylation, and conjugation with GSH are considered to be involved in the cellular metabolism of anthracyclines and transport of their metabolites via the MRP/GS-X pump (41). In this context, increased GSH biosynthesis may also be essential for acquired resistance to a wide spectrum of anticancer drugs and heavy metals.
Mechanisms of Induction of the MRP/GS-X Pump and ␥-GCS Gene Expression by Heavy Metals-One of the major findings presented in this communication is that genes encoding the MRP/GS-X pump and ␥-GCS can be induced by heavy metals. Coordinated regulation of these two genes have been found in ␥-GCS-transfected human small cell lung cancer cells (42), as well as in other drug-resistant variants and in tumor biopsies. 5 These results suggest that certain common factor(s) may be involved in the expression of these two genes. Although the mechanisms underlying the concerted regulation of these two genes are not known at present, possible mechanisms can be speculated. In yeast, the GSH1 gene encoding ␥-GCS and the YCF1 gene encoding the MRP homolog are coordinately regulated by yAP-1, which encodes yeast transcription factor AP-1 (43). Transcriptional activation of these genes mediated by yAP-l is essential for cadmium tolerance in yeast cells (44). Induction of c-myc and c-jun expression in rat L6 myoblasts by cadmium has been reported (45), and we have found that levels of c-jun transcript are increased in HL-60/R-CP cells (data not shown). Whether ␥-GCS and MRP/GS-X pump genes in mammalian cells are also transcriptionally regulated by AP-1 remains to be determined. The promoter regions of human ␥-GCS and MRP genes have been characterized recently, and it is noteworthy that cis-regulatory elements including AP-1 binding sites have been identified for both of the genes (46 -48).
Several recent reports have demonstrated transient inductions of drug resistance gene expression by cytotoxic compounds. The human MDR1 gene encoding P-glycoprotein has been shown to be regulated by heat shock, arsenite, and cadmium (49) as well as other cytotoxic compounds (50). Furthermore, following exposure to several cytotoxic compounds, most of which are known to be substrates for the multidrug transporter, mdr RNA levels in cultured rodent cells were found to increase (51). Both transcriptional and post-transcriptional mechanisms are apparently involved in this regulation (51). The present study showing that ␥-GCS and MRP/GS-X pump can be coordinately induced by anticancer drug cisplatin and heavy metals provide important information to the steadily accumulating evidence that various drug resistance genes can be acutely induced upon drug treatments. These observations may have important clinical implications. Unlike cell culture studies where drug-resistant variants are usually obtained through long-term, continuous drug exposure, such transient induction of drug resistance gene expression is considered to be more relevant to the treatment protocols for cancer patients. Thus, understanding the mechanisms involved in transcriptional and/or post-transcriptional regulation of the expression of MRP/GS-X pump and ␥-GCS genes may facilitate the development of novel approaches to control drug resistance in cancer chemotherapy. These experiments are currently under way in our laboratories.