The role of glutathione S-transferase P1-1 (GSTP1-1) in the cellular detoxification of cisplatin

Abstract

Cells expressing elevated levels of allelic variants of human GSTP1 and/or efflux transporters, MRP 1 or 2, were used to evaluate the role of GSTP1-1 in cisplatin resistance. These studies revealed that GSTP1-1 confers low-level resistance (1.4–1.7-fold) to cisplatin-induced cytotoxicity in MCF7 cells. However, expression of MRP1 (MCF7 cells) or MRP2 (HepG2 cells) failed to augment or potentiate GSTP1-1-mediated resistance in either cell line. To understand the mechanism by which variants of GSTP1-1 confer resistance to cisplatin, their relative abilities to catalyze conjugation of cisplatin with glutathione were examined. Enzymes encoded by all three alleles tested GSTP1a (I₁₀₄A₁₁₃), GSTP1b (V₁₀₄A₁₁₃) and GSTP1c (V₁₀₄V₁₁₃) —increased the formation rate of the mono-platinum-glutathione derivative of cisplatin with relative catalytic activities of 1.0 (GSTP1a-1a variant) and 1.8–1.9 (GSTP1b-1b and GSTP1c-1c variants). While these data are consistent with the idea that very low-level resistance to cisplatin may be conferred by GSTP1-1-mediated cisplatin/glutathione conjugation, two observations indicate that such catalysis plays a minor role in the protection from cisplatin toxicity. First, the rates of GSTP1-1-mediated conjugation are extremely slow (1.7–2.6 hr⁻¹ at 25⁰C). Second, despite an 80–90 % increase in catalysis of cisplatin conjugation by GSTP1b-1b or GSTP1c-1c over GSTP1a-1a, we observed no discernable differences in relative resistances conferred by these alternative variants when expressed in MCF7 cells. We conclude that high-level cisplatin resistance attributed to GSTP1-1 in other studies is not likely due to catalysis of cisplatin conjugation but rather must be explained by other mechanisms — mechanisms which may include GSTP1-mediated modulation of signaling pathways.

INTRODUCTION

Cisplatin (cis-diamminedichloroplatinum (II)) is one of a number of platinum (Pt)-containing compounds used for the treatment of a variety of malignancies including ovarian, testicular and lung cancers (1). It is generally accepted that cisplatin initiates cell death primarily through the formation of intrastrand crosslinks, the majority of which involve neighboring purine bases (2). Accumulation of these adducts can inhibit DNA replication and transcription, triggering cell cycle arrest and apoptosis (3–5). A major limitation of cisplatin treatment is the acquisition of drug resistance. Several molecular and biochemical mechanisms have been implicated in this resistance including decreased cellular uptake, enhanced cellular inactivation, increased tolerance to DNA damage, and increased cellular efflux (6). The association between elevated cellular levels of glutathione (GSH) and cisplatin resistance has lead to the suggestion that inactivation of cisplatin can occur through conjugation with GSH (7, 8). Indeed cisplatin has been shown to complex non-enzymatically with GSH (9–12) and some of the resulting platinum-GSH conjugates (Pt-SG’s) have been characterized (9, 11, 12).

Elevated glutathione S-transferase (GST) P1-1 expression has been associated with resistance to cisplatin-based chemotherapy and with cellular resistance to cisplatin cytotoxicity in many experimental systems (13–17). The finding that GSTP1-1 can enhance the rate of Pt-SG formation suggests a potential catalytic mechanism for GSTP1-1-associated cisplatin resistance (18). In addition to the most common allelic variant of human GSTP1, GSTP1a (I₁₀₄A₁₁₃), at least three other allelic variants have been identified including GSTP1b (V₁₀₄A₁₁₃) and GSTP1c (V₁₀₄V₁₁₃) (19, 20). The particular alternative allelic form(s) expressed is(are) potentially important as the variants are reported to differ in their catalytic efficiencies towards the conjugation with GSH of a number of environmental toxins and anticancer drugs — catalytic differences that lead to apparent variations in protection afforded by the GSTP1-1 variants against the toxicities of several compounds including the diolepoxides of polycyclic aromatic hydrocarbons and anticancer agents such as thiotepa and 4-hydroxyifosphamide (14, 20–23). Moreover, studies evaluating the distribution of the GSTP1 polymorphisms in the population have shown a relationship between expression of specific GSTP1-1 variants and susceptibility to a variety of cancers (24, 25). Lastly, expression of GSTP1 allelic variants in bacterial cells yielded different levels of protection from cisplatin toxicities (14).

The primary goals of the studies described herein were to determine if forced expression of GSTP1-1 could confer resistance to cisplatin in model MCF7 cells and, if so, whether the mechanism of GSTP1-1-mediated resistance involves catalysis of cisplatin/GSH conjugation. Additionally, three variants of GSTP1-1 were examined to determine whether they differed in their relative catalysis of Pt-SG formation and if expression of these alternative allelic variants would be associated with differential protection from cisplatin cytotoxicity.

However, as suggested by Ishikawa (26), complete detoxification may not be attained by GSH conjugate (GS-X) formation alone; rather, this may require active efflux of the GS-X formed intracellularly. Indeed our laboratory has shown that removal of GS-X’s are often required to fully achieve GST-associated resistance (27–29). Accordingly, in the studies presented we examined whether co-expression of GS-X efflux pumps, MRP1 (MCF7 cells) or MRP2 (HepG2 cells), would augment or potentiate GSTP1-1-mediated resistance to cisplatin cytotoxicity.

The results outlined below indicate that while variants of GSTP1-1 can indeed confer low-level resistance to cisplatin in some cell lines, the rates of GSTP1-1-mediated catalysis are extremely slow suggesting that alternative mechanisms are needed to explain GSTP1-1-associated resistance to cisplatin.

MATERIALS AND METHODS

Cell lines and tissue culture

Parental MCF7 (MCF7/WT) and HepG2 cell lines and their transgenic derivatives were cultured in Dulbecco’s modified Eagle medium (DMEM, MCF7 cells) or DMEM/F12 medium (HepG2 cells) containing 10% fetal bovine serum. MCF7/WT cells possess extremely low endogenous GST activity and MRP1 expression (27). Transgenic cell lines engineered to express human GSTP1a-1a (MCF7/P1a-7) and/or MRP1 (MCF7/MRP1-10 and MRP1-10/P1a-4) have been developed, characterized, and cultured as previously described (30, 31). HepG2 parental cells expressing MRP2 were stably transfected with the doxycycline-repressible VP16 transactivator fusion gene (HepG2/tTA) and a doxycycline-repressible GSTP1a-1a expression vector (HepG2/π3) as described (29). In addition, cDNA’s encoding the human alleleic variants, GSTP1b and GST1c (kindly supplied by Dr. Francis Ali-Osman, Duke University), were subcloned into pLHCX. Packaging cells (PA317) were transfected with the pLHCX/GSTP1b and pLNCX/GSTP1c vectors and the viral particle containing supernatants were used to stably transduce parental MCF7 cells as described previously (30, 31). Clones stably expressing allelic variants of GSTP1 were selected in 0.4 mg/ml hygromycin. Transgenic cells were maintained in selecting drugs — 0.2 mg/ml hygromycin (MCF7/P1a-7, MCF7/P1b-9 & -18, and MCF7/P1c-4 & -27), 1 mg/ml G418 (MCF7/WT-LN, MCF7/MRP1-10, and HepG2/tTA), or both drugs (MRP1-10/P1a-4 and HepG2/π3) until just prior to experiments at which time selecting drugs were removed from the medium. For some experiments, GSTP1a-1a expression in HepG2/π3 was repressed by culturing in 1 μg/ml doxycycline.

Cytotoxicity determinations

Cytotoxicity was determined using the sulforhodamine B microtiter plate assay as described (32). Cells were plated at a density of 300 cells/well (MCF7 derivatives) or 1200 cells/well (HepG2 derivatives) in 96-well plates. Twenty-four hours later, the cells were exposed to varying concentrations of cisplatin, doxorubicin (Sigma-Aldrich, St. Louis, MO), or vehicle control for 1 h or 24 h in medium containing 10% fetal bovine serum. The medium was subsequently replaced and the plates were incubated 5–7 days, then fixed and stained. In some experiments, MRP2 was inhibited with 20 or 50 μM MK571 (Alexis, Carlsbad, CA) which was added to the medium 15 min before introduction of cisplatin or doxorubicin and continued throughout the drug exposure. Cytotoxicities are expressed as IC₅₀’s or as relative resistances (IC₅₀ test cell line IC₅₀ control cell line).

HPLC/MS/MS analysis of platinum-GSH (Pt-SG) conjugates

The characterization of platinum/GSH conjugates (Pt-SG’s) was accomplished as follows. Pt-SG’s were formed by incubating equimolar (3 mM) cisplatin and GSH in 10 mM sodium phosphate buffer, pH 7.5, containing 150 mM NaCl for 30 min, at 25 C. The reaction mixture (25 μl) was separated, prior to entering the electrospray source, by HPLC using a Beckman Ultrasphere C₁₈ reverse phase column (25 cm × 4.6 mm, 5 mm) (BeckmanCoulter) and isocratic elution in 15 mM formic acid at 0.7 ml/min. Approximately 95% was sent to a diode array detector and 5% was diverted to a Micromass Quattro II mass spectrometer (Waters, Milford, MA) equipped with a z-spray and triple quadrupole analyzer. Peaks corresponding to cisplatin, mono-platinum-mono-GSH conjugate (Pt₁-SG, m/z 572), and bis-platinum-mono-GSH conjugate (Pt₂-SG, m/z 835) eluted from the HPLC column at 3.6 min, 4.7 min, and 4.2 min, respectively. MS/MS analysis was achieved in positive ion mode with the following instrument settings: capillary voltage, 3.50 kV; cone voltage, 35 V; and collision energy 15 eV.

Recombinant GSTP1-1 variants

Expression and purification of recombinant GSTP1a-1a has been described previously (30). The coding regions of GSTP1b and GSTP1c were amplified by PCR and inserted into pOXO4. Expression and purification of the GSTP1-1 variants were achieved as follows. Briefly, BL21 Star competent E. coli (Invitrogen, Carlsbad, CA) were transformed with the appropriate plasmid vector and expression was induced using isopropyl-β-D-thiogalactopyranoside. Bacteria were centrifuged, washed with PBS, resuspended in lysis buffer (50 mM Tris (pH 7.5) + 50 mM EDTA) and sonicated. GSTP1-1 were purified using GSH-agarose affinity chromatography and final preparations were subjected to high-performance centrifugal concentration (Orbital Biosciences, Topsfield, MA). The purities of GSTP1-1 variants were confirmed by SDS PAGE. Protein concentrations and activities for the GST preparations were determined as described (33, 34). Specific activities were: GSTP1a-1a, 91 U/mg; GSTP1b-1b, 42 U/mg; and GSTP1c-1c, 38 U/mg where U is defined as μmol 1-chloro-2,4-dinitrobenzene (CDNB) conjugated with GSH per minute at 25°C (34).

Analysis of non-enzymatic and GSTP1-1-catalyzed Pt-SG formation

Analysis of Pt-SG formation was accomplished as follows. Reaction mixtures contained 1 mM CDDP, 2 mM GSH in the presence or absence of 33 U/ml of recombinant GSTP1a-1a, GSTP1b-1b, or GSTP1c-1c in 10 mM sodium phosphate, pH 6.8. Mixtures were incubated for 0, 2, or 60 min, at 25 C, after which 30 μl was removed and analyzed by reverse phase HPLC using the system and conditions described above. Elution was monitored at 260 nm: the peaks corresponding to cisplatin (3.7 min), Pt₁-SG (4.7 min) and Pt₂-SG (4.2 min) were identified and quantified by integrating the areas under the peaks. Relative rates of Pt₁-SG formation were determined from quantified peak areas (4.7 min peaks) at 2 and 60 min with relative rates defined as Pt₁-SG formation of the test sample (± GST) ÷ non-enzymatic Pt₁-SG formation (- GST). Actual rates of cisplatin reactions with GSH were estimated from the rates of cisplatin depletion (Δ concentration cisplatin•time⁻¹) using quantified cisplatin peak areas (3.7 min peaks) obtained at 0, 2, and 60 minutes. Enzymatic rates were calculated by subtracting non-enzymatic (− GST) from enzymatic (+ GST) cisplatin depletion data. Dividing these values by molar enzyme concentrations resulted in approximate enzymatic activity towards cisplatin conjugation expressed as time⁻¹.

Western blot Analysis

Western blots for GSTP1-1 expression were prepared from whole cell lysates as described previously (27). For the analysis of p-c-Jun expression, MCF7 derivative cells were plated in 100 mm dishes at a density of 2.5 × 10⁶ cells/plate. Twenty-four hours later, cells were treated with 10, 25, 50 or 100 μM cisplatin or vehicle control for the indicated times. Cells were washed with PBS and total cell extracts were prepared as described (27). Immunoblot analysis of p-c-Jun was performed using 25 μg of whole cell lysate separated by 10% acrylamide SDS-PAGE. Proteins were transferred to a nitrocellulose membrane, blocked with a 5% milk solution and incubated with a 1:500 dilution of polyclonal rabbit anti-p-c-Jun antibody (sc-7981, Santa Cruz), followed by a 1:4000 dilution of goat horseradish peroxidase-conjugated anti-rabbit antibody. Bands corresponding to p-c-Jun were visualized by chemiluminescence (ECL). Equivalency of protein loading and transfer was established by staining the nitrocellulose membranes with Ponceau S.

RESULTS

Expression of allelic variants of GSTP1 in MCF7 cells and their role in cisplatin resistance

To determine the relative abilities of the variants of GSTP1-1 to influence cellular sensitivity to cisplatin, MRP-poor parental MCF7 cells (MCF7/WT) were stably transduced with expression vectors encoding the P1a, P1b, and P1c allelic variants of GSTP1. In addition, the MRP1-expressing transgenic derivative, MCF7/MRP1-10, was stably transduced with the GSTP1a-1a expression vector. Western blot analysis showed similar levels of GSTP1-1 protein expression in all six GST-transduced cell lines whereas GSTP1-1 was undetectable in the parental MCF7/WT or MCF7/MRP1-10 cells (Fig. 1). The GST activities of the 8 MCF7 cell lines are shown in table 1. Because MCF7/WT and /MRP1-10 cells have very little total GST, the activities of the transduced cell lines represent GSTP1-1 expression almost exclusively.

Figure 1. Expression of allelic variants of human GSTP1 in MCF7 cell derivatives.

Shown is a Western blot for GSTP1-1 expression in parental MCF7/WT cells stably transduced with: empty vector control (MCF7/WT-LN; lane 1), GSTP1a-1a expression vector (MCF7/P1a-7, lane 2), GSTP1b-1b expression vector (MCF7/P1b-9 and MCF7/P1b-18, lanes 3 and 4, respectively), GSTP1c-1c expression vector (MCF7/P1c-4 and MCF7/P1c-27, lanes 5 and 6, respectively), MRP1 expression vector alone (MCF7/MRP1-10, lane 7), and MRP1 plus GSTP1a-1a expression vectors (MRP1-10/P1a-4, lane 8). Each lane contained 50 µg total cellular protein. Shown are the positions of 28.2 and 18.8 kDa peptide standards.

Table 1.

GSTP1-1 expression in transduced MCF7 cell lines.

Cell Linea	GSTP1 allele	MRP1	GST activityb	GSTP1-1 (μg/mg)c
MCF7/WT-LN	–	−	< 5	–
MCF7/P1a-7	P1a (I104A113)	−	408 ± 25	4.5 ± 0.3
MCF7/P1b-9	P1b (V104A113)	−	170 ± 14	4.0 ± 0.3
MCF7/P1b-18	P1b (V104A113)	−	287 ± 24	6.8 ± 0.2
MCF7/P1c-4	P1c (V104V113)	−	224 ± 19	5.9 ± 0.5
MCF7/P1c-27	P1c (V104V113)	−	365 ± 25	9.6 ± 0.6
MCF7/MRP1-10	–	+	< 5	–
MRP1-10/P1a-4	P1a (I104A113)	+	300 ± 28	3.3 ± 0.3

^aCell lines designated MCF7/ were derived from parental MCF7/WT cells stably transduced with control vector (pLNCX, MCF7/WT-LN) or expression vectors encoding the P1a, P1b, or P1c alleles of human GSTP1 as indicated. MCF7/MRP1-10 cells were derived from MCF7/WT cells stably transduced with a human MRP1 expression vector. MRP1-10/P1a-4 cells were MCF7/MRP1-10 cells stably transduced with the GSTP1a expression vector.

^bGST activities were determined from whole cell lysates and are expressed as nmol CDNB conjugated•min⁻¹• mg⁻¹ of total cellular protein. Values are the means of 6 to 11 independent determinations ± 1 sem.

^cGSTP1-1 protein expression levels (μg GSTP1-1/mg total cellular protein) were calculated by dividing CDNB activity (^b) by the specific activities of purified GSTP1-1 variants towards CDNB. Values are the means of 6 to 11 independent determinations ± 1 sem.

Cisplatin cytotoxicity was determined in the 8 cell lines (Fig. 2). To ensure ample opportunity for GST to detoxify cisplatin, cells were exposed to drug for 24 hr (Fig. 2B and D) as well as 1 hr (Fig. 2A and C). The IC₅₀ values determined from these experiments were used to calculate relative resistance (Fig. 2A and B). As shown, relative to empty vector transduced control cells (MCF7/WT-LN, WT-LN), expression of all three alternative GSTP1 — alleles GSTP1a (MCF7/P1a-7), GSTP1b (MCF7/P1b-9 & -18), and GSTP1c (MCF7/P1c-4 & -27) — conferred a modest but statistically significant 1.4–1.7 fold resistance to cisplatin cytotoxicity using both durations of drug exposure (Fig. 2A and B). Remarkably, however, there was no consistent relationship between relative resistance and the particular variant expressed or its level of expression (table 1). Similarly, co-expression of GSTP1a-1a with MRP1 (MRP1-10/P1a-4 cells) conferred a low level of resistance to cisplatin that was statistically significant in the 24 hr exposure experiments (Fig. 2C and D). However, also shown in Fig. 2C and D, expression of MRP1 alone (MCF7/MRP1-10) conferred no survival advantage over MRP-poor cells (MCF7/WT-LN) nor was MRP1 able to augment the level of resistance conferred by GSTP1a-1a alone (compare MRP1-10/P1a-4 with P1a-7, Fig. 2C and D).

Figure 2. Expression of GSTP1 allelic variants in MCF7 cells confers resistance to cisplatin cytotoxicity.

Shown are cytotoxicity data from MCF7 derivatives exposed to varying concentration of cisplatin for 1 hr (A and C) or 24 hr (B and D) as described in Materials and Methods. The results in A and B are expressed as relative resistance—a value defined as the IC₅₀ of the test cell line ÷ the IC₅₀ of the MCF7/WT-LN control cell line. Data in C and D are expressed as IC₅₀ values. Cell lines are designated as follows: control cells, WT-LN (MCF7/WT-LN); cells expressing MRP1 alone, MRP1-10 (MCF7/MRP1-10); cells expressing GSTP1a-1a alone, P1a-7 (MCF7/P1a-7), or in combination with MRP1, MRP1-10/P1a-4; cells expressing GSTP1b-1b, P1b-9 and P1b-18 (MCF7/P1b-9 and -18); and cells expressing GSTP1c-1c, P1c-4 and P1c-27 (MCF7/P1c-4 and -27). Bars represent the means of 7–25 independent experiments ± 1 sem. Statistical analysis was accomplished with two-tailed, unpaired Student’s t tests. Asterisks denote values statistically different from WT-LN controls (panels A and B; P < 0.001). Statistical comparisons in panels C and D are indicated by brackets (N.S., not significant, P > 0.05).

Can MRP2 potentiate GSTP1-1-mediated resistance to cisplatin?

The failure of MRP1 to influence cisplatin cytotoxicity (Fig. 2C and D) was not entirely unexpected as previous studies have generally found no association between MRP1 expression and cisplatin resistance (35, 36). However, there is evidence supporting the role of MRP2 expression in cisplatin detoxication (37–39). To determine if MRP2 could potentiate or enhance GSTP1-1-associated cisplatin resistance, we performed additional cytotoxicity studies using HepG2 — cells a cell line expressing MRP2 but not MRP1 (29) — engineered to conditionally express GSTP1a-1a. As summarized in table 2, expression of GSTP1a-1a (HepG2/π3 minus doxycycline) conferred no survival advantage over control cells expressing little or no GST (HepG2/tTA and HepG2/π3 plus doxycycline) when the cells were treated with cisplatin for 1 hr or 24 hr. Moreover, inhibition of MRP2 with 20 or 50 μM MK571 had no effect on cisplatin cytotoxicity in HepG2/ 3 cells expressing GSTP1a-1a (Fig. 3). Control experiments revealed that similar treatment of HepG2/ 3 cells with 50 M MK571 resulted in 1.6 fold sensitization to doxorubicin cytotoxicity (supplemental Fig. 1) — suggesting that such treatment with MK571 effectively inhibits MRP2 function in HepG2 cells. Thus, in contrast to MCF7 cells, expression of GSTP1-1 failed to confer cisplatin resistance in HepG2 cells. Furthermore, expression of MRP2 neither conferred resistance (MK571 inhibition studies) nor potentiated GSTP1-1-mediated resistance to cisplatin (table 2).

Table 2.

GSTP1a-1a expression fails to confer resistance to cisplatin cytotoxicity in HepG2 cells.

	GST activityb	Relative Resistancec
Cell Linea	(nmol•min−1•mg−1)	1 hr exposure	24 hr exposure
HepG2/tTA	< 5	1.0	1.0
HepG2/π3 (−dox)	436 ± 35	1.0 ± 0.12	1.0 ± 0.13
HepG2/π3 (+dox)	10 ± 3	1.1 ± 0.13	1.3 ± 0.11

^aHepG2/tTA are HepG2 cells stably transfected with the doxycycline-repressible TetR-VP16 fusion vector. HepG2/π3 cells are HepG2/tTA cells further transfected with a GSTP1a expression plasmid under the transcriptional control of the Tet-response element. In some experiments, GSTP1a-1a expression was repressed by the inclusion of 1 μM doxycycline (+dox) in the tissue culture medium.

^bGST activity was expressed as in table 1. Values represent the means from ≥3 determinations ± 1 sem.

^cRelative resistance to CDDP cytotoxicity following 1 or 24 hr CDDP exposure was defined as the IC₅₀ for the test cell line ÷ IC₅₀ of HepG2/tTA cells. Values shown are the means from 6 independent experiments ± 1 sem.

Figure 3. Inhibition of MRP2 activity has no effect on cisplatin sensitivity of GSTP1a-1a expressing HepG2 cells.

Cytotoxicity of cisplatin was determined using the sulforhodamine B assay as described in Materials and Methods. HepG2/π3 cells were grown in the absence of doxycyline and express GSTP1a-1a (Table 2). Cells were treated for 1 hr with varying concentrations of cisplatin without MRP1/2 inhibitor MK571 (−MK571, closed circles) or in the presence of MK571 (+ 20 µM MK571, open circles or + 50 µM MK571, open diamonds). Values represent the means of 8 replicate determinations ± 1 sd.

Catalysis of cisplatin/GSH conjugation by variants of GSTP1-1 and its potential role in cisplatin resistance

Non-enzymatic reactions with GSH and cisplatin were characterized by HPLC/MS as described in Materials and Methods. These reactions yielded two major products, the predominant one eluting at 4.7 min with a m/z of 572 and the other at 4.2 min with a m/z 835. Mass spectra of each species showed the characteristic isotopic distribution of platinum (Fig. 4A & C). Further analysis by collision induced dissociation (CID) of the 572 parent ion yielded a fragmentation pattern (Fig. 4B) consistent with the mono-platinum-mono-GSH structure (Pt₁-SG) shown (inset Fig. 4A). The other species, m/z 835, yielded a CID spectrum (Fig. 4D) consistent with the bis-platinum-mono-GSH structure (Pt₂-SG) (inset Fig. 4C) previously described by Bernareggi et al. (9) although these data cannot rule out the alternative Pt₂-SG structure proposed by Townsend et al. (12). We were unable to detect any product consistent with the bis-GSH-mono-platinum adduct reported by Odenheimer and Wolf (40) and Ishikawa and Ali-Osman (11). Only the dominant adduct, Pt₁-SG, showed linear, time-dependent non-enzymatic and GSTP1-1-catalyzed formation; hence, analysis of this species was chosen for the following enzymatic studies.

Figure 4. Analysis of cisplatin/GSH conjugates by mass spectrometry.

Cisplatin was reacted with GSH and the products analyzed by HPLC/ESI/MS and HPLC/ESI/MS/MS as described in Materials and Methods. The major peak eluting at 4.7 min was analyzed by ESI/MS (panel A) and by ESI/MS/MS (panel B), both in the positive ion mode. These spectra are consistent with the mono-Pt-mono-GSH adduct, Pt₁-SG, whose structure is shown as an inset in panel A. Major ions formed by CID (panel B) of the parent 572 ion (panel A) are consistent with the following: 553, M-NH₃, 461, M-glycine-Cl; 442, M-glutamate; and 424, M-glutamate-NH₃. Analysis of the lesser peak eluting at 4.2 min (panels C and D) suggested the bis-Pt-mono-GSH structure, Pt₂-SG, shown as an inset in panel C. Ions identified in the CID spectrum (panel D) are consistent with: 835, precursor ion; 818, M-NH₃; 801, M-2NH₃; 784, M-3NH₃; 764, M-2NH₃-Cl; 726, M-2NH₃-glycine; and 709, M-3NH₃-glycine.

Inclusion of purified recombinant GSTP1-1’s (33 U) in reactions containing 1 mM cisplatin and 2 mM GSH (pH 6.8, 25°C) resulted in very modest but statistically significant increases in relative Pt₁-SG formation after 2 and 60 minute incubations for all three allelic variants of GSTP1 (Fig. 5). The increases in relative Pt₁-SG conjugation rates, over non-enzymatic rates, were greatest for GSTP1b-1b and GSTP1c-1c. When the non-enymatic rates were subtracted from the data shown in Fig. 5 and the resulting relative GST-mediated conjugation rates are expressed as relative rates per mg GSTP1-1, the GSTP1b-1b and GSTP1c-1c variants supported rates of conjugation which were 1.8-1.9 fold higher (80–90% increase) than the rate mediated by GSTP1a-1a. Despite these increased rates of conjugation mediated by GSTP1b-1b and GSTP1c-1c, no discernible differences in the sensitivities toward cisplatin cytotoxicity were observed among MCF7 cells expressing the three alternative GSTP1-1 variants (Fig. 2). These data suggest that GSTP1-1 catalysis of Pt₁-SG conjugate formation may not be the sole, or even major, determinant of cisplatin resistance associated with GSTP1-1 expression in MCF7 cells.

Figure 5. Catalysis of cisplatin/GSH conjugation by variants of GSTP1-1.

Cisplatin (1 mM) and GSH (2 mM) were incubated in the presence or absence of 33 U/ml GSTP1a-1a, GSTP1b-1b, or GSTP1c-1c at pH 6.8, 25°C. Aliquots were removed at 2 min (panel A) or 60 min (panel B) and Pt₁-SG formation was quantified by HPLC analysis as described in Materials and Methods. Pt₁-SG levels formed in the presence of GST were normalized to Pt₁-SG formation in the absence of GST (−GST) and were expressed as relative Pt₁-SG formation. The bars represent mean values obtained from 6–18 independent experiments and the error bars indicate ± 1 sem. Pt₁-SG formation was significantly higher in GSTP1-1 containing reactions than in control (−GST) reactions at both time points (*, P < 0.05; **, P < 0.01; Student’s t test).

To further address this point, the actual enzymatic rates for GSTP1-1-mediated cisplatin conjugation were estimated from cisplatin depletion data as described in Materials and Methods. Under the conditions used (1 mM cisplatin, 2 mM GSH, 7.8–18.6 μM GSTP1-1, pH 6.8, 25°C) the enzymatic rates for the three allelic variants ranged from 1.7 to 2.6 hr⁻¹ —an extremely slow rate, indeed, lending further support to the argument that GSTP1-1 catalysis of cisplatin conjugation is not likely to be a major determinant of cellular sensitivity to cisplatin.

Alternative mechanisms of GSTP1-1-mediated cisplatin resistance: influence of GSTP1-1 expression on JNK/SAPK signaling pathways

Expression of GSTP1-1 is reported to interact with and modulate the activities of several signaling kinases and pathways (41–43), including the c-Jun N-terminal kinase/stress activated protein kinase (JNK/SAPK) pathways, which — in susceptible cells influence cell survival and death (44–46). A comprehensive analysis of these pathways is beyond the scope of the present study; however, we were particularly interested in examining whether GSTP1-1 expression could modulate JNK-associated c-Jun phosphorylation a signaling event linked to cisplatin-mediated apoptosis and cytotoxicity in several studies (44, 46, 47). Using p-c-Jun levels as an indicator of activity in this signaling pathway, treatment of parental MCF7 cells with 100 M cisplatin resulted in a robust time-dependent increase in p-c-Jun; whereas, expression of GSTP1a-1a resulted in significant reductions of cisplatin-induced p-c-Jun levels (supplemental Fig. 2). In contrast to treatment with 100 M cisplatin, treatment at lower concentrations (10–50 μM) demonstrated very little difference between the two cell lines with respect to p-c-Jun increases. Hence, the significance of GSTP1-1-mediated attenuation of JNK activity observed at 100 μM cisplatin is uncertain.

DISCUSSION

In several published reports, increased expression of GSTP1-1 has been associated with cellular resistance to the DNA damaging cancer drug, cisplatin (13–15, 17). The formation of GSH conjugates with cisplatin (Pt-SG’s) renders the drug unable to form DNA strand crosslinks and, hence, is expected to significantly reduce its cytotoxicity. Thus an obvious potential mechanism for GSTP1-1-associated resistance to cisplatin is via the catalysis of Pt-SG formation. Here we have evaluated the role of human GSTP1-1 in cisplatin resistance: using purified variants of GSTP1-1, catalysis of Pt₁-SG formation was examined, quantitative differences in catalysis of Pt₁-SG formation between the variants were identified, and—using model transgenic cell lines expressing the alternative allelic variants of GSTP1—we asked whether these differences in catalysis would be associated with differences in cisplatin sensitivity. In addition, previous results from our laboratory have suggested that, because of product inhibition or residual toxicity of glutathione conjugates, expression of GST is often insufficient to confer significant cellular protection from DNA damaging electrophiles unless a GS-X efflux transporter is also expressed (27–29). Moreover, Pt-SG’s—while unable to form DNA strand crosslinks—may retain some toxicities such as the inhibition of protein synthesis (11). Thus, a secondary goal of the present study was to evaluate the potential contribution of GS-X transporters on cisplatin resistance.

We find that forced expression of GSTP1-1 in MCF7 cells confers low level but statistically significant resistance to cisplatin -cytotoxicity (Fig. 2). Variants of GSTP1-1 enhance the rate of cisplatin conjugation with GSH (Pt₁-SG formation) and the GSTP1b-1b and GSTP1c-1c variants are 80–90% more active in catalyzing these reactions than is the GSTP1a-1a variant. Despite these differences in catalytic activity, there are no discernible differences in cisplatin resistance conferred by expression of the alternative alleles (Fig. 2A and B) even though the levels of GSTP1-1 expression are comparable among the transgenic clones (table 1). These data cast doubt on the importance of GSTP1-1-mediated catalysis of Pt₁-SG formation as a mechanism of cisplatin resistance. More striking is the finding that the catalytic activities of GSTP1-1 variants towards cisplatin conjugation are extremely low at 1.7 to 2.6 hr⁻¹ (25°C, 1 mM cisplatin, 2 mM GSH, pH 6.8)—values that are even less than the k_cat of ~14 hr⁻¹, itself quite low, that can be calculated from the data of Goto et al. (18)¹. Finally, expression of MRP1 (MCF7 cells, Fig. 2C and D) or MRP2 (HepG2 cells, table 2 and Fig. 3) neither confers cellular resistance to cisplatin nor potentiates (or augments) GSTP1-1-mediated resistance to cisplatin. Together these data suggest GSTP1-1 catalysis of Pt₁-SG formation is very unlikely to have a significant impact on cellular sensitivity to cisplatin toxicity with or without the co-expression of GS-X efflux transporters.

Thus while the studies described herein confirm that GSTP1-1 can confer cisplatin resistance in some cell lines, an alternative mechanism that does not involve catalysis of cisplatin/GSH conjugation is necessary to explain GSTP1-1-mediated resistance. There are several reports indicating that expression of GSTP1-1 can modulate the activities of multiple signaling pathways (41–43). A comprehensive analysis of these pathways is beyond the scope of the present work. Nevertheless, reports that GSTP1-1 may inhibit JNK/SAPK signaling were particularly interesting (41): this is because cisplatin-induced activation of JNK signaling is implicated in increased apoptosis and cell death while inhibition of JNK/SAPK is associated with resistance to cisplatin (44, 45). These findings lead to the suggestion that JNK/SAPK pathways can play crucial roles in the execution of cisplatin-mediated cell death. To test the idea that GSTP1-1 may influence cellular sensitivity to cisplatin by interfering with JNK signaling, we examined changes in phosphorylated levels of the JNK substrate, c-Jun, upon cisplatin treatment. Indeed we found that upon exposure of cells to high concentration (100 M) cisplatin, expression of GSTP1-1 is associated with significant reductions in cisplatin-induced p-c-Jun levels (supplemental Fig. 2)—consistent with the suggestion that GSTP1-1 may mediate resistance by inhibition of the JNK pathway. However, as little difference in cisplatin-induced p-c-Jun levels was observed between GSTP1-1-expressing and non-expressing MCF7 cells at lower cisplatin concentrations, the significance of GSTP1-mediated attenuation of JNK signaling at 100 μM cisplatin is uncertain.

The signaling events induced by cisplatin are many and complex. For example, in addition to JNK/SAPK, both ERK and p38 pathways are affected by cisplatin treatment and are reported to be influenced by GSTP1-1 (43). ERK activation, in particular, may oppose the apoptotic response associated with JNK/SAPK activation in cisplatin-treated cells. Moreover, in some cells, in contrast or in addition to its role in cisplatin-induced apoptosis, JNK activation is implicated in protection from cisplatin toxicity via enhanced repair of DNA damage (48, 49) or by alternative mechanisms (50). The balance and magnitude of these potentially opposing signaling events are very likely cell-specific and, hence, the net effect of GSTP1-1 expression on cisplastin sensitivity may well vary between cells of different origin. Another potential determinant of altered JNK signaling is the relative levels of JNK and GSTP1 expressed and, hence, the ratios of JNK:GSTP1 complex formed. Indeed, experiments using GSTP-deficient mice indicate that JNK signaling is constitutively increased in GstP1/P2^(−/−) animals (51). In MCF7 cells, GSTP1-1 expression resulted in only 1.4–1.7 fold resistance to cisplatin (Fig. 2); whereas, in other cell lines—in which the balance of opposing signaling pathways is shifted—inhibition of JNK or modulation of other signaling pathways by GSTP1-1 may have a relatively larger effect on cellular sensitivity to cisplatin toxicity. Regardless, based upon enzymatic data presented herein, the expression of GSTP1-1 is not likely to have a major effect on cellular sensitivity to cisplatin via catalysis of Pt₁-SG formation; rather, an alternative mechanism is required to explain the high level resistance to cisplatin associated with GSTP1-1 expression reported in some studies.

Abbreviations used

CDNB

1-chloro-2,4-dinitrobenzene

JNK

c-Jun N-terminal kinase

GSH

glutathione

GST

GSH-S-transferase

GS-X

GSH conjugate

MRP

multidrug resistance or multidrug resistance-associated protein

Pt-SG

GSH conjugates of cisplatin

SAPK

stress activated protein kinase