Biochemical Journal

Research article

Modulation of doxorubicin resistance by the glucose-6-phosphate dehydrogenase activity

Manuela Polimeni , Claudia Voena , Joanna Kopecka , Chiara Riganti , Gianpiero Pescarmona , Amalia Bosia , Dario Ghigo

Abstract

How anti-neoplastic agents induce MDR (multidrug resistance) in cancer cells and the role of GSH (glutathione) in the activation of pumps such as the MRPs (MDR-associated proteins) are still open questions. In the present paper we illustrate that a doxorubicin-resistant human colon cancer cell line (HT29-DX), exhibiting decreased doxorubicin accumulation, increased intracellular GSH content, and increased MRP1 and MRP2 expression in comparison with doxorubicin-sensitive HT29 cells, shows increased activity of the PPP (pentose phosphate pathway) and of G6PD (glucose-6-phosphate dehydrogenase). We observed the onset of MDR in HT29 cells overexpressing G6PD which was accompanied by an increase in GSH. The G6PD inhibitors DHEA (dehydroepiandrosterone) and 6-AN (6-aminonicotinamide) reversed the increase of G6PD and GSH and inhibited MDR both in HT29-DX cells and in HT29 cells overexpressing G6PD. In our opinion, these results suggest that the activation of the PPP and an increased activity of G6PD are necessary to some MDR cells to keep the GSH content high, which is in turn necessary to extrude anticancer drugs out of the cell. We think that our data provide a new further mechanism for GSH increase and its effects on MDR acquisition.

  • doxorubicin
  • glucose-6-phosphate dehydrogenase (G6PD)
  • glutathione (GSH)
  • multidrug resistance (MDR)
  • pentose phosphate pathway (PPP)

INTRODUCTION

The resistance to anticancer drugs such as the anthracycline doxorubicin is still a major cause of chemotherapy failure in cancer patients [1,2]. The mechanisms involved are complex and multifactorial and may depend on an inadequate drug exposure and/or genetic and epigenetic alterations in the cancer cell itself [3]. The problem of resistance is compounded by the fact that the amount of doxorubicin which can be administered is limited by its dose-dependent cardiotoxicity [4]. Tumour cells that develop drug resistance generally exhibit a cross-resistance to a large number of compounds that are chemically and structurally unrelated. This acquired MDR (multidrug resistance) is determined in part by the overexpression of efflux pumps, such as P-glycoprotein and MRPs (MDR-associated proteins) [1]. These efflux transporters show a different substrate selectivity. In particular, some MRPs are involved in the transport of GSH (glutathione), glucuronate or sulfate conjugates of organic anions that arise from detoxification reactions by phase II conjugating enzymes [5], but MRPs also export GSH, and co-transport with GSH is required for the MRP-mediated extrusion of chemotherapeutic drugs [6]. Furthermore, both carcinogenesis and the MDR phenotype are frequently associated with an increased oxidative stress and activation of cellular redox metabolism [7]. GSH is an ubiquitous tripeptide that represents the most abundant cellular thiol. Owing to its reactivity and high intracellular concentration, GSH functions as an effective detoxification system that plays a pivotal role in cancer [8] and MDR development [9].

The PPP (pentose phosphate pathway) is the most important pathway for cellular GSH recycling. In the PPP oxidative phase, glucose 6-phosphate is irreversibly converted into ribulose 5-phosphate and CO2 leading to the synthesis of NADPH, a redox cofactor for many antioxidant enzymes. G6PD (glucose-6-phosphate dehydrogenase) catalyses the first and rate-limiting step of the pathway, and is mainly regulated by the NADPH/NADP+ ratio. The NADPH produced is used by glutathione reductase to reduce GSSG (glutathione disulfide) to GSH.

The aim of the present study has been to clarify the role of the PPP, and in particular of its key enzyme G6PD, in MDR. After having confirmed that doxorubicin-resistant human colon cancer cells (HT29-DX) overexpress MRP1 and MRP2, accumulate less doxorubicin, and show an increased content of GSH when compared with doxorubicin-sensitive human colon cancer cells (HT29), we investigated whether the increase of intracellular GSH or G6PD may be sufficient to induce the onset of the MDR phenotype in HT29 cells, and whether G6PD inhibition may increase the cytotoxic effect of doxorubicin in drug-resistant cells.

EXPERIMENTAL

Materials

Plasticware for cell culture was from Falcon (Becton Dickinson); [1-14C]glucose and [6-14C]glucose were from Dupont/New England Nuclear; the protein content of cells was assessed with the BCA (bicinchoninic acid) kit. If not otherwise specified, the reagents used were purchased from Sigma.

Cells

Human epithelial lung adenocarcinoma cells (A549), colon doxorubicin-sensitive cancer cells (HT29) (both provided by the Istituto Zooprofilattico Sperimentale ‘Bruno Ubertini’, Brescia, Italy) and doxorubicin-resistant cancer cells (HT29-DX cells), obtained as described previously [10], were cultured in RPMI 1640 medium (Gibco) supplemented with 10% FBS (fetal bovine serum) and 1% penicillin/streptomycin, and maintained in a humidified atmosphere at 37 °C and 5% CO2. HT29-DX cells were cultured in RPMI 1640 containing 34 nmol/l doxorubicin (the ‘maintenance’ dose). Depending on the number of cells required for specific experiments, cells were plated on either 100-mm, 60-mm or 35-mm Petri dishes and cultured until an approximate confluence of 70–80%.

Doxorubicin accumulation

Intracellular doxorubicin accumulation was measured using a PerkinElmer LS-5 fluorimeter as described previously [10]. Fluorescence was converted into nanomoles of doxorubicin per mg of cellular proteins using a calibration curve.

Measurement of extracellular LDH (lactate dehydrogenase) activity

The cytotoxic effect of doxorubicin was measured as leakage of LDH activity into the extracellular medium [11] after 24 h of incubation using a Packard EL340 microplate reader (Bio-Tek Instruments). Both intracellular and extracellular enzyme activities were expressed as micromoles of NADH oxidized/min per dish and then extracellular LDH activity (LDH out) was calculated as the percentage of the total (intracellular+extracellular) LDH activity (LDH tot) in the dish.

Detection of apoptosis

Induction of apoptosis due to the cytotoxic effect of doxorubicin was evaluated using the annexin V-FITC apoptosis detection kit, as described previously [12]. The fluorescence level of annexin V was recorded using a FACSCalibur system and the percentage of cells positive for annexin V was calculated using Cell Quest software (Becton Dickinson). The results were expressed as the average percentage values obtained from triplicate experiments compared with controls.

Measurement of PPP activity

The metabolic fluxes through the PPP and the tricarboxylic acid cycle were measured using a closed experimental system to trap the 14CO2 developed from [14C]glucose over 1 h in the absence or presence of an oxidative stress (menadione), as described previously [11]. The PPP metabolic flux (expressed as picomoles of CO2/h per mg of cell proteins) was obtained by subtracting the amount of CO2 developed from [6-14C]glucose from the CO2 released from [1-14C]glucose. The extent of [6-14C]glucose metabolism, an index of the tricarboxylic acid cycle alone, did not change significantly in the different experimental conditions (results not shown).

Measurement of GSH

The cells were washed with PBS and 600 μl of 0.01 M HCl was added to each cell monolayer. After gentle scraping, the cells were frozen/thawed twice and the proteins were precipitated by adding 120 μl of 6.5% (w/v) 5-sulfosalicylic acid to 480 μl of cell lysate. Each sample was placed in ice for 1 h and centrifuged at 4 °C for 15 min at 13000 g. The content of GSH was measured using a Packard microplate reader EL340 as described previously [13]. The results were expressed as nanomoles of GSH per mg of cellular proteins.

Measurement of G6PD and 6PGD (6-phosphogluconate dehydrogenase) activities

Enzymatic activities were measured by using a Packard microplate reader EL340 as described previously [11]. A first measurement is performed by adding a saturating amount of both 6-phosphogluconate and glucose 6-phosphate to the assay system to ensure that the rate of NADP+ reduction is the result of both G6PD and 6PGD activities. A second assay was performed with 6-phosphogluconate as the only substrate. This assay allowed the measurement 6PGD activity alone. G6PD activity is obtained by subtracting the rate of the second assay from the rate of the first one [11]. Enzymatic activity was expressed as micromoles of NADP+ reduced/min per mg of cell proteins.

Western blotting

Whole-cell extracts were directly solubilized in an ice-cold sample buffer (for G6PD, 63 mmol/l Tris/HCl, 2% SDS and 10% glycerol, pH 6.8) or SDS lysis buffer (for MRP1, MRP2 and RLIP76, 50 mmol/l Tris/HCl, 10 mmol/l EDTA and 1% SDS, pH 8.1) and supplemented with the protease inhibitor cocktail set III {100 mmol/l AEBSF [4-(2-aminoethyl)benzenesulfonyl fluoride], 80 μmol/l aprotinin, 5 mmol/l bestatin, 1.5 mmol/l E-64 [trans-epoxysuccinyl-L-leucylamido-(4-guanidino)butane], 2 mmol/l leupeptin and 1 mmol/l pepstatin; Calbiochem-Novabiochem}, 2 mmol/l PMSF and 1 mmol/l sodium orthovanadate. Samples were sonicated on ice with two 10-s bursts and incubated for 20 min at 4 °C with agitation. Proteins were separated by SDS/PAGE (12% or 7% gels) and probed with antibodies diluted in TBS (25 mmol/l Tris/HCl, pH 7.6, and 150 mmol/l NaCl)/Tween (0.1%) with 5% non-fat dried milk. The primary antibodies were rabbit anti-G6PD (1:250; Santa Cruz Biotechnology), mouse anti-MRP1 (1:50; Abcam), mouse anti-MRP2 (1:50; Abcam), mouse anti-RLIP76 (1:50; Abcam) and rabbit anti-GAPDH (glyceraldehyde-3-phosphate dehydrogenase; used as a housekeeping gene; 1:500; Santa Cruz Biotechnology). The blots were probed with horseradish peroxidase-conjugated goat anti-rabbit (1:4000; Bio-Rad Laboratories) or goat anti-mouse (1:4000; Bio-Rad Laboratories) antibodies. Proteins were detected by enhanced chemiluminescence (PerkinElmer). Densitometric analysis of the bands was performed using ImageJ software (http://rsbweb.nih.gov/ij/). The relative quantification of each sample was performed comparing each band with GAPDH.

qRT-PCR (quantitative real-time PCR)

Total RNA was obtained by the guanidinium thiocyanate-phenol-chloroform method [14]. Total RNA (3 μg) was reverse-transcribed into cDNA and qRT-PCR was carried out using IQ™ SYBR Green Supermix (Bio-Rad Laboratories) for the quantification of G6PD, MRP1, MRP2, RLIP76 and GAPDH according to the manufacturer's instructions. The RT-PCR primers used are illustrated in Table 1. PCR amplification was 1 cycle of denaturation at 95 °C for 3 min, 45 cycles of denaturation at 94 °C for 30 s, annealing (at temperatures indicated in Table 1) for 30 s, and synthesis at 72 °C for 30 s (G6PD and GAPDH) or 1 min (MRP1, MRP2 and RLIP76). The relative quantification of each sample was performed comparing each PCR gene product with the GAPDH product using the Gene Expression Macro (http://www3.bio-rad.com/LifeScience/jobs/2004/04-0684/genex.xls; Bio-Rad Laboratories).

View this table:
Table 1 Details of the primers used

Primers used for qRT-PCR and annealing temperature.

Immunofluorescence staining

Cells (0.5×106) were grown on sterile glass coverslips, rinsed with PBS, fixed with 4% (w/v) paraformaldehyde (diluted in PBS), washed three times with PBS and probed with primary antibodies against mouse anti-MRP1 and -MRP2 (1:50 dilution in 1% PBS + FBS; Abcam) and TRITC (tetramethylrhodamine β-isothiocyanate)-conjugated goat anti-mouse antibodies (diluted 1:400 in 1% PBS + FBS). The cells were then washed with PBS and incubated with DAPI (4′,6-diamidino-2-phenylindole; 1:20000). Fluorescently labelled cells were washed three times with PBS and once with water and then the slides were mounted with 4 μl of Gel Mount Aqueous Mounting medium and examined with a Leica DC100 fluorescence microscope (Leica Microsystems). For each experimental point, a minimum of five microscopic fields were examined.

DNA constructs, retroviral production and cell infection

To obtain the G6PD construct the total cDNA from HT29 cell lines was amplified using specific primers flanking the G6PD-encoding region. The primers used were: G6PD forward 5′-TGAGGCGGGTCCGCTCA-3′ and reverse 5′-TGCCCAGGGCTCAGAGCTT-3′. PCR products were cloned into the PCRII vector using the TA-cloning system (Invitrogen) and sequenced. For retroviral production, G6PD cDNA was then cloned into the retroviral vector Pallino [15], the G6PD construct was digested to obtain a HindIII-XhoI fragment and then it was cloned into the HindIII-XhoI sites of the Pallino vector. G6PD retroviruses were obtained by co-transfection of the Pallino expression vector containing G6PD with the pMD2VSV-G plasmid into the GP-293 packaging cell line (Invitrogen). The original GP-293 cell line was created by stably transfecting the gag/pol genes into HEK (human embryonic kidney)-293 cells and selecting for the Streptoalloteichus hindustanus bleomycin gene. Empty vector Pallino retroviruses were generated as a control. Retroviruses released into culture medium were collected 24 h after transfection. A 500 μl aliquot of filtered (pore size 0.45 mm) supernatants from GP-293 cells were supplemented with 8 μg/ml polybrene and added to 1×105 target cells. After a 12 h incubation, 1 ml of complete medium was added and the cells were cultured for a further 2 days. The percentage of transfected cells were analysed for GFP (green fluorescent protein) expression on a FACSCalibur flow cytometer (Becton Dickinson) (results not shown). The Cell Quest software (Becton Dickinson) was used for the data acquisition and analysis.

Pgp (glycoprotein P) silencing

Cells (150000) were treated for 6 h with 1 ml of siRNA (small interfering RNA) transfection medium, containing 5 μl of siRNA transfection reagent and 50 pmol of Pgp siRNA (Santa Cruz Biotechnology). In each set of experiments, one dish was treated with 50 pmol of control siRNA-A (Santa Cruz Biotechnology), a non-targeting 20–25 nt siRNA designed as a negative control instead of Pgp siRNA. After 72 h the silencing efficacy was verified by Western blotting for Pgp protein. Cell toxicity was assessed by monitoring the extracellular release of LDH and was superimposable to the one observed in untreated cells (results not shown).

Statistical analysis

All of the results in the text and Figures are means±S.E.M. The results were analysed by one-way ANOVA and Tukey's test. P<0.05 was considered significant.

RESULTS

HT29-DX cells are a good model of acquired doxorubicin-resistant cells as they accumulate less doxorubicin than their normal counterpart HT29 cells and are subsequently less sensitive to its cytotoxic effects [10]. In the present paper we have investigated the expression of three MRP drug-efflux pumps (MRP1, MRP2 and RLIP76), which have been implicated in GSH-associated drug resistance [9]. In keeping with the data from the literature obtained in other cell types, HT29-DX cells showed a significantly increased expression (in terms of both mRNA and protein) of MRP1 and MRP2 when compared with their sensitive control (Figure 1). However, RLIP76 expression was similar in both cellular subpopulations (Supplementary Figures S1 and S2 at http://www.BiochemJ.org/bj/439/bj4390141add.htm).

Figure 1 MRP1 and MRP2 expression

Relative expression of MRP1 and MRP2 mRNA and proteins, checked by qRT-PCR and Western blotting in HT29 and HT29-DX cells. Left-hand and middle panels: results are means±S.E.M. for three different experiments. *P<0.0001 compared with HT29 cells. Right-hand panel: representative of three experiments giving similar results.

The basal level of GSH in HT29-DX cells was significantly higher than in HT29 cells (Figure 2, first dataset), whereas the accumulation of doxorubicin was significantly lower than in the controls (Figure 2, second dataset). In the presence of 5 μmol/l doxorubicin, HT29 cells showed a significant decrease in GSH, as was expected due to the well-known ability of doxorubicin to elicit oxidative stress [16]. The same effect was not observed in HT29-DX cells (Figure 2, second dataset).

Figure 2 GSH levels and doxorubicin accumulation

HT29, HT29-DX and A549 cells were incubated for 3 h in the absence (CTRL) or presence of 5 μmol/l doxorubicin (DOXO), with or without 30 mmol/l GSH (n=4), and then intracellular GSH and doxorubicin accumulation were measured as indicated in the Experimental section. GSH versus HT29 CTRL: *P<0.05; **P<0.0001. DOXO versus HT29 DOXO: °P<0.05; °°P<0.001.

A way to increase the intracellular level of glutathione is to maintain it in the reduced form by activating PPP flux via the activation of G6PD. HT29-DX cells showed a significantly higher basal activity of the PPP compared with the control cells (Figure 3). When the PPP activity was stimulated by the presence of menadione, a compound which exerts an oxidative stress by generating superoxide anions through its redox cycling and by forming a conjugate with GSH [17], a very significant increase of the metabolic pathway was observed in both cell cultures (Figure 3). The menadione-stimulated PPP activity is a useful index of maximal activation of the pathway and therefore of its antioxidant potential. HT29 and HT29-DX cells were also incubated in presence of the non-competitive G6PD inhibitor DHEA (dehydroepiandrosterone) [1820] or the competitive inhibitor 6-AN (6-aminonicotinamide) [21,22] for 24 h before the measurement. Both compounds were able to significantly inhibit the PPP activity, both basal and menadione-induced, in HT29-DX and HT29 cells (Figure 3). G6PD is the key regulatory enzyme of the PPP flux. HT29-DX cells showed a significantly increased basal expression of G6PD (mRNA and protein) when compared with their sensitive counterpart (Figure 4A).

Figure 3 Effects of DHEA and 6-AN on PPP activity

Effect of either DHEA or 6-AN on basal and menadione-stimulated PPP activity in HT29 and HT29-DX cells. Cells were incubated for 24 h in the absence (CTRL) or presence of either DHEA (100, 250 or 500 μmol/l) or 6-AN (100, 250 or 500 μmol/l), and the PPP was measured after 1 h of further incubation in the absence or presence of 100 μmol/l menadione (MEN) as indicated in the Experimental section (n=5). In the tests with menadione the concentrations of DHEA and 6-AN were 100 μmol/l. Versus HT29 CTRL: *P<0.05; **P<0.005; ***P<0.0001. Versus HT29-DX CTRL: °P<0.05; °°P<0.0001. Versus HT29+MEN: #P<0.0001. Versus HT29-DX+MEN: §P<0.0001.

Figure 4 G6PD expression and activity.

(A) Relative expression of G6PD mRNA and protein, evaluated by qRT-PCR and Western blotting respectively, in HT29 and HT29-DX cells. Left-hand panel: results are means±S.E.M. for three different experiments. *P<0.02. Right-hand panel: the Figure is representative of three experiments giving similar results. *P<0.002. (B) Effect of either DHEA or 6-AN on G6PD activity and 6PGD activity in HT29 and HT29-DX cells. Cells were incubated for 24 h in the absence (CTRL) or presence of either DHEA (100, 250 or 500 μmol/l) or 6-AN (100, 250 or 500 μmol/l) (n=3). Versus HT29 CTRL: *P<0.05; **P<0.0001. Versus HT29-DX CTRL: °P<0.001; °°P<0.0001.

After a 24 h incubation with either DHEA or 6-AN, the expression of G6PD was not modified significantly in both cell culture types when compared with their controls (results not shown). Resistant cells showed a significantly increased enzymatic activity in comparison with HT29 cells (Figure 4B). A 24 h incubation of cells with DHEA or 6-AN inhibited the G6PD activity in HT29-DX cells and in HT29 cells, but only at the highest concentration tested. This inhibition was quite specific, since the activity of another oxidoreductase of the same pathway, 6PGD, was not modified in comparison with the controls (Figure 4B). The different degree of inhibition of the PPP and G6PD by DHEA and 6-AN can be explained with the different conditions of measurement of these two events. PPP flux is strictly related to in vivo G6PD activity, and is measured at the intracellular concentrations of reagents, which are low and not saturating, so that the G6PD rate is normally far from Vmax. Instead, the G6PD activity is measured in lysed cells in the presence of saturating concentrations of reagents, then the enzyme rate matches the Vmax.

HT29 and HT29-DX cells were incubated for 24 h in the absence or presence of DHEA or 6-AN, then 5 μmol/l doxorubicin was added during the last 3 h of incubation. As described previously (Figure 2, second dataset), the basal drug accumulation in HT29-DX cells was significantly lower than HT29 cells (Figure 5A). DHEA and 6-AN induced a significant increase in intracellular doxorubicin in the resistant HT29-DX cells compared with HT29 cells. A significant increase was observed in the doxorubicin-sensitive HT29 cells incubated with DHEA, but this effect was only observed at the highest concentration of 6-AN tested (Figure 5A). To clarify the eventual role of Pgp in the accumulation of doxorubicin, HT29-DX cells were silenced for 72 h for Pgp [23] and were subsequently incubated with 5 μmol/l doxorubicin for 24 h, in the presence or absence of DHEA (100, 250 and 500 μmol/l). In Pgp-silenced HT29-DX cells, DHEA treatment did not significantly modify the accumulation of doxorubicin compared with the Pgp-containing cells (Supplementary Figure S3 at http://www.BiochemJ.org/bj/439/bj4390141add.htm). LDH is a soluble cytosolic enzyme that is released into the culture medium following loss of membrane integrity and thus it represents a sensitive index of cytotoxicity. The cell damage induced by doxorubicin was also investigated in terms of apoptosis by the fluorescence of cell-associated annexin V. As described previously [10], a 24 h incubation of HT29 cells with doxorubicin caused an increase in LDH release into the extracellular medium and in the percentage of cells positive for annexin V, whereas it did not induce any modification in HT29-DX cells (Figures 5A and 5B). DHEA and 6-AN were able to increase the concentration-dependent cytotoxic effect of doxorubicin, observed after a 24 h incubation, only in HT29-DX cells (Figure 5A), whereas in the absence of doxorubicin they did not modify the LDH release (results not shown). In contrast, apoptosis induced by doxorubicin was detectable after a 24 h incubation with DHEA or 6-AN in both cell types. In the absence of doxorubicin the percentage of annexin V-positive cells was significantly higher only in the control HT29 cells (Figure 5B).

Figure 5 Effects of DHEA and 6-AN on doxorubicin accumulation, cytotoxicity and intracellular GSH levels

Effect of either DHEA or 6-AN on doxorubicin accumulation, LDH release into the extracellular medium (A) and the induction of apoptosis (B) in HT29 and HT29-DX cells. For doxorubicin accumulation, cells were incubated for 24 h in the absence (CTRL) or presence of either DHEA (100, 250 or 500 μmol/l) or 6-AN (100, 250 or 500 μmol/l), then (except for CTRL) in the last 3 h of incubation 5 μmol/l doxorubicin (DOXO) was added (n=5). For LDH release and apoptosis, cells (except CTRL) were incubated for 24 h with 5 μmol/l doxorubicin in the presence of either DHEA (100, 250 or 500 5 μmol/l) or 6-AN (100, 250 or 500 5 μmol/l) (n=5). (A) DOXO versus HT29 DOXO: *P<0.05; **P<0.001; ***P<0.0001. DOXO versus HT29-DX DOXO: °P<0.05; °°P<0.001; °°°P<0.0001; LDH versus HT29 CTRL: §P<0.05. LDH versus HT29-DX DOXO: #P<0.05; ##P<0.001; ###P<0.0001. (B) Cells were incubated for 24h in the absence (CTRL) or presence of either DHEA (100, 250 or 500 μmol/l) or 6-AN (100, 250 or 500 μmol/l) (n=5). Versus HT29 CTRL: *P<0.05; **P<0.001; ***P<0.0001. Versus HT29-DX CTRL: °P<0.05; °°P<0.001; °°°P<0.0001. (C) Results were expressed as the relative values obtained compared with HT29 cells (assumed as value=1) (n=3). Versus HT29 CTRL: *P<0.05; **P<0.01; ***P<0.005. Versus HT29+DOXO: §P<0.05; §§P<0.005. Versus HT29-DX+DOXO: #P<0.05; ##P<0.01.

The PPP flux is very important to provide the cell with the NADPH necessary to keep GSH in the reduced form. We had observed previously that the intracellular content of GSH, in agreement with the PPP activity, shows a significantly higher level in HT29-DX cells than in the sensitive cells. After a 24 h incubation of the cells with DHEA or 6-AN, a significant decrease in GSH was observed in both cell types (Figure 5C).

We attempted to induce doxorubicin resistance in HT29 cells by enhancing their GSH intracellular levels. After a 3 h incubation of HT29 cells with 30 mmol/l GSH, the intracellular GSH in HT29 cells was increased significantly, overcoming the level of GSH in HT29-DX cells in the absence of exogenous doxorubicin (Figure 2). In GSH-enriched HT29 cells doxorubicin caused a slight decrease in GSH, whose content was still significantly higher than in the control cells. The HT29-DX cells did not show a change in the level of GSH after GSH loading. The enrichment of GSH caused a significant decrease of the accumulation of intracellular doxorubicin in HT29 cells, mimicking the behaviour of the HT29-DX cells. This effect was absent in A549 cells, a human epithelial pulmonary cancer cell line that accumulates doxorubicin but is not able to uptake extracellular GSH [24] (Figure 2). As expected, no difference in doxorubicin accumulation was observed in HT29-DX cells, in keeping with the lack of response to GSH loading.

Finally, to demonstrate a correlation between MDR and cellular GSH content we investigated whether MDR could be induced in HT29 cells by overexpression of the G6PD gene. Interestingly, HT29 cells transfected with G6PD showed a significant increase in the expression of G6PD mRNA, protein (Figure 6A) and enzymatic activity (Figure 6B), whereas HT29 cells infected with the retrovirus without G6PD showed a behaviour similar to the control cells. The enhanced enzymatic activity was significantly reduced with the addition of DHEA or 6-AN (Figure 6B). Moreover, the up-regulation of G6PD induced a significant increase of intracellular GSH levels in HT29 cells, similar to that observed in non-transfected HT29-DX cells (Figure 7A). GSH decreased significantly when cells were incubated for 24 h with DHEA or 6-AN (Figure 7A). After a 3 h incubation with 5 μmol/l doxorubicin the G6PD-transfected HT29 cells showed a drug-resistant phenotype, assessed as reduced accumulation of intracellular drug, which was significantly reversed when cells were incubated with DHEA or 6-AN (Figure 7B). However, although G6PD-transfected HT29 cells showed similar cellular GSH levels when compared with resistant cells, they accumulated more doxorubicin than HT29-DX cells, thus suggesting that the GSH content is unlikely to be the only cause of drug resistance and that other molecular mechanisms are involved. In contrast, when we measured the extracellular levels of GSH and the intracellular/extracellular levels of GSSG we did not observe any significant change (results not shown). This means, in our opinion, that glutathione interacts in a way that is not yet understood with the drug and/or a transporter. However, no difference in the expression of MRP1, MRP2 or RLIP76 proteins (results not shown) was shown between HT29 cells transfected with G6PD or with empty virus alone, suggesting that the modified expression of transporters is not the main factor responsible for the MDR.

Figure 6 G6PD expression and activity in HT29-transfected cells

(A) Relative expression of G6PD mRNA and protein in HT29, HT29-DX, HT29+empty virus and HT29+virus containing the G6PD gene. Left-hand panel: results are means±S.E.M. obtained from three different experiments. Versus HT29: *P<0.02; **P<0.0001. Versus HT29-DX: °P<0.0001. Versus HT29+empty virus: §P<0.0001. Right-hand panel: the Figure is representative of three experiments giving similar results. Versus HT29: *P<0.002. Versus HT29+empty virus: §P<0.002. (B) Basal G6PD activity and 6PGD activity and effect of either DHEA or 6-AN on enzymatic activities in HT29+empty virus and HT29+virus containing the G6PD gene. Cells were incubated for 24 h in the absence (CTRL) or presence of either DHEA (100, 250 or 500 μmol/l) or 6-AN (100, 250 or 500 μmol/l) (n=3). Versus HT29+empty virus CTRL: *P<0.05; **P<0.0001. Versus HT29+G6PD-virus CTRL: °P<0.0001.

Figure 7 GSH levels and doxorubicin accumulation in HT29-transfected cells

(A) Intracellular GSH levels and effect of either DHEA or 6-AN in HT29+empty virus and HT29+virus containing the G6PD gene. Cells were incubated for 24 h in the absence (CTRL) or presence of either DHEA (100, 250 or 500 μmol/l) or 6-AN (100, 250 or 500 μmol/l) (n=3). The values in HT29 and HT29-DX cells were reported as a comparison. Versus HT29+empty virus CTRL: *P<0.05; **P<0.001; ***P<0.0001. Versus HT29+G6PD-virus CTRL: °P<0.005; °°P<0.001; °°°P<0.0001. (B) Cells were incubated in the absence (CTRL) or presence of DHEA (100, 250 or 500 μmol/l) or 6-AN (100, 250 or 500 μmol/l) for 24 h. During the last 3 h of incubation 5 μmol/l doxorubicin was added (n=3). The values in HT29 and HT29-DX cells were reported as a comparison. Versus HT29+empty virus CTRL: *P<0.05; **P<0.001; ***P<0.0001. Versus HT29+G6PD-virus CTRL: °P<0.0001.

DISCUSSION

Doxorubicin is commonly used in the therapy of solid tumours. As for many other chemotherapeutic agents, the clinical use of doxorubicin is hampered by several side effects including MDR and cardiotoxicity [25], which impair the clinical response and the survival of patients [26]. Making cancer cells more sensitive to doxorubicin would be an efficient approach to decrease the overall dose of anthracyclines infused and increase the ratio between therapeutic effect and toxicity. In the present study, we used a cell line derived from human colon cancer, one of the most frequent solid tumours that show MDR and from these cells we obtained a model of doxorubicin-resistant cells (HT29-DX) as described previously [10].

MDR is regarded as a multifactorial process [1], but several studies indicate that a cardinal mechanism of resistance is an increased rate of removal of the drug out of the cell [2]. Several transporters such as MRP1, MRP2 and RLIP76 require GSH for the transport activity [2729]. The results of the present study, showing that the G6PD transfection does not modify the MRP1, MRP2 and RLIP76 (results not shown) expression in HT29 cells, suggest that the different expression of redox-modulated transporters is not responsible for MDR in these cells.

The results of the present study show that HT29-DX cells, in comparison with their drug-sensitive counterpart (HT29 cells), have a higher level of intracellular GSH, confirming previous data obtained in other MDR tumour cells [26,30]. Cellular thiols could inactivate chemotherapeutics before they reach their target and enhance DNA repair by providing a reducing environment [31], but it is well known that GSH may also be implicated in the glutathionation of chemotherapeutic agents or as a cofactor in the transport of unconjugated drugs [9].

High cellular GSH concentrations may sometimes be related to an increased expression of γ-GCS (γ-glutamylcysteine synthase), an increased level of which has been observed in MDR tumours and cell lines. Exposure of cells to a specific inhibitor of γ-GCS, buthionine sulfoximine, decreased MDR for doxorubicin and vincristine [32] by reducing cellular GSH levels [33], whereas the exposure of these cells to GSH ethyl ester increased the GSH levels and restored the resistance to these drugs [33].

An alternative and less investigated way to increase the level of GSH is to activate the PPP flux via the activation of G6PD. The PPP flux is one of the metabolic pathways most sensitive to cellular oxidative stress, via G6PD activation, and leads to the synthesis of NADPH that is necessary to keep GSH in the reduced form [34]. Menadione-stimulated PPP activity is a useful index of the maximal activation of the pathway and therefore of its antioxidant potential. Several studies have show that doxorubicin is able to induce oxidant species production [35], promoting the oxidation of GSH to GSSG which induced the production of free radicals and H2O2 [36,37]. We observed that an acute administration of doxorubicin in HT29 cells induces a significant decrease in GSH level and an increased G6PD activation (results not shown), but a prolonged exposure to low concentrations of doxorubicin produces a persistent oxidative stress that could lead to enhanced content of GSH and increased G6PD expression and activity. We observed that HT29-DX cells exposed continuously to doxorubicin show a significantly higher (basal and menadione-stimulated) activity of the PPP when compared with HT29 cells. Since the PPP flux is regulated by the activity of G6PD, this metabolic pathway can be considered a sensitive index of the actual G6PD activity in a whole cell. The results of the present study also showed that HT29-DX cells exhibit a significantly increased G6PD expression, in terms of both mRNA and protein, and a significantly higher G6PD enzymatic activity if compared with doxorubicin-sensitive cells. Although substantial evidence has been provided that oxidative stress plays a role in MDR, a lot of controversial data are still present, probably due to the different cell types investigated. A decreased G6PD activity and PPP flux were observed in adriamycin-resistant MCF-7 cells [38]. An increased G6PD activity in resistant cancer cells, but no correlation between GSH level or G6PD activity and drug resistance, was detected by Yu et al. [39]. Anthracycline-resistant HL60/AR cells showed down-regulation of cellular GSH [40], whereas HEK-293 anthracycline-sensitive and -resistant cells exhibited similar levels of intracellular GSH [39]. In contrast, several studies demonstrated an elevated PPP cycle, G6PD activity and cellular GSH in human MDR cells [30,4144].

In the present study, for the first time to our knowledge, the non-competitive G6PD inhibitor DHEA and the competitive inhibitor 6-AN were used to down-regulate PPP and to see their effect on the MDR phenotype. The inhibition of PPP flux and G6PD by DHEA and 6-AN was more evident in HT29-DX cells, probably because these metabolic activities were higher than in HT29 cells.

The lower ability to activate the PPP in response to oxidative stress was probably responsible for a parallel consumption of GSH, which was more evident in HT29-DX cells. DHEA and 6-AN induced a significant decrease of GSH in HT29-DX cells, reaching a level similar to the one found in sensitive cells. Doxorubicin sensitivity is known to increase in parallel with GSH depletion [45]. The increase of drug accumulation and cytotoxic effect (evaluated as leakage of LDH activity and amount of annexin V-positive cells) suggests that the MDR phenotype was reversed by DHEA and 6-AN. The treatment with DHEA did not significantly change the accumulation of doxorubicin in Pgp-silenced HT29-DX cells, thus suggesting that Pgp is not implicated in the GSH-dependent cellular efflux of doxorubicin.

To confirm our hypothesis of a linkage between GSH and drug resistance, we induced a MDR phenotype in sensitive HT29 cells by enhancing their intracellular GSH levels or by inducing overexpression of G6PD. In GSH-enriched HT29 cells we observed a decreased accumulation of doxorubicin. Moreover, a human epithelial pulmonary cancer cell line (A549) which is unable to uptake extracellular GSH [24] did not show MDR acquisition. HT29 cells transfected with the G6PD gene showed elevated expression of G6PD mRNA, protein and enzymatic activity, a significant decrease of intracellular doxorubicin and a significant increase of intracellular GSH. Moreover, both DHEA and 6-AN incubation decreased significantly the G6PD enzymatic activity and GSH levels reverting the induction of resistance phenotype in G6PD-transduced HT29 cells, thus confirming the involvement of G6PD in the onset of doxorubicin resistance.

On the basis of the results of the present study, we suggest that the improving effects of G6PD transfection or GSH addition on the ability of HT29 cells to extrude cellular GSH conjugates is not due to a modified expression of MRP1, MRP2 or RLIP76, but rather to the increased availability of intracellular GSH, which then activates one or more of these transporters. The higher content of MRP1 and MRP2 in HT29-DX cells, although not playing a primary role, may, however, be important in enhancing the drug extrusion.

In summary, in agreement with other results reported previously regarding the role of redox metabolism in MDR [7], our results confirm that the activation of intracellular PPP may be one of the causes of acquired MDR phenotype, suggesting a further mechanism for GSH increase and its effects on MDR, i.e. the hyperactivity of the PPP and of its rate-limiting enzyme G6PD. Finding new inhibitors of G6PD, more active at lower concentrations than those used for DHEA and 6-AN, could allow a reduction in the overall administered dose of doxorubicin and an increase in the balance between therapeutic benefits and side effects. Moreover, now our intention is to better understand the molecular mechanism leading to altered cellular redox metabolism in HT29-DX cells. The knowledge of this molecular pathway could allow the improvement of chemotherapy by decreasing drug toxicity, increasing therapeutic effects and preventing the up-modulation of G6PD and/or protein pumps such as MRPs in cancer cells.

AUTHOR CONTRIBUTION

Manuela Polimeni, Amalia Bosia and Dario Ghigo designed the study; Manuela Polimeni, Claudia Voena and Joanna Kopecka performed experiments; Amalia Bosia and Chiara Riganti analysed and interpreted data; Manuela Polimeni, Claudia Voena and Dario Ghigo wrote the paper; Gianpiero Pescarmona, Amalia Bosia and Dario Ghigo supervised the study.

FUNDING

This work was supported by the Ministero dell'Università e della Ricerca [grant number 2007498XRF] and by the Regione Piemonte (to D.G.).

Abbreviations: 6-AN, 6-aminonicotinamide; DHEA, dehydroepiandrosterone; FBS, fetal bovine serum; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; G6PD, glucose-6-phosphate dehydrogenase; γ-GCS, γ-glutamylcysteine synthase; GSH, glutathione; GSSG, glutathione disulfide; HEK, human embryonic kidney; LDH, lactate dehydrogenase; MDR, multidrug resistance; MRP, MDR-associated protein; 6PGD, 6-phosphogluconate dehydrogenase; Pgp, glycoprotein P; PPP, pentose phosphate pathway; qRT-PCR, quantitative real-time PCR; RLIP76, Ral-interacting protein 76; siRNA, small interfering RNA

References

View Abstract