The FASEB Journal express article 10.1096/fj.02-0105fje. Published online November 1, 2002.

Glutathione disulfide induces apoptosis in U937 cells by a redox-mediated p38 mitogen-activated protein kinase pathway Giuseppe Filomeni,* Giuseppe Rotilio,* and Maria Rosa Ciriolo� *Department of Biology, University of Rome �Tor Vergata,� Rome, Italy; �Department of Biomedical Sciences, University of Chieti �G. D�Annunzio,� Chieti, Italy Corresponding author: Maria Rosa Ciriolo, Department of Biomedical Sciences, University of Chieti �G. D�Annunzio,� Via dei Vestini 66013, Chieti, Italy. E-mail: ciriolo@bio.uniroma2.it ABSTRACT Changes in the intracellular reduced/oxidized glutathione ratio (GSH/GSSG) are crucial reduction-oxidation (redox) events that trigger downstream proliferation or death responses. We investigated the molecular mechanisms underlying redox-mediated cell signaling upon an oxidative insult by treating U937 cells with exogenous nonpermeable GSSG. This treatment results in a significant decrease of exofacial cell membrane thiol groups and intracellular decrement of GSH content, owing to its engagement in the formation of mixed disulfides. Changes in thioredoxin redox state were also observed, and they may be related to the activation of upstream ASK1 and selective induction of downstream p38 mitogen-activated protein kinase (MAPK) pathway, detectable by phosphorylation of MKK3/6 and p38 MAPK. Moreover, an increase in reactive oxygen species production was detected, and cells were committed to apoptosis along the mitochondrial pathway, evidenced by Bcl-2 down-regulation, cytochome c release from mitochondria, caspase-9 cleavage, and caspase-3 activation. GSH ethyl ester, a precursor of GSH, by counteracting intracellular mixed disulfide formation, canceled both p38 MAPK activation and GSSG-mediated apoptosis via inhibition of thioredoxin oxidation and stabilization of thioredoxin/ASK1 complex, whereas, blockage of p38 MAPK by specific inhibitor SB 203580 allowed apoptosis at a very reduced extent. Results suggest that kinase cascade may serve as a primary transducer of cytoplasmic oxidative signals to the nucleus before apoptosis-inducing signals are activated. Key words: redox signaling • thioredoxin • MAPK • GSH redox state • mixed disulfide

T

he reduction-oxidation (redox) state of the cell is a consequence of the balance between the levels of oxidizing and reducing equivalents. A reducing intracellular environment is fundamental for cell survival; however, redox unbalance is necessary because it represents a

regulatory sensor for several nuclear transcription factors (1). In fact, the reduced/oxidized ratio provides a means of dynamically regulating intracellular redox state and, at the same time, it is suggested as one mechanism by which cells may transduce oxidative stress into the inducible expression of a wide variety of genes implicated in cellular processes such as growth, differentiation, and death. It has been suggested that several factors involved in stress response, such as mitogen-activated protein kinases (MAPKs) and transcription factors, represent downstream modulators of an altered intracellular redox state (2�4). Moreover, many cellular receptors, such as tumor necrosis factor receptors (TNF-R) and CD95/Fas, or protein tyrosine kinases involved in

signal transduction are characterized by several cysteine-rich regions, and it has been shown that they are susceptible to redox modifications (5�7). Redox chemistry involves electron exchanges between molecules that display several possible oxidation states. Within the cellular context, the redox status depends on the relative amounts of the oxidized and reduced partners of the major redox molecules, such as glutathione, which represents the most prevalent non protein intracellular thiol with several fundamental functions ranging from detoxification to messenger molecule (8). Glutathione exists as a reduced predominant form (GSH), as a disulfide form (GSSG), or as mixed disulfide (GSSR) with protein thiols (9). The ratio GSSG/GSH reflects the redox status; thus, the oxidation of a limited amount of GSH into GSSG can drastically change this ratio, greatly affecting the redox status within the cell. GSSG and mixed disulfides are continuously reduced back through the reaction of the NADPH-dependent glutathione disulfide reductase and by the catalysis of cellular enzymes such as glutaredoxin or thioredoxin, respectively (9, 10). The latter process can represent a first line of defense against oxidative stress that, instead, leads to irreversible protein-thiol oxidation of reactive cysteines, resulting in protein structure and function alteration. Indeed, recent reports demonstrate that the process of glutathione-S-thiolation is involved in the regulation of protein activity during free radical-mediated oxidative reactions (2, 11�13). Thioredoxin is another important component among the redox couples that maintain intracellular redox environment (14). By recycling between the oxidized and reduced forms, which differ in the presence of a disulfide bridge between two vicinal cysteines (C32, C35), thioredoxin governs several functions inside the cell (15). In particular, its activity is necessary for the correct protein folding inside the endoplasmic reticulum and for the reduction of GSH mixed disulfide with protein thiols. It is also directly involved in redox regulation of gene expression either by promoting DNA binding of transcription factors such as NF-κB, AP-1, and p53 or by activating MAPK pathway (16). Therefore, the glutathione/thioredoxin system acts as a homeostatic redox buffer; this is an important cellular parameter, because the intracellular redox status reflects the relative amounts of the oxidized and reduced species of each redox system. Malfunctions in the control of this redox balance may lead to cancer, apoptosis, or neurodegenerative diseases (17�19). Among these, programmed cell death represents one of the most assessed redox-regulated processes; in fact, it has been extensively demonstrated that the molecular mechanisms underlying this process are associated with oxidative stress in which GSH depletion, rather than reactive oxygen species (ROS) increase, has a central role (20, 21). From a mechanistic point of view, the primary target of redox regulation can be the sulphydryl group of cysteine residues, which, depending on the surrounding environment, could be easily oxidized to form not only an S-glutathiolated adduct, but also intra- and/or intermolecular disulfide bonds. Such changes have been demonstrated to occur in different redox-sensitive proteins, leading to both �gain� or �loss� of function. In particular, whereas intramolecular disulfide bond formation results in alteration of protein structure and conformation, the intermolecular disulfide bridge could lead to dimer and/or multimer formation. The latter process may represent a regulatory pathway that potentially influences the association between different cellular proteins (22). Among proteins that are regulated in their function upon formation of a disulfide bond, thioredoxin and ASK1 represent well-established examples. Studies, performed using the yeast two-hybrid system, identified

thioredoxin as an ASK1 interacting regulatory protein (23). The thioredoxin/ASK1 association gives rise to the inactivation of kinase activity of ASK1; moreover, this interaction is found only in nonstressed cells and seems to be modulated by intracellular ROS levels. In particular, an increase in ROS concentration causes disulfide bridge formation between thioredoxin molecules, and, as a result, ASK1 can undergo multimerization, which corresponds to the active form of the enzyme (24). ASK1 has been suggested to activate p38- and JNK-upstream kinases, MKK3/6 and MKK4/7, respectively (25). Previous observations indicate that free intracellular GSH content can be modulated by the external redox milieu via a thiol/disulfide exchange mechanism that sequentially involves sulfur-rich proteins spanning across plasma membrane (26, 27). In this study, we examined the effects of GSSG, which cannot cross the cell membrane, on intracellular redox state and survival of U937 promonocytic cells, exploring the underlying mechanisms of its effects. Our data demonstrate that GSSG is able to alter the membrane thiols redox state, thus shifting the intracellular redox homeostasis toward oxidizing conditions and giving rise to a decrease in GSH content in concomitance with a significant increase in the concentration of mixed disulfides. The resulting oxidized environment causes cells to trigger programmed death by activating a p38 MAPK-dependent mechanism via ASK1 modulation. p38 MAPK acts at an early step before dysfunction of mitochondria, caspase activation, and ROS production. These findings suggest a potential pharmacological role for GSSG in regulating cell fate in pathological settings. MATERIALS AND METHODS CHAPS, 3-[cholamindopropyl dimethylammonium]-1-propanesulfonate; D-mannitol; dimethyl sulfoxide (DMSO); dithiothreitol (DTT); HEPES, (N-[2-hydroxyethyl] piperazine-N'-[2-ethanesulfonic acid]); EGTA, ethyleneglycol-bis-(β-amino ethyl ether) N,N,N',N'-tetracetic acid; IGEPAL CA-630, (octyphenoxy)polyethoxyethanol; PMSF, phenylmethylsulfonyl fluoride; PIPES, piperazine-N,N'-bis[2-ethanesulfonic acid] disodium salt; potassium borohydride (KBH4); propidium iodide; protease inhibitors cocktail; tert-butyl hydroperoxide; protein-A; acivicin; reduced glutathione ethyl ester (GSHest); anti-Bcl-2; clone Bcl-2-100; and anti-actin, clone AC-40, monoclonal antibodies were obtained from Sigma (St. Louis, MO). Ac-DEVD-AFC, caspase-3 inhibitor I (Ac-DEVD-CHO); p38 MAPK inhibitor SB 203580, [4-(4-fluorophenyl)-2-(4-methylsulfinylphenyl)-5-(4-pyridyl)-1H-imidazole]; and Hoechst 33342 were purchased from Calbiochem-Novabiochem (La Jolla, CA). Anti-cytochrome c monoclonal antibody, clone 7H8.2c12, was from PharMingen (San Diego, CA). Anti-p-38 (c-20) polyclonal antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). Polyclonal antibodies detecting activation of p38 MAPK pathway, anti-phospho-MKK3/6 (Ser189/207); anti-phospho-p38 MAP kinase (Th180/Tyr182); and anti-rabbit IgG, HP-linked were from Cell Signaling Technology-New England BioLabs (Beverly, MA). Anti-caspase-9 monoclonal antibody, clone 96-2-22, was from Upstate Biotechnology (Lake Placid, NY). Anti-thioredoxin polyclonal antibody was kindly provided by Dr. Roberto Sitia and Dr. Elena Pasqualetto from DiBiT-HSR and Università Vita-Salute San Raffaele (Milan, Italy). Anti-ASK1 (ASK1 1151-ab2004) polyclonal antibody was from Abcam (Cambridge, UK). Goat anti-mouse and anti-rabbit IgG (H+L)-horseradish peroxidase conjugate was from Bio-Rad (Hercules, CA). SuperSignal substrate chemiluminescent reagent was from Pierce (Rockford, IL). 2',7'-dichlorofluorescein diacetate (DCF-DA) and the nonpermeable thiol reactive compound Alexa fluor C5 maleimide were from Molecular Probes (Eugene, OR).

MTS assay kit CellTiter 96 AQueous One Solution Cell Proliferation Assay was purchased from Promega (Madison, WI). Reduced and oxidized glutathione were obtained from Roche Molecular Biochemicals (Monza, Italy). Annexin V-FITC kit was from Bender MedSystems (Vienna, Austria). All other chemicals were obtained from Merck (Darmstadt, Germany). Cell culture Human promonocytic cell line U937 was purchased from American Type Culture Collection (Rockville, MD) and grown in RPMI 1640 medium supplemented with 10% fetal bovine serum at 37°C in an atmosphere containing 5% CO2. Cells were routinely collected by centrifugation at 700g and resuspended in fresh medium at a concentration of 2 × 105/ml. Cell viability was assessed by Trypan Blue exclusion. Peripheral blood mononuclear cells were isolated from human buffy coat blood preparations by centrifugation on ficoll-hypaque and suspended at 2.5 × 106/ml in RPMI 1640 supplemented with 10% fetal bovine serum, 5 mM L-glutamine, and Gentamicine (5 µg/ml). After incubation for 1 h in 10 ml of complete medium at 37°C, nonadherent cells were removed by washing with RPMI 1640. Cells were analyzed by flow cytometry by CD14 staining and morphological parameters (forward scatter vs. side scatter) to assess monocyte purity, which was always >85%. Nonadherent cells were washed away, and adherent monocytes were collected by gently detaching them with a cell scraper after 15 min incubation with 5 mM EDTA in phosphate-buffered saline (PBS) at 4°C. Treatments Glutathione disulfide solution was prepared dissolving GSSG powder in PBS (10 mM phosphate buffer, 2.7 mM KCl, 137 mM NaCl, pH 7.4), buffered at pH 5.5 with NaOH and sterilized by filtration. Treatments were performed with different amounts of GSSG ranging from 0.25 to 2.5 mM for different times at 37°C in medium supplemented with serum. For detection of intracellular ROS, cells were incubated with 50 µM DCF-DA (dissolved in dimethyl sulfoxide) for 30 min at 37°C, pelletted, washed, and resuspended in ice-cold PBS. The fluorescence intensities of 2',7'-dichlorofluorescein, formed by the reaction of DCF with ROS, of more than 10,000 cells from each sample were analyzed by recording FL-1 fluorescence with a flow cytometer. Before data collection, propidium iodide was added to the samples for gating out dead cells. Treatment with 100 µM tert-butylhydroperoxyde was used as a positive control. Experiments were repeated at least thee times with similar results. The data are given as one representative histogram. For detection of external plasma membrane thiols, cells were washed, suspended in RPMI, and incubated with 10 µM Alexa fluor C5 maleimide for 1 h at 37°C. Cells were pelleted, washed, and resuspended in PBS. Labeled thiols were detected either by cytofluorimetric analyses or fluorescence microscope. Images of cells were rapidly digitized with a Cool Snap video camera connected to Nikon Eclipse TE200 epifluorescence microscopy. All images were captured under constant exposure time, gain, and offset.

Acivicin, specific inhibitor of γ-glutamyl transpeptidase (γ-GT), was added at a concentration of 50 µM for 1 h before and throughout the treatment with GSSG. Treatment with GSH ethyl ester (GSHest) was performed at a concentration of 1 mM for 1 h before and during GSSG supplementation. SB 203580, a specific p38 MAPK inhibitor, was dissolved in dimethyl sulfoxide and used at a concentration of 5 µM for 2 h before and during GSSG treatment. Analysis of cell viabilty and apoptosis Cell growth was assessed by a CellTiter 96 AQueous One Solution Cell Proliferation Assay kit. This reagent contains a novel tetrazolium compound (MTS) and an electron-coupling reagent (PES). MTS is bioreduced by viable cells into a colored formazan product that is soluble in cell culture medium. At different times, aliquots of 200 µl of treated and control cells were withdrawn and put into a 96-well plate. Each assay was done in triplicate. Analyses were performed by adding CellTiter 96 solution (1:10 v/v) to culture wells, incubating for 2 h, and then recording the absorbance at 490 nm against appropriate blanks made with culture medium with or without GSSG. Increase in absorbance was considered proportional to cell growth. Data were expressed as arbitrary units. Nuclear fragmentation was detected by optical microscopy on slides of hematoxylin-stained cells or by fluorescent microscopy, using vital staining assay with DNA-specific cell-permeable dye Hoechst 33342. Images of cells were rapidly digitized with a Cool Snap video camera connected to Nikon Eclipse TE200 epifluorescence microscopy. All images were captured under constant exposure time, gain, and offset. Blebbing cells were detected by phase contrast microscopy as the modification of cell shape from nearly round to blackberry-like. The fraction of cells with fragmented nuclei among the total cell population was calculated using Hoechst 33342-stained cells, counting at least 300 cells in at least 10 random-selected fields as previously described (21). Alternatively, cells were washed with PBS, stained with an annexin V-FITC/propidium iodide kit, and analyzed by a FACScan instrument (Becton Dickinson, San Jose, CA). Preparation of cell lysates and Western blot analyses Cells were washed with PBS and collected by centrifugation at 700g for 7 min at 4°C. The cell pellet was resuspended in extraction buffer containing 220 mM mannitol, 68 mM sucrose, 50 mM PIPES-KOH (pH 7.4), 50 mM KCl, 5 mM EGTA, 2 mM MgCl2, 1 mM dithiotheitol (DTT), and protease inhibitors (28). After 30 min incubation on ice, cells were homogenized with a glass Dounce. Cell homogenates were spun at 14,000g for 15 min at 4°C; supernatants and pellets were collected and stored at �80°C until analysis by gel electrophoresis; and 20 µg of cytosolic protein and 10 µg of pellet extracts were loaded onto each lane of a 12% sodium dodecyl sulfate (SDS)-polyacrylamide gel, separated, and then blotted to nitrocellulose membrane (Bio-Rad). Purified mouse anti-cytochome c, anti-caspase-9, and anti-β-actin monoclonal antibodies were used as primary antibodies (1:5000). The specific protein complex, formed upon anti-mouse secondary antibody treatment (1:10,000), was identified using SuperSignal substrate chemiluminescence reagent. Cell pellet was resuspended in lysis buffer containing 10 mM Tris-HCl, pH 7.4, 5 mM EDTA, 150 mM NaCl, 0.5% IGEPAL CA-630, and protease inhibitors. After 30 min incubation on ice, cells

were disrupted by 10 s of sonication. Lysates were then centrifuged at 14,000g for 15 min at 4°C, and supernatants were removed and stored at �80°C. Proteins (20 µg for Bcl-2 or 50 µg for thioredoxin, p38, and β-actin) were loaded on 12% polyacrylamide gel and transferred onto a PVDF membrane (Millipore, Bedford, MA). Monoclonal anti-Bcl-2 and β-actin (1:5000) or polyclonal anti-thioredoxin (1:200) and anti-p-38 (1:500) were used as primary antibodies. For Western blot analyses of thioredoxin redox state, DTT was dissolved in the SDS-containing sample buffer at a concentration of 40 mM. Densitometry analyses of proteins were performed using the Software Gel Pro Analyzer, and protein levels were normalized to the density of related actin bands. Cell pellet was resuspended in lysis buffer containing 62.5 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, 50 mM DTT, 0.01% bromophenol blue and was sonicated for 15 s to shear DNA and reduce sample viscosity. Sample (20 µl) was loaded on 12% polyacrylamide gel and transferred onto a nitrocellulose membrane (Bio-Rad). Polyclonal anti-phospho proteins of p38 MAPK pathway (1:200) were used as primary antibodies. After incubation with anti-rabbit IgG, HP-linked (from Cell Signaling), specific proteins were identified using SuperSignal chemiluminescence substrate. Cell pellet was resuspended in lysis buffer containing 10 mM Tris-HCl (pH 7.4), 5 mM EDTA, 150 mM NaCl, 0.5% IGEPAL CA-630, and protease inhibitors. After 30 min incubation on ice, cells were disrupted by 10 s of sonication. Lysates were then centrifuged at 14,000g for 15 min at 4°C. Proteins (400 µg) were incubated in lysis buffer with 10 µl of anti-ASK1 antibody to a total volume of 200 µl for 2 h at 4°C. Immunocomplexes were adsorbed with 10 µl of protein A-sepharose for 30 min at 4°C. After three washes with lysis buffer containing 0.5% IGEPAL CA-630, immune pellets were boiled in SDS sample buffer. Proteins (50 µg) were loaded on 15% SDS-polyacrylamide gel and transferred to PVDF. Polyclonal anti-thioredoxin antibody was used as the primary antibody (1:200). The specific protein complex, formed upon anti-rabbit secondary antibody treatment (1:10,000), was identified using SuperSignal substrate chemiluminescence reagent. Glutathione determination and other assays Intracellular glutathione was assayed upon formation of S-carboxymethyl derivatives of free thiols with iodoacetic acid, followed by the conversion of free amino groups to 2,4-dinitrophenyl derivatives by the reaction with 1-fluoro-2,4-dinitrobenzene as described by Reed et al. (29). Cells were washed three times with PBS in order to completely eliminate exogenous GSSG, resuspended, and lysed by repeated cycles of freezing and thawing under liquid nitrogen. Lysates were then divided into two aliquots: one was utilized for GSH/GSSG assay, and the other was used to determine mixed disulfides by treating with 300 mM KBH4 at 40°C for 30 min before derivatization, as described previously (30). Data are expressed as nanomoles of GSH equivalents per milligrams of protein. Cells were washed and collected by centrifugation at 700g for 7 min at 4°C. Cell pellets were resuspended in lysis buffer containing 100 mM Hepes (pH 7.5) 0.1% CHAPS, 1 mM PMSF, 10 mM DTT, and 1 mM EDTA and put on ice for 30 min. Following sonication, lysates were centrifuged at 10,000g for 10 min at 4°C. Supernatants were used for caspase-3 assay in lysis buffer containing the specific substrate Ac-DEVD-AFC (5 µM) at 30°C. Substrate cleavage was monitored fluorometrically using a Perkin Elmer (Monza, Italy) luminescence spectrometer LS-5

equipped with a chart recorder model R 100A, with excitation and emission wavelengths of 400 and 505 nm, respectively. Specific activity was demonstrated by the use of caspase-3 inhibitor I (Ac-DEVD-CHO) that completely impedes the development of fluorescence due to cleaved product. Proteins were determined by the method of Lowry et al. (31). All experiments were done with n = 6. Each assay was performed at least twice under identical conditions, unless otherwise indicated. Data are expressed as means ±SD, and significance was assessed by Student�s t test for unpaired data. P<0.05 was considered significant. RESULTS GSSG commits U937 cells to apoptosis Although GSSG is a component of the most important redox couple, GSH/GSSG, the role of this molecule in signal transduction pathways has not yet been fully investigated because the intracellular concentration of this molecule is low, due to its efficient reduction to GSH by the activity of the NADPH-dependent GSSG-reductase. However, under oxidative burst, GSSG content can be raised, and it is extruded from the cell into the extracellular milieu, becoming a pro-oxidizing molecule. To analyze the pro-oxidant role of GSSG, we utilized the promonocytic cell line U937. By using a cell growth and viability assay (MTS assay), we showed that GSSG treatment of U937 cells led to a dose-dependent decrease in their growth, significant as early as after 24 h even with the lowest concentration used (P<0.05; n=6) (Fig. 1A). To verify that GSSG affects cell growth and not the metabolic rate of the cell, direct cell counting was also done using the vital dye Trypan Blue. As shown in Figure 1B, GSSG treatment of U937 cells caused loss of viability that paralleled the results obtained with the MTS assay. Furthermore, prolonged treatment with the lowest concentration of GSSG (0.25 mM) resulted in 60% loss of viability upon 3 days of culture. Based on these results, we decided to use the concentration of 1 mM for the experiments hereafter, unless otherwise specified. Microscopic examination of U937 cells after treatment with 1 mM GSSG revealed that a substantial fraction of the cells showed membrane blebbing, a typical morphological change associated with apoptotic cells (data not shown). Further analyses, performed by vital staining with the DNA-specific cell-permeant dye Hoechst 33342, shows that U937 cells treated with 1 mM GSSG have the characteristic features of apoptotic nuclei (Fig. 1C, upper panels); moreover, the amount of cells with fragmented nuclei increased in a time-dependent manner. Cytofluorimetric analyses of U937 cells were also made after double staining with annexin V-FITC/propidium iodide. Annexin V binds with high affinity to phosphatidylserine of early apoptotic cell membrane surface, and propidium iodide binds to DNA, being a valid marker for late apoptotic or necrotic cells. Figure 1C (bottom panels) shows a time-dependent increase in annexin V-FITC-positive cells upon treatment with 1 mM GSSG; after 24 h and 48 h, the cells positive only to annexin V-FITC were 10 ± 2% and 17 ± 3%, respectively; and 15 ± 3% and 45 ± 5% were the percentages of cells positive to both annexin V-FITC and propidium iodide, indicating cells undergoing late apoptosis/secondary necrosis. To examine the molecular factors involved in the apoptotic process, cytosolic extracts were prepared under conditions that keep mitochondria intact, and cytosolic cytochome c protein level

was measured by immunoblot analysis. The cytosol of untreated U937 cells contained no detectable amounts of cytochome c, whereas it accumulated upon 1 mM GSSG treatment (Fig. 2A), being detectable at 18 h, with a concomitant decrease of the mitochondrial-associated cytochome c (Fig. 2A). Hallmarks of the apoptotic process are the activation of cysteine-proteases (caspases), which represent both initiators and executors of death signals. Activation of caspase-9, an initiator factor, was shown by Western blot analysis, following the appearance of the active proteolyzed caspase band. Figure 2B shows that active caspase-9 was not present in untreated cells in which a quantitatively high level of pro-caspase-9 was detected, whereas it accumulated in a time-dependent manner in U937 cells upon treatment with 1 mM GSSG. Moreover, proteolysed caspase-9 started to accumulate at 18 h, concomitantly with cytochome c release, confirming the role played by these factors as upstream effectors of apoptosis. Proteolytic activity was also measured by testing the cytosolic extracts for their ability to cleave the fluorimetric substrate Ac-DEVD-AFC, which is specific for caspase-3. Active caspase-3 has a pivotal role in the cascade of proteolytic events in that it is considered one of the points of no return in the process leading to cell destruction. The activity of this caspase was significantly different from untreated cells as early as 12 h after treatment with 2.5 mM GSSG, whereas with 1 mM GSSG, the activation was significant after 24 h (Fig. 2C). GSSG alters the intracellular redox buffer system GSSG per se is a nonpermeant pro-oxidizing molecule, thus its effects must be operating from outside the cell membrane. Given our previous observation that reducing/oxidizing impermeant agents could affect intracellular redox state by translocation of their equivalents from external to intracellular space (27), we wanted to test whether GSSG could affect both the thiol redox state of cell membrane and the intracellular redox environment. To this end, upon 12-h treatment with 1 mM GSSG, cells were labeled with 10 µM Alexa fluor C5 maleimide, a specific nonpermeable thiol reactive compound able to covalently bind to membrane surface cysteines, as described in Materials and Methods. Figure 3A (upper panel) shows that GSSG-treated cells were significantly less fluorescent with respect to controls, indicating that reduced thiol groups on the membrane surface were not available for binding with Alexa fluor C5 maleimide. Cytofluorimetric analyses were also performed, and, as shown in Figure 3A (bottom panel), a significant decrease in the green fluorescence, upon GSSG treatment, was detected, thus confirming that GSSG was able to efficiently oxidize extracellular cysteines. We next evaluated whether GSSG induced intracellular oxidative changes by analyzing the intracellular redox buffer in terms of GSH, GSSG, and mixed disulfides. As shown in Figure 3B, significantly lower GSH levels were measured as early as 12 h after treatment, reaching a 95% decrease at 48 h with the highest concentration used. This decrement was not associated with an increase in the corresponding free oxidized form, GSSG (data not shown), but rather to engagement of GSH in the formation of mixed disulfides with intracellular protein thiols (Fig. 3C). Moreover, the lack of an increase in GSSG intracellularly confirmed that externally added GSSG was not able to cross the plasma membrane. However, γ-glutamyl transpeptidase (γ-GT), which is involved in the recycling of GSH from the extracellular milieu into the cytosol, could also utilize GSSG, although with a lower affinity. Thus, through γ-GT, dipeptidase activity, and disulfide exchange, cystine could enter the cells, becoming the real effector of the oxidative events observed. Under our experimental conditions, no increase in cysteine or cystine level was detected by HPLC analyses upon GSSG treatment (data not shown).

Furthermore, treatment with the specific inhibitor of γ-GT, acivicin, did not influence GSSG effects. In fact, Figure 3D shows that the decrement in GSH content, after GSSG treatment, was not affected by acivicin; conversely, it was able to significantly increase the amount of intracellular GSSG, both in untreated or GSSG-treated cells (Fig. 3D). These results clearly demonstrate that externally added GSSG per se was able to directly modify the external membrane thiols redox state, and, as a consequence, the intracellular redox homeostasis was shifted toward oxidizing conditions, resulting in an oxidative challenge to cells.

GSSG treatment results in Bcl-2 down-regulation and increased flux of ROS A decrease in GSH content might result in an increase in the flux of ROS (32, 33); we assayed for ROS content in U937 cells treated with GSSG. Cells were loaded with the peroxide- and hydroxyl radical-sensitive fluorophore DCF-DA and assayed by FACS analysis. Figure 4A shows that U937 cells treated with GSSG had markedly higher levels of DCF fluorescence than untreated cells; this increase was significantly different only 12 h after treatment with 1 mM GSSG, and no further increase was detected during the following incubations (data not shown). Note that at this time a significant decrease in GSH content was determined (see Fig. 1), confirming that an increased ROS flux is operative in the cells when the GSH level starts to decline. Bcl-2 protein modulation is responsible for cell survival or suicide, and the function of Bcl-2 has been associated with both oxidative and antioxidative effects in vivo (34, 35). Moreover, it has been demonstrated that Bcl-2 and GSH are strictly related in various cell systems in the maintenance of the correct intracellular redox state and in the commitment of cells to apoptosis (36, 37). Under our experimental conditions, we detected by Western blot analysis a decrease in the Bcl-2 content (�31 ± 4%) as early as 3 h after treatment with 1 mM GSSG, reaching a decrement of 63 ± 6% at 6 h (Fig. 4B). This in turn can be a detrimental factor as far as ROS generation is concerned. In fact, it is well established that the cytosolic steady state of ROS is increased when Bcl-2 is down-regulated (38). Moreover, Figure 4C shows that immunoreactive Bcl-2 protein levels remained lower with respect to control cells either at 12 or 24 h of GSSG treatment, thus suggesting that GSSG-mediated effects gave rise to persistent oxidative conditions that ultimately lead to cell death.

Effects of GSHest on GSSG-mediated toxicity To clarify the role of glutathione depletion in GSSG-mediated toxicity, we treated U937 cells with a compound that can increase GSH content. We used GSHest, a permeant compound that directly increases the GSH levels without the action of γ-GT (39). GSHest (1 mM) was added to cell cultures 1 h before and maintained during GSSG treatment. Figure 5A shows that treatment with this compound prevents GSSG-mediated cell death. In fact, at 24 h, the number of viable cells treated with 1 mM GSSG, in the presence of GSHest, was equal to that of control cells, whereas a slightly different number of cells were counted at 48 h. Furthermore, no significant difference in GSH and mixed disulfide content was observed between control and GSHest-supplemented cells at 12 h (Fig. 5B), even if a slight increase in the mixed disulfide levels was detected after 24 h. Figure 5C shows the results from cytofluorimetric analyses of cells loaded with GSHest upon annexin V-FITC/propidium iodide staining. The percentage of apoptotic cells measured at 48 h strongly decreased from 62% to 12% in the presence of GSHest. Such protection reflected the inhibition of the molecular markers involved in GSSG-mediated apoptosis. Figure 5D shows that cytosolic

cytochome c, caspase-9 activation, and Bcl-2 down-regulation were almost unaffected at 18 h after GSSG treatment. These results indicate that the decrease in intracellular GSH is causative of GSSG insult and related to apoptotic cell death, because by supplying reducing equivalents to cells, we were able to buffer the oxidative stress mediated by GSSG and, at the same time, to counteract cell death commitment. GSSG treatment affects the thioredoxin/ASK1 complex Thioredoxin represents another component of the intracellular antioxidant defense strictly related to glutathione redox state (14, 40). Thus, we evaluated the thioredoxin expression level under our experimental conditions. Cells were treated with 1 mM GSSG for 12 h, and lysates were used for Western blot analyses as described in Materials and Methods. Figure 6A shows that thioredoxin levels were significantly reduced upon GSSG treatment (�63 ± 5%); however, when electrophoresis was performed under reducing conditions, thioredoxin levels were comparable to untreated cells. Moreover, in the presence of GSHest, higher levels of thioredoxin were also detected under nonreducing conditions (�32 ± 7% with respect to untreated cells), indicating that thioredoxin redox state strictly relies on the intracellular redox environment. It has been recently reported that under physiological conditions, thioredoxin can inhibit apoptosis signal-regulating kinase 1 (ASK1) activity through direct interaction; however, under oxidative stress, thioredoxin dissociates from ASK1 as a result of conformational changes following oxidation of its two proximal cysteine residues (22). Active ASK1 could phospho-activate downstream MAPKs, which ultimately led to cell death. Additional experiments were carried out to assess whether the different levels of thioredoxin detected under nonreducing conditions were related to its binding to ASK1. Cell lysates were immunoprecipitated with a polyclonal antibody against ASK1, as described in Materials and Methods, and used for Western blot analyses. Figure 6B shows that the total amount of ASK1 was unchanged during GSSG treatment; however, the amount of thioredoxin that coprecipitated with ASK1 was significantly different. In particular, GSSG treatment inhibited thioredoxin/ASK1 association, whereas GSHest restored their binding capacity.

GSSG treatment induces cell death by the p38 MAPK pathway Furthermore, in search of the mechanism(s) through which GSSG causes cell death, we focused on the effects of GSSG on downstream kinases of ASK1 such as MKK3/6, that induce the activation of p38 MAPK. Western blot analyses of p38 showed an increase as early as 3 h after GSSG treatment (Fig. 7A), indicating that this molecular factor could be involved in GSSG-mediated toxicity. The activity of MAPKs is mainly regulated by changes in phosphorylation rather than by changes in total protein content (which are usually slight). p38 phosphorylation depends on the activity of upstream MKK3/6 kinases, whereas phospho-p38 in turn (via phosphorylation) regulates gene transcription. To verify the involvement of this pathway in GSSG-mediated toxicity, we measured the phosphorylated forms of these proteins. In particular, as shown in Figure 7A, the p38 upstream kinases MKK3/6 were already activated 3 h after treatment; whereas phospho-p38 was profoundly affected upon 6 h. Overall, these results indicate that the p38 MAPK pathway is operative during GSSG-mediated cytotoxicity, and it seems to represent a very early signaling event in the processes leading to cell death.

The involvement of the p38 MAPK pathway in GSSG-mediated toxicity was further confirmed by the results obtained with the specific inhibitor of p38, the pyridyl imidazole compound SB 203580. Figure 7B shows that treatment of cells with 5 µM SB 203580 abolished the toxicity of 1 mM GSSG, as evaluated by double staining with annexin V-FITC/propidium iodide. Based on cell viability, the molecular factors involved in apoptosis were markedly inhibited upon treatment with SB 203580 (Fig. 7C). These data are in good agreement with the specific SB 203580-mediated blockage of the p38 MAPK pathway at the level of p38 (Fig. 7D). Moreover, GSHest was able to significantly reduce the p38 MAPK pathway at the MKK3/6 levels, supporting that redox unbalance represents an upstream regulator of p38 activation, which ultimately governs the downstream response to GSSG-mediated toxicity. GSSG treatment does not induce cell death in differentiated cells One of the approaches used to selectively kill tumor cells is represented by alteration of the intracellular redox state, because under physiological conditions, differentiated cells possess more efficient antioxidant defenses with respect to tumor cells (41). To test the efficacy and specificity of GSSG-mediated toxicity for cancer cells, we performed additional experiments with primary monocytes. Figure 8A shows that cell viability of monocytes was not affected by GSSG treatment up to 3 days, even at a concentration of 5 mM. Furthermore, no changes were observed in the intracellular GSH content (Fig. 8B). The same results were obtained with phorbol-12-myristate-13-acetate (PMA)-differentiated U937 cells (data not shown). The lack of GSSG toxicity in primary or differentiated cells, even at very high concentrations, indicates that GSSG has potential clinical efficacy in anticancer chemotherapy. DISCUSSION Increasing evidence suggests that ligand-stimulated ROS generation plays a role in the complex machinery of signal transduction by altering the intracellular redox state (3, 4, 42, 43). However, little is known regarding the specific mechanism through which the oxidant acts. One piece of evidence linking the cellular redox state to specific signaling pathways came from the observation that ROS specifically alter the oxidation state of reactive cysteine residues in proteins. Oxidative stress can oxidize cysteine thiol groups to sulfenic ions that readily react to form either inter- and intramolecular disulfide bridges or mixed disulfide with GSH (2, 44, 45). This clearly represents a specific, reversible mechanism for regulation of enzyme activity by oxidants. Taking into account that oxidative stress appears to contribute to a variety of human diseases (46), it seems likely that the information learned from probing redox signaling will provide insight into a wide variety of physiological and pathological conditions. In this study, we have tried to shed light on the signaling events mimicked by GSSG, the physiological oxidation product of GSH. GSSG, without crossing the plasma membrane, was able to commit U937 cells to programmed cell death in a dose-dependent manner. We propose that in the mechanisms leading to death, a signaling process via thiol/disulfide exchange was activated starting from the extracellular environment. This was evidenced by the experiments carried out in the presence of Alexa fluor C5 maleimide, a nonpermeable fluorescent thiol-labeling probe. In particular, we demonstrated a remarkable decrease in the exofacial thiol residues of proteins, after 12 h of GSSG treatment, that correspond to the time point during which changes in the intracellular

redox start to be detectable. GSH was significantly decreased concomitantly with an increase in mixed disulfides, whereas GSSG levels were not changed. These results demonstrated that GSSG was not able to enter the cells; however, the lack of increase in cysteine or cystine that we detected by HPLC analyses, together with the experiments carried out on the presence of acivicin, confirmed that enzymatic cleavage of GSSG by γ-GT, specifically involved in the γ-glutamyl cycle, did not occur. Among the molecular factors that can be modulated by GSH depletion, ROS and the antiapoptotic protein Bcl-2, seem to be affected in several experimental conditions (38). In fact, both Bcl-2 and GSH are strictly related to alterations in the flux of ROS; Bcl-2 has been associated with antioxidative effects (35), and GSH represents a valid antioxidant agent either directly by efficiently scavenging H2O2 and ROOH radical species or by the action of enzymatic catalysis of GSH-peroxidase. In our model, both factors were affected: ROS production increased at 12 h after treatment, which, as described previously, corresponds to the time at which GSH decrease was significant, whereas Bcl-2 was significantly down-regulated as early as 6 h. This points to a sequence of events downstream of GSSG treatment, with increased ROS production being simultaneous to GSH decrement and subsequent to the significant decrease of Bcl-2. Moreover, the increase in ROS flux may represent an amplification tool of the oxidative stress induced by GSSG treatment. The persistent oxidative insult caused by GSSG, which is monitored by both the time-dependent decrease of intracellular GSH and the low levels of Bcl-2, could be responsible for the activation of redox-mediated signaling events that ultimately commit U937 cells to the mitochondrial pathway of programmed cell death (47). In fact, we observed cytochrome c release from mitochondria, caspase-9 cleavage, and finally caspase-3 activation. On the basis of our previous work that demonstrated a direct correlation between apoptotic U937 cells and GSH content (48), we are confident to suggest that the decrease in GSH content occurs before cell commitment to apoptosis. In fact, taking into account that U937 apoptotic cells are completely void of GSH, the remaining viable cells had the same total GSH content as untreated cells; however, the distribution among the redox forms of glutathione are changed, with the mixed disulfide species being the predominant form. In search of the molecular factors involved in the stress-responsive signal cascade of apoptosis induction, we focused on p38 MAPK cascade because it has been demonstrated that this pathway could be selectively activated by low levels of oxidative stress (4). Although various other signal pathways have not been examined, it was evident that the p38 MAPK pathway was selectively activated under the GSSG-mediated oxidative stress through up-regulation and phospho-activation of p38 MAPK. MKK3/6 may at least in part contribute to the selective p38 activation: We found that it was phosphorylated and activated as early as 3 h, whereas p38 activation was evident at 6 h. Moreover, the activation of such a signaling cascade was observed before the changes in GSH homeostasis, suggesting that the p38 MAPK pathway sensed very low changes in the intracellular redox state. However, the decrease in GSH content could be an amplifying cofactor of p38-induced cell death, which, by keeping the intracellular environment in an oxidative state, favors the execution step of programmed cell death. Owing to its hydrophilic nature, GSSG is not able to enter the cells, thus the activation of the p38 MAPK pathway could be mediated by thiol/disulfide exchange reactions with membrane protein thiols belonging to receptors spanning the plasma membrane, which represent the origin from which signal transduction pathways start. We suggested that the molecular link between p38 MAPK

activation and the GSSG-mediated oxidative insult had to be an upstream effector, which is able to �sense� and transduce the alterations of the extracellular redox environment. It has been demonstrated that ASK1 represents the upstream MAPKK kinase, which connects both membrane receptor activation and environmental homeostasis changes to cell response mediated by the phospho-activation of downstream effectors (49). Recent observations have shown that the capability of ASK1 to respond to redox alteration depends on its association with a regulatory protein, thioredoxin. This heterocomplex maintains ASK1 in the inactivated form until oxidative stress alters the redox state of thioredoxin. The consequent oxidation of thioredoxin allows ASK1 multimerization and phospho-activation of downstream effectors (42). p38 MAPK represents one of such molecular factors that, in order to regulate gene transcription, is phospho-activated by MKK3/6, a specific ASK1-dependent downstream protein kinase (50). Under our experimental conditions, the thioredoxin redox state was altered upon GSSG treatment, as evidenced by the lower thioredoxin content detected by Western blot analyses under nonreducing conditions. Moreover, the decrease in its content, upon coprecipitation with ASK1, indicated an involvement of the thioredoxin/ASK1 complex in the GSSG-mediated oxidative insult. Even though this mechanism remains to be elucidated and is currently under investigation in our laboratory, a direct relationship between thioredoxin, ASK1, and GSSG-mediated oxidative stress seems to be reasonable in our experimental model. The activation of the p38 MAPK pathway upon GSSG treatment was found to be reverted, in terms of cell viability and inhibition of the mitochondrial pathway, by the protection achieved in the experiments carried out in the presence of SB 203580. This compound was able to specifically inhibit p38 MAPK phosphorylation, thus showing a characteristic increase in the immunoreactive band of the upstream MKK3/6 phosphorylated form. At the same time, GSSG-mediated toxicity was efficiently counteracted by restoring the intracellular reducing environment through treatment with GSHest. GSHest treatment also affected thioredoxin/ASK1 complex; we showed, by Western blot analyses, an increase in the levels of thioredoxin as well as in the ASK1-associated form. Then, GSHest protective effects were upstream with respect to those exerted by SB 203580, because we demonstrated that the activation of the p38 MAPK pathway was inhibited at the level of MKK3/6, even though a direct inhibition of p38 MAPK by GSHest could not be excluded. The results obtained with both SB 203580 and GSHest demonstrated a significant protection against GSSG-mediated cell death. However, especially in the experiments carried out in the presence of the p38 inhibitor, we noticed that a small amount of cells still underwent apoptosis. It is conceivable that the redox alterations induced by GSSG, which became severe after 24 h, were per se able to induce cell death by other mechanisms not related to the activation of the p38 MAPK pathway. However, the low amount of apoptotic cells detected at 48 h, upon treatment with GSSG in the presence of GSHest, could still be due to p38 MAPK pathway activation, because the intracellular redox state at this experimental time point was not completely restored. Therefore, externally added GSSG, without crossing the cell membrane, was able to trigger cell death by activation of a canonical oxidative stress response, via alteration of the redox cell environment, and increased ROS production, and, at the same time, it can be envisaged as a ligand-like factor. This role is supported by the fact that, during the cellular response to GSSG, several molecular markers of the receptor-mediated pathway, such as those that belong to p38 MAPK-

mediated signaling, are activated. Although the characterization of the membrane proteins, which are potential targets of GSSG, is still under investigation in our laboratory, we can suggest that ASK1 upstream cysteine-rich receptors, such as those belonging to the TNF-R super family (51), could be affected. This hypothesis is strengthened by the lack of effects in intracellular redox state or in cell viability on primary monocytes or PMA-differentiated U937 cells treated with GSSG, and is in line with the finding that tumor cells are generally more sensitive to the detrimental effects mediated by specific TNF-R such as by TNF-related apoptosis-inducing ligand (TRAIL) receptors. (52, 53). In fact, TRAIL receptors, which are present in a wide range of cells but are operative only on tumor cell-membrane surfaces, have been suggested as a novel target for cancer therapy (54, 55). It will be interesting to determine whether tumor cell lines of different origins respond in the same manner to GSSG treatment. This could be of paramount importance in postulating a specific role for GSSG in anticancer therapy. In fact, in the strategy of selectively sensitizing tumor cells to chemotherapy and radiotherapy, the potential of depleting intracellular GSH as an adjuvant-based therapy has been extensively addressed (56�59). However, GSH depletion by an inhibitor of its synthesis or by alkylating agents had contradictory effects on enhancing tumor sensitivity to chemotherapeutic agents (60, 61). GSSG by specifically altering the intracellular redox state of tumor cells without effects on differentiated cells may be regarded as a potential tool for enhancing cell sensitivity to chemotherapeutic agents or radiation. ACKNOWLEDGMENTS This work was partially supported by MURST �Cofinanziamento 2000,� CNR-MIUR, and Ministero della Sanità �Progetto di Ricerca Finalizzata.� REFERENCES 1. Sun, Y., and Oberley, L.W. (1996) Redox regulation of transcriptional activators. Free Radic.

Biol. Med. 21, 335�348 2. Finkel T. (2000) Redox-dependent signal transduction. FEBS Lett. 476, 52�54 3. Herrlich, P., and Bohmer F.D. (2000) Redox regulation of signal transduction in mammalian

cells. Biochem. Pharmacol. 59, 35�41 4. Kurata, S. (2000) Selective activation of p38 MAPK cascade and mitotic arrest caused by low

level oxidative stress. J. Biol. Chem. 275, 23413�23416 5. Schulze-Osthoff, K., Ferrari, D., Los, M., Wesselborg, S., and Peter M.E., (1998) Apoptosis

signaling by death receptors. Eur. J. Biochem. 254, 439�459 6. Trevillyan, J.M., Chiou, X.G., Ballaron, S.J., Tang, Q.M., Buko, A., Sheets, M.P., Smith, M.L.,

Putman, C.B., Wiedeman, P., Tu, N., Madar, D., Smith, H.T., Gubbins, E.J., Warrior, U.P., Chen, Y.W., Mollison, K.W., Faltynek, C.R., and Djuric, S.W. (1999) Inhibition of p56(lck) tyrosine kinase by isothiazolones. Arch. Biochem. Biophys. 364, 19�29

7. Hehner, S.P., Breitkreutz, R., Shubinsky, G., Unsoeld, H., Schulze-Osthoff, K., Schmitz, M.L., and Droge, W. (2000) Enhancement of T cell receptor signaling by a mild oxidative shift in the intracellular thiol pool. J. Immunol. 165, 4319�4328

8. Cotgreave, I.A., and Gerdes, R.G. (1998) Recent trends in glutathione biochemistry-glutathione-

protein interactions: a molecular link between oxidative stress and cell proliferation? Biochem. Biophys. Res. Commun. 242, 1�9

9. Meister, A., and Anderson, M.E. (1983) Glutathione. Annu. Rev. Biochem. 52, 711�760 10. Holmgren, A. (2000) Antioxidant function of thioredoxin and glutaredoxin systems. Antioxid.

Redox Signal. 2, 811�820 11. Klatt, P., and Lamas, S. (2000) Regulation of protein function by S-glutathiolation in response

to oxidative and nitrosative stress. Eur. J. Biochem. 267, 4928�4944 12. Barrett, W.C., DeGnore, J.P., Konig, S., Fales, H.M., Keng, Y.F., Zhang, Z.Y., Yim, M.B., and

Chock, P.B. (1999) Regulation of PTP1B via glutathionylation of the active site cysteine 215. Biochemistry 38, 6699�6705

13. Rainwater, R., Parks, D., Anderson, M.E., Tegtmeyer, P., and Mann, K. (1995) Role of cysteine

residues in regulation of p53 function. Mol. Cell. Biol. 15, 3892�3903 14. Schafer, F.Q., and Buettner, G.R. (2001) Redox environment of the cell as viewed though the

redox state of the glutathione disulfide/glutathione couple. Free Radic. Biol. Med. 30, 1191�1212

15. Nordberg, J., and Arner, E.S. (2001) Reactive oxygen species, antioxidants, and the mammalian

thioredoxin system. Free Radic. Biol. Med. 31, 1287�1312 16. Nishiyama, A., Masutani, H., Nakamura, H., Nishinaka, Y., and Yodoi, J. (2001) Redox

regulation by thioredoxin and thioredoxin-binding proteins. IUBMB Life 52, 29�33 17. Davis, W. Jr, Ronai. Z., and Tew, K.D. (2001) Cellular thiols and reactive oxygen species in

drug-induced apoptosis. J. Pharmacol. Exp. Ther. 296, 1�6 18. Maxwell, S.R. (1995) Prospects for the use of antioxidant therapies. Drugs 49, 345�361 19. Shackelford, R.E., Kaufmann, W.K., and Paules, R.S. (2000) Oxidative stress and cell cycle

checkpoint function. Free Radic. Biol. Med. 28, 1387�1404 20. van den Dobbelsteen, D.J., Nobel, C.S., Schlegel, J., Cotgreave, I.A., Orrenius, S., and Slater,

A.F. (1996) Rapid and specific efflux of reduced glutathione during apoptosis induced by anti-Fas/APO-1 antibody. J. Biol. Chem. 271, 15420�15427

21. Ghibelli, L., Fanelli, C., Rotilio, G., Lafavia, E., Coppola, S., Colussi, C., Civitareale, P., and Ciriolo, M.R. (1998) Rescue of cells from apoptosis by inhibition of active GSH extrusion. FASEB J. 12, 479�486

22. Adler, V., Yin, Z., Tew, K.D., and Ronai, Z. (1999) Role of redox potential and reactive oxygen

species in stress signaling. Oncogene 18, 6104�6111 23. Saitoh, M., Nishitoh, H., Fujii, M., Takeda, K., Tobiume, K., Sawada, Y., Kawabata, M.,

Miyazono, K., and Ichijo, H. (1998) Mammalian thioredoxin is a direct inhibitor of apoptosis signal-regulating kinase (ASK) 1. EMBO J. 17, 2596�2606

24. Gotoh, Y., and Cooper, J.A. (1998) Reactive oxygen species- and dimerization-induced

activation of apoptosis signal-regulating kinase 1 in tumor necrosis factor-alpha signal transduction. J. Biol. Chem. 273, 17477�17482

25. Ichijo, H., Nishida, E., Irie, K., Dijke, P., Saitoh, M., Moriguchi, T., Takagi, M., Matsumoto, K.,

Miyzaono, K., and Gotoh, Y. (1997) Induction of apoptosis by ASK1, a mammalian MAPKKK that activates SAPK/JNK and p38 signaling pathways. Science 275, 90�94

26. Reglinski, J., Hoey, S., Smith, W.E., and Sturrock, R.D. (1988) Cellular response to oxidative

stress at sulfhydryl group receptor sites on the erythocyte membrane. J. Biol. Chem. 263, 12360�12366

27. Ciriolo, M.R., Paci, M., Sette, M., De Martino, A., Bozzi, A., and Rotilio, G. (1993)

Transduction of reducing power across the plasma membrane by reduced glutathione. A 1H-NMR spin-echo study of intact human erythocytes. Eur. J. Biochem. 215, 711�718

28. Ciriolo, M.R., De Martino, A., Lafavia, E., Rossi, L., Carri, M.T., and Rotilio, G. (2000) Cu,Zn-

superoxide dismutase-dependent apoptosis induced by nitric oxide in neuronal cells. J. Biol. Chem. 275, 5065�5072

29. Reed, D.J., Babson, J.R., Beatty, P.W., Brodie, A.E., Ellis, W.W., and Potter, D.W. (1980)

High-performance liquid chomatography analysis of nanomole levels of glutathione, glutathione disulfide, and related thiols and disulfides. Anal. Biochem. 106, 55�62

30. Ciriolo, M.R., Palamara, A.T., Incerpi, S., Lafavia, E., Bue, M.C., De Vito, P., Garaci, E., and

Rotilio, G. (1997) Loss of GSH, oxidative stress, and decrease of intracellular pH as sequential steps in viral infection. J. Biol. Chem. 272, 2700�2708

31. Lowry, O.H., Rosebrough, N.J., Farr, A.L., and Randall, R.J. (1951) Protein measurement with

the Folin-phenol reagent. J. Biol. Chem. 193, 265�275 32. Tan, S., Sagara, Y., Liu, Y., Maher, P., and Schubert, D. (1998) The regulation of reactive

oxygen species production during programmed cell death. J. Cell. Biol. 141, 1423�1432

33. Lizard, G., Gueldry, S., Sordet, O., Monier, S., Athias, A., Miguet, C., Bessede, G., Lemaire, S., Solary, E., and Gambert, P. (1998) Glutathione is implied in the control of 7-ketocholesterol-induced apoptosis, which is associated with radical oxygen species production. FASEB J. 12, 1651�1663

34. Steinman, H.M. (1995) The Bcl-2 oncoprotein functions as a pro-oxidant. J. Biol. Chem. 270,

3487�3490 35. Voehinger, D.W., and Meyn, R.E. (2000) Redox aspects of Bcl-2 function. Antioxid. Redox

Signal. 2, 537�550 36. Hockenbery, D.M., Oltvai, Z.N., Yin, X.M., Milliman, C.L., and Korsmeyer, S.J. (1993) Bcl-2

functions in an antioxidant pathway to prevent apoptosis. Cell 75, 241�251 37. Vahmeijer, A.L., Hoetelmans, R.W., Mulder, G.J., Schutrups, J., van Vlierberghe, R.L., van de

Velde, C.J., and van Dierendonck, J.H. (2000) Development of resistance to glutathione depletion-induced cell death in CC531 colon carcinoma cells: association with increased expression of bcl-2. Biochem. Pharmacol. 59, 1557�1562

38. Voehinger, D.W. (1999) BCL-2 and glutathione: alterations in cellular redox state that regulate

apoptosis sensitivity. Free Radic. Biol. Med. 27, 945�950 39. Martensson, J., Han, J., Griffith, O.W., and Meister, A. (1993) Glutathione ester delays the

onset of scurvy in ascorbate-deficient guinea pigs. Proc. Natl. Acad. Sci. 90, 317�321 40. Casagrande, S., Sonetto, V., Fratelli, M., Gianazza, E., Eberini, I., Massignan, T., Salmona, M.,

Chang, G., Holmgren, A., and Grezzi, P. (2002) from the Cover: Glutathionylation of human thioredoxin: A possible crosstalk between the glutathione and thioredoxin systems. Proc. Natl. Acad. Sci. USA 99, 9745�9749

41. Cleveland, J.L., and Kastan, M.B. (2000) Cancer. A radical approach to treatment. Nature 407,

309�311 42. Kamata, H., and Hirata, H. (1999) Redox regulation of cellular signalling. Cell. Signal. 11, 1�14 43. Marshall, H.E., Merchant, K., and Stamler, J.S. (2000) Nitrosation and oxidation in the

regulation of gene expression. FASEB J. 14, 1889�1900 44. Klatt, P., Molina, E.P., De Lacoba, M.G., Padilla, C.A., Martinez-Galesteo, E., Barcena, J.A.,

and Lamas, S. (1999) Redox regulation of c-Jun DNA binding by reversible S-glutathiolation. FASEB J. 13, 1481�1490

45. Thannickal, V.J., and Fanburg, B.L. (2000) Reactive oxygen species in cell signaling. Am. J.

Physiol. Lung Cell. Mol. Physiol. 279, L1005�L1028

46. Halliwell, B., and Gutteridge, J.M.C. Free radicals and antioxidant protection: mechanisms and significance in toxicology and disease. (1998) Hum. Toxicol. 7, 7�13

47. Desagher, S., and Martinou, J.C. (2000) Mitochondria as the central control point of apoptosis.

Trends Cell. Biol. 10, 369�377 48. Ghibelli, L., Coppola, S., Rotilio, G., Lafavia, E., Maresca, V., and Ciriolo M.R. (1995)

Nonoxidative loss of glutathione in apoptosis via GSH extrusion. Biochem. Biophys. Res. Commun. 216, 313�320

49. Tobiume, K., Matsuzawa, A., Takahashi, T., Nishitoh, H., Morita, K., Takeda, K., Minora, O.,

Miyazono, K., Noda, T., and Ichijo, H. (2001) ASK1 is required for sustained activations of JNK/p38 MAP kinases and apoptosis. EMBO Rep. 2, 222�228

50. Rincon, M., Flavell, R.A., and Davis, R.A. (2000) The JNK and P38 MAP kinase signaling

pathways in T cell-mediated immune responses. Free Radic. Biol. Med. 28, 1328�1337 51. Ichijo, H. (1999) From receptors to stress-activated MAP kinases. Oncogene 18, 6087�6093 52. Daniel, P.T., Wieder, T., Sturm, I., and Schulze-Osthoff, K. (2001) The kiss of death: promises

and failures of death receptors and ligands in cancer therapy. Leukemia 15, 1022�1032 53. Solary, E., Droin, N., Bettaieb, A., Corcos, L., Dimanche-Boitrel, M.T., and Garrido, C. (2000)

Positive and negative regulation of apoptotic pathways by cytotoxic agents in hematological malignancies. Leukemia 14, 1833�1849

54. Bonavida, B., Ng, C.P., Jazirehi, A., Schiller, G., and Mizutani, Y. (1999) Selectivity of TRAIL-

mediated apoptosis of cancer cells and synergy with drugs: the trail to nontoxic cancer therapeutics (review). Int. J. Oncol. 15, 793�802

55. Gibson, S.B., Oyer, R., Spalding, A.C., Anderson, S.M., and Johnson, G.L. (2000) Increased

expression of death receptors 4 and 5 synergizes the apoptosis response to combined treatment with etoposide and TRAIL. Mol. Cell. Biol. 20, 205�212

56. Meister, A. (1983) Selective modification of glutathione metabolism. Science 220, 472�477 57. Estrema, J.M., Obrador, E., Navarro, J., Lasso De la Vega, M.C., and Pellicer, J.A. (1995)

Elimination of Ehlich tumours by ATP-induced growth inhibition, glutathione depletion and X-rays. Nat. Med. 1, 84�88

58. Rahman, I., and MacNee, W. (2000) Regulation of redox glutathione levels and gene

transcription in lung inflammation: therapeutic approaches. Free Radic. Biol. Med. 28, 1405�1420

59. Coffey, R.N., Watson, R.W., Hegarty, N.J., O'Neill, A., Gibbons, N., Brady, H.R., and Fitzpatrick, J.M. (2000) Thiol-mediated apoptosis in prostate carcinoma cells. Cancer 88, 2092�2104

60. van der Kolk, D.M., Vellenga, E., Muller, M., and de Vries, E.G. (1999) Multidrug resistance

protein MRP1, glutathione, and related enzymes. Their importance in acute myeloid leukemia. Adv. Exp. Med. Biol. 457, 187�198

61. Jordan, J., d'Arcy Doherty, M., and Cohen, G.M. (1987) Effects of glutathione depletion on the

cytotoxicity of agents toward a human colonic tumour cell line. Br. J. Cancer. 55, 627�631

Received February 21, 2002; accepted September 23, 2002.

Fig. 1

Figure 1. Glutathione disulfide (GSSG) induces apoptosis in U937 cells. U937 cells were grown in 10% fetal bovine serum RPMI over a 2-day period in the presence or absence of different amounts of GSSG. Cell proliferation was assessed either by the MTS [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium] colorimetric assay (A) or direct counting (B). Data are from three separate experiments carried out in duplicate and expressed as means ±SD. C) U937 cells treated with 1 mM GSSG for 24 and 48 h were washed with phosphate-buffered saline; cell suspensions were stained either with Hoechst 33342 to detect nuclear fragmentation by fluorescence microscopy (upper panels) or with annexin V-FITC and propidium iodide (bottom panels), as described in Materials and Methods. Data are from a typical experiment out of 10 giving comparable results. Fluorescence was analyzed by a FACScan instrument, and percentages of positive-stained cells were calculated using WinMDI version 2.8 software.

Fig. 2

Figure 2. Molecular markers of apoptotic mitochondrial pathway in U937 cells treated with GSSG. U937 cells were treated with 1 mM GSSG for 18 and 24 h. A) Cytosolic proteins (20 µg) and 10 µg of mitochondrial-enriched pellets were loaded onto each lane for Western blot analyses of cytochrome c as described in Materials and Methods. B) Cytosolic proteins (20 µg) were loaded for pro-caspase-9 and caspase-9 protein levels. The results are from one experiment representative of three that gave similar results. β-Actin was used as loaded sample control. C) Caspase-3 activity was determined by a fluorimetric assay on cell extracts, using Ac-DEVD-AFC as substrate, and presented as arbitrary units. Data are expressed as mean ±SD, n=4 (*P<0.001).

Fig. 3

Figure 3. GSSG induces membrane thiols oxidation and intracellular changes of glutathione redox state. A) After 12 h of treatment with 1 mM GSSG, U937 cells were incubated with 50 µM Alexa fluor C5 maleimide at 37°C for 1 h and analyzed by fluorescence microscope and FACScan as described in Materials and Methods. Data are from a typical experiment out of three giving comparable results. For intracellular glutathione redox-state analyses, U937 cells were grown in the presence of different amounts of GSSG. At the indicated time points, cells were collected; washed with phosphate-buffered saline; and used for GSH, GSSG, and mixed disulfides (GS-R) assay by high-performance liquid chromatography (HPLC) as described in Materials and Methods. B) GSH. C) GS-R. Data are expressed as nmol of GSH equivalents/mg total protein. Values represent means ±SD (n=8). All treated cells were significantly different from the control (P<0.001). D) U937 cells were incubated with acivicin for 1 h before and throughout the following 24 h of treatment with 1 mM GSSG. Intracellular levels of GSH and GSSG were measured by HPLC as described in Materials and Methods. Data are expressed as mean ±SD (n=4). *P<0.05 with respect to acivicin untreated cells.

Fig. 4

Figure 4. Increased reactive oxygen species production and Bcl-2 down-regulation as hallmarks of GSSG treatment. A) U937 cells were treated with 1 mM GSSG for 12 h, and then the cell suspension was incubated with 50 µM DCF-DA for 30 min at 37°C and assayed by FACS analysis, as described in Materials and Methods. t-Butyl-OOH; cells treated with 100 µM tert-butyl hydroperoxide as positive control. B) U937 cells were treated with 1 mM GSSG. At the indicated time, aliquots of cell suspension were withdrawn and treated as described in Materials and Methods for Western blot analyses. Proteins (20 µg) from total cell extract were loaded onto each lane. C) Proteins (50 µg) from total cell extract were loaded onto each lane for Western blot analyses of Bcl-2 at 12 and 24 h treatment with 1 mM GSSG as described in Materials and Methods. The results are from one experiment representative of three that gave similar results. The time-dependent decrease in Bcl-2 expression levels is shown as percentages calculated with respect to control, after normalization vs. β-actin.

Fig. 5

Figure 5. GSHest prevents GSSG-induced apoptosis. U937 cells were grown in 10% fetal bovine serum RPMI; 1 mM GSHest was added 1 h before the addition of GSSG and maintained throughout the experiment. A) Cell proliferation was assessed by direct counting. B) GSH and mixed disulfides (GS-R) were assayed by HPLC, as described in Materials and Methods. Data are expressed as nmol of GSH equivalents/mg total protein. Values represent means ±SD (n=4). C) After 24 and 48 h of treatment with 1 mM GSSG, cells were washed with phosphate-buffered saline and stained with annexin V-FITC and propidium iodide for FACScan analyses as described in Materials and Methods. Percentages of positive-stained cells were calculated using WinMDI version 2.8 software. The data reported in A and C are from three separate experiments carried out in duplicate and expressed as means ±SD. Control, control cells; GSSG, cells treated with 1 mM GSSG; + GSHest, cells incubated with 1 mM GSH ethyl ester 1 h before and during GSSG treatment. D) U937 cells were treated with 1 mM GSSG for 18 h and used for Western blot analyses as described in Materials and Methods. Proteins (20 µg) from cytosolic extract were loaded onto each lane for cytosolic cytochrome c and caspase-9; proteins (20 µg) from total cell lysates for Bcl-2. The results are from one experiment representative of three that gave similar results.

Fig. 6

Figure 6. Thioredoxin/ASK1 complex is affected by GSSG treatment. U937 cells were grown in 10% fetal bovine serum RPMI; 1 mM GSHest was added 1 h before the addition of GSSG and maintained throughout the experiment. After 12 h of treatment with 1 mM GSSG, aliquots of cell suspension were withdrawn and treated as described in Materials and Methods for Western blot analyses. A) Proteins (50 µg) from total cell extract were loaded onto each lane. B) After 12 h of treatment with 1 mM GSSG, cells were lysates and immunoprecipitated using ASK1 polyclonal antibody. Immunoprecipitated proteins (50 µg) were loaded onto each lane, and thioredoxin was detected by a specific antibody as described in Materials and Methods. As control, 20 µg of proteins from the lysates, before immunoprecipitation, were loaded and examined for ASK1 and β-actin. The results are from one experiment representative of three that gave similar results.

Fig. 7

Figure 7. p38 mitogen-activated protein kinase signaling pathway is involved in GSSG-induced apoptosis. A) U937 cells were treated with 1 mM GSSG. At the indicated time points, aliquots of cell suspension were withdrawn and treated as described in Materials and Methods for Western blot analyses. Proteins (50 µg) from total cell extract were loaded onto each lane. The results are from one experiment representative of three that gave similar results. B) U937 cells were grown in 10% fetal bovine serum RPMI; 5 µM SB 203580 was added 2 h before the addition of GSSG and maintained throughout the experiment. After 24 and 48 h of treatment with 1 mM GSSG, cells were washed with phosphate-buffered saline and stained with annexin V-FITC and propidium iodide for FACScan analyses as described in Materials and Methods. Control, control cells; GSSG, cells treated with 1 mM GSSG; + SB 203580, cells incubated with 5 µM SB 203580 2 h before and during GSSG treatment. Percentages of positive-stained cells were calculated using WinMDI version 2.8 software. Data are from three separate experiments carried out in duplicate and expressed as means ±SD C) U937 cells were treated with 1 mM GSSG for 18 h. Cell lysates (± SB 203580) were analysed by Western blot analyses as described in Materials and Methods. Cytosolic proteins (20 µg) were loaded onto each lane for cytosolic cytochrome c and caspase-9; proteins (20 µg) from total cell lysates for Bcl-2. The results are from one experiment representative of three that gave similar results. D) GSHest (1 mM) or 5 µM SB 203580 were added 1 h and 2 h, respectively, before the addition of GSSG and maintained throughout the experiment. At 3 h and 6 h, aliquots of cell suspension were withdrawn and treated as described in Materials and Methods for Western blot analyses. Sample (20 µl) was loaded onto each lane. The results are from one experiment representative of three that gave similar results.

Fig. 8

Figure 8. GSSG does not induce cytotoxic effects in primary monocytes. Primary monocytes were grown in 10% fetal bovine serum RPMI and treated with different concentration of GSSG ( control; 1 mM; 2 mM; 5 mM). At the indicated days of culture, cells were starved and used for cell viability assessed by Trypan Blue exclusion (A) and GSH assay (B) as described in Materials and Methods. GSH was reported as the percentage of that present in untreated cells, which corresponds to 100%. Data are expressed as mean ±SD (n=4)