- Email: [email protected]
Hydrogen Peroxide Suppresses U937 Cell Death by Two Different Mechanisms Depending on Its Concentration
Experimental Cell Research 248, 430 – 438 (1999) Article ID excr.1999.4409, available online at http://www.idealibrary.com on
Hydrogen Peroxide Suppresses U937 Cell Death by Two Different Mechanisms Depending on Its Concentration Byung Ryong Lee and Hong-Duck Um 1 Laboratory of Cell Biology, Yonsei Medical Research Center, Yonsei University College of Medicine, CPO Box 8044, Seoul, Korea
To investigate the mechanisms of H 2O 2 adaptation in mammalian cells, we exposed human U937 leukemia cells to 0.05 mM H 2O 2. This treatment significantly suppressed cell death and DNA fragmentation induced by a subsequent challenge with 1 mM H 2O 2. A more dramatic protection was observed when cells were pretreated with 0.25 mM H 2O 2. Pretreatment with either 0.05 or 0.25 mM H 2O 2 also imparted cells with a survival advantage against serum withdrawal and C 2-ceramide treatment. H 2O 2 was found to be a mediator of cell death induced by serum withdrawal, but not by the addition of C 2-ceramide. Interestingly, 0.25 mM H 2O 2 greatly induced glutathione peroxidase, a H 2O 2consuming enzyme, whereas 0.05 mM H 2O 2 did not. Consistent with observation, pretreatment with 0.25 mM H 2O 2 resulted in a great reduction of cellular oxidant levels as determined by 2*7*-dichlorofluorescein fluorescence, and it also prevented elevation of oxidant levels upon subsequent challenge with 1 mM H 2O 2 or with serum withdrawal. These effects were not observed in cells pretreated with 0.05 mM H 2O 2. The sum of the data indicated that H 2O 2 suppresses cell death by two different mechanisms depending on its concentration: Relatively high concentrations enhance cellular antioxidant capacity, and lower concentrations block the lethal action of H 2O 2. © 1999 Academic Press
Key Words: adaptation; hydrogen peroxide; oxidative stress; cell death; antioxidant.
INTRODUCTION
Cells under aerobic conditions continuously generate reactive oxygen intermediates (ROIs) such as hydrogen peroxide (H 2O 2), superoxide anion (O 2. ), and hydroxyl radical (HOz). To counter the toxicity of ROIs, cells possess protective antioxidant systems. However, the generation of ROIs can often exceed cellular defense capacity. This oxidative condition is induced by ligation 1 To whom correspondence and reprint requests should be addressed. Fax: 82-02-362-8647. E-mail: [email protected].
0014-4827/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
of cell surface receptors and also by multiple classes of environmental agents [1]. As ROIs are implicated in the etiology of numerous physiological and pathophysiological states, e.g., inflammation, aging, carcinogenesis, atherosclerosis, and neurodegeneration [2], understanding cellular responses to oxidative stress has been a major goal of diverse fields of biology. An interesting feature of oxidative stress is that it can induce adaptive responses. When bacteria are exposed first to low concentrations of H 2O 2 or O 2. -generating agents, they become resistant to high levels of the same agents which would otherwise abolish their colony formation [for review see 3, 4]. Using similar assays to measure cellular growth competence, others have shown that similar protective responses occur in yeast and mammalian cells [5, 6]. In bacteria, the protective responses are accompanied by the synthesis of 30 – 40 proteins, including enzymes involved in DNA repair and antioxidant pathways and some heat shock proteins [7, 8]. The induction of a subset of the proteins was dependent upon two distinct regulons, oxyR and soxRS, which selectively responded to H 2O 2 and O 2. , respectively [7–9]. OxyR mediates the induction of catalase which reduces H 2O 2 to water, whereas soxRS does that of superoxide dismutase (SOD) which scavenges O 2. . It appears, therefore, that bacteria achieve the protection from ROIs, at least in part, by enhancing their capacity to degrade selective members of ROIs. Compared to bacteria, the mechanism(s) by which mammalian cells adapt to ROIs has not been well characterized. It has been reported that pretreatment of mammalian cells by relatively low concentrations of H 2O 2 induces the synthesis of at least 20 new proteins [6]. Inhibition of protein synthesis by cycloheximide (CHX) suppresses the induction of resistance to a subsequent challenge with high concentrations of H 2O 2 [6]. Such data indicate that protein synthesis is required for mammalian cell adaptation. Although the nature of protective proteins is not clear yet, potential candidates may include antioxidant enzymes. Indeed, vascular endothelial and tracheobronchial epithelial
430
INHIBITION OF CELL DEATH BY HYDROGEN PEROXIDE
cells respond to H 2O 2 by enhancing their mRNA levels and the activities of antioxidant enzymes, particularly catalase and glutathione peroxidase (GPX) [10, 11]. Those observations have led many investigators to believe that antioxidant enzymes play a role in mammalian cell adaptation. However, in an elegantly established adaptive condition, pretreatment of hamster fibroblast cells by H 2O 2 did not increase either the activities or the protein levels of various antioxidant enzymes including catalase, GPX, and SOD [6]. Moreover, the adapted cells did not display an enhanced capacity to degrade H 2O 2 when subsequently challenged with high levels of H 2O 2. One simple explanation for these contradictory observations could be that mammalian cells can adapt to H 2O 2 by two different mechanisms, one dependent on antioxidant enzymes and one independent of such enzymes. To date, however, no information is available concerning the factor(s) responsible for the selective induction of those mechanisms. In the present study, we explored whether the potential two mechanisms of H 2O 2 adaptation can function in a single cell type. To do so, we pretreated human U937 leukemia cells with various concentrations of H 2O 2 and evaluated the relationship between enhancement of cellular antioxidant capacity and induction of resistance to subsequent lethal stimuli. The data indicated that H 2O 2 adapts U937 cells by dual mechanisms: One mechanism enhances the cellular capacity to degrade H 2O 2, while the other suppresses the lethal action of H 2O 2. These two protective mechanisms were distinguished by their sensitivities to H 2O 2. EXPERIMENTAL PROCEDURES Materials. U937 cells were obtained from the American Type Culture Collection (Rockville, MD). Cell culture medium and its supplements were supplied by BioWhittaker (Walkersville, MD). Antibodies against human catalase, CuZn SOD, and rat GPX were generous gifts from Dr. Ho Zoon Chae. Anti-human Bcl-2 and -Bcl-X L antibodies were purchased from DAKO (Carpinteria, CA) and Transduction Laboratories (Lexington, KY), respectively. All other reagents whose suppliers are not indicated were purchased from Sigma (St. Louis, MO). Cell culture and treatments. U937 cells were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated FBS and gentamicin (50 mg/ml) at 37°C and 5% CO 2. To induce adaptive responses, cells at the density of 3.5 3 10 5/ml were exposed to 0.05 mM H 2O 2 (unless otherwise specified). At the end of the indicated incubation periods, pretreated and untreated control cells were challenged with either H 2O 2 (1 mM) or C 2-ceramide (0.02 mM) or washed in PBS and then cultured in serum-depleted medium. When cells were pretreated by 0.25 mM H 2O 2 or where indicated, viable cells were isolated by centrifugation over a layer of Ficoll–Paque (Pharmacia Biotech, Uppsala, Sweden), washed twice in PBS, and then challenged as described. Analysis of viability. Incubated cells received propidium iodide (PI) (5 mg/ml) followed by flow cytometry analysis to simultaneously
431
monitor PI uptake (FL-2 channel) and cell size (forward light scatter). To assess cell death, 10,000 cells were counted, and the cells that displayed both a reduction in cell size and a high permeability to PI were understood to be dead cells, as defined previously [12]. The percentage of cell death was calculated by the following formula: number of dead cells/number of total counted cells 3 100 and expressed as percentage of PI 1 cells. DNA fragmentation assay. A total of 2 3 10 6 cells were lysed in Triton X-100/Tris buffer [12]. The lysates were centrifuged at 13,000g. DNA fragments in the supernatant were extracted and processed for electrophoresis in 2% agarose at 40 V for 4 h [12]. Flow cytometric assay of 29,79-dichlorofluorescein oxidation. Cells were incubated in the absence or presence of serum, C 2-ceramide, or H 2O 2. After the indicated incubation times, the cells were treated for 5 min by 50 mM 29,79-dichlorofluorescein diacetate (DCFH-DA; Eastman Kodak, Rochester, NY). This compound rapidly diffuses into cells, and once within cells is hydrolyzed to oxidation-sensitive DCFH [13]. In the presence of ROIs, DCFH is oxidized to the highly fluorescent DCF, which is retained by living cells. The cell-associated levels of DCF fluorescence were analyzed by flow cytometry [13]. Western blot analysis. H 2O 2-treated and untreated control cells were lysed in Tris–HCl (40 mM, pH 8), 120 mM NaCl, 0.5 % Nonidet P-40, and protease inhibitors (2 mg/ml aprotinin, 2 mg/ml leupeptin, and 100 mg/ml PMSF). After removing cell debris by centrifugation at 13,000g for 5 min, equal amounts of proteins (40 –100 mg) were separated by 12% SDS–PAGE and then electrotransferred to Immobilon membranes (Millipore, Bedford, MA) which were subsequently blotted using the indicated antibodies and visualized by chemiluminescence (ECL; Amersham, Arlington Heights, IL).
RESULTS
H 2O 2 Induces a Survival Pathway H 2O 2 can induce either apoptosis or necrosis depending on its dose. In the case of U937 cells, apoptosis is the sole, or at least a major, mechanism of cell death triggered by 1 mM H 2O 2, while higher concentrations of H 2O 2 increase the necrotic population [14]. Concentrations of H 2O 2 lower than 0.1 mM did not significantly induce cell death under our experimental conditions (Fig. 1), although other investigators have reported cytotoxicity using similar conditions [15]. Therefore, 0.05 mM H 2O 2 was used for the following adaptation experiments. To investigate whether relatively low concentrations of H 2O 2 can adapt U937 against cell death, cells were preincubated in the absence or presence of 0.05 mM H 2O 2 for 24 h and then challenged with 1 mM H 2O 2. When preincubated without H 2O 2, cells rapidly lost their viability on addition of 1 mM H 2O 2 (Fig. 1A). Cell death was accompanied by a dramatic elevation of DNA fragmentation (Fig. 1B), confirming a previous report that 1 mM H 2O 2 kills U937 cells by apoptosis [14]. Cell death and DNA fragmentation were markedly reduced in cells that had been pretreated with 0.05 mM H 2O 2. Therefore, relatively low concentrations of H 2O 2 can trigger a survival pathway which appears to function against apoptosis. To determine a dose response for protection, we pre-
432
LEE AND UM
Time Course of the Protective Response To further establish optimal conditions for the protective response, cells were pretreated with 0.05 mM H 2O 2 for varied periods of time prior to their challenge with 1 mM H 2O 2. The protective effect of H 2O 2 was not observed in cells that had been pretreated for 4 h. Cells preincubated for 8 h displayed some protection, while
FIG. 1. H 2O 2 adaptation functions against apoptosis. (A) U937 cells were incubated in the absence (circles) or presence (squares) of 0.05 mM H 2O 2 for 24 h and then challenged with 1 mM H 2O 2 (closed symbols) or not (open symbols). The viability of U937 was determined by PI uptake at the indicated times. Values are the mean of four separate experiments with an error bar representing standard deviations. (B) After 16 h of exposure to the challenge, fragmented DNA was analyzed by agarose gel electrophoresis.
treated cells with various concentrations of H 2O 2 up to 0.25 mM. Since this relatively high concentration of H 2O 2 induced cell death (20 –30% PI 1 after 24 h), we carried out Ficoll gradient centrifugation to remove the dead cells, and then the collected viable cells received a subsequent challenge with 1 mM H 2O 2. In this experiment, Ficoll gradient was equally performed for the cells pretreated with sublethal concentrations of H 2O 2. As shown in Fig. 2A, pretreatment with 0.01 mM H 2O 2 was not protective, while pretreatment with 0.025 mM or higher concentrations of H 2O 2 offered protection in a dose-dependent manner. Therefore, induction of the survival pathway is dependent on the concentration of H 2O 2 employed in the preincubation. Moreover, the induction of protective response by high lethal concentrations of H 2O 2 indicated that as the concentration of H 2O 2 increases, H 2O 2 triggers both the survival and the death pathways.
FIG. 2. Dose response and time course of H 2O 2 adaptation. (A) Cells were exposed to the indicated concentrations of H 2O 2 for 24 h. Viable cells were isolated on a Ficoll–Paque gradient at the end of preincubation and were then challenged with 1 mM H 2O 2. Cellular viability was determined 48 h after the challenge. Open circles, pretreated but not challenged; closed circles, pretreated and challenged. (B) Cells were incubated in the absence (open circles) or presence (closed circles) of 0.05 mM H 2O 2 for the indicated periods of time and subsequently exposed to 1 mM H 2O 2 for 48 h.
INHIBITION OF CELL DEATH BY HYDROGEN PEROXIDE
FIG. 3. H 2O 2 does not induce Bcl-2 and Bcl-X L. Cells were treated by the indicated concentrations of H 2O 2 for 24 h and then the cellular levels of Bcl-2 and Bcl-X L were analyzed by Western blotting. Similar results were obtained when cells were treated by H 2O 2 for 4, 8, and 16 h.
maximal protection was seen when cells were pretreated overnight (16 –24 h) (Fig. 2B). Thereafter the protection declined. Thus, there is a time lag before the protective effect of H 2O 2 can be materialized, and once established, the protective response is transient. Requirement of Protein Synthesis The time lag likely reflects the need for macromolecular synthesis, possibly an antiapoptotic protein(s). To address this possibility, U937 cells were treated with 0.05 mM H 2O 2 in the absence or presence of 0.25 mM CHX. Neither H 2O 2 nor CHX alone significantly influenced cellular viability at the concentrations used (approximately 10% PI 1 cells at 24 h of incubation). However, their simultaneous presence resulted in a marked induction of cell death (41% PI 1). The synergy between H 2O 2 and CHX supports the premise that the H 2O 2induced survival pathway involves protein synthesis: Inhibition of the protein synthesis renders cells hypersensitive to the lethal action of H 2O 2, as is also seen in bacteria [7].
433
which efficiently induces the protective response (Fig. 4). However, treatment with 0.25 mM H 2O 2 resulted in the strong induction of two isoforms of GPX (19 and 24 kDa). The data suggested that GPX is a survival factor induced by H 2O 2, particularly with concentrations of 0.25 mM. In contrast, none of the tested concentrations of H 2O 2 significantly elevated the protein levels of catalase under our experimental conditions. Levels of SOD were not significantly altered by 0.05 mM H 2O 2, but were decreased by 0.25 mM H 2O 2. To examine the functional significance of GPX induction, we compared the capacity of control and adapted cells to consume H 2O 2. To do so, H 2O 2-pretreated and untreated control cells were exposed to 1 mM H 2O 2, and cell-associated levels of H 2O 2 were compared by DCF fluorescence. When control cells were analyzed, DCF fluorescence was immediately elevated after the H 2O 2 treatment, peaking at 20 min. After this time, the level of DCF fluorescence decreased, reflecting degradation of added H 2O 2 (Fig. 5). A similar pattern of DCF fluorescence was obtained using cells pretreated by 0.05 mM H 2O 2. It appears, therefore, that 0.05 mM H 2O 2 does not enhance cellular capacity to degrade H 2O 2. In contrast, pretreatment by 0.25 mM H 2O 2 dra-
H 2O 2 Does Not Induce Bcl-2 and Bcl-X L Bcl-2 and BCL-X L are well characterized antiapoptotic proteins. Given that overexpression of either Bcl-2 or BCL-X L can suppress H 2O 2-induced cell death [16, 17], it was of interest to ask whether the protective proteins induced by H 2O 2 pretreatment involved Bcl-2 and BCL-X L. However, neither 0.05 nor 0.25 mM H 2O 2 enhanced cellular levels of Bcl-2 and BCL-X L (Fig. 3). Cellular Antioxidant Capacity Is Enhanced by 0.25 mM, but Not 0.05 mM, H 2O 2 We next addressed the possibility that H 2O 2 pretreatment may induce antioxidant enzymes. GPX is a major H 2O 2-consuming enzyme. Numerous isoforms of the enzyme with different molecular masses (18 –25 kDa) have been reported in mammalian cells [18 –20]. In U937 cells, we could barely detect any of the isoforms when cells were not treated with H 2O 2 or when cells were treated with 0.05 mM H 2O 2, a concentration
FIG. 4. Effect of H 2O 2 on levels of antioxidant enzymes. Cells were treated by the indicated concentrations of H 2O 2. After 24 h, Western blotting was performed to analyze cellular levels of GPX, catalase, and SOD.
434
LEE AND UM
FIG. 5. Effects of H 2O 2 pretreatments on cellular capacity to consume H 2O 2. Cells were incubated in the absence (circles) or presence of 0.05 mM (squares) or 0.25 mM H 2O 2 (diamonds) for 24 h and were then subsequently given 1 mM H 2O 2. At the indicated times after the challenge, cellular capacity to oxidize DCFH-DA to DCF was analyzed by flow cytometry. Data are the median values of DCF fluorescence and representative of three similar experiments.
matically reduced the basal level of DCF fluorescence measured before the addition of 1 mM H 2O 2 (Fig. 5, and also see Fig. 8A to compare levels of DCF fluorescence before serum withdrawal). Moreover, that level of fluorescence was not significantly altered even after the addition of 1 mM H 2O 2. These data demonstrated that 0.25 mM H 2O 2 enhances cellular capacity to degrade endogenous and exogenous H 2O 2, a result consistent with the strong induction of GPX. Taking all these observations together, it was suggested that H 2O 2 adaptation can occur by two different mechanisms depending on the concentrations of H 2O 2. While 0.25 mM enhances cellular antioxidant capacity and thus can protect cells by degrading H 2O 2, 0.05 mM H 2O 2 protects cells in a manner independent of H 2O 2 degradation. H 2O 2 Pretreatment Suppresses Both OxidantDependent and Oxidant-Independent Cell Death The mechanisms of protection described above could allow H 2O 2 pretreatment to impart cells with resistance to both oxidant-dependent and oxidant-independent lethal stimuli. To test this possibility, we induced cell death by two alternative means; by removing serum from the culture of U937 cells or by adding C 2ceramide to the culture. Cell death by serum withdrawal was associated with a rapid elevation of cellular
DCF fluorescence (Fig. 6). The elevated DCF fluorescence was observed within 5 min of serum withdrawal and was stable for 3– 6 h, after which time it decreased. Interestingly, the addition of catalase (1000 U/ml) immediately after serum withdrawal blocked the elevation of DCF fluorescence (relative median values of DCF fluorescence, control 2.32 6 0.41; serum withdrawal 7.86 6 2.42; serum withdrawal and catalase 2.47 6 0.33, n 5 3). Such an effect was almost completely abolished when catalase inactivated by 3-amino-1,2,4-triazole (1 mM) was examined (DCF value 6.96 6 1.62). In contrast to catalase, mannitol (30 mM), a HO . scavenger, and nonspecific antioxidants such as N-acetylcysteine (5 mM) and glutathione (5 mM) did not prevent the elevation of DCF fluorescence (data not shown). Therefore, H 2O 2 appears to be the main species of ROIs produced by serum withdrawal. Consistent with this premise, the lethality by serum withdrawal was almost completely abolished by the addition of catalase, but not by the other antioxidants described above (Fig. 7). The sum of the data suggests that H 2O 2 acts as an early mediator of cell death induced by serum withdrawal. In contrast, C 2-ceramide did not elevate DCF fluorescence (Fig. 6). This was true when we performed the assay from 5 min up to 12 h after C 2-ceramide treatment. Moreover, cell death by C 2-ceramide was not suppressed by any of the antioxidants tested (Fig. 7). These data indicated that C 2ceramide kills U937 cells in an oxidant-independent mechanism. We next investigated the possibility that H 2O 2 pretreatment rescues cells from serum withdrawal and C 2-ceramide treatment. Pretreatment with 0.05 mM H 2O 2 did not block the elevation of H 2O 2 levels by serum withdrawal (Fig. 8A), but efficiently suppressed its lethality (Fig. 8B). A better protection was observed when cells were pretreated with 0.25 mM H 2O 2, which abrogated the elevation of H 2O 2 levels by serum withdrawal (Fig. 8). These responses of H 2O 2-pretreated cells to serum withdrawal are quite reminiscent of those to H 2O 2 treatments described in Figs. 2A and 5, indicating that H 2O 2 adaptation can suppress the lethal action of endogenously generated and exogenously supplied H 2O 2 in the same pattern. Pretreatment with 0.05 mM H 2O 2 also suppressed oxidant-independent cell death induced by C 2-ceramide (Fig. 8B), supporting the argument that H 2O 2 can protect cells in an antioxidant-independent manner. In this case, however, increases in the concentration of H 2O 2 up to 0.25 mM did not offer better protection. DISCUSSION
In this study, we demonstrated that pretreatment with H 2O 2 protects U937 cells from subsequent apopto-
INHIBITION OF CELL DEATH BY HYDROGEN PEROXIDE
435
FIG. 6. Serum withdrawal elevates cellular ROI levels, but the addition of C 2-ceramide does not. Cells were treated by 0.02 mM C 2-ceramide (solid line) or not (dotted line). Alternatively, cells were washed and resuspended in serum-free medium (bold line). Immediately after those treatments, cells were given DCFH-DA (50 mM). After 5 min of incubation in 37°C, cell-associated levels of DCF fluorescence were analyzed by flow cytometry. Data are representative of three similar experiments.
duced. The pretreatment also prevented the elevation of oxidant levels upon subsequent challenge with 1 mM H 2O 2 or with serum withdrawal. Therefore, enhanced antioxidant capacity appears to be a mechanism protecting cells from H 2O 2 treatment and also from H 2O 2dependent lethal stimuli such as serum withdrawal. The strong induction of GPX by 0.25 mM H 2O 2 suggests that GPX is a primary enzyme responsible for the enhanced antioxidant capacity. Under our experimental conditions, 0.25 mM H 2O 2 did not significantly alter protein levels of catalase and even reduced levels of SOD. These observations are consistent with the proposal that oxidative stresses differentially regulate antioxidant enzymes in a manner specific to experimental conditions [11]. Given that mRNA levels of SOD can be reduced by oxidative stresses [11], the downregulation of SOD protein appears to reflect transcriptional control. In contrast to 0.25 mM, 0.05 mM H 2O 2 neither induced the tested antioxidant enzymes nor significantly influenced cellular capacity to consume H 2O 2. Nevertheless, 0.05 mM H 2O 2 efficiently induced the protective response, suggesting that the protection can
tic doses of H 2O 2. The data indicate that H 2O 2 can trigger two opposing cellular pathways, one leading to cell death and another leading to cell survival. The protective effect of H 2O 2 was initially observed at 0.025 mM, while the induction of cell death minimally required 0.1 mM of H 2O 2. Thus, the survival pathway is phenotypically fourfold more sensitive to H 2O 2. The finding that a protective response can still be observed when cells are pretreated with apoptotic concentrations of H 2O 2 suggests that the survival pathway functions as a negative feedback mechanism against H 2O 2 lethality. Therefore, the cellular response to H 2O 2 appears to be determined by the balance of two contradictory signals. Clearly, the survival pathway is dominant when cells are treated with relatively low concentrations of H 2O 2, while it is overwhelmed by the death pathway when exposed to higher concentrations of H 2O 2. The H 2O 2-induced survival pathway appears to involve protein synthesis. This was first suggested by the time lag required for the induction of a protective response and was more directly indicated by the synergy between H 2O 2 and CHX to kill U937 cells. Given that micromolar concentrations of H 2O 2 induce the synthesis of 20 –25 new proteins in mammalian cells [6, 10], we have investigated the possibility that H 2O 2 suppresses cell death by multiple mechanisms. Indeed, our data suggest that there exist at least two different mechanisms which are induced by different concentrations of H 2O 2. When cells were pretreated with 0.25 mM H 2O 2, cellular levels of oxidants were greatly re-
FIG. 7. Effects of antioxidants on the lethality by serum withdrawal or C 2-ceramide treatment. Cells received C 2-ceramide (0.02 mM) or were washed and resuspended in serum-free medium. Immediately after these treatments, the indicated antioxidants were added to the concentrations: catalase (1000 U/ml), mannitol (30 mM), N-acetylcysteine (5 mM), and glutathione (5 mM). After 48 h, the viability of cells was determined. Values represent the mean of three separate experiments, and the SD was routinely less than 10% of the mean.
436
LEE AND UM
FIG. 8. Effects of H 2O 2 pretreatments on ROI generation and cell death by serum withdrawal and C 2-ceramide. (A) Cells were treated by the indicated concentrations of H 2O 2 for 24 h and then washed and resuspended in the absence or presence of serum. After 5 min, cellular levels of ROI were determined by DCF fluorescence. (B) Cells were pretreated by the indicated concentrations of H 2O 2 for 24 h and were then given 0.02 mM C 2-ceramide (C 2-ceramide) or not (control). Alternatively, the preincubated cells were washed and resuspended in serum-free medium (serum withdrawal). Cellular viability was determined 48 h after the challenge.
be achieved by an antioxidant-independent mechanism as proposed by other investigators [6]. This argument was further supported by the finding that 0.05 mM H 2O 2 can suppress oxidant-independent cell death induced by C 2-ceramide. This compound killed U937 cells
without increasing ROI levels and its lethality was not reversed by any tested antioxidants. Moreover, 0.05 mM and 0.25 mM H 2O 2 induced similar levels of protection against C 2-ceramide, implying that the antioxidant pathway induced by 0.25 mM H 2O 2 does not contribute to the protection from C 2-ceramide. Our data contrast the observations in WEHI 231 B cells where catalase was shown to antagonize the lethal action of C 2-ceramide [17]. Such differences may reflect the nature of the cell types used in these respective studies. In U937 cells, H 2O 2-induced cell death is associated with a rapid production of ceramide [21]. This would imply that ceramide functions to induce cell death at a point downstream of H 2O 2. Given that cells pretreated by 0.05 mM H 2O 2 withstand all the lethal stimuli we have tested, i.e., H 2O 2, serum withdrawal, and C 2-ceramide, the antioxidant-independent protection most likely results from a block in the pathway common to those lethal stimuli. The events leading to this block are clearly more sensitive to H 2O 2, at least under our experimental conditions, than those of the antioxidant-dependent pathway. This belief is based on our observations that relatively low concentrations of H 2O 2 predominantly induce the antioxidant-independent mechanism, whereas higher concentrations of H 2O 2 are necessary to also trigger the antioxidantdependent one. It is likely that the costimulation of two defense mechanisms by relatively high concentrations of H 2O 2 explains the superior protective effect of 0.25 mM than 0.05 mM H 2O 2, against H 2O 2 and serum withdrawal. The mechanism of antioxidant-independent protection is currently unclear. Given that Bcl-2 and Bcl-X L rescue cells from numerous lethal stimuli, we explored the possible induction of the antiapoptotic proteins by H 2O 2. However, we could not find evidence for such an induction. Although overexpression of the major heat shock protein HSP70 was shown to rescue cells from oxidative stresses and various other lethal stimuli [22– 25], H 2O 2 fails to induce heat shock proteins in U937 cells as well as in other mammalian cells [26, 27]. Moreover, H 2O 2 adaptation did not protect U937 cells from heat shock (data not shown), as reported in yeast cells [28]. Therefore, heat shock proteins do not appear to be responsible for the protective effect of H 2O 2. Cysteine proteases termed caspases have emerged as common mediators of various apoptotic stimuli [29]. It has been recently reported that H 2O 2 can inhibit a caspase 3-like activity in Jurkat T cells [30]. However, such an effect of H 2O 2 was temporary, and cells could recover a normal caspase function by 30 min of H 2O 2 treatment. Since the protective effect of H 2O 2 in our system required at least 8 h of H 2O 2 pretreatment, it is unlikely that an inhibition of caspase activity is responsible for the protective effect of H 2O 2. NF-kB is a transcription
INHIBITION OF CELL DEATH BY HYDROGEN PEROXIDE
437
factor for which activation can protect cells from numerous lethal stimuli [31–34]. In the case of U937 cells, NF-kB did not respond to 0.01 mM H 2O 2, but it was activated when cells were exposed to 0.025 mM or higher concentrations of H 2O 2 (Lee and Um, unpublished data). Given that 0.025 mM is the minimal concentration of H 2O 2 required for the protective response, it is conceivable that NF-kB may act as a mediator of H 2O 2 adaptation, inducing the protective proteins. We are currently developing the experimental tools necessary to explore this possibility. Certain carcinoma cells constitutively generate ROIs to levels higher than their normal counterparts and are highly resistant to chemotherapy and oxidative cytolysis [2, 35, 36]. Our finding that H 2O 2 can suppress cell death is consistent with the possibility that prooxidant states may provide a natural mechanism for the survival of cancer cells. Further unraveling of H 2O 2 signaling pathways, therefore, may not only provide new molecular clues into the regulatory mechanism of apoptosis but may also identify potential targets for enhancing the efficacy of cancer therapy.
11.
Shull, S., Heintz, N. H., Periasamy, M., Manohar, M., Janssen, Y. M. W., Marsh, J. P., and Mossman, B. T. (1991). Differential regulation of antioxidant enzymes in response to oxidants. J. Biol. Chem. 266, 24398 –24403.
12.
Mangan, D., Welch, G. R., and Wahl, S. M. (1991). Lipopolysaccharide, tumor necrosis factora, and IL-1b prevent programmed cell death (apoptosis) in human peripheral blood monocytes. J. Immunol. 146, 1541–1546.
13.
Bass, D. A., Parce, W., Dechatelet, L. R., Szejda, P., Seeds, M. C., and Thomas, M. (1983). Flow cytometric studies of oxidative product formation by neutrophiles: A graded response to membrane response. J. Immunol. 130, 1910 –1917.
14.
Nosseri, C., Coppola, S., and Ghibelli, L. (1994). Possible involvement of poly(ADP-ribosyl) polymerase in triggering stressinduced apoptosis. Exp. Cell Res. 212, 367–373.
15.
Yanagisawa-Shiota, F., Sakagami, H., Kuribayashi, N., Iida, M., Sakagami, T., and Takeda, M. (1995). Endonuclease activity and induction of DNA fragmentation in human myelogenous leukemic cell line. Anticancer Res. 15, 259 –266.
16.
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.
17.
Fang, W., Rivard, J. J., Ganser, J. A., LeBien, T. W., Nath, K. A., Mueller, D. L., and Behrens, T. W. (1995). Bcl-X L rescues WEHI 231 B lymphocytes from oxidant-mediated death following diverse apoptotic stimuli. J. Immunol. 155, 66 –75.
This work was supported by a Faculty Grant of Yonsei University College of Medicine for 1996. We express our thanks to Ms. Eun Sook Cho for her technical assistance. We also thank Drs. Claudette Klein and Diane Dowd for their critical comments.
18.
Avissar, N., Whitin, J. C., Allen, P. Z., Wagner, D. D., Liegey, P., and Cohen, H. J. (1989). Plasma selenium-dependent glutathione peroxidase. J. Biol. Chem. 264, 15850 –15855.
19.
Maiorino, M., Chu, F. F., Ursini, F., Davies, K. J. A., Doroshow, J. H., and Esworthy, R. S. (1991). Phospholipid hydroperoxide glutathione peroxidase is the 18-kDa selenoprotein expressed in human tumor cell line. J. Biol. Chem. 266, 7728 –7732.
20.
Chu, F.-F., Doroshow, J. H., and Esworthy, R. S. (1993). Expression, characterization, and tissue distribution of a new cellular selenium-dependent glutathione peroxidase, GSHPx-GI. J. Biol. Chem. 268, 2571–2576.
21.
Verheij, M., Bose, R., Lin, X. H., Tao, B., Jarvis, W. D., Grant, S., Birrer, M. J., Azabo, E., Zon, L. I., Kyriakis, J. M., Haimovitz-Friedman, A., Fuks, Z., and Kolesnick, R. N. (1996). Requirement for ceramide-initiated SAPK/JNK signaling in stress-induced apoptosis. Nature 380, 75–79.
22.
Jaattela, M., Wissing, D., Bauer, P. A., and Li, G. C. (1992). Major heat shock protein hsp70 protects tumor cells from tumor necrosis factor cytotoxicity. EMBO J. 11, 3507–3512.
23.
Jaattela, M., and Wissing, D. (1993). Heat-shock proteins protect cells from monocyte cytotoxicity: Possible mechanism of self-protection. J. Exp. Med. 177, 231–236.
24.
Simon, M. M., Reikerstorfer, A., Schwarz, A., Krone, C., Luger, T. A., Jaattela, M., and Schwarz, T. (1995). Heat shock protein 70 overexpression affects the response to ultraviolet light in murine fibroblasts. J. Clin. Invest. 95, 926 –933.
25.
Jaattela, M. (1995). Over-expression of hsp70 confers tumorigenicity to mouse fibrosarcoma cells. Int. J. Cancer 60, 689 – 693.
26.
Polla, B. S., Healy, A. M., Wojno, W. C., and Krane, S. M. (1987). Hormone 1a,25-dihydroxyvitamin D 3 modulates heat shock response in monocytes. Am. J. Physiol. 252, C640 –C649.
27.
Bruce, J. L., Price, B. D., Coleman, C. M., and Calderwood, S. K. (1993). Oxidative injury rapidly activates the heat shock transcription factor but fails to increase levels of heat shock proteins. Cancer Res. 53, 12–15.
REFERENCES 1. 2. 3.
4. 5.
6.
7.
8.
9.
10.
Freeman, B. A., and Crapo, J. D. (1982). Biology of disease; free radicals and tissue injury. Lab. Invest. 47, 412– 426. Toyokuni, S., Okamoto, K., Yodoi, J., and Hiai, H. (1995). Persistent oxidative stress in cancer. FEBS Lett. 358, 1–3. Farr, S. B., and Kogoma, T. (1991). Oxidative stress responses in Escherichia coli and Salmonella typhimurium. Microbiol. Rev. 55, 561–585. Demple, B. (1991). Regulation of bacterial oxidative stress genes. Annu. Rev. Genet. 25, 315–337. Davies, J. M. S., Lowry C. V., and Davies, K. J. A. (1995). Transient adaptation to oxidative stress in yeast. Arch. Biochem. Biophys. 317, 1– 6. Wiese, A. G., Pacifici, R. E., and Davies, K. J. A. (1995). Transient adaptation to oxidative stress in mammalian cells. Arch. Biochem. Biophys. 318, 231–240. Christman, M. F., Morgan, R. W., Jacobson, F. S., and Ames, B. N. (1985). Positive control of a regulon for defenses against oxidative stress and some heat shock proteins in Salmonella typhimurium. Cell 41, 753–762. Greenberg, J. T., Monach, P., Chou, J. H., Josephy, P. D., and Demple, B. (1990). Positive control of a global antioxidant defense regulon activated by superoxide-generating agents in Escherichia coli. Proc. Natl. Acad. Sci. USA 87, 6181– 6185. Wu, J., and Weiss, B. (1991). Two divergently transcribed genes, soxR and soxS, control a superoxide response regulon in Escherichia coli. J. Bacteriol. 173, 2864 –2871. Lu, D., Maulik, N., Moraru, I. I., Kreutzer, D. L., and Das, D. K. (1993). Molecular adaptation of vascular endothelial cells to oxidative stress. Am. J. Physiol. 264, C715–C722.
438 28.
29. 30.
31.
32.
LEE AND UM Collinson, L. P., and Dawes, I. W. (1992). Inducibility of the response of yeast cells to peroxide stress. J. Gen. Microbiol. 138, 329 –335. Cohen, G. M. (1997). Caspases: The executioners of apoptosis. Biochem. J. 326, 1–16. Hampton, M. B., and Orrenius, S. (1997). Dual regulation of caspase activity by hydrogen peroxide: implications for apoptosis. FEBS Lett. 414, 552–556. Beg, A. A., and Baltimore, D. (1996). An essential role for NF-kB in preventing TNFa-induced cell death. Science 274, 782–784. Wang, C.-Y., Mayo, M. W., and Baldwin, A. S., Jr. (1996). TNFand cancer therapy-induced apoptosis: Potentiation by inhibition of NF-kB. Science 274, 784 –787.
Received October 21, 1998 Revised version received January 21, 1999
33.
Van Antwerp, D. J., Martin, S. J., Kafri, T., Green, D., and Verma, I. M. (1996). Suppression of TNFa-induced apoptosis by NF-kB. Science 274, 787–789.
34.
Mayo, M. W., Wang, C.-Y., Cogswell, P. C., Rogers-Graham, K. S., Lowe, S. W., Der, C. J., and Baldwin, A. S., Jr. (1997). Requirement of NF-kB activation to suppress p53-independent apoptosis induced by oncogenic Ras. Science 278, 1812– 1815.
35.
O’Donnell-Tormey, J., DeBoer, C. J., and Nathan, C. (1985). Resistance of human tumor cells in vitro to oxidative cytolysis. J. Clin. Invest. 76, 80 – 86.
36.
Yagoda, A. (1989). Chemotherapy of renal cell carcinoma. Semin. Urol. 7, 199 –206.
Hydrogen Peroxide Suppresses U937 Cell Death by Two Different Mechanisms Depending on Its Concentration Byung Ryong Lee and Hong-Duck Um 1 Laboratory of Cell Biology, Yonsei Medical Research Center, Yonsei University College of Medicine, CPO Box 8044, Seoul, Korea
To investigate the mechanisms of H 2O 2 adaptation in mammalian cells, we exposed human U937 leukemia cells to 0.05 mM H 2O 2. This treatment significantly suppressed cell death and DNA fragmentation induced by a subsequent challenge with 1 mM H 2O 2. A more dramatic protection was observed when cells were pretreated with 0.25 mM H 2O 2. Pretreatment with either 0.05 or 0.25 mM H 2O 2 also imparted cells with a survival advantage against serum withdrawal and C 2-ceramide treatment. H 2O 2 was found to be a mediator of cell death induced by serum withdrawal, but not by the addition of C 2-ceramide. Interestingly, 0.25 mM H 2O 2 greatly induced glutathione peroxidase, a H 2O 2consuming enzyme, whereas 0.05 mM H 2O 2 did not. Consistent with observation, pretreatment with 0.25 mM H 2O 2 resulted in a great reduction of cellular oxidant levels as determined by 2*7*-dichlorofluorescein fluorescence, and it also prevented elevation of oxidant levels upon subsequent challenge with 1 mM H 2O 2 or with serum withdrawal. These effects were not observed in cells pretreated with 0.05 mM H 2O 2. The sum of the data indicated that H 2O 2 suppresses cell death by two different mechanisms depending on its concentration: Relatively high concentrations enhance cellular antioxidant capacity, and lower concentrations block the lethal action of H 2O 2. © 1999 Academic Press
Key Words: adaptation; hydrogen peroxide; oxidative stress; cell death; antioxidant.
INTRODUCTION
Cells under aerobic conditions continuously generate reactive oxygen intermediates (ROIs) such as hydrogen peroxide (H 2O 2), superoxide anion (O 2. ), and hydroxyl radical (HOz). To counter the toxicity of ROIs, cells possess protective antioxidant systems. However, the generation of ROIs can often exceed cellular defense capacity. This oxidative condition is induced by ligation 1 To whom correspondence and reprint requests should be addressed. Fax: 82-02-362-8647. E-mail: [email protected].
0014-4827/99 $30.00 Copyright © 1999 by Academic Press All rights of reproduction in any form reserved.
of cell surface receptors and also by multiple classes of environmental agents [1]. As ROIs are implicated in the etiology of numerous physiological and pathophysiological states, e.g., inflammation, aging, carcinogenesis, atherosclerosis, and neurodegeneration [2], understanding cellular responses to oxidative stress has been a major goal of diverse fields of biology. An interesting feature of oxidative stress is that it can induce adaptive responses. When bacteria are exposed first to low concentrations of H 2O 2 or O 2. -generating agents, they become resistant to high levels of the same agents which would otherwise abolish their colony formation [for review see 3, 4]. Using similar assays to measure cellular growth competence, others have shown that similar protective responses occur in yeast and mammalian cells [5, 6]. In bacteria, the protective responses are accompanied by the synthesis of 30 – 40 proteins, including enzymes involved in DNA repair and antioxidant pathways and some heat shock proteins [7, 8]. The induction of a subset of the proteins was dependent upon two distinct regulons, oxyR and soxRS, which selectively responded to H 2O 2 and O 2. , respectively [7–9]. OxyR mediates the induction of catalase which reduces H 2O 2 to water, whereas soxRS does that of superoxide dismutase (SOD) which scavenges O 2. . It appears, therefore, that bacteria achieve the protection from ROIs, at least in part, by enhancing their capacity to degrade selective members of ROIs. Compared to bacteria, the mechanism(s) by which mammalian cells adapt to ROIs has not been well characterized. It has been reported that pretreatment of mammalian cells by relatively low concentrations of H 2O 2 induces the synthesis of at least 20 new proteins [6]. Inhibition of protein synthesis by cycloheximide (CHX) suppresses the induction of resistance to a subsequent challenge with high concentrations of H 2O 2 [6]. Such data indicate that protein synthesis is required for mammalian cell adaptation. Although the nature of protective proteins is not clear yet, potential candidates may include antioxidant enzymes. Indeed, vascular endothelial and tracheobronchial epithelial
430
INHIBITION OF CELL DEATH BY HYDROGEN PEROXIDE
cells respond to H 2O 2 by enhancing their mRNA levels and the activities of antioxidant enzymes, particularly catalase and glutathione peroxidase (GPX) [10, 11]. Those observations have led many investigators to believe that antioxidant enzymes play a role in mammalian cell adaptation. However, in an elegantly established adaptive condition, pretreatment of hamster fibroblast cells by H 2O 2 did not increase either the activities or the protein levels of various antioxidant enzymes including catalase, GPX, and SOD [6]. Moreover, the adapted cells did not display an enhanced capacity to degrade H 2O 2 when subsequently challenged with high levels of H 2O 2. One simple explanation for these contradictory observations could be that mammalian cells can adapt to H 2O 2 by two different mechanisms, one dependent on antioxidant enzymes and one independent of such enzymes. To date, however, no information is available concerning the factor(s) responsible for the selective induction of those mechanisms. In the present study, we explored whether the potential two mechanisms of H 2O 2 adaptation can function in a single cell type. To do so, we pretreated human U937 leukemia cells with various concentrations of H 2O 2 and evaluated the relationship between enhancement of cellular antioxidant capacity and induction of resistance to subsequent lethal stimuli. The data indicated that H 2O 2 adapts U937 cells by dual mechanisms: One mechanism enhances the cellular capacity to degrade H 2O 2, while the other suppresses the lethal action of H 2O 2. These two protective mechanisms were distinguished by their sensitivities to H 2O 2. EXPERIMENTAL PROCEDURES Materials. U937 cells were obtained from the American Type Culture Collection (Rockville, MD). Cell culture medium and its supplements were supplied by BioWhittaker (Walkersville, MD). Antibodies against human catalase, CuZn SOD, and rat GPX were generous gifts from Dr. Ho Zoon Chae. Anti-human Bcl-2 and -Bcl-X L antibodies were purchased from DAKO (Carpinteria, CA) and Transduction Laboratories (Lexington, KY), respectively. All other reagents whose suppliers are not indicated were purchased from Sigma (St. Louis, MO). Cell culture and treatments. U937 cells were cultured in RPMI 1640 medium supplemented with 10% heat-inactivated FBS and gentamicin (50 mg/ml) at 37°C and 5% CO 2. To induce adaptive responses, cells at the density of 3.5 3 10 5/ml were exposed to 0.05 mM H 2O 2 (unless otherwise specified). At the end of the indicated incubation periods, pretreated and untreated control cells were challenged with either H 2O 2 (1 mM) or C 2-ceramide (0.02 mM) or washed in PBS and then cultured in serum-depleted medium. When cells were pretreated by 0.25 mM H 2O 2 or where indicated, viable cells were isolated by centrifugation over a layer of Ficoll–Paque (Pharmacia Biotech, Uppsala, Sweden), washed twice in PBS, and then challenged as described. Analysis of viability. Incubated cells received propidium iodide (PI) (5 mg/ml) followed by flow cytometry analysis to simultaneously
431
monitor PI uptake (FL-2 channel) and cell size (forward light scatter). To assess cell death, 10,000 cells were counted, and the cells that displayed both a reduction in cell size and a high permeability to PI were understood to be dead cells, as defined previously [12]. The percentage of cell death was calculated by the following formula: number of dead cells/number of total counted cells 3 100 and expressed as percentage of PI 1 cells. DNA fragmentation assay. A total of 2 3 10 6 cells were lysed in Triton X-100/Tris buffer [12]. The lysates were centrifuged at 13,000g. DNA fragments in the supernatant were extracted and processed for electrophoresis in 2% agarose at 40 V for 4 h [12]. Flow cytometric assay of 29,79-dichlorofluorescein oxidation. Cells were incubated in the absence or presence of serum, C 2-ceramide, or H 2O 2. After the indicated incubation times, the cells were treated for 5 min by 50 mM 29,79-dichlorofluorescein diacetate (DCFH-DA; Eastman Kodak, Rochester, NY). This compound rapidly diffuses into cells, and once within cells is hydrolyzed to oxidation-sensitive DCFH [13]. In the presence of ROIs, DCFH is oxidized to the highly fluorescent DCF, which is retained by living cells. The cell-associated levels of DCF fluorescence were analyzed by flow cytometry [13]. Western blot analysis. H 2O 2-treated and untreated control cells were lysed in Tris–HCl (40 mM, pH 8), 120 mM NaCl, 0.5 % Nonidet P-40, and protease inhibitors (2 mg/ml aprotinin, 2 mg/ml leupeptin, and 100 mg/ml PMSF). After removing cell debris by centrifugation at 13,000g for 5 min, equal amounts of proteins (40 –100 mg) were separated by 12% SDS–PAGE and then electrotransferred to Immobilon membranes (Millipore, Bedford, MA) which were subsequently blotted using the indicated antibodies and visualized by chemiluminescence (ECL; Amersham, Arlington Heights, IL).
RESULTS
H 2O 2 Induces a Survival Pathway H 2O 2 can induce either apoptosis or necrosis depending on its dose. In the case of U937 cells, apoptosis is the sole, or at least a major, mechanism of cell death triggered by 1 mM H 2O 2, while higher concentrations of H 2O 2 increase the necrotic population [14]. Concentrations of H 2O 2 lower than 0.1 mM did not significantly induce cell death under our experimental conditions (Fig. 1), although other investigators have reported cytotoxicity using similar conditions [15]. Therefore, 0.05 mM H 2O 2 was used for the following adaptation experiments. To investigate whether relatively low concentrations of H 2O 2 can adapt U937 against cell death, cells were preincubated in the absence or presence of 0.05 mM H 2O 2 for 24 h and then challenged with 1 mM H 2O 2. When preincubated without H 2O 2, cells rapidly lost their viability on addition of 1 mM H 2O 2 (Fig. 1A). Cell death was accompanied by a dramatic elevation of DNA fragmentation (Fig. 1B), confirming a previous report that 1 mM H 2O 2 kills U937 cells by apoptosis [14]. Cell death and DNA fragmentation were markedly reduced in cells that had been pretreated with 0.05 mM H 2O 2. Therefore, relatively low concentrations of H 2O 2 can trigger a survival pathway which appears to function against apoptosis. To determine a dose response for protection, we pre-
432
LEE AND UM
Time Course of the Protective Response To further establish optimal conditions for the protective response, cells were pretreated with 0.05 mM H 2O 2 for varied periods of time prior to their challenge with 1 mM H 2O 2. The protective effect of H 2O 2 was not observed in cells that had been pretreated for 4 h. Cells preincubated for 8 h displayed some protection, while
FIG. 1. H 2O 2 adaptation functions against apoptosis. (A) U937 cells were incubated in the absence (circles) or presence (squares) of 0.05 mM H 2O 2 for 24 h and then challenged with 1 mM H 2O 2 (closed symbols) or not (open symbols). The viability of U937 was determined by PI uptake at the indicated times. Values are the mean of four separate experiments with an error bar representing standard deviations. (B) After 16 h of exposure to the challenge, fragmented DNA was analyzed by agarose gel electrophoresis.
treated cells with various concentrations of H 2O 2 up to 0.25 mM. Since this relatively high concentration of H 2O 2 induced cell death (20 –30% PI 1 after 24 h), we carried out Ficoll gradient centrifugation to remove the dead cells, and then the collected viable cells received a subsequent challenge with 1 mM H 2O 2. In this experiment, Ficoll gradient was equally performed for the cells pretreated with sublethal concentrations of H 2O 2. As shown in Fig. 2A, pretreatment with 0.01 mM H 2O 2 was not protective, while pretreatment with 0.025 mM or higher concentrations of H 2O 2 offered protection in a dose-dependent manner. Therefore, induction of the survival pathway is dependent on the concentration of H 2O 2 employed in the preincubation. Moreover, the induction of protective response by high lethal concentrations of H 2O 2 indicated that as the concentration of H 2O 2 increases, H 2O 2 triggers both the survival and the death pathways.
FIG. 2. Dose response and time course of H 2O 2 adaptation. (A) Cells were exposed to the indicated concentrations of H 2O 2 for 24 h. Viable cells were isolated on a Ficoll–Paque gradient at the end of preincubation and were then challenged with 1 mM H 2O 2. Cellular viability was determined 48 h after the challenge. Open circles, pretreated but not challenged; closed circles, pretreated and challenged. (B) Cells were incubated in the absence (open circles) or presence (closed circles) of 0.05 mM H 2O 2 for the indicated periods of time and subsequently exposed to 1 mM H 2O 2 for 48 h.
INHIBITION OF CELL DEATH BY HYDROGEN PEROXIDE
FIG. 3. H 2O 2 does not induce Bcl-2 and Bcl-X L. Cells were treated by the indicated concentrations of H 2O 2 for 24 h and then the cellular levels of Bcl-2 and Bcl-X L were analyzed by Western blotting. Similar results were obtained when cells were treated by H 2O 2 for 4, 8, and 16 h.
maximal protection was seen when cells were pretreated overnight (16 –24 h) (Fig. 2B). Thereafter the protection declined. Thus, there is a time lag before the protective effect of H 2O 2 can be materialized, and once established, the protective response is transient. Requirement of Protein Synthesis The time lag likely reflects the need for macromolecular synthesis, possibly an antiapoptotic protein(s). To address this possibility, U937 cells were treated with 0.05 mM H 2O 2 in the absence or presence of 0.25 mM CHX. Neither H 2O 2 nor CHX alone significantly influenced cellular viability at the concentrations used (approximately 10% PI 1 cells at 24 h of incubation). However, their simultaneous presence resulted in a marked induction of cell death (41% PI 1). The synergy between H 2O 2 and CHX supports the premise that the H 2O 2induced survival pathway involves protein synthesis: Inhibition of the protein synthesis renders cells hypersensitive to the lethal action of H 2O 2, as is also seen in bacteria [7].
433
which efficiently induces the protective response (Fig. 4). However, treatment with 0.25 mM H 2O 2 resulted in the strong induction of two isoforms of GPX (19 and 24 kDa). The data suggested that GPX is a survival factor induced by H 2O 2, particularly with concentrations of 0.25 mM. In contrast, none of the tested concentrations of H 2O 2 significantly elevated the protein levels of catalase under our experimental conditions. Levels of SOD were not significantly altered by 0.05 mM H 2O 2, but were decreased by 0.25 mM H 2O 2. To examine the functional significance of GPX induction, we compared the capacity of control and adapted cells to consume H 2O 2. To do so, H 2O 2-pretreated and untreated control cells were exposed to 1 mM H 2O 2, and cell-associated levels of H 2O 2 were compared by DCF fluorescence. When control cells were analyzed, DCF fluorescence was immediately elevated after the H 2O 2 treatment, peaking at 20 min. After this time, the level of DCF fluorescence decreased, reflecting degradation of added H 2O 2 (Fig. 5). A similar pattern of DCF fluorescence was obtained using cells pretreated by 0.05 mM H 2O 2. It appears, therefore, that 0.05 mM H 2O 2 does not enhance cellular capacity to degrade H 2O 2. In contrast, pretreatment by 0.25 mM H 2O 2 dra-
H 2O 2 Does Not Induce Bcl-2 and Bcl-X L Bcl-2 and BCL-X L are well characterized antiapoptotic proteins. Given that overexpression of either Bcl-2 or BCL-X L can suppress H 2O 2-induced cell death [16, 17], it was of interest to ask whether the protective proteins induced by H 2O 2 pretreatment involved Bcl-2 and BCL-X L. However, neither 0.05 nor 0.25 mM H 2O 2 enhanced cellular levels of Bcl-2 and BCL-X L (Fig. 3). Cellular Antioxidant Capacity Is Enhanced by 0.25 mM, but Not 0.05 mM, H 2O 2 We next addressed the possibility that H 2O 2 pretreatment may induce antioxidant enzymes. GPX is a major H 2O 2-consuming enzyme. Numerous isoforms of the enzyme with different molecular masses (18 –25 kDa) have been reported in mammalian cells [18 –20]. In U937 cells, we could barely detect any of the isoforms when cells were not treated with H 2O 2 or when cells were treated with 0.05 mM H 2O 2, a concentration
FIG. 4. Effect of H 2O 2 on levels of antioxidant enzymes. Cells were treated by the indicated concentrations of H 2O 2. After 24 h, Western blotting was performed to analyze cellular levels of GPX, catalase, and SOD.
434
LEE AND UM
FIG. 5. Effects of H 2O 2 pretreatments on cellular capacity to consume H 2O 2. Cells were incubated in the absence (circles) or presence of 0.05 mM (squares) or 0.25 mM H 2O 2 (diamonds) for 24 h and were then subsequently given 1 mM H 2O 2. At the indicated times after the challenge, cellular capacity to oxidize DCFH-DA to DCF was analyzed by flow cytometry. Data are the median values of DCF fluorescence and representative of three similar experiments.
matically reduced the basal level of DCF fluorescence measured before the addition of 1 mM H 2O 2 (Fig. 5, and also see Fig. 8A to compare levels of DCF fluorescence before serum withdrawal). Moreover, that level of fluorescence was not significantly altered even after the addition of 1 mM H 2O 2. These data demonstrated that 0.25 mM H 2O 2 enhances cellular capacity to degrade endogenous and exogenous H 2O 2, a result consistent with the strong induction of GPX. Taking all these observations together, it was suggested that H 2O 2 adaptation can occur by two different mechanisms depending on the concentrations of H 2O 2. While 0.25 mM enhances cellular antioxidant capacity and thus can protect cells by degrading H 2O 2, 0.05 mM H 2O 2 protects cells in a manner independent of H 2O 2 degradation. H 2O 2 Pretreatment Suppresses Both OxidantDependent and Oxidant-Independent Cell Death The mechanisms of protection described above could allow H 2O 2 pretreatment to impart cells with resistance to both oxidant-dependent and oxidant-independent lethal stimuli. To test this possibility, we induced cell death by two alternative means; by removing serum from the culture of U937 cells or by adding C 2ceramide to the culture. Cell death by serum withdrawal was associated with a rapid elevation of cellular
DCF fluorescence (Fig. 6). The elevated DCF fluorescence was observed within 5 min of serum withdrawal and was stable for 3– 6 h, after which time it decreased. Interestingly, the addition of catalase (1000 U/ml) immediately after serum withdrawal blocked the elevation of DCF fluorescence (relative median values of DCF fluorescence, control 2.32 6 0.41; serum withdrawal 7.86 6 2.42; serum withdrawal and catalase 2.47 6 0.33, n 5 3). Such an effect was almost completely abolished when catalase inactivated by 3-amino-1,2,4-triazole (1 mM) was examined (DCF value 6.96 6 1.62). In contrast to catalase, mannitol (30 mM), a HO . scavenger, and nonspecific antioxidants such as N-acetylcysteine (5 mM) and glutathione (5 mM) did not prevent the elevation of DCF fluorescence (data not shown). Therefore, H 2O 2 appears to be the main species of ROIs produced by serum withdrawal. Consistent with this premise, the lethality by serum withdrawal was almost completely abolished by the addition of catalase, but not by the other antioxidants described above (Fig. 7). The sum of the data suggests that H 2O 2 acts as an early mediator of cell death induced by serum withdrawal. In contrast, C 2-ceramide did not elevate DCF fluorescence (Fig. 6). This was true when we performed the assay from 5 min up to 12 h after C 2-ceramide treatment. Moreover, cell death by C 2-ceramide was not suppressed by any of the antioxidants tested (Fig. 7). These data indicated that C 2ceramide kills U937 cells in an oxidant-independent mechanism. We next investigated the possibility that H 2O 2 pretreatment rescues cells from serum withdrawal and C 2-ceramide treatment. Pretreatment with 0.05 mM H 2O 2 did not block the elevation of H 2O 2 levels by serum withdrawal (Fig. 8A), but efficiently suppressed its lethality (Fig. 8B). A better protection was observed when cells were pretreated with 0.25 mM H 2O 2, which abrogated the elevation of H 2O 2 levels by serum withdrawal (Fig. 8). These responses of H 2O 2-pretreated cells to serum withdrawal are quite reminiscent of those to H 2O 2 treatments described in Figs. 2A and 5, indicating that H 2O 2 adaptation can suppress the lethal action of endogenously generated and exogenously supplied H 2O 2 in the same pattern. Pretreatment with 0.05 mM H 2O 2 also suppressed oxidant-independent cell death induced by C 2-ceramide (Fig. 8B), supporting the argument that H 2O 2 can protect cells in an antioxidant-independent manner. In this case, however, increases in the concentration of H 2O 2 up to 0.25 mM did not offer better protection. DISCUSSION
In this study, we demonstrated that pretreatment with H 2O 2 protects U937 cells from subsequent apopto-
INHIBITION OF CELL DEATH BY HYDROGEN PEROXIDE
435
FIG. 6. Serum withdrawal elevates cellular ROI levels, but the addition of C 2-ceramide does not. Cells were treated by 0.02 mM C 2-ceramide (solid line) or not (dotted line). Alternatively, cells were washed and resuspended in serum-free medium (bold line). Immediately after those treatments, cells were given DCFH-DA (50 mM). After 5 min of incubation in 37°C, cell-associated levels of DCF fluorescence were analyzed by flow cytometry. Data are representative of three similar experiments.
duced. The pretreatment also prevented the elevation of oxidant levels upon subsequent challenge with 1 mM H 2O 2 or with serum withdrawal. Therefore, enhanced antioxidant capacity appears to be a mechanism protecting cells from H 2O 2 treatment and also from H 2O 2dependent lethal stimuli such as serum withdrawal. The strong induction of GPX by 0.25 mM H 2O 2 suggests that GPX is a primary enzyme responsible for the enhanced antioxidant capacity. Under our experimental conditions, 0.25 mM H 2O 2 did not significantly alter protein levels of catalase and even reduced levels of SOD. These observations are consistent with the proposal that oxidative stresses differentially regulate antioxidant enzymes in a manner specific to experimental conditions [11]. Given that mRNA levels of SOD can be reduced by oxidative stresses [11], the downregulation of SOD protein appears to reflect transcriptional control. In contrast to 0.25 mM, 0.05 mM H 2O 2 neither induced the tested antioxidant enzymes nor significantly influenced cellular capacity to consume H 2O 2. Nevertheless, 0.05 mM H 2O 2 efficiently induced the protective response, suggesting that the protection can
tic doses of H 2O 2. The data indicate that H 2O 2 can trigger two opposing cellular pathways, one leading to cell death and another leading to cell survival. The protective effect of H 2O 2 was initially observed at 0.025 mM, while the induction of cell death minimally required 0.1 mM of H 2O 2. Thus, the survival pathway is phenotypically fourfold more sensitive to H 2O 2. The finding that a protective response can still be observed when cells are pretreated with apoptotic concentrations of H 2O 2 suggests that the survival pathway functions as a negative feedback mechanism against H 2O 2 lethality. Therefore, the cellular response to H 2O 2 appears to be determined by the balance of two contradictory signals. Clearly, the survival pathway is dominant when cells are treated with relatively low concentrations of H 2O 2, while it is overwhelmed by the death pathway when exposed to higher concentrations of H 2O 2. The H 2O 2-induced survival pathway appears to involve protein synthesis. This was first suggested by the time lag required for the induction of a protective response and was more directly indicated by the synergy between H 2O 2 and CHX to kill U937 cells. Given that micromolar concentrations of H 2O 2 induce the synthesis of 20 –25 new proteins in mammalian cells [6, 10], we have investigated the possibility that H 2O 2 suppresses cell death by multiple mechanisms. Indeed, our data suggest that there exist at least two different mechanisms which are induced by different concentrations of H 2O 2. When cells were pretreated with 0.25 mM H 2O 2, cellular levels of oxidants were greatly re-
FIG. 7. Effects of antioxidants on the lethality by serum withdrawal or C 2-ceramide treatment. Cells received C 2-ceramide (0.02 mM) or were washed and resuspended in serum-free medium. Immediately after these treatments, the indicated antioxidants were added to the concentrations: catalase (1000 U/ml), mannitol (30 mM), N-acetylcysteine (5 mM), and glutathione (5 mM). After 48 h, the viability of cells was determined. Values represent the mean of three separate experiments, and the SD was routinely less than 10% of the mean.
436
LEE AND UM
FIG. 8. Effects of H 2O 2 pretreatments on ROI generation and cell death by serum withdrawal and C 2-ceramide. (A) Cells were treated by the indicated concentrations of H 2O 2 for 24 h and then washed and resuspended in the absence or presence of serum. After 5 min, cellular levels of ROI were determined by DCF fluorescence. (B) Cells were pretreated by the indicated concentrations of H 2O 2 for 24 h and were then given 0.02 mM C 2-ceramide (C 2-ceramide) or not (control). Alternatively, the preincubated cells were washed and resuspended in serum-free medium (serum withdrawal). Cellular viability was determined 48 h after the challenge.
be achieved by an antioxidant-independent mechanism as proposed by other investigators [6]. This argument was further supported by the finding that 0.05 mM H 2O 2 can suppress oxidant-independent cell death induced by C 2-ceramide. This compound killed U937 cells
without increasing ROI levels and its lethality was not reversed by any tested antioxidants. Moreover, 0.05 mM and 0.25 mM H 2O 2 induced similar levels of protection against C 2-ceramide, implying that the antioxidant pathway induced by 0.25 mM H 2O 2 does not contribute to the protection from C 2-ceramide. Our data contrast the observations in WEHI 231 B cells where catalase was shown to antagonize the lethal action of C 2-ceramide [17]. Such differences may reflect the nature of the cell types used in these respective studies. In U937 cells, H 2O 2-induced cell death is associated with a rapid production of ceramide [21]. This would imply that ceramide functions to induce cell death at a point downstream of H 2O 2. Given that cells pretreated by 0.05 mM H 2O 2 withstand all the lethal stimuli we have tested, i.e., H 2O 2, serum withdrawal, and C 2-ceramide, the antioxidant-independent protection most likely results from a block in the pathway common to those lethal stimuli. The events leading to this block are clearly more sensitive to H 2O 2, at least under our experimental conditions, than those of the antioxidant-dependent pathway. This belief is based on our observations that relatively low concentrations of H 2O 2 predominantly induce the antioxidant-independent mechanism, whereas higher concentrations of H 2O 2 are necessary to also trigger the antioxidantdependent one. It is likely that the costimulation of two defense mechanisms by relatively high concentrations of H 2O 2 explains the superior protective effect of 0.25 mM than 0.05 mM H 2O 2, against H 2O 2 and serum withdrawal. The mechanism of antioxidant-independent protection is currently unclear. Given that Bcl-2 and Bcl-X L rescue cells from numerous lethal stimuli, we explored the possible induction of the antiapoptotic proteins by H 2O 2. However, we could not find evidence for such an induction. Although overexpression of the major heat shock protein HSP70 was shown to rescue cells from oxidative stresses and various other lethal stimuli [22– 25], H 2O 2 fails to induce heat shock proteins in U937 cells as well as in other mammalian cells [26, 27]. Moreover, H 2O 2 adaptation did not protect U937 cells from heat shock (data not shown), as reported in yeast cells [28]. Therefore, heat shock proteins do not appear to be responsible for the protective effect of H 2O 2. Cysteine proteases termed caspases have emerged as common mediators of various apoptotic stimuli [29]. It has been recently reported that H 2O 2 can inhibit a caspase 3-like activity in Jurkat T cells [30]. However, such an effect of H 2O 2 was temporary, and cells could recover a normal caspase function by 30 min of H 2O 2 treatment. Since the protective effect of H 2O 2 in our system required at least 8 h of H 2O 2 pretreatment, it is unlikely that an inhibition of caspase activity is responsible for the protective effect of H 2O 2. NF-kB is a transcription
INHIBITION OF CELL DEATH BY HYDROGEN PEROXIDE
437
factor for which activation can protect cells from numerous lethal stimuli [31–34]. In the case of U937 cells, NF-kB did not respond to 0.01 mM H 2O 2, but it was activated when cells were exposed to 0.025 mM or higher concentrations of H 2O 2 (Lee and Um, unpublished data). Given that 0.025 mM is the minimal concentration of H 2O 2 required for the protective response, it is conceivable that NF-kB may act as a mediator of H 2O 2 adaptation, inducing the protective proteins. We are currently developing the experimental tools necessary to explore this possibility. Certain carcinoma cells constitutively generate ROIs to levels higher than their normal counterparts and are highly resistant to chemotherapy and oxidative cytolysis [2, 35, 36]. Our finding that H 2O 2 can suppress cell death is consistent with the possibility that prooxidant states may provide a natural mechanism for the survival of cancer cells. Further unraveling of H 2O 2 signaling pathways, therefore, may not only provide new molecular clues into the regulatory mechanism of apoptosis but may also identify potential targets for enhancing the efficacy of cancer therapy.
11.
Shull, S., Heintz, N. H., Periasamy, M., Manohar, M., Janssen, Y. M. W., Marsh, J. P., and Mossman, B. T. (1991). Differential regulation of antioxidant enzymes in response to oxidants. J. Biol. Chem. 266, 24398 –24403.
12.
Mangan, D., Welch, G. R., and Wahl, S. M. (1991). Lipopolysaccharide, tumor necrosis factora, and IL-1b prevent programmed cell death (apoptosis) in human peripheral blood monocytes. J. Immunol. 146, 1541–1546.
13.
Bass, D. A., Parce, W., Dechatelet, L. R., Szejda, P., Seeds, M. C., and Thomas, M. (1983). Flow cytometric studies of oxidative product formation by neutrophiles: A graded response to membrane response. J. Immunol. 130, 1910 –1917.
14.
Nosseri, C., Coppola, S., and Ghibelli, L. (1994). Possible involvement of poly(ADP-ribosyl) polymerase in triggering stressinduced apoptosis. Exp. Cell Res. 212, 367–373.
15.
Yanagisawa-Shiota, F., Sakagami, H., Kuribayashi, N., Iida, M., Sakagami, T., and Takeda, M. (1995). Endonuclease activity and induction of DNA fragmentation in human myelogenous leukemic cell line. Anticancer Res. 15, 259 –266.
16.
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.
17.
Fang, W., Rivard, J. J., Ganser, J. A., LeBien, T. W., Nath, K. A., Mueller, D. L., and Behrens, T. W. (1995). Bcl-X L rescues WEHI 231 B lymphocytes from oxidant-mediated death following diverse apoptotic stimuli. J. Immunol. 155, 66 –75.
This work was supported by a Faculty Grant of Yonsei University College of Medicine for 1996. We express our thanks to Ms. Eun Sook Cho for her technical assistance. We also thank Drs. Claudette Klein and Diane Dowd for their critical comments.
18.
Avissar, N., Whitin, J. C., Allen, P. Z., Wagner, D. D., Liegey, P., and Cohen, H. J. (1989). Plasma selenium-dependent glutathione peroxidase. J. Biol. Chem. 264, 15850 –15855.
19.
Maiorino, M., Chu, F. F., Ursini, F., Davies, K. J. A., Doroshow, J. H., and Esworthy, R. S. (1991). Phospholipid hydroperoxide glutathione peroxidase is the 18-kDa selenoprotein expressed in human tumor cell line. J. Biol. Chem. 266, 7728 –7732.
20.
Chu, F.-F., Doroshow, J. H., and Esworthy, R. S. (1993). Expression, characterization, and tissue distribution of a new cellular selenium-dependent glutathione peroxidase, GSHPx-GI. J. Biol. Chem. 268, 2571–2576.
21.
Verheij, M., Bose, R., Lin, X. H., Tao, B., Jarvis, W. D., Grant, S., Birrer, M. J., Azabo, E., Zon, L. I., Kyriakis, J. M., Haimovitz-Friedman, A., Fuks, Z., and Kolesnick, R. N. (1996). Requirement for ceramide-initiated SAPK/JNK signaling in stress-induced apoptosis. Nature 380, 75–79.
22.
Jaattela, M., Wissing, D., Bauer, P. A., and Li, G. C. (1992). Major heat shock protein hsp70 protects tumor cells from tumor necrosis factor cytotoxicity. EMBO J. 11, 3507–3512.
23.
Jaattela, M., and Wissing, D. (1993). Heat-shock proteins protect cells from monocyte cytotoxicity: Possible mechanism of self-protection. J. Exp. Med. 177, 231–236.
24.
Simon, M. M., Reikerstorfer, A., Schwarz, A., Krone, C., Luger, T. A., Jaattela, M., and Schwarz, T. (1995). Heat shock protein 70 overexpression affects the response to ultraviolet light in murine fibroblasts. J. Clin. Invest. 95, 926 –933.
25.
Jaattela, M. (1995). Over-expression of hsp70 confers tumorigenicity to mouse fibrosarcoma cells. Int. J. Cancer 60, 689 – 693.
26.
Polla, B. S., Healy, A. M., Wojno, W. C., and Krane, S. M. (1987). Hormone 1a,25-dihydroxyvitamin D 3 modulates heat shock response in monocytes. Am. J. Physiol. 252, C640 –C649.
27.
Bruce, J. L., Price, B. D., Coleman, C. M., and Calderwood, S. K. (1993). Oxidative injury rapidly activates the heat shock transcription factor but fails to increase levels of heat shock proteins. Cancer Res. 53, 12–15.
REFERENCES 1. 2. 3.
4. 5.
6.
7.
8.
9.
10.
Freeman, B. A., and Crapo, J. D. (1982). Biology of disease; free radicals and tissue injury. Lab. Invest. 47, 412– 426. Toyokuni, S., Okamoto, K., Yodoi, J., and Hiai, H. (1995). Persistent oxidative stress in cancer. FEBS Lett. 358, 1–3. Farr, S. B., and Kogoma, T. (1991). Oxidative stress responses in Escherichia coli and Salmonella typhimurium. Microbiol. Rev. 55, 561–585. Demple, B. (1991). Regulation of bacterial oxidative stress genes. Annu. Rev. Genet. 25, 315–337. Davies, J. M. S., Lowry C. V., and Davies, K. J. A. (1995). Transient adaptation to oxidative stress in yeast. Arch. Biochem. Biophys. 317, 1– 6. Wiese, A. G., Pacifici, R. E., and Davies, K. J. A. (1995). Transient adaptation to oxidative stress in mammalian cells. Arch. Biochem. Biophys. 318, 231–240. Christman, M. F., Morgan, R. W., Jacobson, F. S., and Ames, B. N. (1985). Positive control of a regulon for defenses against oxidative stress and some heat shock proteins in Salmonella typhimurium. Cell 41, 753–762. Greenberg, J. T., Monach, P., Chou, J. H., Josephy, P. D., and Demple, B. (1990). Positive control of a global antioxidant defense regulon activated by superoxide-generating agents in Escherichia coli. Proc. Natl. Acad. Sci. USA 87, 6181– 6185. Wu, J., and Weiss, B. (1991). Two divergently transcribed genes, soxR and soxS, control a superoxide response regulon in Escherichia coli. J. Bacteriol. 173, 2864 –2871. Lu, D., Maulik, N., Moraru, I. I., Kreutzer, D. L., and Das, D. K. (1993). Molecular adaptation of vascular endothelial cells to oxidative stress. Am. J. Physiol. 264, C715–C722.
438 28.
29. 30.
31.
32.
LEE AND UM Collinson, L. P., and Dawes, I. W. (1992). Inducibility of the response of yeast cells to peroxide stress. J. Gen. Microbiol. 138, 329 –335. Cohen, G. M. (1997). Caspases: The executioners of apoptosis. Biochem. J. 326, 1–16. Hampton, M. B., and Orrenius, S. (1997). Dual regulation of caspase activity by hydrogen peroxide: implications for apoptosis. FEBS Lett. 414, 552–556. Beg, A. A., and Baltimore, D. (1996). An essential role for NF-kB in preventing TNFa-induced cell death. Science 274, 782–784. Wang, C.-Y., Mayo, M. W., and Baldwin, A. S., Jr. (1996). TNFand cancer therapy-induced apoptosis: Potentiation by inhibition of NF-kB. Science 274, 784 –787.
Received October 21, 1998 Revised version received January 21, 1999
33.
Van Antwerp, D. J., Martin, S. J., Kafri, T., Green, D., and Verma, I. M. (1996). Suppression of TNFa-induced apoptosis by NF-kB. Science 274, 787–789.
34.
Mayo, M. W., Wang, C.-Y., Cogswell, P. C., Rogers-Graham, K. S., Lowe, S. W., Der, C. J., and Baldwin, A. S., Jr. (1997). Requirement of NF-kB activation to suppress p53-independent apoptosis induced by oncogenic Ras. Science 278, 1812– 1815.
35.
O’Donnell-Tormey, J., DeBoer, C. J., and Nathan, C. (1985). Resistance of human tumor cells in vitro to oxidative cytolysis. J. Clin. Invest. 76, 80 – 86.
36.
Yagoda, A. (1989). Chemotherapy of renal cell carcinoma. Semin. Urol. 7, 199 –206.