Nitroxides Tempol and Tempo Induce Divergent Signal Transduction Pathways in MDA-MB 231 Breast Cancer Cells*

Tempol and tempo are stable free radical nitroxides that possess antioxidant properties. In this study, we examined the effects of these compounds on components of the mitogen-activated protein kinase signal transduction cascade. Tempo treatment (15 min) of MDA-MB 231 human breast cancer cells resulted in significant levels of tyrosine phosphorylation of several as yet unidentified proteins compared with equimolar concentration of tempol (10 mm). Both compounds caused tyrosine phosphorylation and activation of Raf-1 protein kinase (30 min, 2–3-fold). Interestingly, however, only tempol caused increased extracellular signal-regulated kinase 1 activity (2 h, ∼3-fold). On the other hand, tempo, but not tempol, potently activated stress-activated protein kinase (2 h, >3-fold). Consistent with these data, tempol was found to be noncytotoxic, whereas tempo induced apoptotic cell death (2 h, >50%). Tempo treatment also resulted in significant elevation of ceramide levels at 30 min (54% over control) and 1 h (71% over control) posttreatment, preceding stress-activated protein kinase activation and apoptosis. These data suggest that in the absence of an environmental oxidative stress, tempol and tempo elicit distinct cellular signaling pathways. The recognition of the molecular mechanisms of nitroxide action may have important implications for biological effectiveness of these compounds.

A cellular antioxidant defense system composed of enzymes such as catalase, superoxide dismutase, glutathione peroxidase, and glutathione protects cells against toxic oxygen metabolites. Exogenously added free radical scavengers have also been shown to alleviate the deleterious effects of oxygen free radicals (1)(2)(3)(4). Nitroxide compounds, including tempol and tempo ( Fig. 1), are low molecular weight, membrane permeable, stable free radicals that are electron paramagnetic resonance detectable (5) and have been used classically as probes for biophysical and biochemical processes; they have been used as paramagnetic contrast agents in NMR imaging (6,7), as probes for membrane structure (8), and as sensors of oxygen in biological systems (9). However, over the past few years, novel applications of nitroxides have been demonstrated. Nitroxides have been shown to possess antioxidant activity and protect cells against a variety of agents that impose oxidative stress, including superoxide, hydrogen peroxide, and ionizing radiation (10 -20). A variety of chemical mechanisms have been proposed to account for nitroxide antioxidant activity, including superoxide dismutase mimic activity (21), oxidation of reduced metals that would otherwise catalyze the formation of hydroxyl radicals (10) from hydrogen peroxide (10), catalase mimic activity (22), radical-radical interactions (13), and detoxification of secondary organic radicals (13). Although significant research has been conducted at the whole cell level and in animals with nitroxides, little is known at the molecular level of how this novel class of antioxidants affects signal transduction pathways.
Members of the mitogen-activated protein kinase (MAPK) family, including ERKs 1 (p42/44 MAPKs), the stress-activated protein kinases (SAPKs) (also called c-Jun NH 2 -terminal kinases (p46/54 JNKs/SAPK1)), and p38 MAPK (also termed reactivating kinase (p38RK)), are activated in response to a variety of cellular stresses, such as changes in osmolarity and metabolism, DNA damage, heat shock, ischemia, UV radiation, ionizing radiation, or inflammatory cytokines . In many of these instances, free radicals and derived species play an important role in initiating a cellular signal transduction response (45). Unlike the ERK signaling pathway, which primarily promotes growth and proliferation/survival, the SAPK and p38 MAPK pathways result in growth arrest and apoptotic or necrotic cell death. Because nitroxides protect against diverse oxidative insults and may have utility in clinical biomedical research, we have investigated the effects of tempol and tempo on MAPK signal transduction pathways in an attempt to better understand their mechanism of action. Evidence presented here demonstrates that tempol and tempo stimulate distinct pathways of the MAPK signaling cascade. Tempol stimulated the ERK activity and was noncytotoxic, whereas tempo induced ceramide generation, SAPK/JNK activation, and apoptotic death of MDA-MB 231 human breast cancer cells. The cytotoxic effect of tempo was also observed in two other cancer cell types, PCI-04A laryngeal squamous carcinoma cells and PC-3 prostate cancer cells. These findings provide new insight into the mechanism of action of nitroxide antioxidants, which will be valuable in understanding their biological effectiveness.
Cell Culture, Treatments with Tempol and Tempo, and Preparation of Cell Lysates-MDA-MB 231 human breast cancer cells were grown to near confluence in 75-cm 2 tissue culture flasks in improved minimum essential medium (Cellgro) containing 10% fetal bovine serum and 2 mM L-glutamine in a humidified atmosphere of 5% CO 2 :95% air at 37°C. Cells were trypsinized and plated onto a 150-mm tissue culture dish (two dishes per flask) overnight in medium containing 10% fetal bovine serum followed by two washes with phosphate-buffered saline (PBS). Cultures were maintained in serum-free medium overnight prior to tempol (10 mM) or tempo (10 mM) treatment. Both nitroxide radicals were dissolved in ethanol before use. For the extraction of whole cell lysates (WCLs), cells with or without nitroxide treatment were washed three times with ice-cold PBS containing 0.5 M sodium orthovanadate (Na 3 VO 4 ) and lysed in lysis buffer (50 mM HEPES, pH 7.5; 1% Nonidet P-40; 10% glycerol; 4 g/ml each of leupeptin, aprotinin, and pepstatin A; 1 mM Na 3 VO 4 ; 1 mM phenylmethylsulfonyl fluoride; 25 mM sodium fluoride (NaF); and 0.5 mM EDTA). WCLs were agitated for 1 h at 4°C and centrifuged in a microcentrifuge at 15,000 ϫ g at 4°C for 15 min to remove cellular debris. The supernatant was aliquoted and stored at Ϫ70°C until use.
Immunoprecipitation and Immunoblotting-Whole cell lysate (1 mg) was immunoprecipitated with the appropriate agarose-conjugated antibody (1 g/ml of lysis buffer) overnight at 4°C with constant agitation. For SAPK immunoprecipitation, WCLs (1 mg) were immunoprecipitated with anti-SAPK antibody (5 g/ml) overnight followed by addition of protein A-agarose (50 l of 250 l/ml stock) and incubation for 2 h at 4°C. Immune-complex beads were collected by microcentrifugation at 15,000 ϫ g for 5 min followed by three washes with lysis buffer. The beads were resuspended in 2ϫ electrophoresis sample buffer and boiled for 5 min, and proteins were resolved by 10% SDS-PAGE and transferred to an Immobilon-P membrane. The membrane was blocked with 4% bovine serum albumin in PBS-Tween (0.25%) and immunoblotted with the desired primary antibody at 1:2000, followed by 1:10,000 dilution of an appropriate horseradish peroxidase-coupled secondary antibody. The immunoreactive protein bands were revealed by ECL detection system (Amersham Pharmacia Biotech). The bands of interest were quantified by ImageQuant software version 3.3 (Molecular Dynamics Personal Densitometer, Sunnyvale, CA). Prior to reprobing, blots were stripped according to the enhanced chemiluminescence kit protocol (NEN Life Science Products), as described earlier (43).
Raf-1 Kinase Assay-Raf-1 protein kinase activity was measured by a kinase cascade assay according to the manufacturer's procedure (Up-state Biotechnology), with the following modifications. Briefly, Raf-1 immune-complex was washed 3 times with lysis buffer and once with kinase binding buffer (KBB). This was followed by incubation of the immune-complex for 30 min at 30°C in reaction mixture containing 20 l of KBB, 10 l of 0.5 mM ATP/Mg mixture (75 mM magnesium chloride, and 500 M ATP in KBB), 1.6 l of inactive MAPK kinase (0.4 g), and 4 l of inactive MAPK (1 g). At the end of the reaction, 8 l of the sample mixture was transferred to a fresh 1.5-ml microcentrifuge tube, followed by the sequential addition of 10 l KBB, 10 l of myelin basic protein (MBP) substrate (2 mg/ml stock), and 10 l of [␥-32 P]ATP (1 Ci/l generated by 1:10 dilution of the stock 3000 Ci/mmol (NEN Life Science Products) in ATP/Mg 2ϩ mixture). This reaction mixture was incubated for 10 min at 30°C. The immune-complex was then pelleted by brief centrifugation in a benchtop microcentrifuge, and 5 l of the sample was spotted, in triplicate, onto P81 paper. The radioactive filters were transferred onto a 50-ml conical tube (20 filters per tube) and washed four times with 40 ml of 0.75% phosphoric acid (15 min each), followed by a brief acetone wash, and counted using Beckman LS 1801 scintillation counter.
Additionally, the immune-complex-associated Raf-1 activity was measured using Syntide-2 as a substrate. This reaction was initiated by sequential addition of 15 l of KBB, 10 l of ATP/Mg 2ϩ mixture, 5 l of Syntide-2 (5 g), and 10 l of diluted [␥-32 P]ATP followed by incubation of the reaction at 30°C for 20 min. At the end of the incubation, reaction mixture was centrifuged briefly in a benchtop microcentrifuge, and 5 l of the supernatant was spotted in triplicate onto P81 filter paper, air dried, washed, and counted as described above.
ERK and SAPK/JNK Activities-Whole cell lysates prepared as described above were immunoprecipitated (1 mg) with an agaroseconjugated anti-ERK1 antibody or an agarose-conjugated anti-JNK1 antibody for 2 h at 4°C with constant agitation. The immune-complexes were washed three times in lysis buffer and once in KBB as mentioned earlier. ERK or JNK activity assay was carried out according to manufacturer's procedures (Upstate Biotechnology). Briefly, the ERK1 immunoprecipitates were incubated for 10 min at 30°C in a kinase reaction containing 10 l of MBP as substrate (2 mg/ml stock), 10 l of inhibitor mixture (20 M protein kinase C inhibitor peptide, 2 M protein A inhibitor peptide, and 20 M compound R24571), and 10 l of magnesium-ATP mixture (1 Ci of [␥-32 P]ATP generated by 1:10 dilution of stock (3000 Ci/mmol) in 75 mM magnesium chloride and 500 m cold ATP). The immune-complexes were centrifuged briefly in a benchtop centrifuge, and 5-l aliquots of the supernatant were spotted in triplicate onto P81 filter papers. The radioactive filters were washed and counted as described above. Alternatively, to visualize the incorporation of ␥-32 P into MBP, the kinase reaction was stopped by addition of 2ϫ electrophoresis sample buffer and boiled for 5 min, and proteins were resolved by 15% SDS-PAGE, followed by autoradiography. For JNK activity assay, JNK1 immunoprecipitates were incubated for 30 min at 30°C in 40 l of kinase reaction mixture containing 10 l of KBB, 20 l of the GST-c-Jun fusion protein (0.2 g/l stock), and 10 l of the diluted [␥ Ϫ32 P]ATP as described above. The kinase reaction was stopped with 2ϫ electrophoresis sample buffer and boiled for 5 min, and the supernatant was electrophoresed by 12.5% SDS-PAGE. The radiolabeled GST-c-Jun fusion protein was detected by autoradiography.
Cell Viability Assay-Effects of nitroxide compounds on cell viability and proliferation were determined using a cell viability detection kit (4-[3-(4-iodophenyl)-2-(4-nitrophenyl)-2H-5-tetrazolio]-1,3-benzene disulfonate, WST-1) according to the manufacturer's instructions (Boehringer Mannheim). Briefly, MDA-MB 231 cells were seeded onto 96-well plates at density of 10,000 cells/well and maintained overnight in 10% fetal bovine serum-containing medium. The cells were then washed twice with phosphate buffered saline and refed in serum-free medium. The following day, cells were treated for various times with tempol (10 mM) or tempo (10 mM), using six wells per treatment condition. At the end of treatment, medium containing the nitroxide compound was removed and replaced with fresh serum-free medium (100 l), followed by addition of WST-1 (10 l). Plates were incubated for 2 h at 37°C and analyzed at A ϭ 450/600 using a MR 700 microplate reader.
Apoptosis Assay-ApoAlert Annexin V apoptosis detection system (CLONTECH, Palo Alto, CA) was used to measure the relative distribution of apoptotic and necrotic cells. Briefly, cells were seeded at a density of 1 ϫ 10 6 cells per 25-cm 2 tissue culture flask in medium containing 10% fetal bovine serum overnight, followed by washing twice in serum-free medium. Cells were maintained overnight in serum-free medium and exposed to tempol (10 mM) or tempo (10 mM) for various times. This was followed by rinsing twice with serum-free medium prior to trypsinization, dilution in two volumes of serum-free medium, and centrifugation at 10,000 ϫ g. The cell pellet was washed once with PBS Ceramide Generation Assay-Ceramide production in MDA-MB 231 cells was determined by diacylglycerol (DAG) kinase assay according to a previously described procedure (46,47). Briefly, MDA-MB 231 cells were split (1:2), and after 24 h, the cells were washed twice with PBS and serum-free medium was added, followed by incubation for additional 24 h. The cells (ϳ2 ϫ 10 6 /60-mm dish) were treated with tempo (10 mM) or tempol (10 mM) for various times. Following treatment, floating cells were collected and pelleted by centrifugation for 10 min at 1200 rpm, and attached cells were collected by scraping. Lipids were extracted from all cells (floating and attached) by incubation in 1 ml of 100% ice-cold methanol. After a partial purification with chloroform, the extracted lipid in the organic phase was dried under N 2 and was treated with a mild alkaline solution (0.1 N KOH in methanol) for 1 h at 37°C to remove glycerolphospholipids. The organic phase extract was resuspended in 20 l of 7.5% n-octyl-␤-D-glucopyranoside, 5 mM cardiolipin, 1 mM EDTA followed by the addition of 40 l of purified DAG kinase in DAG kinase buffer (20 mM Tris-HCl (pH 7.4), 10 mM dithiothreitol, 1.5 M NaCl, 250 mM sucrose, 15% glycerol). The kinase reaction was initiated by the addition of 20 l of diluted [␥-32 P]ATP (10 mM at 1,000 dpm/pmol in DAG kinase buffer) and incubated for 30 min at 22°C. This reaction was terminated by extraction of lipids with 1 ml of CHCl 3 :CH 3 OH:HCl (100:100:1), 170 l of buffered saline solution (135 mM NaCl, 1.5 mM CaCl 2 , 0.5 mM MgCl 2 , 5.6 mM glucose, and 10 mM HEPES, pH 7.2), and 30 l of 100 mM EDTA. The lower organic phase containing ceramide 1-phosphate was collected and dried under N 2 followed by spotting and run of 40 l (80%) onto a thin layer chromatographic (Whatman silica gel 150A) plate and developing in chamber containing CHCl 3 :CH 3 OH:HAc (65:15:5, v/v) as solvent. The spot containing the ceramide 1-phosphate was visualized by autoradiography, and the incorporated 32 P was removed by scraping and quantified by Cerenkov counting. A standard curve consisting of a known amount of ceramide was used as a comparison to the level of observed ceramide generated in MDA-MB 231 cells. Fig. 2 illustrates a conspicuous increase in the tyrosine phosphorylation of several as yet unidentified protein bands within 15 min after the exposure of MDA-MB 231 cells to 10 mM tempo. These levels remained elevated for the duration of the study (2 h). In parallel experiments, minimal protein tyrosine phosphorylation was observed at various times (15 min to 2 h) following the treatment of cells with an equimolar concentration of tempol. These data show that although both nitroxides induced protein tyrosine phosphorylation, the magnitude of this response was clearly higher in tempo-treated cells.

Effects of Tempol and Tempo on Protein Tyrosine Phosphorylation-
Tempol and Tempo Stimulate Tyrosine Phosphorylation and Activity of Raf-1 in Vivo-Previously, we demonstrated that ionizing radiation, a well known stress-inducing agent, causes tyrosine phosphorylation of Raf-1 in MDA-MB 231 breast cancer cells (43). Here, we examined the possibility of tyrosine phosphorylation and activation of Raf-1 protein kinase in response to tempol and tempo. Interestingly, both tempol and tempo treatments led to an increase in the level of tyrosine phosphorylated Raf-1 (Raf P ) (Fig. 3A, top panel). The level of total Raf-1 protein remained unchanged (Fig. 3A, bottom panel). The immunoreactive Raf P bands were quantified. Densitometric analysis indicated that increase in the level of Raf P detected at 15 min was ϳ5-8-fold, and Raf P content was comparable to the basal level by ϳ60 -120 min (data not shown). The activity of Raf-1 protein kinase was determined by a kinase cascade assay or by the Syntide-2 phosphorylation assay (Fig.  3B). In agreement with the enhanced tyrosine phosphorylation of Raf-1, tempol or tempo treatment resulted in ϳ2-3-fold increase in the Raf-1 protein kinase activity.
Tempol Stimulates ERK Activity-Because Raf-1 activation generally leads to ERK (p42/44 MAPK) activation, we examined the effects of tempol and tempo on ERK1 enzymatic activity. Representative experiments are shown in Fig. 4. An approximately 3-fold increase in the enzymatic activity of ERK1 was detected by 2 h in cells treated with tempol (Fig. 4A). Interestingly, however, no change in ERK1 activity was noted following tempo treatment compared with control cells (Fig.  4A). In addition, ERK1 phosphorylation was seen as a shift to a more slowly migrating phosphorylated form (ERK1 P ) on immunoblots using ERK1 immunoprecipitates at 2 h after tempol exposure (10 mM), but we were unable to identify a shift in the mobility of ERK1 in tempo-treated cells (data not shown).
Tempo Treatment Results in Enhanced Phosphorylation and Activation of SAPK in Vivo-We next measured the effects of tempol and tempo on SAPK/JNK, a well-known component of the stress-induced signal transduction pathway. The time course experiments indicated that tempo treatment resulted in a significant increase in the level of phosphorylated SAPK (ϳ54 kDa, SAPK P ) compared with tempol treatment or untreated controls (Fig. 5, A and B). Consistent with these data, SAPK enzymatic activity was significantly induced in tempo-treated cells as shown by the level of phosphorylated GST-c-Jun (Fig.  5C). Densitometric analysis of three independently performed experiments indicated a 3-7-fold increase in the phosphorylated GST-c-Jun fusion protein detectable after tempo exposure (10 mM, 2 h) compared with tempol (10 mM, 2 h) or control (1% ethanol, 2 h) treatment.
Tempo Induces Apoptotic Cell Death-Several studies have reported that activation of the SAPK signaling cascade is associated with induction of apoptotic cell death (32). To examine the possible cytotoxic effects of tempo, we first used a colorimetric assay to determine the cell viability and proliferation. Treatment of cells with 10 mM tempo resulted in Ͼ50% decrease in the number of viable cells within 2 h. In parallel experiments, the number of viable cells in cultures treated with 10 mM tempol was comparable to control cells treated with 1% ethanol (Fig. 6). These observations prompted us to evaluate whether decrease in the number of viable cells following tempo treatment was due to apoptotic and/or necrotic cell death.
Apoptosis is a process of cell death characterized by cytoplasmic shrinkage, nuclear condensation, and DNA fragmentation (48). Several reports suggest that an early event leading to apoptosis is accompanied by a loss of cell membrane phospholipid asymmetry as a result of translocation of phosphatidylserine from the intracellular membrane to the extracellular membrane while leaving the cell membrane intact (49). A phos- FIG. 2. Effects of tempol and tempo on protein tyrosine phosphorylation. Subconfluent cultures were grown in serum-free medium overnight, followed by treatment with tempol (10 mM) or tempo (10 mM) for the indicated times and lysis. Normalized protein contents (1 mg) were immunoprecipitated with agarose-conjugated anti-PY mAb and then immunoblotted with anti-PY mAb. Data shown are representative of two independent experiments. UT, untreated cells grown overnight in serum-free medium.
phatidylserine-binding protein, annexin V, has been used as a specific probe to detect externalization of this phospholipid in a variety of murine and human cell types undergoing apoptosis (50,51). Cell necrosis, on the other hand, is associated with both the translocation of phosphatidylserine to the external cell surface and the loss of membrane integrity (52). The cell membrane integrity of apoptotic cells can be established with a dye exclusion test using PI. In the following experiments, we used FITC-conjugated annexin V and PI as markers for the evaluation of apoptosis and necrosis. MDA-MB 231 cells treated with tempol or tempo were double-labeled with FITC-conjugated annexin V and PI and then subjected to flow cytometric analysis. Representative cytogram analysis of MDA-MB 231 cells with or without nitroxide compound is shown in Fig. 7A. The lower left quadrant represents viable cells (V), which were negative for annexin V and PI. The lower right quadrant represents apoptotic cells (A), which were positive for annexin V staining. The upper right quadrant represents necrotic cells (N), which were positive for both annexin V and PI stains. Tempo treatment (10 mM for 2 h) resulted in a significant increase in both the annexin V uptake (52.42% apoptotic cells) and the annexin V plus PI uptake (10.90% necrotic cells) compared with tempol (10 mM for 2 h) (6.32% apoptotic cells, 1.31% necrotic cells) and control cells (1% ethanol for 2 h) (6.05% apoptotic cells, 1.06% necrotic cells). Time course analysis indicated that tempo treatment resulted in a steady increase in the number of apoptotic cells for upto 2 h, followed by a considerable increase in the number of necrotic cells by 3 h (Fig.  7B). Tempol treatment did not induce apoptosis or necrosis for the duration of the study (3 h) (Fig. 7B). These data suggest that tempo-stimulated SAPK phosphorylation and activation may be associated with apoptotic cell death in MDA-MB 231 cells.
To determine the generality of the cytotoxic effect of tempo in cancer cells, we have examined two other cancer cell lines: PCI-04A, a human laryngeal squamous carcinoma-derived cell line (53), and PC-3, a human prostate cancer cell line. The data shown in Fig. 7, C and D, demonstrate a significant level of apoptosis and necrosis at 2 h post-tempo treatment (10 mM) in PCI-04A cells. In PC-3 cells, 10 mM tempo treatment resulted in ϳ84% necrotic cells by 2 h, implying that this treatment condition was highly toxic (Fig. 7, E and F). Tempo also induced apoptotic cell death in bovine aortic endothelial cells, as measured by the bisbenzamide trihydrochloride/Hoechst-33258 staining method, as previously described (46)   UT, untreated cells grown in serum-free medium and treated with 1% ethanol. B, Raf-1 protein kinase activity was measured either by a kinase cascade "read-out" assay (top) or using the Syntide 2 phosphorylation assay (bottom). Cells were grown in serum-free medium overnight, followed by treatment with tempol (10 mM) or tempo (10 mM) for indicated times and lysis. WCLs (1 mg) were immunoprecipitated with agarose-conjugated anti-Raf-1 antibody. For the coupled-kinase cascade reaction, Raf-1 immune-complexes were first incubated at 30°C for 30 min with 5 nmol of [␥-32 P]ATP, inactive MAPK kinase (0.4 g), and inactive MAPK (1 g) in 40 l kinase reaction buffer. MBP (20 g) was then added to the first reaction mixture (8 l), and the reaction was continued at 30°C for 10 min in 30 l of kinase reaction buffer. MBP phosphorylation was quantified using a filter binding assay as described under "Experimental Procedures." For the Syntide 2 phosphorylation assay, Raf-1 immune-complexes were incubated at 30°C for 20 min with 5 nmol of [␥-32 P]ATP and 5 g of Syntide-2 in 40 l of kinase reaction buffer, and Syntide-2 phosphorylation was quantified using a filter binding assay. Data shown are mean Ϯ S.D. from two or three independent experiments. Control, cells were grown overnight in serum-free medium and treated with 1% ethanol for 1 h (top) or 2 h (bottom).
FIG. 4. Tempol stimulates ERK1 activity. Cells were grown in serum-free medium overnight, treated with tempol (10 mM) or tempo (10 mM) for 2 h, and lysed. WCLs (1 mg) were immunoprecipitated with agarose-conjugated anti-ERK1 antibody, and in vitro MBP phosphorylation assay was performed as described under "Experimental Procedures." The incorporation of ␥-32 P into MBP was determined in a filter binding assay (A). In other independent experiments, the reaction products were electrophoresed by 15% SDS-PAGE, and MBP (at ϳ18 kDa) was visualized by autoradiography (B). Cont/Control, cells grown overnight in serum-free medium and treated with 1% ethanol for 2 h.
Ceramide Generation in Tempo-treated MDA-MB 231 Cells-Ceramide, a second messenger molecule generated as a result of hydrolysis of the plasma membrane phospholipid sphingomyelin or via de novo synthesis, has been implicated in a variety of biological responses to environmental cues (54). Increase in ceramide has been correlated with increased SAPK/ JNK activity, and ceramide and SAPK/JNK have been shown to participate in a signal transduction pathway leading to cell death (26,44,55,56). To assess the possibility of a role of ceramide in tempo-induced SAPK and apoptosis, we used a DAG kinase assay to quantify the ceramide levels in MDA-MB 231 cells treated with or without the nitroxide compound. A 54% increase over control (normalized to 100%) in ceramide level was observed at 30 min, and ceramide level reached 71% over control at 1 h post-tempo treatment (Fig. 8). The level of ceramide generated in tempol-treated cells was not significantly higher compared with control cells at all time points. A significant increase in ceramide level was also noted in tempotreated (5 mM) bovine aortic endothelial cells, but with slightly different kinetics (data not shown). Ceramide production preceded maximal stimulation of SAPK/JNK and apoptosis, implying its involvement in tempo-induced signaling in MDA-MB 231 cells.

DISCUSSION
This study reports, for the first time to our knowledge, signal transduction mechanisms of cellular response to two nitroxides, tempol and tempo, that are well known for their antioxidant properties. Initially, we hypothesized that because ERK pathway is used by a wide variety of cell types for transducing survival or proliferative signals, the antioxidant effects of tempol and tempo may be complemented by stimulation of the ERK signaling pathway. Previous in vitro studies suggested that at least 5-10 mM tempol is required to provide radioprotection, and a protection factor as high as 2.2 was achieved with 100 mM tempol (15). Our data showing activation of ERK1 by tempol (10 mM) are consistent with these and other reports of a protective role of tempol against radiation-induced mutagenicity and double strand breaks and hydrogen peroxide-induced mutagenicity (15,(57)(58)(59). Surprisingly, however, tempo (10 mM) had no detectable effect on ERK1 activity, suggesting that a dissociation may also exist between ERK signaling and antioxidant activity of certain nitroxides.
Enhanced protein tyrosine phosphorylation, generation of ceramide, activation of SAPK, and induction of apoptosis by tempo are unexpected and novel observations. One possibility for further evaluation is that there may be differential intracellular reduction rate of tempol versus tempo. In this situation, tempo-treated cells may have higher tempo free radical concentration. Free radicals, as second messengers, would then find appropriate cellular targets and turn on a signaling pathway. In this context, it is noteworthy that addition of plateletderived growth factor to vascular smooth muscle cells results in increased intracellular levels of hydrogen peroxide and reactive oxygen species, and these events have been correlated with platelet-derived growth factor-induced tyrosine phosphorylation, MAPK stimulation, and DNA synthesis (60). In other reports, induction of protein tyrosine phosphorylation in neutrophils is dependent on NADPH oxidase activation (61), and stimulation of as yet unidentified protein tyrosine kinases has been linked to apoptotic death of B-lymphocytes (62). The short time required to observe the apoptosis (2 h) (Fig. 7) suggests that cell cycle, DNA synthesis, or significant transcription/ translation may not be a prerequisite for tempo-initiated cell death. It seems possible that posttranslational modification of existing proteins required for the induction of apoptosis is regulated by a free radical-mediated protein kinase pathway(s) involving SAPK.
Endogenous sphingolipid metabolites, such as ceramides and sphingosines, have been recognized as lipid mediators of cell growth, differentiation, and apoptosis (46, 63-67). Apoptosis FIG. 5. Tempo stimulates tyrosine phosphorylation and activity of SAPK. Cells were grown in serum-free medium overnight and treated with tempol (10 mM) or tempo (10 mM) for the indicated times, followed by lysis. A, WCLs (1 mg) were immunoprecipitated with agarose-conjugated anti-SAPK antibody, followed by immunoblotting with anti-PY mAb. B, in other independent experiments, anti-SAPK immunoprecipitates were first probed with anti-PY mAb (top), and then the blot was then reprobed with anti-SAPK antibody (bottom). C, cells were grown in serum-free medium overnight and treated with tempol (10 mM) or tempo (10 mM) for 2 h. WCLs (1 mg) were immunoprecipitated with anti-JNK1 antibody, and the JNK1 activity in immunoprecipitates was measured using GST-c-Jun (ϳ41 kDa) as a substrate. UT, untreated cells grown overnight in serum-free medium; CONT/Control, cells grown in serum-free medium overnight and treated with 1% ethanol for 2 h.
FIG. 6. Effects of tempol and tempo on cell viability. Cells were grown in serum-free medium overnight in 96-well plates and treated with tempol (10 mM) or tempo (10 mM) for indicated times, followed by removal of medium containing the nitroxide compound. Control cells were grown overnight in serum-free medium, followed by treatment with 1% ethanol for various times. Fresh serum-free medium (100 l) was added to each well in all plates, including controls followed by the addition of WST-1 (10 l). Plates were incubated for 2 h at 37°C, and the color solution developed by WST-1 was quantified using a MR 700 microplate reader at A ϭ 450/600. Values shown are mean Ϯ S.D. of six determinations per treatment condition in a representative experiment, and the experiment was repeated three times.
has been suggested to be dependent on or independent of ceramide release (26,44,68,69), and more recently, ceramide has been shown to interact with mitochondria, leading to generation of reactive oxygen species (70). In other studies, activation of a family of cysteine proteases with specificity for aspartic acid residues, also known as caspases, has been tightly linked with apoptotic cell death, and this pathway involves the release of cytochrome C from mitochondria (71,72). Whether ceramide generation in tempo-treated cells is due to activation of sphingomyelinase and/or ceramide synthase or tempo-treatment results in the activation of caspases are important issues currently under investigation in our laboratory.
What components are upstream of ERK and SAPK in the tempol-and tempo-initiated signaling, respectively? Tempol and tempo are uncharged nitroxides in the physiological pH range and readily cross the cell membrane; however, their concentrations in subcellular compartments differ. Tempo is approximately 200 times more lipophilic than tempol (73); hence, tempo would be expected to accumulate in the cell membrane to a greater extent than tempol. Thus, having an agent such as tempo (a stable free radical) localized in the cell membrane and capable of participating in redox reactions may  Logarithmically growing cells were cultured in serum-free medium overnight in 60-mm dishes and treated with tempol (10 mM) and tempo (10 mM) for indicated times, followed by lipid extraction and quantitation of ceramide by DAG kinase assay as described under "Experimental Procedures." The organic phase extract containing the ␥-32 P-labeled ceramide was quantitated. Control cells were grown overnight in serum-free medium, followed by treatment with 1% ethanol for various times ranging from 0.5 to 2 h. Tempol/tempo treatment values shown are the mean Ϯ S.D. of triplicate determinations; the value at each time point is normalized to control (100%). initiate a signal transduction cascade distinct from tempol, which is more water soluble and more evenly distributed throughout the cell. Although both tempol and tempo stimulated Raf-1, ERK1 activity was increased only in tempoltreated cells. Raf activation was short-lived compared with ERK1.
Raf-1 activity peaked at 30 min, whereas ERK activity began to rise at 15 min and continued to rise for at least 120 min. This lack of correlation between the kinetics of Raf-1 activation and ERK activation has been observed earlier (29,43) and may be due to multiple effectors, including Raf-1, upstream of ERK. At present, the significance of Raf-1 activity in a nitroxide-induced response is unclear. Mitogen-activated ERK kinase, a known physiological substrate of Raf-1 and activator of ERK (74 -79), and stress-activated kinase kinase 1, a potent activator of SAPK (34,35,38,41,80), are other potential upstream targets. The regulation of MAPKs, including ERK and SAPK/JNK, involves sequential phosphorylations, often initiated at the cell surface by a receptor or nonreceptor protein tyrosine kinase(s). Other reports have suggested a balance between ERK and SAPK activities as a determinant of cell survival or cell death (31). Based on a significant increase in protein tyrosine phosphorylation within 15 min after tempo treatment compared with tempol, it is plausible to speculate the activation of a lipid-mediated signaling pathway that involves proapoptotic protein tyrosine kinase(s) in tempo-treated cells.
In conclusion, the present studies provide evidence that (a) tempo induces a significant tyrosine phosphorylation of several as yet unidentified proteins as compared with tempol, (b) tempol and tempo stimulate tyrosine phosphorylation and activity of Raf-1 protein kinase, (c) tempol stimulates MAPK (ERK) activity, whereas tempo is a potent inducer of SAPK phosphorylation and activity, (d) tempo, but not tempol, induces apoptotic cell death, and (e) tempo-induced cell death could be associated with ceramide generation in MDA-MB 231 cells. Our findings imply that in the absence of an environmental oxidative stress, such as that induced by ionizing radiation, nitroxides tempol and tempo stimulate distinct signal transduction pathways, perhaps triggered by secondary radicals associated with cellular metabolism and differentially regulated by early events, such as the control of protein tyrosine phosphorylation and generation of ceramide.
The MAP kinase pathway is a widely used signal transduction mechanism that initiates proliferation. Hyperexpression of MAP kinase has been localized to malignant breast epithelium and metastatic cells of patients with breast cancer (81). Identification of compounds activating a cell death pathway(s) should then lead to their rational use in cancer therapy. The finding that tempo induces apoptosis in different cell types warrants further study. It is most interesting that an agent that exerts antioxidant activity can also induce cytotoxicity by apoptosis. Should there be a differential induction of apoptosis in human tumor versus normal cells, the use of tempo may have clinical utility. Studies are presently under way in our laboratories to explore this possibility.