Selective involvement of superoxide anion, but not downstream compounds hydrogen peroxide and peroxynitrite, in tumor necrosis factor-alpha-induced apoptosis of rat mesangial cells.

Tumor necrosis factor-alpha (TNF-alpha) induces reactive oxygen species (ROS) that serve as second messengers for intracellular signaling. Currently, precise roles of individual ROS in the actions of TNF-alpha remain to be elucidated. In this report, we investigated the roles of superoxide anion (O-(2)), hydrogen peroxide (H(2)O(2)), and peroxynitrite (ONOO(-)) in TNF-alpha-triggered apoptosis of mesangial cells. Mesangial cells stimulated by TNF-alpha produced O-(2) and underwent apoptosis. The apoptosis was inhibited by transfection with manganese superoxide dismutase or treatment with a pharmacological scavenger of O-(2), Tiron. In contrast, although exogenous H(2)O(2) induced apoptosis, TNF-alpha-triggered apoptosis was not affected either by transfection with catalase cDNA or by treatment with catalase protein or glutathione ethyl ester. Similarly, although ONOO(-) precursor SIN-1 induced apoptosis, treatment with a scavenger of ONOO(-), uric acid, or an inhibitor of nitric oxide synthesis, N(G)-nitro-L-argininemethyl ester hydrochloride, did not affect the TNF-alpha-triggered apoptosis. Like TNF-alpha-induced apoptosis, treatment with a O-(2)-releasing agent, pyrogallol, induced typical apoptosis even in the concurrent presence of scavengers for H(2)O(2) and ONOO(-). These results suggested that, in mesangial cells, TNF-alpha induces apoptosis through selective ROS. O-(2), but not H(2)O(2) or ONOO(-), was identified as the crucial mediator for the TNF-alpha-initiated, apoptotic pathway.


H 2 O 2 or ONOO ؊ , was identified as the crucial mediator for the TNF-␣-initiated, apoptotic pathway.
Redox reactions regulate a broad array of signal transduction pathways. Reactive oxygen species (ROS), 1 including su-peroxide anion (O 2 . ), hydrogen peroxide (H 2 O 2 ), hydroxyl radical (HO ⅐ ), and peroxynitrite (ONOO Ϫ ) are now thought of as signaling molecules that are mobilized in response to stimuli. ROS modulate Ca 2ϩ signaling and protein phosphorylation events and, thereby, function as regulators for various biological processes, including gene expression, cell growth, differentiation, chemotaxis, and apoptosis (1). ROS have been implicated in the signaling pathways initiated by tumor necrosis factor-␣ (TNF-␣). Treatment of mammalian cells with TNF-␣ triggers generation of various ROS (2)(3)(4)(5). Pharmacological experiments revealed that the mitochondrial respiratory chain is the major source of TNF-␣-induced ROS (6,7). Antioxidants inhibit various actions of TNF-␣, e.g. activation of transcription factors, gene expression, and cytotoxicity, and externally added ROS mimic its biological potential (1,8). These data support the current hypothesis that ROS serve as crucial second messengers for TNF-␣ signaling.
TNF-␣ induces apoptosis by engaging a cell surface receptor, TNF receptor 1 (TNFR1). Trimerization of TNFR1 by TNF-␣ induces association of the receptors' death domains. Subsequently, the adopter protein, namely, TNFR-associated death domain (TRADD) binds to the clustered receptor death domains through its own death domain. TRADD functions as a platform adapter that recruits Fas-associated death domain (FADD). FADD contains a death effector domain that binds to an analogous domain within procaspase-8. Upon recruitment by FADD, oligomerization of procaspase-8 drives its activation via self-cleavage and subsequently activates downstream effector caspases, leading to apoptosis (9). During this signaling process, ROS are supposed to play a critical role. It is based on the following evidence. (i) Stimulation of cells with TNF-␣ results in a rapid rise in the levels of intracellular ROS. (ii) Addition of ROS or depletion of endogenous antioxidants induces cellular death. (iii) TNF-␣-triggered cytotoxicity is inhibited by antioxidants/ROS scavengers, including thioredoxin, N-acetylcysteine, pyrrolidine dithiocarbamate, and superoxide dismutase (SOD) (8,10,11).
Oxygen normally accepts four electrons and is converted to water. In biological systems, however, partial reduction of oxygen occurs, resulting in the generation of cytotoxic ROS. That is, the sequential reduction of oxygen leads to the generation of O 2 . , H 2 O 2 , and OH ⅐ (12 HO ⅐ , and ONOO Ϫ (2-5). All of these ROS have the potential for triggering apoptosis (10,12,13). Currently, however, the involvement of and precise roles for individual ROS in mediating the apoptotic process are not well understood. It is largely unknown which compounds are required or not required for the TNF-␣-initiated apoptosis. In the present report, we dissect the roles of ROS in the TNF-␣ signaling. HO ⅐ is not required to mediate the apoptotic process in mesangial cells.

EXPERIMENTAL PROCEDURES
Cells and Transfectants-Mesangial cells (SM43) were established from isolated glomeruli of a male Harlan Sprague-Dawley rat and identified as being of mesangial cell phenotype as described before (14). Medium containing 1% fetal calf serum (FCS) was generally used for experiments.
The nuclear factor-B (NF-B)-inactive mesangial cell line SM/ IB␣M was created by overexpressing the super-repressor mutant of IB␣, IB␣M (15). SM/IB␣M cells produce IB␣M protein, which is resistant to degradation, and exhibit blunted activation of NF-B in response to interleukin 1␤ (IL-1␤) and TNF-␣ (16). As a control, mocktransfected mesangial cells that express neo alone (SM/Neo) were used.
Pharmacological Manipulation-SM/IB␣M cells (1 ϫ 10 5 /well for 24-well plates; 5 ϫ 10 5 /well for 6-well plates) cultured in the presence of Microscopic Analyses-Morphological examination was performed using phase-contrast microscopy. For fluorescence microscopy, cells were fixed with 4% formaldehyde in PBS for 10 min and stained using Hoechst 33258 (10 g/ml; Sigma) for 1 h. Apoptosis was identified using morphological criteria, including shrinkage of the cytoplasm, membrane blebbing, and nuclear condensation and/or fragmentation. Quantitative assessment was performed using both attached and detached cells.
Ladder Detection Assay-After the induction of apoptosis, both attached and detached cells were harvested and subjected to ladder detection assay, as described previously (24).
Transfection-Using the calcium phosphate co-precipitation method (25), SM/IB␣M cells cultured in 24-well plates (1.0 -1.2 ϫ 10 5 /well, 10% FCS) were co-transfected with pCIneo-catalase (26), pcDNA3-Mn-SOD (11), or an empty plasmid pcDNA3 (Invitrogen, San Diego, CA) (500 ng/well, respectively) together with pCI-␤Gal (167 ng/well; a gift from Promega, Madison, WI). pCI-␤Gal introduces a ␤-galactosidase gene under the control of the immediate-early enhancer/promoter of human cytomegalovirus. After incubation overnight, the medium was replaced with fresh medium containing 1% FCS. After 48 h, cells were treated with TNF-␣ (250 units/ml, 24 h) or H 2 O 2 (100 M, 6 h) and subjected to 5-bromo-4-chloro-3-indolyl ␤-D-galactopyranoside (X-gal) assay (27). Assays were performed in quadruplicate. The percentage of shrunk/rounded blue cells against the total number of blue cells was calculated for each well, and the mean value of four wells was used to compare data among different groups (28). . After the incubation, supernatants were collected and centrifuged, and the absorbance (550 nM) was measured using a spectrophotometer. Samples incubated in the absence of cells were used as blanks. O 2 . production was expressed as nanomoles/10 7 cells (29).
Assays were performed in quadruplicate. Statistical Analysis-Data were expressed as means Ϯ S.E. Statistical analysis was performed using the nonparametric Mann-Whitney U test to compare data among different groups. A p value of Ͻ0.05 was used to indicate a statistically significant difference.

Apoptosis of NF-B-inactive Mesangial Cells in Response to
TNF-␣-In many cell types, TNF-␣ rarely triggers apoptosis because of induction of anti-apoptotic proteins via NF-B-dependent mechanisms (30). TNF-␣-sensitive mesangial cells were created by overexpression of a super-repressor mutant of IB␣, IB␣M (19). Expression of the transgene in the established SM/IB␣M was confirmed by Northern blot analysis (Fig. 1A). To examine whether the overexpressed IB␣M is functional, parental cells (SM), mock-transfected cells (SM/ Neo), and SM/IB␣M cells were cultured in the absence or presence of IL-1␤ (10 ng/ml), and expression of NF-B-dependent genes, including endogenous IB␣ and MCP-1 (31), was examined. Under the basal culture condition, only faint expression of IB␣ and MCP-1 was observed. When stimulated by IL-1␤, expression of these genes was markedly up-regulated in the control cells, SM and SM/Neo. In contrast, the induction of IB␣ and MCP-1 was attenuated in the NF-B-inactive SM/ IB␣M cells (Fig. 1, A and B).

FIG. 2. Apoptosis of NF-B-inactive SM/IB␣M cells in response to TNF-␣.
A, phase-contrast microscopy. SM/Neo and SM/ IB␣M were exposed to TNF-␣ (250 units/ml) for 20 h, and microscopic analysis was performed. B, Hoechst staining. After exposure to TNF-␣, cells were fixed stained by Hoechst 33258, and fluorescence microscopy was performed. C, detection of the DNA ladder. Cells were cultured in the absence (Ϫ) or presence (ϩ) of TNF-␣ for 20 h and subjected to DNA ladder assay.

Selective Involvement of O 2 . in TNF-␣-induced Apoptosis
or an inhibitor of nitric oxide synthesis, L-NAME (1 mM), and stimulated by TNF-␣. Stimulation with the ONOO Ϫ precursor SIN-1 (1 mM) was used as a positive control. SIN-1 simultaneously generates NO and O 2 . , forming ONOO Ϫ (13). Exposure of the cells to SIN-1 induced round-up of the cells, as shown in Fig. 5A (bottom). Hoechst staining showed condensation and fragmentation of nuclei typical of apoptosis (Fig. 5B, bottom). Pretreatment with uric acid abolished the SIN-1-induced morphological changes. In contrast, under the same experimental condition, TNF-␣-induced apoptosis was not affected by uric acid or L-NAME (Fig. 5, A and B, top). The percentages of apoptotic cells were 0.2 Ϯ 0.2% in untreated control, 7.5 Ϯ 0.7% in TNF-␣ alone, 8.8 Ϯ 1.3% in uric acid ϩ TNF-␣, and 6.6 Ϯ 0.9% in L-NAME ϩ TNF-␣ (not statistically different) (Fig. 5C).
The lack of involvement of H 2 O 2 and ONOO Ϫ was further confirmed by concurrent scavenging of these ROS. SM/IB␣M cells were pretreated with uric acid together with catalase and stimulated by TNF-␣, H 2 O 2 , or SIN-1. Phase-contrast microscopy and Hoechst staining showed that the double ROS scavenging attenuated apoptosis induced by H 2 O 2 or SIN-1. However, the TNF-␣-induced apoptosis was not inhibited even in the concurrent presence of the H 2 O 2 scavenger and the ONOO Ϫ scavenger (Fig. 6, A and B). The percentages of apoptotic cells were 0.6 Ϯ 0.2% in untreated control, 18.9 Ϯ 2.9% in TNF-␣ alone, and 27.3 Ϯ 3.0% in uric acid/catalase ϩ TNF-␣ (Fig. 6C). Consistently, TNF-␣-triggered DNA laddering was observed regardless of the presence of uric acid and catalase (Fig. 6D). HO ⅐ is generated by reduction of H 2 O 2 or through ONOO Ϫ (12). The lack of involvement of H 2 O 2 and ONOO Ϫ in the TNF-␣-induced apoptosis was further confirmed using a HO ⅐ scavenger, Me 2 SO. SM/IB␣M cells were pretreated with Me 2 SO (20 -100 mM) and stimulated by TNF-␣. As shown in Fig. 6E, pretreatment with Me 2 SO did not affect the TNF-␣induced apoptosis. The percentages of apoptotic cells were 1.2 Ϯ 0.5% in untreated control, 19 gallol. Hoechst staining showed that pyrogallol induced round-up of the cells and condensation and fragmentation of nuclei typical of apoptosis (Fig. 7A, left). The percentage of apoptotic cells was increased from 0.6 Ϯ 0.2% to 37.8 Ϯ 5.3% by the treatment with pyrogallol (Fig. 7A, right).
To examine whether O 2 . induces apoptosis even without its conversion to H 2 O 2 or ONOO Ϫ , SM/IB␣M cells were stimulated by pyrogallol in the presence of excessive catalase (1000 units/ml) and uric acid (2 mM). Hoechst staining showed that substantial induction of apoptosis was still observed even in the concurrent presence of scavengers for H 2 O 2 and ONOO Ϫ (Fig. 7B, left). The percentages of apoptotic cells were 0.3 Ϯ 0.3% in uric acid/catalase and 11.3 Ϯ 1.1% in uric acid/catalase ϩ TNF-␣ (Fig. 7B, right). production was expressed as nanomoles per 10 7 cells. Assays were performed in quadruplicate. The asterisk indicates a statistically significant difference (p Ͻ 0.05).

DISCUSSION
TNF-␣ induces generation of ROS that may serve as second messengers for cell death signaling. Currently, precise roles of individual ROS in the cytotoxic action of TNF-␣ are not well understood. Previous reports showed that cellular sensitivity or resistance to TNF cytotoxicity is correlated with decreased or increased levels of SOD, respectively (11,32,33). These data indicated that O 2 . has a role in mediating TNF-induced cellular death. However, TNF-␣ induces both necrosis and apoptosis via ROS-dependent mechanisms (6,7,34,35 (12). HO ⅐ is known to be a highly reactive ROS and may be responsible for the oxidative damage of the cells (12). A previ-ous report showed that induction of HO ⅐ is responsible for the cytotoxic effect of TNF-␣ on tumor cells (4). However, based on our current data, the contribution of HO ⅐ to the apoptotic proc-   7). In mesangial cells, the NADPH oxidase system seems to play a crucial role in the generation of ROS. For example, mesangial cells have functional NADPH oxidase components (36), and inhibition of the NADPH oxidase system suppresses expression of chemokines in response to TNF-␣ (37). The NADPH oxidase system may participate in the apoptotic process mediated by O 2 . in TNF-␣stimulated mesangial cells.
In contrast to the knowledge on the sources of ROS, little is understood about molecular targets of ROS during the apoptotic process. Direct DNA damage or formation of oxidized lipids in cell membranes may mediate or facilitate TNF-␣induced apoptosis (10). Alternatively, particular signaling molecules may be affected by ROS directly or indirectly. A recent study suggested that apoptosis signal-regulating kinase 1 (ASK1) is a possible target of ROS (38). ASK1 is a member of the mitogen-activated protein kinase kinase kinase superfamily that activates both the c-Jun N-terminal kinase pathway and the p38 MAP kinase pathway by direct phosphorylation of MKK3, -4, -6, and -7 (39,40). ASK1 is involved in the TNF-␣induced apoptotic pathway, because ASK1 is activated by TNF-␣ in many cell types, overexpression of ASK1 induces apoptosis, and expression of dominant-negative ASK1 inhibits TNF-␣-induced apoptosis (40). Gotoh et al. (38) found that ROS induces activation of ASK1, that TNF-␣-triggered activation of ASK1 is inhibited by antioxidants, and that ROS-induced apoptosis is markedly enhanced by overexpression of ASK1. These results suggested that TNF-␣-induced activation of ASK1 is caused by ROS and contributes to the induction of apoptosis. Of note, ASK1 has a cysteine-rich domain in its N terminus. It might be a direct target for the action of ROS (38).
TNF-␣ induces apoptosis by engaging TNFR1. After ligation of the receptors, TRADD and FADD are recruited, leading to activation of downstream effector caspases (9). A recent report showed that activation of Z-VAD-sensitive caspase(s) is required for mitochondria-dependent ROS production (41). This result is consistent with another recent report that showed that staurosporine-induced apoptosis involves caspase-1-like proteases as initiators upstream of O 2 . production (42). These data indicate a possibility that the ROS-mediated process is an event downstream of serial caspase activation initiated by TNF-␣. Another putative mechanism implicated in TNF-␣-mediated apoptosis is via generation of ceramide, the hydrolyzed product of the phospholipid sphingomyelin (43). A recent study suggested that the sphingomyelin pathway has a role in TNF-␣induced ROS production. Garcia-Ruiz et al. showed that mitochondria isolated from TNF-treated cells exhibit an increase in the amount of ceramide, that addition of C 2 -ceramide to mitochondria leads ROS generation, and that blockade of the electron transport chain prevents the C 2 -ceramide-induced production of ROS (3). In contrast, another recent study indicated that ROS participate in TNF-mediated ceramide production (44), suggesting that the ROS-sensitive pathway may be located upstream of ceramide production. Further investigation is required to disclose the relationship between ROS and the sphingomyelin pathway during TNF-␣-induced apoptosis.
The differential roles of ROS in signal transduction pathways, especially in apoptotic pathways, have not been addressed before. The present study showed the selective involvement of O 2 . , but not its downstream compounds H 2 O 2 , ONOO Ϫ , and HO⅐ in TNF-␣-induced apoptosis. To our knowledge, this is the first study to demonstrate the differential roles of O 2 . and downstream compounds in this particular apoptotic process.