Cellular Non-heme Iron Content Is a Determinant of Nitric Oxide-mediated Apoptosis, Necrosis, and Caspase Inhibition*

In this report, we tested the hypothesis that cellular content of non-heme iron determined whether cytotoxic levels of nitric oxide (NO) resulted in apoptosis versus necrosis. The consequences of NO exposure on cell viability were tested in RAW264.7 cells (a cell type with low non-heme iron levels) and hepatocytes (cells with high non-heme iron content). Whereas micromolar concentrations of the NO donor S -nitroso- N -acetyl- DL -penicilla- mine induced apoptosis in RAW264.7 cells, millimolar concentrations were required to induce necrosis in hepatocytes. Caspase-3 activation and cytochrome c release were evident in RAW264.7 cells, but only cytochrome c release was detectable in hepatocytes following high dose S -nitroso- N -acetyl- DL -penicillamine ex- posure. Pretreating RAW264.7 cells with FeSO 4 increased intracellular non-heme iron to levels similar to those measured in hepatocytes and delayed NO-induced cell death, which then occurred in the absence of caspase-3 activation. Iron loading was also associated with the formation of intracellular dinitrosyl-iron complexes (DNIC) upon NO exposure. Cytosolic preparations containing

It is now clear that nitric oxide (NO) 1 can either induce necrosis or apoptosis or protect cells from death. The consequence of NO exposure on cells depends on a number of poorly characterized factors. Cell type certainly plays an important role. Macrophages, thymocytes, neuronal cells, pancreatic islets, and some tumor cells are very sensitive to NO and undergo apoptosis (1) or necrosis upon exposure to even low levels of NO (2). Other cell types, such as hepatocytes (3,4), human B lymphocytes (5), endothelial cells (6), cardiac myocytes (7), splenocytes (8), and ovarian follicles (9) are resistant to NO toxicity. Many cell types that are resistant to NO toxicity can even be protected from either type of death by physiologically relevant levels of NO. The level of NO exposure is another key factor in determining the effect of NO on cell viability. For example, low concentrations of NO prevent apoptosis in serumstarved PC12 cells while high concentrations lead to necrotic cell death (10).
As part of the study of the role of NO in apoptosis, an interesting relationship between NO and caspase activity has emerged. Caspases are a family of cysteine proteases that play an essential role in the signaling cascade leading to apoptosis. Upon exposure to a proapoptotic signal, zymogen forms of caspases constitutively present in cells became proteolytically cleaved and activated. Initiator caspases such as caspase-8, -9, and -10 can cleave other caspases, while executioner caspases, including caspase-3, -6, and -7, cleave death substrates (11,12). By inhibiting caspase activity through the S-nitrosylation of the cysteine within the active site of the enzyme, NO inhibits apoptosis in hepatocytes (3,4), endothelial cells (6), and several tumor cell lines (13). All caspases tested have been shown to be susceptible to reversible inhibition by NO through this redox modification (14). However, the observation that NO-mediated apoptosis in macrophages, thymocytes, and several tumor cells involves the activation of caspase (15) and consequent caspasedependent cell death has created an unexplained paradox. How can NO inhibit cell death by blocking caspase activity in some cells while inducing cell death through activation of caspases in others? We postulated that one explanation would come from understanding the chemical fate of NO in cells. In order for NO to S-nitrosylate proteins, it must give up an electron to favor the formation of a nitrosonium (NO ϩ )-like species (16). Potential electron acceptors in cells include non-heme iron and oxygen (17,18). Oxygen availability should not differ in vitro between cells susceptible to NO-induced cell death and cells protected from cell death by NO. Because it has been shown that increased cellular non-heme iron levels protect cells from NO-induced cytotoxicity in cultured hepatocytes (19), we focused our attention on the role of cellular non-heme iron content in NO cytotoxicity. We addressed this question using hepatocytes, a cell type protected from apoptosis by NO through the efficient S-nitrosylation of caspases (3,4) and RAW264.7 cells, a murine macrophage cell line susceptible to NO-induced apoptosis (15). We report here that cellular non-heme iron levels determine whether NO inhibits caspase activity and whether NO induces apoptosis or necrosis.
Preparation of Dinitrosyl Iron Complexes (DNIC)-DNIC-containing cytosol was prepared from hepatocytes or RAW264.7 cells treated with 400 M SNAP for 5 h. The cells were harvested, washed twice with ice-cold PBS, and lysed by three cycles of freezing and thawing. The cytosol was obtained by centrifugation at 100,000 ϫ g for 30 min. This solution was dialyzed at 4°C for 6 h and concentrated using a Centricon-3 concentrator (Amicon, Danvers, MA). In addition, DNIC were also chemically synthesized as described previously (20). In brief, degassed solutions of FeSO 4 (5 mg/ml) and neutralized L-cysteine (72 mM) (the molar ratio of Fe 2ϩ to L-cysteine was 1:20) were mixed in a Thunburgtype reaction vessel under pure NO gas (P NO ϭ 500 mm Hg). NO was added 4 min before mixing. The solution immediately turned green. After 1 min, the solution was evacuated for 2 min by a high vacuum to remove excess NO. The solution was immediately frozen and stored in liquid nitrogen until use.
Cell Culture-Purified rat hepatocytes were isolated from male Harlan Sprague-Dawley rats (200 -250 g; Harlan Sprague-Dawley) by a collagen perfusion method (21). Highly purified hepatocytes (Ͼ98%) were obtained by repeated centrifugation at 400 ϫ g followed by further purification over 30% Percoll. Hepatocytes with Ͼ95% viability by trypan blue exclusion were suspended in Williams' medium E supplemented with 1 M insulin, 2 mM glutamine, 100 units/ml penicillin, 100 g/ml streptomycin, and 10% low endotoxin calf serum. Cells were plated in 100-mm Petri dishes (5 ml/dish) and 12-well plates (1 ml/well) at a concentration of 1 ϫ 10 6 and 2 ϫ 10 5 cells/ml, respectively, and precultured in a CO 2 incubator (5% CO 2 , 95% air) at 37°C for 16 h. The RAW264.7 macrophage cell line was maintained with complete Williams' medium E containing 10% calf serum. RAW264.7 cells were plated in 100-mm Petri dishes or 12-well plates with a concentration of 1 ϫ 10 6 cells/ml and then incubated in a CO 2 incubator for 8 h. For some experiments, RAW264.7 cells were pretreated in complete Williams' medium E containing 80 M FeSO 4 for 24 h, and then cells were washed with fresh medium. Cell cultures were treated with different concentrations of SNAP or 2,000 units/ml TNF␣ and 0.2 g/ml ActD in some experiments.
Cell Viability-Cell viability was determined by the crystal violet staining method, as described previously (22). In brief, cells were stained with 0.5% crystal violet in 30% ethanol and 3% formaldehyde for 10 min at room temperature. Plates were washed four times with tap water. After drying, cells were lysed with 1% SDS solution, and dye uptake was measured at 550 nm using a 96-well plate reader. Cell viability was calculated from relative dye intensity compared with untreated samples.
Western Blot Analysis-Cells were harvested, washed twice with ice-cold PBS, and resuspended in 20 mM Tris-HCl buffer (pH 7.4) containing a protease inhibitor mixture (0.1 mM phenylmethylsulfonyl fluoride, 5 g/ml aprotinin, 5 g/ml pepstatin A, and 1 g/ml chymostatin). Cytosolic fractions were prepared for cytochrome c release by homogenization and differential centrifugation, as described previously (11). Whole cell lysates were prepared for immunoblotting analysis of caspase-3 and PARP by sonication and centrifugation at 13,000 ϫ g for 20 min at 4°C. Proteins (20 g) were separated on 8% SDS-PAGE for PARP and 14% SDS-PAGE for caspase and cytochrome c and then transferred to nitrocellulose. The membranes were hybridized with cytochrome c or caspase-3 antibody, and protein bands were visualized by exposing to x-ray film, as described previously (22).
Cytosolic DNA Extraction and Electrophoresis-Cultured hepatocytes were washed with PBS, harvested using a plastic scraper, and pelleted by centrifugation at maximum speed in a microcentrifuge for 10 s at 4°C. Fragmented cytosolic DNA was prepared by the method of Leist et al. (23). Briefly, the pellets were resuspended in 740 l of lysis buffer (20 mM Tris-HCl, 10 mM EDTA, 0.5% Triton X-100, pH 8.0) and occasionally shaken while on ice for 40 min. The cytosolic fraction was collected by centrifugation at 12,000 ϫ g for 20 min at 4°C. Cytosol aliquots containing equal amounts of protein were extracted with a mixture of phenol and chloroform. One-tenth volume of 3 M sodium acetate was added to the solution, and DNA was precipitated by adding an equal volume of isopropyl alcohol. After storing at 20°C overnight, a DNA pellet was obtained by centrifugation at 13,000 ϫ g for 15 min at 4°C and washed twice with 75% ethanol. The pellet was dried and resuspended in 80 l of 20 mM Tris-HCl, pH 8.0. After digesting RNA with RNase (0.1 mg/ml) at 37°C for 40 min, samples (15 l) were electrophoresed through a 1.2% agarose gel in 450 mM Tris borate plus EDTA buffer (pH 8.0). DNA was photographed under visualization with UV light.
Assay for Enzyme Activity and Cleavage of PARP and Inhibitor of Caspase-activated DNase (ICAD)-Cell suspensions in 100 mM Hepes (pH 7.4) containing the protease inhibitor mixture were lysed by three cycles of freezing and thawing. The crude cytosol was obtained by centrifugation at 13,000 rpm for 20 min at 4°C. Recombinant human caspase-3 (8 ng) was mixed with DNIC-containing cytosol (200 g of protein) or chemically synthesized DNIC in Hepes buffer (100 mM Hepes, pH 7.4), 140 mM NaCl, and the protease mixture in a final volume of 20 l and incubated at room temperature. The caspase-3-like activity was measured by colorimetric assay using the peptide-based substrate Ac-DEVD-p-nitroanilide, as described previously (4). For the in vitro cleavage assay of PARP and ICAD, 35 S-labeled PARP and ICAD were prepared by coupled in vitro transcription/translation with T7 polymerase and purified PARP or ICAD gene-containing plasmid in a reticulocyte lysate system (Promega, Madison, WI). The labeled protein (6 l) was mixed with the cytosolic fraction in a final volume of 18 l in 100 mM Hepes (pH 7.4). The mixture was incubated at 37°C for 40 min and then mixed with an equal volume of 2ϫ SDS-sample buffer. The cleaved products were separated on SDS-PAGE. After drying, the gel was exposed to x-ray film at room temperature. Lactate dehydrogenase (LDH) activity was measured in culture medium. For total LDH activity, cell lysate was prepared from control cells in the culture by adding 1 ⁄100 volume of 10% Triton X-100 into the culture medium, incubating at room temperature for 15 min, and centrifuging at 80 ϫ g for 10 min. All samples were dialyzed against PBS at 4°C overnight. The LDH activity was measured by using an automated procedure on a Technitron RA-500 autoanalyzer.
Determination of Non-heme Iron, Heme Iron, DNIC, and S-Nitrosothiols-Total iron was determined using a colorimetric method after acid-permanganate treatment according to Fish (24). Heme iron was analyzed by the pyridine-chromogen method (25). Non-heme iron was calculated from the difference between total iron and heme iron. Using EPR spectroscopy, DNIC were detected in SNAP-treated cells. Cells were harvested, washed twice with ice-cold PBS, and resuspended in a small volume of PBS. Cell concentration was adjusted by measuring total protein using a Lowry protein assay kit (P5656; Sigma). An equal volume of cell suspension was placed in a sample tube and frozen in liquid nitrogen for EPR spectroscopy. Quantitative cellular DNIC were determined by the intensity of the EPR spectrum at g ϭ 2.04 with the standard curve of chemically synthesized DNIC. EPR examination was performed at 77 K as described previously (26). For S-nitrosothiol measurement, cells were harvested, washed twice with ice-cold PBS, and lysed by three cycles of freezing and thawing. Cell extract was prepared by centrifugation at 4,000 rpm for 5 min at 4°C. NO bound to thiols was displaced by the addition of HgCl 2 to the cell lysates (4). Following NO displacement, protein was removed by centrifugation with a Centricon-3 concentrator (Amicon). Total S-nitrosothiols were calculated by measuring the difference in nitrite plus nitrate concentration before and after the addition of HgCl 2 . Nitrite plus nitrate was measured with an NO analyzer (NO-2000, Ki-Woo Biotech). Nitrite and nitrate were not detectable in cell lysates without HgCl 2 addition.
Induction of Apoptosis in a Cell-free Reconstitution System-The cytosolic S-100 fraction and nuclei were isolated as described previously (4). The reaction mixture contained 40 l of S-100 (ϳ10 mg/ml), 10 l of nuclei solution (ϳ1 ϫ 10 6 nuclei), and 400 ng of rh-caspase-3 pretreated with DNIC in a final volume of 80 l of reconstitution buffer (10 mM Hepes, pH 7.4, 40 mM ␤-glycerophosphate, 50 mM NaCl, 2 mM MgCl 2 , 4 mM EGTA, 2 mM ATP, 10 mM creatine phosphate, 50 g/ml creatine kinase, and 0.2 mg/ml bovine serum albumin). The combinations were incubated at 37°C for 140 min and occasionally mixed. The reaction solution was then mixed with 500 l of buffer A (50 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1% SDS, and 0.2 mg/ml of proteinase K) and incubated at 37°C for 1 h. The solution was extracted with phenol/chloroform. DNA isolation and electrophoresis were carried out as described previously (4).
Other Analyses-Protein concentration was determined with the BCA assay (Pierce). Data are presented as means Ϯ S.D. of at least three separate experiments except where results of blots are shown, in which case a representative experiment is depicted in the figure. Comparisons between two values were analyzed using Student's t test. Differences were considered significant at p values Յ 0.05.

RESULTS
High Concentrations of NO Are Required for Hepatic Necrosis-Cell types differ considerably in their sensitivity to the cytotoxic actions of NO, and in many instances exposure to high levels of NO is required for either apoptosis (15,27) or necrosis (2). Concentrations of NO generated by the inducible NO synthase inhibit apoptosis in hepatocytes (3,4); however, it is unknown whether higher levels will lead to cell death. Isolated rat hepatocytes were treated with increasing concentrations of the NO donor SNAP for 24 h, and cell viability was measured. No change in cell viability was observed at SNAP concentrations up to 1 mM, but higher concentrations reduced cell viability, reaching 80% cytotoxicity at 4 mM SNAP (Fig. 1A). Oxidized SNAP (4 mM) did not affect hepatocyte viability, confirming that the cytotoxic factor released from the donor was NO. This cell death was not accompanied by DNA fragmentation (Fig. 1B); however, a known inducer of apoptosis, TNF␣ plus ActD (TNF␣/ActD), resulted in hepatocyte death (Fig. 1, B and C) and DNA fragmentation (Fig. 1D). Medium levels of LDH were measured as a marker of cell lysis and necrosis. SNAP concentrations that caused cell death significantly increased LDH activity in culture medium in a dose-dependent manner, while TNF␣/ActD resulted in much lower LDH release (Fig. 1E) associated with greater cell death (Fig. 1, B and C).
These results indicate that levels of NO generated by millimolar concentrations of an NO donor induce necrosis, not apoptosis, in cultured rat hepatocytes.
SNAP Induces Apoptotic Cell Death in RAW264.7 Cells-Similar studies were carried out in the macrophage cell line RAW264.7, a cell type known to be more sensitive to NOinduced toxicity. The cells were treated with increasing concentrations of SNAP for 18 h. Concentrations of SNAP between 0.5 and 1.25 mM were associated with cell death ( Fig. 2A) accompanied by DNA fragmentation characteristic of apoptosis (Fig. 2B). Oxidized SNAP (1 mM) did not cause cell death or DNA fragmentation.
NO Induces Cytochrome c Release but Does Not Activate Caspase-3 in Hepatocytes-Caspase-3-like activity is known to increase in apoptotic hepatocytes (4). No increase in caspase-3-like activity was observed in SNAP-treated hepatocytes. Interestingly, high concentrations of SNAP reduced the basal activities even lower, and this was reversed by incubating the cytosol with 5 mM DTT for 20 min, suggesting that the caspase was S-nitrosylated following NO exposure (Fig. 3A). As expected, cytosol from hepatocytes treated with TNF␣/ActD had increased caspase-3-like activity at 8 h, but DTT had no effect on activity (Fig. 3B). Consistent with the observations on caspase-3 activity, PARP, a well known caspase-3 substrate, was cleaved into 85-kDa fragments in the cells as well as by the cytosol from TNF␣/ActD-treated hepatocytes (Fig. 3, C and D), but not in SNAP-treated cells. Furthermore, the active 17-kDa fragment of caspase-3 was identified only in the TNF␣/ActDtreated cells (Fig. 3E). Another cellular event associated with cell death is the release of cytochrome c from mitochondria. Therefore, we next examined whether mitochondrial cytochrome c was released into cytosol using Western blot analysis. Treatment with 2 mM SNAP or TNF␣/ActD both resulted in cytochrome c release into cytosol (Fig. 3F).
NO Induces Cytochrome c Release and Caspase-3 Activation in RAW264.7 Cells-We next determined if the NO-induced apoptotic cell death observed in RAW264.7 cells was associated with caspase-3 activation. Fig. 4A shows that SNAP increased caspase-3-like activity in a dose-dependent manner. Incubation of the cytosol with DTT caused only small increases in activity, indicating that caspases were not S-nitrosylated. PARP was cleaved in SNAP-treated RAW264.7 cells as well as by the cytosol from these cells (Fig. 4, B and C). Western blot analysis confirmed the appearance of the active fragment (p17) of caspase-3 only in the SNAP-treated cells (Fig. 4D). Caspase-3 activation occurred in association with mitochondrial cytochrome c release into cytosol (Fig. 4E). Thus, in clear distinction from necrotic death observed in hepatocytes, where caspases are not activated but cytochrome c is released, NO induces both caspase-3 activation and cytochrome c release in apoptotic RAW264.7 cells. inhibitor Ac-DEVD-cho protects cells from caspase-dependent apoptotic cell death (4,28). The treatment of hepatocytes with Ac-DEVD-cho did not prevent cell death induced by 2 mM SNAP; however, removal of NO using a known NO scavenger, red blood cells (RBC; 4 mM of hemoglobin), did prevent cell death (Fig. 5A). As expected, hepatocyte apoptosis induced by TNF␣/ActD was completely inhibited by Ac-DEVD-cho (Fig.  5A). Ac-DEVD-cho also completely inhibited caspase-3-like activity (Fig. 5B) and DNA fragmentation in TNF␣/ActD-treated hepatocytes (Fig. 5C). In RAW264.7 cells, Ac-DEVD-cho partially inhibited SNAP-induced RAW264.7 cell death, while NOscavenging RBC completely inhibited SNAP-induced cell death (Fig. 6A). Both Ac-DEVD-cho and RBC prevented the increase in caspase-3-like activity (Fig. 6B) and completely inhibited SNAP-induced DNA fragmentation (Fig. 6C). These data confirm the finding that NO induces necrosis in hepatocytes and indicates that the cell death in RAW264.7 cells most likely is a combination of apoptosis and necrosis.
Intracellular Iron Regulates Apoptotic Cell Death by Inhibition of Caspase Activity-Hepatocytes are iron-rich compared with RAW264.7 cells, and it is known that NO readily reacts with non-heme iron to form iron-nitrosyl complexes (21,29). These complexes could, in turn, protect cells from NO-induced toxicity by either scavenging NO (29) or converting NO to a potent S-nitrosylating species (10,20). To determine if iron content accounted for the difference in NO sensitivity between hepatocytes and the macrophage cell line, we preloaded RAW264.7 cells with iron. Preincubation of RAW264.7 cells with 80 M FeSO 4 for 24 h increased intracellular non-heme iron to a level comparable with hepatocytes but had no effect on the level of heme iron (Fig. 7A). Upon exposure to 400 M SNAP for 4 h, DNIC (showing the typical EPR spectrum at g ϭ 2.04) and cellular S-nitrosothiols increased in the iron-loaded RAW264.7 cells by 3-fold compared with the control RAW264.7 cells. The levels of these NO adducts in the iron-loaded cells were comparable with the levels measured in SNAP-treated hepatocytes (Fig. 7, B and C). Loaded iron increased cellular DNIC levels in a dose-dependent manner, which directly correlated with cell survival after exposure to 750 M SNAP (Fig.  7D). At 15 h after SNAP addition, there was no detectable cell death in the iron preloaded RAW264.7 cells (Fig. 8A). Furthermore, DNA fragmentation was absent (Fig. 8B). At this time point, there was no increase in caspase-3-like activity (Fig. 8C) or caspase-3 cleavage, and cytochrome c release was absent (Fig. 8D). Importantly, incubation of the cytosol with DTT did not increase caspase-3-like activity (Fig. 8C). By 21 h, some cell death was evident in the iron-loaded cells (Fig. 8A); however, this death took place with almost no DNA fragmentation (Fig.  8B) and with markedly reduced levels of caspase-3-like activity (Fig. 8C), caspase-3 cleavage, and cytochrome c release (Fig.  8D) when compared with the non-iron-loaded RAW264.7 cells exposed to SNAP. DTT did significantly increase caspase-3-like activity in the cytosol from iron-loaded RAW264.7 cells at 21 h (Fig. 8C). These data suggest that iron loading prevents apoptotic death by both removing NO and by converting NO to an S-nitrosylating species.
DNIC Inhibit Caspase Activity-We (4, 13) and others (6) have shown that NO or an NO ϩ -like species prevents caspase activity by modifying the active site cysteine in the enzyme. NO as part of DNIC possesses partially positive charged properties (20), allowing DNIC to act as nitrosating intermediates or alternatively promote the conversion of NO to nontoxic nitrate (30). To examine if cellular DNIC inhibit caspase activity through nitrosylation, we measured caspase-3 activity in a mixture of rh-caspase-3 and cytosol from hepatocytes and RAW264.7 cells exposed to SNAP for 5 h (Fig. 9A). Cytosol from SNAP-treated hepatocytes inhibited caspase-3 activity, while cytosol from SNAP-treated RAW264.7 cells did not. However, cytosol from the iron-supplemented RAW264.7 cells exposed to SNAP contained inhibitory activity (Fig. 9A). The inhibition of caspase activity by cytosol from SNAP-treated hepatocytes or iron-loaded RAW264.7 cells was reversed by DTT. Formation of DNIC in iron-loaded RAW264.7 cells rapidly increased to ϳ2.0 nmol/mg protein until 5 h following SNAP treatment and then slowly decreased to ϳ0.6 nmol/mg protein at 25 h. The intracellular DNIC levels could be restored to over 90% of peak levels by reexposing the cells to SNAP (data not shown). The inhibitory effect of the cytosol from the cells on caspase-3 activity was highly correlated with the level of DNIC (Fig. 9B). Under the same experimental conditions, the half-life of SNAP was ϳ5.4 h (Fig. 9B), and the half-life was not significantly different between cell types (hepatocytes, RAW264.7 cells, and iron-loaded RAW264.7 cells; data not shown). Chemically synthesized DNIC inhibited caspase-3 activity in a dose-and timedependent manner (Fig. 10, A and B). DNIC (200 M) also inhibited caspase-3-dependent cleavage of ICAD (Fig. 10C) and caspase-3-dependent DNA fragmentation in an in vitro reconstitution system (Fig. 10D). The inhibitory effects of DNIC on caspase activity were reversed by the addition of DTT to the DNIC plus caspase mixture (Fig. 10D). The effects of DNIC were similar to the inhibition seen with SNAP alone. Taken together, these data indicate that DNIC have S-nitrosylating activity.

DISCUSSION
This study was undertaken to identify factors that determine whether NO causes necrosis or apoptosis in cells. We focused our attention on the level of intracellular non-heme iron and its relationship to caspase activity. We found that hepatocytes (cells that are NO-resistant) undergo necrosis when exposed to FIG. 5

. Effects of caspase inhibitor on cell viability, caspase-3-like activity, and DNA fragmentation in SNAPand TNF␣/ActD-treated hepatocytes.
Hepatocytes were treated with 2 mM SNAP or TNF␣ (2,000 units/ml)/ActD (0.2 g/ml) in the presence or absence of Ac-DEVD-cho (280 M) or RBC (4 mM hemoglobin). Cell viability (A), caspase-3-like activity (B), and DNA fragmentation (C) were measured by the same methods as described in the legends to Figs. 1 and 3.   , n ϭ 4). B, DNA fragmentation was analyzed by agarose gel electrophoresis after isolation of cytosolic DNA. C, caspase-3-like activity was measured by colorimetric assay following preincubation with or without 5 mM DTT for 20 min. D, caspase-3 activation and cytochrome c release were analyzed by Western blot after separation of cytosolic proteins on SDS-PAGE. *, p Ͻ 0.01 versus without DTT. millimolar concentrations of an NO donor. Caspase activation could not be identified, and these cells were found to have high levels of non-heme iron. In contrast, RAW264.7 cells underwent apoptotic death when exposed to micromolar concentrations of SNAP, and this was associated with caspase-3 activation. More importantly, elevation of non-heme iron in the RAW264.7 cells to levels comparable with that seen in hepatocytes resulted in a significant delay in cell death, which then appeared to be necrosis instead of apoptosis. The delay in cell death was associated with an inhibition of NO-induced caspase-3 activation initially and then an increase in S-nitrosylation of caspase-3-like enzymes. The inhibition of caspase activity and delay in cell death correlated with the formation of DNIC in the cells, and DNIC was able to inhibit caspase activity and caspase-3-mediated DNA fragmentation in the in vitro reconstitution system. These results identify the level of nonheme iron as an important factor in determining the consequence of NO exposure on cell viability. Not only may nonheme iron effectively remove NO, but the formation of DNIC may also regulate caspase activity in cells by facilitating Snitrosylation of the critical -SH in caspase enzymes. The capacity of NO to efficiently S-nitrosylate caspases could dictate whether cells are protected from NO and whether cells, if not protected, undergo apoptosis versus necrosis.
NO toxicity is dependent on the redox environment of target cells and the rate of NO production. NO can induce apoptosis in some tumor cells, and this cell death can be blocked by increasing cellular redox potential (31). NO-mediated cell death requires a high rate of NO generation rather than a large total amount of NO production. Recent evidence shows that NO donors with long half-lives protect cells from apoptotic cell death, but short half-lived NO donors induce apoptotic cell death (32) in the presence of the same concentrations of NO donors, which indicates that the rate of NO production is a critical factor for NO-mediated cell death. In our experiments, the rate of NO release from SNAP (750 M) was ϳ70 M/h, which was not significantly different in hepatocytes and RAW264.7 cells, and the half-life of SNAP was ϳ5.4 h in these culture conditions. Also, there was no difference in the half-life of SNAP in SNAP concentrations from 100 to 1,000 M (data not shown). After a 24-h incubation with 750 M SNAP, 730 M SNAP (97%) had decomposed to release NO. After a 16-h incubation, the rate of NO production was ϳ1.5 M/h. Therefore, additional SNAP was added into the culture medium at this point to induce toxicity (Fig. 8A). The kinetics of NO release from SNAP was not different in RAW264.7 cells with or without iron preloading. However, both iron-rich hepatocytes and iron-loaded RAW264.7 cells were more resistant than iron-poor macrophage RAW264.7 cells to SNAP-mediated cell death, indicating that cellular iron content is an important factor for regulating NO-mediated cell death. Our data suggest that the formation of DNIC is determined by iron content and that DNIC can divert NO into nontoxic or protective pathways.
Participation of caspase family proteases has been demonstrated in apoptosis resulting from growth factor withdrawal (10) or induced by apogenic stimuli such as TNF␣ (4,33), TRAIL (34), Fas (35), and lipopolysaccharide (36). Although it has been shown that blocking caspase activation or activity can prevent apoptosis, inhibition of caspase activity by benzyloxycarbonyl-VAD-fluoromethylketone or benzyloxycarbonyl-Aspfluoromethylketone also induces a switch from apoptosis to necrosis in some cells (37). Similarly, intracellular ATP concentration may also dictate the path to cell death (38). Apoptosis is an energy-dependent process, and ATP is required for the activation of caspase-9 by acting as a cofactor for the interaction of cytochrome c with Apaf-1 (39). We have previously shown that NO inhibits caspase activation/activity by both redox-based S-nitrosylation of the cysteine residue present in the catalytic site of all caspase enzymes and through the cGMPdependent inhibition of caspase activation in hepatocytes (4). That NO induced necrosis but not apoptosis when millimolar concentrations of SNAP were added to cultured hepatocytes is FIG. 9. Effect of DNIC-containing cytosol on caspase-3 activity. A, cytosol was prepared from hepatocytes and RAW264.7 cells treated with 400 M SNAP for 5 h. rh-caspase-3 (4 ng) was incubated with 200 g of cytosolic protein at room temperature for 40 min. Caspase-3-like activity was measured by colorimetric assay following preincubation with or without 5 mM DTT for 20 min. B, ironpreloaded RAW264.7 cells were treated with 400 M SNAP. NO production, DNIC, and caspase-3 activity were measured as described under "Experimental Procedures" (average from n ϭ 3).
FIG. 10. Effect of DNIC on caspase-3-dependent activity and in vitro DNA fragmentation. A, dose-dependent inhibition of caspase-3 activity. rh-caspase-3 (8 ng) was incubated with DNIC (200 M) at room temperature for 30 min. The enzyme activity was measured by colorimetric assay. B, time course inhibition of caspase-3 activity. C, inhibition of caspase-3-dependent ICAD cleavage. 32 P-Labeled ICAD was incubated with rh-caspase-3 pretreated with 200 M DNIC or SNAP at room temperature for 30 min. ICAD cleavage was analyzed by electrophoresis. D, inhibition of caspase-3-dependent DNA fragmentation in a cell-free system by DNIC. rh-caspase-3 (400 ng) pretreated with DNIC was incubated with hepatocyte S-100 cytosolic fraction (400 g) and 10 l of nuclei solution (ϳ1 ϫ 10 6 nuclei) in a final volume of 80 l at 37°C for 140 min. DNA fragmentation was analyzed by electrophoresing on a 2% agarose gel and visualized with ethidium bromide staining. consistent with the capacity of NO to inhibit increases in caspase activity in these cells. In contrast, NO increased caspase-3 activation/activity and induced apoptosis in RAW264.7 cells. Although it is likely that some necrosis also took place, the failure of NO to limit caspase activity in these cells probably contributes to the differential sensitivity of the two cell types to NO. Furthermore, these results are consistent with the previous reports indicating that inhibition of caspase activity can drive the cells exposed to an apoptotic stimulus into necrosis (37,40). In fact, redox-based inactivation of caspases by reactive oxygen species has been shown to suppress apoptosis and lead to necrosis (41).
The release of cytochrome c under both apoptotic and necrotic conditions may come about through different mechanisms. In apoptosis, the NO-induced release is most likely part of a regulated process resulting from upstream caspase activation or due to a direct effect of NO on mitochondria (42). During necrosis resulting from high levels of NO exposure, NO most likely induces mitochondrial dysfunction or even damage leading to cytochrome c release. In support of this, we have shown that NO at high concentrations inhibits mitochondrial electron transport in hepatocytes (43). Similar results have been reported in tumor cell lines where disruption of iron-sulfur centers in aconitase, complex I, and complex II lead to cell death (44). Cytochrome c release in the absence of new ATP formation probably cannot activate caspase-9, and the cell dies via necrosis as energy stores are depleted.
NO can limit apoptosis in a number of cell types by the inhibition of caspase activity through S-nitrosylation (4,13). This S-nitrosylation is reversed under strong reducing conditions such as DTT (4). In intact cells, the presence of a denitrosylase enzyme has been recently proposed and could account for the failure of NO to suppress apoptosis in some cells (45). Although NO can interact with thiol groups, its chemical reactivity is very weak compared with highly reactive NO reaction products with NO ϩ -like characteristics. NO ϩ can be generated by the loss of one electron from NO. Known electron acceptors are transition metal ions such as iron and copper, which readily react with NO in vivo. S-Nitrosylating species can be generated by the interaction of NO with iron-sulfur complexes (20). Hepatocytes are rich in NO-interactable iron-sulfur proteins when compared with many other cell types. DNIC have been shown to carry out S-nitrosylation of albumin and GSH reductase through the formation of NO ϩ (20,45). We reported that conversion of heme to non-heme iron by NO-mediated hemeoxygenase induction increased non-heme iron-nitrosyl complexes (19,47) and protected hepatocytes from NO-induced cell death (19). We recently showed that NO protected iron-supplemented MCF-7 cells, but not control cells, from TNF␣-mediated apoptosis by inhibition of caspase activity through S-nitrosylation (13). Here, preloading RAW264.7 cells with ferrous ion increased the cellular level of non-heme iron, resulting in increased levels of DNIC and inhibition of caspase activity upon NO exposure. Incubation of rh-caspase-3 with DNICcontaining cytosol (ϳ0.4 nmol/200 g of protein) in a final reaction volume of 20 l, which is identical to 20 M DNIC, inhibited enzyme activity by 30% (Fig. 9B). A similar effect was seen using 20 -50 M synthetic DNIC (Fig. 10A). Furthermore, the inhibitory effect of DNIC on ICAD cleavage and in vitro DNA fragmentation by caspase-3 was reversed by DTT, which also restored S-nitrosylated thiol to an active state. Our results provide evidence that DNIC function as S-nitrosylating species in cells by showing that DNIC-containing cytosol and chemically synthesized DNIC inactivated caspase-3 activity. It is also possible that the increased iron-sulfur complexes scavenge NO, thereby reducing SNAP-mediated cell death. This scavenging capacity could account for the delay in cell death noted in the iron-loaded cells, while the formation of DNIC leading to S-nitrosylation of caspases could have favored necrotic death.