Introduction

Nitric oxide (NO) is a free radical that is endogenously produced by the enzyme NO synthase, which catalyzes the oxidation of L-arginine, yielding NO and L-citrulline. Although NO regulates various cell functions,1, 2, 3, 4 its excessive or inappropriate formation might cause deleterious effects relevant in various human pathologies.3, 4, 5 A large proportion of these toxic effects appears to be mediated by peroxynitrite, the coupling product of NO and superoxide.5, 6, 7 The well-established notion that peroxynitrite is a highly reactive nitrogen species has long been considered as a straightforward indication that the deleterious effects mediated by this species are entirely ascribable to its direct interactions with target biomolecules. These effects include lipid peroxidation,8 protein nitration and nitrosylation,9 DNA damage10 and oxidation of thiols,11 and are thought to result in mitochondrial dysfunction and cytotoxicity.12, 13, 14, 15 Since peroxynitrite very rapidly decomposes at physiological pH values,16 its effects on target biomolecules should then be produced within a few seconds after addition to cultured cells, i.e., the time required for its disappearance from the intra–extracellular media.

This prediction, however, did not find experimental support in a recent study performed in our laboratory,17 indicating that the DNA single-strand breakage generated by exposure of U937 cells to a bolus of peroxynitrite is barely detectable at 5 min but then linearly increases with time for up to 30 min of incubation. We demonstrated that an initial event triggered by peroxynitrite, i.e., inhibition of complex III of the mitochondrial respiratory chain, was responsible for the time-dependent formation of superoxide that dismutates to H2O2, a species that can reach the nucleus and cause DNA strand scission via a Fenton type reaction. Thus, these findings are on the one hand in obvious contrast with the notion that all the lesions resulting from exposure to peroxynitrite are directly generated by the oxidant and on the other hand consistent with the notion that various species might be involved in the ensuing lethal response.

Using a toxicity paradigm in which peroxynitrite causes U937 cell necrosis mediated by mitochondrial permeability transition,15 we were indeed able to demonstrate that delayed formation of H2O2 is essential for the occurrence of cell death.18 There are at least two potential mechanisms whereby H2O2 can affect the peroxynitrite-dependent lethal response: (i) H2O2, which is continuously generated after peroxynitrite exposure,18 may cause the accumulation of additional lesions, which include DNA strand scission,17 then resulting in cytotoxicity. Such a role would then be consistent with the conventional view that necrosis is a passive response to overwhelming damage. (ii) H2O2 may interfere with cytoprotective signalling events counteracting the lethal response evoked by peroxynitrite. This scenario would instead be consistent with the notion that reactive species can specifically regulate cell signalling19, 20 and with the emerging hypothesis that necrosis might not always be an obligatory response to a stochastic process of cell damage.21 Our previous findings22 that peroxynitrite promotes an early activation of the cytosolic phospholipase A2 isoform (cPLA2), and that the ensuing release of arachidonic acid (AA) plays an important cytoprotective role, offer important clues in this direction.

The present study provides solid experimental evidence indicating that the mechanism whereby delayed formation of H2O2 promotes toxicity in cells exposed to peroxynitrite involves inhibition of the cPLA2-dependent cytoprotective signalling. It therefore appears that exposure to peroxynitrite primes U937 cells to mitochondrial permeability transition-dependent toxicity and that this event occurs only under conditions in which H2O2 impairs the AA-dependent cytoprotective signalling.

Results

Exogenous H2O2 prevents cytoprotection mediated by rotenone but not by AA

The results illustrated in Figure 1a indicate that a 55 min exposure to concentrations of H2O2 higher than 300 μM promote U937 cell death. The maximal nontoxic concentration of the oxidant was therefore used to ask the question of whether its supplementation is able to prevent cytoprotection afforded by AA, or rotenone (a complex I inhibitor), in cells exposed to authentic peroxynitrite. It was found that exogenous H2O2, while not affecting the toxicity mediated by 1.2 mM peroxynitrite, abolishes the cytoprotective effects of rotenone (0.5 μM) but does not reverse cytoprotection mediated by AA (0.1 μM) (Figure 1b). As shown in Figure 1c, concentrations of H2O2 lower than 300 μM partially reduced the protective effects mediated by rotenone.

Figure 1
figure 1

Hydrogen peroxide abolishes the protective effects mediated by rotenone on peroxynitrite-induced cytotoxicity. No effect on cytoprotection afforded by AA. (a) Cells were exposed for 55 min to increasing concentrations of H2O2 and then analyzed for cytotoxicity using the trypan blue exclusion assay. Results represent the mean±S.E.M. from three separate experiments. (b) Cells were exposed for 5 min to 0 or 1.2 mM peroxynitrite, centrifuged and then postincubated for a further 55 min in the absence or presence of H2O2 (300 μM), with or without rotenone (0.5 μM) or AA (0.1 μM). Cytotoxicity was then determined. Also shown is the lethal response of cells exposed to 300 μM H2O2 alone for 55 min. Results represent the mean±S.E.M. from at least four separate experiments. (*) P<0.0001 as compared to untreated cells; (#) P<0.0001 as compared to cells exposed to H2O2 alone; (∧) P<0.0001 as compared to cells treated with peroxynitrite alone; (†) P<0.0001 as compared to cells treated with peroxynitrite and H2O2 (unpaired t-test). (c) Cells were exposed for 5 min to 1.2 mM peroxynitrite, centrifuged and then postincubated for a further 55 min in the absence (open circle) or presence of rotenone alone or associated with H2O2 (50–300 μM, closed circles). Results represent the mean±S.E.M. from three separate experiments

Full size image

Formation of endogenous H2O2 was then assessed using the fluorescent probe dihydrorhodamine 123 (DHR, 10 μM), added to the cultures after the 5 min treatment with peroxynitrite, a condition that does not allow peroxynitrite to oxidize DHR directly. The results illustrated in Figure 2 indicate that rotenone was as effective as catalase (10 U/ml) in preventing the DHR fluorescence response mediated by peroxynitrite. AA or boiled catalase, however, did not produce detectable effects. Control experiments revealed that rotenone, AA and catalase were not toxic and/or able to promote formation of H2O2, when added alone to the cultures (not shown).

Figure 2
figure 2

Inhibition of PLA2 does not affect the extent of H2O2 formation mediated by peroxynitrite. Cells were exposed for 5 min to 0 or 1.2 mM peroxynitrite, centrifuged and then postincubated for a further 10 min in fresh treatment buffer containing 10 μM DHR, in the absence or presence of rotenone (0.5 μM), AA (0.1 μM), catalase (10 U/ml) or boiled catalase, with or without ETYA (50 μM). The DHR-derived fluorescence was then monitored by confocal microscopy and analyzed as detailed under Materials and Methods. Results represent the mean±S.E.M. from five separate experiments. (*) P<0.001 as compared to untreated cells (unpaired t-test)

Full size image

Thus, rotenone prevents formation of H2O2 as well as toxicity induced by peroxynitrite, and exogenous H2O2 can override these cytoprotective effects. Exogenous AA, while not preventing formation of H2O2 induced by peroxynitrite, affords cytoprotection and further addition of exogenous H2O2 fails to restore toxicity.

Formation of H2O2 is a dispensable event in the peroxynitrite-dependent lethal response of phospholipase A2 (PLA2) inhibitor-supplemented cells

The results illustrated in Figure 3 indicate that rotenone as well as catalase prevented the lethal response mediated by peroxynitrite, that the PLA2 inhibitor arachidonyl trifluoromethyl ketone (AACOCF3, 50 μM) reversed these effects and that a further addition of AA resulted again in cytoprotection. AACOCF3 did not significantly enhance the lethal response evoked by 1.2 mM peroxynitrite. The effects of AACOCF3 were mimicked by two general PLA2 inhibitors, 5,8,11,14-eicosatetraynoic acid (ETYA, 50 μM) and mepacrine (50 μM), and none of these compounds was toxic when added alone to the cultures (not shown). E-6-(bromomethylene) tetrahydro-3-(1-naphthalenyl)-2H-pyran-2-one (HELSS, 30 μM), an inhibitor of the calcium-independent PLA2 (iPLA2), failed to reproduce the effects of AACOCF3. Additional experiments revealed that ETYA, both in the absence or presence of AA, failed to affect formation of H2O2 mediated by peroxynitrite (Figure 2). Furthermore, ETYA neither modified the effects of rotenone or catalase in the peroxynitrite-dependent DHR fluorescence response nor produced detectable formation of H2O2 when added to the cells in the absence of additional treatments (Figure 2). Replacing ETYA with mepacrine led to identical outcomes, although this inhibitor was intrinsically fluorescent (not shown). The high fluorescence observed after exposure to AACOCF3 did not allow us to perform experiments with this inhibitor.

Figure 3
figure 3

Rotenone or catalase fails to prevent peroxynitrite toxicity in PLA2 inhibitor-supplemented cells. Cells were exposed for 5 min to 1.2 mM peroxynitrite, centrifuged and then postincubated for a further 55 min in fresh treatment buffer, in the absence or presence of rotenone (0.5 μM) or catalase (10 U/ml), with or without HELSS (30 μM) or AACOCF3 (50 μM) alone or associated with AA (0.1 μM). Cells were then analyzed for toxicity using the trypan blue exclusion assay. Results represent the mean±S.E.M. from four separate experiments. (*) P<0.0001 as compared to cells treated with peroxynitrite alone; (#) P<0.0001 as compared to cells exposed to peroxynitrite associated with AACOCF3; (∧) P<0.0001 as compared to cells exposed to peroxynitrite and rotenone; (†) P<0.0001 as compared to cells exposed to peroxynitrite and catalase (unpaired t-test)

Full size image

Thus rotenone, or catalase, while preventing the peroxynitrite-dependent formation of H2O2, affords cytoprotection only under conditions in which PLA2 is not pharmacologically inhibited. In addition, the effects of the PLA2 inhibitors detected in toxicity studies were not accompanied by effects at the level of H2O2 formation. Finally, exogenous AA promoted survival also under conditions permissive for H2O2 formation.

As mentioned above, AACOCF3 did not significantly enhance the toxicity induced by 1.2 mM peroxynitrite. In agreement with our previous findings,22 however, a remarkable lethal response is observed after exposure of PLA2 inhibitor-supplemented cells to 100 μM peroxynitrite (Figure 4), a treatment that does not affect viability when added to the cells in the absence of further additions. Interestingly, AA, unlike catalase, prevented the enhancing effects of AACOCF3. The effects of AACOCF3 were mimicked by ETYA and mepacrine (not shown) but not by HELSS.

Figure 4
figure 4

Antimycin A fails to enhance peroxynitrite toxicity in PLA2 inhibitor-supplemented cells. Cells were exposed for 5 min to 100 μM peroxynitrite, centrifuged and then postincubated for a further 55 min in fresh treatment buffer in the absence or presence of HELSS (30 μM), AACOCF3 (50 μM) alone or associated with either AA (0.1 μM) or catalase (10 U/ml). In other experiments, treatment with peroxynitrite was followed by exposure to antimycin A (1 μM) with or without AACOCF3, catalase or AA. Cells were then analyzed immediately for toxicity using the trypan blue exclusion assay. Results represent the mean±S.E.M. from four separate experiments. (*) P<0.0001 as compared to cells treated with peroxynitrite alone; (#) P<0.0001 as compared to cells exposed to peroxynitrite associated with AACOCF3; (∧) P<0.0001 as compared to cells exposed to peroxynitrite and antimycin A (unpaired t-test)

Full size image

As we reported previously,18 the complex III inhibitor antimycin A (1 μM) also promotes toxicity in cells exposed to an otherwise nontoxic concentration of peroxynitrite. The results illustrated in Figure 4 are in keeping with this notion and also indicate that AACOCF3 did not further enhance toxicity induced by peroxynitrite in the presence of antimycin A. Furthermore, both catalase and AA afforded protection in cells exposed to peroxynitrite/antimycin A. The results illustrated in Figure 5 indicate that formation of H2O2 does not take place after treatment with 100 μM peroxynitrite alone or associated with ETYA (as previously mentioned, AACOCF3 cannot be used in experiments measuring DHR oxidation because of its intrinsic fluorescence). This response, however, can be clearly observed upon exposure to peroxynitrite and antimycin A and was insensitive to ETYA or AA, but was prevented by catalase. The complex III inhibitor did not cause detectable DHR oxidation when added to the cells in the absence of additional treatments.

Figure 5
figure 5

Antimycin A enhances the extent of H2O2 formation mediated by peroxynitrite. Cells were exposed for 5 min to 0 or 100 μM peroxynitrite, centrifuged and then postincubated for a further 10 min in fresh treatment buffer containing 10 μM DHR, in the absence or presence of ETYA (50 μM). In other experiments, treatment with peroxynitrite was followed by exposure to antimycin A (1 μM), with or without ETYA, AA or catalase. The DHR-derived fluorescence was then monitored by confocal microscopy and analyzed as detailed under Materials and Methods. Results represent the mean±S.E.M. from at least three separate experiments. (*) P<0.0001 as compared to untreated cells (unpaired t-test)

Full size image

Thus, it appears that an otherwise nontoxic concentration of peroxynitrite promotes cell death under a variety of conditions that are, or are not, associated with H2O2 formation. Addition of AA was always cytoprotective, regardless of whether H2O2 was being formed or not.

Taken together, the results reported in this section indicate that formation of H2O2 is a dispensable event for the occurrence of toxicity in PLA2 inhibitor-supplemented cells exposed to peroxynitrite.

Formation of H2O2 is a dispensable event in the peroxynitrite-dependent lethal response of cPLA2 antisense oligonucleotide-transfected cells

U937 cells were transfected with phosphorothioate-modified cPLA2 antisense oligonucleotides (10 μM) previously employed to prevent successfully the translation and synthesis of the enzyme protein in U937 cells.23 Complementary nonsense oligonucleotides were used as a negative control. Immunocytochemical and Western blot analyses using a monoclonal antibody to human cPLA2 showed that the level of the protein was indeed significantly lower (about 70% decrease) in cells treated with antisense oligonucleotides, as compared to cells treated with nonsense oligonucleotides (Figure 6a). Consistently, the release of AA stimulated by 200 μM peroxynitrite was lower in cells transfected with antisense oligonucleotides, as compared to that observed in nonsense oligonucleotide-transfected cells (Figure 6b). Treatment with 100 μM peroxynitrite did not cause significant AA release in cPLA2-downregulated cells, whereas a 45% increase in AA release was detected in nonsense oligonucleotide-transfected cells (not shown). Importantly, AA release mediated by peroxynitrite was suppressed by AACOCF3 and by the calcium chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetra-acetic acid tetrakis(acetoxymethyl ester (BAPTA/AM, 20 μM) (Figure 6b).

Figure 6
figure 6

Delayed formation of H2O2 and toxicity mediated by peroxynitrite in cells transfected with cPLA2 antisense oligonucleotides. (a) Cells transfected with cPLA2 nonsense or antisense oligonucleotides were stained with (i) or without (ii) monoclonal antibodies against human cPLA2 and visualized with a confocal microscope, as indicated in the Materials and Methods section. Representative fields are shown along with the results of the quantitative analysis of the fluorescence detected. The level of cPLA2 protein in nonsense and antisense oligonucleotide-transfected cells, detected by Western blotting analysis, is also shown. cPLA2 protein expression was quantified by densitometric analysis. The results shown are representative of three separate experiments. (b) RP or RD U937 cells were transfected with cPLA2 nonsense or antisense oligonucleotides, labelled with [3H]-AA, exposed for 5 min to 200 μM peroxynitrite, centrifuged and then postincubated for a further 10 min in a fresh treatment buffer in the absence or presence of AACOCF3 (50 μM). The effect of BAPTA/AM (20 μM) added after peroxynitrite in RP cells transfected with cPLA2 nonsense or antisense oligonucleotides is also shown. After the treatments, [3H]-AA release was quantified as described in the Materials and Methods section. Results are expressed as percentages of [3H]-AA release of the appropriate controls (RP cells: 746±65 cpm; cPLA2 nonsense oligonucleotide-transfected RP cells: 805±74 cpm; cPLA2 antisense oligonucleotide-transfected RP cells: 991±39 cpm; RD cells: 781±56 cpm; cPLA2 nonsense oligonucleotide-transfected RD cells: 815±62 cpm; cPLA2 antisense oligonucleotide-transfected RD cells: 998±32 cpm) and represent the mean±S.E.M. from three separate experiments, each performed in duplicate. (*) P<0.0001 and (#) P<0.001 as compared to untreated cells (unpaired t-test). (c) cPLA2 nonsense or antisense oligonucleotide-transfected cells were exposed for 5 min to 0 or 100 μM peroxynitrite, centrifuged and then postincubated for a further 55 min in the absence or presence of either AACOCF3 (50 μM) or antimycin A (1 μM). Cells were then analyzed for cytotoxicity using the trypan blue exclusion assay. Results represent the mean±S.E.M. from 3–5 separate experiments. (*) P<0.001 as compared to nonsense oligonucleotide-transfected cells treated with peroxynitrite alone; (#) P<0.001 as compared to untreated antisense oligonucleotide-transfected cells (unpaired t-test). (d) cPLA2 nonsense or antisense oligonucleotide-transfected cells were exposed for 5 min to 0 or 100 μM peroxynitrite, centrifuged and then postincubated for a further 10 min in fresh treatment buffer containing 10 μM DHR, in the absence or presence of either ETYA (50 μM) or antimycin A (1 μM). The DHR-derived fluorescence was monitored by confocal microscopy and analyzed as detailed under Materials and Methods. Results represent the mean±S.E.M. from five separate experiments. (*) P<0.001 as compared to cells treated with peroxynitrite alone (unpaired t-test)

Full size image

Toxicity studies were next performed. As expected,22 100 μM peroxynitrite, while not being toxic for nonsense oligonucleotide-transfected cells, caused an AACOCF3-insensitive lethal response in cPLA2 antisense oligonucleotide-transfected cells (Figure 6c). The PLA2 inhibitor, however, promoted toxicity in peroxynitrite-treated cells transfected with nonsense oligonucleotides. These results emphasize the specificity of the effects of AACOCF3. The observation that under none of these conditions oxidation of DHR was detected (Figure 6d) confirms the notion that formation of H2O2 is a dispensable event for the occurrence of toxicity in cells unable to generate AA. Antimycin A promoted formation of H2O2 (Figure 6d) and toxicity (Figure 6c) in nonsense-transfected cells treated with peroxynitrite. In cPLA2 antisense-transfected cells, however, the complex III inhibitor on the one hand caused formation of H2O2 (Figure 6d) but on the other hand failed to further increase toxicity (Figure 6c).

U937 cells were also transfected with iPLA2 antisense or nonsense oligonucleotides. Western blot analyses showed that the level of the protein was lower (about 60% decrease) in cells treated with antisense oligonucleotides, as compared to cells treated with nonsense oligonucleotides (Figure 7a). Under these conditions, peroxynitrite caused an identical AA release that was inhibited by AACOCF3 but not by HELSS (Figure 7b). Toxicity studies revealed that 100 μM peroxynitrite was not toxic for both antisense and nonsense oligonucleotide-transfected cells. Under both conditions, toxicity was observed in the presence of AACOCF3 and hardly any effect was observed upon addition of HELSS (Figure 7c).

Figure 7
figure 7

Transfection with iPLA2 antisense oligonucleotide does not affect [3H]-AA release or toxicity mediated by peroxynitrite. (a) Cells were transfected with iPLA2 nonsense or antisense oligonucleotides. The level of iPLA2 was detected by Western blotting and quantified by densitometric analysis as indicated in the Materials and Methods section. The results shown are representative of three separate experiments. (b) iPLA2 nonsense and antisense oligonucleotide-transfected cells, labelled with [3H]-AA, were exposed for 5 min to 200 μM peroxynitrite, centrifuged and then postincubated for a further 10 min in fresh treatment buffer in the absence or presence of HELSS (30 μM) or AACOCF3 (50 μM). After the treatments, [3H]-AA release was quantified as described in the Materials and Methods section. Results represent the means±S.E.M. from three separate experiments each performed in duplicate. (c) iPLA2 nonsense or antisense oligonucleotide-transfected cells were exposed for 5 min to 0 or 100 μM peroxynitrite, centrifuged and then postincubated for a further 55 min in the absence or presence of AACOCF3 or HELSS. Cells were then analyzed for toxicity using the trypan blue exclusion assay. Results represent the mean±S.E.M. from 3–5 separate experiments

Full size image

These results confirm the notions that formation of H2O2 is a dispensable event for the occurrence of toxicity in PLA2-inhibited cells exposed to peroxynitrite and that cPLA2 is the specific PLA2 isoform involved in the AA-dependent cytoprotective signalling.22

Pharmacological inhibition or genetic depletion of cPLA2 abolishes resistance of respiration-deficient cells to toxicity mediated by peroxynitrite

The last approach utilized to establish the relationships existing between H2O2 and the cPLA2 cytoprotective signalling in cells exposed to peroxynitrite took advantage of our previous observation that respiration-deficient (RD) cells are both unable to generate H2O2 and resistant to killing mediated by peroxynitrite.18 We reasoned that pharmacological inhibition or genetic depletion of cPLA2 should abrogate the resistance phenotype.

For this purpose, the lethal response evoked by increasing concentrations of peroxynitrite in respiration-proficient (RP) and RD cells was investigated. As expected,18 the latter cell type displayed resistance against toxicity mediated by peroxynitrite (Figure 8a). Addition of mepacrine (b), ETYA (c) or AACOCF3 (d) remarkably increased toxicity mediated by low concentrations of peroxynitrite in both RP and RD cells, and the dose-response curves obtained in the two cell types were superimposable. Exogenous AA was cytoprotective under both conditions. Additional experiments were performed using RP or RD cells transfected with cPLA2 antisense or nonsense oligonucleotides. As indicated in Figure 8e, genetic depletion of cPLA2 abolished resistance of RD against peroxynitrite toxicity. Exogenous AA was cytoprotective also under these conditions. In contrast, RD and RP cells transfected with cPLA2 nonsense oligonucleotides responded to peroxynitrite alone, or associated with AA, as their respective nontransfected cells (Figure 8f).

Figure 8
figure 8

Inhibition of PLA2 or transfection with cPLA2 antisense oligonucleotides abolishes resistance of RD U937 cells to peroxynitrite-induced toxicity. RP (open circles) or RD (closed circles) cells were treated with increasing concentrations of peroxynitrite for 5 min and then postincubated for 55 min in fresh treatment buffer in the absence (a) or presence of mepacrine (50 μM) (b), ETYA (50 μM) (c) or AACOCF3 (50 μM) (d). Panels e and f show the results of experiments, similar to those described in (a), conducted with RP and RD cells transfected with cPLA2 antisense (e) or nonsense (f) oligonucleotides. Finally, the effect of addition of exogenous AA during the 55 min of post-treatment incubation is also shown (open squares, RP cells; closed squares, RD cells). Cytotoxicity was determined using the trypan blue exclusion assay. Results represent the mean±S.E.M. from four separate experiments. (*) P<0.001, (**) P<0.0001 as compared to cells exposed to peroxynitrite alone. (#) P<0.0001 as compared to cells treated with peroxynitrite associated with PLA2 inhibitors (unpaired t-test)

Full size image

Thus, cells unable to generate H2O2 because of being devoid of a functional mitochondrial respiratory chain are resistant to killing induced by peroxynitrite but lose this phenotypic feature when the cPLA2-dependent cytoprotective signalling is impaired.

Formation of H2O2 impairs the peroxynitrite-dependent AA release

We previously illustrated the results shown in Figure 6b indicating that cPLA2 antisense oligonucleotide-transfected cells release less AA than cPLA2 nonsense oligonucleotide-transfected cells in response to peroxynitrite. The same figure shows similar results obtained using RD cells transfected with the same oligonucleotides. The extent of AA release obtained with nonsense-transfected cells, however, appears greater in RD than in RP cells. This finding suggests that H2O2 may impair the cPLA2-dependent AA release. We therefore determined the extent of AA release in cells exposed to increasing concentrations of peroxynitrite under conditions permissive or not permissive for H2O2 formation. As previously noted,22 this response linearly increases up to 200 μM peroxynitrite and does not significantly change thereafter. In agreement with the data shown in Figure 6b, the extent of AA release was enhanced by the RD phenotype (Figure 9). In addition, rotenone and antimycin A, respectively, increased or decreased AA release mediated by peroxynitrite in RP cells.

Figure 9
figure 9

Peroxynitrite-dependent AA release in RP or RD cells. [3H]-AA-labelled RP cells were exposed for 5 min to increasing concentrations of peroxynitrite and then postincubated for a further 10 min in fresh treatment buffer in the absence (open circles) or presence of rotenone (0.5 μM) (closed circles), or antimycin A (1 μM) (closed squares). Also shown are the results obtaining using RD cells (open squares). [3H]-AA release was quantified as described in the Materials and Methods section. Results represent the mean±S.E.M. from three separate experiments, each performed in duplicate. (*) P<0.01, (**) P<0.001 as compared to RP cells exposed to peroxynitrite alone (unpaired t-test)

Full size image

Thus, it appears that the release of AA mediated by peroxynitrite is negatively controlled by H2O2. As a consequence, these results are consistent with the notion that the mechanism whereby H2O2 mediates toxicity in U937 cells exposed to peroxynitrite involves inhibition of the cPLA2-dependent cytoprotective signalling.

Discussion

We previously reported that delayed formation of H2O2, resulting from exposure of U937 cells to toxic levels of peroxynitrite, plays a pivotal role in the ensuing lethal response.18 Since the toxicity paradigm adopted is associated with a mitochondrial permeability transition-dependent necrosis, taking place within minutes after treatment with peroxynitrite and rapidly evolving in cell lysis,15 it may be inferred that H2O2 contributes to toxicity by inflicting molecular damage in addition to that directly produced by peroxynitrite. The evidence presented here, however, does not confirm this notion. Rather, the effects of H2O2 seem to be restricted to interference with a cytoprotective signalling pathway, previously identified in our laboratory,22 mediated by AA released upon peroxynitrite-dependent activation of cPLA2.

The first clue in this direction derives from experiments in which we asked the question of whether supplementation with a nontoxic dose of H2O2 promotes toxicity in cells challenged with an otherwise lethal concentration of peroxynitrite but rescued with either rotenone or AA. It is important to note that these two strategies to abolish peroxynitrite-dependent cell death are based on different mechanisms. Rotenone is a complex I inhibitor and, as we previously showed,18 prevents formation of superoxide at the complex III level and the subsequent generation of H2O2. That rotenone prevents delayed formation of H2O2 induced by peroxynitrite is also shown in Figure 2. The fact that exogenous H2O2 was able to override cytoprotection afforded by rotenone (Figure 1b and c) on the one hand emphasizes the specificity of the effects of this inhibitor whereas, on the other hand, confirms a role for H2O2 in this toxicity paradigm. AA is the cPLA2 product capable of preventing toxicity induced by peroxynitrite,22 and the observation that exogenous H2O2 was unable to revert cytoprotection mediated by this lipid messenger (Figure 1b) is consistent with a role of the oxidant in a signalling pathway operating upstream to AA formation. Importantly, AA abolished toxicity (Figure 1b) without interfering with H2O2 formation (Figure 2) and was cytoprotective also in the presence of excess amounts of H2O2 (e.g. endogenous and exogenous, Figure 1b).

The interpretation of the results from these experiments, however, is complicated by the fact that the effects of exogenous H2O2 may or may not entirely reproduce the effects of the oxidant generated in specific microenvironments of the cell. We therefore adopted an array of different strategies to manipulate differentially the intracellular formation of H2O2 and the activity of PLA2 to establish the relationships existing between these two events.

The outcome of these experiments (Figures 2, 3, 4 and 5), regardless of the specific approach utilized, was always the same. Formation of H2O2 was necessary for the occurrence of the peroxynitrite-dependent lethal response in cells in which PLA2 was not pharmacologically inhibited. However, under the latter conditions, formation of H2O2 was a dispensable event. Furthermore, cells were always rescued by AA regardless of whether H2O2 was or was not being generated in the specific toxicity paradigm employed.

It is important to emphasize that these findings were obtained using two general PLA2 inhibitors (mepacrine and ETYA) and AACOCF3, a compound extremely more potent in inhibiting cPLA2 than other calcium-dependent PLA2 isoforms, such as secretory PLA2,24, 25 but that can also inhibit iPLA2.26 It would appear that, in our system, AACOCF3 acts as a cPLA2 inhibitor since its effects were not reproduced by HELSS, an iPLA2 inhibitor.26, 27 The notion that cPLA2 is specifically involved in the cytoprotective signalling opposing peroxynitrite toxicity was further established using antisense oligonucleotides downregulating either cPLA2 (Figure 6) or iPLA2 (Figure 7). Cells transfected with the respective nonsense oligonucleotides were used as a negative control and, in both toxicity and DHR oxidation studies, responded to peroxynitrite alone, or associated with various treatments, as the nontransfected cells. Similar results were obtained using cells transfected with iPLA2 antisense oligonucleotides. The outcome of the experiments using cPLA2 antisense oligonucleotide-transfected cells, however, demonstrated that these cells are killed by 100 μM peroxynitrite and respond to various manipulations as the nontransfected cells supplemented with AACOCF3. Indeed, formation of H2O2 was found to be a dispensable event for the occurrence of the peroxynitrite-dependent lethal response in cells in which cPLA2 was genetically depleted. Thus, the information provided by these experiments is in keeping with that deriving from inhibitor studies and strongly suggests that cPLA2 is the specific PLA2 isoform involved in the AA-dependent survival pathway. Further evidence supporting this notion was obtained from experiments measuring AA release in response to peroxynitrite using both pharmacological inhibitors of PLA2 and cells transfected with either cPLA2 or iPLA2 antisense oligonucleotides (Figures 6 and 7).

The results thus far discussed therefore imply that H2O2 promotes the peroxynitrite-dependent lethal response by interfering with the cytoprotective signalling mediated by cPLA2.

This notion is consistent with the results obtained with RD cells. These cells were more resistant than RP cells to toxicity induced by increasing concentrations of peroxynitrite, because they do not produce H2O2 in response to peroxynitrite.18 Interestingly, we found that pharmacological inhibition or genetic depletion of cPLA2 abolishes this resistance and that AA prevents cell death mediated under all of these conditions (Figure 8).

Thus, it appears that the role of H2O2 is restricted to inhibition of the PLA2-dependent cytoprotective signalling. That this was indeed the case is demonstrated by the observation that RD cells release more AA after challenge with increasing concentrations of peroxynitrite (Figures 6b and 9). In addition, the extent of this response is basically superimposable on that obtained using RP cells supplemented with rotenone, a treatment that suppresses formation of H2O2 (Figure 2). Finally, further support for the notion that H2O2 inhibits PLA2 activity stimulated by peroxynitrite is given by the observation that antimycin A, a treatment enhancing formation of H2O2 (Figure 5), promotes inhibition of the peroxynitrite-dependent AA release (Figure 9).

In conclusion, the results presented in this study lead to the identification of the mechanism whereby delayed formation of H2O2 promotes cell death after exposure to peroxynitrite. It appears that H2O2 does not act by producing nonspecific molecular damage but, rather, by selectively inhibiting the AA-dependent cytoprotective signalling pathway. The observation that the cPLA2 activity stimulated by peroxynitrite is inhibited by endogenously generated H2O2 is an unexpected finding since the latter was previously described as a PLA2 activator.28, 29, 30 It is tempting to speculate that signalling events triggered by peroxynitrite, and leading to cPLA2 activation, are intercepted by an inhibitory signalling triggered by H2O2. The possibility also exists that peroxynitrite and H2O2 may cooperate in inducing ATP depletion, a condition recently reported to be associated with inhibition of cPLA2 activity.31 Treatment with a toxic dose of peroxynitrite is indeed expected to cause a robust decline in cellular ATP concentration since the elevated energy demand for repair processes cannot be compensated by enhanced energy production for at least two separate reasons: (a) peroxynitrite, by inhibiting complex III,12, 17 will decrease the synthesis of mitochondrial ATP; (b) glycolytic ATP may also decrease since both peroxynitrite32 and H2O233 are effective inhibitors of glyceraldehyde-3-phosphate dehydrogenase, a rate-limiting enzyme of glycolysis.

Thus, while the specific mechanism(s) whereby H2O2 inhibits cPLA2 activity stimulated by peroxynitrite is currently being investigated, an integrated view of the results presented in this and our previous studies suggests the following scenario: exposure to a toxic dose of peroxynitrite promotes an early activation of cPLA2 followed by a decline in activity caused by H2O2, which is continuously being formed via peroxynitrite-dependent inhibition of complex III. A direct consequence of cPLA2 inhibition is reduced formation of AA, and thus impairment of the cytoprotective signalling triggered by this lipid messenger. As mentioned above, peroxynitrite kills U937 cells by inducing mitochondrial permeability transition15 and, as a consequence, AA is expected to promote downstream events opposing the opening of permeability transition pores. Our recent work34 demonstrates that this is indeed the case, and future work will investigate the mechanism involved in this survival-promoting response. The notion that AA causes translocation35 and activation of protein kinase C,36 as well as stimulation of extracellular signal-regulated kinases 1/2 phosphorylation,37 is well established and these events were shown to prevent apoptosis.38, 39 Thus, the question to be answered is whether these pathways also mediate the protective effects of AA in the peroxynitrite-dependent necrotic response.

These results strongly support the emerging notion that necrosis is not simply a passive and obligatory response to overwhelming damage. The fine regulation of the balance between cell survival and death, operated by different signalling pathways, was indeed identified using a toxicity paradigm characterized by a very rapid necrotic response associated with immediate cell lysis.

Materials and Methods

Chemicals

Catalase, H2O2, antimycin A, rotenone, mepacrine, AA (sodium salt), ETYA and BAPTA/AM, as well as most of reagent grade chemicals, were obtained from Sigma-Aldrich (Milan, Italy). AACOCF3 was from Calbiochem (San Diego, CA, USA). HELSS was from ICN Pharmaceuticals (Milan, Italy). DHR was from Molecular Probes Europe (Leiden, The Netherlands). [3H]-AA was obtained from Amersham Pharmacia Biotech (Buckinghamshire, England).

Cell culture

U937 cells were cultured in suspension in RPMI 1640 medium (Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal bovine serum (FBS) (Biological Industries, Kibbutz Beit Haemek, Israel), penicillin (50 U/ml) and streptomycin (50 μg/ml) (Sera-Lab Ltd., Crawley Down, England), at 37°C in T-75 tissue culture flasks (Corning, Corning, NY, USA) gassed with an atmosphere of 95% air– 5% CO2.

RD U937 cells were isolated by culturing the cells in RPMI medium containing 400 ng/ml ethidium bromide, 100 μg/ml pyruvate and 5 μg/ml uridine for 6 days with changes of medium at days 2 and 4. These cells were unable to consume oxygen in response to glucose (5 mM) or to the membrane-permeant NADH-linked substrate pyruvate (5 mM; not shown).

Synthesis of peroxynitrite and treatment conditions

Peroxynitrite was synthesized by the reaction of nitrite with acidified H2O2 as described by Radi et al.,40 and MnO2 (1 mg/ml) was added to the mixture for 30 min at 4°C to eliminate excess H2O2. MnO2 was removed by centrifugation and filtration through 0.45 μm pore microfilters. The solution was frozen at −80°C for 24 h. The concentration of peroxynitrite, which forms a yellow top layer due to freeze fractionation, was determined spectrophotometrically by measuring the absorbance at 302 nm in 1.5 M NaOH (ɛ302=1670 M−1 cm1).

Treatments were performed in 2 ml of prewarmed treatment buffer (8.182 g/l NaCl, 0.372 g/l KCl, 0.336 g/l NaHCO3 and 0.9 g/l glucose) containing 5 × 105 cells. The cell suspension was inoculated into 15 ml tubes before the addition of peroxynitrite. Peroxynitrite was rapidly added on the wall of plastic tubes and mixed for few seconds to equilibrate the peroxynitrite concentration on the cell suspension; to avoid changes in pH due to the high alkalinity of the peroxynitrite stock solution, an appropriate amount of 1 N HCl was also added. Stock solutions of AA, mepacrine and catalase were prepared in distilled water. ETYA, AACOCF3, BAPTA/AM, HELSS, DHR and rotenone were dissolved in dimethyl sulfoxide. Antimycin A was dissolved in 99% ethanol. At the treatment stage, the final dimethyl sulfoxide and ethanol concentrations were never higher than 0.05%. Under these conditions, dimethyl sulfoxide and ethanol were not toxic.

Cytotoxicity assay

Cytotoxicity was determined with the trypan blue exclusion assay. Briefly, an aliquot of the cell suspension was diluted 1 : 1 (v/v) with 0.4% trypan blue and the cells were counted with a hemocytometer.

Measurement of extracellular release of [3H]-AA

The cells were labelled with [3H]-AA (0.5 μCi/ml) and grown for 18 h. Before treatments, the cells (2 × 105) were washed twice with the treatment buffer supplemented with 1 mg/ml fatty acid-free bovine serum albumin (BSA) and resuspended in a final volume of 1 ml of treatment buffer. The solution was then separated and centrifuged at 5000 × g for 1.5 min; 500 μl of the resulting supernatant was removed and radioactivity was determined in a Wallac 1409 liquid scintillation counter (Wallac, Turku, Finland).

cPLA2 and iPLA2 antisense oligonucleotides

The human cPLA2 antisense (5′-GTA AGG ATC TAT AAA TGA CAT-3′), the human iPLA2 antisense (5′-GAG GCG TCC AAA GAA CTG CAT-3′), the cPLA2 nonsense (5′-AGT AGA TTG AAT AGA CAC TAT-3′) and the iPLA2 nonsense (5′-ACG CGG ATC GGC AAT AAG TAC-3′) oligonucleotides were phosphorothioate modified and synthesized by MWG Biotech (Florence, Italy). U937 cells were washed twice with serum-free medium and seeded (1 × 106/ml) in serum-free RPMI 1640 for 6 h in the absence or presence of oligonucleotides (10 μM). A final concentration of 5% FBS was then added and the cells were cultured for an additional 24 h and utilized for experiments.

DHR oxidation and confocal imaging

Peroxynitrite-pretreated cells (5 min) were postincubated in the treatment buffer containing 10 μM DHR (10 min). The cells were washed three times and resuspended in 100 μl of phosphate-buffered saline (0.121 M NaCl, 10 mM NaH2PO4, 1.5 mM KH2PO4 and 3 mM KCl). This cell suspension (20 μl, 50 000 cells) was stratified on a slide and cellular fluorescence was then imaged using a confocal laser microscope (DVC 250, Bio Rad, Richmond, CA, USA) equipped with a Hamamatsu 5985 (Hamamatsu Italy, Milan, Italy) charge-coupled device camera. Cells were illuminated with the 488 nm line of the argon laser and the fluorescence emitted was monitored at λ>515 nm. The laser intensity, the shutter aperture and the exposure/integration settings were kept constant to allow quantitative comparisons of relative fluorescence intensity of cells between treatment groups. Laser exposure was limited to brief image acquisition intervals (<5 s) to minimize photo-oxidation of DHR. Confocal images were digitally acquired and processed for fluorescence determination at the single-cell level on a Macintosh G4 computer using the public domain NIH Image 1.63 program (developed at the US National Institutes of Health and available on Internet at http://rsb.info.nih.gov/nih-image/). Mean fluorescence values were determined by averaging the fluorescence of at least 50 cells/treatment condition/experiment.

Western blot analysis

After treatments, the cells were washed twice with phosphate-buffered saline, and incubated on ice for 1 h with the lysis buffer (50 mM Tris, 5 mM EDTA, 150 mM NaCl, 0.5% Nonidet P-40, 1 mM phenylmethylsulfonylfluoride, 1 mM sodium vanadate and 1 mM sodium fluoride, pH 8.0). Cells were then sonicated with a Sonicator Ultrasonic Liquid Processor XL (Heat System-Ultrasonics, Inc., NY, USA) and centrifuged at 21 500 × g for 10 min at 4°C to remove detergent-insoluble material. Supernatants were assayed for protein concentration using the Bio Rad protein assay reagent. Protein samples (25 μg) were resolved in 10% sodium dodecyl sulfate polyacrylamide gel and electrotransferred to polyvinylidene difluoride membranes. The blots were blocked for 1 h at room temperature with 5% milk powder in Tris-buffered saline (140 mM NaCl, 50 mM Tris-HCl, pH 7.2) containing 0.06% Tween-20 and probed with a primary mouse monoclonal antibody against human cPLA2 (1 : 500) or goat polyclonal anti-human iPLA2 (1 : 500) overnight at 4°C. Horseradish peroxidase-conjugated monoclonal (cPLA2) or polyclonal (iPLA2) antibodies (1 : 2000) were used for enhanced chemiluminescence detection. The antibodies were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Densitometric analysis of blots was performed using the electrophoresis documentation and the public domain NIH Image 1.63 program.

Immunocytochemical determination of cPLA2 in U937 cells

Cells treated with cPLA2 antisense or nonsense oligonucleotide (1 × 106/ml) were mounted on slides by cytospinning, fixed with 4% paraformaldehyde and then permeabilized with ethanol : acetic acid (95 : 5) for 1 min at room temperature. The permeabilized cells were blocked with 1% BSA in phosphate-buffered saline. The slide glass was then incubated overnight (4°C) with anti-human cPLA2 antibody (1 : 200 dilution of stock) in phosphate-buffered saline containing 1% BSA. Excess antibody binding was removed by washing the slide glass with phosphate-buffered saline. Exposure to the secondary antibody (rhodamine-labelled anti-mouse IgG), diluted 1 : 200 in phosphate-buffered saline, was for 3 h at 37°C. This antibody was from Molecular Probes Europe (Leiden, The Netherlands). After washing with phosphate-buffered saline, the cells were observed with a confocal microscope and the resulting images were taken and processed as described above.

Statistical analysis

All data in figures are expressed as mean±S.E.M. For comparison between two groups, the Student's unpaired t-test was used.