Different Prooxidant Levels Stimulate Growth, Trigger Apoptosis, or Produce Necrosis of Insulin-secreting RINm5F Cells THE ROLE OF INTRACELLULAR POLYAMINES”

and extensive DNA single strand breakage, which resulted in necrotic cell death. Our results show that a disturbance of polyamine bio- synthesis occurred prior to cell growth or apoptosis elicited by oxidative stress. In addition, we show that effects as opposite as cell proliferation and deletion, by either apoptosis or necrosis, can be induced, in the same sys- tem, by varying the exposure to a prooxidant. a Bio-Rad confocal micro-scope using the excitation line of typical chromatin morphology with whereas

In several pathophysiological conditions, excessive generation of free radicals can lead t o various degrees of oxidative stress. Moderate oxidative stress may selectively alter sensitive physiological processes, including cell signaling and gene activation (11, as well as the balance between DNA damage and repair (2). Radicals can modulate the activity of kinases ( 3 4 , promoting cell growth (6-8), and directly interact with protooncogenes (1). Thus, sustained exposure to moderate prooxidant levels may play an important role at several stages in carcinogenesis (9, 10). Conversely, when radical generation overwhelms cell antioxidant defense, lethal mechanisms are activated, and cell death ensues (11,12). Excessive oxidative stress Science Research Council (NFR, Project 10173-3001, The Regulatory * This work has been supported by grants from The Swedish Natural Board for Research on Laboratory Animals (CFN,, The The Swedish Medical Research Council (MFR 03X-2471). The costs of International Life Sciences Institute (ILSI) Research Foundation, and charges. This article must therefore be hereby marked "uduertisernent" publication of this article were defrayed in part by the payment of page in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ To whom correspondence should be addressed. Tel.: 46-8-728-7586;Fax: 46-8-314-2051. is known t o cause cell necrosis (131, but it has been also implicated in the activation of apoptosis in several systems, including cytokine-dependent killing of human immunodeficiency virus-infected cells (141, neural cell killing following glutathione depletion (15), and macrophage apoptosis induced by the fungal metabolite, gliotoxin (16). More recently, it has also become apparent that genes that suppress apoptosis may encode for antioxidant proteins (17). Thus, it appears that oxidative stress, which alters both signal transduction and genomic processes, may cause either inappropriate growth stimulation or activate the cell death program in different experimental systems.
Increasing evidence suggests that growth stimulation, besides triggering the entering of cells into the cell cycle, can act as surviving factor (18). Thus, cell cycle block can result in apoptosis (19), whereas mitogenic stimuli can prevent apoptosis (20). Consequently, should mitosis and cell death be regulated by related processes, it is conceivable that one or more regulatory proteins are differently expressed or activated in the normal cell cycle progression and in apoptosis. Ornithine decarboxylase (ODC),' a rate-limiting enzyme in polyamine biosynthesis, is induced after mitogenic stimulation of quiescent cells (21) and, in general, in cell proliferation (22). Induction of ODC and the other rate-limiting enzymes in polyamine biosynthesis, S-adenosyl-L-methionine decarboxylase (SAMDC) results in increased levels of intracellular putrescine, spermidine, and spermine. Increased intracellular levels of these polycations in dividing cells may have multiple functions, including a possible antioxidant effect (23) and a modulatory effect on chromatin structure (24). We have previously reported that supplementation of the culture medium with polyamines prevents thymocyte apoptosis, whereas inhibition of polyamine biosynthesis can sensitize cells to apoptosis (25). Accordingly, it has been proposed that mitogenic stimulation may prevent the activation of the cell death program by increasing ODC expression and the intracellular polyamine content (26).

or Necrosis after Oxidative Stress
duration of exposure to the quinone determined whether these cells were stimulated to grow, or to die by either apoptosis or necrosis. When the effects of DMNQ on polyamine biosynthesis were investigated, we found that intracellular polyamine depletion was closely associated with the onset of apoptosis. Preventing polyamine depletion by phorbol ester treatment also prevented apoptosis and restored cell proliferation.
Cell Cultures and Treatments-RINm5F cells were grown in RPMI 1640 medium supplemented with 10% (v/v) inactivated fetal calf serum, 1% (v/v) L-glutamine, and 1% (v/v) penicillin-streptomycin (all media and media supplements were from Life Technologies, Inc.) (28). I n all experiments, cells were seeded at a density of 20 x lo3 cells/cm2 and grown for 72 h to a density of approximately 0.5 x lo6 celldcm'. Cells were subsequently exposed in appropriate volumes of growth medium in cell culture vessels (Costar). DMNQ stocks were prepared by dilution with Me,SO. The final Me,SO concentration during exposure was always ~0 . 1 % . Control cells were exposed to Me,SO alone in all experiments.
Colony-forming E@ciency-RINm5F cells were cultured in 96-well plates as described above and subsequently exposed to DMNQ. After exposure, the cells were washed once in growth medium and then reincubated for 3 h in fresh medium to allow for completion of DNA repair. After that, cells were washed, trypsinized, and subcultured at clonal density (250 cells/cm2) in 60 mm dishes. Feeding was performed every 2nd day for 11 days when the experiments were interrupted. Cells were then washed twice with HBS, fixed with 10% formalin, and stained with 0.25% crystal violet. Colonies that contained more than 25 cells were counted and compared to those formed by untreated cells.
PHlThymidine Zncorporation-RINmSF cells were treated in the same manner as for the colony-forming efficiency assay with the exception that, after subcultivation, cells were incubated with [3Hlthymidine (2.5 pCi/ml) for 20 min before each time point. Cells were then washed four times with HBS, trypsinized, and counted for radioactivity.
Alkaline Elution-RINm5F cells were cultured in six-well plates in the presence of [l4C1thymidine (0.02-0.04 pCiiml) for two generations. A chase period of 24 h followed, when [l4C1thymidine was substituted with cold thymidine. Cells were then exposed to DMNQ. DNA single strand breakage was measured by the alkaline elution technique described by Kohn et al. (29). Results were calculated and expressed as the double logarithmic plot, normalized with respect to the elution of an internal standard [3H]thymidine-labeled L1210 cells irradiated with 3 grays).
oh8&-Formation of oh*dG was determined by HPLC using electrochemical detection according to the procedure described by Richter et al. (30). Agarose Gel Electrophoresis-Agarose gel electrophoresis of fragmented DNA was performed according to Wyllie et al. (31), with minor modifications. Samples were loaded on 1.8% agarose gels and electrophoresis was performed applying a 60-mA current. DNA was visualized under W light (305 nm), after staining with ethidium bromide (6 pgiml).
ODCActiuity-Exposures of RINm5F cells were performed in 60 mm dishes. At the time showed in the figures, cells were scraped off the dishes, pelleted (1000 x g for 5 rnin), and sonicated for 2 min in 50 pl of a lysis buffer containing 50 mM Tris-HC1 (pH 7.4), 0.1% Triton X-100, 4 mM EDTA, 5 mM dithiothreitol, 200 kallikrein inactivation unitdm1 Trasylol", 10 mM benzamidine, and 0.1 mg/ml albumin. The suspensions were centrifuged at 20,000 x g for 10 min at 4 "C. Supernatants were mixed with 50 pl of a Krebs-Ringer bicarbonate buffer supplemented with 0.3% (w/v) Hepes, 2 mM pyridoxalphosphate, 0.4 mM L-ornithine, and 10 pWml ~-[l-'~CIornithine and equilibrated under an atmosphere of 5% C0,:95% 0,. The assay was performed in glass vials inserted into an outer flask tightly sealed with a rubber membrane. Hyamine, (250 pl) was included in the outer flask to trap the 14C0, generated in the reaction. Samples were incubated for 60 min, at 37 "C in a slowly shaking waterbath. The reactions were ended by injecting 100 p1 of 5 M H,SO,. Radioactivity was measured after an overnight postincubation period (32,33). SAMDC Activity-SAMDC activity was determined by the same method described for the ornithine decarboxylase assay except the use of S-adenosyl-~-[carboxyl-~~Clmethionine as substrate (32,33).
termined by HPLC using a modification of the procedure described by Polyamine Determinations-Intracellular polyamine content was de- Wickstrom and Betn6r (34). Briefly, cells were cultured in 60-mm dishes as described above. After exposure, the cells were scraped off the plate, centrifuged a t 800 x g for 3 min and sonicated in 80 p l of 0.3 M perchloric acid at 4 "C. After centrifugation (12,000 x g for 5 min), the supernatant was neutralized by addition of 10 pl of a 1 M Na,CO, stock solution, samples were derivatized with Fmoc 9-fluorenyl-methylchloroformate (Fluka Chemie, Switzerland) and dissolved i n acetonitrile (HPLC grade, LabScan Analytical Sciences, Dublin, Ireland). The derivatization procedure was performed according to Wickstrom and Betner (34). The chromatographic system was a two-pump gradient system (Millipore, Waters, Milford, MA) consisting of model 501 pumps, model 680 gradient controller and a U6K injector. Fluorescence was monitored by a Schimadzu RF-530 detector, where the excitation wavelength was set at 260 nm and the emission wavelength at 315 nm. A Waters Baseline 810 Chromatography data system was used for peak integration. The column was a CT-si1 (C18), 5 p~ (250 x 4.6 mm) with a precolumn packed with the same material (10 x 3 mm; ChromTech, Norsborg, Sweden). The eluent A was a mixture of 70% acetate buffered with NaOH (50 mM, pH 4.2), and 30% acetonitrile (Fisous HPLC-solvent, FSA Laboratory Supplies, Loughborough, UK) and eluent B was 100% acetonitrile. Derivatization yield was determined by using 14C-labeled putrescine, spermidine, and spermine and was higher than 77% for all three polyamines. Standard concentration curves determined with this technique were linear between 5-800 pmol, for each polyamine.
DNA Staining for Confocal Microscopy-Cells detached from the dishes were harvested, fixed in 80% methanol, stained with propidium iodide (50 pg/ml) for 15 min and mounted onto glass-slides in glycerol: HBS (1:l). Cells were examined in a Bio-Rad MRC 600 confocal microscope using the 488 nm excitation line of the kryptodargon laser. Nuclei of untreated cells revealed a typical chromatin morphology with distinct organization, whereas apoptotic nuclei had high fluorescence and appeared condensed often with chromatin polarization.
Intracellular Total Thiol Content-Intracellular total thiol content was determined by the method of Saville (35). Results were expressed as GSH equivalentdlO6 cells and were illustrated as the percentage of control values.
ATP, NAD', and Cellular Energy Charge-The intracellular ATP, ADP, AMP, and NAD+ contents were determined in cell extracts by HPLC as described by Jones (36). Cellular energy charge was calculated according to the formula: EC = (ATP + 0.5 ADP)/(ATP + ADP + AMP).
Cytosolic Ca2+ Measurements-Intracellular Caz+ was measured fluorometrically by loading cells with the Ca2+ indicator Fura-Z/AM (37) as described previously (38) Cell Membrane Damage-Cell membrane damage was assessed by the spectrophotometric measurement of the lactate dehydrogenase leakage (39).
Statistical Analyses-Statistical analyses were performed using Student's t test for unpaired data.

Effects of DMNQ on RINm5F Cell
Growth-Incubation of RINm5F cells with 10 p f DMNQ for 1 h gradually stimulated [3H]thymidine incorporation with a maximum after 24 h. This treatment also enhanced the ability of these cells to form new colonies ( Fig. lA and inset). Maximal stimulation of colony forming efficiency was found when cells were exposed for 1 h, whereas exposure to 10 pv DMNQ for 3 h enhanced growth to a lesser extent. Lower DMNQ concentrations ( i e . 1 and 3 p v ) , had also a moderate stimulatory effect (Fig. lA 1. Incubation of RINm5F cells with a marginally higher DMNQ concentration, 30 PM, for 1 h, did not modify the proliferation rate, whereas incubation with 30 PM DMNQ for 3 h markedly decreased colony-forming efficiency. The decrease in colony-forming efficiency resulted from cell death, which became evident after 9 h and further increased after 12 h (Fig. 1B). Notably, cell death and the decrease in colony forming efficiency were prevented when cells were preincubated for 1 h with the phorbol ester, TPA, and then exposed to 30 p~ DMNQ for 3 h. TPA obviously stimulated growth of untreated cells when used alone. However, the combination of TPA pretreatment with exposure of the cells for 1 h to 30 PM DMNQ stimulated cell growth to an extent higher than that elicited by the single TPA treatment, At 100 p~, DMNQ caused cell killing after 1 h exposure, resulting in decreased colony forming efficiency (Fig. 1, A and B). In this case, pretreatment with TPA had no effect. Intracellular ODC, SAMDC Activities, and Polyamine Levels-When RINm5F cells were exposed to 10 PM DMNQ (Fig. 2 A ) , ODC activity increased. A moderate initial increase was observed within the first 30 min, whereas a second marked activation occurred between 1 and 3 h. The activity of ODC subsequently declined but remained at a level that was 2t o 3-fold higher than that observed before the addition of DMNQ. Pretreatment of cells with cycloheximide, to inhibit protein synthesis, abolished the increase in ODC activity observed after exposure to 10 p~ DMNQ (Fig. 2 A ) . Pretreatment with staurosporin, an inhibitor of protein kinase C-blocked DMNQinduced stimulation of ODC activity (Fig. 2 . 4 ) . Similar findings were obtained for SAMDC (data not shown).
At 30 p~, DMNQ increased ODC activity during the first 30 min. However, now ODC activity declined after 1 h, and by 2 h

!crasis after Oxidative Stress 30555
was reduced to about 50% of that found before exposure (Fig.  2B). Finally, exposure to 100 p~ DMNQ caused a n immediate and persistent inhibition of ODC activity (Fig. 2C). The activity of ODC did not significantly vary in control cells over the 6 h period. Pretreatment of RINm5F cells with the phorbol ester TPA stimulated ODC activity in untreated controls, but it did not prevent ODC inhibition after exposure to 30 1.1~ DMNQ (data not shown). Fig. 3 shows that SAMDC activity decreased only slightly after exposure to 30 p~ DMNQ. However, here, pretreatment with TPA prior to DMNQ exposure markedly increased the enzyme activity.
The changes observed in the activities of ODC and SAMDC were accompanied by alterations in intracellular polyamine content. As shown in Fig. 4A, spermidine and spermine levels increased with a maximum at 3 h following treatment with 10 PM DMNQ. Conversely, in cells exposed to 30 p~ DMNQ, the intracellular content of spermine decreased progressively during the first 3 h to about 50% of the control value (Fig. 4B). Finally, putrescine depletion occurred rapidly, and cells contained only 40% of this polyamine after 1 h (Fig. 4C). Preincubation of the cells with TPA for 1 h prior to the addition of 30 PM DMNQ increased the initial spermine and putrescine levels ( Fig. 4, B and C ) . Moreover, TPA pretreatment prevented spermine depletion and delayed the loss of putrescine caused by 30 PM DMNQ (Fig. 4, B and C ) . The loss of putrescine after 3 h of incubation with DMNQ was not prevented by TPA treatment. Since many cells progressively detached and shrank with a typical apoptotic appearance (see below), during the first 6 h of incubation, we decided to measure the polyamine content in these cells and in the population that had remained attached to the dishes. We harvested detached cells between 3 and 6 h after treatment with 30 PM DMNQ. Virtually all these cells had the typical apoptotic morphology, a condensed cytoplasm and a highly fluorescent pyknotic nucleus (Fig. 5 A ) . A few preapoptotic cells (ie. with patched and partially condensed chromatin) could still be seen among the adherent cell population (Fig. 5 B ) , whereas control cells had none of the above. In the apoptotic population, both putrescine and spermine levels were reduced to about 20% of those measured in the adherent population. In the apoptotic cells the putrescine and spermine contents were 18 2 4 pmol/106 cells and 176 2 20 pmol/106cells, respectively, whereas the adherent population had 108 2 28 pmol putrescine/106 cells and 812 .I-78 pmol spermine/106 cells.
DMNQ Induces DNA Damage in RINm5F Cells-Exposure of RINm5F cells to 10 or 30 p~ DMNQ resulted in base oxidation, as shown by the increased amounts of ohsdG found in the nuclear DNA (Table I). Nevertheless, the accumulation of DNA single strand breaks after exposure to 10 or 30 PM DMNQ for 1 h was relatively modest (51.5 single strand breaks x 10-lo daltons). Extensive single strand breakage was instead found after 1 h exposure to 100 PM DMNQ (Fig. 6A). Extending to 3 h the exposure to either 10 or 30 PM DMNQ caused only a moderate increase in single strand break formation (1.7 single strand breaks x 10-l' daltons with 30 p~ DMNQ). In contrast, agarose gel electrophoresis of DNA extracted from cells exposed to 30 PM DMNQ revealed a DNA fragmentation pattern indicative of internucleosomal cleavage (Fig. 6B), such as that caused by endonucleases during apoptosis (40). DNA fragmentation was evident at 3 and 6 h, as cells progressively lost adherence to the dishes, but did not release lactate dehydrogenase (see Fig. 1B). The events leading to apoptosis occurred between 1 and 3 h, since cells exposed to DMNQ for 1 h, washed, and then reincubated in fresh medium neither detached nor presented nuclear pyknosis and DNA fragmentation. Cells exposed to 10 PM DMNQ had no DNA fragmentation, whereas the DNA extracted from cells treated with 100 PM DMNQ exhibited high degrees of intranucleosomal DNA degradation, which appeared as a smear on the agarose gels (Fig.  6B). This is consistent with the observed accumulation of large amounts of single strand breaks.

Cell Growth, Apoptosis, or
Previous studies have shown that endonuclease-mediated DNA cleavage during apoptosis could be prevented by supplementing cells with polyamines (25) or by treatment with phor-bo1 esters (41). When spermine (10 p~) was added to the incubation medium, DNA fragmentation by 30 p~ DMNQ was partially prevented (Fig. 6C). In contrast, higher spermine concentrations were ineffective and in fact enhanced the nonenzymatic DNA damage (data not shown). Treatment with TPA (10 nM), which increased the intracellular spermine concentration (see above), was rather effective in preventing oligonucleo-soma1 DNA fragmentation and apoptosis of cells exposed to 30 p~ DMNQ (Fig. 6C). These pretreatments did not modify the DNA damage induced by 100 p~ DMNQ, nor did they affect control cultures.

Ca2' Overload and Depletion of GSH, ATP, and NAD+ Precede
RINm5F Cell Necrosis after Exposure to 100 ,LLM DMNQ-While treatment with either 10 or 30 p~ DMNQ did not cause sustained modifications of the resting Ca2+ level (Table II), a progressive intracellular Ca2+ overload followed the exposure of cells to 100 p~ DMNQ (Fig. 7A). The cytosolic free Ca2+ concentration increased up to 6 h, when a sizeable number of cells began to die. Incubation of RINm5F cells with 100 PM DMNQ also caused a progressive loss of GSH and ATP. Thiol levels rapidly declined, and, by 6 h, cells were virtually depleted of GSH (Fig. 7B 1. After 3 h, the ATP content of these cells was also reduced to about 15% of that measured in controls (Fig. 7C). Under these conditions cell energy charge was well outside the range required for cell survival (normal range: 0.85-0.95; the energy charge was 0.91 in controls and 0.66 in cells exposed to 100 p~ DMNQ for 3 h). Notably, both the glutathione level and cell energy charge were not significantly modified in cells exposed to either 10 or 30 JIM DMNQ (Table 11).
Rapid and marked depletion of the NAD+ pool also occurred in cells treated with 100 p~ DMNQ (Fig. 70). In contrast, as shown in Table 11, NAD+ levels were maintained after treatment with 10 p~ DMNQ. A moderate NAD+ decrease was also found in cells treated with 30 PM DMNQ, presumably because of the activation of DNA repair processes that require poly-ADP-ribosylation. Necrotic cell death (i.e. leakage of intracellular enzymes without DNA laddering and/or apoptotic body formation) followed the Ca2+ overload and the disruption of intracellular energy metabolism caused by the treatment with 100 p~ DMNQ. As shown in Fig. lB, after 3 h, cells had begun to release lactate dehydrogenase and by 9 h, only 40% of the population had retained viability.

DISCUSSION
In the present study, we show that different prooxidant concentrations can alternatively stimulate growth or cause deletion of pancreatic RINm5F cells. Growth stimulation and the progression of the cell through the cell cycle are normally associated with the rapid induction of two short-lived enzymes involved in polyamine biosynthesis, ODC and SAMDC (42,43). Protein kinase C activation mediates the induction of these enzymes (44,45). In RINmFS cells exposed to low level oxidative stress, the involvement of protein kinase C in ODC induction is suggested by the effect of staurosporin, which abolished the growth stimulation elicited by 10 PM DMNQ and blocked ODC activation. The sensitivity to cycloheximide pretreatment also suggests that the increased enzyme activity be due to induction of new ODC molecules. The rapid increase of ODC activity correlated with the rise of intracellular polyamine levels up to 3 h after exposure to DMNQ, whereas the subsequent decrease in ODC activity was, most likely, the result of feedback inhibition by the elevated polyamine levels (42, 43).
When the DMNQ concentration was raised to 30 p~, cells died by apoptosis. Several observations have suggested a link between the mechanisms regulating cell proliferation and apoptosis. For example, deregulation of c-myc proto-oncogene ex-FIG. 5. Apoptotic nuclei in RINm5F cells exposed to 30 p M DMNQ. Panel A, following the exposure to DMNQ. cells which lost adherence to the dishes were harvested, collected and treated as indicated under "Experimental Procedures." At the confocal microscope, nuclei appeared highly fluorescent, typically pyknotic and with condensed chromatin. Their average size was half of that measured in untreated cells. Panel B, cells that had remained adherent to the dishes after treatment with DMNQ displayed normal chromatin structure. A potential pre-apoptotic nucleus larger in size, with patched chromatin, is indicated by the arrow.

TABLE I
Formation of ohRdG in RINmSF cells exposed to DMNQ Cells were exposed to DMNQ and processed for ohSdG determination as indicated under "Experimental Procedures." Each value represents the mean t S.D. from seven seDarate determinations. While non-nuclear proteins may produce the characteristic nuclear changes of apoptosis in some systems (47), there is substantial evidence suggesting that modifications of the chromatin structure may be the primary signal for gene degradation and death in others (48). Previous studies in our laboratory have shown that DNA fragmentation and apoptotic body formation in thymocytes could be prevented by manipulating the intracellular content of polyamines (25). The effect of the polyamines, primarily spermine, correlated with their ability to promote chromatin compaction (24). Thus, it appeared conceivable that the suppression of apoptosis in thymocytes was due to their ability to reduce chromatin unfolding (25).
Polyamine depletion seems to play a role in chromatin degradation after exposure of RINm5F cells to 30 p~ DMNQ, as suggested by the following observations: (i) apoptotic cells were

TABLE I1
For Ca2+ measurements, cells exposed to DMNQ were preloaded with Fura 2/AM for 30 min a t 37 "C, as described in detail in Dypbukt et ul. (38). virtually depleted of putrescine and spermine; (ii) pretreatment with TPA before exposure to 30 VM DMNQ restored spermine content and prevented both DNA fragmentation and cell killing; (iii) addition of extracellular spermine reduced DNA fragmentation. The observation that treatment with the phorbol ester restored both the spermine level and the ability of the cells to grow supports the idea that DMNQ-induced depletion of intracellular polyamines was, a t least in part, related to changes in protein kinase C activity. However, ODC is also sensitive to direct oxidation and it is possible that this may have contributed to the inactivation of the enzyme and to its proteolytic cleavage (49). This latter assumption is supported by the finding that TPA pretreatment did not prevent the irreversible inhibition of ODC activity caused by DMNQ. It could then be questioned how TPA pretreatment would prevent spermine depletion without restoration of ODC activity. This can be explained by the finding that treatment with the phorbol ester prevented SAMDC inhibition by 30 PM DMNQ. SAMDC would provide decarboxylated S-adenosylmethionine for spermidine and spermine synthases (42,431 resulting in increased spermidine and spermine levels, at the expense of putrescine. Our findings support this assumption, although we cannot exclude the involvement of additional alterations in polyamine metabolism. Furthermore, exposure for 1 h to 30 PM DMNQ neither depleted intracellular polyamines nor did it trigger apoptosis. It is conceivable that reincubation in growth medium after 1 h may have supplemented mitogenic factors preventing polyamine depletion. In contrast, exposure of RINm5F cells to 30 PM DMNQ for 3 h followed by medium replenishment still caused cell death. This suggests that apoptosis was activated between 1 and 3 h, which coincides with the time when the intracellular polyamine level decreased. In this context, an increased intracellular polyamine content after mitogenic stimulation would provide protection against excessive chromatin exposure, whereas polyamine depletion may favor chromatin unfolding and cleavage during apoptosis.

Determinations
DNA damage induced by 30 VM DMNQ appeared predominantly as oligonucleosomal fragmentation. Both 10 and 30 p~ DMNQ stimulated ohsdG formation and moderate DNA single  Table I and Fig. 6A). However, since both lethal and non-lethal DMNQ concentrations caused similar damage, it may be concluded that direct DNA damage by active oxygen species was not acutely lethal in this system. On the other hand, direct oxidative DNA damage, such as that induced by X-irradiation, can actually trigger apoptosis by inducing wild-type p53 gene expression (50,51). This, in turn, results in the activation of growth arrest genes, cell cycle block, and death (52). Accordingly, alterations in the cell cycle clock due to increased demand for DNA repair, or to unbalanced mitogenic signals may favor gene exposure and chromatin degradation, which would be counteracted by a rise in the intracellular polyamine level. Following induction of the nitric oxide synthase activity and the subsequent nitric oxide accumulation (53), RINm5F cells express p53 protein2 and undergo apoptosis. It is conceivable that p53 expression may also take part in DMNQinduced RINm5F apoptosis, although, at present, we have no conclusive evidence to support this assumption.
Finally, RINm5F cells rapidly died by necrosis when exposed to the highest prooxidant level. Under these conditions, cell * M. Ankarcrona  lysed without DNA laddering or apoptotic body formation. Cell death was preceded by GSH, ATP and NAD+ depletion, intracellular Ca2+ accumulation, and widespread DNA single strand breakage. It is therefore difficult to decide the most lethal event. Most likely, the recruitment of several catabolic events, with high degrees of oxidative stress, led to such swift cell disintegration that the selection of the organized, apoptotic program was prevented. Notably, while energy levels are believed to be preserved in apoptosis, necrosis is often associated with marked ATP depletion. Supporting this assumption is the observation that cell energy charge was preserved in cells exposed to 30 PM DMNQ (Table 11). Similar findings have recently been obtained in experiments where neural cells underwent either apoptosis or necrosis following exposure t o different glutamate concentration^.^ Therefore, one critical factor in deciding the type of death may be the energy level of the cell. Cells depleted of ATP and with a marked ion imbalance swell and rapidly lyse, whereas energy-dependent cell shrinkage and chromatin compaction would require a preserved energy M. Ankarcrona, J. M. Dypbukt, and P. Nicotera, unpublished observations.