Introduction

Top Abstract Introduction Materials & Methods Results Discussion Conclusions References

Hydrogen peroxide generates an array of different lesions within the cell, including damage at the genomic DNA level (Cantoni et al., 1995). These lesions are mainly represented by DNA SSBs because DNA double strand breaks cannot be detected even under conditions in which the oxidant produces an enormous amount of DNA SSBs (Ward et al., 1985 and 1987; Cantoni et al., 1986, 1989, 1992). A large body of experimental evidence indicates that the effects at the DNA level are mediated by the so-called Fenton reaction in which the hydroxyl radical, the ultimate DNA-damaging species, is formed as a consequence of the interaction between the oxidant and divalent iron (Mello Filho and Meneghini, 1984; Mello Filho et al., 1984). Organic hydroperoxides, and in particular model compounds such as tB-OOH or cumene hydroperoxide, have also been shown to promote DNA single strand breakage in the absence of detectable DNA double strand breakage (Guidarelli et al., 1995). Understanding the mechanism whereby tB-OOH generates DNA single strand breakage is of considerable importance since this and other organic hydroperoxides, while inactive as initiators or complete carcinogens (O'Connel et al., 1986; Slaga et al., 1983), are tumor promoters in the skin of SENCAR mice (Taffe et al., 1987). Few studies, however, have investigated the molecular basis for these effects; most importantly, the identity of the species mediating DNA damage in cells exposed to organic hydroperoxides remains largely unexplored. It has been reported that DNA cleavage evoked by tB-OOH is abolished by iron chelators (Guidarelli et al., 1995, 1997; Coleman et al., 1989; Latour et al., 1995), is insensitive to antioxidants (Guidarelli et al., 1995, 1997; Coleman et al., 1989; Latour et al., 1995) and is repaired in a relatively short time (Guidarelli et al., 1995, 1997; Coleman et al., 1989; Sandström, 1991; Baker and He, 1991). The fact that antioxidants abolished cell death induced by tB-OOH without preventing the formation of DNA lesions (Coleman et al., 1989; Guidarelli et al., 1997) suggests that the species involved in the cyto- and geno-toxic responses are different. Consistent with this possibility are our recent results showing that the complex III inhibitor antimycin A reduces the toxicity of tB-OOH as well as the formation of tB-OOH-derived radical species (methyl and tert-butoxyl radicals), while markedly enhancing the accumulation of DNA SSBs (Guidarelli et al., 1996). These findings collectively suggest the existence of similarities between the genotoxic effects of tB-OOH and those of H₂O₂ and imply that iron-dependent processes play a major role in the formation of free-radical intermediate(s) of an as yet unknown nature, which could ultimately generate DNA lesions in cells challenged with tB-OOH. Recent results from our laboratory, however, indicate that DNA cleavage caused by tB-OOH is a rather complicated process in which the formation of DNA-damaging species or, more likely, of their precursors, requires additional mechanisms operating in the cytosol and is modulated by calcium ions. Indeed, tB-OOH, unlike H₂O₂, was unable to elicit DNA strand scission in purified DNA or isolated nuclei (Guidarelli et al., 1997). Furthermore, the DNA damaging efficiency of tB-OOH was remarkably decreased when the basal cytosolic calcium concentrations were lowered with EGTA, whereas only a slight decline in the number of DNA SSBs was observed in cells treated with H₂O₂ (Guidarelli et al., 1997). These data indicate that the biochemical pathways leading to DNA damage in cells exposed to organic or inorganic hydroperoxides coincide only in part.

In our study we investigated the nature of the involvement of calcium ions in the formation of DNA SSBs induced by tB-OOH. The results obtained indicate that tB-OOH releases calcium ions from intracellular stores and that a significant proportion of the cation is cleared by the mitochondria. Mitochondrial calcium plays a pivotal role in the process of DNA strand scission promoted by this organic hydroperoxide. H₂O₂ was found to be one the species that are produced within the mitochondria and which mediate part of the DNA single strand breakage elicited by tB-OOH in cultured U937 cells. The relative amount of DNA lesions mediated by H₂O₂ would appear be dependent on the intramitochondrial calcium content.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion Conclusions References

Materials. Fura-2 AM, Tg and Iono, were purchased from Calbiochem, San Diego, CA. Cf, FCCP, RR, Ry, catalase, SOD, tB-OOH, H₂O₂, menadione and the remaining chemicals were from Sigma-Aldrich, Milano, Italy. RPMI 1640 culture medium was from GIBCO, Grand Island, NY, and fetal bovine serum, penicillin and streptomycin were from Seralab, Sussex, UK. T-75 tissue culture flasks were purchased from Corning, Corning, NY. [¹⁴C]-thymidine was obtained from NEN/Dupont, Boston, MA. Polycarbonate filters and liquid scintillation were purchased from Nuclepore, Pleasanton, CA and Beckman, Fullerton, CA, respectively.

Cell culture and treatments. Human myeloid leukemia U937 cells were cultured in suspension in RPMI 1640 culture medium supplemented with 10% fetal bovine serum, penicillin (50 U/ml), and streptomycin (50 µg/ml), at 37°C in T-75 tissue culture flasks in a humidified atmosphere of 95% air-5% CO₂.

[Ca⁺⁺]_i Measurements. Cells were harvested, washed three times by centrifugation and resuspended in Krebs Ringer Hepes medium containing 125 mM NaCl, 5 mM KCl, 1.2 mM KH₂PO₄, 1.2 mM MgSO₄, 2 mM CaCl₂, 6 mM glucose, 25 mM Hepes-NaOH (pH 7.4). Cell suspensions were loaded with the Ca⁺⁺-sensitive dye fura-2 AM (3 µM final concentration) for 30 min at 25°C in Krebs Ringer Hepes medium and kept at 37°C until use. Cell aliquots (4 × 10⁶ cells) were washed three times and resuspended in saline A, transferred to a thermostatted cuvette in a Perkin Elmer (Norwalk, CT) LS-50 fluorimeter and maintained at 37°C under continuous stirring. The various drugs interfering with Ca⁺⁺ homeostasis here employed (Tg, Iono, tB-OOH, FCCP and Cf) were added as indicated in the figures and maintained throughout the experiment. In the experiments in which Ry was used, preincubations with this drug were for 5 min before the beginning of the recording. Traces were recorded and analyzed as previously described (Grynkiewicz et al., 1985). The results shown are traces representative of 8 to 10 highly consistent experiments.

Measurement of DNA SSBs by alkaline elution. The cells were labeled overnight with [¹⁴C]-thymidine (0.05 µCi/ml) and incubated for a further 6 hr in a medium containing unlabelled thymidine (1 µg/ml). At this stage the cells (2.5 × 10⁵/ml) were either treated in saline or permeabilized and treated for 10 min in permeabilization buffer. Permeabilization was achieved by adding digitonin (10 µM, 12.5 µg/10⁵ cells) to a medium consisting of 0.25 M sucrose, 0.1% bovine serum albumin, 10 mM MgCl₂, 10 mM K⁺-Hepes, 5 mM KH₂PO₄, pH 7.2 at 37°C. Under these experimental conditions, digitonin permeabilizes the plasma membrane but leaves mitochondrial membranes intact (Fiskum et al., 1980). After the treatments the cells were washed with prechilled saline A and analyzed immediately for DNA damage using the alkaline elution technique that was carried out using a procedure virtually identical to that described in (Kohn et al., 1981) with minor modifications (Cantoni et al., 1986). Briefly, 3.5 to 4 × 10⁵ cells were gently loaded onto 25 mm, 2-µm pore polycarbonate filters and then rinsed twice with 10 ml of ice-cold saline A containing 5 mM EDTA (disodium salt). Cells were lysed with 5 ml of 2% sodium dodecyl sulfate, 0.025 M EDTA (tetrasodium salt), pH 10.1. Lysates were rinsed with 7 ml of 0.02 M EDTA (tetrasodium salt) and the DNA was eluted overnight in the dark with 1.5% tetraethyl ammonium hydroxide/0.02 M EDTA (free acid)/0.1% sodium dodecyl sulfate (pH 12.1), at a flow rate of ca. 30 µl/min. Fractions were collected at 2-hr intervals and counted in 7 ml of liquid scintillation containing 0.7% glacial acetic acid. DNA remaining on the filters was recovered by heating for 1 hr at 60°C in 0.4 ml of 1N HCl followed by the addition of 0.4 N NaOH (2.5 ml) and was again determined by scintillation counting. DNA was also recovered from the interior of the membrane holders after vigorous flushing with 3 ml of 0.4 N NaOH. This solution was processed for scintillation counting as described above. Strand scission factor values were calculated from the resulting elution profiles by determining the absolute log. of the ratio of the percentage of DNA retained in the filters of the drug-treated sample to that retained from the untreated control sample (both after 8 hr of elution).

	Results

Top Abstract Introduction Materials & Methods Results Discussion Conclusions References

tB-OOH-induced DNA strand scission is associated with a rise in cytosolic calcium ion concentration and with an increased mitochondrial calcium uptake. A number of studies had previously demonstrated that tB-OOH enhances the intracellular concentration of free calcium ions ([Ca⁺⁺]_i) (Thor et al., 1984; Sakaida et al., 1991; Livingston et al., 1992). Similarly, we found that a short treatment with 200 µM tB-OOH elevated [Ca⁺⁺]_i in U937 cells (fig. 1). This [Ca⁺⁺]_i increase was due to release from internal stores, since the experiments were performed utilizing nominally calcium-free medium (saline A). Furthermore, similar results (not shown) were obtained using 10 µM EGTA-containing saline A (estimated extracellular [Ca⁺⁺] ~ 0⁸ M). The Ca⁺⁺ pool mobilized by tB-OOH appeared to be of neutral pH, because it was dischargeable by the protonophore Iono (fig. 1A), and insensitive to cell pretreatment with either the sarcoplasmic/endoplasmic reticulum Ca⁺⁺ - ATPase (SERCA) blocker Tg (fig. 1A), the IP₃-generating agonist ATP (not shown), or the Ry receptor agonist Cf (fig. 2B). These observations collectively indicate that tB-OOH mobilizes Ca⁺⁺ from pools that are different from the ER-located, SERCA-containing, IP₃- and Ry-sensitive Ca⁺⁺ stores. Application of the protonophore FCCP (10 µM) 5 min after peroxide addition revealed that, under these conditions, mitochondrial calcium uptake is a major route for clearance of the released calcium ions (fig. 1B). To explore the relationships existing between this event and the formation of DNA lesions promoted by tB-OOH we tested the effect of RR, an inhibitor of mitochondrial calcium uptake (Carafoli, 1987). As illustrated in figure 1C, 25 µM RR markedly reduced the extent of DNA strand scission caused by tB-OOH. In these experiments, the inhibitor was given for 5 min before, and maintained during, the 30-min exposure to the hydroperoxide. RR was ineffective when present only during the pre-exposure phase (not shown). In figure 1D, it can be seen that the inhibitor was also ineffective when added to the cultures 10 min (or more) after treatment with tB-OOH, a result that may well be explained by the fact that most of the DNA SSBs are generated within 10 min of exposure to tB-OOH (inset to fig. 1D). It is important to emphasize that, under the experimental conditions used in this study, tB-OOH was not cytotoxic (not shown, see "Materials and Methods"). The protective effects afforded by RR cannot be ascribed to iron-chelating or radical-scavenging mechanisms, since under the same experimental conditions in which it inhibited DNA strand scission induced by tB-OOH, RR did not affect the DNA-damaging response evoked by 50 µM H₂O₂ (fig. 1C). Importantly, under these conditions, H₂O₂ did not produce significant changes in [Ca⁺⁺]_i (not shown). Finally, the results shown in figure 1C demonstrate that the effect of RR was not a consequence of possible interactions with Cf-sensitive intracellular Ca⁺⁺ release channels (Ry receptors, Berridge, 1993) because 20 µM Ry, while able to abolish the elevation in [Ca⁺⁺]_i promoted by 10 mM Cf (fig. 2A), did not modify DNA damage generated by either H₂O₂ or tB-OOH (fig. 1C).

View larger version (28K): [in this window] [in a new window]   Fig. 1.   tB-OOH elevates [Ca++]i and enhances mitochondrial calcium uptake that appears to mediate part of the DNA single strand breakage caused by tB-OOH in intact U937 cells. A, Fura-2 loaded U937 cell suspensions were supplemented with saline A (continuous trace), Tg (30 nM, dashed line) or Iono (500 nM, dotted line) as indicated. tB-OOH (200 µM) was added to all samples 5 min later. The numbers to the left indicate the [Ca++]i values. Traces are representative of eight consistent experiments. B, Fura-2 loaded cell suspensions were supplemented with saline A (dotted line) or tB-OOH (200 µM, continuous trace) where indicated, i.e., 5 min before the addition of FCCP (10 µM). The numbers to the left indicate the [Ca++]i values. Traces are representative of eight to ten consistent experiments. C, The cells were exposed for 5 min in saline A to either 25 µM RR or 20 µM Ry and then treated for further 30 min with 200 µM tB-OOH or 50 µM H2O2. The level of DNA SSBs was measured immediately after the treatments using the alkaline elution technique. Results represent the mean ± S.E.M. calculated from three to five separate experiments, and were significantly different from those for DNA damage generated by the hydroperoxide alone at *P < .001 (unpaired t test). D, The cells were treated for 30 min in saline A with 200 µM tB-OOH in the absence or presence of 25 µM RR, which was added at different times after the hydroperoxide. The abscissa axis indicates the time elapsed between the addition of tB-OOH and subsequent addition of RR. Results represent the means ± S.E.M. of the percent inhibition of DNA SSBs induced by tB-OOH calculated from three to five separate experiments and were significantly different from those for DNA damage generated by the hydroperoxide alone at *P < .001 or **P < .01 (unpaired t test). The inset shows the level of DNA SSBs induced by treatment with 200 µM tB-OOH for increasing time intervals. Results represent the mean ± S.E.M. calculated from three separate experiments.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Caffeine enhances DNA SSBs caused by tB-OOH in intact U937 cells as well as the effects of the hydroperoxide on mitochondrial calcium uptake. A, Fura-2 loaded U937 cell suspensions were supplemented with saline A (dashed line) or Cf (10 mM, continuous trace) where indicated, i.e., 10 min before FCCP (10 µM) addition. The dotted line refers to cells incubated with Ry (20 µM), 5 min before Cf (and FCCP) administration. The numbers to the left indicate the [Ca++]i values. Traces are representative of ten consistent experiments. B, Conditions as in A. Addition of tB-OOH (200 µM) to all samples was 5 min after saline A (dashed line) or Cf (continuous trace) administration. The dotted line refers to cells preincubated with Ry. C, The cells were exposed for 5 min in saline A to 0 or 10 mM Cf and then treated for further 30 min with 200 µM tB-OOH. 25 µM RR or 20 µM Ry were added 5 min before Cf. The level of DNA SSBs was measured immediately after the treatments using the alkaline elution technique. Results represent the mean ± S.E.M. calculated from three to five separate experiments, and were significantly different from those for DNA damage generated by tB-OOH alone at *P < .001 or by tB-OOH associated with Cf at (*)P < .001 (unpaired t test).

Caffeine enhances mitochondrial calcium uptake as well as the extent of DNA cleavage in cells exposed to tB-OOH. We next investigated whether release of Ca⁺⁺ from intracellular stores other than those mobilized by tB-OOH could further enhance the DNA-damaging effects of the hydroperoxide as well as its effects on mitochondrial Ca⁺⁺ uptake. For this purpose we used Cf that, at high concentrations, had previously been shown to promote the efflux of calcium ions from the ER-located, Tg-sensitive Ca⁺⁺ pool via opening of the Ry receptor in a number of different cell lines (Berridge, 1993). This was also true in U937 cells because addition of 10 mM Cf resulted in a transient increase in [Ca⁺⁺]_i (fig. 2A), an effect abolished by prior Tg-treatment (not shown). Cf-induced Ca⁺⁺ release was followed by enhanced mitochondrial calcium accumulation (fig. 2A). Addition of tB-OOH 5 min after the application of Cf further enhanced mitochondrial calcium uptake (fig. 2B). Interestingly Cf, although not producing DNA strand scission (not shown), markedly enhanced the extent of DNA cleavage caused by tB-OOH (fig. 2C). RR abolished this response and the extent of DNA damage detected under these conditions was identical to that observed after treatment with tB-OOH and RR (compare figs. 1C and 2C). Ry also reduced the Cf-mediated enhancement of tB-OOH-induced DNA single strand breakage (fig. 2C) leading to the same level of DNA damage generated by tB-OOH alone (fig. 1C). Consistent with these results, Ry was found to prevent the increase in [Ca⁺⁺]_i and mitochondrial calcium accumulation elicited by Cf (fig. 2A) as well as the effects of Cf on mitochondrial calcium accumulation mediated by tB-OOH (fig. 2B).

The mitochondrial calcium uptake-based mechanism either does not alter the identity of the DNA-damaging species produced by tB-OOH or results in the formation of different species with similar reactivities mediating similar types of DNA lesions. Previous studies demonstrated that lipid peroxidation products may display DNA-damaging properties (Ochi and Cerutti, 1987; Brambilla et al., 1986). These species, however, did not appear to mediate DNA cleavage generated by tB-OOH in U937 cells, because this response was not inhibited by antioxidants under the same experimental conditions in which these agents prevented cell death caused by the hydroperoxide (Coleman et al., 1989; A. Guidarelli, P. Sestili, O. Cantoni, unpublished). Iron chelators suppressed the formation of DNA lesions as well as the toxicity induced by tB-OOH (Guidarelli et al., 1995). The results illustrated in table 1 confirm the above findings and indicate that DNA SSBs generated by the hydroperoxide in the presence of Cf were also insensitive to antioxidants and abolished by the membrane-permeant iron chelator o-pt. Similar results were obtained in cells treated with the combination tB-OOH/RR (table 1) or with H₂O₂ (Guidarelli et al., 1997). In other experiments, the cells were treated for 30 min with 200 µM tB-OOH, alone or associated with 10 mM Cf, and the level of DNA SSBs was measured immediately after the treatments as well as at different time intervals of posttreatment incubation in fresh culture medium. The results illustrated in figure 3 clearly indicate that the DNA SSBs induced under these two experimental conditions are removed with kinetics that are basically identical. Interestingly, the same kinetics of DNA SSB-removal were observed when the initial damage was induced by tB-OOH associated with 25 µM RR or by a 30-min exposure to 50 µM H₂O₂.

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Kinetics of DNA SSB-rejoining after treatment with tB-OOH in the absence or presence of caffeine or ruthenium red. The cells were exposed for 30 min to 200 µM tB-OOH alone or associated with either 10 mM Cf or 25 µM RR and then allowed to repair in fresh prewarmed RPMI medium containing 10% fetal bovine serum. In other experiments the cells were exposed for 30 min to 50 µM H2O2. The level of DNA SSBs was measured immediately after the treatments as well as after various time intervals of repair. Data points are the means of two separate experiments. The kinetics of DNA SSB removal illustrated appear to be first order with respect to time and linear regression analysis of the data points obtained under the different experimental conditions leads to the estimated t1/2 values that are reported in the figure.

Calcium-dependent modulation of DNA single strand breakage in permeabilized cells exposed to tB-OOH. Previous studies from our laboratory demonstrated that tB-OOH is an efficient inducer of DNA SSBs in intact cells but fails to generate strand scission in partially purified DNA or in isolated nuclei (Guidarelli et al., 1997).

View larger version (22K): [in this window] [in a new window]   Fig. 4.   tB-OOH induces DNA single strand breakage in permeabilized U937 cells and this effect is abolished by EGTA and enhanced by caffeine or CaCl2. A, Digitonin-permeabilized cells were treated for 10 min with 200 µM tB-OOH either alone or associated with 10 mM Cf in the absence or presence of 1 mM EGTA. DNA SSBs were measured as detailed in the legend to figure 1C. Treatment with EGTA or Cf did not produce DNA SSBs. The results are the means of three to four separate experiments and were significantly different from those for DNA damage generated by the tB-OOH alone* or by tB-OOH associated with Cf(*) at P < .0001 (unpaired t test). B, Permeabilized cells were exposed for 10 min to increasing concentrations of CaCl2 in the absence or presence of 50 µM H2O2 or 200 µM tB-OOH either alone or associated with 10 mM Cf. Treatment with CaCl2 did not produce DNA SSBs. Data are expressed as % of DNA damage caused by the different treatments in the absence of exogenous calcium ions and represent the means of three separate experiments. Experimental points were significantly different from appropriate controls at *P < .0001, **P < .001 (unpaired t test). C, Permeabilized cells were treated for 10 min with 200 nM RR or 20 µM Ry in the presence of 200 µM tB-OOH either alone or associated with 10 mM Cf or 30 µM CaCl2. Treatment with RR or Ry did not produce DNA SSBs. Data represent the means ± S.E.M. of the percent inhibition of DNA single strand breakage induced by tB-OOH alone or associated with caffeine or Ca++ which was statistically significant at *P < .0001, **P < .001 (unpaired t test).

H₂O₂ mediates part of the DNA strand scission caused by tB-OOH/Ca⁺⁺. Previous reports indicated that mitochondrial production of superoxide anions and hydrogen peroxide is sensitive to mitochondrial Ca⁺⁺ content (Cadenas and Boveris, 1980) and, in particular, it was recently suggested (Valle et al., 1993; Castilho et al., 1995) that mitochondrial calcium accumulation leads to an enhanced formation of H₂O₂ in mitochondria exposed to tB-OOH. It is therefore possible that mitochondrial calcium uptake enhances the accumulation of DNA lesions in cells treated with tB-OOH via enforced mitochondrial formation of superoxides and hydrogen peroxide. To test this hypothesis we took advantage of the permeabilized cell system described above with the specific aim of assessing the effects of catalase and/or SOD on DNA damage induced by tB-OOH both in the absence and presence of CaCl₂. As illustrated in figure 5, 10 Sigma U/ml of bovine catalase were able to significantly reduce the extent of DNA strand scission caused by 200 µM tB-OOH and afforded a much greater protection against DNA cleavage caused by the hydroperoxide in the presence of 30 µM CaCl₂. Under this second experimental condition, however, catalase markedly reduced but did not abolish the enhancing effects of calcium ions. However, catalase prevented DNA strand scission induced by H₂O₂ (50 µM) or by the redox-cycling quinone menadione (25 µM) (fig. 5, inset). Importantly, the protective effects of catalase disappeared when the enzyme was boiled prior to addition to the cultures. SOD (200 U/ml) neither affected the accumulation of DNA lesions in cells challenged with tB-OOH alone or combined with CaCl₂, nor did it modulate the protective effects elicited by catalase (fig. 5).

View larger version (38K): [in this window] [in a new window]   Fig. 5.   Catalase prevents the Ca++-mediated enhancement of DNA SSBs induced by tB-OOH in permeabilized U937 cells. Digitonin-permeabilized cells were treated for 10 min with 200 µM tB-OOH either alone or associated with 30 µM CaCl2 in the absence or presence of 10 Sigma U/ml of catalase or 200 U/ml SOD or with the combination of these two enzymes. The inset shows the effect of catalase on DNA strand scission caused by 50 µM H2O2 or 25 µM menadione. The specificity of the inhibitory effect promoted by catalase is emphasized by the lack of effect of the temperature-inactivated (boiled) enzyme. The level of DNA SSBs was measured immediately after the treatments using the alkaline elution technique. Results represent the mean ± S.E.M. calculated from three to five separate experiments and were significantly different from those for DNA damage generated by menadione, H2O2 and tB-OOH alone or associated with CaCl2 at *P < .0001, **P < .001 or ***P < .01 (unpaired t test).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion Conclusions References

The results presented in this study define a previously unexpected role of mitochondria in the formation of lesions at the level of genomic DNA. In particular, we report experimental evidence demonstrating Ca⁺⁺-dependent mitochondrial formation of species, mainly represented by H₂O₂, which mediate DNA cleavage triggered by tB-OOH. In addition, we demonstrate that agents increasing mitochondrial Ca⁺⁺ accumulation evoke a parallel enhancement of the tB-OOH-induced genotoxic response.

Our results demonstrate that treatment with a subtoxic albeit DNA-damaging concentration of tB-OOH promotes a transient elevation in [Ca⁺⁺]_i (fig. 1A) and that a significant proportion of the cation was cleared by the mitochondria (fig. 1B). The cation was released from internal stores that are of neutral pH (fig. 1A) and different from the ER-located, SERCA-containing, IP₃- and Ry-sensitive Ca⁺⁺ stores (not shown and fig. 2B).

The first experimental evidence providing a link between mitochondrial calcium uptake and the formation of DNA lesions generated by tB-OOH was that RR significantly reduced the DNA-damaging response evoked by the hydroperoxide (fig. 1C). It is important to note that RR, although being a potent inhibitor of the calcium uniporter of mitochondria (Carafoli, 1987), can also inhibit the calcium efflux from the Ry receptor (Berridge, 1993). Furthermore, a number of different studies have reported that RR can exert antioxidant and scavenging as well as redox properties (Bellomo et al., 1984; Bernardes et al., 1986; Vercesi et al., 1988; Weis et al., 1994). These possibilities, however, were ruled out by the observation that Ry and RR did not modify the extent of the DNA-damaging responses evoked by tB-OOH and H₂O₂, respectively (fig. 1C). Thus, the effect of RR on the formation of DNA SSBs induced by tB-OOH appears to be specifically associated with inhibition of mitochondrial calcium uptake. This inference is further supported by the experimental results that will be discussed below.

As a second approach to investigate the effect of an elevation in [Ca⁺⁺]_i in general and more specifically of mitochondrial calcium accumulation on the DNA-damaging effects of the hydroperoxide, we assessed the effects of release of Ca⁺⁺ from intracellular stores other than those mobilized by tB-OOH. On the basis of the results discussed above we used Cf, which specifically releases Ca⁺⁺ from the Ry receptor (Meissner, 1994; fig. 2A). The results obtained indicate that Cf not only elevated [Ca⁺⁺]_i and mitochondrial Ca⁺⁺ uptake (fig. 2A) but also further enhanced mitochondrial accumulation of calcium ions promoted by treatment with the hydroperoxide (fig. 2B), as well as the formation of DNA SSBs (fig. 2C). These effects of Cf appeared to be specific and were due to the ability of the drug to release Ca⁺⁺ via opening of the Ry receptor, since they were prevented by high doses of Ry (fig. 2A-C), a treatment known to block the opening of the Ry receptor (Meissner, 1994).

Taken together, these results are consistent with the possibility that mitochondrial calcium overload is the sole mechanism whereby Cf enhances DNA strand scission caused by tB-OOH. The experiments using permeabilized cells provide a straightforward demonstration causally linking these two phenomena. Indeed, using digitonin-permeabilized cells we were able to directly deliver Ca⁺⁺ in the cytosol and show that the cation enhances DNA cleavage caused by tB-OOH in a concentration-dependent fashion (fig. 4B). The fact that Ca⁺⁺ did not modify the DNA-damaging response evoked by H₂O₂ (fig. 4B) and that its potentiating effects on DNA damage induced by tB-OOH were prevented by as low as 200 nM RR -but not by Ry- (fig. 4C) demonstrates that mitochondrial calcium uptake specifically enhances the formation of DNA-damaging species generated by tB-OOH. The permeabilized cell system was also responsive to Cf that once again potentiated DNA damage caused by tB-OOH (fig. 4A) in a RR- as well as Ry-inhibitable fashion (fig. 4C). Interestingly, addition of exogenous calcium ions did not. however, further increase the accumulation of DNA lesions in cells exposed to tB-OOH and Cf (fig. 4B), strongly suggesting that Ca⁺⁺ and Cf act via common mechanisms, i.e., by promoting mitochondrial calcium overload. It is important to emphasize that these experiments also rule out the possibility that the compounds used to modulate the intramitochondrial calcium content, and the extent of the tB-OOH-induced DNA damage, produced changes in the uptake or intracellular distribution of the hydroperoxide that were coincidental with, but not causally related to mitochondrial calcium. Along the same lines, we can exclude that an enhancement in [Ca⁺⁺]_i per se generates DNA strand scission, making it unlikely that Ca⁺⁺-dependent endonucleases mediate the formation of DNA lesions in cells exposed to tB-OOH. The results thus far discussed clearly indicate that the site in which at least some of the species mediating DNA damage induced by tB-OOH are being formed is the mitochondrion and that mitochondrial Ca⁺⁺ uptake enhances their formation. It was therefore important to investigate the nature of these species that, on the basis of previous results from our (Guidarelli et al., 1996) and other (Coleman et al., 1989) laboratories, appear to be different from those involved in the cytotoxic response. In this study we report experimental evidence indicating that H₂O₂ is one of the species that mediate DNA cleavage induced by tB-OOH. It is conceivable that Ca⁺⁺ is involved at the level of formation of H₂O₂ because mitochondrial production of superoxide anions and hydrogen peroxide was shown to be sensitive to mitochondrial Ca⁺⁺ content (Cadenas and Boveris, 1980). Furthermore, reports from the Vercesi group (Valle et al., 1993; Castilho et al., 1995) indicate that mitochondrial calcium accumulation leads to an enhanced formation of H₂O₂ in mitochondria exposed to tB-OOH. Therefore, it is tempting to speculate that at least part of the DNA strand scission caused by tB-OOH occurs as a result of H₂O₂ formation taking place at the mitochondrial level and that mitochondrial Ca⁺⁺ uptake leads to an increased formation of H₂O₂ and thus to an enhanced DNA strand scission. Our previous results, demonstrating an increased formation of tB-OOH- induced DNA lesions in catalase-depleted cells (Guidarelli et al., 1997), are consistent with this possibility that finds definitive proof in the outcome of the experiments obtained in our study using the permeabilized cell system. Indeed, catalase was found to reduce DNA cleavage caused by tB-OOH and afforded a greater protective effect after treatment with the peroxide in the presence of exogenous calcium ions (fig. 5). However, it is important to note that catalase, although abolishing DNA strand scission caused by H₂O₂ or by the redox cycling quinone menadione (fig. 5, inset), afforded only a partial protection against DNA damage caused by tB-OOH alone or associated with CaCl₂. The fact that SOD neither affected DNA damage induced by tB-OOH nor modulated the protective effects of catalase (fig. 5) rules out the possibility that superoxides migrate into the nucleus before being converted into H₂O₂. This sequence of events, although unlikely (superoxides readily dismutate either spontaneously or enzymatically), would have explained why catalase failed to completely abolish the formation of DNA lesions under conditions in which the only species responsible for this effect was H₂O₂.

Taken together these results, although demonstrating that H₂O₂ is the most relevant DNA-damaging species that is produced within the mitochondria via a Ca⁺⁺-dependent mechanism, would imply the formation of tB-OOH-derived DNA-damaging species different from H₂O₂.

Additional results reported in this study, however, provide circumstantial evidence suggesting remarkable similarities in the types of DNA lesions generated after treatments with tB-OOH alone or associated with Cf (a condition that magnifies the relative amount of the DNA lesions generated via the calcium-based mechanism), or with the peroxide associated with RR (a condition abolishing the formation of DNA lesions generated via the calcium-based mechanism), thus suggesting that the species responsible for their formation were characterized by similar reactivities. Indeed, the formation of DNA lesions was always iron dependent and insensitive to antioxidants (table 1) and their repair was characterized by superimposable kinetics (fig. 3). Even more interesting was the observation that these kinetics were virtually identical to those detected after treatment with H₂O₂ (fig. 3), which also generates DNA strand scission inhibitable by iron chelators (Mello Filho et al., 1984; Mello Filho and Meneghini 1984; Guidarelli et al., 1995; Coleman et al., 1989; Latour et al., 1995; Guidarelli et al., 1997) and insensitive to antioxidants (Coleman et al., 1989; Guidarelli et al., 1997). Thus, it may be hypothesized that the H₂O₂-independent component of DNA strand scission caused by tB-OOH is mediated by different species with reactivities similar to that of the hydroxyl radical, the final DNA-damaging product resulting from the interaction between H₂O₂ and divalent iron. The nature of these species is not readily apparent from the results presented in this study. Ferryl and per-ferryl radicals as well as complexes of iron and oxygen are possible candidates, because they are all sensitive to iron chelators and their reactivities are remarkably similar to that of the hydroxyl radical. The identification of these species, however, does not appear to be an easy task and their involvement in specific reactions can only be suggested on the basis of indirect experimental evidence. In a recent review on the Fenton reaction Goldstein et al. (1993) well summarized the problems that can arise in these types of studies by stating that "identifying oxidizing intermediates in mammalian cells is almost impossible."

	Conclusions

Top Abstract Introduction Materials & Methods Results Discussion Conclusions References

Our results demonstrate that a sub-toxic concentration of tB-OOH elevates [Ca⁺⁺]_i, a large proportion of which is promptly cleared by the mitochondria. Intramitochondrial Ca⁺⁺ promotes the formation of tB-OOH-derived DNA-damaging species mainly represented by H₂O₂. The contribution of this mechanism to the overall DNA-damaging response could be remarkably enhanced by Cf. Thus, it may be hypothesized that the DNA-damaging efficiency of organic hydroperoxides is potentially modulated by agents (hormones, drugs, toxins, etc.) which elevate [Ca⁺⁺]_i, provided that this latter event is associated with mitochondrial clearance of the cation. Finally, tB-OOH also generates DNA-damaging species other than H₂O₂ resulting in DNA lesions remarkably similar to those generated by the hydroxyl radical.

This study identifies the mechanism whereby tB-OOH induces DNA single strand breakage in cultured mammalian cells and, more generally, provides new insights into the mechanism of oxidative stress associated with organic hydroperoxides. Future research should better define the generality as well as the specific aspects of these effects. It will be important to determine the role of the Ca⁺⁺-dependent mitochondrial formation of tB-OOH-derived DNA-damaging species on the tumor promoting properties of the hydroperoxide. These events may also activate specific redox-sensitive signal transduction pathways. Finally, it will be important to investigate the effects at the level of mitochondrial DNA. Indeed, it is reasonable to expect that a large amount of lesions will accumulate in the DNA of those mitochondria in which tB-OOH-derived DNA-damaging species are being formed.