THE EFFECT OF OXIDATIVE STRESS ON LEVELS OF CYTOSOLIC CALCIUM WITHIN AND UPTAKE OF CALCIUM BY SYNAPTOSOMES

-The ability of synaptosomes subjected to oxidative stress, to maintain homeostasis has been evaluated using various indices of cellular integrity. These include levels of cytosolic calcium and leakiness of the plasma membrane. The status of a neural characteristic; depolarization-induced calcium entry into the cytoplasm, has also been studied. The presence of 5 µM FeS0 4 and 0.1 mM ascorbic acid increased peroxidative activity as judged by the rate of thiobarbituric acid reactive material production, and depressed levels of free ionic calcium [Ca 2 +j, as determined using the calcium-sensitive fiouorescent indicator dye fura-2. Depolarization-induced influx of 45 Ca2+ was greatly depressed under these conditions, while basal calcium uptake was inhibited to a much lesser degree. The efflux offura-2 from synaptosomes was enhanced in the oxidizing environment, suggesting increased permeability of the synaptosomal outer limiting mem brane. The treatment of synaptosomes with 25 µM a-tocopherol succinate before and during exposure to the Fe 2 + /ascorbate mixture prevented many of the changes otherwise induced by the oxidizing system. Similar pretreatment with ti-carotene or superoxide dismutase did not have any protective effect. Ganglioside GM 1 pre-exposure did not alter the Fe 2 + /ascorbate-induced changes in calcium-related parameters, but mitigated synaptosomal plasma membrane damage as judged by fura-2 leakage. Thus exogenous agents may be capable of reducing the severity of oxidative stress in nervous tissue.

Free radical formation can occur as a result of oxidative events leading to the formation of various active oxygen species. The central nervous system may be a tissue especially vulnerable to oxidative stress for several reasons : (1) The high and continuous metabolic demand of this tissue is concomitant with a very high rate of oxidative phosphorylation, and aerobic mitochondrial activity is unusually large. Transient interruption of the energy and oxygen supply to the brain can lead to subsequent reperfusion injury, of which a significant component is incomplete substrate combustion and consequent formation of reactive free radicals (Rehncrona et al., 1980;Cao et al., 1988). Similarly, an increased formation of lipid peroxide has been found in brain tissue following temporary interruption of oxidative phosphorylation by cyanide (Johnson et al., 1987).
*Author to whom all correspondence and reprint requests should be addressed.

615
(2) The presence of myelin provides an environment relatively rich in polyunsaturated lipids (Sun and Sun, 1972). These can be a site of sequestration of liphophilic radicals such as the hydrated superoxide radical H0 2 - (Halliwell and Gutteridge, 1986), and can also serve as a substrate for formation oflipid peroxides, leading to a chain reaction production and propagation of free radicals. (3) The rat brain has a relatively low content of enzymes relating to protection against oxidative damage, such as glutathione peroxidase, superoxide dismutase and catalase (Carmagnol et al., 1983;Marklund et al., 1982;Sinet et al., 1980), and these enzymes are predominantly in glia rather than neurons (Savolainen, 1978). (4) Certain catecholamine neurotransmitters such as norepinephrine are susceptible to autooxidation and consequent free radical formation (Cohen and Heikkila, 1974;Adam-Vizin and Seregi, 1982). Metals such as manganese and iron which can catalyze oxidative events are presented in especially high amounts in some brain regions (Halliwell and Gutteridge, 1986). It has been pro-posed that decompartmentalization of metal ions may be of importance in trauma-related brain injury seizure responses (Willmore et al., 1983).
We have previously found that methyl mercury, a neurotoxic agent known to enhance lipid peroxidation, can elevate levels of cytosolic calcium in synaptosomes. This elevation coincided with an increased permeability of the synaptosomal plasma membrane. Both of these indices of synaptosomal damage were reversed by the antioxidant ct-tocopherol (Bondy and McKee, 1990). This led us to consider whether the formation of free radicals might underlie the elevation of cytosolic calcium and leakiness of synaptosomes exposed to oxidative stress.
A role for calcium in many pathological states related to free radical formation has been postulated (Farber, 1982). Abnormal levels of intracellular calcium consequent to injury to the cell membrane or energy-dependent Ca-regulatory homeostatic mechanisms may underlie cell death (Siesjo, 1986). There is evidence for an interactive relation between oxidizing conditions and levels of intracellular calcium (Maridonneau-Parini et al., 1986). Thiol groups may play a role in Ca-sequestration (Thor et al., 1985). Excess oxidative activity can impair mitochondrial function and thus prevent effective sequestration of Ca 2 + or its pumping out of the cell (Bellomo et aL 1982). Conversely, calcium may potentiate free radical injury to mitochondria (Malis and Bonventre, 1986). Such interactions have not always been found to be synergistic. Omission of extracellular Cac+ can elevate oxidative damage in isolated hepatocytes (Thomas and Reed, 1988). This study was conducted in order to examine the effect of free radical-forming conditions upon both the entry of calcium into synaptosomes and levels of free ionic calcium within these structures. The effect of agents known to retard oxidizing conditions upon these parameters was also evaluated.
[Ca'+], was measured using the indicator dye fura-2 AM; I-ml aliquots of the synaptosome suspension were incubated with 5 µM fura-2 AM added in DMSO (final concentration of DMSO, 0.5%) at 3T'C for 10 min. HEPES buffer (9 ml) at 3T'C was then added and the incubation continued for 25 min. Synaptosomes were sedimented (12,300 g, 8 min), resuspended in 5 ml of HEPES buffer and kept on ice. For fluorescence measurement, 500-ml aliquots of the suspensions were centrifuged rapidly in Eppendorf-tubes in a microcentrifuge (16,000 g, 2 min) and the synaptosomal pellet was resuspended in 2 ml of warm (37°C) HEPES buffer. This buffer did not contain NaHCO, or NaH 2 P0 4 in order to prevent precipitation of high lev.els of Ca2+ used subsequently.
The emitted fluorescence of the hydrolyzed fura-2 ester in the sample was measured in a thermostatted (37°C) cuvette using a spectrofluorometer at the excitation wavelengths of 340 and 380 nm (bandpass 3 nm) and at 510 nm for emission (bandpass 20 nm). Samples contained 140-160 µg of protein and were allowed to equilibrate for 10 min before addition of chemicals. Mixing was carried out with a magnetic stirrer 30--60 s before reading fluorescence. Extrasynaptosomal fura-2 was quenched by 4.0 pM MnCl, at the end of each incubation. It was added 15-20 s before recording the fluorescence. This Mn2+ concentration corrected for extrasynaptosomal fura-2 (Jacob et al., 1987;Komulainen and Bondy, 1987a). This method of determination of intrasynaptosomal fura-2 gave the same results as the more direct procedure of fluorescent assay of the supernatant after centrifuging synaptosomes down from the incubation medium (Bondy and Halsall, 1988b). However, this latter procedure has the disadvantage that no subsequent determination of the synaptosomal signal under conditions of zero or saturating Ca2+ is possible. The proportion of fura-2 that was found to be outside the synaptosomes is described in the results as"% leakage." In incubated but untreated synaptosomes basal leakage was 13-17%. In order to calculate [Ca 2 +], before addition of agents, a separate average Mn 2 + correction was made for each batch of synaptosomes. Autofluorescence of synaptosomes was always recorded and subtracted before calculation of fluorescence ratios of fura-2. This constituted around 15% of total fluorescence. For calibration of the synaptosomal fura-2-Ca 2 + signal, (R), Rmin (the ratio of fluorescence at 340/380 nm in the absence ofCa2+) and Rmax (the ratio when all fura-2 of the sample was saturated with Ca2+) were determined for each batch of fura-2 loaded synaptosomes. In order to determine Rmim synaptosomes were lysed with 0.1 % sodium dodecyl sulfate, and Ca2+ and Mn 2 + chelated with 5 mM alkaline EGT A. Rm., was determined by the addition of 9 mM CaCl 2 • [Cai+], was calculated using the formula (Grynkiewicz et al., 1985): where Kd (224 nM) is the dissociation constant of fura-2-Ca2+ complex and Sf 2 and Sb 2 denote fluorescence of fura-2 at zero Ca2+ and Ca2+ saturation, respectively, at the excitation wavelength 380 nm.

45
Ca 2 + transport A synaptosomal suspension (0.1 ml containing 150-170 µg of protein) was mixed with 0.9 ml of calcium-free HEPES buffer and incubated 30 min at 37°C. Then 0.1 ml of 45 Ca2+ (0.5 µCi, 55.9 Ci/mo!) in either 1.0 M KC! or 1.0 M NaCl was mixed into the suspension and incubation continued for an additional 20 s. While uptake was not completely linear over this period, the use of shorter incubation times adversely affected reproducibility. When indicated, various chemicals were added during the preincubation. oc-Tocopherol succinate and /)-carotene were added in DMSO (the final concentration not exceeding I%). 45 Ca 2 + uptake was stopped by the addition of 5 ml of ice-cold wash buffer (HEPES-buffer omitting CaCl 2 and containing I mM EGT A) followed by rapid filtration on glass fiber filters (Type A/E, Gelman Sciences Inc., Ann Arbor, Mich.) (Wu et al., 1982). The filter disks were washed with 2 x 5 ml of buffer and accumulated label counted in 5 ml of Ecoscint (National Diagnostics, Somerville, N.J.) in a liquid scintillation counter. The difference in 45 Ca uptake between incubation in the presence of added 90 mM KC! or 90 mM NaCl was taken to represent depolarization-triggered Ca2+ entry. This was between 2 and 3 times the basal uptake in the presence of NaCl. Blank tubes contained synaptosomes that had been sonicated and heated to J00°C for 10 min and were thus incapable ofCa 2 + transport.

Lipid peroxidation
The formation of thiobarbiturate-reactive material (TBAR) was used as an index of peroxidative activity. The precise method used was based on the modification of Uchiyama and Mihara (1978) using I% phosphoric acid as the protein denaturing agent. Color formation at 535 nm was determined in the supernatant with no organic solvent extraction. The molar extinction coefficient of a malondialdehyde standard (1.56 x 10 5 ) confirmed the calibration of the spectrophotometer. At the low tissue concentration used, color formation was found to be proportional to the amount of tissue present and was linear for at least 2 h at 37°C (Lambert and Bondy, 1989). A 30 min incubation was necessary to allow sufficient color development for accurate assay.

Protein determination
Protein content of synaptosomal suspensions was assayed with the method of Bradford (1976) using bovine serum albumin as a reference.

Statistics
Results were analyzed using Fisher's Least Significant Difference Test after one-way analysis of variance. Throughout results, the symbol * means P < 0.05, using a two-tailed t distribution, that the value differs from the corresponding control value. Each data point presented is the mean of 4-9 individual determinations.

Materials
Chemicals were obtained from Sigma Chemical Corporation (St Louis, Mo.), and 45 CaCl 2 from New England Nuclear (Bedford, Mass.).

Effects of Fe/ascorbate on synaptosomal peroxidation, calcium levels and uptake
Exposure of synaptosomes to the Fe/ascorbate oxidizing system elevated the formation of thiobarbituric acid reactive products (TBAR) in a timedependent manner for at least 2 h (Lambert and Bondy, 1989). Increasing ascorbate levels effected a dose-related but non-linear increase in TBAR formation (Fig. 1).
Synaptosomal ionic free calcium content [Ca 2 +]; was depressed in the presence of 5 µM Fe 2 + and 0.1 mM ascorbate (Fig. 2).1 We previously established that the concentration of Fe 2 + used had no effect on the intensity of fura-2, although higher Fez+ concentrations could quench this fluorescence. Leakage of the calcium-responsive dye was elevated in the presence of oxidizing conditions, suggesting damage to the synaptosomal limiting membrane (Fig. 2). The use of 0.1 mM ascorbate alone did not alter [Ca 2 +]; or fura-2 leakage.
Since ascorbate potentiated the effect of Fe 2 +, the standard oxidizing conditions used in subsequent The presence of 10 µM desferrioxamine in the incubation mixture completely blocked all Fe/ascorbateinduced TBAR formation (Table 1 ). This concentration ofiron-chelator also totally prevented any Fe/ascorbate-induced changes in synaptosomal [Ca 2 +]; and fura-2 leakage (data not shown). This suggested that the observed synaptosomal changes effected by Fe/ascorbate were related to oxidative damage rather than other metabolic changes.
Further evidence for altered function of the synaptosomal plasma membrane came from the study of 45 Ca uptake by synaptosomes (Table 2). This was reduced in the presence of 5 µM Fe 2 + and various   concentrations of ascorbate. Basal uptake was inhibited to a much lesser extent than K + -depolarization induced uptake.

Modulation ol oxidative damage hy antioxidants and other chemicals
Synaptosomes were pretreated with a variety of substances potentially capable of preventing oxidative stress or of stabilization of cell membranes before being exposed to a standard Fe/ascorbate mixture (5 11M FeS0.-100 µM ascorbic acid). Fe/ascorbate (Table 1). This concentration of cx-tocopherol completely prevented the reduction of 45 Ca uptake and of[Ca 2 +]i caused by Fe/ascorbate. cx-Tocopherol alone had no significant effect on [Ca 2 +]i or fura leakage (Fig. 2). However, both basal and Kstimulated synaptosomal 45 Ca uptake were somewhat elevated (Fig. 3). The reversal of changes in Ca parameters effected by cx-tocopherol was paralleled by prevention of Fe/ascorbate-induced elevations of fura-2 leakage into the extra-synaptosomal compartment (Fig. 2). This implied that plasma membrane function could be preserved by cx-tocopherol.
( b) {3-Carotene. Pretreatment of synaptosomes with 25 µM {J-carotene for 15 min did not prevent Fe/ ascorbic acid from exerting its usual effect on TBAR formation or on synaptosomal [Ca 2 +]i (Fig. 4, Table  1 ). {J-Carotene also did not protect the synaptosomal limiting membrane against the effects of subsequent addition of Fe/ascorbate as judged by fura-2 leakage. Furthermore, the depression in depolarizationinduced 45 Ca uptake caused by Fe/ascorbate was not mitigated by {J-carotene (data not shown).
(c) Superoxide dismutase. In view of the ability of superoxide dismutase to degrade the superoxide radical, 0 2 , a parallel pretreatment to those described above was carried out, by exposing synaptosomes to 300 units of superoxide dismutase/ml, for 15 min. No protective effect was found on TBAR formation (Table 1), [Ca 2 +1 and synaptosomal leakiness changes effected by Fe/ascorbic acid (Fig. 5), or 45 Ca uptake (data not shown).
( d) Ganglioside GM 1 • Preliminary treatment of synaptosomes with JOO µM GM 1 for 15 min at 37°C did not significantly alter their subsequent response in 45 Ca uptake or depression of[Ca2+]i to the standard oxidizing conditions employed here (Fig. 6, data not shown). However, the permeability of synaptosomes  * P < 0.05 that value differs significantly from control value.
to fura-2 was unaltered by subsequent addition of Fe/ascorbate. The effect of oxidizing conditions upon 45 Ca uptake and [Ca 2 +1 was, in this case, separable from effects upon membrane integrity. Since TBAR production was not significantly altered by GM 1 (Table 1), the protective effect of this ganglioside on Fe/ascorbate-induced changes cannot be attributed to a direct antioxidant effect.

DISCUSSION
While cx-tocopherol was able to completely prevent the effects of a subsequent addition of the Fe/ ascorbate mixture upon 45 Ca uptake, [Ca 2 +ji and leakiness of synaptosomes, malondialdehyde production was only partially inhibited (Table 2). Several reports indicate that cx-tocopherol administration may protect against various types of damage to nervous tissue including FeCl 2 -induced focal edema (Willmore and Rubin, 1984), penicillin-induced focal epilepsy (Nikushkin et al., 1981) and hippocampal slices rendered anoxic by superfusion with oxygen free media (Acosta et al., 1987). However, this protection is not always closely related to the extent of inhibition of lipid peroxidation. A limited inhibition of malondialdehyde formation afforded by cx-tocopherol, toeither to the relative impermeability of synaptosomes function, has been described for hippocampal slices (Acosta et al., 1987).
The reduction of [Ca 2 +]i by oxidizing conditions is in contrast to the effect of various neurotoxic, agents including several organometals and organochlorine insecticides which elevated synaptosomal [Ca 2 +1 (Komulainen and Bondy, 1987a,b; Halsall, 1988b;Joy and Burns, 1988). The increased leakiness of synaptosomes in the presence of oxidizing mixtures is, however, clear evidence of damage. The protection offered by the simultaneous presence of 1X-tocopherol against observed changes also suggests that oxidative activity was the cause of these changes. The data obtained using 45 Ca substantiate this concept. Depolarization-induced 45 Ca entry is known to be depressed in synaptosomes that have been exposed to a variety of neurotoxic agents such as chlordecone and lindane (Komulainen and Bondy ,l 987b;Bondy and Halsall,l 988b ). Basal 45 Ca uptake was inhibited to a much lesser extent than K +-induced uptake by oxidizing conditions. This suggested that reduction of uptake was due to specific effects on a dynamic physiological process rather than interference with overall membrane permeability. While we have found K-stimulated Ca uptake to be reduced by peroxidative conditions, basal 45 Ca uptake by synaptosomes has been shown to be elevated in the presence of200 µM Fe 2 + (Braughler, 1988). The inhibition of depolarized-induced Ca uptake implies a more selective effect of the relatively moderate peroxidizing conditions used here (also indicated by a lower rate of formation of TBAR) upon voltage-sensitive channels. While these data form a consistent pattern, the observed depression of [Ca 2 +]i was unexpected. Since this reduction was not due to any interference with fura fluorescence caused by the Fe/ ascorbate mixture, the effect appeared related to oxidative events in that it was reversed by IX-tocopherol. It may be that some of the apparent prevention of changes induced by oxidizing conditions described here are due to modification of membrane properties unrelated to antioxidant activity. This is clear in the case of ganglioside GM 1 which did not alter stimulated TBAR production (Table 2), and may also partly account for the protective effects of IX-tocopherol. However, TBAR production does not give a total picture of oxidative activity and thus incomplete inhibition of TBAR formation by 1X-tocopherol may not completely reflect its antioxidative properties. Furthermore, the protective effect of IX-tocopherol resembled that produced by desferrioxamine whose mode of action is clearly antioxidative.
The failure of superoxide dismutase to protect against synaptosomal damage (as judged by increased fura-2 leakage) or alter Ca parameters may be due either to the relative impermeability of synaptosomes to this protein or to the possibility that o~ is not the directly harmful active species. Since superoxide dismutase can enter intact neurons (Saez et al., 1987), the latter explanation is more likely. Only in the case of ganglioside GM 1 was there dissonance between results derived from fura-2 leakage and those pertaining to calcium levels and translocations. In this case, the depression of[Ca 2 +]i and 45 Ca uptake effected by Fe/ascorbate was not blocked by GM 1 pretreatment while the elevated fura-2 leakage was protected against. Further separation of leakage and [Ca2+]i-related indices of damage is evidenced by the fact that several neurotoxic agents increase both these parameters (Komulainen and Bondy, l 987b;Bondy and Halsall, 1988b), but Fe/ascorbate-induced oxidative stress increases leakage but decreases [Ca 2 +];.
The relation between calcium and peroxidative activity is unclear, and seemingly conflicting results have been reported. High calcium levels in the incubation medium have been found to enhance basal and induced peroxidative damage in synaptosomes (Braughler et al., 1985). Further evidence for a positive interaction between calcium and oxidative damage comes from the demonstration that hypoxia can both increase rates of peroxidation and elevate [Ca 2 +1 in nerve tissue (Johnson et al., 1986(Johnson et al., , 1987White and Clark, 1988). However, hypoxia has also been reported to reduce both the total calcium content and 45 Ca uptake of synaptosomes (Peterson and Gibson, 1984). In support of an antagonism rather than synergy between peroxidative events and calcium levels is the finding that low levels of extracellular calcium are associated with increased lipid peroxidation in isolated hepatic cells, as well as increased leakiness of these cells (Smith and Sandy, 1985;Thomas and Reed, 1988). A mechanism that could account for our [Ca 2 +1 and 45 Ca-derived data is the reduction of calcium influx by inhibition of voltage-sensitive and receptor-operated channels. Such an inhibition has been described for ethanol (Dildy and Leslie, 1989). The relation between intracellular calcium and the generation of free radicals is not close, since each parameter is subject to relatively independent modulation. While both calcium influx and oxidative stress can be damaging to mitochondria, these processes are clearly dissociable (Stark et al., 1986). Intracellular [Ca2+]i represents an equilibrium of the sum of many processes. Calcium influx into the cytoplasm can occur (i) through specific channels or through non-specific sites in the plasma membrane; (ii) by enzymically induced release from endoplasmic reticulum, and (iii) by the mitochondrial Ca/proton exchanger. Balancing these are the energy dependent pumps and sequestration mechanisms present in many membranes and the Ca/Na exchangers of the plasma membrane. Some of these mechanisms are likely to be more sensitive to oxidative damage than others. A critical aspect is the extent to which cellular energy-generating mechanisms which underlie all anabolic processes are compromised by excess formation of free radicals. However, even when synaptosomal ATP levels are at 50% of their normal level, [Ca2+]. is not significantly elevated (Bondy and Komulainen, 1988). Calcium-removing processes may thus have a priority for available energy when ATP levels are depressed. Pro-oxidative conditions are known to increase the rate of calcium cycling across the mitochondrial outer membrane (Richter and Frei, 1988). When oxidative damage is not too severe, the final equilibrium of such increased flux may result in reduced [Ca2+Ji-The increased cell permeability and partial collapse of mitochondrial potential provoked by oxidative stress (Lambert and Bondy, 1989) may result in an over-response of calcium homeostatic processes leading to a reduction of [Ca2+],.