SENSITIVE AND RAPID QUANTITATION OF OXYGEN REACTIVE SPECIES FORMATION IN RAT SYNAPTOSOMES

-The formation of oxygen reactive species in response to oxidative stimuli was measured in rat synaptosomes. Studies employed the non-fluorescent probe 2', 7' -dichlorofluorescin diacetate (DCFH-DA), which after de-esterification is oxidized in the presence of oxygen reactive species to the highly fluorescent 2',7'-dichlorofluorescein (DCF). Oxygen reactive species formation, as measured by DCF fluorescence, was stimulated by ascorbate and/or FeS0 4, and xanthine/xanthine oxidase under various buffering con ditions. These agents all increased DCF formation in Tris, HEPES and phosphate buffer. Ascorbate also stimulated the formation of DCF in a concentration-dependent manner. The presence of Ca 2 + in HEPES buffer did not enhance or diminish the effects of ascorbate/FeS0 4 on DCF formation. Deferoxamine inhibited the ascorbate/FeS0 4 -induced stimulation of DCF formation, but xanthine/xanthine oxidase induced stimulation was not affected by pretreatment with superoxide dismutase. Results indicate that DCF fluorescence is a sensitive, quantitative and direct measure of oxygen reactive species formation in synaptosomes, providing a rapid method for investigating early neuronal events that occur during oxidative stress.

Numerous studies have documented that formation that argue against the use of thiobarbiturate/ of free radicals plays a major role in events leading to malondialdehyde-like indicators of free-radical biological membrane damage (for reviews see generation report that oxygen radical-induced events Kappus, 1985;Halliwell and Gutteridge, 1984). In such as proteolysis precede, and are independent the central nervous system, much has been written of, lipid peroxidation (Davies and Goldberg, 1987a, b). regarding oxygen radical formation as an early post-Finally, the instability of the thiobarbiturate/ traumatic event following spinal cord injury (Demo-malondialdehyde complex in the presence of H 2 0 2 , a poulos Milvy et al., 1973), and in the highly reactive oxygen species, was recently demondestruction of catecholamine neurons (Cohen, 1984). strated (Kostka and Kwan, 1989). Oxidative stress is reported to be magnified by elev-A direct measurement of oxygen reactive species ations in intracellular Ca 2 + (Braughler et al., 1985), has been reported in cell culture systems using the modulate y-aminobutyric acid (GABA)/barbiturate non-fluorescent probe DCFH-DA. Relative intrareceptor function (Schwartz et al., 1988), affect mem-cellular oxidation has been quantitated via DCF fluobrane potentials (Lambert and Bondy, 1989), and rescence in polymorphonuclear leukocytes (Szejda et other electric membrane properties (Scott and Lew, al., 1984;Bass et al., 1983) and renal epithelial cells 1988). Using electron spin resonance techniques, brain  in flow cytometric studies. DCFHmitochondrial superoxide radical formation has been DA is a stable, non-fluorescent molecule that readily shown to be age-related (Sawada and Carlson, 1987). crosses cell membranes, and is hydrolyzed by intra-Traditionally, free-radical initiated lipid peroxida-cellular esterases to non-fluorescent 2',7'-dichlorotion has been estimated with techniques such as the fluorescin (DCFH) (Bass et al., 1983). DCFH is then reaction between thiobarbituric acid and malon-rapidly oxidized in the presence of oxygen reactive dialdehyde, and by conjugated dienes. However, species to highly fluorescent DCF (Bass et al., 1983; thiobarbituric acid and malondialdehyde have been Szejda et al., 1984;). shown to cross react with several endogenous sub-Since all existing studies investigating the formation stances (deoxyribose, amino acids) and functional of oxygen reactive species have used cell culture moieties such as amino groups (Gutteridge, 1981;systems,ourinterestsweretoadapttheuseofDCFH-Halliwell and Gutteridge, 1981 ). Additionally, studies DA to subcellular systems, widely used by biochemists CARL P. LEBEL and STEPHEN c. BONUY and neurochemists. Furthermore, DCFH-DA may prove useful as a marker of early neurological insult. A recent study demonstrated that dietary vitamin E deficiency resulted in lower basal formation rates of oxygen reactive species in cerebrocortical P2 fractions using DCFH-DA (LeBel et al., 1989).
In order to expand the potential utility of DCFH-DA, it is necessary to consider some possible limitations: (I) Does DCFH-DA truly reflect intracellular (intrasynaptosomal) oxidative events? (2) Can it be accurately quantitated? (3) What oxidative species are assayed? (4) How sensitive is the analysis? (5) Is DCF fluorescence prone to artifact? This study investigates the application of the DCFH-DA probe to measure oxygen reactive species in rat synaptosomes.

Animals
Adult male CR I CD rats (250-300 g) were obtained from Charles River Breeding Laboratories (Wilmington, Mass.), and were maintained in the animal facility in clear polypropylene cages with water and food provided ad lihitum.

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

Assay.fi;r oxyqen reactii-c species.formation
Synaptosomes (0.5 ml) were diluted in 9 vols of either HEPES, HEPES minus CaCl 2 , 40 mM Tris or 0.1 M NaH 2 P0 4 , all at pH 7.4. The diluted synaptosomes (5.0 ml) were then incubated with 5 µM DCFH-DA (added from a stock solution of 1.25 mM in methanol) at 37C for 15 min. To terminate the incubation, the synaptosomes were centrifuged at 12,500 K for 8 min (0-4T), and the pellet was resuspended in 5 ml of the respective ice-cold buffer. Fluorescence was monitored on a Farrand Spectrofluorometer, with excitation wavelength at 488 nm (band width 5 nm), and emission wavelength 525 nm (band width 20 nm). The cuvette holder was thermostatically maintained at37T.
Autofluorescence of synaptosomes was corrected for by the inclusion in each experiment of parallel blanks (unloaded synaptosomes). The correction for autofluorescence was always less than 11 % of the total. Oxygen reactive species formation was quantitated from a DCF standard curve in methanol (0.05-1.0 µM).
DCFH was prepared from DCFH-DA by the method of Cathcart et al. (1984) by mixing 0.5 ml 1.0 mM DCFH-DA in methanol, with 2.0 ml 0.01 N NaOH. This de-esterification of DCFH-DA proceeded at room temperature for 30 min, then was neutralized with 10 ml 25 mM NaH 2 P0 4 , pH 7.4. This solution was kept on ice in a foil wrapped container until analysis. To determine whether the incubation media alone oxidized DCFH, 50 1d of DCFH was placed in 1.95 ml Tris, the fluorescence was recorded, the solution was incubated at 37'T for 60 min and the final fluorescence was determined.

(a) Dependence ol probe fluorescence on hiological tissue
DCFH-DA was placed in Tris buffer in the absence or synaptosomes to determine whether intrasynaptosomal esterases were required for the eventual oxidation or DCFH-DA to DCF. No fluorescence was observed either prior to or following the addition of ascorbate/FeS0 4 in this aqueous medium. These studies were based on the assumption that DCFH oxidation is solely due to biological processes and does not occur significantly in the presence of the incubation medium. In order to verify this assumption, DCFH was prepared from DCFH-DA (sec Experimental Procedures section) and its proneness to oxidative media alone was studied. The rate of intrinsic oxidation ofDCFH was invariably below 3°/., of the corresponding rate in the presence of biological tissue (Table 1). This was the case even in the presence of potentially oxidizing environments containing 0.1 mM ascorbate/5 µM FeS0 4 • To determine the degree ofleakage by DCFH from synaptosomes, DCFH-DA loaded samples were incubated for 60 min and the fluorescence recorded. The samples were recentrifuged for 8 min at 12,500 g, resuspended in the same volume of Tris, and the fluorescence was recorded. Synaptosomes contained from 75 to 82% of the original DCF formed, or an approximate 18-25% leakage of DCFH formed intrasynaptosomally.

(b) Synaptosomal production of oxygen radicals
Unstimulated synaptosomes pre-loaded for 15 min with DCFH-DA in several buffers showed detectable basal levels of DCF fluorescence after a 60 min incubation (Table 2). Basal levels of DCF formation in phosphate and Tris buffers were somewhat higher than in both HEPES buffers. In all of the buffers employed, ascorbate was more efficacious in stimulating DCF formation than was FeS0 4 , while FeS0 4 alone was unable to induce DCF formation in phosphate and HEPES-Ca 2 + buffers.
In all buffers, the combination of ascorbate and FeS0 4 potentiated the formation of DCF, as com- The data are expressed as the mean DCF formed in pmol/2 ml/ min cuvette and were obtained from two independent experiments with differences no larger than 7% between assays.
pared to incubations with ascorbate and FeS0 4 alone ( Table 2). Ascorbate/FeS0 4 -stimulated DCF formation was higher in phosphate and Tris buffers than in Ca 2+ -containing or Ca 2 + -absent HEPES buffers. The presence of Ca 2 + in HEPES did not enhance or diminish the effects of ascorbate/FeS0 4 on DCF formation. Synaptosomes were pre-loaded with DCFH-DA for various times prior to ascorbate/FeS0 4 stimulation. Maximal DCF formation was reached after pre-loading for 15 min (Fig. 1).
The rate of DCF formation was linear (r = 0.986) with respect to the amount of protein employed (Fig.  2). Ascorbate/FeS0 4 -induced DCF formation could be detected using as little as 40 µg synaptosomal protein.
The generation of DCF in synaptosomes, when stimulated by ascorbate, was also concentrationdependent (Fig. 3). This ascorbate/FeS0 4 -stimulated DCF formation was completely blocked by deferoxamine (Fig. 4). While xanthine/xanthine oxidase  17 1338 Synaptosomes were exposed to various oxidizing agents in several buffers at 37°C for 60 min. The data were obtained from two independent experiments and are expressed as the means with differences no larger than 11 % between assays. Protein (mg/ml) Fig. 2. The relation of protein concentration to ascorbate/ FeS0 4 -induced DCF formation. The incubation time was 60 min. Data were obtained from two independent experiments and are expressed as the mean with differences no larger than 11 % between assays. induced excess DCF formation, this stimulation was not inhibited by superoxide dismutase.

()JSCUSSION
Biochemical reactions and pathways that utilize oxygen produce reduced oxygen species such as superoxide anion (O;), hydrogen peroxide (H 2 0 2 ). and hydroxyl radical (·OH) (Kappus. 1985;Halliwell and Gutteridge, 1984). These oxygen reactive species have recently been the subject of much attention in the CNS (Cohen. 1984;Braughler et al., 1985;Schwartz et al .. 1988;Scott and Lew. 1988). To date however, the oxidation of spin trapping agents by 0 2 (measured STEPHEN C. BONDY  by ESR spectroscopy) is the only direct measure for basal levels of oxygen reactive species employed in neuronal tissues (Sawada and Carlson, 1987). The intracellular trapping of DCFH, and its subsequent oxidation to DCF has been reported as a direct measure of oxygen reactive species formation in non-neuronal tissue (Scott ct al., 1988;Szejda et al., 1984;Bass et al., 1983 ). The techniques employed in the present study demonstrate that the non-fluorescent chemical (DCFH) is largely trapped within the synaptosome as a result of intrasynaptosomal deesterification. As noted in the methods. following preloading with DCFH-DA. synaptosomes were centrifuged (pelletted) and resuspended in fresh buffer. which allowed for the removal of extrasynaptosomal DCFH-DA. The presence of synaptosomes was required for de-esterification of DCFH-DA to the activated substrate DCFH since no fiuorochrome was detected in Tris buffer alone. Virtually all the oxygen reactive species described here were genuinely formed within biological tissue, since without such tissue. DCF formation occurred at a very low rate (Table I). DCFH leakage ( 18--25%, ). is similar to that observed for the [Ca'+], indicator dye fura-2 AM in synaptosomes (Bondy and McKee. 1990). These data support the concept that DCFH-DA enters synaptosomes, is de-esterified by intrasynaptosomal esterascs to DCFH. which is oxidized by intrasynaptosomally formed oxygen reactive species to the detected fluorochromc DCF.
Under basal conditions. DCF formation was relatively low (Table 2) while after addition of free radical generating agents, DCF formation increased (Figs l,  3 and 4). The fact that DCF formation showed good linearity with the amount of protein employed (Fig.  2), and was stimulated by ascorbate in a concentration-dependent manner (Fig. 3) further supports the use of DCFH-DA as an oxygen reactive species probe in neuronal tissue preparations.
The optimal dye loading time that provided maximal DCF formation after ascorbate/FeS0 4 stimulation was 15 min (Fig. 1). Interestingly, the extent of DCF formation was : 15 min > 60 min ~ 30 min > 5 min. Since DCF fluorescence reflects intracellular oxygen reactive species formation, lengthy dye loading times (30--60 min) may allow for increased dye leakage from the synaptosome. Therefore, beyond 15 min, DCFH may increasingly leak into the extrasynaptosomal medium and is removed consequently by centrifugation, resulting in lower rates of synaptosomal DCF formation.
Ascorbate/FeS0 4 -induced DCF formation was completely inhibited by deferoxamine (Fig. 4), a specific Fe 3 + chelator. These findings demonstrate that DCF (oxygen reactive species) formation results from iron-dependent radical reactions (Halliwell and Gutteridge, 1986). Ascorbate/FeS0 4 and xanthine/xanthine oxidase are known to produce o;' H 2 0 2 and ·OH (Davies and Goldberg, 1987b;Kuppusamy and Zweier, 1989). Xanthine/xanthine oxidase-induced DCF formation was not blocked by superoxide dismutase (Fig. 4), a result in agreement with a report of Bass and coworkers ( 1983) who were unable to inhibit xanthine/xanthine oxidase-induced DCF formation in neutrophils. Thus, o; does not appear to play a role in the oxidation of DCFH to DCF in oxidatively stressed synaptosomes. Alternatively, exogenously added superoxide dismutase may have been unable to penetrate the intact synaptosomal membrane, preventing its interaction with o;, as suggested by Chan et al. (1988).
Our results suggest that in oxidative stress studies, care must be taken with regard to the choice of buffer. While Tris and phosphate buffers enabled maximal DCF formation after ascorbate/FeS0 4 stimulation, FeS0 4 alone did not stimulate DCF formation in phosphate buffer, perhaps due to formation of an insoluble iron-phosphate complex (Table 2). HEPES buffer attenuated ascorbate/FeSO 4 -stimulated formation of DCF. Recent studies have reported that Tris and Hepes buffers may afford protection against oxygen radical-induced proteolysis of bovine serum albumin (Davies et al., 1987). This study partially confirms the concept that HEPES can quench the reactivity of free radicals. In the present study, Tris provided little protection against oxygen reactive species (Table 2). Although Tris has been suggested to be a good scavenger of ·OH, substitution of Tris with phosphate buffer did not enhance the formation of DCF (Table 2). Furthermore, studies from this laboratory found that well known ·OH scavengers, such as DMSO (l %, v/v) and mannitol (0.1 mM) decreased DCF formation in Tris buffer (data not shown). These findings support the hypothesis that DCFH is oxidized by more than one form of oxygen reactive species, such as H 2 0 2 , ·OH and ferry! ion (Bass et al., 1983;Szejda et al., 1984;Dunford, 1982).
Much has been written regarding the role of Ca 2 + in mediating cytotoxic mechanisms in the brain (Patel et al., 1988;Komulainen and Bondy, 1988). To determine whether Ca 2 + plays a role in DCF formation, ascorbate/FeS0 4 -induced DCF formation in HEPES buffer with and without Ca 2 + was investigated. Ca 2 + did not significantly enhance the effects of ascorbate/FeS04 on DCF formation (Table 2). Therefore, Ca 2 + does not appear to play a role in the intracellular formation of oxygen reactive species, as measured by DCF formation, in synaptosomes. Other studies report that Ca 2 + and free radicals act synergistically in lipid peroxidation to damage cell membranes (Braughler et al., 1985). Braughler and coworkers measured the effects of free radicals on synaptosomal GABA uptake, an indirect measure of lipid peroxidation. The present study makes no statement regarding lipid peroxidation, but directly measures oxygen reactive species generation (DCF formation), which is generally believed to be the initiation step in lipid peroxidation. Therefore, this lack of a synergistic effect by Ca 2+ in ascorbate/FeSO 4 -induced DCF formation supports the hypothesis that early oxygen radical-induced damage may be independent of lipid peroxidation (Davies and Goldberg, l 987a;Richards et al., 1988). On the other hand, our data show no exacerbation of oxidative activity in the absence of Ca 2 + in the incubation medium. Such a Ca 2 +-free milieu has been postulated to enhance free radical formation and thus bring about cell death (Fariss and Reed, 1985).
This study constitutes the first report concerning the application of a fluorescent technique for quantitation of oxygen reactive species formation in synaptosomes. The method is rapid, quantitative, and since DCF formation is in direct molar proportion to oxygen reactive species (i.e. ·OH, H 2 0 2 , ferry! ion), the results may be expressed directly as moles of oxygen reactive species formed. The sensitivity of the fluorescent technique allows quantitation of basal formation rates of oxygen reactive species, and thus has the potential to be adapted for studies involving in vivo exposure to neurotoxic compounds (LeBel et al., 1989). The sensitivity also allows for further adaptation of the method to subcellular preparations in separate brain regions. The use of DCF should provide further insight into the early events that occur under oxidative stress conditions in the brain.