Evaluation of the probe 2',7'-dichlorofluorescin as an indicator of reactive oxygen species formation and oxidative stress.

The use of dichlorofluorescin (DCFH) as a measure of reactive oxygen species was studied in aqueous media. Hydrogen peroxide oxidized DCFH to fluorescent dichlorofluorescein (DCF), and the oxidation was amplified by the addition of ferrous iron. Hydrogen peroxide-induced DCF formation in the presence of ferrous iron was completely inhibited by def eroxamine and partially inhibited by ethylenediaminetetraacetic acid, but was augmented by diethylenetri aminepentaacetic acid. Iron-peroxide-induced oxidation of DCFH was partially inhibited by catalase but not by horseradish peroxidase. Nonchelated iron-peroxide oxidation of DCFH was partially inhibited by several hydroxyl radical scavengers, but was independent of the scavenger concentration, and this suggests that free hydroxyl radical is not involved in the oxidation of DCFH in this system. Superoxide anion did not directly oxidize DCFH. Data suggest that H2 0 2-Fe2+-derived oxidant is mainly responsible for the nonenzymatic oxidation of DCFH. In addition, peroxidase alone and oxidants formed during the reduction of H2 0 2 by peroxidase oxidize DCFH. Since DCFH oxidation may be derived from several reactive intermediates, interpretation of specific reactive oxygen species involved in biological systems should be ap proached with caution. However, DCFH remains an attractive probe as an overall index of oxidative stress in toxicological phenomena. Nephrotoxic cysteine conjugates derived from a variety of halogenated alkenes are enzymatically activated via the ,6-lyase pathway to yield reactive sulfur-containing metabolites which bind covalently to cellular macromolecules. Mitochondria contain ,6-lyase enzymes and are primary targets for binding and toxicity. Previously, mitochondrial protein and/or DNA have been considered as molecular targets for cysteine conjugate metabolite binding. We now report that metabolites of nephrotoxic cysteine conjugates form covalent adducts with rat kidney mito chondrial phospholipids. Rat kidney mitochondria were incubated with the 35 8-labeled conjugates S-(1,1,2,2-tetrafluoroethyl)-L-cysteine (TFEC), S-(2-chloro-1,1,2-trifluoroethyl)-L-cysteine (CTFC), S-(1,2-dichlorovinyl)-L-cysteine, and S-(1,2,3,4,4-pentachlorobutadienyl)-L-cysteine. Quantitation of metabolite binding to whole mitochondria and to mitochondrial protein and lipid fractions revealed that as much as 42% of the 35 8-label associated with the mitochondria was found in the lipid fraction. Total lipids were also extracted from 35 8-treated mitochondria and separated by thin-layer chromatography. 35 8-Containing metabolites were found in the lipid fractions from mitochondria treated with each of the conjugates. Lipids from both [ 35 S]CTFC-


I ntroductlon
The use of 2',7'-dichlorofluorescin diacetate (DCFH-DA)1 was first described as a fluorometric assay of hydrogen peroxide (H 2 0 2 ) in the presence of peroxidase by Keaton and Brandt (1). Additionally, DCFH-DA has been used to measure the formation of lipid hydroperoxides and has been proposed as an alternative to traditional lipid peroxidative techniques such as the thiobarbituric acid method (2). Several studies dealing with the effects of reactive oxygen species in cell culture (3)(4)(5) and in aqueous systems (6) have employed DCFH-DA. Recently, DCFH-DA has been used to investigate the role of reactive oxygen species in various mechanisms of neurotoxicity (7)(8)(9)(10).
Studies using DCFH-DA in either intact cells or subcellular preparations have been based on the premise that the nonpolar, nonionic DCFH-DA crosses cell membranes and is enzymatically hydrolyzed by intracellular esterases to nonfluorescent DCFH ( Figure 1) (11). In the presence of reactive oxygen species, DCFH is rapidly oxidized to highly fluorescent 2',7'-dichlorofluorescein (DCF). However, it remains unclear which reactive oxygen species are responsible for the oxidation of DCFH. While H 2 0 2 and several lipid hydroperoxides, in the presence of hematin, are reported to oxidize DCFH (2), results on the ability of other reactive oxygen species such as superoxide anion (0 2 •-) and hydroxyl radical (•OH) to stimulate the formation of DCF are inconclusive. Conflicting findings regarding the inhibition of the oxidation of DCFH by superoxide dismutase (SOD) have been reported (4,7,11). The endeavor undertaken in this study was to investigate the reactive oxygen species involved in the oxidation of DCFH to DCF.
• Author to whom correspondence should be addressed. 1 Alkermes, Inc. I University of Alabama Birmingham. I University of California.

Experimental Procedures
Chemicals. 2',7'-Dichlorofluorescin diacetate (DCFH-DA) was purchased from Molecular Probes, Inc. (Eugene, OR). All other chemicals were purchased from Sigma Chemical Co. (St. Louis, MO). All solutions were prepared fresh and used immediately for all assays.
Preparation of Dichlorofluorescin. DCFH was prepared from DCFH-DA by the method of Cathcart et al. (2) by mixing 0.5 mL of 1.0 mM DCFH-DA in methanol with 2.0 mL of 0.01 N NaOH. This deesterification of DCFH-DA proceeded at room temperature for 30 min, and the mixture was then neutralized with 10 mL of 25 mM NaH 2 P0 4 , pH 7.4. This solution was kept on ice in the dark until use.
cence was corrected for by the inclusion in each experiment of parallel blanks (DCFH alone in buffer). The correction for autofluorescence was always less than 10% of the total. Benzoate Hydroxylation. All reactions (final volume 2.0 mL) were peformed in 25 mM NaH 2 P0 4 , pH 7.4, adapted from the method of Baker and Gebicki (12). The incubation mixtures contained DCFH (10-40 µL, final concentration 30-120 µM) and benzoate (20 µL of a 5 mM stock solution), and reactions were started by the simultaneous addition of 10 µM H 2 0 2 and 10 µM Fe 2 +. Each mixture was incubated at 37 °C for 5 min, and the reactions were terminated by the addition of 40 µL of 1 mM NaOH. Fluorescence of each solution, a measure of the formation of 2-and 3-hydroxybenzoate, was determined with excitation wavelength 300 nm (bandwidth 3 nm) and emission wavelength 390 nm (bandwidth 20 nm). Autofluorescence, always less than 5%, was accounted for by the inclusion of parallel blanks.

ResuHs and Discussion
The initial studies on DCFH oxidation focused on H 2 0 2 , a strong oxidizing agent. The reaction of H 2 0 2 with many organic compounds proceeds at a slow rate (14). The slow rate of DCFH oxidation by H 2 0 2 was demonstrated in the original studies by Cathcart and co-workers (2), who employed hematin in deaerated phosphate buffer solutions, with reactions that were required to be carried out at 50 °C for 50 min. Under those conditions, the H 2 0rinduced LeBel et al.  •Incubations (10 min) were carried out in the presence of 100 µM H 2 0 2 • The data were obtained from three independent experiments and are expressed as the means ± SE. •Incubations (10 min) were performed in the presence of 67.5 mM H 2 0 2 • The data were obtained from three independent experiments and are expressed as the means ± SE. oxidation of DCFH to DCF was linear. The present study, using ambient atmospheric conditions at 37 °C and incubation times of only 5-15 min, demonstrated good linearity  (Table I). The buffer employed in the reactions of iron-H 2 0 2 was Tris-HCl.
We have recently investigated the effects of several different buffer systems on DCF formation in a rat synaptosomal preparation (7). We demonstrated that optimal formation of DCF was obtained in the Tris buffer in comparison to HEPES and phosphate buffer systems. Tris is both a weak chelator and scavenger of •OH; however, it has modest effects on Fe 2 + autoxidation (15).
We next addressed the potential role of iron chelation in this system by employing various iron chelators. Deferoxamine, a potent and relatively inert chelating agent with a high affinity (Kd = 10- 31 ) for Fe 3 + (14, 16), inhibited the oxidation of DCFH in the presence of large quantities of H 2 0 2 without the addition of iron (Table II). The presence of 10 µM Fe 2 + or 10 µM Fe 3 + without H 2 0 2 did not result in DCFH oxidation (data not shown). In the  presence of H 2 0 2 + Fe 3 + the oxidation of DCFH was not affected by EDTA; however, it was inhibited by DETA-PAC ( Table I). In the presence of H 2 0 2 + Fe 2 +, EDTA partially prevented the formation of DCF, and surprisingly DET AP AC enhanced DCFH oxidation, in which fluorescence intensity readings actually exceeded the upper limits of detection (Table I). There is no ready explanation for the rapid conversion of DCFH to DCF by DETAP AC in this system, although it has been reported that Fe 2 + -DE-TAPAC chelates may actually catalyze the formation of ethylene gas from S-methyl 2-ketothiobutyrate, phenol from benzene, and the formation of the hydroxyl spin adduct of 5,5-dimethyl-1-pyrroline N-oxide (17).
The reaction of H 2 0 2 with Fe 2 + is thought to proceed by two pathways: The •OH generated in reaction 1 is known as an extremely reactive oxidizing agent that is believed to cause damage to a variety of cell constituents (18). This reaction is a Fenton redox chain reaction which is indirectly inhibited by EDTA (19). This is in agreement with our observations. Several reports have also suggested the formation of an iron-oxygen complex from the reaction of H 2 0 2 with nonchelated Fe 2 +, forming the ferryl ion (Feo 2 + or FeOH 3 +) in the reactions (20,21): FeOH 3 + -+ Fe0 2 + + H+ (3) To investigate whether the oxidation of DCFH was due to 0 0H formed from reaction 1, various 0 0H scavengers, listed in order of reactivity toward 0 0H from the reported second-order rate constants, were tested in the reaction mixture. In the H 2 0 2 + Fe 2 + system, DMSO, ethanol, mannitol, and Tris provided partial inhibition, while formate (koH = 2.8 X 10 9 M-1 s-1 ) was without effect ( Figure   3). However, increasing the concentration of these scavengers 10-f old did not provide any further inhibition of DCFH oxidation. For example, mannitol in the range of 5-10 mM showed a 22-26% inhibition ofDCFH oxidation, while scavengers such as Tris, which has a reported second-order rate constant for the reaction with 0 0H of 1.3 x 109 M-1 s-1 , showed greater inhibition than mannitol (koH = 1.8 X 10 9 M-1 s-1 ) at equimolar concentrations (13).
X 10 9 at concentration S = 5 mM, and the estimated concentration D of the detector DCFH = 60 µM, the predicted rate constant for the reaction of DCFH with ·OH is on the order of 10 11 M-1 s-1 , a value indicating that free 0 0H is not involved in the oxidation of DCFH to DCF.
Two alternative assays using the nonchelated H 2 0 2 + Fe 2 + system were employed in which DCFH was used as a scavenger, notably benzoate hydroxylation and deoxyribose oxidation. Results showed that DCFH moderately but significantly inhibited the oxidation of deoxyribose in a concentration-dependent manner (20% at 2 µM DCFH, p ~ 0.05), while benzoate hydroxylation was unaltered.
These data suggest that in the nonchelated H 2 0 2 + Fe 2 + system the oxidation of DCFH is not derived from free "OH radical. Alternatively, the carboxyl group of DCFH may bind Fe 2 + and lead to the formation of site-specific 0 0H, which can oxidize DCFH. The site-specific oxidation of DCFH by 0 0H would not be expected to be affected by "OH scavengers. Thus, the evidence to suggest that 0 0H is the primary oxidant of DCFH remains equivocal.
To determine whether 0 2 ·-was involved in the oxidation of DCFH, we studied the effects of SOD and a protein without SOD activity, lysozyme, in the xanthine + xanthine oxidase system. SOD did not provide any significant inhibition ofDCFH oxidation (Table III). Thus 0 2 ·-may not be involved in the oxidation of DCFH.
Catalase in excess decreased, but did not completely inhibit, the fluorescence intensity resulting from H 2 0 2induced oxidation of DCFH (Table II). The effect of catalase on the oxidation of DCFH even occurred in the presence of deferoxamine. An explanation of this effect on DCFH is based on the decomposition of H 2 0 2 by catalase, which proceeds via the formation of compound I (reaction 4), a ferryl type structure with a porphyrin 11'catalase + H20 2 -+ catalase-porphyrin" +-Fe 1 V=O + H 2 0 (4) catalase-porphyrin" +-Fe 1 V=O-+ catalase-porphyrin-Fe v • + + Feoa+ (5) cation radical (14,22,23). The ability of catalase-iron complexes to carry out these types of oxidations has been previously reported (14,22,23).
Horseradish peroxidase (HRP, type II) also oxidized DCFH in presence and absence of H 2 0 2 ( Figure 4). A comparison of the concentration-response curves demonstrated that HRP, in the presence of 10 µM H 2 0 2 , was twice as potent in its ability to oxidize DCFH than HRP alone. These data indicate that, in the presence of H 2 0 2 , DCFH serves directly as a reducing substrate for HRP in a concentration-dependent manner.
This study suggests that the oxidation of DCFH to DCF in nonbiological systems is iron-peroxide dependent and is related to the oxidation state of iron. DCFH oxidation appears to be mediated not only by H 2 0 2 -iron complexes, u  the role of which in biological syst.ems is not yet complet.ely understood (24)(25)(26), but also by peroxidase and peroxidase-iron complexes. In addition, free ·OH may not be responsible for DCFH oxidation in the nonchelat.ed H 2 0 2 + Fe 2 + system. Therefore, due to the broad range of reactive species which oxidize DCFH, caution should be taken when interpreting data in the context of a particular reactive species. DCFH's lack of specificity toward reactive oxygen species is what makes it pot.entially appealing as a probe in studying toxicological phenomena. Numerous studies exist demonstrating the utility of DCFH as an index of free radical reactions that take place in the living animal following exposure to toxic chemicals (9,(27)(28)(29). Toxicology studies using DCFH have reported the following: (1) significant differences in brain DCF formation rates between vitamin E deficient mice and those provided normal diets (8); (2) significant differences in DCF formation rat.es in brain regions known to be selectively vulnerable to the neurotoxic organometals methylmercury and trimethyltin (9); (3) significant differences in brain DCF formation rates in animals pretreated with deferoxamine prior to exposure to methylmercury (30); (4) significant differences in brain DCF formation rates in animals exposed to toluene, a neurotoxic organic solvent, in contrast to no observable alt.erations in brain DCF formation in animals treated with the structurally related solvent benzene (31).
The present study demonstrated that DCFH may be oxidized by peroxidases alone. This raises the possibility that the presence of peroxidases in vivo would be sufficient to confound data generated from biological systems. However, it is widely accepted that the brain is by nature low in antioxidant protective agents such as catalase, glutathione peroxidase, and glutathione (JO). We have reported that a variety of neurotoxicants increase the oxidation of DCFH in brain subcellular preparations and that these differences are present in animals as early as 1 h postdose (9,30,31). It is possible that exposure to these agents may involve an induction of peroxidase, in response to an oxidative stress, and that this may also directly enhance DCFH oxidation. Use of the probe DCFH may provide a link between reactive oxygen species formation in vitro and the formation of such reactive int.ermediates in the nervous system of the living animal.