Spin Trapping Evidence for Myeloperoxidase-dependent Hydroxyl Radical Formation by Human Neutrophils and Monocytes*

12-myristate

Using the electron spin resonance/spin trapping system, 4-pyridyl l-oxide N-tert-butylnitrone (4-POBN)/ ethanol, hydroxyl radical was detected as the a-hydroxyethyl spin trapped adduct of 4-POBN, 4-POBN-CH(CH3)OH, from phorbol 12-myristate 13-acetatestimulated human neutrophils and monocytes without the addition of supplemental iron. 4-POBN-CH(CH3)0H was stable in the presence of a neutrophilderived superoxide flux. Hydroxyl radical formation was inhibited by treatment with superoxide dismutase, catalase, and azide. Treatment with a series of transition metal chelators did not appreciably alter 4-POBN-CH(CH3)OH, which suggested that hydroxyl radical generation was mediated by a mechanism independent of the transition metal-catalyzed Haber-Weiss reaction. Kinetic differences between transition metal-dependent and -independent mechanisms of hydroxyl radical generation by stimulated neutrophils were demonstrated by a greater rate of 4-POBN-CH(CH3)-OH accumulation in the presence of supplemental iron. Detection of hydroxyl radical from stimulated monocyte-derived macrophages, which lack myeloperoxidase, required the addition of supplemental iron. The addition of purified myeloperoxidase to an enzymatic superoxide generating system resulted in the detection of hydroxyl radical that was dependent upon the presence of chloride and was inhibited by superoxide dismutase, catalase, and azide. These findings implicated the reaction of hypochlorous acid and superoxide to produce hydroxyl radical. 4-POBN-CH(CH3)0H was not observed upon stimulation of myeloperoxidase-deficient neutrophils, whereas addition of myeloperoxidase to the reaction mixture resulted in the detection of hydroxyl radical. These results support the ability of human neutrophils and monocytes to generate hydroxyl radical through a myeloperoxidase-dependent mechanism.
A major component of the neutrophil and mononuclear phagocyte response to a microbial challenge is the generation of reactive oxygen species (1). The cascade of events associated with oxygen-dependent microbicidal activity begins with increased oxygen consumption and the formation of superoxide (OX)' through the action of a membrane associated NADPH-dependent oxidase (2-4). Reduction of OT by dismutation results in the generation of Hz02 (5). In addition, myeloperoxidase catalyzes the reaction of H202 and chloride to form products such as HOC1, which have potent bactericidal activity (6, 7). Although the in vitro transition metalcatalyzed Haber-Weiss reaction of 0; and H202 results in the formation of 'OH (8), the ability of phagocytes to generate 'OH without the addition of exogenous iron has not been clearly demonstrated. Since reactive oxygen species have been implicated in the pathology of tissue injury, resulting from a number of ischemic and inflammatory events (9, lo), it is important to clarify the role of 'OH in phagocyte biology as a basis for rational therapeutic design.
Given the reactive nature of ' OH towards a range of cellular and extracellular targets (11), detection methods require a high level of sensitivity and selectivity. Techniques with variable sensitivity and specificity, such as ethylene generation (12, 13) and deoxyribose oxidation (14), have produced conflicting results (reviewed in Refs. 15 and 16). ESR/spin trapping offers the potential to identify a number of free radicals generated by biological systems, including 0: and 'OH (17)(18)(19). Using the spin trap, DMPO, in conjunction with dimethyl sulfoxide, generation of 'OH by stimulated neutrophils and mononuclear phagocytes was not detected unless exogenous iron was present (20-23). However, spin trapped adducts of DMPO are unstable, particularly in the presence of a 0; flux Therefore, these findings point to the need for improvement in sensitivity and stability of the spin trapping technique in order to detect low levels of 'OH generated in the presence of a continued flux of OX. Recently, we have studied the reaction kinetics of the spin traps, DMPO, PBN, and 4-POBN with .a-hydroxyethyl radical, resulting from the reaction of photolytically generated 'OH with ethanol.2 The rate constant for the reaction of 4-POBN with a-hydroxyethyl radical was nearly an order of magnitude greater than those deter-'The abbreviations used are: O;, superoxide; DMPO, 5,5-dimethyl-1-pyrroline 1-oxide; PBN, N-tert-butyl-a-phenylnitrone; 4-POBN, 4-pyridyl-1-oxide N-tert-butylnitrone; PMA, phorbol 12-myristate 13-acetate; DTPA, diethylenetriaminepentaacetic acid; MDM, monocyte-derived macrophages.
* The second order rate constant for the reaction of a-hydroxyethyl radical with 4-POBN is 8.1 X lo6"' s-' (S. Pou et al., manuscript in preparation

8307
This is an Open Access article under the CC BY license. mined for DMPO and PBN. Recognizing the potential for improved sensitivity in the detection of physiological ' OH by this system, 4-POBN in conjunction with ethanol was used to determine if 'OH is generated by stimulated neutrophils and mononuclear phagocytes.

MATERIALS AND METHODS
Reagents-Deferoxamine mesylate, diethylenetriaminepentaacetic acid (DTPA), EDTA, hypoxanthine, xanthine, phorbol 12-myristate 13-acetate (PMA), ferricytochrome c (type VI) were purchased from Sigma. Superoxide dismutase, catalase and xanthine oxidase were obtained from Boehringer Mannheim; 4-pyridyl 1-oxide N-tert-butylnitrone (4-POBN) from Aldrich; pure ethyl alcohol (U. S. P.) from Warner-Graham, Cockeysville, MD; [l-13C]ethanol from Cambridge Isotope Laboratories, Cambridge, MA sodium azide from Fischer (Fair Lawn, NJ); and Hanks' balanced salt solution (HBSS) without Ca2+, M P , or phenol red from Gibco Laboratories. In order to minimize iron contamination, buffers were treated with Chelex 100 chelating resin (Bio-Rad) according to the method of Buettner (28) and xanthine oxidase was dialyzed against deferoxamine (1 mM) as described by Britigan et al. (29). Purified human neutrophil myeloperoxidase was kindly provided by Dr. William Nauseef (University of Iowa, Iowa City, IA). Activity of myeloperoxidase was quantitated using the o-dianisidine assay in which one unit of activity is defined as the concentration of enzyme which induces an increase of 0.001 absorbance units/min at 460 nm. This is equivalent to 2.66 X 10" IU.
Cell Isolation-Human neutrophils and monocytes were isolated from heparinized venous blood of normal human volunteers using dextran sedimentation and a Ficoll-Hypaque gradient (30). Monocytes were separated from lymphocytes by adherence to sterile culture dishes (21), or the mononuclear layer from the Ficoll-Hypaque gradient was used as a source of monocytes without further purification. Monocyte-derived macrophages were obtained by differentiation of human monocytes in culture for 4-7 days using medium 199 (University of Iowa Cancer Facility) supplemented with 13% autologous plasma or 10% fetal calf serum, penicillin (100 units/ml), and streptomycin (100 pg/ml) (21). Cells were suspended in Chelex-treated HBSS, pH 7.4 (1-20 X lo6 cells/ml) and maintained on ice until use.
Myeloperoxidase-deficient neutrophils were obtained from a subject whose status has been verified by the o-dianisidine myeloperoxidase assay and Western blot analysis (31).
Stability of 4-POBN-CH(CHdOH-The a-hydroxyethyl radical adduct of 4-POBN, 4-POBN-CH(CH3)OH, was obtained by irradiating a solution of HzOz (88 mM), ethanol (17 mM), and 4-POBN (10 mM) for 2 min with an ultraviolet light source (Ultraviolet Products model SCT1, San Gabriel, CA). Catalase (600 units/ml) was then added to eliminate excess H202. Neutrophils (10 X lo6 cells/ml) were added to 0.1 ml of the irradiated solution followed by PMA (100 ng/ml) to a final volume of 0.5 ml. The reaction mixture was transferred to a quartz ESR flat cell open to air and fitted into the cavity of an ESR spectrometer (Century Line model E109, Varian Associates, Palo Alto, CA). Stability at 25 "C was determined by monitoring 4-POBN-CH(CH3)OH peak intensity over time after neutrophil stimulation.
Superoxide Detection-Generation of 0; by neutrophils, mononuclear phagocytes and cell-free systems was measured spectrophotometrically at 550 nm as the superoxide dismutase-inhibitable reduction of ferricytochrome c using an extinction coefficient of 2 1 mM" cm" ESRISpin Trapping-Reaction mixtures consisted of cells (1-20 X 106/ml), 4-POBN (10 mM), ethanol (170 mM), PMA (100 ng/ml) and sufficient HBSS for a final volume of 0.5 ml. Reaction mixtures were transferred to a quartz ESR flat cell open to air, placed in the cavity of an ESR spectrometer (Century Line model E109 or E104A, Varian Associates, Palo Alto, CA) and spectra recorded at 25 "C. Instrument settings, unless noted differently, were as follows: microwave power, 20 milliwatts; modulation frequency, 100 kHz; modulation amplitude, 1.0 G; sweep time 12.5 G/min; and response time, 1 s. The receiver gains are given in each figure legend. Computer simulation of ESR spectra were obtained using an algorithm developed by Oehler and Janzen (33).

RESULTS AND DISCUSSION
In several studies using the spin trapping system DMPO/ dimethyl sulfoxide, generation of 'OH by stimulated human neutrophils was not observed unless supplemental iron was provided (20-23). In a more recent study using the spin trapping system PBN/dimethyl sulfoxide, in which 'OH was detected as PBN-OCH3, neutrophil 'OH formation was again observed only in the presence of exogenous iron (34). Together, these findings have served as the basis for the conclusion that neutrophils require an exogenous transition metal catalyst in order to generate 'OH. However, these results are entirely dependent on the sensitivity of the spin trapping system, which is determined by the relative rates of spin trap adduct formation and decomposition (23). Accordingly, we tested the ability of the 4-POBN/ethanol spin trapping system to detect 'OH from stimulated neutrophils in the presence and absence of supplemental iron.
Addition of PMA (100 ng/ml) to human neutrophils (1-20  POBN-CH(CH3)OH (Fig. lB), whereas catalase (300 units/ ml) inhibited 4-POBN-CH(CH3)0H (Fig. 1C). Previous spin trapping studies (20,23,37) have also reported enhanced .OH generation with low levels of superoxide dismutase in the presence of exogenous iron. This finding can be attributed to an increased rate of formation of HzOz which then undergoes transition metal-mediated reduction to 'OH (38). Treatment with azide (1 mM), in the presence of iron (Fig. lD), resulted in no change and in some experiments an increase in peak intensity. This result is explained by a decrease in HzOz catabolism through azide inhibition of myeloperoxidase and catalase (39). Therefore, one may conclude that in the presence of supplemental iron, conditions are favorable for the reduction of HzOz to *OH (14, 37).
Formation of 4-POBN-CH(CH3)0H was also observed upon PMA stimulation of neutrophils in the absence of supplemental iron (Fig. 2 4 ) . In contrast to the iron-supplemented system, addition of superoxide dismutase (15 units/ ml) abolished and azide (1 mM) significantly inhibited formation of 4-POBN-CH(CH3)0H (Fig. 2, E and D ) . Azide, in concentrations up to 10 mM, did not appreciably either inhibit OX generation from neutrophils or interfere with spin trapping of a-hydroxyethyl radical (data not shown). Therefore, inhibition of 4-POBN-CH(CHs)OH by azide did not occur through inhibition of the respiratory burst or free radical scavenging. Catalase (300 units/ml) also inhibited 4-POBN-CH(CH,)OH (Fig. 2C). As an additional control experiment, [13C]ethanol was substituted for ethanol in the reaction mixture to demonstrate that the reaction of a-hydroxyethyl radical with 4-POBN was responsible for the observed nitroxide. Stimulation of neutrophils with PMA in the presence of 4-POBN (10 mM) and [13C]ethanol (170 mM) resulted in the appearance of a nitroxide with hyperfine splitting constants consistent with 4-P0BN-[l3C]H(CH3)0H (AN = 15.75 G, AH = 2.4 G, A'3c = 3.75 G) (Fig. 2E), thus confirming the spin trapping of a-hydroxyethyl radical. Since the ability to observe the ESR spectrum corresponding to 4-POBN-CH(CH3)OH is also dependent on the stability of this spin trapped adduct, we determined the decomposition rate of preformed 4-POBN-CH(CH3)0H in the presence of neutrophils. Upon PMA stimulation, which resulted in a 0; flux of 2-3 pM/min/l X lo6 cells, there was only a slight reduction in 4-POBN-CH(CH3)0H signal intensity over a period of 24 min.3 Although 'OH was detected from stimulated neutrophils without supplemental iron, the potential involvement of transition metal-catalyzed ' OH formation was more thoroughly evaluated by treating neutrophils with transition metal chelators (40). Neutrophils were suspended in Chelex resintreated HBSS containing either DTPA (0.1 mM), EDTA (0.1 mM), or deferoxamine (0.1 mM). Treatment with these compounds did not significantly effect the spin trapping of ahydroxyethyl radical. Therefore, based on results from treatment with superoxide dismutase, catalase, azide, and chelators, we concluded that OT, HzO2, and possibly myeloperoxidase were involved in the transition metal-independent formation of 'OH by neutrophils. Previous studies have reported neutrophil derived 'OH with potential involvement of myeloperoxidase (13,(41)(42)(43)(44). However, definitive determination of neutrophil 'OH generation and the potential mechanisms responsible were limited by the low specificity of the detection methods and difficulty in differentiating transition metalcatalyzed 'OH formation.
It has been postulated that 'OH is a product of the in uitro, pH-dependent reaction of HOCl and 0; (45). Given that myeloperoxidase catalyzes the formation of HOCl (Reaction l), then this enzyme may be responsible for transition metalindependent ' OH generation by neutrophils.

(1)
HOCl + 0; + 'OH + 0 2 + C1- (2) Using the 4-POBN/ethanol spin trapping system, the ability of this reaction mechanism to generate 'OH was tested. Addition of human neutrophil myeloperoxidase (200 milliunits) to a mixture of hypoxanthine (2 mM) and sufficient xanthine oxidase to generate a prolonged flux of 0; (10 p~/ min) resulted in the accumulation of 4-POBN-CH(CH3)0H (Fig. 3B), which was inhibited by catalase (300 units/ml), superoxide dismutase (30 units/ml), and azide (1 mM) (Fig. 3, C-E). Potential interference by transition metal-mediated 'OH generation was minimized by dialyzing xanthine oxidase against deferoxamine (1 mM) and using Chelex resin-treated HBSS with 0.1 mM DTPA as the reaction buffer. The chloride dependence of 'OH generation in this system was demonstrated by the absence of 4-POBN-CH(CH3)0H when the reaction was performed in a halide-free medium (Chelextreated deionized, distilled water) (Fig. 3F). These findings suggested the ability of myeloperoxidase to drive 'OH generation through the reaction of HOCl and OX. The second order rate constant for the reaction of HOCl with 0; has been determined to be nearly 1 X lo7 M-' s-l for the pH range 5-7.5 (45). The similar pH range for the intraphagosomal and extracellular media of activated granulocytes (46) would provide favorable conditions for this reaction.
In an effort to further delineate a mechanistic role for myeloperoxidase in neutrophil *OH formation, spin trapping was performed with human monocytes, monocyte-derived macrophages (MDM), and neutrophils obtained from a myeloperoxidase-deficient individual. PMA stimulation of human monocytes, which contain myeloperoxidase (47), resulted in the formation of 4-POBN-CH(CH3)0H (Fig. 4A) accumulation was less than observed with neutrophils at a similar cell concentration. Treatment with azide (1 mM) abolished 4-POBN-CH(CH3)0H (Fig. 4B) Stimulation of MDM, which in the course of differentiation from monocytes lose myeloperoxidase activity (47), did not lead to detectable concentrations of 4-POBN-CH(CH3)0H, except in the presence of supplemental iron (Fig. 4, C and D). Likewise, PMA stimulation of myeloperoxidase-deficient neutrophils did not result in detectable 'OH formation (Fig. 5A). Addition of myeloperoxidase, at an amount similar to that in normal neutrophils (200 milliunits), resulted in the formation of 4-POBN-CH(CH3)0H (Fig. 5B) which was inhibited by treatment of the cells with catalase (300 units/ml), superoxide dismutase (30 units/ml), and azide (1 mM) (Fig. 5, C-E).
Taken together, these results offer direct evidence for a myeloperoxidase-dependent mechanism of ' OH generation by human neutrophils.
Upon inspection, the ESR spectrum resulting from the addition of purified myeloperoxidase to myeloperoxidase-deficient neutrophils (Fig. 5 B ) indicated spin trapping of an additional free radical forming a nitroxide with hyperfine splitting constants AN = 14.5 G and AH = 1.8 G. Interestingly, a mixture of the computer simulation of this species (Fig. 5F) with that of 4-POBN-CH(CH3)0H (Fig. 5G) resulted in a simulated spectrum (Fig. 5H) nearly identical to Fig. 5B. We believe that the mixed spectrum most likely resulted from changes in oxygen tension, rather than a physiological source. Since myeloperoxidase-deficient neutrophils generate a significantly greater 0; flux (39), upon addition of myeloperoxidase it is suspected that a greater amount of oxygen would ' evolve from the reaction of HOC1 and 0; (Reaction 2). This could favor formation of an oxygen-centered free radical derived from the reaction of 'OH with ethanol that effectively competes with the carbon-centered a-hydroxyethyl radical for 4-POBN, resulting in an ESR spectrum with hyperfine splitting constants consistent with two spin trapped adducts. The nature of this species is currently under investigation.
Overall, these results emphasize the utility of the 4-POBN/ ethanol spin trapping system for the evaluation of *OH in biological systems. Using this spin trapping system, we have demonstrated *OH generation by stimulated neutrophils and monocytes through a myeloperoxidase-dependent mechanism that does not require the presence of a supplemental Haber-Weiss catalyst. Initially, it would appear that our findings are in conflict with reports that suggest myeloperoxidase acts to limit potential neutrophil 'OH generation by decreasing HzOz available to react with transition metals (14,37). However, the sensitivity of the 4-POBN/ethanol spin trapping system allowed the detection of *OH in the absence of confounding iron. Under these conditions, Hz02 is less likely to undergo direct reduction to 'OH. Therefore, H202 is available for the myeloperoxidase-catalyzed formation of HOC1. Reaction with 0; results in the production of 'OH (Reactions 1 and 2).
A direct comparison between transition metal and myeloperoxidase-catalyzed .OH formation by neutrophils was made by observing the accumulation of 4-POBN-CH(CHJOH with time after PMA stimulation in the presence and absence of supplemental iron. In the presence of exogenous iron (0.1 mM ferrous sulfate), 4-POBN-CH(CH3)0H was detected earlier after activation and accumulated at a greater rate than in the absence of supplemental iron (Fig. 6). Clearly, the presence of supplemental iron drives the generation of 'OH through the Haber-Weiss reaction at a rate and extent that obscures OH formation through the myeloperoxidase-dependent mechanism. However, transition metal regulatory systems, such as lactoferrin and transferrin, limit potential 'OH for-mation through the Haber-Weiss reaction by binding iron in a form unable to catalyze the reduction of HzOz (38,(48)(49)(50)(51). Unless these systems are overwhelmed, as demonstrated in Fig. 6, a relatively low flux of 'OH generated through the transition metal-independent mechanism may be a primary physiological source of phagocyte-derived 'OH. The significance of 'OH formation through the myeloperoxidase-dependent pathway in microbicidal activity and neutrophilrelated tissue injury requires further study.