Quantitative and Temporal Characterization of the Extracellular H202 Pool Generated by Human Neutrophils*

The extracellular HzOz concentration surrounding stimulated human neutrophils was continuously quan- titated with a sensitive, Hz02-detecting electrode. Following stimulation of neutrophils with phorbol myris- tate acetate, opsonized zymosan particles, or N-for-myl-Met-Leu-Phe, the extracellular HzOz concentration rapidly increased and maintained steady state con- ditions before falling to undetectable levels in a manner that was dependent on the triggering agent used. Total extracellular H202 accumulation for each stimulus was quantitated as the integral of the HzOz concentration with respect to time. HzOz accumulation in the extra- cellular milieu was unaffected by the addition of superoxide dismutase, whereas exogenous catalase or myeloperoxidase completely consumed the released HzO2. Analysis of HzOz metabolism by neutrophils re- vealed that stimulus-dependent differences in the size of the extracellular HzOz pool may be partially attrib- utable to differences in hypochlorous acid generation by the HzOz, myeloperoxidase, chloride system. Fi- nally, both the concentration of HzOz in the extracellular space and its utilization by myeloperoxidase could be diminished in the presence of an extracellular target cell. These data indicate that the ability of a triggering agent to stimulate the neutrophil to generate HzOz and release myeloperoxidase, coupled with the characteristics of a target cell population, control HzOz metabo- lism in effector-target cell interactions.

3 Supported by National Institutes of Health Grant 5 R01 HL-28024-02. TO whom all correspondence should be addressed. catalyzed dismutation to produce H202 (16). As the neutrophil's plasma membrane begins to encircle the particle, 0; and HzOz are released directly into the extracellular medium until the particle is completely engulfed within a phagocytic vacuole. Inside this environment, H202 may remain in the vacuole or diffuse either into the cytoplasm or outside the cell. Therefore, the H20z generated by an intact neutrophil should exist in an equilibrium among interconnected vacuolar, cytosolic, and extracellular "pools." In turn, this H202 equilibrium will be regulated by the relative amounts of Hz02 generated or consumed in each of these individual sites (Fig. 1). Within the phagocytic vacuole or in the extracellular space, H,Oz may be utilized by the secreted lysosomal enzyme myeloperoxidase or it may react directly with target molecules (16). In contrast, H202 that has diffused into the cytoplasm is primarily reduced by either catalase or glutathione peroxidase (16). Thus, there are multiple factors controlling the distribution and concentration of H202 in the intracellular and extracellular environment. Nonetheless, it is clear that the ability of a phagocyte to mediate Hz02-dependent damage against an extracellular target is not dictated by the total amount of H20, produced but rather the extracellular H202 concentration as a function of time.
At present, little is known about the size or characteristics of the extracellular pool of H202 surrounding stimulated neutrophils. In order to continuously monitor the extracellular H202 concentration, a detection system must be chosen that ideally does not alter H20, production or H202 catabolism, or perturb the H202 concentration in any of its three pools. To date, no studies have accurately monitored the extracellular pool as a function of time because current techniques either require the scavenging of all the released H202 (thus constantly perturbing the normal equilibrium between intra-and extracellular pools) or require the addition of agents that block normal catabolic pathways of H202 metabolism. Advances in polarographic techniques have led to the development of H202-sensing electrodes that rapidly quantitate H202 concentrations while consuming only miniscule amounts for analysis. Utilizing this approach, we have quantitated and examined both the characteristics of the free pool of H202 released by human neutrophils and its interaction with an extracellular target cell population.

EXPERIMENTAL PROCEDURES
Cell Preparations-Neutrophils were isolated from the peripheral venous blood of normal volunteers by Ficoll-Hypaque density centrifugation followed by dextran sedimentation as previously described (17). Cells were suspended in Dulbecco's phosphate-buffered saline (pH 7.4; Grand Island Biological Co., Grand Island, NY) supplemented with 1 mg/ml of glucose.
Erythrocytes were also isolated from the peripheral venous blood of normal volunteers. In order to minimize hemolysis, all cell washes were performed in the presence of 100 pg/ml of human serum albumin (Sigma) and cells were centrifuged at 4 "C.
HzOz-detecting System-HzOz release was measured continuously with a YSI model 25 oxidase meter fitted with a YSI 2510 oxidase probe (Yellow Springs Instrument Co., Yellow Springs, OH). The electrode was covered with a single layer of polycarbonate membrane (0.015-fim pore size; Nuclepore Corp., Pleasanton, CA). The meter was attached to a Hewlett-Packard 7130A dual channel chart recorder (Hewlett-Packard, Palo Alto, CA) fitted with an external voltage divider circuit. In order to allow simultaneous recording from 3 probemeter assemblies, an ATC model 342 flip-flop timer (Automatic Timing and Controls Co., King of Prussia, PA) was interposed between two of the oxidase meters and one of the recorder inputs. Meters were calibrated before each experiment with dilutions of reagent HZ02 (30% H202, Mallinckrodt Inc., Paris, KY) based on an extinction coefficient at 230 nm of 81 M-' cm" (17). The probe could detect a H2OZ concentration as low as 0.25 GM and the meter response was linear from 0-100 NM ( r = 0.999). The oxidase probe consumes H20z at a rate of 1.81 X lo-" mol/h at a probe current of 1 namp (42). The electrode did not respond to either reagent HOC1 or Nchlorotaurine. In some experiments, H202 was also measured spectrophotometrically by the method of Thurman et al. (18) as previously described (1).
Neutrophil System-Neutrophils (3 X 105/ml, except as otherwise noted) were incubated in siliconized cuvettes (final volume of 4 ml) were prepared in Me2S0 (Sigma) and stored at -20 "C. In some experiments, neutrophils were preincubated for 10 min with cytochalasin B (Aldrich Chemical Co.), which was prepared from stock solutions (10 mg/ml in MezSO) stored at -20 "C.
All results are expressed as the mean ? 1 S.D.

HtOz
Release by Triggered Neutrophils-Neutrophils are known to generate HzOz in response to a variety of soluble and particulate triggering agents (16). The most powerful soluble stimulus is the croton oil derivative PMA. This agent triggers the oxidative metabolism of neutrophils (19), stimulates lysosomal enzyme release (eo), and initiates vacuole formation (21). When 3 x lo5 neutrophils/ml (a total of 1.2 X lo6 cells in a final volume of 4 ml) were stimulated with PMA (25 ng/ml), Hz02 was readily detected in the extracellular space within 40 s (Fig. 2). Continuous monitoring of the extracellular H202 pool revealed that its concentration peaked at 12. min. If aliquots of the supernatant from stimulated cells were removed as the maximal levels were detected and then measured by an independent spectrophotometric assay, the H202 concentration varied by no more than 15% between the two techniques. The effect of varying the neutrophil concentration on the extracellular HzOz pool is shown in Fig. 3. Although the extracellular H202 concentration increased with cell number, the maximal levels of HzO2 did not double with a corresponding doubling of the neutrophil concentration. In addition, with increasing numbers of neutrophils, peak levels of HzOn and their subsequent fall to undetectable levels occurred more rapidly (Fig. 3). Apparently, as neutrophil numbers are increased, there is a relatively greater increase in H202 catabolism than in its generation.
Although PMA-stimulated neutrophils release substantial amounts of Hz02 into the extracellular space, PMA is a nonphysiological trigger and may not accurately reflect normal H,Oz metabolism. In order to examine the neutrophils' response to stimuli more likely to be encountered in uiuo, biologically relevant particulate and soluble triggers were next studied. Serum-treated zymosan particles are coated with the complement fragment C3b which is recognized by specific receptors on the neutrophil membrane (22). Compared to PMA-stimulated cells, the extracellular HzO, concentration peaked more rapidly when cells were triggered with opsonized zymosan but the maximal levels were significantly lower ( 2). In 12 experiments, the peak extracellular HzOz concentration was 3.6 f 1.0 W M after 8.1 k 1.7 min and returned to undetectable levels after 66.0 * 7.9 min ( n = 9). N-Formyl oligopeptides resemble naturally occurring bacterial products and can stimulate neutrophil chemotaxis, lysosomal enzyme release, and oxygen metabolism (23). Although these peptides are considered weak stimuli for oxygen metabolite generation by neutrophils, extracellular HzOZ could readily be detected in our system. In tration generated by PMA-stimulated neutrophils (3 X 106/ml) was measured continuously in the presence or absence of added catalase or myeloperoxidase. Bovine catalase (25 pg/ml) was added either 5 min prior to addition of PMA (---) or when Hz02 reached peak concentration (-). Purified canine myeloperoxidase ("0) (2 milliunits/ml) was added at peak H202 in a simultaneous assay (. . . .). extracellularly by FMLP-triggered cells by inhibiting vacuole formation or closure (24). However, the extracellular H202 concentration actually decreased when cytochalasin B-treated neutrophils were studied (Fig. 4) despite a 2-%fold increase in 0; release (data not shown).
In order to directly compare the total amounts of Hz02   (Fig. 5). These results underline both the specificity and the rapid response time of the HzOz-sensing electrode in this system. As 0; is generated by the neutrophil, it can dismutate to H202 or be consumed in other redox reactions (16). In an attempt to shuttle all the 0; to H2O2, experiments were performed in the presence of exogenous superoxide dismutase (10 pg/ml). For PMA-or zymosan-stimulated neutrophils there was no change in the extracellular H202 concentration as a function of time ( n = 3). Apparently, the pool of 0; available to exogenous superoxide dismutase dismutates spontaneously to H20z under the conditions studied.
If the extracellular H202 detected in our system arises primarily from the dismutation of 0, rather than the direct divalent reduction of 02, we would predict that the 0; scavenger cytochrome c would inhibit H2O2 production. Indeed, the extracellular H202 fell to <lo% of control values when neutrophils were stimulated in the presence of 160 p~ cytochrome c. However, the addition of superoxide dismutase in quantities sufficient to inhibit cytochrome c reduction (10 pg/ ml) failed to regenerate more than 50% of the expected H202. This unexpected result was explained by the finding that ferricytochrome c rapidly consumed HzOz in cell-free preparations. Thus, it is not possible to accurately interpret data with respect to polarographically detected H202 in the presence of cytochrome c.
Effects of Azide on the Extracellular H202 Pool-The size of the extracellular H202 pool at any point in time is regulated by the rate of H202 generation uersus the rates of enzymatic and nonenzymatic consumption. H202 generated by the neutrophil can be enzymatically catabolized via the glutathione peroxidase system, catalase, or myeloperoxidase but the latter two enzymes are heme-containing and their activity can be inhibited by azide (16). If neutrophils were stimulated in the presence of 1 mM azide, the amounts of HzOz detected extracellularly were dramatically increased (see Fig. 2). The peak extracellular Hz02 concentration with PMA-stimulated cells was increased 5-fold to 61.9 f 12.3 p M after 79.5 f 9.3 min ( n = 13) and, for zymosan-triggered neutrophils, it increased 12.3-fold to 44.3 ? 9.1 p~ after 50.7 -1-5.0 min ( n = 3). After the extracellular H202 concentration peaked in the presence of azide, the HzOz slowly disappeared with either population of stimulated cells (Fig. 2). Approximately 75% of the H2O2 could still be detected after a 4-h incubation. In contrast, the addition of azide to FMLP-stimulated neutrophils resulted in an insignificant increase in the HzO, concentration (Fig. 4).
As with PMA and zymosan, there was a slow decay in extracellular H202 after the maximal concentration was reached.
HOCl Generation by Stimulated Neutrophils-The large increases in the extracellular H202 concentration observed in the presence of azide suggest that large amounts of the generated H202 are consumed by either catalase or myeloperoxidase. Although the amounts of H2O2 reduced by catalase are difficult to assess, we have recently described a technique to measure H202 utilization by myeloperoxidase (17). In the presence of H202, myeloperoxidase can oxidize C1-, Br-, or Ito their respective hypohalous acids (16). Based on their in uiuo concentrations, C1-is the most likely halide oxidized and model H202-myeloperoxidase-C1-systems have been demon-  Because HOCl is the rate-limiting step in the chlorination reaction (25), exogenous taurine should not markedly alter the size of the extracellular H202 pool. Indeed, in the presence of 10 mM taurine, there were only small decreases in the peak HzOz concentration detected with PMA-, zymosan-, or FMLP-stimulated cells (TabIe I). If the amounts of N-chlorotaurine generated in these samples were simultaneously quantitated, zymosan-triggered neutrophils generated 2 times more N-chlorotaurine than the PMA-stimulated cells (Table  I).' Thus, neutrophils triggered with zymosan appear to generate a lower extracellular H20z concentration but shuttle larger amounts of H202 through the myeloperoxidase system than those treated with PMA. FMLP-stimulated cells could also chlorinate taurine, thus demonstrating that sufficient amounts of myeloperoxidase were released to catalyze HOCl formation (Table I). We had noted earlier that the extracellular H2O2 pool was smaller and disappeared more rapidly when FMLP-stimulated cells were pretreated with cytochalasin B (see Fig. 4). Because cytochalasin-treated, FMLP-triggered neutrophils release increased quantities of lysosomal enzymes (22,23), we speculated that the changes in the extracellular HzOz concentration could be related to increased H202 utilization by myeloperoxidase. Indeed, with cytochalasin-treated cells, the extracellular H202 concentration fell to base-line levels within 5 min (see  Neutrophils (3 X 106/ml) were incubated in the presence or absence of erythrocytes (3 X 106/ml) and stimulated with either PMA (25 ng/ ml) or zymosan (1.25 mg/ml) in the presence of added taurine (10 mM). After the extracellular H202 returned to base-line, aliquots were removed and N-chlorotaurine formation was quantitated as described under "Experimental Procedures." Results are expressed as mean f Thus, it appears that the extracellular H202 concentration is depleted in the presence of an extracellular target. When neutrophils are triggered to generate oxygen metabolites in the presence of an extracellular target, sufficient quantities of H202 may be utilized by myeloperoxidase to mediate cytotoxicity via HOCl (26). If H,Oz that is normally utilized by myeloperoxidase instead interacts directly with the target cell, then the amounts of H202 available to myeloperoxidase would be decreased. In order to test the ability of the erythrocytes to "steal" H2O2 from myeloperoxidase, neutrophils (3 x 105/ml) were triggered with PMA or zymosan in the presence of taurine (10 mM) with or without erythrocyte targets (3 x 106/ml) and N-chlorotaurine formation quantitated. Indeed, the quantity of N-chlorotaurine was reduced 61.9 f 7.2% (n = 4) and 22.5 f 14.7% (n = 3) for PMA-and zymosan-stimulated cells, respectively (Table 11). These re-ductions were not due to a direct reaction of the generated HOCl with the erythrocyte (rather than the exogenous taurine) because >95% of a bolus of HOCl (100 nmol) added to a mixture of erythrocytes and taurine could be recovered as N-chlorotaurine. Likewise, >95% of synthesized N-chlorotaurine (100 nmol) could be recovered following a 90-min incubation with 3 X 106/ml of erythrocytes. It would appear that the erythrocyte targets are capable of reducing both the extracellular HzOz pool and its subsequent utilization by the neutrophil.

DISCUSSION
Recent studies have clearly demonstrated that phagocytes can generate sufficient quantities of Hz02 to destroy cultured endothelial cells (1, 2), fibroblasts (2, 3), or tumor cells (4-6). In addition, HzOz can mediate a variety of nonlytic effects by altering the function of erythrocytes (7-9), platelets (lo), neutrophils (ll), T and B lymphocytes (12-14), and natural killer cells (15). Although the biochemical processes underlying these HzOz-dependent effects are unclear, several reports have demonstrated that the rate of H202 generation and its concentration as a function of time play a key role in determining target cell damage or destruction (5,27-29). Triggered neutrophils can generate H202, but the extracellular concentration is controlled by both the rates of H20z release and H2O2 catabolism at any point in time. Thus, in order to quantitatively measure the extracellular H20, concentration surrounding the neutrophil, neither of these parameters can be altered. The polarographic technique used in this report allowed us to continuously monitor the extracellular H20z pool while minimally perturbing H,02 metabolism or distribution. For example, 3 x lo5 neutrophils/ml stimulated with PMA generated a peak extracellular H20z concentration of -12 PM for 2 min. During this time, the electrode would have consumed only 2.5 X lo-" mol of H,02 for analysis. Thus, the electrode may be envisioned as an innocent bystander that continuously senses the surrounding pool of H20z without altering its size. In contrast, all the other studies that have examined Hz02 release by phagocytes have been limited to cumulative measurements of H202 generation (18, 30-38).
In the most commonly used technique, phagocytes are stimulated in the presence of exogenous horseradish peroxidase and a hydrogen donor (e.g. scopoletin (35), phenol red (38), diacetyldichlorofluorescein (32, 34)) whose H20n-dependent oxidation can be followed spectrophotometrically or fluorometrically. Optimal quantities of the peroxidase and its substrate are added in an attempt to consume all the H20, in the extracellular space. Thus, by design, this approach circumvents normal routes of H,Oz catabolism and tends to maximize the amounts of Hz02 released by continuously perturbing the normal H20, equilibrium between intracellular and extracellular sites. Although this technique is useful in quantitating the maximal amounts of HzOz that a phagocyte might release, it does not allow one to assess the instantaneous extracellular Hz02 concentration.
In our system, Hz02 could be detected in the extracellular space within 40 s for all three stimuli. Because the polarographic analyses did not depend on the presence of an oxygen metabolite scavenger or trap (e.g. cytochrome c for O;, or peroxidase for HzO2) and consume only trivial amounts for analysis, it appears that H20, is released freely into the extracellular milieu under physiologic conditions. The peak extracellular H202 concentrations reached by PMA-, zymosan, and FMLP-triggered cells varied by a factor of 12 and occurred within 26 min-, 8 min, and 3 min, respectively. Once attained, the maximal extracellular concentrations remained constant for various lengths of time (i.e. steady state conditions existed) before the rate of H202 catabolism began to exceed the rate of release and the extracellular H20, concentration fell to undetectable levels. These results indicate that neutrophils would bathe adjacent targets in a constantly changing Hz02 pool whose concentration and lifetime is regulated by the particular triggering agent used.
PMA and zymosan are both powerful stimuli of oxygen metabolite generation and optimally activate NADPH oxidase activity in human neutrophils (39). Thus, we were surprised that PMA-stimulated cells generated peak extracellular H,Oz concentrations and c x t values 3 . 4~ and 9~ greater, respectively, than those obtained with zymosan-treated cells. However, in the presence of azide, both PMA-and zymosantriggered neutrophils released HzOZ extracellularly at comparable rates and the peak H202 concentration detected with zymosan-stimulated cells was 72% of that observed with PMA. Azide can increase the extracellular concentration of HzOz by inhibiting catalase and myeloperoxidase and it is tempting to estimate rates of HzOz generation and catabolism by comparing HzO, release in the presence and absence of azide. Unfortunately, azide also interferes with the termination of the respiratory burst (probably by inhibiting autoinactivation of the NADPH oxidase by the myeloperoxidase system, Ref. 40) and increases Hz02 generation by artificially prolonging oxygen metabolite generation. Nonetheless, azide uncovered the potential of both zymosan-and PMA-stimulated cells to release large amounts of H202 into the extracellular milieu. It is not surprising that heme-enzyme inhibitors have been reported to enhance phagocyte-mediated, H202dependent effects (1, 5, 41) but the HzOz concentrations attained are clearly much higher and maintained longer than those reached under physiologic conditions. Neutrophils can use H202 directly to mediate cytotoxicity or alter cell function but it is clear that H202 can also be utilized by myeloperoxidase and C1-to generate HOCl (16). Myeloperoxidase, unlike other peroxidases, generates free HOCl and its rate of turnover and H202 utilization should not be altered by the presence of a chlorinateable substrate (25). 3 Indeed, the addition of high concentrations of taurine did not markedly alter the characteristics of the extracellular Hz02 pool. In terms of extracellular cytotoxicity, the ratio of H,Oz to HOCl released will be regulated by the ability of myeloperoxidase to compete with all other routes of H202 utilization. Opsonized zymosan particles are a more potent stimulus for myeloperoxidase release than PMA (16) and we speculated that the low extracellular HzOz concentrations and E X t values obtained with zymosan could be partially related to increased H202 utilization by myeloperoxidase. Indeed, zymosan-triggered cells used at least 2 times more H2O2 for HOCl generation than the PMA-treated neutrophils. Similar relationships between HzOz release and HOCl generation were found when values for FMLP-triggered cells were compared to those for cells preincubated with cytochalasin B before FMLP addition. Thus, there appears to be an inverse relationship between the extracellular Hz02 concentration and the quantities of HOCl detected. Apparently, stimulated neutrophils can simultaneously release H202 and HOCl in ratios determined by the triggering agent used.
The ability to monitor the extracellular H202 concentration without altering catabolism afforded us the opportunity to examine changes in the H202 pool in a model of a neutrophiltarget cell interaction. Continuous monitoring of the extracellular H202 revealed that erythrocytes decreased the peak Chlorine acceptors could increase HzOz utilization indirectly by scavenging HOCl before it inactivates myeloperoxidase.
Hz02 concentration generated by PMA-or zymosan-stimulated cells by 76 and 30%, respectively. The erythrocytes did not mediate this effect by nonspecifically stimulating the neutrophils' catabolism of HzOz because the released H,O, can attack the erythrocyte and oxidize intracellular hemoglobin (7-9). Although the processes involved in HzOz consumption by the erythrocyte have not been identified in this report, we have found that increased amounts of H,On are recovered when either catalase activity or the GSH system are inhibited in the erythrocyte targets? In addition to a reduction in the extracellular H,Oz pool, the presence of erythrocytes decreased Hz02 utilization by myeloperoxidase as reflected in a 61 and 30% inhibition of N-chlorotaurine formation with PMA-and zymosan-triggered cells, respectively. The ability of the erythrocyte to "steal" more H202 from PMA-stimulated cells may be related to the fact that PMA is a weak stimulus for myeloperoxidase release (16) and the erythrocyte can more effectively compete for available H2O2 when only small amounts of the peroxidase are secreted. Depending on the target cells' ability to catabolize H2Oz, our results suggest that it may be directly damaged by released HzOz or the target may protect itself by consuming the released H202 before it is utilized by the myeloperoxidase system to generate a more toxic oxidant. Thus, H2O2 catabolism by certain target cells may represent an unusual defense mechanism aimed at reducing the generation of HOCl by stimulated neutrophils.
In conclusion, we have provided the first quantitative and temporal analysis of the extracellular H202 pool generated by intact neutrophils. Our results indicate that the triggering agents' ability to stimulate the neutrophil to generate HzOz and release myeloperoxidase, coupled with the characteristics of the target cell, all control H202 metabolism in an effectortarget cell interaction. These analyses allow us to directly study the extracellular pool of H20, that is utilized by phagocytes in host defense and the inflammatory response.