Journal
of Leukocyte
Biology
50:131-139
(1991)
Effect of Tumor Necrosis Factor on the Generation of Chlorinated Oxidants by Adherent Human Neutrophils Samuel Children’s
Hospital
Oakland
T. Test
Research
Institute,
Oakland,
California
Human neutrophils adherent to simulated biologic surfaces undergo significant activation of the respiratory burst over prolonged periods of time in response to stimulation with the cytokines tumor necrosis factor-alpha (TNFa) or tumor necrosis factor-beta (TNF3) or with the chemotactic peptide N-formyl-methionylleucylphenylalanine(FMLP). In this study, neutrophils were examined for their ability to generate the highly reactive and powerful oxidant hypochlorous acid (HOCI) and the longer-lived, less reactive endogenous nitrogen-chlorine (N-Cl) derivatives in response to these stimuli either alone or when exposed to recombinant human TNFa (rTNFa) or 3 (rTNF3) prior to addition of FMLP. Neutrophils adherent to fetal bovine serum-coated polystyrene tissue culture wells were able to generate only small quantities of HOCI when incubated with rTNFa, rTNF3, or FMLP individually. However, when neutrophils were first incubated with either rTNFa or rTNF3 prior to addition of FMLP, there was a marked increase in HOCI generation. Neutrophils stimulated in such a manner consumed -i8% of the HOCI generated in the formation of N-Cl derivatives. Further scrutiny of the response to the combination of rTNFa and FMLP revealed that HOCI release was rapid, with 80% of total HOCI accumulation occurring within i5 mm after FMLP addition. The amount of HOCI generated was dependent on the number of cells added and on the concentration of both rTNFa and FMLP. Comparison of HOCI generation with superoxide anion and myeloperoxidase release showed that the amount of HOCI generated was limited primarily by the amount of myeloperoxidase released rather than by the degree of respiratory burst activation. These results demonstrate that human neutrophils stimulated with FMLP after a brief incubation with rTNFa or rTNF3 can generate cytotoxic and microbicidal concentrations of chlorinated oxidants. Key words:
N-formyl-Met-Leu-Phe, rine derivatives
hypochlorous
INTRODUCTION Neutrophils entering an inflammatory focus could potentially encounter a diverse array of soluble mediators capable of triggering the release of toxic granule constituents and/or the generation of a variety of destructive oxygen metabolites. Recent attention has focused on the ability of cytokines released from activated macrophages or T-lymphocytes to stimulate these cytotoxic responses in human neutrophils [reviewed in 3 1]. Tumor necrosis factor-alpha (TNFa), which is released from macrophages exposed to lipopolysaccharide (LPS) [37], is capable of affecting a variety of cellular functions and has been examined for its effects on neutrophils in a number of studies [31]. Tumor necrosis factor-beta (TNF13), a product of activated T cells [10], shares structural features with and has biologic activity similar to that of TNFa [24], but has been used less extensively in studies of neutrophil function [3 1] Both of these cytokines can also be released following the exposure of peripheral blood mononuclear cells to mitogens and other cytokines, including interleukin-2 alone or with interferon-’y [22].
acid, myeloperoxidase,
nitrogen-chlo-
It is becoming increasingly apparent that the microenvironment in which the human neutrophil undergoes stimulation can profoundly influence its response to soluble agonists. For example, it has recently been demonstrated that the triggering of neutrophils adherent to simulated biologic surfaces compared with neutrophils in suspension resulted in a marked enhancement of hydrogen peroxide (H202) generation following an initial lag period of 15-90 mm [20,2 1 ] Respiratory burst activation having these features was characteristic of the neutrophil response to recombinant human tumor necrosis factor a (rTNFa) and 13 (rTNFf3) [20], to the chemotactic peptide N-formyl-methionylleucylphenylalanine (FMLP) [20], and to granulocyte/macrophageand granulocyte-colony stimulating factors [2 1 ] However, .
.
.
© 1991
Wiley-Liss,
Inc.
Received
September
Reprint requests: search Institute.
1 1. 1990:
accepted
December
18,
1990.
Samuel T. Test. Children’s Hospital Oakland 747 Fifty Second Street, Oakland. CA 94609.
Re-
132
Test
the effect of neutrophil adherence to simulated biologic surfaces on the generation of oxygen metabolites which derive from the reaction of H2O2 with other neutrophil constituents has not been examined. The most potent and reactive of these oxidants are the hydroxyl radical (OH ) which is generated by the reaction of superoxide anion (O2) with H,O2, and hypochlorous acid (HOCI), which is the end-product of the H2O2-myeloperoxidasechloride system [4,40]. Whereas unquestionable proof of OH generation by neutrophils does not yet exist [4,40], myeloperoxidase is the primary constituent of the neutrophil primary (azurophil) granule and has been shown to participate in the generation of both HOC1 [9,34,35,38] and a class of long-lived oxidants known as nitrogen-chlorine (N-Cl) derivatives [ 1 1 ,26,33-35, 38,39]. These two species of chlorinated oxidants are capable of mediating diverse biologic effects, including alterations in serum and cellular proteins which can result in functional changes or cell lysis [34]. TNFa and 13have been shown to be weak triggers for O2 or H2O2 production by neutrophils in suspension [3,6, 13-16, 19,23,25,28,36,41]. When used as priming agents for neutrophils in suspension, the TNFs can enhance 02 generation in response to FMLP [3,6, 12,32,41] or unopsonized zymosan [13] and can enhance myeloperoxidase-dependent iodination [13] or HOCI generation [29] in response to the latter triggering agent. Enhanced myeloperoxidase release by FMLP-stimulated neutrophils following rTNFa priming was not observed [6]. In light of the profound increase in respiratory burst activation which occurred when neutrophils were stimulated with the TNFs or FMLP on serum-coated polystyrene surfaces [20], we were interested whether the increase in H,O, generation was accompanied by a sufficient enhancement of myeloperoxidase release to result in an increase in HOCI generation of similar magnitude. In the experiments detailed below, we show that neutrophils adherent to serum-coated surfaces generate only negligible quantities of HOC1 when triggered with rTNFa, rTNF13, or FMLP alone and that this is likely due to minimal release of myeloperoxidase. However, a marked increase in HOC1 generation occurs when neutrophils are first incubated with rTNFa or rTNF13 prior to the addition of FMLP. This is accompanied by a significant increase in the amount of myeloperoxidase released into the extracellular medium. Furthermore, the absence of any lag period in respiratory burst activation when neutrophils are primed with rTNFa before triggering with FMLP (compared with incubation with these agonists alone), coupled with the rapid release of myeloperoxidase, results in the rapid attainment of maximal HOC1 levels over a time course approximating that for O2 production. .
.
MATERIALS Materials rTNFa
AND METHODS
was kindly
supplied
by Genentech,
Inc.
(South
San Francisco, CA). The preparation, as assayed by the supplier, had a specific activity of 5.6 X l0 U/mg and a concentration of 0.5 mg/mi in phosphate-buffered saline (PBS) and was stored at 4#{176}C. Working dilutions of 40 nM in Dulbecco’s PBS with Ca2 and Mg2 (UCSF Cell Culture Facility, San Francisco, CA), pH 7.4, supplemented with 10 mM glucose (PBSG) were prepared daily immediately before addition to incubation mixtures. rTNF13, also provided by Genentech, Inc. had a specific activity of 2 1 X i0 U/mg and a concentration ofO. 1 mg/mi. Working dilutions of2O nM in PBSG were prepared immediately before addition to incubation mixtures. FMLP (Sigma Chemical Co., St. Louis, MO) was stored at -20#{176}Cas a 3 x l02 M solution in dimethyisuifoxide (DMSO). Stock solutions of lO_6 M in PBSG were prepared daily immediately prior to use. Fetal bovine serum (FBS) was purchased from Hyclone Laboratories, Inc. (Logan, UT) and heat-inactivated at 56#{176}C for 30 mm prior to use. Other reagents used included cytochrome c (horse heart type VI), superoxide dismutase (SOD, from bovine erythrocytes), catalase (from bovine liver) 3 ,3 ,5 ,5 ‘-tetramethylbenzidine (TMB) sodium acetate, NADH, sodium pyruvate, disodium ethyldiaminetetraacetic acid (EDTA), taurine, sodium azide, L-methionine (Sigma); hydrogen peroxide (30%, ACS certified), N ,N-dimethylformamide (DMFA) potassium iodide (Fisher Scientific, Fair Lawn, NJ); PBS (Ca2 and Mg2tfree, I X and lOX concentrate), pyrogen-free water (UCSF Cell Culture Facility); Percoil (Pharmacia LKB Biotechnology AB, Uppsala, Sweden); and Triton X-iOO (Bio-Rad Laboratories, Richmond, CA). ,
.
,
‘
,
,
Isolation
of Polymorphonuclear
Neutrophils
Neutrophils were prepared by centrifugation of whole blood on discontinuous Percoil gradients. Isotonic stock solutions of Percoli ( 100% Percoll stock) were prepared by adding 1 part lOX PBS to 9 parts Percoll. The Percoli stock solution was subsequently diluted to 60% and 77% with 1 x PBS: these solutions were allowed to come to room temperature before use. Fifteen milliliters of 77% Percoli were layered under 15 ml of6O% Percoli in sterile 50 ml conical bottom polypropylene tubes (Applied Scientific, San Francisco, CA). Fifteen to 20 ml of fresh whole blood anticoagulated with sterile 1 .5% Na2EDTA was obtained by peripheral venipuncture from normal human volunteers, carefully layered over the Percoli and subsequently centrifuged at 500g for 30 mm at room temperature. The neutrophil layer was aspirated, contammating red blood cells (RBC) were removed by hypotonic lysis, and the remaining cells washed with PBS and
Generation
suspended to 2 X 106 cells/mI in PBSG. preparation was >96% neutrophils by Wright’s viability with >98% by trypan blue exclusion.
Endotoxin
The final stain and
Contamination
The endotoxin concentration of the lot of FBS used in these experiments was assayed by the manufacturer (Hycione Laboratories, Inc.) and determined to be 25 pg/ml. The endotoxin content of other reagents was quantitated with a chromogenic limulus amebocyte lysate assay (Whittaker Bioproducts, Inc. Walkersville, MD) with a lower limit of detection of 10 pg/ml. Endotoxin levels were undetectable in the 1 X and lOx PBS used for the preparation of Percoll dilutions, in the PBSG used for neutrophil suspensions and dilution of reagents, and in the working dilutions of rTNFa, rTNF13, and FMLP. The endotoxin content of Percoll could not be accurately determined because this substance interfered with the endotoxin assay. ,
HOCI Generation HOC1 generation was assayed in a system using neutrophils adherent to FBS-coated plastic tissue culture wells. The wells (4.0 cm2) of 12 well tissue culture plates (Costar, Cambridge, MA) were covered with 0.3 ml heat-activated FBS and incubated for 1-2 h at 37#{176}C in a humidified chamber. The FBS was then aspirated and the wells washed three times with 0.9% saline prepared with pyrogen-free water. Neutrophils (i06/well) were added and incubated for 10 mm at 37#{176}C prior to the addition of either rTNFa or rTNF13. The incubation mixtures were incubated for an additional 10 mm and FMLP was carefully added dropwise. The priming period of 10 mm was chosen on the basis of the optimal priming duration previously reported for respiratory burst activation by FMLP following TNFa priming [3,6,41]. Taurine (15 mM) was present in all incubation mixtures (final volume 1 ml) to scavenge HOCI and permit its subsequent quantification by the iodide oxidation method [38]. When accumulation of endogenous N-Cl derivatives was to be measured, taurine was omitted. Incubations were terminated by the addition of catalase (5 pUg), any nonadherent cells were pelleted by centrifugation, and the supernatants were assayed for HOC1 (as the stable N-chlorotaurine derivative) or endogenous N-Cl derivatives. In some experiments, neutrophils were added to buffer in 12 x 75 mm polypropylene tubes and incubated at 37#{176}C for 10 mm before the addition of rTNFa The tubes were then placed on a rocking platform to keep the cells in suspension for the duration of the incubation. FMLP was added after 10 mm and the incubation continued for 60 mm and terminated as described above. The cells were pelieted and supernatants assayed for HOC1 or N-Cl derivatives as described. .
of Chlorinated
Oxidants
Superoxide
by Adherent
Neutrophils
133
Generation
Activation of the neutrophil respiratory burst was quantitated by measuring the SOD-inhibitable reduction ofcytochrome c [5]. Incubation conditions were identical to those used to measure HOC1 generation except that cytochrome c (2 mg/mi) was substituted for taurine, and catalase ( 10 p.g/ml) was added to all wells to prevent the reoxidation of cytochrome c by any H2O, generated during the course of the incubation. Duplicate wells contained 50 ig/ml ofSOD. Incubations were terminated by removing the supernatants and pelleting any nonadherent neutrophils at 4#{176}C for 5 mm at 800g. The absorbance at 550 nm of the supernatants was determined and the nmoi of O2 calculated using an extinction coefficient for reduced minus oxidized cytochrome c of 2.1 x iO M’ cm [18].
Myeloperoxidase
Activity
Determination of myeioperoxidase release from neutrophiis was performed on supernatants from neutrophils incubated as for assay of HOC1 release, but without the addition of catalase to terminate the incubation. Instead, any remaining cells were removed by centrifugation at 4#{176}C and the supernatants kept on ice until myeioperoxidase activity was measured. Myeloperoxidase activity was assayed by a modification of the method of Andrews and Krinsky [2]. Briefly, a reaction mixture consisting of 0. 1 ml PBS, 0.01 ml Triton X-lOO (2.0% in PBS), 0.04 ml DMFA, 0.01 ml H,O, (15 mM), 0.2 ml sodium phosphate (200 mM, pH 5.4), and 0.04 ml TMB (20 mM in DMFA) was incubated for 5 mm at 37#{176}C. The reaction was initiated by the addition of 0. 1 ml of supernatant to the mixture. After incubation at 37#{176}C for an additional 5 mm, the reaction was stopped by the addition of 1 .75 ml of 200 mM sodium acetate, pH 3.0, and the mixtures were placed on ice. The amount of substrate conversion was determined by reading the absorbance at 655 nm of the reaction mixtures. These were compared with 100% control values determined by peiieting 106 neutrophiis, adding 1 ml of 0.2% Triton X-lOO, and sonicating the suspension for 1 mm. In the assay ofthis iysate, 0.01 ml of PBS was substituted for the Triton X- 100 in the reaction mixture.
Lactate
Dehydrogenase
Activity
Release of lactate dehydrogenase pernatants was determined according Dewald and Baggiolini [8].
Statistical
into neutrophii to the method
suof
Analysis
Results are expressed as the mean ± I SEM for the indicated number of experiments. Data were analyzed by Student’s one-tailed t test for paired samples except as noted (not significant: P > 0.05).
134
Test
TABLE
I
.
Effect
of rTNFa,
rT NFI3, and FMLP
HOC1
Treatment
PMN PMN PMN PMN PMN PMN
+ + + + +
rTNFa rTNF/3 FMLP rTNFa rTNFJ3
aNeutrophils When
terminated the assay bMean ± CNo
+ FMLP + FMLP
(PMN)
incubated
0.02 0.37 0.12 0.71 5.28 2.67
with
were
incubated
combinations
I h afterthe addition of HOCI generation. SEM
significant
of(n)
± ± ± ± ± ±
as described
Supe roxide,
02 1.27 12.56
0.01 (10)b 0.07* (10) 0.08c (3) 0.09** (10) 0.39** (10) 0.54*** (3) in Materials were
8.42 21.46 and Methods
exposedto
Myeloperoxidase(MPO)release
vs.
PMN
in buffer
MPO release (% of maximum)
(nmol) ± 0.29 (4) ± 333*** (4) ND ± 0.85* (4) ± l.51** (4) ND or with
I nM rTNFa,
2.7 ± 0.3 (5) 3.7 ± 0.9c (5) ND 3.8 ± 0.51I (5) 8.4 ± l.7********* ND I nM rTNF/3,
or $ for 10 mm priorto the addition ofFMLP. wasdetermined in supernatantsofincubation
or I0”
(5)
M FMLP.
Incubations mixturesused
were for
alone.
*P 0.005 vs. PMN alone by one-tailed t test **P 0.0005 vs. PMN alone. ***P ; 0.025 vs. PMN alone. ****P 0.01 vs. PMN alone. *****P 0.05 vs. PMN + rTNFa or vs. PMN
for
paired
+ FMLP
values.
by two-tailed
RESULTS Oxidant
rTNFa
Re lease by Neutrophils
experiments.
difference
Effect of TNFa,
and Myeloperoxidase
(nmol)
ofagonists,cells
ofFMLP.
on HOCI,
TNF13, and FMLP on Chlorinated
Generation
by Adherent
Neutrophils
Table 1 summarizes the results of experiments in which HOC1 generation was measured following incubation of neutrophils on FBS-coated polystyrene tissue culture wells for 1 h with the TNFs or FMLP alone or with FMLP added following a short incubation with one of the TNFs. Incubation with rTNFa, rTNF13, or FMLP alone resulted in the generation of very small, but significant, quantities of HOC1 compared with levels produced by neutrophils incubated in the absence of any stimulus. In order to rule out the possibility that greater quantities of HOCI might be generated by neutrophils exposed to these triggering agents following a lag period in either the respiratory burst or myeloperoxidase release which exceeded I h, incubations were extended for as long as 5 h without any further increase in the amount of HOCI detected (data not shown). However, if neutrophils were first incubated with either rTNFa or rTNF13 for 10 mm prior to the addition of FMLP, a marked increase in HOC1 generation by FMLP was seen. If the order of reagent addition was reversed and FMLP added before either rTNFa or rTNF13, HOC1 production was 0.5 nmol. Because the effect of rTNFa preincubation was about twice that seen with rTNF13 (7.4-fold vs. 3.8-fold increase in HOCI generation in response to FMLP), subsequent experiments focused on the characterization of chlorinated oxidant generation by the combination of rTNFa and FMLP. Consistent with results published for other stimuli [38], the detection of HOCI generation by neutrophils preincubated with rTNFa and triggered with FMLP re-
a’test for paired
values.
quired the presence of both H2O2 and functional myeloperoxidase, as the addition ofeither 50 ig/ml catalase (to catabolize H,O,) or I mM sodium azide (as a myeioperoxidase inhibitor) completely inhibited HOC1 generation (data not shown). Further, the addition of 1 mM methionine to .the system as a scavenger of HOCI prevented the detection of any HOC1. The time course of HOC1 generation by FMLPstimulated neutrophils preincubated with rTNFa is shown in Figure 1 HOC1 accumulated rapidly over the first 30 mm (6.00 ± 0.64 nmoi, 92.4% of total release) and reached a maximum of 6.49 ± 0. 13 nmol at 60 mm, remaining relatively constant thereafter. HOC1 generation as a function of neutrophil number was approximately linear up to 1 .0 X l0 neutrophils and reached a maximum of 6.60 ± 0.82 nmol with 2.0 X 106 neutrophils (Fig. 2). Figure 3 shows the relationship between the concentration of FMLP and the quantity of HOC! generated by neutrophils preincubated with differing concentrations of rTNFa ranging from 0-10 nM. FMLP alone induced the generation of only small amounts of HOCI by neutrophils with a maximum of 0.88 ± 0.02 nmol at an FMLP concentration of 10_b M. With FMLP concentrations of 10_s M or greater, preincubation with as little as 0. 1 nM rTNFa resulted in about a quadrupling of HOC! production. For example, at an FMLP concentration of l0 M, HOC! generation increased from 0.83 ± 0. 12 nmol with no other addition to 3.27 ± 0.59 nmol following preincubation with 0. 1 nmol rTNFa. The amount of HOC! trapped did not differ appreciably for concentrations of rTNFa of 1 and 10 nM, with maximal values of 6.04 ± I .03 and 6.37 ± 0.97 nmoi, respectively, at an FMLP concentration of l06 M. Although there was a relatively large increase in HOCI generation .
Generation
of Chlorinated
8.0
Oxidants by Adherent Neutrophils
135
10.0’ 8.0’
6.0 0
0
E
E
C
6.0
C
4.0
0
C)
0
0
x 2.0
4.0 2.0
0.0
0.0 0
15
30
45
Time
60
of incubation
75
90
105
0.0
120
(mm)
Fig. 1 Time course of HOCI generation by neutrophils preincubated with rTNFa prior to stimulation with FMLP. Neutrophils (i06/mI) were incubated at 37#{176}C as described in Materials and Methods. Recombinant TNFa (i nM) was added 10 mm prior to the addition of FMLP (i07 M). Incubations were continued at 37#{176}C for the indicated times and terminated by the addition of catalase (5 .tg). Supernatant HOCI (as the trapped N-chlorotaurine derivative) was then determined. Individual points represent the mean ± i SEM for three or four experiments. .
1 .0
2.0
Neutrophils
3.0
4.0
(millions)
Fig. 2. Relationship between cell number and HOCI generation by neutrophils preincubated with rTNFa prior to stimulation with FMLP. Neutrophils (0.25 x i06/ml-4.00 x 106/mI) were Incubated as described for Figure i except that incubations were terminated 60 mm after the addition of FMLP. Supernatant HOCI was assayed as described. Individual points represent the mean ± I SEM for three experiments.
8.00
with increasing FMLP concentration up to i0 M, there was only a further 1 nmol increase in HOC1 when the FMLP concentration was raised to lO_6 M (with the two highest concentrations of rTNFa). In contrast to the results observed when neutrophils were triggered under conditions permitting adherence, neutrophils (1 .0 x l06/ml) incubated in suspension (n = 4) failed to generate significant levels of HOC! when stimulated with buffer (0.0 ± 0.0 nmoi), rTNFa (0.04 ± 0.04 nmoi), or FMLP (0.04 ± 0.04 nmol) alone. These values differed significantly from those observed for adherent cells incubated with rTNFa (P 0.02 by two-tailed t test) or FMLP (P 0.001) alone. When neutrophils in suspension were incubated with rTNFa prior to the addition of FMLP, HOCI generation (1 .29 ± 0.62 nmol) was about one-quarter that observed for adherent neutrophils (P 0.001). Finally, we wished to determine whether adherent neutrophils triggered under conditions inducing optimal HOC! generation, i.e. with FMLP added following rTNFa, generated detectable quantities of endogenous N-C! derivatives. Indeed, in three experiments, 0.88 ± 0.20 nmo! of stable N-Cl derivatives accumulated in the supernatants of neutrophils (106/ml) stimulated in such a manner. This represented 18. 1 ± 3.2% of the HOCI generated by neutrophils in those experiments, a proportion similar to that previously reported for neutrophils stimulated with PMA or serum-opsonized zymosan [33]. The appearance of N-Cl derivatives in the supernatant was somewhat delayed compared with that of HOC!, ,
6.00 0
E
E.
4.00
0
0 I
2.00
0.00 i8
FMLP
concentration
10
(M)
Fig. 3. Relationship between FMLP concentration and HOCI generation by neutrophils preincubated with rTNFa prior to addition of FMLP. Neutrophils (i06/ml) were primed with buffer (.) or with various concentrations of rTNFa (0, 0.i nM; A i nM; L, iO nM)for 10 mm and then stimulated with FMLP (i09 M) for 60 mm as described in Materials and Methods. Supernatant HOCI was assayed as described. Error bars have been omitted for the sake of clarity. Individual points represent the mean of two experiments for neutrophils not primed with rTNFa and the mean of three or four experiments for cells primed with 0.i-io nM rTNFa. ,
-
with measurable levels first appearing Values reached -95% of maximum maximal values at 90 mm. Adherent bated with rTNFa or FMLP alone detectable quantities of N-Cl derivatives,
at 20-30 mm. by 60 mm and neutrophils incudid not generate nor did neutro-
136
Test
phils incubated the stimuli.
in suspension
with
any
combination
of
Comparison of O2, HOCI, and Myeloperoxidase Release by Adherent Neutrophils The quantity of HOC! generated by neutrophils following exposure to an agonist is a function both of the degree of respiratory burst activation and of the amount of myeioperoxidase released from primary granule stores. In order to determine which of these factors was responsible for the augmentation in HOC! production seen when rTNFa addition preceded that of FMLP, HOC1 generation was compared with 02 production and myeloperoxidase release (Table 1 ). At 1 h after addition of stimulus, the amount of 02 generated by either rTNFa or FMLP alone was close to or greater than twice the amount of HOC1 generated under any incubation conditions, even when rTNFa was added before FMLP. Assuming that one molecule of H202 arises from the dismutation of two molecules of O2 and that one molecule of H2O1 is used to generate one of HOC1, this suggested that the limiting factor in HOC1 generation with these stimuli was the amount of myeloperoxidase released, rather than the amount of H,O, available to react with the enzyme. Indeed, myeloperoxidase release from adherent neutrophils treated with both rTNFa and FMLP was more than twice that which occurred when the cells were stimulated with rTNFa or FMLP alone. In these experiments, there was no significant difference (by two-tailed t test for paired values, n 4) in lactate dehydrogenase release by neutrophi!s treated with rTNFa (5.4 ± 0.9%),FMLP(4.8 ± 0.7%),orrTNFa + FMLP (6.9 ± 0.8%). Although the amount of myeloperoxidase released into the supernatant by rTNFa alone did not differ significantly from that released from neutrophils incubated in buffer alone, it is likely that a small increase in myeioperoxidase release did in fact occur under these conditions, as HOC1 generation induced by either rTNFa (or FMLP) alone was slightly and significantly greater than for neutrophils in buffer only (Table 1). Although the quantity of mye!operoxidase released from granule stores appeared to be the preeminent factor in determining total HOC! production in these experiments, the kinetics of respiratory burst activation could still influence the characteristics of HOC1 generation by adherent neutrophi!s. In fact, the rate of 02 release was markedly accelerated when neutrophils were preincubated with rTNFa prior to triggering with FMLP (Fig. 4). Under these conditions, there was no evidence of an extended lag period in respiratory burst activation. The overall appearance of the time course of O generation by neutrophils stimulated with FMLP after addition of rTNFa was remarkably similar to that for HOC! production seen in Figure 1. This suggested that myeloperoxi-
0
E C
a V x
0 4)
a. Cl)
0
15
30
Time
45
60
75
of mncubation
90
105
120
(mm)
Fig. 4. Time course of superoxide generation by neutrophils incubated with rTNFa, FMLP, or rTNFa + FMLP. Neutrophils (i06/mI) were incubated at 37#{176}C in the presence of cytochrome C (2 mg/mI) with rTNFa (1 nM), or with FMLP (i0 M) alone or added after a iO mm priming incubation with rTNFa (i nM) After the indicated times, supernatants were removed and the reaction was terminated by centrifugation at 4#{176}C to remove any nonadherent cells. Superoxide generation was measured as described in Materials and Methods from the time of addition of rTNFa. FMLP was added at t iO mm. The curves shown are from a representative experiment of two performed. =
dase re!ease occurred rapidly enough to allow HOC! generation to closely follow O2 production. Indeed, 73.4 ± 4.3% (n = 2) of the amount of mye!operoxidase which would ultimately be released by 60 mm of incubation had been released by 15 mm after the addition of FMLP to rTNFa-preincubated cells.
DISCUSSION The ability of neutrophils adherent to surfaces coated with serum or extracellular matrix proteins to generate large quantities of H2O2 in response to activation with rTNFa, rTNF13, or FMLP [20] implies a mechanism whereby these agents might lead to biologically significant tissue damage during inflammation in vivo. Although H,O2 is capable of inducing cell injury by itself, the reaction of H2O, with myeioperoxidase in the presence of C1 leads to the generation of HOCI, a much more powerful and reactive oxidant [34]. Because of its high reactivity, HOC1 is relatively short-lived in biologic systems. However, the oxidizing equivalents of HOC1 can be preserved in the form of endogenous N-C! derivatives, long-lived oxidants which result from the reaction of HOC1 with amine-containing compounds released during the activation of neutrophils [ 1 1,26,3335,38,39]. The production of chlorinated oxidants by a neutrophil would require not only activation of the respiratory burst, but also the coordinated release of
Generation
myeloperoxidase from the primary granules. In the present study, we have shown that neutrophils incubated with rTNFa, rTNF13, or FMLP on polystyrene surfaces precoated with FBS generated only small quantities of HOC1. Further characterization of the activation process for rTNFa and FMLP demonstrated that minimal HOC! production occurred despite sufficient activation of the respiratory burst to potentially result in from 10-20 times the amount of HOC1 that was actually observed. The relative lack of HOC! generation by adherent neutrophils in response to these soluble triggers was subsequently shown to be a function of their ability to induce release of only a small amount of myeloperoxidase from granule stores. In contrast, neutrophils which were first incubated with either rTNFa or rTNF13 for 10 mm before the addition of FMLP showed a marked enhancement of HOC! generation. One hour after the addition of FMLP to 106 neutrophils which had been preincubated with rTNFa, 5.28 ± 0.39 nmol of HOC! had been released into the supernatant, an amount approximately twice that seen when rTNF13 was used in a similar manner. In addition, 1 8% of the HOC! generated was consumed in the formation of long-lived N-C! derivatives, a proportion similar to the 10-15% seen for neutrophils stimuiated with PMA or serum-opsonized zymosan [33] The increase in HOC! which resulted from the use of rTNFa and FMLP together in the neutrophi! incubations was accompanied by a significant increase in mye!operoxidase release into the supernatant. Although mye!operoxidase present in the cell supernatant might not represent all of the enzyme released from granule stores (e.g., some enzyme might remain adherent to the cell membrane), the !evels measured in the present experiments correlated well with the HOC! measurements and thus seemed to provide a reliable index of total mye!operoxidase release The rapid release of myeloperoxidase when rTNFa addition preceded FMLP triggering, coupled with enhanced respiratory burst kinetics (see below) led to rapid accumulation of HOC! in the medium. By 15 mm after the addition of FMLP, 73% of total mye!operoxidase release into the supernatant and 80% of total HOC! accumulation had been achieved (Fig 1 ). The concentrations of rTNFa (1 nM) and FMLP (l0 M) used for the majority of these experiments proved to be nearoptima! (Fig. 3), although enhanced HOC! generation was seen with a priming concentration of rTNFa as low as0.l nM. The failure of rTNFa to induce HOC1 generation by neutrophiis when used as a trigger, but to markedly enhance HOC! release when used as a priming agent for another agonist is consistent with a previously published report in which rTNFa was used by itself or with unopsonized zymosan to stimulate neutrophils in suspension [29]. That study differed from the present one in that -.
.
.
.
of Chlorinated
Oxidants
by Adherent
Neutrophils
137
the neutrophils were incubated in suspension with continuous stirring, the stimulus (unopsonized zymosan) was nonphysiologic and could be internalized by phagocytosis, and the mechanism of enhancement of HOC! was not determined [29] In another study of neutrophils activated with rTNFa and unopsonized zymosan, the cytokine was shown to significantly increase iodination, mye!operoxidase release, and H,O, generation stimulated by zymosan [13]. In both studies, rTNFa enhanced phagocytosis of the unopsonized zymosan, and this was considered a possible, if not likely, explanation for the observed effects on cell function [13,29]. Clearly, enhanced phagocytosis would not explain the effect seen in our system, which involved another soluble stimulus, FMLP, in addition to rTNFa. Nonetheless, a similar increase in myeloperoxidase release occurred when neutrophi!s were exposed to rTNFa before the addition of FMLP. In the only other study for which data are available on myeloperoxidase release in response to these stimuli, rTNFa priming did not cause an increase in extracellular activity of the enzyme in response to FMLP, but the neutrophils were in suspension rather than adherent to a surface [6]. Thus, adherence to a “biologic” surface may be the critical factor in inducing the enhanced mye!operoxidase release which leads to the HOC! response seen when a soluble agonist (FMLP) is used in combination with either TNFa or 13. The data in Table 1 and Figure 4 demonstrate that although total O2 generation (measured at 1 or 2 h) appeared to be additive for adherent neutrophils treated with both agonists vs. either alone, it is clear that rTNFa was behaving as a “priming” agent for FMLP in that the initial rate of O2 generation was markedly accelerated when cells were incubated with rTNFa for 10 mm prior to the addition of FMLP (Fig. 4). Indeed, total O2 accumulation at 15 or 30 mm suggested a synergistic effect from the use of both agents compared with either alone (Fig. 4). Under optimal incubation conditions in the present study, neutrophils were able to generate quantities of HOC! far below their demonstrated capacity. The 5-7.5 nmol of HOC! generated per 106 neutrophils is a small fraction of the potential release of 100-I 50 nmol reported for neutrophils triggered with PMA or serumopsonized zymosan [34,38] Can these amounts of HOC1 mediate important biologic effects? On the basis of published data, low concentrations of HOC1 clearly can induce both lethal and sublethal effects on target cells. The amounts of HOC! released by adherent neutrophils triggered with rTNFa and FMLP could inhibit cell division of -3.9 x i07 Escherichia (‘0/i or induce lysis of -l0 tumor cells [calculated on the basis of results published in ref. 1 and 30]. Recently, similar concentrations of HOC! were also shown to mediate important .
--
.
138
Test
sublethal effects on a murine tumor ce!l line [27]. Concentrations of HOC! as low as 10 pM could disturb plasma membrane functions including uptake of glucose and aminoisobutyric acid, maintenance of intracellular potassium, and maintenance of cell volume [27]. Thus, neutrophils exposed to the soluble agonists used in study, all of which could be encountered in an inflammatory focus, would be able to induce both cytotoxic and microbicida! injury. Neutrophils migrating into a site of inflammation would be more likely to encounter a complex mixture of cytokines and inflammatory mediators rather than any single mediator by itself. We have shown that two of these cytokines, TNFa or 13, in combination with FMLP, a chemotactic peptide similar or identical to N-formylmethionyl peptides released from bacteria [17] and eukaryotic mitochondria [7], can lead to the rapid activation of human peripheral blood neutrophils and that this is accompanied by the release of potentially cytotoxic concentrations of HOC! along with longer-lived N-Cl derivatives. Future investigations aimed at more precise!y defining the composition of the extracellular fluids in inflammatory sites should aid in determining which of the many potential neutrophil-mediator interactions might be appropriate targets for attempts at therapeutic intervention.
ACKNOWLEDGMENTS I thank Dr. B. Lubin for encouraging these studies and for reviewing the manuscript. This work was supported by National Institutes of Health grants DK-32094 and HL-20985.
REFERENCES I . Albrich, J.M. , and Hurst, J.K. Oxidative inactivation of Escherichia coli by hypochlorous acid. Rates and differentiation of respiratory from other reaction sites. FEBS Lett. 144, 157, 1982. 2. Andrews, P.C. , and Krinsky, NI. Human myeloperoxidase and hemi-myeloperoxidase. Methods Enzymol. 132, 369, 1986. 3. Atkinson, Y.H. , Marasco, WA. , Lopez, A.F. , and Vadas, MA. Recombinant human tumor necrosis factor-a. Regulation of N-formylmethionylleucylphenylalanine receptor affinity and function on human neutrophils. J. Clin. Invest. 81, 759, 1988. 4. Babior, B.M. Oxidants from phagocytes: agents of defense and destruction. Blood 64, 959, 1984. 5. Babior. B.M. , Kipnes, R.S. , and Curnutte, iT. Biological defense mechanisms. The production by leukocytes of superoxide, a potential bactericidal agent. J. Clin. Invest. 52, 741, 1973. 6. Berkow, R.L. , Wang, D. , Larrick, 1W. , Dodson, R.W. , and Howard, T.H. Enhancement of neutrophil superoxide production by preincubation with recombinant human tumor necrosis factor. J. Immunol. 139, 3783, 1987. 7. Carp, H. Mitochondrial N-formylmethionyl proteins as chemoattractants for neutrophils. J. Exp. Med. 155, 264, 1982. 8. Dewald, B., and Baggiolini, M. Methods for assessing exocytosis by neutrophil leukocytes. Methods Enzymol. 132, 267, 1986.
9. Foote, CS. , Goyne, I.E. , and Lehrer, RI. Assessment of chlorination by human neutrophils. Nature 301 , 715, 1983. 10. Granger, GA. , and Williams, T.W. Lymphocyte cytotoxicity in vitro: activation and release of a cytotoxic factor. Nature 218, 1253, 1968. 1 1 . Grisham, MB. , Jefferson, MM. , Melton, D.F. , and Thomas, EL. Chlorination of endogenous amines by isolated neutrophils. Ammonia-dependent bactericidal, cytotoxic, and cytolytic activities of the chloramines. I. Biol. Chem. 259, 10404, 1984. 12. Kapp, A. , Zeck-Kapp, G. , and Blohm, D. Human tumor necrosis factor is a potent activator of the oxidative metabolism in human polymorphonuclear neutrophilic granulocytes: comparison with human lymphotoxin. J. Invest. Dermatol. 92, 348, 1989. 13. Kiebanoff, Si. , Vadas, MA., Harlan, J.M. , Sparks, L.H., Gamble, JR. , Agosti, J.M. , and Waltersdorph, AM. Stimulation of neutrophils by tumor necrosis factor. J. Immunol. 136, 4220, 1986. 14. Kownatzki, E. , Kapp, A. , and Uhrich, S. Modulation of human neutrophilic granulocyte functions by recombinant human tumor necrosis factor and recombinant human lymphotoxin. Promotion of adherence, inhibition of chemotactic migration and superoxide anion release from adherent cells. Clin. Exp. Immunol. 74, 143, 1988. 15. Larrick, J.W. , Graham, D. , Toy, K. , Lin, L.S. , Senyk, G., and Fendly, B.M. Recombinant tumor necrosis factor causes activation of human granulocytes. Blood 69, 640, 1987. 16. Laudanna, C. , Miron, S. , Berton, G. , and Rossi, F. Tumor necrosis factor a/cachectin activates the 02 -generating system of human neutrophils independently of the hydrolysis of phosphoinositides and the release of arachidonic acid. Biochem. Biophys. Res. Commun. 166, 308, 1990. 17. Marasco, WA. , Phan, S.H. , Krutzsch, H. , Showell, Hi., Feltner, D.E. , Nairn, R. , Becker, EL. , and Ward, PA. Purification and identification of formyl-methionyl-leucyl-phenylalanine as the major peptide neutrophil chemotactic factor produced by Escherichia co/i. J. Biol. Chem. 259, 5430, 1984. 18. Massey, V. The microestimation of succinate and the extinction coefficient of cytochrome c. Biochim. Biophys. Acta 34, 255, 1959. 19. Meurer, R., and Maclntyre, D.E. Lack ofeffect ofpertussis toxin on TNF-a-induced formation of reactive oxygen intermediates by human neutrophils. Biochem. Biophys. Res. Commun. 159, 763, 1989. 20. Nathan, CF. Neutrophil activation on biological surfaces. Massive secretion of hydrogen peroxide in response to products of macrophages and lymphocytes. J. Clin. Invest. 80, 1550, 1987. 21 . Nathan, CF. Respiratory burst in adherent human neutrophils: triggering by colony-stimulating factors CSF-GM and CSF-G. Blood 73, 301, 1989. 22. Nedwin, G.E. , Svedersky, L.P. , Bringman, T.S. , Palladinno, MA. , Jr. , and Goeddel, DV. Effect of interleukin 2, interferon‘1 and mitogens on the production of tumor necrosis factors a and 23.
24.
25.
13. J. Immunol. 135, 2492, 1985. Neumann, M. , and Kownatzki, E. The effect of adherence on the generation of reactive oxygen species by human neutrophilic granulocytes. Agents Actions 26, 183, 1989. Pennica, D. , Nedwin, G.E. , Hayflick, iS. , Seeburg, PH., Derynck, R. , Palladino, MA. , Kohr, Wi. , Aggarwal, B.B. , and Goeddel, DV. Human tumor necrosis factor: precursor structure, expression and homology to lymphotoxin. Nature 312, 724, 1984. Perussia, B. , Kobayashi, M. , Rossi, ME. , Anegon, I. , and Trinchieri, G. Immune interferon enhances functional properties of human granulocytes: role of Fe receptors and effect of lymphotoxin , tumor necrosis factor, and granulocyte-macrophage colony-stimulating factor. J. Immunol. 138, 765, 1987.
Generation of Chlorinated Oxidants by Adherent 26. Sagone,
27.
28.
29.
30.
31 32.
33.
.
AL. , Jr. , Husney, R.M. . O’Dorisio, MS. , and Metz. EN. Mechanisms for the oxidation of reduced glutathione by stimulated neutrophils. Blood 63, 96, 1984. Schraufst#{228}tter, lU. , Browne, K. , Harris, A. , Hyslop, PA., Jackson, J.H. , Quehenberger, 0. , and Cochrane, C.G. Mechanisms of hypochlorite injury of target cells. J. Clin. Invest. 85, 554, 1990. Shalaby, MR. , Palladino, MA. , Jr. , Hirabayashi, SE. , Eessalu, T.E. , Lewis, GD. , Shepard, H.M. , and Aggarwal, B.B. Receptor binding and activation of polymorphonuclear neutrophils by tumor necrosis factor-alpha. J. Leukoc. Biol. 41 , 196, 1987. She, Z. , Wewers, M.D. , Herzyk, D.J. , Sagone, AL. , and Davis, W.B. Tumor necrosis factor primes neutrophils for hypochlorous acid production. Am. J. Physiol. 257. L338, 1989. Slivka, A. , LoBuglio, A.F. , and Weiss, S.J. A potential role for hypochlorous acid in granulocyte-mediated tumor cell cytotoxicity. Blood 55, 347, 1980. Steinbeck, AM. , and Roth, J.A. Neutrophil activation by recombinant cytokines. Rev. Infect. Dis. I 1 , 549, 1989. Tennenberg, S.D. , and Solomkin, J.S. Activation of neutrophils by cachectin/tumor necrosis factor: priming of N-formyl-methionyl-leucyl-phenylalanine-induced oxidative responsiveness via receptor mobilization without degranulation. J. Leukoc. Biol. 47, 217, 1990. Test, ST., Lampert. MB., Ossanna, P.J., Thoene, J.G., and Weiss, S.J. Generation of nitrogen-chlorine oxidants by human phagocytes. J. Clin. Invest. 74, 1341, 1984.
34.
Neutrophils
139
Test. ST. , and Weiss, S.J. The generation and utilization of chlorinated oxidants by human neutrophils. Adv. Free Radic. Biol. Med. 2, 91, 1986. 35. Thomas, EL. . Grisham, MB. , and Jefferson, MM. Myeloperoxidase-dependent effect of amines on functions of isolated neutrophils. J. Clin. Invest. 72, 441, 1983. 36. Tsujimoto. M. , Yokota, S. , Vilcek, J. . and Weissmann, G. Tumor necrosis factor provokes superoxide anion generation from neutrophils. Biochem. Biophys. Res. Commun. 137, 1094, 1986. 37. Urban, J.L. , Shepard, H.M. , Rothstein, J.L. , Sugarman, B.J., and Schreiber, H. Tumor necrosis factor: a potent effector molecule for tumor cell killing by activated macrophages. Proc. Natl. Acad. Sci. USA 83, 5233, 1986. 38. Weiss, S.J. , Klein, R. . Slivka, A. , and Wei, M. Chlorination of taurine by human neutrophils. Evidence for hypochlorous acid generation. J. Clin. Invest. 70, 598, 1982. 39. Weiss, S.J. , Lampert. MB. . and Test, ST. Long-lived oxidants generated by human neutrophils: characterization and bioactivity. Science 222, 625. 1983. 40. Weiss, S.J. Oxygen as a weapon in the phagocyte armamentarium. In Handbook of Inflammation: Immunology of Inflammation, vol. 4, (Ward, PA. , Ed.). Amsterdam: Elsevier Science Publishers, p. 37, 1983. 41. Yuo, A., Kitagawa, S., Suzuki, I., Urabe, A.. Okabe, T., Saito, M. , and Takaku, F. Tumor necrosis factor as an activator of human granulocytes. Potentiation of the metabolisms triggered by the Ca2-mobilizing agonists. J. Immunol. 142, 1678, 1989.
of Leukocyte
Biology
50:131-139
(1991)
Effect of Tumor Necrosis Factor on the Generation of Chlorinated Oxidants by Adherent Human Neutrophils Samuel Children’s
Hospital
Oakland
T. Test
Research
Institute,
Oakland,
California
Human neutrophils adherent to simulated biologic surfaces undergo significant activation of the respiratory burst over prolonged periods of time in response to stimulation with the cytokines tumor necrosis factor-alpha (TNFa) or tumor necrosis factor-beta (TNF3) or with the chemotactic peptide N-formyl-methionylleucylphenylalanine(FMLP). In this study, neutrophils were examined for their ability to generate the highly reactive and powerful oxidant hypochlorous acid (HOCI) and the longer-lived, less reactive endogenous nitrogen-chlorine (N-Cl) derivatives in response to these stimuli either alone or when exposed to recombinant human TNFa (rTNFa) or 3 (rTNF3) prior to addition of FMLP. Neutrophils adherent to fetal bovine serum-coated polystyrene tissue culture wells were able to generate only small quantities of HOCI when incubated with rTNFa, rTNF3, or FMLP individually. However, when neutrophils were first incubated with either rTNFa or rTNF3 prior to addition of FMLP, there was a marked increase in HOCI generation. Neutrophils stimulated in such a manner consumed -i8% of the HOCI generated in the formation of N-Cl derivatives. Further scrutiny of the response to the combination of rTNFa and FMLP revealed that HOCI release was rapid, with 80% of total HOCI accumulation occurring within i5 mm after FMLP addition. The amount of HOCI generated was dependent on the number of cells added and on the concentration of both rTNFa and FMLP. Comparison of HOCI generation with superoxide anion and myeloperoxidase release showed that the amount of HOCI generated was limited primarily by the amount of myeloperoxidase released rather than by the degree of respiratory burst activation. These results demonstrate that human neutrophils stimulated with FMLP after a brief incubation with rTNFa or rTNF3 can generate cytotoxic and microbicidal concentrations of chlorinated oxidants. Key words:
N-formyl-Met-Leu-Phe, rine derivatives
hypochlorous
INTRODUCTION Neutrophils entering an inflammatory focus could potentially encounter a diverse array of soluble mediators capable of triggering the release of toxic granule constituents and/or the generation of a variety of destructive oxygen metabolites. Recent attention has focused on the ability of cytokines released from activated macrophages or T-lymphocytes to stimulate these cytotoxic responses in human neutrophils [reviewed in 3 1]. Tumor necrosis factor-alpha (TNFa), which is released from macrophages exposed to lipopolysaccharide (LPS) [37], is capable of affecting a variety of cellular functions and has been examined for its effects on neutrophils in a number of studies [31]. Tumor necrosis factor-beta (TNF13), a product of activated T cells [10], shares structural features with and has biologic activity similar to that of TNFa [24], but has been used less extensively in studies of neutrophil function [3 1] Both of these cytokines can also be released following the exposure of peripheral blood mononuclear cells to mitogens and other cytokines, including interleukin-2 alone or with interferon-’y [22].
acid, myeloperoxidase,
nitrogen-chlo-
It is becoming increasingly apparent that the microenvironment in which the human neutrophil undergoes stimulation can profoundly influence its response to soluble agonists. For example, it has recently been demonstrated that the triggering of neutrophils adherent to simulated biologic surfaces compared with neutrophils in suspension resulted in a marked enhancement of hydrogen peroxide (H202) generation following an initial lag period of 15-90 mm [20,2 1 ] Respiratory burst activation having these features was characteristic of the neutrophil response to recombinant human tumor necrosis factor a (rTNFa) and 13 (rTNFf3) [20], to the chemotactic peptide N-formyl-methionylleucylphenylalanine (FMLP) [20], and to granulocyte/macrophageand granulocyte-colony stimulating factors [2 1 ] However, .
.
.
© 1991
Wiley-Liss,
Inc.
Received
September
Reprint requests: search Institute.
1 1. 1990:
accepted
December
18,
1990.
Samuel T. Test. Children’s Hospital Oakland 747 Fifty Second Street, Oakland. CA 94609.
Re-
132
Test
the effect of neutrophil adherence to simulated biologic surfaces on the generation of oxygen metabolites which derive from the reaction of H2O2 with other neutrophil constituents has not been examined. The most potent and reactive of these oxidants are the hydroxyl radical (OH ) which is generated by the reaction of superoxide anion (O2) with H,O2, and hypochlorous acid (HOCI), which is the end-product of the H2O2-myeloperoxidasechloride system [4,40]. Whereas unquestionable proof of OH generation by neutrophils does not yet exist [4,40], myeloperoxidase is the primary constituent of the neutrophil primary (azurophil) granule and has been shown to participate in the generation of both HOC1 [9,34,35,38] and a class of long-lived oxidants known as nitrogen-chlorine (N-Cl) derivatives [ 1 1 ,26,33-35, 38,39]. These two species of chlorinated oxidants are capable of mediating diverse biologic effects, including alterations in serum and cellular proteins which can result in functional changes or cell lysis [34]. TNFa and 13have been shown to be weak triggers for O2 or H2O2 production by neutrophils in suspension [3,6, 13-16, 19,23,25,28,36,41]. When used as priming agents for neutrophils in suspension, the TNFs can enhance 02 generation in response to FMLP [3,6, 12,32,41] or unopsonized zymosan [13] and can enhance myeloperoxidase-dependent iodination [13] or HOCI generation [29] in response to the latter triggering agent. Enhanced myeloperoxidase release by FMLP-stimulated neutrophils following rTNFa priming was not observed [6]. In light of the profound increase in respiratory burst activation which occurred when neutrophils were stimulated with the TNFs or FMLP on serum-coated polystyrene surfaces [20], we were interested whether the increase in H,O, generation was accompanied by a sufficient enhancement of myeloperoxidase release to result in an increase in HOCI generation of similar magnitude. In the experiments detailed below, we show that neutrophils adherent to serum-coated surfaces generate only negligible quantities of HOC1 when triggered with rTNFa, rTNF13, or FMLP alone and that this is likely due to minimal release of myeloperoxidase. However, a marked increase in HOC1 generation occurs when neutrophils are first incubated with rTNFa or rTNF13 prior to the addition of FMLP. This is accompanied by a significant increase in the amount of myeloperoxidase released into the extracellular medium. Furthermore, the absence of any lag period in respiratory burst activation when neutrophils are primed with rTNFa before triggering with FMLP (compared with incubation with these agonists alone), coupled with the rapid release of myeloperoxidase, results in the rapid attainment of maximal HOC1 levels over a time course approximating that for O2 production. .
.
MATERIALS Materials rTNFa
AND METHODS
was kindly
supplied
by Genentech,
Inc.
(South
San Francisco, CA). The preparation, as assayed by the supplier, had a specific activity of 5.6 X l0 U/mg and a concentration of 0.5 mg/mi in phosphate-buffered saline (PBS) and was stored at 4#{176}C. Working dilutions of 40 nM in Dulbecco’s PBS with Ca2 and Mg2 (UCSF Cell Culture Facility, San Francisco, CA), pH 7.4, supplemented with 10 mM glucose (PBSG) were prepared daily immediately before addition to incubation mixtures. rTNF13, also provided by Genentech, Inc. had a specific activity of 2 1 X i0 U/mg and a concentration ofO. 1 mg/mi. Working dilutions of2O nM in PBSG were prepared immediately before addition to incubation mixtures. FMLP (Sigma Chemical Co., St. Louis, MO) was stored at -20#{176}Cas a 3 x l02 M solution in dimethyisuifoxide (DMSO). Stock solutions of lO_6 M in PBSG were prepared daily immediately prior to use. Fetal bovine serum (FBS) was purchased from Hyclone Laboratories, Inc. (Logan, UT) and heat-inactivated at 56#{176}C for 30 mm prior to use. Other reagents used included cytochrome c (horse heart type VI), superoxide dismutase (SOD, from bovine erythrocytes), catalase (from bovine liver) 3 ,3 ,5 ,5 ‘-tetramethylbenzidine (TMB) sodium acetate, NADH, sodium pyruvate, disodium ethyldiaminetetraacetic acid (EDTA), taurine, sodium azide, L-methionine (Sigma); hydrogen peroxide (30%, ACS certified), N ,N-dimethylformamide (DMFA) potassium iodide (Fisher Scientific, Fair Lawn, NJ); PBS (Ca2 and Mg2tfree, I X and lOX concentrate), pyrogen-free water (UCSF Cell Culture Facility); Percoil (Pharmacia LKB Biotechnology AB, Uppsala, Sweden); and Triton X-iOO (Bio-Rad Laboratories, Richmond, CA). ,
.
,
‘
,
,
Isolation
of Polymorphonuclear
Neutrophils
Neutrophils were prepared by centrifugation of whole blood on discontinuous Percoil gradients. Isotonic stock solutions of Percoli ( 100% Percoll stock) were prepared by adding 1 part lOX PBS to 9 parts Percoll. The Percoli stock solution was subsequently diluted to 60% and 77% with 1 x PBS: these solutions were allowed to come to room temperature before use. Fifteen milliliters of 77% Percoli were layered under 15 ml of6O% Percoli in sterile 50 ml conical bottom polypropylene tubes (Applied Scientific, San Francisco, CA). Fifteen to 20 ml of fresh whole blood anticoagulated with sterile 1 .5% Na2EDTA was obtained by peripheral venipuncture from normal human volunteers, carefully layered over the Percoli and subsequently centrifuged at 500g for 30 mm at room temperature. The neutrophil layer was aspirated, contammating red blood cells (RBC) were removed by hypotonic lysis, and the remaining cells washed with PBS and
Generation
suspended to 2 X 106 cells/mI in PBSG. preparation was >96% neutrophils by Wright’s viability with >98% by trypan blue exclusion.
Endotoxin
The final stain and
Contamination
The endotoxin concentration of the lot of FBS used in these experiments was assayed by the manufacturer (Hycione Laboratories, Inc.) and determined to be 25 pg/ml. The endotoxin content of other reagents was quantitated with a chromogenic limulus amebocyte lysate assay (Whittaker Bioproducts, Inc. Walkersville, MD) with a lower limit of detection of 10 pg/ml. Endotoxin levels were undetectable in the 1 X and lOx PBS used for the preparation of Percoll dilutions, in the PBSG used for neutrophil suspensions and dilution of reagents, and in the working dilutions of rTNFa, rTNF13, and FMLP. The endotoxin content of Percoll could not be accurately determined because this substance interfered with the endotoxin assay. ,
HOCI Generation HOC1 generation was assayed in a system using neutrophils adherent to FBS-coated plastic tissue culture wells. The wells (4.0 cm2) of 12 well tissue culture plates (Costar, Cambridge, MA) were covered with 0.3 ml heat-activated FBS and incubated for 1-2 h at 37#{176}C in a humidified chamber. The FBS was then aspirated and the wells washed three times with 0.9% saline prepared with pyrogen-free water. Neutrophils (i06/well) were added and incubated for 10 mm at 37#{176}C prior to the addition of either rTNFa or rTNF13. The incubation mixtures were incubated for an additional 10 mm and FMLP was carefully added dropwise. The priming period of 10 mm was chosen on the basis of the optimal priming duration previously reported for respiratory burst activation by FMLP following TNFa priming [3,6,41]. Taurine (15 mM) was present in all incubation mixtures (final volume 1 ml) to scavenge HOCI and permit its subsequent quantification by the iodide oxidation method [38]. When accumulation of endogenous N-Cl derivatives was to be measured, taurine was omitted. Incubations were terminated by the addition of catalase (5 pUg), any nonadherent cells were pelleted by centrifugation, and the supernatants were assayed for HOC1 (as the stable N-chlorotaurine derivative) or endogenous N-Cl derivatives. In some experiments, neutrophils were added to buffer in 12 x 75 mm polypropylene tubes and incubated at 37#{176}C for 10 mm before the addition of rTNFa The tubes were then placed on a rocking platform to keep the cells in suspension for the duration of the incubation. FMLP was added after 10 mm and the incubation continued for 60 mm and terminated as described above. The cells were pelieted and supernatants assayed for HOC1 or N-Cl derivatives as described. .
of Chlorinated
Oxidants
Superoxide
by Adherent
Neutrophils
133
Generation
Activation of the neutrophil respiratory burst was quantitated by measuring the SOD-inhibitable reduction ofcytochrome c [5]. Incubation conditions were identical to those used to measure HOC1 generation except that cytochrome c (2 mg/mi) was substituted for taurine, and catalase ( 10 p.g/ml) was added to all wells to prevent the reoxidation of cytochrome c by any H2O, generated during the course of the incubation. Duplicate wells contained 50 ig/ml ofSOD. Incubations were terminated by removing the supernatants and pelleting any nonadherent neutrophils at 4#{176}C for 5 mm at 800g. The absorbance at 550 nm of the supernatants was determined and the nmoi of O2 calculated using an extinction coefficient for reduced minus oxidized cytochrome c of 2.1 x iO M’ cm [18].
Myeloperoxidase
Activity
Determination of myeioperoxidase release from neutrophiis was performed on supernatants from neutrophils incubated as for assay of HOC1 release, but without the addition of catalase to terminate the incubation. Instead, any remaining cells were removed by centrifugation at 4#{176}C and the supernatants kept on ice until myeioperoxidase activity was measured. Myeloperoxidase activity was assayed by a modification of the method of Andrews and Krinsky [2]. Briefly, a reaction mixture consisting of 0. 1 ml PBS, 0.01 ml Triton X-lOO (2.0% in PBS), 0.04 ml DMFA, 0.01 ml H,O, (15 mM), 0.2 ml sodium phosphate (200 mM, pH 5.4), and 0.04 ml TMB (20 mM in DMFA) was incubated for 5 mm at 37#{176}C. The reaction was initiated by the addition of 0. 1 ml of supernatant to the mixture. After incubation at 37#{176}C for an additional 5 mm, the reaction was stopped by the addition of 1 .75 ml of 200 mM sodium acetate, pH 3.0, and the mixtures were placed on ice. The amount of substrate conversion was determined by reading the absorbance at 655 nm of the reaction mixtures. These were compared with 100% control values determined by peiieting 106 neutrophiis, adding 1 ml of 0.2% Triton X-lOO, and sonicating the suspension for 1 mm. In the assay ofthis iysate, 0.01 ml of PBS was substituted for the Triton X- 100 in the reaction mixture.
Lactate
Dehydrogenase
Activity
Release of lactate dehydrogenase pernatants was determined according Dewald and Baggiolini [8].
Statistical
into neutrophii to the method
suof
Analysis
Results are expressed as the mean ± I SEM for the indicated number of experiments. Data were analyzed by Student’s one-tailed t test for paired samples except as noted (not significant: P > 0.05).
134
Test
TABLE
I
.
Effect
of rTNFa,
rT NFI3, and FMLP
HOC1
Treatment
PMN PMN PMN PMN PMN PMN
+ + + + +
rTNFa rTNF/3 FMLP rTNFa rTNFJ3
aNeutrophils When
terminated the assay bMean ± CNo
+ FMLP + FMLP
(PMN)
incubated
0.02 0.37 0.12 0.71 5.28 2.67
with
were
incubated
combinations
I h afterthe addition of HOCI generation. SEM
significant
of(n)
± ± ± ± ± ±
as described
Supe roxide,
02 1.27 12.56
0.01 (10)b 0.07* (10) 0.08c (3) 0.09** (10) 0.39** (10) 0.54*** (3) in Materials were
8.42 21.46 and Methods
exposedto
Myeloperoxidase(MPO)release
vs.
PMN
in buffer
MPO release (% of maximum)
(nmol) ± 0.29 (4) ± 333*** (4) ND ± 0.85* (4) ± l.51** (4) ND or with
I nM rTNFa,
2.7 ± 0.3 (5) 3.7 ± 0.9c (5) ND 3.8 ± 0.51I (5) 8.4 ± l.7********* ND I nM rTNF/3,
or $ for 10 mm priorto the addition ofFMLP. wasdetermined in supernatantsofincubation
or I0”
(5)
M FMLP.
Incubations mixturesused
were for
alone.
*P 0.005 vs. PMN alone by one-tailed t test **P 0.0005 vs. PMN alone. ***P ; 0.025 vs. PMN alone. ****P 0.01 vs. PMN alone. *****P 0.05 vs. PMN + rTNFa or vs. PMN
for
paired
+ FMLP
values.
by two-tailed
RESULTS Oxidant
rTNFa
Re lease by Neutrophils
experiments.
difference
Effect of TNFa,
and Myeloperoxidase
(nmol)
ofagonists,cells
ofFMLP.
on HOCI,
TNF13, and FMLP on Chlorinated
Generation
by Adherent
Neutrophils
Table 1 summarizes the results of experiments in which HOC1 generation was measured following incubation of neutrophils on FBS-coated polystyrene tissue culture wells for 1 h with the TNFs or FMLP alone or with FMLP added following a short incubation with one of the TNFs. Incubation with rTNFa, rTNF13, or FMLP alone resulted in the generation of very small, but significant, quantities of HOC1 compared with levels produced by neutrophils incubated in the absence of any stimulus. In order to rule out the possibility that greater quantities of HOCI might be generated by neutrophils exposed to these triggering agents following a lag period in either the respiratory burst or myeloperoxidase release which exceeded I h, incubations were extended for as long as 5 h without any further increase in the amount of HOCI detected (data not shown). However, if neutrophils were first incubated with either rTNFa or rTNF13 for 10 mm prior to the addition of FMLP, a marked increase in HOC1 generation by FMLP was seen. If the order of reagent addition was reversed and FMLP added before either rTNFa or rTNF13, HOC1 production was 0.5 nmol. Because the effect of rTNFa preincubation was about twice that seen with rTNF13 (7.4-fold vs. 3.8-fold increase in HOCI generation in response to FMLP), subsequent experiments focused on the characterization of chlorinated oxidant generation by the combination of rTNFa and FMLP. Consistent with results published for other stimuli [38], the detection of HOCI generation by neutrophils preincubated with rTNFa and triggered with FMLP re-
a’test for paired
values.
quired the presence of both H2O2 and functional myeloperoxidase, as the addition ofeither 50 ig/ml catalase (to catabolize H,O,) or I mM sodium azide (as a myeioperoxidase inhibitor) completely inhibited HOC1 generation (data not shown). Further, the addition of 1 mM methionine to .the system as a scavenger of HOCI prevented the detection of any HOC1. The time course of HOC1 generation by FMLPstimulated neutrophils preincubated with rTNFa is shown in Figure 1 HOC1 accumulated rapidly over the first 30 mm (6.00 ± 0.64 nmoi, 92.4% of total release) and reached a maximum of 6.49 ± 0. 13 nmol at 60 mm, remaining relatively constant thereafter. HOC1 generation as a function of neutrophil number was approximately linear up to 1 .0 X l0 neutrophils and reached a maximum of 6.60 ± 0.82 nmol with 2.0 X 106 neutrophils (Fig. 2). Figure 3 shows the relationship between the concentration of FMLP and the quantity of HOC! generated by neutrophils preincubated with differing concentrations of rTNFa ranging from 0-10 nM. FMLP alone induced the generation of only small amounts of HOCI by neutrophils with a maximum of 0.88 ± 0.02 nmol at an FMLP concentration of 10_b M. With FMLP concentrations of 10_s M or greater, preincubation with as little as 0. 1 nM rTNFa resulted in about a quadrupling of HOC! production. For example, at an FMLP concentration of l0 M, HOC! generation increased from 0.83 ± 0. 12 nmol with no other addition to 3.27 ± 0.59 nmol following preincubation with 0. 1 nmol rTNFa. The amount of HOC! trapped did not differ appreciably for concentrations of rTNFa of 1 and 10 nM, with maximal values of 6.04 ± I .03 and 6.37 ± 0.97 nmoi, respectively, at an FMLP concentration of l06 M. Although there was a relatively large increase in HOCI generation .
Generation
of Chlorinated
8.0
Oxidants by Adherent Neutrophils
135
10.0’ 8.0’
6.0 0
0
E
E
C
6.0
C
4.0
0
C)
0
0
x 2.0
4.0 2.0
0.0
0.0 0
15
30
45
Time
60
of incubation
75
90
105
0.0
120
(mm)
Fig. 1 Time course of HOCI generation by neutrophils preincubated with rTNFa prior to stimulation with FMLP. Neutrophils (i06/mI) were incubated at 37#{176}C as described in Materials and Methods. Recombinant TNFa (i nM) was added 10 mm prior to the addition of FMLP (i07 M). Incubations were continued at 37#{176}C for the indicated times and terminated by the addition of catalase (5 .tg). Supernatant HOCI (as the trapped N-chlorotaurine derivative) was then determined. Individual points represent the mean ± i SEM for three or four experiments. .
1 .0
2.0
Neutrophils
3.0
4.0
(millions)
Fig. 2. Relationship between cell number and HOCI generation by neutrophils preincubated with rTNFa prior to stimulation with FMLP. Neutrophils (0.25 x i06/ml-4.00 x 106/mI) were Incubated as described for Figure i except that incubations were terminated 60 mm after the addition of FMLP. Supernatant HOCI was assayed as described. Individual points represent the mean ± I SEM for three experiments.
8.00
with increasing FMLP concentration up to i0 M, there was only a further 1 nmol increase in HOC1 when the FMLP concentration was raised to lO_6 M (with the two highest concentrations of rTNFa). In contrast to the results observed when neutrophils were triggered under conditions permitting adherence, neutrophils (1 .0 x l06/ml) incubated in suspension (n = 4) failed to generate significant levels of HOC! when stimulated with buffer (0.0 ± 0.0 nmoi), rTNFa (0.04 ± 0.04 nmoi), or FMLP (0.04 ± 0.04 nmol) alone. These values differed significantly from those observed for adherent cells incubated with rTNFa (P 0.02 by two-tailed t test) or FMLP (P 0.001) alone. When neutrophils in suspension were incubated with rTNFa prior to the addition of FMLP, HOCI generation (1 .29 ± 0.62 nmol) was about one-quarter that observed for adherent neutrophils (P 0.001). Finally, we wished to determine whether adherent neutrophils triggered under conditions inducing optimal HOC! generation, i.e. with FMLP added following rTNFa, generated detectable quantities of endogenous N-C! derivatives. Indeed, in three experiments, 0.88 ± 0.20 nmo! of stable N-Cl derivatives accumulated in the supernatants of neutrophils (106/ml) stimulated in such a manner. This represented 18. 1 ± 3.2% of the HOCI generated by neutrophils in those experiments, a proportion similar to that previously reported for neutrophils stimulated with PMA or serum-opsonized zymosan [33]. The appearance of N-Cl derivatives in the supernatant was somewhat delayed compared with that of HOC!, ,
6.00 0
E
E.
4.00
0
0 I
2.00
0.00 i8
FMLP
concentration
10
(M)
Fig. 3. Relationship between FMLP concentration and HOCI generation by neutrophils preincubated with rTNFa prior to addition of FMLP. Neutrophils (i06/ml) were primed with buffer (.) or with various concentrations of rTNFa (0, 0.i nM; A i nM; L, iO nM)for 10 mm and then stimulated with FMLP (i09 M) for 60 mm as described in Materials and Methods. Supernatant HOCI was assayed as described. Error bars have been omitted for the sake of clarity. Individual points represent the mean of two experiments for neutrophils not primed with rTNFa and the mean of three or four experiments for cells primed with 0.i-io nM rTNFa. ,
-
with measurable levels first appearing Values reached -95% of maximum maximal values at 90 mm. Adherent bated with rTNFa or FMLP alone detectable quantities of N-Cl derivatives,
at 20-30 mm. by 60 mm and neutrophils incudid not generate nor did neutro-
136
Test
phils incubated the stimuli.
in suspension
with
any
combination
of
Comparison of O2, HOCI, and Myeloperoxidase Release by Adherent Neutrophils The quantity of HOC! generated by neutrophils following exposure to an agonist is a function both of the degree of respiratory burst activation and of the amount of myeioperoxidase released from primary granule stores. In order to determine which of these factors was responsible for the augmentation in HOC! production seen when rTNFa addition preceded that of FMLP, HOC1 generation was compared with 02 production and myeloperoxidase release (Table 1 ). At 1 h after addition of stimulus, the amount of 02 generated by either rTNFa or FMLP alone was close to or greater than twice the amount of HOC1 generated under any incubation conditions, even when rTNFa was added before FMLP. Assuming that one molecule of H202 arises from the dismutation of two molecules of O2 and that one molecule of H2O1 is used to generate one of HOC1, this suggested that the limiting factor in HOC1 generation with these stimuli was the amount of myeloperoxidase released, rather than the amount of H,O, available to react with the enzyme. Indeed, myeloperoxidase release from adherent neutrophils treated with both rTNFa and FMLP was more than twice that which occurred when the cells were stimulated with rTNFa or FMLP alone. In these experiments, there was no significant difference (by two-tailed t test for paired values, n 4) in lactate dehydrogenase release by neutrophi!s treated with rTNFa (5.4 ± 0.9%),FMLP(4.8 ± 0.7%),orrTNFa + FMLP (6.9 ± 0.8%). Although the amount of myeloperoxidase released into the supernatant by rTNFa alone did not differ significantly from that released from neutrophils incubated in buffer alone, it is likely that a small increase in myeioperoxidase release did in fact occur under these conditions, as HOC1 generation induced by either rTNFa (or FMLP) alone was slightly and significantly greater than for neutrophils in buffer only (Table 1). Although the quantity of mye!operoxidase released from granule stores appeared to be the preeminent factor in determining total HOC! production in these experiments, the kinetics of respiratory burst activation could still influence the characteristics of HOC1 generation by adherent neutrophi!s. In fact, the rate of 02 release was markedly accelerated when neutrophils were preincubated with rTNFa prior to triggering with FMLP (Fig. 4). Under these conditions, there was no evidence of an extended lag period in respiratory burst activation. The overall appearance of the time course of O generation by neutrophils stimulated with FMLP after addition of rTNFa was remarkably similar to that for HOC! production seen in Figure 1. This suggested that myeloperoxi-
0
E C
a V x
0 4)
a. Cl)
0
15
30
Time
45
60
75
of mncubation
90
105
120
(mm)
Fig. 4. Time course of superoxide generation by neutrophils incubated with rTNFa, FMLP, or rTNFa + FMLP. Neutrophils (i06/mI) were incubated at 37#{176}C in the presence of cytochrome C (2 mg/mI) with rTNFa (1 nM), or with FMLP (i0 M) alone or added after a iO mm priming incubation with rTNFa (i nM) After the indicated times, supernatants were removed and the reaction was terminated by centrifugation at 4#{176}C to remove any nonadherent cells. Superoxide generation was measured as described in Materials and Methods from the time of addition of rTNFa. FMLP was added at t iO mm. The curves shown are from a representative experiment of two performed. =
dase re!ease occurred rapidly enough to allow HOC! generation to closely follow O2 production. Indeed, 73.4 ± 4.3% (n = 2) of the amount of mye!operoxidase which would ultimately be released by 60 mm of incubation had been released by 15 mm after the addition of FMLP to rTNFa-preincubated cells.
DISCUSSION The ability of neutrophils adherent to surfaces coated with serum or extracellular matrix proteins to generate large quantities of H2O2 in response to activation with rTNFa, rTNF13, or FMLP [20] implies a mechanism whereby these agents might lead to biologically significant tissue damage during inflammation in vivo. Although H,O2 is capable of inducing cell injury by itself, the reaction of H2O, with myeioperoxidase in the presence of C1 leads to the generation of HOCI, a much more powerful and reactive oxidant [34]. Because of its high reactivity, HOC1 is relatively short-lived in biologic systems. However, the oxidizing equivalents of HOC1 can be preserved in the form of endogenous N-C! derivatives, long-lived oxidants which result from the reaction of HOC1 with amine-containing compounds released during the activation of neutrophils [ 1 1,26,3335,38,39]. The production of chlorinated oxidants by a neutrophil would require not only activation of the respiratory burst, but also the coordinated release of
Generation
myeloperoxidase from the primary granules. In the present study, we have shown that neutrophils incubated with rTNFa, rTNF13, or FMLP on polystyrene surfaces precoated with FBS generated only small quantities of HOC1. Further characterization of the activation process for rTNFa and FMLP demonstrated that minimal HOC! production occurred despite sufficient activation of the respiratory burst to potentially result in from 10-20 times the amount of HOC1 that was actually observed. The relative lack of HOC! generation by adherent neutrophils in response to these soluble triggers was subsequently shown to be a function of their ability to induce release of only a small amount of myeloperoxidase from granule stores. In contrast, neutrophils which were first incubated with either rTNFa or rTNF13 for 10 mm before the addition of FMLP showed a marked enhancement of HOC! generation. One hour after the addition of FMLP to 106 neutrophils which had been preincubated with rTNFa, 5.28 ± 0.39 nmol of HOC! had been released into the supernatant, an amount approximately twice that seen when rTNF13 was used in a similar manner. In addition, 1 8% of the HOC! generated was consumed in the formation of long-lived N-C! derivatives, a proportion similar to the 10-15% seen for neutrophils stimuiated with PMA or serum-opsonized zymosan [33] The increase in HOC! which resulted from the use of rTNFa and FMLP together in the neutrophi! incubations was accompanied by a significant increase in mye!operoxidase release into the supernatant. Although mye!operoxidase present in the cell supernatant might not represent all of the enzyme released from granule stores (e.g., some enzyme might remain adherent to the cell membrane), the !evels measured in the present experiments correlated well with the HOC! measurements and thus seemed to provide a reliable index of total mye!operoxidase release The rapid release of myeloperoxidase when rTNFa addition preceded FMLP triggering, coupled with enhanced respiratory burst kinetics (see below) led to rapid accumulation of HOC! in the medium. By 15 mm after the addition of FMLP, 73% of total mye!operoxidase release into the supernatant and 80% of total HOC! accumulation had been achieved (Fig 1 ). The concentrations of rTNFa (1 nM) and FMLP (l0 M) used for the majority of these experiments proved to be nearoptima! (Fig. 3), although enhanced HOC! generation was seen with a priming concentration of rTNFa as low as0.l nM. The failure of rTNFa to induce HOC1 generation by neutrophiis when used as a trigger, but to markedly enhance HOC! release when used as a priming agent for another agonist is consistent with a previously published report in which rTNFa was used by itself or with unopsonized zymosan to stimulate neutrophils in suspension [29]. That study differed from the present one in that -.
.
.
.
of Chlorinated
Oxidants
by Adherent
Neutrophils
137
the neutrophils were incubated in suspension with continuous stirring, the stimulus (unopsonized zymosan) was nonphysiologic and could be internalized by phagocytosis, and the mechanism of enhancement of HOC! was not determined [29] In another study of neutrophils activated with rTNFa and unopsonized zymosan, the cytokine was shown to significantly increase iodination, mye!operoxidase release, and H,O, generation stimulated by zymosan [13]. In both studies, rTNFa enhanced phagocytosis of the unopsonized zymosan, and this was considered a possible, if not likely, explanation for the observed effects on cell function [13,29]. Clearly, enhanced phagocytosis would not explain the effect seen in our system, which involved another soluble stimulus, FMLP, in addition to rTNFa. Nonetheless, a similar increase in myeloperoxidase release occurred when neutrophi!s were exposed to rTNFa before the addition of FMLP. In the only other study for which data are available on myeloperoxidase release in response to these stimuli, rTNFa priming did not cause an increase in extracellular activity of the enzyme in response to FMLP, but the neutrophils were in suspension rather than adherent to a surface [6]. Thus, adherence to a “biologic” surface may be the critical factor in inducing the enhanced mye!operoxidase release which leads to the HOC! response seen when a soluble agonist (FMLP) is used in combination with either TNFa or 13. The data in Table 1 and Figure 4 demonstrate that although total O2 generation (measured at 1 or 2 h) appeared to be additive for adherent neutrophils treated with both agonists vs. either alone, it is clear that rTNFa was behaving as a “priming” agent for FMLP in that the initial rate of O2 generation was markedly accelerated when cells were incubated with rTNFa for 10 mm prior to the addition of FMLP (Fig. 4). Indeed, total O2 accumulation at 15 or 30 mm suggested a synergistic effect from the use of both agents compared with either alone (Fig. 4). Under optimal incubation conditions in the present study, neutrophils were able to generate quantities of HOC! far below their demonstrated capacity. The 5-7.5 nmol of HOC! generated per 106 neutrophils is a small fraction of the potential release of 100-I 50 nmol reported for neutrophils triggered with PMA or serumopsonized zymosan [34,38] Can these amounts of HOC1 mediate important biologic effects? On the basis of published data, low concentrations of HOC1 clearly can induce both lethal and sublethal effects on target cells. The amounts of HOC! released by adherent neutrophils triggered with rTNFa and FMLP could inhibit cell division of -3.9 x i07 Escherichia (‘0/i or induce lysis of -l0 tumor cells [calculated on the basis of results published in ref. 1 and 30]. Recently, similar concentrations of HOC! were also shown to mediate important .
--
.
138
Test
sublethal effects on a murine tumor ce!l line [27]. Concentrations of HOC! as low as 10 pM could disturb plasma membrane functions including uptake of glucose and aminoisobutyric acid, maintenance of intracellular potassium, and maintenance of cell volume [27]. Thus, neutrophils exposed to the soluble agonists used in study, all of which could be encountered in an inflammatory focus, would be able to induce both cytotoxic and microbicida! injury. Neutrophils migrating into a site of inflammation would be more likely to encounter a complex mixture of cytokines and inflammatory mediators rather than any single mediator by itself. We have shown that two of these cytokines, TNFa or 13, in combination with FMLP, a chemotactic peptide similar or identical to N-formylmethionyl peptides released from bacteria [17] and eukaryotic mitochondria [7], can lead to the rapid activation of human peripheral blood neutrophils and that this is accompanied by the release of potentially cytotoxic concentrations of HOC! along with longer-lived N-Cl derivatives. Future investigations aimed at more precise!y defining the composition of the extracellular fluids in inflammatory sites should aid in determining which of the many potential neutrophil-mediator interactions might be appropriate targets for attempts at therapeutic intervention.
ACKNOWLEDGMENTS I thank Dr. B. Lubin for encouraging these studies and for reviewing the manuscript. This work was supported by National Institutes of Health grants DK-32094 and HL-20985.
REFERENCES I . Albrich, J.M. , and Hurst, J.K. Oxidative inactivation of Escherichia coli by hypochlorous acid. Rates and differentiation of respiratory from other reaction sites. FEBS Lett. 144, 157, 1982. 2. Andrews, P.C. , and Krinsky, NI. Human myeloperoxidase and hemi-myeloperoxidase. Methods Enzymol. 132, 369, 1986. 3. Atkinson, Y.H. , Marasco, WA. , Lopez, A.F. , and Vadas, MA. Recombinant human tumor necrosis factor-a. Regulation of N-formylmethionylleucylphenylalanine receptor affinity and function on human neutrophils. J. Clin. Invest. 81, 759, 1988. 4. Babior, B.M. Oxidants from phagocytes: agents of defense and destruction. Blood 64, 959, 1984. 5. Babior. B.M. , Kipnes, R.S. , and Curnutte, iT. Biological defense mechanisms. The production by leukocytes of superoxide, a potential bactericidal agent. J. Clin. Invest. 52, 741, 1973. 6. Berkow, R.L. , Wang, D. , Larrick, 1W. , Dodson, R.W. , and Howard, T.H. Enhancement of neutrophil superoxide production by preincubation with recombinant human tumor necrosis factor. J. Immunol. 139, 3783, 1987. 7. Carp, H. Mitochondrial N-formylmethionyl proteins as chemoattractants for neutrophils. J. Exp. Med. 155, 264, 1982. 8. Dewald, B., and Baggiolini, M. Methods for assessing exocytosis by neutrophil leukocytes. Methods Enzymol. 132, 267, 1986.
9. Foote, CS. , Goyne, I.E. , and Lehrer, RI. Assessment of chlorination by human neutrophils. Nature 301 , 715, 1983. 10. Granger, GA. , and Williams, T.W. Lymphocyte cytotoxicity in vitro: activation and release of a cytotoxic factor. Nature 218, 1253, 1968. 1 1 . Grisham, MB. , Jefferson, MM. , Melton, D.F. , and Thomas, EL. Chlorination of endogenous amines by isolated neutrophils. Ammonia-dependent bactericidal, cytotoxic, and cytolytic activities of the chloramines. I. Biol. Chem. 259, 10404, 1984. 12. Kapp, A. , Zeck-Kapp, G. , and Blohm, D. Human tumor necrosis factor is a potent activator of the oxidative metabolism in human polymorphonuclear neutrophilic granulocytes: comparison with human lymphotoxin. J. Invest. Dermatol. 92, 348, 1989. 13. Kiebanoff, Si. , Vadas, MA., Harlan, J.M. , Sparks, L.H., Gamble, JR. , Agosti, J.M. , and Waltersdorph, AM. Stimulation of neutrophils by tumor necrosis factor. J. Immunol. 136, 4220, 1986. 14. Kownatzki, E. , Kapp, A. , and Uhrich, S. Modulation of human neutrophilic granulocyte functions by recombinant human tumor necrosis factor and recombinant human lymphotoxin. Promotion of adherence, inhibition of chemotactic migration and superoxide anion release from adherent cells. Clin. Exp. Immunol. 74, 143, 1988. 15. Larrick, J.W. , Graham, D. , Toy, K. , Lin, L.S. , Senyk, G., and Fendly, B.M. Recombinant tumor necrosis factor causes activation of human granulocytes. Blood 69, 640, 1987. 16. Laudanna, C. , Miron, S. , Berton, G. , and Rossi, F. Tumor necrosis factor a/cachectin activates the 02 -generating system of human neutrophils independently of the hydrolysis of phosphoinositides and the release of arachidonic acid. Biochem. Biophys. Res. Commun. 166, 308, 1990. 17. Marasco, WA. , Phan, S.H. , Krutzsch, H. , Showell, Hi., Feltner, D.E. , Nairn, R. , Becker, EL. , and Ward, PA. Purification and identification of formyl-methionyl-leucyl-phenylalanine as the major peptide neutrophil chemotactic factor produced by Escherichia co/i. J. Biol. Chem. 259, 5430, 1984. 18. Massey, V. The microestimation of succinate and the extinction coefficient of cytochrome c. Biochim. Biophys. Acta 34, 255, 1959. 19. Meurer, R., and Maclntyre, D.E. Lack ofeffect ofpertussis toxin on TNF-a-induced formation of reactive oxygen intermediates by human neutrophils. Biochem. Biophys. Res. Commun. 159, 763, 1989. 20. Nathan, CF. Neutrophil activation on biological surfaces. Massive secretion of hydrogen peroxide in response to products of macrophages and lymphocytes. J. Clin. Invest. 80, 1550, 1987. 21 . Nathan, CF. Respiratory burst in adherent human neutrophils: triggering by colony-stimulating factors CSF-GM and CSF-G. Blood 73, 301, 1989. 22. Nedwin, G.E. , Svedersky, L.P. , Bringman, T.S. , Palladinno, MA. , Jr. , and Goeddel, DV. Effect of interleukin 2, interferon‘1 and mitogens on the production of tumor necrosis factors a and 23.
24.
25.
13. J. Immunol. 135, 2492, 1985. Neumann, M. , and Kownatzki, E. The effect of adherence on the generation of reactive oxygen species by human neutrophilic granulocytes. Agents Actions 26, 183, 1989. Pennica, D. , Nedwin, G.E. , Hayflick, iS. , Seeburg, PH., Derynck, R. , Palladino, MA. , Kohr, Wi. , Aggarwal, B.B. , and Goeddel, DV. Human tumor necrosis factor: precursor structure, expression and homology to lymphotoxin. Nature 312, 724, 1984. Perussia, B. , Kobayashi, M. , Rossi, ME. , Anegon, I. , and Trinchieri, G. Immune interferon enhances functional properties of human granulocytes: role of Fe receptors and effect of lymphotoxin , tumor necrosis factor, and granulocyte-macrophage colony-stimulating factor. J. Immunol. 138, 765, 1987.
Generation of Chlorinated Oxidants by Adherent 26. Sagone,
27.
28.
29.
30.
31 32.
33.
.
AL. , Jr. , Husney, R.M. . O’Dorisio, MS. , and Metz. EN. Mechanisms for the oxidation of reduced glutathione by stimulated neutrophils. Blood 63, 96, 1984. Schraufst#{228}tter, lU. , Browne, K. , Harris, A. , Hyslop, PA., Jackson, J.H. , Quehenberger, 0. , and Cochrane, C.G. Mechanisms of hypochlorite injury of target cells. J. Clin. Invest. 85, 554, 1990. Shalaby, MR. , Palladino, MA. , Jr. , Hirabayashi, SE. , Eessalu, T.E. , Lewis, GD. , Shepard, H.M. , and Aggarwal, B.B. Receptor binding and activation of polymorphonuclear neutrophils by tumor necrosis factor-alpha. J. Leukoc. Biol. 41 , 196, 1987. She, Z. , Wewers, M.D. , Herzyk, D.J. , Sagone, AL. , and Davis, W.B. Tumor necrosis factor primes neutrophils for hypochlorous acid production. Am. J. Physiol. 257. L338, 1989. Slivka, A. , LoBuglio, A.F. , and Weiss, S.J. A potential role for hypochlorous acid in granulocyte-mediated tumor cell cytotoxicity. Blood 55, 347, 1980. Steinbeck, AM. , and Roth, J.A. Neutrophil activation by recombinant cytokines. Rev. Infect. Dis. I 1 , 549, 1989. Tennenberg, S.D. , and Solomkin, J.S. Activation of neutrophils by cachectin/tumor necrosis factor: priming of N-formyl-methionyl-leucyl-phenylalanine-induced oxidative responsiveness via receptor mobilization without degranulation. J. Leukoc. Biol. 47, 217, 1990. Test, ST., Lampert. MB., Ossanna, P.J., Thoene, J.G., and Weiss, S.J. Generation of nitrogen-chlorine oxidants by human phagocytes. J. Clin. Invest. 74, 1341, 1984.
34.
Neutrophils
139
Test. ST. , and Weiss, S.J. The generation and utilization of chlorinated oxidants by human neutrophils. Adv. Free Radic. Biol. Med. 2, 91, 1986. 35. Thomas, EL. . Grisham, MB. , and Jefferson, MM. Myeloperoxidase-dependent effect of amines on functions of isolated neutrophils. J. Clin. Invest. 72, 441, 1983. 36. Tsujimoto. M. , Yokota, S. , Vilcek, J. . and Weissmann, G. Tumor necrosis factor provokes superoxide anion generation from neutrophils. Biochem. Biophys. Res. Commun. 137, 1094, 1986. 37. Urban, J.L. , Shepard, H.M. , Rothstein, J.L. , Sugarman, B.J., and Schreiber, H. Tumor necrosis factor: a potent effector molecule for tumor cell killing by activated macrophages. Proc. Natl. Acad. Sci. USA 83, 5233, 1986. 38. Weiss, S.J. , Klein, R. . Slivka, A. , and Wei, M. Chlorination of taurine by human neutrophils. Evidence for hypochlorous acid generation. J. Clin. Invest. 70, 598, 1982. 39. Weiss, S.J. , Lampert. MB. . and Test, ST. Long-lived oxidants generated by human neutrophils: characterization and bioactivity. Science 222, 625. 1983. 40. Weiss, S.J. Oxygen as a weapon in the phagocyte armamentarium. In Handbook of Inflammation: Immunology of Inflammation, vol. 4, (Ward, PA. , Ed.). Amsterdam: Elsevier Science Publishers, p. 37, 1983. 41. Yuo, A., Kitagawa, S., Suzuki, I., Urabe, A.. Okabe, T., Saito, M. , and Takaku, F. Tumor necrosis factor as an activator of human granulocytes. Potentiation of the metabolisms triggered by the Ca2-mobilizing agonists. J. Immunol. 142, 1678, 1989.
