INFECTION AND IMMUNITY, Sept. 1979, P. 912-921 00 19-9567/79/09-0912/ 10$02.()0/0
Vol. 25, No. 3
Endotoxin In Vitro Interactions with Human Neutrophils: Depression of Chemiluminescence, Oxygen Consumption, Superoxide Production, and Killing RICHARD A. PROCTOR Departments of Medical Microbiology and Medicine, University of Wisconsin, Madison, Wisconsin 53706 Received for publication 22 June 1979
Endotoxin was shown to depress neutrophil bactericidal activity while enhancing Nitro Blue Tetrazolium reduction and hexose monophosphate shunt activity. Separation of bactericidal action from oxidative metabolism suggests that the toxic effect of endotoxin might involve the formation of reactive oxygen radicals such as superoxide. Chemiluminescence often accompanies metabolic activation of polymorphonuclear neutrophils (PMNs). However, human PMNs did not show chemiluminescence when challenged with endotoxin (lipopolysaccharide; LPS) or lipid A. Superoxide formation was also unaffected by endotoxin. In contrast, preincubation of PMNs with LPS for 30 min produced significant depression of chemiluminescence, oxygen consumption, and superoxide formation. Decreased chemiluminescence was not the result of complement consumption. In a cell-free system, superoxide was not scavenged by LPS, nor did LPS stimulate superoxide dismutase. Oxidase enzymes for reduced nicotinamide adenine dinucleotide or reduced nicotinamide adenine dinucleotide phosphate harvested from broken cells were not affected by LPS. The toxicity of LPS may reside in its ability to activate the PMNs while simultaneously blocking bactericidal capacity.
There is general agreement that exposure of polymorphonuclear neutrophils (PMNs) to endotoxin (lipopolysaccharide; LPS) in vitro results in enhanced hexose monophosphate shunt (HMPS) activity (10, 11, 23, 41), increased glycolysis (10, 11, 23, 28), increased lysosomal enzyme release (22), and accelerated Nitro Blue Tetrazolium (NBT) reduction (19, 29, 33, 34, 40, 41). In contrast, various authors have found that LPS challenge increases (23, 45), decreases (28), or has no effect (11) on neutrophil oxygen consumption. This disparity of results might be due to the different species used in these experiments (human, guinea pig, rabbit), differences in methods of obtaining PMNs (i.e., from exudates or venous blood), the presence or absence of serum in the system, and the relative insensitivity of the Warburg manometric technique (48). Understanding the oxidative changes that occur in LPS-challenged human PMNs is critical since oxygen-derived radicals are instrumental in neutrophil bactericidal activity (3). This communication describes the effects of endotoxin on human neutrophil oxygen consumption, HMPS activity, chemiluminescence (CL), superoxide production, reduced nicotinamide adenine dinucleotide (phosphate) (NADPH-NADH = [NAD(P)H]) oxidase activity, and bactericidal
activity. 912
MATERIALS AND METHODS
Preparation of neutrophils. Venous blood from healthy donors was heparinized (2 U of aqueous sodium heparin per ml of blood; grade 1; Sigma Chemical Co., St. Louis, Mo.) and combined with 1 to 2 volumes of 3% dextran (average molecular weight, 264,000; Sigma) in isotonic saline (pH 7.4). After 20 min of erythrocyte sedimentation at room temperature, the leukocyte-rich supernatant was separated and then centrifuged (200 x g, 5 min). Contaminating erythrocytes were lysed by addition of 0.155 M NH4Cl, 10 mM ethylenediaminetetraacetic acid, and 0.3 mM NaHCO:3 at 37°C. The preparation was washed twice with 'ianks balanced salt solution without phenol red (HBSS; GIBCO, Grand Island, N.Y.) (pH 7.4). Greater than 95% of PMNs appeared viable as assessed by 1% trypan blue exclusion. Differential and quantitative leukocyte counts were performed. Suspensions of PMNs were adjusted with HBSS to final concentrations as noted in legends of figures. LPS was added to PMNs either simultaneously with or 30 min before challenge with other substances. Some of the PMNs were preincubated with LPS at 37°C for 30 min, and the rest of the PMNs were similarly incubated as controls. After 30 min of incubation of LPS with PMNs, greater than 91% remained viable as determined by trypan blue exclusion. Phagocytizable particles. Escherichia coli (ATCC 25922), E. coli 0:6 (a clinical isolate from urine), and Staphylococcus aureus (ATCC 25923) were grown overnight in tryptic soy broth (Difco Lab-
VOL. 25, 1979
oratories, Detroit, Mich.). (Serotyping was performed in the laboratory of Calvin Kunin, William S. Middleton V.A. Hospital, Madison, Wis.) Suspensions of bacteria were washed three times in HBSS. Single batches of lyophilized E. coli (ATCC 25922) and S. aureus were stored at -20'C and used in oxygen consumption studies. Before lyophilization, bacteria were suspended in sterile distilled water. Those bacteria used for bactericidal assay (E. coli 0:6 and S. aureus) were obtained from overnight cultures and adjusted to an optical density of 0.6 at 620 nm. A 1:1,000 dilution gave 1 x 10' to 4 x 10' bacteria per ml. Both E. coli 0:6 and S. aureus were resistant to serum bactericidal action. Zymosan (Sigma) was prepared using the method of Rosen and Klebanoff (38). Bacteria and zymosan were preopsonized by incubation with serum (37"C, 30 min) and then washed twice. Serum was obtained from healthy donors and stored at -40'C for less than 6 weeks before use. Latex beads (0.81 pm, Difco) were washed twice and then adjusted to 4 x 109 particles per ml. Endotoxincoated latex beads were prepared by incubation of LPS with latex beads (1 mg of LPS per ml of latex suspension). Binding was confirmed by specific agglutination of triply washed latex beads with specific rabbit antisera to E. coli serotype 0111:B4 (antisera donated by Calvin Kunin). Normal rabbit sera failed to agglutinate endotoxin-coated beads. The endotoxin antisera did not agglutinate uncoated latex beads. Endotoxin-coated latex beads were also adjusted to 4 x 109 particles per ml. Endotoxin. A single lot (no. 740856) of E. coli endotoxin 0111:B4 prepared by the Westphal method was obtained commercially (Difco) and used throughout. LPS caused gelation of Limulus amoebocyte lysate at 1 to 5 ng/ml (Sigma) (43) and demonstrated an intraperitoneal 50% lethal dose for C57BL mice (Sprague-Dawley) of 150 Mg. This LPS contained about 1% protein by weight. Lipid A was a gift from Paul Madson (William S. Middleton Memorial V. A. Hospital, Madison, Wis.) which had been prepared from Salmonella in the laboratory of 0. Westphal (Freiburg, Germany) as previously described (17). Lipid A was solubilized by adding bovine serum albumin.
Oxygen measurements. Oxygen consumption by PMNs was measured polarographically with a Clark oxygen electrode (Gibson Medical Electronics, Middleton, Wis.). Measurements were made at 370C for 3 to 5 min after challenge with phagocytizable particles. Experimental conditions are noted in Table 1, footnote a. Bactericidal assay. PMN bactericidal activity was modified using the method of Quie et al. (36). Duplicate 1-ml reaction mixtures in HBSS containing 10% serum, 5 x 10" PMNs, and 1 x 10' to 4 X 10' bacteria were rotated end-over-end for 120 min at 37"C. Samples were taken at 30-min intervals and diluted with distilled water. The number of bacteria per milliliter was determined by the pour-plate method using tryptic soy agar (Difco). Quantitative NBT test. The quantitative NBT dye reduction test was performed as previously described (4). A brief description of the method, includ-
ENDOTOXIN IN VITRO INTERACTIONS
913
ing the anaerobic modification, follows. PMNs were divided into equal portions and held on ice either in room air or in a Coy anaerobic glove box (Ann Arbor, Mich.) for 5 h. After the cells were allowed to warm for 30 min at 37°C, 4 x 10` PMNs, 0.05 ml of 2C mM KCN, 0.4 ml of 0.1% NBT, 100 yg of LPS, and HBSS sufficient to make a volume of 1.1 ml were added and incubated either aerobically or anaerobically for 30 min at 37°C. Reactions were terminated by adding 4.0 ml of ice-cold 0.4 N HCI. Pellets were extracted with 4.0 ml of pyridine, and optical density was read at 515 nm. HMPS activity. HMPS activity was determined as previously described (35). Briefly, 5 x 10" PMNs, 1.0 MuCi of D-[1-'4C]glucose or D-[6-'4C]glucose, 2 mM KCN, and challenge substance (as noted in Table 3) were incubated in glucose-free HBSS (total volume, 1.0 ml) at 37°C on a reciprocating shaker. The reaction was terminated by addition of 1.0 ml of 10% trichloroacetic acid. Released 14C02 was trapped on filter paper impregnated with methyl benzethonium hydroxide and then counted in a Packard Tri-Carb liquid scintillation counter. CL. CL was measured in a Packard Tri-Carb liquid scintillation counter by the method of Allen et al. (1). Reaction mixtures were maintained at 37°C until they were counted for 1 min. Concanavalin A (ConA) was obtained from Sigma. Superoxide measurements. In the cell-free system, superoxide production was measured by using the method of McCord and Fridovich (31). Reaction mixtures containing 0.15 mM xanthine (Sigma), 3.6 x 10 9 M xanthine oxidase (Boehringer Mannheim Corp., New York, N.Y.) and 30 MM horse heart ferricytochrome (type III, Sigma) in 3 ml of 0.05 M potassium phosphate buffer (pH 7.8) were incubated at 20°C. Reaction mixtures containing superoxide dismutase (>3,000 U/mg; Truett Labs., Dallas, Tex.) received sufficient enzyme to cause a 50% decrease in cytochrome c reduction. When superoxide (02-) measurements were performed on PMNs, reaction mixtures contained 75 MM horse ferricytochrome c, 5 x 10" PMNs, and 10% serum in HBSS, according to the method of Goldstein et al. (21). An extinction coefficient of 2.11 x 104 M cm-' at 550 nm (reducedoxidized) was used to calculate micromoles of ferrocytochrome c formed. When present, LPS was at a final concentration of 100 jg/ml. NAD(P)H oxidase activity. Neutrophil oxidase activity was measured by the fluorometric method of DeChatelet (14, 26) and is briefly described below. Reaction mixtures containing 5 x 10' PMNs in phosphate-buffered saline (pH 7.4) were equilibrated for 30 min at 37°C before challenge with preopsonized zymosan particles (final concentration, 3 mg/ml). To those tubes receiving LPS, LPS (100 Mg/ml) was added either at the beginning or the end of the equilibration period. Three minutes after zymosan was added, the reaction was stopped by adding 2.0 ml of iced 0.68 M sucrose. The cells were homogenized (greater than 95% breakage) and then centrifuged at 500 x g, and finally the supernatant was removed and centrifuged at 27,000 x g. The protein concentration in the highspeed pellet was adjusted to 1 mg/ml, and 0.3-ml samples were used for assay in the NAD(P)H oxidase
914
PROCTOR
assay. Assay mixtures (1.0 ml) contained 0.3 ml of pellet protein solution, 2.0 mM KCN, 0.1 M 2-(Nmorpholino)ethanesulfonic acid (pH 6.0) buffer, and either 0.2 mM NADPH or 0.8 mM NADH. After 30 min of incubation at 370C, reactions were stopped by the addition of 1 ml of 0.4 M HCl04, and precipitated protein was removed by centrifugation. A 0.1-ml supernatant was incubated with 0.15 ml of 10 N NaOH for 1 h at 20'C. Then 1.6 ml of water was added, and fluorescence was read on an Aminco-Bowman spectrofluorometer (excitation, 365 nm; emission, 448 nm). Nanomoles of NADH or NADPH formed were determined by comparison to relative fluorescence of standard curves, run in parallel daily. All determinations were performed in duplicate. Direct addition of LPS to the assay mixture was found not to alter oxidase activity.
RESULTS Effect of endotoxin on neutrophil bactericidal activity. The effects of graded doses of LPS on PMN function are shown in Fig. 1. When no LPS was present, approximately 1.5 log10 bacteria were killed in 120 min. Preincubation of LPS with PMNs (Fig. 1A and C) inhibited killing of E. coli at 10 jig of LPS per ml or greater and resulted in a dose-dependent reduction in the killing of S. aureus. When LPS was added simultaneously with E. coli or S. aureus (Fig. 1B and D), PMN bactericidal action was again reduced, but to a lesser extent. Of particular interest were the rates of killing of E. coli during the first 30 min. With simultaneous addition of 10 to 100 pg of LPS per ml and E. coli to PMNs (Fig. 1B), the rates of killing progressed rapidly for 30 min and then plateaued. After 30 min of incubation, the contours of the growth curves in Fig. 1B were nearly identical to those in Fig. 1A. That these bacteria were resistant to killing by 10% serum was demonstrated by their growth when no PMNs were present (Fig. 1). Uptake of bacteria was not affected by the presence of LPS, as determined by the percentage of PMNs that had phagocytized bacteria or by the number of bacteria phagocytized per PMN. Effect of endotoxin on neutrophil oxygen consumption. Neutrophils rapidly consumed oxygen when phagocytizing E. coli, S. aureus, or zymosan (Table 1). Concomitant addition of LPS with these same particles allowed the phagocytically induced increase in oxygen consumption to proceed normally. In contrast, preincubation of LPS with PMNs resulted in a marked reduction of oxygen consumption during phagocytosis. Challenge of PMNs with LPS alone resulted in rates of oxygen consumption nearly identical to that of resting PMNs, even when reaction rates were monitored for as long as 45 min. Addition of 0.1 ml of 0.01 M KCN to the reaction mixtures did not alter oxygen uptake in
INFECT. IMMUN.
two of the experiments done in duplicate. Since the bacteria were in lag phase (lyophilized and reconstituted for 30 to 60 min before use), they did not show oxygen uptake when added to the reaction chamber alone. Effect of LPS on neutrophil HMPS activity. Challenge of neutrophils with E. coli or S. aureus resulted in an approximate 10-fold increase in [1-'4C]glucose metabolism and a 2-fold increase in [6-'4C]glucose metabolism over resting values (Table 2). Addition of endotoxin resulted in a twofold (10 Mg) to threefold (100 Mg) increase in [1-'4C]glucose oxidation and a 40 to 75% increase in [6-'4C]glucose oxidation. Effect of LPS on NBT dye reduction. Reduction of NBT dye by PMNs was enhanced by addition of either latex spheres or LPS (Table 3). This reduction increased with increasing concentrations of LPS. Addition of larger quantities of LPS (up to 1,000 4g/ml) resulted in no greater reduction than occurred at 100 tg/ml. Because particle-induced NBT reduction by neutrophils is reportedly mediated solely by superoxide (an oxygen-derived radical) in whole PMNs (6), we elected to measure LPS-induced NBT reduction anaerobically (Fig. 2). A consistent increase in NBT reduction occurred whether the cells were held aerobically or anaerobically for 5 h. A larger mean increase in optical density was seen under aerobic (0.067 ± 0.021) conditions as compared to anaerobic conditions (0.039 ± 0.009). Also, total change in optical density was smaller after the PMNs were held (both aerobic and anaerobic) for 5 h at 4°C than when cells were used immediately. Cell suspensions were warmed for 30 min at 37°C before use to allow microtubule function to be restored. Effect of LPS on PMN CL. When PMNs were challenged with concentrations of LPS from 0.1 to 100 pg/ml, there was no change in CL as compared to resting cells. Table 4 summarizes the data for E. coli O111:B4 LPS observed at 25 min. Challenge with endotoxins from E. coli 055:B5 (Difco) and Salmonella typhimurium (Difco) also failed to induce CL. Also, when observations were performed at 5- to 15-min intervals over a 1- to 180-min time span, LPS did not stimulate PMN CL. Furthermore, no enhancement of neutrophil CL followed lipid A challenge (Table 4). To be certain that LPS was entering the PMNs, LPS was bound to latex particles. Whether LPS was absent, added simultaneously, or bound to the latex beads, no alteration in latex-induced CL was noted (Table 4). Addition of S. aureus and E. coli to PMNs resulted in a 20- to 30-fold increase in CL (Table 5). This increase was not affected by the simul-
VOL. 25, 1979
ENDOTOXIN IN VITRO INTERACTIONS
915
70 s0o 50#9 /0
()
6.00
1.Efc FIG.~~~
fenooi 0
nnutohlbceicdlatvt. 0
30
60
90
120
0
0 30
60
Ms 10#9 90
eu,
BS
n
120
TIME (min) FIG. 1. Effect of endotoxin on neutrophil bactericidal activity. 5 x 106 PMNs, 10% serum, HBSS, and bacteria (S. aureus or E. coli 0:6) were incubated with LPS (O to 100 pg/ml). (A) and (C) Data when LPS was preincubated with PMNs; (B) and (D) effects of simultaneous additions. Each point represents the mean of five to six experiments done in duplicate. One standard deviation is shown for 120-min values.
taneous addition of LPS, whereas simultaneously added lipid A caused a consistent, but not statistically significant, reduction. In contrast, preincubation of neutrophils with LPS or lipid A for 30 min produced a highly significant reduction in bacteria-induced CL. A similar reduction in ConA-stimulated CL was noted in those PMNs preincubated with LPS (Table 5). Because LPS is known to activate the alternative pathway of complement (18, 32), studies were undertaken to assess the role of complement in the LPS-altered CL response of PMNs. When heated serum replaced fresh serum (Fig. 3), a marked reduction in light production was
obtained. Nevertheless, the PMNs preincubated with LPS still showed less CL in heated serum than those PMNs that were either unexposed to LPS or added to the system simultaneously with LPS. PMNs challenged with preopsonized bacteria in heated serum showed the same CL values as those challenged with bacteria in fresh serum, whereas the LPS-preincubated PMNs showed significantly less light production. Effect of endotoxin on neutrophil superoxide production. To be sure that LPS was not decreasing CL by light quenching, by scavenging light-producing radicals, or by enhancing superoxide dismutase activity, 02 formation
916
INFECT. IMMUN.
PROCTOR
TABLE 1. Effect of endotoxin in particle-induced neutrophil oxygen consumption Oxygen consumption' by: P
Particles
PMNs + LPS-PI
PMNs + LPS
PMNs
4.31 ± 0.4 (13) 4.87 ± 0.5 (13) 4.38 ± 0.5 (14) None
Vol. 25, No. 3
Endotoxin In Vitro Interactions with Human Neutrophils: Depression of Chemiluminescence, Oxygen Consumption, Superoxide Production, and Killing RICHARD A. PROCTOR Departments of Medical Microbiology and Medicine, University of Wisconsin, Madison, Wisconsin 53706 Received for publication 22 June 1979
Endotoxin was shown to depress neutrophil bactericidal activity while enhancing Nitro Blue Tetrazolium reduction and hexose monophosphate shunt activity. Separation of bactericidal action from oxidative metabolism suggests that the toxic effect of endotoxin might involve the formation of reactive oxygen radicals such as superoxide. Chemiluminescence often accompanies metabolic activation of polymorphonuclear neutrophils (PMNs). However, human PMNs did not show chemiluminescence when challenged with endotoxin (lipopolysaccharide; LPS) or lipid A. Superoxide formation was also unaffected by endotoxin. In contrast, preincubation of PMNs with LPS for 30 min produced significant depression of chemiluminescence, oxygen consumption, and superoxide formation. Decreased chemiluminescence was not the result of complement consumption. In a cell-free system, superoxide was not scavenged by LPS, nor did LPS stimulate superoxide dismutase. Oxidase enzymes for reduced nicotinamide adenine dinucleotide or reduced nicotinamide adenine dinucleotide phosphate harvested from broken cells were not affected by LPS. The toxicity of LPS may reside in its ability to activate the PMNs while simultaneously blocking bactericidal capacity.
There is general agreement that exposure of polymorphonuclear neutrophils (PMNs) to endotoxin (lipopolysaccharide; LPS) in vitro results in enhanced hexose monophosphate shunt (HMPS) activity (10, 11, 23, 41), increased glycolysis (10, 11, 23, 28), increased lysosomal enzyme release (22), and accelerated Nitro Blue Tetrazolium (NBT) reduction (19, 29, 33, 34, 40, 41). In contrast, various authors have found that LPS challenge increases (23, 45), decreases (28), or has no effect (11) on neutrophil oxygen consumption. This disparity of results might be due to the different species used in these experiments (human, guinea pig, rabbit), differences in methods of obtaining PMNs (i.e., from exudates or venous blood), the presence or absence of serum in the system, and the relative insensitivity of the Warburg manometric technique (48). Understanding the oxidative changes that occur in LPS-challenged human PMNs is critical since oxygen-derived radicals are instrumental in neutrophil bactericidal activity (3). This communication describes the effects of endotoxin on human neutrophil oxygen consumption, HMPS activity, chemiluminescence (CL), superoxide production, reduced nicotinamide adenine dinucleotide (phosphate) (NADPH-NADH = [NAD(P)H]) oxidase activity, and bactericidal
activity. 912
MATERIALS AND METHODS
Preparation of neutrophils. Venous blood from healthy donors was heparinized (2 U of aqueous sodium heparin per ml of blood; grade 1; Sigma Chemical Co., St. Louis, Mo.) and combined with 1 to 2 volumes of 3% dextran (average molecular weight, 264,000; Sigma) in isotonic saline (pH 7.4). After 20 min of erythrocyte sedimentation at room temperature, the leukocyte-rich supernatant was separated and then centrifuged (200 x g, 5 min). Contaminating erythrocytes were lysed by addition of 0.155 M NH4Cl, 10 mM ethylenediaminetetraacetic acid, and 0.3 mM NaHCO:3 at 37°C. The preparation was washed twice with 'ianks balanced salt solution without phenol red (HBSS; GIBCO, Grand Island, N.Y.) (pH 7.4). Greater than 95% of PMNs appeared viable as assessed by 1% trypan blue exclusion. Differential and quantitative leukocyte counts were performed. Suspensions of PMNs were adjusted with HBSS to final concentrations as noted in legends of figures. LPS was added to PMNs either simultaneously with or 30 min before challenge with other substances. Some of the PMNs were preincubated with LPS at 37°C for 30 min, and the rest of the PMNs were similarly incubated as controls. After 30 min of incubation of LPS with PMNs, greater than 91% remained viable as determined by trypan blue exclusion. Phagocytizable particles. Escherichia coli (ATCC 25922), E. coli 0:6 (a clinical isolate from urine), and Staphylococcus aureus (ATCC 25923) were grown overnight in tryptic soy broth (Difco Lab-
VOL. 25, 1979
oratories, Detroit, Mich.). (Serotyping was performed in the laboratory of Calvin Kunin, William S. Middleton V.A. Hospital, Madison, Wis.) Suspensions of bacteria were washed three times in HBSS. Single batches of lyophilized E. coli (ATCC 25922) and S. aureus were stored at -20'C and used in oxygen consumption studies. Before lyophilization, bacteria were suspended in sterile distilled water. Those bacteria used for bactericidal assay (E. coli 0:6 and S. aureus) were obtained from overnight cultures and adjusted to an optical density of 0.6 at 620 nm. A 1:1,000 dilution gave 1 x 10' to 4 x 10' bacteria per ml. Both E. coli 0:6 and S. aureus were resistant to serum bactericidal action. Zymosan (Sigma) was prepared using the method of Rosen and Klebanoff (38). Bacteria and zymosan were preopsonized by incubation with serum (37"C, 30 min) and then washed twice. Serum was obtained from healthy donors and stored at -40'C for less than 6 weeks before use. Latex beads (0.81 pm, Difco) were washed twice and then adjusted to 4 x 109 particles per ml. Endotoxincoated latex beads were prepared by incubation of LPS with latex beads (1 mg of LPS per ml of latex suspension). Binding was confirmed by specific agglutination of triply washed latex beads with specific rabbit antisera to E. coli serotype 0111:B4 (antisera donated by Calvin Kunin). Normal rabbit sera failed to agglutinate endotoxin-coated beads. The endotoxin antisera did not agglutinate uncoated latex beads. Endotoxin-coated latex beads were also adjusted to 4 x 109 particles per ml. Endotoxin. A single lot (no. 740856) of E. coli endotoxin 0111:B4 prepared by the Westphal method was obtained commercially (Difco) and used throughout. LPS caused gelation of Limulus amoebocyte lysate at 1 to 5 ng/ml (Sigma) (43) and demonstrated an intraperitoneal 50% lethal dose for C57BL mice (Sprague-Dawley) of 150 Mg. This LPS contained about 1% protein by weight. Lipid A was a gift from Paul Madson (William S. Middleton Memorial V. A. Hospital, Madison, Wis.) which had been prepared from Salmonella in the laboratory of 0. Westphal (Freiburg, Germany) as previously described (17). Lipid A was solubilized by adding bovine serum albumin.
Oxygen measurements. Oxygen consumption by PMNs was measured polarographically with a Clark oxygen electrode (Gibson Medical Electronics, Middleton, Wis.). Measurements were made at 370C for 3 to 5 min after challenge with phagocytizable particles. Experimental conditions are noted in Table 1, footnote a. Bactericidal assay. PMN bactericidal activity was modified using the method of Quie et al. (36). Duplicate 1-ml reaction mixtures in HBSS containing 10% serum, 5 x 10" PMNs, and 1 x 10' to 4 X 10' bacteria were rotated end-over-end for 120 min at 37"C. Samples were taken at 30-min intervals and diluted with distilled water. The number of bacteria per milliliter was determined by the pour-plate method using tryptic soy agar (Difco). Quantitative NBT test. The quantitative NBT dye reduction test was performed as previously described (4). A brief description of the method, includ-
ENDOTOXIN IN VITRO INTERACTIONS
913
ing the anaerobic modification, follows. PMNs were divided into equal portions and held on ice either in room air or in a Coy anaerobic glove box (Ann Arbor, Mich.) for 5 h. After the cells were allowed to warm for 30 min at 37°C, 4 x 10` PMNs, 0.05 ml of 2C mM KCN, 0.4 ml of 0.1% NBT, 100 yg of LPS, and HBSS sufficient to make a volume of 1.1 ml were added and incubated either aerobically or anaerobically for 30 min at 37°C. Reactions were terminated by adding 4.0 ml of ice-cold 0.4 N HCI. Pellets were extracted with 4.0 ml of pyridine, and optical density was read at 515 nm. HMPS activity. HMPS activity was determined as previously described (35). Briefly, 5 x 10" PMNs, 1.0 MuCi of D-[1-'4C]glucose or D-[6-'4C]glucose, 2 mM KCN, and challenge substance (as noted in Table 3) were incubated in glucose-free HBSS (total volume, 1.0 ml) at 37°C on a reciprocating shaker. The reaction was terminated by addition of 1.0 ml of 10% trichloroacetic acid. Released 14C02 was trapped on filter paper impregnated with methyl benzethonium hydroxide and then counted in a Packard Tri-Carb liquid scintillation counter. CL. CL was measured in a Packard Tri-Carb liquid scintillation counter by the method of Allen et al. (1). Reaction mixtures were maintained at 37°C until they were counted for 1 min. Concanavalin A (ConA) was obtained from Sigma. Superoxide measurements. In the cell-free system, superoxide production was measured by using the method of McCord and Fridovich (31). Reaction mixtures containing 0.15 mM xanthine (Sigma), 3.6 x 10 9 M xanthine oxidase (Boehringer Mannheim Corp., New York, N.Y.) and 30 MM horse heart ferricytochrome (type III, Sigma) in 3 ml of 0.05 M potassium phosphate buffer (pH 7.8) were incubated at 20°C. Reaction mixtures containing superoxide dismutase (>3,000 U/mg; Truett Labs., Dallas, Tex.) received sufficient enzyme to cause a 50% decrease in cytochrome c reduction. When superoxide (02-) measurements were performed on PMNs, reaction mixtures contained 75 MM horse ferricytochrome c, 5 x 10" PMNs, and 10% serum in HBSS, according to the method of Goldstein et al. (21). An extinction coefficient of 2.11 x 104 M cm-' at 550 nm (reducedoxidized) was used to calculate micromoles of ferrocytochrome c formed. When present, LPS was at a final concentration of 100 jg/ml. NAD(P)H oxidase activity. Neutrophil oxidase activity was measured by the fluorometric method of DeChatelet (14, 26) and is briefly described below. Reaction mixtures containing 5 x 10' PMNs in phosphate-buffered saline (pH 7.4) were equilibrated for 30 min at 37°C before challenge with preopsonized zymosan particles (final concentration, 3 mg/ml). To those tubes receiving LPS, LPS (100 Mg/ml) was added either at the beginning or the end of the equilibration period. Three minutes after zymosan was added, the reaction was stopped by adding 2.0 ml of iced 0.68 M sucrose. The cells were homogenized (greater than 95% breakage) and then centrifuged at 500 x g, and finally the supernatant was removed and centrifuged at 27,000 x g. The protein concentration in the highspeed pellet was adjusted to 1 mg/ml, and 0.3-ml samples were used for assay in the NAD(P)H oxidase
914
PROCTOR
assay. Assay mixtures (1.0 ml) contained 0.3 ml of pellet protein solution, 2.0 mM KCN, 0.1 M 2-(Nmorpholino)ethanesulfonic acid (pH 6.0) buffer, and either 0.2 mM NADPH or 0.8 mM NADH. After 30 min of incubation at 370C, reactions were stopped by the addition of 1 ml of 0.4 M HCl04, and precipitated protein was removed by centrifugation. A 0.1-ml supernatant was incubated with 0.15 ml of 10 N NaOH for 1 h at 20'C. Then 1.6 ml of water was added, and fluorescence was read on an Aminco-Bowman spectrofluorometer (excitation, 365 nm; emission, 448 nm). Nanomoles of NADH or NADPH formed were determined by comparison to relative fluorescence of standard curves, run in parallel daily. All determinations were performed in duplicate. Direct addition of LPS to the assay mixture was found not to alter oxidase activity.
RESULTS Effect of endotoxin on neutrophil bactericidal activity. The effects of graded doses of LPS on PMN function are shown in Fig. 1. When no LPS was present, approximately 1.5 log10 bacteria were killed in 120 min. Preincubation of LPS with PMNs (Fig. 1A and C) inhibited killing of E. coli at 10 jig of LPS per ml or greater and resulted in a dose-dependent reduction in the killing of S. aureus. When LPS was added simultaneously with E. coli or S. aureus (Fig. 1B and D), PMN bactericidal action was again reduced, but to a lesser extent. Of particular interest were the rates of killing of E. coli during the first 30 min. With simultaneous addition of 10 to 100 pg of LPS per ml and E. coli to PMNs (Fig. 1B), the rates of killing progressed rapidly for 30 min and then plateaued. After 30 min of incubation, the contours of the growth curves in Fig. 1B were nearly identical to those in Fig. 1A. That these bacteria were resistant to killing by 10% serum was demonstrated by their growth when no PMNs were present (Fig. 1). Uptake of bacteria was not affected by the presence of LPS, as determined by the percentage of PMNs that had phagocytized bacteria or by the number of bacteria phagocytized per PMN. Effect of endotoxin on neutrophil oxygen consumption. Neutrophils rapidly consumed oxygen when phagocytizing E. coli, S. aureus, or zymosan (Table 1). Concomitant addition of LPS with these same particles allowed the phagocytically induced increase in oxygen consumption to proceed normally. In contrast, preincubation of LPS with PMNs resulted in a marked reduction of oxygen consumption during phagocytosis. Challenge of PMNs with LPS alone resulted in rates of oxygen consumption nearly identical to that of resting PMNs, even when reaction rates were monitored for as long as 45 min. Addition of 0.1 ml of 0.01 M KCN to the reaction mixtures did not alter oxygen uptake in
INFECT. IMMUN.
two of the experiments done in duplicate. Since the bacteria were in lag phase (lyophilized and reconstituted for 30 to 60 min before use), they did not show oxygen uptake when added to the reaction chamber alone. Effect of LPS on neutrophil HMPS activity. Challenge of neutrophils with E. coli or S. aureus resulted in an approximate 10-fold increase in [1-'4C]glucose metabolism and a 2-fold increase in [6-'4C]glucose metabolism over resting values (Table 2). Addition of endotoxin resulted in a twofold (10 Mg) to threefold (100 Mg) increase in [1-'4C]glucose oxidation and a 40 to 75% increase in [6-'4C]glucose oxidation. Effect of LPS on NBT dye reduction. Reduction of NBT dye by PMNs was enhanced by addition of either latex spheres or LPS (Table 3). This reduction increased with increasing concentrations of LPS. Addition of larger quantities of LPS (up to 1,000 4g/ml) resulted in no greater reduction than occurred at 100 tg/ml. Because particle-induced NBT reduction by neutrophils is reportedly mediated solely by superoxide (an oxygen-derived radical) in whole PMNs (6), we elected to measure LPS-induced NBT reduction anaerobically (Fig. 2). A consistent increase in NBT reduction occurred whether the cells were held aerobically or anaerobically for 5 h. A larger mean increase in optical density was seen under aerobic (0.067 ± 0.021) conditions as compared to anaerobic conditions (0.039 ± 0.009). Also, total change in optical density was smaller after the PMNs were held (both aerobic and anaerobic) for 5 h at 4°C than when cells were used immediately. Cell suspensions were warmed for 30 min at 37°C before use to allow microtubule function to be restored. Effect of LPS on PMN CL. When PMNs were challenged with concentrations of LPS from 0.1 to 100 pg/ml, there was no change in CL as compared to resting cells. Table 4 summarizes the data for E. coli O111:B4 LPS observed at 25 min. Challenge with endotoxins from E. coli 055:B5 (Difco) and Salmonella typhimurium (Difco) also failed to induce CL. Also, when observations were performed at 5- to 15-min intervals over a 1- to 180-min time span, LPS did not stimulate PMN CL. Furthermore, no enhancement of neutrophil CL followed lipid A challenge (Table 4). To be certain that LPS was entering the PMNs, LPS was bound to latex particles. Whether LPS was absent, added simultaneously, or bound to the latex beads, no alteration in latex-induced CL was noted (Table 4). Addition of S. aureus and E. coli to PMNs resulted in a 20- to 30-fold increase in CL (Table 5). This increase was not affected by the simul-
VOL. 25, 1979
ENDOTOXIN IN VITRO INTERACTIONS
915
70 s0o 50#9 /0
()
6.00
1.Efc FIG.~~~
fenooi 0
nnutohlbceicdlatvt. 0
30
60
90
120
0
0 30
60
Ms 10#9 90
eu,
BS
n
120
TIME (min) FIG. 1. Effect of endotoxin on neutrophil bactericidal activity. 5 x 106 PMNs, 10% serum, HBSS, and bacteria (S. aureus or E. coli 0:6) were incubated with LPS (O to 100 pg/ml). (A) and (C) Data when LPS was preincubated with PMNs; (B) and (D) effects of simultaneous additions. Each point represents the mean of five to six experiments done in duplicate. One standard deviation is shown for 120-min values.
taneous addition of LPS, whereas simultaneously added lipid A caused a consistent, but not statistically significant, reduction. In contrast, preincubation of neutrophils with LPS or lipid A for 30 min produced a highly significant reduction in bacteria-induced CL. A similar reduction in ConA-stimulated CL was noted in those PMNs preincubated with LPS (Table 5). Because LPS is known to activate the alternative pathway of complement (18, 32), studies were undertaken to assess the role of complement in the LPS-altered CL response of PMNs. When heated serum replaced fresh serum (Fig. 3), a marked reduction in light production was
obtained. Nevertheless, the PMNs preincubated with LPS still showed less CL in heated serum than those PMNs that were either unexposed to LPS or added to the system simultaneously with LPS. PMNs challenged with preopsonized bacteria in heated serum showed the same CL values as those challenged with bacteria in fresh serum, whereas the LPS-preincubated PMNs showed significantly less light production. Effect of endotoxin on neutrophil superoxide production. To be sure that LPS was not decreasing CL by light quenching, by scavenging light-producing radicals, or by enhancing superoxide dismutase activity, 02 formation
916
INFECT. IMMUN.
PROCTOR
TABLE 1. Effect of endotoxin in particle-induced neutrophil oxygen consumption Oxygen consumption' by: P
Particles
PMNs + LPS-PI
PMNs + LPS
PMNs
4.31 ± 0.4 (13) 4.87 ± 0.5 (13) 4.38 ± 0.5 (14) None
