Veterinary Immunology and Immunopathology, 31 ( 1992 ) 241-253
241
Elsevier Science Publishers B.V., Amsterdam
Secretory activity of equine polymorphonuclear leukocytes: stimulus specificity and priming effects of bacterial lipopolysaccharide P h i l i p N. Bochsler, D a v i d O. S l a u s o n a n d N a n c y R. N e i l s e n
Department of Pathobiology, College of Veterinary Medicine, Universityof Tennessee, Knoxville, TN 37901-1071, USA (Accepted 15 April 1991 )
ABSTRACT Bochsler, P.N. Slauson, D.O. and Neilsen, N.R., 1992. Secretory activity of equine polymorphonuclear leukocytes: stimulus specificity and priming effects of bacterial lipopolysaccharide. Vet. lmmunol. ImmunopathoL, 3 h 241-253. Neutrophil (PMN) contributions to the acute inflammatory process and host defense include generation ofbioreactive oxygen metabolites and secretion of granule enzymes. We assessed equine PMN secretion using several PMN stimuli, singly and in combination with bacterial lipopolysaccharide (LPS). LPS avidly associated with equine PMN, as shown by strong PMN labeling with FITC-conjugated LPS. LPS alone (1 or 10/~g m l - ' ) was a weak stimulus for PMN superoxide anion (O~-) generation, but preincubation with LPS followed by phorbol ester (PMA, I 0 ng ml -~ ) significantly augmented (P< 0.01 ) secretion of OF ( 19.38 nmol O~- per 2 X 10 6 PMN per 5 min) over the amount generated by PMA stimulation alone (13.75 nmol OF ). A qualitatively similar, but smaller OFgeneration response occurred when either opsonized zymosan or recombinant human C5a was used as the PMN stimulus. Arachidonic acid (ArA; 50-200/~M ) was a potent stimulus, with secreted OF levels similar to those from PMA-stimulated PMN. Preincubation of PMN with either the formyl peptide, fMLP, or platelet-activatingfactor before stimulation with ArA did not significantly increase O~- generation over levels obtained using ArA alone. Release of PMN granule enzymes was also quantitated. A small amount of lysozyme secretion resulted when PMN were exposed to LPS alone (8.20% of total cell content ), and PMA stimulation caused marked release ofPMN lysozyme (44.45%). Nonspecific proteolytic activity in PMN supernatants, assessed by cleavage of a collagen-rich substrate, was minimal with LPS as a sole stimulus (5.08%). There was significant proteolytic activity (P< 0.01 ) in supernatants from PMA-stimulated PMN (27.21%), and preincubation with LPS followed by PMA stimulation slightly enhanced (P< 0.05 ) the release of PMN proteases (34.62%). The activities offlglucuronidase, acid phosphatase, and alkaline phosphatase were minimal in PMN supernatants when using LPS and PMA as stimuli. The activity of PMN granule enzymes was found to be sensitive to the presence of normal equine serum, and proteolytic activity was markedly reduced (80.13% reduction) in the presence of 10% pooled serum. ABBREVIATIONS ACD, acid citrate dextrose; ArA, arachidonic acid; FITC, fluorescein isothiocyanate; fMLP, formyl peptide, f-Met-Leu-Phe; HBSS, Hanks' Balanced Salt Solution; LDH, lactate dehydrogenase; LPS, lipopolysaccharide; NADH, nicotinamide adenine dinucleotide; OZ, opsonized zymosan; PBS, phos-
© 1992 Elsevier Science Publishers B.V. All rights reserved 0165-2427/92/$05.00
242
P.N. BOCHSLERET AL.
phate-buffered solution; PMA, phorbal 12-rnyristate 13-acetate; PMN, neutrophil; PNP, p-nitrophenyl phosphate; SOD, superoxide dismutase.
INTRODUCTION
Neutrophils (PMN) are important cellular components of host defense and the acute inflammatory process, and PMN-generated bioreactive oxygen species and granule enzymes mediate microbicidal activity (Colditz et al., 1988 ). PMN respond to a plethora of inflammatory mediators, and patterns of functional responses often overlap for various phlogistic stimuli. Bacterial lipopolysaccharide (LPS) is an integral component of the cell wall of gram-negative bacteria, and may be released and act as an inflammatory signal during the course of host infections (Movat et al., 1987). Peripheral blood leukocytes interact with LPS, as has been well-documented by many investigators (Berry, 1985 ). Trace concentrations of LPS will augment metabolic and secretory responses of human PMN when subsequently exposed to a second stimulus, such as the formyl peptide, f-Met-Leu-Phe (fMLP), or the phorbol ester, phorbol 12-myristate 13-acetate (PMA), and this state of heightened PMN responsiveness has been referred to as "priming" or "the primed state" (Guthrie et al., 1984; Fittschen et al., 1988 ). Increased generation of oxygen metabolites upon LPS exposure may be beneficial to the host, since these oxygen species, along with granule enzymes, mediate killing of ingested bacteria. LPS from the intruding bacteria may thus act as a signal for enhanced PMNmediated antibacterial activity. Excessive extracellular release of cytotoxic PMN-derived mediators may, however, result in inappropriate damage to host tissues. Equidae are severely and sometimes lethally affected by systemic release of bacterial LPS (Burrows, 1981; Ward et al., 1987 ). Because LPS has profound effects on human PMN in vitro, we examined the effects of LPS o n secretory functions of equine peripheral blood PMN. We found that LPS, as a sole PMN stimulus, had minimal stimulatory activity in most PMN secretion assays included in this study. However, preincubation of equine PMN with LPS, followed by stimulation with a second PMN agonist resulted in moderately enhanced superoxide anion (O~-) generation. A slight increase in activity of some proteases from supernatants of LPS-exposed PMN also occurred. METHODS
Isolation of equine PMN Blood was collected from normal horses via jugular venipuncture into acid citrate dextrose (ACD) at a ratio of 1:9 ACD to blood. Packed cell volume,
SECRETORY ACTIVITY OF EQUINE NEUTROPHILS
243
total leukocyte count, differential white cell count, and cytologic examination were performed to assure conformance with normal values and morphology. Neutrophils ( P M N ) were isolated from other cells and whole blood constituents using a modified Ficoll-Hypaque density centrifugation method (Ferrante and Thong, 1980). Isolated P M N were resuspended in Hank's Balanced Salt Solution (HBSS) supplemented with Ca 2+ and Mg 2+. Ficoll-Hypaque, HBSS, and sterile saline used in cell isolation were tested with the limulus amoebocyte lysate gel-clot test and found to contain less than 0.1 ng LPS m l - 1. Viability was assessed with trypan blue dye exclusion, and viability and P M N purity were typically 96-99%.
Neutrophil superoxide anion generation Superoxide anion ( 0 7 ) generation was measured as the superoxide dismutase (SOD)-inhibitable reduction of ferricytochrome-c (Babior et al., 1970). Neutrophils were added to cuvettes or test tubes to final concentrations of 2 X 106 P M N ml-1 in HBSS with Ca 2+ and Mg 2+. After warming the P M N to 37°C, warm ferricytochrome-c was added (0.8 mg m l - ~), followed by the stimuli. In assays utilizing LPS or other potential P M N priming agents, the P M N were preincubated with LPS ( 1 or 10/tg m l - ~from E. coli 0111 : B4 ), other stimuli, or with buffer (controls) for 1 h at 37°C before stimulation with a second agonist. P M N were preincubated with cytochalasin B (5/lg ml-~ ) for 10 rain at 37 °C in some assays before subsequent stimulation with recombinant h u m a n complement fragment 5a (rhC5a). Superoxide dismutase (SOD) was also used (40/tg m l - l ) as a control for the specificity of the reaction, and in other control tubes spontaneous O~- generation was quantitared ( < 1.0 nmol O~- ). Most experiments used a fixed time endpoint of 5 min at 37°C (or 15 min when using opsonized zymosan), then samples were rapidly chilled in an ice-bath, centrifuged (600 Xg for 10 min ), the cell-free supernatant transferred to an optical cuvette, and the absorbance of the supernatant determined at 550 nm. The data were expressed as nmol O~-/2 X 106 P M N 5 min-~ and were corrected for any spontaneous O~- generation; an extinction coefficient of 21.1 was used for ferricytochrome-c.
Neutrophil Enzyme Assays For assays of P M N enzyme secretion, P M N suspensions (50X 106 P M N ml-~ ) were incubated for a total of 1 h at 37 °C in HBSS with Ca 2+ and Mg 2+. P M N treatments were as follows: HBSS alone; LPS ( 1/tg m l - ~) from E. coli 0111 : B4; 45 min with HBSS followed by 15 min with PMA ( l/tg m l - ' ); 45 min with LPS ( 1/~g m l - ~) followed by 15 min with PMA ( 1/tg m l - ~). Maximal release of enzymes from control P M N was obtained by Triton X-100 (0. 1% ) lysis of P M N at 4 ° C for 1 h. After treatment all samples were chilled
244
P,N. BOCHSLER ET AL.
to 4°C, centrifuged ( 2 0 0 0 × g for 15 m i n ) and the supernatants were collected. Supernatants were frozen at - 2 0 ° C until assays were conducted (within 1 week), except for LDH samples which were kept at 4 ° C. Values were reported as a percent of total measurable cellular enzyme content as determined from Triton X-100 (0.1%)-lysed PMN. fl-Glucuronidase was assayed in P M N supernatants using phenolphthalein mono-fl-glucuronic acid (PGA) as a substrate (Fishman et al., 1967; Gallin et al., 1982 ). Lysozyme was quantitated in P M N supernatants by the ability to hydrolyze fl-1,4glucosidic linkages that occur in the mucopeptide cell wall structure of Micrococcus lysodeikticus (Litwack, 1955; Gallin et al., 1982) PMN secretory proteolytic activity was measured using Azocoll in a manner similar to the procedure recommended by Calbiochem. Azocoll is insoluble, powdered cowhide linked to a red dye (azo-dye-bound collagen) with a myriad assortment of peptide linkages, and cleavage of any of these linkages by proteolytic enzymes releases the azo-dye into the medium. Bacterial collagenase from Clostridium histolyticum ( 10, 5, 1 or 0.5/~g ml -~ ) was substituted for the PMN supernatants in some tubes to serve as a positive control. After 3 h at 37°C proteolytic activity of supernatants was assessed as cleavage of azo-dye-bound protein by absorbance at 520 nm. Normal pooled equine serum (total biuret p r o t e i n = 6.98 g d l - 1, a l b u m i n = 3.35 g dl- L) was added to some reaction mixtures at 10% final concentration in order to assess the ability of serum to inhibit P M N proteolytic activity. Acid phosphatase and alkaline phosphatase activity of P M N supernatants was measured as the hydrolysis of p-nitrophenyl phosphate ( P N P ) under acid conditions (Bretz and Boggliolini, 1974; Babior and Cohon, 1981 ), and alkaline conditions (DeChatelet and Cooper, 1970; Lindhart and Watler, 1974), respectively. For both assays, the absorbance determined at 400 n m was proportional to the enzymatic activity. Lactate dehydrogenase ( L D H ) was measured in supernatants from LPS-exposed PMN by the LDH-catalyzed conversion of pyruvate to lactate and the concomitant oxidation of nicotinamide adenine dinucleotide (NADH) reduced to NAD (Bergmeyer and Bernt, 1963; Wright et al., 1977 ). The presence of LDH was assayed as the decrease in absorbance at 340 nm due to consumption of NADH.
Opsonized zymosan Zymosan was suspended in saline solution, vortexed repeatedly, boiled for 30 min, pelleted by centrifugation ( 2 0 0 0 × g for 15 m i n ) , and washed twice in saline followed by similar centrifugation prior to opsonization. Prepared zymosan was incubated with plasma for 30 min at 37 °C with constant agitation, cooled to 4 oC, and centrifuged (2000 × g for 15 min). The supernatant plasma was decanted, and the pellet of opsonized zymosan (OZ) was washed
SECRETORY ACTIVITY OF EQUINE NEUTROPHILS
245
twice with phosphate-buffered saline (PBS; 0.01 M, pH 7.3) with 0.1 M EDTA, and subsequently washed twice with PBS.
Association of LPS with PMN Fluorescein isothiocyanate-conjugated LPS (FITC-LPS) from E. coli 0111 :B4 was incubated with PMN ( 10× 106 m1-1 ) at 37°C in HBSS buffer at concentrations of 1, 10 or 100/tg m l - 1. FITC-LPS was also preincubated for 30 min with polymyxin-B sulfate (20/tg m l - ' ) at 22 ° C and then similarly incubated with PMN. Controls included PMN incubated with HBSS, with 1, 10 or 100 #g m l - 1 of non-FITC-conjugated LPS from E. coli 0111 : B4, PMN incubated with unconjugated FITC in HBSS at a concentration equaling that in the FITC-LPS preparation, and PMN incubated with polymyxin-B alone. After incubation, slide preparations were immediately viewed using immunofluorescence microscopy.
Sources of materials Reagents and materials used in superoxide anion generation and most enzyme assays were purchased from Sigma Chemicals, St. Louis, MO. These included: zymosan A, PAF, LPS, FITC-LPS, limulus amoebocyte lysate gelclot test, phorbol 12-myrisate 13-acetate (PMA), ferricytochrome-c, superoxide dismutase, polymyxin-B sulfate, recombinant human C5a, and ficoll. Other sources were: Hypaque (Hypaque sodium 50%, diatrizoate sodium injection USP, Winthrop Labs, New York); Hanks' Balanced Salt Solution (Gibco, Grand Island, New York); arachidonic acid (ArA; Nu-Chek Prep., Elysian, MN); bacterial collagenase from Clostridium histolyticum, 178 units mg-1 (Worthington Biochemicals, Malvern, PA); Azocoll, < 50 mesh (Calbiochem, La Jolla, CA). Equine serum and plasma were obtained from clinically normal horses maintained as blood donors.
Statistical analys& The data were analyzed for statistical significance using a one-sample, twotailed Student's t-test. Each PMN-treatment group was individually compared with the appropriate control group of PMN run in parallel. RESULTS
Association of FITC-LPS with PMN PMN incubated at 37 ° C with 100/tg m l - 1 of FITC-LPS exhibited intense fluorescence (Fig. 1 ), with diminished fluorescence also seen at 10/tg ml-1
246
P.N. BOCHSLER ET AL
Fig. 1. Association of LPS with equine PMN. FITC-conjugated LPS avidly associated with isolated peripheral blood equine PMN. Fluorescence micrograph, X 600. and 1/tg ml-~ FITC-LPS. Fluorescence was intense within the cytosol, but absent in the area of t h e nucleus. F I T C - L P S labeling of P M N was inhibited by preincubation of F I T C - L P S for 30 min with polymyxin-B sulfate; polymyxin-B is known to bind to the lipid-A portion of LPS (Morrison and Jacobs, 1976 ). Labeling with F I T C - L P S was also reduced when P M N were incubated at 4 ° C rather than 37 ° C. F I T C - L P S also similarly associated with equine peripheral blood lymphocytes and monocytes.
Superoxide anion generation PMA at a concentration of 10 ng m l - ~induced O~- generation ( 13.75 nmol O z 2 × 106 P M N - ~ 5 m i n - ~) from PMN, and this was a maximal or near maximal stimulus of P M A (Table 1 ). Increasing the PMA concentration to 2 #g m l - 1 did not augment O~- generation ( 11.07 nmol O~- ) over that seen with the 10 ng ml-1 concentration o f PMA. P M A at 1 ng ml-~ was a very weak P M N stimulus, and LPS at 1 or 10 #g m l - ~ was similarly weak when used as a sole agonist. LPS at either of these concentrations was non-toxic to equine P M N as d e t e r m i n e d by LDH-release assays. Neutrophils preincubated with LPS ( 1/~g m l - ~) and subsequently stimulated with a low concentration of PMA ( 1 ng m l - 1) generated m o r e O~- than with either LPS or PMA alone ( P < 0.05 ), although the total a m o u n t o f O~- generated was very low (2.20 nmol O~- ; Table 1 ). W h e n preincubated with LPS ( 1 #g m l - 1) and
SECRETORY ACTIVITYOF EQUINE NEUTROPHILS
247
TABLE1
Superoxide anion (O~-) generation from equine P M N Stimulus
O~- ~
N
PMA 1 0 n g m l -~ LPS 1/zg m1-1 LPS ( 1/tg) + P M A ( 10 ng) LPS 10/tg ml -~ LPS ( 1 0 / t g ) + P M A ( 1 0 n g ) PMA
13.75_+7.02 0.24_+0.33 19.37_+ 9.30** 0.68_+0.53 17.93_+8.32"*
14 10 14 11 14
0.90_+ 0.81 0.86_+0.21 2.20_+ 1.16" 4.37 _+1.97 1.48_+0.88 7.10_+2.48"* 1.65 _+0.39 1.01 _+0.19 4.09 _+0.21 ** 16.72 _+4.71"*
6 6 6 6 6 6 4 4 4 4
0.67 _+0.63 15.89_+6.22 12.26_+57.4 15.01 _+4.47
5 5 5 5
1 ngml -I LPS 1/tgml - l LPS (1/tg) + P M A (1 ng)
Opsonized zymosan LPS 1/zgml -~ LPS (1/zg) + O Z rhC5a 10 -6 M LPS 1/zg ml - l LPS + rhC5a Cytochalasin B ( 5/~g m l - l ) + rhC5a
Arachidonic acid 5/zM 50/zM 100/zM 200/zM
nmol O~- 2 × ! 06 P M N - t 5 min _+SD, except for OZ = × 15 min. *P< 0.05, compared with PMA alone; **P< 0.01, compared with PMA (or OZ, or rhC5a) alone.
subsequently stimulated with 10 ng m l - 1 PMA, neutrophils generated significantly more O~- (19.38 nmol) than with PMA alone ( P < 0 . 0 1 ) . A similar result was observed when the concentration of LPS was increased to l0 pg m l - 1 ( 17.93 nmol O~- ). In separate experiments, we found that O~- generation from PMA-stimulated equine PMN was unaffected by 5 rain pretreatment with the anti-inflammatory drugs dexamethasone ( 10-5 M ) and phenylbutazone ( 100 pg m l - ~). Opsonized zymosan was a weaker stimulus for O~- generation from equine PMN than was 10 ng ml-1 PMA, and the time period required for O~- generation assays when using OZ was longer (15 min) than that required for PMA ( 5 min ). When preincubated with LPS ( 1/tg m l - 1) and subsequently stimulated with OZ, PMN generated significantly more O~- ( P < 0.01 ) than with OZ alone (Table 1 ). Similarly, rhCSa was a weak stimulus when used alone, but preincubation with LPS or cytochalasin-B before stimulation with rhC5a caused moderate (LPS) or marked (cytochalasin-B) enhancement of O~- generation ( P < 0 . 0 1 ) . ArA ( 50-200/tM) was a potent stimulus, inducing O~- generation at levels
248
P.N. BOCHSLER ET AL.
equal to or greater than did PMA (Table 1 ). Below 50 #M, marked variability in data was observed. Neither PAF nor FMLP showed evidence of stimulatory activity. When used in combination, preincubation of neutrophils with either PAF ( 1 0 - ~ - 1 0 -7 M) or FMLP ( 10-9-10 -4 M) did not significantly augment O~- generation when PMN were then stimulated by ArA. P M N granule enzyme release Equine PMN exposed to LPS as a sole stimulus secreted a small but statistically significant amount ( 8.20%, P < 0.01 ) of total cell content of lysozyme (Table 2). PMA (1 /~g ml -~ ) produced significant lysozyme release ( P < 0.01 ), although preincubation of PMN with LPS before PMA stimulation did not cause significant enhancement. Release offl-glucuronidase from neutrophils tested with LPS ( 1 #g m l - ~), PMA ( 1/tg m l - ~), or LPS followed by PMA, did not increase over the baseline amount for buffer-treated PMN (Table 2 ), although large quantities of fl-glucuronidase were detected in the supernatants of lysed PMN. Similar resuits were observed for acid and alkaline phosphatase assays, with little of the phosphatases detected in supernatants of PMN stimulated with LPS or PMA, but strong activity from lysed PMN supernatants (Table 2 ). When PMN were incubated with LPS there was minimal proteolytic activity in supernatants (5.08%) using the AzocoU assay (Table 2 ). Phorbol ester ( 1 fig ml -~ ) stimulated significant secretion of PMN proteases (27.21%; P
241
Elsevier Science Publishers B.V., Amsterdam
Secretory activity of equine polymorphonuclear leukocytes: stimulus specificity and priming effects of bacterial lipopolysaccharide P h i l i p N. Bochsler, D a v i d O. S l a u s o n a n d N a n c y R. N e i l s e n
Department of Pathobiology, College of Veterinary Medicine, Universityof Tennessee, Knoxville, TN 37901-1071, USA (Accepted 15 April 1991 )
ABSTRACT Bochsler, P.N. Slauson, D.O. and Neilsen, N.R., 1992. Secretory activity of equine polymorphonuclear leukocytes: stimulus specificity and priming effects of bacterial lipopolysaccharide. Vet. lmmunol. ImmunopathoL, 3 h 241-253. Neutrophil (PMN) contributions to the acute inflammatory process and host defense include generation ofbioreactive oxygen metabolites and secretion of granule enzymes. We assessed equine PMN secretion using several PMN stimuli, singly and in combination with bacterial lipopolysaccharide (LPS). LPS avidly associated with equine PMN, as shown by strong PMN labeling with FITC-conjugated LPS. LPS alone (1 or 10/~g m l - ' ) was a weak stimulus for PMN superoxide anion (O~-) generation, but preincubation with LPS followed by phorbol ester (PMA, I 0 ng ml -~ ) significantly augmented (P< 0.01 ) secretion of OF ( 19.38 nmol O~- per 2 X 10 6 PMN per 5 min) over the amount generated by PMA stimulation alone (13.75 nmol OF ). A qualitatively similar, but smaller OFgeneration response occurred when either opsonized zymosan or recombinant human C5a was used as the PMN stimulus. Arachidonic acid (ArA; 50-200/~M ) was a potent stimulus, with secreted OF levels similar to those from PMA-stimulated PMN. Preincubation of PMN with either the formyl peptide, fMLP, or platelet-activatingfactor before stimulation with ArA did not significantly increase O~- generation over levels obtained using ArA alone. Release of PMN granule enzymes was also quantitated. A small amount of lysozyme secretion resulted when PMN were exposed to LPS alone (8.20% of total cell content ), and PMA stimulation caused marked release ofPMN lysozyme (44.45%). Nonspecific proteolytic activity in PMN supernatants, assessed by cleavage of a collagen-rich substrate, was minimal with LPS as a sole stimulus (5.08%). There was significant proteolytic activity (P< 0.01 ) in supernatants from PMA-stimulated PMN (27.21%), and preincubation with LPS followed by PMA stimulation slightly enhanced (P< 0.05 ) the release of PMN proteases (34.62%). The activities offlglucuronidase, acid phosphatase, and alkaline phosphatase were minimal in PMN supernatants when using LPS and PMA as stimuli. The activity of PMN granule enzymes was found to be sensitive to the presence of normal equine serum, and proteolytic activity was markedly reduced (80.13% reduction) in the presence of 10% pooled serum. ABBREVIATIONS ACD, acid citrate dextrose; ArA, arachidonic acid; FITC, fluorescein isothiocyanate; fMLP, formyl peptide, f-Met-Leu-Phe; HBSS, Hanks' Balanced Salt Solution; LDH, lactate dehydrogenase; LPS, lipopolysaccharide; NADH, nicotinamide adenine dinucleotide; OZ, opsonized zymosan; PBS, phos-
© 1992 Elsevier Science Publishers B.V. All rights reserved 0165-2427/92/$05.00
242
P.N. BOCHSLERET AL.
phate-buffered solution; PMA, phorbal 12-rnyristate 13-acetate; PMN, neutrophil; PNP, p-nitrophenyl phosphate; SOD, superoxide dismutase.
INTRODUCTION
Neutrophils (PMN) are important cellular components of host defense and the acute inflammatory process, and PMN-generated bioreactive oxygen species and granule enzymes mediate microbicidal activity (Colditz et al., 1988 ). PMN respond to a plethora of inflammatory mediators, and patterns of functional responses often overlap for various phlogistic stimuli. Bacterial lipopolysaccharide (LPS) is an integral component of the cell wall of gram-negative bacteria, and may be released and act as an inflammatory signal during the course of host infections (Movat et al., 1987). Peripheral blood leukocytes interact with LPS, as has been well-documented by many investigators (Berry, 1985 ). Trace concentrations of LPS will augment metabolic and secretory responses of human PMN when subsequently exposed to a second stimulus, such as the formyl peptide, f-Met-Leu-Phe (fMLP), or the phorbol ester, phorbol 12-myristate 13-acetate (PMA), and this state of heightened PMN responsiveness has been referred to as "priming" or "the primed state" (Guthrie et al., 1984; Fittschen et al., 1988 ). Increased generation of oxygen metabolites upon LPS exposure may be beneficial to the host, since these oxygen species, along with granule enzymes, mediate killing of ingested bacteria. LPS from the intruding bacteria may thus act as a signal for enhanced PMNmediated antibacterial activity. Excessive extracellular release of cytotoxic PMN-derived mediators may, however, result in inappropriate damage to host tissues. Equidae are severely and sometimes lethally affected by systemic release of bacterial LPS (Burrows, 1981; Ward et al., 1987 ). Because LPS has profound effects on human PMN in vitro, we examined the effects of LPS o n secretory functions of equine peripheral blood PMN. We found that LPS, as a sole PMN stimulus, had minimal stimulatory activity in most PMN secretion assays included in this study. However, preincubation of equine PMN with LPS, followed by stimulation with a second PMN agonist resulted in moderately enhanced superoxide anion (O~-) generation. A slight increase in activity of some proteases from supernatants of LPS-exposed PMN also occurred. METHODS
Isolation of equine PMN Blood was collected from normal horses via jugular venipuncture into acid citrate dextrose (ACD) at a ratio of 1:9 ACD to blood. Packed cell volume,
SECRETORY ACTIVITY OF EQUINE NEUTROPHILS
243
total leukocyte count, differential white cell count, and cytologic examination were performed to assure conformance with normal values and morphology. Neutrophils ( P M N ) were isolated from other cells and whole blood constituents using a modified Ficoll-Hypaque density centrifugation method (Ferrante and Thong, 1980). Isolated P M N were resuspended in Hank's Balanced Salt Solution (HBSS) supplemented with Ca 2+ and Mg 2+. Ficoll-Hypaque, HBSS, and sterile saline used in cell isolation were tested with the limulus amoebocyte lysate gel-clot test and found to contain less than 0.1 ng LPS m l - 1. Viability was assessed with trypan blue dye exclusion, and viability and P M N purity were typically 96-99%.
Neutrophil superoxide anion generation Superoxide anion ( 0 7 ) generation was measured as the superoxide dismutase (SOD)-inhibitable reduction of ferricytochrome-c (Babior et al., 1970). Neutrophils were added to cuvettes or test tubes to final concentrations of 2 X 106 P M N ml-1 in HBSS with Ca 2+ and Mg 2+. After warming the P M N to 37°C, warm ferricytochrome-c was added (0.8 mg m l - ~), followed by the stimuli. In assays utilizing LPS or other potential P M N priming agents, the P M N were preincubated with LPS ( 1 or 10/tg m l - ~from E. coli 0111 : B4 ), other stimuli, or with buffer (controls) for 1 h at 37°C before stimulation with a second agonist. P M N were preincubated with cytochalasin B (5/lg ml-~ ) for 10 rain at 37 °C in some assays before subsequent stimulation with recombinant h u m a n complement fragment 5a (rhC5a). Superoxide dismutase (SOD) was also used (40/tg m l - l ) as a control for the specificity of the reaction, and in other control tubes spontaneous O~- generation was quantitared ( < 1.0 nmol O~- ). Most experiments used a fixed time endpoint of 5 min at 37°C (or 15 min when using opsonized zymosan), then samples were rapidly chilled in an ice-bath, centrifuged (600 Xg for 10 min ), the cell-free supernatant transferred to an optical cuvette, and the absorbance of the supernatant determined at 550 nm. The data were expressed as nmol O~-/2 X 106 P M N 5 min-~ and were corrected for any spontaneous O~- generation; an extinction coefficient of 21.1 was used for ferricytochrome-c.
Neutrophil Enzyme Assays For assays of P M N enzyme secretion, P M N suspensions (50X 106 P M N ml-~ ) were incubated for a total of 1 h at 37 °C in HBSS with Ca 2+ and Mg 2+. P M N treatments were as follows: HBSS alone; LPS ( 1/tg m l - ~) from E. coli 0111 : B4; 45 min with HBSS followed by 15 min with PMA ( l/tg m l - ' ); 45 min with LPS ( 1/~g m l - ~) followed by 15 min with PMA ( 1/tg m l - ~). Maximal release of enzymes from control P M N was obtained by Triton X-100 (0. 1% ) lysis of P M N at 4 ° C for 1 h. After treatment all samples were chilled
244
P,N. BOCHSLER ET AL.
to 4°C, centrifuged ( 2 0 0 0 × g for 15 m i n ) and the supernatants were collected. Supernatants were frozen at - 2 0 ° C until assays were conducted (within 1 week), except for LDH samples which were kept at 4 ° C. Values were reported as a percent of total measurable cellular enzyme content as determined from Triton X-100 (0.1%)-lysed PMN. fl-Glucuronidase was assayed in P M N supernatants using phenolphthalein mono-fl-glucuronic acid (PGA) as a substrate (Fishman et al., 1967; Gallin et al., 1982 ). Lysozyme was quantitated in P M N supernatants by the ability to hydrolyze fl-1,4glucosidic linkages that occur in the mucopeptide cell wall structure of Micrococcus lysodeikticus (Litwack, 1955; Gallin et al., 1982) PMN secretory proteolytic activity was measured using Azocoll in a manner similar to the procedure recommended by Calbiochem. Azocoll is insoluble, powdered cowhide linked to a red dye (azo-dye-bound collagen) with a myriad assortment of peptide linkages, and cleavage of any of these linkages by proteolytic enzymes releases the azo-dye into the medium. Bacterial collagenase from Clostridium histolyticum ( 10, 5, 1 or 0.5/~g ml -~ ) was substituted for the PMN supernatants in some tubes to serve as a positive control. After 3 h at 37°C proteolytic activity of supernatants was assessed as cleavage of azo-dye-bound protein by absorbance at 520 nm. Normal pooled equine serum (total biuret p r o t e i n = 6.98 g d l - 1, a l b u m i n = 3.35 g dl- L) was added to some reaction mixtures at 10% final concentration in order to assess the ability of serum to inhibit P M N proteolytic activity. Acid phosphatase and alkaline phosphatase activity of P M N supernatants was measured as the hydrolysis of p-nitrophenyl phosphate ( P N P ) under acid conditions (Bretz and Boggliolini, 1974; Babior and Cohon, 1981 ), and alkaline conditions (DeChatelet and Cooper, 1970; Lindhart and Watler, 1974), respectively. For both assays, the absorbance determined at 400 n m was proportional to the enzymatic activity. Lactate dehydrogenase ( L D H ) was measured in supernatants from LPS-exposed PMN by the LDH-catalyzed conversion of pyruvate to lactate and the concomitant oxidation of nicotinamide adenine dinucleotide (NADH) reduced to NAD (Bergmeyer and Bernt, 1963; Wright et al., 1977 ). The presence of LDH was assayed as the decrease in absorbance at 340 nm due to consumption of NADH.
Opsonized zymosan Zymosan was suspended in saline solution, vortexed repeatedly, boiled for 30 min, pelleted by centrifugation ( 2 0 0 0 × g for 15 m i n ) , and washed twice in saline followed by similar centrifugation prior to opsonization. Prepared zymosan was incubated with plasma for 30 min at 37 °C with constant agitation, cooled to 4 oC, and centrifuged (2000 × g for 15 min). The supernatant plasma was decanted, and the pellet of opsonized zymosan (OZ) was washed
SECRETORY ACTIVITY OF EQUINE NEUTROPHILS
245
twice with phosphate-buffered saline (PBS; 0.01 M, pH 7.3) with 0.1 M EDTA, and subsequently washed twice with PBS.
Association of LPS with PMN Fluorescein isothiocyanate-conjugated LPS (FITC-LPS) from E. coli 0111 :B4 was incubated with PMN ( 10× 106 m1-1 ) at 37°C in HBSS buffer at concentrations of 1, 10 or 100/tg m l - 1. FITC-LPS was also preincubated for 30 min with polymyxin-B sulfate (20/tg m l - ' ) at 22 ° C and then similarly incubated with PMN. Controls included PMN incubated with HBSS, with 1, 10 or 100 #g m l - 1 of non-FITC-conjugated LPS from E. coli 0111 : B4, PMN incubated with unconjugated FITC in HBSS at a concentration equaling that in the FITC-LPS preparation, and PMN incubated with polymyxin-B alone. After incubation, slide preparations were immediately viewed using immunofluorescence microscopy.
Sources of materials Reagents and materials used in superoxide anion generation and most enzyme assays were purchased from Sigma Chemicals, St. Louis, MO. These included: zymosan A, PAF, LPS, FITC-LPS, limulus amoebocyte lysate gelclot test, phorbol 12-myrisate 13-acetate (PMA), ferricytochrome-c, superoxide dismutase, polymyxin-B sulfate, recombinant human C5a, and ficoll. Other sources were: Hypaque (Hypaque sodium 50%, diatrizoate sodium injection USP, Winthrop Labs, New York); Hanks' Balanced Salt Solution (Gibco, Grand Island, New York); arachidonic acid (ArA; Nu-Chek Prep., Elysian, MN); bacterial collagenase from Clostridium histolyticum, 178 units mg-1 (Worthington Biochemicals, Malvern, PA); Azocoll, < 50 mesh (Calbiochem, La Jolla, CA). Equine serum and plasma were obtained from clinically normal horses maintained as blood donors.
Statistical analys& The data were analyzed for statistical significance using a one-sample, twotailed Student's t-test. Each PMN-treatment group was individually compared with the appropriate control group of PMN run in parallel. RESULTS
Association of FITC-LPS with PMN PMN incubated at 37 ° C with 100/tg m l - 1 of FITC-LPS exhibited intense fluorescence (Fig. 1 ), with diminished fluorescence also seen at 10/tg ml-1
246
P.N. BOCHSLER ET AL
Fig. 1. Association of LPS with equine PMN. FITC-conjugated LPS avidly associated with isolated peripheral blood equine PMN. Fluorescence micrograph, X 600. and 1/tg ml-~ FITC-LPS. Fluorescence was intense within the cytosol, but absent in the area of t h e nucleus. F I T C - L P S labeling of P M N was inhibited by preincubation of F I T C - L P S for 30 min with polymyxin-B sulfate; polymyxin-B is known to bind to the lipid-A portion of LPS (Morrison and Jacobs, 1976 ). Labeling with F I T C - L P S was also reduced when P M N were incubated at 4 ° C rather than 37 ° C. F I T C - L P S also similarly associated with equine peripheral blood lymphocytes and monocytes.
Superoxide anion generation PMA at a concentration of 10 ng m l - ~induced O~- generation ( 13.75 nmol O z 2 × 106 P M N - ~ 5 m i n - ~) from PMN, and this was a maximal or near maximal stimulus of P M A (Table 1 ). Increasing the PMA concentration to 2 #g m l - 1 did not augment O~- generation ( 11.07 nmol O~- ) over that seen with the 10 ng ml-1 concentration o f PMA. P M A at 1 ng ml-~ was a very weak P M N stimulus, and LPS at 1 or 10 #g m l - ~ was similarly weak when used as a sole agonist. LPS at either of these concentrations was non-toxic to equine P M N as d e t e r m i n e d by LDH-release assays. Neutrophils preincubated with LPS ( 1/~g m l - ~) and subsequently stimulated with a low concentration of PMA ( 1 ng m l - 1) generated m o r e O~- than with either LPS or PMA alone ( P < 0.05 ), although the total a m o u n t o f O~- generated was very low (2.20 nmol O~- ; Table 1 ). W h e n preincubated with LPS ( 1 #g m l - 1) and
SECRETORY ACTIVITYOF EQUINE NEUTROPHILS
247
TABLE1
Superoxide anion (O~-) generation from equine P M N Stimulus
O~- ~
N
PMA 1 0 n g m l -~ LPS 1/zg m1-1 LPS ( 1/tg) + P M A ( 10 ng) LPS 10/tg ml -~ LPS ( 1 0 / t g ) + P M A ( 1 0 n g ) PMA
13.75_+7.02 0.24_+0.33 19.37_+ 9.30** 0.68_+0.53 17.93_+8.32"*
14 10 14 11 14
0.90_+ 0.81 0.86_+0.21 2.20_+ 1.16" 4.37 _+1.97 1.48_+0.88 7.10_+2.48"* 1.65 _+0.39 1.01 _+0.19 4.09 _+0.21 ** 16.72 _+4.71"*
6 6 6 6 6 6 4 4 4 4
0.67 _+0.63 15.89_+6.22 12.26_+57.4 15.01 _+4.47
5 5 5 5
1 ngml -I LPS 1/tgml - l LPS (1/tg) + P M A (1 ng)
Opsonized zymosan LPS 1/zgml -~ LPS (1/zg) + O Z rhC5a 10 -6 M LPS 1/zg ml - l LPS + rhC5a Cytochalasin B ( 5/~g m l - l ) + rhC5a
Arachidonic acid 5/zM 50/zM 100/zM 200/zM
nmol O~- 2 × ! 06 P M N - t 5 min _+SD, except for OZ = × 15 min. *P< 0.05, compared with PMA alone; **P< 0.01, compared with PMA (or OZ, or rhC5a) alone.
subsequently stimulated with 10 ng m l - 1 PMA, neutrophils generated significantly more O~- (19.38 nmol) than with PMA alone ( P < 0 . 0 1 ) . A similar result was observed when the concentration of LPS was increased to l0 pg m l - 1 ( 17.93 nmol O~- ). In separate experiments, we found that O~- generation from PMA-stimulated equine PMN was unaffected by 5 rain pretreatment with the anti-inflammatory drugs dexamethasone ( 10-5 M ) and phenylbutazone ( 100 pg m l - ~). Opsonized zymosan was a weaker stimulus for O~- generation from equine PMN than was 10 ng ml-1 PMA, and the time period required for O~- generation assays when using OZ was longer (15 min) than that required for PMA ( 5 min ). When preincubated with LPS ( 1/tg m l - 1) and subsequently stimulated with OZ, PMN generated significantly more O~- ( P < 0.01 ) than with OZ alone (Table 1 ). Similarly, rhCSa was a weak stimulus when used alone, but preincubation with LPS or cytochalasin-B before stimulation with rhC5a caused moderate (LPS) or marked (cytochalasin-B) enhancement of O~- generation ( P < 0 . 0 1 ) . ArA ( 50-200/tM) was a potent stimulus, inducing O~- generation at levels
248
P.N. BOCHSLER ET AL.
equal to or greater than did PMA (Table 1 ). Below 50 #M, marked variability in data was observed. Neither PAF nor FMLP showed evidence of stimulatory activity. When used in combination, preincubation of neutrophils with either PAF ( 1 0 - ~ - 1 0 -7 M) or FMLP ( 10-9-10 -4 M) did not significantly augment O~- generation when PMN were then stimulated by ArA. P M N granule enzyme release Equine PMN exposed to LPS as a sole stimulus secreted a small but statistically significant amount ( 8.20%, P < 0.01 ) of total cell content of lysozyme (Table 2). PMA (1 /~g ml -~ ) produced significant lysozyme release ( P < 0.01 ), although preincubation of PMN with LPS before PMA stimulation did not cause significant enhancement. Release offl-glucuronidase from neutrophils tested with LPS ( 1 #g m l - ~), PMA ( 1/tg m l - ~), or LPS followed by PMA, did not increase over the baseline amount for buffer-treated PMN (Table 2 ), although large quantities of fl-glucuronidase were detected in the supernatants of lysed PMN. Similar resuits were observed for acid and alkaline phosphatase assays, with little of the phosphatases detected in supernatants of PMN stimulated with LPS or PMA, but strong activity from lysed PMN supernatants (Table 2 ). When PMN were incubated with LPS there was minimal proteolytic activity in supernatants (5.08%) using the AzocoU assay (Table 2 ). Phorbol ester ( 1 fig ml -~ ) stimulated significant secretion of PMN proteases (27.21%; P
