Veterinary Immunology and Immunopathology, 29 ( 1991 ) 15-30
15
Elsevier Science Publishers B.V., Amsterdam
Cytotoxic tumor necrosis factor activity produced by equine alveolar rnacrophages: preliminary characterization* R.J. MacKay a, R.R. King b, J.R. Dankert c, K.J. R e i s a'm a n d L.A. S k e l l e y a aDepartment of Large Animal ClinicalSciences, bDepartmentof Small Animal ClinicalSciences and CDepartment of Comparative and Experimental Pathology, Universityof Florida, Gainesville, FL 32610, USA (Accepted 7 August 1990)
ABSTRACT MacKay, R.J., King, R.R., Dankert, J.R., Reis, K.J. and Skelley, L.A., 1991. Cytotoxic tumor necrosis factor activity produced by equine alveolar macrophages: Preliminary characterization. Vet. Immunol. Immunopathol., 29:15-30. Blood monocytes and alveolar macrophages (AM) were harvested from foals (aged 46 days to 6 months) and cultured in either medium alone or medium containing l 0/~g/ml bacterial lipopolysaccharide (LPS). After 24 h, culture supernates were collected and analyzed for cytotoxic activity on sensitized L929 cells. Both monocytes and AM that had been treated with LPS produced significantly more cytotoxic activity than the same cell type exposed to medium lacking LPS. LPS-treated macrophages secreted significantly more cytotoxic activity ( 120_+ 17.8 U/ml) than did LPS-treated monocytes (47.3_+ 17.0 U/ml); however, constitutive production of cytotoxin by monocytes was higher ( 16.7 +_4. l versus 1.2 +_1.2 U/ml). The identification of the cytotoxin as tumor necrosis factor (TNF) was strongly suggested by its reactivity with a rabbit antiserum directed against the N-terminal 15 amino acids of human TNF. TNF secretion by AM increased in a dose-dependent manner between LPS concentrations of 0.0001 and l/~g/ml, then leveled off. Most of the cytotoxic TNF activity produced by AM was secreted within the first 8 h after initial contact with LPS. Macrophage supernatant TNF was stable over a pH range of 6-1 l, but lost activity when kept at a pH less than 6. Equine TNF also was destroyed by exposure for l h to temperatures more than 60 ° C. TNF bioactivity was recovered as a single peak after crude macrophage supernate was subjected to analysis by either anion exchange or gel filtration chromatography (molecular weight approximately 56 000). ABBREVIATIONS AM, alveolar macrophages; BA, bronchoalveolar; LPS, lipopolysaccharide; PBS, phosphate-buffered saline; TNF, tumor necrosis factor. *Supported in part by a research grant from the American Quarter Horse Association. Present address: Genex Corporation, Gaithersburg, MD 20877, USA.
0165-2427/91/$03.50
© 1991 m Elsevier Science Publishers B.V.
16
R.J. MACKAY ET AL.
INTRODUCTION
Tumor necrosis factor (TNF) activity originally was identified in the serum of BCG-infected mice that had been injected with large amounts of endotoxin (Carswell et al., 1975). When this serum was transferred into other mice bearing transplantable tumors, there was rapid, hemorrhagic destruction and involution of the tumors. Some, but not all, cultured tumor cell lines also were shown to be susceptible to the cytolytic effects of"tumor necrotizing" serum (Helson et al., 1975; Sugarman et al., 1985). Subsequently, TNF molecules from mice and several other animal species have been purified, and their eDNA cloned and sequenced (Aggarwal et al., 1985; Haranaka et al., 1985; Marmenout et al., 1985; Pennica and Goeddel, 1987). When analyzed under denaturing conditions, purified TNF (either native or recombinant-derived ) is a 17-17.5 kDa polypeptide (Aggarwal et al., 1985; Baglioni et al., 1985). On the basis of amino acid sequence analysis, TNF was found to be identical to cachectin, an activity that has been associated with the wasting diathesis of chronic trypanosomiasis and other diseases characterized by cachexia (Beutier et al., 1985a). In concert with other pleiotropic cytokines (most notably, interleuldns, colony stimulating factors and interferons), TNF influences many aspects of the immune and inflammatory responses. It has been shown to have a pivotal role in experimental endotoxic shock (Beutler et al., 1985c; Tracey et al., 1987) and circulating TNF has been found in humans and other animals with a variety of diseases (Balk;viii et al., 1987; Waage et al., 1987; Maury and Teppo, 1989). On a cellular level, TNF is reported to participate in the activation o f l v m n h n ~ v t a ~ nantraphile a n d r.~ at.r,~nh~,~,~e {;,"iok~,,,-~¢¢--+ ,~1 l a O g . Hackett et al., 1988; Heidenreich et al., 1988 ) and to induce the production and secretion of inflammatory mediators, procoagulant activity and acute phase reactants (Dayer et al., !985; Darlington et al., 1986; Nawroth and Stem, 1986). In addition to its proinflammatory effects, TNF has been shown to posses both antiviral-i~lducing and cell growth factor activities (Mestan et al., 1986; Vilcek et al., 1986). Mononuclear phagocytes are the most important cellular source of TNF in vivo and explanted monocytes and macrophages from several different anatomic sites and from several species have been induced to secrete high levels of TNF in culture (Matthews and Watkins, 1978; Chen et al., 1985; Decker et al., 1987; Becker et al., 1989; Ruco et al., 1989). The most potent known stimulus for TNF production in vitro is endotoxin lipopolysaccharide (LPS) although viruses, plasmodial lysates, Candida albicans and certain Grampositive organisms also are reported to stimulate TNF secretion by macrophages (Beutler et al., 1986a,b; Djeu et al., 1988; Wright et al., 1988). TNF has recently been identified in the serum of horses that had been treated with endotoxin (MacKay et al., 1991 ). In the present study we show
CYTOTOXIC TNF ACTIVITY BY EQUINE ALVEOLAR MACROPHAGES
17
that TNF is secreted in vitro by endotoxin-stimulated equine alveolar macrophages (AM) and blood monocytes and we report prelimin3ry physicochemical characterizations of the secreted molecule.
~IATERIAL$ AND METHODS
Animals Nine healthy suckling horse foals ages 46 days to 6 months and three adult horses of various breeds were used to provide cells for experiments. They were kept at pasture and received routine health care including vaccinations for EEE/WEE and tetanus.
Sample collection Immediately before sampling procedures were performed, foals were sedated with 0.5 mg/kg xylazine (Rompun; Mobay Corp., Shawnee, KS) and 0.05 mg/kg butorphanol tartrate (Torbugesic; Fort Dodge Labs., Fort Dodge, IA), both given intravenously. Forty milliliters of venous blood were collected in a syringe containing sodium heparin (final concentration, 10 U/ml blood), then bronchoalveolar lavage was performed as described (King and Raskin, 1989). In brief, a 0.6 × 145 cm fiberoptic endoscope (Machida Inc., Orangeburg, NY) was advanced via the nasotracheal route until it wedged in a fourth to fifth order bronchus. Bronchoalveolar (BA) cells were harvested through the endoscope's biopsy port. On average, 89% (263.5_+ 11.8 ml, X_+ SEM) of the total 300 ml delivered into the airways was recovered as lavage fluid. Heparinized blood and BA lavage fluid samples were kept on ice until processed (maximum of 30 min ).
Culture of alveolar macrophages BA cells were pelleted by centrifuging lavage fluid at 200×g for 10 min, then washed twice by resuspension and centrifugation in physiologic saline. Washed cells were resuspended in modified Eagle's minimal essential medium (Auto-POW MEM; Flow Labs., McLean, VA) supplemented with 10% newborn calf serum (Gibco Labs., Grand Island, NY), 2 mg/ml sodium bicarbonate, 2 mM glutamine, 100 U/ml potassium penicillin G, 100/lg/ml streptomycin and 15 mM HEPES buffer (HMEM/10% NCS). Differential cell counts were determined after staining with Wright Giemsa. Cell recoveries and differentials (X _+SEM ) for eight lavages (one from each of eight foals )
18
R.J. MACKAY ET AL.
were as follows: total cells 90.9_+ 10.6 × 106; macrophages 72.3 + 3.2%; lymphocytes 23.6_+ 3.6%, neutrophils 1.9_+0.45%; eosinophils 0.81 _+0.68%; mast cells 0.43_+0.24%; epithelial cells 0.68 _+0.35%. BA cells were adjusted to a concentration of 1.3× 106 ml (containing approximately 1 × 106 macrophages/ml) in HMEM/10% NCS and pipetted either into the 16 mm diameter wells of 24-well tissue culture plates (1 ml/well) or into 75 cm 2 flasks (for bulk culture). After 2 h incubation at 37°C in 5% CO2 in air, non adherent cells were removed by vigorously washing the plates three times with warm saline. Adherent cell monolayers prepared in this way were more than 95% AM as determined by differential staining of the bottoms of wells with Wright Giemsa. The composition of the tissue culture media added back to washed AM monolayers varied according to the type of experiment: for bulk TNF production, the stimulating medium was serum-free HMEM containing 10 /~g/ml LPS (Escherichia coli 0111 :B4, Sigma Chemical Co., St. Louis, MO); for dose-response experiments, HMEM/10% NCS ( 1 ml/weU) was supplemented with LPS concentrations from 0.0001 to 100/~g/ml; and for smallscale TNF production and time-course experiments, HMEM/10% NCS containing 10/~g/ml LPS was used. Except during time-course experiments, LPSstimulated AM were incubated at 37°C for 24 h. Plates were then centrifuged at 150×g for 1 min and the uppermost 0.7 ml of medium in each well was removed. Bulk-produced supernates removed from tissue culture flasks were f'lltered at 0.45/zm. All supe.rnates were stored at - 8 0 ° C until assayed (within 2 months).
Culture of blood monocytes • -~.,[oaLlaall-~u u I U U U
WGD
U l D L I I U U L ~ U JLLILLU J U
lUl ~JUly~31Ul3yl~ll~ U~III.IIIU~
tubes ( 12 ml/tube), then mixed with 20 ml phosphate-buffered saline (PBS and underlaid with 12 ml Histopaque-1077 (Sigma). After tubes had been centrifuged at 350 × g for 40 rain, cells were aspirated from the Histopaqueblood interface, then washed three times by centrifugation and rcsuspension in PBS ( 150×g for 10 min). All centrifugation steps were performed at 20°C. Cells collected in this way usually were more than 97% mononuclear cells (range 95-99%). Pelleted, washed, mononuclear cells were resuspended in HMEM/10% NCS and distributed into 100 mm diameter tissue culture dishes at 5 × 107 cells/dish. Each dish was incubated for 4 h at 37 °C, then nonadherent cells were removed by washing with warm PBS using seven cycles of gentle plate agitation followed by aspiration of supernatant medium. After dishes had been held on ice for an additional 45 min, adherent cells were recovered by scraping the surfaces of the dishes with a rubber policeman. By differential staining, more than 85% of these cells were monocytes (range 8593%). Processed cells were resuspended at a concentration of 1 × 106 monocytes/ml in HMEM/10% NCS with or without LPS ( 10/zg/kg) and distributed into the wells of 24-wel! plates at 1 ml/well. Plates were incubated at
CYTOTOXIC TNF ACTIVITY BY EQUINE ALVEOLAR MACROPHAGES
19
37°C for 24 h then supernatant media were harvested from the wells and frozen as described for culture of AM.
Cytotoxicity bioassayfor TNF TNF activity (in U / m l ) was quantified as cytotoxicity of samples for sensitized L929 cells, using procedures previously described (MacKay et al., 1991).
Anion-exchange chromatography Crude TNF-containing supernate, prepared by LPS stimulation of macrophages in serum-free medium, was concentrated 10-12 fold by vacuum dialysis, then dialyzed extensively against 100 mM tris-HCl, pH 7.8. Prior to the concentration step, crude supernates contained an average of 175 U / m l TNF activity and 6/lg/ml protein. Octyl glucoside (Sigma) was added to concentrated, dialyzed, supernate (final concentration 0.01%), then anion exchange chromatography was performed by a modification of a procedure previously described for mouse TNF (Beutler et al., 1985a). In brief, 5 ml of sample were loaded at room temperature onto a 0.5 cm × 5 cm Mono-Q FPLC column (Pharmacia LKB Biotechnology Inc., Piscataway, NJ) equilibrated with 100 mM Tris-HCl, pH 7.8. The column was then washed with o:~e bed volume of equilibration buffer and the sample was eluted with a 20 ml linear 100-500 mM Tris-HCl gradient applied at a flow rate of 2 ml/min. All column buffers contained 0.01% octyl glucoside. The eluate was monitored for protein content by absorbence at 280 am and fractions were assayed immediately on L929 cells far TNF activity
Gelfiltration chromatography Concentrated supernate (produced as for anion exchange chromatography) was dialyzed against column buffer ( 100 mM NaCI, 15 mM imidazole pH 6.8) and 1 ml was applied to a Sepharose CL-6B (Pharmacia) gel filtration column (2.5 c m × 30 cm ) and run at 6.5 ml/h. The eluate was monitored at 280 nm and collected fractions were assayed immediately for TNF activity on L929 cells.
Protein assay Column chromatography fractions which contained cytotoxic activity were pooled and protein content was quantified by the bicinchoninic acid method (Smith et al., 1985) using reagents obtained from Pierce Chemical, Rockford, IL.
Effect of heating One milliliter amounts of crude serum-free supernate were distributed into
20
R.J. MACKAY ET AL.
glass tubes and held for 1 h in a heated water bath at temperatures ranging from 20 to 100°C. After heat treatment, triplicate samples were held overnight at 4 ° C, then assayed for cytotoxic activity on L929 cells.
Effect of pH One millimeter amounts of serum-free supernate were placed into dialysis bags ( 12 000 MW cutoff) and each was dialyzed for 20 h against two changes of 40 ml of one of the following buffers: glycine-HCl pH 1-4, Na acetateacetic acid pH 4-6, Na2HPO4-NaH2PO4 pH 6-8, glycine-NaOH pH 8-11, all 50 mM, and PBS pH 7.2. Each treatment was performed in triplicate. Samples were then dialyzed overnight against three changes of 300 ml PBS and assayed for cytotoxic activity on L929 cells.
Immunosubtraction of TNF activity An antiserum raised in rabbits against a peptide comprising the N-terminal 15 amino acids of human TNI= (anti-Hu-TNF ( 1-15 ) ) was generously provided by Dr. Susan Socher of Merck, Sharp & Dohme Research Laboratories (West Point, PA). Anti-Hu-TNF ( 1-15 ) was prepared as described (Socher et al., 1987) and by ELISA had a binding titer for human TNF of 256 000. In preliminary experiments, this antiserum proved also to have low neutralizing activity (in solution) for equine TNF. A more detailed description of the activity against equine TNF of this and other anti-Hu-TNF antisera will be published elsewhere. Normal rabbit serum and anti-Hu-TNF(1-15) were heated at 56 °C before use in experiments. Equine TNF activity was adsorbed from solution by incubating the test samples with a 1/20 dilution of anti-Hu TNF ( 1-15 ) (or control normal rabbit serum ) for 4 h fallowed by 40 mg of a commercial Staphylococcus aureus preparation (IgGsorb; The Enzyme Center Inc, Malden, MA) for 2 h, all at 4 ° C. The insoluble complexes were pelleted in a microfuge at 12 000 × g for 4 min and the supernates immediately removed and assayed on L929 cells.
Statistical analysis The stimulating effect of LPS and TNF secretion by macrophages and monocytes was analyzed by subjecting the data for each cell type to paired Student's t-tests of the null hypothesis that/Zo < 0 (i.e. l-taii~J t-test) where /Zo was the mean difference between the amount of TNF secreted by LPStreated cells and the amount secreted by cells that had not received LPS. Macrophages and monocytes were treated as independent samples and mean TNF data collected from each cell type were compared by two-tailed t-tests using pooled estimates of variance. In either situation, a P value of 5% or less was considered significant.
21
CYTOTOXIC TNF ACTIVITY BY EQUINE ALVEOLAR MACROPHAGES
RESULTS
Production of TNF activity by alveolar macrophages and blood monocytes Alveolar macrophages and blood monocytes from each of six foals were cultured for 24 h with either medium or medium containing 10/~g/ml LPS. The cytotoxic activity in each of the culture supernates was quantified and the results are summarized in Table 1. The mean activity of supernates from LPS-stimulated macrophages and monocytes was significantly higher than that from cultures that were exposed to medium alone (Table 1 ). On a per cell TABLE 1 Production of cytotoxic activity (X _ SEM ) by alveolar macrophages and blood monocytes Cell source
Cytotoxic activity ( U / m l )
Macrophages Monocytes
Medium
LPS
1.2 + 0.5 16.7 + 4.1
120.0 + 17.8* 47.3 + 17.0**
Comparison of LPS- with medium-treated cells: *P< 0.0001 **P< 0.05.
Cytotoxic activity (Ulml)
.°°I 700
600 I 5OO 4OO
fi
300 200 100 0
med
-4
-3
-2
-1
0
1
2
LPS conc (LoglO/Jglrnl) Fig. 1. Effect of LPS concentretion on T N F production by equine AM. AM were harvested and prepared as monolayers in the ! 6 m m wells of 24-well plates as described in Materials and Methods. After exposure of AM for 24 h to the indicated LPS concentrations, supernates were harvested and assayed on L929 cells for cytotoxic TNF activity° Each solid line is the curve for cells from an individual foal. Each point is the mean of three replicate wells. The dotted line is the mean curve for the data from the three foals shown.
22
R.J. MACKAYET AL. Cytotoxic
activity
(U/ml)
700 600
4OO 300
.......... -------'---~
200
/
100
,.
"
/
0
~ 0
''"'~-~ p 8
4
L 12
i 16
, 20
I. 24
Time (hours)
Fig. 2. Time course for production of TNF activity ~y ~c,~;ne AM exposed to 10/zg/ml LPS. AM were exposed in 16 m m wells to medium containing LPS and supernatant media were harvested at the indicated times and assayed for T N F activity. Each solid line is the curve for the cells from an individual foal and each point on the line is the mean of triplicate wells. The dotted line is the mean curve for the data from the three foals shown.
TNF activity (Ulml) 8000 [ - ' - 00280
0D280 (ABS units X 100)
25
2600 ~ -B- TNF I~ 0.1-0.6M Trle
2000 [ 1500 l" / I ooo~ _ . - ~ 50
S
7
~
35
40
45
50
55 60 65 70 75 80 85 Fraction number Fig~ 3. Anion exchange chromatography of crude concentrated equine TNF. All buffers were 0.01% octyl glucoside. Five millimeter of sample was loaded onto a Mono-Q FPLC column (0.5 c m × 5 cm) equilibrated with 100 mM Tris-HCl, pH 7.8. After the column was washed with equilibration buffer, the sample was eluted in a 20 ml linear 100-500 mM Tris-HCl gradient. The eluate was monitored at 280 nm and collected 0.5 ml fractions were assayed immediately for TNF activity.
CYTOTOXITNF C ACTIVITY BYEQUINE ALVEOLAR MACROPHAGES
23
basis, LPS-treated macrophages secreted significantly more TNF than did LPS-treated monocytes from the same foals (P