Immunology 1990 70 82-87
Human tumour necrosis factor-alpha (TNF-a) directly stimulates arachidonic acid release in human neutrophils Y. H. ATKINSON, A. W. MURRAY,* S. KRILISt M. A. VADAS & A. F. LOPEZ The Division of Human Immunology, Institute of Medical and Veterinary Science, Adelaide, * School of Biological Sciences, Flinders University, and t Department of Medicine, St George's Hospital, New South Wales, Australia Acceptedfor publication 20 December 1989
SUMMARY The ability of tumour necrosis factor-alpha (TNF-a) to directly stimulate phospholipid turnover from human neutrophils was studied. Stimulation with recombinant human (rH)TNF-a induced the release of significant amounts of radioactivity from [3H]arachidonic acid-labelled neutrophils. This stimulation was equipotent to that induced by the bacterial tripeptide formyl-methionyl-leucylphenylalanine (FMLP). The time of maximum stimulated release varied between donors, with the most common maximal stimulation being 45 min. Dose-response experiments indicated that 1001000 U/ml rH TNF-a were required for the maximum stimulatory effect. High-performance liquid chromatography analysis of the supernatants revealed that the radioactivity was associated with arachidonic acid, but not with its metabolites, indicating that TNF-a stimulates the release of arachidonic acid from cellular phospholipids but does not stimulate its metabolism. A comparison of TNF-a with other cytokines indicated that stimulation of arachidonic acid release paralleled the 'priming' of neutrophils for enhanced superoxide production, raising the possibility that phospholipid turnover and priming of neutrophils are causally related.
mented, little is known about its direct effects, which could help explain its mechanism of action. In an effort to define this, it has been previously reported that rHTNF-a directly regulates the surface receptor Cr3 involved in adherence (Gamble et al., 1985) and the surface expression of FMLP receptors (Atkinson et al., 1988b). Since the regulation of membrane receptors and adherence by rHTNF-a all require changes in membrane properties, it was hypothesized that rHTNF-a may directly alter the phospholipid turnover of neutrophils. Here it is shown that rHTNF-a induces the release of arachidonic acid from neutrophils but does not stimulate the metabolism of this fatty acid. The stimulation of phospholipid turnover in neutrophils may play a role in the regulatory effects of rHTNF-a on neutrophil function.
INTRODUCTION Tumour necrosis factor-alpha (TNF-cx) is a macrophage derived protein with multiple biological activities (Beutler & Cerami, 1988). This cytokine is synthesized upon activation of macrophages by certain stimuli, such as endogenous cytokines, and bacterial products, such as lipopolysaccharide (LPS), and may play a central role in inflammation (Movat et al., 1987). For example recombinant (r) human (H) TNF-a has been shown to 'prime' neutrophil function by enhancing the neutrophils' response to an activating agent. Thus rHTNF-a increases the phagocytosis and degranulation of neutrophils in response to both opsonized and unopsonized zymosan (Klebanoff et al., 1986) and enhances the respiratory burst and inhibits the chemotactic responsiveness of neutrophils to formyl-methionylleucyl-phenylalanine (FMLP) (Atkinson et al., 1988b). Although the enhancing effects of rHTNF-a are well docu-
MATERIALS AND METHODS
Purification of human neutrophils Neutrophils were obtained from the peripheral blood of healthy volunteers and prepared essentially as described elsewhere (Atkinson et al., 1988b). The cells (always > 95% pure neutrophils) were prepared in serum-free RPMI-1640 with 20 mm HEPES buffer and antibiotics. Following purification, the neutrophils were resuspended in RPMI-1640+0-1% bovine serum albumin (BSA). The assays were carried out in this medium unless otherwise stated.
Abbreviations: BSA, bovine serum albumin; FMLP, N-formyl-
methionyl-leucyl-phenylalanine; GM-CSF, granulocyte-macrophage colony-stimulating factor; H, human; HETE, hydroxy-eicosatetraenoic acid; HPLC, high-performance liquid chromatography; LPS, lipopolysaccharide; IL-I#, interleukin-l beta; LT, leukotriene; PAF, plateletactivating factor; PG, prostaglandin; r, recombinant; TCA, trichloroacetic acid; TNF-a, tumour necrosis factor-alpha. Correspondence: Dr Y. H. Atkinson, The Division of Human Immunology, Institute of Veterinary Science, Adelaide, South Australia.
82
Human tumour necrosis factor-alpha Cytokines The rHTNF-a used was from lot number N9017AX containing 0 5 mg/ml of protein. This cytokine was generously provided by Genentech, South San Francisco, CA. The molecule was produced in Escherichia coli (Pennica et al., 1985) and was approximately 99% pure. The LPS content of this solution was determined to be less than 0-03 endotoxin U/ml, as measured by the manufacturers using the Limulus amebocyte lysate assay. Dilutions from stock material were made weekly into RPMI1640 plus 0- 1% BSA and stored at 4°. The final concentration of LPS present in incubation suspensions was less than 1-5 x 10-7 endotoxin U/ml. The rH granulocyte-macrophage colony-stimulating factor (GM-CSF) was from lot number C009DI-701, containing 1-4 x 107 U/mg protein (when tested on chronic myeloid leukaemic cells), was 99-8% pure, and was a generous gift from Dr S. Clark, Genetics Institute, Cambridge, MA. This protein was obtained from the supernatant of COS cells and purified as described elsewhere (Wong et al., 1985). Dilutions from stock material were stored at -70° in RPMI-1640+ 10% fetal calf serum (FCS). The rH interleukin-lbeta (IL-I1,) was provided by Immunex, Seattle, WA and contained 108 U/mg protein when assayed by thymocyte mitogenesis (Kronheim et al., 1985). Stock material was stored at -70° and diluted into RPMI1640 +0-1% BSA for use. Chemicals was purchased from New England Nuclear, Boston, MA. Ionophore A23187 was from Behring Diagnostics, La Jolla, CA. Essentially fatty acidfree BSA (Fraction V), FMLP, butylated hydroxytoluene and arachidonic acid were obtained from Sigma Chemical Co., St Louis, MO.
[1-_4C] arachidonic acid (100 Ci/mmol)
[3H]arachidonic acid release Purified neutrophils (107/ml) were incubated in medium with 2 x 106 pCi/ml arachidonic acid, [5, 6, 8, 9, 11, 12, 14, 15-3H(N)] for 60 min at 37°. The cells were then washed three times in cold medium and resuspended to 107/ml. Aliquots of cells (3 x 106 cells) were incubated for various times at 370 in the presence or absence of various stimuli. An aliquot of neutrophils was incubated with 4 ym A23187 for 10 min at 370 as a positive control. The cells were placed on ice to terminate the incubation then centrifuged at 400 g for 5 min at 4°. An aliquot of the supernatant was transferred into a scintillation vial with 10 ml Aqueous Counting Scintillant II (Amersham, Arlington Heights, IL), and radioactivity measured using a Packard TriCarb 2000CA liquid scintillation counter. Results are expressed as d.p.m. per 2 x 106 neutrophils.
Reverse-phase HPLC analysis of [3H]arachidonic acid-labelled neutrophil supernatants Neutrophil supernatants were prepared as described above. The supernatants were removed into fresh tubes and extracted in an ice-cold solution of 80% methanol (final concentration) overnight. The precipitate was centrifuged at 1000 g for 2 min at 40, and the methanol supernatants dried in a Speed Vac (Savant Instruments Inc., Farmingdale, NY). Samples were stored
83
under argon at -20° until analysed by high-performance liquid chromatography (HPLC). The samples were extracted into 80% ethanol, evaporated to dryness using a Speed Vac rotary evaporator (Savant Instruments, Hicksville, NY) and reconstituted in the RP-HPLC mobile phase. The arachidonic acid metabolites were separated by RP-HPLC on a Beckman C18 5 jm Ultrasphere ODS 4-6 mm x 25 cm column (Beckman Intruments Inc., Berkeley, CA) with a Waters HPLC system (two model 510 pumps, model 680 gradient controller and a model U6K injector). The buffer system consisted of a step-wise gradient of methanol (Mallinkrodt, Melbourne) and 0-01% acetic acid in distilled deionized water, pH 5.6, adjusted with ammonia, modified from Henke, Kouzan & Eling (1988) (30 min at 55%, 30 min at 65%, 20 min at 73% and 15 min at 100%) at a flow rate of 0-8 ml/min. This system separated prostaglandin (PG), leukotriene (LT), hydroxy-eicosatetraenoic acid (HETE) groups and arachidonic acid. All solvents were of HPLC grade and were filtered and degased before use. Fractions were collected into plastic minivials (Packard Instrument Co., Downers Grove, IL), and radioactivity measured by liquid scintillation counting. The HPLC system was calibrated using a series of tritiated PG, LT and HETE. 6-keto-PGFoo (6KPGFia) and Thromboxane B2 (TxB2) were from Amersham, and PGD2, PGF200, LTB4, LTD4, 15H, 12H, 5H, and arachidonic acid (AA) were from New England Nuclear. The unlabelled 20-OH LTB4 standard used was a gift from Dr B. Spur (Institut Henri Beaufour, Paris, France), and synthesized according to Soberman et al. (1988). The retention time of this compound was determined by monitoring UV absorbance at 270 nm. The recovery of radioactivity from chromatography was 96 5 + 18 8%. Extraction of neutrophil lipids Cellular lipids from neutrophils loaded and stimulated as described above, were extracted by a slight modification of the method of Wertz & Mueller (1978). The cells were resuspended to 107/ml and precipitated in 5% trichloroacetic acid (TCA) for 20 min at 4°. The TCA insoluble material was then dissolved in 5 ml chloroform/methanol (2: 1, v/v) + 50 Mg/ml butylated hydroxytoluene (Sigma). The aqueous phase was partitioned with 1 ml 0 04% CaCl2. Following this, the organic phase was further extracted with 1 ml solvent and 1 ml CaCl2. The lipid extract was dried under a stream of nitrogen at 250 and stored under argon at -20° until analysis by thin-layer chromatography.
Thin-layer chromatography Total lipids from 107 cells were dissolved in chloroform methanol (2:1, v/v) + 50 gg/ml butylated hydroxitoluene for application to aluminium-backed thin-layer chromatography plates (silica gel 60, 0-2 mm thick; Merck, Schuchardt, FRG). To separate the [3H]arachidonic acid from the cell lipids, a solvent system consisting of hexane: diethylether: formic acid (80:20:2, v/v, room temperature) was used. The plates were sprayed with En3hance (New England Nuclear) and autoradiographed on Hyperfilm-3H (Amersham, Amersham, Bucks, U.K.) Fractions were then identified by visualization with I2 vapour. After evaporation of the I2 from the plate, the fractions were scraped into scintillation vials and quantified by liquid scintillation counting. Result are expressed as a percentage of incorporation into total cellular lipids.
Y. H. Atkinson et al.
84
10,000r TNF-a U) O
0
FMLP c
None
X 0
4000
CLm
6000 8000 10,000 12,000 D.p.m. / 2 x 106 neutrophils
Figure 1. rHTNF-a induces arachidonic acid release from neutrophils. [3H]arachidonic acid-labelled neutrophils (2 x 106) were incubated with medium, 10-7 M FMLP or 100 U/mi rHTNF-a for 45 min at 370. Results are shown as the mean (± SE) of 16 experiments performed in triplicate. The values obtained in the presence of TNF-a were significantly different (P < 0-001) from nil controls.
RESULTS
b
0
30 45 15 Time (min)
Figure 2. Time-dependent stimulation of arachidonic acid release. [3HJ arachidonic acid-labelled neutrophils were incubated with medium (0), 10-7 M FMLP (0) or 100 U/ml rHTNF-a (0) for different times at 37°. Results are shown as the mean (± SE) of 20 experiments performed in triplicate. The values obtained with TNF-a were significantly different from medium control values at 30 min (P < 0 005) and 45 min (P
Human tumour necrosis factor-alpha (TNF-a) directly stimulates arachidonic acid release in human neutrophils Y. H. ATKINSON, A. W. MURRAY,* S. KRILISt M. A. VADAS & A. F. LOPEZ The Division of Human Immunology, Institute of Medical and Veterinary Science, Adelaide, * School of Biological Sciences, Flinders University, and t Department of Medicine, St George's Hospital, New South Wales, Australia Acceptedfor publication 20 December 1989
SUMMARY The ability of tumour necrosis factor-alpha (TNF-a) to directly stimulate phospholipid turnover from human neutrophils was studied. Stimulation with recombinant human (rH)TNF-a induced the release of significant amounts of radioactivity from [3H]arachidonic acid-labelled neutrophils. This stimulation was equipotent to that induced by the bacterial tripeptide formyl-methionyl-leucylphenylalanine (FMLP). The time of maximum stimulated release varied between donors, with the most common maximal stimulation being 45 min. Dose-response experiments indicated that 1001000 U/ml rH TNF-a were required for the maximum stimulatory effect. High-performance liquid chromatography analysis of the supernatants revealed that the radioactivity was associated with arachidonic acid, but not with its metabolites, indicating that TNF-a stimulates the release of arachidonic acid from cellular phospholipids but does not stimulate its metabolism. A comparison of TNF-a with other cytokines indicated that stimulation of arachidonic acid release paralleled the 'priming' of neutrophils for enhanced superoxide production, raising the possibility that phospholipid turnover and priming of neutrophils are causally related.
mented, little is known about its direct effects, which could help explain its mechanism of action. In an effort to define this, it has been previously reported that rHTNF-a directly regulates the surface receptor Cr3 involved in adherence (Gamble et al., 1985) and the surface expression of FMLP receptors (Atkinson et al., 1988b). Since the regulation of membrane receptors and adherence by rHTNF-a all require changes in membrane properties, it was hypothesized that rHTNF-a may directly alter the phospholipid turnover of neutrophils. Here it is shown that rHTNF-a induces the release of arachidonic acid from neutrophils but does not stimulate the metabolism of this fatty acid. The stimulation of phospholipid turnover in neutrophils may play a role in the regulatory effects of rHTNF-a on neutrophil function.
INTRODUCTION Tumour necrosis factor-alpha (TNF-cx) is a macrophage derived protein with multiple biological activities (Beutler & Cerami, 1988). This cytokine is synthesized upon activation of macrophages by certain stimuli, such as endogenous cytokines, and bacterial products, such as lipopolysaccharide (LPS), and may play a central role in inflammation (Movat et al., 1987). For example recombinant (r) human (H) TNF-a has been shown to 'prime' neutrophil function by enhancing the neutrophils' response to an activating agent. Thus rHTNF-a increases the phagocytosis and degranulation of neutrophils in response to both opsonized and unopsonized zymosan (Klebanoff et al., 1986) and enhances the respiratory burst and inhibits the chemotactic responsiveness of neutrophils to formyl-methionylleucyl-phenylalanine (FMLP) (Atkinson et al., 1988b). Although the enhancing effects of rHTNF-a are well docu-
MATERIALS AND METHODS
Purification of human neutrophils Neutrophils were obtained from the peripheral blood of healthy volunteers and prepared essentially as described elsewhere (Atkinson et al., 1988b). The cells (always > 95% pure neutrophils) were prepared in serum-free RPMI-1640 with 20 mm HEPES buffer and antibiotics. Following purification, the neutrophils were resuspended in RPMI-1640+0-1% bovine serum albumin (BSA). The assays were carried out in this medium unless otherwise stated.
Abbreviations: BSA, bovine serum albumin; FMLP, N-formyl-
methionyl-leucyl-phenylalanine; GM-CSF, granulocyte-macrophage colony-stimulating factor; H, human; HETE, hydroxy-eicosatetraenoic acid; HPLC, high-performance liquid chromatography; LPS, lipopolysaccharide; IL-I#, interleukin-l beta; LT, leukotriene; PAF, plateletactivating factor; PG, prostaglandin; r, recombinant; TCA, trichloroacetic acid; TNF-a, tumour necrosis factor-alpha. Correspondence: Dr Y. H. Atkinson, The Division of Human Immunology, Institute of Veterinary Science, Adelaide, South Australia.
82
Human tumour necrosis factor-alpha Cytokines The rHTNF-a used was from lot number N9017AX containing 0 5 mg/ml of protein. This cytokine was generously provided by Genentech, South San Francisco, CA. The molecule was produced in Escherichia coli (Pennica et al., 1985) and was approximately 99% pure. The LPS content of this solution was determined to be less than 0-03 endotoxin U/ml, as measured by the manufacturers using the Limulus amebocyte lysate assay. Dilutions from stock material were made weekly into RPMI1640 plus 0- 1% BSA and stored at 4°. The final concentration of LPS present in incubation suspensions was less than 1-5 x 10-7 endotoxin U/ml. The rH granulocyte-macrophage colony-stimulating factor (GM-CSF) was from lot number C009DI-701, containing 1-4 x 107 U/mg protein (when tested on chronic myeloid leukaemic cells), was 99-8% pure, and was a generous gift from Dr S. Clark, Genetics Institute, Cambridge, MA. This protein was obtained from the supernatant of COS cells and purified as described elsewhere (Wong et al., 1985). Dilutions from stock material were stored at -70° in RPMI-1640+ 10% fetal calf serum (FCS). The rH interleukin-lbeta (IL-I1,) was provided by Immunex, Seattle, WA and contained 108 U/mg protein when assayed by thymocyte mitogenesis (Kronheim et al., 1985). Stock material was stored at -70° and diluted into RPMI1640 +0-1% BSA for use. Chemicals was purchased from New England Nuclear, Boston, MA. Ionophore A23187 was from Behring Diagnostics, La Jolla, CA. Essentially fatty acidfree BSA (Fraction V), FMLP, butylated hydroxytoluene and arachidonic acid were obtained from Sigma Chemical Co., St Louis, MO.
[1-_4C] arachidonic acid (100 Ci/mmol)
[3H]arachidonic acid release Purified neutrophils (107/ml) were incubated in medium with 2 x 106 pCi/ml arachidonic acid, [5, 6, 8, 9, 11, 12, 14, 15-3H(N)] for 60 min at 37°. The cells were then washed three times in cold medium and resuspended to 107/ml. Aliquots of cells (3 x 106 cells) were incubated for various times at 370 in the presence or absence of various stimuli. An aliquot of neutrophils was incubated with 4 ym A23187 for 10 min at 370 as a positive control. The cells were placed on ice to terminate the incubation then centrifuged at 400 g for 5 min at 4°. An aliquot of the supernatant was transferred into a scintillation vial with 10 ml Aqueous Counting Scintillant II (Amersham, Arlington Heights, IL), and radioactivity measured using a Packard TriCarb 2000CA liquid scintillation counter. Results are expressed as d.p.m. per 2 x 106 neutrophils.
Reverse-phase HPLC analysis of [3H]arachidonic acid-labelled neutrophil supernatants Neutrophil supernatants were prepared as described above. The supernatants were removed into fresh tubes and extracted in an ice-cold solution of 80% methanol (final concentration) overnight. The precipitate was centrifuged at 1000 g for 2 min at 40, and the methanol supernatants dried in a Speed Vac (Savant Instruments Inc., Farmingdale, NY). Samples were stored
83
under argon at -20° until analysed by high-performance liquid chromatography (HPLC). The samples were extracted into 80% ethanol, evaporated to dryness using a Speed Vac rotary evaporator (Savant Instruments, Hicksville, NY) and reconstituted in the RP-HPLC mobile phase. The arachidonic acid metabolites were separated by RP-HPLC on a Beckman C18 5 jm Ultrasphere ODS 4-6 mm x 25 cm column (Beckman Intruments Inc., Berkeley, CA) with a Waters HPLC system (two model 510 pumps, model 680 gradient controller and a model U6K injector). The buffer system consisted of a step-wise gradient of methanol (Mallinkrodt, Melbourne) and 0-01% acetic acid in distilled deionized water, pH 5.6, adjusted with ammonia, modified from Henke, Kouzan & Eling (1988) (30 min at 55%, 30 min at 65%, 20 min at 73% and 15 min at 100%) at a flow rate of 0-8 ml/min. This system separated prostaglandin (PG), leukotriene (LT), hydroxy-eicosatetraenoic acid (HETE) groups and arachidonic acid. All solvents were of HPLC grade and were filtered and degased before use. Fractions were collected into plastic minivials (Packard Instrument Co., Downers Grove, IL), and radioactivity measured by liquid scintillation counting. The HPLC system was calibrated using a series of tritiated PG, LT and HETE. 6-keto-PGFoo (6KPGFia) and Thromboxane B2 (TxB2) were from Amersham, and PGD2, PGF200, LTB4, LTD4, 15H, 12H, 5H, and arachidonic acid (AA) were from New England Nuclear. The unlabelled 20-OH LTB4 standard used was a gift from Dr B. Spur (Institut Henri Beaufour, Paris, France), and synthesized according to Soberman et al. (1988). The retention time of this compound was determined by monitoring UV absorbance at 270 nm. The recovery of radioactivity from chromatography was 96 5 + 18 8%. Extraction of neutrophil lipids Cellular lipids from neutrophils loaded and stimulated as described above, were extracted by a slight modification of the method of Wertz & Mueller (1978). The cells were resuspended to 107/ml and precipitated in 5% trichloroacetic acid (TCA) for 20 min at 4°. The TCA insoluble material was then dissolved in 5 ml chloroform/methanol (2: 1, v/v) + 50 Mg/ml butylated hydroxytoluene (Sigma). The aqueous phase was partitioned with 1 ml 0 04% CaCl2. Following this, the organic phase was further extracted with 1 ml solvent and 1 ml CaCl2. The lipid extract was dried under a stream of nitrogen at 250 and stored under argon at -20° until analysis by thin-layer chromatography.
Thin-layer chromatography Total lipids from 107 cells were dissolved in chloroform methanol (2:1, v/v) + 50 gg/ml butylated hydroxitoluene for application to aluminium-backed thin-layer chromatography plates (silica gel 60, 0-2 mm thick; Merck, Schuchardt, FRG). To separate the [3H]arachidonic acid from the cell lipids, a solvent system consisting of hexane: diethylether: formic acid (80:20:2, v/v, room temperature) was used. The plates were sprayed with En3hance (New England Nuclear) and autoradiographed on Hyperfilm-3H (Amersham, Amersham, Bucks, U.K.) Fractions were then identified by visualization with I2 vapour. After evaporation of the I2 from the plate, the fractions were scraped into scintillation vials and quantified by liquid scintillation counting. Result are expressed as a percentage of incorporation into total cellular lipids.
Y. H. Atkinson et al.
84
10,000r TNF-a U) O
0
FMLP c
None
X 0
4000
CLm
6000 8000 10,000 12,000 D.p.m. / 2 x 106 neutrophils
Figure 1. rHTNF-a induces arachidonic acid release from neutrophils. [3H]arachidonic acid-labelled neutrophils (2 x 106) were incubated with medium, 10-7 M FMLP or 100 U/mi rHTNF-a for 45 min at 370. Results are shown as the mean (± SE) of 16 experiments performed in triplicate. The values obtained in the presence of TNF-a were significantly different (P < 0-001) from nil controls.
RESULTS
b
0
30 45 15 Time (min)
Figure 2. Time-dependent stimulation of arachidonic acid release. [3HJ arachidonic acid-labelled neutrophils were incubated with medium (0), 10-7 M FMLP (0) or 100 U/ml rHTNF-a (0) for different times at 37°. Results are shown as the mean (± SE) of 20 experiments performed in triplicate. The values obtained with TNF-a were significantly different from medium control values at 30 min (P < 0 005) and 45 min (P
