135299-3 13 (199 1)
and Leukocyte Elastase inactivate Human Tumor Necrosis Factor and Lymphotoxin’
PHILIP SCUDERI,~ PATRICIA A. NEZ, MELINDA L. DUERR, BETTY J. WONG, AND CECELIA M. VALDEZ Arizona Cancer Center and the Departments of Anatomy and Microbiology and Immunology University of Arizona, Tucson,Arizona 85 721 Received July 6, 1990; accepted February 4, 1991 The addition of either cathepsin-G or leukocyte elastaseto endotoxin-stimulated human peripheral blood monocytes decreasedthe immunoreactive tumor necrosis factor (TNF) detected in culture supernatants in a concentration-dependent manner. Both enzymes also induced a loss of supernatant cytolytic activity as determined on the WEHItarget cell line. Incubation of recombinant human TNF and lymphotoxin (LT) with either cathepsin-G or leukocyte elastase resulted in a loss of cytokine bioactivity. Examination of enzyme-treated recombinant cytokines by gel electrophoresisrevealed that cathepsin-G cleaved LT into a 12.6-l&a fragment and leukocyte elastasefragmented LT into a 14.1-kDa product. On Western blots cathepsin-G and leukocyte elastasedegradedTNF into 1I- and 7.6~kDafragments,respectively. Incubating leukocyte elastase with plasma elastaseinhibitor cu-I-antitrypsin prevented the loss of recombinant TNF hioactivity and blocked the degradation of this cytokine. This study suggeststhat two of the most abundant neutrophil proteases,cathepsin-G and leukocyte elastase,may he important regulators of TNF and LT bioactivity. o 1991 Academic PKSS, h.
INTRODUCTION Tumor necrosis factor (TNF) and lymphotoxin (LT) are closely related cytokines which share a 30% amino acid sequence homology and a broad range of biological activities. TNF is produced primarily by members of the mononuclear phagocyte lineage (1); however, this cytokine can be secreted by interleukin 2 (IL-2)-activated lymphocytes in response to an as yet unidentified second signal provided by tumor cells (2). LT is primarily a product of lymphocytes (3, 4). Both of these pleiotropic cytokines bind to the same cell surface receptor (5) and are inflammatory mediators which can induce tissue injury (6, 7). The control TNF and LT bioactivity at sites of inflammation may be critical to the progress of the inflammatory and wound repair processes.This is suggestedby the capacity of TNF to stimulate fibroblast proliferation (8) and PGE2 secretion (9). In addition, TNF has been shown to activate neutrophils ( 10-12). Neutrophil granules contain both cathepsin-G and leukocyte elastase( 13). I Supported by a grant from the Green Valley Associatesin Cancer Research. r To whom correspondenceand reprint requestsshould be addressedat Miles Inc., 4th and Parker Streets. P.O. Box 1986, Berkeley, CA 94701. 299 0008-8749/91 $3.00 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.
Cathepsin-G has been shown to convert angiotensin I to angiotensin II ( 14, 15), while leukocyte elastaseis known to degrade collagen and as such is believed to induce the destruction of the lung in emphysema and chronic lung diseases(16). Becauseboth mononuclear phagocytes and neutrophils are present at sites of inflammation and have already been shown to interact in a number of ways, it was of interest to test the effectof the neutrophil serine proteasescathepsin-G and leukocyte elastaseon monocyte TNF secretion and bioactivity. Here we show that both proteasesdegradeand inactivate human TNF and LT. METHODS AND MATERIALS Cells and Culture Conditions Human peripheral blood was collected from healthy adult donors by venous puncture. The mononuclear cells were separatedby Ficoll-Hypaque (Pharmacia Inc., Piscataway, NJ) and suspendedat a density of 1 X 1O6cells/ml in complete tissue culture medium consisting of RPM1 1640 supplemented with 10% fetal calf serum, 2 mM Lglutamine, penicillin, and streptomycin. The endotoxin content of the complete medium was determined to be 150 pg/ml using a limulus amoebocyte lysate assay(Associatesof Cape Cod, Woodshole, MA). One hundred microliters of this cell suspension was added to each well of a 96-well assay plate and incubated at 37°C for 2 hr in a humidified incubator. After 2 hr, the nonadherent cells were removed and the adherent monocytes were washed six times with RPM1 1640 medium without fetal calf serum. Finally, 100 ~1of RPM1 1640 containing 1 Kg/ml of LPS 0 111:B4 (Difco Inc., Detroit, MI) and the appropriate concentration of either cathepsin-G or leukocyte elastase (CalBiochem Inc., San Diego, CA) were added. Culture supernatants were assayedfor secretedTNF after 18 hr with an ELISA system. WEHI- 164 cells used in cytotoxicity assayswere obtained from the American Type Culture Collection (Rockville, MD) and grown in RPM1 1640 complete medium. Reagents and Antibodies Recombinant human TNF, LT, and monoclonal antibody 6E with specificity for human TNF were gifts from Genentech Inc. (South San Francisco, CA). Rabbit antiserum specific for human TNF was produced in New Zealand white rabbits as previously described (17). Human cathepsin-G (4 U/mg), leukocyte elastase(40 U/mg), (Y-1-antitrypsin ((Y,-AT) and (Y-1-antichymotrypsin were purchased from CalBiochem Inc. The proteaseswere reconstituted in phosphate-buffered saline, pH 7.2, and stored at -70°C. Recombinant human IL-8 was purchased from R & D Systems,Inc. (Minneapolis, MN). Cytochalasin B, N-formyl-L-methionyl-L-leucyl+phenylalinine (lMLP), the leukocyte elastase substrate methoxysuccinyl-(Ala)2-Pro-Val para-nitroanilide, and the cathepsin-G substrateSue-(Ala)*-Pro-Phe NPhN03 were purchased from Sigma Inc. (St. Louis, MO). Recombinant human IL- 1cuwas a gift from Hoffman LaRoche (Nutley, NJ). TNF ELISA The ELISA assay for the detection of human TNF has been described previously (18). Briefly, 96-well Immulon II (Dynatech, Inc., McLean, VA) plastic plates were
coated with the TNF-specific monoclonal antibody 6E (6.25 rig/well). Culture supernatants were added to the wells and incubated for 1 hr at room temperature. The plates were washedthree times with phosphate-buffered saline containing 0.05% Tween 20, and a rabbit anti-TNF-specific antiserum was subsequently added. Following a similar incubation and wash procedure a peroxidase-conjugated goat anti-rabbit IgGspecific antiserum, which had been adsorbed against murine IgG (American Qualex, Inc., LaMirada, CA), was added. The peroxidase substrate used was 2,2’-azino-bis-3ethylbenzthiazoline-6-sulfonic acid (Sigma Inc.). Optical density readings were made on an ELISA reader set at 405 nm within 10 min of substrate addition. The quantity of TNF in culture supematants was determined by comparison with a set of standards made with recombinant human TNF. Cytotoxicity Assay The cytotoxic activity of the culture supematants was assessedusing a modification of the technique of Mosmann ( 19) by adding 100 ~1of these fluids to 100~1of complete medium containing 1 X lo4 WEHI- 164 target cells in each well of a 96-well plate. After incubation for 72 hr, each well received 50 ~1 of a 2 mg/ml solution of MTT dye substrate (3’-4,5-dimethylthiazoyl-2-yl-2,5-diphenyl-tetrazolium bromide) (Sigma Chemical Co.). After substrate addition the assayplates were incubated at 37°C for 4 hr. The unmetabolized MTT was removed by aspiration at the end of the incubation period, and the crystallized tetrazolium salt was solubilized by the addition of 100 ~1 of dimethyl sulfoxide. Adsorbance measurements were determined with a Titerteck Multiskan (Flow Labs, Inc., McLean, VA) at a wavelength of 570 nm (A&. The percentage of control MTT dye metabolized was calculated as (experimental AsTO) x 100. (control AjTO) (The control AsTOwas produced by incubating WEHI- 1640 tumor cells alone in a pattern on 96-well plates which took into account variations in cell growth in both outer and inner wells. All assayswere performed with a minimum of four replicates for each control and experimental condition. % Control AsTO=
SDS-PAGE Separation of TNF and LT fragments was accomplished on 20% polyacrylamide gels containing 10% glycerol, pH 9.8, which were run at 15 mA for 8 hr without cooling. Samples for electrophoresis were mixed with an equal volume of 2X sample buffer containing 1% SDS, 15%glycerol, 0.0 1% bromphenoi blue dye, and 15 mM 2ME. Each sample was heated to 100°C for 3 min prior to loading on gels. The gels were stained with a solution of 0.125% Coomassie blue R-250 in 50% methanol and 10% acetic acid. Gels were destained in 50% methanol and 10% acetic acid for 1 hr, followed by 12 hr in 7% acetic acid and 5% methanol. In some experiments 18 and 20% SDS gels (Novex Inc., Encinitas, CA) were used and silver stained using the method of Morrissey (20). Western Blot Two micrograms of either recombinant human TNF or LT was admixed with either 17.5 units of leukocyte elastaseor 0.7 units of cathepsin-G and incubated for 3 hr at
37°C. Controls in these experiments included equal amounts of TNF, LT, elastase, and cathepsin-G incubated alone, under identical conditions. Each sample was subsequently admixed with the SDS sample buffer, boiled, and run on 18%reducing gels. The separatedproteins were transferred to nitrocellulose membranes and probed with either a specific rabbit anti-human TNF or anti-LT antiserum using the methods of Towbin (21). Goat anti-rabbit antibody conjugated to alkaline phosphatase (BioRad, Inc., Richmond, CA) followed by BCIP/NBT substrate (p-nitro blue tetrazolium chloride and 5-bromo-4-chloro-3-indolyl-phosphate toluidine salt), prepared according to the instructions of the manufacturer (BioRad), were used to visualize the labeled proteins. Neutrophil Activation Peripheral blood was collected in heparinized tubes from three healthy adults. Neutrophils were separated from nucleated leukocytes and red cells using mono-poly resolving medium (Flow Inc.) and the contaminating red cells were subsequently lysed with sterile distilled water. The neutrophils were counted and resuspended in Hanks’ balanced salt solution (GIBCO Inc., Detroit, MI) containing 1%globulin-free bovine serum albumin (Sigma Inc.) at a density of 5 X lo6 cells/ml. The cells were either incubated alone for 1 hr at 37°C or pretreated by incubation with 2.5 pg/ml cytochalasin B for 5 min. Following this preincubation step the cells were either exposed to lMLP at a final concentration of 5.7 X lop7 M or recombinant human IL-8 at 10 nM for an additional 60 min at 37°C. As a control cytochalasin B and either fMLP or IL-8 were incubated without neutrophils under the sameconditions. The neutrophil culture supernatants and both fMLP and IL-8 controls were harvested, admixed with recombinant human TNF at concentration of 10 rig/ml, and incubated for 1 hr at 37°C. Each fluid was subsequently added to actinomycin D-treated WEHItarget cells and the cytotoxic effect was assessedafter 18 hr. The cytolytic activity of each sample was assessedby comparison to a standard amount of freshly prepared recombinant human TNF. Both human leukocyte elastase(22) and cathepsin-G (23) activity in the neutrophil culture supematants were assessedusing specific chromogenic substrates for each enzyme. RESULTS Cathepsin-G and Leukocyte Elastase Added to Plastic Adherent Human Leukocytes Reduces Immunoreactive TNF in Culture Supernatants Figure 1 shows that adding either purified cathepsin-G or leukocyte elastaseto LPSstimulated plastic adherent human peripheral blood leukocytes from five healthy adults decreasedthe quantity of immunoreactive TNF detected by ELISA. The reduction in supematant TNF was dependent on the concentration of enzyme added. The plastic adherent cells used in these studies were found to be 84% monocytes as determined by morphology (data not shown). Cathepsin-G and Leukocyte Elastase Reduce the Leukocyte Supernatant Cytolytic Activity To determine if the reduction in secreted TNF detected by ELISA in the previous experiment also resulted in a loss of cytolytic activity, monocyte culture supematants
TNF IS INACTIVATED
BY CATHEPSIN-G AND LEUKOCYTE
2.5 -E \ F
FIG. I. Effect of cathepsin-G and leukocyte elastaseon human monocyte TNF secretion. Plastic adherent monocytes were obtained from the Ficoll-Hypaque-separated peripheral blood mononuclear leukocytes of five healthy adults by suspending the cells in complete medium at a density of 1 X 106/ml and incubating 1 X IO5cells/well in 96-well plates for 2 hr at 37°C. The adherent cells were washedwith serum-free medium and resuspendedin 100 ~1of RPM1 1640 without fetal calf serum which contained 1 &ml LPS 01 I l:B4 and the indicated concentration of protease. After 18 hr the culture supematants were assayedfor secreted TNF by ELISA. Each point representsthe mean of four observations f the standard deviation.
were added to WEHItarget cells. Figure 2 shows that the untreated culture supernatant from LPS-stimulated monocytes which contained 3.2 rig/ml TNF by ELISA decreased target cell viability and that the cytolytic effect was lost by dilution. In contrast either the supernatant from cathepsin-G (0.04 U/ml)- or leukocyte elastase (0.02 U/ml)-treated, LPS-stimulated monocytes, had lost all cytolytic activity. Although it was unlikely that either cathepsin-G or leukocyte elastaseremaining in the monocyte supematants would be active after being added to WEHIcells in medium containing fetal calf serum, additional enzyme controls were included in these experiments. The effect of residual enzyme activity was controlled for by adding dilutions of either cathepsin-G beginning at 0.02 U/ml or leukocyte elastasestarting at 0.01 U/ml to the target cells. Cathepsin-G and Leukocyte Elastase Reduce the Bioactivity of Recombinant Human TNF and LT The capacity of cathepsin-G and leukocyte elastaseto reduce the cytolytic activity of recombinant TNF and LT was tested by comparing the effect enzyme treated and untreated recombinant cytokines on WEHI- 164 cells.
SCUDERI ET AL.
FIG. 2. Effect of proteaseson monocyte culture supematant cytolytic activity. Plastic adherent monocytes from five healthy adults were obtained by incubating 100 pl of a 1 X IO6mononuclear cell suspensionfor 2 hr at 37°C in each well of 96-well assay plates. The adherent cells were washed and incubated for 18 hr at 37°C in serum-free RPM1 1640 containing 1 &ml LPS 01 I l:B4 (untreated). Some wells contained, in addition to the LPS, either 0.02 U/ml human cathepsin-G (cathepsin-G treated) or 0.16 U/ml human leukocyte elastase(elastasetreated). The 18-hr culture supematants were removed, diluted as indicated with complete RPM1 1640 containing 10% fetal calf serum, and added to 1 X IO4WEHItarget cells. As a control for the effectsof residual protease in the monocyte culture supernatants, cathepsin-G and leukocyte elastasewere added to medium without cells; after a similar incubation these enzyme-treated media were diluted and added to target cells (cathepsin-G and elastasecontrols). After 48 hr at 37°C MTT dye was added and the WEHIcells were subsequently incubated an additional 4 hr. The culture supematants were removed and the insoluble tetrazolium salt wassolubilized in 100pl of DMSO. Absorbancemeasurements were made at 570 nm in an automated ELISA reader.One representativeexperiment, in which the monocytes from a single individual were used,is shown. Each point representsthe mean of four observations f standard deviation.
Figure 3 shows that both cathepsin-G and leukocyte elastasereduced the bioactivity of recombinant TNF and LT. Cathepsin-G and leukocyte elastasecontrols were included to ensure that the loss of WEHIviability was not due to residual enzyme activity. Cathepsin-G and Leukocyte EIastase Degrade Both TNF and LT To determine if the loss in recombinant cytokine cytolytic activity observed by adding either cathepsin-G or leukocyte elastasewas due to a degradation of TNF and LT, enzyme-treated samples were analyzed by SDS-PAGE using 20% gels containing glycerol. Figure 4 shows that adding cathepsin-G to recombinant TNF resulted in the
TNF IS INACTIVATED
BY CATHEPSIN-G AND LEUKOCYTE
FIG. 3. Effect of cathepsin-G and leukocyte elastaseon the cytolytic activity of recombinant TNF and LT. One microliter of either rHTNF or rHLT containing 2.5 X IO4and 1.2 X IO4units of cytolytic activity, respectively, was admixed with 10 ~1of either cathepsin-G or leukocyte elastase,each containing 2.0 units of enzyme activity. As a control, the recombinant cytokines were also admixed with phosphate-buffered saline used to reconstitute the proteases,and diluted proteaseswithout recombinant cytokines were included in theseexperiments. All solutions were incubated at 37°C for 3 hr, diluted in complete RPM1 1640containing 10%fetal calf serum, and added to 1 X IO4WEHItarget cells at the indicated final concentration. After 48 hr MTT dye was added, followed by an additional 4-hr incubation, solubilization of the tetrazolium metabolite, and adsorbancereadings at 570 nm. Each point representsthe mean of four observations + standard deviation.
appearance of a 14.3-kDa breakdown product which increased with time. CathepsinG treated recombinant LT produced a 12.6-kDa breakdown product. Adding leukocyte elastaseto TNF and LT resulted in a reduction of intact cytokine and the emergence of a minor band at 12.2 kDa. The major bands at 13.9 and 14.1 kDa are produced by the autodegradation of elastase,which was determined by comparison to an elastase alone control (data not shown). Western Blot qf Elastase and Cathepsin-G- Treated TNF To further characterize the degradation of TNF by both leukocyte elastase and cathepsin-G we preformed Western blots on enzyme-treated cytokine samples. Figure 5 shows that elastasecleaved recombinant human TNF into several fragments which retained antigenicity. The major degradation product which is identified by the dark band had a molecular weight of 7.6 kDa. Cathepsin-G also fragmented TNF into a series of immunoreactive products with the major band at 11 kDa. Leukocyte Elastase and Cathepsin-G Fail to Degrade Recombinant Human IL-ICY To determine if the degradation of TNF and LT by either leukocyte elastase or cathepsin-G was specific to these related cytokines, we admixed both enzymes with
SCUDERI ET AL.
3y 6fI 120 I
LT14.3 kD -12.6
LEUKOCYTE EIASTASE TIME(min)
TNF 13.9 kD-
3p ‘j,o j20
- 14.1 kD -12.2
FIG. 4. Degradation of recombinant TNF and LT by cathepsin-G and leukocyte elastase.(Top) Two micrograms of recombinant human TNF was combined with 5.0 pg of cathepsin-G containing 0.02 units of activity and incubated for the indicated length of time at 37°C. Recombinant human LT was treated similarly except that 1.Opg of cytokine was used.At the end of the incubation period sample buffer containing SDS and 2-ME was added, and the samples were boiled for 3 mm, loaded on 20% SDS reducing gels containing glycerol, and electrophoresed. (Bottom) Either 1.0 pg TNF or 1.0 pg LT was combined with 0.5 gg of leukocyte elastasecontaining 0.02 units of activity and incubated for the indicated length of time. Molecular weight determinations were made by comparison to a set of low-molecular-weight standards run on each gel.
TNF IS INACTIVATED
BY CATHEPSIN-G AND LEUKOCYTE
FIG. 5. Western blot of leukocyte elastaseand cathepsin-G-treated TNF. Two micrograms of recombinant human TNF was incubated at 37°C for 3 hr alone and with either 17.5 units of human leukocyte elastase or 0.7 units of human cathepsin-G. Controls included an equal volume of both elastaseand cathepsin-G incubated alone under identical conditions. The proteins were separated on 18% SDS-PAGE gels, blotted onto nitrocellulose, and probed first with a specific rabbit anti-human TNF antiserum. A goat anti-rabbit alkaline phosphatase-specificantibody followed by BCIP/NBT substrate was used to visualize the labeled proteins.
recombinant human IL-la. Figure 6 shows that after a 3-hr incubation with either protease, IL- 1(Yshowed no signs of degradation as detected by the presence of breakdown products on reducing gels. a-1-Antitrypsin Prevents the Elastase-Induced Loss of TNF Bioactivity The primary plasma inhibitor of leukocyte elastase is cu-1-antitrypsin (13). The capacity of 1y,-AT and another cy-globulin, (Y-1-antichymotrypsin (al-ACT), to block the leukocyte elastase-induced loss of TNF bioactivity was tested. Figure 7 shows that adding the elastaseinhibitor al-AT to leukocyte elastaseprevented the loss of recombinant TNF cytolytic activity. In contrast cu,-ACT, which is not known to block elastase activity, had no protective effect. a-ldntitrypsin
Prevents the Elastase-Induced Degradation
The capacity of (u~-ATto preserve the cytolytic activity of recombinant TNF in the previous experiment would most easily be explained by an inhibition of enzymeinduced fragmentation. To test this possibility, a-globulin preincubated with enzyme was added to samples of recombinant TNF and compared to cytokine treated with leukocyte elastasealone by SDS-PAGE.
SCUDERI ET AL.
FIG. 6. Effect of leukocyte elastaseand cathepsin-G on recombinant human IL-la. Two micrograms or recombinant human IL-la was incubated at 37°C for 3 hr alone and with either 17.5 units of human leukocyte elastaseor 0.7 units of human cathepsin-G. Controls included an equal volume of both elastase and cathepsin-G incubated alone under identical conditions. The proteins were separated on 18% SDSPAGE gels and silver stained.
Figure 8 shows that TNF alone runs as a single band on reducing gels. Incubating leukocyte elastasewith TNF for 3 hr eliminates the TNF band. The addition of alACT has no effect on TNF and fails to protect the cytokine when it is added after preincubation with elastase.TNF incubated for 3 hr with al-AT remains unfmgmented. When al-AT, which had been preincubated for 1 hr with leukocyte elastase,is added to the recombinant cytokine, the fragmentation of TNF is blocked. Thus the effect of this plasma protease inhibitor is to protect the soluble cytokine from degradation. Activated Neutrophil Culture Supernatants Reduce the Bioactivity of Recombinant Human TNF Neutrophils contain both leukocyte elastaseand cathepsin-G in cytoplasmic granules which can be releasedafter activation with either fMLP or IL-8 Here we have treated cytocholasin B-primed neutrophils with either fMLP or recombinant human IL-8 for 1 hr and then added recombinant TNF to the culture supematants at a concentration of 10 rig/ml. Controls in these experiments included untreated neutrophils and either fMLP or IL-8 added to Hanks’ salt solution, each of which were incubated under identical conditions and subsequently spiked with recombinant TNF. The supematants and controls were then incubated for an additional 1 hr. All fluids were subsequently added to WEHItarget cells and the cytolytic effect of each is shown in Fig. 9. As can be seenthe culture supematant from untreated cells had lost 25.3% of the bioactive
TNF IS INACTIVATED
BY CATHEPSIN-G AND LEUKOCYTE
rHTNF rHTNF rHTNF rHTNF rHTNF rHTNF
50 E s x
.---. O-O -m
+ + + + +
ELASTASE a, -ACT a, -ACT/ELASTASE q-AT al -AT/ElASTASE
FIG. 7. Effect of o-globulin protease inhibitors on the capacity of leukocyte elastaseto reduce the cytolytic activity of recombinant TNF. Leukocyte elastase(0.5 pg) containing 0.02 units of activity was admixed with either 5.0 gg of human al-ACT or 5.0 r.rgof human ol,-AT and incubated at 37°C for I hr. The enzyme and inhibitors (ol,-ACTIELASTASE) and (ol,-ATIELASTASE) were added to 0.5 rg of recombinant human TNF (rHTNF) containing 2.5 X lo4 units of cytolytic activity and incubated for 4 hr at 37°C. Controls included rHTNF diluted in phosphate-buffered saline, 0.5 fig rHTNF + 0.5 rg elastase,and recombinant cytokine admixed with either ol-globulin atone. All solutions were subsequently diluted in complete RPM1 1640 containing 10% fetal calf serum and added to 1 X lo4 WEHItarget cells for a 48-hr incubation at 37°C. At the end of this incubation period the target cells were given MTT dye and incubated for an additional 4 hr at 37°C. The tetrazolium metabolite was solubilized in DMSO and the adsorbance reading was made at 570 nm. Each point representsthe mean of four observations + SD.
TNF and contained 7.47 rig/ml. Examination of the unactivated neutrophil culture supernatant with sensitive and specific assaysfor both human leukocyte elastaseand cathepsin-G revealed that there was no detectable enzyme activity present in this fluid. As an additional control TNF at 10 rig/ml was incubated in Hanks’ balanced salt solution for 1 hr at 37°C and produced a 28% loss of bioactivity, suggesting that the reduction in TNF seenin the unstimulated neutrophil supernatant was not caused by the activity of the cells. The fMLP control reduced the bioactivity of TNF by 84.7%. In contrast, the fMLP-stimulated neutrophil supernatant decreasedthe cytolytic effect by 93.5% and contained 0.097 units of leukocyte elastaseand 0.832 units of cathepsinG activity/lo6 cells. The IL-8 control reduced the cytolytic effect on the WEHI cells by 87.6% and the IL-8-treated neutrophil supematant decreasedTNF bioactivity by 94.3%. IL-8-stimulated cells secreted 0.053 units of leukocyte elastaseand 1.66 units of cathepsin-G/ 1O6neutrophils. DISCUSSION The present study demonstrates that two of the most abundant neutrophil serine proteases,cathepsin-G and leukocyte elastase,can fragment and inactivate both TNF and lymphotoxin. The significance of these observations is suggestedby the close association between polymorphonuclear neutrophils and mononuclear phagocytes at sites of inflammation. Often monocytes are the first hemopoietic cell type to enter tissues after injury. Monocyte activation results in the releaseof TNF (24), IL-l (25), IL-6 (26), and both TNF and IL-1 can activate fibroblast IL-8 production (27). TNF
SCUDERI ET AL.
FIG. 8. n-1-antitrypsin protects TNF from degradation by leukocyte elastase.Leukocyte elastase(0.5 pg) containing 0.02 units of activity was admixed with either 5.0 pg of human (Y,-ACT or 5.0 rg of human LY,AT and incubated at 37°C for 1 hr. The enzyme and inhibitors (a,-ACT/ELASTASE and a,-AT/ELASTASE) were each added to 1.0 rg of rHTNF and incubated for 4 hr at 37°C. Controls included rHTNF diluted in phosphate-buffered saline, 1.Org rHTNF + 0.5 mg elastase,and recombinant cytokine admixed with either ol-globulin alone. At the end of the incubation period sample buffer containing SDS and 2-ME was added, and the sample was boiled for 3 min, loaded onto 20% SDS reducing gels containing glycerol, and electrophoresed.
has been shown to be chemotactic for neutrophils (12) and may, once secreted by monocytes, induce a significant influx of neutrophils into sites of tissue injury. As neutrophils become localized the inflammatory process may become accelerated by the presence of monocyte-derived IL-8 which induces degranulation (28). The neutrophil azurophilic granules are a source of severalenzymes,including leukocyte elastase and cathepsin-G. Both of these serine proteaseshave been associatedwith inflammation (29) and leukocyte elastasein particular has been linked to tissue injury through the breakdown of collagen (30-33). Degradation and inactivation of TNF by proteases may be one of a seriesof mechanisms controlling inflammation in localized microenvironments. The inactivation of TNF by either cathepsin-G or leukocyte elastasemay serve to slow the influx of neutrophils into an area of tissue injury and inflammation by removing at least one of the chemotactic agents present. Additional homeostatic control mechanisms are likely to include the production of PGE2 (9) by fibroblasts in responseto TNF, an increase in the level of circulating a-globulins induced primarily by the monocyte release of IL-6 (36, 37) and an influx of these plasma proteins into the inflammatory site (38). Among the a-globulins which increase in concentration
TNF IS INACTIVATED
BY CATHEPSIN-G AND LEUKOCYTE
FIG. 9. Effect of activated human neutrophil culture supernatants on the bioactivity of recombinant human TNF. Peripheral blood neutrophils were separated on mono-poly resolving medium and the red cells removed by lysis with distilled water. Neutrophils (2 X IO’) were suspended in 4 ml of Hanks’ salt solution and incubated at 37°C for 1 hr (untreated cells). The same number of cells were also treated with 2.5 fig/ml cytochalasin B for 5 min at 37”C, followed by either fMLP at 5.7 X IO-’ M (fh4LP treated) or recombinant human IL-8 at IO nM (IL-8 treated) for an additional 1 hr. Controls included cytochalasin B and either fMLP (tMLP control) or IL-8 (IL-8 control) incubated without cells under identical conditions. Recombinant human TNF was added to the culture supernatants and controls at a final concentration of 10 rig/ml and the fluids were incubated for 1 hr at 37°C. The cytolytic activity of the treated TNF was then determined on WEHItarget cells by comparison to a set of standards made with freshly prepared recombinant human TNF. The values displayed are the mean bioactive TNF -CSD.
during the acute-phase response (Y-1-acid glycoprotein and (Y-1-antitrypsin may play particularly important roles in regulating both TNF secretion and enzyme-induced tissue destruction. a-l-acid glycoprotein has been shown to suppress a number of immune responses,including the secretion of TNF from monocytes (39). ol-l-antitrypsin is a serine protease inhibitor and the principal inhibitor of leukocyte elastase (32). This protease inhibitor in particular may play a dynamic role in inflammation in that it has been shown to block the secretion of TNF from monocytes (39) and as demonstrated here, preserve the secreted cytokine from enzymatic degradation. Although TNF has been reported to have a short circulating half-life, no definitive mechanisms by which TNF is eliminated from the blood have been described (40). The limited half-life of TNF in the blood could be due to degradation of the cytokine and clearance of the breakdown products from the circulation. Although leukocyte elastase and cathepsin-G are effective inactivators of TNF in vitro the presence of inhibitors like (Y-1-antitrypsin in the blood would render blood-borne enzymes inactive. Thus it is unlikely that either leukocyte elastaseor cathepsin-G limits the TNF halflife while the cytokine is in the systemic circulation. However, these proteases,which are present at sites of inflammation (4 l-46), may act at the microenvironmental level to degrade and inactivate cytokines in tissue spaceswhich contain low levels of active a-globulin protease inhibitors. These conditions may be achieved at sites of inflammation where the concentrations of active cY-globulin protease inhibitors can drop transiently due either to inactivation produced by an abundant supply of secreted
SCUDERI ET AL.
proteases(32) or by reactive oxygen liberated form neutrophils (47). Recently published evidence suggeststhat neutrophil-secreted proteases can also function at the cellsubstrate interface to degrade fibronectin and elastin even in the presence of multiple plasma protease inhibitors (48). In conclusion, our findings suggest a role for the neutrophil serine proteases cathepsin-G and leukocyte elastasein the regulation of TNF and LT bioactivity. Further, this study also suggestsan additional role for a-globulin protease inhibitors in TNF homeostasis. REFERENCES 1. Ramani, A., and Aggatwal, B. B., In “Cytolytic Lymphocytes and Complement Effecters ofthe Immune System,” Vol. 11, p. 105. CRC Serial Texts, Boca Raton, FL, 1988. 2. Chong, A. S.-F., Aleksijevic, A., Scuderi, P., Hersh, E. M., and Grimes, W. J., Cancer Immunol. Immunother. 29, 210, 1989.
3. 4. 5. 6.
Ruddle, N. H., and Waksman, B. H., Science 157, 1060, 1967. Granger, G. A., and Williams, T. W., Nature 218, 1253, 1968. Aggarwal, B. B., Eessalu,T. E., and Hass, P. E., Nature 318, 665, 1985. Tracey, K. J., Lowry, S. F., Fahey, T. J., Albert, J. D., Fong, Y., Hesse,D., Beutler, B., Manogue, K. R., Calvano, S., Wei, H., Cerami, A., and Shires, G. T., Surg. Gynecol. Obstet. 164, 4 15, 1987. 7. Dinareho, C. A., Cannon, J. G., Wolff, S. M., Bemheim, H. A., Beutler, B., Cerami, A., Figari, I. S., Palladino, M. A., and O’Connor, J. V., .I. Exp. Med. 163, 1433, 1986. 8. Sugarman, B. J., Aggarwal, B. B., Hass, P. E., Figari, I. S.;Palladino, M. A., and Shepard, H. M., Science 230,943, 1985. 9. Dayer, J. M., Beutler, B., and Cerami, A., J. Exp. Med. 162, 2163, 1985. 10. Berkow, R. L., and Dodson, M. R., J. Leukocyte Biol. 44, 345, 1988. 11. Reed, D., and Moore, F. D., Arch. Surg. 123, 1333, 1988. 12. Meyer, J. D., Yurt, R. W., Duhaney, R., Hesse,D. G., Tracey, K. J., Fong, Y. M., Verma, M., Shires, G. T., Dinean, P., and Lowry, S. F., Arch. Surg. 123, 1454, 1988. 13. Wieslaw, W., Farley, D., Salversen, G., and Travis, J., In “Advances in Experimental Medicine and Biology” (W. H. Horl and A. Heidland, Eds.), Vol. 240, p. 23. Plenum, New York, 1988. 14. Reilly, C. F., Tewksbury, D. A., Schechter, N. M., and Travis, J., .I Biol. Chem. 257, 8619, 1982. 15. Tonnesen, M. G., Klempner, M. S., Austin, K. F., and Wintroub, B. U., .I Clin. Invest. 69,25, 1982. 16. Buist, S. A., Hosp. Pratt. 24, 51, 1989. 17. Scuderi, P., Sterling, K. E., Raitano, A. B., Grogan, T. M., and Rippe, R. A., J. Interferon Res. 7, 155, 1987. 18. Scuderi, P., Sterling, K. E., Lam, K. S., Finley, P. R., Ryan, K. J., Ray, C. G., Peterson, E., Slymen, D. J., and Salmon, S. E., Lancet 2, 1364, 1986. 19. Mosmann, T., J. Immunol. Methods 65, 53, 1983. 20. Monissey, J. H., Anal. Biochem. 117, 307, 1981. 21. Towbin, H., Staehlin, J., and Gordon, J., Proc. Natl. Acad. Sci. USA 76, 4350, 1979. 22. Castillo, M., Najakima, K., and Zimmerman, M., Anal. Biochem. 99, 53, 1979. 23. Barrett, A., “Methods in Enzymology” (L. Lorand, Ed.), Vol. 80, p. 561. Academic Press,San Diego, 1981. 24. Carswell, E., Old, L. J., Kassel, R. L., Green, S., Fiore, N., and Williamson, B., Proc. Natl. Acad. Sci. USA 72,3666,
25. Sisson, S., and Dinarello, C. A., In “Human Monocytes” (M. Zembala and G. L. Asherson, Eds.), p. 183. Harcourt Brace Jovanovich, New York, 1989. 26. Andersson, V., and Mats&a, T., Eur. J. Immunol. 19, 1157, 1989. 27. Larsen, C. G., Anderson, A. O., Oppenheim, J. J., and Matsushima, K., Immunology 68,31, 1989. 28. Willems, J., Jonian, M., Cingve, S., and Damme, J., Immunology 67, 540, 1989. 29. Travis, J., Giles, P. J., Porcelli, L., Reilly, C. F., Baugh, R., and Powers, J., In “Protein Degradation in Health and Disease,” p. 5 1. Exerpta Medica, Amsterdam, 1980. 30. Mainardi, C. L., Hasty, D. L., Seyer, J. M., and Kang, A. H., J. Biol. Chem. 255, 12,006, 1980.
TNF IS INACTIVATED
BY CATHEPSIN-G AND LEUKOCYTE
31. Mainardi, C. L., Dixit, S. N., and Kang, A. H., J. Biol. Chem. 255, 5435, 1980. 32. Gadek, J. E., Fells. G. A., Wright, D. G., and Crystal, R. G., Biochem. Biophys. Res. Commun. 95, 1815, 1980. 33. Pipoly, D. J., and Crouch, E. C., Biochemistry 26, 5748, 1987. 34. Kohase, M., Henriksen-De Stefano, D., May, L. T., Vilcek, J., and SegaI,P., Ceil 45, 659, 1986. 35. Koj, A., Gordon, A. H., and Gauldie, J., Experientia 44, 9, 1988. 36. Geiger, T., Andus, T., Klapproth, J., Hirano, T., Kishimoto, T., and Heimich, P. C., Eur. J. Immunol. 18, 717, 1988. 37. Andus, T., Geiger, T., Hirano, T., Kishimoto, T., and Heimich, P. C., Eur. J. Immunol. 18,739, 1988. 38. Jayle, M. F., Janiaud, P., Engler, R., Degrelle, H., and Marcais, J., In “Protides ofthe Biological Fluids,” Vol. 18, p. 159. Pergamon, Oxford, 1971. 39. Scuderi, P., Dorr, R. T., Liddil, J. D., Finley, P. R., Meltzer, P., Raitano, A. B., and Rybski, J., Eur. J. Immunol. 19, 939, 1989. 40. Buetler, B. A., Milsark, I. W., and Cerami, A., J. Immunol. 135, 3972, 1985. 41. Daminaon, V. V., Tsant, A., Kucich, U., Abrams, W. R., Rosenbloom, J., Kimbel, P., Fallahnejad. M.. and Weinbaum, G., J. Clin. Invest. 78, 482, 1986. 42. Merritt, T. A., Cochrane, C. G., Holcomb, K., Bohl, B., Hallman, M., Strayer, D., and Edwards, D. K.. J. Clin. Invest. 72, 656, 1983. 43. Abrams, W. R., Fein, A. M., Kucich, U., Kueppers, F., Yamada, H., Kuzmowycz, T., Morgan. L., Lippman, M., Goldberg, S. K., and Weinbaum, G., Amer. Rev. Respir. Dis. 129, 735, 1984. 44. McGuire, W. W., Sprang, R. G., Cohen, A. B., and Cochrane, C. G., J. Cl&. Invest. 69, 543, 1982. 45. Gysen, P., Malaise, M., Gaspar, S., and Franchimont, P., Clin. Rheum. 4, 39, 1985. 46. Velvert, M., and Fehr, K., Rheumatol. Int. 7, 195, 1987. 47. Neumann, M., and Kownatzki, E., Agents Actions 26, 183, 1989. 48. Powers, J., Anal. Biochem. 99, 53, 1979.