Chicken cystatin stimulates nitric oxide release from

Eur. J. Biochem. 266, 1111±1117 (1999) q FEBS 1999

Chicken cystatin stimulates nitric oxide release from interferon-g-activated mouse peritoneal macrophages via cytokine synthesis Ludovic Verdot1,2, Gilles Lalmanach2, Vincent Vercruysse1, Johan Hoebeke3, Francis Gauthier2 and Bernard Vray1 1

Laboratoire d'Immunologie ExpeÂrimentale, FaculteÂ de MeÂdecine, UniversiteÂ Libre de Bruxelles, Brussels, Belgium; Laboratoire d'Enzymologie et de Chimie des ProteÂines, UniversiteÂ FrancËois Rabelais, Tours, France; 3Immunochimie des Peptides et des Virus, UPR9021 du CNRS, Institut de Biologie MoleÂculaire et Cellulaire, Strasbourg, France

2

Cystatins are natural tight-binding, reversible inhibitors of cysteine proteases. We have shown that cystatins also stimulate nitric oxide (NO) production by interferon-g-activated mouse peritoneal macrophages [Verdot, L., Lalmanach, G., Vercruysse, V., Hartman, S., Lucius, R., Hoebeke, J., Gauthier F. & Vray, B. (1996) J. Biol. Chem. 271, 28077±28081]. The present study was undertaken to further document this new function. Macrophages activated with interferon-g and then stimulated with interferon-g plus chicken cystatin generated increased amounts of NO in comparison with macrophages only activated with interferon-g. Interferon-gactivated macrophages must be incubated with chicken cystatin for at least 8 h to upregulate NO production. NO induction was due to increased inducible nitric oxide synthase protein synthesis. Macrophages incubated with chicken cystatin alone or with interferon-g plus chicken cystatin produced increased amounts of both tumor necrosis factor a and interleukin 10. The addition of recombinant murine tumor necrosis factor a alone or in combination with recombinant murine interleukin-10 mimicked the effect of chicken cystatin. The addition of neutralizing anti-(tumor necrosis factor a) antibodies reduced sharply NO production by chicken cystatin/ interferon-g-activated mouse peritoneal macrophages. Taken together, these data suggest that chicken cystatin induces the synthesis of tumor necrosis factor a and interleukin 10. In turn, these two cytokines stimulate the production of NO by interferon-g-activated macrophages. The findings point to a new relationship between cystatins, cytokines, inflammation and the immune response. Keywords: nitric oxide synthase; tumor necrosis factor a; interleukin 10. Nitric oxide (NO) is important in various physiological processes (vasodilatation, smooth muscle regulation, neurotransmission) and particularly in the immune response. It is involved in several inflammatory diseases and is cytotoxic or cytostatic in a wide range of infections [1,2]. NO is synthesized from l-arginine by NO synthase (NOS). Three isoforms of NOS have been identified; they differ in their tissue distributions and regulation. Neuronal NOS (nNOS or NOS1) and endothelial NOS (eNOS or NOS3) types are constitutively present in neurons and endothelial cells, respectively, and are calcium-dependent [2]. The third NOS isoform is an inducible enzyme (iNOS or NOS2) synthesized in response to specific signals by cells such as hepatocytes, fibroblasts and macrophages and is calcium independent [3]. Its synthesis is induced mainly by interferon gamma (IFN-g) and leads to the rapid production of large amounts of NO that protect the organism in many infections [2]. In addition, other cytokines such as tumor Correspondence to B. Vray, Laboratoire d'Immunologie ExpeÂrimentale (CP 615), FaculteÂ de MeÂdecine, UniversiteÂ Libre de Bruxelles, 808 route de Lennik, B-1070 Brussels, Belgium. Fax: + 32 2 555 63 60, Tel.: + 32 2 555 62 60, E-mail: [email protected] Abbreviations: CC, chicken cystatin; GM-CSF:, granulocyte macrophage-colony stimulating factor; HBSS, Hanks' balanced salt solution; IFN-g, interferon gamma; IL-10, interleukin 10; iNOS, inducible nitric oxide synthase; LPS, lipopolysaccharide; MPM, mouse peritoneal macrophages; NOS, nitric oxide synthase; rIL-10, recombinant murine IL-10; rTNF-a, recombinant murine TNF-a; TNF-a, tumor necrosis factor a. mAb, monoclonal antibody; RAb, rabbit polyclonal antibody. (Received 14 June 1999, revised 1 October 1999, accepted 26 October 1999)

necrosis factor-a (TNF-a) cooperate synergistically with IFN-g to increase significantly the production of NO [4] whereas interleukin 10 (IL-10) modulates the IFN-g-induced production of NO [5±9]. Cystatins are natural tight-binding, reversible inhibitors of cysteine proteinases. They are involved in various biological and pathological processes, including protein catabolism, antigen processing, inflammation, dystrophic disorders and metastasis [10]. The cystatin superfamily contains three main families (stefins, cystatins and kininogens) whose members are present in various body fluids [11]. All the cystatins that we have tested to date [human stefin B, chicken cystatin (CC), and T-kininogen] potentiate the NO release from IFN-g-activated mouse peritoneal macrophages (MPMs) in a concentrationdependent manner. By complexing the enzyme-inhibitory site of CC with carboxymethylated papain, the archetypal cysteine proteinase, we have demonstrated that this site is not involved in the augmentation of NO synthesis. This indicates that another site is involved in the stimulation of NO synthesis [12]. But, cystatins do not stimulate NO production by unactivated MPM, indicating that they do not induce the NO synthesis pathway, as does IFN-g. They seem to modulate NO production, as do many cytokines produced during infection or inflammation, such as TNF-a [13], IL-10 [5±8], granulocyte macrophage-colony stimulating factor (GM-CSF) [14], IFN-a/b [15], or bacterial products (lipopolysaccharide, LPS) [16]. In the present work, the kinetics and the synergistic mechanism of CC and IFN-g upon MPM have been further investigated emphasizing the possible role of other cytokines.

1112 L. Verdot et al. (Eur. J. Biochem. 266)

M AT E R I A L S A N D M E T H O D S Reagents RPMI 1640 medium supplemented with 25 mm Hepes and 2 mm glutamine, fetal bovine serum, penicillin, streptomycin, and HBSS were from Gibco (Grand Island, NY, USA). Chicken cystatin [17] and polymyxin B sulfate were from Calbiochem Co. (La Jolla, CA, USA). LPS-free recombinant mouse IFN-g was kindly provided by A. Billiau and H. Heremans (Katholieke Universiteit Leuven, Leuven, Belgium). The rabbit anti-iNOS polyclonal Ig was from Sanvertech (Boechout, Belgium). Goat anti-rabbit peroxidase conjugated antibodies, 4-chloro-1-naphthol, LPS (Escherichia coli, serotype 0127:B8), actinomycin D and cycloheximide were from Sigma Chemical Co. (St Louis, MO, USA). ELISAs for TNF-a and IL-10 were from Genzyme (Cambridge, MA). A monoclonal rat anti(mouse TNF-a) IgG1, neutralizing mouse TNF-a bioactivity, MP6-XT3, and a rat IgG1 isotype control (R3-34) were both from Pharmingen. A neutralizing anti-IL-10 monoclonal IgG1 (mAb; JES5±2A5) was obtained in ascite form [18]. As isotype-matched control, we used ascites of LO-DNP-2 hybridoma cells (kindly provided by H. Bazin, Experimental Immunology Unit, UniversiteÂ Catholique de Louvain, Belgium), secreting a rat IgG1 mAb with anti-dinitrophenyl specificity. A rabbit anti-(mouse TNF-a) polyclonal Ig was kindly provided by W. Buurman (Department of Surgery, University of Limburg, Maastricht, the Netherlands). Rabbit Ig were isolated from the serum of a control rabbit, precipitated with ammonium sulfate and extensively dialysed against NaCl/Pi. rTNF-a was purchased from Genzyme. rIL-10 was obtained as culture supernatants from Sf 9 insect cells stably transfected with the corresponding complementary DNA, as previously described [9,19] using the baculovirus expression vector pBlue Bac2 (Invitrogen, San Diego, CA, USA). Murine peritoneal macrophage culture Male BALB/c mice, 6- to 8-week-old (Bantin and Kingman Universal Ltd, UK) were killed by ether inhalation. MPM were harvested by washing the peritoneal cavities twice with ice-cold HBSS without Ca2 and Mg2 . MPM were collected by centrifugation at 400 g for 10 min at 4 8C. Distilled sterile water (1 mL) was added to the pellet for 30 s to lyse red cells. MPM were immediately suspended in HBSS without Ca2+ and Mg2+ and centrifuged as above. The resulting pellet was suspended in culture medium (RPMI 1640 medium) supplemented with 25 mm Hepes, 2 mm glutamine, 10% fetal bovine serum (mycoplasma free and endotoxin concentration less than 27.5 pg´mL21 ), penicillin (100 IU´mL21 ) and streptomycin (100 mg´mL21 ). MPM were then allowed to adhere (1.5 105 MPM per well) in 96-well microplates (Nunc, Roskilde, Denmark) for 2 h at 37 8C in a 5% CO2 water-saturated atmosphere in culture medium. Non-adherent cells were removed by washing with culture medium at 37 8C before adding appropriate solutions diluted in culture medium. Polymyxin B (10 IU final concentration) was added to all solutions, except those containing LPS, to neutralize traces of contaminating endotoxin. Determination of NO NO production was assayed by measuring nitrite, its stable degradation product, by the Griess reaction [20]. MPM were

q FEBS 1999

incubated with CC (1027 m) or with IFN-g (100 IU´mL21) or a combination of both for 48 h. Culture medium served as control. Aliquots of culture supernatants (100 mL) were removed and mixed with 100 mL of Griess solution (1% sulfanilamide, 0.1% naphthylethylene diamine dihydrochloride, 2% H3PO4) and the absorbance was measured at 540 nm in a microplate ELISA reader (Titertek Multiscan MCC/340, MKII EFLAB, Helsinki, Finland). Sodium nitrite (NaNO2) diluted in culture medium was used as a standard. The detection limit of the assay was 1.25 mm. The production of NO by unstimulated MPM was typically around the detection limit. The concentration of LPS in all the reagents and media was below 80 pg´mL21 according to the colorimetric Limulus Amoebocyte Lysate assay (detection limit: 1 pg´mL21) (Coatest Endotoxin Chromogenix, MoÈlndal, Sweden). Cytokines MPM were incubated for 48 h with CC (1027 m), or with IFN-g (100 IU´mL21), or a combination of both. Culture medium served as control. Aliquots of culture supernatants (50 mL) were removed for TNF-a and IL-10 determination by ELISA, performed according to the manufacturers instructions. Experiments were repeated at least three times and performed in triplicate. rTNF-a (100 or 1000 pg´mL21) and/or rIL-10 (100 or 500 pg´mL21) were added to MPM in the medium containing IFN-g but without CC. MPM were incubated for 48 h and nitrite levels were measured in the culture supernatants. Biological activity of TNF-a was neutralized with 10 mg´mL21 of rat anti-(TNF-a) IgG1 mAb MP6-XT3, or a polyclonal rabbit anti-(TNF-a) (10 mg´mL21) and, as control, an unrelated rat IgG1 isotype-matched control (R3-34) or a rabbit immunoglobulin fraction from the serum of a control rabbit, respectively. Biological activity of IL-10 was neutralized by using 10 mg´mL21 of anti-IL-10 mAb (JES5-2A5) and an unrelated isotype-matched control (LO-DNP). Immunoblotting MPM were cultured in 24-well microplates (2 106 cells/well, Nunc) as described above and incubated with the test solutions in the culture medium for 48 h at 37 8C in a 5% CO2 watersaturated atmosphere. The MPM were then quickly washed twice with 1 mL ice-cold HBSS and lysed with 100 mL icecold lysis buffer (Hepes 5 mm, pH 7.4). Unlysed cells were collected by centrifugation and the lysis step repeated twice. The resulting supernatants were pooled. The complete lysis of the cells was checked under the microscope. An equal volume of reducing electrophoresis buffer (4% sodium dodecylsulfate; 10% glycerol; 10% 2-mercaptoethanol; 0.2% bromphenol blue) was added to the supernatant, thoroughly mixed and stored frozen at 220 8C. Samples for electrophoresis were boiled for 3 min and then electrophoresed in 10% polyacrylamide/SDS gels. They were then actively transferred to nitrocellulose. The nitrocellulose strips were saturated in Tris 50 mm, NaCl 150 mm, 2% Tween 20 buffer pH 7.4 for 1 h at 37 8C, and then incubated overnight at 4 8C, with agitation with anti-iNOS antibodies diluted 1/1000 in the incubation buffer (NaCl/Pi; 5% skimmed milk powder; 0.1% Tween 20). The strips were thoroughly washed (three times for 15 min) with NaCl/Pi containing Tween 20 (0.1%) and then incubated with goat anti-rabbit peroxidase-conjugate antibodies in incubation

q FEBS 1999

Cystatin induces NO via TNF-a and IL-10 synthesis (Eur. J. Biochem. 266) 1113

buffer (1/1000 final dilution). The strips were again washed (3 15 min) and the complexes were revealed with 4-chloro1-naphthol and hydrogen peroxide in 50 mm Tris, 0.5 m NaCl, pH 7.6. Immunofluorescence MPM were cultured on 16-well Lab Tek chamber slide (5 104 cells per well; Nunc) and incubated for 24 h with culture medium, IFN-g (100 IU´mL21), CC (1027 M), IFN-g (100 IU´mL21) plus CC (1027 m), IFN-g (100 IU´mL21) plus LPS (10 ng´mL21) or LPS (10 ng´mL21). The supernatants were removed and the wells were washed twice with 200 mL HBSS. The MPM were then incubated for 5 min in the same buffer and fixed for 10 min at room temperature with freshly prepared HBSS containing 3% paraformaldehyde. The membranes were permeabilized with HBSS containing Triton X 100 (1%). The fixed MPM were washed with HBSS, saturated with HBSS containing 1.5% bovine serum albumin for 30 min and incubated for 2 h at room temperature with anti-iNOS antibodies (1/1000 final dilution) in the saturation buffer. They were washed three times with HBSS and mixed with fluoresceinated goat antirabbit antibodies (diluted 1/100 in saturation buffer) for 45 min. The wells were washed 3 times for 10 min. Lastly, the slides were mounted in NaCl/Pi /glycerol (v/v) and examined under a microscope (Zeiss Universal) equipped for fluorescence.

R E S U LT S Sequence of incubation with IFN-g and CC We previously demonstrated that CC stimulates NO production by IFN-g-activated MPM [12]. We further documented this observation by incubating MPM first with CC or IFN-g (step 1) and then with medium, IFN-g, or CC or a combination of both (step 2). As shown in Fig. 1, there was no significant stimulation of nitrite production when MPM were incubated with culture medium (A) or with CC (B) before incubating with IFN-g alone. This could be because the cells lost the capacity to produce large amounts of NO when the MPM were maintained in the culture medium without IFN-g for 24 h. To exclude this possibility, MPM were first incubated for 24 h in culture medium (C) or medium containing CC (D) and then for 24 h in medium containing IFN-g/CC. The nitrite concentration was increased sixfold when MPM were incubated (C) with both IFN-g and CC, instead of with IFN-g alone (A). Surprisingly, the nitrite concentration was reduced by 50% when MPM were first incubated with CC (D) instead of medium (C) (C vs. D: P , 0.05, Student's t-test). As a control, when MPM were incubated in medium alone in steps 1 and 2, NO was undetectable (data not shown). As CC did not stimulate MPM to produce NO, the reverse sequence was tested. MPM were first activated with IFN-g and then incubated with CC (F). This resulted in a threefold increase in NO production over the control: IFNg-activated MPM incubated with culture medium without CC (E) (E vs. F, P , 0.01, Student's t-test). This enhanced nitrite production was comparable to that obtained by incubating MPM with IFN-g for 48 h (G). The greatest increase in nitrite production was obtained when MPM were activated with IFN-g and then incubated with IFN-g/CC (H). The results confirmed that CC stimulates IFN-g-activated MPM to produce NO and indicates that MPM activated with IFN-g

and then incubated with IFN-g/CC generated the greatest production of NO. Effect of CC on NO release from LPS-activated MPM LPS is a powerful activator of macrophages [16,21,22]. We raised the question of whether CC could stimulate the production of NO by LPS-activated MPM, as it does with IFN-g-activated MPM. Nitrite was produced by LPS-activated MPM in a concentration-dependent manner (from 1029 to 1025 m), but the addition of CC did not significantly increase nitrite production, indicating that CC potentiates the effect of IFN-g but not that of LPS (data not shown). Kinetics of NO upregulation by IFN-g-activated MPM incubated with CC We investigated the rate at which CC upregulated NO production. MPM were incubated with IFN-g for 24 h and then with CC or control medium, for another 24 h. The NO production started 8 h after the addition of CC and increased in a significant way at a later time point (18 h; Fig. 2A). These data show that the induction of the CC-mediated increased production of NO lasts at least 8 h. Incubation of IFN-g-activated MPM with CC plus IFN-g resulted in maximal NO production (Fig. 1). To assess whether this synergistic effect was related to the contact time between CC and IFN-g-activated MPM, MPM were activated with IFN-g for 24 h. Then, CC (1027 m) or control medium were added on top (time 0). At various intervals, over another 24 h incubation period, CC-containing medium was harvested (harvest 1), frozen and replaced by IFN-g-containing medium for the remainder of the 24 h incubation period. The experiment

Fig. 1. Time-sequence of CC-stimulated NO-production by MPM in combination with IFN-g. MPM (1.5 105 per well) were allowed to adhere for 2 h in 96-well plates and then washed twice with warmed culture medium. MPM were first incubated for 24 h in medium containing the appropriate solutions (step 1): medium alone (A and C); CC (1027 m; B and D), or IFN-g (100 IU´mL21; E, F, G, and H). Wells were then washed twice with warmed culture medium and the medium replaced (step 2) with fresh medium (E), CC (1027 m; F), IFN-g (100 IU´mL21; A, B and G), CC (1027 m) plus IFN-g (100 UI´mL21) (C, D and H) for 24 h. Nitrite was then measured. Means ^ SD of one experiment performed in triplicate and representative of five independent experiments. *, P , 0.05 when comparing C to D, **, P , 0.01 when comparing E to F and G; ***, P , 0.01 when comparing H to F and G; Student's t-test.


q FEBS 1999

CC-induced increase of NO production from IFN-g-activated MPM requires mRNA and protein synthesis The long delay (more than 8 h) necessary for CC to induce increased NO production (Fig. 2A) suggests that CC could act by stimulating protein synthesis. This hypothesis was checked using two inhibitors of gene transcription and translation, actinomycin D and cycloheximide, respectively. MPM were first incubated with IFN-g (100 IU´mL21) for 24 h to induce the synthesis of iNOS mRNA and to constitute a pool of iNOS protein. They were then washed and incubated for 30 min with medium, cycloheximide (10 mm) or actinomycin D (1 mm). CC without IFN-g was then added on top for 48 h. CC increased nitrite production (up to 25 mm), but treatment with cycloheximide or actinomycin D prevented this increase. Cycloheximide and actinomycin D alone had no effect. Effect of CC on iNOS synthesis by MPM: immunoblotting and immunofluorescence studies Murine macrophages synthesize iNOS protein following appropriate activation [21,23,24] and IFN-g-activated MPM must be incubated with CC for a long time for NO synthesis to increase. This could mean that CC induces an increase of the iNOS protein concentration. We verified this by preparing immunoblots using anti-iNOS antibodies (Fig. 3). MPM were incubated with culture medium (A), IFN-g (B), IFN-g plus CC (C), or LPS (D) for 48 h and then treated as described in Materials and methods. Untreated MPM incubated with culture medium gave no immunoreactive band (Lane A). MPM Fig. 2. Kinetics of NO upregulation by IFN-g-activated MPM incubated with CC. (A) MPM (1.5 105 per well) were allowed to adhere for 2 h in 96-well plates (see legend of Fig. 1) and incubated with IFN-g (100 IU´mL21) for 24 h. Then, culture medium (open square) or CC (1027 m, black circle) diluted in culture medium were added on top (time 0). The supernatants were harvested at various times and immediatly frozen at 220 8C until use. Data are means ^ SD of one experiment performed in triplicate and representative of four independent experiments. (B) MPM (1.5 105 per well) were allowed to adhere for 2 h in 96-well plates (see the legend of Fig. 1) and incubated with IFN-g (100 IU´mL21) for 24 (h). Then, culture medium (open square), or CC (1027 m, black circle) were added on top to all the wells (time 0). At various time points over a second 24 h incubation period, the culture supernatants were harvested and immediatly frozen at 220 8C until use (harvest 1). The cultures were washed and IFN-g-containing medium was added for the remainder of the 24 h incubation period. The experiment ended by harvesting the supernatants, and freezing them at 220 8C untill use (harvest 2). The nitrite concentrations in the supernatants from harvests 1 and 2 were added and plotted against time of incubation after addition of CC. Data are means ^ SD of one experiment with triplicate cultures. Similar data were generated in 3 additional confirmatory experiments.

ended by harvesting the supernatants and freezing them (harvest 2). The values for nitrite concentrations, measured separately in both harvests, were added and plotted against time of incubation after addition of CC (Fig. 2B). The upregulation of NO production depended on the contact time between CC and IFN-g-activated MPM. However, the delay in NO upregulation was not dependent on duration of interaction between CC and IFN-g-activated MPM. Indeed, MPM incubated for only 30 min with CC and washed thereafter showed a significant upregulation of the total 24 h NO production (Fig. 2B).

Fig. 3. Effect of CC on iNOS synthesis by MPM. MPM (2 106 per well) were allowed to adhere for 2 h in 24 well microplates (see legend to Fig. 1) and incubated for 48 h at 37 8C with culture medium, IFN-g (100 IU´mL21), IFN-g (100 IU´mL21) plus CC (1027 m), or LPS (1 mg´mL21). The MPM were lysed and the supernatants were electrophoresed under reducing conditions in a 10% polyacrylamide sodium dodecylsulfate gel and transferred to a nitrocellulose sheet. The sheet was saturated with skimmed milk powder and incubated overnight at 4 8C under agitation with an anti-iNOS antibody diluted 1 : 1000 in incubation buffer. The nitrocellulose strips were then incubated with a peroxidase-conjugated anti-rabbit antibody diluted 1 : 1000 in incubation buffer and the complexes were detected using 4-chloro-1-naphthol and hydrogen peroxide in 50 mm Tris, 0.5 m NaCl, pH 7.6. Culture medium: lane A; IFN-g: lane B; IFN-g and CC: lane C; LPS: lane D.

q FEBS 1999


Table 1. Nitrite, TNF-a and IL-10 measurements. MPM (1.5 105 per well) were allowed to adhere for 2 h in 96-well microplates (see legend of Fig. 1) and were then incubated for 48 h with medium alone, CC (1027 m), IFN-g (100 IU´mL21) or IFN-g (100 IU´mL21) and CC (1027 m). Data are means ^ SD of three independent experiments performed in duplicate. (The nonparametric Mann±Whitney U-test was used because of the individual variations.) MPM treatment

Nitrite (mm)

TNF-a (ng´mL21)

IL-10 (ng´mL21)

Medium IFN-g CC IFN-g + CC

1.8 12.1 2.8 50.0

0.11 0.14 2.35 3.43

0.12 0.99 0.52 0.18

^ ^ ^ ^

1.3 9.4 2.3 2.0

^ ^ ^ ^

0.07 0.08 1.20a 2.00b

^ ^ ^ ^

0.04 0.04 0.39c 0.11d,e

a CC vs. IFN-g, P , 0.01; b IFN-g plus CC vs. IFN-g, P , 0.01; c CC vs. medium, P , 0.01; d IFN-g plus CC vs. IFN-g, P , 0.01; e IFN-g plus CC vs. CC: P , 0.01.

incubated with IFN-g alone (B), IFN-g (100 IU´mL21) and CC (1027 m) (C), or with LPS (1 mg´mL21) (D), all produced a protein with a relative molecular mass of about 130 kDa, the size of the iNOS protein (lanes B, C and D) [23,24]. The relative intensities of staining were in good agreement with the nitrite concentrations measured (Fig. 1). MPM incubated Table 2. Effects of the addition of cytokines or neutralizing antibodies on NO production by MPM. For cytokines, MPM (1.5 105 per well) were allowed to adhere for 2 h in 96-well microplates (see legend of Fig. 1) and were then incubated for 48 h with IFN-g (100 IU´mL21) or IFN-g (100 IU´mL21) and TNF-a and/or IL-10. Data are means ^ SD of two independent experiments performed in triplicate. For antibodies, MPM (see above) were incubated for 48 h with CC (1027 m) and IFN-g (100 IU´mL21) in the presence of anti-(TNF-a) mAb, or rabbit anti(TNF-a) antibodies, or anti-(IL-10) mAb or a combination of anti-(TNF-a)/ anti-(IL-10) mAb. All these antibodies were used at 10 mg´mL21. Isotypematched control antibodies were also used but were without effect (data not shown). Data are means ^ SD of three independent experiments performed in triplicate. The nonparametric Mann±Whitney U-test was used for both cytokines and antibodies because of the individual variations. Addition

Nitrite (mm)

Cytokines added to IFN-g-activated MPM None + TNF-a (100 pg´mL21) + TNF-a (1000 pg´mL21) + IL-10 (100 pg´mL21) + IL-10 (500 pg´mL21) + TNF-a (100 pg´mL21) + IL-10 (500 pg´mL21) + TNF-a (1000 pg´mL21) + IL-10 (500 pg´mL21)

5.4 ^ 13.7 ^ 47.8 ^ , 1.25 , 1.25 33.1 ^ 73.9 ^

1.4 5.5a 12.9a,b

Antibodies added to CC/IFN-g-activated MPM None Anti-(TNF-a) (mAb) Anti-(TNF-a) (polyclonal) Anti-(IL-10) mAb Anti-(TNF-a)-IL-10 mAb

51.4 21.5 20.8 44.2 24.1

10.2 15.6c 16.0c 16.7 15.1c

^ ^ ^ ^ ^

19.7a,b 7.6

a TNF-a (100 pg´mL 21 ), TNF-a (1000 pg´mL 21 ) and TNF-a (100 pg´mL21) + IL-10 (500 pg´mL21) vs. none: P , 0.05; b TNF-a (1000 pg´mL21) and TNF-a (100 pg´mL21) + IL-10 (500 pg´mL21) vs. TNF-a (1000 pg´mL21) + IL-10 (500 pg´mL21): P , 0.05; c anti-TNF-a mAb, or rabbit anti-TNF-a antibodies, or a combination of anti-TNF-a/ anti-IL-10 mAb vs. none: P , 0.01.

with IFN-g and CC seemed to produce more iNOS protein (lane C) than MPM incubated with IFN-g alone (lane B). The immunofluorescence studies using anti-iNOS antibodies and paraformaldehyde-fixed, Triton X 100-permeabilized MPM confirmed the immunoblot data. No immunofluorescence was detected in unactivated MPM with or without CC. Immunofluorescence was seen only on MPM activated with IFN-g and its intensity was sharply increased when CC was added together with IFN-g (data not shown). Role of cytokines in the CC-stimulated increase of NO release by IFN-g-activated MPM Several cytokines (TNF-a, IL-10, GM-CSF) act on IFN-gactivated MPM to enhance NO production [5±8,13,14]. We examined the capacity of CC to increase cytokine production and hence to enhance the NO released from IFN-g-activated MPM. The TNF-a and IL-10 concentrations in the supernatants of MPM cultured for 48 h with CC, with IFN-g, or with CC and IFN-g were measured. MPM incubated with CC alone produced increased levels of TNF-a and IL-10 (Table 1). The TNF-a concentration was increased 17-fold and that of IL-10 4-fold over the values obtained with MPM incubated with medium or with IFN-g alone. Incubation of MPM with IFN-g plus CC slightly decreased IL-10 production, while no statistically significant difference was seen for TNF-a. NO production remained at the control level (culture medium alone or CC alone) and was slightly enhanced by IFN-g alone, and sharply enhanced by IFN-g followed by CC [12]. The NO concentrations obtained with IFN-g plus CC were similar to those shown in the Fig. 1H. The effects of exogenous rTNF-a and/or rIL-10 on NO production by IFN-g-activated MPM were also investigated (Table 2). NO production by IFN-g-activated MPM was enhanced by adding 100 pg´mL21 or 1000 pg´mL21 rTNF-a. The addition of rIL-10 alone (100 pg´mL21 or 500 pg´mL21) was without effect, except when this cytokine (500 pg´mL21) was added together with rTNF-a Neutralizing anti-TNF-a mAb were added to CC/IFN-gactivated MPM and NO production was measured 48 h later (Table 2). By comparison with CC/IFN-g-activated MPM, NO production was sharply reduced. A rabbit anti-(TNF-a) polyclonal antibody had a similar inhibitory effect. The same inhibitory effect was obtained when adding anti-TNF-a and anti-IL-10 mAb together. No inhibition of NO production was observed in the presence of anti-IL-10 mAb. Isotype-matched control antibodies had no effect (data not shown). These data indicate that TNF-a plays a central role in the NO production by CC/IFN-g-activated MPM.

DISCUSSION IFN-g is so far the only cytokine able to induce by its own the synthesis of iNOS and the release of NO from MPM. However, it has been shown recently that IFN-g-induced TNF-a is a prerequisite for in vitro production of NO released by MPM [2,4]. In addition to TNF-a, many other cytokines and bacterial products including transforming growth factor b, IL-10, TNF-a/b, GM-CSF and LPS [5±8,13,16,22,25] can also trigger NO production by acting in synergistic pairs on IFN-g-activated MPM. We recently demonstrated that representative members of the cystatin superfamily, and particularly CC, can stimulate the release of NO from IFN-g-activated MPM by stimulating the iNOS/NO system [12]. The results reported here show that CC stimulates the synthesis of TNF-a and IL-10 and


upregulates the NO production by IFN-g-activated MPM. The early iNOS induction by IFN-g, followed by a CC stimulation, leads to maximal production of NO by MPM. This suggests that CC acts as an amplifier, but only if the iNOS/IFN-g system is activated. The immunoblotting and immunofluorescence studies show that CC increases the intracellular level of the iNOS protein. CC could upregulate transcription of the iNOS gene directly or indirectly and then increase iNOS mRNA. Actinomycin D and cycloheximide that block protein synthesis at the transcription and translation levels, respectively, abolished the effect of CC, indicating that CC acts via a nuclear mechanism. These results, plus those of the immunofluorescence and immunoblotting studies are in agreement with the long delay (more than 8 h) required for CC-dependent NO production. Many studies have demonstrated that cytokines cooperate with IFN-g to stimulate mRNA synthesis [5,21,26]. TNF-a is a pleiotropic cytokine produced by several cell types (T- and B-cells, mastocytes and natural killer cells) including macrophages in response to a large array of stimuli (virus, parasites, LPS) [27]. By investigating the possible role of TNF-a, we found that MPM incubated with CC or with IFN-g and CC, released increased amounts of this cytokine known to cooperate with IFN-g and to significantly increase iNOS mRNA and nitrite production by binding to its cell membrane receptors [28]. Though CC stimulates TNF-a synthesis, NO production is only increased by CC when MPM are stimulated with IFN-g, probably because iNOS synthesis must be activated for TNF-a to act [26]. The role of this cytokine was assessed by adding rTNF-a to IFN-g-activated MPM; this resulted in a high level of NO production whereas the addition of neutralizing anti-TNF-a sharply reduced the NO production. IL-10 is an activating as well as a deactivating cytokine constitutively produced by several cell types (T-cells, certain B-cells and monocytes) including macrophages [29]. It can enhance but also inhibit NO synthesis by IFN-g-activated mouse macrophages depending on whether it is added before or after the IFN-g- (or LPS)-mediated activation of MPM [5,6,9,30]. High levels of IL-10 were released by unactivated MPM on incubation with CC. However, this was not associated with increasing NO production, as also previously shown [5±7]. However, IL-10 release by unactivated MPM could explain the inhibitory activity on NO upregulation by MPM pretreated with CC before IFN-g activation. An inhibitory effect of IL-10 on NO synthesis by TNF-a and IFN-g-activated MPM has been previously reported (31±33). Finally, the addition of neutralizing anti-IL-10 antibodies had no effect on NO upregulation. This is probably due to the endogenous synthesis of IL-10 [9]. These results are supported by a recent study showing that a cysteine protease inhibitor isolated from a rodent nematode can enhance IL-10 production by murine spleen cells [34]. To summarize, the results reported here indicate that CC stimulates the release of TNF-a and IL-10 by IFN-gactivated MPM. This observation could be of biological importance as cystatin concentrations needed to upregulate TNF-a, IL-10 and NO synthesis are in the physiological range (10±1000 nm) found in human body fluids [11]. The CC-mediated release of TNF-a is probably responsible for the increase in NO production by IFN-g-activated MPM. The synthesis of IL-10 by MPM first incubated with CC and then activated with IFN-g could explain the low production of NO by so-treated cells. Our findings point to a new relationship between cystatins, cytokines, inflammation and immune responses.

q FEBS 1999

ACKNOWLEDGEMENTS We thank Prof. A. Billiau and Dr H. Heremans (Rega Institute, Katholieke Universiteit Leuven, Leuven, Belgium) for the gift of murine recombinant IFN-g. LO-DNP-2 hybridoma cells and the rabbit polyclonal anti-(mouse TNF-a) were kindly provided, respectively, by Dr H. Bazin (Experimental Immunology Unit, UniversiteÂ Catholique de Louvain, Belgium) and by Prof. Wim Buurman (Department of Surgery, University of Limburg, Maastricht, the Netherlands). This work was supported in part by a grant from Action de Recherche ConcerteÂe, U.L.B. 1991, 1994 and 1996 and Fonds Emile Defay. L. V. is the recipient of a postdoctoral grant from the Region Centre (France).

REFERENCES 1. Lowenstein, C.J., Dinerman, J.L. & Snyder, S.H. (1994) Nitric oxide: a physiological messenger. Ann. Intern. Med. 120, 227±237. 2. MacMicking, J., Xie, Q.-W. & Nathan, C. (1997) Nitric oxide and macrophage function. Annu. Rev. Immunol. 15, 323±350. 3. Michel, T., Xie, Q.-W. & Nathan, C. (1996) Molecular biological analysis of nitric oxide synthases. In Methods in Nitric Oxide Research (Feelisch, M. & Stamler, J.S., eds), pp. 161±175. Wiley & Sons, Chichester, UK. 4. Frankova, D. & Zidek, Z. (1998) IFN-g-induced TNF-a is a prerequisite for in vitro production of nitric oxide generated in murine peritoneal macrophages by IFN-g. Eur. J. Immunol. 28, 838±843. 5. Corradin, S.B., Fasel, N., BuchmuÈller-Rouiller, Y., Ransijn, A., Smith, J. & MaueÈl, J. (1993) Induction of macrophage nitric oxide production by interferon-g and tumor necrosis factor-a is enhanced by interleukin-10. Eur. J. Immunol. 23, 2045±2048. 6. Chesrown, S.E., Monnier, J., Visner, G. & Nick, H.S. (1994) Regulation of inducible nitric oxide synthase mRNA levels by LPS, IFN-g, TGF-b, and IL-10 in murine macrophage cells lines and rat peritoneal macrophages. Biochem. Biophys. Res. Comm. 200, 126±134. 7. Appelberg, R. (1995) Opposing effects of interleukin-10 on mouse macrophage functions. Scand. J. Immunol. 41, 539±544. 8. Loftis, L.L., Meals, E.A. & English, B.K. (1997) Differential effects of pentoxifylline and interleukin-10 on production of tumor necrosis factor and inducible nitric oxide synthase by murine macrophages. J. Infect. Dis. 175, 1008±1011. 9. Jacobs, F., Chaussabel, D., Truyens, C., Leclercq, V., Carlier, Y., Goldman, M. & Vray, B. (1998) Interleukin-10 upregulates nitric oxide synthesis by lipopolysaccharide-activated macrophages: improved control of Trypanosoma cruzi infection. Clin. Exp. Immunol. 113, 59±64. 10. Turk, B., Turk, V. & Turk, D. (1997) Structural and functional aspects of papain-like cysteine proteinases and their protein inhibitors. Biol. Chem. Hoppe-Seyler 378, 141±150. 11. Henskens, Y.M.C., Veerman, E.C.I. & Nieuw Amerongen, A.V. (1996). Cystatins in health and disease. Biol. Chem. Hoppe-Seyler. 377, 71±86. 12. Verdot, L., Lalmanach, G., Vercruysse, V., Hartman, S., Lucius, R., Hoebeke, J., Gauthier, F. & Vray, B. (1996) Cystatins up-regulate nitric oxide release from interferon-g-activated mouse peritoneal macrophages. J. Biol. Chem. 271, 28077±28081. 13. Munoz-Fernandez, M.A., Fernandez, M.A. & Fresno, M. (1992) Synergism between tumor necrosis factor-a and interferon-g on macrophage activation for the killing of intracellular Trypanosoma cruzi through a nitric oxide-dependent mechanism. Eur. J. Immunol. 22, 301±307. 14. Olivares Fontt, E. & Vray, B. (1995) Relationship between granulocyte macrophage-colony stimulating factor, tumor necrosis factor-a and Trypanosoma cruzi infection in murine macrophages. Parasite Immunol. 17, 135±141. 15. Riches, D.W. & Underwood, G.A. (1991) Expression of interferon-beta during the triggering phase of macrophage cytocidal activation. Evidence for an autocrine/paracrine role in the regulation of this state. J. Biol. Chem. 266, 24785±24792.

q FEBS 1999


16. Stuehr, D.J. & Marletta, M.A. (1985) Mammalian nitrate biosynthesis: mouse macrophages produce nitrite and nitrate in response to Escherichia coli lipopolysaccharide. Proc. Natl Acad. Sci. USA 82, 7738±7742. 17. Anastasi, A., Brown, M.A., Khembavi, A.A., Nicklin, M.J.H., Sayers, C.A., Sunter, D.C. & Berrett, A.J. (1983) Cystatin, a protein inhibitor of cysteine proteinases. Improved purification from egg white, characterization, and detection in chicken serum. Biochem. J. 211, 129±138. 18. Mosmann, T.R., Schumaker, J.H., Fiorentino, D.F., Leverah, J., Moore, K.W. & Bond, M.W. (1990) Isolation of monoclonal antibodies specific for IL-4, IL-5, IL-6, and a new Th-2specific cytokine (IL-10), cytokine synthesis inhibitory factor, by using a solid phase radioimmunoadsorbent assay. J. Immunol. 145, 2938±2945. 19. GeÂrard, C., Bruyns, C., Marchant, A., Abramowicz, D., Vandenabeele, P., Delvaux, A., Fiers, W., Goldman, M. & Velu, T. (1993) Interleukin 10 reduces the release of tumor necrosis factor and prevents lethality in experimental endotoxemia. J. Exp. Med. 177, 547±550. 20. Green, L.C., Wagner, D.A., Golowski, J., Skipper, P.L., Wishnock, J.S. & Tannenbaum, S.R. (1982) Analysis of nitrate, nitrite and (15N) nitrate in biological fluids. Anal. Biochem. 126, 136±138. 21. Stuehr, D.J. & Marletta, M.A. (1987) Induction of nitrite/nitrate synthesis in murine macrophages by BCG infection, lymphokines, or interferon-g. J. Immunol. 139, 518±525. 22. Weisz, A., Ogushi, S., Cicatiello, L. & Esumi, H. (1994) Dual mechanism for the control of inducible-type NO synthase gene expression in macrophages during acivation by interferon-gamma and bacterial lipopolysaccharide. Transcriptional and post-transcriptional regulation. J. Biol. Chem. 269, 8324±8433. 23. Xie, Q.-W., Cho, H.J., Calaycay, J., Munford, R.A., Swiderek, K.M., Lee, T.D., Ding, A., Troso, T. & Nathan, C. (1992) Cloning and characterization of an inducible nitric oxide synthase from mouse macrophages. Science 256, 225±228. 24. Hevel, J.M., White, K.A. & Marletta, M.A. (1991) Purification of the inducible murine macrophage nitric oxide synthase. J. Biol. Chem. 266, 22789±22791.

25. Ding, A.H., Nathan, C.F. & Stuehr, D.J. (1988) Release of reactive nitrogen intermediates and reactive oxygen intermediates from mouse peritoneal macrophages. J. Immunol. 141, 2407±2412. 26. Deng, W., Thiel, B., Tannenbaum, C.S., Hamilton, T.A. & Stuehr, D.J. (1993) Synergistic cooperation between T cell lymphokines for induction of the nitric oxide synthase gene in murine peritoneal macrophages. J. Immunol. 151, 322±329. 27. Beutler, B.A. & Van Huffel, C. (1994) Unraveling function in the TNF ligand and receptor families. Science 264, 667±668. 28. Drapier, J.-C., Wietzerbin, J. & Hibbs, J.B. (1988) Interferon-g and tumor necrosis factor induce the l-arginine-dependent cytotoxic effector mechanism in murine macrophages. Eur. J. Immunol. 18, 1587±1592. 29. Howard, M., O'Garra, A., Ishida, H., de Waal Malefyt, R. & de Vries, J. (1992) Biological properties of interleukin 10. J. Clin. Immunol. 12, 239±247. 30. Cunha, F.Q., Moncada, C. & Liew, F.Y. (1992) Interleukin-10 (IL-10) inhibits the induction of nitric oxide synthase by interferon-g in murine macrophages. Biochem. Biophys. Res. Commun. 182, 1155±1159. 31. Oswald, I.P., Gazzinelli, T.R., Sher, A. & James, S.L. (1992) IL-10 synergizes with IL-4 and transforming growth factor-b to inhibit macrophage cytotoxic activity. J. Immunol. 148, 3578±3582. 32. Gazzinelli, T.R., Oswald, I.P., James, S. & Sher, A. (1992) IL-10 inhibits parasite killing and nitrogen oxide production by IFN-gactivated macrophages. J. Immunol. 148, 1792±1796. 33. Gazzinelli, R.T., Oswald, I., Hieny, S., James, S. & Sher, A. (1992) The microbicidal activity of interferon-g treated macrophages against Trypanosoma cruzi involves an l-arginine-dependent nitrogen oxide mediated mechanism inhibitable by interleukin 10 and transforming growth factor-b. Eur. J. Immunol. 22, 2501±2506. 34. Hartmann, S., Kyewski, B., Sonnenburg, B. & Lucius, R. (1997) A filarial cysteine protease inhibitor down-regulates T cell proliferation and enhances interleukin-10 production. Eur. J. Immunol. 27, 2253±2260.