ur Resplr J 1992, 5, 67-72 Dexamethasone modulation of tumour nerosis fator-a (ahetin) release by ativated normal human alveolar marophages N. Martinet, P. Vaillant, Th. Charles, J. Lambert, Y. Martinet* Dexamethasone modulation of tumour nerosis fator-a (ahetin) release by ativated normal human alveolar marophages. N. Martinet, P. Vaillant, Th. Charles, J. Lambert, Y. Martinet. ABSTRACT: Reurrent infetions of the lower respiratory trat are a frequent and serious side-effet of hroni ortiosteroid treatment. Sine alveolar marophages () are urrently thought to play a entral role In the protetion of t.he lower respiratory trat against infetious agents, it Is likely that a steroid Indued defiieny of Is Involved In this proess. In this respet, when ativated, are major produers of tumour nerosis fator-a (TNF or ahetln), a versatile ytokine with several biologial properties Inluding antiviral and anti-infetious ativities. A defiit of TNF prodution Indued by ortlosterolds may be one mehanism of the sensitivity to infetions. Thus normal human obtained by bronhoalveolar lavage were pretreated with dexamethasone (DXM) before ativation with llpopolysaharides (LPS) and the amounts of TNF released In ulture were quantified. Pretreatment with DXM resulted In a marked derease of TNF release In a dose-dependent fashion. In ontrast, when were ativated with LPS before DXM treatment, TNF release by was suppressed in a more limited fashion. Thus DXM suppression of LPS-atlvated ability to release TNF may play a role in the suseptibility to Infetions of patients hronially treated with ortlosterolds. ur Respir J, 1992, 5, 67-72. JNSRM Ul4, and Servie de Pneumologie Vandoeuvre-les-Nany Frane. Correspondene: Y. Martinet, INSRM Ul4, C.O. n 1 54511 Vandoeuvre-les-Nany Cedex Frane. Keywords; Alveolar marophages ortiosteroids endotoxi shok tumour nerosis fator-a ( ahetin). This work was supported, in part, by grants from Ligue National entre le Caner, Federation Nationale and Comites de Meurthe et Moselle et de Me use. Cahetin, or tumour nerosis fator-a (TNF), a ytokine produed by mononulear phagoytes, an stimulate numerous biologial ativities (1-4] inluding: ytotoxiity for ertain tumour ell lines, indution of ahexia and endotoxi shok, antiviral [5-7] and anti-infetious ativities [8-12]. In this respet, the human lower respiratory trat is physiologially sterile and this property is urrently thought to be related to the presene of alveolar marophages () [13, 14]. Thus, the ability of to release in situ large amounts of TNF may partiipate in the protetion of the lower respiratory trat against infetions [15, 16]. Impairment of lung defenes against infetions is a ommon and serious side-effet of hroni ortiosteroid treatment [17] and, sine TNF is a ytokine likely to play a role in the protetion of the lower respiratory trat, lipopolysaharide(lps)-ativated normal were evaluated with or without pretreat ment with ortiosteroids (dexamethasone (DXM)), demonstrating that pretreatment with DXM suppressed TNF release, while, in ontrast, when DXM treatment was applied after LPS ativation, TNF release was redued in a limited fashion. Materials and methods Isolation of normal alveolar marophages Fourteen normal subjets (9 men, 5 women, mean age 37±5 yrs), all nonsmokers, with no urrent or past history of lung disorders and with normal hest X-ray and lung funtion test, underwent bronhoalveolar lavage (BAL), as desribed previously (18]. Briefly, after loal anaesthesia, suessive aliquots of 5 ml of sterile saline solution were infused and subsequently reovered by gentle aspiration. For eah subjet, three different lobes were lavaged with a total maximum of 25 ml of saline solution instilled. The reovered ells were separated from the epithelial lining fluid by entrifugation and resuspended for total ell ount and lavage ell differential (by ytoentrifugation preparation): : 92±3%, lymphoytes: 6±2%, neutrophils 1±1 %, and eosinophils 1 %. were purified by adherene to plasti dish (1 h, 37 C, 1% CO in 24- well Falon plates (Beton Dikinson Labware, Lin oln Park, NJ, USA) in Dulbeo's modified agle's medium (DMM, Sigma, St Louis, MO, USA). In order to have a relatively fixed number of in
68 N. MARTINT T AL. eah well for every subjet (1±.3 x 1Q6 ml' 1 ), the number of lavage ells plated depended on the initial number of ounted by ell differential. After adherene, non-adherent ells were eliminated by several washes. purity was >97% and viability >93% (assessed by trypan blue exlusion). Ativation of alveolar marophages Purified were ativated by addition of lipopolysaharides (LPS, sherihia oli 127; B8, Sigma) and ultured (24 h, DMM, 1% COJ. As an initial step, different onentrations of LPS were tested and subsequently, if not otherwise stated, an optimal standard onentration of 1 J.tg mt-1 was used. At this onentration, LPS had no signifiant toxi effet on and did not interat with TNF quantifiation. Tumour nerosis fator-a quantifiation TNF present in supernatants was deteted and quantified by its ytotoxi ativity on mouse L-929 ells (CCLl; Amerian Tissue Culture Colletion, Rokville, MD, USA) using the assay desribed by AooARWAL and KoHR [19]. Reombinant TNF was used as a standard, and anti-tnf antibody was added in positive samples for speifiity. Both TNF and anti TNF antibody were generous gifts from Boehringer Ingelheim (Vienna, Austria). Media alone was the negative ontrol and guanidium hydrohloride (3 M, Sigma) was used for maximum lysis. After inubation of L-929 ells (18h, 37, 1% COJ in different onditions, the ells were washed with old phosphate buffered saline (PBS), and ell viability was assessed by rystal violet (Sigma) inorporation (3 min, 24 q by viable L-929 ells [19], and quantified at 54 nm after ell lysis in distilled water as shown in figure la..5 "t I().4.!!2 Cl) >- 16 IU :.3 Cl).2 ;::.1 A Suessive dilutions (1 to 1 in DMM) of eah sample of supernatant were tested in dupliate and the amount of TNF present was determined in regard to TNF standard urves (fig. lb). The results are expressed as number of TNF units released by 16. Several ontrols were used to assure that the ytotoxiity observed was due to the speifi presene of TNF: 1) sine interleukin-1 (IL-l), another ytokine produed (but in small amounts) by, an have, in some experimental onditions, a ytotoxi ativity on L-929 ells [2], IL-l-a and IL-l- were tested for ytotoxiity on the ells used. In our hands, neither IL-l-a nor IL-1-13 had ytotoxi ativity in the test, onsistent with the fat that some sublones of L-929 ells are not sensitive to IL-l ytotoxiity; 2) to further onfirm the speifi detetion of TNF by this test, eah positive sample was retested in parallel, straight and after inubation with anti-tnf antibody (1 h, 24 q demonstrating that only 4±2% of the ytotoxiity measured was related to the presene of mediator(s) other than TNF-a. Dexamethasone treatment of alveolar marophages Dexamethasone (DXM, Sigma) if not otherwise stated, was added to for 1 h, at 37, in 1% C ; 2 the ells were then washed three times to remove DXM present in the supernatants before ativation by addition of LPS, and ulture (24 h, DMM, 1% COJ. On the other hand, when DXM was applied to after LPS ativation, LPS (1 J.tg ml 1) was added for 1 h, at 37, in 1% C, 2 before treatment with DXM applied as desribed above. Several dilutions of DXM (1 5 to 1Q 1 M) were tested. However, DXM at 1Q 5M onentration had some effets on ell viability. At the onentrations used (1-6 to 1Q 9M), DXM had no speifi effet on TNF assay. "t I() Cl).5.4 16.3 IU ;::.2 -Cl) ;::.1 B +anti-tnf-a antibody 2 4 6 8 1 Celllysis % r-tnf-a Fig. 1. - Quantifiation of tumour nerosis fator-a (TNF-a) by its ytotoxi ativity on mouse L-929 ells. A) Colorimetri assessment by rystal violet inorporation of L-929 ell lysis as desribed in Materials and Methods. B) xample of standard urve of ytotoxiity of inreasing onentrations of reombinant TNF (r-tnf-a u ml 1) tested straight or after inubation with anti-tnf-a antibody. 2 3 4 8
Statistial analysis All data are presented as mean±standard error of the mean. All statistial omparisons were made using a two-tailed Student's t-test. DXTHASON SUPRSSION OF TNF- a RLAS BY 69 A LPS: 1 149 B preliminary observation, were subsequently ativated with 1 14g mp of LPS. LPS-ativated release of TNF was onfirmed for all subjets (fig. 2B): 839±72 TNF units vs 2±1 1NF units released by inativated, po.ol. 1 1 1 alone CJ +LPS Cultural duration h Fig. 2. - Release of tumour nerosis fator-a (TNF a) by ativated normal human alveolar marophages (). were ultured alone or with addition of lipopolysaharides (LPS). The amounts of TNF present in the supernatants after ulture were quantified as desribed in Materials and Methods. A) Time-dependent and LPS dose-dependent prodution of TNF by : eah point represents the mean for two evaluations; the data shown orrespond to one representative experiment using alveolar marophages of one subjet. B) TNF release over 24 h in ulture by from all subjets with or without ativation by LPS (1 J.Lg ml 1). 1 f LPS 1 A DXM 1" 9 1-19 B 1o 8 149 If... ::) 1 ts I u. 1"6 1!9 If... ::) ts 11. 1 4 12 24 Culture duration 1 r +LPS +LPS +DXM Fig. 3. - Suppression by pretreatment with dexamethasone of tumour nerosis fator-a (TNF-a) release b> ativated normal human alveolar marophages () were pretreated by serial dilutions of dexamethasone (DXM) before being ativated by addition of Hpopolysabnides (LPS, 1 ll& mj 1) and ultured for various durations of time. The amounts of TNF present in the supernatants after ullu.re were quantified as desribed in Materials and Methods. A) Dose-dependent inhibition by dexamethasone of TNF release by : eah point represents the mean for two evaluations; the data shown orrespond to one representative experiment using of one subjet. B) Dexamethasone (1" M) inhibition of TNF release by LPS (1 J.Lg mj 1) ativated. Results In order to define TNF optimal release by ativated human, from normal subjets were ultured alone or in the presene of inreasing onentrations of LPS (fig. 2A). Whilst normal did not spontaneously release any signifiant amount of TNF, the addition of LPS indued a dose-dependent and time-dependent release of TNF. Maximum release of 1NF was observed with an LPS onentration of 1 J.tg m1 1 without ell toxiity and without effet on 1NF quantifiation by L-929 test. As a result of this
7 N. MARTINT T AL. To address the question of DXM effets on TNF release, normal were pretreated by inubation with DXM and, after removal of DXM, the same were ativated by LPS and ultured. TNF release was suppressed in a dose-dependent fashion with a maximal suppression, without ytotoxiity and without effet on TNF assay, for a onentration of DXM of 1Q 6M (fig. 3A), and onentrations lower than 1 8M were without effet on TNF release. As a result of this preliminary observation, from all subjets were pretreated with 1Q 6M of DXM before LPS ativation. This proedure resulted in a onstant and marked suppression of TNF release (fig. 3B): 86±13 vs 839±72 TNF units, po.ol. A 1 To evaluate whether this suppression was only speifi for optimally ativated (1 -tg ml 1 of LPS) or was also observed with ultured in the presene of lower onentrations of LPS, from several normal subjets were pretreated with 1Q 6M DXM before ativation with dilutions of L PS (fig. 4). Interestingly, for eah onentration of LPS, there was a suppression of TNF release in relation with DXM pretreatment: LPS 1 J.Ag ml 1: 839±72 vs 86±13 TNF units, p.1; LPS 5 -tg ml 1 : 666±36 vs 58±23 TNF units, p.1; LPS 2.5 J.Ag ml 1: 224±57 vs 53±26 TNF units, p.1; and LPS 1 -tg ml 1 : 127±51 vs 23±14 TNF units, p.5. B 1 CO b "':" 1 :::> Cl I l.l. z f- 1 +LPS 11-19 5 -tg 2.5J.t9 1 1-19 CO b Cl I l.l. z r- 1 1 +DXM +DXM +DXM +DXM +LPS 1!l9 5 -tq 2.5 J.t9 1 J.A9 Fig. 4. - Suppression by pretreatment with dexamethasone of tumour nerosis fator-a (TNF-a) release by normal human alveolar marophages () ativated with several onentrations of lipopolysaharides. obtained from normal subjets (n=s) were ultured (24 h) after ativation by addition of dilutions of lipopolysaharides (LPS). The amounts of TNF present in the supernatants after ulture were quantified as desribed in Materials and Methods. A) Absene of pretreatment with dexamethasone. B) were pretreated with dexamethasone (DXM; 1 6 M) before LPS ativation. 1 1 1 1 Finally, in order to define whether DXM treatment ould suppress TNF release only when applied before LPS ativation or also when applied after LPS ativation, normal were isolated and ultured in parallel either with DXM treatment before LPS ativation or with LPS ativation before DXM treatment. As a result (fig. 5), DXM treatment was shown to have a limited effet on TNF release by when applied after LPS ativation: 5347 TNF units vs 839± 72 TNF units with LPS alone (po.ol ), and vs 86±13 TNF units with DXM treatment prior to LPS ativation (p.1), suggesting that, one ativated, are less sensitive to DXM treatment. +LPS +DXM +LPS +LPS +DXM Fig. 5. - Limited suppression of tumour nerosis fator-a (TNF-a) release by normal alveolar marophages () ativated w ith lipopolysahrides before dexamethasone treatment. from normal subjets (n=6) were ultured for 24 h, alone (), with addition of lipopolysabarides (1 -tg mt-t) (+LPS), with addition of LPS after pretreatment with dexamethasone (1 6 M) (+DXM+LPS), or with treatment with DXM after preativation with LPS (+LPS+DXM). The amounts of TNF present in the supernatants after ulture were quantified as desribed in Materials and Methods. Disussion Ativated human release high amounts of TNF, and the suppression of this release by pretreatment with DXM is of interest in respet to this ytokine 1 S biologial ativities and the urrent understanding of its role in lung disorders. T NF is a monokine initially purified as a fator induing in vitro and in vivo ell lysis of some aner ell lines [1-4], and haraterized by other major
DXTIIASON SUPRSSION OF TNF-a RLAS BY 71 ativities involved with: ahexia, anti-infetious defenes, regulation of aute phase protein gene expression, artilage and bone resorption, and the modulation of endothelial ell properties [1-4). In the lung, ativated normal produe elevated amounts of TNF in omparison to autologous blood monoytes, suggesting that high in situ levels of TNF an be present in the lung when are ativated [15, 16). Furthermore, TNF release has been suggested to play a role in several lung disorders inluding: the adult respiratory distress syndrome (ARDS) ompliating the endotoxi shok (21], saroidosis [22), and pneumooniosis [23). Cortiosteroids exert their effets by interating with speifi reeptors present on target ells. In this respet, express reeptors for ortiosteroids [24), and DXM modulation of TNF release has previously been shown to be exerted at both transriptional and post-transriptional levels for mouse marophages, and human blood monoytes [25, 26). Furthermore, DXM, at onentrations above 1 5M, has been suggested to redue TNF prodution by normal human [16]. Our observations an be related to the alteration of the anti-infetious defenes indued by hroni ortiosteroid treatment [17), and to the pathogeniity of the endotoxi shok. TNF is likely to play a role in lung protetion against infetions by eliiting neutrophils and by ativating them by indution of their phagoytosis, ytotoxi ativities, degranulation, and release of reative oxygen radials in parallel with a stimulation of marophage ytotoxiity in an autorine fashion [8-12, 27, 28). Furthermore, TNF an indue an antiviral state in some target ells [5-7]. Thus, the absene of a high loal release of TNF by ativated by infetious bodies present in the lower respiratory trat an lead, in parallel with the steroid-indued suppression of a number of biologial proesses, to a derease of lung defenes and the subsequent development of an infetious ondition. This information is also relevant to the pathogeniity of the endotoxi shok, knowing that, at least in experimental models, TNF has been shown to be able to indue a shok [1-4) and that pretreatment by anti-tnf antibody prevents the animal from developing a toxi shok after endotoxin injetion [29]. Furthermore, pretreatment with ortiosteroids an prevent the endotoxi shok whhe, one the endotoxinindued pathologial state has been started, ortiosteroid treatment is unable to stop the evolution (3]. These data and linial observations [21) suggest that TNF release may play a entral role in the pathogeniity of the ARDS ompliating the endotoxi shok. In this respet, the marked suppression by DXM of TNF release by, only before th eir ativation, is onsistent with the fat that ortiosteroids would be effiient in preventing an endotoxi shok with ARDS only when presribed as a preventive treatment in patients likely to develop this ondition. Referenes 1. Ruff MR, Gifford G. - Tumor nerosis fator. Lymphokines, 1981, 2, 235-272. 2. Beutler B, Greenwald D, Hulmes JD, Chang M, Pan Y-C, Mathison J, Ulevith R, Cerami A. - Identity of tumor nerosis fator and the marophage-sereted fator ahetin. Nature, 1985, 316, 552-554. 3. Beutler B, Cerami A. - Cahetin and tumor nerosis fator as two sides of the same biologial oin. Nature, 1986, 32, 584-588. 4. Dinarello CA. - Cytokines: interleukin-1 and tumor nerosis fator (ahetin). /n: Inflammation: Basi Priniples and Clinial Correlates. J.I. Gallin, I.M. Goldstein, R. Snyderman, eds., Raven Press Ltd, New York, 1988, pp. 195-28. 5. Mestan J, Digel W, Mittnaht S, Hillen H, Blohm D, Muller A, Jaobsen H, Kirhner H. - Antiviral effets of reombinant tumor nerosis fator in vitro. Nature, 1986, 323, 816-819. 6. Wong GHW, Goeddel DV. - Tumor nerosis fators a and inhibit virus repliation and synergize with interferons. Nature, 1986, 323, 819-822. 7. Koff WC, Fann A V. - Human tu m or nerosis fator-a kills herpes virus-infeted but not normal ells. Lymphokine Res, 1986, 5, 215-221. 8. Klebanoff SJ, Vadas MA, Harlan JM, Sparks LJ, Gamble JR, Agosti JM, Waltersdorph. - Stimulation of neutrophils by tumor nerosis fator. J lmmunol, 1986, 136, 422-4225. 9. Tsujimoto M, Yokota S, Vilek J, Weissman C. - Tumor nerosis fator provokes superoxide anion generation from neutrophils. Biohem Biophys Res Commun, 1986, 137, 194-11. 1. Djeu JY, Blanhard DK, Halkias D, Friedman H. - Growth inhibition of Candida albians by human polymorphonulear neutrophils: ativation by interferongamma and tumor nerosis fator. J lmmunol, 1986, 137, 298-2984. 11. Ming WJ, Bersani L, Mantovani A. - Tumor nerosis fator is hemotati for monoytes and polymorphonulear leukoytes. J /mmunol, 1987, 138, 1469-1474. 12. Philip R, pstein LB. - Tumor nerosis fator as immunomodulator and mediator of monoyte ytoxiity indued by itself, gamma-interferon, and interleukin-1. Nature, 1986, 323, 86-89. 13. Hoking WG, Golde DW. - The pulmonary-alveolar marophage. New ng/ J Med, 1919, 31, 58-587, and 639-645. 14. Bowden DH. - The alveolar marophage. nviron Health Perspet, 1984, 55, 327-341. 15. Martinet Y, Yamauhi K, Crystal RG. - Differential expression of the tumor nerosis fator/ahetin gene by blood and lung mononulear phagoytes. Am Rev Respir Dis, 1988, 138, 659-665. 16. Striter RM, Remik DG, Lynh JP Ill, Genord M, Raiford C, Spengler R, Kunkel SL. - Differential regulation of tumor nerosis fator-a in human alveolar marophages and peripheral blood monoytes: a ellular and moleular analysis. Am J Respir Cell Mol Bioi, 1989, 1, 57-63. 17. Siege! SC. - Cortiosteroid agents; overview of ortiosteroid therapy. J Allergy C/in lmmunol, 1985, 76, 312-32. 18. Kleh H, Pohl W. - Tehnial reommedations and guidelines for bronhoalveolar lavage (BAL). ur Respir J, 1989, 2, 561-585.
72 N. MARTINT T AL. 19. Aggarwal BB, Kohr WJ. - Human tumor nerosis fator. Methods nrymol, 1985, 116, 448-456. 2. Onozaki K. Matsushima K, Aggarwal BB, Oppenheim JJ. - Human interleukin-1 is a ytoidal fator for several tumor ell lines. J Immunol, 1985, 135, 3962-3968. 21. Marks JD, Marks CB, Lue JM, Montgomery AB, Turner J, Metz CA, Murray JF. - Plasma tumor nerosis fator in patients with septi shok. Mortality rate, inidene of adult respiratory distress syndrome, and effets of methylprednisolone administration. Am Rev Respir Dis, 199, 141, 94-97. 22. Bahwih PR, Lynh JP III, Larrik J, Spengler M, Kunkel SL. - Tumour nerosis fator prodution by human saroid alveolar marophages. Am J Pathol, 1986, 125, 421-425. 23. Borm PJA, Palmen N, ngelen JJM, Buurrnan WA. - Spontaneous and stimulated release of tumor nerosis fator-a (TNF) from blood monoytes of miners with oal workers pneumooniosis. Am Rev Respir Dis, 1988, 138, 1589-1594. 24. Werb Z, Foler R, Munk A. - Interation of glyoortioids and marophages. Identifiation of glyoortioid reeptors in monoytes and marophages. J xp Med, 1978, 147, 1684-1694. 25. Beutler B, Krohin N, Milsark lw, Lued.ke C, Cerami A. - Control of ahetin (tumor nerosis fator) synthesis: mehanisms of endotoxin resistane. Siene, 1986, 232, 977-98. 26. Sariban, lmamura K, Leubbers R, Kufe D. Transriptional and posttransriptional regulation of tumor nerosis fator gene expression in human monoytes. J Clin Invest, 1988, 81, 15151. 27. Shalaby MR, Aggarwal BB, Rinderkneht, Sverdesky LP, Finkle BS, Palladino MA Jr. - Ativation of human polymorphonulear neutrophil funtions by interferon-gamma and tumor nerosis fators. J Immunol, 1985, 135, 269-273. 28. Larrik JW, Graham D, Toy K, Lin LS, Senyk G, Fendly BM. - Reombinant tumor nerosis fator auses ativation of human granuloytes. Blood, 1987, 69, 64-644. 29. Beutler B, Milsark lw, Cerami A. - Passive immunization against ahetin/tumor nerosis fator protets mie from lethal effet of endotoxin. Siene, 1985, 229, 869-871. 3. Brigham KL, Bowers R, MKeen CR. - Methylprednisolone prevention of inreased lung vasular permeability following endotoxemia in sheep. J Clin Invest, 1981, 67, 113-111.