CELLULAR
IMMUNOLOGY
14,249-257
(1992)
Sensitization of Rat Alveolar Macrophages to Enhanced TNF-CY Release by in Vivo Treatment with Dexamethasone H. RENZ,*” A. HENKE,* P. HOFMANN,* L. J. WoLFF,t A. SCHMIDT,* J. R~SCHOFF,$ AND D. GEMSA* *Institute of Immunology and *Institute of Pathology, Philipps University, Marburg, Germany; and tDepartment ofPediatrics, Oregon Health Sciences University, Portland, Oregon 97201 Received December 28, 1990; accepted July 12, I992
Treatment of rats with dexamethasone rapidly induced a marked weight loss which occurred within 3 days and persisted for several weeks. The cachectic state was paralleled by increased serum levels of triglycerides, albumin, and protein and a strong reduction of blood mononuclear leukocytes. In lung sections, an increased number of mononuclear giant cells was found but no bacteria, fungi, or Pneumocystis carinii organisms. Quite strikingly, alveolar macrophages from dexamethasone-treated rats, but not from control animals, were highly sensitive to LPS and released large amounts of TNF-a ex vivo. Also under in vivo conditions, high TNF-a serum concentrations were found in dexamethasone-treated but not control rats when examined 14 hr after an intravenous LPS injection. These data suggestthat the glucocorticoid-induced cachexia of rats may be linked, at least in part, to readily inducible TNF-(Yreleasefrom primed macrophages. 0 1992 Academic
Press, Inc.
INTRODUCTION Tumor necrosis factor-a (TNF-(Y) is a cytokine which is primarily produced by macrophages (l-4). It displays a wide variety of actions, including the destruction of tumor cells, activation of leukocytes, modulation of endothelial cell function, and induction of fibroblast stimulation (5-l 1). In addition, TNF-a is a potent mediator of fever ( 12) and plays a major role in the pathogenesisof endotoxin-induced shock ( 13). TNF-a levels have been shown to be elevated in a variety of infectious diseases. Cerami and co-workers reported that cattle and rabbits with trypanosomiasis displayed a wasting syndrome which was characterized by severeweight loss, hypertriglyceridemia, and significant depletion of fat and protein reserves( 14- 17). It was of particular interest that this cachexia accompanying parasitic infections was closely correlated with incre ed serum levels of the cytokine cachectin which was later found to be identical, ito TNF-a. In an ongoing study to induce a Pneumocystis carinii infection in rats by continuous glucocorticoid-treatment we found a rapid weight loss during the first 2 to 4 weeks and this preceded a P. carinii pneumonitis 8 to 12 weeks later. In this respect, the glucocorticoid treatment of rats resembled trypanosomiasis in cattle and rabbits, and the clinical picture suggestedan involvement of TNF-cu/cachectin. ’ Present address: National Jewish Center for Immunology and Respiratory Medicine, Department of Pediatrics, Denver, CO 80206. 249 0008-8749/92 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form resawd.
250
RENZ
ET AL.
MATERIALS AND METHODS Maintenance of Animals Female Lewis rats between 6 and 8 weeks old (160-200 g) were purchased from Savo (Kisslegg,Germany). The animals were delivered naturally, not barrier-sustained, and received no antibiotics prior to shipping. Rats were housed one per nonfiltered cage. The rodent quarters were maintained at 20-22°C and 50-60% humidity. The rats were divided into three groups. Group A received 0.2 mg dexamethasone (Serva, Heidelberg, Germany) and 50 mg tetracycline (Serva) per 100 ml of acidified, filtered tap water. Group B received tetracycline in acidified water. Group C received acidified drinking water only. Rats drank between 20 and 50 ml of water per day. This amounted to between 0.2 and 0.5 pg dexamethasone/g/day. The rats received standard rat chow (Hoveler, Langenfeld-Immigrath, Germany). Rats were observed daily for signs of infection and were weighed every other day. Preparation of Alveolar Macrophages Before the start of therapy and on Days 3, 7, 14, 21, and 28 of treatment, three animals from each group were sacrificed using intraperitoneal pentobarbital (60 mg/ kg) and exsanguination. Rat alveolar macrophagesmere obtained by gentle lavage of the lungs with phosphate-buffered saline without C$‘/Mg2’. Cells were washed three times with serum-free RPM1 1640 supplemented with L-glutamine (2 mJ@, penicillin ( 100 U/ml), streptomycin ( 100 pg/ml), and mercaptoethanol ( 10 mM). After plating for 1 hr in 24-well culture plates in RPM1 1640 medium plus 10% heat-inactivated (56’C, 30 min) fetal calf serum, the monolayers were washedseveraltimes with medium to remove the nonadhering cells. The remaining monolayer (0.5 X 106/ml) contained >95% macrophagesas determined by phagocytosis of carbon particles and nonspecific esterasestaining (17, 18). The incubation was continued for 20 hr, with or without the addition of Escherichia coli 0127:B8 lipopolysaccharide (LPS, 1 pg/ml). At the end of the incubation, cell-free supernatant was harvested and stored at -20°C until assayedfor TNF activity. In Vivo Injection of LPS Three rats of each of the groups A, B, and C were injected on Day 7 with LPS at a dose of 0.6 mg/kg body wt. After 11 hr, the animals were sacrificed and serum was carefully collected in order to determine the contents of TNF-a, interleukin-la (IL1(Y),and interleukin-6 (IL-6). Determination of TNF-a The amount of TNF-(U in culture supernatants was measured by using the cytotoxicity against TNF-a-sensitive L929 cells as previously described (19,20). Briefly, L929 cells (6 X 104/0.2 ml) were grown in RPM1 1640 to establish a dense monolayer in 96-well microtiter plates. Macrophage culture supernatants or human recombinant TNF-a (kindly provided by BASF/Knoll AG, Ludwigshafen, Germany) or murine recombinant TNF-ar (kindly provided by Dr. G. R. Adolf, Ernst-Boehringer-Institute, Vienna, Austria) were diluted in RPM1 1640 and added to L929 cells in the presence of actinomycin D (1 pug/ml)to enhance L929 sensitivity to TNF. After an incubation for 18 hr, the viability of L929 cells was measured by staining for 1 hr with 3-(4,5-
DEXAMETHASONE
IN VZVO AND TNF-a
RELEASE
251
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, 20 &well of a 5 mg/ ml stock solution). After completion of the MTT incubation, culture supernatants were removed and cells were lysed by an isopropanol solution containing 40 nut4 HCl. Plates were read in a microplate-autoreader MR 600 (Dynatech, Denkendorf, Germany) at 540 nm. TNF-a activity in macrophage culture supematants is expressed in units per milliliter, based on a standard curve established with either murine or human recombinant TNF-cr. A unit (U) was defined as the TNF-a amount required to lyse 50% of L929 target cells. TNF-(U specificity was corroborated by using both TNF+ insensitive L929 cells and neutralization of TNF-a by a polyclonal rabbit antimurine TNF-(U antiserum (kindly provided by Dr. D. MPnnel, German Cancer Research Center, Heidelberg, Germany). The content of TNF-a in serum from dexamethasone-treated and control rats was determined by a highly specific ELISA for murine TNF-cu. The assay system was established in our laboratory by using a monoclonal hamster anti-murine TNF-(Y antiserum (Genzyme, Munich, Germany), a polyclonal rabbit anti-murine TNF-(U antiserum (kindly provided by Dr. G. R. Adolf, Ernst-Boehringer-Institute, Vienna, Austria), and a peroxidase-labeledgoat anti-rabbit IgG antiserum (Dianova, Hamburg, Germany). Determination of IL-la and IL-6 The concentration of IL- 1crin serum from dexamethasone-treated and control rats was determined by a sensitive and specific radioimmunoassay for the detection of rat IL- 1a (Cytokine Sciences,Inc., Boston, MA). The content of IL6 serum was measured by a sensitive ELISA that was capable of detecting murine as well as rat IL-6 (Endogen, Inc., Boston, MA). Determination of Prostaglandin E2 (PGEj The content of PGE2 in macrophage culture supematants was measured after 20 hr of incubation by radioimmunoassay as previously described (19, 20). Hematological and Chemistry Determinations From all animals, blood was obtained and absolute numbers of leukocytes and differentiation into neutrophil granulocytes and mononuclear cells were determined by standard clinical laboratory methods. The contents of triglycerides, cholesterol, and protein in serum were also determined by established routine methods. Histopathological Evaluation At Day 28 of the study, the lungs of control and dexamethasone-treated animals were aseptically removed and fixed in PBS containing 2% formalin. The lung tissue was embedded in paraffin, sectioned, and stained with hematoxylin and eosin. Tissues were also stained with Giemsa and Gomori methanemine silver to assesslung sections for the presence of P. carinii organisms. RESULTS Rats receiving dexamethasone displayed a weight decreaseof 38% during the initial 2 weeks of treatment, with a particularly sharp decreaseduring the first 3 days (Fig.
252
RENZ ET AL.
Dexamethasone
Days FIG. 1. Loss of body weight of rats treated continuously with dexamethasone. Values show the means + SD of three rats per group and per time point.
1). Control rats, receiving either tetracycline or no medication, showed a regular weight gain during the study. Despite a pronounced weight loss, dexamethasone-treated rats exhibited no apparent signs of apathy, restlessness,respiratory distress, or infections. After 3 days of dexamethasone treatment, the blood mononuclear cells (approximately 80% lymphocytes) dramatically declined to 18% of the control group (Fig. 2) and remained at a very low level during the observation period of 28 days. In contrast, no statistically significant changes in absolute numbers of erythrocytes and thrombocytes were noted between treated and untreated rats, except for neutrophil granulocytes which declined during the initial 3 days of dexamethasone treatment but thereafter returned to normal absolute numbers. When alveolar macrophages were harvested, no difference in cell yield (from 2 to 3 X 106/animal) was found. Upon in vitro culture, the spontaneous release of TNF(Ywas low and did not differ between macrophages from treated and untreated rats. However, in responseto LPS stimulation, macrophages from dexamethasone-treated
-
V’
0
3
7
14
21
28
Days FIG. 2. Decline of blood mononuclear leukocytes during dexamethasone treatment. Values show the means f SD of three rats per time point.
DEXAMETHASONE
IN
253
VZVO AND TNF-a RELEASE
70 60
0
3
7
14
21
20
Days FIG. 3. Enhanced TNF-ol release from alveolar macrophages of rats treated with dexamethasone. Macrophages from dexamethasone (0) or untreated control rats (0) were harvested at indicated time periods and incubated for 20 hr at a concentration of 0.5 X 106/ml in the presenceof LPS (1 &ml), and the culture supematant was assayedfor TNF-(Y activity. Values are the means f SD of macrophages from three rats per time point.
animals displayed a particularly high sensitivity and releasedS- to 1O-fold more TNFa than control macrophages (Fig. 3). The enhanced LPS sensitivity was detected as early as Day 3, remained extremely high for 14 days of dexamethasone medication, and then declined, although it still persisted at a higher level in in vivo dexamethasonetreated macrophages.It was also noted that alveolar macrophagesfrom dexamethasonetreated rats were capable of spontaneously releasing more PGE2 than control macrophages (Fig. 4). LPS was found to be an inefficient stimulus for further PGE2 production in alveolar macrophages. In peritoneal macrophages, an identical pattern of spontaneous PGE2 release was observed and, in addition, those cells were LPS responsive (data not shown). To examine an in vivo production of TNF-a, dexamethasone-treated and control rats received a single intravenous injection of LPS on Day 7. The serum was harvested
Control
Dexometho aone
-
FIG. 4. Enhanced PGEz release from alveolar macrophages of dexamethasone-treated rats. Seven days after dexamethasone or control treatment, macrophages were harvested and incubated for 20 hr at a concentration of 0.5 X 106/ml, and PGEz content was determined in the culture supematant. Values represent the means + SD of macrophages from three rats per group.
254
RENZ ET AL. A
0.5 1
0 m
without LPS with LPS
Control
Dexomethosone
0 without LPS EZl with LPS
200
0 without LPS 6SQ with LPS
60
” t
x
z 60 g ‘p 40 d
150
-p 100
d
20
50
Control
Control
Dexomathosone
Dexomethosone
FIG. 5. Elevated TNF-ol and reduced IL- 101and IL-6 in vivo releasein dexamethasone-treated rats. Seven days after dexamethasone or control treatment, LPS (0.6 mg/kg) was intravenously injected and after 14 hr, the contents of TNF-(U (A), IL- 101(B), and IL-6 (C) in the serum were determined. Values represent the means +- SD of three rats per group.
11 hr later and was assayedfor TNF-a content by a highly specific ELISA (Fig. 5A). Although dexamethasone-treated animals did not spontaneously release TNF-a in circulation, an LPS injection induced a high TNF-(U production, whereas control rats remained rather unresponsive. In striking contrast, the LPS-inducible release of ILla and IL-6 into circulation was markedly reduced in dexamethasone-treated rats (Figs. 5B and 5C). Already after 3 days of treatment, the dexamethasone group showed a marked increase in serum triglyceride levels, approximately six times higher than levels in the TABLE 1 Effect of Dexamethasone on SelectedSerum Components’ Treatment
Triglycerides (mg/dl)
Cholesterol bw/dl)
Protein WW
Albumin Wdl)
Control Dexamethasone (Day 3)
128 +- 24 739 f 12
59.5 + 5 102.0 +- 15
6.0 + 0.2 10.3 + 0.6
3.9 * 0.1 7.0 + 0.3
r?All values are the means f SD of three rats and the tests were performed in triplicate.
DEXAMETHASONE
IN VZVO AND TNF-a RELEASE
255
control group (Table 1). Serum triglycerides remained high during the entire observation period and closely paralleled LPS-inducible TNF-(I! production (Fig. 3). Rather similarly, although not as pronounced, serum concentrations of cholesterol, total protein, and albumin showed significant increases at Day 3 of dexamethasone treatment (Table 1) and remained elevated up to Day 28. In the lung tissue of dexamethasone-treated rats, a marked giant mononuclear cell infiltration was noted (Fig. 6). Despite this cell infiltration, the bronchoalveolar lavage produced no higher cell yield (seeabove). No granulocytes were seenin this infiltrate nor were any P. carinii organisms, fungi, or bacteria observed. DISCUSSION This study provides evidence that TNF-a may play a major role in the marked weight loss which is induced in rats by continuous treatment with dexamethasone. The weight loss occurred rapidly during the first 3 days of dexamethasone therapy, persisted up to 28 days, and was closely associated with a hypertriglyceridemia and enhanced serum cholesterol, albumin, and protein levels. These clinical observations were reminiscent of the cachexia caused by cachectin release in parasitic infections (14-16, 2 1) and the anorexia and severe weight loss induced in rats by injection of TNF-a! ( 13). As a primary source of TNF release,macrophagesand, to a lesserdegree, lymphocytes must be considered. Circulating monocytes or lymphocytes could not be examined in our study since the dexamethasone treatment led to a marked decline of mononuclear cell counts to 5% of the control group (Fig. 2). Therefore, we examined alveolar macrophages which were easily accessibleand were found to be major TNFLYproducers. Alveolar macrophagesfrom dexamethasone-and untreated rats were both inefficient in releasing spontaneously significant amounts of TNF-(Y in the absenceof exogenous stimuli. However, macrophages from dexamethasone-treated rats were clearly preconditioned or “primed” for copious TNF-a! releasewhen they were in vitro exposed to LPS (Fig. 3). Furthermore, not only in vitro but also in vim, an LPS injection led to a marked release of TNF-(U into circulation of dexamethasone-treated rats (Fig.
FIG. 6. Histopathology of cross sections of lungs from control (A, 30X) and dexamethasone-treated (B, 30X; C, 80X) rats. Sections were done 28 days after dexamethasone treatment.
256
RENZ ET AL.
5A). These striking findings represent an entirely novel and, in fact, unexpected observation since glucocorticoids are not regarded as priming factors for TNF-ar release such as interferon-y (22) or GM-CSF (23). On the contrary, glucocorticoids have been shown to reduce cytokine releasewhich has been reconfirmed in this study by parallel measurements of IL-la (Fig. 5B) and IL-6 (Fig. 5C) in serum of dexamethasonetreated rats. Thus, on first sight, our TNF-a data may stand in contradiction to hitherto established models of glucocorticoid actions. Previous in vitro studies have shown that glucocorticoids inhibit TNF-a at both the transcriptional and the translational level by blocking gene transcription and mRNA mobilization (24). Also, when dexamethasone was intraperitoneally given to rats 2 to 8 hr prior to intravenous LPS injection, a 70 to 90% reduction in the serum peak concentration of TNF-a! was observed (25). Similarly, in vitro studies with human monocytes revealed an 80% reduction of TNFa! releasewhen monocytes were preincubated for 48 hr with dexamethasone (26). In contrast to those studies, our results were obtained with rats that had been treated for a prolonged time period (up to 28 days) with dexamethasone, and the TNF-a releasing capacity was examined both ex vivo in alveolar macrophages and in vivo in intact animals. It was of additional interest that alveolar macrophages from dexamethasonetreated animals displayed an enhanced spontaneous PGE2production when compared to control macrophages(Fig. 4). Thus, analogousto TNF-a production, the arachidonic acid metabolism was also shifted to an augmented prostanoid production. It remains to be elucidated whether a common underlying mechanism preactivated both mediator systems or whether they mutually affected each other’s functional role as previously demonstrated (19). How prolonged dexamethasone treatment of rats may selectively enhance LPSinducible release of TNF-a, but not of IL-la or IL-6, from alveolar macrophages remains unclear. A histologic examination of glucocorticoid-treated rats revealed an accumulation of macrophage-like giant cells in the alveolar spacesbut the lung sections showed no neutrophil granulocyte infiltration which argues against a typical bacterial infection. The concomitant administration of tetracycline to dexamethasone-treated rats appears to have protected against bacterial infections. Additionally, a specific staining showed no indication of infection by P. carinii or fungi. However, an infection by viruses cannot be ruled out and has to be taken into consideration, although all clinical signs and preliminary virological tests do not support this possibility. Apart from clinically inapparent infections, it is also feasible that glucocorticoid treatment of rats may have induced a shift of monocyte/macrophage migration in such a manner that TNF-LUproducing subpopulations may have preferentially emigrated to the lung (27, 28). The low numbers of mononuclear cells in circulation may indicate not only that glucocorticoid-sensitive lymphocytes have disappeared, but also that a redistribution of monocytes/macrophages may have taken place. A priming for enhanced TNF-a! releaseindicates that ample TNF-a mRNA should be available for ready translation into the bioactive protein following stimuli such as LPS or others. Work is currently in progress to examine TNF-a and other cytokine gene expression in alveolar macrophages following in vivo glucocorticoid treatment. An enhanced TNF-a mRNA accumulation may be the sequelaeof an enhanced gene expression, possibly mediated by glucocorticoid interference with resynthesis of repressor proteins (29), a prolonged stabilization of mRNA by inhibition of normal degradation mechanisms, or a glucocorticoid-induced production of unknown mediators which in turn causes enhanced TNF-a gene transcription. It may even be speculated that priming of alveolar macrophages for enhanced TNF-(Y release in re-
DEXAMETHASONE
IN VIVO AND TNF-(Y RELEASE
257
sponse to microbial products such as LPS may represent a protective mechanism in the lung against infections that could easily occur during glucocorticoid-induced immunosuppression. ACKNOWLEDGMENTS We thank Amy Beermann for competent secretarial help. We are grateful to Dr. G. R. Adolf, ErnstBoehringer-Institute, Vienna, Austria, for donating murine recombinant TNF-cx and anti-TNF-cYantiserum, and to BASF/Knoll AG, Ludwigshafen, Germany, for providing us with human recombinant TNF-(Y. This study was supported by the Volkswagen-Stiftung (I/65-535).
REFERENCES 1. Old, L. J., Science 230, 630, 1985. 2. Beutler, B., and Cerami, A., Nature 320, 584, 1986. 3. Urban, J. L., Shepard, H. M., Rothstein, J. L., Sugarman, B. J., and Schreiber, H., Proc. Natl. Acad. Ski. USA 83, 5233, 1986. 4. Kornbluth, R. S., and Edgington, T. S., J. Immunol. 137,2585, 1986. 5. Carswell, E. A., Old, L. J., Kassel, R. L., Green, S., Fiore, N., and Williamson, B., Proc. Natl. Acad. Sci. USA 72, 3666, 1975. 6. Esparza, I., Mannel, D., Ruppel, A., Falk, W., and Krammer, P. H., J. Exp. Med. 166, 589, 1987. 7. Heidenreich, S., Weyers, M., Gong, J.-H., Sprenger, H., Nain, M., and Gemsa, D., J. Immunol. 140, 1511, 1988. 8. Philip, R., and Epstein, L. B., Nature 323, 86, 1986. 9. Mantovani, A., and Dejana, E., Zmmunol. Today 10, 370, 1989. 10. Sugarman, B. J., Aggarwal, B. B., Haas, P. E., Figari, I. S., Palladino, M. A., and Shepard, H. M., Science 230,943, 1985. 11. Sherry, B., and Cerami, A., J. Cell Biol. 107, 1269, 1988. 12. Dinarello, 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., J. Exp. Med. 163, 1433, 1986. 13. Tracey, K. J., Beutler, B., Lowry, S. F., Merryweather, J., Wolpe, S., Milsark, I. W., Hariri, R. J., Fahey, T. J., III, Zentella, A., Albert, J. D., Shires, G. T., and Cerami, A., Science 234, 470, 1986. 14. Rouzer, C. A., and Cerami, A., Mol. Biochem. Parasitol. 2, 3 1, 1980. 15. Kawakami, M., and Cerami, A., J. Exp. Med. 154,631, 198 I. 16. Kawakami, M., Pekala, P. H., Lane, M. D., and Cerami, A., Proc. Natl. Acad. Sci. USA 79,9 12, 1982. 17. Gemsa, D., Woo, C. H., Fudenberg, H. H., and Schmid, R., J. C/in. Invest. 52, 8 12, 1973. 18. Lovett, D., Kozan, B., Hadam, M., Resch, K., and Gemsa, D., J. Immunol. 136, 340, 1986. 19. Renz, H., Gong, J.-H., Schmidt, A., Nain, M., and Gemsa, D., J. Immunol. 141, 2388, 1988. 20. Rem, H., Gentz, U., Schmidt, A., Dapper, T., Nain, M., and Gemsa, D., Infect. Immunol. 57, 3172, 1989. 21. Hotez, P. J., Le Trang, N., Fairlamb, A. H., and Cerami, A., Parasite Immunol. 6, 203, 1984. 22. Pace, J. L., Russel, S. W., Torres, B. A., Johnson, H. M., and Gray, P. W., J. Immunol. 130, 2011, 1983.
23. 24. 25. 26. 27.
Heidenreich, S., Gong, J.-H., Schmidt, A., Nain, M., and Gemsa, D., J. Immunol. 143, 1198, 1989. Beutler, B., Krochin, N., Milsark, I. W., Luedecke, C., and Cerami, A., Science 232, 977, 1986. Waage, A., Clin. Immunol. Immunopathol. 45, 348, 1987. Waage, A., and Bakke, O., Immunology 63,299, 1988. Rinehart, J. J., Sasone,A. L., Balcerzak, S. P., Ackerman, G. A., and LoBuglio, A. F., N. Engl. J. Med. 292, 236, 1975. 28. Boggs, D. R., Arthens, J. W., and Cartwright, C. E., Am. J. Pathol. 44,763, 1964. 29. Collart, M. A., Belin, D., Vassalli, J.-D., de Kossodo, S., and Vassalli, P., J. Exp. Med. 164, 2 I 13, 1986.
IMMUNOLOGY
14,249-257
(1992)
Sensitization of Rat Alveolar Macrophages to Enhanced TNF-CY Release by in Vivo Treatment with Dexamethasone H. RENZ,*” A. HENKE,* P. HOFMANN,* L. J. WoLFF,t A. SCHMIDT,* J. R~SCHOFF,$ AND D. GEMSA* *Institute of Immunology and *Institute of Pathology, Philipps University, Marburg, Germany; and tDepartment ofPediatrics, Oregon Health Sciences University, Portland, Oregon 97201 Received December 28, 1990; accepted July 12, I992
Treatment of rats with dexamethasone rapidly induced a marked weight loss which occurred within 3 days and persisted for several weeks. The cachectic state was paralleled by increased serum levels of triglycerides, albumin, and protein and a strong reduction of blood mononuclear leukocytes. In lung sections, an increased number of mononuclear giant cells was found but no bacteria, fungi, or Pneumocystis carinii organisms. Quite strikingly, alveolar macrophages from dexamethasone-treated rats, but not from control animals, were highly sensitive to LPS and released large amounts of TNF-a ex vivo. Also under in vivo conditions, high TNF-a serum concentrations were found in dexamethasone-treated but not control rats when examined 14 hr after an intravenous LPS injection. These data suggestthat the glucocorticoid-induced cachexia of rats may be linked, at least in part, to readily inducible TNF-(Yreleasefrom primed macrophages. 0 1992 Academic
Press, Inc.
INTRODUCTION Tumor necrosis factor-a (TNF-(Y) is a cytokine which is primarily produced by macrophages (l-4). It displays a wide variety of actions, including the destruction of tumor cells, activation of leukocytes, modulation of endothelial cell function, and induction of fibroblast stimulation (5-l 1). In addition, TNF-a is a potent mediator of fever ( 12) and plays a major role in the pathogenesisof endotoxin-induced shock ( 13). TNF-a levels have been shown to be elevated in a variety of infectious diseases. Cerami and co-workers reported that cattle and rabbits with trypanosomiasis displayed a wasting syndrome which was characterized by severeweight loss, hypertriglyceridemia, and significant depletion of fat and protein reserves( 14- 17). It was of particular interest that this cachexia accompanying parasitic infections was closely correlated with incre ed serum levels of the cytokine cachectin which was later found to be identical, ito TNF-a. In an ongoing study to induce a Pneumocystis carinii infection in rats by continuous glucocorticoid-treatment we found a rapid weight loss during the first 2 to 4 weeks and this preceded a P. carinii pneumonitis 8 to 12 weeks later. In this respect, the glucocorticoid treatment of rats resembled trypanosomiasis in cattle and rabbits, and the clinical picture suggestedan involvement of TNF-cu/cachectin. ’ Present address: National Jewish Center for Immunology and Respiratory Medicine, Department of Pediatrics, Denver, CO 80206. 249 0008-8749/92 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form resawd.
250
RENZ
ET AL.
MATERIALS AND METHODS Maintenance of Animals Female Lewis rats between 6 and 8 weeks old (160-200 g) were purchased from Savo (Kisslegg,Germany). The animals were delivered naturally, not barrier-sustained, and received no antibiotics prior to shipping. Rats were housed one per nonfiltered cage. The rodent quarters were maintained at 20-22°C and 50-60% humidity. The rats were divided into three groups. Group A received 0.2 mg dexamethasone (Serva, Heidelberg, Germany) and 50 mg tetracycline (Serva) per 100 ml of acidified, filtered tap water. Group B received tetracycline in acidified water. Group C received acidified drinking water only. Rats drank between 20 and 50 ml of water per day. This amounted to between 0.2 and 0.5 pg dexamethasone/g/day. The rats received standard rat chow (Hoveler, Langenfeld-Immigrath, Germany). Rats were observed daily for signs of infection and were weighed every other day. Preparation of Alveolar Macrophages Before the start of therapy and on Days 3, 7, 14, 21, and 28 of treatment, three animals from each group were sacrificed using intraperitoneal pentobarbital (60 mg/ kg) and exsanguination. Rat alveolar macrophagesmere obtained by gentle lavage of the lungs with phosphate-buffered saline without C$‘/Mg2’. Cells were washed three times with serum-free RPM1 1640 supplemented with L-glutamine (2 mJ@, penicillin ( 100 U/ml), streptomycin ( 100 pg/ml), and mercaptoethanol ( 10 mM). After plating for 1 hr in 24-well culture plates in RPM1 1640 medium plus 10% heat-inactivated (56’C, 30 min) fetal calf serum, the monolayers were washedseveraltimes with medium to remove the nonadhering cells. The remaining monolayer (0.5 X 106/ml) contained >95% macrophagesas determined by phagocytosis of carbon particles and nonspecific esterasestaining (17, 18). The incubation was continued for 20 hr, with or without the addition of Escherichia coli 0127:B8 lipopolysaccharide (LPS, 1 pg/ml). At the end of the incubation, cell-free supernatant was harvested and stored at -20°C until assayedfor TNF activity. In Vivo Injection of LPS Three rats of each of the groups A, B, and C were injected on Day 7 with LPS at a dose of 0.6 mg/kg body wt. After 11 hr, the animals were sacrificed and serum was carefully collected in order to determine the contents of TNF-a, interleukin-la (IL1(Y),and interleukin-6 (IL-6). Determination of TNF-a The amount of TNF-(U in culture supernatants was measured by using the cytotoxicity against TNF-a-sensitive L929 cells as previously described (19,20). Briefly, L929 cells (6 X 104/0.2 ml) were grown in RPM1 1640 to establish a dense monolayer in 96-well microtiter plates. Macrophage culture supernatants or human recombinant TNF-a (kindly provided by BASF/Knoll AG, Ludwigshafen, Germany) or murine recombinant TNF-ar (kindly provided by Dr. G. R. Adolf, Ernst-Boehringer-Institute, Vienna, Austria) were diluted in RPM1 1640 and added to L929 cells in the presence of actinomycin D (1 pug/ml)to enhance L929 sensitivity to TNF. After an incubation for 18 hr, the viability of L929 cells was measured by staining for 1 hr with 3-(4,5-
DEXAMETHASONE
IN VZVO AND TNF-a
RELEASE
251
dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, 20 &well of a 5 mg/ ml stock solution). After completion of the MTT incubation, culture supernatants were removed and cells were lysed by an isopropanol solution containing 40 nut4 HCl. Plates were read in a microplate-autoreader MR 600 (Dynatech, Denkendorf, Germany) at 540 nm. TNF-a activity in macrophage culture supematants is expressed in units per milliliter, based on a standard curve established with either murine or human recombinant TNF-cr. A unit (U) was defined as the TNF-a amount required to lyse 50% of L929 target cells. TNF-(U specificity was corroborated by using both TNF+ insensitive L929 cells and neutralization of TNF-a by a polyclonal rabbit antimurine TNF-(U antiserum (kindly provided by Dr. D. MPnnel, German Cancer Research Center, Heidelberg, Germany). The content of TNF-a in serum from dexamethasone-treated and control rats was determined by a highly specific ELISA for murine TNF-cu. The assay system was established in our laboratory by using a monoclonal hamster anti-murine TNF-(Y antiserum (Genzyme, Munich, Germany), a polyclonal rabbit anti-murine TNF-(U antiserum (kindly provided by Dr. G. R. Adolf, Ernst-Boehringer-Institute, Vienna, Austria), and a peroxidase-labeledgoat anti-rabbit IgG antiserum (Dianova, Hamburg, Germany). Determination of IL-la and IL-6 The concentration of IL- 1crin serum from dexamethasone-treated and control rats was determined by a sensitive and specific radioimmunoassay for the detection of rat IL- 1a (Cytokine Sciences,Inc., Boston, MA). The content of IL6 serum was measured by a sensitive ELISA that was capable of detecting murine as well as rat IL-6 (Endogen, Inc., Boston, MA). Determination of Prostaglandin E2 (PGEj The content of PGE2 in macrophage culture supematants was measured after 20 hr of incubation by radioimmunoassay as previously described (19, 20). Hematological and Chemistry Determinations From all animals, blood was obtained and absolute numbers of leukocytes and differentiation into neutrophil granulocytes and mononuclear cells were determined by standard clinical laboratory methods. The contents of triglycerides, cholesterol, and protein in serum were also determined by established routine methods. Histopathological Evaluation At Day 28 of the study, the lungs of control and dexamethasone-treated animals were aseptically removed and fixed in PBS containing 2% formalin. The lung tissue was embedded in paraffin, sectioned, and stained with hematoxylin and eosin. Tissues were also stained with Giemsa and Gomori methanemine silver to assesslung sections for the presence of P. carinii organisms. RESULTS Rats receiving dexamethasone displayed a weight decreaseof 38% during the initial 2 weeks of treatment, with a particularly sharp decreaseduring the first 3 days (Fig.
252
RENZ ET AL.
Dexamethasone
Days FIG. 1. Loss of body weight of rats treated continuously with dexamethasone. Values show the means + SD of three rats per group and per time point.
1). Control rats, receiving either tetracycline or no medication, showed a regular weight gain during the study. Despite a pronounced weight loss, dexamethasone-treated rats exhibited no apparent signs of apathy, restlessness,respiratory distress, or infections. After 3 days of dexamethasone treatment, the blood mononuclear cells (approximately 80% lymphocytes) dramatically declined to 18% of the control group (Fig. 2) and remained at a very low level during the observation period of 28 days. In contrast, no statistically significant changes in absolute numbers of erythrocytes and thrombocytes were noted between treated and untreated rats, except for neutrophil granulocytes which declined during the initial 3 days of dexamethasone treatment but thereafter returned to normal absolute numbers. When alveolar macrophages were harvested, no difference in cell yield (from 2 to 3 X 106/animal) was found. Upon in vitro culture, the spontaneous release of TNF(Ywas low and did not differ between macrophages from treated and untreated rats. However, in responseto LPS stimulation, macrophages from dexamethasone-treated
-
V’
0
3
7
14
21
28
Days FIG. 2. Decline of blood mononuclear leukocytes during dexamethasone treatment. Values show the means f SD of three rats per time point.
DEXAMETHASONE
IN
253
VZVO AND TNF-a RELEASE
70 60
0
3
7
14
21
20
Days FIG. 3. Enhanced TNF-ol release from alveolar macrophages of rats treated with dexamethasone. Macrophages from dexamethasone (0) or untreated control rats (0) were harvested at indicated time periods and incubated for 20 hr at a concentration of 0.5 X 106/ml in the presenceof LPS (1 &ml), and the culture supematant was assayedfor TNF-(Y activity. Values are the means f SD of macrophages from three rats per time point.
animals displayed a particularly high sensitivity and releasedS- to 1O-fold more TNFa than control macrophages (Fig. 3). The enhanced LPS sensitivity was detected as early as Day 3, remained extremely high for 14 days of dexamethasone medication, and then declined, although it still persisted at a higher level in in vivo dexamethasonetreated macrophages.It was also noted that alveolar macrophagesfrom dexamethasonetreated rats were capable of spontaneously releasing more PGE2 than control macrophages (Fig. 4). LPS was found to be an inefficient stimulus for further PGE2 production in alveolar macrophages. In peritoneal macrophages, an identical pattern of spontaneous PGE2 release was observed and, in addition, those cells were LPS responsive (data not shown). To examine an in vivo production of TNF-a, dexamethasone-treated and control rats received a single intravenous injection of LPS on Day 7. The serum was harvested
Control
Dexometho aone
-
FIG. 4. Enhanced PGEz release from alveolar macrophages of dexamethasone-treated rats. Seven days after dexamethasone or control treatment, macrophages were harvested and incubated for 20 hr at a concentration of 0.5 X 106/ml, and PGEz content was determined in the culture supematant. Values represent the means + SD of macrophages from three rats per group.
254
RENZ ET AL. A
0.5 1
0 m
without LPS with LPS
Control
Dexomethosone
0 without LPS EZl with LPS
200
0 without LPS 6SQ with LPS
60
” t
x
z 60 g ‘p 40 d
150
-p 100
d
20
50
Control
Control
Dexomathosone
Dexomethosone
FIG. 5. Elevated TNF-ol and reduced IL- 101and IL-6 in vivo releasein dexamethasone-treated rats. Seven days after dexamethasone or control treatment, LPS (0.6 mg/kg) was intravenously injected and after 14 hr, the contents of TNF-(U (A), IL- 101(B), and IL-6 (C) in the serum were determined. Values represent the means +- SD of three rats per group.
11 hr later and was assayedfor TNF-a content by a highly specific ELISA (Fig. 5A). Although dexamethasone-treated animals did not spontaneously release TNF-a in circulation, an LPS injection induced a high TNF-(U production, whereas control rats remained rather unresponsive. In striking contrast, the LPS-inducible release of ILla and IL-6 into circulation was markedly reduced in dexamethasone-treated rats (Figs. 5B and 5C). Already after 3 days of treatment, the dexamethasone group showed a marked increase in serum triglyceride levels, approximately six times higher than levels in the TABLE 1 Effect of Dexamethasone on SelectedSerum Components’ Treatment
Triglycerides (mg/dl)
Cholesterol bw/dl)
Protein WW
Albumin Wdl)
Control Dexamethasone (Day 3)
128 +- 24 739 f 12
59.5 + 5 102.0 +- 15
6.0 + 0.2 10.3 + 0.6
3.9 * 0.1 7.0 + 0.3
r?All values are the means f SD of three rats and the tests were performed in triplicate.
DEXAMETHASONE
IN VZVO AND TNF-a RELEASE
255
control group (Table 1). Serum triglycerides remained high during the entire observation period and closely paralleled LPS-inducible TNF-(I! production (Fig. 3). Rather similarly, although not as pronounced, serum concentrations of cholesterol, total protein, and albumin showed significant increases at Day 3 of dexamethasone treatment (Table 1) and remained elevated up to Day 28. In the lung tissue of dexamethasone-treated rats, a marked giant mononuclear cell infiltration was noted (Fig. 6). Despite this cell infiltration, the bronchoalveolar lavage produced no higher cell yield (seeabove). No granulocytes were seenin this infiltrate nor were any P. carinii organisms, fungi, or bacteria observed. DISCUSSION This study provides evidence that TNF-a may play a major role in the marked weight loss which is induced in rats by continuous treatment with dexamethasone. The weight loss occurred rapidly during the first 3 days of dexamethasone therapy, persisted up to 28 days, and was closely associated with a hypertriglyceridemia and enhanced serum cholesterol, albumin, and protein levels. These clinical observations were reminiscent of the cachexia caused by cachectin release in parasitic infections (14-16, 2 1) and the anorexia and severe weight loss induced in rats by injection of TNF-a! ( 13). As a primary source of TNF release,macrophagesand, to a lesserdegree, lymphocytes must be considered. Circulating monocytes or lymphocytes could not be examined in our study since the dexamethasone treatment led to a marked decline of mononuclear cell counts to 5% of the control group (Fig. 2). Therefore, we examined alveolar macrophages which were easily accessibleand were found to be major TNFLYproducers. Alveolar macrophagesfrom dexamethasone-and untreated rats were both inefficient in releasing spontaneously significant amounts of TNF-(Y in the absenceof exogenous stimuli. However, macrophages from dexamethasone-treated rats were clearly preconditioned or “primed” for copious TNF-a! releasewhen they were in vitro exposed to LPS (Fig. 3). Furthermore, not only in vitro but also in vim, an LPS injection led to a marked release of TNF-(U into circulation of dexamethasone-treated rats (Fig.
FIG. 6. Histopathology of cross sections of lungs from control (A, 30X) and dexamethasone-treated (B, 30X; C, 80X) rats. Sections were done 28 days after dexamethasone treatment.
256
RENZ ET AL.
5A). These striking findings represent an entirely novel and, in fact, unexpected observation since glucocorticoids are not regarded as priming factors for TNF-ar release such as interferon-y (22) or GM-CSF (23). On the contrary, glucocorticoids have been shown to reduce cytokine releasewhich has been reconfirmed in this study by parallel measurements of IL-la (Fig. 5B) and IL-6 (Fig. 5C) in serum of dexamethasonetreated rats. Thus, on first sight, our TNF-a data may stand in contradiction to hitherto established models of glucocorticoid actions. Previous in vitro studies have shown that glucocorticoids inhibit TNF-a at both the transcriptional and the translational level by blocking gene transcription and mRNA mobilization (24). Also, when dexamethasone was intraperitoneally given to rats 2 to 8 hr prior to intravenous LPS injection, a 70 to 90% reduction in the serum peak concentration of TNF-a! was observed (25). Similarly, in vitro studies with human monocytes revealed an 80% reduction of TNFa! releasewhen monocytes were preincubated for 48 hr with dexamethasone (26). In contrast to those studies, our results were obtained with rats that had been treated for a prolonged time period (up to 28 days) with dexamethasone, and the TNF-a releasing capacity was examined both ex vivo in alveolar macrophages and in vivo in intact animals. It was of additional interest that alveolar macrophages from dexamethasonetreated animals displayed an enhanced spontaneous PGE2production when compared to control macrophages(Fig. 4). Thus, analogousto TNF-a production, the arachidonic acid metabolism was also shifted to an augmented prostanoid production. It remains to be elucidated whether a common underlying mechanism preactivated both mediator systems or whether they mutually affected each other’s functional role as previously demonstrated (19). How prolonged dexamethasone treatment of rats may selectively enhance LPSinducible release of TNF-a, but not of IL-la or IL-6, from alveolar macrophages remains unclear. A histologic examination of glucocorticoid-treated rats revealed an accumulation of macrophage-like giant cells in the alveolar spacesbut the lung sections showed no neutrophil granulocyte infiltration which argues against a typical bacterial infection. The concomitant administration of tetracycline to dexamethasone-treated rats appears to have protected against bacterial infections. Additionally, a specific staining showed no indication of infection by P. carinii or fungi. However, an infection by viruses cannot be ruled out and has to be taken into consideration, although all clinical signs and preliminary virological tests do not support this possibility. Apart from clinically inapparent infections, it is also feasible that glucocorticoid treatment of rats may have induced a shift of monocyte/macrophage migration in such a manner that TNF-LUproducing subpopulations may have preferentially emigrated to the lung (27, 28). The low numbers of mononuclear cells in circulation may indicate not only that glucocorticoid-sensitive lymphocytes have disappeared, but also that a redistribution of monocytes/macrophages may have taken place. A priming for enhanced TNF-a! releaseindicates that ample TNF-a mRNA should be available for ready translation into the bioactive protein following stimuli such as LPS or others. Work is currently in progress to examine TNF-a and other cytokine gene expression in alveolar macrophages following in vivo glucocorticoid treatment. An enhanced TNF-a mRNA accumulation may be the sequelaeof an enhanced gene expression, possibly mediated by glucocorticoid interference with resynthesis of repressor proteins (29), a prolonged stabilization of mRNA by inhibition of normal degradation mechanisms, or a glucocorticoid-induced production of unknown mediators which in turn causes enhanced TNF-a gene transcription. It may even be speculated that priming of alveolar macrophages for enhanced TNF-(Y release in re-
DEXAMETHASONE
IN VIVO AND TNF-(Y RELEASE
257
sponse to microbial products such as LPS may represent a protective mechanism in the lung against infections that could easily occur during glucocorticoid-induced immunosuppression. ACKNOWLEDGMENTS We thank Amy Beermann for competent secretarial help. We are grateful to Dr. G. R. Adolf, ErnstBoehringer-Institute, Vienna, Austria, for donating murine recombinant TNF-cx and anti-TNF-cYantiserum, and to BASF/Knoll AG, Ludwigshafen, Germany, for providing us with human recombinant TNF-(Y. This study was supported by the Volkswagen-Stiftung (I/65-535).
REFERENCES 1. Old, L. J., Science 230, 630, 1985. 2. Beutler, B., and Cerami, A., Nature 320, 584, 1986. 3. Urban, J. L., Shepard, H. M., Rothstein, J. L., Sugarman, B. J., and Schreiber, H., Proc. Natl. Acad. Ski. USA 83, 5233, 1986. 4. Kornbluth, R. S., and Edgington, T. S., J. Immunol. 137,2585, 1986. 5. Carswell, E. A., Old, L. J., Kassel, R. L., Green, S., Fiore, N., and Williamson, B., Proc. Natl. Acad. Sci. USA 72, 3666, 1975. 6. Esparza, I., Mannel, D., Ruppel, A., Falk, W., and Krammer, P. H., J. Exp. Med. 166, 589, 1987. 7. Heidenreich, S., Weyers, M., Gong, J.-H., Sprenger, H., Nain, M., and Gemsa, D., J. Immunol. 140, 1511, 1988. 8. Philip, R., and Epstein, L. B., Nature 323, 86, 1986. 9. Mantovani, A., and Dejana, E., Zmmunol. Today 10, 370, 1989. 10. Sugarman, B. J., Aggarwal, B. B., Haas, P. E., Figari, I. S., Palladino, M. A., and Shepard, H. M., Science 230,943, 1985. 11. Sherry, B., and Cerami, A., J. Cell Biol. 107, 1269, 1988. 12. Dinarello, 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., J. Exp. Med. 163, 1433, 1986. 13. Tracey, K. J., Beutler, B., Lowry, S. F., Merryweather, J., Wolpe, S., Milsark, I. W., Hariri, R. J., Fahey, T. J., III, Zentella, A., Albert, J. D., Shires, G. T., and Cerami, A., Science 234, 470, 1986. 14. Rouzer, C. A., and Cerami, A., Mol. Biochem. Parasitol. 2, 3 1, 1980. 15. Kawakami, M., and Cerami, A., J. Exp. Med. 154,631, 198 I. 16. Kawakami, M., Pekala, P. H., Lane, M. D., and Cerami, A., Proc. Natl. Acad. Sci. USA 79,9 12, 1982. 17. Gemsa, D., Woo, C. H., Fudenberg, H. H., and Schmid, R., J. C/in. Invest. 52, 8 12, 1973. 18. Lovett, D., Kozan, B., Hadam, M., Resch, K., and Gemsa, D., J. Immunol. 136, 340, 1986. 19. Renz, H., Gong, J.-H., Schmidt, A., Nain, M., and Gemsa, D., J. Immunol. 141, 2388, 1988. 20. Rem, H., Gentz, U., Schmidt, A., Dapper, T., Nain, M., and Gemsa, D., Infect. Immunol. 57, 3172, 1989. 21. Hotez, P. J., Le Trang, N., Fairlamb, A. H., and Cerami, A., Parasite Immunol. 6, 203, 1984. 22. Pace, J. L., Russel, S. W., Torres, B. A., Johnson, H. M., and Gray, P. W., J. Immunol. 130, 2011, 1983.
23. 24. 25. 26. 27.
Heidenreich, S., Gong, J.-H., Schmidt, A., Nain, M., and Gemsa, D., J. Immunol. 143, 1198, 1989. Beutler, B., Krochin, N., Milsark, I. W., Luedecke, C., and Cerami, A., Science 232, 977, 1986. Waage, A., Clin. Immunol. Immunopathol. 45, 348, 1987. Waage, A., and Bakke, O., Immunology 63,299, 1988. Rinehart, J. J., Sasone,A. L., Balcerzak, S. P., Ackerman, G. A., and LoBuglio, A. F., N. Engl. J. Med. 292, 236, 1975. 28. Boggs, D. R., Arthens, J. W., and Cartwright, C. E., Am. J. Pathol. 44,763, 1964. 29. Collart, M. A., Belin, D., Vassalli, J.-D., de Kossodo, S., and Vassalli, P., J. Exp. Med. 164, 2 I 13, 1986.
