CELLULAR
IMMUNOLOGY
14,
5 13-5 19 (1992)
Endotoxin Tolerance: In Vivo Regulation of Tumor Necrosis Factor and Interleukin-1 Synthesis Is at the Transcriptional Level STEVEN H. ZUCKERMAN’
AND GLENN F. EVANS
Lilly Research Laboratories, Indianapolis, Indiana 46285 Received August 12, 1991; acceptedDecember 17, 1991 Endotoxin tolerance is associatedwith a decreasedproduction of many inflammatory mediators following stimulus-induced desensitization. The mechanisms involved in the regulation of tumor necrosis factor (TNF) and interleukin-1 (IL-l) j3 expression were investigated in peritoneal macrophages from endotoxin tolerant mice. The absence of a serum TNF peak in tolerant mice correlated with a significant reduction in both TNF and IL-1 ,f3mRNA accumulation in macrophages from these animals. The reduction in both IL- 1 and TNF mRNA was consistent with a decreasein transcriptional activity by nuclear run-on assaysfor both cytokines and a reduced amount of the nuclear-associated transcription factor NF-KB. Therefore, the hyporesponsive state associated with endotoxin tolerance is characterized by transcriptional regulation of both TNF and IL- 1 synthesis. o 1992 Academic press IIIC.
INTRODUCTION The host response to lipopolysaccharide (LPS) is characterized by the induction of inflammatory mediators including cytokines, arachidonic acid metabolites, and reactive oxygen intermediates which contribute to the pathology associated with endotoxin shock. The ability of the mammalian host to desensitize to a repetitive endotoxin stimulus has been reported and is characterized by a reduction or absence of these inflammatory mediators in vitro or in vivo, a diminished febrile response,and increased resistance relative to naive animals to a lethal LPS challenge ( l-7). Whether a second LPS stimulus results in a diminished inflammatory response as evident during endotoxin tolerance or an augmented responseas in the Shwartzman phenomena appears dependent on the concentration and route of LPS administration and may be modulated by endogenous factors including glucocorticoids in the former and interferony in the latter (8, 9). While endotoxin tolerant rodents and rabbits have reduced levels of cytokines including tumor necrosis factor (TNF), interferon, and colony-stimulating factor, the mechanisms involved in cytokine regulation are not well understood. In vitro studies have focused on an endotoxin refractory state using macrophage-like cell lines or primary peritoneal macrophages and have suggestedboth transcriptional ( 10, 11) and post-translational ( 12) mechanisms involved in the regulation of TNF. In the present study, the regulation of TNF expression has been investigated in an in vivo model of ’ To whom correspondence should be addressed. 513 0008-8749192$3.00 Copyright 0 1992 by Academic F’res, Inc. All rights of reproduction in any form reserved.
514
SHORT
COMMUNICATIONS
endotoxin tolerance. We report that peritoneal macrophages from endotoxin tolerant mice have reduced levels of both TNF and IGl @mRNA and that regulation was at the transcriptional level by both nuclear run-on assaysand by quantitation of nuclearassociatedNF-KB. METHODS AND MATERIALS In Vivo Endotoxin Model Thioglycollate-primed female BALB/c mice were injected intravenously with 10 pg of LPS (Escherichia coli 055:B5; Difco Lab., Detroit, MI) and were sacrificed at 1 or 20 hr post-LPS or again injected at 20 hr with 10 pg of LPS and sacrificed 1 hr later. These time points have been chosen to quantitate serum TNF or mRNA based on the kinetics of serum TNF and peritoneal macrophage TNF and IL-1 @mRNA accumulation in both LPS sensitive and tolerant mice (13). Both sera and peritoneal macrophages were harvested at each time point. Macrophages from three to six animals per group were pooled and processed for subsequent assays.Serum TNF was quantitated by cytotoxicity of actinomycin D-treated L929 fibroblasts as described previously ( 14). RNA Analysis Peritoneal macrophages were lysed in 5 M guanidine thiocyanate, centrifuged through 5.7 A4 CsCl, and RNA pellets were dissolved in water and 5 pg applied on nitrocellulose filter slot blots (Schleicher and Schuell, Keene, NH). Oligonucleotide probes specific for murine TNF (Y,IL-1 /I (Amgen, Thousand Oaks, CA), and rat ,f3 actin (provided by Nancy Mayne, Lilly Research Labs) were end labeled with 32Py ATP (New England Nuclear, Boston, MA) at the 5’ position with an end labeling kit (Boehringer Mannheim, Indianapolis, IN). Nitrocellulose filters were hybridized to the oligonucleotide probes at 42’C in 50% formamide for 20 hr and then washed and autoradiographed (Kodak Corp., Rochester, NY). The extent of oligonucleotide hybridization was quantitated by the Betagen Model 603 blot analyzer (Waltham, MA). Nuclear Run-on Transcription Assays Macrophages were lysed in 0.5% NP-40 in 10 mMTris, pH 7.4, 10 mMNaC1, and 3 mM MgC12and nuclei were isolated and stored at -70°C. Nuclei were labeled with 32PUTP ( 15) and the resulting RNA was hybridized at 65 “C with nitrocellulose slot blots containing 7.5 pg of alkali denatured plasmids for 48 hr. Nitrocellulose blots were washed, analyzed by Betagen, and autoradiographed. Plasmids used included ~55 1 containing the SmaI to EcoRI fragment of the murine TNF gene inserted into pGEM 1 (16) a mouse (Y actin cDNA plasmid in which the 3’ Pstl fragment was inserted into pSP64 ( 17), and pMIL-/3 containing a 1400-bp cDNA fragment encompassing the murine IL- 1 @gene (18). Nuclear Extract and Gel Shift Assays Macrophages were lysed in 0.5% NP-40 in 10 mM Tris-HCl, pH 7.4, and the resulting nuclei were extracted in a Hepes-buffered glycerol hypertonic salt solution containing protease inhibitors as described (19). Gel shift assayswere performed by
515
SHORT COMMUNICATIONS
incubating 5 pg of nuclear extract protein with 1 ng, 0.2 &i, of a 5’ end-labeled oligonucleotide containing the NF-KB recognition sequenceGGGGGCTTTCC from the murine TNF (Ypromoter (20) which had previously been converted to a doublestranded molecule by hybridizing with the complementary oligonucleotide. Following a 15-min incubation at 30°C in the presence of 25 pg poly[d(I-C)], samples were applied to an 8% native polyacrylamide gel and electrophoresed at 250 V for 2.5 hr at 4°C in 50 mM Tris, 100 rnM glycine, pH 8.5, running buffer (21). Gels were subsequently dried onto nitrocellulose filters and bands resulting from gel retardation were quantitated by the Betagen blot analyzer prior to autoradiography. RESULTS Injection of LPS into mice results in a rapid rise in serum TNF which peaks 1 hr post-LPS. This increase in serum TNF correlated with an increase in TNF mRNA in peritoneal macrophages (Fig. 1A). TNF serum levels return to control levels 3-5 hr later (14) and by 20 hr, there was no detectable serum TNF and no TNF mRNA in peritoneal macrophages. These animals have become endotoxin tolerant as a second injection of LPS at 20 hr did not result in any increase in serum TNF or TNF mRNA in peritoneal macrophages. The absenceof a secondary increase in TNF mRNA was not unique to this cytokine as a similar pattern was observed for IL-1 p while actin mRNA levels remained relatively constant at all time points evaluated (Fig. 1B). A profile similar to IL-l fl was also observed for IL-l CYwith peak mRNA levels 1 hr after LPS and no increasesdetected 1 hr after the secondLPS stimulus (data not shown).
A
c 1000
30' WWW f
20-
_ 800
0 Counts _
p
800 400
% loI-
9 s0
200 0 0
1
20
21
Hr
FIG. 1. Peritoneal macrophage TNF mRNA and serum TNF levels during endotoxin tolerance. Thioglycollate-primed mice were injected with 10 gg of LPS iv and at designated intervals animals were bled and peritoneal macrophageswere harvested and processedfor RNA. Twenty hours after the initial LPS stimulus animals received a second LPS injection and were sacrificed 1 hr later. Serum TNF levels and peritoneal macrophage TNF mRNA were quantitated (A) and the actual autoradiogram for the mRNA analysis for TNF, actin, and IL- 1 fl is presented (B). Representative experiment of four.
516
SHORT COMMUNICATIONS
The correlation between serum TNF and mRNA levels in peritoneal macrophages from endotoxin tolerant mice suggesteda pretranslational regulation of TNF expression. Nuclear run-on transcriptional assayswere then performed to more precisely determine the mechanism involved in cytokine regulation during endotoxin tolerance. As demonstrated in Fig. 2, while macrophage nuclei isolated 30 min after the LPS injection had increased transcriptional activity at both the TNF and IL- 1 ,8loci, nuclei isolated 30 min following the second LPS injection had significantly reduced transcriptional activity at these loci. Constitutive transcription at the (Y actin locus in distinction was observed at all points evaluated. Similar results were observed if nuclei were isolated at 40 or 60 min post-primary or secondary LPS injection, although the overall transcriptional activity at both the TNF and IL- 1 p loci were decreasedrelative to the 30 min time point (data not shown). Theseresults suggestedthat in vivo endotoxin tolerance correlated with the transcriptional regulation at both the TNF and IL- 1 loci. The role for NF-RR in the transcriptional regulation of TNF expression during endotoxin tolerance was then evaluated by DNA gel shift mobility experiments. As demonstrated in Fig. 3, nuclei from macrophages 1 hr post-LPS injection had a fivefold increase (lane 3) over macrophages from nonstimulated mice (lane 2) in the intensity of a gel shift band using a 32Poligonucleotide sequence recognized by NF-KB. Macrophages from endotoxin tolerant mice in contrast failed to demonstrate a significant increase in nuclear-associated NFXR-like activity following a second LPS stimulus (lanes 4 and 5). The inability of the macrophages in endotoxin tolerant mice to translocate NF-RI3 into the nucleus following a repetitive LPS stimulus is consistent with the regulation of TNF and IL-l expression occurring at the transcriptional level. A
1200 1000 4 800 g 600 0 400 200 n 0
0.5
20
20.5
Hr
6
TNF IL-1
.,
Actin pGEM1 0
0.5
20
20.5
Hr FIG. 2. Nuclear run-on transcriptional assay during endotoxin tolerance. Mice were injected iv with 10 r.rgof LPS and peritoneal macrophages were isolated 30 min later and the resulting nuclei were allowed to incorporate ‘*P UTP in vitro. At 20 hr mice were sacrificed or again injected with LPS and macrophages were harvested 30 min later. RNA was isolated from each nuclear preparation and lo6 cpm were hybridized to cDNA plasmids for TNF, IG 10, actin, and the TNF parental vector pGEM 1.Hybridization was quantitated by Betagen (A) and autoradiographed (B). Representative experiment of three.
SHORT COMMUNICATIONS
517
12345
FIG. 3. Absenceof NF-KB induction during endotoxin tolerance. Macrophageswere lysed at the designated intervals post-LPS and nuclear extracts were incubated with an oligonucleotide from the TNF promoter containing a NF-KB recognition sequence.Retardation of the electrophoretic migration of this oligonucleotide was evaluated on a native 8% polyacrylamide gel and the band intensity was quantitated by Betagen. Lanes l-5 represent no nuclear extract (lane l), and extracts from macrophages 0, 1, and 20 hr (lanes 2-4) post LPS and 1 hr after the second LPS stimulus (lane 5). Hybridization counts for the gel shift band (arrow) were 17,012, 89,199, 14,964, and 15,775 for lanes 2-5, respectively. Representative experiment of three.
DISCUSSION Endotoxin tolerance is characterized by a diminished host inflammatory response following a repetitive LPS challenge. Whereas a decreasein cytokine levels has been demonstrated in the sera of endotoxin tolerant animals and a reduced production of these mediators is also observed in LPS refractory macrophage cultures, the mechanisms involved in the establishment of tolerance remain poorly understood. Animal models characterized by an enhanced sensitivity to the lethal effectsof endotoxin such as adrenalectomized (14) or galactosamine-sensitized mice (22) do not become endotoxin tolerant (23) using protocols which induce tolerance in normal animals. Furthermore, tolerance does not result in the reduction in the levels of all serum cytokines as LPS has been demonstrated to increase serum levels of both IL- 1 (13) and IL-6 (9) in endotoxin tolerant mice. Previous studies have focusedon the level at which cytokine expressionwas regulated in in vitro cultures of LPS refractory macrophagesor macrophage-like cell lines. These studies have reported transcriptional regulation in murine RAW 264 cells (10) and the human monocyte-macrophage line Mono-Mac-6 (1 l), post-transcriptional regulation related to mRNA stability for IL- 1 fi in THP- 1 cells (24) and post-translational regulation in primary cultures of murine peritoneal macrophages ( 12). Differences in the cell systemsused and the nature of the protocols employed to induce an endotoxin refractory state may explain the absence of a consensus concerning the mechanisms involved in LPS desensitization. In the present study, the absenceof a secondary peak in serum TNF was due to transcriptional regulation of cytokine gene expression as evidenced by decreasednuclear run-on activity and a reduction in nuclear-associated
518
SHORT COMMUNICATIONS
NF-KB. These observations clearly demonstrate that transcriptional regulation of macrophage cytokine gene expression is fundamentally involved in the process of endotoxin tolerance. Whether cytokine expression during endotoxin tolerance in other animal models is regulated at the transcriptional level remains to be determined. The difference in transcriptional regulation in this study and post-translational regulation of TNF expression in our previous in vitro study on LPS refractory macrophage cultures demonstrates multiple levels of regulating TNF synthesis during LPS desensitization. Whether the regulation observed at the transcriptional level in endotoxin tolerant animals was mediated by endogenous glucocorticoids or other factors which are absent in in vitro cultures remains to be determined. The role of NF-KB has been considered in the human Mono-Mac-6 cell line in in vitro endotoxin refractory studies. While TNF mRNA was not apparent in response to a second LPS stimulus, these cells did demonstrate a significant increase in NF-KB activity (11). In the present study, the absence of a significant increase in NF-KB after the second LPS stimulus was consistent with reduced run-on transcriptional activity for TNF and IL-l ,& The differences in NF-KB observed between the present study and the previous study (11) may reflect fundamental differences in the mechanisms involved in the regulation of endotoxin tolerance between the in vitro studies with the human Mono-Mac-6 cell line and the in vivo studies where other regulatory factors, including endogenous glucocorticoids, are involved in the regulation of cytokine gene expression. These studies suggestthat the mechanisms by which endotoxin tolerance results in transcriptional regulation of cytokine gene expression are diverse and may be dependent on species, the state of macrophage differentiation, as well as the involvement of other soluble factors, the expression of which may be dependent upon the appropriate in vitro and in vivo environments. Soluble TNF receptors, for example, may represent one such factor which could explain the absenceof a second peak of biologically active serum TNF in endotoxin tolerant mice (25). However, attempts to demonstrate the presence of soluble receptors or other blocking factors by performing TNF dose-response cytotoxicity assayswith sera from control and endotoxin tolerant mice have not been successful(unpublished observations). In addition to transcriptional regulation of cytokine genes in macrophages from endotoxin tolerant mice, other metabolic changes associated with a LPS refractory statehave been describedin macrophages.Introna et al. (26), for example, have reported decreasedtranscriptional activity by nuclear run-on assaysfor c-fos in response to LPS in endotoxin refractory murine peritoneal macrophages.The demonstration that secretion of arachidonic acid metabolites is reduced in macrophages from endotoxin tolerant rats (27) and that these macrophageshave a diminished thromboxane response to guanine nucleotide analogues (28), as well as the inability in the present study to detect a LPS-mediated increase in NF-KB in macrophages from endotoxin tolerant mice, suggeststhat endotoxin tolerance may be associated with the uncoupling of signal transduction mechanisms associated with endotoxin activation. ACKNOWLEDGMENTS The authors acknowledgeLaura Guthrie for her superb technical assistance,Drs. Bill Roeder (Lilly Research Labs) and David Chapman (Washington University of St. Louis) for the TNF and IL-1 j3plasmids,respectively, and Dr. Chandrasekhar for his review of our manuscript.
SHORT COMMUNICATIONS
519
REFERENCES 1. Greisman, S. E., In “Beneficial Effects of Endotoxins” (A. Nowotny, Ed.), p. 149. Plenum, New York, 1983. 2. Heyman, A., Vener. Dis. Znf 26,5 1, 1945. 3. Beeson, P. B., J. Exp. Med. 86, 39, 1947. 4. Watson, D., and Kim, Y. B., J. Exp. Med. 118, 425, 1963. 5. Quesenberry, P., Halperin, J., Ryan, M., and Stohlman, F., Blood 45, 789, 1975. 6. Henricson, B. E., Benjamin, W. R., and Vogel, S. N., Infect. Immun. 58, 2429, 1990. 7. Williams, Z., Hertogs, C. F., and Pluznik, D. H., Infect. Immun. 41, 1, 1983. 8. Zuckerman, S. H., and Bendele, A. M., In@. Immun. 57, 3009, 1989. 9. Heremans, H., Van Damme, J. V., Dillen, C., Dikmans, R., and Billiau, A., J. Exp. Med. 171, 1853, 1990. 10. Virca, G. D., Kim, S. Y., Glaser, K. B., and Ulevitch, R. J., J. Biol. Chem. 264, 21,951, 1989. 11. Haas, J. G., Baeuerle, P. A., Riethmuller, G., and Ziegler-Heitbrock, H. W. L., Proc. Natl. Acad. Sci. USA 81,9563, 1990. 12. Zuckerman, S. H., Evans, G. F., Snyder, Y. M., and Roeder, W. D., J. Immunol. 143, 1223, 1989. 13. Zuckerman, S. H., Evans, G. F., and Butler, L. D., Infect. Immun. 59, 2774, 1991. 14. Zuckerman, S. H., Shellhaas, J., and Butler, L. D., Eur. J. Immunol. 19, 301, 1989. 15. Greenberg, M. E., and Ziff, E. B., Nature 311,433, 1984. 16. Gardner, S. M., Mock, B. A., Hilgers, J., Huppi, K. E., and Roeder, W. D., J. Immunol. 139,476, 1987. 17. Minty, A. J., Caravatti, M., Robert, B., Coben, A., Daubas, P., Weydert, A., Gros, F., and Buckingham, M. E., J. Biol. Chem. 256, 1008, 1981. 18. Gray, P. W., Glaister, D., Chen, E., Goeddel, D. V., and Pennica, D., J. Immunol. 137, 3644, 1986. 19. Dignam, J. D., Lebovitz, R. M., and Roeder, R. G., Nucleic Acids Res. 11, 1475, 1983. 20. Shakhov, A. N., Collart, M. A., Vassalli, P., Nedospasov, S. A., and Jongeneel, C. V., J. Exp. Med. 171, 35, 1990. 2 1. Chodosh, L. A., In “Current Protocols in Molecular Biology” (F. M. Ausubel, R. Brent, R. E. Kingston,
22. 23. 24.
25. 26.
D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl, Eds.), Vol. 2, p. 12.2.1.Greene Publishing Associates,New York, 1988. Galanos, C., Freudenberg, M. A., and Reutter, W., Proc. Natl. Acad. Sci. USA 76, 5939, 1979. Evans, G. F., and Zuckerman, S. H., Eur. J. Immunol. 21, 1973, 1991. Fenton, M. J., Vermeulen, M. W., Clark, B. D., Webb, A. C., and Auron, P. E., J. Immunol. 140,2261, 1988. Nophar, Y., Kemper, O., Brakebusch, C., Engelmann, H., Zwang, R., Aderka, D., Holtmann, H., and Wallach, D., EMBO J. 9, 3269, 1990. Introna, M., Bast, R. C., Johnston, P. A., Adams, D. O., and Hamilton, T. A., J. Cell. Physiol. 131, 36,
1987. 27. Rogers, T. S., Halushka, P. V., Wise, W. C., and Cook, J. A., Prostaglandins 31, 639, 1986. 28. Coffee, K. A., Halushka, P. V., Wise, W. C., and Cook, J. A., Biochim. Biophys. Acta 1035,201, 1990.
IMMUNOLOGY
14,
5 13-5 19 (1992)
Endotoxin Tolerance: In Vivo Regulation of Tumor Necrosis Factor and Interleukin-1 Synthesis Is at the Transcriptional Level STEVEN H. ZUCKERMAN’
AND GLENN F. EVANS
Lilly Research Laboratories, Indianapolis, Indiana 46285 Received August 12, 1991; acceptedDecember 17, 1991 Endotoxin tolerance is associatedwith a decreasedproduction of many inflammatory mediators following stimulus-induced desensitization. The mechanisms involved in the regulation of tumor necrosis factor (TNF) and interleukin-1 (IL-l) j3 expression were investigated in peritoneal macrophages from endotoxin tolerant mice. The absence of a serum TNF peak in tolerant mice correlated with a significant reduction in both TNF and IL-1 ,f3mRNA accumulation in macrophages from these animals. The reduction in both IL- 1 and TNF mRNA was consistent with a decreasein transcriptional activity by nuclear run-on assaysfor both cytokines and a reduced amount of the nuclear-associated transcription factor NF-KB. Therefore, the hyporesponsive state associated with endotoxin tolerance is characterized by transcriptional regulation of both TNF and IL- 1 synthesis. o 1992 Academic press IIIC.
INTRODUCTION The host response to lipopolysaccharide (LPS) is characterized by the induction of inflammatory mediators including cytokines, arachidonic acid metabolites, and reactive oxygen intermediates which contribute to the pathology associated with endotoxin shock. The ability of the mammalian host to desensitize to a repetitive endotoxin stimulus has been reported and is characterized by a reduction or absence of these inflammatory mediators in vitro or in vivo, a diminished febrile response,and increased resistance relative to naive animals to a lethal LPS challenge ( l-7). Whether a second LPS stimulus results in a diminished inflammatory response as evident during endotoxin tolerance or an augmented responseas in the Shwartzman phenomena appears dependent on the concentration and route of LPS administration and may be modulated by endogenous factors including glucocorticoids in the former and interferony in the latter (8, 9). While endotoxin tolerant rodents and rabbits have reduced levels of cytokines including tumor necrosis factor (TNF), interferon, and colony-stimulating factor, the mechanisms involved in cytokine regulation are not well understood. In vitro studies have focused on an endotoxin refractory state using macrophage-like cell lines or primary peritoneal macrophages and have suggestedboth transcriptional ( 10, 11) and post-translational ( 12) mechanisms involved in the regulation of TNF. In the present study, the regulation of TNF expression has been investigated in an in vivo model of ’ To whom correspondence should be addressed. 513 0008-8749192$3.00 Copyright 0 1992 by Academic F’res, Inc. All rights of reproduction in any form reserved.
514
SHORT
COMMUNICATIONS
endotoxin tolerance. We report that peritoneal macrophages from endotoxin tolerant mice have reduced levels of both TNF and IGl @mRNA and that regulation was at the transcriptional level by both nuclear run-on assaysand by quantitation of nuclearassociatedNF-KB. METHODS AND MATERIALS In Vivo Endotoxin Model Thioglycollate-primed female BALB/c mice were injected intravenously with 10 pg of LPS (Escherichia coli 055:B5; Difco Lab., Detroit, MI) and were sacrificed at 1 or 20 hr post-LPS or again injected at 20 hr with 10 pg of LPS and sacrificed 1 hr later. These time points have been chosen to quantitate serum TNF or mRNA based on the kinetics of serum TNF and peritoneal macrophage TNF and IL-1 @mRNA accumulation in both LPS sensitive and tolerant mice (13). Both sera and peritoneal macrophages were harvested at each time point. Macrophages from three to six animals per group were pooled and processed for subsequent assays.Serum TNF was quantitated by cytotoxicity of actinomycin D-treated L929 fibroblasts as described previously ( 14). RNA Analysis Peritoneal macrophages were lysed in 5 M guanidine thiocyanate, centrifuged through 5.7 A4 CsCl, and RNA pellets were dissolved in water and 5 pg applied on nitrocellulose filter slot blots (Schleicher and Schuell, Keene, NH). Oligonucleotide probes specific for murine TNF (Y,IL-1 /I (Amgen, Thousand Oaks, CA), and rat ,f3 actin (provided by Nancy Mayne, Lilly Research Labs) were end labeled with 32Py ATP (New England Nuclear, Boston, MA) at the 5’ position with an end labeling kit (Boehringer Mannheim, Indianapolis, IN). Nitrocellulose filters were hybridized to the oligonucleotide probes at 42’C in 50% formamide for 20 hr and then washed and autoradiographed (Kodak Corp., Rochester, NY). The extent of oligonucleotide hybridization was quantitated by the Betagen Model 603 blot analyzer (Waltham, MA). Nuclear Run-on Transcription Assays Macrophages were lysed in 0.5% NP-40 in 10 mMTris, pH 7.4, 10 mMNaC1, and 3 mM MgC12and nuclei were isolated and stored at -70°C. Nuclei were labeled with 32PUTP ( 15) and the resulting RNA was hybridized at 65 “C with nitrocellulose slot blots containing 7.5 pg of alkali denatured plasmids for 48 hr. Nitrocellulose blots were washed, analyzed by Betagen, and autoradiographed. Plasmids used included ~55 1 containing the SmaI to EcoRI fragment of the murine TNF gene inserted into pGEM 1 (16) a mouse (Y actin cDNA plasmid in which the 3’ Pstl fragment was inserted into pSP64 ( 17), and pMIL-/3 containing a 1400-bp cDNA fragment encompassing the murine IL- 1 @gene (18). Nuclear Extract and Gel Shift Assays Macrophages were lysed in 0.5% NP-40 in 10 mM Tris-HCl, pH 7.4, and the resulting nuclei were extracted in a Hepes-buffered glycerol hypertonic salt solution containing protease inhibitors as described (19). Gel shift assayswere performed by
515
SHORT COMMUNICATIONS
incubating 5 pg of nuclear extract protein with 1 ng, 0.2 &i, of a 5’ end-labeled oligonucleotide containing the NF-KB recognition sequenceGGGGGCTTTCC from the murine TNF (Ypromoter (20) which had previously been converted to a doublestranded molecule by hybridizing with the complementary oligonucleotide. Following a 15-min incubation at 30°C in the presence of 25 pg poly[d(I-C)], samples were applied to an 8% native polyacrylamide gel and electrophoresed at 250 V for 2.5 hr at 4°C in 50 mM Tris, 100 rnM glycine, pH 8.5, running buffer (21). Gels were subsequently dried onto nitrocellulose filters and bands resulting from gel retardation were quantitated by the Betagen blot analyzer prior to autoradiography. RESULTS Injection of LPS into mice results in a rapid rise in serum TNF which peaks 1 hr post-LPS. This increase in serum TNF correlated with an increase in TNF mRNA in peritoneal macrophages (Fig. 1A). TNF serum levels return to control levels 3-5 hr later (14) and by 20 hr, there was no detectable serum TNF and no TNF mRNA in peritoneal macrophages. These animals have become endotoxin tolerant as a second injection of LPS at 20 hr did not result in any increase in serum TNF or TNF mRNA in peritoneal macrophages. The absenceof a secondary increase in TNF mRNA was not unique to this cytokine as a similar pattern was observed for IL-1 p while actin mRNA levels remained relatively constant at all time points evaluated (Fig. 1B). A profile similar to IL-l fl was also observed for IL-l CYwith peak mRNA levels 1 hr after LPS and no increasesdetected 1 hr after the secondLPS stimulus (data not shown).
A
c 1000
30' WWW f
20-
_ 800
0 Counts _
p
800 400
% loI-
9 s0
200 0 0
1
20
21
Hr
FIG. 1. Peritoneal macrophage TNF mRNA and serum TNF levels during endotoxin tolerance. Thioglycollate-primed mice were injected with 10 gg of LPS iv and at designated intervals animals were bled and peritoneal macrophageswere harvested and processedfor RNA. Twenty hours after the initial LPS stimulus animals received a second LPS injection and were sacrificed 1 hr later. Serum TNF levels and peritoneal macrophage TNF mRNA were quantitated (A) and the actual autoradiogram for the mRNA analysis for TNF, actin, and IL- 1 fl is presented (B). Representative experiment of four.
516
SHORT COMMUNICATIONS
The correlation between serum TNF and mRNA levels in peritoneal macrophages from endotoxin tolerant mice suggesteda pretranslational regulation of TNF expression. Nuclear run-on transcriptional assayswere then performed to more precisely determine the mechanism involved in cytokine regulation during endotoxin tolerance. As demonstrated in Fig. 2, while macrophage nuclei isolated 30 min after the LPS injection had increased transcriptional activity at both the TNF and IL- 1 ,8loci, nuclei isolated 30 min following the second LPS injection had significantly reduced transcriptional activity at these loci. Constitutive transcription at the (Y actin locus in distinction was observed at all points evaluated. Similar results were observed if nuclei were isolated at 40 or 60 min post-primary or secondary LPS injection, although the overall transcriptional activity at both the TNF and IL- 1 p loci were decreasedrelative to the 30 min time point (data not shown). Theseresults suggestedthat in vivo endotoxin tolerance correlated with the transcriptional regulation at both the TNF and IL- 1 loci. The role for NF-RR in the transcriptional regulation of TNF expression during endotoxin tolerance was then evaluated by DNA gel shift mobility experiments. As demonstrated in Fig. 3, nuclei from macrophages 1 hr post-LPS injection had a fivefold increase (lane 3) over macrophages from nonstimulated mice (lane 2) in the intensity of a gel shift band using a 32Poligonucleotide sequence recognized by NF-KB. Macrophages from endotoxin tolerant mice in contrast failed to demonstrate a significant increase in nuclear-associated NFXR-like activity following a second LPS stimulus (lanes 4 and 5). The inability of the macrophages in endotoxin tolerant mice to translocate NF-RI3 into the nucleus following a repetitive LPS stimulus is consistent with the regulation of TNF and IL-l expression occurring at the transcriptional level. A
1200 1000 4 800 g 600 0 400 200 n 0
0.5
20
20.5
Hr
6
TNF IL-1
.,
Actin pGEM1 0
0.5
20
20.5
Hr FIG. 2. Nuclear run-on transcriptional assay during endotoxin tolerance. Mice were injected iv with 10 r.rgof LPS and peritoneal macrophages were isolated 30 min later and the resulting nuclei were allowed to incorporate ‘*P UTP in vitro. At 20 hr mice were sacrificed or again injected with LPS and macrophages were harvested 30 min later. RNA was isolated from each nuclear preparation and lo6 cpm were hybridized to cDNA plasmids for TNF, IG 10, actin, and the TNF parental vector pGEM 1.Hybridization was quantitated by Betagen (A) and autoradiographed (B). Representative experiment of three.
SHORT COMMUNICATIONS
517
12345
FIG. 3. Absenceof NF-KB induction during endotoxin tolerance. Macrophageswere lysed at the designated intervals post-LPS and nuclear extracts were incubated with an oligonucleotide from the TNF promoter containing a NF-KB recognition sequence.Retardation of the electrophoretic migration of this oligonucleotide was evaluated on a native 8% polyacrylamide gel and the band intensity was quantitated by Betagen. Lanes l-5 represent no nuclear extract (lane l), and extracts from macrophages 0, 1, and 20 hr (lanes 2-4) post LPS and 1 hr after the second LPS stimulus (lane 5). Hybridization counts for the gel shift band (arrow) were 17,012, 89,199, 14,964, and 15,775 for lanes 2-5, respectively. Representative experiment of three.
DISCUSSION Endotoxin tolerance is characterized by a diminished host inflammatory response following a repetitive LPS challenge. Whereas a decreasein cytokine levels has been demonstrated in the sera of endotoxin tolerant animals and a reduced production of these mediators is also observed in LPS refractory macrophage cultures, the mechanisms involved in the establishment of tolerance remain poorly understood. Animal models characterized by an enhanced sensitivity to the lethal effectsof endotoxin such as adrenalectomized (14) or galactosamine-sensitized mice (22) do not become endotoxin tolerant (23) using protocols which induce tolerance in normal animals. Furthermore, tolerance does not result in the reduction in the levels of all serum cytokines as LPS has been demonstrated to increase serum levels of both IL- 1 (13) and IL-6 (9) in endotoxin tolerant mice. Previous studies have focusedon the level at which cytokine expressionwas regulated in in vitro cultures of LPS refractory macrophagesor macrophage-like cell lines. These studies have reported transcriptional regulation in murine RAW 264 cells (10) and the human monocyte-macrophage line Mono-Mac-6 (1 l), post-transcriptional regulation related to mRNA stability for IL- 1 fi in THP- 1 cells (24) and post-translational regulation in primary cultures of murine peritoneal macrophages ( 12). Differences in the cell systemsused and the nature of the protocols employed to induce an endotoxin refractory state may explain the absence of a consensus concerning the mechanisms involved in LPS desensitization. In the present study, the absenceof a secondary peak in serum TNF was due to transcriptional regulation of cytokine gene expression as evidenced by decreasednuclear run-on activity and a reduction in nuclear-associated
518
SHORT COMMUNICATIONS
NF-KB. These observations clearly demonstrate that transcriptional regulation of macrophage cytokine gene expression is fundamentally involved in the process of endotoxin tolerance. Whether cytokine expression during endotoxin tolerance in other animal models is regulated at the transcriptional level remains to be determined. The difference in transcriptional regulation in this study and post-translational regulation of TNF expression in our previous in vitro study on LPS refractory macrophage cultures demonstrates multiple levels of regulating TNF synthesis during LPS desensitization. Whether the regulation observed at the transcriptional level in endotoxin tolerant animals was mediated by endogenous glucocorticoids or other factors which are absent in in vitro cultures remains to be determined. The role of NF-KB has been considered in the human Mono-Mac-6 cell line in in vitro endotoxin refractory studies. While TNF mRNA was not apparent in response to a second LPS stimulus, these cells did demonstrate a significant increase in NF-KB activity (11). In the present study, the absence of a significant increase in NF-KB after the second LPS stimulus was consistent with reduced run-on transcriptional activity for TNF and IL-l ,& The differences in NF-KB observed between the present study and the previous study (11) may reflect fundamental differences in the mechanisms involved in the regulation of endotoxin tolerance between the in vitro studies with the human Mono-Mac-6 cell line and the in vivo studies where other regulatory factors, including endogenous glucocorticoids, are involved in the regulation of cytokine gene expression. These studies suggestthat the mechanisms by which endotoxin tolerance results in transcriptional regulation of cytokine gene expression are diverse and may be dependent on species, the state of macrophage differentiation, as well as the involvement of other soluble factors, the expression of which may be dependent upon the appropriate in vitro and in vivo environments. Soluble TNF receptors, for example, may represent one such factor which could explain the absenceof a second peak of biologically active serum TNF in endotoxin tolerant mice (25). However, attempts to demonstrate the presence of soluble receptors or other blocking factors by performing TNF dose-response cytotoxicity assayswith sera from control and endotoxin tolerant mice have not been successful(unpublished observations). In addition to transcriptional regulation of cytokine genes in macrophages from endotoxin tolerant mice, other metabolic changes associated with a LPS refractory statehave been describedin macrophages.Introna et al. (26), for example, have reported decreasedtranscriptional activity by nuclear run-on assaysfor c-fos in response to LPS in endotoxin refractory murine peritoneal macrophages.The demonstration that secretion of arachidonic acid metabolites is reduced in macrophages from endotoxin tolerant rats (27) and that these macrophageshave a diminished thromboxane response to guanine nucleotide analogues (28), as well as the inability in the present study to detect a LPS-mediated increase in NF-KB in macrophages from endotoxin tolerant mice, suggeststhat endotoxin tolerance may be associated with the uncoupling of signal transduction mechanisms associated with endotoxin activation. ACKNOWLEDGMENTS The authors acknowledgeLaura Guthrie for her superb technical assistance,Drs. Bill Roeder (Lilly Research Labs) and David Chapman (Washington University of St. Louis) for the TNF and IL-1 j3plasmids,respectively, and Dr. Chandrasekhar for his review of our manuscript.
SHORT COMMUNICATIONS
519
REFERENCES 1. Greisman, S. E., In “Beneficial Effects of Endotoxins” (A. Nowotny, Ed.), p. 149. Plenum, New York, 1983. 2. Heyman, A., Vener. Dis. Znf 26,5 1, 1945. 3. Beeson, P. B., J. Exp. Med. 86, 39, 1947. 4. Watson, D., and Kim, Y. B., J. Exp. Med. 118, 425, 1963. 5. Quesenberry, P., Halperin, J., Ryan, M., and Stohlman, F., Blood 45, 789, 1975. 6. Henricson, B. E., Benjamin, W. R., and Vogel, S. N., Infect. Immun. 58, 2429, 1990. 7. Williams, Z., Hertogs, C. F., and Pluznik, D. H., Infect. Immun. 41, 1, 1983. 8. Zuckerman, S. H., and Bendele, A. M., In@. Immun. 57, 3009, 1989. 9. Heremans, H., Van Damme, J. V., Dillen, C., Dikmans, R., and Billiau, A., J. Exp. Med. 171, 1853, 1990. 10. Virca, G. D., Kim, S. Y., Glaser, K. B., and Ulevitch, R. J., J. Biol. Chem. 264, 21,951, 1989. 11. Haas, J. G., Baeuerle, P. A., Riethmuller, G., and Ziegler-Heitbrock, H. W. L., Proc. Natl. Acad. Sci. USA 81,9563, 1990. 12. Zuckerman, S. H., Evans, G. F., Snyder, Y. M., and Roeder, W. D., J. Immunol. 143, 1223, 1989. 13. Zuckerman, S. H., Evans, G. F., and Butler, L. D., Infect. Immun. 59, 2774, 1991. 14. Zuckerman, S. H., Shellhaas, J., and Butler, L. D., Eur. J. Immunol. 19, 301, 1989. 15. Greenberg, M. E., and Ziff, E. B., Nature 311,433, 1984. 16. Gardner, S. M., Mock, B. A., Hilgers, J., Huppi, K. E., and Roeder, W. D., J. Immunol. 139,476, 1987. 17. Minty, A. J., Caravatti, M., Robert, B., Coben, A., Daubas, P., Weydert, A., Gros, F., and Buckingham, M. E., J. Biol. Chem. 256, 1008, 1981. 18. Gray, P. W., Glaister, D., Chen, E., Goeddel, D. V., and Pennica, D., J. Immunol. 137, 3644, 1986. 19. Dignam, J. D., Lebovitz, R. M., and Roeder, R. G., Nucleic Acids Res. 11, 1475, 1983. 20. Shakhov, A. N., Collart, M. A., Vassalli, P., Nedospasov, S. A., and Jongeneel, C. V., J. Exp. Med. 171, 35, 1990. 2 1. Chodosh, L. A., In “Current Protocols in Molecular Biology” (F. M. Ausubel, R. Brent, R. E. Kingston,
22. 23. 24.
25. 26.
D. D. Moore, J. G. Seidman, J. A. Smith, and K. Struhl, Eds.), Vol. 2, p. 12.2.1.Greene Publishing Associates,New York, 1988. Galanos, C., Freudenberg, M. A., and Reutter, W., Proc. Natl. Acad. Sci. USA 76, 5939, 1979. Evans, G. F., and Zuckerman, S. H., Eur. J. Immunol. 21, 1973, 1991. Fenton, M. J., Vermeulen, M. W., Clark, B. D., Webb, A. C., and Auron, P. E., J. Immunol. 140,2261, 1988. Nophar, Y., Kemper, O., Brakebusch, C., Engelmann, H., Zwang, R., Aderka, D., Holtmann, H., and Wallach, D., EMBO J. 9, 3269, 1990. Introna, M., Bast, R. C., Johnston, P. A., Adams, D. O., and Hamilton, T. A., J. Cell. Physiol. 131, 36,
1987. 27. Rogers, T. S., Halushka, P. V., Wise, W. C., and Cook, J. A., Prostaglandins 31, 639, 1986. 28. Coffee, K. A., Halushka, P. V., Wise, W. C., and Cook, J. A., Biochim. Biophys. Acta 1035,201, 1990.
