- Email: [email protected]
The effects of glucocorticoids on phorbol ester and cytokine stimulated transcription factor activation in human lung
Life Sciences, Vol 55, No 14, pp. 1147-1153, 1994 Copyright © 1994 Elsevier Science Ltd Pnnted m the USA All rights ~ t v e d 0024-3205/94 $6.00 + 00
Pergamon
0024-3205(94)00243-6
THE EFFECTS OF GLUCOCORTICOIDS ON PHORBOL ESTER AND CYTOKINE STIMULATED TRANSCRIPTION FACTOR ACTIVATION IN HUMAN LUNG
I.M. Adcock, H. Shirasaki, C.M Gelder, M. J Peters, C R. Brown & P.J. Barnes. Department of Thoracic Medicine, National Heart and Lung Institute, London SW3 6LY, U.K. (Received in final form July 27, 1994) Summary
Glucocorticoids have a wide variety of effects which result in the dampening of inflammatory and immune responses and other challenges to homeostasis. An important site of steroid action may be on the control of transcription factor binding to DNA. The interaction of the transcription factors, activator protein 1 (AP-1) and nuclear factor kappa from B cells 0'eFt,B) with DNA and glucocorticoid receptors was analysed by gel mobility shift assays following stimulation by tumour necrosis factor alpha (TNFcx) and a phorbol ester (PMA) that activates protein kinase C PMA and TNFot both caused significant (1g0340%) increases in AP-1 and NFr,_B DNA binding which peaked at 15 minutes and decreased to a constant elevated level at between 1-3 hrs and was sustained for 24hrs. Dexamethasone (llxM) caused a rapid and long lasting 40-50% decrease in both AP-1 and NFr,B DNA binding lasting over 24hrs Combined treatment with dexamethasone and PMA or TNFct prevented the increase in both AP-1 and NFr,B binding due to PMA and TNFct returning levels to those seen in control untreated samples This suggests that in human lung, the glucocorticoid receptor functionally interacts within the nucleus with other transcription factors that are induced by inflammatory mediators such as cytokines. This may be an important molecular site of steroid action in chronic inflammatory lung diseases such as asthma. Key Words: lung, glucocorticoidreceptors, activation protein-l, NFKB,cross coupling
Inhaled glucocorticosteroids are the most effective therapy currently available for the treatment of asthma Despite their efficacy in controlling all types of asthma their mode of action remains uncertain Glucocorticoids inhibit many steps in the inflammatory process and limit the proliferative responses of cells in chronic destructive diseases (1). In general steroids appear to dampen the defence reactions activated by challenges to homeostasis In asthma inhaled steroids reduce airway hyperresponsiveness and suppress the inflammatory response in asthmatic airways (2). Recent reports have indicated that glucocorticoid receptors (GR) antagonise the actions of many inflammatory mediators which act through various second messenger systems that all converge on the transcription factor, activating protein-1 (AP-1) which is a heterodimer of the proto-oncogene proteins c-Fos and c-Jun (3-5) This suggests that an important site of steroid action may be on the control of transcription factor binding to specific cis-acting elements on D N A These trans-acting proteins affect the rate of mRNA synthesis by binding to discrete sequences or promotors in the 5' upstream controlling region of genes (6) Interaction of distinct regulatory pathways may occur via cross-coupling of transcription factors (3-5) Cytokines have been implicated in the pathogenesis of Correspondenceto: ProfessorP J Barnes,Departmentof ThoractcMedicine,,NattonalHeart & Lung Institute, DovehouseStreet, LondonSW3 6LY, U K
1148
Transcription Factor Interactions in Lung
Vol. 55, No. 14, 1994
many chronic inflammatory diseases, including arthritis and asthma (7), and their molecular effects include the induction of several specific transcription factors including AP-1 and nuclear factor kappa B (NFrd3). Tumour necrosis factor-cx (TNFc0 is such a cytokine and is synthesized in many cells upon stimulation with endotoxin, inflammatory mediators or other cytokines such as IL-1 and, in an autocrine manner, by TNFct itself Furthermore TNFcx induces the synthesis of other cytokines such as IL-2 and its receptor via the activation of NFr,B (8) TNFcx has also been implicated in the pathological events occurring in cachexia, endotoxic shock and chronic inflammation (9). Protein kinase C may also be involved in chronic regulation of cell function (10) and may be activated by phorbol esters such as phorbol mysterate acetate (PMA) We have therefore examined in human lung the interaction of glucocorticoids with AP-1 and NFr,.B, which have been implicated in the chronic inflammatory responses to TNFct and other inflammatory stimuli in many cell types (11,12) Methods
Chemicals All chemicals used were of at least Analar grade and all solutions and glassware were sterilised either by autoclaving or by dry heat Enzymes were supplied by Northumberland Biologicals Ltd (Washington, U K ) and Promega (Southampton, U K ) [32p] yATP (3000Ci/mmole) was purchased from Amersham International (Amersham, U K.) Tissues
Lungs were obtained from 13 donors (7-45yr) undergoing cardiac transplantation and maintained on ventilation prior to transplantation Electrophorettc Mobthtv Shift A s s ~ Nuclear proteins were extracted from chopped human lung tissue (5-10mm 3) that had been incubated at 20°C in oxygenated Krebs solution containing 0 llxM PMA or 10pM TNFcx in the presence or absence of lp.M dexamethasone according to the method of Osborn et a/.(13) In some experiments the tissue was finely chopped using a Mcllwain tissue chopper and used for incubation with cytokines (IL-1B, IL-2, IL-6, TNFa, TNFB) at a range of concentrations (10"14-10"lli) at 37°C Tissue was lysed using NP-40 lysis at 4°C for 15 minutes and soluble nuclear extracts obtained following osmotic lysis of the nuclear envelope. Incubation of nuclear proteins with 50,000 cpm 32p-labelled double-stranded oligonucleotides containing a tandem repeat of the consensus sequence for the AP-1 DNA binding site (-GATCCTTCGTGACTCACrCGCK~ATC CTTCGTGACTCAGCGG-) and the NFv._B DNA binding site (-GATCCAAGGGGACTTTCC ATGGATCCAAGGGGACTTTCCATG-) (Gibco-BRL, Uxbridge, UK) as described by Schiale et aL (3) DNA-protein complexes were resolved on a 6% non-denaturing polyacrylamide gel (37:1 acrylamide: bis-acrylamide) in 0 25x TBE Gels were dried and auto-radiographed at -70°C using Kodak XAR-1 film. The retarded bands were quantified by laser densitometry and band density measurements were then expressed as a percentage of binding at t=0. Specificity was determined by addition of a 100-fold excess unlabelled double-stranded AP-1 or NFkB oligonucleotides and also the addition of 5lag antibody to human c-Fos, c-Jun or the p65 subunit of NFv,~B. Results
Mobility shift assays on nuclear extracts showed a 2-300% increase in AP-1 activity present in the lung within 15 minutes of incubation with PMA (0 ltxM) (Figs. la and 2a) This increase was reduced to a 95+15% increase by 30 minutes and this level was maintained for the remainder of the experiment (24h). Dexamethasone (l~tM) caused a marked decrease in AP-1 activity at 15 minutes This reduction was found throughout the time course of the experiment reaching a value of 60a:6%
Voi. 55, No. 14, 1994
Transcription Factor Interactions in Lung
1149
at 3h of dexamethasone treatment Combined treatment with PMA and dexamethasone abolished this enhancement in AP-1 binding induced by PMA throughout the 24h time course with levels of AP-1 binding being reduced to those of control samples or in some cases to levels seen in dexamethasone treated samples. Specificity of the AP-1/DNA binding was confirmed by the use of an excess of cold unlabelled double-stranded oligonucleotide (Fig. l a) and antibodies to c-Fos and cJun that greatly reduced the AP-1 band (data not shown) Similar results were seen with NFr.B in that there was an initial 3-400% increase in DNA binding which was maintained at a slightly reduced level for 24h following 0. lttM PMA treatment (Figs lb and 2b). Dexamethasone treatment (lttM) for time points up to 24h produced a marked reduction in NFr.B binding which reached a maximum at 30 minutes and remained at this level throughout the experiment. Combined PMAdexamethasone treatment produced a marked reduction in NFr.B binding as compared to the levels seen in PMA treated samples at all time points studied The specificity of NFr.B binding was again determined by the addition of a 100-fold excess of unlabelled double-stranded oligonucleotide (Fig. lb) or 5lag of anti-p65 antibody. To check for the specificity of transcription factor effects, the level of GRE binding was also measured in these tissues. Dexamethasone produced a marked increase in GKE binding, whilst PMA and TNFct caused a 50% reduction in the level of GRE binding (data not shown). PMA
Control
Dex
g~ o o
~ o
o ~
o cq
o ~
c~
ooo-°
~ o ~
o
. . . .
r~
Control
o o
~ o
o ~
O,
PMA
o ¢~
o g
o o
m o
o ~
Dex
o tcs
o g
o o
o
m o
o ~
4~- A P - I
XS ohgo
o o3
o g
44-
NF•B
L
Fig. 1. (a) Gel mobility assay showing an increase in AP-1 binding after treatment with PMA (0.1 gtM) over time, decreases following dexamethasone (I~M) and no change in basal level following control treatment of human lung parenchyma. (b) Gel mobility assay showing an increase in NFkB binding after PMA treatment (0 lgM) over time, decreases following dexamethasone (lvM) and no change in basal level following control treatment of human lung parenchyma. In all experiments 10gg human lung nuclear extract was used. Arrows indicate the position of the AP-1 or NFkB/oligo complex that has been retarded in the gel matrix Specificity of binding is shown by incubation of nuclear cell extract with and without excess unlabelled double-stranded oligo.
1150
Transcription Factor Interactions in Lung
A
B :~ +450 -
+300 ~ ¢= +250
~ +400 .~ +350 ~ +300 el .o +250
-¢$ +200 +150
...~
T +100 o_ ~ i
Voi. 55, No. 14, 1994
~
]
PMA
~
+50 0
SHAM P/D DEX
-50 a~ -100 0
1
, 2
, 3
, 4
II
, 24
~
~
+2oo
~, "~
SHAM DEX
o
P
-5o -100 1
Time (hours)
PMA
u. Z +150 ¢__ +1oo ~ +5o -
2
, 3
II 4
, 24
Time (hours) Fig. 2.
Time course of transcription factor binding The effect of dexamethasone (IlIM, A), PMA (0 llaM, V), sham ([3) or combined treatment (0) on (a) AP-1 activity and (b) on NFr,.B activity over 24 hr incubation In all experiments 10lag human lung nuclear extract was used. Values shown are means + SEM (n=6-12) Using finely chopped human lung a rapid increase in AP-I binding was seen following cytokine incubation at 37°C (IL-1i3, IL-2, IL-6 and TNFa, all at 10"llM) although this increase was not sustained over the time course of the experiment The level of cytokine-induced AP-1 binding at the lh time point was up to 125% greater than that in control untreated tissue Binding returning to control levels alter 24h with all cytokines examined except TNF0t and IL-2 that still enhanced AP-1 binding at 24h by up to 50% of control levels (Fig 3a) A dose response curve to both TNFo~ and IL-113 after incubation for lh showed a dose dependent increase in AP-1 band density that reached peak levels similar to those seen in the previous experiment (Fig 3b). The dose range used suggests that these increases in AP-1 binding occur at concentrations known to occur m w v o . Similar results were seen with NFv,.B following TNFcq IL-113 and IL-2 treatment of human lung NFr,B was rapidly induced by TNFct, IL-lf5 and IL-2 reaching a peak at 60 minutes at a level up to 157% greater than in control samples Within 3hrs this was reduced to 87°/0 of control levels and this was sustained for up to 24hrs in the case of TNFcx, and IL-2 Within 24hrs the II_.-18 induction of NFv,_B was reduced to control levels (data not shown) Dose responses of NFvd3 induction following TNFa and IL-1B showed maximal induction by 10pM cytokine Treatment of human lung tissue with 10pM TNF~ produced an 82% and 106% increase in AP-1 and NFrB DNA binding This was maximal at lh and was sustained for the duration of the experiment Dexamethasone (1 laM) alone produced a 30-50% decrease in the band density of AP-1 and NFrA3 that peaked at 3h and was still maintained at 24h When dexarnethasone was added to the incubation media along with TNFtx, the TNFtx induced increase in AP-1 and NFrA3 binding was severely attenuated by up to 95% in the case of AP-1 and up to 80% in the case of NFrd3 over the whole time course of the experiment The levels seen were no different from those of control untreated samples for AP-1 (Fig. 3c+d) This suggests that the ability of GR to modulate the DNA binding of NFr._B induced by TNFot is reduced compared to that of its ability to inhibit PMA-induced NFrd3 bindmg Thus TNFot may produce other factors that can sequester GR, limiting GR effects on NFr,.B binding
Vol. 55,
Transcription Factor Interactions in Lung
N o . 14, 1 9 9 4
1151
B
A
.--,_ ~ +150
o L-la • IL-2
>.+100
"o
"~c +100 E <[
+50
c_
& N
-50
I
C
2
3
Time
>, +125 +100 i ÷75
4
14 13 1211
24 D
IHrs)
÷50 "1"25
TNF/OEX SHAM
0
IL-1B
+150 +125
--~ +50 +25
TNF/DEX SHAM
0 • _¢
-25
13 12 11 t 0 9 -log(M)
TNFa
OEX
-SO
-25 -50 •
o
- 75
~t -100 f
, 2
Treatment
, 3
, 4
Z~
II~
2•
- 75 -IO0
. . . . 1
T i m e (hours)
2
Treatment
II-----1 3
•
2•
T i m e (hours)
Fi~. 3. Time and dose dependence of cytokine action on AP-1 and NFrB DNA binding in human peripheral lung. (a) IL-113 (0), IL-6 (V), and TNFI~(*) (10-11M) caused a transient rise in AP-1 binding which peaked at lh and returned to baseline levels at 24h. TNFtx (¢) and IL2 (A)(10-11M) induced a long-term induction of AP-1 activity. (b) Concentrationdependent increase in AP-1 activation following TNFtx and IL-113 incubation for lh. (n=26). (c) The effect of TNFct (10 -11M), in the absence or presence of dexamethasone (1 laM) on AP-1 activity over 24hr incubation. (d) Effect of TNFct (10-11M), in the absence or presence of dexamethasone (llaM) on NFr,_B activity over 24hr incubation. In all experiments 10pg human lung nuclear extract was used Values shown are the means + SEM of 6-12 observations Discussion Control of chronic inflammation is important in the treatment of several inflammatory conditions, including asthma and fibrosing lung diseases. The regulatory pathways that control chronic inflammation are complex and multifactorial. Glucocorticoids inhibit the inflammatory response of many cells to a wide variety of stimuli and so an understanding of its mode of action is of central importance Cytokines may be important mediators of chronic inflammation. Although the molecular pathways involved in cytokine action are not well understood, several cytokines act on responsive cells to induce the activation or synthesis of a few well characterised transcription factors The finding that both TNFct and the phorbol ester PMA induced a rapid-long lasting activation of both AP-1 and NFr.B in the human lung, and that the ability of these DNA binding
1152
Transcription Factor Interactions in Lung
Vol. 55, No. 14, 1994
proteins to bind to their target sequences is abolished by dexamethasone suggests that this may be an important molecular mechanism of steroids, that could account for their multiple antiinflammatory actions and for their efficacy in chronic inflammatory diseases Several mechanisms have been proposed to account for the interaction between transcription factors One factor could replace another at its site of action on the upstream regulatory site of the target genes, the two factors may compete for another factor at either the gene or the protein level, or their may be a direct interaction between one transcription factor (e.g GR) and other transcription factors at a protein-protein level (14) Direct protein-protein interaction is the more likely mechanism in these studies, at least m the short term, as AP-1 binding is reduced in the nuclear extract, although no reductions in the components of AP-1 are seen in the same extracts by western blotting until 24h (I Adcock, unpublished observations). Furthermore, the oligonucleotides used in these studies show no specific binding for other DNA binding proteins being highly selective for AP-1 and NFr,_B In the present studies the effect of actinomycin D is not addressed. Although is unlikely to affect the short term results, those for the 24hr time point may be influenced by the transcriptional activation of various genes by PMA, TNFct or dexamethasone on the components of the AP-1 complex. The different models proposed to account for the interaction between AP-1 and GR differ by virtue of their expected effect on A_P-1 DNA binding of GR activation With the tethering mechanism (4,15) proposed by Herrlich, an interaction between AP-1 and GR occurs without any effect on AP1 DNA binding. Indeed this group has analysed the AP-1 binding on the collagenase gene by Footprinting analysis and show no lack of protection following GR stimulation although the transcriptional readout was markedly reduced In contrast, the mutual repression model proposes that the interaction occurs in solution and both proteins block the DNA binding of the other In other studies investigating the interaction between GR and AP-1 Kerppola et al (16) again using the collagenase gene promotor region have shown a repression of AP-1 DNA binding by the DNA binding domain of the GR Furthermore, we have previously shown a reduction in GR binding to DNA in human lung following PMA treatment, suggesting that mutual repression can occur in this tissue (17) These results and the results presented in this study suggest that the exact interaction between GR and AP-I may depend upon the cell type used and the exact type of stimulus given and also suggest that more than one mechanism of interaction may occur at any one time. Some of the contradictions between the two models may be explained by the relative affinity of the GR for the Fos-Jun heterodimer and the Jun-Jun homodimer composing two forms of AP-1 family Interaction may occur between the GK and the Fos-Jun heterodimer that displaces the Fos-Jun dimer from the AP-1 site allowing the lower affinity Jun-Jun dimer to bind, since GR has a lower affinity for Jun-Jun homodimer than for Fos-~'un heterodimer (18). The Jun-Jun homodimer is less potent at activating transcription resulting in repression of gene transcription with no concomitant loss of AP-1 DNA site occupancy Therefore cell specificity of GR inhibition of AP-1 activity may be due to the presence of different endogenous AP-1 family proteins in different cell types. Yamamoto and colleagues (19) have proposed an alternative model for AP-1/GR interaction in the nucleus involving overlapping atypical GREs and TREs termed a composite GRE C-Jun homodimers interact with GR to cause a synergistic effect increasing transcription of responsive genes, whilst c-Fos/c-Jun heterodimers interact with DNA bound GR to switch off transcription Thus transcription rate is controlled by the level of c-Fos in the cell Glucocorticoids inhibit the synthesis of several cytokines, including IL-1 and TNFa (20,21). Although steroids may inhibit the inflammatory process at several sites, the direct interaction with transcription factors, such as AP-1 and NFrd3 may be of particular relevance, since this could provide a mechanism for switching off inflammation driven by cytokines known to activate these transcription factors.
Vol. 55, No. 14, 1994
Transcription Factor Interactions in Lung
1153
The results of our studies indicate that in human peripheral lung the transcription factors AP-1 and NFr.B are induced by phorbol esters and cytokines in a time and concentration dependent manner. Gel mobility shift assays indicate that the activated GR may interact with these transcription factors within the nucleus causing an inhibition of DNA binding by the respective proteins. In this way steroids may inhibit the synthesis chronic inflammatory mediators such as cytokines or inhibit the effects of agents that activate protein kinase C. The anti-inflammatory properties of glucocorticoids in lung could be explained if glucocorticoids repressed cytokine-induced AP-1 and NFr,.B mediated induction of mRNA coding for other cytokines, enzymes and receptor genes. Cross-coupling of distinct and opposing regulatory pathways provides another level of regulation in determining the inflammatory response of the cell This inhibition of transactivation may be an important molecular site of steroid action in asthma and other chronic inflammatory diseases
Acknowledgements CMG is an MRC Training Fellow. MJP is the Allen & Hanburys/Thoracic Society of Australia and New Zealand Travelling Fellow. We would like to thank Professor Sir Magdi Yacoub for providing transplant tissue This work was supported by The Medical Research Council and Glaxo Group Research
References 1 A.MUNCK, D.B MENDEL, L.I SMITH & E ORTI Am.Rev.Respw.Dls. 141 $2-S10. (1990). 2. P J.BARNES N.Engl.JMed. 321 1517-1527 (1989) 3. R.SCHULE, P.RANGARAJAN, S KLIEWER, L J.RANSOME, l BOLADO, N.YANG, I M.VERMA & R M EVANS Cell 6__2_21217-1226 (1990). 4. C.JONAT, H.J RAHMSDORF, K-K PARK, A C B.CATO, S GEBEL, H.PONTA & P.HERRLICH Cell 6._.221189-1204 (1990). 5. H-F.YANG-YEN, J-C.CHAMBARD, Y-L SUN, T.SMEAL, T J.SCHMIDT, J.DROUIN & M.KARIN Cell62 1205-1215 (1990) 6. D S LATCHMAN Blochem.J. 270 281-289 (1990) 7 J.KELLEYAm.Rev.Respir.Dts. 141 765-788 (1990) 8. G CAMUSSI, E.ALBANO, C TETTA & F.BUSSOLINO Eur.J.Btochem. 202 3-14 (1991). 9 K.J.TRACEY, H.VLASSARA& A.CERAMILancetl 1122-1126 (1989). 10. Y NISHIZUKA Science 233 305-312 (1986) 11 M.J LENARDO & D.BALTIMORE Cell 58 227-229 (1989) 12. J.AUW RX, B.STAELS & P.SASSONE-CORSI NucLActds Res. 18 221-228 (1990). 13. L OSBORN, S.KUNKEL & G J NABEL Proc.NatLAcad.Sci. (USA) 86 2336-2340 (1989). 14. M.E.JACKSONd.CellSci. 100 1-7 (1991) 15 H.KONIG, H PONTA, H J RAHMSDORF & P HERRLICH EMBO d. 1__!12241-2246 (1992) 16 T K KERPPOLA, D LUK & T.CURRANMoL Cell. Biol 13 3782-3791 (1993) 17 I M. ADCOCK, C M. GELDER, H SHIRASAKI, M YACOUB & P.J. BARNES Am. Rev. Respw. Dis. 145 A834 (Abstract)(1992) 18. L.SHEMSHEDINI, R KNAUTHE, P.SASSONE-CORSI, A.PORNON & H GRONMEYER EMBO J. 1__03839-3849 (1991) 19. M I.DIAMOND, J.N.MINER, S K YOSHINAGA & K.R YAMAMOTO Science 249 12661272 (1990). 20. B BEUTLER, N.KROCHIN, I.W.MILSARK, C LUEDKE & A.CERAMI Science 232 977980.(1986). 21. S.W.LEE, A-P.TSOU, H.CHAN, J.THOMAS, K PETRIE, E M EUGUI & A.C.ALLISON Proc.Natl.Acad.Sct.(USA) 8__551204-1208 (1988)
Pergamon
0024-3205(94)00243-6
THE EFFECTS OF GLUCOCORTICOIDS ON PHORBOL ESTER AND CYTOKINE STIMULATED TRANSCRIPTION FACTOR ACTIVATION IN HUMAN LUNG
I.M. Adcock, H. Shirasaki, C.M Gelder, M. J Peters, C R. Brown & P.J. Barnes. Department of Thoracic Medicine, National Heart and Lung Institute, London SW3 6LY, U.K. (Received in final form July 27, 1994) Summary
Glucocorticoids have a wide variety of effects which result in the dampening of inflammatory and immune responses and other challenges to homeostasis. An important site of steroid action may be on the control of transcription factor binding to DNA. The interaction of the transcription factors, activator protein 1 (AP-1) and nuclear factor kappa from B cells 0'eFt,B) with DNA and glucocorticoid receptors was analysed by gel mobility shift assays following stimulation by tumour necrosis factor alpha (TNFcx) and a phorbol ester (PMA) that activates protein kinase C PMA and TNFot both caused significant (1g0340%) increases in AP-1 and NFr,_B DNA binding which peaked at 15 minutes and decreased to a constant elevated level at between 1-3 hrs and was sustained for 24hrs. Dexamethasone (llxM) caused a rapid and long lasting 40-50% decrease in both AP-1 and NFr,B DNA binding lasting over 24hrs Combined treatment with dexamethasone and PMA or TNFct prevented the increase in both AP-1 and NFr,B binding due to PMA and TNFct returning levels to those seen in control untreated samples This suggests that in human lung, the glucocorticoid receptor functionally interacts within the nucleus with other transcription factors that are induced by inflammatory mediators such as cytokines. This may be an important molecular site of steroid action in chronic inflammatory lung diseases such as asthma. Key Words: lung, glucocorticoidreceptors, activation protein-l, NFKB,cross coupling
Inhaled glucocorticosteroids are the most effective therapy currently available for the treatment of asthma Despite their efficacy in controlling all types of asthma their mode of action remains uncertain Glucocorticoids inhibit many steps in the inflammatory process and limit the proliferative responses of cells in chronic destructive diseases (1). In general steroids appear to dampen the defence reactions activated by challenges to homeostasis In asthma inhaled steroids reduce airway hyperresponsiveness and suppress the inflammatory response in asthmatic airways (2). Recent reports have indicated that glucocorticoid receptors (GR) antagonise the actions of many inflammatory mediators which act through various second messenger systems that all converge on the transcription factor, activating protein-1 (AP-1) which is a heterodimer of the proto-oncogene proteins c-Fos and c-Jun (3-5) This suggests that an important site of steroid action may be on the control of transcription factor binding to specific cis-acting elements on D N A These trans-acting proteins affect the rate of mRNA synthesis by binding to discrete sequences or promotors in the 5' upstream controlling region of genes (6) Interaction of distinct regulatory pathways may occur via cross-coupling of transcription factors (3-5) Cytokines have been implicated in the pathogenesis of Correspondenceto: ProfessorP J Barnes,Departmentof ThoractcMedicine,,NattonalHeart & Lung Institute, DovehouseStreet, LondonSW3 6LY, U K
1148
Transcription Factor Interactions in Lung
Vol. 55, No. 14, 1994
many chronic inflammatory diseases, including arthritis and asthma (7), and their molecular effects include the induction of several specific transcription factors including AP-1 and nuclear factor kappa B (NFrd3). Tumour necrosis factor-cx (TNFc0 is such a cytokine and is synthesized in many cells upon stimulation with endotoxin, inflammatory mediators or other cytokines such as IL-1 and, in an autocrine manner, by TNFct itself Furthermore TNFcx induces the synthesis of other cytokines such as IL-2 and its receptor via the activation of NFr,B (8) TNFcx has also been implicated in the pathological events occurring in cachexia, endotoxic shock and chronic inflammation (9). Protein kinase C may also be involved in chronic regulation of cell function (10) and may be activated by phorbol esters such as phorbol mysterate acetate (PMA) We have therefore examined in human lung the interaction of glucocorticoids with AP-1 and NFr,.B, which have been implicated in the chronic inflammatory responses to TNFct and other inflammatory stimuli in many cell types (11,12) Methods
Chemicals All chemicals used were of at least Analar grade and all solutions and glassware were sterilised either by autoclaving or by dry heat Enzymes were supplied by Northumberland Biologicals Ltd (Washington, U K ) and Promega (Southampton, U K ) [32p] yATP (3000Ci/mmole) was purchased from Amersham International (Amersham, U K.) Tissues
Lungs were obtained from 13 donors (7-45yr) undergoing cardiac transplantation and maintained on ventilation prior to transplantation Electrophorettc Mobthtv Shift A s s ~ Nuclear proteins were extracted from chopped human lung tissue (5-10mm 3) that had been incubated at 20°C in oxygenated Krebs solution containing 0 llxM PMA or 10pM TNFcx in the presence or absence of lp.M dexamethasone according to the method of Osborn et a/.(13) In some experiments the tissue was finely chopped using a Mcllwain tissue chopper and used for incubation with cytokines (IL-1B, IL-2, IL-6, TNFa, TNFB) at a range of concentrations (10"14-10"lli) at 37°C Tissue was lysed using NP-40 lysis at 4°C for 15 minutes and soluble nuclear extracts obtained following osmotic lysis of the nuclear envelope. Incubation of nuclear proteins with 50,000 cpm 32p-labelled double-stranded oligonucleotides containing a tandem repeat of the consensus sequence for the AP-1 DNA binding site (-GATCCTTCGTGACTCACrCGCK~ATC CTTCGTGACTCAGCGG-) and the NFv._B DNA binding site (-GATCCAAGGGGACTTTCC ATGGATCCAAGGGGACTTTCCATG-) (Gibco-BRL, Uxbridge, UK) as described by Schiale et aL (3) DNA-protein complexes were resolved on a 6% non-denaturing polyacrylamide gel (37:1 acrylamide: bis-acrylamide) in 0 25x TBE Gels were dried and auto-radiographed at -70°C using Kodak XAR-1 film. The retarded bands were quantified by laser densitometry and band density measurements were then expressed as a percentage of binding at t=0. Specificity was determined by addition of a 100-fold excess unlabelled double-stranded AP-1 or NFkB oligonucleotides and also the addition of 5lag antibody to human c-Fos, c-Jun or the p65 subunit of NFv,~B. Results
Mobility shift assays on nuclear extracts showed a 2-300% increase in AP-1 activity present in the lung within 15 minutes of incubation with PMA (0 ltxM) (Figs. la and 2a) This increase was reduced to a 95+15% increase by 30 minutes and this level was maintained for the remainder of the experiment (24h). Dexamethasone (l~tM) caused a marked decrease in AP-1 activity at 15 minutes This reduction was found throughout the time course of the experiment reaching a value of 60a:6%
Voi. 55, No. 14, 1994
Transcription Factor Interactions in Lung
1149
at 3h of dexamethasone treatment Combined treatment with PMA and dexamethasone abolished this enhancement in AP-1 binding induced by PMA throughout the 24h time course with levels of AP-1 binding being reduced to those of control samples or in some cases to levels seen in dexamethasone treated samples. Specificity of the AP-1/DNA binding was confirmed by the use of an excess of cold unlabelled double-stranded oligonucleotide (Fig. l a) and antibodies to c-Fos and cJun that greatly reduced the AP-1 band (data not shown) Similar results were seen with NFr.B in that there was an initial 3-400% increase in DNA binding which was maintained at a slightly reduced level for 24h following 0. lttM PMA treatment (Figs lb and 2b). Dexamethasone treatment (lttM) for time points up to 24h produced a marked reduction in NFr.B binding which reached a maximum at 30 minutes and remained at this level throughout the experiment. Combined PMAdexamethasone treatment produced a marked reduction in NFr.B binding as compared to the levels seen in PMA treated samples at all time points studied The specificity of NFr.B binding was again determined by the addition of a 100-fold excess of unlabelled double-stranded oligonucleotide (Fig. lb) or 5lag of anti-p65 antibody. To check for the specificity of transcription factor effects, the level of GRE binding was also measured in these tissues. Dexamethasone produced a marked increase in GKE binding, whilst PMA and TNFct caused a 50% reduction in the level of GRE binding (data not shown). PMA
Control
Dex
g~ o o
~ o
o ~
o cq
o ~
c~
ooo-°
~ o ~
o
. . . .
r~
Control
o o
~ o
o ~
O,
PMA
o ¢~
o g
o o
m o
o ~
Dex
o tcs
o g
o o
o
m o
o ~
4~- A P - I
XS ohgo
o o3
o g
44-
NF•B
L
Fig. 1. (a) Gel mobility assay showing an increase in AP-1 binding after treatment with PMA (0.1 gtM) over time, decreases following dexamethasone (I~M) and no change in basal level following control treatment of human lung parenchyma. (b) Gel mobility assay showing an increase in NFkB binding after PMA treatment (0 lgM) over time, decreases following dexamethasone (lvM) and no change in basal level following control treatment of human lung parenchyma. In all experiments 10gg human lung nuclear extract was used. Arrows indicate the position of the AP-1 or NFkB/oligo complex that has been retarded in the gel matrix Specificity of binding is shown by incubation of nuclear cell extract with and without excess unlabelled double-stranded oligo.
1150
Transcription Factor Interactions in Lung
A
B :~ +450 -
+300 ~ ¢= +250
~ +400 .~ +350 ~ +300 el .o +250
-¢$ +200 +150
...~
T +100 o_ ~ i
Voi. 55, No. 14, 1994
~
]
PMA
~
+50 0
SHAM P/D DEX
-50 a~ -100 0
1
, 2
, 3
, 4
II
, 24
~
~
+2oo
~, "~
SHAM DEX
o
P
-5o -100 1
Time (hours)
PMA
u. Z +150 ¢__ +1oo ~ +5o -
2
, 3
II 4
, 24
Time (hours) Fig. 2.
Time course of transcription factor binding The effect of dexamethasone (IlIM, A), PMA (0 llaM, V), sham ([3) or combined treatment (0) on (a) AP-1 activity and (b) on NFr,.B activity over 24 hr incubation In all experiments 10lag human lung nuclear extract was used. Values shown are means + SEM (n=6-12) Using finely chopped human lung a rapid increase in AP-I binding was seen following cytokine incubation at 37°C (IL-1i3, IL-2, IL-6 and TNFa, all at 10"llM) although this increase was not sustained over the time course of the experiment The level of cytokine-induced AP-1 binding at the lh time point was up to 125% greater than that in control untreated tissue Binding returning to control levels alter 24h with all cytokines examined except TNF0t and IL-2 that still enhanced AP-1 binding at 24h by up to 50% of control levels (Fig 3a) A dose response curve to both TNFo~ and IL-113 after incubation for lh showed a dose dependent increase in AP-1 band density that reached peak levels similar to those seen in the previous experiment (Fig 3b). The dose range used suggests that these increases in AP-1 binding occur at concentrations known to occur m w v o . Similar results were seen with NFv,.B following TNFcq IL-113 and IL-2 treatment of human lung NFr,B was rapidly induced by TNFct, IL-lf5 and IL-2 reaching a peak at 60 minutes at a level up to 157% greater than in control samples Within 3hrs this was reduced to 87°/0 of control levels and this was sustained for up to 24hrs in the case of TNFcx, and IL-2 Within 24hrs the II_.-18 induction of NFv,_B was reduced to control levels (data not shown) Dose responses of NFvd3 induction following TNFa and IL-1B showed maximal induction by 10pM cytokine Treatment of human lung tissue with 10pM TNF~ produced an 82% and 106% increase in AP-1 and NFrB DNA binding This was maximal at lh and was sustained for the duration of the experiment Dexamethasone (1 laM) alone produced a 30-50% decrease in the band density of AP-1 and NFrA3 that peaked at 3h and was still maintained at 24h When dexarnethasone was added to the incubation media along with TNFtx, the TNFtx induced increase in AP-1 and NFrA3 binding was severely attenuated by up to 95% in the case of AP-1 and up to 80% in the case of NFrd3 over the whole time course of the experiment The levels seen were no different from those of control untreated samples for AP-1 (Fig. 3c+d) This suggests that the ability of GR to modulate the DNA binding of NFr._B induced by TNFot is reduced compared to that of its ability to inhibit PMA-induced NFrd3 bindmg Thus TNFot may produce other factors that can sequester GR, limiting GR effects on NFr,.B binding
Vol. 55,
Transcription Factor Interactions in Lung
N o . 14, 1 9 9 4
1151
B
A
.--,_ ~ +150
o L-la • IL-2
>.+100
"o
"~c +100 E <[
+50
c_
& N
-50
I
C
2
3
Time
>, +125 +100 i ÷75
4
14 13 1211
24 D
IHrs)
÷50 "1"25
TNF/OEX SHAM
0
IL-1B
+150 +125
--~ +50 +25
TNF/DEX SHAM
0 • _¢
-25
13 12 11 t 0 9 -log(M)
TNFa
OEX
-SO
-25 -50 •
o
- 75
~t -100 f
, 2
Treatment
, 3
, 4
Z~
II~
2•
- 75 -IO0
. . . . 1
T i m e (hours)
2
Treatment
II-----1 3
•
2•
T i m e (hours)
Fi~. 3. Time and dose dependence of cytokine action on AP-1 and NFrB DNA binding in human peripheral lung. (a) IL-113 (0), IL-6 (V), and TNFI~(*) (10-11M) caused a transient rise in AP-1 binding which peaked at lh and returned to baseline levels at 24h. TNFtx (¢) and IL2 (A)(10-11M) induced a long-term induction of AP-1 activity. (b) Concentrationdependent increase in AP-1 activation following TNFtx and IL-113 incubation for lh. (n=26). (c) The effect of TNFct (10 -11M), in the absence or presence of dexamethasone (1 laM) on AP-1 activity over 24hr incubation. (d) Effect of TNFct (10-11M), in the absence or presence of dexamethasone (llaM) on NFr,_B activity over 24hr incubation. In all experiments 10pg human lung nuclear extract was used Values shown are the means + SEM of 6-12 observations Discussion Control of chronic inflammation is important in the treatment of several inflammatory conditions, including asthma and fibrosing lung diseases. The regulatory pathways that control chronic inflammation are complex and multifactorial. Glucocorticoids inhibit the inflammatory response of many cells to a wide variety of stimuli and so an understanding of its mode of action is of central importance Cytokines may be important mediators of chronic inflammation. Although the molecular pathways involved in cytokine action are not well understood, several cytokines act on responsive cells to induce the activation or synthesis of a few well characterised transcription factors The finding that both TNFct and the phorbol ester PMA induced a rapid-long lasting activation of both AP-1 and NFr.B in the human lung, and that the ability of these DNA binding
1152
Transcription Factor Interactions in Lung
Vol. 55, No. 14, 1994
proteins to bind to their target sequences is abolished by dexamethasone suggests that this may be an important molecular mechanism of steroids, that could account for their multiple antiinflammatory actions and for their efficacy in chronic inflammatory diseases Several mechanisms have been proposed to account for the interaction between transcription factors One factor could replace another at its site of action on the upstream regulatory site of the target genes, the two factors may compete for another factor at either the gene or the protein level, or their may be a direct interaction between one transcription factor (e.g GR) and other transcription factors at a protein-protein level (14) Direct protein-protein interaction is the more likely mechanism in these studies, at least m the short term, as AP-1 binding is reduced in the nuclear extract, although no reductions in the components of AP-1 are seen in the same extracts by western blotting until 24h (I Adcock, unpublished observations). Furthermore, the oligonucleotides used in these studies show no specific binding for other DNA binding proteins being highly selective for AP-1 and NFr,_B In the present studies the effect of actinomycin D is not addressed. Although is unlikely to affect the short term results, those for the 24hr time point may be influenced by the transcriptional activation of various genes by PMA, TNFct or dexamethasone on the components of the AP-1 complex. The different models proposed to account for the interaction between AP-1 and GR differ by virtue of their expected effect on A_P-1 DNA binding of GR activation With the tethering mechanism (4,15) proposed by Herrlich, an interaction between AP-1 and GR occurs without any effect on AP1 DNA binding. Indeed this group has analysed the AP-1 binding on the collagenase gene by Footprinting analysis and show no lack of protection following GR stimulation although the transcriptional readout was markedly reduced In contrast, the mutual repression model proposes that the interaction occurs in solution and both proteins block the DNA binding of the other In other studies investigating the interaction between GR and AP-1 Kerppola et al (16) again using the collagenase gene promotor region have shown a repression of AP-1 DNA binding by the DNA binding domain of the GR Furthermore, we have previously shown a reduction in GR binding to DNA in human lung following PMA treatment, suggesting that mutual repression can occur in this tissue (17) These results and the results presented in this study suggest that the exact interaction between GR and AP-I may depend upon the cell type used and the exact type of stimulus given and also suggest that more than one mechanism of interaction may occur at any one time. Some of the contradictions between the two models may be explained by the relative affinity of the GR for the Fos-Jun heterodimer and the Jun-Jun homodimer composing two forms of AP-1 family Interaction may occur between the GK and the Fos-Jun heterodimer that displaces the Fos-Jun dimer from the AP-1 site allowing the lower affinity Jun-Jun dimer to bind, since GR has a lower affinity for Jun-Jun homodimer than for Fos-~'un heterodimer (18). The Jun-Jun homodimer is less potent at activating transcription resulting in repression of gene transcription with no concomitant loss of AP-1 DNA site occupancy Therefore cell specificity of GR inhibition of AP-1 activity may be due to the presence of different endogenous AP-1 family proteins in different cell types. Yamamoto and colleagues (19) have proposed an alternative model for AP-1/GR interaction in the nucleus involving overlapping atypical GREs and TREs termed a composite GRE C-Jun homodimers interact with GR to cause a synergistic effect increasing transcription of responsive genes, whilst c-Fos/c-Jun heterodimers interact with DNA bound GR to switch off transcription Thus transcription rate is controlled by the level of c-Fos in the cell Glucocorticoids inhibit the synthesis of several cytokines, including IL-1 and TNFa (20,21). Although steroids may inhibit the inflammatory process at several sites, the direct interaction with transcription factors, such as AP-1 and NFrd3 may be of particular relevance, since this could provide a mechanism for switching off inflammation driven by cytokines known to activate these transcription factors.
Vol. 55, No. 14, 1994
Transcription Factor Interactions in Lung
1153
The results of our studies indicate that in human peripheral lung the transcription factors AP-1 and NFr.B are induced by phorbol esters and cytokines in a time and concentration dependent manner. Gel mobility shift assays indicate that the activated GR may interact with these transcription factors within the nucleus causing an inhibition of DNA binding by the respective proteins. In this way steroids may inhibit the synthesis chronic inflammatory mediators such as cytokines or inhibit the effects of agents that activate protein kinase C. The anti-inflammatory properties of glucocorticoids in lung could be explained if glucocorticoids repressed cytokine-induced AP-1 and NFr,.B mediated induction of mRNA coding for other cytokines, enzymes and receptor genes. Cross-coupling of distinct and opposing regulatory pathways provides another level of regulation in determining the inflammatory response of the cell This inhibition of transactivation may be an important molecular site of steroid action in asthma and other chronic inflammatory diseases
Acknowledgements CMG is an MRC Training Fellow. MJP is the Allen & Hanburys/Thoracic Society of Australia and New Zealand Travelling Fellow. We would like to thank Professor Sir Magdi Yacoub for providing transplant tissue This work was supported by The Medical Research Council and Glaxo Group Research
References 1 A.MUNCK, D.B MENDEL, L.I SMITH & E ORTI Am.Rev.Respw.Dls. 141 $2-S10. (1990). 2. P J.BARNES N.Engl.JMed. 321 1517-1527 (1989) 3. R.SCHULE, P.RANGARAJAN, S KLIEWER, L J.RANSOME, l BOLADO, N.YANG, I M.VERMA & R M EVANS Cell 6__2_21217-1226 (1990). 4. C.JONAT, H.J RAHMSDORF, K-K PARK, A C B.CATO, S GEBEL, H.PONTA & P.HERRLICH Cell 6._.221189-1204 (1990). 5. H-F.YANG-YEN, J-C.CHAMBARD, Y-L SUN, T.SMEAL, T J.SCHMIDT, J.DROUIN & M.KARIN Cell62 1205-1215 (1990) 6. D S LATCHMAN Blochem.J. 270 281-289 (1990) 7 J.KELLEYAm.Rev.Respir.Dts. 141 765-788 (1990) 8. G CAMUSSI, E.ALBANO, C TETTA & F.BUSSOLINO Eur.J.Btochem. 202 3-14 (1991). 9 K.J.TRACEY, H.VLASSARA& A.CERAMILancetl 1122-1126 (1989). 10. Y NISHIZUKA Science 233 305-312 (1986) 11 M.J LENARDO & D.BALTIMORE Cell 58 227-229 (1989) 12. J.AUW RX, B.STAELS & P.SASSONE-CORSI NucLActds Res. 18 221-228 (1990). 13. L OSBORN, S.KUNKEL & G J NABEL Proc.NatLAcad.Sci. (USA) 86 2336-2340 (1989). 14. M.E.JACKSONd.CellSci. 100 1-7 (1991) 15 H.KONIG, H PONTA, H J RAHMSDORF & P HERRLICH EMBO d. 1__!12241-2246 (1992) 16 T K KERPPOLA, D LUK & T.CURRANMoL Cell. Biol 13 3782-3791 (1993) 17 I M. ADCOCK, C M. GELDER, H SHIRASAKI, M YACOUB & P.J. BARNES Am. Rev. Respw. Dis. 145 A834 (Abstract)(1992) 18. L.SHEMSHEDINI, R KNAUTHE, P.SASSONE-CORSI, A.PORNON & H GRONMEYER EMBO J. 1__03839-3849 (1991) 19. M I.DIAMOND, J.N.MINER, S K YOSHINAGA & K.R YAMAMOTO Science 249 12661272 (1990). 20. B BEUTLER, N.KROCHIN, I.W.MILSARK, C LUEDKE & A.CERAMI Science 232 977980.(1986). 21. S.W.LEE, A-P.TSOU, H.CHAN, J.THOMAS, K PETRIE, E M EUGUI & A.C.ALLISON Proc.Natl.Acad.Sct.(USA) 8__551204-1208 (1988)