sion is active in MCF-7 cells. Keywords: nuclear factor KB ; CAAT/enhancer-binding protein; interleukin-6; gene expression ; cytokine. Interleukin-6 (IL-6) is a ...
Eur. J. Biochem. 2.9, 624-631 (1996) 0 FEBS 1996
Nuclear factor KB (NF-KB),nuclear factor interleukin-6 (NFIL-6 or CIEBPP) and nuclear factor interleukin-6P (NFIL6-P or CIEBPG) are not sufficient to activate the endogenous interleukin-6 gene in the human breast carcinoma cell line MCF-7 Comparative analysis with MDA-MB-231 cells, an interleukin-6-expressinghuman breast carcinoma cell line Laura FAGGIOLI', Chiara COSTANZO', Marcello MEROLA', Ercolina BIANCHINI ', Adriana FURIA2, Antonella CARSANA2 and Marta PALMIERI' ' Istituto di Chimica Biologica, Universiti di Verona, Verona, Italy Dipartimento di Chimica Organica e Biologica, Universiti di Napoli, Napoli, Italy (Received 22 January/22 March 1996) - EJB 96 0081/1
A comparative study of the molecular mechanism of interleukin-6 (IL-6) gene induction on two breast-carcinoma-derived cell lines has been performed. MDA-MB-231 cells produce constitutive detectable levels of both secreted 1L-6 and mRNA which, as expected. are dramatically enhanced following induction by either IL-Ip or tumor necrosis factor-a (TNF-a). The levels of both secreted IL-6 and IL-6 mRNA are significantly higher in response to IL-1p in spite of the fact that stimulation by TNF-a alone enhances the half life of IL-6 mRNA. The protein synthesis inhibitor cycloheximide is also a fairly strong inducer of IL-6 in these cells. In contrast, MCF-7 cells fail to produce detectable IL-6 protein or mRNA, even after stimulation with proper inducers. Analysis of transcription factors NF-KB, NFIL6 and NFIL6p, which have been described to be sufficient to activate the IL-6 gene in other cell systems, shows a similar pattern of expression in both MCF-7 and MDA-MB-231 cells. Furthermore, transfection of a recombinant plasmid carrying the IL-6 promoter linked to a luciferase reporter gene shows that both cell lines are able to drive IL-1p or TNF-a activation of this construction in a very similar manner. Finally, when MCF-7 cells were treated with IL-1p or TNF-a in the presence of cycloheximide, transcription of IL-6 mRNA from the endogenous IL-6 gene was observed. These data suggest that a mechanism of IL-6 gene repression is active in MCF-7 cells. Keywords: nuclear factor KB ; CAAT/enhancer-binding protein; interleukin-6; gene expression ; cytokine.
Interleukin-6 (IL-6) is a pleiotropic cytokine which plays a central role in the homeostasis of the immune system and in the modulation of acute-phase protein synthesis [ l , 21. Many different cell types such as monocytes, fibroblasts, keratinocytes, endothelial cells, T-cells, B-cells and some tumor cells, synthesize IL-6 either constitutively or following treatment with a proper inducer [3]. Accumulating evidence has suggested that deregulation of IL-6 expression is involved in the pathogenesis of various immunologically mediated diseases and malignancies such as rheumatoid arthritis, cardiac myxoma and myeloma [ I , 31. Under physiological conditions, the most potent inducers of IL-6 gene expression described to date are the cytokines interleukin-lp (IL-1p) and tumor necrosis factor-a (TNF-a). L I P , TNF-a and IL-6 form the group of alarm hormones as they are key regulators of the host defense with hormonelike activity and alter global carbohydrate, lipid and amino Correspondence to M. Palmieri, Istituto di Chimica Biologica, Universiti di Verona, Strada le Grazie 8, 1-37134 Verona, Italy Fax: + 39 45 8098170. Abbreviations. IL-6, interleukin-6; TL-18, interleukin-la; TNF-a, tumor necrosis factor-a; LPS, lipopolysaccharide; TPA, tumor-promoting agent 12-0-tetradecanoylphorbol-13-acetate; IFN-tr, interferon-a; IFN-8, interferon-/I; IFN-y, interferon-7.
acid metabolism. It is currently accepted that the biological effects of IL-lp and TNF-a are mediated, at least in part, by IL-6 [4]. The regulation of IL-6 gene expression occurs mainly at the tranrcriptional level, though mechanisms of post-transcriptional regulation have also been described [5-71. For all the inducers analysed to date, two responsive elements have been identified at positions -173 to -145 and -73 to -64 of the transcriptional regulatory region of the human IL-6 gene [8-131. The furthest upstream element contains a 14-bp consensus sequence for the binding of NFIL6 (CEBPP) [8, 141 and NFIL6p (C/ EBPS) [IS], two transcription factors belonging to the C/EBP family 1161. The truns-activation potential, as well as the expression of these latter transcription factors, are up-regulated in several cell lines by inducers such as IL-I/J, TNF-a, lipopolysaccharide (LPS) and IL-6 18, 14, 15, 17-21]. The responsive element at position -73 to -64 on the IL-6 promoter corresponds to a KB consensus sequence that has been shown to bind purified NF-KB [9], a transcription factor belonging to the Re1 family [25]. NF-KB is an heterodimer of SO-kDa (pSO/KBFl) and 65-kDa (p6S/Rel A) proteins which is activated by dissociation of the subunits from an inhibitor molecule (IKB) whose function is to prevent the translocation of the factor into the nucleus [22251. It has been shown that cotransfection of NFIL6 with the
Faggioli et al. (Eur: J. Biochern. 239)
NF-KB p65 subunit synergistically activates an IL-6 promoterreporter construct in the murine embryonic carcinoma cell line P19 [26), indicating that the two factors are sufficient to sustain the activation of the IL-6 gene. These results, however, cannot account for the diversity of tissue-specific expression of the IL-6 gene following treatment with IL-lp or TNF-a. Normal liver, in fact, does not express IL-6, while hepatocytes require active forms of NF-KB, NFIL6 and NFIL6p to regulate several genes of the acute-phase response [2, 27, 371. Furthermore, it has been shown that polymorphonuclear cells are not able to produce IL-6 following stimulation with LPS, although they can express the IL-8 gene which is under the control of the two transcription factors NF-KB and NFIL6 [28]. This behaviour would suggest an additional level of transcriptional control for the IL-6 gene. To investigate in more detail the differential transcriptional regulation of the IL-6 gene, we have chosen two breast carcinoma cell lines which show opposing behaviour in terms of IL-6 production. Analysis of secreted IL-6 by ELISA and IL-6 mRNA by Northern blotting have shown that, while MDA-MB231 cells produce high levels of IL-6 in response to all inducers tested, the MCF-7 cell line fails to produce both IL-6 protein and mRNA under the same experimental conditions. The activation of NF-KB, NFIL6 and NFIL6P following induction with TNF-a and IL-lp is similar in both cell lines, indicating that the pattern of transcription factors activated by the two inducers is similar in MDA-MB-231 as well as in MCF-7 cells. Moreover, in a transient expression assay, an IL-6 promoter-reporter construct is functional in both cell lines. Thus, it is likely that the lack of IL-6 gene expression in MCF-7 cells is due to transcriptional repression which depends on de novo protein synthesis, since the expression of the endogenous IL-6 gene is restored by treating the cells with cycloheximide in addition to IL-1p or TNF-a.
MATERIALS AND METHODS
Materials. Recombinant human TNF-a (specific activity
-- 2 X 1Oi to 2 X lo7 units/mg protein) and recombinant human IL-lp (specific activity =2X1Ox to 1X1o9 units/mg protein) were purchased from Genzyme. The IL-6 ELISA kit was from Amersham Corp. Oligonucleotide probes were synthesised with an Applied Biosystem oligosynthesiser. Polyclonal anti-p65 and anti-NFIL6P sera were purchased from Santa Cruz Biotechnology, anti-NFIL6 serum was kindly provided by Dr S. Akira and T. Kishimoto (Osaka, Japan). Cell lines. Human carcinoma cell lines MDA-MB-231 and MCF-7 were kindly provided by Dr del Senno (Ferrara, Italy). Both cell lines were propagated in Eagle MEM medium (GibcoBRL) containing 10% fetal calf serum, antibiotics and L-glutamine, in 5 % CO,. IL-6 determination. IL-6 production was measured by ELISA in cell-culture supernatants after treatment with IL-1P (10 ng/ml), TNF-a (10 ng/ml), LPS (1 pg/ml), tumor-promoting agent 12-0-tetradecanoylphorbol-13-acetate(TPA: 50 ng/ml), interferon-a (IFN-a; 100 U/ml), interferon$ (IFN-p; 100 U/ml) or interferon-? (IFN-y; 100 U/ml). RNA extraction and Northern-blot analysis. Total cellular RNA was prepared from lo7 cells by a one-step guanidinium thiocyanate/cesium chloride centrifugation method. For blot analysis, 25 pg total RNA samples were electrophoresed on 1% agarose gels containing formaldehyde and transferred onto nylon membranes. The filters were hybridized with IL-6 or p-actin cDNA probes labeled with 3zPby random priming (Amersham). mRNAs were detected and quantified by scanning filters with a
625
Molecular Dynamics PhosphorImager using ImageQuant software. Nuclear extracts and electrophoretic mobility shift assay (EMSA). Cells were induced with cytokines for 15 min at 37 “C prior to the preparation of nuclear extracts. The cells were then scraped, washed twice with 137 mM NaCI, 3 mM KCl, 9 mM Na,HPO,, 1 mM KH,PO,, pH 7.2 (NaClP,) and transferred to a 12-ml microfuge tube. Nuclear extracts were prepared according to the method of Osborn et al. [29], with minor modifications [i.e. lo7cells were resuspended in a volume of 50-80 p1 instead of 20 pl: nuclear proteins were extracted by 10 min ultracentrifugation at 30000Xg (Beckman table ultracentrifuge Optima TLX) and the supernatants were frozen undiluted instead of adding 3 vol. buffer D]. Protein concentration was measured by the Bradford method (Biorad), using BSA as standard protein. Binding reactions were carried out in a final volume of 20 pl using 8-10 pg nuclear extract in 20 mM Tris/HCl, pH 7.5, 100 mM KCl, 1 mM dithiothreitol, 1 mM EDTA, 1 mg/ml BSA, 0.1 % Nonidet P-40 4% glycerol containing 1-2 pg poly[(dIdC) . (dI-dC)] (Pharmacia). Radiolabeled probe (0.02 pmol : specific activity -- 3 X 10”cpdpmol) was added last to each reaction mixture and samples were incubated at room temperature for 30 min. In competition assays, a 100-200-fold molar excess of unlabeled double-stranded oligonucleotides was added to the reaction mixture. In the supershift experiments, antibodies were incubated for 1 hour on ice in the absence of radiolabeled probe prior to proceeding as described above. In order to establish the specificity of the antibody reaction, controls have been performed with specific antibody plus free probe or with irrilevant antibody. Samples were then loaded on a 5 % (30: 1.2) native polyacrylamide gel in 45 mM Tris base, 45 mM Sodium borate, 10 mM EDTA, pH 8.0 (0.5XTBE) and run at 150 V, dried, and exposed to Kodak films. The sequences of the oligonucleotides used in the study are shown below (only the sequence of the sense strand is indicated) : I L ~ - K B ,5’-AATGTGGGATTTTCCCATG-3’; Mut,1L6-~B,S’-AATGT&ATTTTCCCATG-3’; l g - ~ B ,5’CAGAGGGGACTTTCCGAGT-3‘. The consensus sequences for each factor are indicated in bold type, and mutations are underlined. Oligonucleotides were synthesized with 5’-ATG protruding ends on each extremity of both chains for labeling purposes. Hybridized oligonucleotides were labeled using Sequenase (USB) and ”P-labeled dATP and dCTP. Unincorporated nucleotides were removed by Sephadex G-50 column chromatography. DNA transfection assays. MDA-MB-231 and MCF-7 cell lines were transfected with 40 pg pIL6{ -592)Luc plasmid by the DNNcalcium phosphate coprecipitation technique. The plasmid was obtained by cloning the 5’ flanking region of the human IL-6 gene spanning from nucleotides -592 to +I0 into the luciferase reporter vector pGL2-Basic (Promega). Cells were routinely cultured as monolayers and plated at a cell density of 3X10h/10-cm plastic Petri dish. After 24 hours, cells were exposed to the DNAkalcium phosphate coprecipitate for 14 hours. The cultures were then washed extensively with serum-free medium and incubated for another 6 hours in serum containing medium alone, or in the presence of IL-lp or TNF-a. Cell lysates containing equal amounts of proteins were assayed for luciferase activity using the Luciferase Assay System (Promega) according to the instructions of the manufacturer. Western-blot analysis. Nuclear extracts were separated by electrophoresis on 12.5 % SDS/polyacrylamide gels, transferred onto nitrocellulose, and probed with the appropriate antibodies. ECL, a light emitting method (Amersham), was used for detection of immobilized specific antigens conjugated directly with horseradish-peroxidase-labeled antibodies.
Faggioli et al. (Eu,: J. Biochern. 239)
626
1
20001 I
1750-
15004 1500-
,"
1250-
1000X
+
750-
-C-
control
-A-
IL-IP TNF-a
0
6
12
18
24
30
36
42
48
Incubation time (h)
Fig. 1. Time course of secreted IL-6 by MDA-MB-231 cells in response to IL-lP or TNF-a. MDA-MB-231 breast carcinoma-derived cells were incubated in serum-containing medium alone (control, white boxes) or in the presence of IL-lfl (10 ng/ml, white triangles) or TNF-n (10 ng/ml, dark circles). At the indicated time intervals, aliquots of the supernatants were removed and the IL-6 concentration determined by ELISA. The values obtained were normalized to IL-6 secreted by lo5 cells.
RESULTS Cytokine regulation of IL-6 production in MDA-MB-231 and MCF-7 cells. The levels of secreted IL-6 by two human carcinoma cell lines MDA-MB-231 and MCF-7, untreated or stimulated with 1L-Ip or TNF-a, was assessed by ELISA on an aliquot of the culture medium. Either constitutively or after treatment with IL-lp, TNF-(x, or with several other inducers tested (IFN-n, IFN-P, IFN-y, TPA and LPS), MCF-7 cells fail to produce detectable IL-6 (data not shown). In contrast, MDA-MB-23 1 cells constitutively secrete IL-6 (approximately 30 pg/lOs cells). This basal release of TL-6 is dramatically induced by stimulation with TNF-G( (up to 1231 pg/105 cells) or IL-lP (up to 1907 pg/105 cells) (Fig. 1). IL-6 activity continues to accumulate with a similar pattern over long incubation periods (up to 45 hours) in the supernatant of both TNF-a-treated and IL-1B-treated cells.
Cytokine regulation of IL-6 mRNA accumulation in MDAMB-231 and MCF-7 cells. Northern-blot analysis performed on total RNA extracted from uninduced and IL-1P-treated or TNF-a-treated MCF-7 cells shows that IL-6 mRNA is undetectable in this cell line (data not shown). In MDA-MB-231 cells, IL-6 mRNA accumulated (Fig. 2). In agreement with the ELISA data, unstimulated cells constitutively express IL-6 mRNA which is strongly enhanced after treatment with cytokines. The pattern of IL-6 gene induction by IL-lp differs from that stimulated with TNF-n. TL-6 mRNA production induced by IL-1p shows a peak at 6 hours after treatment, followed by slow decrease. With TNF-a, two peaks of IL-6 mRNA are observed: the first at 1.5 h and the second about 24 hours after induction. Moreover, the IL-6 mRNA level obtained following TNF-a induction is about six-times lower than that observed with IL-1p. Differential effects of the two cytokines were also observed when the stability of IL-6 mRNA was determined in the presence of IL-1p or TNF-rx and actinomycin D over a period of two hours of induction. The turn-over of the IL-6 mRNA is about twice as fast in IL-1fl-induced cells (half-life about 1 hour) com-
pared to in TNF-a induced cells (half-life about 2 hours; Fig. 3). In unstimulated cells, the stability of the IL-6 transcript is similar to that determined in TL-1p-treated cells (half-life about 1 hour).
Cytokine induction of NF-KB, NFIL6 and NFIL6P in MDAMB-231 and MCF-7 cells. In order to establish whether the lack of expression of the IL-6 gene in MCF-7 cells might be due to the absence of the factors known to be responsible for the IL-6 gene expression, the presence of NF-KB, NFIL6 and NFIL6P in the two carcinoma cell lines was verified. EMSA were performed using an oligonucleotide corresponding to the NF-KB-binding site on the 5'-regulatory region of the IL-6 gene as a probe and nuclear extracts prepared from cells treated for various times with either IL-IB or TNF-a. Three protein-DNA complexes are observed i n both MDA-MB-231 and MCF-7 cell lines (Fig. 4) ; two higher molecular-mass complexes (BI and B2) are IL-lp/TNF-a inducible, whereas a lower molecular-mass complex (B3) is constitutive. The intensity of the bands corresponding to the three complexes does not change appreciably by increasing the time of induction. Competition with excess unlabeled oligonucleotide of wild-type probe, a mutant probe or the oligonucleotide corresponding to the NF-KBbinding site of the immunoglobulin K chain enhancer (IgK) shows that all three bands are the result of specific binding of nuclear proteins, but only bands B1 and B2 contain factors belonging to the Re1 family. Furthermore, a complex supershifting with antibodies raised against the p6.5 subunit of the NF-KB shows that both complexes B1 and B2 contain p65 (Fig. 5). Western-blot analysis of NFIL6 and NFIL6iI was performed on nuclear extracts obtained from MDA-MB-231 and MCF-7 cells before or after stimulation with IL-lp or TNF-a, in order to determine the pattern of expression of the two factors. A single band was detected using specific antibodies against either NFIL6 or NFIL6P (Fig. 6 a and b). This band corresponds to proteins of 2 4 3 kDa for NFIL6 and =29 kDa for NFIL6p and appears to be constitutive and non inducible in both cell lines.
627
Faggioli et al. (Eul: J. Biochenz. 239) 3501 I
-0-
300-
250
IL-lP
.--t TNF-a
-
200 -
150-
0
2
4
6
8
10
12
14
16
10
20
22
24
Incubation time (h)
Fig. 2. Time-course of IL-6 mRNA accumulation in MDA-MB-231 cells in response to IL-lP or TNF-a. MDA-MB-231 breast-carcinomaderived cells were grown in serum-containing medium supplied with IL-lP (10 ng/ml, white boxes) or TNF-o(l0 nglml, dark circles). At the indicated time intervals, total cellular RNA was prepared and Northern-blot analysis was performed using IL-6 and p-actin as probes. The amount of IL-6 mRNA was quantified by scanning filters with a phosphorimager and normalized relative to the level ofp-actin mRNA. Results are expressed as a ratio between IL-6 mRNA levels in induced and uninduced cells. The values shown are representative of four independent experiments.
100~
u noninduced
+ IL-lP 50
--f-
TNFa
10
0
1
2
3
4
Time after addition of actinomycin D (h)
Fig. 3. Stability of IL-6 mRNA levels in MDA-MB-231 cells cultured with or without IL-lP or TNF-a. MDA-MB-231 breast carcinoma cells were incubated in the presence or absence of L I B (10ng/ml) or TNF-u (10 ng/ml) for 15 hours before addition of actinomycin D (10 pg/ ml) to the cultures. After 1 , 2 and 4 hours of incubation, cells were collected, total cellular RNA was prepared and Northern-blot analysis was performed using IL-6 and p-actin as probes. The amount of 1L-6 mRNA was quantified by scanning filters with a phosphorimager and normalized relative to the level of p-actin mRNA. Values are expressed as actual mRNA levels relative to baseline mRNA levels ( i s . at time = 0). The values shown are representative of three independent experiments.
Activation of the exogenous IL-6 promoter in MDA-MB-231 and MCF-7 cells. In order to establish whether NF-KB, NFIL6 and NFIL6P were sufficient to activate an exogenous IL-6 pro-
moter in MCF-7 cells, a fusion gene was constructed by linking IL-6 5’ flanking DNA from positions -592 to + 10 to the luciferase reporter gene. The activation of this chimeric gene was tested by measuring luciferase activity in transient expression assays with MDA-MB-231 or MCF-7 cells uninduced or induced with IL-1P or TNF-a; the inducibility of the IL-6 promoter is similar in the two cell lines (Fig. 7).
Activation of the endogenous IL-6 gene by cycloheximide in MDA-MB-231 and MCF-7 cells. To determine whether the absence of transcription of the endogenous IL-6 gene in MCF-7 cells under uninduced or induced conditions requires de now protein synthesis, MCF-7 and MDA-MB-231 cells were treated with 20 pglml cycloheximide before induction with IL-IB or TNF-(1. Cycloheximide is able to prevent, with a similar efficiency, the inhibition of IL-6 gene transcription in MCF-7 cells in the presence of either IL-1p or TNF-a (Fig. 8B). The cycloheximide treatment of MDA-MB-231 cells enhanced both basal and IL-1P-induced or TNF-a-induced levels of IL-6 transcript by about sixfold and twofold, respectively (Fig. 8A).
DISCUSSION A key role in various aspects of the host defense system is played by cytokines, whose rapid induction is responsible for the inflammatory reaction [2]. An important aspect of this activity is the ability of cytokine-producing cells to suppress their own response to these effectors before any damage can occur beyond that required to eliminate the foreign organism o r repair the altered tissue [30]. in order to obtain this effect, both cytokines and their mRNAs have high turnover rates. The biosynthesis and secretion of cytokines is thus controlled predominantly at the transcriptional level [31-331. Alterations in the rate of inRNA translation and in the stability of proteins and mRNA may contribute only marginally to the control of the biosynthesis rate. Much of the recent work involving cytokines and, in particular, IL-6 transcriptional regulation has been devoted to activator
628
Faggioli et al. (Eui: J. Biochem. 239)
A
MDA-MB-231
MCF-7
Fig.4. NF-KB activation in MDA-MB-231 and MCF-7 cells. Nuclear extracts were prepared from MDA-MB-231 and MCF-7 breast carcinoma cell lines uninduced (NI) or induced with either IL-ID or TNF-ix for 15 min (A), 1 hour or 6 hours (B). EMSA were performed using "P-labeled I L - ~ - K Bprobe. The arrows indicate the position of the NF-KB and other retarded protein-DNA complexes. A 200-fold excess of a particular unlabeled competitor oligonucleotide, specified on the top left-hand part of the figure, was added to the binding mixture where indicated. The oligonucleotide Ig-rcR comprises the KB consensus sequence of the immunoglobulin K chain enhancer; Mut,IL6-/cB contains a GGG+AAT mutation that destroys the KB binding site (refer to Materials and Methods for oligonucleotides sequences). The gel was run for about 4 hours to allow an optimal separation of the bands. Under these conditions, labeled free probe is not included o n the gel.
proteins which are typically bound to the DNA upstream from the promoter. However, very little attention has been given to transcriptional repression. Indeed, transcriptional repression is the most likely mechanism which would explain the lack of IL-6 expression in normal liver 12, 271 or in polymorphonuclear cells (281, as both systems possess active forms of NFIL6 and NF-KB. To investigate in more detail 1L-6 gene activation, we have compared two human breast carcinoma cell lines for their ability to produce IL-6. We have shown that MDA-MB-231 cells produce high level of IL-6 following induction with IL-IP and TNF-a, while MCF-7 cells do no1 express any IL-6 under the experimental conditions tested. Both IL-1P and TNF-cx induce the production and secretion of IL-6 in MDA-MB-232 cells in similar ways, differing only in the amount of the IL-6 produced. In fact, we observed a stronger inducing effect of IL-lP compared to TNF-u, as already described in other cell systems [5, 341. In contrast, IL-6 message accumulated to different extents after treatment with IL-lP or TNF-cx, varying both i n its cellular concentration as well as its kinetics. IL-1P induced one peak of IL-6 mRNA, while two peaks of induction were observed with TNF-a. Similar results have been reported for human fibroblasts [34]. In this case, the authors have also observed that by treating TNF-cx-activated human fibroblasts with an excess of neutralizing anti-IL-I serum, only the second peak of IL-6 induction is
partially inhibited [34]. In our system, instead, we have observed that the amount of IL-lP transcript after treatment of the cells with TNF-a reinains at the basal level (data not shown). These observations suggest that the two cytokines act independently in the activation of the IL-6 gene. To date, however, no data are available which demonstrate different mechanisms of IL-6 gene activation by IL-lp or TNF-a; further studies will be necessary to elucidate this aspect. We also investigated the stability of IL-6 mRNA in uninduced and IL-l/I-induced or TNF-a-induced MDA-MB-231 cells. Our data suggest that IL-6 mRNA is apparently more stable in TNF-a-treated than in IL-latreated or uninduced MDA-MB-231 cells. Thus, regulation of IL-6 mRNA stability by IL-lB can be ruled out, although such a mechanism may occur in TNF-a-induced cells. However, the control of mRNA stability cannot be the sole factor responsible for the high level of IL-6 message accumulation relative to the control observed following treatment with TNF-a. The comparison between the two human breast carcinoma cell lines was extended to the cytokine-mediated activation of the three transcription factors NF-KB, NFIL6 and NFIL6/1, which are known to be responsible for the 1L-6 gene expression in other cell systems. No measurable difference in the expression of these transcription factors was observed between MDA-MB231 and MCF-7 cells in response to IL-1P or TNF-a induction. EMSA(s) performed on nuclear extracts from MDA-MB-231
Faggioli et al. ( E m J . Biochern. 239)
Fig. 5. Identification of NF-KB/Rel related proteins in ILQKB binding complexes by specific antibodies. Nuclear extracts were prepared from MDA-MB-231 and MCF-7 breast carcinoma cell lines either uninduced or induced with IL-1p or TNF-a for 15 min. EMSA were performed using "P-labeled IL-~-KBprobe. The arrows on the left of the figure indicate the position of the retarded complexes denominated in the same manner as in Fig. 4. Where indicated anti-p65 serum(np65) was used.
MDA-MB-231 nuclear extract TNF-a IL-lP hours
MCF-7 nuclear extract
- -
-
- -
+
+
-
+ +
0
1.5 3
1.5
3
0 1.5 3 1.5 3
-
-
-
+ +
+ +
- -
A
B Fig. 6. Western-blot analysis of NFIL6 and NFJL6P in MDA-MB231 and MCF-7 cells. Nuclear extracts were prepared from MDA-MB231 and MCF-7 breast carcinoma cell lines uninduced (NI) or induced with either IL-lp or TNF-a for 1.5 hours or 3 hours. Proteins were separated on a 12.5 % SDS/polyacrylamide gel, transferred to a nitrocellulose membrane and the two transcription factors revealed by anti-NFIL6 (a) or anti-NFIL6iI (b) sera.
and MCF-7 cells, using the IL-6 UB site as a probe, show a pattern of retarded complexes equivalent for the two cell lines. Interestingly, i n both extracts, a faster migrating complex is present that behaves as a non-Re1 protein in respect to its DNAbinding activity, as determined by competition with an excess of unlabeled Ig-K-B probe. It is not yet clear whether or not this factor plays a role in IL-6 gene expression and only preliminary data are available concerning its identity. The NFIL6 gene is known to encode three proteins (43, 36 and 18 kDa) translated in-frame from different initiation codons [35]. In our cell systems, we detected only the 43-kDa protein, which is thought to be an active form of the factor [36]. Moreover, NFIL6 as well as NFIL6j5' appear to be constitutively present in the nuclei of both cell lines and their levels do not seem to be induced in response to treatment of the cells with either IL-lp or TNF-a. While NFIL6 has already been described as a constitutive factor in both nucleus and cytoplasm whose endogenous level varies
629
Fig. 7. pIL-6(-592)-luciferase induction in MDA-MB-231 and MCF-7 cells by IL-lP and TNF-a. MDA-MB-231 and MCF-7 breast carcinoma cell lines were transiently transfected with 40 pg plasmid containing a region of 592 bp upstream to the start transcription site of the 1L-6 gene linked to the luciferase gene (see Materials and Methods). After 14 hours exposure to the DNAkalcium phosphate coprecipitate, cells were incubated at 37 "C for 6 hours in the presence of IL-1p (1 0 ng/ ml) or TNF-a (10 ng/ml) in growth medium. After extensive washing with NaCUP,, cells were lysed with 200 p1 luciferase lysis buffer for 15 min at room temperature. The lysate was then scraped and centrifuged to eliminate cellular debris. Aliquots of the supernatant were then assayed for luciferase activity and protein concentration. Transfections were repeated at least three times. The values shown are representative of at least three independent transfections.
according to cell type [14, 17, 371, previous reports refer to NFIL6p as undetectable in non-induced cells and transcriptionally induced following proper stimulation [15, 371. For this latter factor, the difference we observed may be due to the cell types used in this study. An unexpected result was obtained from transient expression assays in MDA-MB-231 and MCF-7 cells, performed with a recombinant plasmid carrying the luciferase reporter gene under the control of the IL-6 promoter. We show that the activation level of the exogenous promoter by IL-1p or TNF-a is similar in the two human carcinoma cell lines, suggesting that the array of transcription factors that normally activates the IL-6 promoter is equally functional in MCF-7 cells as in MDA-MB-231 cells. Additionally, MCF-7 cells were able to transcribe IL-6 mRNA from the endogenous IL-6 gene when treatment with IL-1p or TNF-a was performed in the presence of cycloheximide. Furthermore, the production of the IL-6 transcript in these conditions was inhibited by the addition of actinomycin D (data not shown), suggesting that in MCF-7 cells a control of IL-6 expression by cycloheximide and L I B or TNF-a may occur at the level of transcription. These results would suggest that in MCF-7 cells the chromatin surrounding the 1L-6 gene is in a repressed state which can be partially relieved by the action of the cycloheximide. In contrast, the presence of a soluble titratable repressor is unlikely since the repression of the endogenous IL-6 gene is not relieved by transfection of the 592 bp upstream to the start transcription site of the gene (data not shown). It has to be pointed out, however, that as the actinomycin D experiment provides an indirect evidence for the action of cyclo-
630
Faggioli et al. (EUKJ. Biochem. 239)
REFERENCES
Fig.8. IL-6 gene activation in MDA-MB-231 and MCF-7 cells following treatment with cycloheximide. MDA-MB-231 (A) and MCF-7 (B) breast carcinoma cell lines were incubated for 15 min alone (NI), or in medium supplemented with 20 pg/ml cycloheximide (CHX). Where indicated, IL-lp or TNF-a were added to the medium and cells were further incubated for 1 .S hours. Total cellular RNA was prepared and Northern-blot analysis was performed using IL-6 and p-actin cDNAs as probes.
heximide at the level of transcription, the possibility remains that an other regulatory mechanism occurs. In MDA-MB-231 cell line, cycloheximide is able to enhance the accuinulation of the IL-6 transcript even in the absence of other inducers. Since in the presence of actinomycin D, cycloheximide has only a minor effect on 1L-6 mRNA accumulation in MDA-MB-231 cells (unpublished data), transcriptional control of IL-6 gene expression by cycloheximide can be suggested. Cycloheximide activation might depend on 11cBa synthesis inhibition. In fact, it has recently been reported that in unstimulated cells of lymphoid and epithelial lineages, IKB~xprotein has a relatively short half-life of approximately 30 min 1381. In contrast, the presence in unstimulated MDA-MB-232 cells of a labile repressor acting at the DNA level can be hypothesized. However, the level of induction obtained with cycloheximide is about twofold lower compared to that obtained with TNF-a, suggesting that the mechanism of derepression could consist in the removal of the inhibitor blocking the activator rather then removal of a repressor molecule directly binding DNA. This hypothesis is supported by the fiiiding that both NFIL6 and NFIL6D are constitutively present in MDA-MB-231 cells. To our knowledge, our results provide the first demonstration that a pattern of activators able to induce an 1L-6 promoterreporter construct does not activate a functional IL-6 endogenous gene. It is tempting to speculate that, in MCF-7 cells, as well as in other more physiological systems, chromatin organization plays an important role in IL-6 expression. We are at present investigating the possibility to test this hypothesis by experiments of DNase hypersensitivity and in vivo footprinting. We are indebted Lo Prof. M. Libonati for his constant support and encouragement throughout this study and thank Dr P. Moore for critical reading of the manuscript. This work was partially supported by the C.N.R. (Target Project on Biotechnology and Bioinstrumentation) and C.I.S.M.I.
1. Hirano, T. & Kishimoto, T. (1992) 46th Forum in immunology: Molecular biology and immunology of interleukin-6, R e x Irnmunol. 143, 722-783. 2. Baumann, H. & Gauldie, J. (1994) The acute phase response, ltnmun. Today 1.5, 74-80. 3. Akira, T., Taga, T. & Kishimoto, T. (1993) Interleukin-6 in biology and medicine, Adv. frnmun. 54, 1-78. 4. Beg, A. A,, Finco, T. S., Nantermet, P. V. & Baldwin, A. S. Jr (1993) Tumor necrosis factor and interleukin-1 lead to phosphorylation and loss of IkBa: a mechanism for NF-kB activation, Mol. Cell. B i d . 13, 3301-3310. 5 . Elias, J. A. & Lentz, V. (1990) 1L-1 and Tumor necrosis factor synergistically stimulate fibroblast IL-6 production and stabilize IL-6 messenger RNA, J . Irnmunol. 145, 161-166. 6. Le, P. T., Lazorick, S . , Whichard, L. P., Haynes, B. F. & Singer, K. H. (1991) Regulation of cytokine production in the human thymus: Epidermal growth factor and Transforming growth factor a regulate mRNA levels of interleukin l a (IL-la), IL-lp, and 1L-6 in human thymic epithelial cells at a post-transcriptional level, J. Exp. Med. 174, 1147-1157. 7. Rola-Pleszczynski, M. & Stankova, J. (1992) Leukotriene B, enhances interleukin-6 (IL-6) production and IL-6 messenger RNA accumulation in human monocytes in vitru: transcriptional and posttranscriptional mechanisms, Blood 80, 1004- 1011. 8. Isshiki, H., Akira, S., Tanabe, O., Nakajima, T., Shimamoto, T., Hirano, T. & Kishimoto, T. (1990) Constitutive and interleukin-I (It,-l)-inducible factors interact with the IL-1-responsive element in the IL-6 gene, Mol. Cell. Biol. 10, 2757-2764. 9. Libermann, T. A. & Baltimore, D. (1990) Activation of interleukin6 gene expression through the NF-KB transcription factor, Mol. Cell. Biol. 10, 2327-2334. 10. Ray, A,, Tatter, S . B., May, L. T. & Sehgal, P. B. (1988) Activation of the human ‘p2-interferon/hepatocyte-stimulatingfactorhnterleukin-6’ promoter by cytokines, viruses and second messenger agonists, Proc. Nut/ Acad. Sci. USA 85, 6701 -6705. 11. Ray, A., Sassone-Corsi, P. & Sehgal, P. B. (1989) A multiple cytokine- and second messenger-responsive element in the enhancer of the human interleukin-6 gene: similarities with c-fos gene regulation, Mol. Cell. B i d . 9, 5531-5547. 12. Shimizu, H., Mitomo, K., Watanabe, T., Okamoto, S. & Yamamoto, K. (1990) Involvement of a NF-iiB-like transcription factor in the activation of the interleukin-6 gene by inflammatory lymphokines, Mol. Cell. Biol. 10, 561-568. 13 Zhang, Y., Lin, J. & Vilcek, J. (1990) Interleukin-6 induction by tumor necrosis factor and interleukin-I in human fibroblast involves activation of a nuclear factor binding to a KB-like sequence, Mol. Cell. B i d . 10, 3818-3823. 14. Akira, S., Isshiki, H., Sugita, T., Tanabe, O., Kinoshita, S., Nishio, Y., Nakajima, T., Hirano, T. & Kishimoto, T. (1990) A nuclear factor for IL-6 expression (NF-IL6) is a member of a C E B P family, EMBU J. 9, 1897-1906. 15. Kinoshita, S., Akira, S. & Kishimoto, T. (1992) A member of the C E B P family, NFIL6/j, forms a heterodimer and transcriptionally synergizes with NFIL6, Pro(.. Natl Acad. Sci. USA 89, 14731476. 16. Hurst, H. C. (1994) Transcription factors 1 : bZIP Proteins, Protein Profile 1, 123-168. 17. Metz, R. & Tiff, E. (1991) CAMPstimulates the C/EBP-related transcription factor rNFIL-6 to trilns-locate to the nucleus and induce c-fos transcription, Genes & Dev. 5, 1754-1766. 18. Nakajima, T., Kinoshita, S., Sasagawa, T., Sasaki, K., Naruto, M., Kishimoto, T. & Akira, S. (1993) Phosphorylation at threonine235 by a ras-dependent mitogen-activated protein kinase cascade is essential for transcription factor NF-IL6, Proc. Natl Acad. Sci. USA 90, 2207-221 1. 19. Trautwein, C., Caelles, C., Van der Geer, P., Hunter, T., Karin, M. & Chojkler, M. (1993) Transactivation by NF-ILhILAP is enhanced by phosphorylation of its activation domain, Nature 364, 544547. 20. Wegner, M., Cao, Z. & Rosenfeld, M. G. (1992) Calcium-regulated phosphorylation within the leucine zipper of CEBPP, Science 256. 370-373.
Faggioli et al. (Eur. J. Biochem. 239) 21. Ray, B. K. & Ray, A. (1994) Expression of the gene encoding nlacid glycoprotein in rabbit liver under acute-phase conditions involves induction and activation of /3 and 6 CCAAT-enhancer-binding proteins, Eur. J. Biochem. 222, 891-900. 22. Beg, A. A. & Baldwin, A. S. Jr (1993) The IKB proteins: multifunctional regulators of ReVNF-icB transcription factors, Genes & Dev. 7, 2064-2070. 23. Liou, H.-C. & Baltimore, D. (1993) Regulation of the NF-KB/rel transcription factor and IKB inhibitor system, Cell Biol. 5, 477487. 24. Baeuerle, I? A. (1991) The inducible transcription activator NF-KB: regulation by distinct protein subunits, Biochim. Biophys. Acfa 1072, 63-80. 25. Grilli, M., Chiu, J. J.-S. & Lenardo, M. J. (1993) NF-KB and Rel: partecipants in a multiform transcriptional regulatory system, Znt. Rev. Cyt. 143, 1-62. 26. Matsusaka, T., Fujikawa, K., Nishio, Y.,Mukaida, N., Matsushima, K., Kishimoto, T. & Akira, S. (1993) Transcription factors NFIL6 and NF-KB synergistically activate transcription of the inflammatory cytokines, interleukin 6 and interleukin 8, Proc. Nut1 Acad. Sci. USA YO, 10193-10197. 27. Kishimoto, T., Taga, T. & Akira, S. (1994) Cytokine signal transduction, Cell 76, 253-262. 28. Wang, P., Wu, P., Anthes, J. C., Siege], M. I., Egan, R. W. & Billah, M. M. (1994) Interleukin-I0 inhibits interleukin-8 production in human neutrophils, Blood 83, 2678-2683. 29. Osborn, L., Kunkel, S. & Nabel, G. J. (1989) Tumor necrosis factor a and interleukin 1 stimulate the human immunodeficiency virus enhancer by activation of the nuclear factor KB, Proc. Nufl Acud. Sci. USA 86. 2336-2340.
631
30. Arai, K., Lee, F., Miyajima, A,, Miyatake, S., Arai, N. & Yokota, T. (1990) Cytokines: coordinators of immune and inflammatory responses, Annu. Rev. Biochem. 59, 783 - 836. 31. Sehgal, P. B., Grieninger, G. & Tosato, G. (1988) Regulation of the acute phase and immune response: interleukin-6, Ann. NY Acud. sci. 557, 1-583. 32. Camussi, G., Albano, E., Tetta, C . & Bussolino, F. (1991) The molecular action of Tumor necrosis factor-alpha, Eur. J. Biochem. 202, 3-14. 33. Fenton, M. J. (1992) Review: transcriptional and post-transcriptional regulation of interleukin 1 gene expression, fnt. J . fmmunopharmac. 14, 401-411. 34. Ng, S. B., Tan, Y. H. & Guy, G. R. (1994) Differential induction of the interleukin-6 gene by Tumor necrosis factor and interleukin-1, J. Biol. Clzem. 269, 19021-19027. 35. Hsu, W. & Chen-Kiang, S. (1993) Convergent regulation of NF-IL6 and Oct-1 synthesis by interleukin-6 and retinoic acid signaling in embryonal carcinoma cells, Mol. Cell. Biol. 13, 2515-2523. 36. Descombes, P. & Schibler, U. (1991) A liver-enriched transcriptional activator protein, LAP, and a transcriptional inhibitory protein, LIP, are translated from the same mRNA, Cell 67, 569-579. 37. Ramji, D. P., Vitelli, A., Tronche, F., Cortese, R. & Ciliberto, G. (1993) The two C E B P isoforms, IL-6BPNF-IL6 and ClebplilNFIL6/3, are induced by IL-6 to promote acute phase gene transcription via different mechanisms, Mucleic Acid Res. 21, 289294. 38. Rice, N. R. & Ernst, M. K. (1993) In vivo control of NF-KB activation by k B a , EMBO J. 12, 4685-4695.
