Transforming Growth Factor-β1 Induces Interleukin-6 Expression via Activating Protein-1 Consisting of JunD Homodimers in Primary Human Lung Fibroblasts*

Transforming growth factor (TGF)-β1 induces extracellular matrix deposition and proliferation of mesenchymal cells. We recently reported that interleukin (IL)-6 is an essential mediator of growth factor-induced proliferation of lung fibroblasts. Here, we demonstrate by reverse transcriptase polymerase chain reaction and enzyme-linked immunoassay that TGF-β1 is a potent inducer of IL-6 mRNA and protein in primary human lung fibroblasts. Transient transfections of fibroblasts with a luciferase reporter gene construct containing nucleotides −651 to +1 of the human IL-6 promoter revealed that TGF-β1 also potently activated IL-6 promoter activity. Progressive 5′-deletions and site-directed mutagenesis of the parental construct located the TGF-β1-responsive cis-regulatory element to a known activating protein-1 (AP-1) sequence (nucleotides −284 to −276). Gel shift analyses revealed that AP-1 DNA binding activity in nuclear extracts was increased 30 min after stimulation with TGF-β1. In contrast, neither CCAAT enhancer-binding protein-β, NF-κB, nor Sp1 were activated by TGF-β1. Supershift analyses demonstrated that the AP-1 complex induced by TGF-β1 was composed of Jun isoforms and absent of Fos isoforms. Moreover, this complex was found to be a JunD homodimer. Our data thus demonstrate that TGF-β1 is a potent inducer of IL-6 in primary human lung fibroblasts. The TGF-β1-activated JunD homodimer may be essential for a majority of the biological effects induced by TGF-β1 in this cell type, such as proliferation and extracellular matrix synthesis.

The transforming growth factor (TGF) 1 -␤ isoforms play an essential role during organ development and tissue homeostasis. Biological effects induced by TGF-␤s are highly cell type-and tissue-specific, depending on the state of differentiation of a given cell type (1)(2)(3)(4). The TGF-␤ superfamily consists of more than 20 members, most important of which are the TGF-␤s themselves, activins, and bone morphogenic proteins. Among the five described TGF-␤ isoforms, TGF-␤1 is most abundant in human tissues (5,6). Signal transduction of TGF-␤1 requires ligand binding to TGF-␤ receptors (T␤R). Three T␤R have been characterized (T␤RI to T␤RIII), T␤RI and T␤RII belong to the family of serine/threonine kinase receptors. Upon ligand binding to T␤RII, T␤RII recruits T␤RI to form a heterooligomeric ligand-receptor complex. T␤RI is then phosphorylated at glycine and serine residues by the constitutively active kinase T␤RII, and downstream signaling is initiated (6,7). Studies investigating the cis-and trans-acting elements of TGF-␤-responsive genes have so far identified several transcription factors directly involved in TGF-␤ signaling, such as Sp1 (8,9), AP-1 (10,11), CTF/NF-1 (12), and SMAD proteins (13)(14)(15).
In mesenchymal derived cells such as fibroblasts, TGF-␤s have two major effects, (i) an increase in extracellular matrix deposition and (ii) an increase in cellular proliferation (16,17). The proliferative response of fibroblasts to mitogenic stimuli, such as TGF-␤1, is preceded by increased DNA synthesis. In this respect, we have previously shown that growth factorinduced proliferation of lung tumor cell lines and primary human lung fibroblasts is essentially mediated by interleukin (IL)-6 (18,19). Down-regulation of IL-6 levels by antisense treatment resulted in abrogation of the proliferative response of fibroblasts to mitogenic stimuli, whereas the response was not affected by antisense oligonucleotides against other interleukins (18). IL-6 expression is known to be affected by a variety of cytokines and growth factors. Up-regulation of IL-6 is reportedly controlled by the activity of several transcription factors with known consensus sequences in the IL-6 promoter region, including AP-1, C/EBP-␤, and NF-B (20,21). In a different cell culture model, TGF-␤1 has been reported to up-regulate IL-6 expression (22,23); however, the molecular mechanism of increased IL-6 expression in response to TGF-␤1 is unknown.
This study addresses the question how IL-6 expression is regulated by TGF-␤1 in primary human lung fibroblasts. We found that upon stimulation with TGF-␤1, IL-6 is rapidly upregulated at the mRNA and protein level. This genomic effect of TGF-␤1 is reflected at the promoter level, demonstrating rapid activation of AP-1, as shown by luciferase and electrophoretic mobility shift assays. The TGF-␤1-induced AP-1 complex is found to be a JunD homodimer, while Fos proteins are absent in this complex. Thus, our data provide new insight into the signaling pathway induced by TGF-␤1, a major growth factor due to its prominent biological effects during induction of fibrosis, in primary human lung fibroblasts.
Cell Culture-Primary human cell lines of fibroblasts (n ϭ 10) were established from human lung tissue biopsies obtained from patients undergoing lobectomy or pneumectomy, as described previously (18,24). Tissues were cut in small slices, placed in cell culture flasks precoated with fetal calf serum, and incubated for 1 week in RPMI 1640 supplemented with 10% fetal calf serum, 8 mM L-glutamine, and 20 mM HEPES. Fibroblasts had regularly grown out from the tissue slices after 1 week and were passaged by standard trypsinization. For all experiments, cells were used before their seventh passage. Fibroblasts were plated onto 150-mm cell culture dishes until confluent. Cells were then starved in low serum medium (0.1% fetal calf serum) for 24 h prior to stimulation with TGF-␤1 at the indicated concentrations. No antibiotics or antimycotics were added to the culture conditions at any time.
RT-PCR-Total RNA was extracted from TGF-␤1-stimulated or control lung fibroblasts (3 ϫ 10 7 cells) with Trizol reagent as described (19). RNA was suspended in 30 l of sterile deionized water, and RNA concentrations were determined spectrophotometrically at 260 nm. All RNA preparations had an A 260 /A 280 ratio of Ͼ1.75. Aliquots of 1 g of total RNA were transcribed into cDNA at 42°C for 15 min in a total volume of 20 l containing 20 mM Tris HCI, pH 8.3, 50 mM KCl, 5 mM MgCl 2 , 1 mM of each dNTP, 1 unit of RNase inhibitor, and 2.5 units of murine leukemia virus reverse transcriptase. PCRs were then performed with 10 l of the reverse transcription reactions for amplification of IL-6 and ␤-actin cDNA. Amplifications were performed in a total volume of 50 l containing 0.5 units of Taq polymerase and 15 pmol of primers specific for IL-6 (5Ј-GCCCAGCTATGAACTCCTTCTC-3Ј and 5Ј-GAGTTGTCATGTCCTGCAGCC-3Ј) or ␤-actin (5Ј-GTACGTTGC-TATCCAGGCTGTGC-3Ј and 5Ј-TCAGGCAGCTCGTAGCTCTTCTC-3Ј). Amplifications were performed with 25 cycles for ␤-actin and 35 cycles for IL-6 cDNA. The amplification profile included denaturation at 98°C for 15 s, primer annealing at 62°C for 15 s, and extension at 72°C for 30 s, followed by a final extension at 72°C for 5 min. 12 l of each PCR were analyzed by agarose gel electrophoresis on 3.0% agarose gels.
Enzyme-linked Immunoassay-Concentrations of IL-6 protein secreted into culture supernatants of fibroblasts were determined by enzyme-linked immunoassay (Amersham Pharmacia Biotech). Cells were plated onto 24-well plates until confluent and starved in 0.1% fetal calf serum for 24 h. Fibroblasts were stimulated with different concentrations of TGF-␤1 as indicated, and aliquots of the supernatants were removed at several time points. Measurements of IL-6 were carried out according to the manufacturer's instructions.
Plasmid Construction, 5Ј-Deletion Constructs, and Site-directed Mutagenesis-A plasmid containing a 651-bp fragment of the human IL-6 gene promoter located directly upstream of the transcriptional start site was kindly provided by Shigeru Katamine (Nagasaki University, Nagasaki, Japan) (25). The 651-bp insert was subcloned (5Ј-KpnI, 3Ј-XhoI) into pGL3 basic luciferase reporter gene vector to give the parental pIL6-luc651 construct. From this parental construct, 5Ј-deletion mutants were made by deleting fragments containing the consensus sequence for transcription factor AP-1 (pIL6-luc220) or CREB (pIL6-luc160) using internal restriction sites for NheI or AatII, respectively.
Cell Transfection and Luciferase Assays-Two days before transfec-tion, fibroblasts were seeded into 24-well plates (1 ϫ 10 4 cells/well) precoated with 1% gelatin. After 24 h, cells were serum-deprived for 24 h and subjected to liposomal transfection using the cationic lipid Tfx-50 at a DNA:lipid ratio of 1:3 (1 g of plasmid/well) for 2 h. Cells were then overlaid with low serum medium with or without TGF-␤1 at the indicated concentrations. After 36 h of incubation, cells were harvested by active lysis, and equal amounts of lysates were analyzed for firefly luciferase expression according to the directions of the manufacturer. In brief, 20-l aliquots of cell lysates were mixed with 100 l of luciferase reagent buffer, and luminescence of the samples was integrated over a time period of 10 s in a LUMAC Biocounter M1500P (Landgraaf, The Netherlands). As an internal control for transfection efficiency, expression plasmids encoding Renilla luciferase driven by the thymidine kinase promoter were used (0.4 g/well). Firefly and Renilla Luciferase have distinct substrate properties, and thus activity of both enzymes can be assessed in the same sample using two substrates sequentially (Dual Luciferase Assay, Promega). Preparation of Cytosolic and Nuclear Extracts-Nuclear and cytosolic extracts from TGF-␤1-stimulated or -unstimulated control cells were prepared for gel shift analyses at the indicated time points. Cells were washed twice in ice-cold phosphate-buffered saline and harvested in 1 ml of phosphate-buffered saline with a rubber policeman. Samples were centrifuged for 1 min at 6,000 ϫ g (4°C), and the resulting cell pellets were resuspended in 100 l of low salt buffer (20 mM Hepes, pH 7.9, 10 mM KCl, 0.1 mM NaVO 4 , 1 mM EDTA, 1 mM EGTA, 0.2% Nonidet P-40, 10% glycerol, supplemented with a set of proteinase inhibitors, Complete TM ). After 10 min of incubation on ice, the samples were centrifuged at 13,000 ϫ g for 2 min (4°C), and the supernatants (cytosolic extracts) were immediately frozen in a dry ice/ethanol bath. Pelleted nuclei were resuspended in 60 l of high salt buffer (20 mM Hepes, pH 7.9, 420 mM NaCl, 10 mM KCl, 0.1 mM NaVO 4 , 1 mM EDTA, 1 mM EGTA, 20% glycerol, supplemented with Complete TM ), and nuclear proteins were extracted by shaking on ice for 30 min. Samples were centrifuged at 13,000 ϫ g for 10 min (4°C), and the supernatants were taken as nuclear extracts.
Statistical Analysis-All data were obtained from at least four different cell lines of primary human lung fibroblasts. Enzyme-linked immunoassay and RT-PCR were performed in duplicate using three independent sets of experiments. For statistical analysis, Student's t test and analysis of variance were performed. p Ͻ 0.01 was estimated significant.
Ethics Committee Approval-The protocol for establishing primary human cell cultures from biopsies obtained during lung surgery was approved by the Ethics Committee of the School of Medicine, University of Basel, Switzerland (approval number M75/97).

RESULTS
TGF-␤1 Increases IL-6 mRNA and Protein Synthesis in Primary Human Lung Fibroblasts-We analyzed whether human lung fibroblasts, in culture, up-regulate IL-6 expression in response to TGF-␤1 by qualitative RT-PCR. Fig. 1A demonstrates a characteristic agarose gel of RT-PCR products amplifying IL-6 and ␤-actin mRNA from cells treated with TGF-␤1 at different concentrations for 8 h. The expected size of IL-6 mRNA was 565 bp, and the expected size of the constitutive message for ␤-actin serving as an internal control was 225 bp. RT-PCR indicated that human lung fibroblasts up-regulated IL-6 mRNA in response to TGF-␤1 (0.01-1.0 ng/ml) in a concentration-dependent manner (Fig. 1A). As shown in Fig. 1A, IL-6 mRNA was clearly up-regulated as compared with control levels 8 h after stimulation with TGF-␤1.
RT-PCR, although not performed quantitatively, thus indicated that TGF-␤1 led to a dose-dependent increase in IL-6 message. We therefore analyzed whether the increased IL-6 mRNA levels coincided with enhanced IL-6 secretion. IL-6 protein levels in culture supernatants of lung fibroblasts treated with TGF-␤1 for 12 h were measured by enzyme-linked immunoassay. As shown in Fig. 1B, the levels of IL-6 protein were similarly increased in a concentration-dependent manner as compared with controls. Unstimulated quiescent human lung fibroblasts secreted 260 Ϯ 13 pg/ml of IL-6 protein within 12 h. IL-6 secretion increased to 886 Ϯ 23 pg/ml in response to TGF-␤1 at a concentration of 1 ng/ml (p Ͻ 0.001), corresponding to a 341% increase. When lung fibroblasts were stimulated with TGF-␤1 at concentrations of 0.1 or 0.01 ng/ml, IL-6 protein concentrations in culture supernatants increased to 723 Ϯ 43 pg/ml (278% increase) and 435 Ϯ 23 pg/ml (167% increase), respectively (p Ͻ 0.001) (Fig. 1B).
In order to identify cis-regulatory sequences that were responsible for the up-regulation of IL-6 expression by TGF-␤1, we generated several deletion mutants of the parental pIL6-luc651 construct. Two deletion mutants were screened for their TGF-␤1 inducibility and compared with the parental pIL6-luc651: (i) pIL6-luc220, deficient in a known AP-1 consensus sequence and (ii) pIL6-luc160, deficient in both a known AP-1 and a known CREB consensus sequence (see schematic diagram in Fig. 2B). As demonstrated in Fig. 2B, IL-6 promoter inducibility by TGF-␤1 was only observed with the parental pIL6-luc651, suggesting that the TGF-␤1-responsive cis-regulatory sequences are located between nucleotides Ϫ651 to Ϫ220 . Fibroblasts were plated into 24-well plates at a density of 5 ϫ 10 4 cells/well and cotransfected with pIL6-luc651 and pRL-TK (encoding Renilla luciferase driven by the thymidine kinase promoter as an internal control) by lipofection for 2 h. The DNA/lipid ratio used for all experiments was 1:3. Transfected cells were stimulated with the indicated amounts of TGF-␤1, harvested after 24 h, and processed for luciferase assays. RLU representing luciferase activity were measured over 10 s in a luminometer and corrected for transfection efficiency using the readout for Renilla luciferase. Data represented were obtained from four independent experiments and are representative for four different cell lines. B, deletions of the parental pIL6-luc651, as well as mutants for the indicated consensus sites, were generated as described under "Experimental Procedures." Constructs were transiently transfected into human lung fibroblasts along with the pRL-TK as described above. Induction of promoter activity by TGF-␤1 was assessed for the respective constructs, corrected for transfection efficiency, and calculated as percentage of induction compared with unstimulated base-line activity. Measurements were done in triplicate and are representative of four different cell lines. CRE, cAMP response element.
of the IL-6 promoter.
To further analyze this observation, we generated mutants of known consensus sequences within the IL-6 promoter by sitedirected mutagenesis (Fig. 2B). We tested the parental pIL6-luc651 with mutations in the AP-1 binding sequence, the C/EBP-␤ binding sequence, and the NF-B binding sequence as well as a double mutant of the C/EBP-␤ and NF-B sequence for their TGF-␤1 inducibility. Interestingly, only the AP-1 mutant was found to lack TGF-␤1 inducibility compared with the parental pIL6-luc651 (Fig. 2B). Neither the C/EBP-␤, the NF-B, nor the double mutant demonstrated loss of TGF-␤1 inducibility as compared with the parental pIL6-luc651. Thus, the results obtained from luciferase assays strongly suggest that AP-1 is specifically required for TGF-␤1-induced IL-6 gene expression in primary human lung fibroblasts.
AP-1 Is Specifically Activated by TGF-␤1-We assessed whether AP-1 is specifically activated by TGF-␤1 by performing electrophoretic mobility shift analyses (EMSA) with nuclear extracts of TGF-␤1-treated and -untreated control fibroblasts. Time course analyses with ␥-32 P-labeled oligonucleotides spanning the consensus sequences of transcription factors important for IL-6 promoter activity were performed. Fig. 3 includes characteristic EMSA demonstrating time course analyses of the transcription factors AP-1 (Fig. 3A), C/EBP-␤ (NF-IL-6) (Fig. 3B), and NF-B (Fig. 3C) in nuclear extracts of fibroblasts treated with TGF-␤1 at 2 ng/ml. Over a representative time course, only AP-1 DNA binding activity in nuclear extracts of TGF-␤1treated fibroblasts was significantly increased (Fig. 3A); none of the aforementioned transcription factors were activated (Fig.  3, B and C). AP-1 activation increased as early as 30 min after stimulation with TGF-␤1, peaked at 5 h, and returned to baseline levels over a time frame of 16 h (Fig. 3A). Both C/EBP-␤ (Fig. 3B) and NF-B (Fig. 3C) exhibited constitutive DNA binding in nuclear extracts of quiescent lung fibroblasts. However, in contrast to AP-1, neither of the two was affected by TGF-␤1 treatment. To further demonstrate specificity of AP-1 activation, we also performed EMSA investigating the transcription factors Sp1 and CREB. Although Sp1 is reported to be activated by TGF-␤1 in different cell types, none of these two transcription factors were induced by TGF-␤1 in human lung fibroblasts (data not shown).
Characterization of the AP-1 Complex Induced by TGF-␤1-AP-1 is a dimeric transcription factor that either consists of Jun homodimers or Jun/Fos heterodimers. We therefore analyzed the composition of the TGF-␤1-induced AP-1 complex by supershift analyses with the addition of antibodies recognizing all Jun isoforms (pan-c-Jun/AP-1) or all Fos isoforms (pan-c-Fos). As demonstrated in Fig. 4A, the antibody recognizing all Jun isoforms clearly diminished the appearance of the induced FIG. 3. EMSA characterizing nuclear protein binding to AP-1, C/EBP-␤, and NF-B consensus site oligonucleotides. Human lung fibroblasts were treated with medium alone or TGF-␤1 at 2 ng/ml for the indicated times, and nuclear proteins were extracted as described under "Experimental Procedures." Total protein was calculated by Bradford assay, and equal amounts of protein (2 g) were used for EMSA. A, characteristic EMSA that demonstrates increased binding of nuclear proteins from TGF-␤1-treated fibroblasts to an AP-1 consensus site-containing oligonucleotide. The specific AP-1 band is indicated on the right, as is an unspecific band representing constitutive binding of nuclear proteins to the oligonucleotide. Free probe is shown at the bottom of the gel. B, the same extracts were prepared for binding to a C/EBP-␤ oligonucleotide. On the left side, the labeled oligonucleotide alone without cell extract and a positive control using nuclear extracts from HeLa cells are shown. Afterward, supershift analyses of the HeLa extract using an antibody specific for C/EBP-␤ (NF-IL6) and competition analyses with a 50-fold excess of unlabeled probe are shown. Analyses of nuclear extracts of TGF-␤1-treated lung fibroblasts (from 0 to 16 h) revealed no change in C/EBP-␤ (NF-IL6) binding. C, EMSA demonstrating constitutive NF-B binding activity in nuclear extracts of lung fibroblast, irrespective of treatment with TGF-␤1. The left part of Fig. 3C demonstrates binding activity of NF-B in nuclear extracts of human lung fibroblasts that is not affected by TGF-␤1 over a time course of 16 h. The far left lane represents unstimulated extract (0 h) with the addition of a 50-fold excess of unlabeled oligonucleotide. The right part of Fig. 3C is a supershift analyses of NF-B activity in control synovial fibroblasts known for TNF-␣-induced NF-B activation. Specific NF-B binding activity is indicated on the right as well as unspecific binding of the extracts to the oligonucleotide. Supershift analyses are performed with extracts of TNF-␣-treated synovial fibroblasts using a p65-specific antibody at the indicated titers. The far right lane represents TNF-␣-stimulated extract with the addition of a 50-fold excess of unlabeled oligonucleotide.
AP-1 complex, whereas the unspecific band was unaffected. In contrast, the antibody specific for all Fos isoforms did not cause a reduction of the specific AP-1 band, thereby suggesting the absence of Fos isoforms in this specific AP-1 complex in human lung fibroblasts.
We further analyzed the AP-1 complex by using antibodies specific for distinct Jun isoforms (c-Jun, JunB, and JunD). As demonstrated in Fig. 4B, only the pan-c-Jun and the JunD antibody resulted in a significant decrease of the specific AP-1 band, thus suggesting the AP-1 complex to be a JunD homodimer. This was further confirmed by the fact that only these two antibodies resulted in a clear supershifted band. As a serum control, an antibody specific for the glucocorticoid receptor (GR) supplied by the same manufacturer was used (Fig. 4B). Since the addition of this antibody did not affect the specific band, unspecific effects of the antibodies used for supershift analyses can be reasonably excluded. Furthermore, the addition of a 50-fold excess of unlabeled competitor oligonucleotides diminished the appearance of the specific AP-1 band (Fig. 4B). DISCUSSION TGF-␤s play an essential role in the pathophysiology of lung fibrosis. They have been shown to be causally responsible for the changes associated with the fibroproliferative response in fibrotic lungs, such as fibroblast proliferation and increased extracellular matrix synthesis (28,29). Studies investigating signal transduction by TGF-␤s have thus far revealed several transcription factors that are activated upon stimulation with TGF-␤s in a tissue-and cell type-specific manner (2,30), but although human lung fibroblasts are highly responsive to TGF-␤s, signal transduction pathways activated by TGF-␤ in this cell type remain to be characterized.
In this report, we investigated the molecular mechanisms of TGF-␤1-induced IL-6 expression, a target gene known to be involved in fibroblast proliferation (18,19). Using cultures of primary human lung fibroblasts derived from different patients undergoing pneumectomy/lobectomy, we demonstrate that TGF-␤1 is a potent inducer of IL-6 mRNA and protein expression in this cell type. Luciferase and electrophoretic mobility shift assays demonstrate that the transcription factor AP-1 is required for the observed induction of IL-6 by TGF-␤1. In human lung fibroblasts, none of the other transcription factors that have previously been reported to be activated by TGF-␤s, such as Sp1, were affected in their DNA binding activity. Thus, these results strongly indicate that, in addition to its effect on IL-6 expression, AP-1 is a major signal transducer of TGF-␤1 and is required for most of its biological effects in human lung fibroblasts. This notion is further strengthened by the fact that all results were obtained with fibroblast cultures derived from multiple patients, thus excluding genotype-specific effects and implicating a general role for AP-1 in the human lung. AP-1 is composed as a homodimer of Jun isoforms or as a heterodimer of Jun/Fos isoforms, both members of the immediate early gene family (31,32). DNA binding of AP-1 occurs after activation and dimerization of the described isoforms. Interestingly, the specificity of biological effects induced by upstream stimuli is determined through the actual composition of AP-1 in a certain signal transduction cascade. In this respect, JunB-containing complexes are predominantly described to be negative regulators of AP-1 function and inhibit transcription by negative interaction with AP-1 consensus sequences (33). In contrast, increased gene expression by AP-1 is mostly attributed to c-Jun-containing complexes.
In our experiments, AP-1 was shown to up-regulate IL-6 expression, and the composition of AP-1 was unequivocally determined by supershift analyses with antibodies directed against distinct members of the Jun/Fos family. These supershift analyses revealed that the TGF-␤1-induced AP-1 complex exclusively consisted of JunD homodimers, which have thus far not been described to be a dominant part of AP-1 in TGF-␤ signaling. Although several investigations have found AP-1 activation upon TGF-␤ treatment in different cell types (30,34,35), supershift analyses have revealed that, for the most part, c-Jun isoforms constituted these AP-1 complexes. Consistent with our results, Fos isoforms were generally not affected by TGF-␤.
In an elegant study, Mauviel and colleagues have described how TGF-␤ induces JunB in dermal fibroblasts, leading to decreased expression of the matrix metalloproteinase-1 gene (36). In the same study, the authors found that TGF-␤ upregulated matrix metalloproteinase-1 expression in epidermal keratinocytes. In this cell type, the increase in matrix metalloproteinase-1 expression was mediated by c-Jun, thereby emphasizing the importance of different Jun isoforms in cell typespecific effects of TGF-␤. Investigation of a different target gene, the matrix metalloproteinase-3 gene, revealed that upregulation of this metalloproteinase by TGF-␤ was due to an AP-1 complex that consisted of c-Fos, c-Jun, and JunD (37). Moreover, AP-1 has been found to transduce gene induction by TGF-␤s in the cases of ␣2(I) collagen (10), clusterin (38), plasminogen activator inhibitor-1 (39), retinoic acid receptors (35), and TGF-␤1 (11) itself.
In addition to AP-1, Sp1 is another major transcription factor frequently described as being activated by TGF-␤s, as in the case of ␣1(I) collagen (40) and p21 gene expression (41). We therefore determined the effects of TGF-␤1 on Sp1 in human lung fibroblasts to assess specificity and to generate a more complete analysis of TGF-␤1 signaling in this cell type. Although fetal human skin fibroblasts have been reported to FIG. 4. Supershift analyses of the TGF-␤1-induced AP-1 complex. Human lung fibroblasts were treated with 2 ng/ml TGF-␤1 for 5 h, and nuclear extracts were prepared as outlined before. 2 g of nuclear proteins were incubated with an AP-1 oligonucleotide and analyzed by supershift analyses. A, antibodies reactive against all isoforms of either Jun or Fos, as indicated, were added to the binding conditions at a titer of 1:10. Binding of nuclear proteins was then analyzed by EMSA. The appearance of the specific AP-1 band is indicated on the right. EMSA is representative for six independent experiments with nuclear extracts from six different fibroblast lines. B, the AP-1 complex was further analyzed by the addition of Jun isoform-specific antibodies, the addition of a glucocorticoid receptor (GR) antibody from the same supplier as a serum control, and the addition of a 50-fold excess of unlabeled oligonucleotide, as indicated. A specific band representing bound AP-1 complex is indicated on the right, as well as supershifted bands and unspecific binding of nuclear proteins. The EMSA shown is representative for six experiments using nuclear extracts of different fibroblast lines. activate Sp1 in response to TGF-␤1 (40), we did not detect any up-regulation of Sp1 in adult lung fibroblasts (data not shown). These different observations, however, may result (i) from the different tissue compartments from which fibroblasts were generated (skin versus lung) or (ii) from the fact that fetal fibroblasts react developmentally distinct from adult fibroblasts.
Besides AP-1 and Sp1, we investigated activation of C/EBP-␤, CREB, and NF-B in this study, transcription factors known to be important in the regulation of IL-6 gene expression (20,27). None of these factors were affected by TGF-␤1, and mutations of their respective consensus sequences within the IL-6 promoter did not affect IL-6 gene induction by TGF-␤1. In our model, AP-1 was the only transcription factor of all those investigated found to be activated by TGF-␤1 in adult human lung fibroblasts.
Although these results imply that AP-1 is the immediate effector molecule of TGF-␤1 signaling, we cannot exclude the possibility that other factors contribute to AP-1-driven gene expression in this model. In this respect, CBP/p300 or SMADs are of significant interest. In recent years, SMAD proteins have evolved as essential elements in TGF-␤ signaling (2,13,14). They become phosphorylated and activated by ligand-bound T␤R family members, associate in specific heterodimers, and translocate to the nucleus, where they control gene transcription. Interestingly, control of gene expression by SMADs is achieved either by association with already DNA-bound factors or by direct interaction with cis-regulatory elements (13,15,16,42,43). SMADs have been found to potentiate AP-1-dependent gene transcription, but their precise nuclear function in this context remains unknown (15,16). Thus, the contribution of SMADs in the described TGF-␤1-induced transcriptional activation by AP-1 needs to be fully resolved.
CBP/p300, in contrast, are coactivators that need to bind to activating transcription factors before potentiation of gene transcription can occur (44). CBP is described to associate with AP-1 (45), thereby leading to increased gene transcription that is controlled by AP-1 consensus sites within the promoter sequences investigated (44,46). It is therefore possible that TGF-␤1 may increase CBP activity or its association with AP-1, which would also lead to an increase in AP-1-driven gene transcription, as observed for IL-6 expression in the above described studies.
Taken together, our findings thus underline the importance of AP-1 as a major mediator of TGF-␤1 signaling in human lung fibroblasts. This can also apply to physiological and pathophysiological conditions in the lung in vivo, such as in diseases evoked by increased TGF-␤ expression. In lung fibrosis, fibroblasts are known to be the key cell type responsible for the transition toward a fibrotic extracellular matrix. In this disease, AP-1 inhibition with special regard to JunD might thus be of crucial importance toward the development of novel therapeutic strategies.