The interleukin-6-type cytokine oncostatin M induces aryl hydrocarbon receptor expression in a STAT3-dependent manner in human HepG2 hepatoma cells



The aryl hydrocarbon receptor (AHR) is a ligand-dependent transcription factor that mediates the toxicity of dioxins, polycyclic aromatic hydrocarbons and related environmental pollutants. Besides drug metabolism, several studies have provided evidence that the AHR and its downstream targets trigger important developmental, physiological and pathophysiological processes. However, in contrast to the molecular mechanisms of AHR-dependent signaling pathways, the transcriptional regulation of the AHR gene itself is as yet only marginally understood. We found that the pleiotropic interleukin (IL)-6-type cytokine oncostatin M (OSM) is an inducer of AHR mRNA and protein expression in human HepG2 hepatocarcinoma cells. Analyses of the human AHR promoter revealed the existence of a putative signal transducer and activator of transcription (STAT)-binding element 5′-upstream of the transcription start site. By means of site-directed mutagenesis, inhibitor experiments and electrophoretic mobility shift assays, we demonstrated that this STAT motif is recognized by STAT3 to regulate basal and cytokine-inducible AHR expression in HepG2 cells. The identification of the AHR as a downstream target of IL-6-type cytokine-stimulated STAT3 signaling may contribute to a better understanding of the multiple facets of AHR during development, physiology and disease.


aryl hydrocarbon receptor


cytochrome P450




oncostatin M


signal transducer and activator of transcription


transforming growth factor-β1


vascular endothelial growth factor


Oncostatin M (OSM) is a multifunctional cytokine of the interleukin (IL)-6 family, whose members trigger multiple cellular processes, such as hematopoiesis, angiogenesis, immune responses and acute-phase reactions [1, 2]. IL-6-type cytokines initiate signal transduction by binding to specific cell-surface receptors, consisting of a cytokine-specific binding domain (e.g. OSM-receptor) and a common signaling domain, the glycoprotein 130. Ligand binding leads to dimerization of two cytokine receptor subunits followed by activation of janus kinases, which in turn phosphorylate specific tyrosine residues of the cytokine receptor subunits to provide distinct binding sites for signal transducer and activator of transcription (STAT) molecules and SHP2. Both SHP2 tyrosine phosphatases and STATs are immediately phosphorylated by janus kinases, leading to: (a) SHP2-mediated activation of MAPK signaling, and (b) the formation of transcriptional active STAT homodimers [1, 2]. The STAT dimers shuttle into the nucleus and induce target gene expression through binding to STAT elements (5′-CTGGGAA-3′) within the promoter region [1, 2]. Beside their physiological function, both OSM and its main effector molecule, STAT3, are often dysregulated in several inflammatory diseases, such as rheumatoid arthritis and inflammatory bowel disease [1-3].

The AHR is a ligand-activated transcription factor known to mediate the biochemical and toxic effects of environmental pollutants, such as dioxins and polycyclic aromatic hydrocarbons [4, 5]. In absence of a ligand, the AHR is trapped in a cytosolic multiprotein complex consisting of heat-shock protein 90, AHR-interacting protein and p23. Upon ligand-binding the cytosolic AHR sheds its co-chaperones and translocates in the nucleus, where it dimerizes with the AHR nuclear translocator and binds to dioxin-responsive elements in the promoter of target genes to enforce transcription [4, 5]. The AHR gene battery encodes for several drug-metabolizing enzymes, such as cytochrome P450 (CYP) 1A1 and 1A2, as well as for proteins involved in regulation of apoptosis, cell growth and differentiation [4, 5]. Besides its role in drug metabolism, several studies on rodents have provided evidence that the AHR is involved in developmental processes [6, 7], inflammation [8] and tumor development [9, 10].

Although the molecular mechanisms of AHR signaling are well characterized, factors and stimuli regulating AHR expression remain quite enigmatic in most cell types and tissues. The purpose of this study was to investigate whether the IL-6-type cytokine OSM can modulate AHR expression in human HepG2 hepatocarcinoma cells and, if so, by which molecular pathway.


To investigate whether IL-6-type cytokines are capable of inducing AHR expression, we treated HepG2 cells with different concentrations of OSM and analyzed AHR mRNA expression using quantitative real-time PCR. As shown in Fig. 1(A), OSM exposure of HepG2 cells resulted in a significant induction of AHR transcription. A time-course study revealed a significant twofold induction of AHR transcription 2 h after exposure to 20 and 40 ng·mL−1 OSM (Fig. 1A). At later time points, AHR gene induction was only slightly enhanced. Treatment of the cells with increasing concentrations of OSM (5–40 ng·mL−1) resulted in a dose-dependent induction of AHR expression after 16 h (Fig. 1B). A first significant elevation of AHR mRNA expression was observed after administration of 10 ng·mL−1 OSM. To test whether enhanced AHR gene expression was also translated to the protein level, we performed western blot analyses. Twenty-four hours after exposure of HepG2 cells to 10 and 20 ng·mL−1 OSM, we observed a slight (1.4-fold) but significant induction of AHR protein (Fig. 1C), indicating that the OSM-mediated induction of AHR gene expression is translated to the protein level and thus probably of functional relevance. Besides OSM, a 16 h exposure to 20 ng·mL−1 IL-6 also led to a significant elevation of AHR mRNA expression, revealing that AHR expression is inducible by other IL-6-type cytokines also (Fig. 1D).

Figure 1.

Induction of AHR mRNA and protein expression in OSM-treated HepG2 cells. (A) HepG2 cells were treated for 2, 4, 8, and 20 h with 20 and 40 ng·mL−1 OSM or solvent, and AHR expression was investigated by real-time PCR analyses. AHR transcripts are shown in ratio to β-actin. n = 3; *, significantly increased compared with solvent-treated controls. (B) HepG2 cells were treated for 16 h with 5, 10, 20 and 40 ng·mL−1 OSM or solvent and AHR expression was investigated by real-time PCR analyses. AHR transcripts are shown in ratio to β-actin. n = 3; *, significantly increased compared with solvent-treated controls. (C) HepG2 cells were treated with 20 and 40 ng·mL−1 OSM or solvent. After 24 h, total protein was isolated and AHR expression was measured using western blot techniques. For densitometry, AHR signal intensity was adjusted to the respective β-actin signals. n = 3; *, significantly increased compared with solvent-treated controls. (D) HepG2 cells were treated for 16 h with 20 ng·mL−1 IL-6 or solvent, and AHR expression was investigated by real-time PCR analyses. AHR transcripts are shown in ratio to β-actin. n = 3; *, significantly increased compared with solvent-treated controls.

In silico promoter analyses revealed the existence of a putative STAT motif (5′-CTGGGAA-3′) located from bases -1975 to -1969 in the human AHR gene locus. Interestingly, this STAT motif is located in a region that we have previously identified to be crucial for basal AHR promoter activity [11]. In order to test whether this STAT element is of functional relevance, we transfected HepG2 cells with pAHR-1980 construct and subsequently treated the cells with 20 ng·mL−1 OSM or solvent. As shown in Fig. 2, OSM exposure resulted in a twofold increase in luciferase activity, indicating that the putative STAT-binding site is active. To confirm the functional relevance of the STAT-binding site for constitutive and OSM-inducible AHR promoter activity, we mutated two core bases of the STAT element and performed comparative reporter gene analyses. A site-directed mutagenesis of the core bases [pAhR-1980(G-1971C/A-1970C)] resulted in a significant decrease in both basal and OSM-stimulated luciferase activity (Fig. 2). In comparison with the pAHR-1980 construct, the constitutive promoter activity of the mutated AHR construct was nearly completely abolished, indicating that STAT3 may be involved in regulation of basal AHR gene expression. Importantly, HepG2 cells were shown to constitutively express OSM [12], a finding that may explain the involvement of STAT3 in basal AHR gene regulation in this cell line. In order to verify these results, we transfected HepG2 cells with the pAhR-1980 reporter plasmid and treated the cells with 20 ng·mL−1 OSM ± 10 μm of the specific STAT3 inhibitor stattic [13]. As shown in Fig. 2, stattic exposure led to a clear decrease in constitutive as well as cytokine-induced AHR promoter activity, thereby proving that STAT3 controls AHR expression. To further ensure a functional involvement of STAT3 in basal AHR gene expression in HepG2 cells, we performed a siRNA-mediated STAT3 knockdown. Indeed, transient gene targeting of STAT3 expression (down to 13% compared with nonsilencing siRNA-transfected controls) resulted in a significant 33% decrease in constitutive AHR transcription (Fig. 3A). As expected, basal expression of the STAT3 target gene vascular endothelial growth factor (VEGF) [14] was also significantly reduced by STAT3 knockdown (Fig. 3A).

Figure 2.

Basal and OSM-induced regulation of AHR promoter activity depends on a functional STAT element. HepG2 cells were transfected with either 1 μg·well−1 pAHR-1980 or pSTAT-mut and subsequently treated with 20 ng·mL−1 OSM, 10 μm stattic or solvent. After 24 h luciferase activity was determined and normalized to protein content. n = 3; *, significantly increased compared with solvent-treated pAHR-1980-transfected controls; $, significantly reduced compared with solvent-treated pAHR-1980-transfected controls; #, significantly reduced compared with OSM-treated pAHR-1980-transfected samples.

Figure 3.

STAT3 regulates the constitutive AHR expression and binds to the STAT element in the human AHR gene. (A) HepG2 cells were transiently transfected with 50 pmol STAT3-siRNA or control siRNA. After 24 h, mRNA expression of STAT3, VEGF and AHR was analyzed and normalized to β-actin copy numbers. n = 3; *, significantly decreased compared with control siRNA-transfected cells. (B) Nuclear extracts from OSM-treated and untreated HepG2 cells were used for EMSA. EMSA was performed using a double-stranded, 32P-labeled oligonucleotide containing the STAT-binding site located on the human AHR promoter. The band corresponding to the STAT3 complex is indicated by an arrow. Excess (100-fold) of an unlabeled oligonucleotide was added as competitor. For supershift assays 1, 2 and 4 μg of anti-STAT3 Ig was added to the reaction mixture. A representative EMSA from two independent experiments is shown.

Next, we performed EMSA analyses to prove the binding of STAT3 to its putative binding site in the AHR promoter. As shown in Fig. 3B, we already observed a band shift in unexposed HepG2 cells (lane 2), again pointing to the idea that STAT3 controls basal AHR expression. The intensity of the band shift further increased upon OSM treatment (lane 3) and completely disappeared upon administration of an excess of cold competitor (lane 4). Interestingly, we did not observe a supershift but a complete disappearance of the shifted band upon addition of increasing amounts of anti-STAT3 IgG (lane 5–7). This observation implied that the antibody has somehow disturbed the protein–probe binding. A possible explanation might be that the antibody and the labeled probe compete for the same epitope on the STAT3 protein, a phenomenon that has been previously observed in shift assays with an anti-STAT3 Ig from another source [15]. However, in combination with the data from the reporter gene assays and the siRNA experiments, our results strongly indicate that IL-6-type cytokines can regulate constitutive as well as inducible AHR expression via STAT3 signal transduction. Interestingly, other cytokines and hormones known to activate STAT3 signaling, namely IL-22 (20 ng·mL−1) and leptin (100 ng·mL−1) [16], did not affect AHR transcription in HepG2 cells (Fig. S1). The molecular reasons for this remain to be elucidated, but may, for instance, involve low or absent expression of the respective membrane receptors. Noteworthy, we cannot completely exclude that, besides STAT3, other transcription factors, e.g. other STAT family members, also bind to the respective STAT-binding site to regulate AHR transcription.

We have previously shown that exposure of HepG2 cells to transforming growth factor-β1 (TGF-β1) stimulated AHR expression [11]. As discussed below, a possible synergistic or additive effect of TGF-β1 and IL-6-type cytokines on AHR expression may be relevant for certain pathophysiological processes. Therefore, we transiently transfected HepG2 cells with pAhR-1980 and subsequently exposed the cells to 20 ng·mL−1 IL-6 and TGF-β1 (50 and 100 pm) alone or in combination. A 24 h treatment with IL-6 resulted in a twofold induction of AHR promoter activity; with 100 pm TGF-β1 the increase was 2.5-fold (Fig. 4). Coexposure of the cells to IL-6 and TGF-β1 did not further elevate AHR promoter activity (Fig. 4), indicating that IL-6- and TGF-β1-stimulated signaling pathways do not cooperate in modulating AHR gene expression in HepG2 cells.

Figure 4.

Influence of single and coexposure to IL-6 and TGF-β1 on human AHR promoter activity. HepG2 cells were transfected with 0.75 μg·well−1 pAHR-1980 and on the next day treated with 20 ng·mL−1 IL-6, 50 or 100 pm TGFβ1, IL-6 plus TGFβ1, or solvent. After 24 h luciferase activity was determined and normalized to protein content. Results are shown as fold of solvent controls. n = 3; *, significantly increased compared with solvent-treated pAHR-1980-transfected controls.


In an earlier study on HepG2 and A549 lung adenocarcinoma cells, we identified TGF-β1 and downstream Smad proteins as critical mediators of cell-specific AHR expression [11]. In addition, only a few other transcription factors, including peroxisome proliferator-activated receptor-α [17], β-catenin [18, 19] and Sp1-related zinc fingers [20], were reported to regulate AHR gene expression. In this study, we identified for the first time a novel molecular mechanism of AHR gene regulation initiated by proinflammatory cytokines of the IL-6 family. Expression analyses, reporter gene assays and EMSA experiments revealed that OSM increased AHR expression through stimulation of STAT3 proteins, which subsequently bind to a STAT-binding element located in the human AHR promoter to enforce transcription. Our results further indicate that an endogenously expressed level of OSM maintains constitutive AHR expression in HepG2 cells. This is probably of relevance regarding the responsiveness of the hepatoma cells to endogenous and exogenous AHR-modulating compounds, which may influence numerous cellular processes, such as metabolism, membrane integrity, adhesion, cell-cycle control, and apoptosis [21, 22].

Interestingly, OSM exposure of HepG2 cells resulted in a transcriptional repression of constitutive CYP1A1 expression (data not shown). Previously, it was reported that treatment of human hepatocytes and HepG2 cells with IL-6-type cytokines resulted in a decrease of basal as well as inducible expression of AHR-dependent (CYP1A1, CYP1A2) and AHR-independent (CYP3A4) CYP monooxygenases [23-25]. The underlying molecular mechanisms are controversially discussed, but probably occur in a janus kinase/STAT-independent manner [26]. As postulated previously, this downregulation of hepatic CYP enzymes in response to infectious and inflammatory stimuli may prevent the CYP-mediated production of reactive oxygen species, anti-inflammatory arachidonic acid metabolites and nitric oxide [27]. Even though the OSM-mediated inhibition of CYP1A1 transcription occurred via an as yet unknown mechanism, it implies that the cytokine-driven increase in AHR expression may not elevate drug metabolism but influence other cellular processes.

OSM is an important cytokine in the context of liver development and physiology. In previous studies, it has been shown to: (a) stimulate the production of acute-phase proteins in hepatocytes and hepatoma cells, (b) regulate the differentiation and functional maturation of fetal hepatocytes, and (c) induce hepatocyte proliferation and tissue remodeling during liver regeneration [1, 28, 29]. Because STAT3 is the main effector of OSM during most of these processes, it is tempting to speculate that STAT3, activated by OSM or other IL-6-type cytokines, is responsible for the variable AHR expression pattern observed in the developing liver [30, 31]. AHR-null mice display a reduced liver weight, alterations in the hepatic vasculature, hepatic fibrosis and an increased rate of hepatocyte apoptosis [7, 32-34], underscoring the urgent requirement of AHR for proper liver development. RNAi-mediated inhibition of AHR expression in HepG2 cells resulted in a significant reduction in proliferation, which was due to decreased progression from the G0/G1-phase to the S-phase, accompanied by reduced expression rates of cyclin D1, cyclin E and cyclin-dependent kinases 2 and 4 [35]. Similar inhibition of cell division was observed in AHR-defective murine Hepa1c1c7 cells [36]. In addition, these AHR-defective hepatoma cells turned out to be more susceptible to UV- and H2O2-induced cell death [37], demonstrating that the AHR serves a pro-proliferative and anti-apoptotic function in hepatoma cells. In vivo, these properties may contribute to the promoting effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin on hepatocarcinogenesis [5]. Just recently, aberrantly high expression rates of AHR have been detected in samples from human hepatocellular carcinomas [38]. Notably, AHR expression height correlated with tumor cell invasion, which might be due to an AHR-dependent modulation of extracellular matrix-degrading enzymes and adhesion factors [39]. Importantly, STAT3 was shown to be activated in 60% of human hepatocellular carcinomas, with STAT3-positive tumors being more aggressive [40]. Even though the reason for this abnormal activation of STAT3 is not clear, it was proposed that NF-κB-driven elevated levels of IL-6-type cytokines (e.g. IL-6, IL-11) may be causative [41]. Thus, one might hypothesize that the AHR is upregulated in human hepatocellular carcinomas through a STAT3-dependent mechanism, similar to that identified in our current in vitro study. Interestingly, increased levels of IL-6 [42, 43], activated STAT3 [44] and AHR [45] have also been noted in tissue samples from human breast cancer. However, because ectopic overexpression of the AHR was shown to be sufficient to induce malignant transformation of mammary epithelial cells [46], it is difficult to determine whether the enhanced AHR expression observed in mammary or hepatic tumors is a consequence of carcinogenesis or vice versa.

Recently, experimental evidence has been provided that both STAT3 [47, 48] and AHR [49, 50] are important for the induction of IL-17-producing inflammatory T cells (TH17). The differentiation of human and murine TH17 cells from naïve T cells strongly depends on the presence of IL-6 and TGF-β1 [51, 52], two factors that are – according to the results of our present study and previously published data from our lab [11] – capable of inducing AHR expression via STAT3- and Smad-dependent pathways, respectively. Indeed, human and murine TH17 cells express significantly higher amounts of AHR than other effector T cells or regulatory T cells [49, 50], pointing to the potential relevance of the newly identified mechanism of AHR regulation for T-cell biology and associated immune responses. However, coexposure to IL-6 and TGF-β1 did not lead to a significant additive effect on AHR promoter activity in HepG2 cells. In contrast, exposure of naïve T cells to IL-6 and TGF-β1, but neither TGF-β1 nor IL-6 alone, was reported to induce AHR expression in naïve T cells [50, 53], indicating putative cell- or tissue-specific differences in AHR gene regulation.

Notably, the identified STAT motif is highly conserved throughout the group of Hominoidea, but not retrievable in the mouse genome, implying putative species-specific differences in AHR gene regulation. However, a further database-driven analysis revealed the existence of two other putative STAT motifs located 5′-upstream (-6342 and -6756) of the transcription start site of the murine AHR gene (accession number NT_039548.8). Even though, the functionality of these binding sites remains to be proven, the abovementioned requirement of IL-6 for AHR expression during differentiation of murine TH17 cells strongly implies involvement of STAT3 signal transduction [53].

In conclusion, the IL-6-type cytokine-driven and STAT3-dependent regulation of AHR identified in this study may explain the often observed alterations in AHR expression during embryonic development and under certain physiological, as well as pathophysiological, conditions. Future in vitro and in vivo investigations are urgently required to clarify whether this regulatory mechanism may serve as a suitable target to prevent or treat malignant, inflammatory or autoimmune diseases.

Material and methods

Cell culture and treatment

The HepG2 cell line was cultured in RPMI medium (PAA, Coelbe, Germany) supplemented with 10% FCS Gold (PAA), 3.7% (w/v) NaHCO3 (PAA) and penicillin/streptomycin in a humidified atmosphere of 5% CO2 at 37 °C. OSM (Sigma-Aldrich, Munich, Germany) was dissolved in NaCl/Pi plus 0.1% BSA, IL-6 (Santa Cruz Biotechnology, Santa Cruz, CA, USA) in NaCl/Pi plus 0.1% fetal calf serum, IL-22 and leptin (PeproTech, Hamburg, Germany) in water plus 0.1% BSA, TGFβ1 (PeproTech) in water plus 0.1% fetal calf serum, and stattic (Merck Millipore, Darmstadt, Germany) in dimethylsulfoxide.

Plasmid constructs

The construction of the AHR gene reporter plasmid AHRΔ-(1980) was described previously [11]. The generation of the pAHR-1980/G-1971C/A-1970C construct containing a point-mutated STAT motif (in the following referred to as pSTAT-mut), was achieved by using the Quick Change site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA) according to the manufacturer's instructions. For mutagenesis, the sense oligonucleotide 5′-CTCTATCGATAGAATTCTGGCCATTTTGTTGATATTC-3′and the corresponding antisense oligonucleotide were used. The exchanged nucleotides (underlined) were verified by automated DNA sequencing.

Reporter gene assays

HepG2 cells (2 × 105 per well) were seeded into six-well plates, and maintained overnight under standard conditions. Cells were transiently transfected with the indicated amounts of AHR promoter constructs pAHR-1980 or pSTAT-mut using JetPEI transfection reagent (Polyplus Transfection, Illkirch, France) according to the manufacturer's instructions. Next day, transfected cells were treated with OSM or solvent as indicated. Cells were lysed and luciferase activities were determined in a Multi-Bioluminat LB 9505C (Berthold Technologies, Bad Wildbad, Germany) using the luciferase assay system (Promega, Mannheim, Germany). Firefly luciferase activity was normalized to protein content, which was determined using the BC assay protein quantitation kit (Interchim, Frankfurt, Germany) according to the manufacturer's instruction.

Quantitative real-time PCR

HepG2 cells (2 × 105 per well) were seeded into six-well plates, maintained overnight under standard conditions and treated as indicated. Isolation of total RNA, cDNA synthesis and quantitative real-time PCR were carried out as described previously [54]. AHR transcripts were normalized to β-actin expression. The sequences of the oligonucleotides used were: 5′-CCCCAGGCACCAGGGCGTGAT-3′ and 5′-GGTCATCTTCTCGCGGTTGGCCTTGGGGT-3′ for β-actin, 5′-TGGTCTCCCCCAGACAGTAG-3′and 5′-TTCATTGCCAGAAAACCAGA-3′for AHR, 5′-CAGGATGGCCCAATGGAATC-3′and 5′-CCCAGGAGATTATGAAACACC-3′for STAT3, and 5′-TGCAAAAACACAGACTCGCG-3′and 5′-TGTCACATCTGCAAGTACGTTCG-3′for VEGF.

RNA interference

For transient gene silencing, HepG2 cells (2.5 × 105 per well) were seeded into six-well plates and maintained overnight under standard conditions. Next day, cells were transfected with 50 pmol STAT3–siRNA (Dharmacon Inc., Lafayette, CO, USA) or nonsilencing control siRNA (Santa Cruz Biotechnology) using INTERFERin siRNA transfection reagent (Polyplus Transfection). After 24 h cells were harvested and mRNA expression was determined as described above.

SDS/PAGE and western blot analyses

Cells were lysed in RIPA buffer (25 mm Tris/HCl, pH 7.4; 150 mm NaCl; 0.1 mm EDTA, pH 8.0; 1% Nonidet P-40; 1% desoxycholate; 0.1% SDS; 0.025% NaN3; protease inhibitors) on ice and subsequently centrifuged for 5 min at 4 °C at maximum speed. Protein samples were separated by 12% SDS/PAGE and blotted onto poly(vinylidene difluoride) membranes (GE Healthcare, Freiburg, Germany). Blots were blocked with 5% skim milk in TBS-Tween 20 0.1% (TBS-T) for 1 h at room temperature. Blots were incubated overnight at 4 °C with antibodies against AHR (Antibodies online, Aachen, Germany; #ABIN802869) or GAPDH (Cell Signaling Technology, Danvers, MA, USA; #2118), according to the manufacturer's instructions. Blots were washed and subsequently incubated for 1 h with a 1 : 5000 dilution of horseradish peroxidase-conjugated secondary antibodies in 5% skim milk in TBS-T at room temperature. Bands were visualized using the chemi-luminescence ECL Prime detection system (GE Healthcare) and X-ray films. Densitometric analyses were performed using the alphaeasefc Software (Alpha Innotech, San Leandro, CA, USA).


HepG2 cells were harvested 10 min after treatment with OSM or solvent and nuclear extracts were prepared as described previously [11].

The double-stranded oligonucleotides were end-labeled with [33P]dATP[γP] by Hartmann Analytic GmbH (Braunschweig, Germany) using T4 polynucleotide kinase. Binding reactions containing 20 μg of nuclear extracts and 10–20 fmol of labeled oligonucleotides (5′-CGATAGAATTCTGGGAATTTTGTTG-3′) were performed for 60 min at 25 °C in a final volume of 25 μL binding buffer [20 mm Hepes pH 7.9, 50 mm KCl, 4 mm MgCl2, 0.1 mm EDTA, 0.5 mm dithiothreitol, 5% glycerol, 4 mm spermidine, 2.5 μg poly(dI–dC)]. Protein–DNA complexes were resolved in 5% polyacrylamide gels containing 0.5 × TGE (25 mm Tris base, 190 mm glycine, 1 mm EDTA pH 8.3) under constant voltage (80 V) for 3.5–4 h. For supershift analyses nuclear extracts were first preincubated for 20 min with the labeled probe, followed by 40 min incubation with indicated amounts of anti-STAT3 (StressGen, Victoria, Canada; #KAP-TF003E).

Statistical analyses

All data shown are mean (± SD) from three independent experiments (except EMSA, n = 2). Data were analyzed using two-sided Student's t-test. values of P ≤ 0.05 were considered significant.


NSM was supported by the Graduate School 1427 of the German Research Foundation (DFG).