The Designer Cytokine Hyper-Interleukin-6 Is a Potent Activator of STAT3-dependent Gene Transcription in Vivoand in Vitro *

Interleukin-6 (IL-6) triggers pivotal pathwaysin vivo. The designer protein hyper-IL-6 (H-IL-6) fuses the soluble IL-6 receptor (sIL-6R) through an intermediate linker with IL-6. The intracellular pathways that are triggered by H-IL-6 are not defined yet. Therefore, we studied the molecular mechanisms leading to H-IL-6-dependent gene activation. H-IL-6 stimulates haptoglobin mRNA expression in HepG2 cells, which is transcriptionally mediated as assessed by run-off experiments. The increase in haptoglobin gene transcription correlates with higher nuclear translocation of tyrosine-phosphorylated STAT3 and its DNA binding. As H-IL-6 stimulates STAT3-dependent gene transcription, we compared the molecular mechanism between IL-6 and H-IL-6. Transfection experiments were performed with a STAT3-dependent luciferase construct. The same amount of H-IL-6 stimulated luciferase activity faster, stronger, and for a longer period of time. Dose response experiments showed that a 10-fold lower dose of H-IL-6 stimulated STAT3-dependent gene transcription comparable with the higher amount of IL-6. Cotransfection with the gp80 and/or gp130 receptor revealed that the effect of H-IL-6 on STAT3-dependent gene transcription is restricted to the gp80/gp130 receptor ratio. High amounts of gp130 increased and high amounts of gp80 decreased the effect on H-IL-6-dependent gene transcription. To investigate the in vivo effect of H-IL-6 on gene transcription in the liver, H-IL-6 and IL-6 were injected into C3H mice. H-IL-6 was at least 10-fold more effective in stimulating the DNA binding and nuclear translocation of STAT3, which enhances haptoglobin mRNA and protein expression. Thus H-IL-6 stimulates STAT3-dependent gene transcription in liver cells in vitro and in vivo at least 10-fold more effectively than IL-6. Our results provide evidence that H-IL-6 is a promising designer protein for therapeutic intervention during different pathophysiological conditions also in humans.

Cytokines are important to maintain several physiological and pathophysiological functions in vivo. Binding to a specific membrane receptor activates intracellular signaling pathways, which trigger a variety of intracellular events. IL-6 is a cyto-kine with pleiotropic functions in the immune system, hematopoietic cells, hepatocytes, and the nervous system. It belongs to a family consisting of IL-6, 1 CNTF, LIF, OSM, IL-11, and CT-1. All these family members interact with a ligand binding domain that confers specificity. The different ligand binding domains are all linked to the signal transducer gp130 (1,2). Three members of the Janus tyrosine kinase family, Jak1, Jak2, and Tyk, are constitutively associated with the intracellular domain of gp130 (3,4).
IL-6 associates with the IL-6 receptor (IL-6R, gp80). Binding of IL-6 to its receptor induces homodimerization of gp130 resulting in autophosphorylation of the associated Jak kinases, which in turn phosphorylate gp130. Phosphorylation of gp130 creates docking sites for SH2 domains containing signaling molecules (5). At least two signaling cascades have been characterized that either trigger STAT3 or mitogen-activated protein kinase activation (6). After association with the gp130 receptor, STAT3 becomes phosphorylated at tyrosine 705 through Jak kinases (7). Tyrosine phosphorylation of STAT3 results in homo-or heterodimerization of STAT3 and its nuclear translocation (8). In the nucleus STAT3 binds to target sequences in different promoters and enhances gene transcription (8). In pro-B cells, STAT3 is required for bcl-2 induction, and thus has an anti-apoptotic effect. The mitogen-activated protein kinase pathway, which also diverges at the intracellular domain of gp130, involves the activation of ras and is essential for cell cycle progression and induction of DNA synthesis (6).
In hepatocytes, IL-6 induces the synthesis of acute phase plasma proteins, which play a protective role during the acute phase response (9). Binding sites for STAT3 are located in most of the promoters of the acute phase genes, and originally STAT3 was also called acute phase response factor (APRF) (10). Besides the role of IL-6 in inducing the acute phase response, there is evidence that through IL-6, STAT3 is also activated during liver regeneration (11,12). Additional, recent experiments in IL-6 and tumor necrosis factor receptor 1 knockout mice show that IL-6 is essential to induce cell cycle progression during liver regeneration. In both mice there is a lack of DNA synthesis after hepatectomy that can be prevented by IL-6 injection (13,14). Therefore, IL-6 and, thus, the activation of the intracellular-signaling cascades that diverge at the gp130 transducer molecule are essential to maintain and restore liver function during different conditions.
Assembly of IL-6 with the IL-6 receptor and the gp130 molecule requires the association of three different molecules (15,16). The on-off rate on the cell membrane and the availability of each partner might be a rate-limiting step. A designer pro-tein (H-IL-6), which fuses IL-6 to the soluble IL-6 receptor through an intermediate linker, might also be of therapeutic benefit in humans, for example during infection or liver injury (17). Therefore, we were interested in studying the effect of H-IL-6 on gene transcription in vivo and in vitro in liver cells. We show that H-IL-6 is severalfold more potent in vivo and in vitro in activating STAT3-dependent gene transcription. The effect on gene transcription is dependent on the gp80/gp130 ratio on a given cell. Therefore, H-IL-6 could be of therapeutic benefit during different pathophysiological conditions also in vivo.

MATERIALS AND METHODS
Recombinant H-IL-6 and IL-6 -Recombinant H-IL-6 was synthesized in yeast cells as described earlier (17). The H-IL-6 protein was purified from yeast supernatants by anion-exchange chromatography and gel filtration. The purified H-IL-6 fraction was visualized by running on a SDS-polyacrylamide gel electrophoresis followed by silver staining. Human IL-6 was produced in Escherichia coli and purified as described previously (18).
IL-6 Determination-IL-6 serum concentrations obtained from the liver vein were measured essentially as described previously (12).
Stimulation of Mice and Preparation of Liver Nuclear Extracts-C3H mice were either stimulated with H-IL-6 or IL-6 intraperitoneally by the doses indicated. Tissue for Northern blot analysis or preparation of liver nuclear extracts was perfomed at different time points after injection. At each time point at least three animals were used in parallel.
For preparation of nuclear extracts the pooled livers were rinsed in freezing phosphate-buffered sulfate, and liver nuclear proteins were prepared as described previously (19). All the steps were performed at 4°C. Nuclear proteins were aliquoted and frozen immediately in liquid nitrogen.
Cell Culture, Transfection Experiments, and Luciferase Assays-HepG2 cells (ATCC) were cultured in Dulbecco's modified essential medium supplemented with 10% fetal calf serum. DNA transfection into HepG2 cells was performed as described previously (19). The reporter construct used in these experiments represents the promoter of the ␣-macroglobulin gene linked to a luciferase reporter gene (20). For H-IL-6 and IL-6 stimulation experiments, cells were cultured with 1% fetal calf serum for 24 h after transfection. Cotransfection of the reporter gene construct with the gp130 and IL-6 receptor expression vector was performed with the amounts indicated. Cells were stimulated with H-IL-6 or IL-6 in the amounts and for the time points indicated. Luciferase activity was measured as described earlier (21) using a Lumat LB 9501 (Berthold, Germany). The relative luciferase activity was normalized by cotransfecting a Rous sarcoma virus-␤galactosidase expression vector. The results, which are shown, represent the normalized luciferase activity. Nuclear extracts were prepared from HepG2 hepatoma cells using the Dignam C method as described previously (22).

SDS-polyacrylamide Gel Electrophoresis and Western Blot
Analysis-Liver nuclear proteins or 20 g of serum were separated on a 10% SDS-polyacrylamide gel (19) and blotted onto a nitrocellulose membrane (Schleicher & Schuell) in 1% SDS, 20% methanol, 400 mM glycine, 50 mM Tris-HCl, pH 8.3, at 4°C for 2 h at 200 mA. STAT3 and phospho-STAT3 was visualized using polyclonal antibodies supplied by Santa Cruz Biotechnology (Santa Cruz, CA) and haptoglobin by using a polyclonal rabbit antibody supplied from DAKO (Hamburg, Germany). The antigen-antibody complexes were visualized using the ECL detection system as recommended by the manufacturer (Amersham, Braunschweig, Germany).
Gel Retardation Assays-For gel retardation assays, liver nuclear extracts were used as indicated. The binding reaction was performed for 20 min on ice (22). For binding assays, an oligonucleotide spanning the STAT3 site in the ␣2-macroglobulin promoter was used as a 32 P-labeled probe. The oligonucleotides (sense 5Ј-GATCCTTCTGGGAATTCCTA-3Ј and antisense 5Ј-GATCTAGGAATTCCCAGAAG-3Ј) were purchased from Naps (Göttingen, Germany). Free DNA and DNA-protein complexes were resolved on a 6% polyacrylamide gel as described previously (22). "Supershift" experiments were performed with nuclear extracts prepared at the time points indicated. Complex formation for supershift experiments was performed with either an antibody directed against STAT3 purchased from Santa Cruz Biotechnology or antibodies directed against STAT1, which were a generous gift from Thomas Decker (Vienna, Austria).
Northern Blot Analysis-Total RNA was isolated by the guanidium isothiocyanate method from the liver of the mice or from HepG2 cells at the time points indicated. Northern blot analysis was performed as described before, according to standard procedures (19). 15 g of total RNA was analyzed through a 1% agarose formaldehyde gel, followed by transfer to Hybond N membranes (Amersham). The haptoglobin and GAPDH cDNA probes were labeled with [␣-32 P]ATP according to random priming (Boehringer Mannheim). The hybridization procedure was performed as described previously (19). Run-off Assays-Run-off experiments were performed essentially as described before (12). 5 ϫ 10 7 isolated nuclei were incubated in reaction buffer (5 mM Tris-Cl, pH 8.0, 2.5 mM MgCl 2 , and 0.15 M KCl) containing 10 ml of 100 mM ATP, 10 ml of 100 mM CTP, 10 ml of 100 mM GTP, and 10 ml of 10 mCi/ml [␣-32 P]UTP for 30 min at 30°C. RNA was extracted, precipitated, and dissolved in TES buffer exactly as described. 2 l of the labeled RNA was counted in a scintillation vial.
For hybridization, 6 g of cDNA was denatured and fixed on nylon membrane in a total dilution volume of 100 ml TES buffer. DNA was fixed for 1 h at 80°C.
For quantification, exposure was performed on a Fuji image plate. After subtraction of the vector control signal, counts of the haptoglobin signal were distributed through the GAPDH value and set to 1. The values at different time points after IL-6 or H-IL-6 treatment were shown as fold activation compared with the pretreatment level.
Quantification-Quantification of results was performed with a Fuji imager as described before (19).

H-IL-6 Stimulates Transcription of the Haptoglobin Gene-
Several intracellular pathways are located downstream of the intracellular domain of the gp130 molecule (6). We were interested in studying whether H-IL-6 can stimulate gene expression on a transcriptional level in liver cells. Therefore HepG2 cells were stimulated with H-IL-6, and the mRNA expression was studied after different time points. Northern blot analysis showed that after 10 ng/ml H-IL-6 haptoglobin mRNA expression was increased for up to 96 h (Fig. 1A). Quantitative analysis was performed by comparing the haptoglobin signal with the GAPDH signal. These results showed that 1 h after H-IL-6 stimulation, haptoglobin mRNA expression was already more than 5-fold enhanced to the pretreatment level. The most prominent increase was found 15 h after stimulation exceeding the pretreatment level more than 120-fold (Fig. 1B).
In further experiments, run-off analysis was performed to demonstrate that the effect on haptoglobin mRNA expression is transcriptionally mediated. As shown in Fig. 1, C and D, the transcription rate of the haptoglobin gene was already more than 16-fold enhanced 45 min after H-IL-6 stimulation. At later time points, the transcription rate of the haptoglobin gene decreased. However 15 h after stimulation, the transcription rate was still more than 6-fold higher compared with the pretreatment level. Thus these results showed that H-IL-6 stimulates haptoglobin mRNA expression via higher gene transcription.
H-IL-6 Induces the Nuclear Translocation, Tyrosine Phosphorylation, and DNA Binding of STAT3-H-IL-6 induces higher transcription of the haptoglobin gene. The transcription rate of the haptoglobin gene during the acute phase response in the liver is controlled via STAT3, and therefore its nuclear expression was studied following H-IL-6 stimulation ( Fig. 2A). Higher STAT3 expression was immediately found after H-IL-6 stimulation, and the effect lasted for up to 9 h. Increased nuclear translocation of STAT3 was associated with an increase in tyrosine phosphorylation of STAT3 (Fig. 2B). Before stimulation, no tyrosine-phosphorylated STAT3 was detected in the nucleus and a clear increase in intensity of the STAT3 tyrosine-phosphorylated band was obvious 5 min after stimulation. Strong STAT3 tyrosine phosphorylation was observed for up to 1 h after H-IL-6 stimulation. A weak signal, higher than the pretreatment level, was still detected for up to 12 h.
H-IL-6 stimulation of HepG2 cells also increased the DNAbinding of STAT3 versus its cognate DNA. As shown in Fig. 2C by gel shift experiments, new complex formation was detected 5 min after stimulation and no DNA binding was detected again 6 h after treatment with H-IL-6. STAT3 binding to the cognate DNA was confirmed by supershift experiments (Fig. 2D).
The results shown in Figs. 1 and 2 indicate that H-IL-6 triggers STAT3 translocation and haptoglobin gene transcription to maximal levels during the first hour after stimulation.
However, because of the accumulation of haptoglobin mRNA levels, maximal haptoglobin mRNA expression was found only 15 h after stimulation. These results indicate that up to this time point the increase in haptoglobin mRNA levels is higher than its subsequent degradation.
Stimulation of STAT3-dependent Gene Transcription via H-IL-6 and IL-6 -H-IL-6 stimulates nuclear STAT3 translocation, and this increase is associated with higher haptoglobin mRNA transcription. We performed further experiments to investigate whether H-IL-6 and IL-6 have comparable effects on STAT3-dependent gene transcription. Thus we transfected HepG2 cells with a STAT3-dependent reporter gene construct and stimulated the cells with 10 ng/ml of IL-6 or H-IL-6. Cells were harvested at different time points after stimulation (Fig.  3A). 1 h after stimulation the reporter activity in the cells The carrier solution consisted of 0.9% NaCl. The same membranes were hybridized with a probe for haptoglobin (HAP) and GAPDH. In panel B the relative changes in haptoglobin mRNA expression are shown in comparison with the GAPDH signal. The haptoglobin to GAPDH ratio before treatment was set to 1 and the relative changes compared with the pretreatment level were calculated accordingly. The transcription rate of the haptoglobin gene was studied by run-off experiments in HepG2 cells before and at different time points after 10 ng/ml H-IL-6 (panel C) stimulation. In panel D the relative changes are shown as fold induction compared with the pretreatment level.

FIG. 2. Nuclear STAT3 expression and DNA-binding in HepG2 cells after H-IL-6 injection.
Western blot analysis was performed with 10 g of nuclear extracts prepared before or at different time points after treatment of HepG2 cells with 10 ng/ml H-IL-6. Nuclear STAT3 (panel A) or tyrosine-phosphorylated nuclear STAT3 (panel B) expression was detected with antibodies directed against STAT3 or tyrosine-phosphorylated STAT3. Gel shift experiments were performed with 1 g of liver nuclear extracts prepared from HepG2 cells at different time points after 10 ng/ml H-IL-6 (panel C) stimulation. The nuclear extracts were incubated with a probe spanning the STAT3 site of the ␣2-macroglobulin promoter. The position of the bound (panel B) DNA is indicated. In panel D, supershift experiments were performed using HepG2 cell nuclear extracts 30 min after H-IL-6 stimulation. In lane 1 no antibody was added, whereas, in lanes 2 and 3, nuclear extracts were either incubated with 1 l of an anti-STAT1 (1 S1) or 1 l of an anti-STAT3 (1 S3) antibody.
activated with H-IL-6 was more than 10-fold higher compared with the pretreatment level and at least 2-fold higher compared with the level found in the IL-6-stimulated cells at the same time point. At all later time points H-IL-6 leads to at least 2-fold higher luciferase activity compared with the cells stimulated with IL-6. Because the half-life of the luciferase protein is around 3 h, the high luciferase activity found 24 and 48 h after H-IL-6 administration indicates that there is still an increased transcription rate after these intervals. Additionally, when the luciferase activity of the H-IL-6 treated cells was compared with those treated with IL-6, these results showed that at later time points the difference increased even more. Therefore, the absolute luciferase activity after H-IL-6 treatment at the different time points was set at 100%. The activity after IL-6 stimulation was expressed as percentage of the H-IL-6-treated cells at the same points. As shown in Fig. 3B, the relative percentage of the luciferase activity of IL-6-treated cells decreased with time (51% after 4 h versus 38% after 48 h), which indicates that after H-IL-6 stimulation the effect on gene transcription was not only stronger but also more prolonged compared with the IL-6 treatment.
In further experiments the dose-dependent effect of H-IL-6  (20) were transfected into HepG2 cells and stimulated with either 10 ng/ml IL-6 or H-IL-6 (panel A). The luciferase activity before stimulation was set to 1. The activity at different time points after stimulation is shown as fold induction compared with the pretreatment level. In panel B the luciferase activity after H-IL-6 stimulation at each time point was set to 100%. The relative activity after IL-6 stimulation was calculated in percent compared with the activity at each time point after H-IL-6 stimulation. In panel C increasing amounts of either H-IL-6 or IL-6 were administered and luciferase activity was measured 4 h after stimulation. The changes are shown as fold induction in comparison with the pretreatment level, which was set to 1. on STAT3-dependent gene transcription was assessed. The cells were harvested 4 h after stimulation. As shown in Fig. 3C in the range between 0.1 and 10 ng, a 10-fold higher dose of IL-6 compared with H-IL-6 leads to similar effect on gene transcription. When 100 ng/ml of each protein were used there was no further increase in gene transcription compared with 10 ng/ml of IL-6 or H-IL-6. However the difference on gene transcription between IL-6 and H-IL-6 remained the same (Fig. 3C).
STAT3-dependent Gene Transcription Activated Through H-IL-6 or IL-6 Depends on the gp130 and gp80/IL-6R Status-To further understand the mechanism that is responsible for the increased effect of H-IL-6 compared with IL-6 on gene transcription, we performed cotransfection experiments with either the gp130, gp80/IL-6R or both receptors (Fig. 4, A and B).
First, the cells were treated with H-IL-6 when increasing amounts of the gp130 receptor was cotransfected with the STAT3-dependent reporter gene construct (Fig. 4A). These experiments revealed that by increasing the amount of gp130, the luciferase activity could be further stimulated. A 30% increase in luciferase activity was found when 1 g of the gp130 plasmid was used. When the highest amount (4 g) of the gp130 plasmid was cotransfected the maximal luciferase activity dropped again but was still 15% higher compared with the level when no gp130 was cotransfected. In contrast to the gp130 experiments, cotransfection of the gp-80/IL-6R construct with the STAT3-dependent reporter plasmid leads to a direct reduction of luciferase activity after H-IL-6 stimulation compared with the level when no gp80/IL-6R was added. When gp130 and gp80/IL-6R were cotransfected and stimulated with H-IL-6, low amounts of the plasmids led to a reduction in luciferase activity compared with the cells that were not transfected with both receptors.
The same cotransfection experiments with the different receptor combinations were performed when the cells were stimulated with IL-6 ( Fig. 4B). Cotransfection of increasing amounts of the gp130 receptor with the STAT3-dependent reporter gene showed the following results compared with the initial activity when no receptor was added (control). 500 ng of gp130 receptor resulted in a 19% increase compared with control. In contrast, when 100, 1,000, or 4,000 ng were transfected, luciferase activity was Ϫ14%, Ϫ4%, and Ϫ28%, respectively.
Cotransfection experiments of lower amounts (100, 500, and 1,000 ng) of the gp80/IL-6R plasmid led to an increase in luciferase activity that was maximal with 500 ng (ϩ77%). However, when 4,000 ng of gp80/IL-6R were cotransfected, the activity of the STAT3-dependent reporter construct decreased again. A similar kinetic was found when both receptors were cotransfected. Low amounts of gp80 and gp130 increased, and high amounts decreased luciferase activity (Fig. 4B).
H-IL-6 Is more Active than IL-6 in Stimulating Acute Phase Gene Expression in Vivo-The in vivo relevance of H-IL-6 was still unclear. As our results in cell culture experiments showed, mainly that gp130 is essential for the interaction with H-IL-6, additional cells may bind the molecule. Therefore it seemed possible that in vivo activation of STAT3-dependent gene transcription in liver cells might be less effective. 4 g of H-IL-6 or 40 g of IL-6 were injected intraperitoneally into C3H mice according to the in vitro results where H-IL-6 was 10-fold more active. At different time points after injection IL-6 serum levels were determined in the liver vein, and haptoglobin mRNA expression was measured in the liver.
As shown in Fig. 5, A and B, haptoglobin mRNA increased to similar levels when 10-fold higher amounts of IL-6 were in-jected compared with H-IL-6. After normalization with the GAPDH signals, some of the time points (1, 6, and 12 h) determined after IL-6 stimulation were higher compared with the corresponding time points after H-IL-6 injection (Fig. 5C).
As H-IL-6 might have a different half-life in vivo compared with IL-6, the IL-6 serum concentration was measured in the liver vein (Fig. 6, A and B). The IL-6 serum concentration after H-IL-6 injection was more than 10-fold lower compared with the levels found after IL-6 administration 0.5 h after injection. At later time points, the difference between the H-IL-6-and IL-6-treated animals was even more dramatic. For example, after 3 h the IL-6 concentration in the H-IL-6-treated animals was more than 20-fold lower compared with the IL-6 treated mice.
The haptoglobin protein expression in the serum of the mice was determined by Western blot analysis. Time and dose kinetic experiments were performed with 20 g of serum. Four different doses of either IL-6 or H-IL-6 were used. As H-IL-6 was 10-fold more active in inducing haptoglobin mRNA expression, the dose kinetics were started with a 10-fold lower amount of H-IL-6 compared with IL-6. As shown in Fig. 7, A and B, the mice were stimulated for 4 h with 0.04, 0.4, 4, or 40 g of IL-6 and 0.004, 0.04, 0.4, or 4 g of H-IL-6. These results showed that a 10-fold lower amount of H-IL-6 had a similar effect on haptoglobin serum expression as did the higher IL-6 concentration (see Fig. 7, A and B).
The dose kinetic experiments indicated that H-IL-6 was 10fold more active than IL-6, therefore time kinetic experiments were performed with either 40 g of IL-6 or 4 g of H-IL-6. Fig.  7, C and D, shows that in vivo the 10-fold lower amount of H-IL-6 induced at least the same strength in haptoglobin expression in the time kinetic experiment as the higher amount of IL-6. Denistometric evaluation revealed that after 6 h, haptoglobin expression was nearly 2-fold after H-IL-6 compared with IL-6 stimulation. At later time points, the kinetic of 40 g of IL-6 and 4 g of H-IL-6 was not significantly different. Loading of the gels was checked by Coomassie staining (data not shown). Therefore, the experiments as shown in Figs. 5-7 demonstrated that in vivo H-IL-6 was at least 10-fold more active in inducing haptoglobin mRNA and protein levels.
In Vivo H-IL-6 Induces Stronger Nuclear Translocation of STAT3 Compared With IL-6 -In further experiments, we investigated whether mainly STAT3 is involved in regulating higher haptoglobin mRNA levels. Thus, the nuclear translocation and tyrosine phosphorylation of STAT3 was studied in liver nuclei. Liver nuclear extracts were prepared from the mice that were either treated with 4 g of H-IL-6 or 40 g of IL-6.
Western blot analysis was performed using anti-STAT3 and anti-phospho-STAT3 antibodies. H-IL-6 and IL-6 induced an increase in nuclear STAT3 expression and tyrosine-phosphorylated STAT3 30 min after injection (Fig. 8). In the H-IL-6 treated animals at later time points, the nuclear concentration of STAT3 continuously fell and reached pretreatment levels 6 h after injection. After IL-6 injection, the nuclear concentration of phospho-STAT3 remained high for the first 2 h. At later time points the nuclear concentration decreased, and no nuclear tyrosine-phosphorylated STAT3 was found after 12 h again. The time kinetic of the nuclear expression of tyrosine-phosphorylated STAT3 was, therefore, closely linked to the kinetic found for the IL-6 serum levels (Fig. 6).
The nuclear translocation of STAT3 is associated with an increased affinity toward its cognate DNA motif. Perhaps additionally posttranscriptional modifications might be involved in mediating DNA binding of STAT3 and target gene transcrip- The same membranes were hybridized with a probe for haptoglobin (HAP) and GAPDH. In panel C the relative changes in haptoglobin mRNA expression are shown in comparison with the GAPDH signal. The haptoglobin to GAPDH ratio before treatment was set to 1, and the relative changes compared with the pretreatment level were calculated accordingly. tion (23,24). Therefore, gel shift experiments were performed with nuclear extracts to study whether H-IL-6 also triggers higher DNA binding. New complex formation was found 30 min after H-IL-6 and IL-6 injection (Fig. 9, A and B). As shown by supershift experiments, the complex was completely supershifted by an anti-STAT3 antibody (Fig. 9C). Increased DNA binding of STAT3 after IL-6 and H-IL-6 injection closely followed the kinetic of the nuclear translocation of tyrosine-phosphorylated STAT3. After H-IL-6 injection STAT3 showed its strongest affinity toward its cognate DNA after 30 min, and a decrease in DNA-binding was found at later time points. No complex formation was detected at the 6 h time point. After IL-6 treatment, a different time kinetic was found. The intensity of the STAT3 complex was lower compared with H-IL-6 stimulation after 1 h and remained high for up to 4 h. The DNA-binding of STAT3 slightly decreased only after this time point, and no complex formation was detected after 12 h. Additionally, after H-IL-6 stimulation a second faster migrating complex was found 1 h after injection. As IL-6 can also activate STAT1 (25), DNA-binding competition experiments were performed with antibodies that either detect STAT1-␤ or both STAT1 isoforms (␣ and ␤) (Fig. 9D). The STAT1 antibody which detects both STAT1 isoforms completely competed the new complex, which was found 1 h after H-IL-6 treatment. In contrast, the anti-STAT1-␤ antibody had only a minor influence on the intensity of the new complex. Therefore, these results indicated that in vivo H-IL-6 also stimulated nuclear translocation and DNA-binding of STAT1-␣. DISCUSSION IL-6 is a pleiotropic cytokine with several essential functions involved in the induction of immunoglobulin secretion of B cells, the maturation of megakaryocytes, colony formation of hematopoietic stem cells, and growth and differentiation of T cells (9,17,26). IL-6 is essential for the activation of the acute phase response in the liver and for triggering liver regeneration (13,27). Therefore, designer molecules that activate IL-6-dependent intracellular signals would offer the possibility of therapeutic intervention during different pathophysiological conditions.
In this study, we investigated whether H-IL-6 might activate STAT3-dependent gene transcription in vitro and in vivo and whether the gp130 to gp80/IL-6R status on the cell surface might influence STAT3-dependent gene transcription. H-IL-6 enhances transcription of the haptoglobin gene in HepG2 cells, which is associated with higher nuclear translocation, tyrosine phosphorylation, and DNA binding of STAT3. These results show that the designer protein H-IL-6 activates homodimerization of two gp130 molecules and phosphorylation of the intracellular tyrosines, which trigger STAT3 activation. In run-off experiments and by using a STAT3-dependent reporter gene construct, we show that the same amount of H-IL-6 induces 2-3-fold higher gene transcription compared with IL-6. When the cells were transfected with the STAT3-dependent reporter construct and stimulated with increasing amounts of H-IL-6 or IL-6, a 10-fold lower amount of H-IL-6 led to the same luciferase activity as the higher IL-6 concentration. Therefore H-IL-6 leads to 10-fold stronger activation of STAT3-dependent gene transcription. Additionally, differences were found in the time kinetic. 1 h after stimulation the luciferase activity in H-IL-6treated cells was 3-fold higher compared with IL-6. After 4 h the difference was 2-fold and at later time points the difference between the two molecules increased again. These differences in the kinetic of STAT3-dependent gene transcription after H-IL-6 and IL-6 stimulation is most likely explained on the receptor level.
Recent experiments published by Somers et al. (28) suggested that the first event in IL-6 signal transduction is the binding of IL-6 through site1 to gp80/IL-6R. The second event is the binding of this heterodimer to the gp130 molecule, which results in the formation of a trimeric complex on the surface of the cell. The third event is the association of two trimeric to a hexameric complex, which results in intracellular dimerization of two gp130 molecules and the activation of intracellularsignaling cascades (28). The first step is characterized by binding of two specific and low affinity partners. In the case of H-IL-6, this first event is not necessary. This advantage may trigger faster assembly and gp130 dimerization, which results in a more rapid STAT3 phosphorylation and luciferase activity as shown in the cell culture experiments. A second mechanism, which might explain the more rapid effect on gene transcription after H-IL-6 compared with IL-6 stimulation, is the observation that IL-6 might form IL-6/IL-6 dimers. The IL-6 homodimer has higher affinity toward gp80/IL-6R; however, a complex consisting of two IL-6 and gp80/IL-6R molecules is less potent in complex formation with gp130 and inducing STAT3 phosphorylation (29). Therefore, H-IL-6 might preferentially first lead to a monomeric IL-6⅐gp80⅐gp130 complex, which more rapidly induces hexamer formation, STAT3 phosphorylation, and thus gene activation.
Besides the faster early events, H-IL-6 has also a more pronounced effect on gene transcription compared with IL-6 as assessed by haptoglobin Northern blot analysis, run-off assays, and STAT3-dependent luciferase reporter gene assays in HepG2 cells. The hexameric IL-6⅐gp80⅐gp130 complex at the surface of hepatocytes is dissolved by internalization and degradation of IL-6 (30,31). A di-leucine internalization motif in the cytoplasmic domain of gp130 seems essential for controlling IL-6 internalization and degradation (32)(33)(34). The internalization process of H-IL-6 is most likely comparable with the one of IL-6 and soluble IL-6R/gp80. However in H-IL-6 the linker peptide, which fuses IL-6 and soluble IL-6R/gp80, could lead to less effective internalization that results in a more prolonged effect on gene transcription. The second mechanism that may account for the longer effect on gene transcription is based on the three-step model during receptor assembly. The first step with low affinity binding is not necessary for H-IL-6. The sec- For time kinetic studies, serum was used before treatment (0) or at different time points (6,12,24, and 48 h) after stimulation with either IL-6 (C) or H-IL-6 (D). The position of the specific haptoglobin (HAP) signal is shown.
FIG. 8. Detection of nuclear STAT3 and nuclear tyrosine phosphorylated STAT3 after IL-6 and H-IL-6 injection in vivo. Western blot analysis was performed with 10 g of liver nuclear extracts prepared before and at different time points after treatment of C3H mice with 40 g of IL-6 (A) or 4 of g H-IL-6 (B). Nuclear STAT3 or tyrosine-phosphorylated nuclear STAT3 expression was detected with antibodies directed against STAT3 or tyrosine-phosphorylated STAT3. ond step is the essential interaction for H-IL-6 and results in high affinity binding between H-IL-6 and gp130. Because of the high affinity between the two partners the on/off rate at the receptor level is shifted toward longer assembly of an active complex on the cell membrane, which results in a more prolonged effect on STAT3-dependent gene transcription.
The interaction of IL-6 with the soluble gp80/IL-6R is also relevant in vivo (35,36). The association of IL-6 and soluble gp80/IL-6R renders cells that express only gp130 sensitive for IL-6 and this mechanism has been termed transignaling (37,38). H-IL-6 closely resembles the association between soluble gp80/IL-6R and IL-6; therefore, the experiments where gp130 and/orthegp80/IL-6RmoleculewerecotransfectedwithaSTAT3dependent reporter gene construct might also have direct implications for the physiological assembly of soluble gp80 and IL-6. These results show that the receptor status of a specific cell influences the effect on gene transcription induced by H-IL-6. When the STAT3-dependent reporter gene construct was cotransfected with the gp80/IL-6R molecule, the H-IL-6-dependent effect on gene transcription was reduced, whereas higher luciferase activity was found after IL-6 stimulation. In contrast, cotransfecting the gp130 molecule had the opposite effect. Therefore the availability of gp130 molecules on the cell surface is the rate-limiting step for inducing STAT3-dependent gene transcription. These results also suggest that in cells which only express gp130 molecules, STAT3-dependent gene transcription can be stimulated, which might explain the observation that hematopoietic progenitor cells strongly respond to H-IL-6 but not to IL-6 stimulation (17). Thus our cotransfection experiments with the gp80 and gp130 receptors clearly show that the target cells, which can be stimulated by IL-6 and H-IL-6, are not absolutely comparable. A higher ratio toward the gp130 receptor would render the cell more susceptible toward H-IL-6 but not IL-6, which results in higher STAT3-de- FIG. 9. DNA binding of STAT3 after IL-6 or H-IL-6 stimulation in vivo. Gel shift experiments were performed with 1 g of liver nuclear extracts prepared from animals before and at different time points after 40 g of IL-6 (A) or 4 g of H-IL-6 (B) stimulation. The nuclear extracts were incubated with a probe spanning the STAT3 site of the ␣2-macroglobulin promoter. The position of the free (F) and bound (B) DNA is indicated. For supershift experiments, nuclear extracts were used from animals either treated for 1 h with IL-6 (C) or H-IL-6 (D). Nuclear extracts were incubated with increasing amounts of a STAT3 (S3) antibody and antibodies that either detect STAT1-␤ (S1-␤) or the C-terminal domain of STAT1-␣ ϩ ␤ (S1). pendent gene transcription. In contrast, cells with a 1:1 gp130/ gp80 ratio would be less ideal target cells for H-IL-6.
H-IL-6 triggers STAT3-dependent gene transcription also in vivo. The IL-6 serum levels show that the decay of the H-IL-6 protein is not dramatically reduced compared with IL-6. Additionally, these experiments indicate that even without the specificity of IL-6, which needs gp80 and gp130 for trimer formation, most of the protein is still very active in the liver. A 10-fold lower dose and serum levels of H-IL-6 have a similar effect on activating STAT3 and haptoglobin mRNA and protein levels compared with IL-6. Earlier reports showed that injection of the gp80/IL-6R protein into rats lead to an enrichment of the protein mainly in liver, muscle, skin, and kidney (39). Because H-IL-6 consists of IL-6 and soluble gp80/IL-6R it may have a similar tissue distribution as the gp80/IL-6R molecule alone. This characteristic would help to specify potential target tissues in vivo, where H-IL-6 might have a potential therapeutic effect.
Because H-IL-6 is more active for inducing STAT3-dependent gene transcription in the liver, several therapeutic implications are evident. The induction of acute phase proteins contributes to a first line of defense during infection (9,26). H-IL-6 would induce these defense mechanisms stronger and more rapidly, and therefore might help to eliminate infectious agents, especially in immunosuppressed patients. Besides mediating the activation of acute phase genes, IL-6 seems involved in activating pathways that are important to trigger cell proliferation and anti-apoptotic mechanism (6,40,41). Hepatocytes of IL-6-deficient and type I tumor necrosis factor receptor-deficient mice are unable to proliferate after partial hepatectomy and recombinant IL-6 is able to restore cell proliferation and thus liver regeneration (13,14). Additionally, administration of IL-6 before Con A injection prevents hepatocytes from undergoing tumor necrosis factor-induced apoptosis (42). Therefore, H-IL-6 seems a promising designer protein to influence several pathophysiological conditions of the liver in vivo and thus might also be important for later therapy in humans.