Abrogation of Viral Interleukin-6 (vIL-6)-Induced Signaling by Intracellular Retention and Neutralization of vIL-6 with an Anti-vIL-6 Single-Chain Antibody Selected by Phage Display

Cells and reagents.COS-7 and HepG2 cells were bought from the ATCC (Rockville, MD) and were routinely grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS). The vIL-6 DNA sequence was originally obtained from F. Neipel and B. Fleckenstein (University of Erlangen, Germany) (48) and has been used in previous studies (26, 46). The PEL-derived BCBL-1 cells were a kind gift from M. Gramatzki's group at the University of Kiel (Germany) and were maintained in RPMI 1640 medium (BioWhittaker,Walkersville, MD) supplemented with 2 mmol/liter l-glutamine (Life Technologies, Grand Island, NY), 10% fetal calf serum (BioWhittaker), and 5 μg/ml gentamicin (Life Technologies). The Ba/F3 cell lines (Ba/F3-gp130, IL-6-R, and Ba/F3-gp130) used in this study were described previously (22, 55). EBNA-293 cells were used as described elsewhere previously (71). The human single-framework scFv libraries Tomlinson A and Tomlinson B (T. Tomlinson, MRC, Centre for Protein Engineering, Cambridge, United Kingdom) were provided by the MRC (see below). The anti-mouse alkaline phosphatase conjugate was bought from Sigma-Aldrich (Munich, Germany), the polyclonal rabbit antibody anti-phospho-STAT3 (Tyr705) and mouse anti-STAT3 were obtained from Cell Signaling Technology (Beverly, Calif.), and polyclonal horseradish peroxidase coupled to goat anti-rabbit and anti-mouse antibodies was obtained from Pierce (Erlangen, Germany). Anti-mouse Alexa Fluor 594 antibody was obtained from Molecular Probes (Invitrogen). The anti-vIL-6 rabbit serum was described previously (26). [³H]thymidine was obtained from Amersham-Pharmacia (Freiburg, Germany). Recombinant cytokines used in this study were produced as described previously (21). Soluble gp130Fc is a fusion protein of the extracellular portion of gp130 and the Fc portion of IgG1 and was expressed and purified as previously described (31).

Cloning of MAV-KDEL, vIL-6-green fluorescent protein (GFP), vIL-6-KDEL, and vIL-6.Standard cloning procedures were performed as described previously by Sambrook et al. (70). The ER retention signal KDEL was added to the C terminus of MAV by PCR (MK404 [5′-GTGGTTTAAACCTAGAGTTCGTCTTTCCGCGGATTCAGATCCTCTTCTGA-3′] and MK304 [5′-ATG GCC GAG GTG CAG CTG-3′]). The signal peptide needed for eukaryotic expression of MAV-KDEL was amplified from human IL-6 (MK104 [5′-CA CGT TTA AAC GCC ACC ATG AAC TCC TTC TCC ACA-3′] and MK704 [5′-CAA CAG CTG CAC CTC GGC CAT CAC GTG TAC TGG GGC AGG GAA GGC-3′]), fused to MAV-KDEL by slicing by overlapping extension PCR (25, 28), and cloned into vector p409B. The resulting plasmid was named p409B-MAV-KDEL. vIL-6 was amplified by PCR (MK504 [5′-CA CTA GCT AGC GCC ACC ATG TGC TGG TTC AAG TTG-3′] and MK604 [5′-CCG CTC GAG CTT ATC GTG GAC GTC AGG-3′], template p409-vIL-6) and inserted into plasmid pcDNA3.1-GFP via NheI and XhoI digestion, leading to pcDNA3.1-vIL-6. To generate an ER-retained vIL-6 expression construct, vIL-6 was amplified by PCR (5′-TAC TAC TCG AGG CCA CCA TGT GCT GGT TCA AGT TG-3′ and 5′-TAT TGC GGC CGC CTT ATC GTG GAC GTC AGG-3′) and inserted into pDCC-MAV-KDEL opened with XhoI and NotI, leading to pDCC-vIL-6-KDEL. As a control, vIL-6 was also amplified by PCR (5′-TAC TAC TCG AGG CCA CCA TGT GCT GGT TCA AGT TG-3′ and 5′-CGCGGATCCCTACAGATCCTCTTCTGAGAT-3′) and cloned into pDCC-vIL-6-KDEL opened with XhoI and BamHI, leading to pDCC-vIL-6.

Expression of recombinant vIL-6.EBNA cells expressing vIL-6 were grown in culture medium containing 10% FCS to 80 to 100% confluence. The medium was then changed to FCS-free medium, and expressed vIL-6 protein was collected in the supernatant for 7 days. Fifty milliliters of cell culture supernatant was centrifuged and incubated at room temperature (RT) with 300 μl Ni-nitrilotriacetic acid (NTA) agarose (QIAGEN, Hilden, Germany) for 2 h. The agarose was washed with phosphate buffer containing 500 mM NaCl at pH 7.4, and recombinant vIL-6 was eluted with the same buffer containing 0.5 M imidazole and dialyzed to 50 mM Tris buffer, pH 7.4, without imidazole.

Phage display.The phage display screening was done as described previously (41). Briefly, panning of two phage libraries (A and B; MRC) was done in 96-well microtiter plates (Nunc Maxisorb) coated at 4°C overnight with 1 μg/well purified Flag-vIL-6-His. Three rounds of biopanning were carried out against vIL-6 immobilized on microtiter plates. Eluted phages were used to infect TG1 cells and were rescued with M13Ko7 helper phage. The culture supernatant, containing the polyclonal phage populations from each round of biopanning, was used to assess the binding to vIL-6 by enzyme-linked immunosorbent assay (ELISA). To assess the binding of monoclonal phage scFvs, single colonies from titration plates of the third round were cultured and analyzed by ELISA. Clones that gave the highest signals were selected to infect Escherichia coli HB2151 cells to express soluble scFvs. Infected HB2151 cells were induced by IPTG (isopropyl-β-d-thiogalactopyranoside), and the supernatant containing the soluble scFvs was analyzed by ELISA.

ELISAs.Microtiter plates were coated with 10 μg/ml vIL-6 (or other antigens for cross-reactivity ELISA) or 1 μg/mls gp130Fc for the sandwich ELISA and incubated at RT overnight. The wells were blocked with 3% BSA for 2 h at RT. Specific binding of scFvs to vIL-6 or to the vIL-6/sgp130Fc complex (sandwich ELISA) or of Flag-tagged vIL-6 to sgp130Fc was detected by ELISA with an anti-c-myc antibody or with an anti-Flag antibody, respectively, and anti-mouse antibodies conjugated with alkaline phosphatase followed by incubation with substrate (1 mg/ml p-nitrophenylphosphate in diethanolamine buffer). The color intensity was measured as extinction at 405 nm using an ELISA reader.

Production and purification of soluble MAV.The MAV coding sequence was cut out from pIT1-MAV (library vector) by NcoI and NotI and inserted into the pET22b vector via NcoI and NotI. pET22b encodes the PelB leader sequence, which is a signal peptide for periplasmic expression of the gene of interest. The antibody sequence together with the PelB leader was then cloned into vector pET23a-SM12 (our unpublished results) by XbaI and NotI, so that MAV obtained C-terminal c-myc and His tags. E. coli BL21 cells were transformed with vector pET23a-MAV. Transformed E. coli BL21 cells were grown with shaking at 37°C to an optical density at 600 nm of 0.5. Periplasmic expression of MAV was induced by adding 0.4 mM IPTG to cells and shaking them overnight at 30°C. The bacterial culture was centrifuged at 6,000 × g for 20 min. The supernatant containing soluble scFv was purified with Ni-NTA agarose. One hundred milliliters of the bacterial supernatant was incubated with 0.5 ml Ni-NTA agarose (QIAGEN, Germany) at RT for 2 h. The matrix was then washed with 50 mM phosphate buffer, pH 7.4, containing 500 mM NaCl, and bound MAV was eluted with 0.5 M imidazole in the same buffer and dialyzed to 50 mM Tris buffer, pH 7.4, without imidazole.

CD spectroscopy.Circular dichroism (CD) measurements were carried out using a Jasco 720 CD spectrometer equipped with a temperature control unit and calibrated with a 0.1% aqueous solution of d-10-camphorsulfonic acid according to the method described previously by Chen and Yang (14). The spectral bandwidth was 1.5 nm. The time constant was 2 s, and the cell path length was 0.1 mm.

Calculation of protein concentrations.Protein concentrations were calculated from absorption spectra in the range of 240 to 320 nm using the method described previously by Waxman et al. (79).

Western blot analysis.For Western blotting of cell lysates, cells from a 10-cm dish were lysed with 1 ml 2× Laemmli buffer (36), 10 μl of which was applied to a sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel. To analyze the proteins of the cell supernatant, 15 μl out of 5 ml was used. Proteins separated by SDS-PAGE were transferred for 1 h at 1 mA/cm² to polyvinylidene difluoride membranes. The membranes were blocked for 1 h in 6% bovine serum albumin (BSA)-Tris-buffered saline (TBS). The filters were then incubated with an appropriate antibody before being labeled with a secondary antibody coupled to peroxidase. Subsequently, the membranes were developed with the Amersham enhanced chemiluminescence kit in accordance with the manufacturer's instructions. Band intensities were estimated from photographic films by densitometry using ImageJ software (National Institutes of Health).

Immunoprecipitation of vIL-6 with MAV.COS-7 cells were transiently transfected with p409-vIL-6 using the DEAE-dextran method (42, 64). The next day, DMEM containing 10% FCS was changed to 5 ml FCS-free DMEM. After 24 h, the supernatant was removed and centrifuged at 10,000 × g for 2 min to remove cell debris. One milliliter of supernatant containing vIL-6 was incubated with 15 μg of MAV at 4°C overnight under continuous rolling. Immunocomplexes were collected with 25 μl Talon beads (Dynal) during incubation at 4°C for 2 h. After washing with phosphate-buffered saline (PBS)-Tween and PBS buffers, the MAV/vIL-6 complexes were eluted with 20 μl 0.5 M imidazole in PBS. The eluates were separated by SDS-PAGE, and Western blotting was done using anti-vIL-6 serum as described above.

Surface plasmon resonance.Forty micrograms of recombinant vIL-6 in 10 mM Na-acetate, pH 5.5, was immobilized onto a surface plasmon resonance sensorchip (IAsys). The binding experiment was performed at room temperature with different concentrations of MAV using the IAsys optical biosensor. Affinograms were analyzed by nonlinear regression with FASTfit software, which uses the Marguard-Levenburg algorithm for iterative data fitting (54).

Immunostaining of transfected COS-7 cells.Approximately 40,000 COS-7 cells grown on glass coverslips for 24 h were washed with PBS containing 0.1% acid and fixed with cold methanol at −20°C for 10 min. The cells were then washed two times with TBS and incubated with blocking buffer (1% BSA, 0.2% glycerine, TBS) at room temperature for 2 h. A 1:100 dilution of the anti-c-myc antibody was applied at 4°C overnight. The cells were washed with TBS-Tween and TBS and incubated with the Alexa Fluor 594 anti-mouse antibody at 37°C for 1 h. The washing was repeated as before, and the glass coverslips were mounted onto slides with SlowFade Light (Molecular Probes) and analyzed using fluorescence microscopy. A ×543 magnification was used to photograph the cells.

Proliferation assay.BCBL-1 and Ba/F3-gp130 cells were washed three times with sterile PBS and resuspended in RPMI medium containing 10% FCS at 5 × 10³ cells per well of a 96-well plate. The cells were cultured in a final volume of 100 μl for 72 h and subsequently pulse-labeled with 0.25 μCi [³H]thymidine for 4 h. The specific activity of [³H]thymidine was 25 Ci/mmol. Cells were harvested on a glass fiber filter (Filtermat A; Wallac, Turku, Finland), and the filter was baked in a microwave oven for 2 min. Subsequently, the filter was soaked in 4.5 ml liquid scintillation cocktail (Betaplate Scint; Wallac, Turku, Finland). [³H]thymidine incorporated into cellular DNA was determined by scintillation counting using a MicroBeta TriLux counter from Perkin-Elmer (Wellesley, MA). Bioassays were performed, with each value being determined in triplicate.

Isolation of human recombinant antibody binding to viral IL-6.The modular episomal vector pCEP-Pu (34), based on the Epstein-Barr virus genome, was used for the expression of vIL-6. The expression constructs contained the vIL-6 sequence with either an N-terminal His tag or with two affinity tags, an N-terminal Flag tag and a C-terminal His tag. Both variants of vIL-6 were produced by stably transfected EBNA-HEK293 cells, and the vIL-6 variants were purified from the cell culture supernatants with Ni-NTA (Fig. 1A). The correct folding of the recombinant vIL-6 proteins was confirmed by CD spectroscopy. vIL-6 expressed and purified from EBNA-293 cells displayed a typical α-helical spectrum (Fig. 1B). The purified vIL-6 variants were biologically active, as shown by the induction of proliferation of the cytokine-dependent murine pre-B-cell line Ba/F3-gp130 (Fig. 1C). Native Ba/F3 cells proliferate in the presence of the cytokine IL-3. After stable transfection of a cDNA coding for gp130, these cells (Ba/F3-gp130), in the absence of IL-3, gain the ability to proliferate in the presence of IL-6/sIL-6-R, hyper-IL-6 (a fusion protein of IL-6 and sIL-6-R) (21), or viral IL-6. Previous studies have shown that Ba/F3-gp130 cells proliferate in response to hyper-IL-6 at a concentration below 1 ng/ml, whereas 100 ng/ml of viral IL-6 was needed to induce Baf/3-gp130 cell proliferation (26, 46).

In this study, Flag-vIL-6-His and His-vIL-6 expressed in EBNA-293 cells induced the proliferation of Ba/F3-gp130 cells at similar concentrations. In contrast, bacterially expressed and refolded viral IL-6 protein was not correctly folded, did not show the typical CD spectrum of a four-helical protein (Fig. 1B), and did not induce proliferation of Ba/F3-gp130 cells (Fig. 1C). Biologically active vIL-6 has been produced in E. coli previously with a different expression strategy (26). In EBNA-293 cells, the expression level of Flag-vIL-6-His was threefold higher than that of His-vIL-6. Thus, Flag-vIL-6-His was produced in milligram amounts to be used for the screening of a scFv phage library. The correct processing of the signal peptide was verified by N-terminal sequencing of the purified secreted protein (data not shown).

For the production of recombinant antibodies against vIL-6, phage display technology was chosen (for a review, see reference 17), and the Tomlinson A and B libraries (T. Tomlinson, MRC, Center for Protein Engineering, Cambridge, United Kingdom), comprising 10⁹ scFv fragments, were used. The libraries are based on a single human framework for the variable regions of heavy (V_H) and light (V_L) human immunoglobulin chains, in which the CDR3 loops have been randomized in sequence and length (for a detailed protocol, see http://www.mrc-cpe.cam.ac.uk/g1p.php?page=1808 ). After three rounds of panning and selection, clones were tested to bind to vIL-6 as soluble scFvs. Only two clones did not cross-react with other irrelevant proteins (data not shown). Sequencing of these two independently selected clones revealed that they were identical, suggesting that in the course of two different selection experiments, the same antibody fragment had been isolated. This underlines the specificity of the screening process. Thus, we obtained one vIL-6-specific scFv antibody, named MAV.

Expression, purification, and characterization of MAV by Western blot, immunoprecipitation, and ELISA and in a vIL-6-phospho-STAT3 assay.Next, MAV was produced in milligram amounts by periplasmic expression of the recombinant protein in E. coli and purified using Ni-NTA agarose (Fig. 2A).The purified recombinant antibody fragment MAV was analyzed by ELISA for specific binding to vIL-6. Different antigens were coated at a concentration of 10 μg/ml and 0.5 μg MAV per well. MAV bound to Flag-vIL-6-His and His-vIL-6 but did not exhibit any cross-reactivity with other antigens such as huIL-6, His-tagged OSM, BSA, lysozyme, ABI3 (66), and sgp130Fc (Fig. 2B). Because Flag-vIL-6-His was used as a coating antigen for the panning, it was important to test the MAV reactivity with the Flag and His tags. Specific binding of MAV to His-vIL-6 suggested that MAV did not recognize the Flag tag. Only Flag-vIL-6-His and His-vIL-6 were recognized by the scFv antibody MAV. Also, the lack of binding to the His-tagged OSM suggested that there was no cross-reactivity with the His tag.

MAV binding to vIL-6 was also shown by Western blotting. MAV was used as the primary antibody recognizing vIL-6. One hundred nanograms of vIL-6 protein could be detected by MAV (Fig. 2C). Next, we investigated whether MAV bound to vIL-6 in solution. Therefore, COS-7 cells were transiently transfected with an expression plasmid encoding vIL-6, and the supernatants of transfected as well as untransfected cells were incubated with MAV. vIL-6 bound to MAV was coprecipitated by the use of Talon beads binding the His-tagged proteins. The precipitates were analyzed by Western blotting (Fig. 2D) with a vIL-6-specific antiserum (26). The fact that vIL-6 was precipitated with MAV suggested that MAV recognized vIL-6 in solution.

The kinetics of MAV binding to vIL-6 immobilized on a solid phase were analyzed by surface plasmon resonance. Hence, a plot of multiple K_on values, derived from a full interaction experiment, allowed the determination of k_ass (gradient) and k_diss (intercept), which were 14,500.5 M⁻¹ s⁻¹ and 0.005012 s⁻¹, respectively. In this way, the affinity of MAV for vIL-6 was defined as follows: dissociation equilibrium constant = k_diss/k_ass = 340 nM (data not shown).

To identify the epitope of vIL-6 that is recognized by MAV, several chimeric proteins of human and viral IL-6 were used (B. Rabe and J. Scheller, unpublished results). huIL-6 as well as vIL-6 have a four-helical structure (24). gp130 receptor binding sites of these cytokines are defined as site II, composed of helix A and helix C, and site III, consisting of the initial part of the AB loop and helix D. The chimeric proteins represent the huIL-6 sequence with the exchange of sites II and III of vIL-6 (Fig. 3). Because MAV specifically binds to vIL-6 but not to huIL-6, these chimeras can be used to define which region of vIL-6 is recognized by MAV (1, 2). As shown by Western blotting, MAV recognized vIL-6 and the chimera II+III, which has the complete receptor binding sites II and III of vIL-6 (Fig. 3). Because MAV did not bind to the chimeras A, C, and A+C, one can conclude that the binding epitope of vIL-6 to MAV is not on site II. MAV binding to chimera II+III suggests that site III of vIL-6 can be the recognition site for MAV. However, the AB loop of site III does not seem to be involved in the binding of MAV, because MAV does not bind to the AB chimera. Therefore, we assume that the D helix of site III of vIL-6 is necessary for recognition by MAV. The lower signal intensity of MAV recognizing the chimera II+III compared to that of wild-type vIL-6 indicates that MAV might bind to a conformational epitope rather than to a simple linear epitope. This was further supported by ELISA data, in which folded vIL-6 and the folded chimera II+III were detected with comparable efficiencies (data not shown).

To ask whether MAV binds to a vIL-6-sgp130Fc complex, a sandwich ELISA was performed. First, binding of Flag-tagged vIL-6 to sgp130Fc was shown by anti-Flag antibodies. After coating with sgp130Fc protein, vIL-6 and then MAV were added. Interestingly, vIL-6 bound to sgp130Fc was not detected by MAV, whereas vIL-6 bound directly to the surface of the well was readily detected (Fig. 4A). We conclude from these experiments that the binding sites of MAV and of sgp130Fc overlap and that MAV might therefore have neutralizing properties.

Therefore, we tested the effect of MAV on vIL-6-induced STAT3 phosphorylation of HepG2 cells. vIL-6 binds to the gp130 receptor, leading to JAK/STAT activation (46). Activated STAT3 was detected with an anti-phospho-STAT3 antibody. HepG2 cells were stimulated with 150 ng/ml vIL-6 in the presence or absence of increasing concentrations of MAV. In fact, quantification by densitometry revealed that MAV reduced STAT3 phosphorylation from 100% to 18% in a dose-dependent manner, indicating the neutralizing potential of the scFv MAV protein (Fig. 4B). It should, however, be noted that such a quantification is only an approximation and not an absolutely accurate method to measure the amount of phospho-STAT3 protein.

The proliferation of the vIL-6-dependent PEL-derived cell line BCBL-1 can be inhibited by recombinant MAV.PEL is a distinct type of lymphoma associated with Kaposi's sarcoma-associated herpesvirus infection. Among others, vIL-6 has been described to be a factor responsible for the unrestrained proliferation of the PEL-derived cell line BCBL-1. BCBL-1 cells were found to release huIL-6 and vIL-6 in culture supernatant (30). Here, we have tested whether recombinant MAV can inhibit the proliferation of BCBL-1 cells. For a control, we have used recombinant sgp130Fc, which has previously been shown to specifically inhibit vIL-6-mediated signaling and transsignaling mediated by huIL-6 in complex with soluble IL-6-R (46). sgp130Fc consists of the complete extracellular part of gp130 (domains 1 to 6) fused to a dimerization domain (Fc part of a human IgG antibody) (31). As depicted in Fig. 5, the proliferation of BCBL-1 cells could be blocked by MAV and sgp130Fc by more than 50%. We conclude from this experiment that MAV specifically inhibits vIL-6-mediated growth of the PEL-derived BCBL-1 cell line.

ER retention of viral IL-6 by scFv MAV-KDEL abrogates vIL-6 signaling.Since it was previously reported that vIL-6 can act from within the cell, presumably by binding to gp130 in the ER (43), we aimed to use MAV to retain and neutralize vIL-6 within the ER. To clone a construct for eukaryotic expression of MAV-KDEL, the signal peptide (SP) of human IL-6 was added to the N terminus and a DNA sequence coding for the ER retention signal KDEL was added to the C terminus of MAV. SP-MAV-KDEL was transfected into COS-7 cells, and MAV-KDEL expression was analyzed by Western blotting of the cell lysate and the cell supernatant. As expected, MAV-KDEL was detected only in the cell lysate and was not detectable in the cell supernatant (Fig. 6A).

We further constructed a vIL-6-GFP fusion protein in which GFP was fused at the C terminus of vIL-6. Using an anti-vIL-6 serum for Western blot analysis, vIL-6-GFP was detected in the supernatant as well as in the cell lysate of transfected COS-7 cells (Fig. 6B). The double band corresponding to vIL-6-GFP in the supernatant most likely reflects differentially glycosylated forms of vIL-6-GFP (compare Fig. 6B and 2D).

Cotransfection of vIL-6-GFP and MAV-KDEL into COS-7 cells should result in the intracellular binding of vIL-6-GFP to MAV-KDEL and in the retention of MAV-KDEL/vIL-6-GFP in the endoplasmic reticulum. This was demonstrated by Western blotting of the lysates and supernatants of the cotransfected COS-7 cells (Fig. 6B). Only a very faint band corresponding to vIL-6-GFP was detected in the cell supernatant, suggesting that a very small amount was secreted in the presence of MAV-KDEL.

The expression of MAV and vIL-6-GFP and their localization within the cell were monitored by immunostaining of transfected COS-7 cells. As expected, MAV-KDEL was seen in the region of the endoplasmic reticulum, and no plasma membrane staining was detected (Fig. 6C, panel IV). Strong fluorescence of the ER suggested a high expression level of vIL-6-GFP (Fig. 6C, panel II). Secretion of vIL-6-GFP was seen in vesicles as green dots (Fig. 6C, panel II, white arrows) and observed as green fluorescence in the region of the plasma membrane.

In double-transfected cells, intracellular colocalization of vIL-6 and MAV-KDEL could be observed as yellow staining caused by the overlay of green (vIL-6-GFP) and red (MAV-KDEL-Alexa Fluor 594) fluorescence (Fig. 6C, compare panels V, VI, and VII).

Finally, we asked whether ER retention of vIL-6 in MAV-KDEL cotransfected cells would lead to reduced STAT3 phosphorylation (Fig. 6D). Therefore, we transfected COS-7 cells with 0.5 μg and 0.25 μg of the expression plasmids coding for vIL-6 and 3 μg of expression plasmids coding for MAV-KDEL or GFP. We found a clear reduction of vIL-6-induced STAT3 phosphorylation in the presence of MAV-KDEL, indicating that the retention of vIL-6 in the ER led to the inhibition of STAT3 phosphorylation. Quantification of the phospho-STAT3 Western blots by densitometry showed that we achieved a reduction of 30% when 0.5 μg vIL-6 and MAV-KDEL plasmids were transfected and a reduction of 69% when 0.25 μg vIL-6 and MAV-KDEL plasmids were transfected. It should, however, be noted that such a quantification is only an approximation and not an absolutely accurate method to measure the amount of phospho-STAT3 protein.

We conclude from these experiments that the neutralization of vIL-6 is feasible from within the cell and that such a strategy can lead to the neutralization of vIL-6 activity.

Recently, it was reported that vIL-6 has the ability to transduce signals from within the cell. It was speculated that only a small proportion of synthesized vIL-6 is secreted and that intracellular signal activation via a vIL-6-gp130 interaction in the ER is predominant (43). To evaluate this model and to demonstrate the relevance of our scFv-vIL-6-ER retention strategy, we generated a vIL-6 cDNA C-terminally fused to the ER retention signal KDEL. Interestingly, vIL-6 protein could be detected in the lysate of HepG2 cells transiently transfected with a cDNA coding for vIL-6 or vIL-6-KDEL (Fig. 7A). Due to the ER retention signal, vIL-6-KDEL could be detected at higher amounts in cell lysates than vIL-6, which was also secreted into the supernatant (Fig. 7C and data not shown). Interestingly, cells transfected with a cDNA coding for vIL-6 or vIL-6-KDEL were found to have identical STAT3 phosphorylation intensities (Fig. 7B), indicating that both intracellular and secreted vIL-6 can induce gp130 signal transduction. We stimulated untransfected HepG2 cells with 4-h conditioned medium of vIL-6- or vIL-6-KDEL-transfected HepG2 cells and analyzed STAT3 phosphorylation to see whether any biologically active vIL-6-KDEL was secreted. As expected, untransfected HepG2 cells treated with conditioned medium of vIL-6-transfected HepG2 cells showed phosphorylated STAT3, whereas untransfected HepG2 cells treated with conditioned medium of vIL-6-KDEL-transfected HepG2 cells showed only unphosphorylated STAT3 (Fig. 7C), indicating that the ER retention of vIL-6-KDEL was complete. The concentration of vIL-6 in the conditioned medium was under the detection limit by Western blotting (data not shown). These data support the model that an active vIL-6-gp130 signal transduction complex can be formed within the cell.