Interleukin-6 is a direct mediator of T cell migration

2.1 IL-6-induced migration of human primary T cells and T cell-derived cell lines

To clarify whether IL-6 functions as a migration factor, we isolated activated human T cells (CD4⁺/CD45RO⁺). The isolated T cell population was monitored by FACS analysis and found to be 98.5% pure (CD4⁺/CD45RO⁺) (Fig. 1B).

IL-6-dependent migration of T cells was measured in a modified Boyden chamber assay. Fig. 1A shows significantly enhanced migration of primary human T cells upon stimulation with IL-6 compared to non-stimulated cells. Human T cell lymphoma-derived Karpas-299 cells were also tested for IL-6-mediated migration. Similar to primary T cells, Karpas-299 cells migrated in response to IL-6 (Fig. 1C). For our ongoing experiments, we used human primary CD4⁺/CD45RO⁺ T cells and Karpas-299 cells, which also express CD4 and CD45RO.

To further confirm that stimulation-induced migration is a response to IL-6 signaling and not due to a contamination within the cytokine preparation, antagonistic anti-gp130 antibodies were used to block migration (Fig. 2A). Control antibodies raised against c-Src did not inhibit IL-6-dependent migration (data not shown). We asked whether IL-6 induces chemokinesis or chemotaxis. Fig. 2B shows that efficient migration was only observed when the cytokine was added to the lower compartment of the Boyden chamber, but not when applied to the upper chamber (Fig. 2B, far right bar). Thus, Karpas-299 cells migrate towards IL-6, suggesting that an IL-6 gradient between the two reservoirs is required for IL-6-induced migration. Following this idea, we tested whether migration decreases with the sequential loss of the IL-6 gradient between the upper and lower reservoir of the Boyden chamber. The right part of Fig. 2B confirms that increasing amounts of IL-6 in the upper chamber inhibit migration towards the lower IL-6-containing compartment. We also asked whether it is sufficient to prime the cells with IL-6 to initiate migration. Therefore, the cells were pretreated with IL-6 for 30 min before being transferred to the Boyden chamber. Fig. 2C shows that pretreatment with the cytokine alone neither induces migration nor enhances IL-6-induced migration. These data show that Karpas-299 cell cannot be primed and further support the observation made in Fig. 2B that the cells require an IL-6 gradient to migrate. All these results indicate chemotactic migration in response to IL-6.

2.2 IL-6-induced migration depends on the presence of fibronectin

IL-6-initiated migration of Karpas-299 cells depends on the extracellular matrix, since coating the membrane with fibronectin was crucial for IL-6-dependent migration. In contrast, stromal cell-derived factor-1 (SDF-1) acts as a potent chemokine even in the absence of fibronectin (Fig. 3A). This requirement does not seem to be specific for fibronectin, since collagen IV and laminin can substitute for fibronectin. Poly-L-lysine, which does not activate integrins, was not potent to support IL-6-dependent migration (Fig. 3B), suggesting that cell attachment (without integrin activation) is not sufficient for migration. Importantly, we did not observe altered cell attachment in response to IL-6 (data not shown), indicating that IL-6 does not mediate detachment of cells instead of inducing chemotactic migration. To corroborate these observations, we analyzed the surface expression of integrins known to be expressed on T cells and to bind fibronectin. FACS analysis was performed using antibodies specific for the integrin subunits α₄, α₅, and β₁ (Fig. 3C, left panels). Additionally, surface expression prior and after stimulation with IL-6 was compared (Fig. 3C, right panels). FACS analysis demonstrates that the integrins very-late antigen (VLA)-4 and VLA-5 are expressed on Karpas-299 cells and that their expression is not influenced by IL-6 stimulation. To further examine the contribution of VLA-4 and VLA-5 on IL-6-mediated migration, we analyzed migration in the presence of antibodies raised against the integrin subunits α₄, α₅, and β₁. As Fig. 3D shows, IL-6-stimulated Karpas-299 cell migration can specifically be blocked by antibodies against α₄ and β₁ integrin, but not α₅ integrin. These data indicate a requirement of VLA-4 for IL-6-induced migration.

Consistently, we proved whether IL-6 affects integrin activation. 9EG7 antibodies recognize the activated β₁ integrin subunit and therefore represent a good tool to monitor VLA-4 activation in response to IL-6 (Fig. 3E). Mn²⁺, known to activate VLA-4, was used as a positive control 18 (upper right panel). In the presence of Mn²⁺, antibody binding to the β₁ subunit was significantly increased [compare closed histogram (without Mn²⁺) and open histogram (with Mn²⁺)]. Similarly, IL-6 initiated enhanced 9EG7 antibody binding (open histogram in lower left panel), which could be reverted by addition of IL-6-blocking antibodies (open histogram in the lower right panel). These data suggest that IL-6 activates β₁ subunit-containing integrins such as VLA-4, shown above to be crucial for IL-6-induced cell migration.

2.3 Migration in response to IL-6-type cytokines

IL-6-type cytokines signal through gp130 homodimers (IL-6, IL-11), gp130/leukemia inhibitory factor (LIF) receptor heterodimers [LIF, ciliary neurotrophic factor (CNTF), oncostatin M (OSM), cardiotrophin 1 (CT-1), CT-1-like factor (CLC)], the gp130/OSM receptor heterodimer, or a gp130/WSX-1 heterodimer (IL-27). Signaling of IL-6, IL-11, CT-1, CNTF, and CLC depends on the presence of additional α receptors. Stimulation with IL-6 in the presence of the soluble IL-6Rα (sIL-6Rα) allows signaling independently of endogenous IL-6Rα expression. Although addition of sIL-6Rα slightly enhanced the chemotactic activity of IL-6, it is not required (Fig. 4A), since IL-6Rα is expressed on the plasma membrane of Karpas-299 cells (Fig. 4C).

To elaborate whether migration can only be induced by activating gp130 homodimers or whether it can also be induced by gp130/LIF receptor- or gp130/OSM receptor-utilizing cytokines, Karpas-299 cells were stimulated with IL-6, IL-11, LIF, or OSM. Fig. 4B illustrates that only IL-6 and IL-11 are potent chemotactic factors. Stimulation with LIF did not lead to migration, due to the lack of cell surface expression of the LIF receptor, as monitored by FACS analysis (Fig. 4C). OSM also did not induce migration, although the OSM receptor is expressed on the plasma membrane (Fig. 4C). These observations suggest that cytokines utilizing gp130 homodimers are more potent to stimulate migration of Karpas-299 cells than those binding to gp130-containing heteromeric receptor complexes.

2.4 IL-6-induced migration is a direct response

In order to exclude that IL-6 stimulates migration indirectly by inducing the release of autocrinally acting chemokines, we analyzed whether pertussis toxin (PTX), which inhibits G protein-coupled receptor signaling, affects migration in response to IL-6. As shown in Fig. 5A, IL-6-initiated migration of primary human CD4⁺/CD45RO⁺ T cells is insensitive to PTX. SDF-1 is a classical chemokine and signals through the G protein-coupled receptor CXCR4. Fig. 5B shows SDF-1 to be much more potent to induce migration of Karpas-299 cells than IL-6 (compare also with Fig. 2B). SDF-1-dependent migration was significantly inhibited by PTX, but the toxin had no effect on IL-6-induced migration (Fig. 5C). These data suggest that IL-6 does not act via chemokines that signal through PTX-sensitive G protein-coupled receptors.

To further evaluate whether migration is directly mediated by IL-6, we analyzed IL-6-conditioned medium for the presence of secondary chemotactic activities (Fig. 5D). Karpas-299 cells were stimulated with conditioned medium that had been depleted of IL-6 with neutralizing antibodies (Fig. 5D, far right bar). The capability to induce migration was significantly reduced as compared to cells stimulated with IL-6-containing conditioned medium (second bar from the right), indicating that induction of migration is triggered by IL-6 alone.

2.5 Contribution of individual signaling pathways for migration

ERK activation is well known to be important for migration in response to many different stimuli. Thus, we tested whether IL-6-initiated migration of human primary CD4⁺/CD45RO⁺ T cells is sensitive to U0126, the low-molecular weight inhibitor of the ERK pathway. Fig. 6A shows a total block of T cell migration in the presence of 10 µM U0126, suggesting a crucial role of ERK also for IL-6-induced migration of T cells.

To further dissect which of the signaling pathways initiated by IL-6 or integrin activation contribute to IL-6-mediated cell migration, we analyzed to which extent well-characterized low-molecular weight inhibitors for JAK (AG490 and JAK Inhibitor 1), PI3K (LY294002 and Wortmannin), or ERK (U0126 and PD98059) affect IL-6-dependent migration of Karpas-299 cells. Fig. 6B demonstrates a crucial role for JAK, which is in line with the upstream function of JAK in IL-6 signaling. Interestingly for our studies on IL-6-induced migration, the PI3K and ERK pathways were found to be crucial for migration. The inhibitors used did not affect the mortality of the cells, as monitored by trypan-blue staining (data not shown).

2.6 Migration through chimeric erythropoietin receptor/gp130 receptors

The above-mentioned experiments carried out with low-molecular weight inhibitors do not allow an analysis of the specific contribution of gp130 in the migration process, but they give a first hint at the general contribution of specific signaling cascades, irrespective of whether they are initiated by integrins or gp130. To focus on the contribution of IL-6-initiated signaling cascades, experiments were designed that allow the specific activation of either the JAK/STAT pathway or the MAPK cascade in response to IL-6. gp130 mutants have been shown to specifically activate either pathway: gp130 mutants lacking Tyr759 are unable to activate the SHP2/MAPK cascade after IL-6 stimulation 19. On the other hand, gp130 mutants lacking the four membrane-distal tyrosine residues (Tyr767, Tyr814, Tyr905 and Tyr915) do not activate STAT1 and STAT3 transcription factors 20, 21. We used these mutants to create chimeric receptors harboring the extracellular part of the erythropoietin receptor (EpoR) and the transmembrane and cytoplasmic parts of gp130. Chimeric receptors are useful and established tools to analyze signaling through exogenous receptor mutants, independently from endogenous receptors.

We infected the cells with a retroviral expression system that does not interfere with the migratory activity triggered by endogenous receptors. The control experiments demonstrate that naive Karpas-299 cells, as well as those infected with recombinant green fluorescent protein (GFP)-encoding retrovirus, migrate in response to IL-6, but not in response to Epo. Furthermore, Epo stimulation does not mediate migration of Karpas-299 cells, a prerequisite for the use of retrovirally expressed EpoR/gp130 chimeric receptors for the following studies (data not shown). Karpas-299 cell clones were generated, stably expressing chimeric EpoR/gp130 constructs containing (a) all six cytoplasmic tyrosine residues [EpoR/gp130(YYYYYY)], (b) a substitution of all six tyrosine residues by phenylalanine [EpoR/gp130(FFFFFF)], as a negative control, (c) a single phenylalanine for tyrosine substitution at position 759 [EpoR/gp130(YFYYYY)], or (d) an exchange of the four membrane-distal tyrosine residues by phenylalanine [EpoR/gp130(YYFFFF)] (Fig. 7A).

Unfortunately, no antibody against the extracellular domains of the murine EpoR part of the chimeric receptors appropriate for FACS analysis is available. Instead, expression levels of GFP and the chimeric receptors were monitored by FACS and Western blot analysis, respectively (Fig. 7B). Expression of EpoR/gp130(YYFFFF) was slightly reduced in comparison to the other chimeric receptors.

Migration in response to IL-6 and Epo was studied, comparing different clones of these four cell lines (Fig. 7C). The left panel shows that cells expressing the wild-type chimeric receptor, EpoR/gp130(YYYYYY), migrate in response to IL-6 (signaling through the endogenous receptor complex) as well as in response to Epo (signaling through the exogenous chimeric receptor). IL-6- and Epo-induced migration was similarly sensitive to the inhibitors tested in Fig. 6 (data not shown). Expression of a chimeric receptor lacking all cytoplasmic tyrosine motifs [EpoR/gp130(FFFFFF)] led to a specific loss of migration in response to Epo, whereas migration in response to IL-6 was unaffected. These observations indicate that the cytoplasmic tyrosine motifs of gp130 are crucial for the gp130-dependent migration of T cells (Fig. 7C, panel 2).

A single mutation of Tyr759 [EpoR/gp130(YFYYYY)] also specifically impairs Epo-dependent migration, indicating that the tyrosine motif responsible for the activation of SHP2 and the MAPK cascade in response to IL-6 stimulation 19 is essential for gp130-mediated cell migration (panel 3). Finally, cells expressing the receptor mutant EpoR/gp130(YYFFFF) unable to initiate STAT1 and STAT3 activation 20, 21 still migrate in response to Epo. The reduced migration observed results very likely from reduced receptor expression in these cell clones. Our data suggest no essential function of STAT activation for the migration of human T cells.

4.1 Materials and cell lines

Recombinant human (h)IL-6 and sIL-6R were prepared as described 30. hIL-11, hOSM and hSDF-1 were purchased from TEBU (Frankfurt/Main, Germany); hLIF was from Sigma (Taufkirchen, Germany). Recombinant Epo was a generous gift of Drs. B. Hilger and K. H. Sellinger (Roche, Mannheim, Germany).

For FACS analysis, PE-conjugated F(ab′)₂ fragments recognizing mouse IgG or rat IgG, respectively, were used (Dianova, Hamburg, Germany). Anti-LIF receptor antibodies (10.B2) were kindly provided by Y. Jacques (Nantes, France); anti-OSM receptor (AN-E2) antibodies were a gift of H. Gascan (Angers, France). Anti-IL-6Rα (B-N12), anti-gp130 (B-P8, B-R3), and isotype control (B-Z2) antibodies were donated by J. Wijdenes (Diaclone, Besançon, France). Antagonistic anti-IL-6 antibodies (mAb206) were from R&D Systems (Minneapolis, MN). Anti-IL-11R antibodies were kindly donated by F. A. Montero-Julian (Marseille, France); antibodies raised against the cytoplasmic part of gp130 were from Upstate Laboratories (Lake Placid, NY). Anti-integrin α₄ (HP1/2) and α5 (A5-PUJ1) antibodies were kindly provided by M. E. Hemler (Boston, MA). Anti-fibronectin, -integrin β₁, -CD4-, and -CD45RO antibodies were purchased from BD Biosciences PharMingen (Heidelberg, Germany). Blocking antibodies to integrin α₄, α₅ and β₁ were from Chemicon (Hofheim, Germany). The anti-IRF-1 antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA).

Collagen IV, poly-L-lysine, and pertussis toxin were from Sigma (Taufkirchen, Germany). U0126 was purchased from Promega (Mannheim, Germany). All other inhibitors were from Calbiochem (Bad Soden, Germany). Laminin was from Harbor Bio-Products (Norwood, MA). The Karpas-299 cells were from DSMZ in Braunschweig, Germany.

4.2 Isolation of primary human T cells

Human PBMC were isolated from fresh blood (200 ml). Activated (CD4⁺/CD45RO⁺) T cells were isolated with a MACS cell negative isolation kit (indirect magnetic labeling system for CD4⁺ and CD45RO⁺ cells) (Miltenyi Biotec, Bergisch Gladbach, Germany), as described by the manufacturer. Purity of the cells was monitored by FACS analysis using FITC-coupled anti-CD4 antibodies and PE-coupled anti-CD45RO antibodies.

4.3 Cell culture

Karpas-299 cells were grown in RPMI supplemented with 10% FCS, streptomycin (100 mg/l), and penicillin (60 mg/l). For migration assays, FCS and antibiotics were omitted from the medium. HEK293T cells were grown in DMEM supplemented as described above.

4.4 Construction of expression vectors

The retroviral vector pM5-EpoR/gp130-NeoGFP is based on the myeloproliferative Sarcoma virus. cDNA were cloned in a bicistronic configuration into the multiple cloning site: the cDNA encoding an EpoR/gp130 chimeric receptor is followed by the IRES of the human NRF gene and the cDNA for an enhanced-GFP/neomycin resistance fusion protein. Four different retroviral vectors were generated to express EpoR/gp130(YYYYYY), EpoR/gp130(YFYYYY), EpoR/gp130(YYFFFF), and EpoR/gp130(FFFFFF) 21. These chimeric receptors contain tyrosine to phenylalanine substitutions within the cytoplasmic part of the receptor, as indicated. Additionally, a retroviral vector for the expression of GFP without any chimeric receptor has been used as a control.

4.5 Modified Boyden chamber assay

Cells (7.5×10⁵/ml) were incubated overnight in RPMI containing 2% FCS. Transwell inserts (6.5-mm diameter, 8-µm pore size; Corning Costar, Bodenheim, Germany) were coated with human fibronectin. Into a transwell, 1×10⁵ cells/100 µl were transferred. The transwell was inserted into a 24-well plate containing 800 µl medium. Cell migration to the bottom chamber was assessed 4 h later. For each experiment, the cells from four different fields were enumerated. Every migration assay was performed three times in triplicate. Approximately up to 2% of the cells migrate spontaneously. Migration was induced by addition of 20 ng/ml IL-6 (and 1 µg/ml sIL-6Rα where indicated) into the lower compartment. Up to 8% of the total number of Karpas-299 cells were found to migrate into the lower compartment of the Boyden chamber within the first 4 h post stimulation. The migration was expressed as means ± SEM. The amount of migrating cells is given in relation to that under standard conditions (100%) to indicate how specific treatment of the cells affects their chemotactic activities. For statistical analysis of significance, we applied one-way analysis of variance (one-way ANOVA).

4.6 FACS analysis

Cells (5×10⁵) were harvested and washed with FACS buffer (PBS containing 5% FCS, 0.1% NaN₃). The cells were resuspended in 100 µl ice-cold FACS buffer and incubated with 1 µg of antibodies raised against the extracellular part of the respective receptors. After washing, mouse IgG-specific PE-conjugated goat antibodies were added to monitor receptor expression, using a FACSCalibur (Becton Dickinson, Heidelberg, Germany). GFP expression was analyzed in the absence of antibodies. For monitoring activated β₁ integrin, Karpas-299 cells were grown overnight in RPMI supplemented with 2% FCS and starved in 25 mM Hepes buffer supplemented with 150 mM NaCl, 500 µg/ml soluble fibronectin, and 1 mM MgCl₂ (or 5 mM MnCl₂ for positive control of activated integrins 18) for 30 min. Stimulation was performed with 15 ng/ml IL-6 for 30 min. The cells were incubated at 4°C for 30 min with 9EG7 antibodies, and after washing, rat IgG-specific PE-conjugated goat antibodies were added. After washing with FACS buffer, cells were monitored for activated β₁ integrin subunits. Negative control was performed by adding blocking anti-IL-6 antibodies (2 µg/ml) 30 min prior to stimulation with IL-6.

4.7 Immunoprecipitation and immunoblot analysis

For immunoprecipitation, 6×10⁷ cells were lysed in 500 µl RIPA lysis buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 0.5% Igepal, 1 mM EDTA, 1 mM Na₃VO₄, 1 mM NaF, 15% glycerol). Buffers were supplemented with 10 µg/ml of each aprotinin, pepstatin, and leupeptin. After precipitation with antibodies against gp130, immune complexes were separated by SDS-PAGE and transferred to a poly(vinylidene difluoride) (PVDF) membrane. Antigens were detected by incubation with primary antibodies specific for the cytoplasmic part of the EpoR/gp130 chimeric receptor and horseradish peroxidase (HRP)-coupled secondary antibodies (1:5,000) (Dako, Hamburg, Germany). The membranes were developed with an enhanced chemiluminescence kit (Amersham-Pharmacia, Uppsala, Sweden).

4.8 Retroviral gene transfer

Production of retroviruses was performed with pV-Pack-Vectors from Stratagene (La Jolla, CA), as described in the manufacturer's protocol. HEK293T cells were transfected using FuGene (Roche Diagnostics GmbH, Mannheim, Germany) with pVPackGP, pVPack-10A1 and pM5-EpoR/gp130-NeoGFP. The retrovirus-containing supernatant was added to Karpas-299 cells two days after transfection, together with 10 µg/ml polybrene (Sigma, Taufkirchen, Germany). After infection (1 day), the cells were selected in RPMI containing 10% FCS and 2 mg/ml neomycin. Single clones were isolated, the expression of GFP was examined using FACS analysis, and the expression of the EpoR/gp130 chimeric receptors was monitored by immunoprecipitation and subsequent Western blot analyses. A representative of at least two clones for each receptor mutant is shown.