IL-1 receptor accessory protein is essential for IL-33-induced activation of T lymphocytes and mast cells
- Shafaqat Ali*,
- Michael Huber†,
- Christian Kollewe*,
- Stephan C. Bischoff‡,
- Werner Falk§, and
- Michael U. Martin*,¶
- *Immunology FB08, Justus-Liebig-University Giessen, Winchesterstrasse 2, D-35394 Giessen, Germany;
- †Molecular Immunology, Institute of Biology III, University of Freiburg and Max-Planck-Institute for Immunobiology, Stübeweg 51, D-79108 Freiburg, Germany;
- ‡Nutritional Medicine/Prevention, University of Hohenheim, Fruwirthstrasse 12, 70593 Stuttgart, Germany; and
- §Department of Internal Medicine I, University of Regensburg, D-93042 Regensburg, Germany
-
Edited by Charles A. Dinarello, University of Colorado Health Sciences Center, Denver, CO, and approved October 3, 2007 (received for review June 25, 2007)
Abstract
Lack of the IL-1 receptor accessory protein (IL-1RAcP) abrogates responses to IL-33 and IL-1 in the mouse thymoma clone EL-4 D6/76 cells. Reconstitution with full-length IL-1RAcP is sufficient to restore responsiveness to IL-33 and IL-1. IL-33 activates IL-1 receptor-associated kinase-1, cJun-N-terminal kinase, and the NF-κB pathway in an IL-1RAcP-dependent manner and results in IL-2 release. IL-33 is able to induce the release of proinflammatory cytokines in bone marrow-derived (BMD) mast cells, indicating that IL-33 may have a proinflammatory potential like its relatives IL-1 and IL-18, in addition to its Th2-skewing properties in the adaptive response described previously. Blocking of murine IL-1RAcP with the neutralizing antibody 4C5 inhibits response of mouse thymoma cells and BMD mast cells to IL-33. The interaction of either membrane-bound or soluble forms of IL-1RAcP and IL-33Rα-chain depends on the presence of IL-33, as demonstrated by coimmunoprecipitation assays. These data demonstrate that IL-1RAcP is indispensable for IL-33 signaling. Furthermore, they suggest that IL-1RAcP is used by more than one α-chain of the IL-1 receptor family and thus may resemble a common β-chain of that family.
IL-33 is a recently identified member of the family of IL-1-like cytokines that binds to the IL-33 receptor, formerly known as the orphan receptor ST2 of the IL-1 receptor family (1). Like IL-18 and IL-1β, the closely related IL-33 is synthesized as a 31-kDa precursor that can be processed to the mature cytokine by caspase 1. Similarly to the precursor of IL-1α, IL-33 can translocate to the nucleus (2). Although these members of the IL-1 family are potent inducers of cytokine production, the profile induced by the respective molecules differs. IL-1 is an important proinflammatory cytokine of the innate arm of immunity and also promotes proliferation of thymocytes (Th0), thus playing a role in adaptive immunity as well. IL-18 induces proinflammatory cytokines; however, its major feature, in collaboration with IL-12, is the induction of IFN-γ. Thus, IL-18 skews Th0 cells toward Th1 differentiation. Little is yet known about proinflammatory properties of IL-33. However, treatment of mice with IL-33 results in an increase in IL-4, IL-5, and IL-13 production, which suppresses Th1 differentiation and favors Th2 differentiation (1).
The structural and functional similarities of IL-1, IL-18, and IL-33 can be extended to their receptors. IL-1 signal transduction is initiated by binding of either form of IL-1 to IL-1 receptor type I (IL-1Rα-chain), which undergoes a conformational change allowing the IL-1 receptor accessory protein (IL-1RAcP) to recognize the ligated IL-1RI. IL-1RAcP does not recognize IL-1. Both transmembrane proteins form a heterodimer, which results in the close association of the cytosolic Toll-like IL-1R (TIR) homology domains (3). IL-1RAcP is essential for IL-1 signaling (4–6), and the TIR domains of both chains are required to facilitate signaling (7, 8). Ligand-mediated heterodimerization with subsequent intracellular association of strongly homologous, but discrete, TIR domains is a typical feature of the IL-1 receptor subfamily of TIR-domain-containing receptors. In accordance, the IL-18 receptor complex consists of IL-18Rα- and β-chains, both of which carry a TIR domain (9, 10). Similar mechanisms have been described for IL-1Rrp2, the receptor for IL-1F6, IL-1F8, and IL-1F9, which also associates with IL-1RAcP (11, 12). Although ST2 has been known as an orphan receptor in the IL-1 receptor family for many years (13, 14), its ligand remained elusive until IL-33 was identified (1). It has been proposed that the IL-33Rα-chain also requires a β-chain containing a TIR domain and that this may be a member of the IL-1 receptor family (15).
Here we demonstrate by different means that IL-1RAcP represents the IL-33Rβ-chain necessary for IL-33 to induce responses in T cells and mast cells. This finding shows that IL-1RAcP may resemble a common β-chain used by more than one α-chain of the IL-1 receptor family.
Results
Murine EL-4 D6/76 (D6/76) Cells Do Not Respond to IL-33 Because of the Lack of IL-1RAcP.
EL-4 6.1 (EL-4) cells respond to IL-1 (16), IL-18, and IL-33 by release of IL-2 (Fig. 1 A). This finding demonstrates that EL-4 cells express functional receptors for all three members of the IL-1 cytokine family. Although picomolar concentrations of rhIL-1β were sufficient to induce IL-2 synthesis, lower nanomolar concentrations were required of either rmIL-18 or rmIL-33. D6/76 cells, derived from EL-4 (17), are unable to respond to IL-1. Reconstitution of D6/76 cells with full-length (fl) mIL-1RAcP restored IL-1 responsiveness (4, 5, 18). D6/76 cells do not respond to IL-33 (Fig. 1 B Left), but do respond to IL-18 (Fig. 2 B Left). Reconstitution with fl mIL-1RAcP restored responsiveness not only to IL-1, but also to IL-33 (Fig. 1 B Left), demonstrating that lack of IL-1RAcP causes D6/76 cells' inability to respond to both cytokines. Reconstitution was not possible when using a C-terminally truncated version of mIL-1RAcP (ΔC-AcP), which lacks the TIR domain. The response to either IL-33 or IL-1β could be increased by overexpressing mIL-1RAcP (Fig. 1 B Right). This result demonstrates that fl IL-1RAcP is equally indispensable for IL-33- and IL-1-induced signaling by their respective receptor complexes.
Cytokine production is induced by IL-33 and IL-1 only in the presence of IL-1RAcP. (A) EL-4 cells respond to IL-1 and IL-18 and by producing IL-2 in a concentration-dependent manner. Cells were stimulated with rhIL-1β (light gray triangles), rmIL-18 (dark gray squares), or rmIL-33 (black circles) as indicated for 18 h, and mIL-2 was measured by ELISA. (B) D6/76 cells respond to IL-33 or IL-1 only if reconstituted with fl IL-1RAcP. Cells were transiently cotransfected with a plasmid encoding mIL-33Rα-chain and either empty vector (e.v.), plasmids encoding C-terminally truncated mIL-1RAcP(ΔC-AcP), fl mIL-1RAcP (Left) (1 μg of plasmid for every 5 × 106 cells), or increasing amounts of plasmid encoding fl mIL-1RAcP (Right). Cells were stimulated with 100 ng/ml rmIL-33 (black bars) or 100 pg/ml rhIL-1β (gray bars), and mIL-2 was measured by ELISA 18 h later. Data depicted are means ± SD from one experiment (triplicates or quadruplicates) of a series of at least three similar ones with comparable results.
IL-33 activation of signal transduction elements of the IL-1 receptor family depends on IL-1RAcP. (A) C-terminally truncated IL-1RAcP reduces IL-33 activation of IRAK-1. 6-IRAK-19 cells (a clone of EL-4 cells stably transfected with hIRAK-1) were transiently transfected with 1 μg of plasmid encoding mIL-33Rα-chain for every 107 cells. In addition, some of the cells were cotransfected with 1 μg of plasmid encoding C-terminally truncated mIL-1RAcP lacking the TIR domain (ΔC-AcP). One day later, transfectants were either stimulated with 10 ng/ml rhIL-1β or 100 ng/ml rmIL-33 for 15 min. IRAK-1 was immunoprecipitated from cell lysates by using mAb 2A9 (a kind gift from Z. Cao). Autophosphorylation of IRAK-1 was measured in an in vitro kinase assay and visualized by autoradiography. Transfection efficiency was on the order of 40% as ascertained by parallel transfections with EGFP plasmid and cytofluorometry. The data shown are of one representative experiment out of a series of three with comparable results. (B) Activation of the transcription factor NF-κB by IL-33 depends on IL-1RAcP. D6/76 (Left) and EL-4 (Right) cells were transiently cotransfected with plasmids encoding mIL-33Rα-chain and an NF-κB-dependent luciferase reporter plasmid (D6/76 cells, 0.5 μg of plasmid each for 5 × 106 cells; EL-4 cells, 0.5 μg of reporter plasmid plus the indicated concentrations of plasmid for mIL-33Rα or empty vector). In addition, in D6/76 cells, fl mIL-1RAcP was overexpressed. Cells were stimulated with either 100 ng/ml rmIL-33 (black bars) or 100 pg/ml rhIL-1β (gray bars) and either 100 ng/ml rmIL-18 or 100 ng/ml rhTNFα, and luciferase activity was determined in the cell lysates the next day. Preincubation with 50 μg/ml mAb 4C5 (white bars) was for 1 h before stimulation. Depicted is fold induction calculated by dividing relative light unit (RLU) values of stimulated samples by the RLU values of controls. Data shown are duplicates of one representative experiment out of a series of three with comparable results. (C) IL-1RAcP lacking the TIR domain (ΔC-AcP) is dominant-negative in IL-33 signaling as measured in an NF-κB reporter gene assay. EL-4 cells were transiently transfected with increasing amounts of a plasmid encoding ΔC-AcP in addition to an NF-κB-dependent luciferase reporter plasmid (1 μg for every 1 × 107 cells). The next day, cells were stimulated with 100 ng/ml rmIL-33 for a further 18 h. Luciferase activity was measured in the cell lysates. Data shown are duplicates of one representative experiment out of a series of three with comparable results. (D) The neutralizing anti-mIL-1RAcP mAb 4C5 reduces IL-1β- and IL-33-stimulated NF-κB activity. EL-4 cells were transiently cotransfected with a plasmid encoding mIL-33Rα-chain and an NF-κB-dependent luciferase reporter. Subsequently, cells were preincubated for 30 min with either 4C5 (black bars) or the isotype control antibody RA3–6B2 (white bars) at 0, 0.1, 1, or 10 μg/ml and stimulated with either 100 pg/ml rhIL-1β (gray bars) or 50 ng/ml rmIL-33 (black bars). The next day, luciferase activity was determined in the cell lysates. Data shown are triplicates of one representative experiment out of a series of three with comparable results.
IL-33 did not induce the same level of IL-2 production in EL-4 or transfected D6/76 cells as was reached with IL-1. We transiently transfected EL-4 cells with increasing concentrations of a plasmid encoding fl mIL-33Rα-chain. In addition, D6/76 cells were transfected with a constant concentration of plasmid encoding fl mIL-1RAcP. In both situations, the response to IL-33 was enhanced approaching the IL-1 response with respect to IL-2 production (Fig. 1 B Left) and NF-κB activation (Fig. 2 B Right). This finding suggests that low expression of the endogenous IL-33Rα-chain is limiting IL-33 responses in these T cell lines.
IL-1RAcP Is Required in Mast Cells to Allow IL-33-Stimulated Cytokine Release.
Freshly isolated and differentiated murine bone marrow-derived (BMD) mast cells express ST2 (IL-33Rα-chains) on their surface (data not shown). Stimulation with rmIL-33 enhanced the constitutive production of the proinflammatory cytokine IL-6 to an extent comparable to LPS (Fig. 3 A). IL-33 also stimulated rapid release of TNFα and IL-1β, but did not affect the low constitutive release of IL-13 (Fig. 3 B). Pretreatment with the monoclonal antibody (mAb) 4C5, which neutralizes IL-1RAcP (19, 20), resulted in a concentration-dependent reduction of the IL-33-induced cytokine production. If mast cells were stimulated with low concentrations of rmIL-33 (1–3 ng/ml), a complete inhibition of stimulated IL-6 production could be achieved (Fig. 3 C) (TNFα release was inhibited 27% at 25 μg/ml or 12% at 50 μg/ml mAb 4C5, respectively).
Murine BMD mast cells respond to IL-33 with cytokine production in an IL-1RAcP-dependent manner. (A) Mast cells produce IL-6 after stimulation with rmIL-33. Cells were stimulated with 0, 1, 10, or 100 ng/ml rmIL-33 or 1 μg/ml LPS. The next day, mIL-6 was detected by ELISA. (B) Mast cells produce several proinflammatory cytokines after stimulation with rmIL-33. The 1 × 106 cells per milliliter were seeded and stimulated with 2 ng/ml rmIL-33 for 0, 3, 6, and 24 h as indicated. Cytokine profiles were detected in the supernatants by using Beadlyte MultiCytokine Flex kits in a Luminex 100 instrument. (C) The neutralizing anti-mIL-1RAcP mAb 4C5 reduces IL-33-stimulated IL-6 release in mast cells. Mast cells were preincubated for 30 min with three different concentrations of 4C5 mAb as indicated and then left either untreated (con) or stimulated with 1 or 5 ng/ml rmIL-33. The next day, IL-6 was measured by ELISA. Results are from one representative experiment of a series with comparable results (B) or are means of triplicates or quadruplicates of one typical experiment out of a series of at least three similar experiments with comparable results (A and C). For mast cell stimulation, only commercial rmIL-33 was used (LPS < 0.1 units/μg protein).
Activation of Signaling Events by IL-33 Is Dependent on the Presence of IL-1RAcP.
The IL-1 receptor family uses the MyD88/IRAK/TRAF6 module to activate downstream signaling elements, such as c-Jun N-terminal kinase (JNK) or the transcription factor NF-κB.
IRAK-1 activity was measured in 6-IRAK-19 cells (a stably transfected EL-4 cell clone that expresses hIRAK-1). 6-IRAK-19 cells were transiently transfected with mIL-33Rα-chain to increase IL-33 responses. Stimulation with either rhIL-1β or rmIL-33 resulted in enhanced autophosphorylation of IRAK-1 (Fig. 2 A). Cotransfection with ΔC-AcP reduced IRAK-1 activation. This result demonstrates that IRAK-1 is activated by the IL-33R complex as previously described (1) and that this activation depends on the presence of the TIR domain in fl mIL-1RAcP.
The activation of NF-κB is a central element of the signaling pathways activated by the IL-1 receptor family. Although D6/76 cells responded to IL-18 and TNFα, they did not respond to IL-1 or IL-33 by activation of NF-κB. Upon transfection with fl mIL-1RAcP, NF-κB activation was restored to both cytokines with no effect on TNFα or IL-18 response. Preincubation with mAb 4C5 inhibited IL-1- and IL-33-stimulated NF-κB activation, but did not affect IL-18- or TNFα-stimulated activation of the transcription factor (Fig. 2 B Left). ΔC-AcP was not able to restore IL-1 or IL-33 responsiveness with respect to NF-κB activation in D6/76 cells (data not shown). Transfection of EL-4 cells with increasing amounts of the plasmid-encoding ΔC-AcP resulted in a concentration-dependent reduction of NF-κB activation after IL-33 stimulation, demonstrating that the truncated form of mIL-1RAcP behaved in a dominant-negative fashion (Fig. 2 C). Transfection of EL-4 cells with increasing amounts of the plasmid encoding fl mIL-33Rα-chain resulted in an increase in NF-κB activation after IL-33 stimulation while leaving the IL-1 response unaffected (Fig. 2 B Right). This finding proves the specificity of the observed effect for IL-33 and that the endogenous expression level of mIL-33Rα-chain limited the response to its ligand. Preincubation of EL-4 or IL-1RAcP-transfected D6/76 cells with mAb 4C5 resulted in a reduction of IL-1β- and IL-33-stimulated NF-κB activation (Fig. 2 D). NF-κB activation was nearly completely inhibited by 4C5 after stimulation with rhIL-1β, whereas rmIL-33-stimulated activation was reduced to ≈50%. The isotype-matched control antibody had no effect on IL-1- or IL-33-stimulated NF-κB activation (Fig. 2 B Left and D, open bars).
The activation of JNK by IL-1 depends on the presence of IL-1RAcP in D6/76 cells (5). IL-1 and IL-33 activated JNK in EL-4 cells in a transient fashion. The maximum activity was measured at 12–16 min in an in vitro kinase assay by using GST-c-Jun as substrate. The kinetics of JNK activation was identical between IL-1 and IL-33 (data not shown).
In summary, these results demonstrate that key elements of the classical signaling pathway used by the IL-1 receptor family such as IRAK-1, NF-κB, and JNK were specifically activated by IL-33 only in the presence of fl IL-1RAcP.
The Interaction of IL-33Rα-Chain and IL-RAcP Is Dependent on IL-33.
After binding of IL-1 to either IL-1RI or IL-1RII, IL-1 receptors and IL-1RAcP form heterodimeric receptor complexes on the cell surface (7). To clarify whether such a complex also forms between IL-1RAcP and IL-33Rα-chain, epitope-tagged versions of IL-1RAcP and IL-33Rα were coexpressed in HEK293RI cells. After stimulation with rmIL-33, complexes were coimmunoprecipitated. IL-1RAcP was coprecipitable with IL-33Rα-chain, and vice versa (data not shown), only in the presence of rmIL-33, but not if the cells were incubated with rhIL-1β (Fig. 4 A). In some experiments, we observed a weak interaction in the absence of ligand, which is because of the strong protein expression in HEK293RI cells. This result demonstrates that the interaction of the membrane-inserted forms of IL-1RAcP and IL-33Rα-chain is specific for and dependent on IL-33.
The interaction of IL-1RAcP with the IL-33Rα chain depends on IL-33. (A) Membrane IL-33Rα-chain and IL-1RAcP form a complex in the presence of IL-33. HEK293RI cells were transiently transfected with plasmids encoding FLAG-tagged mIL-1RAcP and/or Myc-tagged mIL-33Rα. The next day, transfected cells were left unstimulated or stimulated with 10 ng/ml rhIL-1β or 100 ng/ml rmIL-33 for 15 min. Cells were lysed in the presence of 10 ng/ml rhIL-1β or 100 ng/ml rmIL-33, respectively. FLAG-tagged mIL-1RAcP was immunoprecipitated by using anti-FLAG M2 agarose beads. Precipitated proteins were visualized after SDS/PAGE by Western blot/ECL reaction with anti-Myc (Upper) and anti-FLAG (Lower) antibodies. The result shown is representative of five independent experiments with similar results. (B) Soluble forms of IL-33Rα-chain and IL-1RAcP form a complex in the presence of IL-33. srmIL-33Rα:Fc fusion protein was incubated in the absence or presence of rmIL-33 with FLAG-epitope-tagged srmIL-1RAcP fixed on FLAG-affinity gel. Coprecipitated fusion protein was visualized by Western blot after SDS/PAGE by using an anti-human IgG secondary antibody (see Materials and Methods).
It has been shown that soluble IL-1R and IL-1RAcP interact in the presence of IL-1 (21, 22). Employing a soluble IL-33Rα:Fc fusion protein and an epitope-tagged soluble IL-1RAcP, we could demonstrate a specific interaction in solution in the presence of IL-33, which was stable enough to be immunoprecipitated (Fig. 4 B). This result extends our finding that IL-1RAcP is the coreceptor for IL-33Rα-chain from the membrane-inserted forms to the soluble molecules.
Discussion
The IL-33Rα-chain, formerly known as ST2/T1, is a member of the IL-1 receptor family of TIR-domain-containing receptors. IL-33Rα-chain possesses a TIR domain in its cytoplasmic moiety, which is required for signaling (23). The IL-33Rα-chain interacts with IL-33, which results in the recruitment of signaling molecules also found in the IL-1RI complex (1). This result demonstrates the close relationship of IL-33R and IL-1R complexes as suggested by the use of chimeric molecules (24). Thus, it has been proposed that the IL-33Rα-chain also requires a β-chain containing a TIR domain and that this may be a member of the IL-1 receptor family (1, 15).
The signaling IL-1 receptor complex consists of a heterodimer of IL-1RI that binds IL-1 and IL-1RAcP, which recognizes the ligated IL-1RI (4, 5, 18, 20). The signaling IL-18 receptor complex also consists of a heterodimer in which the IL-18Rα-chain binds IL-18 (25) and the IL-18Rβ-chain (26) functions as a coreceptor. Two additional heterodimeric receptor complexes in the IL-1R system have been described. IL-1F9 was reported to activate NF-κB in an IL-1Rrp2-dependent manner (11). This finding was extended by showing that three members of the IL-1 family of cytokines (IL-1F6, IL-1F8, and IL-F9) bind to IL-1Rrp2 and induce signaling events practically identical to IL-1 (12). IL-1F6, IL-1F8, and IL-1F9 were not able to induce signaling in mouse embryonic fibroblasts derived from IL-1RAcP knockout animals. In addition, a neutralizing antibody against IL-1RAcP inhibited the response of cells to these ligands, strongly suggesting that IL-1RAcP could function as a coreceptor for ligated IL-1Rrp2 (12). Independently, chimeric receptor molecules of the then-orphan receptor ST2 were generated to investigate the signaling pathways activated by its cytoplasmic TIR domain. In D6/76 cells, IL-1RAcP served as a coreceptor for the ST2-chimera; however, no activation of NF-κB was detectable (27).
Here we show that IL-33Rα (ST2) requires IL-1RAcP to function as a signaling receptor for IL-33 in two different types of cells. We also show that central signaling pathways activated by IL-33 are comparable to IL-1. We detected NF-κB activation by IL-33 and IL-1, which is in contrast to the data obtained in the chimeric model (27), but in accordance with transiently transfected HEK293 cells (1). This discrepancy may be because chimeric molecules were used, which may behave differently to the wild-type IL-33Rα-chain ligated with IL-33 (27). We also detected transient JNK activation by IL-33 in a time frame comparable to the transient activation of this protein kinase by IL-1. In transiently transfected HEK293 cells, a slight shift in JNK activation was observed compared with IL-1 (1). In EL-4 cells, we could not see this effect, suggesting subtle differences in different cell types. Interestingly, although rhIL-1β was active in the picomolar range in EL-4 or transfected D6/76 cells, we required nanomolar concentrations of rmIL-33 (or rmIL-18) irrespective of the source of the rmIL-33 to achieve an effect. This result demonstrates the great potency of IL-1 when compared with its sister cytokines, IL-18 and IL-33.
Besides Th2 lymphocytes, mast cells are known to express high levels of IL-33Rα (ST2) (28). Therefore, we included murine BMD mast cells, which are nontransformed and nontransfected primary cells, in our studies. These cells responded readily to IL-33 by producing IL-6 like they did when TLR4 was activated by LPS. In addition, upon IL-33 stimulation, they produced TNFα and IL-1β, two proinflammatory cytokines, whereas the constitutive production of IL-13 was not altered. We used mAb 4C5, which recognizes mIL-1RAcP (19, 20) and neutralizes IL-1β-mediated signal transduction, to inhibit IL-33-mediated signaling in primary mast cells. 4C5 behaves similarly to the mAb M49, which neutralizes hIL-1RAcP and was used to inhibit IL-1RAcP-dependent signaling of IL-1F6 by IL-1Rrp2 in human cells (12). We argued that if 4C5 recognizes mIL-1RAcP and functions by inhibiting the association of the coreceptor with the IL-1-ligated IL-1RI, it also should work in the IL-33 receptor system. We observed a concentration-dependent inhibition of IL-33 effects in the T cell lines EL-4 or D6/76 (after reconstitution with mIL-1RAcP) and, most important, also in mast cells. At present, we do not know why 4C5 was less effective in inhibiting IL-33-stimulated responses when compared with IL-1β. However, it has been described that this anti-IL-1RAcP antibody differs with respect to its neutralizing capacity between receptor complexes containing either IL-1α or IL-1β (19). Thus, it is conceivable that this antibody also distinguishes between IL-1β in IL-1RI and IL-33 in IL-33Rα.
Summarizing the results achieved with anti-IL-1RAcP mAb in different systems (IL-1F6,-8,-9/IL-1Rrp2; IL-1/IL-1R; and IL-33/IL-33R) (12, 19, 20), we must conclude that recognition of the receptor α-chain by IL-1RAcP is largely independent of the cytokine bound to the respective α-chain. A possible explanation may be that binding of ligand creates a conformational neoepitope on its respective receptor α-chain, which is structurally related such that it is recognized by IL-1RAcP. Presumably, the site in IL-1RAcP required for this recognition is affected by neutralizing anti-IL-1RAcP mAbs. No crystal structures of the ligated receptor-like members of the IL-1 receptor family with IL-1RAcP are available. However, a model was calculated in which two possible interactions of IL-1RAcP with IL-1-ligated IL-1RI were accommodated (29). It was proposed that IL-1RAcP could bind to IL-1-ligated IL-1RI by embracing IL-1 in the IL-1RI in a so-called FRONT model. In this model, residues of both ligand and receptor contribute to the association with IL-1RAcP. In light of the new results, which show a promiscuity of IL-1RAcP with respect to its ligated receptor α-chains, we favor the BACK model, in which IL-1RAcP recognizes the backbone of the ligated receptor that changes its conformation after IL-1 binding (30, 31). In this model, the ligand contributes few or maybe no residues to the interaction.
In the EL-4 cell system, we investigated the requirement of the TIR domain for signal transduction initiated by IL-33 by overexpressing a C-terminally truncated version of IL-1RAcP lacking the TIR domain. IL-33 signaling was totally dependent on the presence of the TIR domain of the β-chain. We did not generate a C-terminally truncated version of the IL-33Rα-chain in this study to prove stringently that in fact both TIR domains are essential for IL-33 signaling. However, similar studies (12, 27) show that signaling is only possible when the TIR domain is present in ST2. Therefore, we must conclude that the TIR domains of the α- and β-chains are required for signal initiation.
We show here that the binding of IL-33 to the IL-33Rα chain allows the formation of a complex with IL-1RAcP in the plasma membrane, as described for IL-1 binding to IL-1RI or IL-1RII (7). In addition, we show that a soluble IL-33Rα-chain can interact with soluble IL-1RAcP in the presence of IL-33. The formation of soluble receptor complexes has been previously described in the IL-1 system (21), where soluble IL-1RAcP seems to stabilize the ligated soluble IL-1R, allowing neutralization and sequestration of the cytokine in vivo. It has been proposed that sIL-1RAcP may be helpful to neutralize IL-1β in chronic inflammatory diseases such as rheumatoid arthritis (21, 22). It is tempting to speculate that sIL-1RAcP also could be used to neutralize and sequester IL-33 in the presence of sIL-33Rα-chain in Th-2-biased diseases (e.g., IgE-mediated allergic reactions).
In conclusion, our results demonstrate that the functional IL-33R complex consists of IL-33Rα and IL-1RAcP. In addition, we show that the soluble forms of IL-33Rα and IL-1RAcP can form complexes in the presence of IL-33. Thus, IL-1RAcP can act as a common β-chain for four discrete members of the IL-1 receptor family: IL-1RI, IL-RII, IL-1Rrp2, and IL-33Rα. This situation is reminiscent of the role of common receptor chains in other cytokine families, like gp130 in the IL-6 receptor system or the common γ-chain in the IL-2 receptor family. In addition, the sharing of IL-1RAcP and the fact that it also can interact with receptors in its soluble form open new levels of cross-talking among the different members of the IL-1 receptor family.
Materials and Methods
Expression Plasmids.
pFLAG-mIL33Rα encodes amino acids 18–567 of mIL-33Rα with N-terminal 3×FLAG epitope tag under control of the CMV promoter in p3×FLAG-CMV-9 vector (Sigma–Aldrich). pMyc-mIL33Rα encodes amino acids 18–567 of mouse IL-33Rα (without signal peptide) with N-terminal c-Myc epitope tag under control of the CMV promoter in pcDNA3.1/Zeo(+) vector (Invitrogen), modified by insertion of mIL-2 signal sequence and c-Myc epitope sequence. pFLAG-mIL1RAcP encodes mIL-1RAcP with N-terminal FLAG epitope tag under control of the CMV promoter in pFLAG-CMV-1 vector (Sigma–Aldrich). ΔC-mIL-1RAcP (ΔC-AcP) encodes amino acids 1–384 of mIL-1RAcP (lacking the TIR domain) in pEF-Bos vector (generated by R. Hofmeister, University of Regensburg).
Others.
The 3×NF-κB-Luc reporter plasmid expresses firefly luciferase after binding of activated NF-κB to three consecutive NF-κB-binding sites in the promoter region of the plasmid.
Expression and Preparation of Purified Biotinylated IL-33 Protein.
Escherichia coli BL21 (DE3) cells were transformed with a bicistronic vector based on pETDuet (Novagen), which encodes the HA-tagged E. coli biotin ligase BirA and mature rmIL-33 comprising amino acids 109–266 (1) fused to an N-terminal minimal BirA substrate (GLNDIFEAQKIEWH) (32). Biotin is added by BirA to the ε-amino group of lysine. Biotinylated rmIL-33 was affinity-purified by using UltraLink immobilized monomeric avidin (Pierce). IL-33 was eluted with 2 mM biotin in PBS containing protease inhibitors. The product was analyzed by using a protein 80 chip in a 2100 Bioanalyzer (Agilent) and shown to be essentially free of contaminants. Purified biotinylated rmIL-33 showed a biological activity comparable with commercially obtained material (Alexis).
Cytokines.
rhIL-1β was a kind gift from D. Boraschi (Consiglio Nazionale delle Ricerche Institute for Biomedical Technology, Pisa, Italy). rmIL-33 was obtained from Alexis (ALX 522–101; <0.1 units LPS per μg of protein). For stimulation of mast cells, only commercial IL-33 was used. For most experiments with EL-4, D6/76, or HEK293RI cells, biotinylated rmIL-33 produced in E. coli was used. Neither EL-4, D6/76, nor HEK293RI cells express TLR4 or respond to LPS, which might be present in minute amounts in our preparations despite extensive purification steps. rmIL-18 was obtained from Peprotech. rhTNFα was a kind gift from BASF Aktiengesellschaft (Ludwigshafen, Germany).
Transient Transfection.
D6/76 cells were transiently transfected with plasmids as indicated in figure legends by using the DEAE dextran-chloroquine transfection method as previously described (33). Total amount of plasmid DNA was always adjusted by using the appropriate empty vector. HEK293RI cells were transfected by a slightly modified polyethylenimine (PEI; Aldrich) transfection method (34).
Reporter Gene Assay.
For reporter gene assays, 5 × 106 D6/76 cells were cotransfected with 0.5 μg of pFLAG-mIL1RAcP and 0.5 μg of 3×NFκB-Luc. EL-4 cells were transfected with 3×NFκB-Luc alone or in combination with pMyc-mIL33Rα. Eighteen hours after transfection, 0.5 × 106 cells were kept either unstimulated or stimulated for 16 h with rhIL-1β, rmIL-33, rmIL-18, or rhTNFα. Cells were harvested, washed with PBS, and lysed with 35 μl of passive lysis buffer (Promega). One second after the addition of 100 μl of substrate solution (436 μM d-luciferin, 436 μM NaOH, 20 mM Tricin, 2.67 mM MgSO4, 1.07 mM Mg carbonate hydroxide, 33 mM DTT, 530 μM ATP, and 290 μM CoA) to 25 μl of cell lysate, luciferase activity was measured for 10 sec by using a microplate luminometer (MicroLumatePlus LB 96V; Berthold Technologies).
Coimmunoprecipitation and Western Blotting.
First, 3.6 × 106 HEK293RI cells were seeded 24 h before transfection with 6 μg of pMyc-mIL33Rα and/or 6 μg of pFLAG-mIL1RAcP after the PEI transfection method. Transfected cells were kept either unstimulated or stimulated with 10 ng/ml rhIL-1β or 100 ng/ml rmIL-33 for 15 min at 37°C, washed, and lysed with lysis buffer (35) for 30 min at 4°C. Nuclei were removed before the supernatant was incubated with anti-FLAG M2 agarose beads (Sigma–Aldrich) for 16 h at 4°C with gentle rotation. For anti-Myc immunoprecipitation, 0.6 μg of anti-c-Myc antibody (Santa Cruz Biotechnology) with protein A Sepharose beads (GE Healthcare) was used instead. Washed beads were resuspended in Laemmli sample buffer and heated to 95°C for 10 min. After SDS/PAGE, Myc- or FLAG-tagged proteins were detected in a Western blot with anti-c-Myc (Santa Cruz Biotechnology) or anti-FLAG bioM2 (Sigma–Aldrich) antibodies.
Interaction of sAcP and sIL-33Rα.
N-terminally FLAG- and His (6)-tagged rm-soluble AcP (Genbank accession no. NM_134103) was prepared by conventional techniques. cDNA was obtained from EL-4 5D3 mRNA by RT-PCR, cloned into pIRESneo3 (BD Biosciences), and expressed in stably transfected HEK293 cells. FLAG affinity gel (Sigma–Aldrich) was loaded with sAcP by incubation with cell supernatant and washed. Then 50 μl of loaded gel was resuspended in 1 ml of PBS with 2% BSA and incubated at room temperature for 1 h with or without rmIL-33 and with or without mouse sIL-33Rα:Fc (1004-MR; R&D Systems) at 1 μg/ml. Gel pellets were again washed, and bound proteins were resolved by SDS/PAGE and blotted onto nitrocellulose. Coprecipitated sIL-33Rα:Fc was visualized by using HRPO-coupled anti-human IgG antibody (A0170; Sigma–Aldrich).
Measurement of Cytokine Production.
EL-4 or D6/76 cells were seeded into 96-well plates at 2.5 × 104 cells per well in 200 μl of medium. Then, 24 h after transfection, cells were stimulated with various concentrations of rmIL-33 (or rhIL-1β or rmIL-18) in the presence of 0.5 μM A23187 for 16 h at 37°C. Mouse IL-2 was determined in the supernatants by ELISA (OptEIA mIL-2 Set; BD Biosciences). Mouse IL-6 was measured in supernatants from BMD mast cells (seeded at a density of 1 × 106 cells per milliliter) by using an IL-6 ELISA kit (BD Biosciences). If the anti-mIL-1RAcP mAb 4C5 (kind gift from N. Dimoudis, Roche, Penzberg, Germany) was used in inhibition studies, the cells were always preincubated with the mAb for 30 min at 37°C before cytokines were added. The rat IgG mAb RA3–6B2 (anti-B220) served as an isotype control. To establish the IL-33-stimulated cytokine profile of BMD mast cells, a panel of cytokines was measured by using Beadlyte MultiCytokine Flex kits (Upstate Biotechnology) according to the manufacturer's instructions and a Luminex 100 instrument.
Mammalian Cell Culture.
EL-4 cells were originally obtained from H. R. MacDonald (Ludwig Institute for Cancer Research, Epalinges, Switzerland). D6/76 cells were generated as previously described (17). 6-IRAK-19 cells were generated by stably transfecting EL-4 cells with a plasmid encoding hIRAK-1 (J. Knop and M.U.M., unpublished data). All murine thymoma cell lines were cultured in RPMI 1640, with the addition of 2 mM l-glutamine and 5% FCS (PAA Laboratories) at 5% CO2 in a humidified incubator at 37°C. HEK293RI cells were originally obtained from Z. Cao (Tularik, South San Francisco, CA) and maintained in DMEM plus 10% FCS at 10% CO2. Murine BMD mast cells were generated as described recently (36).
Measurement of Signaling Pathways.
EL-4 cells were transfected with mIL-33Rα by using the DEAE dextran method. Then, 20 h after transfection, 2 × 106 cells were stimulated with 10 ng/ml rhIL-1β or 200 ng/ml rmIL-33 for 0, 4, 8, 12, 16, 20, 24, 28, 32, and 36 min at 37°C. JNK activity was measured in an in vitro kinase assay by using GST-cJun as a substrate (amino acids 1–79; Stratagene) as described previously (5). The total amount of JNK was measured in parallel samples by Western blot.
IRAK-1 Autophosphorylation Activity.
6-IRAK-19 cells were transiently transfected with a plasmid encoding mIL-33Rα-chain alone or in combination with a plasmid encoding ΔC-AcP. Then, 24 h after transfection, 2 × 107 cells were stimulated for 15 min at 37°C with either 10 ng/ml rhIL-1β or 100 ng/ml rm IL-33. Subsequently, cells were lysed, and IRAK-1 was immunoprecipitated by using the mAb 2A9 (a kind gift from Z. Cao and H. Wesche, Tularik) and subjected to an in vitro kinase assay as described previously (37).
Acknowledgments
We thank Dr. N. Dimoudis (Penzberg, Germany) for providing the neutralizing anti-mIL-1RAcP mAb 4C5, Dr. J. Strouboulis (Rotterdam, The Netherlands) for BirA cDNA, and Daniela Vogl, B. Reitz, and S. Franke for excellent technical assistance. This work was supported by a Deutscher Akademischer Austausch Dienst stipend (to S.A.).
Note Added in Proof.
While our manuscript was under review, four papers were published addressing different aspects presented here. Chakerian et al. (38) demonstrated that IL-1RAcP is the coreceptor for the IL-33Rα-chain, Allakhverdi et al. (39) and Iikura et al. (40) demonstrated IL-33 effects on human mast cells, and H. Hayakawa et al. (41) demonstrated the effect of soluble ST2 on IL-33 signaling.
Footnotes
- ¶To whom correspondence should be addressed. E-mail: michael.martin{at}bio.uni-giessen.de
-
Author contributions: W.F. and M.U.M. contributed equally to this work; C.K., W.F., and M.U.M. designed research; S.A., M.H., W.F., and M.U.M. performed research; M.H., S.C.B., and W.F. contributed new reagents/analytic tools; S.A., M.H., C.K., and M.U.M. analyzed data; and C.K. and M.U.M. wrote the paper.
-
The authors declare no conflict of interest.
-
This article is a PNAS Direct Submission.
- © 2007 by The National Academy of Sciences of the USA