Structure and signalling in the IL-17 receptor superfamily

IL-17A and IL-17F drive inflammation and autoimmunity

IL-17A and IL-17F, eponymous cytokines of the T_H17 cell lineage, are by far the best characterized. Both are covalent homodimers, and recent findings show that they also form IL-17A–IL-17F heterodimers (Figure 1)^{17, 18}. IL-17A–IL-17F heterodimers are produced at higher levels than IL-17A homodimers by human peripheral blood mononuclear cells in vitro ¹⁷, but the form that predominates in vivo is not well defined ¹⁹. IL-17A, IL-17F and IL-17A–IL-17F all signal through the same receptor subunits, IL-17RA and IL-17RC, which together form a heteromeric complex ^{20, 21}. Nonetheless, IL-17A and IL-17F have distinct biological effects. Studies comparing Il17a^-/- mice with Il17f^-/- mice indicate that IL-17A seems to have a more central role in driving autoimmunity than IL-17F, probably due to its more potent strength of signalling ^{19, 22},. The role of the IL-17A–IL-17F heterodimer in vivo will be more challenging to evaluate. The basis for functional differences between IL-17A and IL-17F could be the markedly decreased strength of signalling triggered by IL-17F compared with IL-17A, as IL-17F responses are 10–30-fold weaker in terms of downstream gene activation than those of IL-17A, with IL-17A–IL-17F heterodimers acting at an intermediate level ^{17, 23, 24}. The precise nature of their respective receptor complexes has not been defined (see also “IL-17RC” section below). As IL-17RA and IL-17RC have very different affinities for IL-17A and IL-17F, it is plausible that different complexes exist with varying ratios of IL-17RA and IL-17RC, which manifest different ligand preferences (Figure 1).

IL-17E is a T_H2 cell-promoting cytokine

IL-17E, commonly known as IL-25, is produced by mucosal epithelial cells, as well as by multiple immune cell types (Table 1). Transgenic overexpression of IL-25 promotes eosinophilia and stimulates the production of T_H2 cell-specific cytokines. Moreover, IL-25 limits T_H17 cell development by inducing the expression of IL-13 by dendritic cells (DCs) or by inhibiting IL-23 production by macrophages (reviewed in Ref. 25). IL-25 binds a receptor complex composed of IL-17RB (also known as IL-25R), which partners with IL-17RA (Figure 1) ²⁶. Il17rb^-/- and Il17ra^-/- mice fail to respond to IL-25, and both knockout strains are refractory to pulmonary inflammation induced by intranasal application of IL-25 ²⁷. Consequently, there is intriguing crosstalk between members of the IL-17 cytokine family.

IL-17B, -C and -D functions are poorly defined

IL-17B, IL-17C and IL-17D remain poorly characterized. IL-17B binds to IL-17RB with fairly high affinity ²⁸, and IL-17C was recently reported to bind IL-17RE and to activate nuclear factor-κB (NF-κB) (S. Levin, personal communication). The receptor(s) for IL-17D remains unknown. Ectopic expression of IL-17B and IL-17C by CD4+ T cells exacerbates collagen-induced arthritis [G] (CIA) ²⁹, and intranasal administration of adenoviruses expressing IL-17C, IL-17A or IL-17F triggers comparable neutrophilic responses ³⁰, suggesting that these cytokines mediate common biological effects, probably through shared signalling pathways. Antibodies specific for IL-17B also partially suppress CIA ²⁹. IL-17B and IL-17C stimulate expression of an array of pro-inflammatory genes similar to those induced by IL-17A and IL-17F ^{29, 31}. IL-17D was reported to promote a pro-inflammatory gene expression profile in endothelial cells, and it had a mild inhibitory effect on myeloid progenitor cell proliferation in vitro ³². As more knockout mice are generated and the target receptors for IL-17B, IL-17C and IL-17D become more clearly defined, we will gain an improved understanding of these newer IL-17 family members.

IL-17RA: expression and complex formation

IL-17RA was initially identified as the receptor for mammalian IL-17A and vIL-17 ¹³. IL-17RA also binds IL-17F, albeit weakly, and is necessary for IL-17A-, IL-17A–IL-17F- and IL-17F-mediated signal transduction ¹⁵. Early studies showed that the affinity of IL-17RA for IL-17A is lower than the concentration required to mediate responses, which indicates that an additional subunit is involved in binding ligand and/or eliciting signalling ³⁴. Indeed, IL-17RA partners with IL-17RC to induce responses to IL-17A and IL-17F ³⁵. Similarly, Il17ra^-/- mice are refractory to the effects of IL-25, suggesting that IL-17RA is also a component of this receptor complex ²⁷. Moreover, IL-17RA might associate with IL-17RD, as these subunits co-localize, and overexpression of a mutant IL-17RD lacking a cytoplasmic tail inhibits IL-17A-dependent signals ³⁶. However, this finding has yet to be validated, and no ligand has been identified for an IL-17RA–IL-17RD pairing (Figure 1). Based on its usage by multiple cytokines, IL-17RA may thus be analogous to common cytokine receptor subunits such as gp130 in the IL-6 family ³⁷.

IL-17RA is expressed ubiquitously, with particularly high levels in haematopoietic tissue ^{13, 22}. This expression pattern is curious, as the main responses to IL-17A occur in epithelial, endothelial and fibroblast cells, although macrophages and DCs are also responsive (reviewed in Ref. 38). Only a limited number of IL-17A-induced genes in lymphocytes have been documented, which are quite distinct from genes induced by IL-17A in other cell types ^{22, 39}. Although its expression is widespread, IL-17RA can be dynamically regulated. For example, IL-15 and IL-21 upregulate expression of IL-17RA in CD8+ T cells ^{40, 41}, and phosphoinositide 3-kinase (PI3K) limits IL-17RA expression in T cells. This could be biologically significant, because IL-17A-induced signalling strength correlates with cell surface expression levels of IL-17RA; in contrast to most cytokine receptors, high levels of IL-17RA seem to be required for effective responses ^{42, 43}. Another function of IL-17RA might be to limit signalling by receptor-mediated internalization of ligand. Surface expression of IL-17RA rapidly decreases after IL-17 binding, theoretically internalizing IL-17A and clearing it from the inflammatory milieu ⁴¹.

The composition of IL-17RA-containing complexes is poorly defined. The textbook paradigm of cytokine receptor signalling is that individual subunits reside in the membrane as monomers. Ligand binding promotes subunit oligomerization, thereby juxtaposing receptor-associated signalling intermediates such as kinases or adaptors. Inconsistent with this view are studies showing that many receptors actually exist as pre-associated, multi-subunit complexes (such as TLRs, tumour necrosis factor receptors (TNFRs), and erythropoietin receptor (EPOR); reviewed in Refs. ^{38, 44}). Mechanistically, pre-assembly poises a receptor to respond rapidly and specifically to ligand. In the case of the IL-17R, studies using fluorescence resonance energy transfer [G] (FRET) to analyse interactions between IL-17RA subunits showed that IL-17RA forms ligand-independent complexes ^45-47. Moreover, IL-17RA co-immunoprecipitated with IL-17RC in overexpression studies in a ligabd-dependent manner, although it is not known whether IL-17RC is pre-assembled with IL-17RA to any degree ³⁵. The precise stoichiometry of the IL-17A-binding complex has not been determined, but native gel analyses of IL-17RC are consistent with a trimeric complex containing two IL-17RA subunits and one IL-17RC subunit ⁴⁸.

For pre-assembled receptors such as the EPOR, ligand binding induces large conformational changes in the relative positions of the receptor subunits. This re-orients receptor-bound cytoplasmic signalling molecules such as JAKs, facilitating interactions with appropriate substrates ⁴⁹. This might be the case for IL-17RA, because FRET between IL-17RA molecules decreases after interaction with ligand ^{45, 46}. One possible scenario is that IL-17RA dimers dissociate and are replaced by IL-17RA–IL-17RC heterodimers. Alternatively, IL-17RC might be recruited to the IL-17RA dimer to create a trimer or larger multimer, together with cytoplasmic signalling molecules such as ACT1 (also known as TRAF3IP2 and CIKS). A better understanding of this process awaits the IL-17R crystal structure.

In the TNFR system, pre-assembly is dictated by an extracellular cysteine-rich motif known as a pre-ligand assembly domain (PLAD). Soluble PLAD peptides have been shown to prevent TNFR assembly and thus block signalling in CIA ⁵⁰. Although it is structurally distinct from TNF receptors, IL-17RA encodes a fibronectin III-like domain (FN) PLAD that dimerizes in a yeast two-hybrid system and is required for co-immunoprecipitation of IL-17RA dimers ⁴⁵ Unlike the TNF system, where the PLAD and ligand-binding site of the receptor are physically discrete, the IL-17RA FN2 motif is also an essential component of the IL-17A-binding domain ⁴⁵. Strategies to block IL-17A or its receptor could therefore include soluble PLADs as well as antibodies or soluble receptors ⁴⁷

IL-17RA signalling: the NF-κB pathway

As IL-17RA lacks homology to known receptors ¹³, it was not possible to investigate signal transduction simply by comparison with other cytokine receptor systems. However, studies defining IL-17A-induced genes showed that IL-17A activates a highly pro-inflammatory programme of gene expression, typical of that induced by innate immune receptors such as IL-1R and TLRs ^{51, 52}. Similar to these receptors, IL-17A activates nuclear factor-κB (NF-κB), a hallmark transcription factor associated with inflammation (Figure 3A) ¹³. Correspondingly, DNA elements that bind NF-κB in the promoters of several IL-17A target genes are required for IL-17A-induced gene expression (reviewed in Ref. 38). Gel shift studies indicated that IL-17A activates p50 and p65, which are components of the classical NF-κB pathway [G] ⁵³. So far, there is no evidence that the non-classical NF-κB pathway [G] is involved in IL-17RA signalling (S.L.G., unpublished observations). However, NF-κB-inducing kinase (NIK), which can mediate events in the non-classical pathway, was reported to be activated in response to IL-17A ⁵⁴. IL-17A also induces the expression of IκBζ ^{51, 55}, which is a positive regulator of transcription by the TLR and IL-1R pathways ⁵⁵. Although it has not been clearly defined for IL-17A, it is likely that IκBζ is involved in IL-17A-dependent gene expression.

A wide diversity of inflammatory stimuli activate NF-κB, but they accomplish this through distinct proximal signalling pathways ⁵⁶. For considerable time, the proximal activators of NF-κB in the IL-17R pathway remained obscure. Hints came from observations that IL-17A signalling had parallels to IL-1R and TLR signalling cascades. For example, TRAF6 is a key adaptor in the IL-1R/TLR signalling cascade, and fibroblast cells from Traf6^-/- mice are defective in IL-17A-mediated induction of NF-κB ⁵⁷. However, the lack of an obvious TRAF6-binding motif in IL-17RA suggested that a signalling intermediate might be required to bridge IL-17RA and TRAF6. However, molecules involved in proximal IL-1R signalling, such as MYD88 (myeloid differentiation primary response gene), TRIF (TIR domain containing adaptor-inducing interferon-β), IRAK4 (IL-1R-associated kinase 4) and IRAK1, are dispensable for IL-17A-dependent signalling ^{43, 58} (F. Shen, L. Li and S.L.G., unpublished data). Consequently, early IL-17R signalling events are distinct from classical activators of innate immune responses (Figure 3B).

A key insight into IL-17R signaling emerged from a bioinformatics analysis that identified a conserved motif in the cytoplasmic domains of IL-17R-family members with homology to the Toll/IL-IR (TIR) domain [G] ⁵⁹. TIR domains provide essential and specific docking sites for intracellular adaptors such as MYD88. This study named the conserved IL-17R region a SEFIR domain, for SEF (similar expression to FGF receptor) /IL-17R. SEFIR domains lack certain elements found in prototypical TIR domains, potentially explaining why they do not engage TIR-containing adaptors. Specifically, SEFIR domains lack the TIR Box 3 subdomain and the BB-loop [G] ⁵⁹, which are crucial specificity determinants directing TIR-based protein–protein interactions ⁶⁰. Strikingly, a region C-terminal to the SEFIR domain in IL-17RA has marked sequence homology to BB-loops. Deletion of or point mutations in this region render IL-17RA non-functional, and so this motif is referred to as a “TIR-like loop” (TILL) ^{43, 61}. It is probable that the TILL provides a unique interaction surface for an as-yet-unknown signalling molecule, potentially serving as the Box 3 equivalent. Surprisingly, although all IL-17R receptors have a SEFIR motif, the TILL domain appears to be unique to IL-17RA, perhaps explaining why IL-17RA functions as a shared subunit (Figure 1).

A SEFIR domain is also present in ACT1, a signalling adaptor that was previously linked to NF-κB activation through BAFF (B cell activating factor) and CD40L ⁵⁹. Indeed, an ACT1 deficiency in knockout mice or generated by RNA interference-mediated knockdown impairs IL-17A- and IL-17F-induced activation of NF-κB ^{25, 62} (Figure 3). Co-immunoprecipitation studies indicate that ACT1 is recruited to IL-17RA within 5 minutes of IL-17A stimulation ⁶³. Supporting these observations, deletion of the SEFIR domain in IL-17RA impairs activation of NF-κB by IL-17A ^{43, 63}. Unlike IL-17RA, ACT1 contains a TRAF6-binding motif, and accordingly can bind TRAF6, TRAF3 and TAK1 ⁶³. However, Act1^-/- cells are still capable of delayed activation of ERK1/2 ⁶³, indicating that ACT1-independent pathways are also mediated by IL-17RA. The crucial role of ACT1 in IL-17A signalling has recently been verified in a second knockout mouse model ²⁶. Intriguingly, a respiratory pathogen, Chlamydia pneumoniae, encodes an Act1-binding protein that appears to inhibit IL-17A-mediated signalling, presumably providing a survival advantage to this organism ⁶⁴. Collectively, these studies illustrate that the proximal events in IL-17R-mediated activation of NF-κB are different from TLR/IL-1R due to unique structural components of IL-17 receptors (Figure 3B).

IL-17RA signalling: AP1, MAPK and mRNA stability

Another typical event induced by pro-inflammatory mediators is activation of the mitogen-activated protein kinase (MAPK) pathway, which leads to the activation of AP1 transcription factors. IL-17A activates various MAPKs, but ERK is generally the most strongly and rapidly phosphorylated. Although AP1 binding sites are enriched in IL-17A target promoters ⁴², the AP1 site in the mouse Il6 promoter is dispensable for IL-17A-mediated activation ⁵³. In general, the MAPK pathway has a more important role in regulating the expression of IL-17A-induced genes through control of mRNA transcript stability. MAPK stabilizes mRNA through the inhibition of destabilizing proteins such as tristetraprolin (TTP). TTP binds to AU-rich elements (AREs) in mRNA transcripts and delivers them to the exosome complex [G], whereupon they are degraded. Phosphorylation of TTP by MAPK blocks its ability to recruit the degradative machinery, thereby increasing mRNA transcript half-life (Reviewed in Ref 65). Many IL-17A target genes are chemokines or cytokines whose transcripts are stabilized by AREs located in the 3′ UTR ⁶⁵. Numerous studies show that activation of MAPK by IL-17A increases the concentration of these transcripts signiflcantly ³⁸. ACT1 is required for IL-17A-mediated stabilization of CXCL1 (also known as Groα or KC) mRNA, but TRAF6 is surprisingly dispensable ⁶⁶. Unexpectedly, the ability of IL-17A to stabilize TNF-induced CXCL1 mRNA is not compromised in Ttp-deficient fibroblasts (T. Hamilton, personal communication). In addition to TTP, other proteins are involved in mediating cytokine mRNA stability such as the recently described Zc3hl2a gene ⁶⁷. Thus, there is clearly much more to learn about the mechanisms by which IL-17A regulates transcript stability, but this is likely to be a very important means by which inflammation is controlled by IL-17.

IL-17RA signalling: C/EBP

IL-17A is consistently found to be a weak activator of NF-κB, suggesting the involvement of additional transcription factors in the response to IL-17. A microarray screen for IL-17A-induced genes identified the CCAAT/enhancer binding protein transcription factors C/EBPβ and C/EBPδ ⁵³. The promoters of genes upregulated by IL-17A are enriched for C/EBP binding elements ⁴², and activation of the Il6 and lcn2 (lipocalin 2, also known as 24p3) promoters have an absolute requirement for C/EBPβ/γ following IL-17A stimulation ⁵³. So, C/EBPβ and C/EBPδ are targets of IL-17A signalling and also important mediators of IL-17A-induced signalling pathways.

In some respects, C/EBPβ and C/EBPδ seem to function redundantly, as reconstitution of C/EBPβδ-deficient cells with either transcription factor can restore IL-17A-dependent induction of IL-6 or lipocalin 2 expression ⁵³. Nonetheless, there are essential differences in how C/EBPβ and C/EBPδ are regulated by IL-17RA. ACT1 and the SEFIR and TILL domains of IL-17RA are required for the induction of C/EBPδ expression, suggesting a role for NF-κB ^{43, 58}. This observation is consistent with a study showing that NF-κB can bind directly to the C/EBPδ promoter after LPS stimulation ⁶⁸.

C/EBPβ expression is regulated by IL-17A in a more complex manner. C/EBPβ exists in three isoforms that are generated by alternative translation ⁶⁹. The largest is a full length isoform known as liver enriched activator protein (LAP*). A shorter form, LAP, is generated from an alternative start codon and is thought to be the most transcriptionally active form of C/EBPβ. LIP (liver-enriched inhibitory protein) is also generated by alternative translation, and acts as a dominant negative inhibitor. The mechanism underlying these alternative translation events is not understood. IL-17A treatment of fibroblast cell lines results in a mild induction of LAP, a large increase in LAP* and no induction of LIP. Induction of LAP* depends on the SEFIR/TILL domain of IL-17RA, suggesting that this process is downstream of ACT1. Generation of LAP* also requires a poorly-characterized C-terminal domain in IL-17RA known as a C/EBPβ-activation domain (C-BAD) ⁴³, which is not found in any other IL-17R family members ⁶¹.

As with many signaling effectors, post-translational modifications of C/EBPβ such as phosphorylation are important for its activity ^{69, 70}. IL-17A triggers dual, sequential phosphorylation of a regulatory site within C/EBPβ. ERK phosphorylates one site (Thr-188) within 15 minutes of IL-17A stimulation. However, GSK (glycogen synthase kinase)-3β phosphorylates Thr-179 only after 1 hour, an event that requires prior phosphorylation at Thr-188 ⁶¹. Interestingly, these phosphorylation events are mediated by distinct subdomains of IL-17RA (Figure 3A). Regulation of ERK and hence phosphorylation at Thr-188 is mediated by the SEFIR–TILL region. ⁴³. By contrast, regulation of GSK3β is mediated by the C-BAD motif. In other systems PI3K signalling is upstream of GSK3β, but standard PI3K inhibitors have no effect on IL-17A signalling in this setting. Unexpectedly, the consequence of C/EBPβ phosphorylation is to downregulate its transcriptional capacity, and so this pathway is one of the few known inhibitory signalling events mediated by IL-17A ⁶¹. Because induction, alternative translation and phosphorylation of C/EBPβ all occur during a time frame in which C/EBPδ is also upregulated by IL-17A ⁵¹, there is a dynamic and complex interplay between members of the C/EBP family that could allow for detailed transcriptional control of C/EBP-dependent genes. Finally, C/EBPβ can be regulated by subcellular localization and association with numerous transcription factors, which provides future avenues of investigation in the context of IL-17R signalling ⁶⁹.

More broadly, these molecular events help to explain the synergism between IL-17 and other cytokines, particularly TNF ⁷¹. The mechanism underlying such synergy occurs partly through enhanced effects on mRNA stability ^{38, 72}, but C/EBP transcription factors help to mediate cooperative activation of target promoters. For example, TNF and IL-17A have an additive effect on the induction of the Il6 promoter, and overexpression of either C/EBPβ or C/EBPδ can replace the contribution of IL-17A to this cooperative signal ⁵³.

Other pathways activated by IL-17RA

Additional events have been implicated in IL-17RA signalling, although generally there is less evidence for these other pathways. Pharmacological inhibitors of Janus kinases (JAKs) have been reported to limit IL-17A signalling ^{73, 74}, but these results should be interpreted with caution due to the non-specific effects of such compounds. Weak activation of signal transducer and activator of transcription (STAT) factors has been reported, but secondary effects from IL-17A-induced secretion of cytokines such as IL-6 were not ruled out satisfactorily ³⁸. PI3K pathways have similarly been implicated, and IL-17A signalling in lung epithelial cells is associated with increased lipid phosphorylation, weak AKT phosphorylation and inhibition by specific PI3K inhibitors ⁷³. However there are few solid biochemical data indicating which PI3K isoforms are involved in this process. Additionally, IL-17RA, and hence downstream signalling, might be controlled by inducible degradation. An overexpression study showed that IL-17A triggers the ubiquitylation and degradation of IL-17RA, which might involve the E3 ubiquitin ligase TRAF6 ⁷⁵.

IL-17RC

IL-17RC (also known as IL-17RL) was identified by homology searches of mammalian expressed sequence tag (EST) databases. Complementation studies showed that IL-17RC is required for IL-17A- and IL-17F-mediated signalling, although it is insufficient to signal without IL-17RA. This study also revealed an unexpected species-dependent interaction between IL-17R subunits. Specifically, mouse and human IL-17A show no species dependence in signalling ¹³, whereas mouse but not human IL-17RA can reconstitute IL-17A-dependent signalling in murine Il17ra^-/- fibroblasts. Furthermore, co-expression of human but not mouse IL-17RC with human IL-17RA can restore signalling ³⁵. Consistently, fibroblasts from Il17rc^-/- mice fail to induce IL-6 or GM-CSF production in response to either IL-17A or IL-17F ⁷⁶. The reason for this species dependence is not clear, but could be important in evaluating anti-cytokine therapeutics in pre-clinical models.

The findings discussed above argue that the IL-17A-binding complex is an obligate multimeric receptor containing both IL-17RA and IL-17RC. However, these subunits have reciprocal expression patterns, suggestive of tissue-dependent activities ^{21, 22}. In contrast to IL-17RA, IL-17RC expression is low in haematopoietic tissues and high in non-immune cells of the prostate, liver, kidney, thyroid and joints ^{21, 77}. In terms of biological signalling, it is conceivable that differential expression of IL-17R subunits may provide an avenue for tissue-specific signalling by IL-17A and/or IL-17F. Whereas IL-17RA binds with an extremely low affinity to IL-17F, IL-17RC binds with higher affinity to IL-17F than to IL-17A ²¹. Accordingly, cells with high IL-17RC expression could be more responsive to IL-17F, whereas those with low IL-17RC expression might respond better to IL-17A. Consistent with this idea, studies in Il17a^-/- and Il17f^-/- mice indicated that IL-17A but not IL-17F mediates signals in murine T cells, correlating with undetectable levels of IL-17RC ²². This raises the intriguing question of whether IL-17RA signals as a homo-multimer in T cells, or whether there is another subunit that might pair with IL-17RA in certain cell types such as T cells. This information can also be valuable therapeutically; soluble IL-17RA can efficiently block IL-17A-dependent but not IL-17F-dependent responses, providing an avenue to selectively target individual cytokines (Figure 2) ²¹. Defining the precise nature of IL-17-binding complexes in different tissues is thus important for understanding signalling but also in terms of how to optimally target this system for clinical benefit.

With the exception of IL-17RA, other members of the IL-17R superfamily are highly spliced at sites in the extracellular domain, potentially giving rise to both agonistic and antagonistic (soluble) forms of the receptors. Indeed, more than 90 splice isoforms of IL-17RC were identified in human prostate cancer lines ⁷⁸. In mice, only four splice variants of IL-17RC are found in EST databases ²¹. Interestingly, there is ligand preference of IL-17RC splice isoforms, as certain forms bind preferentially to IL-17A or IL-17F; moreover, some isoforms do not bind either cytokine, suggesting that there might be additional ligands for this receptor subunit ²¹. The mechanism by which splicing is controlled has not been determined, but may be important for directing the ability of a particular cell or tissue to respond to different IL-17 family cytokines.

As expected, the IL-17RC cytoplasmic tail is required for signal transduction ³⁵, but not for inducible co-immunoprecipitation with IL-17RA (W. Ouyang, personal communication). So far, little is known about exactly how IL-17RC participates in signalling. IL-17RC contains a SEFIR but not an obvious TILL domain, and it is not known whether IL-17RC binds directly to ACT1, TRAF6 or other downstream signalling molecules. It is tempting to speculate that IL-17RC might recruit unique molecules to the receptor complex, but it is possible that IL-17RC simply increases the number of ACT1 molecules within the complex to facilitate more efficient signal transduction.

IL-17RB

IL-17RB (also known as IL-25R, Evi-27 or IL-17rh-l) binds both IL-17B and IL-17E/IL-25. This subunit is expressed by various endocrine tissues as well as kidney, liver and T_H2 cells ⁷⁹. IL-17RB is also highly spliced in the extracellular domain ⁸⁰, although no evaluation of the role of each splice variant has been reported. Recent evidence indicates that IL-17RB pairs with IL-17RA to form a functional IL-25R complex, but the specific contribution of each subunit to downstream signalling is unclear ²⁷. This finding raises issues regarding conclusions drawn from studies of Il17ra^-/- mice, as effects attributed to IL-17A and IL-17F could involve additional contributions from IL-25. Unlike IL-17RA, IL-17RB encodes a TRAF6-binding motif in its cytoplasmic tail. Indeed, antibody-mediated crosslinking of the receptor activates NF-κB, which can be blocked by a dominant-negative form of TRAF6 but not TRAF2 ⁸¹. IL-25 has been shown to activate NFATc1 and JUNB, leading to increased IL-4 expression by T_H2 cells ⁸². Like the rest of the family, the IL-17RB cytoplasmic tail contains a SEFIR domain, and was recently shown to bind ACT1 in a SEFIR-dependent manner. Accordingly, Act1^-/- mice show deficits in responding to IL-25 in vivo ^{26, 83}.

IL-17RD

IL-17RD has no known ligand (Fig. 1). It was first identified in zebrafish (Danio rerio) based on a similar expression pattern to the fibroblast growth factor receptor (FGFR), and thus is commonly known as SEF (similar expression to the FGF receptor). IL-17RD seems to be the most evolutionarily ancient member of the IL-17R family, as it has homologues in sea lamprey, frogs and C. elegans ⁸⁴. Its role in zebrafish suggests that the original function of the IL-17R family may have been to control development. Interestingly, this represents another parallel to the TLR family, as Drosophila Toll acts to control dorsoventral polarity during embryonic development. IL-17RD inhibits FGF-mediated RAS–MAPK and PI3K signalling in both zebrafish and Xenopus laevis development ⁸⁵. IL-17RD blocks FGF in part by physically associating with FGFR1 and FGFR2. The cytoplasmic tail of IL-17RD is required for its ability to inhibit signalling and associate with FGFR, but the extracellular domain is dispensable ⁸⁵. The human homologue of IL-17RD also co-immunoprecipitates with FGFR1 and can inhibit FGF-dependent ERK activation and FGF-dependent proliferation ^86-88. Although the mechanism by which IL-17RD-mediated inhibition occurs is not defined, murine IL-17RD has been shown to interact with TAK1 to activate the MKK4–JNK pathway ⁸⁹. In addition to associating with FGFR, human IL-17RD has been shown to form homodimers, although it is not clear whether these are functional ⁸⁶. IL-17RD can also interact with IL-17RA, although the biological significance of this association remains unclear ³⁶. IL-17RD-deficient mice are viable and have no obvious developmental abnormalities, but detailed phenotypic analysis has not yet been carried out (J. Tocker, personal communication).

IL-17RE

IL-17RE is the least understood member of the IL-17R family, but recent studies suggest that its ligand is IL-17C (S. Levin, personal communication). IL-17RE is highly spliced, with six isoforms identified in EST databases. Ectopic expression of an EPOR–IL-17RE construct in the myeloid BaF3 cell line indicated that this receptor might promote proliferation, but this occurred in a ligand-independent manner, which raises concerns about the validity of the assay system ⁹⁰.