Inhibition of IL-6 family cytokines by SOCS3

Abstract

IL-6 a multi-functional cytokine with important effects in both inflammation and haematopoiesis. SOCS3 is the primary inhibitor of IL-6 signalling, interacting with gp130, the common shared chain of the IL-6 family of cytokines, and JAK1, JAK2 and TYK2 to control both the duration of signalling and the biological response. Recent biochemical and structural studies have shown SOCS3 binds to only these three JAKs, all of which are associated with IL-6 signalling, and not JAK3. This specificity is determined by a three residue “GQM” motif in the kinase domain of JAK1, JAK2 and TYK2. SOCS3 binds to JAK and gp130 simultaneously, and inhibits JAK activity in an ATP-independent manner by partially occluding the kinase’s substrate binding groove with its kinase inhibitory region. We therefore propose a model in which each of gp130, JAK and SOCS3 are directly bound to the other two, allowing SOCS3 to inhibit IL6 signalling with high potency and specificity

Introduction: IL-6 Signalling

IL-6 is a pleiotropic cytokine that exerts both inflammatory and anti-inflammatory effects depending upon its cellular context and is an important differentiation factor during haematopoiesis (reviewed in [1]). IL-6 belongs to a family of cytokines that also include IL-11, IL-27, LIF, OSM, CT-1 and CNTF. These cytokines are structurally similar[2] and signal via association with cell-surface trans-membrane receptors that each consist of a dimer (or higher-order oligomer) of the common shared chain, gp130 and a cytokine-specific alpha chain[3, 4].

In classical IL-6 signalling, IL-6 first associates with its specific receptor alpha chain, IL-6Rα, and this dimer then associates with gp130 to form a hexameric signalling competent complex with 2:2:2 stoichiometry[5, 6]. Whilst gp130 is expressed on the surface of most cell-types, IL-6Rα expression is more restricted. However, many cells which do not express IL-6Rα still respond to IL-6 by virtue of circulating soluble IL-6Rα (sIL-6Rα). This is termed trans-signalling and is often associated with the pro-inflammatory effects of IL-6[7, 8]

In both classical and trans-signalling, once the gp130 dimer is occupied there is an autoactivation of associated JAKs (Janus Kinases) which are found in an inactive state prior to cytokine exposure [9]. Based on similarity to receptor tyrosine kinase (RTK) signalling (for example insulin signalling) [10], activation is thought to occur by auto-phosphorylation in trans. In more detail, according to this model one JAK molecule bound to one cytokine receptor chain is phosphorylated by the JAK molecule bound to the other receptor chain (and vice-versa) within the receptor homo- or hetero-dimer. Activation involves phosphorylation of specific tyrosine(s) within the activation loop of the kinase [9] which causes the activation loop to translocate out of the active site in order to allow ATP and substrate to bind [11]. JAK1, JAK2 and TYK2 have all been found associated with gp130[12] in certain contexts however genetic deletion of these kinases has implicated JAK1 as the most important member of the family for gp130 induced signalling[13]. Upon activation, JAKs then phosphorylate five specific tyrosines on the cytoplasmic domain of gp130. Four of these phosphotyrosines are recruitment sites for STAT1 and/or STAT3 (Signal Transducer and Activator of Transcription-1 and −3) which are then activated by phosphorylation, again through the kinase activity of JAK1, JAK2 or TYK2[14]. STAT1 and STAT3 are latent transcription factors and once activated, they translocate into the nucleus and induce the transcription of appropriate IL-6-responsive genes. Thus STATs are the primary drivers of the biological response (See Figure 1, left). However, in addition to the JAK/STAT cascade, the MAP kinase and PI3 kinase pathways are also activated. This is via the fifth tyrosine, Y⁷⁵⁹, which, once phosphorylated, is a docking site for SHP2. SHP2 is activated by phosphorylation after binding and this leads to stimulation of both the MAPK/ERK and PI3 kinase pathways[15].

Figure 1. IL-6 signalling and its inhibition.

Schematic diagram illustrating activation (left) of the JAK/STAT signalling cascade in response to IL-6 and the termination of signalling (right) catalysed by SOCS3. IL-6 signals via a cell-surface receptor that consists of a 2:2 gp130(blue):IL-6Rα (orange) tetramer. Interaction between the cytokine and its receptor induces the autoactivation (in trans) of Janus Kinases (JAKs,: JAK1, JAK2, TYK2; shown in beige) bound to the cytoplasmic domain of gp130. Activated JAK then phosphorylates five tyrosines within gp130^cyt. Four of these phosphotyrosines recruit STAT3 or STAT1/STAT3 which are then themselves phosphorylated, and thereby activated, by JAK, translocate to the nucleus and begin inducing the transcription of IL-6-responsive genes. STATs also upregulate the transcription of SOCS3 (red) which binds to the fifth phosphotyrosine in gp130^cyt (pY⁷⁵⁹) and shuts down the JAK/STAT signalling cascade by binding to JAK and directly inhibiting its catalytic activity, forming a negative feedback loop. This phosphotyrosine also recruits SHP-2, which leads to activation of the MAPK/ERK and PI3K pathways (not shown here) and therefore SOCS3, which competes for this site, is also capable of inhibiting those signalling cascades. Signalling and inhibition is symmetric with respect to both gp130 chains and is shown here divided into left and right for ease of illustration. The structures shown are those solved and/or modelled for components of the signalling cascade, Note that the pseudokinase and SH2-like domains of JAK are omitted for clarity in this figure.

In addition to driving the biological response, activated STAT3 also induces expression of SOCS3 (Suppressor of Cytokine Signalling-3). SOCS3 in turn terminates the JAK/STAT signalling cascade, forming a negative feedback loop that allows the cell to return to its basal (unstimulated) state (Figure 1, right). This action of SOCS3 appears to be the primary mechanism by which IL-6 signalling is regulated within the organism. This review will focus on the mechanism by which SOCS3 inhibits IL-6 (and IL-6 family) signalling.

Discovery of the SOCS proteins

In 1997 the SOCS family of proteins were discovered concurrently by the groups of Hilton (Walter and Eliza Hall Institute, Australia), Yoshimura (Kurume University, Japan) and Kishimoto (Osaka University, Japan)[16–18]. Each group used a different approach. Hilton et al., used an expression cloning methodology to identify proteins capable of inhibiting the IL-6-induced differentiation of the mouse M1 myelomonocytic cell-line and discovered, and named, SOCS1 (Suppressor of Cytokine Signalling 1). Yoshimura’s group discovered the same entity via a yeast two-hybrid screen aimed at identifying proteins that bind to JAK and termed the protein JAB (JAK-binding protein). Finally, Kishimoto et al isolated a protein (SSI-1) on the basis of a short region of sequence similarity with STAT3. SSI-1 was found to be related to the SH2 domain-containing protein CIS (Cytokine inducible SH2 domain containing protein) and identical in sequence to SOCS1/JAB identified by the other two groups. Collectively, these three manuscripts described the major attributes of the SOCS1 protein: (A) That its expression is induced by a variety of cytokines; (B) it then inhibits the signalling cascade initiated by those same cytokines, forming a negative feedback loop; (C) it acts by binding to, and directly inhibiting, JAK with (D) the end result that STAT activation is curtailed.

At the same time as they discovered SOCS1, the group of Hilton et al., cloned two other proteins with similar domain architectures, termed SOCS2 and SOCS3. These three proteins, in addition to the already identified protein CIS[19–21], all contained an SH2 domain (responsible for binding phosphotyrosine residues) and a short, approximately 40 amino acid domain at their C-terminus that the authors termed the SOCS box. Subsequently, an extensive search of the genome databases discovered a further four proteins that shared this common domain structure (SH2 domain-SOCS box domain) and these were termed SOCS4-7[22].

The SOCS family

Evolutionarily, SOCS proteins are first seen in animals having bilateral symmetry[23]. Recent analyses suggest the existence of three SOCS proteins in these organisms: orthologues of CIS/SOCS1/SOCS2/SOCS3 as well as SOCS4/5 and SOCS6/7[23, 24]. Whilst certain species, most notably the fruit fly D. melanogaster have lost one or more of these three subgroups, they are all represented in vertebrates where they have expanded to form the eight family members seen in mammals. Of these three subgroups it is the CIS/SOCS1/SOCS2/SOCS3 class which are strictly associated with the control of cytokine signalling whereas the function of the SOCS4-7 homologues may also include regulation of RTK (Receptor Tyrosine Kinase) signalling such as that initiated by epidermal growth factor [25, 26], [27, 28].

The most notable feature of the SOCS family is their SOCS box domain[22]. The SOCS box is responsible, via an interaction with elonginsB/C and Cullin5, for promoting the ubiquitination of target proteins[29–31]. In general, these target proteins are cytokine receptors and thus SOCS proteins mostly function to control cytokine action by inducing the degradation of specific cytokine receptors[29, 31–34].

However the two most potent SOCS family members, SOCS1 and SOCS3 act primarily via a different mechanism[35, 36], distinct from that common to other SOCS proteins[26, 37–42]. They function by directly inhibiting the enzymatic activity of the JAKs, the initiators of the intracellular signalling cascade induced upon cytokine exposure and this mode-of-action is the major subject of this review.

SOCS3 is the primary regulator of IL-6 signalling

Despite SOCS1 being discovered on the basis of its ability to inhibit IL-6 action when overexpressed[17, 18], genetic deletion studies have surprisingly shown that SOCS1 plays little, if any, role in inhibiting IL-6 in vivo[43]. Rather it is SOCS3 that is the family member responsible for inhibiting IL-6 under physiological conditions[44–46]. This is a cautionary tale regarding the interpretation of the effects of individual SOCS proteins on various cytokines; whilst many SOCS proteins inhibit a number of different cytokines when artificially over-expressed, under normal conditions their activity is usually highly specific for only a few cytokines. This has been made clear by genetic deletion of SOCS1 and SOCS3 in mice which has highlighted their true role as regulators of signalling by interferon-α/γ^36,40–43 and IL-6/G-CSF/Leptin/LIF[44, 46–50] respectively.

SOCS3 controls the duration of IL-6 signalling

Genetic deletion of SOCS3 in mice is lethal due to placental insufficiency as a result of dysregulated signalling by LIF[48]. Therefore, confirmation of the important role that SOCS3 plays in regulating signalling by other IL-6 family members, including IL-6 itself, has been via conditional knockout of the Socs3 gene. The first such studies were knockouts of SOCS3 in hepatocytes[46] and macrophages[44] (using cre recombinase under control of the Albumin or LysM promoters respectively) and the use of Socs3^−/− fetal liver cells to repopulate wild-type mice[51]. These experiments showed that the loss of SOCS3 had a profound effect on the duration of signalling induced by exposure of these cells to IL-6. For example, when wild type mice are injected with IL-6, activated (phosphorylated) STAT1 and STAT3 are detectable in liver cells from approximately 15 minutes after IL-6 exposure, but return to basal (undetectable) levels after approximately 30 minutes and 2 hours respectively[46] (Figure 2a). Loss of SOCS3 has no effect on the magnitude or time of initiation of JAK/STAT signalling after IL-6 exposure but led to a four- and two-fold increase in the persistence of activated STAT1 and STAT3 respectively. Whilst a 2–4 fold increase may seem like only a mild molecular defect it has drastic consequences for the animal. Mice lacking SOCS3 in their haematopoietic system (vavCre) develop a lethal inflammatory disease, largely due to dysregulated IL-6 signalling[45].

Figure 2. SOCS3 inhibits the duration of IL-6 signalling by direct inhibition of JAK1, JAK2 and TYK2 but not JAK3.

(A) Schematic diagram showing the effect of SOCS3 on STAT1 and STAT3 activation post IL-6 stimulation. Shown is a representation of the data from[46] (B). The structure of SOCS3 (PDB 4GL9) with an explanation of the major functional motifs shown as a schematic below. Note that the PEST motif is absent from the structure and that the SOCS box has been modelled based on the structure of the SOCS2 SOCS box (PDB 2C9W). (C) SOCS3 (red) binds the kinase domains of JAK1, JAK2 and TYK2 and inhibits its catalytic activity by blocking the substrate binding site with its kinase inhibitory region (black). Note that SOCS3 remains bound to gp130 (blue) whilst in complex with JAK (beige) and that ATP binding is unaffected.

SOCS3 shapes the cells response to IL-6

As well as controlling the duration of IL-6 signalling, SOCS3 also helps shape the cell’s response to IL-6. For example, the transcriptional output of Socs3^−/− macrophages stimulated with IL-6 differs not just quantitatively but also qualitatively from that of wild-type cells. In particular, loss of SOCS3 leads to an IL-6-induced transcriptional response that in part resembles that for interferon-γ with a number of interferon-inducible genes being switched on by IL-6 in these cells[51]. Likewise, IL-6 stimulation of Socs3^−/− haematopoietic progenitor cells skewed differentiation toward the macrophage lineage rather than neutrophil lineage seen with wild-type cells, again indicating that it shapes the response to IL-6 rather than simply inhibiting it[52]. One explanation for this phenomenon is that, in addition to inducing the phosphorylation of STAT3, IL-6 also induces low level STAT1 activation. In the presence of SOCS3, this activation of STAT1 is even more effectively curtailed than is the activation of STAT3[46, 53] thus preventing induction of a STAT1 (interferon-γ-like) transcriptional response. In the absence of SOCS3 therefore, STAT1 is “on” for long enough to induce the transcription of interferon-inducible genes leading to a qualitatively different cellular response.

SOCS3 interacts with gp130, the shared receptor for IL-6 family cytokines

STAT3 is activated by a number of different cytokines and is a powerful inducer of SOCS3 expression[54] . However, SOCS3 only feeds back to inhibit STAT3 that is activated in response to particular cytokines (for example IL-6) and not others (for example IL-10 or interferon-γ)[44, 51, 55]. The key to this specificity is that SOCS3 directly interacts with gp130, the co-receptor for IL-6 family cytokines[56–58]. This allows SOCS3 to specifically target the IL-6 signalling cascade and not those induced by other cytokines.

SOCS3 binds gp130 with high affinity [56, 57, 59, 60]. The interaction occurs via the SH2 domain of SOCS3 that binds to a motif on gp130 (centered upon pTyr⁷⁵⁹) only once it has been phosphorylated. As Tyr⁷⁵⁹ is only phosphorylated by JAK after IL-6 stimulation, SOCS3 cannot bind to an unstimulated IL-6 receptor and this ensures that a cell can still respond to the first wave of IL-6 stimulation, even if SOCS3 is present in the cytoplasm (for example if a different cytokine has already induced its expression). The interaction between the SOCS3 SH2 domain and gp130 has been well described both biochemically[56, 57, 59] and structurally[61] (see Figure 2b) and there are a number of features of the interaction responsible for its high affinity that are worth noting.

SH2 domains bind phosphotyrosine residues in peptides and proteins[62–66] but usually only when embedded within a particular sequence motif. This specificity is generally achieved by an interaction with the so-called BG loop of the SH2 domain (sometimes termed the specificity-determining loop) and amino-acids located 2–4 residues downstream (i.e. C-terminal) of the phosphotyrosine on the target molecule. For the vast majority of SH2 domains therefore, specificity and affinity for a target sequence are encoded only by the pTyr residue and the amino-acids immediately downstream of it. However, SOCS3 is unusual in that it also has a requirement for specific residues upstream of the phosphotyrosine on its target molecules[56]. In particular, SOCS3 contacts a hydrophobic residue at the pY-2 position on gp130, a valine. The interaction with this valine adds 10-fold to the affinity with which SOCS3 targets gp130[56] (K_D=100nM) compared to a typical SH2-target interaction which is usually of micromolar affinity[66]. The SH2 domain of SHP-2 is similar in this regard and SHP-2 is known to bind a number of SOCS3 targets[56, 57, 67]. In addition to the pY-2 interaction, the BG loop of SOCS3 makes extensive contacts with the Val-Val-His sequence in the pY +3 to +5 region of gp130[61] which also contributes to the high affinity. In addition to gp130, SOCS3 also interacts with the receptors for leptin[58], G-CSF[68] and potentially EPO[69] although all of these interactions are at least 10-fold weaker than that seen for gp130. These interactions reveal a minimal consensus motif of V/L-X-pY-X-X-V/L-V/L-X.

The other major feature of the SOCS3 SH2 domain is that it contains a large (35 amino acid) unstructured loop inserted immediately prior to the specificity determining BG loop[60]. This loop is a PEST motif (Pro, Glu, Ser, Thr rich motif), a motif first described by Rogers on the basis of their being found in a number of intra-cellular proteins with very short half-lives[70]. The PEST motif does not effect the structure of the SH2 domain, as a comparison between the NMR structure of wild-type SOCS3[71] and the crystal structure of a PEST-deleted construct[61] clearly shows. Neither does it interfere with the SH2 domain function as both wild-type and PEST-deleted SOCS3 bind a gp130 phosphopeptide with similar affinties[71]. Rather it reduces the stability of SOCS3 inside the cell and leads to its proteolytic degradation in a mostly non-proteasome dependent fashion when tested in 293T cells[71]. Several other SOCS proteins are predicted to contain a PEST motif although this has not yet been verified by half-life studies. The most obvious is CIS, which contains a predicted PEST motif in the same position within its SH2 domain as SOCS3[60]. SOCS3 is known to have an extremely short half-life in certain cell-types[72]. This is likely the result of several mechanisms, including ubiquitination of lysine6[72], the presence of the SOCS box[73, 74] and PEST motif itself. SOCS3 turnover appears necessary for the cell to be able to respond to subsequent rounds of cytokine stimulation and therefore its half-life is strictly regulated.

The interaction of SOCS3 with gp130 at pTyr⁷⁵⁹ is competitive with the binding of SHP2. It has been shown that SOCS3 deletion (but not just deletion of its SOCS box) leads to hyper-phosphorylation of SHP2, presumably via increased accessibility to its binding site on gp130, in response to LIF stimulation of embryonic stem (ES) cells^32,[75]. Phosphorylation of SHP2 appears to be the major mechanism of activating the MAPK/ERK pathway in IL-6/LIF signalling[76]. Consequently, Socs3^−/− ES cells display extended activation of pERK1/2 in response to LIF signalling³²,^[75]. Socs3^−/− ES cells, unlike wild-type ES cells, display reduced self-renewal and spontaneous differentiation into primitive endoderm in the presence of LIF and this differentiation could be prevented by the use of MAPK/ERK pathway inhibitors[75]. This indicates that this mode of inhibition by SOCS3 (attenuation of MAPK signalling) is independent of its E3 ligase activity and has an important role in the biological outcome of IL-6/LIF signalling.

The interaction of SOCS3 with gp130 is a key molecular determinant of its specificity and its ability to inhibit cytokine signalling. However, whilst the SH2 domain-gp130 interaction is sufficient to inhibit the MAPK/ERK pathway post IL-6 stimulation it is not sufficient to inhibit the JAK/STAT signalling cascade. Its ability to inhibit the JAK/STAT pathway relies upon an interaction with, and inhibition of, JAK whilst both entities are scaffolded on gp130. This is largely due to the “kinase inhibitory region” (KIR) of SOCS3 and this mechanism will now be discussed in detail.

The Kinase Inhibitory Region (KIR) of SOCS3 allows it to directly inhibit JAK’s catalytic activity

SOCS3 in mice and humans is a 225 amino acid protein that, like all SOCS proteins, contains an SH2 domain (residues 45–185) and a SOCS box domain (residues 186–225)[18, 22]. SOCS3 also contains a short N-terminal segment (residues 1–44), the most notable feature of which is the so-called Kinase Inhibitory Region (KIR)[77–80], an 8–12 amino acid sequence that allows it to directly inhibit JAK’s catalytic domain and is absolutely required for function (Figure 2b).

The existence of the KIR was first identified in both SOCS3 and SOCS1 in two seminal papers by Yoshimura’s group in 1999[78, 79]. These manuscripts defined the KIR as a 12 amino-acid sequence (residues 22–33), upstream of the SH2 domain. The KIR is only found in SOCS1 and SOCS3 and has been shown to be required for both interaction with, and inhibition of, JAK[78]. Although they are unstructured in the absence of JAK[60, 61, 71], the first eight residues of the KIR adopt an extended conformation that occupies the interface between the JAK kinase and the SOCS3 SH2 domain when the two proteins are in complex[81]. The four C-terminal residues of the KIR are also structured upon JAK binding, forming an extra, N-terminal, turn on the ESS (extended SH2 subdomain) helix. The ESS helix is a 14-residue alpha-helix immediately prior to the N-terminus of the SH2 domain and is a feature that is shared by all eight SOCS proteins. This helix is integral to the stability of the SOCS3 SH2 domain as, when deleted, the protein becomes unstable. The ESS covers a large hydrophobic surface on the under-side of the central β-sheet of the SH2 domain which gives it a very fixed geometry relative to the rest of the domain. This geometry may be important for positioning KIR in the case of SOCS1 and SOCS3. Now that the structure of SOCS3 in complex with JAK has been solved, we favour a redefinition of the KIR as consisting of residues 22–29 of SOCS3 and the ESS helix as residues 30–44 (Figure 2b).

The KIR inhibits JAK by partially blocking the substrate binding groove on the surface of the kinase (Figure 2c). This prevents substrates (for example STAT3) from accessing the active site. Tyrosine kinases catalyse the transfer of the terminal (γ) phosphate from ATP to a tyrosine hydroxyl moiety and are therefore two-substrate enzymes: ATP (the phosphate donor) and the tyrosine-containing protein/peptide (the phosphate acceptor). Whilst the tyrosine-containing protein/peptide is occluded from its binding site when SOCS3 is bound to JAK, ATP binding remains unperturbed (Figure 2c). This makes SOCS3 a non-competitive inhibitor (with regards to ATP) of JAK and may be an important aspect to its function as it does not need to compete with the high concentrations of ATP found in the cytoplasm[82].

SOCS3 inhibits JAK1, JAK2 and TYK2 but not JAK3

Whilst the KIR is required to inhibit JAK, it is not sufficient. There is no detectable inhibition of JAK using a SOCS3 KIR peptide[82]. The structure of SOCS3 bound to JAK2 shows that only approximately 20% of the buried surface area within the complex involves the KIR. The majority of the SOCS3:JAK affinity is derived from an interaction between the SH2 domain of SOCS3 and JAK. It is important to note that this does not involve the classical phosphotyrosine binding groove on the SH2 domain, which remains accessible for binding to gp130. Rather it is on the opposing face of the domain and also involves residues on the ESS. This surface wraps around helix αG of JAK, in particular a three residue “GQM” (Gly-Gln-Met) motif at its N-terminal end.

Interestingly, this GQM motif is only found in JAK1, JAK2 and TYK2 but not JAK3. In vitro inhibition assays have shown that SOCS3 can only directly inhibit JAK1, JAK2 and TYK2 and that the GQM motif is responsible for this specificity[82]. Thus all three JAKs found associated with gp130 are susceptible to SOCS3 inhibition whilst JAK3 is not. The GQM motif is conserved in JAK1, JAK2 and TYK2 throughout vertebrate evolution and is always absent in JAK3 suggesting an important distinction.

A model of SOCS3 inhibition of IL-6 signalling

Knockout studies have shown that SOCS3 is a highly potent and specific inhibitor of IL-6 family cytokines, G-CSF and leptin, despite the fact that its expression is induced by a much larger number of cytokines. Any model of SOCS3 action must explain this specificity. Our model is centred upon the fact that SOCS3 binds JAK and the IL-6 receptor simultaneously via two opposing surfaces. Thus it is a particular JAK/Receptor complex that is the true target of SOCS3, rather than an individual JAK or receptor alone. This allows SOCS3 to inhibit IL-6 signalling (in addition to G-CSF and leptin) with both (A) high specificity and (B) high potency (affinity).

Specificity is derived from the fact that only particular JAK/receptor pairs are targeted, overcoming the redundancy caused by the fact that all cytokines signal through only four different JAK kinases and six different STAT transcription factors. SOCS3 can inhibit JAK1, JAK2 and TYK2 however it does so effectively only when they are already bound to a cytokine receptor that contains a SOCS3 binding site such as gp130, G-CSFR and lepR.

High affinity is derived from the formation of an unusual three-way complex (JAK/gp130/SOCS3) in which each member is directly bound to the other two (see figure 3). JAK binds gp130 through its FERM domain and SOCS3 through its kinase domain. gp130 binds JAK via its Box1 motif and SOCS3 via pY⁷⁵⁹. Finally, SOCS3 binds gp130 via its phosphotyrosine binding groove and JAK via the KIR and a surface on the opposing face of its SH2 domain. This creates a SOCS3:JAK/gp130 avidity that is higher than the mere sum of the individual SOCS:JAK and SOCS:gp130 affinities, reminiscent of certain antibody-antigen interactions. This avidity arises from SOCS3 containing two independent binding sites for the JAK/gp130 complex (see Figure 3).

Figure 3.

SOCS3 inhibits IL-6 family signalling by targeting a gp130:JAK dimer.

There are two predictions from this model of SOCS3 action that can be made: (1) it will inhibit a broad range of cytokines, in fact any cytokine that uses JAK1, JAK2 or TYK2, when overexpressed and (2) when present at physiological levels it will only inhibit of cytokines whose receptors contain SOCS3 binding sites. The former phenomenon has been well described elsewhere[83–85], whilst the latter phenomenon correlates well with the known specificity of SOCS3 for IL-6, LIF, G-CSF and leptin (all of which use receptors with SOCS3 binding sites) signalling. To date, the highest affinity SOCS3 binding site is found on gp130 (pY759). SOCS3 binds gp130 with a >10-fold higher affinity than it does G-CSFR and LepR. This explains why SOCS3 is such an effective inhibitor of all IL-6 family cytokines.

The mechanism of SOCS3 is reminiscent of the inhibition of insulin signalling by Grb14[86]. Like SOCS3, Grb14 inhibits insulin signalling by binding to a specific phosphotyrosine on the insulin receptor cytoplasmic domain and then inhibiting its associated kinase activity by blocking the kinase’s substrate binding site with a short inhibitory region. The difference is that rather than being bound to a kinase, the insulin receptor is the kinase and thus both interactions made by Grb14 (scaffolding via its SH2 domain and kinase inhibition by it kinase inhibitory sequence) are with the same molecular entity, rather than a dimeric kinase/receptor complex consisting of two separate chains. Given this conservation of mechanism between these major inhibitors of cytokine and RTK (receptor tyrosine kinase) signalling pathways it will be interesting to determine whether other kinase-based signalling pathways are similarly controlled.

• SOCS3 is the primary feedback inhibitor of IL-6 family signalling

• SOCS3 controls the duration of IL-6 signalling and shapes the cells response to it.

• SOCS3 binds to gp130, the shared IL-6 family co-receptor

• SOCS3 directly inhibits JAK1, JAK2 and TYK2 but not JAK3

• SOCS3 targets gp130/JAK complexes

Abbreviations

JAKs

Janus Kinases

SOCS

Suppressor of Cytokine Signalling)

STAT

Signal Transducers and Activators of Transcription

IL-6

Interleukin-6

IL-10

Interleukin-10

IL-11

Interleukin-11

IL-27

Interleukin-27

gp130

glycoprotein 130

OSM

Oncostatin M

LIF

Leukemia Inhibitory Factor

CNTF

cillary neurotrophic factor

CT-1

Cardiotrophin 1

IL-6Rα

Interleukin-6 Receptor alpha-chain

GCSF

Granulocyte colony-stimulating factor

KIR

kinase inhibitory region

SH2

Src homology 2

SHP2

SH2 domain containing phosphatase

PI3K

Phosphoinositide 3-kinase

MAPK

Mitogen-activated protein kinase

ERK

extracellular-signal-regulated kinase