IL-23-IL-17 immune axis: Discovery, Mechanistic Understanding, and Clinical Testing

Initiating cytokines

Studies in 2003 and 2005 showed that IL-23 promotes the development and expansion of a pathogenic T cell population with a unique inflammatory gene signature^{1, 4, 6, 7}. However, IL-23 alone cannot drive differentiation of Th17 cells from naïve CD4+ T cell precursors, indicating that additional factors are required for their lineage fate determination. Shortly thereafter, Stockinger, Weaver, and Kuchroo simultaneously showed that addition of TGFβ and IL-6 during initial T cell recognition of cognate antigen promotes Th17 cell differentation^27–29. IL-6 has an essential role in this process by activating signal transducer and STAT3³⁰, which directly drives transcription of T_H17 lineage-specific genes³¹ including Rorc, Il17, and Il23r. STAT3 also suppresses TGFβ-induced forkhead box P3 (FOXP3) expression, and thereby inhibit the generation of regulatory T (T_Reg) cells²⁷. Consequently, Il6−/− mice are unable to generate Th17 cells and are protected from developing experimental autoimmune encephalomyelitis (EAE)³² and collagen-induced arthritis (CIA)³³. Although IL-6 was identified as a therapeutic target long before the recognition of Th17 cells, the recent advances have validated IL-6 as a target for the treatment of rheumatoid arthritis and other inflammatory conditions.

IL-1β signaling also has a critical role during the initial stages of Th17 cell differentiation. Il1r1−/− mice fail to develop antigen-specific Th17 cells and are resistant to EAE³⁴. Expression of IL-1 receptor type 1 (IL-1R1) by Th17 cells is induced by IL-6, and signaling through the IL-1R promotes the transcription factor interferon-regulatory factor 4 (IRF4), which reinforces the expression of RORγt³⁵. Mechanistically, IL-1β induces phosphorylation of the mammalian target of rapamycin (mTOR) in Th17 cells and thereby enhances the metabolic fitness of rapidly dividing Th17 cells during inflammation. Indeed, rapamycin treatment or mTOR-deficiency completely abolishes IL-1β-induced Th17 cell proliferation³⁶. These results suggest that IL-6 functions to direct differentiation of Th17 cells, whereas IL-1β functions to enhance expansion of Th17 cells in competition with other T cell subsets in the context of a resource-limited tissue environment.

In contrast to IL-1β and IL-6, the role of TGFβ in Th17 cell differentiation is more complex. Genetic approaches that used a dominant negative TGFβ receptor II, which inhibits the formation of a functional TGFβ signaling complex, and a T cell-specific deletion of TGFβ confirmed that endogenous TGFβ induces Th17 cell development in vivo^37–39. However, deletion of TGFβ also led to excessive production of IFNγ and IL-4, which inhibited Th17 cell development and expansion^{37, 38}. Thus, TGFβ can either function as an obligate Th17-inducing factor or could indirectly suppress alternate cell fates. Notably, in the absence of TGFβ, IL-6 alone can promote Th17 development in T cells that lack transcription factors T-bet and STAT6⁴⁰, which suggest that TGFβ represses Th17 differentiation indirectly by repressing T-bet and transcription factor GATA-binding protein 3 (GATA-3). This complexity is further highlighted by human studies; three separate research groups show that TGFβ is required for in vitro Th17 cell development, whereas two other research groups found that a cocktail of IL-1β, IL-6 and IL-23 is sufficient to drive Th17 differentiation^41–45. These discrepancies could be due to the source of the human T cells (for example peripheral blood versus umbilical cord blood), which may influence their developmental status and hence, the requirement for TGFβ. However, under in vivo inflammatory conditions, the presence of TGFβ is probably necessary for optimal Th17 cell differentiation.

Although the combination of TGFβ and IL-6 efficiently generates Th17 cells in vitro, the cells generated by this cocktail are only weakly pathogenic, and subsequent exposure to IL-23 is essential to drive pathogenicity^{21, 46–50}. A recent study further supported the notion that TGFβ is not always required for in vitro Th17 differentiation, as Th17 cells could be generated in response to IL-1β, IL-6 and IL-23 in a TGFβ-independent manner⁵⁰. Interestingly, these TGFβ-independent IL-23R⁺, RORγt⁺ Th17 cells are also T-bet⁺ and co-express IFNγ and GM-CSF^{4, 19, 20, 50}; such “hybrid” T cells with diverse functional plasticity (see below) are frequently detected in lesions from patients with multiple sclerosis (MS) and mice with EAE ^{51, 52}.

The studies mentioned above specifically examined TGFβ1. In a recent report, it was proposed that TGFβ3 promotes pathogenic T_H17 cells by upregulating IL-23R expression⁵³. Interestingly, TGFβ3 was induced by IL-23 and acted endogenously in a feed forward loop to enhance IL-23R expression, thereby amplifying the IL-23 signal⁵³. Taken together, these studies indicate that TGFβ together with IL-6 can drive development of IL-17- and IL-10-producing cells that have mucosal defensive functions; however, subsequent exposure to IL-23 is required for development of autoimmune-associated Th17 cells.

Differentiating and stabilizing cytokines

TGFβ- and IL-6-driven Th17 cells have limited inherent pathogenicity, and exposure to IL-23 is essential for the maturation of inflammatory Th17 cells¹. This is supported by genetic studies that link IL23R polymorphisms with susceptibility to autoimmune diseases such as psoriasis, psoriatic arthritis, ankylosing spondylitis, multiple sclerosis and Crohn’s disease^54–58. Accumulating data clearly indicates that IL-23 promotes pathogenicity of mature Th17 cells by several mechanisms, including maintenance of Th17 signature genes (Rorc and Il17), induction of effector genes (Il22, Csf2 and Ifng), downregulation of repressive factors (IL-2, IL-27 and IL-12) and amplification of its own signal by upregulation of Il23r expression.

By using an approach that restricts IL-23R deficiency to T cells, we have shown that without IL-23 signaling, Th17 cells are arrested at an early activation stage and are unable to mediate encephalitogenicity in EAE ⁴⁸. Subsequent studies revealed that IL-23R is needed for co-expression of GM-CSF; accordingly, GM-CSF-deficient mice are resistant to EAE and GM-CSF-deficient T_H17 cells cannot transfer EAE to naive recipients ^{19, 20}. IL-23 also induces IFNγ expression in Th17 cells, and IFNγ⁺IL-17⁺ cells are highly pathogenic^{52, 59}. To gain inflammatory function, Th17 cells must mitigate the impact of inhibitory cytokines such as IL-27 and IL-2 ^{60, 61}. Indeed, committed IL-23R⁺ Th17 cells are highly resistant to IL-27- and IL-2-mediated suppression⁶². This phenomenon could be due to downregulation of IL-27R in mature Th17 cells, and IL-23R signaling may also induce transcription factors that partner with RORγt to inhibit repressive signals, and thereby stabilize the Th17 lineage. Pathogenic Th17 cells induced by IL-23 in the absence of TGFβ also express reduced levels of aryl hydrocarbon receptor (AHR), C-MAF and IL-10 compared with TGFβ driven Th17 cells^{53, 59}. AHR and C-MAF are induced under TGFβ- and IL-6-driven Th17 differentiation conditions and transactivate IL-10 expression^63–65. However, it is unclear whether IL-23 directly suppresses these factors to drive pathogenicity. The ability of IL-23 to drive differentiation and stabilization of pathogenic Th17 cells is upregulation of IL-23R expression, which is STAT3-dependent ^{10, 59}. Taken together, these studies establish that TGFβ, IL-6 and IL-1β are essential elements that initiate Th17 cell development, whereas IL-23 serves as a pivotal factor that drives both the differentiation and inflammatory functions of pathogenic Th17 cells.

Positive activation of signaling

When IL-17R was cloned in 1995 it was noteworthy for its striking lack of similarity to known receptor families⁸³. Subsequently, it became clear that the IL-17R (now known as IL-17RA) was the founding member of a new subclass of cytokine receptors, which comprises five receptor subunits, IL-17RA–IL-17RE, that have homology to one another but not to other receptor families⁸² (Figure 3). All subunits of the IL-17R family exhibit a broad expression pattern; IL-17RA is most highly expressed in hematopoietic cell types and IL-17RC in non-hematopoietic cells⁸⁴. Most of the homology within the IL-17R family is located in the conserved cytoplasmic motif SEF/IL-17R (SEFIR), which suggests a common mode of signaling that is probably distinct from other cytokine receptors. A landmark bioinformatics study demonstrated that this domain was related to the Toll/IL-1R (TIR) domain. Recent studies have revealed that the IL-17RA subunit is shared among several members of this family (Figure 3). IL-17RA is essential for signaling through IL-17E (also known as IL-25) when partnered with IL-17RB, and also for signaling through IL-17C when partnered with IL-17RE^{85, 86}. The only other protein identified to have a SEFIR domain is the adaptor ACT1 (also known as CIKS)⁸⁷; indeed, ACT1 is essential for signal transduction by all analyzed IL-17R family members^{82, 88, 89}.

Despite the similarity between IL-17R and TLR/IL-1Rs, studies of IL-17 have revealed several differences in how IL-17R operates. For example, ACT1 is not a positive activator of signaling in any system outside the IL-17R family⁹⁰. ACT1 serves both as a signaling adaptor, to recruit tumour necrosis factor receptor (TNFR)-associated factor (TRAF) proteins to direct activation of several downstream signaling cascades⁸⁸, and as an E3 ligase to mediate ubiquitination of TRAF6⁹¹. Recruitment and ubiquitination of TRAF6 leads to activation of the canonical NF-κB pathway and the mitogen-activated protein kinase (MAPK) pathways, which include extracellular signal-regulated kinase (ERK), p38 and JUN N-terminal kinase (JNK). Consequently, activation of these pathways by IL-17 is impaired in the absence of either ACT1 or TRAF6^{88, 89, 92}. Alternatively, ACT1 can recruit TRAF2 and TRAF5, which promote activation of an mRNA stability pathway via the SF2/ASF or HuR complexes^{93, 94}. Regulation of these alternative pathways occurs via inducible ACT1 phosphorylation, which is mediated by Ikki (Ikkε) and TBK1 (Figure 4)^{95, 96}. ACT1 function and stability is in turn regulated by the chaperone protein hsp90. Inhibition of hsp90 results in proteasomal degradation of ACT1, which in turn leads to defective IL-17 signaling⁹⁷. ACT1 appears to be required for most IL-17-dependent signaling events⁹⁸ and the majority of phenotypes Act1^−/− mice are similar to Il17ra^−/− mice. Notably, a single-nucleotide polymorphism (SNP) in ACT1 has been associated with psoriasis⁹⁹. Thus, in the IL-17R family all known signaling emanates from ACT1, which leads to an activation of pro-inflammatory signaling cascades.

IL-17 induces signaling mainly in cells of non-hematopoietic origin, including epithelial and mesenchymal cell types. The genes that are induced by IL-17 largely explain its pro-inflammatory activities; for example, IL-17 induces expression of chemokines such as CXC-chemokine ligand 1 (CXCL1), CXCL2 and CXCL5 that promote neutrophil chemotaxis¹⁰⁰. IL-17 also induces expression of CC-chemokine ligand 20 (CCL20), which promotes trafficking of mucosal-associated cells expressing CC-chemokine receptor 6 (CCR6); notably, CCR6 is a characteristic receptor on IL-17-expressing cells such as Th17 cells and ILC3s ¹⁰¹. Other cytokines induced by IL-17, such as IL-6 and G-CSF, regulate myeloid lineages and especially neutrophils. Additionally, IL-17 is a strong inducer of antimicrobial peptides (AMPs), particularly β-defensins, which can prevent infection at mucosal surfaces and skin and that may also exert chemotactic activity¹⁰². Lipocalin-2 (LCN2; also known as 24p3) is another IL-17-induced antimicrobial protein that controls bacterial growth by restricting access to dietary iron¹⁰³.

The CCAAT/enhancer-binding protein (C/EBP) family of transcription factors is also activated by IL-17 signal transduction, and is often just as important as NF-κB in the regulation of target genes. For example, the Il6 and Lcn2 promoters require both NF-κB and C/EBP elements, as deletion of either element renders their promoters insensitive to IL-17-mediated signals^{104, 105}. IL-17 induces expression of C/EBPδ mRNA and protein, and to a lesser extent C/EBPβ^{104, 106}. The latter C/EBPδ protein is regulated extensively by posttranscriptional or posttranslational modifications, including alternative translation and phosphorylation¹⁰⁷. Phosphorylation of C/EBPβ depends on a GSK3β- and ERK-dependent pathway, which functions to dampen IL-17 dependent signaling (see below)¹⁰⁸.

On their own, IL-17 (and IL-17F) are modest activators of signaling, but they function cooperatively with other pro-inflammatory molecules, particularly TNFα but also IFNγ, IL-22, lymphotoxin, IL-1β and lipopolysaccharide¹⁰⁹. The molecular basis for this synergy is not completely understood, and probably involves multiple mechanisms. In synovial tissue, IL-17 upregulates TNFRII expression, and thereby enhances responsiveness to TNFα signaling¹¹⁰. For some genes, cooperativity between IL-17 and TNFα occurs at the level of the promoter (for example Il6 and Lcn2) and/or mRNA stability¹¹¹. IL-17 also upregulates expression of IκBζ — a positive regulator of the NF-κB family — which in turn promotes expression of at least some target genes¹¹². The ability of IL-17 to signal cooperatively with other cytokines is probably an important aspect of its biology, since inflammatory environments contain multiple cytokines that can act in concert.

Negative regulation of IL-17 signaling

Dysregulated IL-17 production and signaling is associated with the development of several autoimmune diseases, thus it is not surprising that there are multiple layers of restriction that control IL-17-mediated inflammatory signaling. As outlined below, a variety of inhibitory mechanisms have been identified (Figure 4).

Deubiquitylating enzymes (DUBs) can temper IL-17-mediated signal transduction by targeting TRAFs; ubiquitin C-terminal hydrolase 25 (USP25) removes K63-linked ubiquitin chains on both TRAF6 and TRAF5, which inhibits transcription of IL-17 target genes as well as IL-17-induced mRNA stability USP25 deficiency is associated with enhanced IL-17-dependent pulmonary inflammation and increased susceptibility to EAE¹¹³.

We recently identified A20 (Tnfaip3) as a vital feedback inhibitor of IL-17 signaling¹¹⁴. SNPs in TNFAIP3 are associated with numerous autoimmune and lymphoproliferative diseases, including psoriasis, rheumatoid arthritis (RA) and certain B cell malignancies¹¹⁵. IL-17 induces expression of Tnfaip3 mRNA and A20 protein levels, which triggers deubiquitination of TRAF6 and downregulation of NF-κB and MAPK activation (particularly JNK). Notably, A20 associates directly with IL-17RA through the receptor’s C-terminal “C/EBPβ activation domain” (CBAD)¹¹⁴. The CBAD is also an interaction site for TRAF3, another negative regulator of IL-17 signaling¹¹⁶, and deletion of the CBAD was shown several years ago in a structure-function analysis of IL-17RA to enhance IL-17 signaling¹¹⁷. Intriguingly, A20 appears to inhibit IL-17 in a noncanonical fashion. In other systems, A20 needs to interact with accessory factors, such as TAX1-binding protein 1 (TAX1BP1) and E3 ubiquitin protein ligase ITCH, in order to hydrolyze K63-linked substrates efficiently¹¹⁸ and deletion of TAXBP1 or ITCH results in a phenotype similar to the lack of A20, albeit less severe^{119, 120}. However, knockdown of either TAX1BP1 or ITCH do not inhibit IL-17 signaling, which suggests that A20 might use alternate adaptors. Indeed, we found that A20 associates with anaphase-promoting complex subunit 5 (APC5)^{121, 122}, which is a subunit of a multi-protein E3 ubiquitin ligase complex. Together, these data show that reversal of K63 ubiquitination on TRAF6 via A20 is an important downregulatory strategy for restricting IL-17-mediated inflammation.

K48-linked ubiquitination and proteasomal degradation of adaptor proteins is another means of regulating inflammatory signaling. IL-17R signaling is inhibited by the Ub ligase complex β-TrCP, which induces K48 ubiquitination on phosphorylated forms of ACT1 that in turn leads to its degradation through the proteasome and consequent dampening of IL-17-dependent signaling¹²³. TRAF proteins also play negative roles in regulating inflammatory signaling pathways by IL-17 ¹²⁴. For example, TRAF3 associates with IL-17RA through a consensus TRAF binding site in the CBAD of the receptor. TRAF3 competes with ACT1 for binding to IL-17RA and thereby dampens IL-17-dependent signals¹¹⁶. Similarly, TRAF4 directly binds Act1 upon IL-17 stimulation and competes for its association with TRAF6. Consistently, TRAF3 and TRAF4 limit IL-17-mediated development of EAE^{116, 126}.

Some data implicate microRNAs in regulating both Th17 development and IL-17 signaling. The microRNA MiR-155 expressed in CD4+ T cells positively regulates Th17 differentiation¹²⁷, which may be due to blockade of suppressor of cytokine signalling 1 (SOCS1), a negative regulator of JAK/STAT signaling¹²⁸. Consequently, MiR155−/− mice or mice in which this miRNA is silenced are resistant to EAE^{127, 129}. In terms of signaling, MiR-23b keeps IL-17-induced NF-κB activation in check by targeting the degradation of TAB2/3 and IKKα. Consistently, patients with RA and systemic lupus erythematosus (SLE) show reduced expression of miR-23b ¹³⁰.

As noted above, C/EBPs regulate IL-17-induced gene expression. Interestingly, C/EBPβ is both a positive and negative regulator of the IL-17 pathway. Although C/EBPβ can positively transactivate IL-17 target genes¹⁰⁴, C/EBP phosphorylation of C/EBPβ by the kinases Erk and GSK3β correlates with a decrease in IL-17-induced gene expression. Notably, phosphorylation through GSK3β is dependent on the CBAD, another negative regulatory event mediated through the IL-17RA distal domain ¹⁰⁸.

Th17 cells in steady state

Naturally, Th17 cells did not evolve to cause autoimmunity, but to provide effective host defense against pathogens. Experiments using IL-17- and IL-17R-knockout mouse models have demonstrated increased susceptibility to a wide variety of infectious organisms, including bacteria, parasites, fungi and viruses. However, examination of families or individuals with specific mutations in the Th17 pathway indicated a surprisingly limited role for Th17 cells and IL-17 in immunity to infections, with susceptibility to commensal fungi (mainly the commensal Candida albicans) by far the dominant disease seen in these settings (Summarized in Table 1). Although the reasons for the dichotomy between mice and humans are still unclear, this is similar to findings in MyD88-deficient humans, who show a relative restricted susceptibility to pyogenic bacteria, in contrast to the broadly susceptible MyD88-knockout mice¹³¹.

As noted above, patients with AD-HIES have autosomal dominant-negative mutations in STAT3 that are associated with reduced frequency of Th17 cells and susceptibility to chronic mucocutaneous candidiasis (CMC), Staphylococcus aureus infections and pulmonary infections^{66, 132}. The reduced frequency of Th17 cells in CMC is associated with mutations that impair fungal sensing (CARD9), IL-23 signaling (IL12RB1, IL12B) or enhance negative regulation of Th17 differentiation (STAT1) ^133–135. Similarly, CMC but not other infections occurs patients with APS-1/APECED, who have naturally-occurring neutralizing antibodies against Th17 cytokines^{136, 137}. Null mutations in IL17RA or ACT1 and dominant-negative mutations in IL17F also promote CMC and some staphylococcal infections^{138, 139}. Therefore, deficiencies in Th17 and/or IL-17 are strongly linked to defects in fungal immunity and these association studies support data from mouse studies while revealing the apparent resilience of the human immune system in the face of immunodeficiency¹⁴⁰.

Th17 cells in disease states

While Th17 cells did not evolve to cause disease, they are programed to provide formidable host protective responses—often associated with rapid recruitment and activation of granulocytes and macrophages capable of producing tissue damaging reactive oxygen species (ROS) ¹⁴¹. During prolonged and dysregulated exposure to IL-1 and IL-23, Th17 cells recruit inflammatory myeloid cells to cause severe local tissue injury^{35, 142}. Experiments using Il17-, Il22-, Il23a- or Il23ra-deficient mice as well as antibody-mediated inhibition have demonstrated a requirement for these cytokines in EAE, CAIA, CIA, IBD, AS, and psoriasis disease models^{4, 5, 25, 142, 143}. Based on these preclinical models and human GWAS studies and pharmacogenomic disease association analyses, psoriasis, IBD, and AS have emerged as the leading disease indications for anti-IL-17 and anti-IL-23 treatments^{54–58, 144}.

Psoriasis is an immune-mediated inflammatory disorder characterized by the activation of both adaptive and innate Type-17 responses¹⁴⁵. The pathogenic involvement of IL-23 in psoriasis is supported by the increased expression of IL-23 in psoriatic lesional skin and its ability to induce psoriasiform characteristics in a preclinical model of intradermal IL-23 administration¹⁴³. Injection of IL-23 promoted dermal acanthosis, neutrophil recruitment, and infiltration of IL-17- and IL-22-producing T cells. In addition, inhibition of IL-23 led to significant reduction of IL-17 and IL-22 in psoriatic lesional skin, which supports that these cytokines are key regulatory factors for this disease. IL-17-producing CD4⁺ T cells, CD8⁺ T cells, γδT cells and ILCs can all be found in the psoriatic skin¹⁴⁶. Additional skin resident cells such as keratinocytes, fibroblasts and endothelial cells respond to IL-17 and produce autocrine factors that promote rapid cell division and neo-angiogenesis, which strongly contribute to the disease process. Inappropriate secretion of anti-microbial peptides (such as β-defensins, lipocalin, LL-37, S100A7) and chemokines (CXCL 1,2,3,5, IL-18, CCL20) by keratinocytes further recruit and activate mast cells, neutrophils, and inflammatory macrophages¹⁴⁷. This interaction between Type-17 lymphoid cells and tissue resident keratinocytes/fibroblasts creates amplification loops that drive epidermal hyperplasia and pro-inflammatory conditions—all of which are hallmarks of psoriasis pathogenesis.

The seronegative spondyloarthropathies (SPA) are a heterogeneous group of immunological diseases that follows specific types of viral and bacterial infections and include ankylosing spondylitis, psoriatic arthritis and the ‘reactive’ arthritis¹⁴⁸. In contrast to rheumatoid arthritis where inflammation is accompanied by bony erosion and destruction, spondyloarthropathy is characterized by new bone formation where ligaments attach onto bone—the entheseal organ. We have recently uncovered the presence of RORγt+ IL-23R+CD3+CD4-CD8-Sca1+ entheseal resident cells and that their production of IL-17 and IL-22 are the major cause of STAT3-dependent osteoblast-mediated bone remodeling¹⁴². The new bone appears as inappropriate bony outgrowths from spinal vertebrae, which can fuse and immobilize the spine causing the ultimate morbidity for this disease¹⁴⁹. This condition is commonly associated with uveitis, aortic valve inflammation, psoriasis and inflammatory bowel disease. Indeed, uveitis accounts for significant comorbidity in ankylosing spondylitis, affecting 40% of patients. In addition, up to 70% of patients with ankylosing spondylitis have microscopically evident bowel inflammation¹⁵⁰. These observations suggest that there is a fundamental underlying pathology linking these immune disorders. Recently GWAS have identified multiple SNPs in IL-23R, which are associated with either protection from ankylosing spondylitis, psoriasis, IBD or increased susceptibility to these autoimmune conditions^54–58.

The same IL-23R SNPs found in individuals with ankylosing spondylitis and psoriasis were present in patients with Crohn’s disease along with Nod2 / Card15 (sensor for bacterial cell wall)^{54, 151, 152}. This is consistent with the body’s immune system having an elevated Th17 response attacking the intestinal mucosa, possibly against commensal bacterial antigens. Preclinical animal studies have also identified IL-23 as a key inflammatory mediator that promotes chronic tissue injury responses. In Crohn’s disease, increased expression of IL-17A mRNA and intracellular protein in the intestinal mucosa has been reported¹⁵³. Elevated fecal IL-17A levels were also described in active Crohn’s disease¹⁵⁴. Taken together, these observations implicate the IL-23-IL-17 immune axis in the pathogenesis for Crohn’s disease and the related immune disorders such as psoriasis and spondyloarthropathies and provide a rationale for targeted-antibody therapy.