Introduction: the IL-23/IL-17 axis in spondyloarthritis

In the present issue of Arthritis Research & Therapy, Heiner Appel and colleagues from Berlin present an important histological study on IL-17-producing cells in spondyloarthritis (SpA) [1]. The IL-23/IL-17 axis has become a major focus of SpA research based on three key observations. Firstly, the ankylosing spondylitis subset of SpA is associated with genetic polymorphisms of the IL-23 receptor - the protective R381Q gene variant has recently been demonstrated to impair Th17 responses [2, 3]. Secondly, the unfolded protein response - a cellular stress programme that can be initiated by HLA-B27 misfolding - strongly increases the production of IL-23 [4]. Thirdly, a recent proof-of-concept trial with an anti-IL-17A monoclonal antibody showing good efficacy in active ankylosing spondylitis provided the first direct evidence that IL-17 plays an important role in SpA pathogenesis [5].

Identification of IL-17-producing cells

The IL-23/IL-17 axis was originally described in T-cell biology, where IL-23 is a crucial factor in the so-called Th17 response. Th17 cells have been demonstrated to play a pathogenic role in many experimental autoimmune diseases. In human SpA, several studies enumerating Th17 cells in SpA have yielded partially conflicting results - but a recent report convincingly demonstrated an increased presence and IL-17 production of KIR3DL2+ T cells responsive to HLA-B27 homodimers in blood and synovial fluid [6]. Before postulating that these cells are the major IL-17 producers in SpA, however, one should consider the fact that many other cell types besides canonical Th17 cells express the IL-23 receptor and respond to IL-23 by producing IL-17.

Based on the original demonstration that IL-23 is able to induce IL-17 production in RAG-deficient mice that lack both B cells and T cells [7], there is now also ample evidence in humans that IL-17 is produced by γδ T cells, mast cells, neutrophils, and innate lymphoid cells between others. This concept urged translational researchers to reassess the cellular source of IL-17 in the primary target tissues of a various diseases. Several seminal studies have identified noncanonical IL-17-producing cells in tissue inflammation, including mast cells in rheumatoid synovitis and psoriasis and innate lymphoid cells in inflammatory bowel disease [8–10]. It is in this context that Appel and colleagues made the crucial observation that innate immune cells rather than canonical T cells express IL-17 in axial spondyloarthritis. Their extensive histological analysis of zygoapophyseal joints clearly demonstrates that CD15+ neutrophils and myeloperoxidasepositive myeloid cells, but not classical T cells, are the major cellular sources of IL-17 in the inflamed bone marrow. These data corroborate with emerging evidence that mast cells and neutrophils, but not T cells, are also the major cellular source of IL-17 in peripheral SpA. Taken together, these studies emphasise the relevance of direct analysis of affected tissues and postulate a role for innate IL-17-producing cells in SpA.

Production of IL-17-related cytokines

The observations of Appel and colleagues urge us to revisit the pathobiology of the IL-23/IL-17 axis in SpA. Firstly, we will have to investigate which cytokines and transcription factors are responsible for this innate IL-17 response. Although it is tempting to propose a role for IL-23 and RAR-related orphan receptor C based on the Th17 biology, this remains to be formally demonstrated and would not suffice to explain why SpA displays an innate rather than acquired IL-17 response.

Secondly, it will be crucial to determine which cytokines are exactly produced by these innate IL-17-producing cells. This includes better characterisation of the IL-17 isotypes, as the reagents used in Appel and colleagues' study and most other reports do not allow one to formally discriminate IL-17A from IL-17F. A related cytokine of major interest is IL-22 as its production is also IL-23 dependent and closely related to IL-17 in T cells. IL-22 provides a unique signal directionality from the immune system to the stromal tissue as the IL-22 receptor is exclusively expressed on mesenchymal cells. The effect of IL-22 on enterocytes can be protective in colitis whereas it contributes to keratinocyte hyperplasia in psoriasis models. Considering the prototypical stromal alterations in SpA, it will be important to determine which innate IL-17-expressing cells can also produce IL-22 [8].

Implications for targeted therapies

The previously mentioned issues may have direct relevance for targeted therapies. If innate IL-17-expressing cells produce other IL-23-dependent pathogenic factors, would IL-23 targeting then be more useful than an anti-IL-17A monoclonal antibody in SpA [5]? Or, in contrast, would it be wiser to target selectively IL-17 without impairing IL-22 production as suggested by some colitis models? Could we imagine trying to target a specific pathogenic IL-17-producing cell population - for example, using imatinib to deplete c-kit-positive mast cells and innate lymphoid cells - rather than the cytokines they produce in order to leave the protective IL-17 production untouched [11]? It is mainly in the context of selective therapeutic targeting that the work of Appel and colleagues set a first, but crucial, step in characterisation of the phenotype, developmental and transcriptional requirements, and function of innate IL-17-producing cells in SpA.