Exploring IL-17 in spondyloarthritis for development of novel treatments and biomarkers

Spondyloarthritis (SpA) is an umbrella term describing a family of chronic inflammatory rheumatic diseases. These diseases are characterised by inflammation of the axial skeleton, peripheral joints, and entheseal insertion sites throughout the body which can lead to structural joint damage including formation of axial syndesmophytes and peripheral osteophytes. Genetic evidence, preclinical and clinical studies indicate a clear role of interleukin (IL)- 23 and IL-17 as mediators in SpA pathogenesis. Targeting the IL-23/ (cid:0) 17 pathways seems an efficient strategy for treatment of SpA patients, and despite the remaining challenges the pathway holds great promise for further advances and improved therapeutic opportunities. Much research is focusing on serological markers and imaging strategies to correctly diagnose patients in the early stages of SpA. Biomarkers may facilitate person-alised medicine tailored to each patient ’ s specific disease to optimise treatment efficacy and to monitor thera- peutic response. This narrative review focuses on the IL-17 pathway in SpA-related diseases with emphasis on its role in pathogenesis, current approved IL-17 inhibitors, and the need for biomarkers reflecting core disease pathways for early diagnosis and measurement of disease activity, prognosis, and response to therapy.


Introduction
Spondyloarthritis (SpA) represents a group of related but phenotypically distinct inflammatory autoimmune diseases. SpA includes axial SpA (axSpA, summarizing radiographic (ankylosing spondylitis, AS); ="" axialSpA) and non-radiographic disease, peripheral SpA (pSpA) and psoriatic arthritis (PsA) [1,2]. These subforms of SpA have overlapping clinical features that reflect shared genetic risk factors and pathophysiology. Genetics plays a role in the aetiology of SpA sharing a close association with the major histocompatibility complex (MHC) class I surface antigen HLA-B27 (human leukocyte antigen B27), which is believed to be the major genetic susceptibility factor [3].
AxSpA involves inflammation, bone and cartilage loss which is followed by new local bone formation that can lead to progressive ankylosis including syndesmophytes formation in the spine. The inflammatory processes and new bone formation take place in the entheses, axial skeleton, and peripheral joints through mechanisms that are poorly understood [2]. Extra-articular manifestations can appear with anterior uveitis, as well as co-occurrence of psoriasis (PsO) and inflammatory bowel disease (IBD) [4]. As mentioned, axSpA consists of two diagnostic entities, non-radiographic axSpA (nr-axSpA) and AS [5]. The latter is the prototype of axSpA and is characterised by definite radiographic damage in the sacroiliac joints and/or in the spine as defined by the Assessment of SpondyloArthritis international Society (ASAS) classification criteria [5,6]. On x-ray, patients with nr-axSpA are visualised without characteristic damage and may represent an early form of AS. However, some patients may never develop structural damage on radiographs However, some patients may never develop structural damage on radiographs [7]. The pSpA category covers inflammation predominantly in peripheral joints rather than axial, commonly affecting the lower extremities asymmetrically. pSpA also includes enthesitis, anterior uveitis and/or dactylitis. Patients diagnosed with reactive arthritis and IBD-associated arthritis often fit into the pSpA category [6] (Fig. 1). In addition, while PsA is usually considered a separate disease entity, PsA often fulfils the criteria for pSpA.
AxSpA affects 0.5-1.5% of the western population [8], and it has been suggested that the entire spectrum of SpA may be more prevalent than rheumatoid arthritis (RA) in the United States [9]. Although TNF inhibitor therapy improves symptoms, functional mobility and reduces disease activity [10][11][12][13][14], almost 40% of axSpA patients do not sufficiently respond to anti-TNF treatment [15,16]. Therapies directed against TNF may be successful in suppressing the inflammation, however, axSpA involves both inflammation, bone loss and new bone formation.
Bone erosion seen in AS is a frequent complication but is often of limited severity. Eventually, the principal cause of structural damage is the aberrant new bone formation which can lead to fusion of the sacroiliac joints and zygapophyseal joints, including outgrowth of syndesmophytes in the spine bridging the vertebral bodies resulting in ankylosis [17] (Fig. 2). Until now, no treatment has shown improvement in long-term outcomes in axSpA, particularly loss of movement caused by bone ankylosis [18]. Hence, treatment that only suppresses inflammation may not resolve bone loss and aberrant bone formation.
In the past decade, strong evidence from genetic, animal, translational and clinical studies has confirmed the role of the IL-23/− 17 axis in the pathogenesis of SpA. Clinical evidence suggests that TNF blockade in AS patients does not influence the IL-23/− 17 axis [19]. Instead, therapeutics have been directed towards the IL-23 and IL-17 which are considered as key disease-modifying targets in the treatment of SpA. This has recently contributed to development of a range of novel therapeutics targeting IL-17 and IL-23 pathway in SpA, PsA and PsO [20]. IL-17 inhibitors have shown beneficial effects on symptoms by significantly improving psoriatic skin lesions, reducing pain and inflammatory joint lesions [21]. Furthermore, recent results have showed that IL-17  Healthy spine compared to spine with ankylosing spondylitis. Severe inflammation can lead to fusion of vertebrae bodies which can result in a less flexible spine with loss of normal curvature seen in a healthy S-curved spine. Additionally, ankylosing spondylitis is characterised by inflammation in the sacroiliac joints. Figure has been modified from [23].
inhibition arrests the progression of catabolic and anabolic bone changes in the joints of PsA patients and may potentially slow disease progression [22]. Yet, comprehensive evidence is still needed to determine whether these therapeutics control the SpA disease and succeed as effective disease-modifying anti-rheumatic drugs (DMARDs).
In recent years, studies have focused on clinically applicable biological markers to overcome the challenges in the management of SpA related to unclear pathogenesis, lack of early diagnosis and accurate monitoring of disease activity and inability to predict joint damage or response to treatment. Applying biomarkers to address the many challenging aspects of SpA may enhance the development of novel treatments which may offer more specialised and effective treatment of patients [24].
This narrative review will explore the role of IL-17 in homeostasis and its heterogeneous effects in inflammation and structural damage to justify for the potential advantageous effect of IL-17 inhibition in AS and PsA. Furthermore, the review will assess the current challenges in the management of SpA and summarise promising biomarkers of inflammation and tissue remodelling including parameters of angiogenesis and autoantibodies that may improve the diagnosis, prognosis and treatment outcomes in SpA.

IL-17 signalling and production
The IL-17 superfamily consists of six structurally related proinflammatory cytokines IL-17A, IL-17B, IL-17C, IL-17D, IL-17E and IL-17F, which are involved in the immune-mediated antimicrobial response. Excessive production of IL-17 during chronic inflammation has been associated with development of inflammatory diseases [20]. IL-17A is the best characterised member of the IL-17 family and generally referred to as IL-17. Of the six cytokine family members, IL-17F is the closet related member to IL-17A with 55% sequence identity [25] and often co-produced with IL-17A by Th17 cells [26]. Both IL-17A and IL-17F exist as disulphide-linked homodimers or as IL-17A/-17F heterodimers. All forms of IL-17A and IL-17F cytokines signal through a heterodimeric IL-17RA and IL-17RC receptor complex. The cytoplasmic tail of IL-17RA possesses discrete signalling motifs, a conserved SEFIR domain (similar expression of fibroblast growth factor and IL-17Rs) with an additional 'SEFIR-extension' domain (SEFEX) which are both required for functional downstream IL-17 signalling [27]. Upon IL-17A binding to the IL-17R complex, the IL-17R undergoes a conformational change enabling interactions between the signalling adaptor protein Act1 and SEFIR domains mediating downstream events [28]. The recruitment of Act1 represents a hallmark of IL-17 signalling and is an essential signalling component of the IL-17R signalling event, not seen in any other known class of receptors [29]. Act1 contains a TNF receptor associated factor (TRAF)-binding site which enables association with different TRAF family proteins. Hereof, Act1 operates as a docking station for the different TRAF proteins being able to activate several independent signalling pathways. The engagement with TRAF6 drives the activation of the canonical pathway (de novo gene transcription) which includes the nuclear factor kappa light chain enhancer of activated B cells (NFKB) pathways activated by the transforming growth factor β-activated kinase 1 (TAK1), mitogen-activated protein kinase activator protein-1 (MAPK/AP1) pathways and C/EBP pathways [30].
enhanced IL-17 signalling [31,33]. Besides activating the transcription of inflammatory genes, the binding of TRAF2 and TRAF5 to Act1 activates the noncanonical pathway, which mediates post-transcriptional mRNA stabilisation, particularly CXCL1 mRNA encoding chemokines [34]. The post-transcriptional stabilisation is controlled through multiple RNA-binding proteins, including the human antigen R (HuR) and ATrich interactive domain-containing protein 5a (Arid5a) [35,36]. The stabilised transcripts are deposited in cytoplasmic granules, where they can be translated or degraded instantly [27]. The IL-17-mediated signalling not only activates proinflammatory cascades but also promotes numerous regulatory pathways. Upon IL-17A stimulation, TRAF4 which acts as a negative modulator of IL-17-mediated signalling, can be recruited and compete with TRAF6 for Act-1 binding [36,37]. Additionally, TRAF3 also acts as a negative regulator by similar competing action on the IL-17 signalling [36,38] (Fig. 3). The IL-17 signal transduction overall consists of many non-redundant mechanisms to finetune IL-17-induced inflammation, however, these mechanisms are not fully understood [27].
The cellular production of IL-17 is complex and heterogeneous. Both innate cell populations and different adaptive T cell subsets can produce IL-17. Notably, naïve CD4+ T cells differentiated into Th17 cells are the main producing cells of IL-17. However, during inflammation most IL-17 is secreted by innate immune cells including innate-acting lymphocytes such as γδ T cells, invariant natural killer T (iNKT) cells [69], lymphoid tissue inducer (LTi) cells [70], type 3 innate lymphoid cells (ILC3s) [71], macrophages [72], neutrophils [73] and Paneth cells [74]. Additionally, IL-17 can be secreted by Tc17 cells subsets of CD8+ T cells [75,76] and resident memory T cells [77]. Several animal and human studies have demonstrated that these cells produce IL-17 in response to inflammation, but their respective contributions to the SpA pathogenesis remain unexplored [30].
Although the pathogenic mechanisms underlying SpA are not fully elucidated, multiple areas of investigation have recently converged to provide strong evidence implicating the IL-23/− 17 axis in rheumatic diseases including AS and PsA. The following section will describe several lines of evidence from genetics, preclinical and clinical studies emphasising a pivotal role of the IL-23/− 17 axis in SpA.

Genetic evidence
IL-23 is responsible for promoting the expansion and survival of Th17 subsets including the production of IL-17 from Th17 cells [78]. IL-23R gene polymorphisms with functional relevance in T cell immune response have been strongly associated with the development of AS [79], PsO [80], and IBD [81]. Additionally, a genome-wide association study (GWAS) in AS has identified several variants in genes including CARD9, IL-12B, PTGER4, IL-6R, IL-27, TYK2 and STAT3, which are implicated directly or indirectly in the IL-23/− 17 axis [82,83].

Evidence from preclinical studies
The importance of IL-17 in the pathogenesis of SpA has further been defined in several animal studies. In 2012, Sherlock et al. systemically overexpressed IL-23 in transgenic mice and identified a unique population of CD3+/CD4-/CD8-T cells in the enthesis which expressed the Th17 transcription factor RAR-related orphan receptor-γt (ROR-γt), and the IL-23 receptor. Upon activation with IL-23, these T cells produced IL-17A, IL-17F, IL-6, and IL-22 resulting in enhanced axial and peripheral enthesis and sacroiliitis followed by IL-22 dependent osteoproliferation in the absence of synovial inflammation. In addition, blockade of IL-17 and IL-22 with neutralising antibodies significantly diminished the disease severity, especially when the two antibodies were administered together [84]. The existence of this unique T cell population may explain the generation of the SpA phenotype in the early phases of the disease, without activating CD8+ T cell mediated autoreactivity or other IL-17 secreting cells [18]. IL-17-deficient mice have shown to exhibit less inflammation in a type II collagen-induced arthritis (CIA) model [85] and the IL-17R-deficient mice also demonstrated reduced secretion of proinflammatory cytokines into the synovium preventing cartilage degradation [86].

Evidence from clinical studies
IL-17 and IL-23 levels have been measured in serum of AS patients and healthy controls by enzyme linked immunosorbent assay (ELISA). Results demonstrated significantly elevated IL-17 and IL-23 levels in AS patients compared to controls [87][88][89]. Furthermore, flow cytometry measurements showed increased numbers of circulating Th17 cells found in the peripheral blood (PB) of PsA and AS compared to RA patients, which showed no increase of circulating Th17 cells in PB [90,91]. In addition, Th17 levels in synovial fluid samples from PsA patients were analysed by flow cytometry. In PsA patients, levels of Th17 cells were significantly elevated compared to RA patients and correlated with disease activity measures and progression of radiographic joint erosion [91]. Increased mRNA expressions of IL-17 and IL-22 were found in skin lesions from PsO patients including elevated levels of Th17 cells in PB measured by fluorescence-activated cell sorting (FACS) [92]. In mast cells, IL-17 levels measured by ELISA were upregulated in the synovium of peripheral joints in patients with SpA [93]. By immunohistochemical analysis, enhanced expression levels of IL-17 were found in CD15+ neutrophils located in the subchondral bone marrow of inflamed spine from SpA patients with low abundance of Th17 cells indicating that IL-17 producing cells other than the Th17 cells are relevant in local inflammation in SpA [94]. IL-23, but not IL-17, has been found overexpressed in inflamed ileum in AS patients analysed by RT-PCR [95]. IL-23 positive cells within the subchondral bone marrow and fibrous tissue in facet joints of AS patients were significantly elevated compared to osteoarthritis (OA) patients measured by immunohistochemistry [96]. Altogether, these results are strong evidence of both the innate and adaptive mechanisms including the IL-23/− 17 axis may drive the pathogenesis of SpA.
In 2011, Bowness et al. found elevated numbers of CD4+ Th17 cells expressing the receptor KIR3DL2 in PB and synovial fluid from AS and SpA patients. KIR3DL2 responds to the HLA-B27 homodimer molecule by promoting survival of Th17 cells and upregulation of IL-17 secretion in AS patients [97]. Nevertheless, levels of CD4+ Th17 cells were not found elevated in other studies [94,98] including the expression of KIR3DL2 [99], whereas IL-23R+ γδ T cells were observed instead [100]. γδ T cells have shown to be involved in large productions of IL-17 in a murine model accumulating enthesitis with enhanced bone regeneration [101,102]. Analysis of the PB and synovial fluid in SpA has been considered inadequate to uncover whether pathologic processes were present or not in the sacroiliac joints, vertebral bodies, zygapophyseal joints including bone marrow next to the axial inflammatory lesions. For instance, TNF-α serum levels were not consistently elevated in SpA patients and even found to be lower than in healthy controls [18]. Thus, this viability of the T cell subtypes seems to be an important factor compared to TNF-α levels and may be an essential therapeutic target.
The important role of IL-23/− 17 in SpA has not only been evidenced as mentioned above, but the significant clinical efficacy of IL-17 inhibitors in treating AS, PsA and PsO has also been acknowledged and will be discussed in the next section.

Targeting the IL-23/¡17 axis in SpA
As evidence from genetic, animal and human studies support the importance of IL-23 and IL-17 in the pathogenesis of SpA, therapeutics inhibiting IL-23 would be expected to show similar clinical efficacy as drugs inhibiting IL-17 in SpA. Interestingly, clinical studies with IL-12/ IL-23 inhibitor, ustekinumab, showed lack of efficacy in axSpA, while effective in PsA and PsO [103]. Furthermore, the efficacy of risankizumab, an IL-23 p19 subunit inhibitor, was evaluated in patients with active AS in a phase II study. Risankizumab did not reduce the signs and symptoms of AS patients compared to placebo, questioning whether IL-23 is a relevant driver of AS pathogenesis and symptoms [104]. Conversely, IL-17A inhibition has demonstrated efficacy in AS, PsA, and PsO in clinical trials. A fully human anti-IL-17A monoclonal antibody, secukinumab, has been approved for treatment of AS, PsA and PsO based on several large randomized controlled trials [105][106][107][108][109][110]. Another IL-17A inhibitor, ixekizumab, a humanized anti-IL-17A antibody has been approved for treatment of PsA and PsO and has shown significant efficacy in AS in two large phase III trials [111][112][113][114][115]. These results suggest that IL-17A is a key cytokine facilitating disease pathogenesis in AS, PsA and PsO. In AS, IL-17 seems to be produced in a largely IL-23independent way. Furthermore, in human interspinous entheseal tissue IL-23 receptor positive and negative subpopulations of γδ T cells were found signifying the role of IL-23-independent IL-17 production [116]. A broader understanding of the divergent roles of IL-23 and IL-17 in the pathophysiology of axSpA is still needed. Evidence towards anatomical and immunological differences between axial and peripheral disease may be one of the reasons for the IL-23/− 17 axis pathway divergence. For instance, more peripheral enthesis is observed in PsA with inflammation in the entheseal soft tissue [117] while in AS, inflammation is observed in the peri-entheseal bone in the spine and more severe axial bone edema has been observed with much higher frequency in HLA-B27 positive patients [118].

Implication of IL-17 in bone damage
Currently, the focus has primarily been on the mechanisms that trigger chronic inflammation in joints and entheses. However, as a result of chronic inflammation, structural damage evolves which is a critical contributor to morbidity in patients suffering from AS and PsA [119].

Bone homeostasis and abnormal turnover
The bone remodelling cycle is a tightly regulated balance between osteoclasts, which are responsible for bone resorption involving removal of mineralised bone matrix, and osteoblasts, which form and replace the resorbed bone [120]. To maintain bone homeostasis, the resorption and formation processes are tightly coupled ensuring that the amount of bone removed equals the amount of bone formed. Within the bone matrix, bone growth factors, such as transforming growth factor-β (TGFβ) and bone morphogenetic proteins (BMPs) are stored. These growth factors are released during the resorptive phase promoting bone formation to maintain bone homeostasis [121].
Abnormal bone turnover is a feature shared by several rheumatic diseases including AS and PsA, and these bone alterations are displayed in different forms [17,120]. Bone turnover in AS and PsA is distinguished from RA by occurrence of prominent focal areas of new bone formation involving endochondral ossification. This process is normally active during the period of skeleton growth and development and involves the replacement of the cartilaginous matrix with mature mineralised bone. However, during bone healing after e.g. fracture or pathological bone growth associated with the ankylosing process, endochondral bone formation can be reactivated through a series of cell differentiation steps ultimately leading to new postnatal bone formation. Although the mechanism behind the reactivation of endochondral bone formation is incompletely understood, it is believed that the mesenchymal progenitor cells commit to chondrogenic differentiation ending up differentiating into hypertrophic chondrocytes which begin to calcify the cartilage tissue matrix. Consequently, through direct bone formation osteoblast precursor cells invade and progressively degrade the matrix and replace it with bone [122]. In addition, hypertrophic chondrocytes have been shown to transdifferentiate into osteoblasts in endochondral bone during development [123]. This indicates that endochondral and direct bone formation are both involved in the progressive ankylosis with the outlay of a cartilage template as a key element in the development of syndesmophytes [122]. This aberrant new bone arises at specific localised anatomical sites and may be the effect of prolonged entheseal inflammation [124]. Both IL-23 and IL-17 are involved in the pathway that expands entheseal inflammation and are closely linked to the new bone formation by promoting catabolic and anabolic bone alterations [124].

The effects of IL-17 on bone
Bone loss seen in PsA, and in some AS patients, is believed to be related to activation of the IL-23/− 17 axis by stimulating differentiation of osteoclasts. Kotake et al. have demonstrated that IL-17 influences bone homeostasis by showing that IL-17 promoted osteoclastogenesis in an osteoblast-osteoclast co-culture model [125]. IL-17 can induce the expression of receptor activator of NFKB ligand (RANKL) on osteoblasts which stimulates RANK signalling on osteoclastic precursors vital for the production of mature osteoclasts [126][127][128]. Several studies have demonstrated pro-osteoclastogenic effect of IL-17 [129,130] including Th17 cells being the main T cell subset to stimulate osteoclast ifferentiation by releasing IL-17 and increasing expression of RANKL on osteoblasts [131]. During chronic inflammation, IL-17 has been shown to activate MMPs and promote matrix turnover leading to cartilage loss in joint tissue [132]. Chabaud et al. have demonstrated that adding IL-17 to human bone samples increased the production of bonedestructive cytokines and bone resorption in vitro [133]. In contrast, the inhibition of IL-17 showed an anti-inflammatory effect and reduced bone destruction and an even more efficient anti-inflammatory effect was obtained when IL-17 was inhibited in combination with TNF-α and IL-6 [67]. These studies suggest that IL-17 activity affects the bone homeostasis by inducing osteoclastogenesis which partially explains the development of destructive bone lesions and loss seen in PsA and AS.
In addition to the effects of IL-17 on bone loss, IL-17 also critically orchestrates the activation and differentiation of osteoblasts. Studies have shown that IL-17 enhances osteogenic differentiation of mesenchymal stem cells (MSCs) [134][135][136]. Furthermore, Osta et al. have demonstrated that IL-17 combined with TNF-α can stimulate osteogenic differentiation of human MScs [137]. Although these findings suggest that IL-17 is involved in the osteogenesis, the role of IL-17 on osteogenic MSCs is still not clear.
Besides the effects of IL-17 on bone remodelling, the development of enthesitis and entheseal new bone formation have been shown to depend on IL-22 [138]. Entheseal resident T cells producing IL-22 have been shown to activate osteoblast-mediated bone remodelling [84].
Furthermore, in vitro IL-22 has been shown to enhance MSC osteogenesis and increase proliferation and migration of human MSCs in inflammatory environments [139]. Hereof, IL-22 may play an important role of the IL-23/− 17 axis in bone remodelling providing a novel pathway for studying the pathological and inflammatory osteogenesis in SpA [30].

Aberrant bone formation in AS
AS is mainly characterised by ankylosis of the sacroiliac joints and vertebral bodies which involves excessive bone formation encompassing both endochondral and direct bone formation [17]. Inflamed sacroiliac joint tissue from patients with AS has been examined by Braun et al. who observed localised nodules of endochondral ossification within the inflamed tissue [140]. TGF-β and BMPs have been proposed as potential key players of the enhanced bone formation in AS. The endochondral bone formation during development and postnatal fracture repair are induced by TGF-β and BMPs and have been implicated in this process [141,142]. TGF-β and BMPs signal through the wingless (Wnt) pathway which is a main regulator of skeletal patterning and post-natal bone remodelling [143]. The activation of the Wnt/ β-catenin pathway results in transcriptional activation of the osteoprotegerin (OPG) gene, which is a potent inhibitor of RANK ligand, the principal regulator of osteoclast differentiation and activation [144].
Lories et al. have elaborated on the concept of a potential relation between biomechanical factors and the ankylosing process [122]. The skeleton consists of mechanosensitive tissue which can adapt biomechanical loading controlled by the bone remodelling cycle. In axSpA patients, osteitis in vertebral bodies and sacroiliac joints influences the normal bone homeostasis by cytokine-mediated osteoclast activation increasing net bone loss. Lories et al. suggested that increased bone net loss leads to instability of the axial skeleton and that the loss of stability triggers tissue remodelling which includes syndesmophyte formation [122]. Studies have suggested that the bone may not be able to compensate for the bone loss in a normal physiological way due to osteoblast activation being suppressed by proinflammatory cytokines or bone formation inhibited by the Wnt signalling antagonist Dickkopf-1 (Dkk1) driven by TNF-α [145].
In summary, bone remodelling seen in AS seems much more complex compared to primary erosive forms of arthritis such as rheumatoid arthritis (RA). Although synovial inflammation in the peripheral joints of patients with AS exhibits histopathologic characteristics similar to those seen in the RA joints, differences have been observed in the excess bone formation and repair pattern [120]. IL-17 has showed to promote excessive osteogenesis at sites of entheses under conditions of mechanical stress, resulting in formation of new bone and syndesmophytes, the key difference between RA and AS [146]. Determining the heterogeneous effects of IL-17 in new bone formation remains a central avenue of future research.

Bone erosion and formation in PsA
Whereas AS is mainly characterised by extensive aberrant bone formation primarily in the axial joints, PsA manifests in the peripheral joints with both increased bone destruction and increased bone formation [147] and differs from RA by its characteristic calcification of entheses [148]. Despite main involvement of peripheral joints, certain PsA patients may have axial involvement with inflammatory back pain and sacroiliitis. In PsA, the bone remodelling homeostasis is disrupted by several cytokines and growth factors altering the differentiation and function of osteoclasts and osteoblasts [149]. These cytokines include both IL-17, IL-6, TNF-α, RANKL, and the growth factors BMPs and TGFβ, similar to AS [150].
Prolonged effect of inflammatory cytokines such as IL-17 has been linked to increased fracture risk in PsA patients [151]. As previously mentioned, IL-17 induces osteoclast differentiation and facilitates bone erosion [126][127][128]131,152] and these observations provide basis for focusing on IL-17 inhibition in PsA to protect from bone erosion. Consistent with these results, clinical studies have showed that PsA patients treated with IL-17A inhibitors, exhibit significantly reduced inflammation and structural damage versus placebo in the short-and long-term [106,153,154]. Recently, the PSARTROS study demonstrated no progression of catabolic or anabolic alterations in the joints of PsA patients treated with secukinumab for 24 weeks [22].
Comparing RA and PsA, similar cytokine expressions have been observed suggesting related mechanisms involved in joint inflammation and destruction [155,156]. However, significant differences between RA and PsA are seen in the bone resorptive phenotype. In RA, the In combination with other proinflammatory cytokines such as TNF-α, IL-1, and IL-6, IL-17 can also directly promote osteoclastogenesis leading to bone erosion. Moreover, the pathological bone growth includes both direct bone formation and reactivation of the endochondral bone formation. IL-17 and IL-22 promote differentiation of bone marrow-derived mesenchymal stem cells (MSCs) which further differentiate either into hypertrophic chondrocytes that form calcified cartilage tissue matrix, or osteoblasts promoting osteoproliferation. Eventually, the calcified cartilage will be progressively degraded and replaced with bone matrix by osteoblasts. This exaggerated new bone formation can result in ankylosis. RANK, receptor activator of NFKB; RANKL, receptor activator of NFKB ligand. Adapted from [18,122,124].
phalangeal joints are most frequently involved in a symmetric manner with U-shaped lesions involving bone destruction and insufficient repair leading to local and systemic bone loss [157,158]. In contrast, PsA exhibits a tubular or omega-like asymmetric bone erosion pattern less severe in size and depth [159]. This suggests that the bone repair mechanism and pathogenesis may be different between PsA and RA.
New bone formation has been observed in PsA patients with entheseophytes and ankylosis mainly manifested in the peripheral joints. However, bone formation in the axial skeleton has also been observed in PsA thus, axial disease can occur both in AS and PsA patients. Likewise, PsO can occur in AS patients which has initiated a debate on whether axial PsA is essentially AS with PsO, or if these conditions are two separate diseases with overlapping features [160]. Like in AS, the mechanisms behind the new bone formation are currently unknown. Hereof, the aberrant bone formation seems to be like AS, but they differ somewhat in their pathology [161].
According to radiographic differences between AS and PsA, the syndesmophytes are symmetrically distributed with a primary marginal shape in AS, while the distribution in PsA is rather asymmetric along the spine with equal amounts of marginal and paramarginal syndesmophytes. From radiographs, the axial severity appears to be worse in AS than in PsA by the number of syndesmophytes and the Bath AS Radiology Index (BASRI) [162]. Like in AS, anabolic growth factors are considered to play one of the key roles in the bone formation in PsA by mediating the formation of hypertrophic chondrocytes and osteoblasts.
Yet, the connection between inflammation and new bone formation is not clearly understood. In mouse models of PsA, BMP signalling in DBA/ 1 mice showed to trigger spontaneously PsA-like arthritis indicating that BMP signalling may be related to the abnormal development of bone in PsA [141,163].
Unlike RA, decreased levels of Dkk1 have been reported in PsA serum compared to healthy subjects which agrees with the increased bone formation and indicates that Wnt/β-catenin signalling plays an important role in directing osteoblast differentiation and function in PsA [164]. Interestingly, in serum samples from AS patients high levels of Dkk1 showed to protect from developing syndesmophytes [165]. Other studies have reported additional high levels of Dkk1 in PsA patients highlighting the complexity of PsA as a disease involving the combination of extensive bone destruction and aberrant bone formation [166,167] (Fig. 4).

IL-17 therapy options for management of SpA
In early 2015, secukinumab was approved in the US and the EU as the first IL-17A inhibiting drug developed for treatment of patients with plaque PsO [168] and in 2016 secukinumab was approved as treatment of AS and PsA patients [169]. Subsequently, another IL-17 inhibitor, ixekizumab, was approved for the treatment of plaque PsO and PsA and is yet under investigation for treatment of AS [170]. In 2017, the latest IL-17 receptor inhibitor class medication, brodalumab, was approved for  [177]. treatment of patients with plaque PsO. Brodalumab, however, has been associated with increased risk of suicidal behaviour [171].
Several novel drug candidates of anti-IL-17 are currently investigated in the development phase. Netakimab, tackling IL-17A is under investigation for treatment of plaque PsO and is now in phase III clinical trials [172]. Bimekizumab, tackling both IL-17 A and IL-17F has shown promising results in phase II trial in AS, PsA and PsO and is currently undergoing phase III trials [173]. Another bispecific agent, afasevikumab, which neutralises IL-17A and IL-17F, is in early phases of clinical trials [174]. One more novel approach is the development of the nanoantibody, ALX-0761, which is undergoing phase II clinical trials for treatment of PsO [175] (Fig. 5). For future treatment of SpA, bispecific biologic therapies which inhibit both TNF-α and IL-17 are of great interest. Dual inhibition of these cytokines in a type II collagen-induced arthritis model has been shown to improve protection against joint damage compared to TNF-α inhibition alone [68,176]. Thus, the combination of blocking TNF-α and IL-17 may provide additional therapeutic value for treatment of SpA. Dual IL-17/TNF-α inhibitors, such as remtolumab and COVA322, are in phase II clinical trials [173].

Biomarkers in drug discovery and development
Over the past decade it has become clear that biomarkers constitute a high priority for the drug discovery process and understanding of the pathogenesis of SpA. Although the ASAS classification and response criteria are useful clinical tools, they are built on patient-reported outcomes including questions about pain and stiffness which is purely subjective and can have high levels of individual variability and bias. Hereof, the usage of more objective measures should be considered in combination with the ASAS criteria [170].
The rheumatoid factor (RF), anti-citrullinated protein antibodies (ACPA), erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) levels are well-established serological biomarkers to evaluate disease activity in RA patients [178]. However, SpA patient are seronegative by the lack of induced levels of serum autoantibodies including RF and ACPA, and ESR and CRP only are used as serum biomarkers, directly reflecting inflammation. CRP is a relative insensitive marker in SpA diseases and correlates poorly with disease activity and is not present in all AS patients thus, may not be useful [179,180]. The onset of disease in SpA patients typically manifests between the ages of 20-30 years with slow progression and is associated with vague symptoms [181]. Altogether, with the lack of confirmatory blood tests and an average delay in diagnosis of AS of 5-10 years, there is a growing interest in the discovery of new biomarkers [181]. Earlier diagnosis and initiation of appropriate therapy would improve the success rates of clinical remission and may prevent the development of structural lesions including ankylosis.
In recent years, several new and expensive therapeutic options such as TNF and IL-17 inhibitors have been approved for treatment of PsA and AS. These agents have shown to improve symptoms, quality of life, and slow down radiographic progression. However, new data on the genetic profiles of the different types of arthritis have appeared with clear evidence of little or no genetic overlap between the different types of arthritis, suggesting that therapies may not be efficient on all these forms [182]. For example, despite the positive effect of IL-17A inhibition in AS and PsA, clinically secukinumab showed to be ineffective in SpA patients with IBD and even exacerbated the symptoms in some patients [183]. Due to the improved understanding of AS and PsA together with the new therapeutic options, more accurate analysis tools are needed to predict early diagnosis, disease activity, disease progression and response to therapy significantly increasing the need for new biological markers. Several biomarkers have been proposed for SpA but they are still in the research phase and not yet widely utilised in clinical practice. This section reviews well-known and new promising biological markers.

CRP
CRP is an acute-phase protein which reflects systemic inflammation. Although CRP is not elevated in a substantial proportion of patients with active SpA, it is commonly used as a reliable acute inflammatory marker of disease activity. Levels of CRP are frequently higher in patients with full-blown radiographic forms of AS compared to nr-axSpA. CRP correlates with Ankylosing Spondylitis Disease Activity Score (ASDAS) and moderately correlates with MRI inflammation which is the preferred method to assess disease activity in SpA, though very costly [184]. In a study by Poddubnyy et al. elevated levels of CRP were observed as a solid predictor of radiographic sacroiliitis progression in AS and nr-axSpA patients [185]. Subsequently, in 2016 Poddubnyy et al. investigated the link between several disease activity parameters (CRP, ASDAS and BASDAI) and radiographic spinal progression in early AS patients. Here, five time points of the mentioned parameters were available during 2 years of follow up. The study confirmed an association between high disease activity and development of structural damage in the spine in patients with early AS [186]. In several studies, SpA patients have been treated with TNF inhibitors which has been shown to decrease the elevated CRP levels, and thus, CRP has been demonstrated as a reliable biomarker for therapeutic response [187][188][189]. In a study by Pedersen et al. the development of new syndesmophytes resulted in a more significant decrease in CRP levels while chronic systemic inflammation was associated with absence of radiographic progression throughout a brief period [188]. These findings support the hypothesis that resolved inflammation results in an increased rate of ossification of the repaired tissue whereas persistent inflammation prevents the formation of syndesmophytes [184].

Anti-CD74
CD74, also known as human lymphocyte antigen (HLA) class II gamma chain or invariant chain, is involved in the assembly and transport of human histocompatibility complex class II (MHC II) molecules [190]. It is believed to mediate B cell proliferation and bind the macrophage migration inhibitory factor (MIF) leading to NFKB activation and thus, expression of pro-inflammatory mediators [191]. ELISA studies performed by Baerlecken et al. 2013 observed antibody reactivity against full-length human CD74 in 56% of 41 patients sera with AS and 5% of 100 healthy subjects. Furthermore, antibody reactivity against CLIP domain of the CD74 protein was observed in 67% of the 216 SpA patients with axial (n = 156) and peripheral (n = 60) SpA. Antibody reactivity was also observed in 45% of 40 PsA patients, 11% of the 80 RA patients, 15% of the 40 patients with systemic lupus, and 0.8% of 125 healthy blood donors. In the 67% of SpA patients, 97% experienced inflammatory back pain of <1 year and hereafter, the reactivity declined by 65% after >20 years after onset and no association with disease activity was observed indicating a potential biomarker for early diagnosis of SpA [192]. In an additional European study, similar results were observed with reactivity in 85.1% of 94 axSpA patients and in 7.8% of 51 controls [193]. However in another study, anti-CD74 IgG and IgA autoantibodies have no diagnostic value in early axSpA demonstrating reactivity with IgG antibodies in 46.4% of 274 axSpA patients and 47.9% of 319 non-SpA chronic back pain controls [194]. This points towards two opposite conclusions in the role of anti-CD74 autoantibody in the diagnosis of SpA. Furthermore, anti-CD74 was measured by ELISA in sera from 71 Chinese SpA patients and 70 controls. The results demonstrated reactivity of 14.1% in SpA patients and 2.9% in healthy controls. Although these results showed higher reactivity in SpA compared to controls, the prevalence of anti-CD74 was lower than in the European studies. This suggests that CD74 may not be a useful diagnostic biomarkers in an Asian population [195]. Based on these results, further investigation is needed to test the efficiency of CD74 as a good target biomarker molecule for axSpA diagnostics.

PPM1A
Protein phosphatase magnesium-dependent 1A (PPM1A) is a serine/ threonine protein phosphatase contributing to the formation of bone including syndesmophytes which is a key feature of AS. PPM1A participates in the bone formation by regulating the BMP and Wnt signalling pathways, including suppressing TGF-β signalling [196,197]. By using high-density protein microarrays, antibody reactivity to PPM1A was tested in sera from patients with PsA (n = 34), PsO (n = 6), and AS (n = 16), RA (n = 21), juvenile idiopathic arthritis (n = 15), and pulmonary artery hypertension (n = 23). Here, higher levels of antibodies against PPM1A were detected in AS patients compared to the other autoimmune diseases. Moreover, in AS patients with grade 3 and 4 radiographic sacroiliitis, higher levels of anti-PPM1A autoantibody were observed compared to those with grade 2 radiographic sacroiliitis. Furthermore, the AS patients were treated with TNF inhibitor which resulted in a significant decrease in the serum levels of anti-PPM1A, which also showed to correlate positively with degree of sacroiliitis in AS by the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) score. PPM1A was also found to be highly expressed in synovial tissue taken from AS patients and driving osteoblast differentiation, which suggests that PPM1A may participate in the pathogenesis of AS. In Korean patients, significantly higher levels of anti-PPM1A were also observed in AS patients (n = 45) compared to RA patients (n = 20), and healthy controls (n = 30). In rats, transgenic for HLA-B27 and human β2microglobulin, anti-PPM1A levels were also increased in sera [198]. Despite these remarkable observations and the potential association with SpA pathogenesis, more research is needed to examine PPM1A as a diagnostic biomarker.

miRNA
Diagnostic biomarker candidates have been considered from gene expression of transcriptomic markers. A meta-analysis performed on gene expression arrays discovered 905 cases of differentially expressed genes in axSpA patients compared to healthy subjects. Here, the most significantly expressed genes represented pathways linked to antigen processing and presentation [199]. In another report, RNA sequence analysis revealed upregulated expression of microRNA (miR)-146a-5p, miR-125a-5p, miR-151a-3p and miR-22-3p, and downregulated expression of miR-150-5p and miR-451a in AS (n = 53) compared to healthy subjects (n = 57). Furthermore, syndesmophyte development in patients with AS was associated with the expression of miR-125a-5p, miR-151a-3p, miR-150-5p and miR-451a [200]. Another miRNA, miR-10b-5p, has shown to be a Th17 regulator present in Th17 cells from AS patients. Thus, miR-10b-5p decreases the IL-17A production by inhibiting the mitogen-activated protein kinase 7 (MAP3K7) in CD4+ T lymphocytes [201]. Other miRNAs found in diseased tissue of hip ligaments from AS patients showed differential expression of 22 miRNAs from AS patients compared to health controls [202]. Altogether, these miRNAs could be attractive promising biomarker candidates for AS diagnosis which may help to clarify the disease pathogenesis, however, further validation is required.

CRPM
The MMP-cleaved protein fragment of CRP (CRPM) has been associated with local inflammation [203] and appears to be a more specific inflammation biomarker for AS compared to measuring the full-length CRP [204]. CRPM was measured in serum from axSpA patients (nr-axSpA = 121, AS = 72) and CRPM levels were higher compared to asymptomatic controls (n = 100). In addition, CRPM levels were increased in AS patients compared to nr-axSpA, indicating that CRPM reflects axSpA associated inflammation. CRPM correlated with ASDAS-CRP in AS patients and with CRP in nr-axSpA patients. In summary, this study indicated that the CRPM biomarker seems to have potential as prospective laboratory tool to assess disease activity in axSpA patients. However, this study had several limitations such as the lack of wholebody MRI in AS and nr-axSpA to examine the correlations between CRPM and local inflammation [205].

Connective tissue remodelling markers
In inflammatory joints disease, the tissue remodelling homeostasis is disrupted leading to altered extracellular matrix (ECM) remodelling. The ECM consists of several collagens including type I, II, III and IV collagens and when degraded by proteases, it generates new collagen metabolites fragments called neoepitopes, which have been suggested as useful future serological biomarkers of SpA [206].
The neoepitope biomarker of type I collagen (C1M) is associated with destruction of soft tissue and high levels have been observed in patients with AS [207]. The neoepitope biomarker of type II collagen (C2M) reflects the degradation of cartilage [208], and metabolite of type III collagen is linked to degradation of soft tissue (C3M). Elevated levels of C2M and C3M have been observed in patients with AS compared to healthy subjects [207]. In addition, C3M has been found in inflamed tissue of liver fibrosis [209]. The neoepitope biomarker of type IV collagen (C4M) has also been found elevated in soft tissue destruction [210]. Recently, C1M, C2M, and C3M have been related with the response to biologic therapy in AS [206] and have shown correlation with CRP, ESR, and radiographic severity [211]. Thus, it has been proposed that the pathological process in AS involves a higher ECM turnover with elevated levels of MMP-cleaved collagen products. In a study by Hušáková et al. C1M, C2M, C3M and C4M were quantified by ELISA in serum from AS (n = 72) and nr-axSpA (n = 121) and controls (n = 100). All collagen metabolite biomarkers showed elevated levels in axSpA patients compared to the controls. Furthermore, the serum levels of C1M, C3M and C4M were significant higher in AS compared to nr-axSpA. The best biomarker to distinguish between AS and nr-axSpA was C3M, but more investigation of C3M is needed before validation as a diagnostic tool within axSpA. Both C1M, C3M, and C4M were correlated with ASDAS-CRP in AS and nr-axSpA reflecting association with disease activity. Altogether, the current study showed that serological MMP-mediated metabolites from collagen may have potential as novel disease activity biomarkers in axSpA however, more comprehensive studies are needed to confirm these results [212]. In another study, Gudmann et al. measured the same serological neoepitope biomarkers, C1M and C3M, in serum from patients with axSpA (n = 110), PsA (n = 101), and healthy controls (n = 120). The C1M and C3M levels were both significantly elevated in axSpA and PsA patients compared to controls. Furthermore, C1M and C3M correlated with ASDAS and Disease Activity Score 28 (DAS28). In summary, this study demonstrated that C1M and C3M could be used to distinguish between diseased and healthy subjects and the increase in biomarker levels was associated with disease activity in axSpA and PsA [213]. In addition, Siebuhr et al. measured C1M, C2M and C3M including CRP, ESR, CRPM in serum from AS patients during their first year of treatment with etanercept, a TNF inhibitor. Study showed that ESR, CRP, BASDAI and C1M were significantly decreased with treatment. Furthermore, C1M and CRP showed to predict treatment efficacy. Overall, this study demonstrated that quantification of acute inflammation and connective tissue remodelling can determine which AS patients will most likely benefit from the etanercept treatment [206].
Besides measuring tissue biomarkers from collagen degradation, PRO-C2, a marker of mature cartilage collagen type IIB formation, and C-Col10, a marker of type X collagen turnover exclusively expressed by hypertrophic chondrocytes, were measured in 2016 by Gudmann et al.
The study aimed to investigate the cartilage metabolism in axSpA (n = 110) and PsA (n = 101) patients by measuring and evaluating the diagnostic utility of PRO-C2 and C-Col10. The PRO-C2 levels showed to be significantly increased in serum samples of both axSpA and PsA compared to healthy subjects. Likewise, C-Col10 was significantly increased in PsA patients compared to healthy subjects, and a similar trend was seen in axSpA patients. Interestingly, a small group of axSpA patients (n = 34) naïve to anti-TNF therapy were significantly lower in C-Col10 levels compared to the axSpA patients undergoing TNF inhibiting treatment. This supports the previously mentioned theory that resolved inflammation as a result of TNF inhibiting treatment leads to an increased rate of ossification which is seen by the increase in C-Col10 levels reflecting hypertrophic chondrocytes. In conclusion, these findings suggest that PRO-C2 and C-Col10 may be promising markers for investigating the cartilage collagen metabolism in SpA patients [214].  Fig. 6. Summary of potential biomarkers for diagnosis, monitoring disease activity, treatment response, and prognostics in patients with spondyloarthritis. P, Patients; C, Controls.

Markers of bone remodelling
As previously mentioned, the Wnt/β-catenin signalling pathway including its inhibitors, Dkk proteins and sclerostin, are believed to be involved in the pathogenesis of SpA. Dkk1 and sclerostin have been suggested as markers, however, conflicting data exists with regards to bone formation, Dkk1 and sclerostin correlation to disease progression. Sclerostin has shown to be significantly lower in serum and in local expression in patients with AS compared to healthy controls [215,216]. Low serum levels of sclerostin have been related to the development of syndesmophytes in AS patients [216]. AS patients treated with anti-TNF therapy showed gradually increased levels of sclerostin over 12 months, while the sclerostin levels stayed substantially lower in AS patients compared to controls [217]. Higher baseline CRP levels were observed in AS patients with lower serum sclerostin. Hereof, AS patients with low serum sclerostin had increased risk of having high CRP levels after 12  206 Type II collagen formaƟon markers, PRO-C2 and PRO-C10.

P=211, C=118
The PRO-C2 was significantly increased in axSpA and PsA compared to healthy subjects. Likewise, C-Col10 was significantly increased in PsA compared to healthy subjects, and similar trend was seen in axSpA paƟents. axSpA paƟents naïve to anƟ-TNF therapy were significantly lower in C-Col10 levels compared to axSpA paƟents undergoing TNF inhibiting treatment. 214

P=204, C=80
Lower sclerosƟn levels were found in AS serum and correlated with high CRP levels. 215 P=46, C=50 Low sclerosƟn serum levels have been found in paƟents with AS and months of treatment compared to patients with higher baseline levels of sclerostin [215,217]. These results may indicate that baseline levels of sclerostin could serve treatment predictive potential of anti-TNF therapy and most likely other types of treatment. Serum levels of Dkk1 have shown to be lower in AS patients compared to healthy controls [145,218]. In AS patients with no syndesmophyte growth, the Dkk1 levels were significantly higher compared to patients with formation of syndesmophytes over 2 years. These results suggest that high serum levels of Dkk1 are protective by reducing Wnt signalling and inhibiting new bone growth, subsequently syndesmophyte growth and spinal ankylosis [165]. Thus, low serum levels of sclerostin and Dkk1 may explain the syndesmophyte growth and ankyloses. However, conflicting data exists of the Dkk1 and sclerostin levels. In contrast to the low levels of Dkk1 observed in AS patients, in another study high levels of serum Dkk1 were found in AS patients compared to healthy subjects including other types of arthritis. Furthermore, in a study employing meta-analysis the serum levels of sclerostin in AS patients were not significantly different to controls [219]. More studies are needed before knowing whether sclerostin and Dkk1 levels can be used as potential biomarkers for structural progression in patients with AS.

Adipokines
The adipokine, visfatin, is released by various cells and can affect bone metabolism by directly activating osteoblasts and promoting angiogenesis [220]. In a study by Syrbe et al., increased baseline levels of visfatin have shown to predict radiographic spinal damage progression in AS over 2 years [221]. Leptin, another adipokine, together with high molecular weight form of adiponectin (HMW-APN) showed to significantly predict protection from spinal radiographic progression in AS patients. In general, females with AS displayed higher levels of leptin and HMW-APN compared to males, which may explain the decreased structural damage in the spine compared to male patients with AS [222]. In addition, serum leptin levels were higher in men with AS compared to healthy male controls and significantly correlated with serum IL-6 levels and BASDAI [223]. From these results, leptin might have a protective effect by inhibition of the extensive aberrant bone formation and a possible role in the inflammatory reactions of AS.

MIF
Macrophage migration inhibitory factor (MIF) is a pro-inflammatory cytokine which has previously been shown to be elevated in serum from AS patients [224]. In a study by Ranganathan et al. serum levels of MIF were significantly elevated in AS patients compared to healthy controls, particularly in patients with disease progression and with appearing syndesmophytes. In addition, increased levels of MIF were found in synovial fluid collected from inflamed peripheral joints of patients with AS. Furthermore, MIF have showed to increase the TNF-α secretion by monocytes and mediated stabilisation of β-catenin and promoted mineralisation of osteoblasts. Thus, MIF has been proposed to have a direct role on the mineralisation of osteoblasts [225]. More studies are needed associated with the formaƟon of new syndesmophytes. P=30, C=30 Low baseline levels of sclerosƟn in serum are associated with high CRP levels in AS paƟents receiving anƟ-TNF therapy. 217
P=204, C=80 Low Dkk1 levels are correlated with high CRP levels in AS paƟents. 215

P=65
Higher levels of Dkk1 in paƟents with no formaƟon of syndesmophyte compared to those with clear syndesmophytes progression. 165 PROGNOSTIC BIOMARKERS VISFATIN Adipokine involved in bone metabolism, shown to promote angiogenesis.
P=86, C=25 VisfaƟn levels were higher in AS paƟents than in healthy subjects. 221

LEPTIN
Adipokine involved in bone metabolism.

P=120
Serum levels of lepƟn are inversely related to radiographic spinal growth in paƟents with AS. 222 P=42, C=42 Higher serum lepƟn levels in men with AS. LepƟn levels correlated with serum IL-6 levels and BASDAI. 223
P=147, C=61 Serum levels of MIF were elevated in AS paƟents, parƟcularly in paƟents with disease progression and with appearing syndesmophytes. Increased levels of MIF were found in synovial fluid collected from inflamed peripheral joints of paƟents with AS. to investigate the prognostic biomarker potential of MIF. Although these biomarkers may have potential value in axSpA, most of these studies have not yet been reproduced or evaluated in prospective studies for the validation as gold standard biomarkers (Fig. 6).

Conclusion
TNF inhibition has long been the only targeted approach to treat SpA, showing good efficacy in a fraction of SpA patients yielding mixed results in preventing ossification of damaged bone or cartilage [226]. This situation has created an emerging need to discover new treatment targets for development of novel therapies for axSpA patients. Consequently, the role of IL-17 in inflammatory process and structural damage in axSpA has opened new opportunities to treat SpA. Secukinumab, the first IL-17 inhibiting agent approved for treatment of axSpA patients, has showed high clinical efficacy. Subsequently, several other IL-17 blocking agents such as ixekizumab, brodalumab, bimekizumab, netakimab, remtolumab, and COVA322 are licensed or under development and expected to be available for treatment in the next couple of years due to positive results observed of phase II/III trials. The introduction of anti-IL-17 therapy has expanded the number of options for targeting cytokine pathways in SpA beyond TNF and marked a step forward in the management of SpA. However, despite the positive results of IL-17 inhibitors on the symptoms which are mostly related to pain of SpA, only limited data exist on their effect on such drugs on the inflammatory disease process in itself and the related structural damage, especially in axial disease. To approach these challenges, a search for objective biomarkers reflecting disease activity including dynamic structural progression are of key interest. Moreover, combinations of biomarkers together with imaging data may provide a future analysis platform for novel drug discovery, improvement of clinical trial design, and patient inclusion when testing new treatments, which can aid in the improvement of individualised care.

Declaration of Competing Interest
Solveig Skovlund Groen, Dovile Sinkeviciute, Simon Francis Thomsen, Georg Schett have no competing interest. Morten A. Karsdal, Anne-Christine Bay-Jensen, Christian S. Thudium and Signe Holm Nielsen are employees of Nordic Bioscience. Morten A. Karsdal and Anne-Christine Bay-Jensen hold stock in Nordic Bioscience.