Spondyloarthritis (SpA) is a family of chronic rheumatic diseases that share many clinical and pathogenic features. A main characteristic of SpA is the involvement of the spine and sacroiliac joints in the inflammatory process ^[1][2][1,2]. SpA includes ankylosing spondylitis, psoriatic arthritis, reactive arthritis, enteropathic arthritis, and undifferentiated spondyloarthritis [3]. More recently, the term axial spondyloarthritis (axSpA) has been coined. AxSpA comprises a radiographic form (r-axSpA), previously referred to as ankylosing spondylitis, and a non-radiographic form (nr-axSpA). The latter is characterized by sacroiliitis, whose signs are not detectable by conventional radiography but only by MRI of the sacroiliac joints [4]. The Assessment of SpondyloArthritis International Society (ASAS) has identified a set of criteria for the diagnosis of both radiographic and non-radiographic axSpa [4]. These include the presence of a major radiologic criterion, such as radiographic changes in the sacroiliac joints or the presence of bone edema on MRI, plus a minor criterion or a major criterion, such as the presence of the HLA-B27 allele [5] plus two minor criteria. AxSpA is believed to originate from an inflammatory process at the level of the enthesis that can subsequently extend to the joint structures. Joint manifestations may be associated with extra-articular involvement, including the presence of psoriasis, uveitis, or chronic inflammatory bowel disease, particularly Crohn’s disease. Another feature of spondyloarthritis is the presence of dactylitis, characterized by severe inflammation of the finger or toe tendons [6]. All these non-joint conditions are, in many cases, the most important pathological components of axSpA. Traditional therapy of axSpA has been based on the use of nonsteroidal anti-inflammatory drugs (NSAIDs) [7]; however, new innovative drugs have become available in the past 20 years. These drugs, known as biologics, are monoclonal antibodies or soluble receptors capable of blocking with high specificity several cytokines involved in disease pathogenesis. The main targets of these drugs are tumor necrosis factor-α (TNFα), interleukin (IL)-17, and IL-23 [8]. Although the use of these new drugs has greatly improved the quality of life of patients and can slow the radiological progression of the disease, a significant number of subjects do not respond or partially respond to therapy. Therefore, the availability of new drugs for these difficult-to-treat patients is urgently needed ^[9][10][9,10]. In recent years, thanks to intensive research in rheumatic diseases, a new class of drugs called Janus kinase inhibitors (JAKinibs) have become available. These drugs were initially used successfully in rheumatoid arthritis [11].

After recognizing their receptor on the target cell, cytokines send their signal to the nucleus through biochemical interactions between different molecules in the cytoplasm. JAK combination with signal transducer and activator of transcription (STAT) is one of the most important cytokine signal transduction pathways ^[12][13][14][45,46,47]. The JAK–STAT pathway involves those cytokines that bind to type I/II cytokine receptors ^[15][48]. There are four members of the JAK family, namely JAK1, JAK2, JAK3, and tyrosine kinase 2 (TYK2). Each cytokine receptor is associated with a homodimeric or heterodimeric pair of these JAK molecules ^[16][49]. After the cytokines bind to their receptors, JAK molecules associated with the intracellular portion of the receptor undergo autophosphorylation and then phosphorylate, in turn, the receptor tail on tyrosine ^[17][50]. Seven members of the STAT family have been identified, namely STAT1, STAT2, STAT3, STAT4, STAT5A, STAT5B, and STAT6. Dimers of STAT bind to phosphorylated docking sites on receptors, where they are phosphorylated by JAK. Once phosphorylated, STAT molecules dissociate from receptors, form homo- or heterodimers, and migrate into the nucleus regulating the expression of target genes ^{[15][16][17][18][19]}[48,49,50,51,52]. Regulation of gene expression requires the recruitment of coactivators by STAT dimers. These coactivators interact with histone proteins with which nuclear DNA is associated, making specific regions of DNA more accessible to STAT and the nuclear transcriptional machinery ^[20][21][53,54]. STAT molecules are then de-phosphorylated and once dissociated from DNA leave the nucleus. As pointed out earlier, JAK molecules are responsible for the signaling of several cytokines. JAK1, in combination with JAK3, is involved in the signaling of cytokine that recognizes receptors formed by common γ chain (γc), such as IL-2, IL-4, IL-7, IL-9, IL-7, and IL-21 ^[15][16][48,49]. These cytokines are involved in the growth/maturation of lymphoid cells and the differentiation/homeostasis of T cells and natural killer cells ^{[22][23][24][25][26]}[13,44,55,56,57]. IL-7, in particular, modulates innate lymphoid cells (ILC), which are strongly implicated in the pathogenesis of axSpA through production of Il-17 ^[27][58]. JAK2 in homodimeric form is associated with receptors recognized by growth factors, including erythropoietin and granulocyte colony-stimulating factor (G-CSF), and with β-chain (βc) receptors recognized by IL-3 and IL-5. Homodimers of JAK2 regulating signaling downstream of erythropoietin and G-CSF play a key role in erythropoiesis and myelopoiesis ^[28][29][59,60]. Importantly, GM-CSF has recently been linked to the pathogenesis of axSpA ^[30][61]. TYK2, in combination with JAK2 and JAK1, may be associated with receptors that share gp130 molecules recognized by different cytokines, including IL-6 and IL-11. IL-6 is also involved in the activation of ILC ^{[13][22][28][29][31][32]}[13,46,59,60,62,63]. TYK2, in combination with JAK1/JAK2, is associated with type I and type II interferon receptors. It is noteworthy that JAK2 and TYK2 regulate IL-12 and IL-23 signaling. The latter cytokine is required to maintain the differentiation state of T-helper 17 cells, which are mainly involved in axSpA immunopathogenesis ^[33][64]. 

Biologic drugs are highly effective in the treatment of axSpA. However, a significant number of patients do not respond to the inhibition of anti-TNFα therapy or have secondary failure to such therapy ^[34][66]. Anti-IL-17 monoclonal antibodies, while effective in treating patients with axSpa, are significantly less effective in those patients who have been previously treated with anti-TNFα ^[35][67]. Therefore, there is an urgent need for new treatments for these difficult-to-treat patients. It has been shown that a large number of cytokines recognize multiple receptors that can induce intracytoplasmic signaling through the activation of different JAK molecules. The cytokines most implicated in the genesis of spondyloarthritis are IL-23 and IL-17. These two cytokines are so functionally connected that the term “IL-23/IL-17 axis” has been coined. IL-23 can maintain the differentiative state of T helper-17 (Th17) cells, impeding these cells’ trans-differentiation into T-regulatory cells (Treg) and inducing Th17 cells to produce IL-17 family members ^[36][68]. Specifically, IL-17A and IL-17F mediate tissue damage by stimulating target cells that express receptors for IL-17. These cells, in turn, produce potent soluble proinflammatory factors. This leads to joint erosion, enthesitis, and disorders of bone proliferation ^[37][38][69,70]. In addition to Th17 cells, other cell types belonging to both adaptive and innate immunity can produce IL-17, and thus, participate in the pathogenesis of spondyloarthritis. Among them are IL17⁺CD8⁺ T cells ^[30][39][40][61,71,72], γδT cells ^[41][42][73,74], mucosal-associated invariant T (MAIT) cells ^[43][44][75,76], invariant natural killer T (iNKT) cells ^[45][77], and group 3 innate lymphocyte cells (ILC3) ^[46][78]. Four JAK inhibitors are currently available for rheumatic diseases, namely tofacitinib, baricitinib, upadacitinib, and filgotinib. The selectivity of different JAK has been evaluated in vitro through several laboratory tests, including biochemical assays using recombinant JAK molecules and cellular assays in which cell lines are treated with JAK inhibitors and then stimulated with cytokines to assess their ability to prevent phosphorylation of STAT molecules. In these assays, tofacitinib demonstrated preferential inhibition of JAK1 and JAK3, with 5- to 100-fold selectivity over JAK2 ^[47][79]. Filgotinib demonstrated 30-fold selectivity for JAK1 vs. JAK2-dependent signaling. Upadacitinib showed higher selectivity for JAK1 than for JAK2, JAK3, and TYK2, demonstrating 60-fold selectivity for JAK1 vs. JAK2 and >100- vs. JAK3 in cellular assays ^[48][80]. There are numerous differences among JAK inhibitors, in addition to their selectivity, such as chemical structure, inhibition potency, metabolism, and excretion profiles. These variables indicate that the clinical effect of JAKinibs may show significant clinical differences. Consistent with the mode of action of JAKinibs, biomarker analyses have shown that tofacitinib inhibits preferentially JAK1 and JAK3 but also, in part, JAK2 inhibiting the broadest array of cytokines as compared with other JAKinibs. Upadacitinib exerts direct inhibitory activity on several JAK1-dependent factors (IFN-α/β, IFN-γ, IL-6, IL-2, IL-5, and IL-7) and indirectly on several JAK1-independent pathways (IL-1, IL-23, IL-17, IL-18, and TNFα ^[32][63] resulting in inhibition of key cytokine-induced events, such as leukocyte activation and mobility, inflammatory response, and connective tissue damage. Filgotinib has also been shown to reduce circulating proinflammatory cytokines and chemokines, adhesion molecules, and markers of matrix remodeling associated with axSpA ^[49][86]. In addition, preclinical models have demonstrated the impact of the JAK–STAT blockade on PsA manifestations ^[50][51][87,88] also through a TNF-dependent mechanism and TNF-independent mechanism ^[50][87]. As discussed earlier, a single JAK inhibitor allows simultaneous inhibition of multiple cytokines, with possibly greater efficacy. While baricitinib, as mentioned above, has been approved exclusively for rheumatoid arthritis, tofacitinib, upadacitinib, and filgotinib have also been approved for the treatment of psoriatic arthritis and ulcerative colitis. Tofacitinib and upadacitinib have recently received approval for the treatment of rx-SpA, and the indication of upadacitinib has also been extended to nr-ax-SpA based on the results of the SELECT-AXIS 2 study, as explained in more detail below ^[52][89]. Therefore, the following section will discuss the different clinical trials aimed at establishing the efficacy and safety of different JAK inhibitors in the treatment of axSpA. 

As discussed above, tofacitinib is a relatively nonselective JAKinib that can mainly inhibit JAK3, JAK2, and JAK1 ^[19][53][52,90]. In a phase II study, tofacitinib proved effective in a small sample of patients with axSpA ^[54][91]. This 12-week study looked at 207 r-axSpA patients who received different doses of tofacitinib, 2 to 10 mg twice daily or placebo. The primary endpoint was ASAS20 (Assessment in Ankylosing Spondylitis 20% improvement) response rate at week 12. Patients treated with 5 mg twice daily achieved an ASAS20 response rate of 80.5%, significantly higher than that of the control group (41.2%). Secondary endpoints such as ASAS40 and Bath ankylosing spondylitis disease activity index 50 (BASDAI50), meaning 50% improvement in BASDAI compared with baseline, as well as a change in the ankylosing spondylitis disease activity score (ASDAS), showed a significant improvement with tofacitinib 5 and 10 mg twice daily compared with placebo. Patients with objective signs of inflammation (elevated CRP or spine edema on MRI) of the sacroiliac joint presented greater treatment efficacy compared with placebo. Changes in MRI scores were analyzed at week 12, with a significantly greater reduction from baseline with tofacitinib 5 and 10 mg compared with placebo. Adverse events were similar between treatment groups ^[54][91]. Overall, at week 16, there was a greater percentage of ASAS20 response in the tofacitinib group than in the placebo group (56.4% vs. 29.4%). ASAS40 response, universally considered a highly significant secondary endpoint, was greater with tofacitinib than with placebo (40.6% vs. 12.5%). Treatment efficacy was maintained until week 48 ^[55][93]. Regarding upadacitinib, it was evaluated in r-axSpA in the SELECT-Axis 1 study ^[56][94]. The 187 patients with r-axSpA who were naive to treatment with biologic drugs and had not responded satisfactorily to at least two NSAIDs were included. The patients were then randomized to receive upadacitinib 15 mg daily or placebo for 14 weeks. Patients who completed the first period were admitted to the second phase of the study, where they received upadacitinib open-label until week 104. The primary endpoint for the first part of the study was the response to ASAS40 at week 14. The response was significantly greater in the upadacitinib group than in the placebo group (52% vs. 26%). Several secondary endpoints were also achieved in the upadacitinib group, but not in patients who received placebo. Secondary endpoints included improvement in ASDAS scores, spine MRI radiology score, and percentage of patients with BASDAI50 and ASAS partial remission. The interim analysis of the SELECT-Axis 1 extension study reported efficacy and safety data at 1 year ^[57][95]. Results showed sustained treatment efficacy for 1 year and increased ASAS40 response throughout the study. Study. The percentage of patients who responded to ASAS40 was higher at week 64 than at week 14. Patients who switched from placebo to upadacitinib showed a similar level of response to those initially randomized to upadacitinib. No significant side effects occurred in one year. The SELECT-AXIS 2 study examined the efficacy of upadacitinib in nr-axSpA ^[52][89]. In this multicenter, randomized, double-blind, placebo-controlled, phase 3 study, adult subjects with objective MRI-based signs of inflammation or elevated C-reactive proteins and an inadequate response to NSAIDs were enrolled. Patients were randomly assigned to receive upadacitinib 15 mg orally once daily or placebo, based on MRI of sacroiliac joint inflammation and high-sensitivity C-reactive protein screening status and previous exposure to disease-modifying biologics. The primary endpoint was the proportion of patients with an ASAS40 response at week 14. Out of a total of 313 patients, 156 received upadacitinib and 157 received placebo. A significantly higher ASAS40 response rate was observed with upadacitinib compared with placebo at week (45% vs. 23%). The rate of adverse events up to week 14 was similar in the upadacitinib group and the placebo group.

Finally, filgotinib, a selective JAK1 inhibitor ^[58][96], was evaluated in patients with axSpA in a double-blind, placebo-controlled phase-two study (TORTUGA) ^[59][97]. The primary endpoint was met, with greater improvement in ASDAS score at week 12 in the filgotinib group than in the control group. Secondary endpoints, which included improvements in ASAS20, ASAS40, at least 20 percent improvement in function, pain, inflammation, global patient, CRP, and spinal mobility, ASAS partial remission, and functional ankylosing spondylitis index (BASFI) at week 12, were achieved in the filgotinib arm compared with placebo. There was also a significant reduction in inflammation scores on MRI of the spine and sacroiliac joint. Safety was considered satisfactory ^[59][97]. Although this analysis showed a greater reduction in inflammatory segments of the spine evaluated, no improvement in bone erosion or new bone formation was observed ^[60][98]. On the other hand, the sacroiliac joint study showed a significant reduction in erosion score in the filgotinib group compared with placebo, demonstrating significant drug activity at this level observable as early as 12 weeks ^[61][99]. However, despite these promising results, filgotinib has not yet received approval for the treatment of axSpa.