Pathogenesis of ligaments ossification in spondyloarthritis: insights and doubts

Abstract

Despite intensive research in spondyloarthritis pathogenesis, some important questions still remain unanswered, particularly concerning enthesis new bone formation. Several evidences suggest that it prevalently occurs by endochondral ossification, however it remains to identify factors that can induce and influence its initiation and progression. Recent progress, achieved in animal models and in vitro and genetic association studies, has led us to hypothesize that several systemic factors (adipokines and gut hormones) and local factors (BMP and Wnt signaling) as well as angiogenesis and mechanical stress are involved.

We critically review and summarize the available data and delineate the possible mechanisms involved in enthesis ossification, particularly at spinal ligament level.

KEY MESSAGES

• Complete understanding of spondyloarthritis pathophysiology requires insights into inflammation, bone destruction and bone formation, which are all located in entheses and lead all together to ankylosis and functional disability.

• Several factors probably play a role in the pathogenesis of bone formation in entheses including not only cytokines but also several systemic factors such as adipokines and gut hormones, and local factors, such as BMP and Wnt signaling, as well as angiogenesis and mechanical stress.

• Data available about pathophysiology of new bone formation in spondyloarthritis are limited and often conflicting and future studies are needed to better delineate it and to develop new therapeutic approaches.

Keywords:

• Spondyloarthritis
• enthesis
• spinal ligament
• ossification

Introduction

Spondyloarthritis (SpA) is a cluster of inflammatory arthritis that consists of ankylosing spondylitis (AS), reactive arthritis, arthritis/spondylitis associated with psoriasis (PsA), arthritis/spondylitis associated with inflammatory bowel diseases (IBD) and “undifferentiated” SpA (uSpA). These five subtypes, in turn, have been split into “axial” and “peripheral” SpA according to whether the involvement is mainly in the spine or in the extremities (1). Even if they may manifest with very heterogeneous clinical features, they share the same genetic background with HLA-B27 positivity (ranging from 50 to 95% in SpA patients according to the clinical presentation), sacroiliitis and spinal inflammation, peripheral arthritis and enthesitis. This last one, the inflammation of entheses, the hallmark of SpA, distinguishes it from other inflammatory rheumatic disorders, such as rheumatoid arthritis (2).

Entheses, the attachment sites of tendon, ligament, joint capsule, fascia or muscle to bones, with role to dissipate mechanical stress and to provide optimal myofascial stability, are hypothesized to be the primary target tissue for inflammation in SpA, responsible for many of the symptoms, and for the multitude of locations of pain in SpA patients. Two major types of entheses have been histologically described by Benjamin’s Group: the fibrous type composed of pure dense fibrous connective tissue and the fibro cartilaginous type, the most common, composed of an intermediate tissue between dense connective tissue and cartilage (3,4). Specifically, four histologically distinct zones have been described in the fibro cartilaginous type entheses: fibrous zone with fibers of type 2 collagen and versican (zone 1), as the predominant extracellular matrix proteoglycan, uncalcified fibrocartilage zone (zone 2), in which aggrecan prevails, calcified fibrocartilage zone (zone 3) and subchondral bone (zone 4). Zone 2, avascular and separated from zone 3 by a regular calcification front, tidemark, represents the site where the inflammation initially occurs and from which extends to the synovium and adjacent bone tissue (5).

Pathological enthesis in SpA is characterized by CD4+ and CD8+ T lymphocyte cell infiltration, edema, angiogenesis, fibrosis, osteitis, erosion and new bone formation (6). Ossification of enthesis can occur everywhere, both at axial and peripheral level and can also involve the insertion of spinal ligaments, as demonstrated by X-ray imaging, the classical tool for identification of structural changes in SpA.

The presence of hypertrophic chondrocytes at entheseal site suggests that bone formation occurs as an endochondral bone formation process (7). Unfortunately, because of the relative inaccessibility of these sites we still lack detailed human studies to define the cause of new bone formation and factors that can induce and influence its initiation and progression. We will summarize the available data from animal models and genetic studies and delineate the possible mechanisms involved in enthesis ossification, particularly at spinal ligament level.

Spinal ligaments ossification and inflammation

In 1983, Ball has reported that enthesis inflammation causes destruction of the attachment of ligaments to bone, which in turn leads to reactive new bone formation (8). More recently, the detailed histological analysis of axial joint sections obtained from rats with increased human beta-2 microglobulin transgene copy numbers, a well-validated SpA rat model (B27/hβ₂m-transgenic rat) (9), revealed a significant correlation between the degree of inflammation and tissue destruction, between the degree of inflammation and osteoproliferation, and the degree of tissue destruction and osteoproliferation (10).

Overexpression of tumor necrosis factor (TNF)-α, obtained by deletion of adenosine-uridine rich elements (ARE) from the mouse genome, leads to destructive polyarthritis and enthesitis, supporting the key role of TNF-α in SpA pathogenesis (11). Additionally tmTNF transgenic rats, which specifically overexpress the transmembrane form (tmTNF) but not the soluble form of TNF (12), showed pronounced axial and peripheral inflammation and new bone formation (13). Taken together, these data strongly support that in rats both destruction and osteoproliferation occur simultaneously with inflammation and that all processes are driven by the tmTNF. Unfortunately, no scientific evidence supports this hypothesis in humans. The ultrasonography analysis of AS enthesis revealed that bone erosion and bone formation are two processes topographically and temporally distinct (14).

It is well known that TNF-α plays a central role in initiating and regulating the cytokine cascade during an inflammatory response, by inducing other cytokines such as interleukin (IL)-1 and IL-6, recruiting immune and inflammatory cells and up-regulating adhesion molecules. This cytokine is produced by a number of key cells, including macrophages, monocytes, mast cells, osteoclasts, dendritic cells, and T-cells in the joint and several of these same cell types as well as dermal dendritic cells, Langerhans cells, and keratinocytes in the skin, but it remains to elucidate which cells mainly contribute to its production in SpA spinal ligaments, whether the soluble and membrane bound fractions. The high levels of TNF founded in sacroiliac joint biopsy specimens from patients with AS and peripheral joint tissue and fluid in PsA patients, support the central role of this cytokine in SpA pathogenesis (15,16), but it is worth reminding that TNF directly and indirectly induces osteoclasts differentiation and activity, while concomitantly inhibits osteoblasts formation and activation.

In SpA patients treated with anti-TNF-α, a good control of clinical and biological inflammation and joint destruction, despite the persistent ossification, has been showed by several clinical trials, suggesting that TNF-α could contribute to inflammation and bone resorption, but another pathogenic pathway is involved in driving SpA bone formation. In fact, Baraliakos et al. (17) have showed that occurrence of spinal inflammation and fatty degeneration at baseline, as detected by magnetic resonance, and development of fatty degeneration after 2 years without previous signs of inflammation, were significantly correlated with syndesmophyte formation, as detected by radiography, after 5 years of anti-TNF therapy. Nevertheless, the greater part of these syndesmophytes did not show corresponding magnetic resonance lesions at baseline and the sequence “inflammation-fatty degeneration-new bone formation” was rarely observed. Moreover, Van der Heijde et al. (18) failed to demonstrate effects of anti-TNF therapy on the structural progression in AS patients after 2 years of treatment and compared with an historical cohort of patients not treated by anti-TNF but with similar parameters of disease activity. Nevertheless, a more recent study conducted by Haroon et al. (19) demonstrated on a multivariate analysis that anti-TNF treatment was associated with a reduced radiographic progression in AS patients, and this effect was even more pronounced if the delay before exposure to the treatment was short and if exposure to TNF-blockers was longer. This suggests that blocking TNF and limiting inflammation could be sufficient to limit bone formation.

Recently, new insights have been gained into the pathogenesis of SpA. Although data are limited at this point, the involvement of IL-23/IL-17 axis seems to play a key role in SpA pathogenesis (20). IL-23 is sufficient to drive enthesitis and promotes IL-17 and IL-22 expression by entheseal resident cells in mice with collagen-antibody-induced arthritis (CAIA) (21). In agreement, the IL-23 receptor R381Q gene variant protects against psoriasis, Crohn's disease and AS and decreases IL-23-dependent IL-17 and IL-22 production (22). Even if it remains unclear which immune cells, besides T-helper (Th)-17, express IL-23 receptor and respond to IL-23 by producing IL-17 and IL-22, several evidences emphasize the pathogenic effects of these cytokines in SpA. In fact, a prevalent expression of IL-23 and an increased number of innate lymphoid cells-3 (ILC3s) has been demonstrated in the gut of patients with AS (23,24). ILCs are specialized immune cells which are divided into three subsets (ILC1, ILC2, ILC3) in order to their specific expression of cytokines and chemokines. In particular, since ILC3s respond to IL-23 by producing IL-17 and IL-22, these cells seem to play a role in AS pathogenesis. Moreover, recently, Ciccia et al. (25) have demonstrated an increased number of gut-derived ILC3s in the peripheral blood, synovial fluid, and inflamed bone marrow of patients with AS, suggesting the existence of a strong link between gut and sacroiliac joint inflammation.

Clinical trials in AS and/or PsA sub-forms, using secukinumab (a fully human anti-IL-17A monoclonal antibody) (26,27), revealed the reduction of signs and symptoms of disease. Significantly higher baseline circulating Th17 and serum IL-17 and IL-23 were observed in active AS patients than in healthy controls, with a strong positive association between their levels and disease activity (28). Elevated IL-17 producing cells together with a marked proliferation of fibroblast-like cells positive for bone morphogenetic protein (BMP)-2 in association with heterotropic formation of cartilages and bones in hyperplastic entheseal tissues, was found in male mice that spontaneously develop seronegative ankylosing enthesitis (29). Overexpression of IL-17 drives synovial inflammation and joint destruction in vivo (30) and, conversely, IL-17 deficiency or inhibition protects from joint inflammation and damage (31). Thus, IL-17 could be involved in SpA enthesitis and bone erosion, whereas IL-22 could explain the new bone formation. IL-22, in turn, induces genes that regulate bone formation, specifically those that encode Wnt family members, bone morphogenic proteins and alkaline phosphatase (ALP) in vitro and activation of the signal transducer and activator of transcription 3 (STAT3) in vivo. The importance of STAT3 signaling has been extensively described: STAT3 mutations reduce bone mass and increase incidence of minimal trauma fractures in humans, as well as its inactivation in bone induces less responsivity to mechanical loading in mice (32). Interestingly, in vitro, STAT3 signaling is responsible for promoting bone formation by runt-related transcription factor 2 (Runx2) and ALP up-regulation and by DKK down-regulation in human mesenchymal stem cells (MSCs) (33). Noteworthy, Asari and co-workers have recently demonstrated the isolation of MSCs from human spinal ligaments and their localization in situ (34). These cells are able to differentiate into osteogenic, adipogenic, or chondrogenic cells. Particularly, after treatment with osteogenic medium for 21 days, MSCs have changed from a spindle-shaped morphology, forming a mineralized matrix, as shown by Alizarin Red S staining, and over-expressing Runx2 and ALP mRNA (34).

Thus IL-17 and Il-22 could explain the concurrence of inflammation, bone erosion and bone formation, and could be potential new therapeutic targets to treat SpA. Nevertheless, the research in this field remains at the discovery stage and many outstanding questions regarding their involvement in SpA pathogenesis need answers. No data are available concerning the role of these cytokines at spinal ligaments levels, neither in vivo nor in vitro.

A role for chemokine (C-C motif) ligand 19 (CCL19) and 21 (CCL21) has been described in SpA. Studies have demonstrated that CCL19, CCL21 and their corresponding receptor CCR7 have a role in angiogenesis and in chemotaxis of monocytes/macrophages, mature dendritic cells and naive T cells (35–39). Recently, Qin et al. (40) have found serum levels of CCL19/CCL21 higher in AS patients than in healthy controls. Moreover, mRNA levels of CCL19/CCL21 in AS hip ligament tissue were found significantly higher than in osteoarthritis ligament tissue, with a significant up-regulation of the expression of bone markers, such as ALP, osteocalcin, Runx-2 and osterix, after exogenous CCL19/CCL21 treatment of AS ligament fibroblast cultures (40). Considering that fibroblasts are the principal cell type in ligament tissues and are directly associated with ligament ossification in AS patients (41,42), Qin et al. (40) hypothesized a role in ligament ossification for CCL19 and CCL21 by promoting fibroblast osteogenic differentiation in AS patients.

Local factors

The transforming growth factor-β (TGF-β) has been demonstrated to be involved in cell proliferation and differentiation and extracellular matrix protein synthesis. Even if genetic studies have not found a strong role for TGF-β gene polymorphisms in SpA (43), several evidences support its role in SpA pathogenesis. TGF-β has been immunohistochemically localized in the ligament tissue of patients with ossification of the posterior longitudinal ligament (44). Moreover, TGF-β has showed an inhibitory role on proliferation of cultured cells derived from the human spinal ligament (45). Additionally, higher TGF-β amounts have been found in the connective tissues and cartilage of patients with more advanced AS when compared to patients with early disease (46). Thus, TGF-β seems to be involved in the late stage of disease and induces bone formation.

BMPs are multi-functional growth factors, belonging to the TGF-β superfamily, which play an important role in osteoblast differentiation and bone formation in an autocrine and paracrine manner. The binding of BMPs to their receptors induces the Smad family proteins phosphorylation which then translocate into the nucleus to direct transcriptional response (47). Nevertheless, there is evidence that BMPs can also act through various kinase pathways such as Smad-independent p38 mitogen activated protein kinase (MAPK) signaling pathway, ERK, JNK, PI3-K, Wnt, and NF-kB (48). Murine fibroblasts of spinal ligaments express BMP receptors (BMPRs), and BMP-2 injection induces proliferation and differentiation into ALP-positive chondrocytes surrounded by an extracellular matrix composed of type I and II collagen (49). Consistent with these data, the expression of BMPRs (BMPRs) has been demonstrated in the ligamentum flavum of the patients with ossification, by immunohistochemical staining (50).

A human in vitro study has demonstrated BMP-2 mRNA expression in cultured spinal ligament cells, obtained from patients with pathological ossification process, and osteogenic differentiation induced by exogenous BMP-2 (51). Interestingly, pro-inflammatory cytokines induced BMPs production in human chondrogenic cells (52) suggesting a link between inflammation and bone formation.

The involvement of BMPs in ligament ossification is also emphasized by recent genetic studies that have indicated BMP-2 as a candidate gene to modify the susceptibility of ossification of posterior longitudinal ligaments of spine. Thymine–guanine base substitution at position 109 causing an arginine–leucine amino acid change in BMP-2 gene, affects the spatial conformation of BMP-2, causing abnormal protein function; this leads to increased level of Smad4 protein expression and the activity of ALP.

The same polymorphism was associated with the occurrence of ligament spinal ossification (53). Similarly, thymine–cytosine transition at nucleotide −726 in exon 3 of BMP-2 gene was associated with genetic susceptibility to ossification of spinal ligament in Chinese patients (54).

Discovered in the 1980s from mouse breast tumors induced by the mouse mammary tumor virus, the Wnt signaling pathway has been investigated extensively and has been characterized as one of the most essential regulator of cell proliferation, cell fate determination, cell differentiation, and cell polarity, which functions by regulating the amount of the transcriptional co-activator β-catenin (55). Mouse genetics have confirmed its importance in the regulation of bone homeostasis, with activation of the pathway leading to increase, and inhibition of the pathway leading to decrease, of bone mass and strength. An in vivo ectopic bone formation study has revealed that BMP-2 is responsible for inducing β-catenin-mediated signaling through Wnt ligands, and β-catenin, in turn, is required for both chondrogenesis and osteogenesis (56). Recently, expression in joints of patients with SpA of Wnt5a, a member of the Wnt family involved in cell proliferation, differentiation, and organogenesis, has been seen to be more associated with ossification than with inflammation. Moreover, Wnt5a seems to play a role in the ossification process by favoring mineralization in cultured osteoblasts and in bone explants. On the contrary, Wnt5a has showed an inhibitory role in the mineralization process in chondrocytes cultures and enthesis explants (57).

DKK-1, an antagonist of Wnt pathway, whose production is enhanced by TNF, has been proposed as potential regulator of balance between bone formation and resorption in animal models of inflammatory arthritis (58). Additionally DKK-1 treatment is involved in osteophyte formation induction (59). Nevertheless, the role of DKK-1 in SpA pathogenesis is highly controversial. A clinical study, investigating serum levels of DKK-1 in patients with AS and healthy controls has provided evidence of reduced levels in SpA (60). Thus, these findings suggest that DKK can regulate bone formation in SpA lesions. Nevertheless, clinical data are really poor and recently no significant differences in DKK levels have been reported between SpA patients and healthy subjects (61). Moreover, Nocturne et al. (62) and Sakellariou et al. (63) have found increased total DKK-1 in serum from SpA patients, anyway a result in line with the possible DKK-1 dysfunction in SpA as proposed by Daoussis et al. (64).

Similarly, discrepancies have been shown in regards to serum levels of sclerostin. Produced by osteocytes, sclerostin antagonizes canonical Wnt signaling by binding to Wnt coreceptors. In 2009, a low serum level of sclerostin has been reported in AS patients and, more interestingly, the level was significantly higher in AS patients without syndesmophyte growth than in AS patients with syndesmophyte growth (65). Conversely, Korkosz’s study group has reported significantly higher sclerostin serum level in AS patients with high disease activity, assessed with bath AS disease activity index (BASDAI), in comparison to healthy subjects and no differences between patients with low disease activity and control subjects (66). In the same study, negative correlation between sclerostin and Dkk-1 in high disease activity group has been revealed, in contrast with the evidence that Dkk-1 triggers sclerostin production (67). Another cross-sectional study, which examined changes in bone mineral density (BMD) and in levels of various bone turnover-related biomarkers and cytokines in a cohort of 55 AS patients, has revealed similar sclerostin levels between patients and controls (61). More recently, Aschermann et al. (68) have demonstrated that sclerostin and Dkk-1 levels were significantly lower in HLA-B27+ subjects compared to HLA-B27 negative controls, independent if they were healthy or affected by SpA or uveitis. Therefore, discrepancies between results of various studies, probably due to the moderate sample size, require other research in this field to clarify the role of Wnt signaling in SpA bone formation. More recently, Nocturne et al. (62) have found significantly decreased sclerostin levels in a large cohort of early SpA compared with healthy controls. Sclerostin levels were significantly associated with age, CRP and DKK-1 serum levels in a multivariate analysis. In fact, higher total serum DKK-1 levels but lower serum levels of sclerostin were found in SpA patients compared to controls, with an association between DKK-1 and sclerostin levels and systemic inflammation and between sclerostin levels and age (62). This study confirmed the presence of decreased sclerostin level in SpA patients as previously described in other studies (65) and as expected in a disease associated with new bone formation.

In recent years, there has been increasing interest for the role of microRNAs (miRNAs) in osteogenic differentiation. miRNAs are fundamental post-transcriptional regulators of gene expression involved in various cellular processes, including proliferation, differentiation, cell cycle, invasion and apoptosis, which act by binding to the seed sequences of the 3′-untranslated region (UTR) of target mRNA sequences and mediating the degradation of mRNA in the RNA-induced silencing complex (69,70). Among miRNAs, miR-29a plays a role in osteogenesis by activating canonical Wnt signaling via inducing β-catenin protein production, and by negative regulation of Dkk-1 and GSK3β (71). A possible involvement of miR-29a dysregulation in inducing new bone formation in AS patients has been hypothesized. Nevertheless, data are limited to few studies which have evaluated miR-29a expression in peripheral blood mononuclear cells (PBMCs) from AS patients with contrasting results (72,73).

Systemic factors

An increased frequency of metabolic syndrome (MetS), a condition characterized by central obesity, hypertension, insulin resistance, and atherogenic dyslipidemia, has been described in patients with rheumatologic disease. Particularly, it has been reported that 45.8% of AS patients has risk of experiencing MetS versus 10.5% in healthy controls (74). Accordingly, higher prevalence of MetS has been found in men with AS, with a direct correlation with disease activity, when compared with controls (75). Recent studies have proposed adipokine profile as novel biomarker and regulator of MetS, given the association of adipokines plasma levels and MetS (76). Adipokines, secreted from adipose tissue, provide an extensive network of communication both within adipose tissue and with other organs, controlling various physiological systems. Several lines of evidence support the notion that among the different adipokines, leptin, and adiponectin are relevant factors, which influence both metabolic disorders and rheumatic diseases. Considering that these adipokines also regulate bone metabolism, it is plausible to hypothesize that they can control SpA bone formation.

Leptin exerts anabolic effects on osteoblasts in vitro and in vivo (77,78). Nevertheless data available concerning the role of leptin in SpA are almost limited and conflicting. Any correlation between serum leptin concentrations and markers of disease activity has not been found by Hulejová and Toussirot study group (79,80). However, one study has reported leptin levels significantly lower in AS patients when compared with controls, but correlated with markers of disease activity (81). Other authors, on the other hand, have reported significantly higher serum levels of leptin in patients with AS than in healthy subjects, with a direct correlation with IL-6 levels and disease activity (82). Nevertheless, at the last EULAR congress Poddubnyy et al. (83) have reported that in SpA patients increased levels of leptin among women might be responsible for a lower extent of structural damage in the spine in women as compared with men.

Besides to regulate energy homeostasis and insulin sensitivity, adiponectin inhibits osteoclastogenesis and stimulates bone formation (84). Higher levels of adiponectin have been reported in AS patients, when compared to normal subjects (85).

Moreover, recently it has been demonstrated the key role of another adipokine, named visfatin in SpA. Syrbe et al. (86) have found that levels of visfatin at baseline were significantly higher in patients with subsequent radiographic spinal progression. Therefore, visfatin could be directly responsible for remodeling of AS joints and could be predictive of subsequent progression of radiographic damage in AS patients (86).

Ghrelin is a gut hormone, with an important role in the stimulation of food intake and gut motility. Its effects on osteoclast are still unclear, but ghrelin treatment enhances in vivo and in vitro osteoblast differentiation and activity (87,88). Even if it is still poorly documented, circulating ghrelin levels result higher in SpA patients when compared to control (80).

Other factors

Even if all fibrocartilages associated with normal entheses are typical avascular, increased vascularity has been observed within enthesis in SpA group (2) and, in agreement, elevated levels of vascular endothelial growth factor (VEGF) have been reported in human ossified ligamentum flavum (89). Another immunohistochemical analysis in AS patients has showed increased microvessel density at sites of acute inflammation, both at the bone–cartilage interface and in the subchondral bone marrow, suggesting that neoangiogenesis is a relevant part of the local immunopathology (90). Considering that vascular invasion of cartilage is necessary for proper bone formation, it is plausible to hypothesize that it contributes to spinal ligaments ossification. There are evidences that in SpA patients, serum VEGF levels are higher than in healthy subjects (91). Moreover, Lin et al. (92) have found increased levels of VEGF in synovial fluid and serum from AS patients with peripheral arthritis, with synovial fluid levels of VEGF significantly higher than serum levels. A recent study has demonstrated VEGF levels significantly higher in AS patients with current and ever smoking, elevated inflammatory indices, including ESR and CRP, and high disease activity and poor patients' functional capacity, measured by BASDAI and Bath AS Functional Index (BASFI) (63). Moreover VEGF levels have been significantly correlated with ESR, CRP, BASDAI, and BASFI (63). Conversely, VEGF levels were not found increased in AS patients with hip involvement and eye involvement (92). A significant longitudinal association between disease activity, measured with AS disease activity score (ASDAS) and BASDAI, and VEGF serum levels have been demonstrated in SpA patients treated with TNF-α inhibitors (93). Moreover, several studies have demonstrated that anti-TNF-α treatment is responsible for decreasing serum VEGF levels in SpA patients (91,94,95). Thus, VEGF has been considered as a potential biomarker for monitoring disease activity and treatment response in SpA patients. Nevertheless, no significant changes in serum levels of VEGF have been found in a study following a large cohort of TNF-α blocker-naïve AS patients for 2 years (94). Thus, further studies are required to clarify utility of VEGF serum levels as clinical biomarker. Moreover, in a recent study serum VEGF levels have been found to be not correlated with radiographic progression and spinal inflammation in AS patients receiving anti-TNF-α (96).

It is also possible that bone formation arises from mechanical stress. A recent study, performed in a mouse model in which chronic and deregulated TNF production leads to arthritis with peripheral and axial joints involvement and extra-articular manifestations, has suggested that mechanical stress drives inflammation and bone formation in entheseal sites (97).

Conclusions

The complete understanding of SpA pathophysiology requires insights into inflammation, bone destruction and bone formation, which are all located in entheses, the primary organ of the disease, and lead all together to ankylosis and functional disability. Despite considerable progress made in understanding cellular and molecular aspects of inflammation and bone destruction in SpA, very less is known concerning the last phenomenon. It prevalently occurs by endochondral ossification, however it remains to identify factors that can induce and influence its initiation and progression. Contrary to what is shown in animal models, the human SpA ossification does not seem to be a simple secondary process to inflammation. Accordingly, several clinical trials have reported good control of clinical and biological inflammation and joint destruction, despite the persistent ossification, in SpA patients treated with biologicals.

Thus, several factors probably play a role in the pathogenesis of bone formation in entheses including not only cytokines but also several systemic factors such as adipokines and gut hormones, and local factors, such as BMP and Wnt signaling, as well as angiogenesis and mechanical stress.

Nevertheless, data available are limited and often conflicting and future experimental and translational studies are needed to better delineate the cause of new bone formation in SpA and to develop new therapeutic approaches, effective in controlling inflammation as well as bone remodeling.

Disclosure statement

All authors state that there are no conflicts of interest.