Inflammatory involvement of tendon, ligament, and joint capsule attachments (enthesitis) is a characteristic, but poorly defined, lesion in ankylosing spondylitis (AS) and the related spondylarthropathies (SpA). The inaugural workshop on enthesitis was a gathering of interested parties, convened in Berlin following recognition of the importance of enthesitis in synovial joint disease in SpA (1); proceedings of that workshop have been reported elsewhere (2). The Second International Enthesitis Workshop, which was held in Leeds March 15–16, 2002, emphasized recent advances in the study of enthesitis, in particular the relationship of biomechanical factors and osteitis to enthesitis. Additionally, the emerging evidence for a unifying anatomic basis for SpA based on the concept of enthesitis was reviewed. The following topics were covered: Anatomical Considerations and New Concepts of the Enthesis; The Role of Joint Biomechanics in Enthesitis; The Relationship Between Enthesitis and Osteitis; The Relationship Between Enthesitis and Synovitis; Mechanisms of Bone Formation in Enthesitis; The Relationship Between HLA–B27 and Enthesitis; An Update on Animal Models of Enthesitis; Therapy of Resistant Enthesitis; and Putative Molecular Mechanisms of Enthesopathy.
The meeting began with a review of some historical aspects of enthesopathy (Dr. M. A. Khan, Cleveland, OH). The adjective “enthetic” was applied in the nineteenth century to diseases “inoculated or implanted into the body from external sources”—particularly infections (3). By the twentieth century, the noun “enthesis” was used to indicate sites of bony attachment of tendons or ligaments, with the term “enthesopathy” being coined to encompass all pathologic conditions affecting them (4).
Histologic studies of enthesitis in SpA, some of which span 4 decades, were reviewed (Dr. D. McGonagle). The importance of enthesitis in AS was recognized (5) shortly before the HLA–B27 disease association was reported (6). However, the discovery of the B27 disease association and explosion in knowledge on immune system function, rather than stimulating research on enthesitis, paradoxically had the opposite effect for almost 3 decades. Although Ball (7) emphasized the importance of enthesitis in SpA, he also considered synovitis to be an equally important and independent pathologic finding. Bywaters (8) made note of enthesitis, but considered that the characteristic lesion of AS was not enthesitis, but subchondral osteitis of the sacroiliac and hip joints. This has subsequently been confirmed by other investigators (9). More recently, François et al (10) have reaffirmed Bywaters' observations that AS pathology represents more than enthesitis. The Second International Enthesitis Workshop sought to establish the relationship between enthesitis and the other diverse pathologies in AS and SpA.
The concept of enthesitis
The conceptual understanding of entheses and enthesitis has changed considerably, with new developments coming from a synthesis of functional anatomic, histologic, and imaging data (Dr. M. Benjamin). There are two types of entheses—a purely fibrous insertion and one containing fibrocartilage (Figure 1). Fibrocartilage is the chief functional adaptation in resisting shear and compressive mechanical stress at entheses (11) (Figure 1). When stress levels are elevated (e.g., in jumper's knee), the quantity of fibrocartilage is increased (12). While enthesitis has been traditionally viewed in relation to focal insertional abnormality (5, 10), it is now clear that certain entheses, e.g., the Achilles tendon insertion, can be viewed as part of an “enthesis organ,” i.e., the enthesis itself is associated with other structures that also reduce stress concentration at the soft–hard tissue interface (13). Thus, in addition to enthesis fibrocartilage at the insertion, there is a further fibrocartilage covering the surface of the bone, immediately adjacent to the osteotendinous junction (called periosteal fibrocartilage because it is a modified periosteum) and on the anterior surface of the tendon (called sesamoid fibrocartilage because it is within the tendon itself—like a sesamoid bone) (11, 14). The periosteal and sesamoid fibrocartilages protect the tendon and bone from compression during dorsiflexion of the foot (11).
Synovial joint disease in SpA is associated with florid extracapsular changes ostensibly not linked to enthesitis (1, 15). These changes cannot be explained in terms of focal insertional abnormality or synovitis, but can be recognized if viewed in the context of an “enthesis organ.” At the interphalangeal joints, the sesamoid fibrocartilage actually forms part of the joint capsule and extensor tendon apparatus (16, 17). Such an enthesis organ is thus also an integral part of the synovial joint and may explain the presence of extracapsular changes, seemingly unrelated to enthesitis or synovitis, in SpA (Figure 2).
Tenosynovitis is a well-recognized feature of SpA, but a relationship with enthesitis has not been shown, although we have suggested that it could emanate from inflammatory changes at the insertion (18). However, the anatomic, biomechanical, and histologic features of the regions where tendons wrap around bony pulleys have striking similarities to those at entheses (19). The fibrocartilages present where tendons change direction around bone could be considered equivalent to the sesamoid and periosteal fibrocartilages in an enthesis organ (13), with these structures being viewed as “functional entheses,” because there is contact with bone but no attachment (13). The effects of these structures on the adjacent bone are discussed below.
Flexor tenosynovitis has been reported as the predominant abnormality in dactylitis (20). It has been suggested that flexor tendon enthesitis at the distal interphalangeal (DIP) joint insertion site was responsible for this, but recent imaging studies failed to show any abnormalities at that location (Dr. I. Olivieri, Potenza, Italy). However, tendons also form “functional entheses” with retinacula or pulleys that hold them in place to prevent them from bowstringing in regions where they change direction. These sites can be recognized by the presence of fibrocartilage (13) (Figure 3). This may explain why tenosynovitis is such a prominent feature of dactylitis in patients with psoriatic arthritis (13), and it should be possible to test this hypothesis with high-resolution imaging.
Pathology and microtrauma at insertions
The basis for disease localization to entheses and related structures in SpA was reviewed. In the course of normal joint function, strain levels at insertion sites may be 4 times those in the tendon midsubstance (21), so microtrauma and degenerative changes at entheses should be anticipated. It is significant that striking parallels between the histopathologic features of entheses (in the Achilles tendon and plantar fascia) and those of synovial joints with early osteoarthritis have been reported (11).
The spine has been described as the “hinterland” of pathology in AS, with much of the spinal disease being enthesitis (8). Apophyseal joint disease has been considered the “essential lesion” that leads to restricted spinal motion prior to spinal fusion in AS (22). Computed tomography studies have confirmed apophyseal joint injury as a significant cause of spinal immobility in both established (23) and early (24) AS. The histologic features of apophyseal joints were reported (Dr. S. Milz, Munich, Germany). Pronounced hypertrophy of both the sesamoid and enthesis fibrocartilages, reflecting increased mechanical stressing, in the dorsal capsule of the apophyseal joints has been noted in patients with degenerative instability of the spine (25). The link between mechanical load, fibrocartilage hypertrophy, and ossification may have parallels with changes that occur in the numerous other spinal ligaments involved in SpA (26). Therefore, findings in the limited studies of the entheses involved in SpA support the concept that these are common sites of tissue microtrauma.
Relationship between enthesitis and osteitis
It is now clear that enthesitis is intimately associated with adjacent osteitis: 50% of patients with plantar fasciitis, for example, have an associated osteitis as determined by magnetic resonance imaging (MRI) (27). It is also evident that osteitis is a feature of mechanically induced plantar fasciitis (27) and mechanically induced enthesopathy in athletes, suggesting common mechanisms of osteitis in SpA and mechanically induced disease (E. Schilders, Bradford, UK). Preliminary sonographic and MRI data have shown that calcaneal osteitis as determined by MRI is invariably associated with sonographically detected swelling at the attachment of the plantar fascia, confirming a close relationship between changes at the insertion and adjacent bone (Dr. A. L. Tan, Leeds, UK).
The osteotendinous junction may play an important role in perientheseal osteitis, with the pathogenesis of osteitis relating to bone microfractures or an abnormal response to bone stress adjacent to the insertions. The osteotendinous junction itself is strong: there is a very elaborate 3-dimensional interlocking of bone and tendon at fibrocartilaginous entheses that maintains its physical integrity (28). Hence, when entheses are under acute trauma, they usually fail within the subchondral bone and not at the bone–tendon/ligament junction itself (Figure 4). This could explain why enthesitis and osteitis are closely linked in SpA.
The importance of joint biomechanics in determining bone microarchitecture was reviewed (Dr. J. Currey, York, UK). One of two patterns of bone microfracture is seen, depending on whether the bone is subject to tensile or compressive loading. When bone is loaded in tension, microcracks that are 2–20 μm in length develop. However, when the bone is subject to compressive loading, cracks of up to 200 μm long develop. When there are more complex patterns of bone stressing (including both tensile and compressive forces), bone strength can be significantly reduced (29). This biomechanical scenario may be present in bone adjacent to entheses and could contribute to damage at these sites. Animal studies have confirmed bone microfractures adjacent to apparently normal insertions (30), but direct evidence for bone microdamage in human SpA has not yet been obtained.
Furthermore, osteitis adjacent to functional entheses (i.e., where there is contact, but no attachment to bone) has been reported in SpA (31). Mechanically related ankle and foot pain is also associated with bone edema seen by MRI in wraparound regions of tendons (32). Like true insertions, these sites are also prone to adjacent periostitis (33). These observations have important implications for understanding the pathogenesis of SpA, because they show that neither entheses nor synovial joints are a sine qua non for bone changes. The pattern of mechanical stressing with shear and compression may induce bone abnormality at such sites without contact (Figures 5–7).
The primary site of pathology in the sacroiliac joints remains a matter of controversy, but the earliest change may be osteitis (34). The fibrocartilaginous sternoclavicular and manubriosternal joints are also prone to enthesitis and osteitis in SpA (35). All these sites, including entheses, functional entheses, and fibrocartilaginous synovial joints, share common anatomic features (i.e., the presence of fibrocartilage) and similar patterns of bone biomechanical stressing. Thus, osteitis here is likely to be related to common anatomic and biomechanical factors (13) (Figures 5–7). In knee disease associated with SpA, osteitis invariably occurs adjacent to fibrocartilaginous entheses rather than articular cartilage (1), but the converse may apply to hip disease (36). Consequently, Khan has proposed that the attachment of articular cartilage to bone should also be viewed as an enthesis (2). The putative mechanisms of osteitis in SpA in humans have not been studied in detail, but tumor necrosis factor α (TNFα) (37) and interleukin-1 contribute to osteoclast activation and may play important roles (Dr. N. Athanasou, Oxford, UK).
In SpA-associated plantar fasciitis, the HLA–B27 gene appears to determine the extent and severity of the condition, but not the susceptibility to osteitis (27). While the role of B27 has traditionally been considered in relation to autoimmunity against a synovial or cartilage antigen, these findings suggest that autoimmunity to a bone antigen may be important in the pathogenesis of spinal disease in AS (27). It is possible that elucidation of the genetic basis for susceptibility to osteitis could help unravel the role of HLA–B27 in disease.
Imaging of enthesitis
Imaging of enthesitis may be critical for an anatomic classification of inflammatory arthritis and could have implications for prognosis. Enthesitis-related changes (including bone erosion, bone formation, periostitis, periarticular sclerosis, and cyst formation) can be imaged by conventional radiography, but this is insensitive in early disease (38). MRI studies demonstrate that enthesitis and extracapsular changes are common adjacent to synovial joints in SpA (15). In the absence of perientheseal osteitis, SpA-associated synovitis cannot be differentiated from rheumatoid arthritis (RA) by MRI.
Recently defined MRI features of arthropathies in SpA were reviewed. The only arthropathy unique to psoriatic arthritis is that in the DIP joint, and the MRI features of DIP joint disease were presented (Dr. H. Marzo-Ortega). The findings at the DIP joints indicate an enthesis- and capsular-based pathology. In AS-associated shoulder disease, the most striking feature was enthesitis with prominent bone changes adjacent to the insertion of the supraspinatus and within the acromion at the origin of the deltoid (Dr. W. Maksymowych, Edmonton, Alberta, Canada). The MRI features of axial disease in ulcerative colitis and Crohn's disease were reported, with sacroiliitis noted in 35% of cases (Professor P. Wordsworth, Oxford, UK). The radiographic, MRI, and histopathologic basis of juvenile SpA with involvement of the foot was reported (Dr. R. Burgos-Vargas, Mexico City, Mexico). Patients with established juvenile SpA had prominent enthesophyte formation (new bone formation at enthesis), bony fusion, and a paucity of inflammatory cells. It is noteworthy that the anatomic and biomechanical features of the midfoot in juvenile SpA, where ankylosis occurs, are reminiscent of those in the sacroiliac joints and entheses. Ankylosing tarsitis may occur in patients with no clinical or radiographic evidence of sacroiliitis.
Ultrasound was first used to evaluate enthesitis almost a decade ago, amidst claims that it is more sensitive than clinical examination for defining enthesitis in the lower limbs (39). The superior sensitivity of sonography over clinical examination for detecting lower limb enthesitis was confirmed (Dr. D. Kane, Glasgow, Scotland, UK). The sonographic pattern of spur formation and erosion described was identical to that described in pathologic studies of the Achilles tendon, with spurs common at the superficial portion of the calcaneal insertion, and erosions more typical of its deeper portion, i.e., in a position corresponding to the site of the periosteal fibrocartilage (11). These observations are best understood in the context of the “enthesis organ.” The use of power Doppler ultrasound for the assessment of peripheral enthesitis in patients with SpA was reported, and a distinctive pattern of vascularization adjacent to entheses was seen in SpA (Dr. M. D'Agostino, Paris, France).
Sonography may also have a role in monitoring of the regression of enthesitis following therapy with infliximab (40). However, recent studies in early RA and SpA suggest that the value of sonographically detected enthesitis in diagnosing arthritis remains unproven (Dr. M. Leirisalo-Repo, Helsinki, Finland). Surprisingly, enthesitis in RA was detected with the same frequency as in SpA, and thus sonography was of little value in discriminating between the diseases. The basis for these findings is unclear, but they are in accordance with MRI studies showing prominent extracapsular changes in a proportion of RA patients (1). The changes in RA may be related to severe inflammation extending across the capsular tissues to involve adjacent entheses (16). Sonography offers a potentially useful, clinically accessible, and inexpensive method for diagnosing and monitoring enthesitis, but further validation is needed.
Animal models of enthesitis
A new model of SpA, induced by immunizing BALB/c mice with the G1 globular domain of versican, which leads to spondylitis and enthesitis (but not peripheral synovitis), was described (Dr. Y. Zhang, San Diego, CA). Versican is present in the annuli fibrosi, sacroiliac joints, ligaments, and tendons, but not in articular cartilage. The features of the mouse disease are reminiscent of AS and suggest that versican autoimmunity may have a role in enthesitis. DBA/1 mice develop a naturally occurring arthritis without overt immunologic abnormalities, which culminates in bone formation leading to joint ankylosis. In this model, entheseal endochondral ossification may depend on bone morphogenetic protein 2 (BMP-2) and BMP-7. The model may have implications for understanding the basis of joint ankylosis, and hence loss of function, in AS (Dr. R. Lories, Leuven, Belgium).
Enthesitis and joint ankylosis
The mechanisms of new bone formation in SpA and the relationship between enthesitis and joint ankylosis have not been fully defined. Immunohistochemical studies in human AS showed increased synovial expression of BMP-4 and transforming growth factor β, suggesting that molecules within the synovium could contribute to joint ankylosis (Dr. T. Duffy, Dublin, Ireland). Histologic studies by Bywaters suggested that bone ankylosis may be independent of enthesitis (8), and this is supported by findings in the ANKENT mouse model of SpA, in which joint fusion without much discernible inflammation has been observed (41). It is noteworthy, too, that diffuse idiopathic skeletal hyperostosis, a disease that can sometimes resemble AS, is not regarded as having an inflammatory component. However, radiographic (42) and recent MRI (43) studies in early human AS indicate that MRI bone changes predate radiographic abnormalities, suggesting that inflammation predates bony fusion. The effects of rapid and sustained suppression of entheseal inflammation on new bone formation remain to be determined (44).
Relationship between enthesitis and synovitis
Synovial cavities can be associated anatomically with both entheses and functional entheses, and it follows that an inflammatory reaction against a component of the fibrocartilage would be expected to trigger synovitis. The putative link between enthesitis and synovitis in diseased synovial joints was reviewed (Professor O. FitzGerald, Dublin, Ireland). T cell repertoire analysis has revealed that synovial CD8 T cells are clonally expanded to a much greater degree than are synovial CD4 T cells (45). Following treatment of psoriatic arthritis with methotrexate, clonal CD8 T cells remain in the tissue. These changes are reminiscent of the predominant CD8 T cell infiltration adjacent to entheses in chronic SpA (46). Taken together, they could represent immunity to common antigens shared by the synovium and enthesis. Another potential mechanism linking synovitis and enthesitis includes the release of proinflammatory cytokines from entheses and related structures, triggering a secondary synovitis (18).
Treatment of enthesitis
Historically, the therapeutic options for enthesitis in SpA have been limited. One of the most significant developments in recent years has been treatment with TNFα blockade. The majority of enthesitis lesions identified on MRI either regressed completely or improved following biologic blockade with etanercept (47). Similar results have been reported with infliximab, with impressive clinical and MRI improvement in indices of enthesopathy (Professor J. Braun, Herne, Germany). (48). However, continuous treatment is required, because disease flares occur after cessation of therapy, both with etanercept and with infliximab.
The bisphosphonates appear to have a role in the therapy of osteitis (Dr. W. Maksymowych). Intravenous pamidronate treatment has been associated with symptomatic improvement and regression of periarticular osteitis as determined by MRI, in an open-label study (36). A randomized controlled trial of low versus therapeutic doses of pamidronate in AS showed efficacy with therapeutic doses (49). Further MRI studies are needed to investigate the basis for the significant declines in C-reactive protein levels with biologic agents, but not with pamidronate, in parallel with improvement in osteitis.
Molecular mechanisms pertaining to immune activation in enthesitis
A number of experimental models and clinical studies indicate that autoimmunity directed against fibrocartilage proteins/proteoglycans, including aggrecan, can induce enthesitis and spondylitis (50–53). The strong HLA–B27 association with AS indicates that autoimmunity is important in human AS, especially at sites of fibrocartilage (31). Using T cell epitope identification techniques in humans, it is possible to show CD4+ and strong CD8+ T cell reactivity against peptides derived from aggrecan (Dr. J. Sieper, Berlin, Germany).
Fibrocartilage is an avascular structure, so autoimmunity against it should be preceded by the initial activation of the innate immune system at entheses. The alternative hypothesis is that the localization of SpA to fibrocartilaginous entheses and related sites suggests an “injury model” of disease, because the presence of fibrocartilage is an adaptive response to mechanical load (shear and/or compression). A “response-to-injury” model to describe immunologic aspects of other diseases, especially atherosclerosis, is well established (54). One potential mechanism of immune activation at sites of mechanical stress and cell injury is up-regulation of heat-shock protein 60 (55) and other “danger signals” for immune activation (56, 57). A “multi-hit” model for immune activation in SpA has been proposed, in which endogenous danger signals, combined with those from exogenous bacteria or bacterial molecules (including lipopolysaccharide), lead to immune activation at entheses and related sites (58). However, the molecular basis for enthesitis remains to be defined. Recently, microarray analysis has provided clues to the genes involved in synovial disease in SpA and in assessing responses to therapy (59). Although no studies were completed at the time of the meeting, it was envisaged that this approach could yield important insights into the transcript pattern at the enthesis in healthy individuals and in those with SpA (Dr. D. T. Yu, Los Angeles, CA).
Conclusions and future directions
Observations stemming from the anatomic similarity between entheses and other targeted sites suggest that the basis for disease localization in SpA involves common underlying biomechanical factors at all disease-prone sites. The consensus view of the meeting attendees was that SpA represents either a disorder relating to fibrocartilage autoimmunity at entheses, functional entheses, and fibrocartilaginous synovial joints (31), or one in which there is a type of “stress arthritis” with subsequent immunity against common fibrocartilage antigens at these sites (60). In fact, a biomechanical basis for disease has been suggested previously, with Bywaters concluding that pathology had little extra to offer in terms of elucidating AS and SpA pathogenesis (8). Based on observations not only of skeletal changes, but also of ascending aortitis in AS, he suggested that the role of biomechanical factors as a unifying basis for disease localization merited consideration (8). The increased understanding of enthesitis strongly supports that assertion. Future research directions that emerged from the conference include further studies into the anatomic and biomechanical basis for disease, additional imaging studies to define the full spectrum of disease, studies exploring the molecular basis of inflammation at insertion sites, molecular studies of the HLA–B27 gene in relation to osteitis, and studies of the effect of inflammation suppression on the prevention of entheseal ankylosis.