Histopathologic Changes at "Synovio-Entheseal Complexes"

Home , Synovial membrane

Histopathologic Changes at “Synovio–Entheseal Complexes” Suggesting a Novel Mechanism for Synovitis in Osteoarthritis and Spondylarthritis

Michael Benjamin1 and Dennis McGonagle2

Objective. To determine the extent to which dif- mechanism by which synovitis could develop in both ferent entheses form part of a “synovio–entheseal com- degenerative joint disease and spondylarthritis. plex” (SEC) and whether such SECs are commonly associated with the presence of inflammatory cells and The central importance of the enthesis in under- evidence of enthesis microdamage. standing the pathophysiology of inflammation in the Methods. Specimens from 49 cadaveric entheses spondylarthritides (SpA) has reemerged during the last were processed for histologic study, and all soft tissue decade (1). It has been aided by the development of the components of the entheses or enthesis organs were “enthesis organ concept” (2), which explains why tissues examined. To exclude articular cartilage degeneration adjacent to the enthesis itself are also subject to patho- as a triggering factor for synovitis, the selected entheses logic change. While the link between enthesitis and included 17 that were not immediately adjacent to such osteitis in SpA has been clarified in recent studies that cartilage. demonstrate a close functional integration of the enthe- Results. An SEC was present at 82% of entheses. sis with neighboring bone (3), the connection between These included 47% of the attachments not adjacent to enthesitis and synovitis (the other cardinal manifestation articular cartilage, where the synovium was that of of SpA) remains a subject of debate, and synovial bursae or tendon sheaths. One or more of a wide variety inflammation has consistently been viewed as being of degenerative changes were noted on the soft tissue independent of the inflammation occurring at entheses side of every enthesis; the most common changes were (4). cell clustering and/or fissuring (in 76% of entheses). In addition to the obvious relationship between Synovial villus formation or inflammatory cell infiltra- the enthesis and SpA, there may be a link between the tion was seen in 85% of entheses, and in 73% of enthesis organ and the development of generalized attachments there were also inflammatory cells in the nodal osteoarthritis (OA), since enthesis and ligament enthesis organ itself. The changes included synovial abnormalities as determined by magnetic resonance invasion (pannus formation) of the enthesis. imaging (MRI) are common in early disease (5,6). This Conclusion. Entheses are frequently juxtaposed observation is further supported by MRI studies showing to synovium, thus forming SECs. They are also often similar patterns of involvement of entheses in early associated with both degenerative and inflammatory psoriatic arthritis and generalized nodal OA of the distal changes, and the latter may involve the immediately interphalangeal joints, although inflammation was much adjacent synovium. These findings suggest a novel more severe in the former (7). Preliminary histologic studies of involved structures suggest that these entheseal regions are common sites of microdamage (6). 1Michael Benjamin, PhD, Cardiff University, Cardiff, UK; Furthermore, MRI studies have occasionally shown that 2Dennis McGonagle, FRCPI, PhD, University of Leeds, Leeds, UK, prominent synovitis can occur in early generalized nodal and Calderdale Royal Hospital, Halifax, UK. Address correspondence and reprint requests to Michael OA, at a stage when the articular cartilage appears Benjamin, PhD, Professor of Musculoskeletal Biology and Sports relatively normal (5). This suggests that some other joint Medicine Research, Cardiff University, Museum Avenue, Cardiff structure(s) could be contributing to synovitis. The fact CF10 3US, UK. E-mail: [email protected]. Submitted for publication March 29, 2007; accepted in revised that synovitis may occur in retrocalcaneal bursae and at form August 13, 2007. other entheseal sites devoid of articular cartilage (8)

3601 3602 BENJAMIN AND McGONAGLE

raises the possibility that pathologic changes at the we have proposed that entheses form a functional unit enthesis could be a contributing factor in joint synovitis. with adjacent synovium, termed the “synovio–entheseal Collectively, these findings suggest that tissue complex” (SEC) (9). While SECs could be important in microdamage or aberrant repair responses at entheses fibrocartilage homeostasis under normal conditions, it is may contribute to synovitis adjacent to attachment sites possible that they are detrimental in the setting of in a variety of diseases or types of injury. Consequently, enthesis microdamage and contribute to the propensity

Table 1. Anatomic sites of the entheses examined Presence or absence of a synovial– Adjacent to entheseal articular Enthesis complex cartilage 1. Abductor pollicis brevis insertion ϩ Yes 2. Abductor pollicis longus insertion ϩ Yes 3. Achilles tendon insertion ϩ No 4. Adductor longus origin Ϫ No 5. Adductor magnus insertion (hamstring part) Ϫ No 6. Annulus fibrosus (L4–L5) Ϫ No 7. Anterior cruciate ligament (tibial attachment) ϩ Yes 8. Anterior cruciate ligament (femoral attachment) ϩ Yes 9. Common tendon of biceps brachii and coracobrachialis ϩ No 10. Biceps brachii insertion ϩ No 11. Biceps femoris insertion ϩ Yes 12. Collateral ligaments of the interphalangeal joints of fingers ϩ Yes 13. Common extensor muscle origin ϩ Yes 14. Common flexor muscle origin ϩ Yes 15. Extensor carpi radialis brevis insertion ϩ Yes 16. Extensor carpi radialis longus insertion ϩ Yes 17. Extensor carpi ulnaris insertion ϩ Yes 18. Extensor digitorum ϩ Yes 19. Extensor hallucis longus insertion ϩ Yes 20. Extensor pollicis longus insertion ϩ Yes 21. Facet joint capsule ϩ Yes 22. Flexor carpi ulnaris insertion plus origin of hypothenar muscles ϩ Yes 23. Flexor digitorum profundus insertion ϩ Yes 24. Flexor hallucis longus insertion ϩ Yes 25. Flexor pollicis longus insertion ϩ Yes 26. Gastrocnemius insertion (lateral head) ϩ Yes 27. Gluteus medius insertion ϩ No 28. Gluteus minimus insertion ϩ No 29. Gracilis insertion Ϫ No 30. Iliopsoas insertion Ϫ No 31. Interspinous ligament (L5–S1) Ϫ No 32. Medial collateral ligament of the 1st metatarsophalangeal joint ϩ Yes 33. Medial collateral ligament of the knee (femoral attachment) ϩ Yes 34. Obturator internus insertion ϩ No 35. Patellar tendon insertion ϩ No 36. Patellar tendon origin ϩ Yes 37. Peroneus brevis insertion/capsule of tarsometatarsal joint ϩ Yes 38. Peroneus longus insertion ϩ Yes 39. Pes anserinus insertion ϩ No 40. Popliteus tendon origin/lateral collateral ligament ϩ Yes 41. Posterior cruciate ligament (tibial attachment) ϩ Yes 42. Pronator teres insertion Ϫ No 43. Quadriceps tendon insertion ϩ Yes 44. Sacroiliac joint (interosseous ligament) Ϫ Yes 45. Sartorius (origin) Ϫ No 46. Subscapularis insertion ϩ No 47. Supraspinatus insertion ϩ Yes 48. Tibialis anterior insertion ϩ Yes 49. Triceps brachii insertion ϩ Yes “SYNOVIO–ENTHESEAL COMPLEXES” IN OA AND SpA 3603

for synovial inflammation in a wide range of joint 17 sites not immediately adjacent to such cartilage diseases and injuries. To explore this issue further, we (47%). In these instances, this was synovium lining investigated 1) the extent to which different entheses can either a bursa or a tendon sheath (Figures 1a and b). form part of an SEC, 2) whether inflammatory cells can Thus, at the trochanteric insertion of the gluteus mini- be found at “normal” SECs, and 3) whether SECs are mus (which is a considerable distance from the hip joint frequently associated with evidence of soft tissue micro- cartilage), for example, the synovium contributing to the damage. We addressed these issues by examining enthe- SEC lined bursae that were both deep and superficial to ses from 49 different anatomic sites in the limbs and the tendon near its attachment site (Figure 1a). The spine of elderly, dissecting-room cadavers. Achilles tendon enthesis was a further clear example in which the synovium of the SEC was remote from any articular cartilage. The synovium covered the tip of MATERIALS AND METHODS Kager’s fat-pad as the latter protruded into the bursa, Specimens were collected from 60 cadavers (35 male, thus bringing it very close to the enthesis itself. 25 female; mean age 84 years [range 49–101]) donated to It is important to recognize that although SECs Cardiff University for anatomic investigation under the provi- could indeed be distant from articular cartilage, they sion of the 1984 Anatomy Act and the 1961 Human Tissue Act. could nevertheless still be associated with synovium that The cadavers had been perfused with an embalming fluid containing 4% formaldehyde and 25% alcohol. They were was continuous with that of the joint. This is because selected according to the quality of preservation and the some bursae are extensions of the joint cavity that absence of grossly visible abnormalities in the region of protrude a considerable distance from the joint. Thus, at interest, although access to medical histories was not possible. the insertion of the subscapularis on the lesser tuberosity The 49 anatomic sites examined histologically are listed in of the humerus, the synovium of the SEC originated Table 1 and included 19 entheses from the upper limb, 26 from from a neighboring bursa and from the biceps tendon the lower limb, and 4 from the spine. Up to 10 specimens were examined from each site. In order to exclude articular cartilage sheath, both of which communicate directly with the degeneration as a triggering factor in synovitis, the 49 selected glenohumeral joint cavity. SECs adjacent to articular entheses included 17 that were not immediately adjacent to cartilage were most typical of areas in which a tendon/ such cartilage. ligament replaced or reinforced part of the joint capsule, Details of the histologic processing procedures and of thus bringing the synovium of the joint into immediate the scoring system used to assess enthesis microdamage and proximity with the enthesis (e.g., the digital tendons). synovial changes are similar to those described previously (3). Briefly, formalin-fixed tissue was embedded in paraffin wax, However, they also included entheses where the tendon and longitudinal 8-␮m sections were stained with Masson’s or ligament fused with the joint capsule only at its trichrome, toluidine blue, or Hall and Brunt’s quadruple stain enthesis (Figure 1c). (10). Sections were examined at 1-mm intervals throughout the Histopathologic abnormalities in SECs. A wide blocks, and a positive finding of a feature was recorded variety of histopathologic features were noted in the whenever it was present in 1 or more specimens. In order to avoid the possibility of confusing inflammatory infiltrates with SECs, either at the attachment site itself or in other populations of bone marrow cells that invaded the soft tissue parts of enthesis organs. At least 1 histopathologic side of the enthesis via holes in the subchondral bone plate (3), abnormality was seen at every anatomic site examined cell infiltrates immediately next to subchondral plate defects (though not necessarily in every specimen). Although that appeared identical to the adjacent marrow itself were degenerative and inflammatory changes at entheses excluded from the evaluation. Synovial tissue that was associated with invasion and destruction of fibrocartilage was termed (both at the attachment sites themselves and in the “pannus.” adjacent synovium) may be intimately linked, they are described separately below, in the interest of clarity. Degenerative changes at entheses (Figure 2). De- RESULTS generative changes were particularly striking at large Normal composition of SECs. At 82% of attach- entheses (e.g., the Achilles, quadriceps, adductor longus, ment sites in the histologic sections, a synovial mem- and patellar tendons) and included intratendinous cal- brane was seen very close to the enthesis, thus contrib- cification, fissuring, cell clustering, fibrocartilage cell uting to the formation of an SEC (Table 1 and Figures hypertrophy, hyper- and hypocellularity, fibrillation and 1a–c). As expected, this was evident in entheses adjacent delamination of sesamoid and periosteal fibrocartilage, to articular cartilage, where the synovium was often part necrosis, and cyst formation. Among the most wide- of the joint itself (Figure 1c). However, synovium that spread histopathologic features that were suggestive of was not part of a synovial joint was also present at 8 of microtrauma and subsequent degeneration were clusters 3604 BENJAMIN AND McGONAGLE

Figure 2. Histopathologic analyses of the synovial–entheseal complex Figure 1. Low-magnification views of 3 different sites at which a demonstrating degenerative changes at entheses resembling those seen synovial–entheseal complex is evident. Each panel shows the position in articular cartilage in osteoarthritis. a, Fissures (arrows) in the zone of the synovium (S) relative to the enthesis (E). a, Insertion of the of uncalcified fibrocartilage at the insertion of the quadriceps tendon. tendon of the gluteus minimus (GMi). The synovium lines 2 adjacent The fissures contain necrotic material (asterisk) and are associated bursae (B). GMe ϭ gluteus medius. Inset, Higher-magnification view with numerous fibrocartilage cell clusters (C). b, Fissuring (F) and cell of the deep bursa. V ϭ synovial villus. b, Fibular insertion site of the cluster formation in the periosteal fibrocartilage (PF) associated with tendon of the biceps femoris (BF). Synovium is associated both with the insertion of the biceps brachii, in conjunction with surface fibril- the lateral collateral ligament (LCL) that passes through a bifurcation lation and delamination of tissue fragments (arrows) into the adjacent in the tendon near its enthesis and with an adjacent bursa. c, Insertion bursa (B). c, Fissuring of a sesamoid fibrocartilage (SF) at the insertion of the flexor hallucis longus tendon (FHL). The synovium is both that of the biceps brachii tendon. Note the presence of detached tissue of the interphalangeal joint (J) and that of a tendon sheath (arrow) fragments (arrows) in the bursa and the extensive invasion of the extending beneath the bony volar plate (VP). All sections were stained periosteal fibrocartilage by blood vessels (BV). All sections were stained with Masson’s trichrome. Bars ϭ 3 mm; bar in inset ϭ 100 ␮m. with Masson’s trichrome. Bar in a ϭ 200 ␮m; bars in b and c ϭ 400 ␮m. “SYNOVIO–ENTHESEAL COMPLEXES” IN OA AND SpA 3605

of hypertrophied fibrocartilage cells and matrix fissuring (Figure 2a). However, necrotic areas at the enthesis were most typically associated with inflammatory cells, whereas cell clustering and fissuring were not. Many of the degenerative changes described above represent typical features of degenerative arthritis, which is more commonly recognized in relation to articular cartilage. The different degenerative changes were often seen in association with one another, and clusters and/or fissures were typical of 76% of entheses examined. The fissures could be empty or filled with degeneration products of the extracellular matrix, and large numbers of cells were seen in sections of the clusters (Figure 2a). Of note, there were marked degenerative changes at the enthesis of the annulus fibrosus from the lumbar region—consistent with the fact that this is a common site of clinically discernible spondylosis. Particularly elongated fissures were observed at the insertion of the biceps brachii tendon and the origin of adductor longus, extending several millimeters away from the enthesis itself. Degenerative changes were found not only at the enthesis itself, but also at associated sesamoid and/or periosteal fibrocartilages, e.g., at the insertions of the biceps brachii, tibialis anterior, and extensor pollicis longus tendons. In the “accessory” fibrocartilages of these enthesis organs, there was also frequent evidence of fibrillation and fissuring, cell cluster formation, and detachment of tissue fragments into the bursae or joint cavities (Figures 2b and c). On rare occasions, large rounded or oval cysts (maximum 2–4 mm), rather than elongated fissures, were found on the soft tissue side of entheses. Such cysts were filled with necrotic material. Inflammatory changes in the synovial component of the SEC (Figure 3). There was evidence of synovial changes in 85% of SECs, including both SECs adjacent to and SECs distant from articular cartilage. The abnormalities included inflammatory cell infiltration, lining cell hyperplasia, and formation of synovial villi (Figures 3a–c). We investigated whether synovial changes were maximal adjacent to areas where synovial membrane Figure 3. Histopathologic analyses of the synovial–entheseal complex demonstrating inflammatory changes within the synovium itself. a, came in close proximity to enthesis fibrocartilage. Of Local concentration of mononuclear cells associated with small blood note, at the Achilles tendon insertion, where evidence of vessels (BV) within a synovial villus at the tibial enthesis of the anterior mild villus formation was seen in 57% of the specimens cruciate ligament. Most of the cells are lymphocytes (L), but occasional examined, villi were most typical of the tip of Kager’s putative macrophages (M) are present. b, Extensive formation of synovial villi (arrows) at the common extensor origin (CEO)ofthe fat-pad, i.e., of the region closest to the enthesis. Syno- forearm muscles. c, Synovial villi (arrows) that are restricted to the tip vial changes (typically mild villus formation) without of Kager’s fat-pad (KFP) in association with the Achilles tendon obvious local soft tissue damage to the enthesis were enthesis, i.e., the villi are in the region nearest to the enthesis itself. B ϭ bursa; PF ϭ periosteal fibrocartilage; SF ϭ sesamoid fibrocarti- occasionally observed, e.g., in the collateral ligaments of lage. All sections were stained with Masson’s trichrome. Bar in a ϭ 20 the interphalangeal joints. ␮m; bars in b and c ϭ 300 ␮m. 3606 BENJAMIN AND McGONAGLE

Figure 4. Histopathologic analyses of the synovial–entheseal complex demonstrating inflammatory cells at the enthesis itself. a, Collection of lymphocytes (L) and putative macrophages (M) in the zone of dense fibrous connective tissue at the tibial attachment of the posterior cruciate Figure 5. Histopathologic analyses of the synovial–entheseal complex ligament. b, Macrophages (arrows) at the insertion of the adductor longus demonstrating synovial membrane invasion of the enthesis. a, Blood tendon. c, Inflammatory cells at the enthesis of the tendon of the tibialis vessels in the adipose synovium covering the tip of Kager’s fat-pad anterior. The cells are located in the zone of dense fibrous connective (KFP) have invaded the deep part of the Achilles tendon (arrows). tissue (DT) and end abruptly (arrows) at the junction of that zone with the SV ϭ synovial villi; TF ϭ tendon fascicles. b, The synovial membrane enthesis fibrocartilage (FC). Inset, Higher-magnification view of the (S) itself has invaded the enthesis of the tibial attachment of the inflammatory cells. d, Collection of lymphocytes, associated with blood anterior cruciate ligament. LF ϭ adjacent ligament fascicles. Arrow vessels (BV), in an area of collagen disruption (asterisk) immediately next indicates a group of inflammatory cells. c, Synovial membrane has to the bone (B) at the insertion of the tendon of the triceps brachii. All eroded the enthesis fibrocartilage (EF) of gluteus minimus (arrow). sections were stained with Masson’s trichrome. Bars in a, b, and inset ϭ This is interpreted to be pannus. All sections were stained with 30 ␮m; bar in c ϭ 400 ␮m; bar in d ϭ 100 ␮m. Masson’s trichrome. Bars ϭ 400 ␮m. “SYNOVIO–ENTHESEAL COMPLEXES” IN OA AND SpA 3607

Inflammatory cells in the enthesis organ (Figures 4 joint (Figure 5c). However, it is important to note that and 5). Small numbers of inflammatory cells (typically inflammatory cells were still evident in 8 of 10 entheses Ͻ10), frequently accompanied by numerous, dilated without SECs, supporting the impression that the in- venules often packed with blood cells, were seen in 73% flammatory infiltrates seen at other sites were not just a of attachment sites. Large collections of inflammatory secondary consequence of synovitis. cells (i.e., with many hundreds of cells) were rare and always associated with marked necrosis at the enthesis. DISCUSSION The predominant cell types in most inflammatory infiltrates (both enthesis- and synovium-related) were The results of this study support the notion that lymphocytes (Figure 4a), but putative macrophages were SECs are common at tendon or ligament attachment also present (Figures 4a and b). Neutrophils were scarce. sites and that degenerative and mild inflammatory Inflammatory infiltrates in the zone of dense fibrous change on the soft tissue side of entheses is a feature of connective tissue (e.g., those seen in Figure 4a) could be attachment sites in elderly cadavers. We thus suggest present several millimeters away from the attachment that SECs could be important in the pathogenesis of site itself, but did not extend into the zone of uncalcified synovitis in both SpA and degenerative arthritis. In the fibrocartilage. Indeed, at several sites (biceps brachii, latter, a link between articular cartilage degeneration iliopsoas, supraspinatus, and tibialis anterior), an inflam- and synovitis is well established (11). We now propose matory cell infiltrate in the zone of dense fibrous that there is also a link between enthesis degeneration connective tissue largely ended abruptly near the zone of and inflammatory changes at or near attachment sites, uncalcified fibrocartilage, so that the latter region was probably reflecting high levels of mechanical stress and virtually devoid of such cells in those particular speci- “wear and tear” at entheses. This concept could be mens (Figure 4c). It should be noted, however, that important in understanding the early stages of general- inflammatory cells were seen within the zone of uncalci- ized nodal OA, in which articular cartilage may appear fied fibrocartilage in other specimens (e.g., triceps to be normal (11,12). brachii), invariably in association with necrosis in that We proposed the SEC concept not only to ex- zone and very close to the bone (Figure 4d). plain the basis for diarthrodial joint synovitis in SpA, but Inflammatory cells within the zone of dense also to suggest an additional possible mechanism that fibrous connective tissue lay in the endotenon or endo- could contribute to synovitis in generalized nodal OA. ligament (i.e., the loose connective tissue sheaths asso- The present findings strongly support the notion of close ciated with the fascicles themselves that allow the fasci- integration and functional interdependence of synovium cles to move independently of one another). Occasional and enthesis in humans, since 82% of specimens had inflammatory infiltrates and (more frequently) numer- entheses that were close to synovia. Furthermore, the ous dilated blood vessels were also seen in the sesamoid association with synovium was observed in nearly half and/or periosteal fibrocartilage, and this caused local the sites without articular cartilage—reflecting the com- disruption of their normal structure (Figure 2c). Inflam- mon occurrence of bursae at entheses that are juxta- matory infiltrates were also present in adipose tissue posed to synovial joints (12), or the presence of neigh- adjacent to entheses, e.g., Hoffa’s fat-pad at the origin of boring synovial tendon sheaths. Finally, it should be the patellar tendon and the insertional angle fat of the acknowledged that SECs could also be crucial in re- flexor hallucis longus tendon. gional mechanical disorders, including trochanteric bur- At some sites, it was evident from the proximity sitis, retrocalcaneal bursitis, and pes anserine bursitis. of neighboring synovium that inflammatory cells and/or Abnormal mechanical loads are likely to result in dam- blood vessels had indeed migrated into the enthesis from age to one or more components of many different the synovium, as at the insertion of the subscapularis and enthesis organs. However, whether synovitis itself or Achilles tendons. At the Achilles tendon, the invading degenerative change at the enthesis is the primary factor blood vessels originated from the adipose synovium that triggers bursitis is unclear. covering the tip of Kager’s fat-pad (Figure 5a). Further- A striking finding in the current study was the more, synovial membrane itself invaded the tibial attach- large number of sites at which there was evidence of ment of the anterior cruciate ligament and the insertions changes at entheses mirroring those seen in articular of the gluteus minimus and tibialis anterior tendons cartilage in OA, i.e., fibrocartilage cell clusters, cell (Figures 5b and c). The invasive nature of this tissue was hypertrophy, fissuring, fibrillation, and delamination. characteristic of pannus in the milieu of an inflamed Such changes at certain individual entheses have been 3608 BENJAMIN AND McGONAGLE

reported previously (13–15), but their occurrence at such stress in the dynamically functioning enthesis remains to a large number of attachment sites from a wide variety be determined. It must also be pointed out that focal of locations highlights just how typical they are of areas of microscopic inflammation in normal synovium entheses in general: 76% of entheses from elderly, have been reported (20), although data on the proximity dissecting-room cadavers showed evidence of cell clus- of the sampled tissue to adjacent entheses was not tering and/or fissuring. The presence of these features presented. reveals clear parallels between the nature of the degen- The tissue-specific factors that regulate immune erative changes affecting articular cartilage and those activation, or lack thereof, at enthesis organs could be an affecting entheses. This is pertinent to the issue that important consideration in future studies of SpA (21). early events responsible for triggering some types of OA Although we use the term “synovitis” in the context of may occur in adjacent ligaments, rather than in articular these diseases, it is likely that the inflammatory cells cartilage or subchondral bone (6,16). As in articular noted at “normal” entheses in the current study are cartilage, fissuring is likely to reflect mechanical damage involved in tissue repair. Indeed, it is now well recog- to collagen, which is followed by local loss of proteogly- nized that under certain circumstances, macrophages cans. Some fissures extended several millimeters away can adopt an alternatively activated phenotype that is from the attachment site, and this together with degen- antiinflammatory and associated with repair responses erative changes seen in some sesamoid and periosteal (22). We do not mean to imply, however, that inflam- fibrocartilages emphasizes the importance of consider- matory cells appearing in association with synovium ing the whole “enthesis organ” (2,17) when addressing necessarily migrate to the enthesis in order to promote enthesopathies. It remains to be determined why gener- tissue repair. Most of the inflammatory cells seen at the alized nodal OA has a certain characteristic clinical entheses examined probably do not derive from syno- pattern, even though entheseal degenerative changes are vium at all; the most common source (judging from their ubiquitous. anatomic proximity) is likely to be the bone marrow, The finding of immune cells associated with although some may also derive from blood vessels in the necrotic areas within the zone of enthesis fibrocartilage, endotenon or endoligament. This explains the presence adjacent connective tissue, and synovium is intriguing of inflammatory changes occurring at entheses that have and indicates that the presence of these cells at subclin- no neighboring synovium at all. The marrow-derived ical levels is more common than generally supposed, at inflammatory cells are likely to be associated with the least in elderly people. The presence of pannus tissue previously described local repair of damaged fibrocarti- invading entheses also provides strong evidence against lage (3). The finding of inflammatory changes at enthe- the idea that pannus is a specific feature of rheumatoid ses without neighboring synovium supports the notion arthritis–related periarticular joint destruction (18). that entheseal microdamage and inflammatory cell infil- These inflammatory changes were observed at entheses tration contribute to inflammation in the immediately that did not border articular cartilage, thus excluding the adjacent synovium in many instances. possibility that they were merely secondary to OA A limitation of this study was the use of elderly, elsewhere in the joint. Synovitis was more obvious dissecting-room cadavers for which information on med- immediately adjacent to the insertion of the Achilles ical history was not available, and we fully recognize that tendon, as suggested by the number of synovial villi at the extent to which the present histopathologic findings the tip of Kager’s fat-pad. In some specimens, this was in older subjects are relevant to younger subjects prone associated with the presence of invading blood vessels to SpA is unclear. However, it is impossible to obtain from the fat-pad, accompanied by entheseal damage. such a comprehensive collection of entheses from young Inflammatory cell invasion of entheses in the zone of SpA patients, and the occasional entheses that may dense fibrous connective tissue tended to be in the plane become available (e.g., discarded surgical waste from hip of the loose endotenon (or endoligament) rather than replacements) inevitably represent late-stage disease. the tendon or ligament fascicles themselves. This is in Thus, work on elderly cadaveric material is still a accordance with the view that tendons may have an valuable way of establishing novel anatomic and his- intrinsic ability to resist inflammatory cell invasion from topathologic principles that provide greater insight into the synovium (19). the fundamental nature of enthesopathies. The use of On occasion, synovitis adjacent to insertions animal models of SpA overcomes the “age factor,” but is without ostensible soft tissue damage was noted, but subject to the criticism that the findings may not be whether this is directly related to synovial mechanical extrapolatable to humans. We do believe, however, that “SYNOVIO–ENTHESEAL COMPLEXES” IN OA AND SpA 3609

our observations could be directly relevant to older synovium of patients with psoriatic arthritis: implications for subjects with joint symptoms due to degenerative dis- treatment. Ann Rheum Dis 2006;65:1551–7. 5. Tan AL, Grainger AJ, Tanner SF, Shelley DM, Pease C, Emery P, ease, and that some of the symptoms could be emanating et al. High-resolution magnetic resonance imaging for the assess- from the SEC. Furthermore, the frequency of degener- ment of hand osteoarthritis. Arthritis Rheum 2005;52:2355–65. ative and inflammatory changes that we observed indi- 6. Tan AL, Toumi H, Benjamin M, Grainger AJ, Tanner SF, Emery P, et al. Combined high-resolution magnetic resonance imaging cates that they are likely a normal, age-related phenom- and histological examination to explore the role of ligaments and enon. However, generalized nodal OA is also a disease tendons in the phenotypic expression of early hand osteoarthritis. that becomes increasingly prevalent with age. Ann Rheum Dis 2006;65:1267–72. 7. Tan AL, Grainger AJ, Tanner SF, Emery P, McGonagle D. A In conclusion, we have shown that the concept of high-resolution magnetic resonance imaging study of distal inter- a “synovio–entheseal complex” is widely applicable at phalangeal joint arthropathy in psoriatic arthritis and osteoarthri- many sites in the body. Furthermore, soft tissue micro- tis: are they the same? Arthritis Rheum 2006;54:1328–33. damage, which is widespread at enthesis organs, is often 8. Olivieri I, Barozzi L, Padula A, De Matteis M, Pierro A, Cantini F, et al. Retrocalcaneal bursitis in spondyloarthropathy: assessment accompanied by synovial changes and the presence of by ultrasonography and magnetic resonance imaging. J Rheumatol immune cells within the enthesis. These findings support 1998;25:1352–7. the idea that biomechanical factors related to the enthe- 9. McGonagle D, Lories RJ, Tan AL, Benjamin M. The concept of a “synovio-entheseal complex” and its implications for understand- sis and the wider SEC could play an important role in ing joint inflammation and damage in psoriatic arthritis and the genesis of synovial inflammation in both degenera- beyond. Arthritis Rheum 2007;56:2482–91. tive and inflammatory arthritis in humans. 10. Hall BK. The role of movement and tissue interactions in the development and growth of bone and secondary cartilage in the clavicle of the embryonic chick. J Embryol Exp Morphol 1986;93: ACKNOWLEDGMENTS 133–52. 11. Loeuille D, Chary-Valckenaere I, Champigneulle J, Rat AC, We thank S. Redman and K. Hayashi for cutting and Toussaint F, Pinzano-Watrin A, et al. Macroscopic and micro- staining the sections and D. Scarborough for processing the scopic features of synovial membrane inflammation in the osteo- arthritic knee: correlating magnetic resonance imaging findings histologic material. with disease severity. Arthritis Rheum 2005;52:3492–501. 12. Standring S, editor. Gray’s anatomy: the anatomical basis of clinical practice. 38th ed. Oxford: Churchill Livingstone; 1995. AUTHOR CONTRIBUTIONS 13. Rufai A, Ralphs JR, Benjamin M. Structure and histopathology of Professor Benjamin had full access to all of the data in the the insertional region of the human Achilles tendon. J Orthop Res study and takes responsibility for the integrity of the data and the 1995;13:585–93. accuracy of the data analysis. 14. Moriggl B, Kumai T, Milz S, Benjamin M. The structure and Study design. Benjamin, McGonagle. histopathology of the “enthesis organ” at the navicular insertion of Acquisition of data. Benjamin, McGonagle. the tendon of tibialis posterior. J Rheumatol 2003;30:508–17. Analysis and interpretation of data. Benjamin, McGonagle. 15. Kumai T, Benjamin M. Heel spur formation and the subcalcaneal Manuscript preparation. Benjamin, McGonagle. enthesis of the plantar fascia. J Rheumatol 2002;29:1957–64. Statistical analysis. Benjamin, McGonagle. 16. Brandt KD, Radin EL, Dieppe PA, van de Putte L. Yet more evidence that osteoarthritis is not a cartilage disease. Ann Rheum Dis 2006;65:1261–4. REFERENCES 17. Benjamin M, McGonagle D. The anatomical basis for disease localisation in seronegative spondyloarthropathy at entheses and 1. Ritchlin CT. Therapies for psoriatic enthesopathy: a systematic related sites. J Anat 2001;199:503–26. review. J Rheumatol 2006;33:1435–8. 18. Rhodes LA, Conaghan PG, Radjenovic A, Grainger AJ, Emery P, 2. Benjamin M, Moriggl B, Brenner E, Emery P, McGonagle D, McGonagle D. Further evidence that a cartilage-pannus junction Redman S. The “enthesis organ” concept: why enthesopathies may synovitis predilection is not a specific feature of rheumatoid not present as focal insertional disorders. Arthritis Rheum 2004; arthritis. Ann Rheum Dis 2005;64:1347–9. 50:3306–13. 19. Aspenberg P. Is inflammation harmless to loaded tendons? J Appl 3. Benjamin M, Toumi H, Suzuki D, Redman S, Emery P, McGo- Physiol 2007;102:3–4. nagle D. Microdamage and altered vascularity at the 20. Singh JA, Arayssi T, Duray P, Schumacher HR. Immunohisto- enthesis–bone interface provides an anatomic explanation for chemistry of normal human knee synovium: a quantitative study. bone involvement in the HLA–B27–associated spondyloarthrides Ann Rheum Dis 2004;63:785–90. and allied disorders. Arthritis Rheum 2007;56:224–33. 21. McGonagle D, McDermott MF. A proposed classification of the 4. Van Kuijk AW, Reinders-Blankert P, Smeets TJ, Dijkmans BA, immunological diseases. PLoS Med 2006;3:e297. Tak PP. Detailed analysis of the cell infiltrate and the expression 22. Gordon S. Alternative activation of macrophages. Nat Rev Immu- of mediators of synovial inflammation and joint destruction in the nol 2003;3:23–35.