ARTHRITIS & RHEUMATISM Vol. 50, No. 12, December 2004, pp 3762–3771 DOI 10.1002/art.20631 © 2004, American College of Rheumatology REVIEW Proinflammatory S100 Proteins in Arthritis and Autoimmune Disease Dirk Foell and Johannes Roth range of pro- and antiinflammatory activities. This heterogeneity is important for propagation or down-regulation of the initiated inflammatory process and the resulting tissue damage (3). Pro- and antiinflammatory properties of macrophages are mainly dependent on the stage of differentiation, as well as on distinct mechanisms of activation. One important signaling mechanism regulating these processes in phagocytes is the elevation of free intracellular calcium concentrations. Calcium ions induce conformational changes of calcium-binding proteins which allow their interaction with specific intracellular target structures. The major calcium-binding proteins, which are specifically expressed in granulocytes and early differentiation stages of monocytes, are S100A8, S100A9, and S100A12. In this review, we bridge the current understanding of the biologic functions assigned to these proinflammatory S100 proteins with a prospect for potential diagnostic and therapeutic strategies for arthritis and autoimmune disorders. Introduction Aberrant adaptive immune responses and disturbed activation of innate immune mechanisms are involved in the pathogenesis of arthritis and other autoinflammatory diseases (1). The innate immune system is equipped with a large repertoire of cellular and humoral components which are normally capable of establishing rapid effector mechanisms that protect the host organism from various challenges. The parallel engagement of different signaling pathways and the extensive interconnection between many proinflammatory mediators lead to a redundancy and selfamplification of innate immune responses. Disturbed regulation of this complicated network involving proand antiinflammatory factors may cause autoinflammation (2). An important cell population integrating many of these mechanisms is the population of phagocytes, especially macrophages (3), due to a virtually endless source of soluble mediators and cell surface receptors. Neutrophilic granulocytes and macrophages are abundant in the inflamed synovial membrane, and their activation correlates significantly with the severity of inflammatory arthritis. Macrophages may be involved in the initiation of arthritis as antigen-presenting cells, but they also contribute to disease progression since they exhibit widespread proinflammatory, destructive, and remodeling capabilities (4,5). However, macrophages are not a homogeneous cell population, since they encompass distinct phenotypes which exhibit a wide The S100 family of calcium-binding proteins Among the several groups of calcium-binding proteins, the family of S100 proteins comprises the largest group. Three phagocyte-specific S100 proteins comprise the group of calgranulins. The members of this protein family, S100A8, S100A9, and S100A12, are characterized by a unique expression pattern, with strong prevalence in cells of myeloid origin. S100A8 and S100A9. S100A8 and S100A9 represent the majority of the calcium-binding capacity in phagocytes (6,7). In neutrophilic granulocytes, S100A8 and S100A9 represent ⬃40% of the soluble cytosolic protein content (7). The human S100A8 protein (also known as myeloid-related protein 8 [MRP-8], calgranulin A, or L1 light chain) is made up of 93 amino acids and has a molecular weight of 10.8 kd. The human S100A9 protein (MRP-14, calgranulin B, and L1 heavy Supported by a grant from the Interdisciplinary Center for Clinical Research (Fo2/26/04) of the University of Muenster. Dirk Foell, MD, Johannes Roth, MD: University of Muenster, Muenster, Germany. Address correspondence and reprint requests to Johannes Roth, MD, Institute of Experimental Dermatology, University of Muenster, Roentgenstrasse 21, D-48149 Muenster, Germany. E-mail: email@example.com. Submitted for publication April 14, 2004; accepted in revised form August 16, 2004. 3762 S100 PROTEINS IN ARTHRITIS AND AUTOIMMUNITY chain) consists of 114 amino acids with a molecular weight of 13.2 kd (8,9). The amino acid sequence of S100A9 reveals highest homology with the other phagocyte-specific proteins S100A8 (30%) and S100A12 (46%). The genes for both proteins are found in a cluster on human chromosome 1q21. The gene structure is typical for members of the S100 family, characterized by a pattern of 3 exons separated by 2 introns. The expression of S100A8 and S100A9 is highly tissue specific. Both proteins are present in circulating granulocytes and monocytes, but neither in resting tissue macrophages nor in lymphocytes (10,11). In general, S100A8 and S100A9 show a parallel regulation of gene expression, but differential expression of S100A8 and S100A9 has been found to indicate a chronic type of inflammation, e.g., in granulomatous lung diseases, some forms of glomerulonephritis, and chronic renal allograft rejection (12–14). Expression of these S100 proteins in monocytes and macrophages is restricted to early differentiation stages and declines rapidly during maturation (11,15). Different inflammatory stimuli suppress transcription of S100A8 and S100A9 messenger RNA restricting the expression of these proteins to early stages of inflammatory processes. Expression in some epithelial cells, e.g., keratinocytes under inflammatory conditions or mucous epithelium of the esophagus, is the single exception to this myeloid expression pattern (16,17). Most of the S100 proteins exist in homodimers, but some also assemble to heterodimers. The preferred form for human S100A8 and S100A9 is the S100A8/ S100A9 heterodimer. This complex was also called leukocyte protein L1 or cystic fibrosis antigen, and is still designated as calprotectin by some research groups (18). Trimer formation of S100A8 and S100A9 has been suggested from studies using gel chromatography, but is contradicted by structural data obtained by nuclear magnetic resonance analysis (19,20). Homodimerization, as well as dimerization with other S100 proteins, especially with S100A12, has been excluded for human S100A8 and S100A9 (21,22). The S100A8/S100A9 heterodimers associate with (S100A8/S100A9)2 heterotetramers in a calcium-dependent manner (20). Calcium binding results in an exposure of hydrophobic domains to the surface of the protein complex, which may alleviate interaction with target proteins (19,23). Elevation of intracellular calcium concentrations induces interactions of S100A8/S100A9 complexes, which are mainly localized in the cytosol of resting phagocytes, with membrane and cytoskeletal structures indicating a role in cytoskeletal rearrangement and cell 3763 migration after activation of phagocytes (6,11,24). In this context, S100A9 acts as a regulatory subunit in the S100A8/S100A9 complex and integrates inputs from 2 major signaling pathways: the p38 microtubuleassociated protein kinase cascade, and calciumdependent signal transduction (6). Accordingly, targeted deletion of the S100A9 gene induces changes in the migratory properties of murine phagocytes. Stimulation of S100A9-deficient neutrophils with interleukin-8 (IL-8) failed to provoke an up-regulation of CD11b and to stimulate migration upon a chemotactic gradient (24). Deletion of the S100A8 gene is not compatible with life, which confirms the biologic relevance of the S100A8/ S100A9 complex (25). S100A12. More recently, S100A12 has been described as an additional phagocyte-specific S100 protein (26,27). Calgranulin C, calcium-binding protein in amniotic fluid 1, and extracellular newly identified receptor for advanced glycation end product–binding protein (EN-RAGE) are synonyms of S100A12. The primary structure of human S100A12 consists of 91 amino acids with a molecular mass of 10.4 kd. Human S100A12 shows highest homologies with S100A9 (46%) and to S100A8 (40%). The human S100A12 gene, localized within the S100 gene cluster on chromosome 1q21 between the S100A8 and S100A9 genes, is composed of 3 exons which are divided by 2 introns. Unlike proteins S100A8 and S100A9, homologs with protein S100A12 have only been identified in some species. There is evidence that chromosomal rearrangements during rodent evolution damaged the murine S100A12 gene, indicating the absence of the protein in mice (28). The occurrence of human S100A12 in the cytoplasm of granulocytes resembles the distribution pattern of S100A8/S100A9, with the difference being that it is less abundant (29,30). In the presence of calcium, S100A12 forms homodimers, but there is no complex formation with S100A8 or S100A9 (22). Thus, S100A12 acts individually during calcium-dependent signaling, independent of S100A8/S100A9 (22). Extracellular functions of phagocyte-specific S100 proteins Release by activated phagocytes. Since S100 proteins lack the leader sequence typical for intracellular transport via the endoplasmic reticulum and Golgi complex (mechanisms by which they pass the plasma membrane), an alternative secretory pathway is implied. It has been demonstrated that S100A8, S100A9, and S100A12 are released by stimulated phagocytes (31–33). 3764 Figure 1. Secretion of S100A8/S100A9 (A8 and A9). Activation of protein kinase C (PKC) through stimulation with lipopolysaccharide (LPS), interleukin-1␤ (IL-1␤), tumor necrosis factor (TNF), or other activators causes secretion of S100A8/S100A9. However, additional calcium signals following contact with stimulated endothelium or extracellular matrix proteins are necessary to induce secretion of S100A8/S100A9. Active secretion of the heterodimer by human monocytes and activated granulocytes uses an alternative pathway, which bypasses the Golgi route and depends on intact microtubules (MT). However, the mechanism of crossing the lipid bilayer of the plasma membrane is completely unknown. Released S100A8/S100A9 interacts with surface receptors on target cells, thereby inducing cellular responses through as-yet-unknown signaling pathways. However, reported effects include up-regulation of adhesion molecules on leukocytes (CD11b) and induction of apoptosis. RAGE ⫽ receptor for advanced glycation end products. Secretion of the S100A8/S100A9 heterodimer by human monocytes is an energy-dependent process which requires intact microtubules (32). Both activation of protein kinase C and additional calcium signals through contact with stimulated endothelium are necessary to induce secretion. In contrast, contact of monocytes with resting endothelium was found to inhibit protein kinase C–induced secretion (34). These data confirm specific release of S100A8/S100A9 during the interaction of phagocytes with activated endothelium (Figure 1). The exact mechanisms of S100A12 secretion remain to be defined, but published observations indicate a similar release by granulocytes which adhere to endothelium at sites of local inflammation (33,35). Effects on endothelium. Since secretion of phagocyte-specific S100 proteins is associated with adherence and transendothelial migration, cells within the endothelial layer are among the first cells targeted by released S100A8/S100A9 or S100A12. S100A12 has been implicated in a novel inflammatory axis involving RAGE, a multiligand receptor of the immunoglobulin FOELL AND ROTH superfamily expressed on endothelium and cells of the immune system (36,37). S100A12 binds to RAGE; whether this is the case for S100A8 and S100A9 has not been established. The binding of S100A12 and possibly S100A8/S100A9 to RAGE or other surface receptors results in activation of various intracellular signaling pathways. As a consequence, surface expression of vascular cell adhesion molecule 1 (VCAM-1) increases on endothelial cells after S100A12 stimulation. Enhanced binding of integrin very late activation antigen 4–bearing mononuclear cells to S100A12-stimulated endothelium has been demonstrated (36). In addition, S100A12 increased expression of intercellular adhesion molecule 1 (ICAM-1), thereby providing a mechanism by which polymorphonuclear leukocytes might be attracted to S100A12-stimulated endothelium as well (36). Induction of VCAM-1 and ICAM-1 expression by S100A12 is, at least in part, mediated by activation of NF-B (38). Other cell surface molecules have been described that could act independently from RAGE or as coreceptors to enhance the affinity of RAGE for S100A12, S100A8, and S100A9. The S100A8/S100A9 complex binds to endothelial cells via the S100A9 molecule by specific interaction with heparan sulfate proteoglycans (39). In addition, both molecules modulate transendothelial migration of leukocytes by binding of novel carboxylated glycans which are expressed by inflammatory activated endothelial cells (40). However, the exact effects of S100A8/S100A9 on endothelium remain to be described in greater detail. In contrast to human S100A8, direct chemotactic potency of the murine counterpart may contribute to transendothelial migration in mice (41). The different function of S100A8 with obvious chemotactic properties in mice compared with humans may be due to substitution of lacking murine S100A12 by S100A8 in mice. The release of S100A8/ S100A9 and/or S100A12 at sites of inflammation observed in vivo is feasible to constitute a positive feedback loop in which interaction of primed phagocytes with endothelial cells facilitates the further recruitment of even more leukocytes (32,42–44). Activation of immune cells. The complex of S100A8/S100A9 increases binding activity of the integrin receptor CD11b/CD18 on neutrophils, which enhances adhesion of these cells to endothelial cells and extracellular matrix proteins (44). S100A12 exhibits direct chemotactic effects on phagocytes (42). Furthermore, activation of RAGE by S100A12 up-regulates expression of proinflammatory cytokines, such as tumor necrosis factor (TNF) and IL-1␤, by murine macrophage-like BV-2 cells (36). Peripheral blood mononuclear cells (PBMCs) S100 PROTEINS IN ARTHRITIS AND AUTOIMMUNITY 3765 lished observations). Antimicrobial properties of S100A8, S100A9, and S100A12 provide further evidence that these proteins contribute to unspecific host defense mechanisms (18). Deleterious effects on fibroblasts and on the recovery of inflammatory tissue from S100A8/ S100A9 have been demonstrated (45,46). Thus, S100A8/ S100A9 released at sites of inflammation may directly contribute to tissue damage mediated by macrophages probably along with other factors, e.g., oxidative stress. RAGE-independent pathways may be responsible for the induction of apoptosis by S100 proteins (47). Due to zinc-binding properties of S100A8/S100A9, zinc depletion has been discussed as a mechanism underlying proapoptotic effects of these proteins (48). The role of S100 proteins in synovial inflammation Figure 2. Proinflammatory effects of S100A12 (A12) on target cells. S100A12 secretion by granulocytes stimulated with TNF or LPS has been demonstrated. The binding of S100A12 to RAGE or to other surface receptors results in activation of various intracellular signaling pathways. NF-B–driven up-regulation of adhesion molecule surface expression on endothelial cells has been demonstrated for vascular cell adhesion molecule 1 (VCAM-1) and intercellular adhesion molecule 1 (ICAM-1). Activation of RAGE by S100A12 also results in upregulated expression of proinflammatory cytokines, such as TNF and IL-1␤, by other phagocytes. Released TNF may stimulate granulocytes to secrete S100A12, thereby establishing a self-amplifying positive feedback loop. S100A12-activated peripheral blood mononuclear cells release IL-2 and show enhanced expression of cytokines. See Figure 1 for other definitions. exposed to S100A12 displayed enhanced release of IL-2 into culture supernatants. In addition to these findings, an enhanced mitogenic response to crosslinking CD3/ CD28 after stimulation of PBMCs with S100A12 was noted (36). Taken together, the results of these in vitro studies demonstrated that released S100 proteins activate immune cells critical to pathogenesis of inflammation by triggering proliferation and generation of cytokines (Figure 2). Cytotoxic effects. It has been suggested that S100A8 and S100A9 exert regulatory activities in inflammatory processes through effects on the survival or growth of cells participating in the inflammatory reaction and tissue homeostasis. The complex of S100A8/ S100A9 exhibits proapoptotic effects on fibroblasts, myocytes, and lymphocytes in vitro, but not on neutrophils or monocytes (ref. 45 and Foell D, et al: unpub- A direct role of S100A8 and S100A9 in inflammatory processes has been confirmed by identification of a new inflammatory syndrome characterized by extraordinarily high expression of these two calciumbinding molecules (49). Arthritis is a symptom in most affected patients, and there is accumulating evidence for a role of proinflammatory S100 proteins in the pathogenesis of synovial inflammation and autoimmune diseases (Tables 1 and 2). Experimental models of arthritis. Data from experimental models of arthritis indicate a direct role of phagocyte-specific S100 proteins in synovitis. Use of a DNA subtraction technique for identification of arthritis-specific genes led to the identification of S100A8 and S100A9 in rats. Expression correlated with chronic inflammation in arthritis-susceptible LEW/n rats, but was absent in F344/n rats, which do not develop arthritis in response to streptococcal cell wall peptidoglycan–polysaccharide complexes (50). There is one contradictory report showing a slight attenuating effect of S100A8/S100A9 on avridine-induced arthritis in rats which, however, was not statistically significant (51). Experimental data obtained from mice lacking the proinflammatory Fc␥ receptor I (Fc␥RI) and Fc␥RIII, or the antiinflammatory Fc␥RII, show a clear correlation of disease activity with expression of S100A8 and S100A9 in complex-mediated as well as antigeninduced arthritis (52,53). Interestingly, this effect was not simply due to quantitative differences of infiltration, but rather reflected the state of activation of synovial macrophages (52–54). There was also an accumulation of S100A8- and S100A9-expressing macrophages in close proximity to the cartilage surface, and a clear correlation of S100A8 and S100A9 expression with signs 3766 FOELL AND ROTH Table 1. Disease associations of S100A8/S100A9 Condition Species Ref. Immune complex–mediated arthritis Antigen-induced arthritis Immune complex–induced arthritis Avridine-induced arthritis Rheumatoid arthritis Psoriatic arthritis Reactive arthritis Juvenile idiopathic arthritis Inflammatory bowel disease Systemic lupus erythematosus/glomerulonephritis Systemic sclerosis Dermatomyositis/polymyositis Hyperzincemia/systemic inflammation Mouse Mouse Rat Rat Human Human Human Human Human Human Human Human Human 52, 53 54 50 51 9, 10, 58, 60–62, 80 60 62 34, 67–70 55, 75 13, 14, 71 77 46 49 of cartilage destruction as shown by matrix metalloproteinase–mediated proteoglycan damage, chondrocyte death, and development of erosions (52– 54). These data indicate a direct role of S100A8 and S100A9 in the destructive process of inflammatory arthritis (Figures 3A–D), which is supported by the finding that these markers are associated with production of IL-1 and oxidative radicals in inflamed tissues (55,56). In addition, these experiments demonstrate that expression of S100A8 and S100A9 in vivo is somehow regulated by the expression of pro- and antiinflammatory Fc␥R on phagocytes. Analysis of the course of experimental arthritis in S100A9⫺/⫺ mice is the subject of current investigations. S100A12 has also been demonstrated to trigger synovial inflammation in mice with collagen-induced arthritis (57). The reported proinflammatory effects involving cytokine production and metalloproteinase activation within the synovium depend on interaction of S100A12 with RAGE (36,37). Rheumatoid arthritis. S100A8 and S100A9 have been initially identified in the context of rheumatoid arthritis (RA). Activated phagocytes expressing these S100 proteins are among the first cells infiltrating inTable 2. Disease associations of S100A12 Condition Species Ref. Collagen-induced arthritis Colitis Hypersensitivity Vasculitis Rheumatoid arthritis Psoriatic arthritis Juvenile idiopathic arthritis Vasculitis Inflammatory bowel disease Mouse Mouse Mouse Mouse Human Human Human Human Human 37, 57 36 36 65 30, 35, 57, 63 35 66 73 33 flammatory lesions in the synovium (9,10). Early recruited phagocytes expressing S100A8 and S100A9 have been found in the sublining layer of inflamed synovial tissue (58). In patients with active arthritis, S100A8/ S100A9 is also expressed in macrophage-like cells within the lining layer, which show altered activation and differentiation under inflammatory conditions (Figure 3E). The expression of S100A8 and S100A9 was found to be strongest at the cartilage–pannus junction, which is the prime site of cartilage destruction and bone erosion in arthritis (58). Direct effects of S100A8/S100A9 on cartilage cells still remain to be analyzed, but there is evidence for release of S100A8/S100A9 in synovial fluid obtained from inflamed joints (34,59,60). S100A8/ S100A9 concentrations in synovial fluid are ⬃10-fold higher than concentrations in serum obtained in parallel from individual patients (34,60). Elevated S100A8/ S100A9 serum concentrations therefore reflect release of these proteins from activated phagocytes within the synovium and the synovial fluid. There are 3 major consequences of this synovial expression pattern and the in vitro data presented above: 1) S100A8/S100A9 expression is associated with inflammation in the synovial tissue; 2) S100A8/S100A9 at the site of inflammation activates endothelium and thereby facilitates further recruitment of inflammatory cells into the synovial tissue; and 3) the resulting increase of S100A8/S100A9 serum concentrations may be an important serum marker of the extent of local inflammation in the affected joints. In clinical practice, laboratory parameters most frequently used in the monitoring of disease activity in chronic arthritis are C-reactive protein level and erythrocyte sedimentation rate, although sensitivity and specificity for changes in synovial inflammation are limited. S100 PROTEINS IN ARTHRITIS AND AUTOIMMUNITY Figure 3. Expression of S100A9 at sites of inflammation. A and B, In an experimental model of immune complex–mediated arthritis in mice, S100A9 was extensively expressed in inflamed synovial tissue (solid arrow), especially at the cartilage–pannus junction (open arrow). C and D, In mice treated with an interleukin-1 receptor antagonist, decreased local expression of S100A9 correlated well with reduced arthritis (arrows mark sites corresponding to the arrows in A and B). E, In inflamed synovial tissue obtained from a patient with rheumatoid arthritis (RA), S100A9 expression was found in the lining layer (solid arrow) as well as in infiltrating cells in the sublining (open arrow). F, In contrast to RA, S100A9 expression was more pronounced in the perivascular tissue of inflamed synovium (solid arrows) in a patient with psoriatic arthritis. G, S100A9 expression by infiltrating macrophages (solid arrow) showed a clear association with muscle fiber necrosis (indicated by the central localization of nuclei in degenerating fibers) (open arrows) in a biopsy specimen from inflamed muscle in a patient with dermatomyositis. H, In a patient with glomerulonephritis accompanying systemic lupus erythematosus, the majority of infiltrating macrophages expressed S100A9 in affected glomeruli. In all stainings, S100A8 colocalized with S100A9 and showed an identical expression pattern (not shown). (Original magnification ⫻ 100 in A and C; ⫻ 200 in B and D; ⫻ 400 in E–H.) 3767 Numerous studies on serum levels of S100A8/S100A9 in patients with RA have confirmed the excellent correlation of serum S100A8/S100A9 concentrations with the inflammatory activity of arthritis (60–62). The diagnostic capacity of S100A8/S100A9 as a marker of synovial inflammation is superior to that of C-reactive protein or erythrocyte sedimentation rate (60–62). More recently, the proinflammatory role of S100A12 in synovitis has been emphasized. S100A12 levels have been found to be increased in the synovial fluid and serum of patients with RA but undetectable in patients with osteoarthritis (63). S100A12 is strongly expressed in inflamed synovial tissue, whereas it is nearly undetectable in synovia of control subjects or patients after successful treatment (35). Inflamed synovium was found to contain S100A12-positive neutrophils in the sublining and interstitial regions, often surrounding the perivascular tissue, but to a lesser extent in the synovial lining layer (30). Serum levels of S100A12 correlate well with disease activity (35). Indirect evidence indicated involvement of S100A12 binding to RAGE in the context of synovial inflammation. A RAGE variant with a glycine–serine substitution at amino acid position 82 within the ligand binding domain of this receptor is associated with an amplified inflammatory response upon engagement by S100A12. Individuals with the genetic variant of the RAGE G82S polymorphism exhibited increased susceptibility for the development of arthritis (57). Thus, exaggerated signaling by different receptor variants in response to S100A12 ligation may contribute to enhanced proinflammatory mechanisms in synovitis. Psoriatic arthritis. Psoriatic arthritis (PsA) is not as destructive as RA, possibly due to a lower degree of synovial macrophage infiltration. Nevertheless, freshly recruited phagocytes contribute to synovial inflammation in PsA. The expression pattern of S100A8 and S100A9 in PsA is clearly distinct from that in RA, as these molecules are predominantly found in perivascular areas of the synovial sublining layer (Figure 3F). This expression is significantly reduced in serum and synovium from patients with PsA after successful methotrexate treatment (60). Similar results have been obtained with S100A12 (35). The perivascular distribution of S100 proteins is particularly intriguing, since angiogenesis and altered function of microvascular endothelium have been reported to be an important mechanism of synovial inflammation in PsA (64). Activation of endothelial cells is a key property of S100A12, and the enhanced angiogenesis in models of diabetic mice was found to be 3768 inhibited by blocking the interaction of S100A12 with RAGE (65). Juvenile idiopathic arthritis. S100A8, S100A9, and S100A12 have also been detected in serum and synovial fluid of patients with juvenile idiopathic arthritis (JIA) (34,66). Serum concentrations of S100A8/ S100A9 correlated well with individual disease activity in long-term studies of children with JIA (34,66–68). Findings of preliminary studies suggested that patients with clinically inactive JIA but elevated levels of S100A8/ S100A9 or S100A12 may be at risk for disease flares (66,69). Children with systemic-onset JIA, characterized by massive neutrophil activation, have serum concentrations of phagocyte-specific S100 proteins up to 20-fold higher than those found in patients with sepsis or other inflammatory disorders (66,68,70). Therefore, S100A8/ S100A9 and S100A12 may be useful to distinguish systemic-onset JIA from systemic infections, which represent the most important differential diagnosis. Interestingly, patients with systemic onset of JIA also show extensive expression of S100A8/S100A9 in the dermal epithelium, indicating for the first time an active role of this cell type during the initiation of a systemic autoimmune disorder (68). S100 proteins in other autoimmune disorders Dermatomyositis. In dermatomyositis, polymyositis, and inclusion body myositis there is a clear association of S100A8 and S100A9 expression by infiltrating macrophages with degeneration of myofibers (Figure 3G). S100A8 and S100A9 inhibited proliferation and differentiation of myoblasts in vitro and induced apoptosis of these cells via activation of caspase 3, indicating that activated macrophages promote destruction and impair regeneration of myocytes in the course of inflammatory myopathies via release of these S100 proteins. Systemic lupus erythematosus. In a crosssectional study of 100 patients with systemic lupus erythematosus (SLE), elevated serum concentrations of S100A8/S100A9 were detected, which correlated significantly with the SLE Disease Activity Index (71). Immunohistochemical analysis of renal biopsy samples demonstrates that expression of S100A8/S100A9 by infiltrating macrophages in the glomeruli of patients with SLE corresponds to the severity of the inflammatory process (Figure 3H). Macrophages in the renal interstitium expressed S100A8 and S100A9 without concomitant formation of their heterodimer, indicating a FOELL AND ROTH chronic type of inflammatory reaction in the interstitium in SLE glomerulonephritis (13). Vasculitis. Enhanced expression of S100 proteins in diabetic apolipoprotein E⫺/⫺ mice with vascular inflammation has been reported (72). Elevated serum concentrations of S100A12 have been detected in Kawasaki disease, which correlated with the inflammatory disease activity in this acute vasculitis syndrome in children (73). S100A12 is more reliable than conventional parameters to determine response to gamma globulin treatment. Patients with coronary artery abnormalities had higher initial and maximal S100A12 concentrations than patients without cardiac complications. Accumulation of S100A8- and S100A9-expressing macrophages was found in antineutrophil cytoplasmic antibody–positive renal vasculitis, especially in areas of glomerular active lesions (74). Chronic inflammatory bowel disease. S100A8/ S100A9-expressing phagocytes have been detected as proinflammatory cells at sites of intestinal inflammation (55,75). S100A8/S100A9 serum levels are useful in monitoring inflammation in patients with Crohn’s disease or ulcerative colitis (75). S100A12 is another proinflammatory promoter of inflammatory bowel disease, since blocking the interaction of S100A12 with RAGE reduced inflammation in a murine model of colitis (36). S100A12 has been found to be massively expressed in inflamed tissue from patients with active inflammatory bowel disease. Serum concentrations of S100A12 correlate well with disease activity in individual patients (33). Other autoimmune disorders. Elevated serum concentrations of S100A8/S100A9 have also been found in patients with Sjögren’s syndrome (76) and in patients with systemic sclerosis (77). The function of S100 proteins in patients with these conditions remains undefined. Opportunities for therapeutic intervention As a consequence of the various proinflammatory properties of phagocyte-specific S100 proteins, strategies targeting these molecules might represent a novel option for antiinflammatory therapies. There is evidence from animal models that the S100A12dependent activation of endothelium and the recruitment of inflammatory cells can be inhibited in vivo by administration of either anti-S100A12 antibodies or soluble RAGE constructs. These blocking agents also inhibited acute inflammation in a murine model of methylated bovine serum albumin–induced hypersensitivity, as well as chronic bowel inflammation in IL-10⫺/⫺ S100 PROTEINS IN ARTHRITIS AND AUTOIMMUNITY mice. Most interestingly, S100A12 antagonists used in mice immunized and challenged with bovine type II collagen suppressed clinical and histologic evidence of arthritis and diminished expression of TNF, IL-6, and matrix metalloproteinases 3, 9, and 13 in affected tissues (37,57). Thus, S100A12 is an attractive target for novel antiinflammatory therapies in arthritis and autoimmune disorders. Data from in vitro studies indicated that the propagation of transendothelial leukocyte migration mediated by S100A8/S100A9 may be inhibited by blocking their binding to N-glycans on endothelial cells (40). Another study tracked successful inhibition of S100A8/ S100A9-dependent inflammation in murine colitis by blocking binding to cells via N-glycans in vivo (78). In addition, blockade of lipopolysaccharide-induced proinflammatory effects of S100A8/S100A9 has been reported to be effective in a murine model (79). Taking into account the abundance of S100A8 and S100A9 in the inflammatory processes, immunologic intervention targeting these proteins may be a promising therapeutic option. However, due to their high extracellular concentrations at sites of inflammation, it seems unlikely that conventional therapeutic interventions, such as application of neutralizing antibodies or soluble receptors, will effectively block proinflammatory activities of S100A8 and S100A9 in the clinical setting. The most promising approach may be the inhibition of the release of these proteins, which is strongly enhanced at sites of active inflammation. This strategy may be very specific, since both proteins are secreted via a so-called “alternative pathway.” The inhibition of this alternative transport mechanism will not affect classic secretion of other proteins via endoplasmic reticulum and Golgi complex, and thus may not be complicated by major side effects. S100A8/S100A9 complex formation and phosphorylation of S100A9 seem to be the important factors for intracellular transport during secretion and may, therefore, be primary targets for S100A8- and S100A9-directed antiinflammatory strategies. Conclusions There is a large and continuously growing body of evidence for the proinflammatory functions of S100A8, S100A9, and S100A12 in the pathogenesis of arthritis and autoinflammatory disorders. These proteins are expressed and secreted by phagocytes within inflamed tissue, and they interact specifically with molecules on target cells. The subsequent activation of endothelium 3769 and immune cells results in the further recruitment of leukocytes at sites of inflammation. In addition, direct cytotoxic effects may contribute to tissue damage in inflammatory lesions. The proinflammatory S100 proteins are valuable diagnostic marker proteins, due to a close correlation with local inflammatory processes that involve phagocyte activation. Diagnostic analyses of these proteins provide additional information that is not accessible from other markers of inflammation. In addition, proinflammatory S100 proteins are attractive therapeutic targets for immune interventions in the treatment of arthritis and autoimmune diseases. REFERENCES 1. Firestein GS. Evolving concepts of rheumatoid arthritis. Nature 2003;423:356–61. 2. Firestein GS, Zvaifler NJ. How important are T cells in chronic rheumatoid synovitis? II. T cell–independent mechanisms from beginning to end. Arthritis Rheum 2002;46:298–308. 3. Mosser DM. The many faces of macrophage activation. J Leukoc Biol 2003;73:209–12. 4. Arend WP. The innate immune system in rheumatoid arthritis. Arthritis Rheum 2001;44:2224–34. 5. Kinne RW, Brauer R, Stuhlmuller B, Palombo-Kinne E, Burmester GR. Macrophages in rheumatoid arthritis. Arthritis Res 2000;2:189–202. 6. Van den Bos C, Roth J, Koch HG, Hartmann M, Sorg C. Phosphorylation of MRP14, an S100 protein expressed during monocytic differentiation, modulates Ca(2⫹)-dependent translocation from cytoplasm to membranes and cytoskeleton. J Immunol 1996;156:1247–54. 7. Edgeworth J, Freemont P, Hogg N. Ionomycin-regulated phosphorylation of the myeloid calcium-binding protein p14. Nature 1989;342:189–92. 8. Lagasse E, Clerc RG. Cloning and expression of two human genes encoding calcium-binding proteins that are regulated during myeloid differentiation. Mol Cell Biol 1988;8:2402–10. 9. Odink K, Cerletti N, Bruggen J, Clerc RG, Tarcsay L, Zwadlo G, et al. Two calcium-binding proteins in infiltrate macrophages of rheumatoid arthritis. Nature 1987;330:80–2. 10. Zwadlo G, Bruggen J, Gerhards G, Schlegel R, Sorg C. Two calcium-binding proteins associated with specific stages of myeloid cell differentiation are expressed by subsets of macrophages in inflammatory tissues. Clin Exp Immunol 1988;72:510–5. 11. Roth J, Burwinkel F, van den Bos C, Goebeler M, Vollmer E, Sorg C. MRP8 and MRP14, S-100-like proteins associated with myeloid differentiation, are translocated to plasma membrane and intermediate filaments in a calcium-dependent manner. Blood 1993;82: 1875–83. 12. Delabie J, de Wolf-Peeters C, van den Oord JJ, Desmet VJ. Differential expression of the calcium-binding proteins MRP8 and MRP14 in granulomatous conditions: an immunohistochemical study. Clin Exp Immunol 1990;81:123–6. 13. Frosch M, Vogl T, Waldherr R, Sorg C, Sunderkotter C, Roth J. Expression of MRP8 and MRP14 by macrophages is a marker for severe forms of glomerulonephritis. J Leukoc Biol 2004;75: 198–206. 14. Goebeler M, Roth J, Burwinkel F, Vollmer E, Bocker W, Sorg C. Expression and complex formation of S100-like proteins MRP8 and MRP14 by macrophages during renal allograft rejection. Transplantation 1994;58:355–61. 3770 15. Roth J, Goebeler M, van den Bos C, Sorg C. Expression of calcium-binding proteins MRP8 and MRP14 is associated with distinct monocytic differentiation pathways in HL-60 cells. Biochem Biophys Res Commun 1993;191:565–70. 16. Kunz M, Roth J, Sorg C, Kolde G. Epidermal expression of the calcium binding surface antigen 27E10 in inflammatory skin diseases. Arch Dermatol Res 1992;284:386–90. 17. Thorey IS, Roth J, Regenbogen J, Halle JP, Bittner M, Vogl T, et al. The Ca2⫹-binding proteins S100A8 and S100A9 are encoded by novel injury-regulated genes. J Biol Chem 2001;276:35818–25. 18. Steinbakk M, Naess-Andresen CF, Lingaas E, Dale I, Brandtzaeg P, Fagerhol MK. Antimicrobial actions of calcium binding leucocyte L1 protein, calprotectin. Lancet 1990;336:763–5. 19. Hunter MJ, Chazin WJ. High level expression and dimer characterization of the S100 EF-hand proteins, migration inhibitory factor-related proteins 8 and 14. J Biol Chem 1998;273:12427–35. 20. Strupat K, Rogniaux H, van Dorsselaer A, Roth J, Vogl T. Calcium-induced noncovalently linked tetramers of MRP8 and MRP14 are confirmed by electrospray ionization-mass analysis. J Am Soc Mass Spectrom 2000;11:780–8. 21. Propper C, Huang X, Roth J, Sorg C, Nacken W. Analysis of the MRP8-MRP14 protein-protein interaction by the two-hybrid system suggests a prominent role of the C-terminal domain of S100 proteins in dimer formation. J Biol Chem 1999;274:183–8. 22. Vogl T, Propper C, Hartmann M, Strey A, Strupat K, van den Bos C, et al. S100A12 is expressed exclusively by granulocytes and acts independently from MRP8 and MRP14. J Biol Chem 1999;274: 25291–6. 23. Groves P, Finn BE, Kuznicki J, Forsen S. A model for target protein binding to calcium-activated S100 dimers. FEBS Lett 1998;421:175–9. 24. Manitz MP, Horst B, Seeliger S, Strey A, Skryabin BV, Gunzer M, et al. Loss of S100A9 (MRP14) results in reduced interleukin-8induced CD11b surface expression, a polarized microfilament system, and diminished responsiveness to chemoattractants in vitro. Mol Cell Biol 2003;23:1034–43. 25. Passey RJ, Williams E, Lichanska AM, Wells C, Hu S, Geczy CL, et al. A null mutation in the inflammation-associated S100 protein S100A8 causes early resorption of the mouse embryo. J Immunol 1999;163:2209–16. 26. Ilg EC, Troxler H, Burgisser DM, Kuster T, Markert M, Guignard F, et al. Amino acid sequence determination of human S100A12 (P6, calgranulin C, CGRP, CAAF1) by tandem mass spectrometry. Biochem Biophys Res Commun 1996;225:146–50. 27. Wicki R, Marenholz I, Mischke D, Schafer BW, Heizmann CW. Characterization of the human S100A12 (calgranulin C, p6, CAAF1, CGRP) gene, a new member of the S100 gene cluster on chromosome 1q21. Cell Calcium 1996;20:459–64. 28. Fuellen G, Foell D, Nacken W, Sorg C, Kerkhoff C. Absence of S100A12 in mouse: implications for RAGE-S100A12 interaction. Trends Immunol 2003;24:622–4. 29. Robinson MJ, Hogg N. A comparison of human S100A12 with MRP-14 (S100A9). Biochem Biophys Res Commun 2000;275: 865–70. 30. Yang Z, Tao T, Raftery MJ, Youssef P, Di Girolamo N, Geczy CL. Proinflammatory properties of the human S100 protein S100A12. J Leukoc Biol 2001;69:986–94. 31. Boussac M, Garin J. Calcium-dependent secretion in human neutrophils: a proteomic approach. Electrophoresis 2000;21: 665–72. 32. Rammes A, Roth J, Goebeler M, Klempt M, Hartmann M, Sorg C. Myeloid-related protein (MRP) 8 and MRP14, calcium-binding proteins of the S100 family, are secreted by activated monocytes via a novel, tubulin-dependent pathway. J Biol Chem 1997;272: 9496–502. 33. Foell D, Kucharzik T, Kraft M, Vogl T, Sorg C, Domschke W, et al. Neutrophil derived human S100A12 (EN-RAGE) is strongly FOELL AND ROTH 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. expressed during chronic active inflammatory bowel disease. Gut 2003;52:847–53. Frosch M, Strey A, Vogl T, Wulffraat NM, Kuis W, Sunderkotter C, et al. Myeloid-related proteins 8 and 14 are specifically secreted during interaction of phagocytes and activated endothelium and are useful markers for monitoring disease activity in pauciarticular-onset juvenile rheumatoid arthritis. Arthritis Rheum 2000;43: 628–37. Foell D, Kane D, Bresnihan B, Vogl T, Nacken W, Sorg C, et al. Expression of the pro-inflammatory protein S100A12 (ENRAGE) in rheumatoid and psoriatic arthritis. Rheumatology (Oxford) 2003;42:1383–9. Hofmann MA, Drury S, Fu C, Qu W, Taguchi A, Lu Y, et al. RAGE mediates a novel proinflammatory axis: a central cell surface receptor for S100/calgranulin polypeptides. Cell 1999;97: 889–901. Schmidt AM, Yan SD, Yan SF, Stern DM. The multiligand receptor RAGE as a progression factor amplifying immune and inflammatory responses. J Clin Invest 2001;108:949–55. Neish AS, Williams AJ, Palmer HJ, Whitley MZ, Collins T. Functional analysis of the human vascular cell adhesion molecule 1 promoter. J Exp Med 1992;176:1583–93. Robinson MJ, Tessier P, Poulsom R, Hogg N. The S100 family heterodimer, MRP-8/14, binds with high affinity to heparin and heparan sulfate glycosaminoglycans on endothelial cells. J Biol Chem 2002;277:3658–65. Srikrishna G, Panneerselvam K, Westphal V, Abraham V, Varki A, Freeze HH. Two proteins modulating transendothelial migration of leukocytes recognize novel carboxylated glycans on endothelial cells. J Immunol 2001;166:4678–88. Lackmann M, Cornish CJ, Simpson RJ, Moritz RL, Geczy CL. Purification and structural analysis of a murine chemotactic cytokine (CP-10) with sequence homology to S100 proteins. J Biol Chem 1992;267:7499–504. Miranda LP, Tao T, Jones A, Chernushevich I, Standing KG, Geczy CL, et al. Total chemical synthesis and chemotactic activity of human S100A12 (EN-RAGE). FEBS Lett 2001;488:85–90. Eue I, Pietz B, Storck J, Klempt M, Sorg C. Transendothelial migration of 27E10⫹ human monocytes. Int Immunol 2000;12: 1593–604. Newton RA, Hogg N. The human S100 protein MRP-14 is a novel activator of the ␤2 integrin Mac-1 on neutrophils. J Immunol 1998;160:1427–35. Yui S, Nakatani Y, Mikami M. Calprotectin (S100A8/S100A9), an inflammatory protein complex from neutrophils with a broad apoptosis-inducing activity. Biol Pharm Bull 2003;26:753–60. Seeliger S, Vogl T, Engels IH, Schroder JM, Sorg C, Sunderkotter C, et al. Expression of calcium-binding proteins MRP8 and MRP14 in inflammatory muscle diseases. Am J Pathol 2003;163: 947–56. Sorci G, Riuzzi F, Agneletti AL, Marchetti C, Donato R. S100B causes apoptosis in a myoblast cell line in a RAGE-independent manner. J Cell Physiol 2004;199:274–83. Yui S, Mikami M, Tsurumaki K, Yamazaki M. Growth-inhibitory and apoptosis-inducing activities of calprotectin derived from inflammatory exudate cells on normal fibroblasts: regulation by metal ions. J Leukoc Biol 1997;61:50–7. Sampson B, Fagerhol MK, Sunderkotter C, Golden BE, Richmond P, Klein N, et al. Hyperzincaemia and hypercalprotectinaemia: a new disorder of zinc metabolism. Lancet 2002;360:1742–5. Imamichi T, Uchida I, Wahl SM, McCartney-Francis N. Expression and cloning of migration inhibitory factor-related protein (MRP)8 and MRP14 in arthritis-susceptible rats. Biochem Biophys Res Commun 1993;194:819–25. Brun JG, Haland G, Haga HJ, Fagerhol MK, Jonsson R. Effects of calprotectin in avridine-induced arthritis. APMIS 1995;103: 233–40. S100 PROTEINS IN ARTHRITIS AND AUTOIMMUNITY 52. Nabbe KC, Blom AB, Holthuysen AE, Boross P, Roth J, Verbeek S, et al. Coordinate expression of activating Fc␥ receptors I and III and inhibiting Fc␥ receptor type II in the determination of joint inflammation and cartilage destruction during immune complex– mediated arthritis. Arthritis Rheum 2003;48:255–65. 53. Nabbe KC, van Lent PL, Holthuysen AE, Kolls JK, Verbeek S, van den Berg WB. Fc␥RI up-regulation induced by local adenoviralmediated interferon-␥ production aggravates chondrocyte death during immune complex-mediated arthritis. Am J Pathol 2003;163: 743–52. 54. Van Lent P, Nabbe KC, Boross P, Blom AB, Roth J, Holthuysen A, et al. The inhibitory receptor Fc␥RII reduces joint inflammation and destruction in experimental immune complex-mediated arthritides not only by inhibition of Fc␥RI/III but also by efficient clearance and endocytosis of immune complexes. Am J Pathol 2003;163:1839–48. 55. Rugtveit J, Nilsen EM, Bakka A, Carlsen H, Brandtzaeg P, Scott H. Cytokine profiles differ in newly recruited and resident subsets of mucosal macrophages from inflammatory bowel disease. Gastroenterology 1997;112:1493–505. 56. Rugtveit J, Haraldsen G, Hogasen AK, Bakka A, Brandtzaeg P, Scott H. Respiratory burst of intestinal macrophages in inflammatory bowel disease is mainly caused by CD14⫹L1⫹ monocyte derived cells. Gut 1995;37:367–73. 57. Hofmann MA, Drury S, Hudson BI, Gleason MR, Qu W, Lu Y, et al. RAGE and arthritis: the G82S polymorphism amplifies the inflammatory response. Genes Immun 2002;3:123–35. 58. Youssef P, Roth J, Frosch M, Costello P, Fitzgerald O, Sorg C, et al. Expression of myeloid related proteins (MRP) 8 and 14 and the MRP8/14 heterodimer in rheumatoid arthritis synovial membrane. J Rheumatol 1999;26:2523–8. 59. Berntzen HB, Olmez U, Fagerhol MK, Munthe E. The leukocyte protein L1 in plasma and synovial fluid from patients with rheumatoid arthritis and osteoarthritis. Scand J Rheumatol 1991; 20:74–82. 60. Kane D, Roth J, Frosch M, Vogl T, Bresnihan B, FitzGerald O. Increased perivascular synovial membrane expression of myeloidrelated proteins in psoriatic arthritis. Arthritis Rheum 2003;48: 1676–85. 61. Brun JG, Jonsson R, Haga HJ. Measurement of plasma calprotectin as an indicator of arthritis and disease activity in patients with inflammatory rheumatic diseases. J Rheumatol 1994;21: 733–8. 62. Berntzen HB, Munthe E, Fagerhol MK. A longitudinal study of the leukocyte protein L1 as an indicator of disease activity in patients with rheumatoid arthritis. J Rheumatol 1989;16:1416–20. 63. Rouleau P, Vandal K, Ryckman C, Poubelle PE, Boivin A, Talbot M, et al. The calcium-binding protein S100A12 induces neutrophil adhesion, migration, and release from bone marrow in mouse at concentrations similar to those found in human inflammatory arthritis. Clin Immunol 2003;107:46–54. 64. Reece RJ, Canete JD, Parsons WJ, Emery P, Veale DJ. Distinct vascular patterns of early synovitis in psoriatic, reactive, and rheumatoid arthritis. Arthritis Rheum 1999;42:1481–4. 65. Park L, Raman KG, Lee KJ, Lu Y, Ferran LJ Jr, Chow WS, et al. Suppression of accelerated diabetic atherosclerosis by the soluble receptor for advanced glycation endproducts. Nat Med 1998;4: 1025–31. 66. Foell D, Wittkowski H, Hammerschmidt I, Wulffraat N, Schmel- 3771 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. ing H, Frosch M, et al. Monitoring neutrophil activation in juvenile rheumatoid arthritis by S100A12 serum concentrations. Arthritis Rheum 2004;50:1286–95. Berntzen HB, Fagerhol MK, Ostensen M, Mowinckel P, Hoyeraal HM. The L1 protein as a new indicator of inflammatory activity in patients with juvenile rheumatoid arthritis. J Rheumatol 1991;18: 133–8. Frosch M, Vogl T, Seeliger S, Wulffraat M, Kuis W, Viemann D, et al. Expression of myeloid-related proteins 8 and 14 in systemiconset juvenile rheumatoid arthritis. Arthritis Rheum 2003;48: 2622–6. Foell D, Frosch M, Schulze zur Wiesch A, Vogl T, Sorg C, Roth J. Methotrexate treatment in juvenile idiopathic arthritis: when is the right time to stop? Ann Rheum Dis 2004;63:206–8. Wulffraat NM, Haas PJ, Frosch M, de Kleer IM, Vogl T, Brinkman DM, et al. Myeloid related protein 8 and 14 secretion reflects phagocyte activation and correlates with disease activity in juvenile idiopathic arthritis treated with autologous stem cell transplantation. Ann Rheum Dis 2003;62:236–41. Haga HJ, Brun JG, Berntzen HB, Cervera R, Khamashta M, Hughes GR. Calprotectin in patients with systemic lupus erythematosus: relation to clinical and laboratory parameters of disease activity. Lupus 1993;2:47–50. Kislinger T, Tanji N, Wendt T, Qu W, Lu Y, Ferran LJ Jr, et al. Receptor for advanced glycation end products mediates inflammation and enhanced expression of tissue factor in vasculature of diabetic apolipoprotein E-null mice. Arterioscler Thromb Vasc Biol 2001;21:905–10. Foell D, Ichida F, Vogl T, Yu X, Chen R, Miyawaki T, et al. S100A12 (EN-RAGE) in monitoring Kawasaki disease. Lancet 2003;361:1270–2. Rastaldi MP, Ferrario F, Crippa A, Dell’Antonio G, Casartelli D, Grillo C, et al. Glomerular monocyte-macrophage features in ANCA-positive renal vasculitis and cryoglobulinemic nephritis. J Am Soc Nephrol 2000;11:2036–43. Lugering N, Stoll R, Kucharzik T, Schmid KW, Rohlmann G, Burmeister G, et al. Immunohistochemical distribution and serum levels of the Ca(2⫹)-binding proteins MRP8, MRP14 and their heterodimeric form MRP8/14 in Crohn’s disease. Digestion 1995; 56:406–14. Brun JG, Cuida M, Jacobsen H, Kloster R, Johannesen AC, Hoyeraal HM, et al. Sjogren’s syndrome in inflammatory rheumatic diseases: analysis of the leukocyte protein calprotectin in plasma and saliva. Scand J Rheumatol 1994;23:114–8. Kuruto R, Nozawa R, Takeishi K, Arai K, Yokota T, Takasaki Y. Myeloid calcium binding proteins: expression in the differentiated HL-60 cells and detection in sera of patients with connective tissue diseases. J Biochem (Tokyo) 1990;108:650–3. Srikrishna G, Turovskaya O, Shaikh R, Newlin R, Foell D, Murch S, et al. Novel N-glycans on antigen-presenting cells mediate Th-1 dependent murine colitis. 2003;13:226. Vandal K, Rouleau P, Boivin A, Ryckman C, Talbot M, Tessier PA. Blockade of S100A8 and S100A9 suppresses neutrophil migration in response to lipopolysaccharide. J Immunol 2003;171: 2602–9. Madland TM, Hordvik M, Haga HJ, Jonsson R, Brun JG. Leukocyte protein calprotectin and outcome in rheumatoid arthritis: a longitudinal study. Scand J Rheumatol 2002;31:351–4.