Activation of FLS in RA

In non-diseased tissue, the physiological function of synovial fibroblasts (FLS) is to build a lining membrane that secretes both hyaluronan and lubricin, key components of the synovial fluid and important in joint lubrication as well as to provide the joint cavity and the adjacent cartilage with nutritive plasma proteins . Moreover, FLSs are involved in continuous matrix remodelling by producing matrix components, such as collagen, and thus maintaining the homeostasis of synovium.

Under disease conditions such as rheumatoid inflammation, cellular alterations following activation through inflammatory stimuli are profound and result in a tumour-like transformation of FLS from fairly ‘innocent’ mesenchymal cells to destructive aggressors. These transformed RA-FLSs then play a leading role in establishing and driving RA. Supporting this notion, RA-FLSs exhibit a special phenotype that is characterised by a number of unique and remarkable features such as reduced susceptibility to apoptosis, expression of adhesion molecules and unbalanced production of matrix-degrading enzymes. At a morphological level, these cells are characterised by abundant rough endoplasmatic reticuli, large nuclei with prominent nucleoli and a more rounded shape. As a consequence of their altered morphology and cellular activation, RA-FLSs exhibit a distinct, anchorage-independent growth pattern and a resistance to contact inhibition at proliferation along with an aggressive, invasive behaviour . Of note, FLSs found in the most superficial lining layer of the synovial membrane and, thus, in direct contact with the extracellular matrix of specific joint components show the most pronounced and aggressive characteristics at cellular, structural and morphological levels, and hence most strongly facilitate the degrading and destroying processes. A number of co-implantation studies of RA-FLS into mice with severe combined immunodeficiency (SCID mice) were performed to investigate the ‘transformed-appearing’ activated phenotype of RA-FLSs. It could be shown that RA-FLSs not only exhibit an invasive phenotype but are also able to maintain their state of cellular activation over prolonged periods even in the absence of continuous stimulation by inflammatory environment . Searching for an explanation of this phenomenon, several molecules and regulatory mechanisms that contribute to this pathological ‘imprinting’ or ‘transformation’ have been identified. To date, there is no conclusive explanation as to what ultimately initiates and triggers the phenotypic switch in RA-FLS. However, it appears that in certain, perhaps predisposed, individuals, the specific combination of chronic stimulation with inflammatory mediators and the exposure to joint-specific molecules of the extracellular matrix results in a self-perpetuating cycle of changes that comprise both the permanent activation of tumour-associated signalling pathways and epigenetic changes. In particular, the hypothesis that epigenetic modifications contribute to the activation of RA-FLS has recently gained support by the demonstration that changes in DNA hypomethylation are part of the characteristics of these cells. Thus, in rheumatoid synovial tissue, the ratio of the predominant DNA methyltransferase Dnmt1 to the proliferating cell nuclear antigen (PCNA) are significantly lower than in synovial tissue of degenerative joint disease, osteoarthritis (OA), suggesting that DNA hypomethylation contributes to the aggressive characteristics of RA-FLS .

On the contrary, it has been understood for quite some time that the stable activation of RA-FLS results in changes of early response genes, proto-oncogenes and transcriptional factors, which clearly indicates the activated nature of these cells . An increased expression of some oncogenes such as egr-1, c-fos, ras, raf, sis, myb and myc has been demonstrated in RA patients and activated RA-FLS already 10 years ago and, in the meantime, the functional relevance of the underlying signalling pathways has been clearly shown . Thus, the oncogene ras is predominantly expressed in the synovial lining layer and associated with the expression of the proteolytic enzymes such as cathepsin L at the sites of invasive growth. Conversely, gene-transfer-based inhibition of ras using dominant-negative mutants of ras, raf and myc ameliorated inflammation and reduced bone destruction in adjuvant arthritis as well as cartilage destruction and RA-FLS invasiveness in the SCID mouse model of RA . Although systemic inhibition of aforementioned signalling molecules most likely would result in severely adverse effects, these results demonstrate that targeting kinases that are involved in FLS activation may substantially influence their invasive properties.

Consistent with the over-expression of proto-oncogenes, alterations in expression and function of different tumour suppressor genes have been associated with fibroblast activation and survival in RA. Thus, Firestein and colleagues described somatic mutations of p53 in RA-FLS, which are commonly observed in a variety of tumours . Although such mutations apparently exhibit a great variability and most likely are caused by the inflammatory environment, they may contribute to an extended life span of activated cells aggressively destroying cartilage and bone. This notion is supported also by experiments demonstrating that the inhibition of p53 by gene transfer of p53-inactivating viral proteins can transform normal synovial fibroblast into an RA-like phenotype . Complementing these observations, the pro-apoptotic effector molecule p53 up-regulated modulator of apoptosis (PUMA) is expressed in low levels in RA synovial tissue. Gene transfer with PUMA rapidly increased apoptosis in cultured FLS even though p53 over-expression was ineffective , suggesting that PUMA could be a potential therapeutic target to induce synoviocyte death in RA. These data are supported by studies demonstrating that aggressive RA-FLS lack the expression of the p53-regulated tumour suppressor PTEN (phosphatase and tensin homologue) . As PTEN is an essential mediator of Fas-/CD95-induced apoptosis , these findings provide a link between invasiveness and resistance to apoptosis in activated cells. Their implications are supported strongly by recent data that have demonstrated beneficial effects of the gene delivery of PTEN into rats with collagen-induced arthritis . Moreover, the PTEN data can be taken as additional argument for targeting PI3K pathways in RA-FLS, because PTEN is a critical regulator of PI3Ks.

Resistance to apoptosis

Synovial hyperplasia is characteristic of RA and in addition to the infiltration of inflammatory cells attributed to the proliferation of RA-FLS on the one hand and to the resistance of RA-FLS against apoptosis on the other. As the resistance of RA-FLS against apoptosis has been identified as one key characteristic of these cells, the underlying mechanism has gained significant interest.

Usually, programmed cell death can be induced either by intrinsic mitochondrial pathways as induced by genotoxic stress or triggered through cell surface death receptors which are activated by exogen factors, particularly cytokines . These receptors include members of the tumour necrosis factor (TNF) receptor family. RA-FLS have been demonstrated to express a variety of death-inducing surface receptors of TNF receptor family such as Fas/CD95 , TRAIL-R1 and -R2 and also TNFR1 . Several studies suggest that despite their expression of Fas/CD95, RA-FLS are relatively resistant against FasL-induced apoptosis . The detailed mechanisms by which RA-FLS are protected from receptor-induced cell death are not completely understood. It most likely is a combination of cytokine effects, expression of specific interleukins and alterations in some internal signalling pathways. One pathway triggering apoptosis is the PI3K/Akt axis influencing cell proliferation and survival. So, in line with the above data, it has been demonstrated that in the study by Wang and colleagues intra-articular treatment of CIA rats with adenoviral PTEN resulted in blocking of Akt, a negative regulation of PI3K signalling and an increase in the apoptosis rate in the ankle joints .

Increased expression of soluble Fas (sFas), which through binding of FasL can inhibit FasL-induced apoptosis, in joints of RA patients has been suggested to contribute to their resistance to FasL-induced apoptosis . However, the maintenance of this feature in vitro suggests that the resistance of RA-FLS against programmed cell death can be linked prominently to the function of specific anti-apoptotic molecules such as FLICE inhibitory protein (FLIP) and the small ubiquitin-like modifier 1 (SUMO-1) that are markedly expressed in RA synovium . FLIP inhibits the apoptosis-triggering enzyme caspase 8 . Down-regulation of FLIP sensitises RA-FLS to Fas-mediated apoptosis and may be a valuable tool for targeting RA-FLS hyperplasia . SUMO-1 belongs to a family of small molecules that in a similar way as ubiquitin can modify target proteins. This process is called SUMOylation and can profoundly alter the way of how target proteins are activated, degraded or how they interact with subsequent signalling molecules. Recent data suggest that the increased expression of SUMO-1 and subsequently altered SUMOylation in RA-FLS, for example, of the nuclear promyelocytic leukaemia (PML) results in an altered recruitment of the pro-apoptotic adaptor molecule DAXX and, thus, to decreased apoptosis . Although important aspects of the mechanisms by which SUMOylation events regulate disease processes such as fibroblast activation in RA are unknown, post-translational protein modification not only constitutes an important novel level at which cellular processes are regulated but also offered a new way of therapeutic intervention.

Members of the Bcl family are other potent inhibitors of apoptosis that have been shown to be up-regulated in RA-FLS. They are important regulators of the mitochondrial pathway of apoptosis. Therefore, bcl-2 inhibits one of the terminal steps of apoptosis and it was found that the enhanced expression of Bcl-2 in RA-FLS correlates with the synovial lining thickening and inflammation . Furthermore, stimulation of RA-FLS with IL-15 reduces the expression of Bcl-2 and Bcl-x mRNA, and it was shown by Kurowska and colleagues that the resistance of apoptosis is at least partially related to the autocrine activation of IL-15 receptors by RA-FLS-derived IL-15 . Besides, the anti-apoptotic Bcl family member Mcl-1 is also up-regulated in RA-FLS and counteracts the effects of the pro-apototic intracellular factors Bax, Bak and Bim . The expression of Mcl-1 is induced by TNFalpha and IL-1beta in RA-FLS, and knockdown of Mcl-1 by siRNA induces apoptosis in RA-FLS as well as in synovial macrophages . These data suggest a significant role for Bcl-proteins in regulating the cell death of RA-FLS.

Resistance to apoptosis

Synovial hyperplasia is characteristic of RA and in addition to the infiltration of inflammatory cells attributed to the proliferation of RA-FLS on the one hand and to the resistance of RA-FLS against apoptosis on the other. As the resistance of RA-FLS against apoptosis has been identified as one key characteristic of these cells, the underlying mechanism has gained significant interest.

Usually, programmed cell death can be induced either by intrinsic mitochondrial pathways as induced by genotoxic stress or triggered through cell surface death receptors which are activated by exogen factors, particularly cytokines . These receptors include members of the tumour necrosis factor (TNF) receptor family. RA-FLS have been demonstrated to express a variety of death-inducing surface receptors of TNF receptor family such as Fas/CD95 , TRAIL-R1 and -R2 and also TNFR1 . Several studies suggest that despite their expression of Fas/CD95, RA-FLS are relatively resistant against FasL-induced apoptosis . The detailed mechanisms by which RA-FLS are protected from receptor-induced cell death are not completely understood. It most likely is a combination of cytokine effects, expression of specific interleukins and alterations in some internal signalling pathways. One pathway triggering apoptosis is the PI3K/Akt axis influencing cell proliferation and survival. So, in line with the above data, it has been demonstrated that in the study by Wang and colleagues intra-articular treatment of CIA rats with adenoviral PTEN resulted in blocking of Akt, a negative regulation of PI3K signalling and an increase in the apoptosis rate in the ankle joints .

Increased expression of soluble Fas (sFas), which through binding of FasL can inhibit FasL-induced apoptosis, in joints of RA patients has been suggested to contribute to their resistance to FasL-induced apoptosis . However, the maintenance of this feature in vitro suggests that the resistance of RA-FLS against programmed cell death can be linked prominently to the function of specific anti-apoptotic molecules such as FLICE inhibitory protein (FLIP) and the small ubiquitin-like modifier 1 (SUMO-1) that are markedly expressed in RA synovium . FLIP inhibits the apoptosis-triggering enzyme caspase 8 . Down-regulation of FLIP sensitises RA-FLS to Fas-mediated apoptosis and may be a valuable tool for targeting RA-FLS hyperplasia . SUMO-1 belongs to a family of small molecules that in a similar way as ubiquitin can modify target proteins. This process is called SUMOylation and can profoundly alter the way of how target proteins are activated, degraded or how they interact with subsequent signalling molecules. Recent data suggest that the increased expression of SUMO-1 and subsequently altered SUMOylation in RA-FLS, for example, of the nuclear promyelocytic leukaemia (PML) results in an altered recruitment of the pro-apoptotic adaptor molecule DAXX and, thus, to decreased apoptosis . Although important aspects of the mechanisms by which SUMOylation events regulate disease processes such as fibroblast activation in RA are unknown, post-translational protein modification not only constitutes an important novel level at which cellular processes are regulated but also offered a new way of therapeutic intervention.

Members of the Bcl family are other potent inhibitors of apoptosis that have been shown to be up-regulated in RA-FLS. They are important regulators of the mitochondrial pathway of apoptosis. Therefore, bcl-2 inhibits one of the terminal steps of apoptosis and it was found that the enhanced expression of Bcl-2 in RA-FLS correlates with the synovial lining thickening and inflammation . Furthermore, stimulation of RA-FLS with IL-15 reduces the expression of Bcl-2 and Bcl-x mRNA, and it was shown by Kurowska and colleagues that the resistance of apoptosis is at least partially related to the autocrine activation of IL-15 receptors by RA-FLS-derived IL-15 . Besides, the anti-apoptotic Bcl family member Mcl-1 is also up-regulated in RA-FLS and counteracts the effects of the pro-apototic intracellular factors Bax, Bak and Bim . The expression of Mcl-1 is induced by TNFalpha and IL-1beta in RA-FLS, and knockdown of Mcl-1 by siRNA induces apoptosis in RA-FLS as well as in synovial macrophages . These data suggest a significant role for Bcl-proteins in regulating the cell death of RA-FLS.

Cytokines and signalling pathways

The chronic exposure to inflammatory cytokines and growth factors most likely belongs to the key factors that mediate the transformation of mesenchymal cells or their progenitors into stably activated RA-FLS. Under normal conditions, the secretion of most cytokines is a very short and tightly regulated process, and there is a balance between pro- and anti-inflammatory cytokines. However, in chronic inflammatory conditions such as RA, inflammatory cytokines are produced at constitutively high levels and there is a massive shift towards inflammatory and catabolic cytokines ultimately leading to joint destruction. Majority of these cytokines are derived from inflammatory cells such as macrophages or lymphocytes, and prominent examples are TNFalpha, IL-1 and interferon-gamma. Several studies have shown that these cytokines contribute to the aggressive behaviour of RA-FLS in RA by inducing the expression of adhesion molecules and matrix-degrading enzymes. The notion that inflammatory cytokines play a central role in cellular activation during chronic destructive arthritis is supported by several in vivo models in which the constitutive over-expression of inflammatory cytokines results in an RA-like disease. The human TNFalpha transgenic (hTNFtg) mouse is a most prominent example , and this model has been used widely to study the mechanisms of inflammatory cartilage and bone destruction by fibroblast-like cells . The activation of RA-FLS, in turn, also induces the synthesis of pro-inflammatory mediators including cytokines, growth factors and lipid mediators that act in an autocrine and paracrine fashion to further stimulate RA-FLS as well as other cells, thus promoting inflammation, angiogenesis, chemoattraction and dysbalanced tissue homeostasis in a vicious circle . IL-1, IL-6, IL-8 and TGF-beta constitute prominent examples of cytokines that act on RA-FLS and are produced by RA-FLS themselves . IL-17, a major Th1 cytokine produced by activated T cells, has been identified as an important mediator of the induction of IL-6 and IL-8 in RA-FLS via activation of the PI3K/Akt signalling pathway . The chemokine ligands CCL2, CCL5 and CXCL12 also enhance IL-6 and IL-8 production by RA-FLS and their corresponding receptors CCR2, CCR5 and CXCR4 . IL-6 is an acute-phase protein engaged in proliferation and differentiation of immune cells and the augmentation of bone erosion through activation of osteoclast precursor cells . Due to the broad spectrum of IL-6 effects on the pathogenesis of RA, inhibition of IL-6 signalling through administration of antibodies against the IL-6R has been established as a new way of therapeutic intervention . Although some questions regarding the mechanisms and effects of anti-IL-6R antibodies in RA remain open, several lines of evidence suggest that RA-FLS are important target cells of this novel treatment. Most recent data suggest that in addition to IL-6, other members of the IL-6 family such as oncostatin M (OSM) are involved in the specific activation of RA-FLS. Thus, it was demonstrated by Hintzen and colleagues that OSM induces the expression of MCP-4 (CCL13) in RA-FLS but not in control fibroblasts . These data not only confirm the unique phenotype of RA-FLS but also contribute to the notion that RA-FLS are involved strongly in the perpetuation of the disease process in RA. They also point to alternative targets for cytokine inhibition in the disease.

In a similar way as IL-6, production of IL-8 by RA-FLS contributes to the perpetuation of disease. This is because IL-8 not only promotes the recruitment of neutrophils and dendritic cells but also has strong pro-angiogenetic activity by induction of blood vessel formation. Both processes ultimately promote neovascularisation and influx of immune cells into inflamed synovia and thus result in the maintenance of synovitis.

In this context, the direct interaction of RA-FLS with inflammatory cells in the inflamed synovium, and particularly the interaction of CD40L (CD154) with CD40 on RA-FLS, has increased interest. Thus, it has been demonstrated that CD40 is expressed also on activated RA-FLS and that activation of CD40 causes the release of growth factors such as vascular endothelial growth factor (VEGF) and chemokines such as SDF-1/CXCL12 by RA-FLS . Further, the influx of CD4 + T cells into the proliferating synovium is enhanced by RA-FLS because of their production of CXCL16 and IL-16 . RA-FLS are also an important source of cytokines with IL-2-like activity, IL-15 and IL-7, in RA joints. IL-15 may be mainly responsible for local T-cell activation and expansion in the presence of deficient IL-2 production by T cells . RA-FLS have been also implicated in direct attraction and accumulation of B cells in inflamed tissue. For example, SDF-1 secretion by RA-FLS promotes B-cell migration and activation to the synovium during arthritis . The survival of synovial B cells at the site of inflammation in RA is increased through induction of expression of mitochondrial apoptosis inhibitors such as Bcl-X(L) and interaction of VLA-4 with VCAM-1 expressed on RA-FLS . In turn, B cells promote RA-FLS activation through IgG binding to the high-affinity FcgammaR (FcgammaR I) on RA-FLS. This interaction increases the production of IL-16 and RANTES by RA-FLS provoking chemo-attractants of T cells . In summary, the interaction between mesenchymal, immune and inflammatory cells is mediated by a tightly regulated, extremely complex network of cytokines that results in the maintenance of inflammation. In this network, RA-FLSs are far more than passively reacting cells but contribute significantly to the perpetuation of disease and to the switch from acute to chronic and destructive inflammation ( Fig. 1 ).

In RA synovium, RA-FLS interact with different cell types promoting inflammation and joint destruction . Within the synovial lining, the cell-to-cell adhesion molecule cadherin-11 links RA-FLS to other RA-FLS. This interaction is critical for synovial lining formation. β1-integrins enable cell-extracellular matrix (ECM) interactions of RA-FLS in the synovial lining. Furthermore, RA-FLS interact within the synovial lining with macrophages by secretion of soluble factors (especially IL-1, TNFalpha and IL-6) and cell surface receptor interactions for promoting their recruitment and activation. RA-FLS can bind via ICAM-1 to β2-integrins and via VCAM-1 to VLA-4 expressed on macrophages.

Sublining RA-FLS interact with a large number of different cell types including T cells, B cells and macrophages resulting in their recruitment, activation and differentiation. This process is mediated by both cell surface receptor interactions and different secreted factors in autocrine as well as paracrine mechanisms. T cell-RA-FLS interactions include recruitment and retention by secretion of SDF-1 for example, and T cell activation via costimulatory receptors such as ICAM-1 and the release of cytokines. Among others, IL-6 and IL-15 are important for differentiation of the Th17 T cell subset. B cells are also recruited and retained by SDF-1 release of RA-FLS and by interactions through cell surface adhesion (e.g. VLA-4 and VCAM-1). Macrophages in the synovial sublining support the activation of RA-FLS by secretion of cytokines like IL-1 and TNFalpha.

The fact that systemic inhibition of inflammatory cytokines leads to impressive and sustained improvement not only of synovitis but also of joint destruction has shown convincingly that interfering with cytokine-triggered pathways is a most interesting approach also with respect to targeting synovial fibroblasts. In this context, several studies show that inhibiting the inflammatory cell response highly up-stream in the signalling cascade could be a feasible approach. In particular, a number of cytokines activate adapter proteins through binding of receptor complexes. For example, formation of TNFalpha/TNFR1 complexes results in the recruitment of TNF receptor associated factor 2 (TRAF 2), and it could be shown that JAB1, a cytoplasmatic regulator of the phosphorylation of c-Jun, is required for ubiquitination of this factor that allows as well further signal transduction as pro-survival responses of RA-FLS .

A number of inflammatory cytokines such as TNFalpha, IL-1 beta and IL-6 activate the mitogen-activated proteinkinases (MAPKs) c-Jun, ERK and p38, all of which constitute important integrating points of incoming inflammatory signals. Although there is considerable complexity and overlap in and among these three kinases and their respective signalling pathways, each of these kinases can be assigned a distinct role in maintaining synovial inflammation: for example, phosphorylation and activation of p38MAPK mainly triggers cytokine production, c-Jun triggers MMP expression and joint destruction and stimulation of ERK influences RA-FLS proliferation and migration . Besides, there are also other effectors such as heat-shock protein 90 (Hsp90), hypoxia, osmotic stress and DNA damage that can modify the protein kinases’ activity. Therefore, the development of inhibitors to these kinases has been a major focus of several drug development programmes. However, initial studies have revealed various side effects due to the important role of MAPKs in a variety of cellular processes such as tissue homeostasis and regulation of immune responses. The development of more specific blockers and their improved bioavailability has resulted in the testing of new promising substances in animal models of inflammatory arthritis. For example, it could be shown that the p38MAPK inhibitor FR167653 reduced serum levels of TNFalpha and IL-1 beta and prevented CIA rats from joint disability . Similarly, blocking of ERK through FR180204 led to an improvement of CIA .

In addition to blocking MAPKs, strategies have been developed that prevent the activation of the transcription factor NF-kappaB, which triggers various immune and inflammatory processes during RA. Based on the central role of the IkappaB kinase (IKK) complex consisting of IKKalpha, IKKbeta and IKKgamma in activating NF-kappaB, several studies were performed to validate this principle and identify specific inhibitors against these kinases. Mbalaviele et al. developed the specific inhibitor PH-408 of IKK-2 that leads to suppression of IkappaB alpha phosphorylation and degradation . Furthermore, it could be shown that ML-120B, a potent, highly selective and reversible inhibitor of IKKbeta blocked various NF-kappaB-regulated cell responses important during inflammatory joint destruction including the expression of matrix metalloproteinases (MMPs), monocyte chemotactic protein-1 (MCP-1) and prostaglandin (PGE2) . Thus, inhibition of activating pathways in RA-FLS constitutes an important element of strategies that aim at interfering with the IKK–NF-kappaB pathway and it may be envisaged that NFkappaB inhibition will reduce fibroblast-mediated joint destruction when applied at sufficient levels. This idea is supported also by several studies demonstrating that RA-FLS have the ability to influence activation and differentiation of macrophages into osteoclasts by producing the receptor activator of NF-kappaB (RANKL) . Despite this general concept it is unlikely that effective systemic inhibition of NF-kappaB can be achieved without considerable adverse effects, and some recent early trials with NF-kappaB inhibitors appear to support this notion. Therefore, it will be of interest to see if local administration of inhibitors to the IKK-NF–kappaB pathway may be of value to treat disease progression in individual joints and perhaps even have systemic benefits.

Adhesion molecules

One crucial consequence of the activation of RA-FLS is the expression and up-regulation of different adhesion molecules that mediate the attachment of these cells to the cartilage. The adhesion to cartilage is an important prerequisite and one of the initial steps for the destruction of extracellular matrix by RA-FLS. In this context, integrins have been associated most strongly with the attachment of RA-FLS to cartilage and also bone. Integrins are heterodimers of a number of different alpha and beta chains, and the expression of beta1-integrins is highly increased on RA-FLS . As beta1-integrins facilitate cell–matrix interactions, for example, by acting as fibronectin receptors, it has been suggested that the fibronectin-rich environment of RA cartilage surface permit the adhesion of RA-FLS to the cartilage. The high expression of CS-1, a spliced isoform of fibronectin, in the RA synovium supports this notion . Consequently, antibody blockade of specific beta1-integrins CD49dCD29 (alpha4beta1) and CD49eCD29 (alpha5beta1) as well as application of beta1 antibodies was shown to inhibit the invasion of RA-FLS into cartilage matrix in vitro . These data suggest that inhibition of integrins may be a promising strategy to interfere with the attachment of RA-FLS to extracellular matrix and consequently with cartilage destruction. Such strategy is further supported by the fact that integrins not only function as ‘glue-like’ receptor molecules for extracellular matrix, but also interact with several signalling pathways involved in the activation of RA-FLS . The expression of early cell-cycle genes such as c-fos and c-myc is specifically stimulated by integrin-mediated cell adhesion . Furthermore, some members of the Ig superfamily of adhesion molecules are expressed within the synovial membrane. Vascular adhesion molecule 1 (VCAM-1) expression is up-regulated in the subpopulation of activated lining fibroblasts and can be induced by different pro-inflammatory cytokines . In addition, the intercellular adhesion molecule 1 (ICAM-1) is expressed on RA-FLS and facilitate the interaction with leucocytes. The role of ICAM-1 has been studied in several animal models of inflammatory arthritis. Antibodies specific for ICAM-1 block AIA in rats and CIA in mice , and ICAM-1-null mice exhibit a decrease susceptibility to collagen-induced arthritis .

To understand the mutual interactions between cell adhesion and matrix degradation, several studies have been performed to identify the respective mediators and effectors. So it is more than 10 years ago that the ECM glycoprotein, tenascin, has been identified to mediate cell adhesion and migration, stimulated by IL-1 . These investigations revealed that the expression is localised in the areas of inflammation and tissue damage. Interestingly, these data have gained substantial functional impact recently, when it was shown by Midwood and colleagues that tenascin knockout mice are protected from erosive arthritis . This study is of particular importance not only because it classifies the protein as an important mediator of cell–matrix interactions during rheumatoid joint destruction but also because it demonstrates that tenascin-c is another endogenous ligand of the toll-like receptor 4 and thus functionally links the activation of RA-FLS to innate immunity. In line with the general concept that cell–matrix interactions are of importance for fibroblast transformation in RA, the hyaluronate receptor CD44, which is a specific adhesion receptor for the extracellular matrix glycoprotein hyaluron, has been found to be up-regulated in RA-FLS . In particular, Bauer et al. detected a possible link in synovial fibroblast adhesion and ECM degradation when they investigated the expression of the fibroblast activation protein (FAP). From the high expression of this type II cell surface serine protease in RA-FLS and its co-localisation with two CD44 splicing variants v3 and v7/v8, the authors conclude that FAP may be a new mediator of cartilage destruction in RA .

The question of how RA-FLSs interact with the extracellular matrix of both cartilage and the surrounding synovial membrane has recently gained interest when it was shown that synovial fibroblasts contribute to the spreading of RA from one joint to another . As demonstrated by Lefévre and colleagues in the SCID mouse model of rheumatoid cartilage destruction, RA-FLSs are capable of detaching from one site, migrating through the blood stream and then re-initiating the destructive process at another, distant site. While many aspects of this process remain unclear, it is very evident that the re-arrangement of cell–matrix and cell–cell contacts is needed to achieve a metastasis-like spreading of RA-FLS. In the context of cell-to-cell contacts, the role of cadherin-11 on RA-FLS has been established recently and cadherin-11 has been identified as a key regulator of cell invasion through mediating homophilic, calcium-dependent cell-to-cell adhesion . As shown by Kiener and colleagues, interactions of RA-FLS through cadherin-11 in the lining layer contribute significantly to pannus formation, whereas the absence of cadherin-11 in mice results in a hypoplastic synovium, characterised by both reduced lining cell number and extracellular matrix . While the specific utility of these findings for the development of novel treatment approaches targeting RA-FLS remains to be established, these data clearly demonstrate that in addition to interfering with inflammatory mediators, cell–cell and cell–matrix interactions are keys to understanding and potentially modulating the disease process in RA.