In a healthy joint the synovium covers the joint cavity and regulates the transport of nutrients and other molecules between the joint cavity and the adjacent tissue. The synovium consists of few cell layers of fibroblast-like synoviocytes and macrophage-like synoviocytes 1. Of these, the fibroblast-like synoviocytes, also termed synovial fibroblasts (SF), synthesize and secrete a rich but balanced variety of products, including cytokines, matrix metalloproteinases (MMPs), hyaluronan and proteoglycans into the synovial fluid.

In joints affected by rheumatoid arthritis (RA), the synovial membrane becomes hyperplastic due to the proliferation of genuine synovial cells such as SF 2 and a massive infiltration of inflammatory cells 3. As a consequence, the inner layer of the synovium, the lining layer, increases in size up to 10 cell layers and more. Similarly, in the normal state the lining layer contains only a small number of blood vessels, and oxygenation and nutrition is facilitated by the blood vessels from the sublining. In diseased synovium, the proliferation of cells (for example, SF) and the infiltration of blood-borne cells (for example, macrophages, B cells, T cells, plasma cells) subsequently result in hypoxic conditions in the tissue because of the increasing distance to a blood vessel and the increased demand for oxygen in the hyperplastic tissue. Neovascularization is thus a prerequisite in the formation and maintenance for the pannus, and intensive neovascularization with blood vessels close to the ultimate lining layer can be observed 45. In addition, there is a close association between rheumatoid synovitis and the formation of complex lymphoid microstructures 6.

At the site of invasion into the adjacent cartilage and bone, the pannus consists mainly of activated fibroblasts. SF mediate also the perichondrocytic cartilage degradation and promote bone destruction by influencing osteoclastogenesis in cooperation with macrophages 7.

As summarized in the present review, numerous researchers have addressed the details of this transition of healthy tissue to diseased synovial tissue. One of the key examples is the work by Steenvorden and colleagues, who have shown that the original epithelial-like phenotype of SF is replaced by a cell showing mesenchymal/fibrotic characteristics, which includes the expression of collagen type I and α-smooth muscle actin 8.

As outlined below and shown in Figure 1, the stimulating factors for this development, which have been examined intensively and have revealed numerous novel aspects in the past year, are cell-derived microparticles, hitherto unknown cytokines and chemoattractive molecules.

Microparticles are a heterogeneous population of small membrane-coated vesicles that can be released from all cell types, including macrophages, monocytes, epithelial cells as well as B cells and T cells. Microparticles in synovial fluid were first described by Berckmans and coworkers, who showed that these particles originate mainly from monocytes and granulocytes 9. The potential function of microparticles in inflammation and as part of mechanisms of the innate immunity was recently reviewed by Distler and colleagues 10.

Microparticles emerge by budding from their parental cells upon apoptosis or activation. The composition of the membrane of the microparticles therefore depends on the cell type of origin. Microparticles inherit all characteristics of the parental cell, including the respective cell surface molecules and receptors, and can therefore act as mediators of cellular interactions. Whether intracellular contents such as cytosolic or nuclear proteins are present within the microparticles or even contribute to their biologic activity remains largely unclear.

In this context, recent findings shed light on these phenomena by indicating that microparticles derived from leukocytes can play a role in inflammatory arthritis by inducing the synthesis of MMPs, chemokines and cytokines in SF 1112. It could also be shown that, in particular, the synthesis of MMP-1, MMP-3, MMP-9 and MMP-13 was strongly induced by microparticles but the expression of MMP-2, MMP-14 and the tissue inhibitors TIMP-1, TIMP-2 and TIMP-3 was unaffected 13. Moreover, in the same study it could be demonstrated that microparticles increased the synthesis of IL-6, IL-8 and the monocyte chemoattractant proteins MCP-1 and MCP-2. As was demonstrated recently, bound complement components and activator molecules are present on microparticles ex vivo 14. In RA synovial fluid, therefore, microparticles might modulate the increased complement activation.

Increasing evidence was provided in 2006 that the more recently discovered 'novel' cytokines are also involved in promoting joint inflammation in RA. For example, IL-32, which is intensively expressed in RA synovial tissue, resulted in joint inflammation and a mild cartilage damage when injected intraarticularly in murine knee joints 15. IL-32 was first described by Kim and colleagues 16. They demonstrated that IL-32 was able to induce the expression of TNFα, IL-1β, IL-6 and several other chemokines in a human acute monocytic leukemia cell line (THP-1), for example, through the cytokine signaling pathways of NF-κB and p38 mitogen-activated protein kinase (MAPK). IL-32 can therefore be considered a proinflammatory mediator in RA.

IL-1F8, a new member of the IL-1 family that is known to play a pivotal role in immune and inflammatory reactions, also exerts proinflammatory effects in primary human joint cells 17. Another 'novel' cytokine, IL-17, which is synthesized primarily by T cells and exhibits proinflammatory activities, has also been associated with RA. Interestingly, it could be shown that IL-17, which is a potent inducer of TNFα and IL-1β, acts independently of TNFα in RA and was able to enhance inflammation and cartilage damage in a TNF-deficient mouse model 18.

A new member of the IL-10 family, IL-20, is known to play a role in skin inflammation and the development of hematopoetic cells. The potential role of IL-20 in RA and arteriosclerosis was recently analyzed by Wei and coworkers 19. In this context, IL-20 was shown to be upregulated in the synovial fluid of RA patients. Furthermore, in a collagen-induced arthritis model in rats it could be shown that both IL-20 and its receptor IL-20RI are present, which confirmed the involvement of IL-20 in the pathogenesis of RA 20. IL-21 is a CD4⁺ T-cell-derived cytokine being involved in innate and adaptive immune response. Overexpression of IL-21 and its receptor IL-21R could also be identified in the inflamed synovial membrane and in synovial fluid leukocytes of RA patients 21. Stimulation of peripheral-blood T cells or synovial-fluid T cells isolated from RA patients with IL-21 resulted in enhanced T-cell activation, proliferation and secretion of proinflammatory cytokines, including TNFα and IFNγ. Furthermore, it is known that activated macrophages produce central inflammatory cytokines such as TNFα and IL-1. As the anti-inflammatory cytokine IL-10 suppresses the macrophage-dependent synthesis of both TNFα and IL-1 in nonmalignant conditions, it was most interesting to see that the responses to IL-10 are dysregulated in RA macrophages, resulting in an inefficient suppression of inflammation 22.

Besides their role in energy metabolism, cytokines derived from adipocytes (for example, adiponectin and resistin) appear to play a pivotal role in the pathogenesis of RA. A strong stimulatory effect of adiponectin on rheumatoid arthritis synovial fibroblasts (RASF) could be detected. Hereby, adiponectin induced IL-6 and MMP-1 in a p38 MAPK pathway-dependent manner 23. Similarly, it could be shown that an increased concentration of adiponectin in the synovial fluid of RA patients is negatively correlated with the local inflammatory process 24. Resistin, another so-called adipokine, may also influence proinflammation in RA. For example, an upregulated concentration of resistin was found at local sites of inflammation in arthritis, and the serum resistin levels correlated with inflammation and the activity of the disease in RA 25. In contrast, circulating levels of the prototype adipocytokine leptin appear not to have a correlation with RA activity 2627.

Chemokines are small chemotactic proteins that play a role in the migration of circulating cells into tissue and migration of cells within the tissue. As recently reviewed by Vergunst and colleagues and by Tarrant and Patel, chemokines play a substantial role in the inflammatory process of RA by promoting leukocyte trafficking into the synovium 2829. The regulation of chemokine ligand CCL18, a T-cell-attracting chemokine, was described by van Lieshout and coworkers 30. These authors showed that IL-10 in combination with IL-4 and IL-13 induced synergistically the secretion of CCL18 in monocytes and monocyte-derived cells. This finding supported the idea that CCL18 is involved in the regulation of the immune system in health and disease.

In several diseases including RA or osteoarthritis, however, chemokines and their receptors are considered potential future therapeutic targets. Based on this idea, a recent study by Haringman and colleagues investigated the expression of the ligands of chemokine receptors CCR1 and CCR5 in the inflamed synovium 31. They found an abundant expression of both receptors CCR1 and CCR5 in the synovial tissue of RA patients, whereas the percentages of CCR1-positive and CCR5-positive monocytes in the peripheral blood of RA patients were found to be decreased. The blockade of CCR1 and CCR5 could therefore be part of an effective future therapy for RA.

'Receptor cells' of the disease-promoting factors outlined above are mainly synovial fibroblasts and macrophages, which are also the predominant cell types in the inflamed synovium 832. With regard to SF it is not known what initiates the initial proliferation of these cells in the early stages of RA, but this pivotal event can occur prior to the onset of inflammation 33. In this regard, the investigation of the mode of proliferation revealed an upregulation of the metastatic lymph node MLN51 gene in hyperactive RASF 34. Even growth-retarded SF showed a significant up-regulation of MLN51 when treated with granulocyte–macrophage colony-stimulating factor or with synovial fluid. As MLN51 was originally identified in breast cancer, this observation once more emphasizes distinct similarities of the mechanisms of cellular activation in RA and in malignant diseases.

Besides unrestricted proliferation, the increasing number of RASF in the synovial lining layer may be also due to an altered apoptosis. It is known that the deficiency or the lack of tumor suppressor genes such as p53, the 'phosphatase and tensin homolog deleted on chromosome 10' PTEN, small ubiquitin-like modifier and p21 leads to long-term cell growth, to extended survival and potentially to tumor formation. Woods and colleagues demonstrated that the cell-cycle inhibitor p21 is significantly reduced in RA synovial lining, particularly in RASF. In addition, p21 is able to repress migration of SF – and, vice versa, loss of p21, which occurs also in RASF, may contribute to the excessive invasion and extended survival of these cells 35. Moreover, although overexpression of p53 is found in RA synovial tissue, only few synoviocytes undergo apoptosis 36. This effect could be explained in part by a low expression of proapoptotic genes. In a study from Cha and coworkers using synovial tissue and SF, it could be shown that a deficient p53-upregulated modulator of apoptosis can inhibit apoptosis of SF 37. Furthermore, the lack of the 'phosphatase and tensin homolog' PTEN in the RA synovial lining was able to contribute to the survival of RASF at sites of destruction 38. Connor and coworkers showed also that this phenomenon could be due to the PTEN-dependent effect on IκB/NF-κB interactions and other nuclear factors (for example, akt/protein kinase B).

As recent data indicated a role of protein geranylgeranylation and RhoA/RhoA kinase blocking in regulation of apoptosis, Nagashima and colleagues suggested lipophilic statins as therapeutic agents for RA, since they are able to induce apoptosis in RASF; for example, through mitochondrial-dependent and caspase-3-dependent pathways and the inhibition of mevalonate pathways 39. Moreover, the antiapoptotic molecule myeloid cell leukemia Mcl-1, which is known to be critical for the survival of T lymphocytes and B lymphocytes and of macrophages 40, appears also to be relevant in the survival of RASF 41.

TNF, being one of the key molecules in driving the inflammatory process in RA synovium, is also linked directly to SF apoptosis. For example, a study conducted by Wang and colleagues 42 revealed that the antiapoptotic effect of TNFα in RASF is regulated by the Jun activating binding protein JAB1, because specific knockdown of JAB1 with an antisense RNA construct resulted in TNFα-induced apoptosis response in RASF. Moreover, Wang and coworkers showed that this antiapoptotic signaling might be due to a JAB1-mediated ubiquitination of TNF-receptor-associated-factor 2. The potential role of the TNF ligand receptor superfamily in the antiapoptotic pathways in RA was recently reviewed by Hsu and colleagues 43.

In 2006 numerous groups supported the idea of RASF being key players in the pathogenesis of RA 83244. For example, a recent study showed that the expression of the extracellular matrix metalloproteinase inducer CD147 was more intensively expressed on RASF than on osteoarthritic SF 45. The authors concluded that the increased expression of CD147 might be responsible for both the elevated secretion of MMPs and the invasive potential of SF. Of the subsequently activated family members, the collagenases MMP-1 and MMP-13, the gelatinases MMP-2 and MMP-9, the stromelysin MMP-3 and the membrane-type MMPs can be found in active RA synovium. Of these, the expression of MMP-3 mRNA is higher in diseased RA pannus tissue compared with adjacent nondiseased RA synovium 46. Most strikingly, although MMP-1 appears to have a function in degrading cartilage collagen type II, it does not appear to derive from pannus tissue but to be secreted by chondrocytes 46. In addition, in vitro inhibition of the membrane type I MT1-MMP with an antisense RNA construct resulted in a significant reduction of cartilage degradation by RASF 47. Also, a study by Bauer and colleagues addressing the expression of fibroblast activation protein by RASF revealed that the expression of FAP is colocalized with MMP-1 and MMP-13, indicating that fibroblast activation protein might be an additional factor in cartilage and bone destruction in RA joints 48.

An effective future therapy for RA could be the selective inhibition of MAPK kinases 49. MKK3 and MKK6 play key roles in the activation of p38 MAPK, which in turn upregulates the expression of cytokines and MMPs in SF. Inoue and colleagues investigated the potential of MKK3 as a therapeutic target. They could show that MKK3 deficiency significantly decreases synovial inflammation and cytokine production in a mouse model of arthritis.

As recently reviewed by Ruocco and Karin, the Iκβ kinase IKKβ is essential for the inflammatory cytokine-induced activation of NF-κB 50. Blocking of IKKβ could therefore be part of a therapeutic strategy for the treatment of inflammation. In this context, it was demonstrated by Wen and coworkers that the inhibition of IKKβ with the β-carboline derivative ML120B inhibits NF-κB signaling in human SF, chondrocytes and mast cells 51. Moreover, it could be shown that ML120B administration reduces NF-κB activity in rats with induced polyarthritis 52.

RASF play an important role in osteoclast formation 53. The molecular basis for this property is the synthesis of the ligand for the receptor activator of nuclear factor β (RANKL) 54. Binding of receptor activator of nuclear factor (RANK) with its ligand RANKL regulates the differentiation of bone-resorbing osteoclasts from monocytes/macrophages progenitor cells. In addition, Lee and coworkers revealed that RASF produce actively RANKL, and thus are part of the RANK/RANKL interaction system 55. Interestingly, a study by Pettit and colleagues demonstrated a focal RANKL, RANK and osteoprotegerin expression in the RA bone microenvironment 56. Taking these results together, RASF most probably perpetuate actively osteoclastogenesis and bone destruction in RA.

In synovial homeostasis, the physiologic function of macrophages is the induction and regulation of inflammation after infection. Similar to RASF, macrophages are key players in promoting inflammation and joint destruction in RA by secreting proinflammatory cytokines such as IL-1 and TNFα and by the induction and perpetuation of osteoclastogenesis 57. Macrophages migrate out of the bloodstream as monocytes and accumulate in the synovial membrane. As demonstrated recently, this migratory capacity is dependent on distinct enzymes. Miyata and colleagues showed that, in contrast to osteoarthritis, patients with RA have a significant increase in cathepsin G activity 58. Interestingly, cathepsin G was able to induce the migration of monocytes in a microchemotaxis chamber and thus cathepsin G appears to promote synovial inflammation in addition to the hydrolytic function of cathepsins in matrix degradation.

With regard to cellular accumulation, Gregory and coworkers showed that macrophage migration inhibitory factor induces the release of CC chemokine ligand 2 from primary microvascular cells 59. This function of macrophage migration inhibitory factor might therefore further promote the pathogenesis of RA by inducing monocyte migration into the synovium.

Activation of macrophages in the RA synovium can take place by several mechanisms 57; for example, activation by T cells that secrete stimulatory cytokines such as IFNγ and IL-2. Direct cell–cell contact between macrophages and T cells can also result in macrophage activation. In a recent study from Beech and coworkers it was shown that RA synovial T cells are able to induce the chemokine production by monocytes in a cell-contact-dependent manner 60. Moreover, there appears also to be a correlation between B cells and macrophage activation. Treatment of RA patients with rituximab, a chimeric antibody against CD20-expressing B cells, resulted in a significant decrease of TNFα in the supernatant of isolated human monocyte-derived macrophages 61.

Not only fibroblasts but also macrophages appear to be resistant to apoptosis, thereby increasing further the number of macrophages in the synovium. In this context, the antiapoptotic B-cell leukemia Bcl-2 family member Mcl-1 may contribute to the survival of these cells. This was supported by a recent study from Liu and coworkers who revealed an increased expression of Mcl-1 in CD14⁺ macrophages derived from the synovial fluid of RA patients. Furthermore, the same group was able to show the induction of apoptosis in synovial macrophages by blocking the PI 3-kinase/Akt-1 or STAT-3 pathways 40.

Macrophages/monocytes are not only involved in inflammatory reactions, but also in the remodeling processes of the bone. Two studies published in the past year provided new insights in the role of monocytes in osteoclast formation. CD14⁺ synovial macrophages isolated from patients with osteoarthritis, RA and pyrophosphate arthropathy have been shown to differentiate into osteoclasts when treated with RANKL 62. Stimulation with TNFα and IL-1α resulted in osteoclast formation of macrophages from RA and pyrophosphate arthropathy patients. Another study confirmed the involvement of monocytes as potential precursors of macrophages and osteoclasts. Komano and colleagues revealed that CD16 monocytes, a subset of human peripheral blood monocytes, bear the potential to differentiate into osteoclasts when stimulated with RANKL and macrophage colony-stimulating factor 63.

In 2006, different groups investigated the potential role of the tyrosine kinase inhibitor imatinib in the treatment of RA 646566. For example, Ando and colleagues showed that imatinib inhibits the proliferation of macrophage colony-stimulating factor-dependent osteoclast precursor cells and the formation of osteoclasts in vitro 64. Moreover, they showed that the administration of imatinib suppressed joint destruction in a collagen-induced arthritis model in rats. Similarly, it was demonstrated that imatinib enhances osteoclast apoptosis in a cell culture model using rabbit osteoclasts 66. A recent study by Paniagua and coworkers demonstrated that, in a collagen-induced arthritis model in mice, imatinib affects the proliferation of B cells and monocytes/macrophages, and inhibits several tyrosine kinases that are directly implicated in the pathogenesis of RA 65. Taken together, the selective inhibition of tyrosine kinases by imatinib could be a promising future therapy for RA.