Osteoclasts and Arthritis


  • The authors state that they have no conflicts of interest

  • Published online on May 25, 2009


Inflammatory arthritis typically presents as joint pain, swelling, and stiffness, which may be of short duration. For example, parvovirus B 19 infection, attended by painful polyarticular swelling, generally resolves spontaneously within 2 wk without joint destruction. Chronic manifestations of these diseases, however, typically represent a self-perpetuating hyperimmune, pathological phenomenon and yield progressive structural damage. Rheumatoid (RA) and psoriatic arthritis, in particular, are joint destructive consequent to chronic inflammation of the synovial membrane and its bone-associated ligaments and tendons.


The dominance of bone destruction in inflammatory arthritis is best documented in RA, a condition affecting roughly 1% of western society.(1) It is characterized by painful, symmetrical swelling, particularly of small, peripheral joints. The disease is insidious, appearing at any age, with woman most frequently afflicted. Autoantibodies, such as rheumatoid factor and those recognizing citrullinated proteins, develop in most patients. The skeletal consequences of RA are usually evident by conventional radiography, manifesting as periarticular bone erosions and systemic osteoporosis, the latter a common component of virtually all forms of chronic inflammatory disease, including those that are extraskeletal.(2,3) The clinical importance of local bone erosions is underscored by their association with impaired joint function.(4,5) As bone damage accumulates in a rheumatoid joint, the likelihood of full functional restoration diminishes, even in the face of optimal therapy.(6) Although these lesions are usually more pronounced with time, initial bone erosions emerge early after disease onset. By 6 mo, periarticular osteolysis is evident in approximately one half of RA patients(7) and, in fact, erosions may appear within <12 wk.(8)


The rapid degradation of periarticular bone is achieved by osteoclasts, the unique skeletal resorptive cells derived from monocyte/macrophage precursors. The initial description of osteoclasts in the inflamed joint dates to the 19th century,(9,10) but the first detailed molecular characterization of these cells in RA patients was published by Gravallese et al. in the late 1990s.(11) They showed that osteoclasts, as multinucleated cells expressing CD68, TRACP, cathepsin K, and the calcitonin receptor, appear at the interface of inflammatory synovial tissue and periosteal surface and eventually form resorption lacunae. Mononuclear macrophage-lineage cells, which are less differentiated but committed to the osteoclast phenotype, also migrate to inflamed joints, predominantly in the vicinity of, but not necessarily directly attached to, the bone surface. These observations establish that osteoclasts populate the chronically inflamed joint, in abundance, and are the product of local differentiation of their precursors. Osteoclasts also appear in the joints of patients with psoriatic arthritis, confirming these cells are a common component of multiple forms of chronic inflammatory joint disease.(12)

Joint anatomy likely dictates location of bone erosions. Such lesions, for example, usually localize where the synovial membrane inserts into the periosteal surface and at the edges of bones, which confine the joint space. This area is also close to the proximal rim of articular cartilage and the attachment sites of ligaments and tendon sheaths. These anatomical properties provide insight into the erosive mechanism because they indicate how inflamed synovial membrane, as well as that lining tendon sheaths, contact the bone surface.(13) These features also explain why bone erosions typically occur at the radial sites of the metacarpophalangeal joints juxtaposed to tendons, collateral ligaments, and insertion sites of the synovial membrane.(14) Animal models of arthritis also confirm that osteoclasts and erosions initially develop where inflamed tendon sheaths pass bony joint edges.(15)

The presence of osteoclasts within resorptive lacunae provides compelling circumstantial evidence these cell are responsible for structural damage attending inflammatory arthritis. This concept conflicts with the hypothesis that fibroblast-like synoviocytes have the capacity to directly degrade bone through metalloproteinase production.(16) These activated mesenchymal cells do invade and destroy superficial, nonmineralized articular cartilage. However, experiments using arthritic osteopetrotic mice, in whom osteoclast formation or function is compromised, established that degradation of bone or mineralized cartilage is mediated by the bone-resorptive polykaryon. In fact, failure of osteoclasts to resorb cartilage is responsible for the pathognomonic histological feature of osteopetrosis, namely persistence of cartilaginous “bars” in metaphyseal bone.(17,18) Osteopetrotic animals protected from destruction of bone, but not of superficial cartilage, in face of severe joint inflammation include mice lacking the key osteoclastogenic cytokine, RANKL, or its receptor, RANK, and those deleted of the essential osteoclast-forming transcription factor, c-fos. Thus, abrogation of joint destruction in rheumatoid patients requires direct or indirect inhibition of osteoclast formation or function.


Osteoclastogenesis is an early event in inflammatory joint disease as committed mononuclear precursors appear within 2 days, and mature resorptive polykaryons within 5 days, of induction of adjuvant- or collagen-induced arthritis.(19) This rapid and pronounced osteoclast formation in inflamed joints is based on synergy of massive influx of monocyte/macrophage osteoclast precursors into the diseased joint space and their maturation into bone-degrading cells by stimulated expression of essential osteoclastogenic factors within the synovium (Fig. 1). Macrophages are a hallmark of inflammation, and the abundance of these cells in RA joints mirrors disease activity.(20) Intra-articular macrophage influx is controlled by chemokines, which promote the cell's attachment to synovial microvessels and their invasion into the synovial membrane. In the context of RA, chemokines such, as CXCL-12 (SDF-1), MIP-1α, and CCL20 likely influence mobilization of marrow mononuclear cells into the circulation and finally to the synovial membrane.(21–23)

Figure Figure 1.

Mechanism of inflammatory osteolysis. Inflammatory macrophages within synovium express cytokines, particularly TNF, which promotes synthesis of IL-1. TNF and IL-1 stimulate synovial fibroblast-like cells to synthesize RANKL and M-CSF. IL-23, in conjunction with IL-6 and TGF-β, commit CD4+ T-cells to the TH17 phenotype. TH17 cells produce IL-17, which also stimulates synovial-like fibroblasts to synthesize RANKL. TH17 cells produce the cytokine themselves, although the biological implications of lymphocyte-synthesized RANKL are unresolved. RANKL and M-CSF, in synergy with TNF and IL-1, directly stimulate osteoclast precursors to differentiate into mature, bone-resorbing polykaryons that destroy the rheumatic joint by degrading periarticular bone and mineralized cartilage.

Differentiation of osteoclasts from their mononuclear precursors requires macrophage-colony stimulating factor (M-CSF), engaging its receptor CD115 (c-Fms), and RANKL, binding to its receptor RANK on the surface of monocyte/macrophage precursors. Both cytokines are abundant in the rheumatoid synovial membrane, providing the prerequisite milieu for osteoclastogenesis.(24–26) In fact, the shift in balance between RANKL and its decoy receptor, osteoprotegerin (OPG), may yield the negative balance in systemic bone remodeling attending inflammatory arthritis.(27)

RANKL is increased in the synovium of RA patients, and its genetic inhibition or OPG-mediated blockade prevents bone erosions in experimental inflammatory arthritis.(28–30) Similar results are achieved by blocking M-CSF or c-Fms.(31) Importantly, a recent clinical trial using denosumab, a humanized anti-RANKL antibody, established that arrest of this signaling pathway is also protective of human RA joint destruction.(32)

M-CSF and RANKL are expressed by synovial fibroblast-like cells, which reinforces their role in promoting the structural damage of RA. Because these cells also approximate osteoclasts in the RA synovium,(33) they are likely a major source of the two essential osteoclastogenic cytokines, particularly in response to the abundance of intra-articular proinflammatory molecules such as TNF, IL-1, and IL-17. Activated T lymphocytes, which are present in the RA synovium, also express RANKL, postulated, but not proven to be an important mechanism in the pathogenesis of inflammatory osteolysis.(30) Initially, the TH1 subset of CD4+ T helper cells was considered the source of lymphocyte-produced RANKL. This concept is challenged, however, by the demonstration that TH1 cells actually suppress osteoclastogenesis, probably through interferon-γ production.(34) In fact, current evidence indicates that a novel subset of CD4+ T cells, namely those producing IL-17 under the aegis of IL-23, IL-6, and TGF-β, are the lymphocyte mediators of osteoclastogenesis in the context of inflammatory arthritis. IL-17 exerts its principal osteoclastogenic properties through stimulated RANKL expression by mesenchymal cells.(35,36) Finally, IL-23 directly prompts CD4+ T cells to produce both IL-17 and RANKL, fortifying the concept that TH17 cells are the lymphocyte mediators of inflammatory osteoclastogenesis.(37)TNF and IL-1, which are abundant in rheumatic joints, also play a key role in inflammatory osteolysis. Their importance is underscored by the therapeutic success achieved by inhibiting either cytokine. TNF promotes osteoclast formation through its type 1 (p55) receptor, whereas its type 2 (p75) receptor is anti-osteoclastogenic.(38,39)

In the context of high levels of TNF, the cytokine exerts its osteoclastogenic effects by directly targeting osteoclast precursors and by promoting mesenchymal cell expression of RANKL and M-CSF. TNF's induction of these two cytokines is an essential component of the osteoclastogenic process in circumstances in which TNF is only moderately increased. Alternatively, high ambient TNF, as occurs in inflammatory joint disease, may recruit osteoclasts solely by directly targeting macrophage precursors. This process, however, requires at least constitutive amounts of RANKL, which “primes” macrophages to commit to the osteoclast phenotype in response to TNF.(40) In a reciprocal fashion, TNF “primes” these precursors to undergo osteoclast differentiation in response to relatively small amounts of RANKL. Thus, TNF and RANKL enjoy a synergistic relationship in the osteoclastogenic process. Whereas the capacity of TNF to induce osteoclastogenesis independent of RANKL has been proposed, failure of RANK knockout mice to develop inflammatory osteolysis speaks against this posture.

TNF also facilitates marrow macrophage trafficking, thereby increasing the pool of cells that, when mobilized to joints, undergoes or stimulates osteoclast formation.(41) Proinflammatory cytokines similarly drive expression of surface molecules on macrophages, which assist osteoclastogenesis. For instance, osteoclast-associated receptor (OSCAR), a member of the leukocyte receptor complex (LRC) protein family, is inducible in peripheral monocytes of patients with RA.(42) These data suggest mononuclear cells entering rheumatic joints are predisposed to commit to the osteoclast lineage.

IL-1 induces expression of RANKL and RANK and thus, like TNF, promotes osteoclast formation.(43,44) In consequence, IL-1–deficient mice are protected from osteolysis in the face of TNF-driven synovial inflammation, indicating that TNF triggers bone erosion through induction of IL-1.(43) The abundance of soluble osteoclastogenic factors produced within inflamed joints is likely to “spill-over” into the circulation, accounting for the frequency of glucocorticoid-independent osteoporosis experienced by RA patients. On the other hand, increased systemic cytokine levels, which persist early in the course of steroid treatment of RA patients, may contribute to the rapid bone loss experienced within the first months of such therapy.

Several checks and balance mechanisms are known that control osteoclast activation and may help to limit structural bone damage in RA.(45) As previously outlined, interferon-γ is a potent antagonist of osteoclast formation, but also other cytokines of the TH1 lineage, such as IL-12, suppress rather than promote osteoclastogenesis.(34,46) Interestingly, IL-23, which is considered a key differentiation factor for TH17 cells, can downregulate osteoclast formation and induce expression of GM-CSF.(47) However, these cytokines, despite being activated in the synovium of RA patients, seem to be insufficient to achieve full control of bone resorption. Why interferon-γ, IL-12, and IL-23, and also regulatory factors, such as osteoprotegerin, do not turn off osteoclast formation in RA is yet unknown, but they may simply be over-ruled by the strong osteoclastogenic potential of TNF RANKL and IL-1.

Physiological bone resorption is typically followed by formation, permitting the essential replacement of old skeletal tissue with new. This relationship between osteoclast and osteoblast function is, however, disrupted in inflammatory osteolysis. RA, for example, is characterized by limited bone formation despite aggressive resorptive activity. At best, some sclerosis of bone erosions appears years after onset of arthritis, especially when inflammation is therapeutically controlled. These observations are in keeping with the paucity of osteoblasts in inflammatory lacunae despite an abundance of osteoclasts.(48) Whereas the molecular control of osteoblasts in RA is incompletely understood, proinflammatory cytokines, particularly TNF, inhibit their differentiation and bone-forming activity. For instance, TNF downregulates expression of the key osteoblastogenic transcription factor, Runx-2, and increases expression of Dickkopf (DKK)-1, a potent Wnt antagonist in inflamed joints.(49) Because Wnt signaling promotes OPG expression, which limits osteoclastogenesis, the abundance of DKK-1 in inflammatory arthritis obviating this regulatory mechanism likely enhances RANKL sensitivity and consequent bone loss.(50)

Treatment of inflammatory arthritis and its osteolytic consequences represents a great therapeutic advance of molecular-based medicine. This achievement reflects insights into the mechanisms of bone loss in conditions such as RA and psoriatic arthritis. Osteoclasts, which exclusively mediate the bone destructive process, are the product of macrophage precursors differentiating in the milieu of articular inflammation, ultimately under the influence of M-CSF and RANKL. Targeted blockade of osteoclast differentiation and/or function in the joint, by agents likely to appear in the near future, may further facilitate prevention of joint damage in these diseases.