Osteoclasts: what do they do and how do they do it?

As Americans live longer, degenerative skeletal diseases, such as osteoporosis, become increasingly prevalent. Regardless of cause, osteoporosis reflects a relative enhancement of osteoclast activity. Thus, this unique bone resorptive cell is a prominent therapeutic target. A number of key observations provide insights into the mechanisms by which precursors commit to the osteoclast phenotype and how the mature cell degrades bone. The osteoclast is a member of the monocyte/macrophage family that differentiates under the aegis of two critical cytokines, namely RANK ligand and M-CSF. Tumor necrosis factor (TNF)-alpha also promotes osteoclastogenesis, particularly in states of inflammatory osteolysis such as that attending rheumatoid arthritis. Once differentiated, the osteoclast forms an intimate relationship with the bone surface via the alphavbeta3 integrin, which transmits matrix-derived, cytoskeleton-organizing, signals. These integrin-transmitted signals include activation of the associated proteins, c-src, syk, Vav3, and Rho GTPases. The organized cytoskeleton generates an isolated microenvironment between the cell's plasma membrane and the bone surface in which matrix mineral is mobilized by the acidic milieu and organic matrix is degraded by the lysosomal protease, cathepsin K. This review focuses on these and other molecules that mediate osteoclast differentiation or function and thus serve as candidate anti-osteoporosis therapeutic targets.

Noninvasive imaging of osteoclasts in parathyroid hormone-induced osteolysis using a 64Cu-labeled RGD peptide

Bone diseases are often a result of increased numbers of osteoclasts, or bone-resorbing cells. Bone metastases are a significant cause of morbidity in many types of cancer. An imaging agent targeting osteoclasts, which are upregulated in osteolytic lesions, may facilitate earlier follow-up in patients with osteolytic or mixed bone metastases. Osteoclasts express high levels of alpha(v)beta3 integrin, to which peptides containing the Arg-Gly-Asp (RGD) sequence are known to bind. We proposed that radiolabeled RGD peptides could be used to detect osteoclasts in lytic bone lesions.

METHODS

The cross-bridged macrocyclic chelator 4,11-bis(carboxymethyl)-1,4,8,11-tetraazabicyclo[6.6.2]hexadecane (CB-TE2A) was conjugated to c(RGDyK) for radiolabeling with 64Cu (t(1/2), 12.7 h; beta+, 17.4%; E(beta+ max), 656 keV; beta-, 39%; E(beta- max), 573 keV). The in vitro affinity of Cu(II)-CB-TE2A-c(RGDyK) for alpha(v)beta3 and alpha(v)beta5 was evaluated in a heterologous competitive binding assay. Ex vivo uptake was examined in osteoclasts prepared from bone marrow macrophages. As a proof of principle, biodistribution and imaging studies were performed on parathyroid hormone (PTH)-induced osteolysis in the calvarium.

RESULTS

Cu-CB-TE2A-c(RGDyK) was shown to have a 30-fold higher affinity for alpha(v)beta3 than for alpha(v)beta5. Osteoclasts were shown to specifically take up (64)Cu-CB-TE2A-c(RGDyK). However, bone marrow macrophages showed only nonspecific uptake. PTH treatment increased calvarial uptake of 64Cu-CB-TE2A-c(RGDyK), compared with uptake in mice receiving a sham treatment. In addition, calvarial uptake correlated linearly with the number of osteoclasts on the bone surface.

CONCLUSION

These results suggest that 64Cu-CB-TE2A-c(RGDyK) selectively binds alpha(v)beta3 on osteoclasts and may potentially be used to identify increased numbers of osteoclasts in osteolytic bone diseases such as osteolytic bone metastasis and inflammatory osteolysis.

SHIP1 negatively regulates proliferation of osteoclast precursors via Akt-dependent alterations in D-type cyclins and p27

Osteoclasts arise from macrophage progenitors in bone marrow (BMMs) as a consequence of signaling events elicited by M-CSF and receptor activator of NF-kappaB ligand, acting on their unique receptors, via c-Fms and receptor activator of NF-kappaB. Both receptors activate the PI3K and MAPK pathways, which promote cell proliferation and survival. SHIP1 is essential for normal bone homeostasis, as mice lacking the protein exhibit osteoporosis resulting from increased numbers of hyper-resorptive osteoclasts. In this study, we show that BMMs from SHIP1 null mice respond to M-CSF, but not receptor activator of NF-kappaB ligand, by increasing Akt activation. In consequence, there are up-regulation of D-type cyclins, down-regulation of the cyclin-dependent kinase inhibitor p27, and, therefore, increased phosphorylation of the retinoblastoma protein and cell proliferation. Surprisingly, cell survival of wild-type and knockout BMMs is unaltered. Finally, osteoclastogenesis and periarticular bone erosions are markedly increased in SHIP1(-/-) mice with inflammatory arthritis, a condition characterized by increased M-CSF expression. The SHIP1/Akt pathway therefore suppresses bone loss in pathological states associated with an excess of the cytokine.

The LIM Protein, LIMD1, Regulates AP-1 Activation through an Interaction with TRAF6 to Influence Osteoclast Development

Increasingly a number of proteins important in the regulation of bone osteoclast development have been shown primarily influence osteoclastogenesis under conditions of physiologic or pathologic stress. Why basal osteoclastogenesis is normal and how these proteins regulate stress osteoclastogenic responses, as opposed to basal osteoclastogenesis, is unclear. LIM proteins of the Ajuba/Zyxin family localize to cellular sites of cell adhesion where they contribute to the regulation of cell adhesion and migration, translocate into the nucleus where they can affect cell fate, but are also found in the cytoplasm where their function is largely unknown. We show that one member of this LIM protein family, Limd1, is uniquely up-regulated during osteoclast differentiation and interacts with Traf6, a critical cytosolic regulator of RANK-L-regulated osteoclast development. Limd1 positively affects the capacity of Traf6 to activate AP-1, and Limd1(-/-) osteoclast precursor cells are defective in the activation of AP-1 and thus induction of NFAT2. Limd1(-/-) mice, although having normal basal bone osteoclast numbers and bone density, are resistant to physiological and pathologic osteoclastogenic stimuli. These results implicate Limd1 as a potentially important regulator of osteoclast development under conditions of stress.

Osteoclasts and integrins

The osteoclast is the unique bone resorptive cell that accomplishes its mission by forming an isolated acidified microenvironment between itself and the bone surface. Creation of this compartment is the first step in bone degradation and establishes that an intimate physical relationship must exist between the osteoclast and bone. Thus, identification of the mechanisms by which the osteoclast attaches to bone is essential to understanding how the cell degrades skeletal tissue. Our studies have investigated whether absence of the alphavbeta3 integrin modifies the ability of c-Fms to induce Rho GTPases, and the implications for formation of the osteoclast cytoskeleton.

Glucocorticoids suppress bone formation via the osteoclast

The pathogenesis of glucocorticoid-induced (GC-induced) bone loss is unclear. For example, osteoblast apoptosis is enhanced by GCs in vivo, but they stimulate bone formation in vitro. This conundrum suggests that an intermediary cell transmits a component of the bone-suppressive effects of GCs to osteoblasts in the intact animal. Bone remodeling is characterized by tethering of the activities of osteoclasts and osteoblasts. Hence, the osteoclast is a potential modulator of the effect of GCs on osteoblasts. To define the direct impact of GCs on bone-resorptive cells, we compared the effects of dexamethasone (DEX) on WT osteoclasts with those derived from mice with disruption of the GC receptor in osteoclast lineage cells (GRoc-/- mice). While the steroid prolonged longevity of osteoclasts, their bone-degrading capacity was suppressed. The inhibitory effect of DEX on bone resorption reflects failure of osteoclasts to organize their cytoskeleton in response to M-CSF. DEX specifically arrested M-CSF activation of RhoA, Rac, and Vav3, each of which regulate the osteoclast cytoskeleton. In all circumstances GRoc-/- mice were spared the impact of DEX on osteoclasts and their precursors. Consistent with osteoclasts modulating the osteoblast-suppressive effect of DEX, GRoc-/- mice are protected from the steroid's inhibition of bone formation.

alphavbeta3 and macrophage colony-stimulating factor: partners in osteoclast biology

Osteoclasts, the sole bone-resorbing cells, arise by fusion and differentiation of monocyte/macrophage precursors. Matrix degradation requires adhesion of the osteoclast to bone, an integrin alphavbeta3-mediated event that also stimulates signals which polarize the cell and secrete resorptive molecules such as hydrochloric acid and acidic proteases. Two cytokines are necessary and sufficient for osteoclastogenesis, receptor activator of nuclear factor kappaB ligand (RANKL) and macrophage colony-stimulating factor (M-CSF), both produced by mesenchymal cells in the bone marrow environment. M-CSF promotes survival and proliferation of osteoclast precursors. It also contributes to their differentiation and regulates the cytoskeletal changes that accompany bone resorption. Binding of M-CSF to c-Fms, its receptor, recruits adapter proteins and cytosolic kinases, thereby activating a variety of intracellular signals. We herein review how alphavbeta3 and M-CSF, alone and in concert, impact production, survival, and function of the osteoclast, thereby controlling skeletal mass. Signals from alphavbeta3 and/or c-Fms activate Syk and Vav3, originally defined by their function in lymphoid cells. Genetic depletion of either protein generates a strong bone phenotype, underscoring the promise of osteoimmunobiology.

Critical role of beta3 integrin in experimental postmenopausal osteoporosis

We show that mice lacking beta3 integrin are protected from OVX-induced bone loss. Using a lentiviral-based strategy to express beta3 mutants in beta3(-/-) mice, we also show that beta3(S752), but not beta3(Y747/Y759), is important for osteoclastic bone resorption in vivo.

INTRODUCTION

Mice lacking the beta3 integrin have dysfunctional osteoclasts and therefore accumulate bone mass with age. Thus, the alphavbeta3 integrin is a potential anti-osteoporosis target. Identifying components of the beta3 integrin that determine its function in vivo is essential for therapeutically exploiting the antiresorptive properties of alphavbeta3.

MATERIALS AND METHODS

We used DXA and histomorphometry to assess bone loss after ovariectomy in wildtype and beta3 integrin null mice. We used lentiviral vectors carrying various human beta3 (hbeta3) integrin constructs to transduce beta3(-/-) bone marrow and reconstituted lethally irradiated beta3(-/-) mice with the transduced marrow. The expressed constructs include the intact integrin and two mutants, namely hbeta3(Y747F/Y759F) and hbeta3(S752P), each of which induces the bleeding dyscrasia, Glanzmann's thrombasthenia, in humans. Two months after transplantation, the expression of hbeta3 was measured by flow cytometry of marrow-derived macrophages. Osteoclast differentiation and function were assessed ex vivo by TRACP and actin-ring staining, respectively. Reconstituted mice were ovariectomized, and bone loss was assessed by DXA, histomorphometry, and serum TRACP5b assay.

RESULTS

beta3(-/-) mice are protected from ovariectomy-induced bone loss, showing no difference in BMD compared with sham-operated controls. We successfully expressed hbeta3 integrins in beta3(-/-) hosts using lentiviral transduction of bone marrow. Two months after transplantation, 25-35% of marrow-derived macrophages expressed the hbeta3 constructs. Similar to its effect in vitro, hbeta3(WT) completely rescued the osteoclast and platelet phenotype of beta3(-/-) mice. Whereas platelet function remained deranged in beta3(-/-) mice overexpressing hbeta3(Y747F/Y759F), osteoclast function was fully restored. In contrast, beta3(-/-) mice expressing hbeta3(S752P) continued to exhibit prolonged bleeding times and dysfunctional osteoclasts in vitro and ex vivo. Most importantly, hbeta3(WT) and hbeta3(Y747F/Y759F) transplanted mice underwent equivalent ovariectomy-induced bone loss, whereas, like those bearing the control vector, hbeta3(S752P) transplanted mice were protected.

CONCLUSIONS

Functional beta3 integrin is required for ovariectomy-induced bone loss. beta3(S752), but not beta3(Y747/Y759), is critical for osteoclast function in vivo.

Osteoclasts; culprits in inflammatory osteolysis

Periarticular osteolysis, a crippling complication of rheumatoid arthritis, is the product of enhanced osteoclast recruitment and activation. The osteoclast, which is a member of the monocyte/macrophage family, is the exclusive bone resorptive cell, and its differentiation and activation are under the aegis of a variety of cytokines. Receptor activator of NF-kappaB ligand (RANKL) and macrophage colony-stimulating factor are the essential osteoclastogenic cytokines and are increased in inflammatory joint disease. Tumor necrosis factor-alpha, which perpetrates arthritic bone loss, exerts its osteoclastogenic effect in the context of RANKL with which it synergizes. Achieving an understanding of the mechanisms by which the three cytokines affect the osteoclast has resulted in a number of active and candidate therapeutic targets.

M-CSF mediates TNF-induced inflammatory osteolysis

TNF-alpha is the dominant cytokine in inflammatory osteolysis. Using mice whose BM stromal cells and osteoclast precursors are chimeric for the presence of TNF receptors, we found that both cell types mediated the cytokine's osteoclastogenic properties. The greater contribution was made, however, by stromal cells that express the osteoclastogenic cytokine M-CSF. TNF-alpha stimulated M-CSF gene expression, in vivo, only in the presence of TNF-responsive stromal cells. M-CSF, in turn, induced the key osteoclastogenic cytokine receptor, receptor activator of NF-kappaB (RANK), in osteoclast precursors. In keeping with the proproliferative and survival properties of M-CSF, TNF-alpha enhanced osteoclast precursor number only in the presence of stromal cells bearing TNF receptors. To determine the clinical relevance of these observations, we induced inflammatory arthritis in wild-type mice and treated them with a mAb directed against the M-CSF receptor, c-Fms. Anti-c-Fms mAb selectively and completely arrested the profound pathological osteoclastogenesis attending this condition, the significance of which is reflected by similar blunting of the in vivo bone resorption marker tartrate-resistant acid phosphatase 5b (TRACP 5b). Confirming that inhibition of the M-CSF signaling pathway targets TNF-alpha, anti-c-Fms also completely arrested osteolysis in TNF-injected mice with nominal effect on macrophage number. M-CSF and its receptor, c-Fms, therefore present as candidate therapeutic targets in states of inflammatory bone erosion.

FHL2 inhibits the activated osteoclast in a TRAF6-dependent manner

TNF receptor-associated factor 6 (TRAF6) associates with the cytoplasmic domain of receptor activator of NF-kappaB (RANK). This event is central to normal osteoclastogenesis. We discovered that TRAF6 also interacts with FHL2 (four and a half LIM domain 2), a LIM domain--only protein that functions as a transcriptional coactivator or corepressor in a cell-type--specific manner. FHL2 mRNA and protein are undetectable in marrow macrophages and increase pari passu with osteoclast differentiation in vitro. FHL2 inhibits TRAF6-induced NF-kappaB activity in wild-type osteoclast precursors and, in keeping with its role as a suppressor of TRAF6-mediated RANK signaling, TRAF6/RANK association is enhanced in FHL2-/- osteoclasts. FHL2 overexpression delays RANK ligand-induced (RANKL-induced) osteoclast formation and cytoskeletal organization. Interestingly, osteoclast-residing FHL2 is not detectable in naive wild-type mice, in vivo, but is abundant in those treated with RANKL and following induction of inflammatory arthritis. Reflecting increased RANKL sensitivity, osteoclasts generated from FHL2-/- mice reach maturation and optimally organize their cytoskeleton earlier than their wild-type counterparts. As a consequence, FHL2-/- osteoclasts are hyperresorptive, and mice lacking the protein undergo enhanced RANKL and inflammatory arthritis-stimulated bone loss. FHL2 is, therefore, an antiosteoclastogenic molecule exerting its effect by attenuating TRAF6-mediated RANK signaling.

Rab3D regulates a novel vesicular trafficking pathway that is required for osteoclastic bone resorption

Rab3 proteins are a subfamily of GTPases, known to mediate membrane transport in eukaryotic cells and play a role in exocytosis. Our data indicate that Rab3D is the major Rab3 species expressed in osteoclasts. To investigate the role of Rab3D in osteoclast physiology we examined the skeletal architecture of Rab3D-deficient mice and found an osteosclerotic phenotype. Although basal osteoclast number in null animals is normal the total eroded surface is significantly reduced, suggesting that the resorptive defect is due to attenuated osteoclast activity. Consistent with this hypothesis, ultrastructural analysis reveals that Rab3D(-/-) osteoclasts exhibit irregular ruffled borders. Furthermore, while overexpression of wild-type, constitutively active, or prenylation-deficient Rab3D has no significant effects, overexpression of GTP-binding-deficient Rab3D impairs bone resorption in vitro. Finally, subcellular localization studies reveal that, unlike wild-type or constitutively active Rab3D, which associate with a nonendosomal/lysosomal subset of post-trans-Golgi network (TGN) vesicles, inactive Rab3D localizes to the TGN and inhibits biogenesis of Rab3D-bearing vesicles. Collectively, our data suggest that Rab3D modulates a post-TGN trafficking step that is required for osteoclastic bone resorption.

IL-1 mediates TNF-induced osteoclastogenesis

TNF-induced receptor activator NF-kappaB ligand (RANKL) synthesis by bone marrow stromal cells is a fundamental component of inflammatory osteolysis. We found that this process was abolished by IL-1 receptor antagonist (IL-1Ra) or in stromal cells derived from type I IL-1 receptor-deficient (IL-1RI-deficient) mice. Reflecting sequential signaling of the cytokines TNF and IL-1, TNF induces stromal cell expression of IL-1 and IL-1RI. These data suggest that TNF regulates RANKL expression via IL-1, and, therefore, IL-1 plays a role in TNF-induced periarticular osteolysis. Consistent with this posture, TNF-stimulated osteoclastogenesis in cultures consisting of WT marrow macrophages and stromal cells exposed to IL-1Ra or in cocultures established with IL-1RI-deficient stromal cells was reduced approximately 50%. The same magnitude of osteoclast inhibition occurred in IL-1RI-deficient mice following TNF administration in vivo. Like TNF, IL-1 directly targeted osteoclast precursors and promoted the osteoclast phenotype in a TNF-independent manner in the presence of permissive levels of RANKL. IL-1 is able to induce RANKL expression by stromal cells and directly stimulate osteoclast precursor differentiation under the aegis of p38 MAPK. Thus, IL-1 mediates the osteoclastogenic effect of TNF by enhancing stromal cell expression of RANKL and directly stimulating differentiation of osteoclast precursors.

Vav3 regulates osteoclast function and bone mass

Osteoporosis, a leading cause of morbidity in the elderly, is characterized by progressive loss of bone mass resulting from excess osteoclastic bone resorption relative to osteoblastic bone formation. Here we identify Vav3, a Rho family guanine nucleotide exchange factor, as essential for stimulated osteoclast activation and bone density in vivo. Vav3-deficient osteoclasts show defective actin cytoskeleton organization, polarization, spreading and resorptive activity resulting from impaired signaling downstream of the M-CSF receptor and alpha(v)beta3 integrin. Vav3-deficient mice have increased bone mass and are protected from bone loss induced by systemic bone resorption stimuli such as parathyroid hormone or RANKL. Moreover, we provide genetic and biochemical evidence for the role of Syk tyrosine kinase as a crucial upstream regulator of Vav3 in osteoclasts. Thus, Vav3 is a potential new target for antiosteoporosis therapy.

Mice lacking the integrin beta5 subunit have accelerated osteoclast maturation and increased activity in the estrogen-deficient state

Integrin alphavbeta5 is expressed on osteoclast precursors and is capable of recognizing the same amino acid motif as alphavbeta3. Three-month-old beta5(-/-) female OVX mice had increased osteoclastogenesis ex vivo, and microCT assessment of trabecular bone volume was 53% lower than WT-OVX animals. These preliminary data suggest alphavbeta5 integrin's presence on osteoclast precursors may inhibit of osteoclast formation.

INTRODUCTION

Osteoclasts are unique resorptive skeletal cells, capable of degrading bone on contact to the juxtaposed matrix. Integrin alphavbeta5 is expressed on osteoclast precursors, structurally similar to alphavbeta3, and capable of recognizing the same amino acid motif. Given the structural relationship and reciprocal regulation of alphavbeta3 and alphavbeta5, the purpose of this study was to evaluate how alphavbeta5 might contribute to osteoclast maturation and activity.

MATERIALS AND METHODS

Three-month-old wildtype (WT) and beta5(-/-) female mice had ovariectomy (OVX) or sham operations. The osteoclastogenic capacity of marrow-derived precursors, the kinetic, the circulating, and structural parameters of bone remodeling, was determined after 6 weeks of paired feeding.

RESULTS AND CONCLUSIONS

OVX increased osteoclastogenesis ex vivo and in vivo. Osteoclast formation and prolonged pre-osteoclast survival were substantially enhanced in cultures containing beta5(-/-) cells whether obtained from sham-operated or OVX mice. Expression of cathepsin K, beta3 integrin subunit, and calcitonin receptor were accelerated in cultured beta5(-/-)osteoclasts. beta5(-/-) osteoclasts from OVX animals showed a 3-fold enhancement of net resorptive activity, with quantitative muCT showing trabecular bone volume loss after OVX 53% greater in beta5(-/-) OVX compared with similarly treated WT OVX mice (p < 0.05). alpha5beta3 seems to be an inhibitor of osteoclast formation, in contrast to alphavbeta3. In addition, loss of alphavbeta5 seems to accelerate osteoclast formation in the OVX model. Further examination of alphavbeta5 signaling pathways may enhance our understanding of the activation of bone resorption.

Unoccupied alpha(v)beta3 integrin regulates osteoclast apoptosis by transmitting a positive death signal

Cell/matrix detachment is a general inducer of programmed cell death, an event mediated by loss of integrin/ligand association. Because alpha(v)beta3 is the major integrin expressed by the osteoclast, we asked whether its occupancy promotes survival of the resorptive cell. Thus, we generated wild-type preosteoclasts and placed them on selective matrix proteins. Consistent with the posture that alpha(v)beta3 occupancy promotes survival, preosteoclasts plated on native collagen, a matrix not recognized by the integrin, undergo apoptosis 4-fold faster than those on the alpha(v)beta3 ligand, vitronectin. To further explore the role of alpha(v)beta3 in osteoclast apoptosis, wild-type and beta3-/- preosteoclasts were suspended and apoptosis determined, with time. Beta3-/- preosteoclasts, in suspension, undergo a rate of apoptosis only 40-60% of that of their wild-type counterparts, indicating that unoccupied alpha(v)beta3 transmits a positive death signal that we find regulated by caspase-8. Attesting to specificity of the unoccupied integrin-transmitted death signal, apoptosis in the absence of alpha(v)beta3 is mediated by capsase-9. We have shown that the resorptive defect of beta3-/- osteoclasts is rescued by wild-type beta3 cDNA but not by one bearing a S752P mutation. To determine whether the same holds true regarding osteoclast apoptosis, we constructed lentivirus vectors encoding green fluorescent protein, wild-type beta3, or beta3S752P. Once again, native beta3-/- preosteoclasts were protected against apoptosis. Similar to its effect on bone resorption, transduced wild-type beta3 normalizes the apoptotic rate of beta3-/- preosteoclasts. Unexpectedly, however, beta3S752P transductants also die at a rate indistinguishable from wild type. Thus, unoccupied alpha(v)beta3 integrin regulates osteoclast apoptosis via a component of the integrin that is different than that regulating resorption.

Marrow Stromal Cells and Osteoclast Precursors Differentially Contribute to TNF-α-Induced Osteoclastogenesis In Vivo

The marrow stromal cell is the principal source of the key osteoclastogenic cytokine receptor activator of NF-kappaB (RANK) ligand (RANKL). To individualize the role of marrow stromal cells in varying states of TNF-alpha-driven osteoclast formation in vivo, we generated chimeric mice in which wild-type (WT) marrow, immunodepleted of T cells and stromal cells, is transplanted into lethally irradiated mice deleted of both the p55 and p75 TNFR. As control, similarly treated WT marrow was transplanted into WT mice. Each group was administered increasing doses of TNF-alpha. Exposure to high-dose cytokine ex vivo induces exuberant osteoclastogenesis irrespective of in vivo TNF-alpha treatment or whether the recipient animals possess TNF-alpha-responsive stromal cells. In contrast, the osteoclastogenic capacity of marrow treated with lower-dose TNF-alpha requires priming by TNFR-bearing stromal cells in vivo. Importantly, the osteoclastogenic contribution of cytokine responsive stromal cells in vivo diminishes as the dose of TNF-alpha increases. In keeping with this conclusion, mice with severe inflammatory arthritis develop profound osteoclastogenesis and bone erosion independent of stromal cell expression of TNFR. The direct induction of osteoclast recruitment by TNF-alpha is characterized by enhanced RANK expression and sensitization of precursor cells to RANKL. Thus, osteolysis attending relatively modest elevations in ambient TNF-alpha depends upon responsive stromal cells. Alternatively, in states of severe periarticular inflammation, TNF-alpha may fully exert its bone erosive effects by directly promoting the differentiation of osteoclast precursors independent of cytokine-responsive stromal cells and T lymphocytes.