Department of Pathology, Washington University School of Medicine, St. Louis, Missouri, USA.
Address correspondence to: Steven L. Teitelbaum, Department of Pathology and Immunology, Washington University School of Medicine, Campus Box 8118, 660 South Euclid Avenue, St. Louis, Missouri 63110, USA. Phone: (314) 454-8463; Fax: (314) 454-5505; E-mail: email@example.com.
First published August 16, 2004 - More info
Pathological bone loss always reflects enhanced net osteoclastic activity. Recognition and binding of the receptor activator of NF-κB (RANK) by RANK ligand (RANKL) is the key osteoclastogenic event, and the signaling cascades induced by this reaction therefore contain potential anti-osteoporosis therapeutic targets. A study reported in this issue of the JCI documents that a pivotal component of RANKL/RANK-mediated osteoclast recruitment involves sequential induction of the transcription factors c-Jun and nuclear factor of activated T cells 2 .
Osteoporosis always represents an imbalance in favor of osteoclast-mediated bone resorption relative to the bone-forming capacity of osteoblasts. In some conditions of accelerated skeletal loss, such as Paget disease, osteoclasts are greatly enlarged, which enhances the individual cell’s resorptive activity. Most osteopenic disorders, however, develop as a consequence of accelerated bone degradation due to increased osteoclast number. Thus, inhibition of osteoclast formation by estrogen supplementation, or osteoclast function by bisphosphonate administration, has been the most effective means of arresting pathological bone loss. However, both forms of therapy are not without drawbacks, and new anti-resorptive targets are rapidly emerging. In this issue of the JCI, Ikeda et al. elegantly show that the JNK-1–activated c-Jun signaling pathway is key to osteoclastic bone resorption (1). The authors find that mice bearing a dominant-negative osteoclast-specific c-Jun transgene have impaired osteoclastogenesis and, because of failed bone resorption, develop increased bone mass in the form of osteopetrosis. Importantly, the failure of dominant-negative c-Jun transgenic mice to generate osteoclasts is due to arrested activation and expression of members of the nuclear factor of activated T cells (NFAT) family of transcription factors in osteoclast precursors.
The osteoclast is a huge cell whose phenotype is characterized by its capacity to attach to bone and polarize its resorptive machinery toward the cell’s interface with mineralized matrix (2). It is a terminally differentiated polykaryon whose multinucleation is a manifestation of fusion rather than endomitosis, and it is the unique resorbing cell of the skeleton.
Donald Walker, in the 1970s, performed the first meaningful experiments with osteopetrotic mice, which have, from that time forward, served as the essential resource for understanding how osteoclasts are born, resorb bone, and die. Using cross-circulation and marrow cell transplantation, Walker established that the enigmatic polykaryon is of hematopoietic origin (3). A decade later, Suda’s group documented that the osteoclast can be generated, in vitro, from mononuclear members of the monocyte/macrophage family, thus establishing their identity as osteoclast precursors (4). Curiously, however, Suda was unable to generate osteoclasts unless their mononuclear progenitors were in physical contact with marrow-derived mesenchymal stromal cells, including osteoblasts. Discovery of the molecule expressed on the surface of stromal cells that mediates commitment of macrophages to the bone resorptive phenotype remained the principal enigma of osteoclast biology until 1998, when this key osteoclastogenic moiety was identified as receptor activator of NF-κB (RANK) ligand (RANKL) (5, 6). RANKL is a member of the TNF superfamily whose crystal structure reveals unique components that enable it to activate its osteoclastogenic receptor, RANK, on osteoclast precursors (7) (Figure 1). While other cytokines, particularly TNF-α, may be profoundly synergistic, only activated RANKL/RANK signaling suffices and is essential for osteoclast differentiation (8, 9).
The AP-1/NFAT transcription complex mediates osteoclast precursor differentiation. Osteoclastogenesis is initiated by RANKL occupying RANK on the surface of osteoclast precursors. Subsequent recruitment of TRAF6 initiates the 3 depicted signaling cascades, in addition to other pathways not shown here. Phosphorylation (P) activates c-Jun in an MKK7/JNK-1–dependent manner, and NFAT1 is activated by dephosphorylation via calcium-mediated induction of calcineurin. RANKL/RANK also induces c-Fos expression by an incompletely understood mechanism. NFAT1 partners with the AP-1 proteins of the Fos/Jun families to transactivate the NFAT2 gene, the product of which forms a similar ternary transcription complex on osteoclastic genes eventuating in appearance of the mature osteoclast phenotype.
The discovery of the osteoclastogenic capacity of RANKL enables generation of large numbers of essentially pure populations of bona fide osteoclasts in vitro, which in turn permits meaningful biochemical and molecular definition of this cell. We now know that RANKL is a homotrimer (7), which not only promotes differentiation of osteoclast precursors but also activates the mature bone-degrading polykaryon (10). Thus, RANKL and RANK, as well as their intracellular signaling pathway, are presently the most promising anti-resorptive therapeutic targets.
The importance of activator protein-1 (AP-1) transcription factors, specifically dimers of the Fos and Jun families of proteins, in the osteoclastogenic process was first documented in the laboratory of Erwin Wagner, wherein c-Fos knockout mice were shown to be osteopetrotic due to a complete absence of osteoclasts (11). Consequently, the discovery of RANKL as the key osteoclastogenic cytokine prompted interest in the mechanisms by which RANK activation regulates AP-1 transcription factors.
Just why RANKL is unique among TNF superfamily members in its capacity to induce osteoclast differentiation is still unresolved but probably involves its interaction with TNF receptor–associated factor 6 (TRAF6) (12, 13). Other RANKL-stimulated intracellular signaling molecules essential to the osteoclast phenotype include the p50/p52 NF-κB subunits and the PI3K/AKT axis. Similarly, the MAPKs extracellular regulated kinases 1 and 2 and p38 are required for osteoclast differentiation or function.
Association of RANKL and TRAF6 activates all key events involving AP-1–mediated transcription of osteoclast specific genes. Expression of the c-Fos gene in 393T cells requires TRAF 6, but whether the same holds true in osteoclasts is unresolved. On the other hand, c-Jun is clearly RANKL activated via TRAF6 by a process involving JNK1 but not JNK2 (14). c-Jun is not the only member of its family to regulate osteoclastogenesis, as Jun-B–deficient mice also have arrested bone resorption (15).
When transcriptionally active, c-Jun associates with members of the Fos family of transcription factors, a number of which — such as Fra-1, itself a c-Fos target — may substitute for c-Fos in the osteoclastogenic process (16). Although c-Jun also homodimerizes, the fact that c-Jun overexpression does not rescue the c-Fos–/– osteopetrotic phenotype indicates that c-Jun is necessary, but not sufficient, to transactivate osteoclast specific genes.
In 2002 Takayanagi et al. identified NFAT2 by genome-wide screening as the predominant gene activated in osteoclast precursors under the influence of RANKL and documented the transcription factor’s essential role in osteoclastogenesis (17). RANK occupancy mobilizes intracellular calcium, a requisite for calcineurin-mediated NFAT activation (Figure 1). Moreover, RANKL not only induces the transcription factor’s expression but facilitates its nuclear translocation, where NFAT binds to its DNA response element via a ternary complex with AP-1 proteins, including Fos/Jun, to transactivate target genes (18). Thus, it is not surprising, in retrospect, that RANKL-induced osteoclastogenesis involves partnering of Fos/Jun with NFAT2. The fact that Ikeda et al. (1) found that expression of NFAT2 itself is NFAT1/Fos/Jun–dependent is in keeping with the presence of NFAT and AP-1 response elements in the NFAT2 promoter (19). Reflecting its role as an essential RANKL-activated signaling molecule, NFAT2, when overexpressed in wild-type macrophages, prompts osteoclast differentiation in the absence of the osteoclastogenic cytokine (17). RANKL-stimulated gene transcription is, therefore, a reflection of NFAT partnering with Fos/Jun.
Thus, Ikeda et al. (1) provide evidence of important components of RANK-mediated osteoclastogenesis specifically as it pertains to the AP-1 transcription complex. Clearly, TRAF6 is an essential player in RANKL-mediated c-Jun and NFAT activation and, perhaps, c-Fos expression. NFAT/Fos/Jun is a critical osteoclastogenic complex, and deletion of any of the three arrests osteoclast formation. The pivotal role that Ikeda et al. document for c-Jun in osteoclast formation is in keeping with the fact that the anti–bone resorptive effects of estrogen are substantially mediated by c-Jun repression (20). Given that RANKL expression is also enhanced in estrogen-deficient women (21), RANKL → TRAF6 → MAPK kinase 7 (MKK7) → JNK1 → Jun → NFAT signaling is likely pivotal to the pathogenesis of post-menopausal osteoporosis, and inhibition of any of the components will theoretically arrest accelerated bone resorption (Figure 1). The therapeutic challenge is how to specifically target these intracellular signaling molecules in osteoclasts.
See the related article beginning on page 475.
Nonstandard abbreviations used: activator protein-1 (AP-1); MAPK kinase 7 (MKK7); nuclear factor of activated T cells (NFAT); receptor activator of NF-κB (RANK); RANK ligand (RANKL); TNF receptor–associated factor 6 (TRAF6).
Conflict of interest: S.L. Teitelbaum is a Scientific Founder of Auxeris Therapeutics Inc., which develops anti-osteoporosis drugs.