Ten per cent of our adult bone mass is replaced each year — a constant remodelling process in which osteoclasts, cells that destroy and resorb old bone, and osteoblasts, which deposit new bone in its place, have a key role. Several molecules have been proposed to coordinate the activities of osteoblasts and osteoclasts, but as a recent report in Nature reveals, osteoclasts might also keep themselves under strict control. Targeted use of the key molecule responsible — interferon-β (IFN-β) — might be of benefit in diseases characterized by inappropriately high bone resorption, such as osteoporosis and rheumatoid arthritis.

Osteoclast maturation from precursors to functional bone-resorbing cells is stimulated when a protein on the surface of precursors — RANK — is activated by its ligand RANKL, which is usually present on the surface of osteoblasts. Activation of RANK leads to an increase in the expression of certain genes, including that coding for the transcription factor c-Fos, which is known to be crucial for osteoclast differentiation.

How are the effects of RANKL regulated? Various molecules are known to control the level and availability of RANKL on osteoblast surfaces; however, it was not known whether RANKL signalling itself activates a negative regulatory mechanism to maintain bone homeostasis. To answer this question, Takayanagi et al. set out to identify genes induced when RANKL binds to RANK. Initial observations highlighted the IFN-β signalling pathway, and its key role was confirmed when the authors showed that mice lacking IFN-β or its cell-surface receptor have low bone mass and increased bone resorption by osteoclasts.

Further experiments indicated that the IFN-β gene is induced by c-Fos, and that IFN-β inhibits RANKL-induced osteoclast differentiation by inhibiting the expression of c-Fos. Bringing all the observations together, it seems that IFN-β and c-Fos are part of a negative feedback loop initiated by RANKL. In one half of the loop, RANK induces IFN-β expression through c-Fos. In the other half, secreted IFN-β binds to its receptor on osteoclast precursors, which causes a decrease in c-Fos levels, and thus inhibition of osteoclast differentiation.

So, could manipulating this loop be of therapeutic benefit? Initial indications seem encouraging — the authors showed that IFN-β prevented bone resorption in a mouse model of arthritis, and intriguingly, some patients with multiple sclerosis being treated with IFN-β have been reported to show improvements in their rheumatoid arthritis.