TSG-6 is an inflammation-induced protein that is produced at pathological sites, including arthritic joints. In animal models of arthritis, TSG-6 protects against joint damage; this has been attributed to its inhibitory effects on neutrophil migration and plasmin activity. Here we investigated whether TSG-6 can directly influence bone erosion. Our data reveal that TSG-6 inhibits RANKL-induced osteoclast differentiation/activation from human and murine precursor cells, where elevated dentine erosion by osteoclasts derived from TSG-6(-/-) mice is consistent with the very severe arthritis seen in these animals. However, the long bones from unchallenged TSG-6(-/-) mice were found to have higher trabecular mass than controls, suggesting that in the absence of inflammation TSG-6 has a role in bone homeostasis; we have detected expression of the TSG-6 protein in the bone marrow of unchallenged wild type mice. Furthermore, we have observed that TSG-6 can inhibit bone morphogenetic protein-2 (BMP-2)-mediated osteoblast differentiation. Interaction analysis revealed that TSG-6 binds directly to RANKL and to BMP-2 (as well as other osteogenic BMPs but not BMP-3) via composite surfaces involving its Link and CUB modules. Consistent with this, the full-length protein is required for maximal inhibition of osteoblast differentiation and osteoclast activation, although the isolated Link module retains significant activity in the latter case. We hypothesize that TSG-6 has dual roles in bone remodeling; one protective, where it inhibits RANKL-induced bone erosion in inflammatory diseases such as arthritis, and the other homeostatic, where its interactions with BMP-2 and RANKL help to balance mineralization by osteoblasts and bone resorption by osteoclasts.