(Pre-)osteoclasts induce retraction of osteoblasts before their fusion to osteoclasts.

Abstract

Precursors of osteoclasts seeded on top of a confluent layer of osteoblasts/bone lining cells induced retraction of the latter cells. The (pre)osteoclasts then migrated in the formed cell-free areas and fused to form osteoclast-like cells. Retraction of the osteoblasts/bone lining cells proved to depend on activity of matrix metalloproteinases, and TGF-beta1 prevented the retraction. It is well known that osteoblasts have a profound effect on (pre)osteoclasts in inducing the formation of bone-resorbing osteoclasts. Whether, on the other hand, (pre)osteoclasts also modulate osteoblast activity is largely unknown. Because osteoblasts/bone lining cells have to retract from the surface before resorption of bone by osteoclasts, we addressed the question of whether (pre)osteoclasts have the capacity to induce such an activity. Rabbit calvarial osteoblasts/bone lining cells or periosteal fibroblasts were cultured until confluency, after which rabbit peripheral blood mononuclear cells (PBMCs) were seeded on top of them. The co-cultures were maintained for up to 15 days in the presence or absence of the cytokines transforming growth factor (TGF)-beta1 and TNF-alpha and selective inhibitors of matrix metalloproteinases and serine proteinases. The formation of cell-free areas and the number of TRACP+ multinucleated osteoclast-like cells were analyzed. In addition, formation of cell-free areas was analyzed in co-cultures of osteoblasts with mature osteoclasts. The seeding of PBMCs on a confluent layer of osteoblasts/bone lining cells resulted in the following sequence of events. (1) A low number of PBMCs strongly attached to osteoblasts. 2) At these sites of contact, the osteoblasts retracted, thus forming cell-free areas. (3) The PBMCs invaded these areas and attached to the surface of the well, after which they fused and formed multinucleated TRACP+ osteoclast-like cells. Retraction was only seen if the cells were in direct contact; conditioned media from cultured PBMCs added to osteoblasts had no effect. Mature osteoclasts seeded on osteoblasts similarly induced retraction, but this retraction occurred at a much faster rate (within 2 days) than the retraction effectuated by the osteoclast precursors (after 8 days in co-culture). Inhibition of matrix metalloproteinase activity, but not of serine proteinases, strongly reduced retraction of the osteoblasts, thus indicating that this type of cell movement depends on the activity of matrix metalloproteinases. A similar inhibitory effect was found with TGF-beta1. TNF-alpha had no effect on osteoblast retraction but enhanced the formation of multinucleated osteoclast-like cells. Addition of PBMCs to confluent layers of periosteal fibroblasts resulted in similar phenomena as observed in co-cultures with osteoblasts. However, the cell-free areas proved to be significantly smaller, and the number of multinucleated cells formed within cell-free areas was three to four times lower. Our results indicate that osteoclast precursors and mature osteoclasts have the capacity to modulate the activity of osteoblasts and that, yet unknown, membrane-bound signaling molecules are essential in inducing retraction of osteoblasts and the subsequent formation of cell-free areas.