It is well known that bone remodeling starts with a resorption event and ends with bone formation. However, what happens in between and how resorption and formation are coupled remains mostly unknown. Remodeling is achieved by so-called basic multicellular units (BMUs), which are local teams of osteoclasts, osteoblasts, and reversal cells recently proven identical with osteoprogenitors. Their organization within a BMU cannot be appropriately analyzed in common histology. The originality of the present study is to capture the events ranging from initiation of resorption to onset of formation as a functional continuum. It was based on the position of specific cell markers in longitudinal sections of Haversian BMUs generating new canals through human long bones. It showed that initial resorption at the tip of the canal is followed by a period where newly recruited reversal/osteoprogenitor cells and osteoclasts alternate, thus revealing the existence of a mixed "reversal-resorption" phase. Three-dimensional reconstructions obtained from serial sections indicated that initial resorption is mainly involved in elongating the canal and the additional resorption events in widening it. Canal diameter measurements show that the latter contribute the most to overall resorption. Of note, the density of osteoprogenitors continuously grew along the "reversal/resorption" surface, reaching at least 39 cells/mm on initiation of bone formation. This value was independent of the length of the reversal/resorption surface. These observations strongly suggest that bone formation is initiated only above a threshold cell density, that the length of the reversal/ resorption period depends on how fast osteoprogenitor recruitment reaches this threshold, and thus that the slower the rate of osteoprogenitor recruitment, the more bone is degraded. They lead to a model where the newly recognized reversal/resorption phase plays a central role in the mechanism linking osteoprogenitor recruitment and the resorption-formation switch.

Osteoclasts (OCs) seeded on bone slices either drill round pits or dig long trenches. Whereas pits correspond to intermittent resorption, trenches correspond to continuous and faster resorption and require a distinct assembly of the resorption apparatus. It is unknown whether the distinction between pits and trenches has any biological relevance. Using OCs prepared from different blood donors, we found that female OCs achieved increased resorption mainly through pit formation, whereas male OCs did so through trench formation. Trench formation went along with high collagenolytic activity and high cathepsin K (CatK) expression, thereby allowing deeper demineralization. A specific CatK inhibitor abrogated the generation of trenches, while still allowing the generation of pits. OCs obtained from bone marrow were more prone to generate trenches than those obtained from blood. Scanning electron microscopy of bone surfaces eroded in vivo showed trenches and pits of similar size as those made by OCs in culture. We conclude that the distinction between trench- and pit-forming OCs is relevant to the differences among OCs from different skeletal sites, different individuals, including gender, and results from differences in collagenolytic power. This indicates a biological relevance and highlights the importance of discriminating between pits and trenches when assessing resorption.