Currently, it is believed that osteoclasts are immobile during bone resorption and move without resorbing to a new resorption site. This model explains well how round pits are generated, but not how elongated resorption cavities, trenches, are formed. It has been suggested that trenches reflect osteoclasts that resorb bone while moving, thus pointing to an additional resorption model.

Using time-lapse (40 to 70 h, 252 osteoclasts), the present study demonstrates that trenches are indeed made by human osteoclasts that resorb bone parallel to the bone surface while they move for several days non-stop (86% continue to resorb for the duration of the time-lapse) (5 experiments). These osteoclasts represent >50% of the active osteoclasts. Osteoclasts in trench-mode erode twice as fast as in pit-mode (100 µm²/h vs 50 µm²/h). Trenches are also 50% deeper than pits. Overall, the trench-mode is therefore more aggressive than the pit-mode. Furthermore, we have also reported that osteoclasts in trench-mode have a more effective collagenolysis, are much more sensitive to odanacatib than in pit-mode and that glucocorticoid enhances the rate of collagenolysis resulting in more trenches. We now report additional characteristics of the trench-mode. It is 3-fold more sensitive to zoledronic acid than the pit-mode (15 experiments). The anti-floatation-barrier molecule, septin9, has a unique localization in osteoclasts making trenches (11 experiments). A septin antagonist acts ≈3-times stronger on OCs in trench-mode than pit-mode (3 experiments). The same holds true when inhibiting tubulin depolymerization (3 experiments). Finally, the level of trenches made by osteoclasts in vitro correlates with the age of the blood donor (18 donors).

Thus, we show that osteoclasts in trench- and pit-mode are fundamentally different both with respect to resorption orientation, aggressiveness, molecular characteristics and drug sensitivity. We believe that this deserves more attention both in vivo and in vitro.