FSTL3 Mediates Exercise Driven Bone Formation

Introduction: Timely healing of fractures or boney defects of the hand is of utmost importance, both to reduce the economic burden of lost work days, as well as to maximize function of the post-injured hand.  Thus, adjuvant therapies that expedite healing are much needed.  Exercise is known to promote bone remodeling in bones that are directly biomechanically stimulated, yet little is known about its secondary effects at non-stimulated sites.  Furthermore, the role of follistatin-like 3 (FSTL3; a mediator of exercise-driven bone formation) in this process has not been previously shown.  The objective of this study was to examine the influence of biomechanical signals generated by exercise and FSTL3 in inducing bone formation in distant, non-stimulated regions.

Methods: Four millimeter diameter, full thickness calvarial defects were made in female C57/BL and FSTL3 knockout (FSTL3 -/-) mice. Subsequently, tissue engineered scaffolds containing bone marrow derived mesenchymal stem cells (BMSCs) were placed in the defects and the incisions closed.  Mice (n=5/gr) were either subjected to: (i) no treatment; (ii) gentle treadmill walking (TW; 7m/min for 45min/day); (iii) subcutaneous injection of BMP-2 (0.5 µg/mouse); or (iv) injection with FSTL3 (300 ng/mouse). The calvaria were harvested 6 or 12 weeks later and examined for bone formation by µCT, histology, and immunohistochemistry.

Results: Bone formation to some extent was observed in all defects implanted with BMSC-containing scaffolds. However, TW induced significantly greater (4 ± 0.6 fold) vascularized bone formation. A 3.2 ± 0.7 fold increase in bone formation was observed in mice injected with FSTL3; likewise, a similar increase of bone formation (2.9 ± 0.6 fold) was observed in mice injected with BMP-2. Histologically, the bone formation in mice subjected to TW demonstrated a well-integrated bone that was homogeneously connected with the calvarial bone. However, bone integration in response to FSTL3 or BMP-2 was incomplete 12 wks following implantations.  Importantly, TW failed to induce bone formation in Fstl3-/- mice.

Conclusions: Our data suggest that biomechanical stimulation is a potent inducer of bone formation, even in bones that are not directly stimulated.  Patients could possibly benefit from prescribed exercise programs. Furthermore, the identification of FSTL3 as an inducer of bone formation provides a novel paradigm for future therapies aimed at augmenting bone formation, especially in large or difficult to heal bone defects.