br Results br Discussion Our studies demonstrate

Results
Discussion Our studies demonstrate that legumain inhibits late stages of ex vivo OB differentiation and the formation of mineralized ECM, suggesting that legumain regulates the ECM-hBMSC interaction that is required for the expression of the mature OB phenotype (Mathews et al., 2012). We identified fibronectin degradation as a mediator of legumain effects, similar to its role in renal proximal tubular cells, which is required for normal renal function (Morita et al., 2007). Fibronectin is known to enhance OB differentiation through interaction with the α5β1 integrin receptor and is required for OB maturation, survival, and matrix mineralization (Brunner et al., 2011; Linsley et al., 2013; Mathews et al., 2012; Moursi et al., 1997). In contrast, fibronectin exerts inhibitory effects on lipid accumulation and AD differentiation (Antras et al., 1989; Rodriguez Fernandez and Ben-Ze\'ev, 1989; Spiegelman and Ginty, 1983). Our data suggest that the proteolytic activity of legumain is important for its effects on hBMSC differentiation and bone formation. Legumain has been reported to regulate bone resorption through inhibition of osteoclast formation and function by its C-terminal fragment (17 kDa), which is enzymatically inactive, suggesting that legumain exerts protease-independent functions (Choi et al., 1999). In addition, legumain exhibits carboxypeptidase (Dall and Brandstetter, 2013) and peptide ligase activity (Dall et al., 2015). The possible involvement of these actions of legumain in regulating skeletal homeostasis requires further studies. We observed that genetic loss- and gain-of-function of legumain were associated with changes in BMSC proliferation. However, regulation of BMSC proliferation by legumain is independent of its enzymatic activity, since small-molecule inhibition of legumain activity did not alter BMSC proliferation. This observation corroborates previous reports showing regulation of cell proliferation by legumain, independent of its enzymatic activity (Andrade et al., 2011). Further support of the potential role of legumain in osteoprogenitor cell proliferation is based on its colocalization with the proliferating osteoblastic Schaftoside in human bone biopsies. The proliferation status of osteoprogenitor cells near the bone-forming surfaces, as determined by immunohistochemical analysis of Ki-67, has been previously reported in human bone specimens (Kristensen et al., 2014). Interestingly, the pattern of Ki-67 immunoreactivity coincided with legumain expression, as the legumain-positive osteoprogenitor cells (such as canopy cells) are also Ki-67 positive, whereas the mature osteoblastic cells (such as bone-lining cells) are both legumain negative and Ki-67 negative. We employed the zebrafish model to investigate the developmental and pharmacological effects of legumain inhibition in vivo. Zebrafish is an attractive model for in vivo screening studies, due to the molecular and cellular conservation of skeletal development and its predictive value when studying human diseases (Fisher et al., 2003; Grimes et al., 2016; Hayes et al., 2014; Li et al., 2009). For example, mutations in zebrafish collagen type IA1 reproduce many aspects of osteogenesis imperfecta (Fisher et al., 2003), and ptk7 mutant zebrafish have been identified as suitable models for scoliosis (Grimes et al., 2016). We extend the usefulness of this model by showing that late developmental events such as the mineralization of vertebrae are amenable to analysis in TALEN-injected animals. Our finding that legumain-deficient zebrafish exhibit enhanced OB differentiation and bone mineralization is consistent with the in vitro effects of legumain knockdown. In addition, we show that post-embryonic pharmacological inhibition of legumain recapitulates the osteogenic effects of genetic legumain ablation. Osteoporosis is a systemic bone disease, characterized by decreased bone formation, reduced bone mass, and disruption of normal bone architecture, resulting in bone fragility and increased risk of fractures (Compston, 2010). Legumain expression was elevated in hBMSCs from osteoporotic patients and, at single-cell resolution, legumain overexpression in ADs inversely correlated with local trabecular bone volume. Bioinformatic analysis revealed the presence of NF-κB binding sites in the legumain promoter, suggesting that the proinflammatory cytokines (e.g., TNF-α and IL6), known to be upregulated in osteoporotic bone marrow microenvironment (Charatcharoenwitthaya et al., 2007), could regulate legumain expression.