During osteoporosis, osteoclastic bone resorption overwhelms bone formation leading to loss of bone mass and increased fragility (Feng and McDonald, 2011). Dynamic bone resorption requires mature osteoclasts with intact actin belts. Here we examined the role of Sirt1 activators (RSV, SRT2183 and SRT1720) on osteoclastic bone resorption, ex vivo. We found that SRT2183 and SRT1720 inhibit the formation of osteoclasts and actin belts in primary BMCs derived from both WT and OC-Sirt1KO mice, while RSV does not. We also observed SRT2183 and SRT1720 inhibit bone resorption by disrupting actin belts of mature osteoclasts. Our findings contrast with previous studies that showed reduced osteoclast formation with RSV treatment in various cell types, such as monocytes from human peripheral blood mononuclear cells, primary bone derived cells from canine bone fragments, and RAW264.7 cells (Boissy et al., 2005; He et al., 2010; Shakibaei et al., 2011; Feng et al., 2018). To our knowledge, no previous ex vivo studies have shown the effect of RSV on osteoclastogenesis in mouse primary BMCs. We chose a 5 µM concentration of RSV based on our cell viability assay data. To our surprise, 5 µM RSV did not reduce osteoclastogenesis and instead appears to increase osteoclast formation in both BMCs and RAW264.7 cells. In contrast, He et al. showed that 3 µM RSV reduced osteoclastogenesis in RAW264.7 cells (He et al., 2010). We note that although RSV did not inhibit osteoclast and actin belt formation in our experiments, it increased phosphorylation of AMPKα (S4 Fig) in both BMMs and RAW264.7 cells, which indicates that RSV activated Sirt1, consistent with previous reports (Price et al., 2012; Lan et al., 2017). Thus, the differences between our findings and those of He et al. may be due to the number of RAW264.7 cells plated (20,000 cells versus 12,500 cells per well) and the concentration of RANKL used to differentiate RAW264.7 cells to osteoclasts (40 ng/ml of RANKL versus 3 ng/ml of RANKL) (He et al., 2010).
However, we observed that both SRT2183 and SRT1720 significantly inhibit RANKL induced differentiation of primary BMCs from young male mice and RAW264.7 to osteoclasts. To our knowledge, no studies have tested the ex vivo effect of SRT1720 on osteoclastogenesis in primary BMCs. Nevertheless, a recent in vivo study by Zainabadi et al., demonstrated that SRT1720 treatment improved bone health in middle-aged male mice and in overiectomized female mice (Zainabadi et al., 2017). Furthermore, consistent with our findings, Gurt et al., observed fewer osteoclasts after treatment of primary BMCs harvested from 8-week-old female 129/Sv mice with 1 uM SRT2183 ex vivo (Gurt et al., 2015).
As Sirt1 is known to regulate bone mineral metabolism, we studied osteoclasts generated from BMCs harvested from osteoclast specific Sirt1 knockout (OC-Sirt1KO) mice. We found that OC-Sirt1KO mice exhibited less bone mineral density, and BMCs harvested from these mice yielded more osteoclasts as compared to WT mice. In support of these observations, previous studies have also observed poor bone mass and greater osteoclastogenesis in OC-Sirt1KO mice as compared to the littermate controls (Edwards et al., 2013; Zainabadi et al., 2017). This supports a role for Sirt1 in osteoclastogenesis. Surprisingly, although SRT2183 and SRT1720 are considered as Sirt1 activating compounds (Feige et al., 2008; Smith et al., 2009; Yamazaki et al., 2009), both of these agents inhibited osteoclast and actin belt formation in BMCs harvested from the OC-Sirt1KO male mice. These findings are consistent with the previous study by Gurt et al., showing that SRT2183 inhibits osteoclast formation and resorption activity in BMCs from constitutive Sirt1 knockout female mice (Gurt et al., 2015). In addition, Huber et al., and Pacholec et al., demonstrated Sirt1 independent effects of SRT2183 and SRT1720 on mouse embryonic fibroblasts and human bone osteosarcoma epithelial cells (U2OS), but not on BMCs or osteoclasts (Huber et al., 2010; Pacholec et al., 2010). In contrast, studies have also demonstrated Sirt1-dependent beneficial impacts of these Sirt1 activating compounds (Hubbard et al., 2013; Mercken et al., 2014). Our in vivo data shows that Sirt1 may take part in bone mineral metabolism by altering BMD in OC-Sirt1KO mice, but is not required for SRT2183 or SRT1720 to inhibit actin belt organization and osteoclastogenesis ex vivo.
Organized actin rings form a stable actin belt in mature osteoclast, which is an important factor for osteoclastic bone resorption. Actin belts generate a sealing zone under which energy consuming dynamic bone resorption occurs (Saltel et al., 2008), and it has been shown that the targeted disruption of genes responsible for actin cytoskeleton organization leads to poor bone resorption (Soriano et al., 1991; Lowe et al., 1993). In our study, SRT2183 and SRT1720 consistently disrupted the actin belts of mature osteoclasts and efficiently inhibited resorption pit formation in cortical bovine bone slices. Gurt et al., also demonstrated poor pit formation by SRT2183 treated osteoclasts derived from mouse BMCs (Gurt et al., 2015).