The nicotinic acetylcholine receptor α7 subunit is an essential negative regulator of bone mass

The nicotinic receptor α7nAchR reportedly regulates vagal nerve targets in brain and cardiac tissue. Here we show that nAchR7−/− mice exhibit increased bone mass due to decreased osteoclast formation, accompanied by elevated osteoprotegerin/RANKL ratios in serum. Vagotomy in wild-type mice also significantly increased the serum osteoprotegerin/RANKL ratio, and elevated bone mass seen in nAchR7−/− mice was reversed in α7nAchR/osteoprotegerin-doubly-deficient mice. α7nAchR loss significantly increased TNFα expression in Mac1-positive macrophages, and TNFα increased the osteoprotegerin/RANKL ratio in osteoblasts. Targeting TNFα in nAchR7−/− mice normalized both serum osteoprotegerin/RANKL ratios and bone mass. Administration of nicotine, an α7nAchR ligand, to wild-type mice increased serum RANKL levels. Thus, vagal nerve stimulation of macrophages via α7nAchR regulates bone mass by modulating osteoclast formation.

Scientific RepoRts | 7:45597 | DOI: 10.1038/srep45597 reduced bone mass 20,21 . Meanwhile, local osteoclastogenesis increases following stimulation of progenitor cells by receptor activator of nuclear factor kappa B ligand (RANKL) produced by osteoblastic cells and osteocytes 22,23 . By contrast, osteoclastogenesis is inhibited by osteoprotegerin (OPG), a decoy receptor of RANKL, which is also produced by osteoblasts 24 . OPG is encoded by the TNF receptor superfamily 11b (Tnfrsf11b) gene, and Tnfrsf11b −/− mice reportedly exhibit severe osteoporosis due to accelerated osteoclast formation 25,26 . In contrast, OPG-overexpressing transgenic mice show elevated bone mass due to inhibited osteoclastogenesis 27 , and RANKL (encoded by the Tnfsf11 gene)-deficient mice exhibit severe osteopetrotic phenotypes and a complete absence of osteoclast formation 28 . Thus, the OPG/RANKL system plays a crucial role in maintaining bone homeostasis by regulating osteoclast differentiation. OPG expression in osteoblastic cells is down-regulated by treatment with active vitamin D3, 1,25(OH) 2 D 3 , while RANKL is up-regulated 24 . RANKL expression is stimulated by inflammatory cytokines such as TNFα and IL-6 29 .
Here, we show that α 7nAchR is required to stimulate osteoclast formation and reduce bone mass by inhibiting circulating levels of OPG and elevating levels of RANKL. α 7nAchR-deficient mice exhibited increased bone mass due to an elevated OPG/RANKL ratio in serum, and increased bone mass seen in α 7nAchR-deficient mice was rescued in α 7nAchR/OPG double-knockout mice. TNFα expression was aberrantly high in Mac1-positive macrophages among bone marrow cells in α 7nAchR-deficient mice, and increases in bone mass and the OPG/ RANKL ratio seen in α 7nAchR-deficient mice were rescued in α 7nAchR/TNFα double-knockout mice. Our results indicate overall that vagus nerve activity is required to maintain bone homeostasis by controlling systemic levels of OPG and RANKL via regulation of macrophage TNFα expression.

Results
Nicotinic acetylcholine receptor α7-deficient mice exhibit increased bone mass. α 7nAchR is one of several subunits of nicotinic acetylcholine receptors 1 . Evaluation of bone phenotypes using DEXA and micro CT analysis revealed elevated bone mass in α7nAchR −/− relative to WT mice ( Fig. 1a and b). Toluidine blue and TRAP staining of tibial tissues from α7nAchR −/− mice revealed elevated trabecular bone mass and inhibited osteoclast formation, respectively ( Fig. 1c and d). Bone morphometric analysis also demonstrated significantly elevated bone volume per tissue volume (BV/TV), trabecular thickness (Tb. Th) and trabecular number (Tb. N) as well as significantly reduced trabecular separation (Tb. Sp) in α7nAchR −/− compared with WT mice, confirming elevated bone mass seen in the mutants. Since osteoclastic parameters such as eroded surface per bone surface (ES/BS), number of osteoclasts per bone perimeter (N. Oc/B. Pm) and osteoclast surface per bone surface (Oc. S/BS) were significantly reduced in mutants, while osteoblastic parameters such as matrix apposition rate (MAR), bone formation rate (BFR) and osteoblast surface per bone surface (Ob. S/BS), were normal (Fig. 1e), we concluded that elevated bone mass in mutant mice resulted from inhibition of osteoclastogenesis.
To determine whether α 7nAchR regulates osteoclastogenesis directly, we isolated osteoclast progenitor cells from α7nAchR −/− and WT mice and cultured them in the presence of M-CSF and RANKL, with or without the ligands acetylcholine or nicotine (Fig. 2). We found that α 7nAchR was expressed in bone marrow macrophages (BMM) and osteoclasts (OCL) (Fig. 2a). However, osteoclastogenesis in WT cells was not stimulated by either ligand (Fig. 2b,c and d), and osteoclast formation was comparable in α7nAchR −/− and WT cells ( Fig. 2b-g). These observations suggest α 7nAchR expressed in osteoclast progenitor cells does not regulate osteoclastogenesis directly.
Vagal nerve activity regulates circulating OPG and RANKL levels via α7nAchR. To determine how α 7nAchR regulates osteoclastogenesis, we analyzed levels of circulating OPG, a RANKL antagonist, given that OPG inhibits osteoclast differentiation and thus positively regulates bone mass 25,26 . Serum OPG levels as determined by ELISA significantly increased in α7nAchR −/− relative to WT mice (Fig. 3a). To determine whether vagus activity regulates circulating OPG levels, we performed vagotomy in WT mice and observed significantly increased serum OPG levels in vagotomy compared with sham-operated mice (Fig. 3b). Meanwhile, systemic administration of nicotine, a bio-mimetic of acetylcholine, to WT mice downregulated serum OPG levels (Fig. 3c). In contrast to serum OPG, circulating levels of RANKL significantly decreased in α7nAchR −/− relative to WT mice (Fig. 3d). Systemic RANKL levels significantly decreased following vagotomy and increased following nicotine administration in WT mice ( Fig. 3e and f), suggesting that RANKL and OPG levels are regulated by the vagus nerve via α 7nAchR.
Macrophage TNFα regulated by α7nAchR-dependent vagal nerve signaling controls systemic OPG and RANKL levels. Next, we asked which cells were responsible for regulating circulating OPG and RANKL levels in response to vagus nerve signaling (Fig. 5). α 7nAchR is abundantly expressed in brain and heart 30,31 , and thus we initially asked whether these tissues might regulate OPG and RANKL levels. To do so, we undertook analysis of bone-related phenotypes using bone marrow (BM) transfer of WT or α7nAchR −/− BM cells into α 7nAchR −/− or WT mice. We observed that serum OPG levels significantly decreased following transfer of WT BM into mutant mice relative to transplantation with α 7nAchR −/− BM cells (Fig. 5a). In contrast, α7nAchR −/− BM cell transfer into WT mice significantly elevated serum OPG levels compared to WT mice transplanted with WT BM cells (Fig. 5a). These observations suggest that BM cells are responsible for OPG production via α 7nAchR. To determine which cells regulated OPG and RANKL expression in BM cells, we fractionated WT and α7nAchR −/− BM cells into CD3-positive T cells, B220-positive B cells and Mac1-positive monocyte/ macrophage cells, and analyzed Tnfrsf11b (OPG) expression by realtime PCR in each; however, Tnfrsf11b expression was not detected in any of the three populations (data not shown). Serum RANKL levels were downregulated by transfer of α7nAchR −/− BM cell into either WT or α7nAchR −/− mice (Fig. 5b), strongly suggesting that BM cells regulate both OPG and RANKL levels via α 7nAchR indirectly.
IL-1β and TNFα are both reportedly upregulated in α 7nAchR −/− mice 4 , and we found that Tnfrsf11b (OPG) expression was significantly elevated by TNFα but not IL-1β in MC3T3-E1 osteoblastic cells in vitro (Fig. 5c). Similarly, we also found that Tnfsf11 (RANKL) expression was significantly downregulated by TNFα in MC3T3-E1 cells (Fig. 5d). Tnfα mRNA expression was specifically and significantly higher in whole BM cells and in Mac1-positive cells sorted from α 7nAchR-deficient compared to WT mice (Fig. 5e). We then administered the TNFα -inhibitor, Etanercept, a soluble receptor of TNFα , to mice of either genotype and observed  downregulation of serum OPG levels in α7nAchR −/− but not in WT mice (Fig. 5f). By contrast, RANKL levels were significantly elevated following Etanercept administration to α7nAchR −/− but not WT mice (Fig. 5g).
Taken together, our results suggest that vagus nerve stimulation enhances osteoclast formation and reduces bone mass by suppressing serum OPG levels and elevating serum RANKL levels through inhibition of α 7nAchR-dependent macrophage TNFα expression.

Discussion
Bone homeostasis is orchestrated by a combination of local and systemic factors physiologically and perturbed in pathological inflammatory conditions. Here, we demonstrate that bone homeostasis is regulated by vagus nerve activity via α 7nAchR. Our model suggests that osteoclast formation is enhanced by down-regulation of OPG ). (f) Levels of serum RANKL in WT mice administered nicotine or control saline (n = 6). All data represent mean serum OPG or RANKL concentrations ± SD (*p < 0.05; **p < 0.01; ***p < 0.001; n = 6). Representative data of at least two independent experiments are shown.
Scientific RepoRts | 7:45597 | DOI: 10.1038/srep45597 levels and up-regulation of serum RANKL levels following inhibition of macrophage TNFα expression, the latter controlled by α 7nAchR-dependent vagus nerve activity (Fig. 7). α 7nAchR loss elevated TNFα expression in macrophages, leading to elevated OPG and reduced RANKL levels in sera, and turn inhibiting osteoclast formation and increasing bone mass. Inhibition of TNFα in α7nAchR −/− mice reversed the perturbed OPG/RANKL ratio status and normalized bone mass. Vagotomy elevated OPG and decreased RANKL levels in WT mouse sera, a phenotype similar to that seen in α7nAchR −/− mice. Finally, nicotine administration reduced OPG and increased RANKL levels in serum of WT mice.
The vagus nerve controls heart, brain and gut function and/or development, but no obvious abnormality is seen in these organs in α 7nAchR-deficient mice 1 . Instead, α 7nAchR reportedly functions to inhibit inflammation and septic reactions 4,8,9 , and thus vagus nerve control of an anti-inflammatory axis likely suppresses excessive immune responses. Here, we show that vagal nerve activity is required to regulate bone homeostasis by controlling Mac1-positive cell expressing TNFα -mediated OPG and RANKL expression in physiological conditions. Elevated serum OPG levels seen in α 7nAchR-deficient mice were partially but significantly restored by deletion of the gene encoding TNFα , suggesting that α 7nAchR regulates OPG levels through factors other than TNFα and IL-1β . Since our bone marrow transplantation analysis demonstrated that bone marrow-derived Mac1-positive cells regulate serum OPG levels, cholinergic anti-inflammatory signaling in those cells must control those levels. This cholinergic anti-inflammatory pathway reportedly functions in macrophages 4 . Mac1 is reportedly expressed in monocytes/macrophages 32 , and these cells are considered regulate serum OPG levels via α 7nAchR. Mac1 is also expressed in neutrophils and natural killer cells 33,34 . Further studies are needed to determine the lineage of Mac1-positive cells that regulates TNFα via α 7nAchR.
Moreover, TNFα promotes RANKL expression in osteoblastic cells 35,36 ; however, in this study, we showed that RANKL expression in osteoblastic cells is inhibited by TNFα . TNFα activates MAPK and NFκB pathways 35,36 , and our data suggest that the MAPK rather than the NFκB pathway contribute controls OPG by TNFα signaling (Fig. S1).
α 7nAchR is expressed in osteoblasts 37 , an observation we confirmed here ( Fig. S2a and S2b). Whether α 7nAchR-deficient mice exhibit bone phenotypes is controversial: some authors reported normal bone mass based on 3D micro CT analysis 10 , while others have shown elevated bone mass based on micro CT analysis 11 . Here, we show relatively elevated bone mass in α 7nAchR-deficient based on DEXA, micro CT and bone histomorphometric analysis. Furthermore, in the absence of nicotine or acetylcholine stimulation, others previously showed that osteoclastogenesis is inhibited in α 7nAchR-deficient compared to wild-type cells in vitro 11 . However, our data indicates that bone homeostasis is regulated via α 7nAchR systemically rather than locally through serum OPG and RANKL. Nicotine is known to stimulate calcium signals in chondrocytes 37 . We have preliminary data showing that nicotine or acetylcholine stimulates calcium signals in macrophages; however, nicotine stimulation reportedly inhibits RANKL-induced calcium oscillation 11 . Thus, mechanisms used by α 7nAchR to regulate a cholinergic anti-inflammatory pathway in macrophages have not been clarified, and further studies are required to determine how α 7nAchRs inhibit inflammation.
The sympathetic and parasympathetic nervous systems cooperatively regulate functions of various organs. However, sympathetic activity reportedly inhibits osteoblastogenesis, leading to bone loss 18 . Here, we showed  that parasympathetic activity stimulates osteoclast formation and reduces bone mass. Thus both the sympathetic and parasympathetic systems apparently reduce bone mass via inhibiting bone-formation and stimulating bone-resorption, respectively.
Inflammation is a risk factor for bone loss, and patients with inflammatory disease such as RA exhibited osteoporosis 37 . TNFα is a major inflammatory cytokine upregulated in RA patients and is implicated in RA pathogenesis 38 . Transgenic mice overexpressing human TNFα exhibit RA-like phenotypes, such as joint inflammation and bone erosion due to increased osteoclast formation 17 . Indeed, treatment with biologics targeting TNFα is effective in treating RA patients 15,39 . TNFα expression is also promoted by pathologic conditions such as bacterial or viral infection or injury, while negative feedback regulation of TNFα -induced inflammation is mediated by induction of anti-inflammatory cytokines 40 . We show here that vagus nerve activity inhibits TNFα expression by macrophages via α 7nAchR, and the pathology in both collagen-induced arthritis or sepsis models is more severe in α7nAchR7 −/− than in control mice 8,9 . Our study suggests that in the bone system, TNFα -mediated OPG induction and RANKL suppression likely represent a feedback mechanism to protect from excessive bone loss.
TNFα reportedly promotes osteoclast formation directly without RANKL 32,41 . However, bone phenotypes have not been reported in TNFα -deficient mice, suggesting that TNFα does not play a central role in regulating bone mass in physiological conditions but rather enhances bone erosion in pathological contexts. We show here that bone mass is elevated even under high TNFα conditions in α7nAchR −/− mice, indicating that TNFα -induced OPG expression and RANKL suppression is dominant over TNFα -promoted bone loss in physiological conditions. OPG is a decoy soluble receptor of RANKL and blocks osteoclast differentiation signals to the RANK receptor. In physiological conditions, OPG and RANKL expression is controlled in osteoblastic cells by osteotropic factors such as 1,25(OH) 2 D 3 and PGE2 24 . In contrast, OPG expression reportedly increases in pathological conditions such as LPS administration 42 . RANKL expression is also stimulated by inflammatory cytokines 29 . Recently, OPG expression was reported to be induced via the PHD-HIF2α axis 43 . Here, we show that the vagus nerve-TNFα axis regulates serum OPG and RANKL levels.
The natural ligand of nAchRs is acetylcholine, an activity mimicked by nicotine. Smoking is a known risk factor for development of osteoporosis/osteopenia 12 , and is implicated as a negative factor for fracture healing 44 . In the current study, we show for the first time that administration of nicotine to WT mice significantly reduced levels of circulating OPG and elevated levels of RANKL. This observation may explain, at least in part, the negative effect of nicotine on bones. Smoking is also seen as a risk factor for RA development 45 .
Taken together, our data present new insight into regulation of bone homeostasis by the vagus nerve and how this interaction is mediated by inflammatory cytokine signaling.
Animals were maintained under specific pathogen-free conditions in animal facilities certified by the Keio University animal care committee. Animal protocols were approved by that committee and carried our in accordance with the committee's guidelines.
Vagotomy. Bilateral subdiaphragmatic vagotomy or sham surgery was performed in 8-week-old WT mice as described 49 . Briefly, the stomach and lower esophagus were exposed by laparotomy. The stomach was gently retracted downward beneath the diaphragm to visualize both vagal trunks and then the bilateral vagus nerve was removed. For pyloroplasty to prevent gastric stasis, an incision was made parallel to the axis of the pylorus through the pyloric sphincter, and the pylorus wall was reconstructed by sutures perpendicular to the pylorus axis. The stomach was returned to its normal position, and incisions were closed. For sham-operated animals, after opening the abdominal cavity, pyloroplasty only was performed.