Oxytocin is an anabolic bone hormone

Abstract

We report that oxytocin (OT), a primitive neurohypophyseal hormone, hitherto thought solely to modulate lactation and social bonding, is a direct regulator of bone mass. Deletion of OT or the OT receptor (Oxtr) in male or female mice causes osteoporosis resulting from reduced bone formation. Consistent with low bone formation, OT stimulates the differentiation of osteoblasts to a mineralizing phenotype by causing the up-regulation of BMP-2, which in turn controls Schnurri-2 and 3, Osterix, and ATF-4 expression. In contrast, OT has dual effects on the osteoclast. It stimulates osteoclast formation both directly, by activating NF-kappaB and MAP kinase signaling, and indirectly through the up-regulation of RANK-L. On the other hand, OT inhibits bone resorption by mature osteoclasts by triggering cytosolic Ca(2+) release and NO synthesis. Together, the complementary genetic and pharmacologic approaches reveal OT as a novel anabolic regulator of bone mass, with potential implications for osteoporosis therapy.

Fig. 1.

Direct skeletal action of OT. Loss-of-function studies demonstrated severe trabecular bone loss in 3- or 6-month-old OT- and Oxtr-null mice and heterozygotes compared with wild-type littermates. Von Kossa-stained sections of the 5th lumbar vertebrae (A; representative) revealed significant reductions in trabecular bone volume (BV/TV, %) at both 3 and 6 months (C), but with no sex-specific differences at 6 months (D). This profound trabecular bone loss was confirmed in Oxtr-deficient mice by μ-CT of the femoral epiphysis (B; representative). Statistics: Student t test, P values as indicated comparing OT- and Oxtr-deficient mice against wild-type littermates (n = 8 mice/group). Indirect skeletal effects of OT through hypothalamic OT-ergic neurones were examined by ICV injection of OT (25 ng/day) via Alzet pumps for 5 days. We found no significant differences in plasma C-telopeptide (E) or OC (F) level; moreover, TRAP-positive osteoclasts (G and H) or alkaline phosphatase-positive CFU-f (I) did not enhance in ex vivo bone marrow cell cultures of ICV-injected mice. The finding that food intake was stimulated by ICV injections (K; percentage of control at day 1) indicates that OT does exert a central action, but that this action does not affect the skeleton. In contrast, 2 i.p. injections of OT (4 μg/mouse) given to 2-month-old mice 12 h apart resulted in a significant increase in TRAP-positive osteoclast formation (J), further supporting a peripheral, rather than a central, action of OT. Statistics: Student t test; P values shown; n = 8 mice per group; 2 experiments pooled. The peripheral action of OT was consistent with the presence of Oxtr on both mouse and human osteoblasts (L) and osteoclasts (M). The cells were first labeled with fluorescent phalloidin (green) to delineate the actin filaments, and thereafter with an anti-Oxtr antibody (red). The experiment was repeated with 3 different batches of human or mouse cells; representative cells are shown. The bottom panels represent Oxtr internalization, seen as intracellular staining, after a 12-hour exposure to OT (10 nM), further confirming functional Oxtr specificity.

Fig. 2.

OT stimulates osteoblastic bone formation. OT deficiency reduced MAR, an index of bone formation in calcein-labeled calvaria from 7-week-old mice (A and B), as well as mineralization (C) and proliferation (D) in ex vivo bone marrow cell cultures, evident on von Kossa staining and the MTT assay, respectively. OT (10 nM) partly rescued the ex vivo mineralization defect (C). Statistics: Student t test comparing OT^−/− with wild-type mice; n = 4 mice per group; *, P < 0.05. In contrast, OT (at stated doses) stimulated the proliferation (E) and differentiation (F) of wild-type murine osteoblast precursors in the MTT assay and qPCR for the osteoblast markers OPN and OC, respectively. OC and OPN mRNA was expectedly low in ex vivo murine OT^−/− osteoblast cultures compared with wild-type cultures (G). This result was associated with reduced Osx protein [H; Western blot analysis, at time 0 (confluence)] and mRNA [I; qPCR, at time 0 (confluence) and at 1 and 2 weeks postdifferentiation induction], but with elevated Runx2 protein and mRNA in OT^−/− osteoblasts. Likewise, BMP-2 and ATF4 (J), as well as Schnurri (Shn) isoforms 1, 2, and 3 (K and L), were reduced dramatically at 0 weeks in OT^−/− osteoblasts compared with wild-type controls, with the exception of Shn3 at 1 and 2 weeks. Statistics: Student t test comparing OT^−/− with wild-type mice at every time point; *, P < 0.05; **, P < 0.01 (in triplicate).

Fig. 3.

OT acutely inhibits resorption in the face of increased osteoclastogenesis. OT deficiency reduced TRAP-positive osteoclast formation at day 6 (A), as well as precursor proliferation (B; MTT assay) and markers of osteoclast differentiation (C; as shown, qPCR) at day 12 of ex vivo murine bone marrow cell culture. As expected, recombinant OT (≤100 nM) stimulated TRAP-positive osteoclasts (D) and enhanced both precursor proliferation [E; human peripheral blood mononuclear cells (PBMCs), MTT assay] and differentiation, as evidenced by elevated osteoclast marker mRNA (F; as stated, qPCR). Reciprocal effects of OT deficiency (G) and recombinant OT (H) on RANK-L and OPG mRNA (qPCR) and protein (I; Western blot analysis) were seen in osteoblasts. RANK-L expression was stimulated and OPG expression was inhibited by OT, suggesting that OT-induced osteoclastogenesis may be exerted in part through an altered RANK-L/OPG ratio. We explored the relevance of enhanced RANK-L through a transwell coculture experiment, in which osteoclast precursors were on the filter and osteoblasts were on the plate. In the absence of added RANK-L (i.e., with M-CSF alone; 50 ng/mL), TRAP-positive osteoclasts were formed (J) only in the presence of added OT, suggesting the release of diffusible soluble RANK-L in response to OT. Whereas osteoclastogenesis was stimulated by OT in precursors plated on dentine (K, Upper), bone resorption, as assessed by toluidine-blue staining for pits, was sharply reduced within 48 h (K, Lower). The acute inhibition of resorption was consistent with a ≈30% decline in the resorption of ³H-proline-labeled bone fragments, which was reversed by a specific OT antagonist, atosiban (L; doses as stated), as well as with a reduction in supernatant cross-laps (see Fig. 4I). Statistics: Student t test comparing OT^−/− with wild-type mice or vehicle (zero dose) with OT (various doses); qPCR and MTT in triplicate; bone resorption, 4 slices per treatment (representative); ³H-proline resorption assay in quadruplicate; cultures in triplicate (representative); *, P < 0.05; **, P < 0.01.

Fig. 4.

OT triggers MAP kinase and NF-κB activation and Ca²⁺ release in osteoclasts. In Western blot analyses by using whole-cell lysates of osteoclasts, OT (10 nM) stimulated the phosphorylation of Erk within 5 min (A) and that of Akt (B) and IκBα (C) within 15 min. The respective loading controls were total Erk, total IκBα, and actin, all of which are pro-osteoclastogenic signals (21). In addition, cytosolic Ca²⁺ concentration (nM) was measured in single, isolated, fura-2-loaded, mature human osteoclasts treated with OT (10 nM) in the presence or absence of thapsigargin (4 μM) to block Ca²⁺ uptake or EGTA (3 mM) to chelate extracelular Ca²⁺. (D) The rise of cytosolic Ca²⁺ (a, Top) was inhibited by thapsigargin (b, Middle) but not by extracellular Ca²⁺ chelation (c, Bottom), indicating that the OT-induced Ca²⁺ transient is attributed mainly to the release of intracellularly stored Ca²⁺. OT also stimulated expression of the Ca²⁺-sensitive isoform of NOS (eNOS) in a time-dependent manner, as shown by qPCR (E) and Western blot analysis (F; actin as loading control). The time course of the elevated eNOS expression was consistent with the time course of the increase in nitrite (NO₂) production when osteoclasts were cultured on plastic (G) or on bone (H). The functional role of NO in OT-induced resorption inhibition was assessed by using a specific eNOS blocker, L-NAME (I). Statistics: Student t test comparing zero dose with OT (various doses); qPCR in triplicate; cytosolic Ca²⁺ (representative); NO₂ and cross-lap assay in quadruplicate; *, P < 0.05; **, P < 0.01.