Neuronal Induction of Bone-Fat Imbalance through Osteocyte Neuropeptide Y

Abstract

A differentiation switch of bone marrow mesenchymal stem/stromal cells (BMSCs) from osteoblasts to adipocytes contributes to age- and menopause-associated bone loss and marrow adiposity. Here it is found that osteocytes, the most abundant bone cells, promote adipogenesis and inhibit osteogenesis of BMSCs by secreting neuropeptide Y (NPY), whose expression increases with aging and osteoporosis. Deletion of NPY in osteocytes generates a high bone mass phenotype, and attenuates aging- and ovariectomy (OVX)-induced bone-fat imbalance in mice. Osteocyte NPY production is under the control of autonomic nervous system (ANS) and osteocyte NPY deletion blocks the ANS-induced regulation of BMSC fate and bone-fat balance. γ-Oryzanol, a clinically used ANS regulator, significantly increases bone formation and reverses aging- and OVX-induced osteocyte NPY overproduction and marrow adiposity in control mice, but not in mice lacking osteocyte NPY. The study suggests a new mode of neuronal control of bone metabolism through the ANS-induced regulation of osteocyte NPY.

Keywords: adipogenesis; autonomic nervous system; bone marrow mesenchymal stem/stromal cells; neuropeptide Y; osteocyte; osteogenesis.

Figure 1

Osteocytes secrete excess NPY during skeletal aging and osteoporosis to shift BMSC fate from osteogenesis towards adipogenesis via Y1R. A) ARS and ORO staining images of BMSCs treated with the culture media (OCY‐CM) from different donor mice‐derived osteocytes or un‐cultured medium (Control) under osteogenic or adipogenic induction. OVX: ovariectomy. Scale bar: 50 µm. B) Quantitation of the percentages of ARS⁺ and ORO⁺ areas. n = 3 per group. Western blotting for NPY protein C) in different tissues from 4‐month‐old male wild‐type mice, or D) in the bone marrow‐depleted femurs from normal male mice at different ages or female mice subjected to OVX or sham operation. NPY expression in the bone from 4‐month‐old male wild‐type mice was tested by E) immunofluorescence double staining for DMP1/NPY and F) immunohistochemical staining for NPY. Arrows indicate the representative NPY⁺/DMP1⁺ or NPY⁺ signals. BM: bone marrow; CB: cortical bone. Scale bar: 50 µm. G) Semiquantitative PCR analysis of mRNA expression of NPY and its five receptors, and H) western blotting for NPY protein in different bone and bone marrow cell types from 8‐month‐old male wild‐type mice. OB: osteoblast; OCY: osteocyte; Mo: monocyte/macrophage; OC: osteoclast. NPY protein in the osteocyte lysates and OCY‐CM was measured by I) western blotting and J) ELISA, respectively. n = 5 per group. K) ARS and ORO staining images and L) quantification of the positively stained areas in BMSCs treated with vehicle or NPY protein under osteogenic or adipogenic induction. Scale bar: 50 µm. n = 3 per group. M,N) Knockdown of NPY in osteocytes using siRNAs blocked the capacities of their culture media (OCY^siNPY‐CM) to inhibit osteogenesis and promote adipogenesis of BMSCs. Scale bar: 50 µm. n = 3 per group. O, P) Blockade of Y1R on the differentiating BMSCs by the Y1R antagonist BIBO3304 abolished the anti‐osteogenic and pro‐adipogenic effects of OCY‐CM. Scale bar: 50 µm. n = 3 per group. Q) ELISA for NPY protein. n = 3 per group. Data are presented as mean ± SD. For panels (L) and (P): unpaired, two‐tailed student's t‐test. For other dot plots: one‐way ANOVA with Bonferroni post hoc test. *P < 0.05, **P < 0.01, and ***P < 0.001.

Figure 2

Deletion of NPY in osteocytes attenuates aging‐ and OVX‐induced bone loss and marrow adiposity. A) Western blotting for NPY protein in osteocytes and osteoblasts from 4‐month‐old male Npy^fl/fl mice and Dmp1‐iCre; Npy^fl/fl mice. B) Immunohistochemical staining for NPY in the bone tissues. Scale bar: 50 µm. C) µCT reconstruction images and D) quantification of bone microarchitecture parameters in femurs from aged‐matched male Npy^fl/fl mice and Dmp1‐iCre; Npy^fl/fl mice. Tb. BV/TV: trabecular bone volume fraction; Tb. N: trabecular number; Tb. Th: trabecular thickness; Tb. Sp: trabecular separation; Ps. Pm: periosteal perimeter; Ct. Th: cortical thickness. Scale bar: 1 mm. n = 10 per group. E) Immunofluorescence staining images for perilipin and F) quantification of adipocyte number in distal femurs. Scale bar: 50 µm. n = 5 per group. G) OCN immunohistochemical staining images and H) the number of OCN‐stained osteoblasts (N. OBs) on trabecular BS in distal femurs. Scale bar: 25 µm. n = 5 per group. I) ELISA for serum OCN. n = 5 per group. J) Calcein double labeling of trabecular bones and K) quantification of BFR/BS and MAR. Scale bar: 25 µm. n = 5 per group. L) µCT reconstruction images and M) quantification of bone microarchitecture parameters in femurs from sham‐ or OVX‐operated female Npy^fl/fl mice and Dmp1‐iCre; Npy^fl/fl mice. Scale bar: 1 mm. n = 10 per group. N) Immunostaining images for perilipin in distal femurs and O) quantification of adipocyte number. Scale bar: 50 µm. n = 5 per group. P) Quantification of OCN⁺ osteoblast number in distal femurs. n = 5 per group. Q) ELISA for serum OCN. n = 5 per group. R) Quantification of BFR/BS and MAR. n = 3 per group. Data are presented as mean ± SD. Two‐way ANOVA combined with Bonferroni post hoc test. *P < 0.05, **P < 0.01, and ***P < 0.001.

Figure 3

The impact of osteocyte NPY deletion on the ability of OCY‐CM to regulate BMSC fate and on the differentiation potential of BMSCs. A) ARS and ORO staining images and B) quantification of the positively stained areas showing the effects of OCY‐CM from Npy^fl/fl mice and Dmp1‐iCre; Npy^fl/fl mice on BMSC differentiation towards osteogenesis and adipogenesis, respectively. Scale bar: 50 µm. n = 3 per group. C) ARS and ORO staining images and D) quantification of the positively stained areas showing the in vitro osteogenic and adipogenic differentiation potential of BMSCs from Npy^fl/fl mice and Dmp1‐iCre; Npy^fl/fl mice. Scale bar: 50 µm. n = 3 per group. Data are presented as mean ± SD. For panel (B): unpaired, two‐tailed student's t‐test (differences between Dmp1‐iCre; Npy^fl/fl OCY‐CM and Dmp1‐iCre; Npy^fl/fl OCY‐CM + NPY groups) or one‐way ANOVA combined with Bonferroni post hoc test (differences among other groups except Dmp1‐iCre; Npy^fl/fl OCY‐CM + NPY group). For panel (D): unpaired, two‐ailed student's t‐test. ***P < 0.001.

Figure 4

Osteocyte NPY‐induced BMSC fate switching is mediated by the inhibition of TEAD1 and JUNB through cAMP/PKA/CREB signaling. qRT‐PCR analysis of Tead1 and Junb expression in BMSCs receiving different treatments under A) osteogenic or B) adipogenic induction for 24 h, or C) in undifferentiated BMSCs with different treatments for 24 h. n = 3 per group. D) qRT‐PCR and E) western blotting confirmed the overexpression efficiency of recombinant Tead1 and Junb lentiviruses in BMSCs. n = 3 per group. F) ARS and ORO staining images and G) quantification of the positively stained areas showing the reduced activity of NPY to inhibit osteogenesis and promote adipogenesis of BMSCs overexpressing Tead1 or Junb. Scale bar: 50 µm. n = 3 per group. ELISA for cAMP in BMSCs receiving different treatments under H) osteogenic or I) adipogenic induction for 24 h. n = 3 per group. PKA activity assay for cell homogenates of BMSCs receiving different treatments under J) osteogenic or K) adipogenic induction for 24 h. n = 3 per group. L) Western blotting for CREB and phosphorylated CREB (p‐CREB) in BMSCs receiving different treatments under osteogenic or adipogenic induction for 24 h. qRT‐PCR analysis of M) Tead1 and N) Junb expression in BMSCs receiving different treatments under osteogenic induction for 24 h. n = 3 per group. O) ARS and ORO staining images and P) quantification of the positively stained areas in BMSCs receiving different treatments under osteogenic or adipogenic induction. Scale bar: 50 µm. n = 3 per group. Q) ChIP‐qRT‐PCR analysis shows enrichment of the CREB antibody (CREB Ab)‐immunoprecipitated Tead1 or Junb promoter region relative to the input DNA. Normal rabbit anti‐IgG served as a negative control. Histone H3 antibody (Histone H3 Ab) pulldown for the enrichment of Rpl30 gene served as a positive control. n = 3 per group. Data are presented as mean ± SD. For panels (A), (B), (G), (H–K), (M), (N), and (P): unpaired, two‐tailed student's t‐test (differences between un‐induced and control groups in (A), (B), and (H–K), or between control and NPY groups in (G), (M), (N), and (P)) or one‐way ANOVA combined with Bonferroni post hoc test (differences among other groups except un‐induced group in (A), (B), and (H–K), or except control group in (M), (N), and (P)). For panel (C) and (D): one‐way ANOVA combined with Bonferroni post hoc test. For panel (Q): unpaired, two‐tailed student's t‐test (differences between control and NPY groups for Rpl30) or two‐way ANOVA combined with Bonferroni post hoc test (differences among other groups for Tead1 or Junb). *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 5

Opposite modulation of osteocyte NPY production and osteocyte‐induced regulation of BMSC fate by NE and ACh through β2AR and M3R. A) Semiquantitative PCR analysis of mRNA expression of α‐ and β‐adrenergic receptors (ARs), muscarinic ACh receptors (mAChRs), and nicotinic (nAChRs) ACh receptors in BMSCs, osteoblasts, and osteocytes from 8‐month‐old male wild‐type mice. B) ARS and ORO staining images of BMSCs with different treatments under osteogenic or adipogenic induction and quantification of the percentages of C) ARS⁺ and D) ORO⁺ areas. Scale bar: 50 µm. n = 3 per group. E) ELISA for NPY in un‐cultured medium (Control) and OCY‐CM from osteocytes treated with vehicle (OCY‐CM), NE (OCY^NE‐CM), or ACh (OCY^ACh‐CM). n = 6 per group. F) ELISA for NPY in un‐cultured medium (Control) and OCY‐CM from osteocytes treated with control siRNAs (OCY^siCon‐CM), β2AR siRNAs (OCY^siβ2AR‐CM), siCon + NE (OCY^siCon+NE‐CM), or siβ2AR + NE (OCY^siβ2AR+NE‐CM). n = 6 per group. G) ARS and ORO staining images of BMSCs with different treatments under osteogenic or adipogenic induction and H) the percentages of ARS⁺ and ORO⁺ areas. Scale bar: 50 µm. n = 3 per group. I) ELISA for NPY in un‐cultured medium (Control) and OCY‐CM from osteocytes treated with siCon (OCY^siCon‐CM), M3R siRNAs (OCY^siM3R‐CM), siCon + ACh (OCY^siCon+ACh‐CM) or, siM3R + ACh (OCY^siM3R+ACh‐CM). n = 6 per group. J) ARS and ORO staining images of BMSCs receiving different treatments under osteogenic or adipogenic induction and K) the percentages of ARS⁺ and ORO⁺ areas. Scale bar: 50 µm. n = 3 per group. Data are presented as mean ± SD. One‐way ANOVA combined with Bonferroni post hoc test. ***P < 0.001.

Figure 6

Osteocyte NPY mediates the ANS‐induced regulation of bone‐fat balance. ELISA for NE and ACh in homogenates of the marrow‐depleted femurs A) from normal male mice at different ages (n = 6 per group) or B) from sham‐ and OVX‐operated female mice (n = 5 per group). C) ELISA for NPY in the marrow‐depleted femur homogenates from 15‐month‐old male Npy^fl/fl mice receiving different treatments for two months. n = 5 per group. Cle: clenbuterol; Riv: rivastigmine. D) µCT reconstruction images and E) quantification of bone microarchitecture parameters in femurs from 15‐month‐old male Npy^fl/fl mice and Dmp1‐iCre; Npy^fl/fl mice receiving different treatments for two months. Scale bar: 1 mm. n = 10 per group. F) Immunofluorescence staining images for perilipin and G) quantification of adipocyte number in distal femurs. Scale bar: 50 µm. n = 5 per group. H) Quantification of osteoblast number in distal femurs. n = 5 per group. I) ELISA for NPY protein in the marrow‐depleted femur homogenates from 3‐month‐old female Npy^fl/fl mice subjected to sham or OVX operation with different treatments for two months. n = 5 per group. J) µCT reconstruction images and K) quantification of bone microarchitecture parameters in femurs from 3‐month‐old female Npy^fl/fl mice and Dmp1‐iCre; Npy^fl/fl mice subjected to sham or OVX operation with different treatments for two months. Scale bar: 1 mm. n = 10 per group. L) Immunofluorescence staining images for perilipin and M) quantification of adipocyte number in distal femurs. Scale bar: 50 µm. n = 5 per group. N) Quantification of osteoblast number in distal femurs. n = 5 per group. O) ELISA for serum OCN. n = 5 per group. P) Calcein double labeling of trabecular bones and Q) quantification of BFR/BS and MAR. Scale bar: 25 µm. n = 5 per group. Data are presented as mean ± SD. For panel (C): one‐way ANOVA combined with Bonferroni post hoc test. For panel (I): unpaired, two‐tailed student's t‐test (differences between Sham + Vehicle and OVX + Vehicle groups) or one‐way ANOVA combined with Bonferroni post hoc test (differences among other groups except Sham + Vehicle group). For other dot plots: two‐way ANOVA combined with Bonferroni post hoc test. *P < 0.05, **P < 0.01, and ***P < 0.001.

Figure 7

γ‐Oryzanol attenuates ANS dysregulation, NPY overproduction, and bone‐fat imbalance in aging‐ and OVX‐induced osteoporotic mice. A) ELISA for NE and ACh in homogenates of the marrow‐depleted femurs from 15‐month‐old male Npy^fl/fl mice and Dmp1‐iCre; Npy^fl/fl mice with solvent or γ‐Oryzanol treatments for two months. ORZ: γ‐Oryzanol. n = 5 per group. B) ELISA for NPY in homogenates of the marrow‐depleted femurs from 15‐month‐old male Npy^fl/fl mice in different treatment groups. n = 5 per group. C) µCT reconstruction images and D) quantification of bone microarchitecture parameters in femurs. Scale bar: 1 mm. n = 10 per group. E) Immunofluorescence staining images for perilipin in distal femurs and F) quantification of adipocyte number. Scale bar: 50 µm. n = 5 per group. G) Quantification of osteoblast number in distal femurs. n = 5 per group. H) ELISA for NE and ACh in homogenates of the marrow‐depleted femurs from 3‐month‐old female Npy^fl/fl mice and Dmp1‐iCre; Npy^fl/fl mice subjected to sham or OVX operation with solvent or γ‐Oryzanol treatments for two months. n = 6 per group. I) ELISA for NPY in homogenates of the marrow‐depleted femurs from female Npy^fl/fl mice in different treatment groups. n = 6 per group. J) µCT reconstruction images and K) quantification of bone microarchitecture parameters in femurs. Scale bar: 1 mm. n = 10 per group. L) Three‐point bending measurement of femur ultimate load. n = 5 per group. M) Immunofluorescence staining images for perilipin in distal femurs and N) quantification of adipocyte number. Scale bar: 50 µm. n = 5 per group. O) The number of osteoblasts in distal femurs. n = 5 per group. P) ELISA for serum OCN. n = 5 per group. Q) Calcein double labeling of trabecular bones and R) quantification of BFR/BS and MAR. Scale bar: 25 µm. n = 5 per group. Data are presented as mean ± SD. For panel (B) and (I): unpaired, two tailed student's t‐test. For other dot plots: two‐way ANOVA combined with Bonferroni post hoc test. *P < 0.05, **P < 0.01, and ***P < 0.001.

Figure 8

Schematic diagram showing an ANS‐osteocyte NPY‐mediated mechanism by which aging and estrogen deficiency switch BMSC differentiation fate from osteogenesis towards adipogenesis and induce bone loss and marrow adiposity.