Sympathetic Innervation Regulates Osteocyte-Mediated Cortical Bone Resorption during Lactation

Abstract

Bone undergoes constant remodeling by osteoclast bone resorption coupled with osteoblast bone formation at the bone surface. A third major cell type in the bone is osteocytes, which are embedded in the matrix, are well-connected to the lacunar network, and are believed to act as mechanical sensors. Here, it is reported that sympathetic innervation directly regulates lacunar osteocyte-mediated bone resorption inside cortical bone. It is found that sympathetic activity is elevated in different mouse models of bone loss, including lactation, ovariectomy, and glucocorticoid treatment. Further, during lactation elevated sympathetic outflow induces netrin-1 expression by osteocytes to further promote sympathetic nerve sprouting in the cortical bone endosteum in a feed-forward loop. Depletion of tyrosine hydroxylase-positive (TH⁺ ) sympathetic nerves ameliorated osteocyte-mediated perilacunar bone resorption in lactating mice. Moreover, norepinephrine activates β-adrenergic receptor 2 (Adrb2) signaling to promote secretion of extracellular vesicles (EVs) containing bone-degrading enzymes for perilacunar bone resorption and inhibit osteoblast differentiation. Importantly, osteocyte-specific deletion of Adrb2 or treatment with a β-blocker results in lower bone resorption in lactating mice. Together, these findings show that the sympathetic nervous system promotes osteocyte-driven bone loss during lactation, likely as an adaptive response to the increased energy and mineral demands of the nursing mother.

Keywords: cortical bone remodeling; lactation; osteocytes; osteocytic bone resorption; sympathetic nerves.

Figure 1

Sympathetic activity is increased in lactating mice, as well as those treated with glucocorticoid or after OVX. a–c) Schematic diagrams illustrating the three mouse models used in the study: lactation (a), glucocorticoid treatment (methylprednisolone, MPS) (b), and OVX (c). d) Schematic diagram showing the procedure for the isolation of the bone chips which were used in ELISA. e–g) Norepinephrine levels of isolated cortical bone chips in lactation (e), glucocorticoid treatment (f), and OVX (g) mouse models compared to their corresponding control groups. n = 9 in the lactation model, and n = 5 in the glucocorticoid treatment and OVX models. h, i) Tartrate‐resistant acid phosphatase (TRAP) staining (purple) (h) and Ctsk staining (green) (i) in cortical bone (femur mid‐shaft) of virgin and lactating mice. Boxed areas are shown at higher magnification in corresponding panels to the right. Scale bars, 40 (left panels), 20 µm (right panels). j,k) Quantification of the percentage of TRAP⁺ osteocytes (j) and the percentage of Ctsk⁺ osteocytes (k) in total osteocytes. n = 7. l,m) Immunofluorescence staining of femoral sections by osteocalcin (OCN) (red) (l) in virgin and lactating mice, and quantification of average OCN⁺ cell number per mm² of tissue area (m). Scale bars, 200 µm (left panels), 100 µm (right panels). n = 4. n–t) Representative micro‐computed tomography (µCT) images (n) and quantitative analysis of total porosity (Po.tot) (o), close porosity (Po.cl) (p), open porosity (Po.op) (q), tissue mineral density (TMD) (r), cortical thickness (Cor.Th) (s) of femoral cortical bone and trabecular bone fraction (BV/TV) (t) in virgin and lactating mice. Scale bar, 1 mm. n = 9. u–w) Alcian Blue staining (u) and quantification of osteocyte lacuna area (v) and osteocyte lacuna perimeter (w) in cortical bone of virgin and lactating mice. n = 7. Data are presented as mean ± standard error of the mean (SEM). Statistical significance was determined by two‐tailed Student's t‐test.

Figure 1

Sympathetic activity is increased in lactating mice, as well as those treated with glucocorticoid or after OVX. a–c) Schematic diagrams illustrating the three mouse models used in the study: lactation (a), glucocorticoid treatment (methylprednisolone, MPS) (b), and OVX (c). d) Schematic diagram showing the procedure for the isolation of the bone chips which were used in ELISA. e–g) Norepinephrine levels of isolated cortical bone chips in lactation (e), glucocorticoid treatment (f), and OVX (g) mouse models compared to their corresponding control groups. n = 9 in the lactation model, and n = 5 in the glucocorticoid treatment and OVX models. h, i) Tartrate‐resistant acid phosphatase (TRAP) staining (purple) (h) and Ctsk staining (green) (i) in cortical bone (femur mid‐shaft) of virgin and lactating mice. Boxed areas are shown at higher magnification in corresponding panels to the right. Scale bars, 40 (left panels), 20 µm (right panels). j,k) Quantification of the percentage of TRAP⁺ osteocytes (j) and the percentage of Ctsk⁺ osteocytes (k) in total osteocytes. n = 7. l,m) Immunofluorescence staining of femoral sections by osteocalcin (OCN) (red) (l) in virgin and lactating mice, and quantification of average OCN⁺ cell number per mm² of tissue area (m). Scale bars, 200 µm (left panels), 100 µm (right panels). n = 4. n–t) Representative micro‐computed tomography (µCT) images (n) and quantitative analysis of total porosity (Po.tot) (o), close porosity (Po.cl) (p), open porosity (Po.op) (q), tissue mineral density (TMD) (r), cortical thickness (Cor.Th) (s) of femoral cortical bone and trabecular bone fraction (BV/TV) (t) in virgin and lactating mice. Scale bar, 1 mm. n = 9. u–w) Alcian Blue staining (u) and quantification of osteocyte lacuna area (v) and osteocyte lacuna perimeter (w) in cortical bone of virgin and lactating mice. n = 7. Data are presented as mean ± standard error of the mean (SEM). Statistical significance was determined by two‐tailed Student's t‐test.

Figure 2

Lactation‐induced osteocyte perilacunar resorption is ameliorated by a β‐blocker. a–f) Representative micro‐computed tomography (µCT) images (a) and quantitative analysis of open porosity (Po.op) (b), total porosity (Po.tot) (c), tissue mineral density (TMD) (d), and cortical thickness (Cor.Th) (e) of femoral cortical bone and trabecular bone fraction (BV/TV) (f) in different treatment groups. Scale bar, 1 mm. n = 6. g–i) Alcian Blue staining (g) and quantification of osteocyte lacuna area (h) and osteocyte lacuna perimeter (i) of cortical bone in different treatment groups. n = 6. j,k) Representative TRAP staining images of cortical bone (femur mid‐shaft) (j) and quantification of TRAP⁺ osteocytes on femur cortical bone (k) in virgin and lactating mice with different treatments. Boxed areas are shown at a higher magnification in corresponding panels to the right. Scale bar, 40 µm (left panels), 20 µm (right panels). n = 7. l,m) Immunofluorescence staining of Ctsk (green) at mid‐shaft of the femur (l) and quantification of Ctsk⁺ osteocytes in the cortical bone (m) in virgin and lactating mice with different treatments. Boxed areas are shown at a higher magnification in corresponding panels to the right. Scale bar, 40 µm (left panels), 20 µm (right panels). n = 6. Data are presented as mean ± SEM. Statistical significance was determined by two‐way repeated measures analysis of variance (ANOVA) with Tukey post hoc test.

Figure 3

Knockout of Adrb2 in Dmp1⁺ osteocytes attenuates lactation‐induced osteocytic resorption in cortical bone. a) mRNA expression of Adrb2, Adrb1, Adra1, and Adra2 in femoral cortical bone of 3‐month‐old WT mice measured via RT‐qPCR. n = 5. b) Representative images of immunofluorescence staining (left) and quantitative analysis (right) of Adrb2⁺ (green) and Adrb1⁺ (red) osteocytes in cortical bone of 3‐month‐old WT mice. Scale bar, 40 µm. n = 7. c) mRNA expression of Adrb2 in cortical bone of WT virgin mice and lactating mice measured by RT‐qPCR. n = 8. d,e) Immunofluorescence staining (d) and quantification (e) of the percentage of Adrb2 positive osteocytes (green) in cortical bone of Adrb2^flox/flox and Adrb2 ^−/− mice. Scale bar, 40 µm. n = 7. f) mRNA expression of Adrb2 in cortical bone of Adrb2^flox/flox and Adrb2 ^−/− mice after 12 days of lactation measured via RT‐qPCR. n = 7. g–l) Representative micro‐computed tomography (µCT) images (g) and quantitative analysis of total porosity (Po.tot) (h), open porosity (Po.op) (i), tissue mineral density (TMD) (j), and cortical thickness (Cor.Th) (k) of femoral cortical bone and trabecular bone fraction (BV/TV) (l) in Adrb2^flox/flox and Adrb2 ^−/− mice after 12 days of lactation. Scale bar, 1 mm. n = 7. m–o) Alcian Blue staining (m) and quantification of osteocyte lacuna area (n) and osteocyte lacuna perimeter (o) of cortical bone in Adrb2^flox/flox and Adrb2 ^−/− mice after 12 days of lactation. n = 7. Data are presented as mean ± SEM. Statistical significance was determined by one‐way ANOVA with Dunnet post hoc test and two‐tailed Student's t‐test.

Figure 4

Lactation‐induced osteocyte perilacunar resorption occurs through NE‐mediated Adrb2 signaling. a,b) Representative images of TRAP staining at femur mid‐shaft (a) and quantification of TRAP⁺ osteocytes (b) in Adrb2^flox/flox and Adrb2 ^−/− mice after 12 days of lactation. Boxed areas are shown at a higher magnification in corresponding panels to the right. Scale bar, 40 µm (left panels), 20 µm (right panels). n = 7. c,d) Representative images of immunofluorescence staining of Ctsk (green) at femur mid‐shaft (c) and quantification of Ctsk⁺ osteocytes (d) in Adrb2^flox/flox and Adrb2 ^−/− mice after 12 days of lactation. Boxed areas are shown at a higher magnification in corresponding panels to the right. Scale bar, 40 µm (left panels), 20 µm (right panels). n = 7. e,f) Immunofluorescence staining of pCREB (green) in primary osteocytes treated with different concentrations of NE, NE+ICI‐118551, and control group (f), and quantification of the percentage of pCREB⁺ osteocytes (e). DAPI stains nuclear blue. Scale bar, 20 µm. n = 5. g) Representative immunoblots of pCREB and CREB in primary osteocytes treated with different concentrations of NE, NE+ICI‐118551, and control group. n = 5. h) Bioinformatics analysis of scRNA‐seq of isolated cortical bone cells from virgin and lactating mice with clustering of osteoblasts in (Figure S2d, Supporting Information). i) Dot plot of osteoblast marker genes in osteoblast populations (cluster 11) of virgin and lactating mice. Data are presented as mean ± SEM. Statistical significance was determined by unpaired, two‐tailed Student's t‐test (b, d) and one‐way ANOVA with Dunnet post hoc test (f).

Figure 5

NE stimulates the release of EVs by osteocytes. a) Schematic diagram illustrating the procedure for isolating EVs from primary osteocytes. b) Examination of the purity of the isolated primary osteocytes obtained in (a) by flow cytometry. c,d) Representative images showing EVs visualized by transmission electron microscopy (TEM) (c) and quantification of percentage of different sized EVs (d). e,f) Immunoblots of Alix, Tsg101, and Ctsk from protein lysates of the isolated EVs (e) and quantification of expression levels after normalized to GAPDH (f). g) Immunofluorescence staining of Tsg101 (green) in primary osteocytes after NE stimulation with different concentrations. DAPI stains nuclear blue. Scale bar, 20 µm. h) Calcium concentration over time in primary osteocytes after NE stimulation. i,j) Immunoblots for Alix, Tsg101, and Ctsk in protein lysates from EVs after stimulation with 100 µm of NE with/without addition of 10 mm of neomycin (i), and quantification of Alix, Tsg101, and Ctsk expression levels (j) after being normalized to GAPDH. Data are presented as mean ± SEM. Statistical significance was determined by one‐way ANOVA with Dunnet post hoc test (f, j).

Figure 5

NE stimulates the release of EVs by osteocytes. a) Schematic diagram illustrating the procedure for isolating EVs from primary osteocytes. b) Examination of the purity of the isolated primary osteocytes obtained in (a) by flow cytometry. c,d) Representative images showing EVs visualized by transmission electron microscopy (TEM) (c) and quantification of percentage of different sized EVs (d). e,f) Immunoblots of Alix, Tsg101, and Ctsk from protein lysates of the isolated EVs (e) and quantification of expression levels after normalized to GAPDH (f). g) Immunofluorescence staining of Tsg101 (green) in primary osteocytes after NE stimulation with different concentrations. DAPI stains nuclear blue. Scale bar, 20 µm. h) Calcium concentration over time in primary osteocytes after NE stimulation. i,j) Immunoblots for Alix, Tsg101, and Ctsk in protein lysates from EVs after stimulation with 100 µm of NE with/without addition of 10 mm of neomycin (i), and quantification of Alix, Tsg101, and Ctsk expression levels (j) after being normalized to GAPDH. Data are presented as mean ± SEM. Statistical significance was determined by one‐way ANOVA with Dunnet post hoc test (f, j).

Figure 6

Adrb2 signaling promotes the netrin‐1⁺ osteocyte number in the endosteum and increases sympathetic tone during lactation. a) NE levels in mice's cortical bone chips at different time points during lactation as measured by ELISA. b) Schematic diagram illustrating the anatomy of cortical bone. c–e) Distribution of TRAP⁺ osteocytes in middle‐shaft of the femur in virgin and lactating mice treated with vehicle (upper panel) or ICI‐18551 (bottom panel) (c) and quantification of the percentage of TRAP⁺ osteocytes (d) and their distance to the endosteum (e). n = 5. f–h) Distribution of TRAP⁺ osteocytes in middle‐shaft of the femur in Adrb2^flox/flox and Adrb2 ^−/− mice after 12 days of lactating (f) and quantification of the percentage of TRAP⁺ osteocytes (g) and their distance to the endosteum (h). n = 7. i,j) Immunofluorescence staining of tyrosine hydroxylase (TH, red) (i) and quantification of the density of TH⁺ nerves (red) in the endosteum area (j) in femoral bone marrow. E, endosteum; BM, bone marrow. n = 6. k–m) Double‐immunofluorescence images of Ctsk (red) and Netrin‐1 (green) in femoral bone sections from virgin and lactating mice treated with vehicle or ICI‐18551 (k) and quantification of the percentage of Netrin‐1⁺ osteocytes (l) and Netrin‐1⁺/ Ctsk⁺ osteocytes in the femur mid‐shaft (m), respectively. Scale bar, 20 µm. n–p) Double‐immunofluorescence images of Ctsk (red) and netrin‐1 (green) in femoral bone sections from virgin and lactating Adrb2^flox/flox/ Adrb2^{−/ −} mice (n) and quantification of the percentage of Netrin‐1⁺ osteocytes (o) and Netrin‐1⁺/ Ctsk⁺ osteocytes in the femur mid‐shaft (p), respectively. Scale bar, 20 µm. n = 7. q) mRNA expression of Ntn1 by RT‐qPCR in tibial cortical bone of Adrb2^flox/flox and Adrb2^−/− virgin mice. n = 5. Data are presented as mean ± SEM. Statistical significance was determined by one‐way ANOVA with Dunnet post hoc test (a, j), two‐way repeated measures ANOVA with Tukey post hoc test (l, m, o, p), and two‐tailed Student's t‐test (q).

Figure 6

Adrb2 signaling promotes the netrin‐1⁺ osteocyte number in the endosteum and increases sympathetic tone during lactation. a) NE levels in mice's cortical bone chips at different time points during lactation as measured by ELISA. b) Schematic diagram illustrating the anatomy of cortical bone. c–e) Distribution of TRAP⁺ osteocytes in middle‐shaft of the femur in virgin and lactating mice treated with vehicle (upper panel) or ICI‐18551 (bottom panel) (c) and quantification of the percentage of TRAP⁺ osteocytes (d) and their distance to the endosteum (e). n = 5. f–h) Distribution of TRAP⁺ osteocytes in middle‐shaft of the femur in Adrb2^flox/flox and Adrb2 ^−/− mice after 12 days of lactating (f) and quantification of the percentage of TRAP⁺ osteocytes (g) and their distance to the endosteum (h). n = 7. i,j) Immunofluorescence staining of tyrosine hydroxylase (TH, red) (i) and quantification of the density of TH⁺ nerves (red) in the endosteum area (j) in femoral bone marrow. E, endosteum; BM, bone marrow. n = 6. k–m) Double‐immunofluorescence images of Ctsk (red) and Netrin‐1 (green) in femoral bone sections from virgin and lactating mice treated with vehicle or ICI‐18551 (k) and quantification of the percentage of Netrin‐1⁺ osteocytes (l) and Netrin‐1⁺/ Ctsk⁺ osteocytes in the femur mid‐shaft (m), respectively. Scale bar, 20 µm. n–p) Double‐immunofluorescence images of Ctsk (red) and netrin‐1 (green) in femoral bone sections from virgin and lactating Adrb2^flox/flox/ Adrb2^{−/ −} mice (n) and quantification of the percentage of Netrin‐1⁺ osteocytes (o) and Netrin‐1⁺/ Ctsk⁺ osteocytes in the femur mid‐shaft (p), respectively. Scale bar, 20 µm. n = 7. q) mRNA expression of Ntn1 by RT‐qPCR in tibial cortical bone of Adrb2^flox/flox and Adrb2^−/− virgin mice. n = 5. Data are presented as mean ± SEM. Statistical significance was determined by one‐way ANOVA with Dunnet post hoc test (a, j), two‐way repeated measures ANOVA with Tukey post hoc test (l, m, o, p), and two‐tailed Student's t‐test (q).

Figure 7

TH⁺ sympathetic nerve fibers are essential for osteocyte lacunar bone resorption. a) Schematic diagram illustrating the procedure of intra‐femoral injection of vehicle, F127/vehicle, F127/6‐OHDA, and F127/guanethidine in mice after 3 days of lactation. b) NE levels in tibial cortical bone chips determined by ELISA in virgin and lactating mice with different treatments. n = 5. c) Immunofluorescence staining of TH (red) in the endosteum area in femoral bone marrow. E, endosteum; BM, bone marrow. n = 5. d–h) Representative micro‐computed tomography (µCT) images (d) and quantitative analysis of open porosity (Po.op) (e), total porosity (Po.tot) (f), cortical thickness (Cor.Th) (g) of femoral cortical bone and trabecular bone fraction (BV/TV) (h). Scale bar, 1 mm. n = 5. i–o) Representative images for tartrate‐resistant acid phosphatase (TRAP) staining (i), immunofluorescence staining of Ctsk (green) (k) and Alcian blue staining (m) on femoral cortical bone of virgin and lactating mice with different treatments, quantification of the percentage of TRAP⁺ osteocytes (j) and Ctsk⁺ osteocytes (l), and calculation of osteocyte lacuna area (n) and osteocyte lacuna perimeter (o). Scale bar, 40 µm. n = 5. Data are presented as mean ± SEM. Statistical significance was determined by one‐way repeated measures ANOVA with Dunnet post hoc test.

Figure 7

TH⁺ sympathetic nerve fibers are essential for osteocyte lacunar bone resorption. a) Schematic diagram illustrating the procedure of intra‐femoral injection of vehicle, F127/vehicle, F127/6‐OHDA, and F127/guanethidine in mice after 3 days of lactation. b) NE levels in tibial cortical bone chips determined by ELISA in virgin and lactating mice with different treatments. n = 5. c) Immunofluorescence staining of TH (red) in the endosteum area in femoral bone marrow. E, endosteum; BM, bone marrow. n = 5. d–h) Representative micro‐computed tomography (µCT) images (d) and quantitative analysis of open porosity (Po.op) (e), total porosity (Po.tot) (f), cortical thickness (Cor.Th) (g) of femoral cortical bone and trabecular bone fraction (BV/TV) (h). Scale bar, 1 mm. n = 5. i–o) Representative images for tartrate‐resistant acid phosphatase (TRAP) staining (i), immunofluorescence staining of Ctsk (green) (k) and Alcian blue staining (m) on femoral cortical bone of virgin and lactating mice with different treatments, quantification of the percentage of TRAP⁺ osteocytes (j) and Ctsk⁺ osteocytes (l), and calculation of osteocyte lacuna area (n) and osteocyte lacuna perimeter (o). Scale bar, 40 µm. n = 5. Data are presented as mean ± SEM. Statistical significance was determined by one‐way repeated measures ANOVA with Dunnet post hoc test.

Figure 8

Summary diagram of SNS‐mediated regulation of lactation‐related osteocyte perilacunar resorption. During lactation, NE levels increase in the cortical bone. In response, osteocytes increase Ca²⁺ signaling to promote the secretion of bone resorptive EVs and Ctsk to degrade the perilacunar area bone. In parallel, NE signaling drives the phosphorylation of CREB to promote Ctsk and netrin‐1 expression, which enhances SNS tone signal, thus leading to an increased innervation of TH⁺ sympathetic nerves in the endosteum in a feedback loop manner.