Bone controls browning of white adipose tissue and protects from diet-induced obesity through Schnurri-3-regulated SLIT2 secretion

Abstract

The skeleton has been suggested to function as an endocrine organ controlling whole organism energy balance, however the mediators of this effect and their molecular links remain unclear. Here, utilizing Schnurri-3^-/- (Shn3^-/-) mice with augmented osteoblast activity, we show Shn3^-/-mice display resistance against diet-induced obesity and enhanced white adipose tissue (WAT) browning. Conditional deletion of Shn3 in osteoblasts but not adipocytes recapitulates lean phenotype of Shn3^-/-mice, indicating this phenotype is driven by skeleton. We further demonstrate osteoblasts lacking Shn3 can secrete cytokines to promote WAT browning. Among them, we identify a C-terminal fragment of SLIT2 (SLIT2-C), primarily secreted by osteoblasts, as a Shn3-regulated osteokine that mediates WAT browning. Lastly, AAV-mediated Shn3 silencing phenocopies the lean phenotype and augmented glucose metabolism. Altogether, our findings establish a novel bone-fat signaling axis via SHN3 regulated SLIT2-C production in osteoblasts, offering a potential therapeutic target to address both osteoporosis and metabolic syndrome.

Fig. 1. Shn3-deficiency protects mice from diet-induced-obesity.

All samples were harvested from C57BL6/J background male WT (n = 7) and Shn3^−/−mice (n = 6) after HFD feeding for 12 weeks. a Body weight curve obtained during 12 weeks HFD feeding starting from 6-weeks of age (P values directly labeled above the time points). b Representative general view of liver, BAT, iWAT and eWAT of WT and Shn3^−/− mice. c Representative histological sections of liver, BAT, iWAT and eWAT with H&E staining from 3 individual repeated experiments. Scale bar=200um. d Tissue wet weight of iWAT, eWAT, BAT and liver. e Body composition quantified through Echo-MRI. f Results of intraperitoneal glucose tolerance test, insulin tolerance test and quantification of area under curve (AUC) (n = 6 for each group). g Representative histological sections of pancreas islets with H&E staining from 3 individual slides. Scale bar=100um. Results were presented as mean ± s.d. and analyzed by two-way ANOVA with Bonferroni’s multiple comparisons test (a, f) or unpaired, two-tailed Student’s t test (d–e). *P < 0.05, **P < 0.005, ***P < 0.0005. Source data are provided as a Source Data file.

Fig. 2. Increased energy expenditure in Shn3-deficient mice.

Individuals measured by indirect calorimetry were WT (n = 7) and Shn3^−/− mice (n = 6) after 12-week-HFD exposure. a Calculated energy expenditure of WT and Shn3^−/−mice under room temperature within a 120-hour-period. b Bar graph of overall, light cycle and dark cycle energy expenditure of WT and Shn3^−/−mice under room temperature. c Regression-based analysis of energy expenditure against total body mass. d Gross view of iWAT from WT and Shn3^−/−mice after 12-week-HFD exposure, black arrow pointing to inguinal lymph nodes. e Representative H&E staining of iWAT sections obtained from WT and Shn3^−/− mice after 12-week-HFD exposure. Scale bar=200um. Results were selected from 3 repeated experiments. f Thermogenic gene expression level relative to Gapdh of iWAT quantified by qPCR in WT and Shn3^−/− mice. g Average adipocyte surface area of WT and Shn3^−/− mice. Results were presented as mean ± s.d. and analyzed by unpaired, two-tailed Student’s t test (b and f) or two-way ANOVA with Bonferroni’s multiple comparisons test (a and g). Source data are provided as a Source Data file.

Fig. 3. Shn3 depletion in early-stage osteoblasts but not adipose tissue produces anti-obesity effects.

a Body weight curve of male Shn3^OSX and Shn3^f/f mice after HFD for 12 weeks (n = 10 for each group, P values directly labeled above the time points). b Body weight curve of Shn3^Adipoq and Shn3^f/f mice (n = 10) after HFD for 12 weeks. c Tissue wet weight of iWAT, eWAT, BAT and liver (Shn3^OSX and Shn3^f/f, upper, n = 10 for Shn3^f/f group and n = 8 for Shn3^OSX group. Shn3^Adipoq and Shn3^f/f, lower, n = 8 for Shn3^f/f group and n = 7 for Shn3^Adipoq group.). d Representative histological sections of iWAT, eWAT and liver from 3 slides of Shn3^OSX and Shn3^f/f mice, respectively. Scale bar=200 um. e–f Results of intraperitoneal glucose tolerance and insulin tolerance tests (P values directly labeled above the time points) and the area under the curve for Shn3^Osx and Shn3^f/f mice is shown (n = 6 for each group). g Serum C-peptide level measured by sandwich ELISA of Shn3^Osx (n = 8) and Shn3^f/f (n = 10) individuals. h Representative H&E and insulin IHC of pancreas islet from 3 individual slides. Scale bar=200 um. i–j Histomophometric quantification result of islet area/tissue area and islet number (n = 5 for each group). (k) Quantification of serum triacylglycerol (TG, n = 7 for Shn3^Osx group and n = 8 for Shn3^f/f group), free fatty acid (FFA, n = 8 for each group) level and adiponectin (n = 8 for each group) of Shn3^OSX and Shn3^f/f mice through ELISA assays. l Representative H&E sections (Scale bar=500 um) and Ucp-1 IHC results (Scale bar=200 um) of iWAT from 3 individual slides of Shn3^Osx and Shn3^f/f individuals, respectively. m Representative immunoblots of UCP1 of iWAT obtained from HFD-fed Shn3^Osx and Shn3^f/f mice. n Quantification of Ucp-1 expression level in iWAT from HFD-fed Shn3^Osx (n = 7) and Shn3^f/f (n = 8) mice. o Thermogenic gene expression level relative to Gapdh in iWAT of Shn3^Osx (n = 10) and Shn3^f/f mice (n = 7). p Lipolytic gene expression in iWAT harvested from HFD-fed Shn3^Osx and Shn3^f/f mice (n = 5 for each group). Results were presented as mean ± s.d., analyzed by unpaired, two-tailed Student’s t test (lower of c, g, i–k, n–p), one-way ANOVA with Tukey’s test (upper of c), two-way ANOVA with Bonferroni’s multiple comparisons test (a, b, e, f). *P < 0.05, **P < 0.005, ***P < 0.0005, ****P < 0.0001 (Shn3^f/f vs. Shn3^Osx). ^#P < 0.05, ^##P < 0.005, ^###P < 0.0005 (Osx Cre vs. Shn3^Osx). Source data are provided as a Source Data file.

Fig. 4. Skeletal Shn3-ablation augments adipose tissue browning in an endocrine manner.

a Schematic diagram of the co-culture assay. b Thermogenic gene expression in co-cultured differentiated beige adipocytes (n = 4 for each group). Genotypes of osteoblasts are indicated. Results were collected from 3 repeated experiments. c Schematic diagram of transplantation assay. d Graft weight of fat graft transplanted into Shn3^Osx and Shn3^f/f hosts. e Representative H&E sections and UCP-1 IHC results of fat allografts implanted into Shn3^Osx and Shn3^f/f hosts for 4 weeks. respectively. Results were collected from 3 individual slides. f Ucp1 gene expression level relative to Gapdh of fat graft transplanted into Shn3^Osx and Shn3^f/f hosts (n = 4 for each group). Results were presented as mean ± s.d., analyzed by unpaired, two-tailed Student’s t test (b, d, f). Diagram (a) and (c) were generated through utilizing BioRender Apps. Source data are provided as a Source Data file.

Fig. 5. Slit2 is an osteokine secreted by osteoblasts that protects mice from DIO.

a Heatmap showing relative RNA levels of several cytokines in primary osteoblasts isolated from WT and Shn3^−/− mice (n = 3). b Immunoblotting for SLIT2 in osteoblast lysates, serum, osteoblast supernatant. The red and bold blue boxes indicate regions that were cut for further LC-MS studies. Results were selected from 3 repeated experiments. c Matched peptides of SLIT2 and the related amino acid sequence detected in murine sera. d ELISA measurement of SLIT2 in conditioned media from WT and Shn3^−/− osteoblasts (n = 4 for each group). (e) Immunoblotting demonstrating the levels of SLIT2-C in the serum of HFD-fed Shn3^−/− and WT mice and (f) ELISA measurement of circulating SLIT2 levels in Shn3^−/− and WT mice after 12-weeks of HFD exposure (n = 9 for WT group and n = 11 for Shn3^−/−group). g Immunoblotting for SLIT2-C in the serum of HFD-exposed Slit2^Osx and Slit2^f/f mice. h ELISA results of circulating SLIT2 levels in Slit2^f/f and Slit2^Osx mice (n = 6 for each group). i Body weight curve of HFD-exposed Slit2^Osx and Slit2^f/f mice (n = 6 for each group). j Representative gross images of iWAT obtained from Slit2^Osx and Slit2^f/f mice. k tissue wet weight of iWAT, eWAT, liver and BAT obtained from Slit2^Osx and Slit2^f/f mice (n = 6 for each group). l Representative histological sections of iWAT, eWAT and liver from 3 individual slides of Slit2^Osx and Slit2^f/f mice after a 10-week-HFD challenge. Scale bar=100 um. m Thermogenic gene expression level relative to Gapdh in iWAT of HFD-fed Slit2^Osx and Slit2^f/f mice (n = 4 for each group). Results were presented as mean ± s.d., analyzed by unpaired, two-tailed Student’s t test (e, f, h, k and m) or two-way ANOVA with Bonferroni’s multiple comparisons test (i). *P < 0.05. Source data are provided as a Source Data file.

Fig. 6. AAV-9 mediated SHN3 depletion protected mice from development of DIO.

Representative IVIS imaging of (a) whole body and (b) major organs from murine models after the injection of a single dose of 4 × 10¹¹ GC rAAV9-amiR-Shn3 or rAAV9-amiR-Ctrl (Fluorescence signal intensity labeled as colored symbols on the right. Yellow represents strong signal and red represents dim signal). μCT reconstruction and quantification of (c) femoral trabecular and (d) cortical bone of rAAV9-amiR-Shn3 or rAAV9-amiR-Ctrl administered mice (n = 5 for each group). Scale bar=1 mm. e Body weight curve of rAAV9-amiR-Shn3 and rAAV9-amiR-Ctrl injected mice over a 12-week HFD exposure (n = 8 for each group). f General view of inguinal adipose fat pad and intrascapular brown adipose tissue. g Tissue wet weight of iWAT, eWAT, liver and BAT obtained from rAAV9-amiR-Shn3 and rAAV9-amiR-Ctrl injected mice (n = 8 for each group). h Representative histological H&E results of iWAT and liver from rAAV9-amiR-Shn3 and rAAV9-amiR-Ctrl group. Scale bar=100 um. Results were selected from 3 repeated experiments. i Thermogenic gene expression level relative to Gapdh in iWAT of mice treated with rAAV9-amiR-Shn3 or rAAV9-amiR-Ctrl (n = 6 for each group). j IGTT and ITT results at 10 weeks after rAAV injection (n = 8 for each group). Results were presented as mean ± s.d., analyzed by unpaired, two-tailed Student’s t test (c–d, g, i) or two-way ANOVA with Bonferroni’s multiple comparisons test (e, j). *P < 0.05, **P < 0.005, ***P < 0.0005, ****P < 0.0001. Source data are provided as a Source Data file.