Insulin receptor signaling in osteoblasts regulates postnatal bone acquisition and body composition

Abstract

Global energy balance in mammals is controlled by the actions of circulating hormones that coordinate fuel production and utilization in metabolically active tissues. Bone-derived osteocalcin, in its undercarboxylated, hormonal form, regulates fat deposition and is a potent insulin secretagogue. Here, we show that insulin receptor (IR) signaling in osteoblasts controls osteoblast development and osteocalcin expression by suppressing the Runx2 inhibitor Twist2. Mice lacking IR in osteoblasts have low circulating undercarboxylated osteocalcin and reduced bone acquisition due to decreased bone formation and deficient numbers of osteoblasts. With age, these mice develop marked peripheral adiposity and hyperglycemia accompanied by severe glucose intolerance and insulin resistance. The metabolic abnormalities in these mice are improved by infusion of undercarboxylated osteocalcin. These results indicate the existence of a bone-pancreas endocrine loop through which insulin signaling in the osteoblast ensures osteoblast differentiation and stimulates osteocalcin production, which in turn regulates insulin sensitivity and pancreatic insulin secretion.

Figure 1. Insulin receptor signaling promotes osteoblast differentiation

(A) Immunohistochemical staining for insulin receptor (a) and an isotype control (b) in trabecular bone of the distal metaphysis. Filled arrows denote osteoblasts, open arrows denote osteocytes. Muscle was used as positive control (c). (B) Western Blot analysis of the insulin receptor expression in primary osteoblasts isolated from IR^flox/flox mice after infection with adenovirus expressing Cre or GFP as a control. (C) Quantification of osteoblast proliferation by BrdU uptake in control and ΔIR osteoblasts. (D) Quantification of osteoblast apoptosis via Annexin V staining in control and ΔIR osteoblasts treated with 8ng/ml staurosporine. (E–G) Examination of osteoblast differentiation following the deletion of IR. (E) Alkaline phosphatase (ALP) and Alizarin Red (ARS) staining after 7 and 14 days of differentiation, respectively. (F) RT-PCR analysis of IR, Runx2, ATF4, and Osteocalcin (OCN) expression after 7 days of differentiation. (G) ALP and ARS staining after culture in the presence or absence of 13nM IGF-1. (H) Osteocalcin promoter activity assessed in Mc3t3-E1 cells transfected with OC1050-Luc and treated with insulin for 6 hours. (I) ChIP analysis of Runx2 binding to the osteocalcin promoter after 6 hours of 10nM insulin treatment. (J) ChIP analysis of Runx2 binding to the osteocalcin promoter in control and ΔIR osteoblasts. (K) RT-PCR analysis of Osteocalcin (OCN) and Runx2 expression in primary osteoblasts treated with 10nM insulin for 6 hours. (L) Microarray analysis of insulin-regulated genes 4-fold over- or under-expressed and related to bone remodeling in ΔIGF-1R osteoblasts as identified by gene ontology analysis (see also Table S1). (M) Twist1 and Twist2 expression in ΔIGF-1R osteoblasts 24 hours after 10nM insulin treatment. (N) Expression of regulators of Runx2 activity in control and ΔIR osteoblasts after 7 days of differentiation. (O) Co-immunoprecipitation of Runx2 and Twist2 in control and ΔIR osteoblasts.

Figure 2. Insulin receptor signaling is necessary for postnatal bone acquisition

(A) PCR analysis of insulin receptor (IR) allele recombination in tissues for Ob-ΔIR mice. (B–F) Quantitative micro-CT analysis of the distal femur in control and Ob-ΔIR mice at 3-, 6-, and 12-weeks of age (n=4–5mice). (B) Representative images. (C) Bone volume/tissue volume, BV/TV (%). (D) Trabecular number, Tb. N (no./mm), (E) Trabecular thickness, Tb. Th (μm). (F) Trabecular spacing, Tb. Sp (μm). (G–K) Static and Dynamic histomorphometric analysis of the distal femoral metaphysis in control and Ob-ΔIR mice at 3- and 6-weeks of age (n=5–7mice). (G) Osteoblast Number per bone perimeter, Ob. N/BPM (no./100mm) at 3-weeks of age. (H) Osteoblast Number per bone perimeter at 6 weeks of age. (I) Bone formation rate per bone surface, BFR/BS, mm³/cm²/yr. (J) Osteoclast number per bone perimeter, Oc. N/BPM (no./100mm) (K) Erosion Depth, EDE (μm) (L) Serum CTx (ng/ml). (M) RankL and Opg expression in control and ΔIR osteoblasts after 7 days of differentiation.

Figure 3. Ob-ΔIR mice, but not Ob-ΔIGF-1R mice, have increased peripheral adiposity

(A–E) Assessment of body composition in control and Ob-ΔIR mice (n=5–7mice). (A) Body weight. (B) Representative DEXA images at 24-weeks. (C) Fat Mass by qMR. (D) Lean Mass by qMR. (E) Mass of individual fat pads at 12-weeks. (F) Mass of individual fat pads at 12-weeks. (G) Food intake over 4 days at 12-weeks (n=4). (H–I) Indirect calorimetry at 12-weeks (n=4). (H) VO₂ (ml/kg/hr). (I) Respiratory exchange ratio (RER). (J) Energy expenditure (kcal/kg/hr). (K–N) Assessment of body composition in control and Ob-ΔIGF-1R mice at 24-weeks of age (n=5–11mice). (K) Representative DEXA images. (L) Body weight. (M) Fat mass by qMR. (N) Lean mass by qMR.

Figure 4. Ob-ΔIR mice are insulin insensitive and glucose intolerant

Measurements of non-fasted serum glucose (A) and serum insulin (B) in 12- and 24-week old control and Ob-ΔIR mice (n=4–5mice). Glucose tolerance test after fasting overnight at 12-weeks (C) and 24-weeks (E) (n=4–6mice). (E) Insulin tolerance test after fasting for 4 hours at 12-weeks (D) and 24-weeks (F) (n=4–6mice). (G–I) Histomorphometric analysis of pancreatic β-cells in control and Ob-ΔIR mice (n=4–5mice). (G) Representative images of islets stained for insulin. (H) β-cell area. (I) β-cell mass. (J–M) qPCR in tissues isolated from control and Ob-ΔIR mice (n=3). (J) Ins1 and Ins2 expression in pancreas. (K) Pparg expression in white adipose. (L) UCP1 expression in brown adipose. (M) Pck1 and G6pase expression in the liver.

Figure 5. Undercarboxylated osteocalcin improves glucose tolerance and insulin sensitivity in Ob-ΔIR mice

Measurements of total (A) and undercarboxylated osteocalcin (B) in the serum of 24-week old control and Ob-ΔIR mice (n=5mice). (C) Total and (D) undercarboxylated osteocalcin in media conditioned by control and ΔIR osteoblasts. (E) Glucose tolerance test and (F) insulin tolerance test in Ob-ΔIR mice infused with saline or undercarboxylated osteocalcin (30ng/g BW/h) for 2 weeks (n=3). (G) Model for the regulation of bone acquisition and body composition by insulin. Insulin signaling stimulates osteoblast differentiation and osteocalcin expression by relieving Twist2’s suppression of Runx2. Undercarboxylated osteocalcin increases tissue insulin sensitivity and insulin secretion.