Negative Skeletal Effects of Locally Produced Adiponectin

Abstract

Epidemiological studies show that high circulating levels of adiponectin are associated with low bone mineral density. The effect of adiponectin on skeletal homeostasis, on osteoblasts in particular, remains controversial. We investigated this issue using mice with adipocyte-specific over-expression of adiponectin (AdTg). MicroCT and histomorphometric analysis revealed decreases (15%) in fractional bone volume in AdTg mice at the proximal tibia with no changes at the distal femur. Cortical bone thickness at mid-shafts of the tibia and at the tibiofibular junction was reduced (3-4%) in AdTg mice. Dynamic histomorphometry at the proximal tibia in AdTg mice revealed inhibition of bone formation. AdTg mice had increased numbers of adipocytes in close proximity to trabecular bone in the tibia, associated with increased adiponectin levels in tibial marrow. Treatment of BMSCs with adiponectin after initiation of osteoblastic differentiation resulted in reduced mineralized colony formation and reduced expression of mRNA of osteoblastic genes, osterix (70%), Runx2 (52%), alkaline phosphatase (72%), Col1 (74%), and osteocalcin (81%). Adiponectin treatment of differentiating osteoblasts increased expression of the osteoblast genes PPARγ (32%) and C/ebpα (55%) and increased adipocyte colony formation. These data suggest a model in which locally produced adiponectin plays a negative role in regulating skeletal homeostasis through inhibition of bone formation and by promoting an adipogenic phenotype.

Fig 1. Cancellous bone measurements.

Bone tissue volume refers to the entire volume of bone scanned (TV), bone volume refers to the volume of mineralized bone within the scanned region (BV) and fractional bone volume referring to the fraction of mineralized bone relative to the total bone volume (BV/TV) was assessed by A) μCT and B) three-dimensional reconstruction μCT renderings at the distal femur and C) TV, BV, and BV/TV at the proximal tibia D) three-dimensional reconstruction images at the proximal tibia D). E) Trabecular thickness (Tb.Th), spacing (Tb.Sp.), and number (Tb.N) was assessed at the proximal tibia by μCT analysis. F) Histomorphometric analysis of distal femur and proximal tibia using Bioquant Software. G) Expression level of osbteoblast marker genes: Osterix, Runx2, Alkaline Phosphatase (Al. Phosph.), Osteocalcin (Osteocal.) and Collagen Type I (Col 1). All expression data were obtained by RT-qPCR analysis of RNA and calculated based on the ddCt method with GAPDH as the reference gene and WT as the control. All data are shown as mean ± SEM from 12 week old female mice (n = 5–8). Statistical significance * P<0.05, ** P<0.01, WT compared with AdTg bones.

Fig 2. Cortical bone parameters of the midshaft of femurs and tibia and histomorphometry.

A) μCT and B) three-dimensional reconstruction μCT renderings. C) Cortical thickness assessed at the Tibiofibular junction (TFJ) and D) three-dimensional reconstruction μCT renderings of the (TFJ). E) Mineral apposition rates and F) bone formation rates measured by dynamic histomorphometry at the proximal tibia. All data are shown as mean ± SEM from 12 week old female mice (n = 5–8). Statistical significance * P<0.05, WT compared with AdTg bones.

Fig 3. Effects of adiponectin on adipogenesis.

Hemotoxylin and Eosin (H & E) staining of A) proximal tibia and B) number of adipocytes measured to be ≤ 10 μm from trabeculi. C) H & E staining of distal femur. D) Adiponectin concentration measured in the marrow from tibia and femurs. E) Expression level of the adipocyte marker genes: peroxisome proliferator-activated receptor gamma (PPARγ) and CCAAT/enhancer-binding protein alpha (C/ebpα). Assessment of adipogenesis F) by Oil Red O staining, from BMSCs isolated from the tibia of WT and AdTg mice. All expression data were obtained by RT-qPCR analysis of RNA and calculated based on the ddCt method with GAPDH as the reference gene and WT as the control. All data are shown as mean ± SEM from 12 week old female mice (n = 3–6). Statistical significance * P<0.05, WT compared with AdTg bones.

Fig 4. Adiponectin regulates osteoblastogenesis in bone marrow stromal cells.

A) Effect of adiponectin treatment (2.5 μg/ml) from day 0–5 on osteoblast differentiation from BMSCs as assessed by Von Kossa (VK) and alkaline phosphatase staining. B) mRNA expression levels of the adiponectin receptor 1 (AdipoR1) throughout 28 days of osteoblastic differentiation. C) Effect of adiponectin treatment (2.5, 5, and 10 μg/ml) from day 14–28 on osteoblast differentiation from BMSCs as assessed by VK and alkaline phosphatase staining and on D) expression levels of osteoblast markers: osterix, Runx2, alkaline phosphatase (Al. Phosph.), osteocalcin (Osteocal.) and collagen type I (Col 1). E) Expression levels of RANKL:OPG. Colony numbers were calculated using Image J Software. All expression data were obtained by RT-qPCR analysis of RNA and GAPDH as the relative control. All data are shown as mean ± SEM from BMSCs isolated from 12 week old WT mice (n = 3–7). Statistical significance * P<0.05.

Fig 5. Effects of Adiponectin on adipogenesis.

A) Effect of adiponectin treatment (2.5 μg/ml) from day 0–5 on adipogenic differentiation from BMSCs as assessed by Oil Red O staining. Effect of adiponectin treatment (2.5 μg/ml) from day 14–28 on adipogenesis as assessed by B) quantification and C) visualization of Oil Red O staining of adipocyte-like cell colonies and D) Nile Red staining with or without 3 mM oleate. Red staining represents lipid droplets and blue staining represents DAPI stained nuclei. E) Expression level of the adipocyte marker genes Pparg and C/ebp alpha. F) Co-localization studies of BMSCs isolated from Col-1 2.3 promoter driven GFP reporter mice treated with 2.5 μg/ml of adiponectin in adipogenic media from day 14–21 following osteoblastic differentiation for 7 days. Green staining represents osteoblast lineage cells, red staining indicates adipogenic marker FABP4, and blue staining indicates DAPI stained nuclei. Cells were visualized with a spinning disk confocal microscope. Colony numbers were calculated using Image J Software. All expression data were obtained by RT-qPCR analysis of RNA and GAPDH as the relative control. All data are shown as mean ± SEM from BMSCs isolated from 12 week old WT mice (n = 3–7). Statistical significance * P<0.05.