Bone Lining Cells Could Be Sources of Bone Marrow Adipocytes

Abstract

Background: Recently, lineage-tracing studies demonstrated that parathyroid hormone and anti-sclerostin antibody (Scl-Ab) can convert bone lining cells (BLCs) into active osteoblasts. However, BLCs might also be differentiated into other lineages. Here we investigated whether BLCs could differentiate into bone marrow adipocytes (BMAds) and whether Scl-Ab could suppress this process.

Methods: Dmp1-CreERt2:mTmG mice were injected with 0.5 mg of 4-hydroxytamoxifen once weekly from postnatal week 4 to week 8. The mice were treated with either vehicle or rosiglitazone for 8 weeks (weeks 12-20). Moreover, they were administered either vehicle or Scl-Ab (50 mg/kg) twice weekly for 4 weeks (weeks 16-20, N = 4-6/group). We chased the GFP+ cells from the endosteal surface to the bone marrow (BM) of the femur. Using immunohistochemical staining, the numbers of perilipin+ or GFP+/perilipin double+ cells in the BM were quantified. In addition, serum N-terminal propeptide of type I procollagen (P1NP) levels were measured at each time point, and bone mass was analyzed at 20 weeks using micro-computed tomography.

Results: Scl-Ab administration significantly reversed the decreases in bone parameters induced by rosiglitazone. Plump GFP+ cells, presumably active osteoblasts, and extremely flat GFP+ cells, presumably BLCs, were present on the endosteal surface of the femur at 8 and 12 weeks, respectively, in line with prior findings. When we chased the GFP+ cells, rosiglitazone significantly increased the number of GFP/perilipin double+ BMAds compared to the effects of the vehicle (P < 0.001), and overlapping Scl-Ab administration decreased the number of GFP/perilipin double + BMAd compared to rosiglitazone alone (P < 0.001). In addition, we found that osteoblast lineage cells such as BLCs might express PPARγ on immunohistochemical staining. When rosiglitazone was administered to Rip-Cre:mTmG mice, GFP+ cells were not present on the endosteal surface or in the BM of the femur; however, they were present in the pancreas.

Conclusion: BLCs could be sources of BMAds, and rosiglitazone could stimulate the differentiation of osteoblast lineage cells into BMAds. Suppression of the differentiation of osteoblast lineage cells into BMAds might contribute to anabolic effects resulting from the pharmacologic inhibition of sclerostin.

Keywords: anti-sclerostin antibody; bone lining cell; bone marrow adipocyte; osteoblast; rosiglitazone.

Figure 1

Experimental design. Dmp1-CreERt2:mTmG mice were injected with 0.5 mg 4-OHTam on postnatal weeks 4, 5, 6, 7, and 8. The mice were treated with either vehicle or rosiglitazone for 8 weeks (weeks 12–20). Moreover, they were administered either vehicle or Scl-Ab (50 mg/kg) twice weekly for 4 weeks (weeks 16–20, N = 4–6/group). We chased the GFP+ cells from the endosteal surface to the bone marrow of the femur. Animals were euthanized on postnatal week 8 or 12 (2 days or 4 weeks after the last 4-OHTam treatment), as well as on week 16 and 20 to evaluate the impact of rosiglitazone or Scl-Ab administration.

Figure 2

Osteoblast lineage cells are labeled in 10-kb Dmp1- CreERt2:mTmG mice Representative confocal microscopy images of endocortical surface of the femur from Dmp1- CreERt2:mTmG mice at the indicated experimental groups(the left panel at 16 weeks shows low power field). BM, bone marrow; C, cortical bone; white arrow, bone lining cell; yellow arrow, active osteoblast; asterisk, osteocyte; red, mTomato; green, mGFP; blue, DAPI staining of nuclei. Red and green protein fluorescence were captured directly. Scale bar: 50 µm.

Figure 3

10-kb Dmp1- CreERt2 targets osteoblast lineage cells and skeletal muscle, but not bone marrow cells in 2-week-old mice. (A–C) Confocal microscopy images of direct fluorescence on longitudinal sections of the tibia at the indicated experimental groups. (D–F) Images at a higher magnification for different areas of bone as indicated. C, cortical bone; GP, growth plate; M, skeletal muscle; yellow arrow, osteoblast; white arrow, muscle; red, mTomato; green, mGFP; blue, DAPI staining of nuclei. Red and green protein fluorescence were captured directly.

Figure 4

Scl-Ab rescues the negative effects of rosiglitazone on bone parameters. (A–E) MicroCT quantification of trabecular bone. Each data represents the mean, and error bars represent the standard error (*, P < 0.05 vs. rosiglitazone; **, P < 0.01 vs. rosiglitazone, N = 6/group). BV, bone volume; TV, total volume; BS, bone surface, Tb.N, trabeculae number; Tb.Sp, trabeculae separation; Tb.Th, trabeculae thickness. (F) Serum P1NP measurements from mice in the indicated groups. Each data represents the mean, and error bars represent the standard error (**, P < 0.01 vs. 8 week; a, P < 0.05 vs. vehicle; b, P < 0.05 vs. rosiglitazone, N = 6–8/group). Vehicle, white; rosiglitazone, gray; Scl-Ab, dark gray; Scl-Ab/rosiglitazone, gray pattern.

Figure 5

Rosiglitazone increases GFP/perilipin double positive bone marrow adipocytes, whereas Scl-Ab suppresses the effect of rosiglitazone on them. (A) Representative immunostaining against GFP and perilipin on sections of femoral bone marrow from Dmp1- CreERt2:mTmG mice in the indicated experimental groups. Scale bar: 50 µm. Red, perilipin+ bone marrow adipocytes (BMAds); yellow arrow, GFP/perilipin double positive BMAds; blue, DAPI staining of nuclei. (B) Quantification of total perilipin+ or GFP/perilipin double+ BMAds. a, P < 0.001 vs. vehicle or Scl-Ab; b, P < 0.001 vs. rosiglitazone + Scl-Ab; N=4–6/group. (C) Quantification of perilipin+ BMAd size in the indicated experimental groups (N = 4–6/group). (D) Representative confocal images showing direct fluorescence in cells on endocortical surfaces of the femur from the indicated experimental groups. BM, bone marrow; C, cortical bone; white arrow, bone lining cell; yellow arrow, active osteoblast; asterisk, osteocyte; red, mTomato; green, mGFP; blue, DAPI staining of nuclei. Red and green protein fluorescence were captured directly. Scale bar: 50 µm.

Figure 6

Rosiglitazone does not increase endogenous GFP expression. (A) Representative immunostaining against GFP and perilipin on sections of femoral bone marrow from Rip-Cre:mTmG mice that had received rosiglitazone for 8 weeks. Red, perilipin+ bone marrow adipocytes; blue, DAPI staining of nuclei. (B, C) Representative confocal images of the pancreas from Rip-Cre:mTmG mice or Rip-Cre(-):mTmG mice that had received rosiglitazone for 8 weeks. Red, mTomato; green, mGFP; blue, DAPI staining of nuclei. Rip-Cre transgene, a 668 bp fragment of the rat insulin II promoter. Red and green protein fluorescence were captured directly. Scale bar: 100 µm.

Figure 7

PPARγ expression in osteoblasts and bone lining cells. Representative images for PPARγ immunohistochemical staining of osteoblast lineage cells (arrow) and osteocytes (asterisk) in the femur at 8 and 12 weeks. Immunostaining is shown as a brown color in white adipose tissue. Staining was absent in a negative control section. The primary antibody was replaced with nonimmune serum (1: 100) for negative control slides. Tissues were counterstained with methyl green (N = 3/each group).