Antenatal Corticosteroid Therapy Attenuates Angiogenesis Through Inhibiting Osteoclastogenesis in Young Mice

Abstract

Antenatal corticosteroid therapy (ACT) has been shown to reduce morbidity and mortality rates in preterm delivery, but the fetus is more likely to face the risk of low bone mineralization and low fetal linear growth. However, the mechanism of ACT inducing low bone mineralization remains largely unknown. Pre-osteoclasts, which play an important role in angiogenesis and osteogenesis, are specifically regulating type H vessels (CD31^hiEmcn^hi) and vessel formation by secreting platelet-derived growth factor-BB (PDGF-BB). We find that the number of pre-osteoclasts and POC-secreted PDGF-BB is dramatically decreased in ACT mice, contributing to the reduction in type H vessels and bone mineralization during the mouse offspring. Quantitative analyses of micro-computed tomography show that the ACT mice have a significant reduction in the mass of trabecular bone relative to the control group. Mononuclear pre-osteoclasts in trabecular bone decreased in ACT mice, which leads to the amount of PDGF-BB reduced and attenuates type H vessel formation. After sorting the Rank+ osteoclast precursors using flow cytometry, we show that the enhancer of zeste homolog 2 (Ezh2) expression is decreased in Rank+ osteoclast precursors in ACT mice. Consistent with the flow data, by using small molecule Ezh2 inhibitor GSK126, we prove that Ezh2 is required for osteoclast differentiation. Downregulating the expression of Ezh2 in osteoclast precursors would reduce PDGF-BB production. Conditioned medium from osteoclast precursor cultures treated with GSK126 inhibited endothelial tube formation, whereas conditioned medium from vehicle group stimulated endothelial tube formation. These results indicate Ezh2 expression of osteoclast precursors is suppressed after ACT, which reduced the pre-osteoclast number and PDGF-BB secretion, thus inhibiting type H vessel formation and ACT-associated low bone mineralization.

Keywords: ACT; EZH2; angiogenesis; osteoclastogenesis; pre-osteoclasts.

FIGURE 1

Prenatal dexamethasone exposuredecreases bone volume and angiogenesis in long bone. Representative μCT images of distal femur in 12-week-old female mouse offspring are shown in panel (A). Quantitative analyses of femoral microstructural parameters of female mouse offspring including trabecular bone volume fraction (BV/TV) (B), trabecular number (Tb. N) (C), trabecular thickness (Tb. Th) (D), and trabecular separation (Tb. Sp) (E). (F,G) Representative images of double-immunofluorescence staining for CD31 (green) and endomucin (Emcn, red) in femoral metaphysis from 12-week-old female offspring. DAPI stains nuclei blue. Images in panel (F) are lower power with boxes outlining the area of higher power in panel (G). Quantification of the relative fluorescence area of CD31⁺Emcn⁺ cells per tissue area in femoral metaphysis (CD31⁺Emcn⁺ area per mm²) (H). (I) Representative images of immunofluorescence staining for osteocalcin (red) and (J) quantitative analysis of osteocalcin⁺ cells in femoral metaphysis from 12-week-old female offspring. DAPI stains nuclei blue. *p < 0.05.

FIGURE 2

Pre-osteoclast number and its synthesis of PDGF-BB decreased in ACT mice. (A,C) Representative TRAP staining (magenta) images (A) and immunofluorescence images of TRAP (red) and PDGF-BB (green) costaining (yellow) (C) in femoral metaphysis from 12-week-old female offspring. DAPI stains nuclei blue. Scale bar = 40 μm. (B,D) Quantitative analysis of the N. POCs (B) and the N. TARP⁺ PDGF-BB⁺ cells (D) in distal femur. (E–H) Representative flow cytometry plots showing the sorting profile used to separate POCs and OCs in femoral bone from 12-week-old female mouse offspring. RANK⁺ cells with 1-2 nuclei (POCs) and greater than three nuclei (OCs) are sorted for gene expression analysis. (I–K) Quantitative analysis of the percentage of RANK⁺ cells (I), POCs (J), and OCs (K) in distal femur. *p < 0.05.

FIGURE 3

Prenatal dexamethasone exposure decreases Ezh2 expression in osteoclast precursors and attenuates osteoclastogenesis. (A) Diagram showing the procedure for the isolation of OCs from femoral metaphysis (see the detailed description in section “Materials and Methods”). (B–E) RANK+ cells were isolated from femoral metaphyses using the same method as in Figure 2. (B) Representative flow cytometry plots showing the sorting profile used to separate POCs and OCs in femoral bone from 12-week-old female mouse offspring. Quantitative RT-PCR analyses of Ezh2 (C) and MafB (D) mRNA expression and Western blot analysis of Ezh2 (E) protein expression in the sorted cells are shown. β-Actin serves as control to ensure appropriate loading. *p < 0.05.

FIGURE 4

Ezh2 plays a critical role in osteoclastogenesis by activating osteoclast regulators and repress the osteoclast inhibitory factors. BMMs were cultured with MCSF for 3 days (days 3 to 0) to form OCL precursors (OCLp) before addition of RANKL/MCSF at day 0 to generate OCL. (A) OCLp were cultured with RANKL/MCSF in the presence of the indicated concentrations of GSK126 for 5 days. TRAP+ OCL formed were analyzed by microscopy (representative images on left) and the counts for multinucleated (n ≥ 3) TRAP+ OCL/well graphed (right). GSK126 (10 μM) was added at the indicated time intervals during the processes of OCLp and OCL (day 4) formation, and representative images and counts for multinucleated (n ≥ 3) TRAP+ OCL/well are shown. (B) OCLp (day 0) or stimulated with RANKL/MCSF ± GSK126 (10 μM; GSK) for 5 days were analyzed for Ezh2 mRNA by quantitative PCR (qPCR). (C) OCLp cultures (day 0) or after stimulation with RANKL/MCSF ± GSK126 (10 μM) for 1 day were analyzed for total cellular levels of Ezh2, H3K27me3, H3, and β-actin by WB. (D–G) OCLp (day 0) or after stimulation with RANKL/MCSF ± GSK126 (10 μM) for 5 days were analyzed for the expression of OCL-specific genes NFATc1, RANK, CatK, ITGB3 by qPCR. (H,I) OCLp (day 0) or after stimulation with RANKL/MCSF ± GSK126 (10 μM) for 1 day were analyzed for the expression of OCL-negative regulators MafB and Irf8 by qPCR. WB analyses show representative of three independent experiments and error bars in mRNA experiments represent SEM for 4 biological replicates. *p < 0.05.

FIGURE 5

Loss of Ezh2 suppress osteoclast precursors PDGF-BB expression and bone vasculature in ACT mice. (A) Diagram showing the procedure for the collection of condition media of OCs under denoted conditions and the co-culture with the EPCs. (B,C) Representative images (B) and tube length (C) from Matrigel endothelial tube formation assay with addition of OCLp (red) and after stimulation with RANKL/MCSF ± GSK126 (10 μM) (green) culture media. (D,E) Quantification of Pdgfb mRNA by real-time PCR (D) and total cellular levels of PDGF-BB and β-actin by WB (E). mRNA normalized to GAPDH. *p < 0.05.