MicroRNA-29a represses osteoclast formation and protects against osteoporosis by regulating PCAF-mediated RANKL and CXCL12

Abstract

Osteoporosis deteriorates bone mass and biomechanical strength, becoming a life-threatening cause to the elderly. MicroRNA is known to regulate tissue remodeling; however, its role in the development of osteoporosis remains elusive. In this study, we uncovered that silencing miR-29a expression decreased mineralized matrix production in osteogenic cells, whereas osteoclast differentiation and pit formation were upregulated in bone marrow macrophages as co-incubated with the osteogenic cells in transwell plates. In vivo, decreased miR-29a expression occurred in ovariectomy-mediated osteoporotic skeletons. Mice overexpressing miR-29a in osteoblasts driven by osteocalcin promoter (miR-29aTg/OCN) displayed higher bone mineral density, trabecular volume and mineral acquisition than wild-type mice. The estrogen deficiency-induced loss of bone mass, trabecular morphometry, mechanical properties, mineral accretion and osteogenesis of bone marrow mesenchymal cells were compromised in miR-29aTg/OCN mice. miR-29a overexpression also attenuated the estrogen loss-mediated excessive osteoclast surface histopathology, osteoclast formation of bone marrow macrophages, receptor activator nuclear factor-κ ligand (RANKL) and C-X-C motif chemokine ligand 12 (CXCL12) expression. Treatment with miR-29a precursor improved the ovariectomy-mediated skeletal deterioration and biomechanical property loss. Mechanistically, miR-29a inhibited RANKL secretion in osteoblasts through binding to 3'-UTR of RANKL. It also suppressed the histone acetyltransferase PCAF-mediated acetylation of lysine 27 in histone 3 (H3K27ac) and decreased the H3K27ac enrichment in CXCL12 promoters. Taken together, miR-29a signaling in osteogenic cells protects bone tissue from osteoporosis through repressing osteoclast regulators RANKL and CXCL12 to reduce osteoclastogenic differentiation. Arrays of analyses shed new light on the miR-29a regulation of crosstalk between osteogenic and osteoclastogenic cells. We also highlight that increasing miR-29a function in osteoblasts is beneficial for bone anabolism to fend off estrogen deficiency-induced excessive osteoclastic resorption and osteoporosis.

Fig. 1. Analysis of miR-29a actions to osteogenic and osteoclastogenic differentiation.

Increased Runx2 and osteocalcin expression and von Kossa-stained mineralized matrix formation in miR-29a-transfected bone marrow mesenchymal cells (a); scale bar: 40 μm. Schematic drawing of bone marrow macrophages and mesenchymal cells co-incubated in transwell plates (b). Increasing miR-29a in osteoblasts decreased TRAP-stained osteoclast differentiation (scale bar: 8 μm) and pit formation (scale bar: 20 μm) of bone marrow macrophages (c). The data of in vitro model are expressed as mean ± SEM calculated from four experiments. Asterisks * indicate significant difference from the scrambled control. Sparse trabecular microstructure (scale bar: 5 mm) along with decreased BMD, BV/TV (d), and miR-29a expression in bone tissue (e) occurred in ovariectomized mice. Investigations are expressed mean ± SEM calculated from six mice. Asterisks * indicate significant difference from sham controls

Fig. 2. Skeletal phenotypes and bone microstructure of miR-29aTg/OCN mice.

Increased miR-29a expression solely occurred in bone tissue in miR-29aTg/OCN mice (a). Macroscopic morphology of 7-day-old miR-29aTg/OCN mice was similar to wild-type mice (b); scale bar: 20 mm. Increased serum osteocalcin and reduced TRAP5b levels (c) and strong PCNA immunostaining (d) in osteoblasts in miR-29aTg/OCN mice; scale bar: 30 μm (low magnification); 10 μm (high magnification). Increased fluorescent calcein accumulation and MAR along with femur length in 8-week-old miR-29aTg/OCN mice (e); scale bar: 30 μm. Significant increases in cortical BMD and Ct.Th (f), trabecular BMD and BV/TV (g) together with spinal BMD and BV/TV (h) in miR-29aTg/OCN mice; scale bar: 5 mm. The data are expressed mean ± SEM calculated from six mice. Asterisks * indicate significant difference from wild-type mice

Fig. 3. Analysis of bone mass and biomechanical property of skeletal tissue.

miR-29a overexpression reduced the ovariectomy-induced excessive body weight gain (a) and retained abundant trabecular microstructure (b); scale bar: 5 mm. The ovariectomy-induced loss of BMD (c), bone area (d), B.Ar/T.Ar (e), Tb.Th (f), Tb.N (g), and connectivity (h) were attenuated in miR-29aTg/OCN mice, whereas Tb.Sp (i) and SMI (j) were improved. Load-displacement curve (k) showed minor responses to ovariectomy-induced downregulation of maximum force (l) and breaking force (m) in miR-29aTg/OCN mice. The data are expressed mean ± SEM calculated from six mice. Asterisks * and hashtags # indicate significant difference from sham controls and OVX, respectively. Ampersands & indicate significant difference from Tg-Sham and Tg-OVX

Fig. 4. Analysis of bone histology and ex vivo osteogenesis and osteoclastogenesis.

miR-29a overexpression repressed the ovariectomy-mediated loss of trabecular histology (scale bar: 120 μm), BV/TV (a), fluorescent calcein labeling (scale bar: 30 μm), MAR (b) and Ob.N (c), as well as compromised excessive TRAP-positive osteoclast distribution (scale bar: 8 μm) and Oc.N (d). miR-29aTg/OCN mice showed minor response to the ovariectomy-induced loss of mineralized matrix formation (scale bar: 7 mm) (e), Runx2 and osteocalcin expression (f) of primary bone marrow mesenchymal cells and downregulated osteoclast formation of primary bone marrow macrophages (g); scale bar: 8 μm. The data are expressed mean ± SEM calculated from six mice. Asterisks * and hashtags # indicate significant difference from sham controls and OVX, respectively. Ampersands & indicate significant difference from Tg-Sham and Tg-OVX

Fig. 5. miR-29a precursor treatment for bone microstructure and histology.

miR-29a precursor attenuated the ovariectomy-induced loss of trabecular microarchitecture (a); scale bar: 5 mm; BMD, BV/TV, and Tb.Th (b), as well as improved load-displacement profile (c), maximum force and breaking force (d). The treatment retained fluorescent calcein accumulation (scale bar: 40 μm) and MAR (e) and compromised the ovariectomy upregulation of osteoclast distribution (scale bar: 8 μm) and Oc.N (f). The data are expressed mean ± SEM calculated from eight mice. Asterisks * and hashtags # indicate significant difference from sham controls and OVX, respectively

Fig. 6. Osteoclast-formation capacity of bone marrow macrophages in miR-29aTg/OCN mice.

Schematic drawing of bone marrow macrophages co-incubated with bone marrow mesenchymal stem cells (a). Decreased osteoclast differentiation (scale bar: 8 μm) and pit formation (scale bar: 20 μm) of wild-type bone marrow macrophage and increased mineralized matrix formation (scale bar: 40 μm) of wild-type bone marrow cells co-incubated with miR-29aTg/OCN bone marrow mesenchymal cells (b, c). Investigations are expressed mean ± SEM calculated from six mice. High CD14 + macrophages (d) rather than CD11 + CD115 + monocytes (e) existed in bone marrow in miR-29aTg/OCN mice. Investigations are expressed mean ± SEM calculated from four mice. Asterisks * indicate significant difference from wild-type mice

Fig. 7. Analysis of miR-29a inhibition of RANKL and CXCL12 expression in bone marrow mesenchymal cells.

miR-29a compromised the ovariectomy upregulation of RANKL and CXCL12 expression in bone marrow mesenchymal cells (a). The data are expressed mean ± SEM calculated from six mice. Asterisks * and hashtags # indicate significant difference from sham controls and OVX, respectively. Forced miR-29a expression decreased luciferase activity of 3′-UTR of RANKL mRNA (b) and downregulated RANKL mRNA expression and protein abundances, whereas miR-29 interference upregulated RANKL expression (c) of mesenchymal stem cell cultures. miR-29a overexpression reduced PCAF and H3K27ac abundances (d) and decreased H3K27ac enrichment in CXCL12 promoters (e). Forced PCAF expression repressed the miR-29a downregulation of H3K27ac levels and RANKL expression (f). It increased H3K27ac occupancy in CXCL12 promoters and CXCL12 expression (g) in bone marrow mesenchymal cells. Investigations are expressed mean ± SEM calculated from three experiments. Asterisks * and hashtags # indicate significant difference from WT and miR-29aTg/OCN, respectively

Fig. 8. Schematic drawing of miR-29a signaling in osteoblasts repressed osteoclastic activity and osteoporosis.

miR-29a targets RANKL expression and downregulates PCAF signaling-mediated CXCL12 expression to reduced osteoclastic activity, sustaining bone mass (a). Estrogen loss decreases miR-29a loss, which augments RANKL and CXCL12 expression in osteogenic cells, accelerating osteoclast formation and osteoporosis development (b)