Postnatally induced inactivation of Osterix in osteoblasts results in the reduction of bone formation and maintenance

Abstract

Osterix (Osx) is a zinc-finger-containing transcription factor that is highly specific to osteoblasts in vivo. Because Osx homozygous null mutants die in the immediate perinatal period showing a complete absence of bone formation, it is impossible determine the role that Osx plays in bones that have already formed after birth. To determine whether Osx is essential for bone maintenance and homeostasis, we conditionally inactivated the Osx gene in adult bone using the Cre/loxP recombination system. In previous reports, 2.3-kb Col1a1-CreERT2 mice that expressed a Cre recombinase that is transiently inducible by 4-hydroxytamoxifen (4-OHT) were intercrossed with Rosa26R (R26R) reporter mice, which resulted in the production of Cre-expressing osteoblasts that were detected upon X-gal staining. In the present study, inducible Col1a1-CreERT2 transgenic mice and conditional Osx mice (Osx(flox/+)) were used to generate Osx(flox/-);Col1a1-CreERT2 mice. The Osx gene in Osx(flox/-);Col1a1-CreERT2 mice was inactivated in the osteoblasts of already formed bones by active Cre recombinase after the administration of 4-OHT. The bones from 4-OHT-treated Osx(flox/-);Col1a1-CreERT2 mice and oil-treated control mice were analyzed by radiography, histology, and histomorphometry. Even though no significant difference was observed in the radiographic images of the whole mouse skeletons, the mineralized trabecular bone volume and number in lumbar vertebrae were remarkably reduced in 4-OHT-treated Osx(flox/-);Col1a1-CreERT2 mice. In addition, the rate of bone formation and area of mineralized surface were also reduced in 4-OHT-treated Osx(flox/-);Col1a1-CreERT2 mice. Osx inactivation in already formed bones during the postnatal period caused a functional defect in osteoblasts that was followed by a reduction of bone formation, even though there were no apparent differences in osteoblast proliferation and osteoclast formation. Taken together, these results indicate that Osx is required to maintain osteoblast function following adult bone maintenance.

Fig. 1

EGFP expression that recapitulates the Osx expression in bones after Cre-mediated Osx excision. (A) EGFP expression in Osx^Δex/+ heterozygous embryo and pups at embryonic day 18.5 (E18.5) and postnatal days 2 and 15 (P2 and P15). Heterozygotes carrying the Osx^Δex allele expressed EGFP to recapitulate Osx expression in all bones from embryos and pups. (B) Skeletons of Osx^Δex/+ and Osx^Δex/− mice were stained with alcian blue and alizarin red at E18.5, indicating an absence of mineralization in Osx^Δex/− null mutants. Bone deformity was remarkably observed in Osx^Δex/− null mutants by EGFP expression.

Fig. 2

Generation of the Osx^flox/−; Col1a1-CreERT2 mice. (A) Breeding scheme to generate Osx^flox/−; Col1a1-CreERT2 mice by crossing conditional Osx^flox/+ mice with Osx^+/− mice with a LacZ knock-in in the Osx locus and a Col1a1-CreERT2 transgenic line. Inducible Col1a1-CreERT2 transgenic mice were used to excise the Osx gene from osteoblasts by the administration of 4-OHT. (B) PCR genotyping to detect the flox, LacZ, Cre allele, and deleted Osx exon 2 (Δex) before and after the administration of 4-OHT. Primers for each PCR genotyping were indicated in structure of the genomic Osx locus: black arrows for the floxed allele, blue arrows for the LacZ allele, and black and red arrows for the Δex allele. In PCR with primers of black arrows, the wild-type and the floxed alleles were amplified to generate a 300-bp and 390-bp fragments, respectively. (C, D) 4-OHT-induced expression of EGFP in Osx^flox/−; Col1a1-CreERT2 embryos and pups. (C) Pregnant females from the breeding scheme shown in (A) were injected with 4-OHT. Only Osx^flox/−; Col1a1-CreERT2 embryos from females expressed EGFP in bone. (D) Osx^flox/−; Col1a1-CreERT2 pups were injected with 1 mg of 4-OHT for 5 consecutive days starting at postnatal day 12, while control pups were injected with oil. EGFP expression was observed in bone, including the jaw, digits, and tibia/fibula. No EGFP expression was detected in oil-injected control animals.

Fig. 3

Induction of recombination in the bones of Osx^flox/−; Col1a1-CreERT2 mice. (A) Scheme for 4-OHT administration to inactivate Osx. Mice were intraperitoneally injected with 1 mg of 4-OHT or oil for 5 consecutive days, twice. The treated mice were sacrificed 4, 8, or 12 weeks after the final injection and their bones were then analyzed. (B) X-ray radiography of whole skeletons of 4-OHT or oil-treated Osx^flox/−; Col1a1-CreERT2 mice at 18 weeks of age. No difference in the lucency of the entire skeleton was observed between 4-OHT-treated mice and the oil-treated controls.

Fig. 4

Reduced bone mass in 4-OHT-treated Osx^flox/−; Col1a1-CreERT2 mice. (A) Histological analysis with von Kossa staining of the lumbar vertebrae from Osx^flox/−; Col1a1-CreERT2 mice. Osx was inactivated in the osteoblasts of intact bones after birth by the administration of 4-OHT using an inducible Cre system. Decreased bone mass was observed in 4-OHT-treated Osx^flox/−; Col1a1-CreERT2 mice compared to oil-treated controls at 18 weeks of age. (B, C) Histomorphometric analysis in Osx^flox/−; Col1a1-CreERT2 mice. At 18 weeks of age, the significant decrease of bone mass and trabecular numbers was observed in 4-OHT-treated (black bar) mice compared to oil-treated Osx^flox/−; Col1a1-CreERT2 mice (white bar), whereas trabecular separation was increased in 4-OHT-treated mice. BV/TV, bone volume per tissue volume; Tb.Th, trabecular thickness; Tb.N, trabecular number; Tb.Sp, trabecular separation; *p<0.05.

Fig. 5

Decreased bone formation in 4-OHT-treated Osx^flox/−; Col1a1-CreERT2 mice at 18 weeks of age. (A) Fluorescent micrographs of calcein double labeling with the distance indicating osteoblast functional activity. A reduced distance between the two labels was observed in 4-OHT-treated mice compared to oil-treated Osx^flox/−; Col1a1-CreERT2 mice. (B) Histomorphometric analysis of calcein-labeled lumbar vertebrae in oil-and 4-OHT-treated Osx^flox/−; Col1a1-CreERT2 mice. MS, MAR, and BFR were reduced remarkably in the bones of 4-OHT-treated Osx^flox/−; Col1a1-CreERT2 mice (black bar). MS/BS, mineralized surface per bone surface; MAR, mineral apposition rate; BFR/BS, bone forming rate per bone surface; *p<0.05.

Fig. 6

Osteoblast differentiation in 4-OHT-treated Osx^flox/−; Col1a1-CreERT2 mice. (A) No overt histological phenotype in 4-OHT-treated Osx^flox/−; Col1a1-CreERT2 mice. Longitudinal sections of tibiae were subjected to H and E, alcian blue, and TRAP staining. No morphological differences were observed in H and E staining. In differentiating chondrocytes and mature osteoclasts by alcian blue and TRAP staining, respectively, no significant differences were observed between 4-OHT-treated and oil-treated Osx^flox/−; Col1a1-CreERT2 mice. In vivo cell proliferation was analyzed based on the incorporation of BrdU into the mice tibia. BrdU-positive cells were not altered in the tibia of 4-OHT-treated Osx^flox/−; Col1a1-CreERT2 mice compared to the controls. The numbers of osteoblasts, osteoclasts, and BrdU-positive cells were quantified in both mice. No significant differences were observed. Representative images and analysis were shown in mouse bones at 10 weeks of age. Scale bar=200 μm. (B) Immunohistochemical analysis using anti-Osx antibody. Osteoblasts with Osx expression were detected in black. No signal was observed in osteoblasts of 4-OHT-treated Osx^flox/−; Col1a1-CreERT2 mice. Representative images were shown in mouse bones at 10 weeks of age. Scale bar=50 μm. (C) Expression of marker genes related to osteoblastic cell differentiation by quantitative real-time RT-PCR analysis. The expression of osteogenic markers, BSP and Col1a1, and a late marker of osteoblast differentiation, OCN, were obviously reduced, whereas the expression of an early marker gene, ALP, was not significantly changed in 4-OHT-treated Osx^flox/−; Col1a1-CreERT2 compared to oil-treated controls. *p<0.05.