Global gene deletion studies in mice and humans have established the pivotal role of runt related transcription factor-2 (Runx2) in both intramembranous and endochondral ossification processes during skeletogenesis. In this study, we for the first time generated mice carrying a conditional Runx2 allele with exon 4, which encodes the Runt domain, flanked by loxP sites. These mice were crossed with α1(I)-collagen-Cre or α1(II)-collagen-Cre transgenic mice to obtain osteoblast-specific or chondrocyte-specific Runx2 deficient mice, respectively. As seen in Runx2−/− mice, perinatal lethality was observed in α1(II)-Cre;Runx2flox/flox mice, but this was not the case in animals in which α1(I)-collagen-Cre was used to delete Runx2. When using double-staining with Alizarin red for mineralized matrix and Alcian blue for cartilaginous matrix, we observed previously that mineralization was totally absent at embryonic day 15.5 (E15.5) throughout the body in Runx2−/− mice, but was found in areas undergoing intramembranous ossification such as skull and clavicles in α1(II)-Cre;Runx2flox/flox mice. In newborn α1(II)-Cre;Runx2flox/flox mice, mineralization impairment was restricted to skeletal areas undergoing endochondral ossification including long bones and vertebrae. In contrast, no apparent skeletal abnormalities were seen in mutant embryo, newborn, and 3-week-old to 6-week old-mice in which Runx2 had been deleted with the α1(I)-collagen-Cre driver. These results suggest that Runx2 is absolutely required for endochondral ossification during embryonic and postnatal skeletogenesis, but that disrupting its expression in already committed osteoblasts as achieved here with the α1(I)-collagen-Cre driver does not affect overtly intramembranous and endochondral ossification. The Runx2 floxed allele established here is undoubtedly useful for investigating the role of Runx2 in particular cells. © 2013 American Society for Bone and Mineral Research.
There are two types of skeletal ossification processes during embryonic and postnatal skeletogenesis. In the endochondral ossification process, the cartilaginous template composed of chondrocytes differentiated from mesenchymal stem cells is eventually replaced with bone containing osteoblasts and osteoclasts. In a process called intramembranous ossification, by contrast, mesenchymal stem cells directly differentiate into osteoblasts. Runt-related transcription factor-2 (Runx2) is a member of the Runt family of transcription factors, which plays a critical role in cellular differentiation processes in osteoblasts and chondrocytes in mice and in humans. Runx2 can trigger an osteoblast gene expression program when increased in mesenchymal cells. Furthermore, genetically modified mouse models have been generated and analyzed with Runx2 for years. These models include global Runx2 deletion (Runx2−/−) mice, transgenic mice overexpressing Runx2 or dominant negative Runx2 (dnRunx2) in osteoblasts under the control by a α1(I)-collagen promoter or in chondrocytes under the control by a α1(II)-collagen promoter. Together these experiments showed that Runx2 is both necessary to induce osteoblast differentiation. In effort to further improve our understanding of when Runx2 becomes necessary during bone formation in vivo, we have generated conditional knockout mice lacking exon 4 of the Runx2 gene, which was crossed, in proof-of-principle initial experiments, with α1(I)-collagen-Cre (α1(I)-Cre;Runx2flox/flox) or α1(II)-collagen-Cre transgenic mice (α1(II)-Cre;Runx2flox/flox).
Subjects and Methods
Twelve splice variants have been shown to exist for Runx2 at protein and mRNA levels according to an Ensemble database search (http://www.ensembl.org/index.html). Because all 12 splice variants contain exon 4 (chromosome 17: 44,724,695–44,724,951), which encodes a functional domain of Runx2 referred to as Runt domain, we decided to generate conditional knockout mice with a floxed exon 4 of the Runx2 locus (Acc. No. CDB0832K: http://www.cdb.riken.jp/arg/mutant%20mice%20list.html) in place of exon 8 targeted previously. The targeting vectors harboring loxP sites and the neomycin resistant (NeoR) cassette were electroporated into TT2 embryonic stem (ES) cells. DNA specimens were extracted from different ES cell clones after neomycin selection for the mutant allele search by Southern blotting using [32P]-labeled probe for the target region of genomic DNA digested with BglII. The wild-type (WT) allele yielded a 6.4-kb fragment, whereas the homologous targeted mutant allele yielded an 8.4-kb fragment. ES cells containing the floxed allele were injected into CD-1 eight-cell stage embryos to generate chimeric mice. To remove the NeoR franked by flippase recombinase target (FRT) sequences, the offspring was crossed with flippase (FLPe) transgenic mice. Runx2flox/+ mice were crossed with either α1(I)-collagen-Cre or α1(II)-collagen-Cre mice to generate α1(I)-Cre;Runx2flox/+ or α1(II)-Cre;Runx2flox/+ mice, and their progeny intercrossed to obtain α1(I)-Cre;Runx2flox/flox or α1(II)-Cre;Runx2flox/flox. Runx2−/− mice were generated as reported. Genotyping was performed by polymerase chain reaction (PCR) using tail genomic DNA. The presence of the 3′ loxP site was verified by PCR using primers (5′-TAAATCCAGATGCCCCTGAG-3′; 5′-TTGAAACCATCCACAGGTGA-3′). The protocol employed here meets the guidelines of the Japanese Society for Pharmacology and was approved by the Committee for Ethical Use of Experimental Animals at Kanazawa University (permit numbers: 71061, 71066).
Primary osteoblasts were prepared from calvarial of 5-day-old mouse pups and cultured in α-minimum essential medium (α-MEM) with 10% fetal bovine serum (FBS) in the presence of 100 μg/mL ascorbic acid and 5 mM β-glycerophosphate for 7 days. Primary chondrocytes were prepared from cartilage of neonatal mouse ribs and maintained in Dulbecco's modified Eagle medium (DMEM) containing 10% FBS and 50 μg/mL ascorbic acid, 1 mM pyruvate, and 1 mM cysteine for 7 days. Western blotting was performed by using anti-Runx2 (C-20) antibody (MBL, Nagoya, Japan) or anti–glyceraldehyde-3-phosphate dehydrogenase (GAPDH) antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA).
Embryos at embryonic day 15.5 (E15.5) were eviscerated and the skins were removed. After fixation in 95% ethanol overnight, embryos were immersed in Alcian blue solution (150 mg Alcian blue, 800 mL 98% ethanol, and 200 mL acetic acid) overnight. After several hours in 95% ethanol, they were kept in 2% KOH for 24 hours, followed by staining in Alizarin red solution (50 mg/L Alizarin red in 2% KOH) overnight and subsequent clearing of skeletons in 1% KOH/20% glycerol for storage in 50% ethanol/50% glycerol.
Histological, histomorphometric, and micro–computed tomography analyses
Neonatal mouse tibias were fixed with 10% formalin neutral buffer solution, embedded in paraffin and then sectioned at a thickness of 6 to 8 μm. Sections were stained with hematoxylin and eosin (H&E), von Kossa, and Alcian blue staining. Bone histomorphometric analyses were performed on undecalcified sections using Osteomesure Analysis Systems (Osteometrics) according to the standard protocol. Trabecular bone architecture was assessed in femur using micro–computed tomography (µCT).
Results and Discussion
The mutant mouse strain expressing a floxed allele of Runx2 using Cre-loxP strategies was generated as shown in Fig. 1A. The WT allele yielded a 6.4-kb fragment, whereas the homologous targeted mutant allele yielded an 8.4-kb fragment (Fig. 1B). Positive ES cells were injected into ICR eight-cell stage embryos to generate chimeric mice for subsequent crossing with C57BL/6J females. This step provided heterozygous animals for the floxed allele of Runx2 (Runx2flox-neo). Mating these mice with FLPe transgenic mice resulted in the heterozygous floxed Runx2 (Runx2flox/+) with the frt-franked neomycin resistance cassette eliminated. Runx2flox/+ mice were next crossed with either α1(I)-collagen-Cre or α1(II)-collagen-Cre transgenic mice to exclusively delete the gene in either osteoblasts or chondrocytes, respectively. The α1(I)-collagen-Cre driver deletes genes in cells already committed to the osteoblast lineage, whereas the α1(II)-collagen-Cre driver deletes genes in chondrocytes; ie, earlier during skeletogenesis. Genotyping for floxed allele and Cre transgene was performed by genomic PCR (Fig. 1C). Western blotting analysis revealed drastically decreased Runx2 protein expression in osteoblasts and long bone from α1(I)-Cre;Runx2flox/flox and chondrocytes from α1(II)-Cre;Runx2flox/flox mice, respectively (Fig. 1D). When α1(I)-Cre;Runx2flox/+ mice were crossed with Runx2flox/flox mice, α1(I)-Cre;Runx2flox/flox mice were obtained according to the Mendelian ratio at 2 weeks old (4 homozygotes out of 15 pups). Although α1(II)-Cre;Runx2flox/flox embryos were alive at E15.5 (4 homozygotes out of 20 embryos); however, α1(II)-Cre;Runx2flox/flox mice died at birth due to breathing difficulties. Perinatal lethality of α1(II)-Cre;Runx2flox/flox mice was almost similar to that seen in Runx2−/− mice.
Alizarin and Alcian blue staining of skeletal preparations was performed in E15.5 embryos and newborns at 1 day old (P1) in Runx2−/−, α1(I)-Cre;Runx2flox/flox, and α1(II)-Cre;Runx2flox/flox mice (Fig. 1E). In E15.5 embryos, no patterning defects were observed in either Runx2−/− or α1(II)-Cre;Runx2flox/flox mice. As reported, E15.5 Runx2−/− mice exhibited a complete absence of mineralized matrix in developing skeletal elements bones. Likewise, E15.5 α1(II)-Cre;Runx2flox/flox mice exhibited an almost complete absence of mineralized tissues stained with Alizarin red, at the exception of facial bone, skull bone, and clavicle, which are all formed through an intramembranous ossification process (Fig. 2A). In contrast, α1(I)-Cre;Runx2flox/flox mice generated with this specific Cre driver showed normal skeletal development based on this assay. In newborn Runx2−/− mice, weak positive Alizarin red staining was seen in tibia, fibula, radius, ulna, dorsal arch of vertebrae, and dorsal root of rib as shown. In newborn α1(II)-Cre;Runx2flox/flox mice, positive Alizarin red staining was found in facial bone, skull bone, clavicle, femur and part of the sternum, in addition to the aforementioned areas, including tibia, fibula, radius, ulna, dorsal arch of vertebrae, and dorsal root of rib. However, no marked aberrancy was seen in skeletal development in newborn mice in which Runx2 had been deleted with the α1(I)-collagen-Cre driver (Fig. 2B).
Histological analyses were performed by staining sections of tibias with H&E, von Kossa for calcified tissues, and Alcian blue for cartilage. In α1(II)-Cre;Runx2flox/flox mice, the middle part of the tibia remained as calcified cartilage without formation of the bone marrow cavity. However, there was no apparent difference in histological analyses with H&E, von Kossa, and Alcian blue between WT and α1(I)-Cre;Runx2flox/flox mice (Fig. 2C). Even when assessed in 6-week-old mice, the ablation of Runx2 by the α1(I)-collagen-Cre driver did not significantly affect bone phenotypes measured by histomorphometric analysis in vertebrae (Fig. 2D) and µCT analysis in femurs (Fig. 2E).
We present here the first description of a floxed Runx2 allele and show that deleting this gene at various stages of skeletogenesis has markedly different consequences. Indeed, when this deletion occurs early during skeletogenesis through the use of α1(II)-collagen-Cre, it reproduces virtually all the defects of endochondral ossification seen in Runx2−/− mice, establishing that those are due to a cell-autonomous defect. This was not the case when using α1(I)-collagen-Cre to achieve Runx2 deletion. Although no apparent skeletal phenotypes were observed in E15.5, newborn, and 3-week-old to 6-week-old α1(I)-Cre;Runx2flox/flox mice, this analysis does not mean or imply that Runx2 is not needed in osteoblasts. The driver gene α1(I)-collagen is weakly expressed by preosteoblasts at the first phase of osteoblastogenesis, whereas its expression is highly upregulated in the next phase of differentiation into immature osteoblasts. The use of a driver gene other than α1(I)-collagen could give us a clue for the clarification of postnatal bone phenotypes. A previous study showed that conditional deletion of exon 8 of Runx2 with the α1(II)-collagen-Cre driver produced similar phenotypes and impaired endochondral ossification to the ones reported here. This only means that the differentiation role of Runx2 in the osteoblast lineage is best seen if Runx2 is deleted before osteoblast differentiation has been initiated. Indeed it has already been shown that modulating Runx2 expression in osteoblasts once they are differentiated can affect bone mass accrual later in life. Hence, this study serves to illustrate how useful a floxed allele of Runx2 may become to studying its function at different stages of development and postnatally.
All authors state that they have no conflicts of interest.
This work was supported in part by Grants-in-Aid for Scientific Research (22500330 to TT) from the Ministry of Education, Culture, Sports, Science and Technology, Japan and by a Research Grant from The Naito Foundation, Japan (to TT).
Authors' roles: Study design: TT, EH, ST, GK, and YY. Study conduct: TT, EH, RN, HO, CX, AT, TA, HK. Data collection: TT, RN, HO, CX, AT, TA, and HK. Data analysis: TT and EH. Data interpretation: TT, EH, ST, GK, and YY. Drafting manuscript: TT, EH, ST, GK, and YY. Approving final version of manuscript: TT, EH, ST, GK, and YY. YY takes responsibility for the integrity of the data analysis.