Maxillofacial Orthognathics, Department of Maxillofacial Reconstruction and Function, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, Tokyo, Japan
Hard Tissue Genome Research Center, Tokyo Medical and Dental University, Tokyo, Japan
Correspondence to: Keiji Moriyama, Maxillofacial Orthognathics, Department of Maxillofacial Reconstruction and Function, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8549, Japan. E-mail: firstname.lastname@example.org
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.
Apert syndrome (AS) (OMIM: 101200) is an autosomal dominantly inherited syndrome characterized by craniosynostosis, midfacial hypoplasia, and symmetric bony syndactyly of the hands and feet. It occurs in approximately 1/65,000 births and accounts for 4.5% of all craniosynostosis (Cohen et al., 1992). AS is caused by one of two missense mutations of the fibroblast growth factor receptor (FGFR) 2 gene involving the following amino acid substitutions: S252W or P253R (Wilkie et al., 1995; Oldridge et al., 1997, 1999), resulting in gain-of-function of fibroblast growth factor (FGF) signaling (Anderson et al., 1998; Ibrahimi et al., 2001).
FGFRs constitute a family of four membrane-spanning tyrosine kinases, and 18 FGF ligands have been identified (Powers et al., 2000; Ornitz and Itoh, 2001; Coumoul and Deng, 2003). FGFR signaling induces the proliferation, migration, differentiation, and survival of many cell types (Ornitz and Itoh, 2001; Eswarakumar et al., 2005). FGFRs contain a hydrophobic leader sequence, three immunoglobulin (Ig)-like domains, an acidic box, transmembrane region, and a divided tyrosine kinase domain. The IgIII domains of FGFR1–3 are encoded for by exons 7–9. Inclusion of exons 8 and 9 is mutually exclusive, producing the IIIb and IIIc splice isoforms. FGFR2IIIb is located in the epithelium and mainly binds to FGF7 and 10, whereas FGFR2IIIc is present in the mesenchyme and has a high affinity for FGF2, 4, 6, 8, and 9. AS mutations cause loss of ligand specificity of FGFR2 isoforms (Yu et al., 2000).
Recently, two independent mouse models of Fgfr2+/S252W (Chen et al., 2003; Wang et al., 2005) and Fgfr2+/P253R (Yin et al., 2008; Wang et al., 2010) have been established. Similar to AS patients, both models exhibit premature coronal suture (CS) fusion, midfacial hypoplasia, and a domed skull. Furthermore, Wang et al. have reported enhanced chondrogenic markers and ectopic cartilage formation in the Fgfr2+/S252W mice at sagittal sutures, while the interfrontal sutures (IFSs) showed defects without any cartilage formation (Wang et al., 2005). They proposed that the incorrect migration or localization of neural crest cells caused this ectopic cartilage formation given the difference in tissue origin between the sagittal suture and IFS (Wang et al., 2005). In humans, infants with AS frequently show bone defects in the calvarial midline with a varying degree of heterotopic bone formation between the osteogenic fronts (Cohen and Kreiborg, 1996). Some reports have described the effects of the AS mutations on downstream signaling pathways in vivo (Shukla et al., 2007; Yin et al., 2008; Holmes et al., 2009; Wang et al., 2010). These mouse models are used to evaluate novel targets and strategies for treatment of AS. The MEK1 inhibitor PD98059 was shown to reduce CS fusion in calvarial explants from Fgfr2+/P253R mice (Yin et al., 2008). Both pre- and post-natal administration of the MEK1/2 inhibitor U0126 inhibited AS-like phenotypes in Fgfr2+/S252W male mice, as did gene therapy to express a short hairpin RNA against the Fgfr2+/S252W allele (Shukla et al., 2007). However, synostosis resumed after withdrawal of the chemical inhibitor, which reflects the postoperative re-synostosis phenomenon frequently observed in syndromic craniosynostosis patients. Inhibitors of other pathways that promote osteoblastic differentiation may provide alternatives for direct inhibition of FGF signaling (Holmes, 2012), but they have not been evaluated in an AS mouse model.
Our previous study demonstrated that a soluble form of the Apert mutant FGFR2 (sFGFR2IIIcS252W), which lacks the transmembrane and cytoplasmic domains, acted as a decoy receptor by competing for ligand binding with FGFRs. This mutant inhibited enhanced osteoblastic differentiation in the MG63 osteosarcoma cell line transfected with FGFR2IIIcS252W (Tanimoto et al., 2004). Furthermore, calvarial osteoblasts derived from FGFR2IIIcS252W transgenic mice proliferated and differentiated via highly activated MEK, ERK, and p38 pathways, whereas these pathways were suppressed in calvarial osteoblasts expressing sFGFR2IIIcS252W (Suzuki et al., 2012). To evaluate the inhibitory effect of FGFR2IIIcS252W in vivo, we examined the effects of sFGFR2IIIcS252W on the Fgfr2+/S252W background in this study.
Generation of Mutant Mice
To examine the therapeutic effects of sFGFR2IIIcS252W on the Apert mouse model, we generated triple mutant mice with EIIa-Cre, Fgfr2+/Neo-S252W, and sFGFR2IIIcS252W. Mating of Fgfr2+/Neo-S252W with EIIa-Cre sFGFR2IIIcS252W mice produced various genotypes such as wild-type, EIIa-Cre, sFGFR2IIIcS252W, Fgfr2+/Neo-S252W, EIIa-Cre Fgfr2+/Neo-S252W (Ap), sFGFR2IIIcS252W Fgfr2+/Neo-S252W, EIIa-Cre sFGFR2IIIcS252W, and EIIa-Cre sFGFR2IIIcS252W Fgfr2+/Neo-S252W (Ap/Sol) (Table 1). Genotypic analysis of 242 offspring showed that the genotype distribution was different from the theoretical values. The frequencies of Fgfr2+/Neo-S252W, EIIa-Cre sFGFR2IIIcS252W, and Ap/Sol mice were lower than expected, while higher values were observed for EIIa-Cre and sFGFR2IIIcS252W mice.
A previous study has indicated that transcript levels of the exogenous sFGFR2IIIcS252W gene in calvarial tissue were marginal (Suzuki et al., 2012). Therefore, we measured the protein levels of sFGFR2IIIcS252W in the blood of sFGFR2IIIcS252W mice because sFGFR2IIIcS252W could be present in blood as a result of its transcription in tissues other than the calvaria. Western blot analysis demonstrated that FLAG-tagged sFGFR2IIIcS252W proteins with molecular masses of about 60 and 35 kDa were present in the serum of sFGFR2IIIcS252W mice (Fig. 1C). Wild-type mice did not show expression of these proteins. The 60-kDa protein was likely native sFGFR2IIIcS252W−3xFLAG, while the 35-kDa protein might have been a degradation product of full-length sFGFR2IIIcS252W−3xFLAG in vivo.
Macroscopic Examination of Fgfr2+/S252W and sFGF R2IIIcS252W Fgfr2+/S252W Mice
General inspection of the appearance of the whole body indicated that Ap mice at P1 were apparently smaller than wild-type mice as reported previously (Fig. 2A) (Chen et al., 2003; Shukla et al., 2007). Ap/Sol mice had a body size between that of Ap and wild-type mice (Fig. 2A). Measurement of body weight supported the observations based on appearance. Body weights of Ap/Sol mice were higher than those of Ap mice but lower than those of wild-type mice, although the differences were not statistically significant (Fig. 2B).
Rescue of the Apert Mouse Model Phenotype by sFGFR2IIIcS252W
To examine the therapeutic effects of sFGFR2IIIcS252W on AS-like phenotypes, triple mutant (Ap/Sol) mice were subjected to micro-CT analysis for morphological characterization and bone mineral density (BMD) analysis, then compared with Ap and wild-type mice. Ap mice had the same BMD at the equivalent site of the coronal suture as the parietal and frontal bones, presumably indicating premature fusing of the CS (white arrowheads in Fig. 3Ab, e), as shown previously in human AS. A wide bony defect at the IFS was also observed in Ap mice at P1 (white arrow in Fig. 3Ah) and P8 (red arrowheads in Fig. 3Bb). Introduction of sFGFR2IIIcS252W to Ap mice altered the CS phenotypes to those that were between the phenotypes of the wild-type and Ap mice. In addition, the IFS of Ap/Sol mice at P1 was narrower, and the width was close to that of wild-type mice (white arrow in Fig. 3Ai). At P8, there was little difference between Ap/Sol and wild-type mice (Fig. 3Bc). BMD of the parietal bone and IFS width were estimated based on micro-CT data. No significant difference was observed among wild-type, Ap, and Ap/Sol mice in BMD (Fig. 3C). IFS widths in Ap/Sol mice were close to those in wild-type mice, but statistically different from those in Ap mice (Fig. 3D).
In Ap mice (at P1), histological analysis demonstrated that the osteogenic fronts in the CS were irregular and fused prematurely (arrow in Fig. 4Ab). Ap/Sol mice exhibited a patent osteogenic front in the CS very close to that of the wild-type phenotype (Fig. 4Ac). Interestingly, ectopic bone formation and greatly thickened calvarial cartilage in wide IFS were obvious in Ap mice (arrowheads in Fig. 4Bb, e, h), while heterotopic skeletal tissues were not observed in Ap/Sol mice (Fig. 4Bc, f, i). The incidences of the AS-like phenotypes are summarized in Table 2. The frequencies of CS irregularity and IFS with ectopic bones of Ap/Sol mice were significantly different compared to those of Ap mice, while there were no significant differences between those of wild-type and Ap/Sol mice. The frequencies of widened IFS of Ap/Sol mice were lower than those of Ap mice but higher than those of wild-type mice, although the differences were not statistically significant.
Table 2. Incidence of AS-Like Phenotypes
Wild-type vs Ap
Ap vs Ap/Sol
wild-type vs Ap/Sol
Statistically significant at P < 0.05 after Bonferroni correction.
Statistically significant at P < 0.01 after Bonferroni correction.
Statistically significant at P < 0.001 after Bonferroni correction.
We investigated the effects of sFGFR2IIIcS252W on Ap mice. Our results showed, for the first time, that sFGFR2IIIcS252W partially rescued the AS-like phenotypes, including low body weight, CS synostosis, ectopic bone formation, thickened cartilage, and widened IFSs in vivo. Previous studies have described these syndrome phenotypes in Ap mice, with the exception of ectopic bone formation and thickened cartilage in IFSs, which were observed in this study (Chen et al., 2003; Wang et al., 2005; Shukla et al., 2007; Holmes et al., 2009).
Ectopic cartilage formation in IFSs, which are derived from neural crest cells (Jiang et al., 2002), presumably results from aberrant FGF signaling in pre-migratory neural crest cells at an early embryonic stage as previously indicated in vitro (Petiot et al., 2002). This mechanism may support our observation of thickened cartilage formation in the IFS of Ap mice. Interestingly, we observed for the first time ectopic bone formation in the wide IFS of Ap mice, which has been demonstrated in human AS patients (Kreiborg and Cohen, 1990; Cohen and Kreiborg, 1996).
We hypothesize that the reversal of AS-like phenotypes is, at least in part, attributable to repression of aberrantly enhanced FGF signaling in Ap mice by the dominant negative effects of sFGFR2IIIcS252W, which acts as a decoy receptor in vivo. This notion is supported by our observations and previous studies by others. We have observed that sFGFR2IIIcS252W can bind a broad range of ligands and dimerize with normal membrane FGFR2 (Yokota et al., unpublished data); sFGFR2IIIcS252W also inhibited aberrant mineralization of MG63 cells overexpressing FGFR2S252W (Tanimoto et al., 2004). In addition, we observed that sFGFR2IIIcS252W repressed the enhanced signaling of the MEK, ERK, and p38 pathways in osteoblasts expressing FGFR2S252W (Suzuki et al., 2012). A previous study has shown the recovery of AS-like phenotypes in Ap mice by administration of an inhibitor U0126 for the MEK-ERK pathway, which is downstream of FGFR (Shukla et al., 2007). Collectively, Ap/Sol mice appeared to have partial recovery of the AS-like phenotypes. The MEK-ERK pathway, which is inhibited by U0126, is involved in the pathogenesis of AS. However, the inhibitor may influence MEK-ERK pathways downstream of receptors other than FGFR2, such as epidermal growth factor receptor (Herbst, 2004) and vascular endothelial growth factor receptor (Ferrara et al., 2003), leading to undesirable effects. In addition, since U0126 only inhibits the MEK-ERK pathway, it would not inhibit other signaling pathways that are downstream of FGFR2. In contrast, all signaling pathways downstream of FGFR2 are possibly affected by sFGFR2IIIcS252W because sFGFR2IIIcS252W functions at the ligand-receptor binding level. As such, sFGFR2IIIcS252W would be more advantageous than U0126 for the potential treatment of AS. On the other hand, sFGFR2S252W would be difficult to use practically in terms of bioavailability and pharmacokinetics, compared to small molecule inhibitors that target FGFR2 signaling. It is difficult to envisage sFGFR2IIIcS252W being delivered in utero to correct developmental defects associated with AS.
To further elucidate the mineralization status of the calvaria in the Apert mouse model, we examined the BMD of the parietal bone of Ap and Ap/Sol mice by quantitative micro-CT analysis. Kreiborg and Cohen (1990) indicated that human Apert calvaria was hypomineralized based on CT examinations. We did not observe any significant differences in the BMD among wild-type, Ap, and Ap/Sol mice. However, calvaria at P1 may be too small and thin to detect differences in BMD. It is also difficult to configure equal regions of interest (ROIs) because there was considerable variability in the volume and shape of individual calvaria, particularly in the mutant strains. To evaluate BMD more precisely, we will need an alternative method to micro-CT analysis for future studies.
Our analysis of genotypes of mutant mice indicated that the birth prevalence of EIIa-Cre, sFGFR2IIIcS252W, Fgfr2+/Neo-S252W, EIIa-Cre sFGFR2IIIcS252W, and Ap/Sol mice was different from the theoretical ratio. While we do not know the exact mechanism for the difference between the observed prevalence and theoretical ratio, the critical role of FGF signaling during early embryogenesis is apparent.
In conclusion, our study indicates that sFGFR2IIIcS252W can potentially prevent AS-like phenotypes in vivo. Some questions remain unanswered, such as determination of the optimum dosage for treatment, an efficacious drug delivery system, and evaluation of other potential side effects.
Generation of the Apert Mouse Model Expressing the Soluble Form of FGFR2IIIcS252W
Fgfr2+/Neo-S252W, sFGFR2IIIcS252W, and EIIa-Cre mice have been described previously (Lakso et al., 1996; Chen et al., 2003; Suzuki et al., 2012). sFGFR2IIIcS252W mice were first crossed with EIIa-Cre mice to generate EIIa-Cre sFGFR2IIIcS252W mice. Crossing between male EIIa-Cre sFGF R2IIIcS252W mice and female Fgfr2+/Neo-S252W mice generated EIIa-Cre, sFGFR2IIIcS252W, Fgfr2+/Neo-S252W, Ap, sFGFR2IIIcS252W Fgfr2+/Neo-S252W, EIIa-Cre sFGFR2IIIcS252W, and Ap/Sol mice. These mice were bred on a mixed background, and then genotyped by PCR analysis of genomic DNA prepared from tail tips with primers specific for Fgfr2, Fgfr2Neo-S252W, Fgfr2S252W, Cre, and sFGFR 2IIIcS252W−3xFLAG. The primers for Fgfr2, Fgfr2Neo-S252W, Fgfr2S252W, and sFGFR2IIIcS252W−3xFLAG have been described previously (Holmes et al., 2009; Suzuki et al., 2012) (Fig. 1A, B). The primers for Cre were 5′-CCTGTTTTGCACGTTCACCG-3′ and 5′-ATGCTTCTGTCCGTTTGCCG-3′ (290 bp) (Fig. 1B). All experiments were performed in accordance with protocols certified by the Institutional Animal Care and Use Committee of Tokyo Medical and Dental University (no. 0120272A).
Protein Preparation and Western Blot Analysis
Serum was collected from wild-type and sFGFR2IIIcS252W mice. FLAG-tagged sFGFR2IIIcS252W protein was purified using a FLAG M Purification Kit (Sigma, St. Louis, MO) according to the manufacturer's instructions. Samples were loaded onto a 10% SDS-polyacrylamide gel and transferred to a polyvinylidene difluoride membrane (Amersham Biosciences, Piscataway, NJ). The membrane was incubated with an anti-FLAG antibody (Sigma), and then probed with a horseradish peroxidase-conjugated anti-mouse IgG (Cell Signaling, Danvers, MA). Bound antibodies were detected using an ECL Plus Western Blotting Detection System (Amersham Biosciences) according to the manufacturer's instructions.
μCT Imaging Protocols of Mouse Skulls
All mouse samples were scanned at 45 kV, 200 μA, and 12 μm/voxel using a high-resolution X-ray micro-CT system (SMX-100CT; Shimadzu, Kyoto, Japan). BMD, calvaria, and IFS width were calculated from the raw data by three-dimensional image analysis software (TRI/3-D-BON; Ratoc, Tokyo, Japan). Portions were considered bones with a BMD of more than 120 mg/cm3. The position of the skull images was calibrated based on the skull base and midline. In coronal section images that showed the narrowest sagittal suture, ROIs were defined as 20 pixels in width and 30 pixels away from the tip of the parietal bone. The IFS width was measured in frontal view with a BMD threshold in 300 mg/cm3.
Tissues were fixed in 4% paraformaldehyde/PBS or 10% formalin and embedded in paraffin. Serial transverse sections (10 μm) were stained with hematoxylin and eosin (H&E). To visualize calcium deposits and acid mucosubstances, the IFS was stained with Alizarin red and Alcian blue, respectively.
Kruskal-Wallis and Dunn's post hoc tests were used to assess the differences in body weight, BMD, and IFS width. Chi-square and Fisher's exact tests were used to evaluate phenotypic differences. Chi-square tests were used to evaluate differences in the birth prevalence of each genotype from the expected ratio. A P value of <0.05 was considered statistically significant. Bonferroni corrections were applied for multiple comparisons.
We thank Dr. Shigetaka Kitajima (Tokyo Medical and Dental University) for providing EIIa-Cre mice and all of the members of our laboratories for assistance. This study was supported by Grants-in-Aid for Scientific Research (nos. 18209060 and 23390471) and the Global (21st Century) Center of Excellence (COE) Program, International Research Center for Molecular Science in Tooth and Bone Diseases from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.