Expression of FAM20C in the Osteogenesis and Odontogenesis of Mouse

Preliminary studies in our laboratory revealed that FAM20C was first detectable in the osseous tissues of CD-I mice at embryonic day 14.5 (E14.5), and its expression in the bone and teeth began to fade away after postnatal 7 weeks. In this report, we describe the findings obtained from analyses of CD-I mice (Harlan Laboratory; Houston, TX) at the developmental stages of E13.5, E14.5, E15.5, E16.5, newborn, and 1, 3, 5, 7, and 8 weeks after birth. These mice were used to systematically analyze the expression of FAM20C in the skeleton and tooth. The whole embryos at the stages of E13.5, E14.5, and E15.5 were processed for ISH and IHC analyses; the femurs and heads of the E16.5, newborn, and 1-week-old mice were dissected for specimen preparation, and the femurs and mandibles of the 3-, 5-, 7-, and 8-week-old mice were dissected and processed for the same analyses.

The acquired tissues were fixed with 4% paraformaldehyde in PBS solution at 4C overnight. The femurs and mandibles from the 1-, 3-, 5-, 7-, and 8-week-old mice were decalcified in 15% EDTA (pH 7.4) at 4C for 2, 4, 6, 8, and 9 days, respectively. The PBS and EDTA solutions were prepared using water pretreated with 0.1 % diethylpyrocarbonate. The tissues were processed for paraffin embedding, and serial 5-μm sections were prepared. The animal protocol was approved by the Baylor College of Dentistry Institutional Animal Use and Care Committee, Texas A&M Health Science Center (Dallas, TX).

In Situ Hybridization

With the full-length mouse FAM20C cDNA (Hao et al. 2007) as a template, a 900-bp fragment was obtained by PCR amplification using forward primer 5′-GCGGCCATGAAGATGATACT-3′ and reverse primer 5′-CTCCTGCTCTCTCGTCTGCT-3′. The PCR product was then subcloned into the pCRII-TOPO vector (Invitrogen; Carlsbad, CA), which was linearized with HindIII to synthesize the antisense RNA probes using the T7 RNA polymerase or with XhoI to synthesize the sense RNA probes using the Sp6 RNA polymerase. The probes were labeled with digoxigenin (DIG) using an RNA Labeling Kit (Roche; Indianapolis, IN). DIG-labeled RNA probes were detected by an enzyme-linked immunoassay using a specific anti-DIG-alkaline phosphatase antibody conjugate (Roche) and an improved substrate, which produces a red color for positive signals (Vector Laboratories; Burlingame, CA), according to the manufacturer's instructions. The sense probe was used in place of the antisense probe in the negative control experiments. A detailed description of the protocols of ISH and methyl green counterstaining can be found in our previous reports (Feng et al. 2002; Baba et al. 2004).

To generate polyclonal antibodies against FAM20C, we designed three oligopeptides: one from the N-terminal region and the others from the C-terminal region of the mouse FAM20C. These three oligopeptides were used to immunize rabbits to obtain polyclonal antibodies (YenZym Antibodies; South San Francisco, CA). Among the three anti-FAM20C polyclonal antibodies, the one generated using the oligopeptide CSSWEDDLATEH-RASTER from the C-terminal region of mouse FAM20C demonstrated a high titer and appropriate specificity and was used for the IHC and Western immunoblotting in this study. This rabbit antiserum was purified by an affinity column made of the oligopeptide CSSWEDDLATEH-RASTER. The titer of the affinity-purified antibody, revealed by ELISA using the immunizing oligopeptide as the antigen, was 1:50,000 (8 ng/ml). The specificity of this polyclonal antibody was confirmed by carefully designed control IHC experiments and Western immunoblotting (described later).

The IHC experiments were carried out using an ABC kit and a DAB kit (Vector Laboratories), according to the manufacturer's instructions. The anti-FAM20C antibody was used at a concentration of 1 μg IgG/ml. Two sets of control experiments were performed to confirm the specificity of the affinity-purified anti-FAM20C antibody. In the first set, rabbit IgG (Abcam; Cambridge, MA) at a concentration of 1 μg IgG/ml was used to replace the anti-FAM20C antibody in the IHC experiments. In the second set of control experiments, the anti-FAM20C antibody solution (1 μg IgG/ml) was preincubated overnight with the immunizing oligopeptide at an antibody/oligopeptide ratio of 1:50 before it was applied to the sections. In the IHC experiments, methyl green was used for counterstaining.

To examine if FAM20C localizations overlap with the mineralization areas, we performed von Kossa staining on E14.5 mouse head sections serial to those used for the IHC of FAM20C. For von Kossa staining, the paraffin sections of E14.5 mouse head (undecalcified) were de-paraffinized and hydrated with water. After being rinsed in several changes of distilled water, the sections were incubated with 1% silver nitrate solution under a 60-W light bulb for 2 hr. The sections were then rinsed in several changes of distilled water and incubated in 5% sodium thiosulfate for 5 min to remove the unreacted silver. Finally, the specimens were counterstained with nuclear fast red for 5 min.

The coding sequence of mouse FAM20C cDNA was sub-cloned into the bicistronic pMES vector (Swartz et al. 2001) in front of the internal ribosome entry site-enhanced green fluorescent protein sequence and downstream to a chicken β-actin promoter. A STOP cassette flanked by LoxP sequences was inserted between the β-actin promoter and the FAM20C sequence. This conditional transgenic construct was named “pMES-STOP-FAM20C.” To generate recombinant FAM20C, the pMES-STOP-FAM20C construct and PBS 185 vector (plasmid 11916; Addgene, Cambridge, MA) (Sauer and Henderson 1990) were co-transfected into human embryonic kidney 293-Epstein-Barr virus nuclear antigen (FIEK293-EBNA) cells. The Cre recombinase expressed by the PBS185 vector removed the floxed STOP cassette in pMES-STOP-FAM20C and thus initiated the transcription of FAM20C in the cells. Samples from non-transfected cells or cells transfected with only the pMES-STOP-FAM20C construct were used as negative controls. After transfection, the cells were cultured in conditional (serum-free) DMEM for 48 hr before both cells and medium were collected for analysis. The cell extract (lysate) and an equal portion of culture medium were loaded onto 12% SDS-PAGE for Western immunoblotting analyses. The FAM20C polyclonal antibody was used at a concentration of 400 ng/ml in the Western immunoblotting experiments. An anti-β-actin monoclonal antibody (clone AC-15; Sigma, St Louis, MO) was used at a concentration of 0.5 μg/ml to detect the mouse β-actin, which served as an internal control for this experiment.

To test the specificity of the anti-FAM20C antibody and to examine if FAM20C is a secretory protein, we performed Western immunoblotting with this antibody on the cell lysates and culture medium from HEK293-EBNA cells co-transfected with the pMES-STOP-FAM20C and the PBS185 constructs.

In the Western immunoblotting analyses, a protein band migrating at ~63 kDa (just below the 75-kDa molecular mass marker) was recognized by the anti-FAM20C antibody (Figure 1). This protein band was present in the cell lysates and in the culture medium from the cells co-transfected with the pMES-STOP-FAM20C and the PBS185 constructs, but was completely absent in the samples from the non-transfected cells or cells transfected with only the pMES-STOP-FAM20C construct. It is worth noting that the protein band from the culture medium had a remarkably greater intensity compared with the cell lysates, indicating that the majority of the FAM20C protein was secreted into the culture medium. The results of these experiments have not only confirmed the specificity of the anti-FAM20C antibody but also provided clear evidence that FAM20C is a secretory protein.

Neither FAM20C mRNA nor its protein was detected in any chondrogenic or osteogenic tissues at E13.5 (data not shown). At E14.5, FAM20C was undetectable in any long bones by either ISH or IHC; however, both FAM20C mRNA and protein were detected in the osteoblasts at the ossification sites of the craniofacial membrane bones (see later).

The mouse femur at E15.5 was used to illustrate the expression of FAM20C in the osseous and cartilaginous tissues. In the E15.5 embryo femur, FAM20C mRNA and protein were observed in the osteoblasts and cells that appeared to be osteoprogenitors, whereas the protein was also detected in the extracellular matrix (ECM) in the primary ossification center (POC) (Figures 2A-2D). The expression level of FAM20C at E16.5 relatively increased and was observed in broader areas in the osseous tissues (Figures 2E-2H) compared with its distribution at E15.5.

In the newborn (Figures 2I-2L) and 1-week-old mice (Figures 3A-3D), the expression of FAM20C could not be detected in the POC, but was found in the osteoblasts and newly formed osteocytes along or within the osteoid matrix of the trabecular bone and cortical bone.

The expression of FAM20C in the femur reached its maximal level at 3 weeks after birth (Figures 3E and 3H). At this stage, both the mRNA and protein for FAM20C were abundant in the articular chondrocytes, proliferative chondrocytes, and prehypertrophic and hypertrophic chondrocytes (Figures 3F and 31) and in the osteoblasts lining the trabecular bone and osteoid (Figures 3G and 3J). FAM20C protein was also detected in the mature osteocytes and in the ECM of the femur.

At 7 weeks after birth, FAM20C mRNA and its protein were observed in the osteoblasts lining the trabecular bone in the femur (Figures 3M and 3P), but the expression level of FAM20C was dramatically reduced (Figures 3K and 3N), when compared with earlier time points. At this stage, FAM20C mRNA was undetectable in the articular chondrocytes and mature osteocytes of the femur (Figure 3L), whereas its protein was observed in these cells and in the ECM of the bone (Figure 3O).

In this study, we systematically evaluated the expression of FAM20C in the molars and incisors of the mouse mandible. For reporting purposes, the results from the first lower molar and lower incisor, along with their surrounding periodontal ligament (PDL), were used to illustrate the expression of FAM20C during odontogenesis.

Neither FAM20C mRNA nor its protein was detected in any tissues of the mandible at E13.5 (data not shown). At E14.5, both FAM20C mRNA and protein were detected in the osteoblasts at the ossification sites of the craniofacial membrane bones. In addition, the localizations of FAM20C in these craniofacial bones overlapped with the mineralization areas as clarified by von Kossa staining (Figures 4A and 4B). When the first lower molar reached its early cap stage at E14.5, FAM20C mRNA and its protein were observed in the dental epithelium of the enamel organ (Figures 4C, 4D, 4F, and 4G) and in the osteoblasts of the alveolar bone (Figures 4E and 4H). FAM20C protein was also observed in the ECM of the alveolar bone osteoid, although the signals for the protein appeared diffused in these epithelial cells. At E15.5 (late cap stage) and E16.5 (bell stage), the FAM20C was seen in the ameloblast layer and the prospective odontoblast layer (data not shown).

The expression level of FAM20C in the odontoblasts and ameloblasts in the newborn mice (Figures 5A-5G) increased dramatically, when compared with the earlier stages. At this stage, FAM20C was observed in the newly formed dentin and enamel matrix (Figure 5F); it was clearly detected in the dentin matrix in the 1-week-old mice (Figure 6D). The tendency of increasingly intense FAM20C signals in the molar continued until 3 weeks after birth (data not shown). After postnatal 3 weeks, the expression of FAM20C in the first lower molar decreased with age (Figures 6E and 6G), and nearly faded away at 7 weeks (Figures 7E and 7F). However, the expression in the incisor and PDL fibroblasts remained at consistently high levels after postnatal 3 weeks (Figures 6E, 7C, 7E, and 7G).

FAM20C was also detected at a high level in the cerebrum cortex and cranial nerve ganglia in the E15.5 and E16.5 embryos (data not shown) and in the newborn mice (Figure 5A). These observations provide clues into the intracranial calcification phenotype in patients with lethal osteosclerotic bone dysplasia caused by mutations in the FAM20C gene.

In the negative control experiments of the ISH analyses, the sense probe did not detect appreciable signals (Figure 6F). In the first set of IHC control experiments, in which the anti-FAM20C antibody was replaced by normal rabbit IgG, no immunostaining signals were observed (data not shown). In the second set of control experiments, in which the anti-FAM20C antibody was preincubated with the immunizing oligopeptide, the immunostaining signals were negligible (Figure 6H).