Amelogenin expression in long bone and cartilage cells and in bone marrow progenitor cells



The amelogenin protein is considered as the major molecular marker of developing ectodermal enamel. Recent data suggest other roles for amelogenin beyond structural regulation of enamel mineral crystal growth. Here we describe our novel discovery of amelogenin expression in long bone cells, in cartilage cells, in cells of the epiphyseal growth plate, and in bone marrow stromal cells. Anat Rec, 2007. © 2007 Wiley-Liss, Inc.

The major enamel proteins produced by the ameloblast cells (a specialized cell layer of the enamel organ) are the amelogenins. They comprise 90% of the developing extracellular enamel matrix proteins (Termine et al., 1980) and play a major role in the biomineralization and structural organization of enamel (Robinson et al., 1998; Fincham et al., 1999). Amelogenins are hydrophobic molecules that self-assemble in vitro and in vivo into nanospheric structures, which regulate the oriented and elongated growth, shape, and size of the enamel mineral crystal (Fincham et al., 1994; Du et al., 2005). During enamel development and mineralization, the abundant secreted amelogenins in the extracellular enamel are sequentially and discretely degraded by specific proteases, the metalloprotease Enamelysin (MMP20) and the serine protease EMSP1 (KLK4) (Simmer and Hu, 2002). The amelogenins are eventually, together with other enamel matrix proteins, replaced by mineral ions calcium and phosphorus, the enamel finally becoming hard, fully mineralized (96%), and mature (Deutsch et al., 1995).

The amelogenin sequence is highly conserved throughout evolution (Sire et al., 2005). The gene contains 7 exons, which undergo alternative mRNA splicing. The most abundant isoform of the native protein secreted into the enamel matrix lacks the internal region encoded by exon 4 (Salido et al., 1992). Two additional exons downstream of exon 7 have been identified in a rare alternatively spliced RNA transcript of amelogenin (Li et al., 1998; Baba et al., 2002). The relatively large number of amelogenin alternatively spliced mRNA isoforms, code for several proteins with different structures and characteristics, which might have different functions (Veis, 2003).

Abnormal enamel formation and mineralization have been demonstrated by knockout of amelogenin expression. The teeth of the knockout (KO) mice expressed a hypoplastic enamel phenotype with reduced enamel thickness (Gibson et al., 2001). Interestingly, a progressive deterioration of cementum (a mineralized tissue covering the tooth root surface) was observed in the amelogenin KO mice. The defects in cementum were characterized by increased presence of osteoclasts. These defects were also associated with an increased expression of receptor activator of nuclear factor-κB ligand (RANKL) near the cementum, suggesting that amelogenin may play a key role in osteoclastogenesis through the RANKL/RANK-mediated pathway (Hatakeyama et al., 2003, 2006).

Although for over 4 decades, the amelogenins were thought to be tissue-specific and exclusively expressed by the ectodermal enamel-producing ameloblast cells, more recently, amelogenin has also been implicated in dentin matrix (Nebgen et al., 1999) and odontoblasts (Papagerakis et al., 2003), during cementogenesis, and in periodontal ligament (PDL) cells (Fong and Hammarstrom, 2000; Janones et al., 2005). Smaller amelogenin gene-spliced products, e.g., leucine-rich-amelogenin peptide (LRAP), have been suggested to be associated with cell signaling and to have chondrogenic and osteogenic potential (Veis et al., 2000; Viswanathan et al., 2003). Recently, it was shown that LRAP binds to lysosome-associated membrane protein 1 (LAMP-1), a transmembrane protein, which might serve as a cell surface receptor for LRAP (Tompkins et al., 2006).

The present study describes a novel finding of amelogenin expression in different long bone cells, in their precursors, in bone marrow stromal cells, which are mainly mesenchymal stem cells, in cartilage cells, and in specific cell layers of the epiphyseal growth plate. These tissues originate from the embryonic mesoderm and not from embryonic neural crest cells, which gives rise to the mesenchymal tissues of the calvaria and tooth attachment apparatus (cementum, periodontal ligament, and alveolar bone).


Tissue Preparation

All experiments were approved by Hadassah Medical School Animal Care Ethical Committee, Hebrew University (Jerusalem, Israel). Femurs and tibias were dissected from 5-week-old Sabra male rats weighting 100 g, and 10-week-old weighting 350 g (five rats from each group of age were used). Three 2-month-old mixed-type dogs were scarified by anesthesia with a mixture of ketamin and penthotal, and penthotal injection to the heart, immediately followed by the dissection of the dog epiphysis from femur, tibia, and patella. For immunohistochemistry, rat tissues were fixed in 4% paraformaldehyde (PFA) for 1 hr and dog tissues for 4 hr at room temperature. For in situ hybridization, both rat and dog tissues were incubated in 4% PFA for 24 hr at room temperature. The tissues were decalcified at 4°C for 2 months in 0.5 M EDTA, pH 7.4, replaced every 2 days, then washed in PBS and dehydrated. The tissues were embedded in paraffin and sectioned (5 μm).

Indirect Immunohistochemistry

Slides were deparaffinized, hydrated, rinsed in PBS, and endogenous peroxidase activity was blocked by 3% H2O2 (diluted in methanol) for 10 min. Slides were blocked in nonimmune goat serum for 15 min (Histostain-SP kit; Zymed Laboratories, San Francisco, CA), followed by overnight incubation in the first antibody (diluted in PBS) at 4°C in a humidified chamber. The first antibodies used were LF-108, a polyclonal rabbit amelogenin antibody, raised against a synthetic peptide corresponding to 10 amino acids at the C-terminus of human amelogenin (cross-reacts with dog and rat), conjugated to horseshoe crab hemocyanin (LPH), and affinity-purified using protein G (diluted to 1/1,000 for rat tissues and 1/2,000 for dog tissues); and 110BQ, a monoclonal mouse antibody raised against human amelogenin (diluted to 1/1,000) (Catalano-Sherman et al., 1994). After rinsing, slides were treated according to the Histostain-SP kit protocol, Picture plus kit or Super Picture kit (Zymed Laboratories). Negative controls included the corresponding preimmune mouse serum (diluted to 1/1,000), nonimmune rabbit serum (diluted to 1/1,000 for rat and 1/2,000 for dog tissues), and PBS replacing the first antibody. All slides were examined by Axioskop (Zeiss, Göttingen, Germany), and pictures were taken using Coolpix 990 digital camera (Nikon, Tokyo, Japan).

Colocalization of Amelogenin and Cell-Specific Markers Using Confocal Microscopy

Double immunofluorescence staining reactions using the mouse monoclonal antibody 110BQ (diluted to 1/200) and polyclonal antibodies (diluted to 1/200) against known cell-specific markers were performed. The cell-specific markers included human bone-sialoprotein antibody (LF-6) (Fisher et al., 1995), a marker for mesenchymal mineralizing tissues cells, and CD105, Endoglin (H-300; Santa Cruz Biotechnology), a marker for mesenchymal stem cells. Slides were deparaffinized, hydrated, rinsed, and incubated for 2 hr in ammonium chloride (75 mM) at 4°C. After rinsing, the slides were blocked for 20 min in 1% BSA and then incubated with a mixture of the two first antibodies (diluted in PBS) at 4°C overnight. The slides were then rinsed and incubated in the second antibodies (diluted in PBS to 10 μg/ml, according to the manufacturer's recommendations) for 2 hr at room temperature. The second fluorescent antibodies were Alexa Fluor 647 (red) goat antirabbit IgG (for the cell markers) and Alexa Fluor 488 (green) goat antimouse IgG (for amelogenin; Molecular Probes, Eugene, OR). The slides were rinsed, mounted, and examined by laser confocal microscopy (LSM 410; Zeiss).

In Situ Hybridization

For the production of antisense and sense (control) RNA probes, male Sabra rat enamel organ was dissected, the cells were lysed, and total cDNA was produced using Cells-to-cDNA II (Ambion, Austin, TX). The rat LRAP sequence was amplified by PCR (Table 1) and cloned into pGEM-T Easy Vector (Promega, Madison, WI). Antisense and sense RNA probes were generated using a DIG RNA labeling kit (SP6/T7; Roche Diagnostics, Mannheim, Germany). Keeping RNase-free conditions, slides were deparaffinized, hydrated, and treated according to the InnoGenex universal ISH kit protocol (InnoGenex, San Ramon, CA). The slides were incubated at 80°c for 5 min, then cooled to the hybridization temperature at 54°c for 12 hr.

Table 1. Rat and dog amelogenin primers
Sequence amplifiedForwardReverse
  1. According to: rat amelogenin sequence, gi: 9506380, dog amelogenin exon 6 sequence gi: 19110511.

Rat LRAP, enamel organ and bone-marrowExon 2 (25–47): 5′TGGATCTTGTTTGCCTGCCTCCT3′End of exon 6 (622–645): 5′ATCCACTTCTTCCCGCTTGGTCTT3′
Amelogenin, dog enamel organA) Rat exon 2 (25–47): 5′TGGATCTTGTTTGCCTGCCTCCT3′A) Dog exon 6 (325–347): 5′AAGCTTCCAGAGGCAGGTCAGGA3′
B) Dog, exon 2–3: 5′GCCTTCAGTATGCCTCT3′B) Rat exon 7 (712–734): 5′GATTGTAGGCACAAATCATTGTG3′
Amelogenin, dog bone marrowDog, exon 2–3: 5′GCCTTCAGTATGCCTCT3′Dog exon 6: 5′AGGTCAGGAAGCATGG3′

Isolation and Sequencing of Amelogenin mRNA From Long Bone Marrow and Enamel Organ

Dog mandibles were kept in solid CO2 immediately after their dissection. The mandibles were viewed by X-rays in order to identify the permanent unerupted teeth. Using a diamond disk, the mandibles were dissected and the unerupted teeth were carefully isolated. The isolation of the enamel organ from dog teeth and from rat teeth was performed using microsurgery tools under a stereodissecting microscope. Dog long bone marrow was aspirated from the hip bones, femurs, and tibias, and rat bone marrow was aspirated from the femurs of Sabra male rats (weighing 100 and 350 g). Dissected tissues were frozen immediately in solid CO2 and stored at −80°C. Rat RNA sample isolation and RT-PCR were performed using Cells-to-cDNA II (Ambion). Dog RNA sample isolation was performed using TRI-Reagent kit (Molecular Research Center, Cincinnati, OH). RT-PCR was performed using SuperScript II reverse transcriptase (Invitrogen). PCR reactions were performed using specific primers (Table 1). Since only part of exon 6 of the dog amelogenin sequence was previously published (gi: 19110511), the forward primer for PCR amplification was designed according to the rat amelogenin sequence (exon 2). The reverse primer was designed according to the published part of dog exon 6. The dog amelogenin sequence was completed (including the stop codon in exon 7) using a reverse primer designed according to the rat sequence (exon 7; Table 1, dog reverse primer B) and forward primer according to the sequence of dog exons 2–3 (Table 1, dog forward primer B). Amplification of amelogenin sequence from dog bone marrow was performed with primers designed according to the dog enamel organ sequence (Table 1). PCR products were extracted from 2% agarose gels using Qiaquick gel extraction kit (Qiagen, Valencia, CA) and sequenced at the Center for Genomic Technologies, Hebrew University (Jerusalem, Israel).

Sequencing of Amleogenin From mRNA Obtained From Cultured Human Bone Cells and Bone Marrow Stromal Cells

mRNA was isolated from cultured human bone cells and from human bone marrow stromal cells, as was previously described (Kuznetsov et al., 1997), and sent to us by Marian Young (National Institute of Dental and Craniofacial Research, National Institutes of Health). RT-PCR and PCR amplification were performed using primers designed according to the human amelogenin sequence (gi: 178528; Table 2), and PCR products were treated and sequenced as was described above.

Table 2. Human amelogenin primers
Sequence amplifiedForwardReverse
  1. Human amelogenin sequence, gi: 178528.

Amelogenin, human bone cells and stromal cellsHuman, exon 5–6 (197–220): 5′GAGCATAAGGCCACCGTACCCTTC3′Exon 6 (382–405): 5′GGATTGGAGTCATGGAGTGTTGGC3′

Isolation and Western Blot Analysis of Amelogenin Protein From Rat Bone Marrow and Cartilage

Articular cartilage from the tibias and femurs were dissected from Sabra male rats weighting 300 g, and bone marrow was aspirated from the same bones. Total protein was isolated according to Sambrook et al. (1989), separated on Tris-Glycine gel 4–20% (Novex, precast gels; Invitrogen Life Sciences, Paisly, UK), and transferred to nitrocellulose. Western blot analysis was performed according to Sambrook et al. (1989) using the polyclonal amelogenin antibody LF108 (diluted to 1/500).


We assumed that since amelogenin is expressed in tooth supporting tissue cells (cementoblasts and PDL cells) of mesenchymal origin, it might also be expressed in long bones, though they differ in embryonic origin and pattern of development. We therefore examined the expression, localization, and distribution of amelogenin in long bone tissues.

Amelogenin mRNA and protein were found to be expressed in rat (Fig. 1a–c) and dog (Fig. 2a and b) articular cartilage chondrocytes, epiphyseal bone cells, and differentially in specific cell layers of the epiphyseal growth plate. The epiphyseal growth plate is responsible for elongation and growth of long bones through endochondral ossification. Veis (2003) showed that LRAP, a specific amelogenin splice product, induced Cbfa1 and collagen type II expression (chondrogenic and osteogenic differentiation markers) in rat muscle fibroblasts in vitro, indicating a possible role for amelogenin in cartilage formation and/or in its transition into bone.

Figure 1.

Amelogenin expression in rat long bone: a) Immunohistochemical analysis of the head of tibia, using amelogenin monoclonal antibody (110BQ). Amelogenin protein expression (red) is demonstrated in articular cartilage (AC) chondrocytes, in cells lining bone trabecules and differentially in epiphyseal growth plate cells. No staining was observed in the control, pre-immune serum replacing the first antibody. b) In-situ hybridization (ISH) analysis on the epiphyseal plate, using rat amelogenin probe. Amelogenin mRNA expression (blue) was detected in specific cells of epiphyseal growth plate. No staining was observed in the control, sense probe. c) Immunohistochemical analysis of the epiphyseal bone, using amelogenin polyclonal antibody (LF108). Amelogenin protein expression was detected in specific bone marrow cells (BMc), osteoblasts (Ob) and osteoclasts (Ocl). No staining was observed in the control, non-immune serum replacing the first antibody. d) Co-localization (yellow) of amelogenin (green) and CD105 (mesenchymal stem cells marker- red) in long bone marrow. Bone marrow cell (BMc), Osteoblast (Ob) and Osteocyte (Oc) e) Coomasie blue staining of bone marrow protein extract and Western blot analysis using amelogenin polyclonal antibody (LF108), revealing expression of amelogenin polypeptides in the bone marrow. f) Rat LRAP amino acid sequences deduced from mRNA samples extracted from bone marrow (BM) and enamel organ (EO).

Figure 2.

Amelogenin expression in dog long bone: a) Immunohistochemical analysis of different regions in the dog epiphyseal growth plate, using amelogenin monoclonal antibody (110BQ). Amelogenin is differentially expressed in the epiphyseal growth plate cells. No staining was observed in the control, PBS replacing the first antibody. b) Immunohistochemical analysis of dog articular cartilage, using amelogenin monoclonal antibody (110BQ). Amelogenin was found to be expressed in part of the articular cartilage cells through the different stages of their differentiation. No staining was observed in the control, PBS replacing the first antibody. c) Immunohistochemical analysis of periosteum and bone, using amelogenin polyclonal antibody (LF108). Amelogenin expression was detected in periosteum cells (PO), osteoblsts (Ob) and osteocytes (Oc). No staining was observed in the control, non-immune serum replacing the first antibody. d) Co-localization (yellow) of amelogenin (green) and bone sialoprotein (BSP-mesencymal mineralized tissues cell marker-red) in epiphyseal trabecules. e) Co-localization (in yellow) of amelogenin (in green) and CD105 (in red) in bone marrow. f) Dog amelogenin amino acid sequences deduced from mRNA samples extracted from bone marrow (BM) and enamel organ (EO).

Amelogenin mRNA and protein were also expressed in long bone periosteum, which is composed of progenitor cells (shown for dog, Fig. 2c), in rat and dog osteoblasts and osteoclasts lining the bone trabecules and by some osteocytes (Figs. 1a–c and 2c and d). Some of these bone cells coexpress proteins specific to the formation of mesenchymal mineralized tissues, such as bone sialoprotein (BSP; Fig. 2d). This strengthens the finding that amelogenin is not only an ectodermal ameloblast-specific protein, but is also expressed in mesenchymal cells.

Amelogenin expression in both osteoblasts and osteoclasts at the mineralization front of long bones (Fig. 1c), and the reported increase in the number of osteoclasts due to lack of amelogenins in the amelogenin KO mouse, may suggest that amelogenin has a role in the cross-talk between these two cell types controlling the process of osteoclastogenesis. It might be that amelogenin inhibits differentiation toward osteoclasts, as was suggested by Hatakeyama et al. (2003, 2006) for osteoclasts in the tooth supporting tissues.

In long bone marrow, amelogenin was found to be expressed by distinct cells (Fig. 1c), some of them coexpress CD105, suggesting these are mesenchymal stem cells (Figs. 1d and 2e). Western blot analysis confirmed the expression of amelogenin polypeptides in rat long bone marrow (Fig. 1e). cDNA sequencing revealed the expression of the full-length dog amelogenin (excluding exon 4, which is the most abundant amelogenin isoform in ameloblasts) in bone marrow. This sequence was identical to the amelogenin sequence from dog enamel organ. This is the first time the full translated sequence of dog amelogenin is published (Fig. 2f). The LRAP splice product, thought to have intercellular signaling activity that induces ectopic bone formation (Veis et al., 2000), was also identified in rat bone marrow (Fig. 1f). Furthermore, sequencing of mRNA obtained from stromal cell samples from different human individuals confirmed that some of the pluripotent mesenchymal stem cells in the bone marrow (i.e., stromal cells) were indeed expressing amelogenin (Fig. 3). Partial human amelogenin sequences obtained from bone and stromal cells (Fig. 3) were identical to the corresponding reported human enamel organ amelogenin sequence (Salido et al., 1992).

Figure 3.

Human amelogenin partial amino acid sequences deduced from mRNA samples extracted from human bone cells samples (Bo) and bone marrow stromal cells samples (SC), compared to the reported human enamel organ amelogenin amino acid sequence.

Periodontal tissues regeneration could often be achieved by application of enamel matrix proteins (EMP) (Hammarstrom et al., 1997), a heterogeneous mixture of over 200 polypeptides, encoded by at least seven genes. About 90% of the EMP are amelogenins, and hence it was suggested that amelogenin is the inducer of periodontal tissues regeneration, including the alveolar bone, periodontal ligament, and cementum (Gestrelius et al., 2000). The process of bone regeneration requires recruitment of mesenchymal stem cells to the injured area and tight control over osteogenic and osteoclastogenic activity. Amelogenin expression in long bone marrow mesenchymal stem cells and its assumed ability to induce regeneration of alveolar bone may suggest a possible role for amelogenin in inducing mesenchymal stem cells recruitment during the processes of periodontal and long bone repair. Goldberg et al. (2006) showed that LRAP has the capability to recruit cells with the potential to proliferate and differentiate into osteoblast-like cells, further strengthening the assumption that amelogenin might induce mesenchymal stem cells recruitment.

Amelogenin expression in osteocytes, osteoblasts, osteoclasts, bone marrow cells, and cartilage cells, together with the accumulating data indicating amelogenin induction of osteogenesis (Veis et al., 2000) and inhibition of osteoclastogenesis (Hatakeyama et al., 2003), suggest that amelogenin has a crucial role in the processes of bone development and remodeling.


The authors thank M. Young for providing mRNA from human bone and stromal cells and L.W. Fisher for the LF108 and LF-6 antibodies.