Role of Hertwig's epithelial root sheath cells in tooth root development

Authors

  • Margarita Zeichner-David,

    Corresponding author
    1. Center for Craniofacial Molecular Biology, University of Southern California School of Dentistry, Los Angeles, California
    • Center for Craniofacial Molecular Biology, University of Southern California School of Dentistry, 2250 Alcazar Street, CSA 106, Los Angeles, CA 90033
    Search for more papers by this author
  • Keiji Oishi,

    1. Center for Craniofacial Molecular Biology, University of Southern California School of Dentistry, Los Angeles, California
    2. Tokushima University, Department of Periodontology and Endodontics, Tokushima, Japan
    Search for more papers by this author
  • Zhengyan Su,

    1. Center for Craniofacial Molecular Biology, University of Southern California School of Dentistry, Los Angeles, California
    Search for more papers by this author
  • Vassili Zakartchenko,

    1. Center for Craniofacial Molecular Biology, University of Southern California School of Dentistry, Los Angeles, California
    Search for more papers by this author
  • Li-Sha Chen,

    1. Center for Craniofacial Molecular Biology, University of Southern California School of Dentistry, Los Angeles, California
    Search for more papers by this author
  • Higinio Arzate,

    1. Universidad Nacional Autónoma de México, Facultad de Odontología, México DF, Mexico
    Search for more papers by this author
  • Pablo Bringas Jr.

    1. Center for Craniofacial Molecular Biology, University of Southern California School of Dentistry, Los Angeles, California
    Search for more papers by this author

Abstract

During tooth development, after the completion of crown formation, the apical mesenchyme forms the developing periodontium while the inner and outer enamel epithelia fuse below the level of the crown cervical margin to produce a bilayered epithelial sheath termed Hertwig's epithelial root sheath (HERS). The role of HERS cells in root formation is widely accepted; however, the precise function of these cells remains controversial. Functions suggested have ranged from structural (subdivide the dental ectomesenchymal tissues into dental papilla and dental follicle), regulators of timing of root development, inducers of mesenchymal cell differentiation into odontoblasts and cementoblasts, to cementoblast cell precursors. The characterization of the HERS phenotype has been hindered by the small amount of tissue present at a given time during root formation. In this study, we report the establishment of an immortal HERS-derived cell line that can be maintained in culture and then induced to differentiate in vitro. Characterization of the HERS phenotype using reverse transcriptase-polymerase chain reaction and Western blot immunostaining suggests that HERS cells initially synthesize and secrete some enamel-related proteins such as ameloblastin, and then these cells appear to change their morphology and produce a mineralized extracellular matrix resembling acellular cementum. These studies suggest that the acellular and cellular cementum are synthesized by two different types of cells, the first one by HERS-derived cementoblasts and the later by neural crest-derived cementoblasts. Developmental Dynamics 228:651–663, 2003. © 2003 Wiley-Liss, Inc.

INTRODUCTION

During tooth development, after the completion of crown formation, the apical mesenchyme continues to proliferate to form the developing periodontium while the inner and outer enamel epithelia (devoid of stellate reticulum and stratum intermedium), fuse below the level of crown cervical enamel to produce a bilayered epithelial sheath termed Hertwig's epithelial root sheath (HERS). The classic theory of root formation states that, as these cells divide, there is apical migration of HERS cells through the underlying dental ectomesenchymal tissues (dividing them into dental papilla and dental follicle). As the root develops, the first radicular mantle dentin is formed, the epithelial sheath fenestrate, and individual cells migrate away from the root into the region of the future periodontal ligament to form the rests of Malassez (Wentz et al., 1950).

It is well accepted that HERS plays an important role in root development; however, the precise nature of this role remains unclear. Amongst the different functions attributed to these cells are that of inducers and regulators of root formation, including the size, shape, and number of roots (Ten Cate, 1996). It has also been suggested that HERS cells deposit chemotactic proteins into the basement membrane to direct the migration of precementoblast cells (Owens, 1978; Thomas and Kollar, 1989; MacNeil and Thomas, 1993). Another role attributed to HERS cells is that of inducers of differentiation of odontoblast cells to form root dentine (Selvig, 1963, Ten Cate, 1978) or dental sac cells to differentiate into cementoblasts (Paynter and Pudy, 1958).

It is thought that after the uncoupling of the epithelial sheath, HERS cells migrate to the periodontal ligament where they re-associate to form the rests of Malassez, and these cells can influence cementum repair (Spouge, 1980). However, studies by Wesselink and Beertsen (1993) disrupting the homeostasis of the periodontal ligament with 1-hydroxyethylidene-1,1-bisphosphonate (HEBP), suggest that the rests of Malassez do not participate in periodontal ligament repair or maintenance as previously believed. Ultrastructural and immunohistochemical studies have shown that not all HERS cells migrate away from the forming root, but some cells remain on the developing root surface (Thomas, 1995; Kaneko et al., 1999) and perhaps serve to regulate the amount of acellular cementum deposited (MacNeil and Thomas, 1993). During second stage cementogenesis (when the tooth reaches occlusion and cellular cementum is formed), proliferation of HERS is considerably reduced, some of these cells are entrapped in the forming mineral and may influence phenotypic changes in the dental sac cells (Thomas, 1995). One more possibility to consider is that HERS cells undergo an epithelial–mesenchymal transformation (EMT) and become functional cementoblasts as proposed by Thomas (1995). EMT has been demonstrated to be a common event in several developmental processes (reviewed by Hay, 1995).

One of the major problems encountered while trying to determine the phenotype of HERS and understand their role in root development has been the very limited amount of tissue available. To solve this problem, we developed an immortal HERS cell line that allows us to grow these cells in sufficient quantities for biochemical characterization and manipulation. To achieve this goal, we took advantage of transgenic mice harboring the SV40 strain tsA58 early region coding sequences under the control of the mouse major histocompatibility complex H-2Kb class I promoter (Jat et al., 1991). This transgenic mouse (H-2Kb-tsA58), also known as Immortomouse, has the potential of producing cell lines that behave like immortal cells while in the presence of γ-interferon (IFN) at permissive temperatures and then differentiate by removing the inducer and being maintained at nonpermissive temperatures. In this study, we report the phenotypical characterization of HERS cells derived from the Immortomouse, which suggests that HERS cells maintained in vitro produce a mineralized extracellular matrix resembling cementum.

RESULTS

Establishment of Immortomouse-Derived HERS Cell Cultures

The presence of HERS in Immortomouse mandibular first molars is apparent at 7 days of postnatal development as can be seen in the histologic preparation shown in Figure 1A,B. The epithelial tissue was isolated and plated in tissue culture dishes and determined to be pure HERS based on cell size and shape. HERS can be seen as a double layer of cuboidal cells (Fig. 1E,F) easily differentiated from the ameloblast cell layer, which shows a single layer of elongated, polarized epithelial cells (Fig. 1C,D). After plating, cell proliferation was very slow taking weeks for the cells to reach confluence. Cells were then grown for several passes in KGM medium to eliminate any possible contamination with mesenchymal cells, which do not proliferate in this medium. Immortomouse-derived periodontal ligament (PDL) and dental pulp mesenchyme (DPM) placed in KGM did not proliferate and died after a few days (data not shown).

Figure 1.

Preparation of Hertwig's epithelial root sheath (HERS). HERS formation can be seen in histologic sections of 7 days postnatal mice first mandibular molars as can be seen in A (4×) or at higher magnification in B (20×). The isolated tissue placed in culture can be seen in E and the purity of the dissections can be seen in an H&E stained representative sample in F. Differences between dissected HERS (E,F) and ameloblast cells (C,D) can be clearly seen. Abbreviations: i, incisor; p, pulp; od, odontoblasts; d, dentin; am, ameloblasts; df, dental follicle; e, enamel; b, bone.

Because the cells harbor the temperature-sensitive mutant of SV-40 large T antigen under control of the H-2Kb promoter, the cells expressed the SV40 gene only under permissive conditions, and under these conditions, cells grew with a doubling time of 23.3 hr and did not stop dividing even after reaching confluence at a density of 6.8 × 104 cells/cm2 (Fig. 2A). Under nonpermissive conditions, cell growth was contact-inhibited at a density of 2.5 × 104 cells/cm2. Cells grown under permissive conditions until confluence and then changed to differentiation conditions also stop dividing (Fig. 2B). The typical cuboidal appearance characteristic of epithelial cells can be seen when HERS cells were grown under permissive conditions (Fig. 2C). When cells were grown under differentiation conditions for up to 30 days in culture, the cells appeared to change their morphology to a more fibroblastic type and arranged themselves around “puddles” of what appeared to be an extracellular matrix (Fig. 2D).

Figure 2.

Growth rate of Hertwig's epithelial root sheath (HERS) cells. Under permissive conditions (33°C + γ-IFN), cells continue to divide and do not stop dividing even after reaching confluence at a density of 6.8 × 104 cells/cm2. Under nonpermissive conditions (39°C) cell growth was contact-inhibited at a density of 2.5 × 104 cells/cm2. The typical cuboidal appearance characteristic of epithelial cells can be seen when HERS cells were grown under permissive conditions in KGM for a several passes and then maintained in DMEM (C). When cells were grown under differentiation conditions for up to 30 days in culture, the cells formed “puddles” of extracellular matrix (D).

To determine the nature of the extracellular matrix produced by the HERS cells, we first analyzed whether these nodules were mineralized. Cells were grown under differentiation conditions for different days in culture, and the presence of calcium deposits was determined by using Von Kossa staining. Von Kossa–positive grains appeared as early as 15 days in culture and increased considerably as the cells remained in culture and appear to be associated with the presence of the nodules, as can be seen in Figure 3D,E. To eliminate the possibility that the mineralization was due to the high concentration of glycerophosphate (10 mM) as reported by Chung et al. (1992), cells were incubated in differentiation medium in the presence of lower concentrations (2 mM) or no glycerophosphate was added. This strategy gave the same results, although the cells differentiated a little bit slower (data not shown). These results indicate that the extracellular matrix deposited by the cells is mineralized regardless of the concentration of glycerophosphate. The HERS mineralized matrix was also seen macroscopically in the culture flasks as a crystalline-appearing deposit strongly attached to the plastic.

Figure 3.

Characterization of the Hertwig's epithelial root sheath (HERS) -produced extracellular matrix (ECM). The ECM produced by HERS cells maintained in culture for 30 days can be seen in D and stained positive for Von Kossa (E). This ECM was compared with the acellular cementum from a section of a 16-day postnatal mouse mandibular first molar (A). The area marked by the rectangle can be seen at higher magnification (40×) in B. Transmission electron microscopy analysis of the mineral present in the area marked by the arrow can be seen in C. The mineral deposited by the HERS cells can be seen in F. Scale bar = 100 nm. b, bone; p, pulp; pdl, periodontal ligament; od, odontoblasts; d, dentin; ac, acellular cementum; n, nodule.

Further characterization of the HERS-deposited extracellular matrix (ECM) was done using transmission electron microscopy (TEM). Cells were grown for 30 days in culture, fixed, and embedded in Epon. Ultrathin sections were prepared and analyzed with a JEOL 1200 electron microscope and compared with the in vivo formed acellular cementum from a section of a 16-day postnatal mouse developing root (Fig. 3A,B). Both samples showed the presence of short and disorganized crystals grouping together in clusters, although the sample from the in vivo acellular cementum (Fig. 3C) appeared to be more dense than the sample obtained from the HERS cells in vitro (Fig. 3F).

Characterization of the Immortomouse-Derived HERS Cells Phenotype

Initial characterization was done by reverse transcriptase-polymerase chain reaction (RT-PCR) using primers for specific genes associated with mineralized tissues as shown in Table 1. First, we tested the expression of SV40, and as shown in Figure 4, these transcripts were present in cells maintained under permissive conditions and after 1 day under differentiation conditions, they were no longer present. Because HERS is derived from the joining of the inner and outer enamel epithelia, we analyzed the expression of enamel-associated proteins. Our results indicate that HERS cells do not express amelogenin or enamelin transcripts, whereas ameloblastin transcripts were clearly seen from 0 to 5 days in culture and faint bands were apparent up to 20 days in culture. The expression of ameloblastin by HERS cells in vitro was confirmed in vivo using in situ hybridization as can be seen in Figure 5 where ameloblastin is only expressed by ameloblast cells and proliferating HERS cells in the cervical region (Fig. 5D) and in the furcae region (Fig. 5E). These results suggest that once the inner and outer enamel epithelia get together to form HERS, the expression of amelogenin and enamelin is down-regulated but they continue to express ameloblastin, which can be used as a marker to localize HERS cells during root formation.

Table 1. Sequence of Primers Used for RT-PCRa
GeneSequenceTm
  • a

    BSP, bone sialoprotein; DMP-1, dentin matrix protein-1; DSPP, dentin sialophosphoprotein; EGF, epidermal growth factor; LEF-1, lymphoid enhancer factor-1; S100, Ca-binding protein; TGF, transforming growth factor.

β-ActinCGGTTGGCCTTAGGGTTCAGGGGG50
 CATCGTGGGCCGCTCAGGCACCA 
AlkalineCCTGACCAAAAACCTCAAAGGC58
PhosphataseACATTTTCCCGTTCACCGTCC 
SV40 T LargeGCTTGGCTACACTGTTTGTTGC60
AntigenTGTTTCATGCCCTGAGTCTTCC 
AmelobiastinACTTAGATCTATGTCAGCATCTAAGATTCCA58
 ACTTCCGCGGTCAGGGCTCTTGGAAACGCCA 
AmelogeninATGGGGACCTGGATTTTGTTT50
 CTCATAGCTTAAGTTGATATAACC 
BSPACGCCACACTTTCCACACTCTC58
 GTTCCTTCTGCACCTGCTTCAG 
E-CadherinAACCGATTCAAGAAGCTGGCG58
 ATGTTGCTGTCCCCAAGTTTGG 
CollagenGGACTTCTACAGAGCTGACC58
type I (α2)TTCAACATCGTTGGAACCCTGC 
CollagenTTAATGGACAAATAGAGAGTCT56
type IIIAATGTCATAGGGTGCGATAT 
DMP-1CAGGTCGGAAGAATCTAAAGG50
 TCTCAGTAACTGTCAGGTTGG 
DSPPAAAACACCGCTGCAACTACTGG54
 ATCACAACCTCGATGAGTGGG 
EGFTGGTTTGTGGTCCTAGAGAAACACCT58
 CCTCTGTCACTTGATGGTGGAATC 
EGF ReceptorGTGGAGAAATGGAACATCCTG52
 CCATAGGTACAGTTGGCGTG 
EnamelinGAAAGAACAGTCACTCCTACC56
 GAAGTGTAGTCTTGTAGAGCC 
FibromodulinAAATCACACACTTACCCCC52
 TACAACTGCTTCTTGCCACTCG 
Integrin α5ATCCTGTACGTGAAATCCCTGC66
 TGGATAAACTGAGACTGCTGGG 
Integrin α6GCCCAAGGAGATTAGCAATGG54
 TATCGGGGAATGCTGTCATCG 
KeratinTGACTCTGGCTAAGACTGACC52
 TCTTGTATTCCTGGTTCTGGC 
mLEF1TCAGCCTGTTTATCCCATCACG50
 GAAGGAGCTTCTCTTACCACC 
LumicanCAATGAACTGGCTGATAGTGG54
 GGAACATACACATGACAGGGG 
OsteocalcinTCTCTCTGACCTCACAGATCC50
 AGAGTTTGGCTTTAGGGCAGC 
OsteopontinTCCCTCGATGTTCATCCCTGTTG58
 ACTAGCTTGTCCTTGTGGCTGTG 
OsteonectinGATCCATGAGAATGAGAAGCG50
 CTATGTGAGCACCTTATCCCC 
mS100a3AAGGAGTTGTTGCAGAAGGAGC52
 CTGAGCAGAGATGTTCTTCCC 
TGF-β1GCGGACTACTATGCTAAAGAGG56
 GTTGTGTTGGTTGTAGAGGGC 
TGF-β2TTGTGAAAACCAGAGCGGAGG58
 TGGCTTTCCCAAGGACTTTAGC 
TGF-β3CTGGACACCAATTACTGCTTCC56
 GAGAACCAATTCTGACCTCTGC 
VimentinTGCGAGAGAAATTGCAGGAGG62
 GCAGTAAAGGCACTTGAAAGC 
Figure 4.

Expression of transcripts for mineralized tissues-associated proteins. Hertwig's epithelial root sheath (HERS) mRNA was extracted at different days in culture under differentiation conditions and transcripts for SV40 and the enamel-associated proteins amelogenin, ameloblastin, and enamelin were determined by using RT-PCR. Bone-associated proteins like Collagen type I (Col. Type I), collagen type III (Col. Type III), osteopontin, osteonectin, osteocalcin, bone sialoprotein, and alkaline phosphatase (ALP) were also determined. Dentin-associated proteins (DMP-1= dentin matrix protein-1 and DSPP= dentin sialophosphoprotein) and cementum-associated proteins like lumican and fibromodulin were determined using the specific primers shown in Table 1.

Figure 5.

Ameloblastin expression in vivo. In situ hybridization of 8 days postnatal mice half-mandibles with digoxigenin-labeled ameloblastin probes. Ameloblastin expression can be seen in the ameloblast layer of the developing first and second molars as well as the incisor (A) and in the proliferating Hertwig's epithelial root sheath (HERS) in the cervical area (C,D), as well as in the forming furcae (C and E). B: No label can be seen using the sense probe. e, enamel; d, dentin; am, ameloblasts; p, pulp; od, odontoblasts.

Screening for other genes associated with dentin, bone, or cementum formation was also first done by RT-PCR. The results shown in Figure 4 indicate that HERS cells express transcripts for collagen type I and III at all stages, whereas the expression of transcripts for osteopontin, osteonectin, osteocalcin, bone sialoprotein, and alkaline phosphatase (ALP) appear to increase as the cells remain in culture for longer periods. The expression of transcripts for dentin matrix protein 1 and lumican were evident only after the cells have been in culture for 10 days, and the expression of fibromodulin appeared to be stronger at the initial stages of culture while disappearing by 25 days in culture. No expression of dentin sialophosphoprotein (DSPP) transcripts were detected at any time tested, thus suggesting that the ECM deposited by HERS cells is not dentin.

To correlate the presence of transcripts with the expression of proteins, cells were grown in culture under differentiation conditions and the proteins secreted in the medium or extracted with ethylenediaminetetraacetic acid (EDTA) from the ECM after different days in culture were collected and subjected to polyacrylamide gel electrophoresis (PAGE). Staining of the medium proteins with Coomassie blue or Stains-All indicates changes in the protein profile expressed at different days in culture, particularly the appearance of proteins in the 70- and 35-kDa range (Fig. 6). Almost no proteins were detected using Stains-All, indicating that these proteins were not highly phosphorylated. Western blot analysis demonstrated the presence of ameloblastin (∼52 kDa) at 0 and 1 days of culture and disappearing thereafter. The presence of the cementum-associated protein cementum attachment protein (CAP) was analyzed in the medium, and the data show that HERS cells synthesize and secrete this protein as a 56-kDa protein secreted into the medium at the initial stages of culture, slowly disappearing, and no longer noticeable by 20 days in culture. No amelogenins were detected at any stage of cell differentiation in vitro (data not shown).

Figure 6.

Expression of mineralized tissue-associated proteins. Proteins secreted by Hertwig's epithelial root sheath (HERS) cells at different days in culture (0, 1, 5, 10, 15, 20, and 25 days) were fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and either stained with Coomassie blue (CB), Stains-All (SA), or transferred to a polyvinylidene difluoride membrane and incubated with antibodies against ameloblastin (AMBN) or cementum attachment protein (CAP). Extracellular matrix proteins deposited by HERS cells were extracted with ethylenediaminetetraacetic acid (EDTA), processed in a similar manner, and incubated with antibodies against bone sialophosphoprotein (BSP) or osteopontin (OPN). A: The presence of alkaline phosphatase (ALP) at different days of HERS culture under differentiation conditions was determined by its activity. The data are expressed as the means ± SD of three samples. B: The presence of osteocalcin (OCN) was determined by using a radioimmunoassay. The data are expressed as the means ± SD of three samples.

When ECM-proteins extracted with EDTA were analyzed (Fig. 6), the Coomassie blue-stained profile showed very few proteins present at the initial stages of culture. An increase in the presence of proteins in the 50- to 70-kDa area was seen after the cells have been in culture for 15 days. A similar pattern was seen with Stains-All, thus suggesting that several of these proteins are phosphorylated. Western blot immunostaining identified some of these proteins as bone sialoprotein (BSP, ∼80 kDa) and osteopontin (∼55–60 kDa). These proteins were also present in the medium (data not shown), and the incidence of these proteins closely follows the pattern observed for their mRNA expression. The presence of ALP was determined by measuring its activity (Fig. 6A), which started to increase after 6 days in culture, reaching its maximum activity after 24 days in culture. The presence of osteocalcin was determined using a radioimmunoassay, and the results shown in Figure 6B showed that osteocalcin is detected after 15 days in culture and continues to increase thereafter, correlating with the data obtained for the mRNA expression. No ameloblastin, amelogenin, or DSPP were present in the HERS extracellular matrix EDTA extract (data not shown).

Taken all together, the data presented indicate that HERS cells deposit a mineralized extracellular matrix, different than enamel because it does not contain amelogenin and enamelin, different than dentin because it does not contain DSPP but perhaps resembling cementum (presence of CAP) or bone. These results suggest the possibility that HERS might undergo EMT. To further examine this possibility, the expression of genes that have been associated with the EMT process in other developmental systems was analyzed at the mRNA level using RT-PCR (Fig. 7). The expression of mRNA for epithelial proteins like keratin was observed from the initial stages up until 21 days in culture, while expression of the mesenchymal associated vimentin was seen at all times. No changes in mRNA expression were detected for epidermal growth factor (EGF) and its receptor, transforming growth factor (TGF) β1, TGFβ2, TGFβ3, and Ca-binding protein (S100). Expression of the “epithelial” α6 integrin was strongly expressed at the initial stages of culture and disappeared after 21 days, as did the “mesenchymal” α5 integrin. The only changes in mRNA expression were associated with the transcription factor lymphoid enhancer factor-1 (LEF-1), which was not expressed in the early stages of culture and then was strongly expressed after 21 days in culture.

Figure 7.

Expression of epithelial, mesenchymal, and epithelial–mesenchymal transformation (EMT) -associated proteins. Hertwig's epithelial root sheath (HERS) mRNA was extracted at different days in culture under differentiation conditions, and transcripts for several different intermediate filament proteins, adhesion proteins, growth factors, and transcription factors associated with EMT were analyzed by reverse transcriptase-polymerase chain reaction (RT-PCR) using the specific primers shown in Table 1. Membrane and intermediate filament proteins synthesized by HERS cells at different days in culture were fractionated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, transferred to a polyvinylidene difluoride membrane, and incubated with antibodies against keratin, vimentin, E-cadherin, OB-cadherin, zonula occludens protein (ZO1), epithelial–membrane protein (MEM), cementum attachment protein (CAP), and α5-integrin. EGF, epidermal growth factor; EGFR, EGF receptor; TGF, transforming growth factor; LEF-1, lymphoid enhancer factor-1; S100, Ca-binding protein.

Analysis at the protein level was done by using HERS intermediate filament- or membrane-extracted proteins obtained at different times in culture (Fig. 7). In the intermediate filaments extract, immunoreactivity to keratins (52–58 kDa) appears to be more prevalent after 5 days in culture and then it almost disappears up to 25 days in culture when it can be seen again. Immunoreactivity against vimentin appears after 15 days in culture. In the membrane-extracted proteins, E-cadherin, the adhesion molecule associated with epithelial cells, was detected as a 120-kDa protein after 5 days in culture and its presence increased thereafter. OB-cadherin, the mesenchymal-associated adhesion molecule, was present as a ∼115 kDa protein at all stages of culture. The epithelial-associated protein zonula occludens protein (ZO1), which is mainly present in tight junctions of epithelial cells, was present as a 150-kDa protein at the initial stages of culture and disappeared thereafter. Another epithelial-associated protein, epithelial–membrane protein (MEM), was equally distributed at all stages of culture as an ∼125-kDa protein. CAP was also found as a 56-kDa membrane-associated protein with concentrations stronger at the initial stages of culture. The attachment protein α5- integrin was also present at all stages as a 120- to 130-kDa protein.

DISCUSSION

The availability of cell lines has been central to the study of developmental, cellular, and molecular biology, ideally; these cell lines should be able to proliferate easily in culture while preserving their functional phenotype. While many immortal cell lines have been produced by isolating the cells and immortalizing them in vitro, we took advantage of the Immortomouse. Several H-2Kb-tsA58 mice cell lines derived from tissues such as skin fibroblasts (Jat et al., 1991), osteoclast cells (Chambers et al., 1993), colon and small intestine epithelial cells (Whitehead et al., 1993), mesenchymal progenitor cells from bone marrow (Dennis and Caplan, 1996), and several epithelial cell lines from cystic fibrosis mice (Takacs-Jarrett et al., 2001) have been created. In the present study, we report the establishment of a HERS cell line derived from the Immortomouse.

The HERS cells were subcloned by stringent selection (culturing in different growth medium favoring epithelial over mesenchymal growth). At present, this method is the only available technique to purify this subpopulation of cells as there are no definitive markers of HERS cells with the exception of ameloblastin and the transcription factor Dlx-2, which marks early HERS cells (Lezot et al., 2000). However, we are very confident that HERS cells were selected and the purity and identity of these cells was initially established by morphologic and biochemical criteria. Morphologically, the founder cells have the typical appearance of epithelial cells and were easily distinguishable from the ameloblast and preameloblast cell layers, and contamination with these cells was ruled out by the lack of expression of amelogenin, enamelin, and DSPP mRNAs. Contamination with mesenchymal cells (pulp, odontoblast, periodontal ligament, dental follicle, or osteoblast cells), if any, was eliminated by growing the cells in KGM medium for several passes to select for epithelial cells, because mesenchymal cells do not survive under these conditions (Tsao et al., 1982; Varani et al., 1990). This lack of contamination was confirmed when Immortomouse-derived mesenchymal cells like PDL and pulp cells were placed in KGM, because these cells did not proliferate and eventually died after a few days. Of interest, although the Immortomouse-derived HERS cells can proliferate indefinitely while under the control of the T antigen, in its absence, the biological clock that limits the number of mitotic cycles continues to function normally and the cells maintain their senescence as if they were normal cells. A similar finding has been reported for Immortomouse-derived fibroblast cells (Ikram et al., 1994).

HERS Express Enamel Proteins

Years ago, it was proposed that HERS-derived products were related to enamel molecules and they function to initiate acellular cementum formation (Slavkin and Boyd, 1974; Slavkin et al., 1988, 1989). The presence of enamel proteins in root development has received considerable attention, because of the possibility that they might have other functions, particularly amelogenin, besides that of enamel formation (Zeichner-David, 2001). In this study, using RT-PCR and Western blot immunostaining, we showed that HERS cells in vitro do not synthesize amelogenin or enamelin but do synthesize ameloblastin. The expression of ameloblastin (amelin) by HERS cells was validated in vivo using in situ hybridization confirming previous studies by Fong et al. (1996) and supporting the identity of our cell line as HERS. Furthermore, ameloblastin represents an excellent marker to follow HERS cells during apical migration and root formation. The function of ameloblastin in enamel as well as in root formation remains unknown. The absence of amelogenin expression is more intriguing since Hamamoto et al. (1996) reported the capacity of HERS cells to produce enamel and express amelogenins in response to pulp inflammation. However, this finding was achieved under pathologic conditions, and these were not HERS cells per se but rather rests of Malassez, which although thought to be derived from HERS, their function and origin has not been fully demonstrated. In this study, we showed that HERS cells do not synthesize amelogenin, thus supporting previous in vivo studies (Luo et al., 1991; Fong et al., 1996; Diekwisch, 2001).

HERS Produce a Cementum-Like Mineralized Extracellular Matrix

Our data showed that after several days in culture, under differentiation conditions, HERS cells change their morphology and produce Von Kossa-positive nodules coincidental with the expression of dentin matrix protein-1, bone sialoprotein, osteopontin, osteocalcin, and high levels of ALP activity. Although dentin, bone, and cementum contain these proteins, the absence of DSPP ruled out the nature of this mineralized matrix as dentin. TEM comparison of the mineralized extracellular matrix deposited by HERS cells in vitro with in vivo deposited acellular cementum suggests that this ECM might be cementum. A precise characterization of the mineralized matrix as cementum is difficult, because the presence of cementum-specific proteins remains questionable. Although some putative cementum-specific proteins have been invoked such as the 56-kDa CAP (McAllister et al., 1990; Pitaru et al., 1995; Wu et al., 1996), a mitogenic factor (Nakae et al., 1991) and a 72-kDa protein (CEM-1, Slavkin et al., 1988), characterization of these proteins has been sparse and no particular cell type has been associated with their expression. In the present study, we showed that HERS cells synthesize a secreted as well as membrane-bound protein cross-reactive with a polyclonal antibody against CAP, thus supporting the idea that HERS cells are capable of producing cementum. Moreover, adhesion molecules that have been found in cementum such as lumican, fibromodulin, and α5 integrin (Steffensen et al., 1992; Cheng et al., 1996; Ivanovski et al., 1999; Grzesik et al., 2000) are also expressed by differentiated HERS cells in vitro.

Our HERS cell line showed a very high activity of ALP, several-fold higher than any other of our mineralized extracellular matrix-producing cell lines tested (data not shown). This finding provides further support that the ECM deposited by these cells is acellular cementum, since it has been reported that isolated “cementoblasts” in vitro do not synthesize ALP (Gao et al., 1999). Although the in vivo expression of ALP by HERS has not been analyzed, ALP-deficient animals present delayed tooth eruption (2 to 3 days), delayed onset of mineralization of the roots dentin, and a defective formation of acellular cementum along the molar roots (thin and irregularly shaped patches around the bases of the periodontal ligament fibers). Alveolar bone, periodontal ligament, and cellular cementum are unaffected (Beertsen et al., 1999), thus indicating that ALP is a very important component of acellular cementum and suggesting an important chemical difference between acellular and cellular cementum and the cells that deposit these ECMs.

Do HERS Undergo EMT?

The initial epithelial characteristic of HERS cells, the synthesis of enamel proteins followed by morphologic and phenotypical changes together with the expression of transcripts for several cementum-associated proteins, support the idea that HERS cells undergo EMT to become functional cementoblasts (Thomas, 1995). This possibility was further analyzed by determining the expression of markers associated with EMT and/or epithelial versus mesenchymal cells.

The expression of keratin by HERS in vivo has been demonstrated by several investigators (Alatli et al., 1996; Kaneko et al., 1999; Onishi et al., 1999), and as expected, HERS in vitro express keratin. Unexpected was the simultaneous expression of vimentin mRNA, although coexpression of vimentin and keratin has been reported for other epithelial cell lines maintained in culture (Zuk et al., 1989). It has been postulated that down-regulation of E-cadherin is necessary for EMT and is accompanied by down-regulation of SAMs like α6β1 and α6β4 integrins while the synthesis of α5β1 integrin is up-regulated (Hay, 1995). The expression of E-cadherin by HERS cells in vitro was expected since expression of this protein by HERS in vivo has been demonstrated (Terling et al., 1998; Obara et al., 1999). According to this information, we expected that HERS will show high expression of E-cadherin at early stages of culture and then be down-regulated at later stages of culture; however, our data showed the opposite: higher levels of E-cadherin as HERS remained in culture. Both, α5 and α6 integrins were expressed simultaneously at all stages of culture. This finding was also true for other mesenchymal or epithelial markers with the exception of ZO1 (epithelial marker), which was down-regulated as the cells remained in culture, and the transcription factor LEF-1, which has been associated with EMT as a regulator of E-cadherin expression (Huber et al., 1996; Behrens et al., 1996) and was expressed just at later stages of HERS cultures.

Although some changes of epithelial and mesenchymal markers were present, no clear changes in the expression of EMT-associated factors were seen. These results could suggest that our cultures consist of a mixed population of both epithelial and mesenchymal cells. Although possible, this prospect is highly unlikely given that the cells were maintained in KGM for several passes and other mesenchymal cell lines grown in this medium did not proliferate and died. Other explanations are that EMT does not take place, other factors are responsible for EMT of HERS cells, or that these cells maintain both epithelial and mesenchymal characteristics to produce the acellular cementum.

Role and Fate of HERS?

It has been postulated that, as the root develops, the first radicular mantle dentin is formed, the epithelial sheath fenestrates, and it is believed that HERS cells migrate away from the root into the region of the future periodontal ligament to form the rests of Malassez (Wentz et al., 1995). However, recent studies suggest that not all HERS cells migrate away; a few of them undergo apoptosis, and some of them remain on the root surface (Kaneko et al., 1999). Furthermore, previous analysis of the expression of Dlx-2 (using a lac Z reporter) and ameloblastin (using antibodies) indicate that some HERS cell express both markers, whereas others express only one, suggesting that there may be more than one cell type present in this tissue (Lezot et al., 2000).

Studies suggest that the basement membrane remains intact on the root dentine surface (MacNeil and Thomas, 1993), and it is disrupted when the cells start secreting osteocalcin (Kagayama et al., 1998). These investigators assumed that the cells producing the osteocalcin were mesenchymal-derived cementoblasts, based on the assumption that only mesenchymal cells express osteocalcin. However, our experiments indicate that HERS cells in vitro express osteocalcin, thus leaving the possibility that the basement membrane disruption is caused by HERS cells when they start depositing the acellular cementum. Furthermore, it has been suggested that the basement membrane contains chemotactic proteins, deposited by the HERS cells, which serve to direct the migration of precementoblast cells (Owens, 1978; Slavkin et al., 1989; Thomas and Kollar, 1989; MacNeil and Thomas, 1993) or induce cementoblast differentiation (Paynter and Pudy, 1958; Slavkin et al., 1988). Recent studies suggest that the collagenous-like protein CAP isolated from cementum, and as demonstrated in this study is present in HERS cells both as a membrane-associated protein and as a secreted protein, possesses chemotactic activity and is capable of recruiting putative cementoblast precursors (McAllister et al., 1990; Pitaru et al., 1995; Metzger et al., 1998; Barkana et al., 2000). The attachment of periodontal ligament cells to CAP appears to be mediated by α5β1 integrin (Ivanovski et al., 1999); however, it has not been demonstrated that HERS cells in vivo express this protein. Based on transplantation and autoradiographic experiments, it has been suggested that cells from the dental follicle are the precursors of cementoblast cells (Yoshikawa and Kollar, 1981; Palmer and Lumsden, 1987; Osborn and Price, 1988; Cho and Garant, 1988; Diekwisch, 2001). Recent experiments by Chai et al. (2000) indicate a common neural crest origin for dental follicle and cementoblast cells. This last possibility would suggest that there are two types of cementoblasts: those derived from HERS and those derived from cranial neural crest cells (Chai et al., 2000). Lezot et al. (2000) suggested that the acellular and cellular cementum have different origins, and our studies support this idea, suggesting that HERS (or HERS-derived cementoblasts) produce acellular cementum and the cranial neural crest-derived cementoblast produce cellular and reparative cementum.

EXPERIMENTAL PROCEDURES

Isolation of the HERS Cells

Homozygous Immortomouse males and CD-1 (ICR) BR normal females (Charles Rivers, Wilmington, MA) were mated overnight, and females were inspected for a vaginal plug (day 0 of gestation). Seven days postnatal heterozygous mice were used for isolation of HERS cells. All animals were treated and killed under humane conditions by using carbon dioxide asphyxiation (procedure approved by the USC Animal Ethics Committee). Mandibles were isolated under sterile conditions and placed in Hanks' buffer (GIBCO-BRL) supplemented with 50 units per ml of penicillin–streptomycin. The mandibular first molars were excised and placed in dispase grade II (Boehringer Mannheim) and incubated at 4°C for 60 min. After digestion, the medium was removed and replaced with calcium-free phosphate buffer solution (CMF-PBS) and HERS tissues were dissected out of the apical end of the molar with tungsten needles. The isolated HERS were placed in a solution of Hanks' containing 0.25% trypsin for 30 min a 37°C. The cells were dispersed by carefully drawing them up and down through an 18-gauge needle. Cells were placed in Hanks' containing 10% fetal calf serum (FCS) and centrifuged a 3,000–5,000 rpm at 4°C for 5 to 10 min. All dissection and culture procedures are performed under sterile conditions using positive pressure hoods.

Cell Cultures

Cells were plated in tissue culture dishes coated with laminin (50–70 mg/ml) and collagen (50–70 mg/ml) in Dulbecco's modified Eagles medium (DMEM) supplemented with 10% FCS, 100 U/ml penicillin/streptomycin, and 10 units/ml of murine IFN (GIBCO-BRL). Cell cultures were incubated at 33°C in a humidified atmosphere of 95% air and 5% CO2, and the medium was changed every 3 days. After cells attained confluence, they were treated with trypsin/EDTA and plated again in larger flasks, and this time, they were grown in keratinocyte growth medium (KGM, Clonetics) containing IFN to eliminate any possible contamination with mesenchymal cells, which do not proliferate well in this medium. The cell concentration (determined by using a hemocytometer) was adjusted to 1 × 105 cells/0.1 ml with medium and frozen in liquid nitrogen. These cells were the source of HERS cells for all subsequent studies. For determination of proliferation rates and differentiation phenotype, cells were plated at a density of 1 × 104 to 5 × 105 in 16-mm 24-well flat-bottom culture dishes containing DMEM supplemented with 10% FCS, 100 U/ml penicillin/streptomycin, and 2 units/ml of IFN. Cultures were incubated at 33°C in a humidified atmosphere of 95% air and 5% CO2. The flasks were washed free of nonadherent cells and differentiation was induced by removing the IFN and incubating the cells in DMEM supplemented with 10% FCS, 100 U/ml penicillin/streptomycin, 50 mg/ml of ascorbic acid, and 2 mM of sodium β-glycerophosphate. Cells were incubated at 39.5°C in a humidified atmosphere of 95% air and 5% CO2. Medium was changed every other day, and samples for analysis were collected as needed. The progress of the cell cultures was continuously monitored by using an inverted microscope.

RT-PCR

Poly(A)-RNA was extracted from cells grown in culture at different days by using the Microfastrack method (Invitrogen, San Diego, CA), following the directions of the manufacturers. The mRNA was converted into cDNA by using Moloney leukemia virus reverse transcriptase and oligo (dT) as primer. The cDNA produced was stored at −80°C until ready to be used. For PCR, the cDNA was amplified with Taq-polymerase (Perkin-Elmer Cetus) using 2 mM of synthetic oligonucleotides containing unique sequences characteristic for the different gene products to be tested. β-Actin was used as internal controls. The reaction proceeded for 30 cycles under the optimal conditions for each primer pair. The reaction mixtures were fractionated on agarose gels and stained with ethidium bromide. Further confirmation that the PCR amplified was the expected gene product was obtained by subjecting the sample to automated DNA sequencing at the USC Microchemical Core Facility.

Extracellular Matrix Analysis

Cells at different days in culture were processed for Von Kossa histochemistry to identify the deposition of Ca++ salt (Thompson and Hunt, 1966). The mineral produced by these cultures was also assayed by TEM. Tissues and cells were fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) for 2 hr at 4°C and post-fixed in 1% osmium tetroxide for 1 hr at 4°C. Samples were dehydrated in graded alcohols and embedded in Epon 812 resin. One-micron sections were cut for orientation, and ultra thin sections (gray–silver) were cut by using a Reichert OM-U3 ultramicrotome and a diamond knife (Diatome). Specimens were viewed by using a JEOL-1200 EX electron microscope operating at 80 kV. Sections were unstained for visualization of the crystals.

In Situ Hybridization

Eight days postnatal heterozygous mice were killed as previously described. Mandibles and maxillae were isolated under sterile conditions and placed in 4% buffered paraformaldehyde for 4 hr. Samples were decalcified with 10% EDTA for 10 days, and the specimens were embedded in paraffin and sectioned according to routine procedures. After clarification with saline, sections were treated with proteinase K (20 μg/ml) for 20 min, washed in PBS, fixed in 4% paraformaldehyde for 1 hr, and washed with 2× standard saline citrate (SSC). Sections were incubated with 0.2 N HCl for 15 min to inhibit any endogenous ALP activity and rinsed with triethanolamine, with acetic anhydride, and then 2× SSC. Incubation with hybridization buffer containing the digoxigenin-labeled sense or antisense ameloblastin probe (denaturated at 80°C for 1 min before hybridization) was done overnight at 50°C. Sections were washed in 2× SSC and then incubated with RNAse A (40 μg/ml) at 37°C for 30 min. Sections were washed in 2× SSC containing 50% formamide at 50°C for 5 min, then in 1× SSC at room temperature, 0.5× SSC, and then Tris buffer. Sections were incubated with 2% sheep serum for 30 min, anti-digoxigenin antibody conjugated with ALP (1:500 dilution) for 1 hr at room temperature and washed in buffer several times. The ALP substrate (nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate) was added, and slides were incubated in the dark at 4°C overnight. The reaction was stopped by adding buffer containing EDTA; slides were rinsed, mounted with Aquamont, and analyzed by light microscopy.

Protein Extraction

Proteins were collected as secreted proteins, EDTA soluble proteins, membrane proteins, or intermediate filament proteins. For collecting secreted proteins, the medium was changed to serum-free DMEM containing ITS for 24 hr. After a second medium change with serum-free DMEM, the conditioned medium was harvested after 48 hr in culture. The collected medium was desalted by using Ultrafree-4 centrifugal ultrafiltration devices (Millipore, Bedford, MA) and lyophilized. Samples were dissolved in sodium dodecyl sulfate (SDS) -PAGE sample buffer. For collecting mineral-associated proteins in the ECM, cell layers were washed with PBS and incubated for 24 hr with 0.5 M EDTA buffer (50 mM Tris, pH 7.4) containing 1 mM phenyl methyl sulfonyl fluoride (PMSF), 5 mM ϵ-aminocaproic acid, and 5 mM benzamidine HCl. Harvested samples were desalted by ultrafiltration, lyophilized, and prepared for SDS-PAGE. For isolation of cell membrane proteins, cells were washed with PBS containing protease inhibitors (1 mM PMSF, 5 mM ϵ-aminocaproic acid, 5 mM benzamidine HCl, 2 μg/ml aprotinin, 2 μg/ml leupeptin, and 1 μg/ml pepstatin) and scraped into the same buffer by using a cell scraper. Cells were disrupted by ultrasonication, and homogenates were centrifuged for 5 min at 14,000 rpm at 4°C to remove large particles. The supernatant was centrifuged for 60 min at 100,000 × g at 4°C to pellet membrane fragments, which were dissolved in SDS-PAGE sample buffer. Intermediate filament (IF) associated proteins were isolated by the method of Steinert et al. (1982). Briefly, cells were washed with PBS containing protease inhibitors and dissolved in lysis buffer (PBS containing 1% Triton X-100, 0.6 M KCl, 10 mM MgSO4, and protease inhibitors) for 5 min. After treating with DNase, homogenates were centrifuged for 30 min at 14,000 rpm at 4°C. The pellets were homogenized in disassembly buffer (8 M urea [pH 8.0], 0.2% β-mercaptoethanol, 5 μM EDTA, and protease inhibitors) by stirring for 30 min at room temperature. After centrifugation, the supernatant was dialyzed against PBS to promote the assembly of IF. The polymerized IF was recovered by centrifugation at 100,000 × g for 30 min and dissolved in SDS-PAGE sample buffer.

Western Immunoblot Analyses

HERS cells grown for 0, 1, 5, 10, 15, 20, and 25 days in differentiation medium were used. Sample aliquots of extracted proteins were fractionated by SDS-PAGE on 10% or 4–12% gradient gels and electrophoretically transferred to a polyvinylidene difluoride membrane (Millipore). The membrane was blocked with 5% BLOTTO and incubated with primary antibody solution at dilutions of 1:800–1:64,000. The antibody against ameloblastin was a gift from Dr. Paul Krebsbach, Michigan. The antibody against CAP was produced by Dr. Arzate (Arzate et al., 2002). The antibody against amelogenin was produced in our laboratory. All other antibodies were commercially purchased from Chemicon International, Inc. (BSP), Santa Cruz Biotechnology, Inc. (cadherin, E- and OB-cadherins, vimentin, keratin, integrin, MEM, and ZO1), R&D Systems, Inc. (osteopontin), and Biomedical Technologies, Inc. (osteocalcin). All antibodies were tested with either recombinant proteins or tissue extracts to ensure that they work and to determine their specificity. Secondary antibodies conjugated with horseradish peroxidase were used at dilutions of 1:1,000–1:4,000. Immunoreactivity was determined by using the enhanced chemiluminescence reaction (Amersham, Arlington Heights, IL).

Alkaline Phosphatase Assay

Cells were scraped into ice-cold 50 mM Tris-HCl buffer and sonicated for 30 sec. After centrifugation for 30 min at 5,000 rpm, ALP activity in the supernatant was assayed in 96-well microtiter plates according to Lowry et al. (1954) using p-nitrophenyl phosphate as a substrate and measuring the absorbance at 405 nm. Protein concentration was determined by using the method of Bradford (1976) assay using BSA as standard. Assays were repeated at least three times.

Osteocalcin Assay

The expression of osteocalcin was measured by radioimmunoassay in the medium by using the kit distributed by Biomedical Technologies, Inc. (Stoughton, MA) and following the procedures described by the vendors.

Ancillary