| Sign In to gain access to subscriptions and/or personal tools. |
Hereditary Dentin Defects
1 Seoul National University, School of Dentistry Department of Pediatric Dentistry & Dental Research Institute, 28-2 Yongon-dong, Chongno-gu, Seoul, Korea 110-749; and Correspondence: * corresponding author, jsimmer{at}umich.edu
By the Shields classification, articulated over 30 years ago, inherited dentin defects are divided into 5 types: 3 types of dentinogenesis imperfecta (DGI), and 2 types of dentin dysplasia (DD). DGI type I is osteogenesis imperfecta (OI) with DGI. OI with DGI is caused, in most cases, by mutations in the 2 genes encoding type I collagen. Many genes are required to generate the enzymes that catalyze collagen’s diverse post-translational modifications and its assembly into fibers, fibrils, bundles, and networks. Rare inherited diseases of bone are caused by defects in these genes, and some are occasionally found to include DGI as a feature. Appreciation of the complicated genetic etiology of DGI associated with bony defects splintered the DGI type I description into a multitude of more precisely defined entities, all with their own designations. In contrast, DD-II, DGI-II, and DGI-III, each with its own pattern of inherited defects limited to the dentition, have been found to be caused by various defects in DSPP (dentin sialophosphoprotein), a gene encoding the major non-collagenous proteins of dentin. Only DD-I, an exceedingly rare condition featuring short, blunt roots with obliterated pulp chambers, remains untouched by the revolution in genetics, and its etiology is still a mystery. A major surprise in the characterization of genes underlying inherited dentin defects is the apparent lack of roles played by the genes encoding the less-abundant non-collagenous proteins in dentin, such as dentin matrix protein 1 (DMP1), integrin-binding sialoprotein (IBSP), matrix extracellular phosphoglycoprotein (MEPE), and secreted phosphoprotein-1, or osteopontin (SPP1, OPN). This review discusses the development of the dentin extracellular matrix in the context of its evolution, and discusses the phenotypes and clinical classifications of isolated hereditary defects of tooth dentin in the context of recent genetic data respecting their genetic etiologies.
Key Words: dentin • dentin sialophosphoprotein • osteogenesis imperfecta • dentinogenesis imperfecta • dentin dysplasia
Dentin is the mineralized tissue constituting the body of a tooth, serving as a protective covering for the pulp and as a support for overlying enamel and cementum. On a weight basis, mature dentin is about 70% mineral, 20% organic matrix, and 10% water. Dentin is the product of specialized, end-differentiated, cells called odontoblasts. Odontoblasts comprise a sheet of columnar cells that line the pulpal surface of dentin and extend cell processes partly or all the way through dentin. Odontoblasts are intimately associated with the formation and maintenance of dentin, communicate with pulp afferent nerves, and serve as the first biological line of defense against environmental injury, such as in caries (Nanci, 2003). Our goal is to understand how normal dentin forms and functions. To achieve this goal, we are interested in the evolution of dentin and other mineralized tissues, how odontoblasts differentiate and control the expression and secretion of proteins, the composition and structural/functional properties of dentin extracellular matrix constituents, how odontoblasts monitor the extracellular matrix and respond to feedback, and, finally, how specific genetic defects lead to the observed patterns of inherited dental malformations. It is hoped that insights gained by improving our understanding in these areas will lead to improvements in the way we diagnose and treat pathologies affecting dentin, whether they arise from genetic or environmental factors, injury, or disease. Here we present a perspective and a review of the hereditary defects of tooth dentin that are classified under the designations of dentinogenesis imperfecta (DGI) and dentin dysplasia (DD) (Shields et al., 1973).
In this section, we discuss development of the dentin extracellular matrix in the context of the evolution of extracellular matrices and biomineralization, to provide insights into the genetic etiologies of DGI and DD. The emergence of multicellular animals from single-celled organisms is associated with the expansion and enhancement of cell-signaling pathways, which include growth factors and receptors capable of transmembrane signal transduction, and homeotic transcription regulatory systems that mediate cell differentiation (David, 2001; Kaiser, 2001). Multicellularity requires strengthening of the linkages that bind cells together, and the construction of extracellular matrices (ECM) to provide structural integrity and to act as a substrate for cell adhesion, migration, and growth. Cells form an intimate relationship with the ECM they secrete (Humphries et al., 2004). Interactions between ECM components and membrane receptors and the sampling of the ECM by endocytosis are ways in which the ECM influences gene expression (Exposito et al., 2002). Fibrillar collagens are expressed by invertebrates and vertebrates and correlate with the emergence of multicellularity. In addition to collagen, extracellular matrices contain proteins, glycoproteins, and proteoglycans that commonly possess cell-binding and/or collagenbinding domains (Myllyharju and Kivirikko, 2001). These molecules organize collagen into fibrillar networks, appraise the cell of conditions in the ECM, and mediate signals that stimulate cellular responses to external stimuli. Besides collagen, hyaluronan is an important carbohydrate polymer of unbranched repeating disaccharide units that is able to bind cell receptors and connect proteins with glycan attachments into proteoglycan assemblies that are major structural and functional constituents of extracellular matrices. The evolution of complex and dynamic extracellular matrices occurred long before the advent of biomineralization. Molecular-clock analyses estimate that invertebrates diverged from chordates between one billion (Wray et al., 1996) and 615 million years ago (Peterson et al., 2004), indicating that collagen-based extracellular matrices existed by that time. Biomineralization of extracellular matrices evolved independently in many different taxa (Knoll, 2003). The earliest fossil evidence of mineralization in vertebrates is pharyngeal tooth-like structures (conodonts) from jawless fish that appear in the fossil record at ~ 540 million years ago (mya), but these organisms are believed to have diverged already from the line leading to tetrapods. The jawless fish (pteraspidomorphs) with dermal armor (~ 470 mya) are more likely inventors of biomineralization in the line ancestral to jawed vertebrates (Kawasaki et al., 2004), suggesting that collagen-based extracellular matrices evolved for tens to hundreds of millions of years prior to the onset of biomineralization. The major inference from the evolutionary perspective is that the biomineralization of bone and dentin is built upon an organic matrix that is involved in many important functions and interactions that are necessary for proper assembly and functioning of the matrix, but are not necessarily also involved in the deposition of mineral. The biomineralization of vertebrate extracellular matrices evolved through a more recent series of gene duplications and the origin or expansion of the secretory calcium-binding phosphoprotein (SCPP) family from the 5' region of the SPARC (secreted protein, acidic and rich in cysteine) gene (Delgado et al., 2001). The human SCPP family is comprised mainly of a cluster of genes on chromosome 4 that encode the major non-collagenous extracellular matrix proteins of bone, dentin, and enamel, as well as proteins secreted in milk and saliva (Kawasaki and Weiss, 2003, 2006). SIBLINGs (small integrin-binding ligand N-linked glycoproteins) are a subfamily of 5 SCPP genes involved in bone and dentin formation (Fisher and Fedarko, 2003), and are the primary candidate genes for isolated inherited dentin defects: dentin sialophosphoprotein (DSPP), dentin matrix protein 1 (DMP1), integrin-binding sialoprotein (IBSP), matrix extracellular phosphoglycoprotein (MEPE), and secreted phosphoprotein-1, or osteopontin (SPP1, OPN). In humans, the SIBLING genes form a cluster on chromosome 4q21–q25 (Fig. 1  ). Core-binding factor a1 (Cbfa1) is a transcription factor required for osteoblast and chondrocyte maturation that regulates the expression of the SIBLINGs and other genes. The Cbfa1–/– knockout mice lack skeletal mineralization and die at birth (Komori et al., 1997).
Hereditary conditions affecting dentin have long been evident in human populations (Gray, 1970). "Hereditary opalescent dentin" was first proposed to describe inherited dentin defects that occurred in the absence of systemic, non-dental, manifestations (Hodge et al., 1936). "Dentinogenesis imperfecta" was used to describe the dental phenotypes associated with osteogenesis imperfecta (OI) (Roberts and Schour, 1939). The classification system currently in use was designed to discriminate among various patterns of dentin defects and to specifically include a designation for a mild dentin phenotype classified as dentin dysplasia type II (Shields et al., 1973). This system recognizes 3 types of dentinogenesis imperfecta [DGI-I (MIM 166240), DGI-II (MIM #125490), and DGI-III (MIM #125500)] and 2 types of dentin dysplasia [DD-I (MIM %125400) and DD-II (MIM #125420)]. When this classification system was conceived, it was appreciated that, as the genetic causes of inherited dentin defects became known, revising the nomenclature for hereditary dentin defects would be necessary (Bixler, 1976). We are now in an uncomfortable transition period where the current system is increasingly at odds with genetic findings, but knowledge of the genes associated with inherited dentin defects has not yet advanced to a point where a comprehensive etiology-based classification system can be proposed (Bixler, 1976; Dean et al., 1997). The dental phenotypes of the 5 divisions of Shields’ classification system are briefly summarized, followed by a more in-depth review of each division and what has been learned of its etiology.
DGI-I This is the dental phenotype in persons afflicted with OI. The teeth show marked discoloration and attrition in both the deciduous and permanent dentitions. Pulpal obliteration occurs soon after eruption or prior to tooth eruption. The degree of expressivity is variable, even within a single individual, ranging from total pulpal obliteration to normal dentin.
DGI-II
DGI-III
DD-I
DD-II
In human populations, there exists a broad spectrum of inherited dentin malformations. Shields’ classification attempted to compartmentalize this phenotypic variation into groups (Fig. 2
DGI-I is the dental phenotype associated with OI. OI is a genetic disorder featuring increased bone fragility, low bone mass, and other connective tissue manifestations usually caused by defects in the 2 genes encoding type I collagen (COL1A1, 17q21.31–q22; COL1A2, 7q22.1). The type I collagen triple helix is a fibril comprised of two alpha 1 chains and one alpha 2 chain. It is the most abundant protein in bone, skin, and other connective tissues. Because many enzymes are involved in catalyzing collagen post-translational modifications (3 hydroxylases, 2 glycosyltransferases, 2 proteinases, 2 isomerases, and an oxidase), and many proteins function through interactions with collagen, defects in many genes can cause conditions resembling OI, and specific type I collagen defects can cause diseases other than OI (Myllyharju and Kivirikko, 2004). OI is generally classified into 4 clinical types, although 3 additional types have been added, to include distinct features (Rauch and Glorieux, 2004). Mild forms of OI are usually caused by a premature stop codon or deletion of a single COL1A1 allele, which reduces the amount of normal type I collagen. Severe forms are caused by dominant-negative mutations in COL1A1 or COL1A2 that lead to structural defects in the assembled collagen fibril (Gajko-Galicka, 2002); however, genotype-phenotype correlations are often complex and unpredictable (Roughley et al., 2003), and OI is found in individuals with no apparent defects in the type I collagen genes (Rauch and Glorieux, 2004). Genetic heterogeneity in the etiology of osteogenesis imperfecta is established, since homozygous OI unlinked to type I collagen genes (Aitchison et al., 1988) was demonstrated, and OI type VII was linked to chromosome 3p22–24 (Labuda et al., 2002). Recently, a gene defect in the mouse osteogenesis and dentinogenesis imperfecta model, fragilitas ossium (fro), was identified in the gene encoding neutral sphingomyelin phosphodiesterase 3 (Smpd3) (Aubin et al., 2005). While OI with DGI (OI/DGI) is usually associated with collagen-I defects, the clinical expression and genetic etiology of OI/DGI are complex. Collagen plays non-identical roles in bone and dentin, since the severity of the dentin and bone defects displayed by individuals with defined collagen mutations varies over a wide range (O’Connell and Marini, 1999). Some persons with OI displaying obvious DGI show no detectable bone phenotype (Pallos et al., 2001). In contrast, about half of all OI cases show no obvious clinical signs of DGI. In some OI cases, the DGI phenotype is not clinically evident, but can be detected radiologically (Lund et al., 1998). In other cases, the DGI is discovered only by histologic examination (Malmgren and Norgren, 2002), and then, the histological appearance of the dysplastic dentin is often less severe in the OI persons having clinical or radiological signs of DGI compared with those who do not (Malmgren and Lindskog, 2003). Mild DGI has been associated with Bruck syndrome 1 (OI with congenital joint contractures, MIM %259450), an autosomal-recessive disorder (Brenner et al., 1993). Bruck syndrome 1 mapped to 17p12 region at the site of the gene encoding bone telopeptide lysyl hydroxylase (Bank et al., 1999). Bruck syndrome 2 (MIM #609220) maps to 3q23–q24 and is caused by mutations in the lysyl hydroxylase 2 gene (PLOD2) (van der Slot et al., 2003).
The inclusion of DGI-I in Shields’ classification is unfortunate. DGI-I is an example of syndromic DGI (where the dentin defects are not the most predominant or consistent manifestation in most kindreds). All of the other inherited dentin defects in Shields’ classification are isolated (dentin defects are the predominant and only consistent manifestation). Increasingly, the DGI phenotype is recognized as a variable feature in many syndromes. Ehlers-Danlos syndrome (EDS) is a heterogeneous group of generalized connective tissue disorders, the major features of which are tissue fragility, skin extensibility, and joint hypermobility (Uitto, 2005). Some forms of EDS have dental phenotypes such as dysplastic dentin and obliterated pulp chambers (Barabas, 1969), dental features mimicking DD-I (Pope et al., 1992), and characteristic DGI-II with variable expressivity (Komorowska et al., 1989). EDS has a wide phenotypic spectrum, consisting of 6 major classification types that can be caused by molecular defects in types I, III, and V collagen, tenascin-X, and 2 collagen-modifying enzymes (lysyl hydroxylase and procollagen N-peptidase) (Mao and Bristow, 2001; Schalkwijk et al., 2001). Goldblatt syndrome (MIM 184260) is a form of spondylometaphyseal dysplasia with joint laxity and DGI. The deciduous teeth display typical DGI, but the permanent teeth appear normal. Aberrant mobility of type II collagen chains by gel electrophoresis suggested a point mutation in COL2A1 (12q13) (Bonaventure et al., 1992). Schimke immuno-osseous dysplasia (SIOD, MIM #242900), an autosomal-recessive disorder, is characterized by a combination of spondylo-epiphyseal dysplasia, progressive renal disease, and lymphopenia with defective cellular immunity (Saraiva et al., 1999). A person with this disorder has characteristic DGI features, such as a grey-yellowish discoloration of the dentin, bulbous crowns with a marked cervical constriction, and small or obliterated pulp chambers (da Fonseca, 2000). Recently, SIOD was linked to mutations in SMARCAL1, encoding a chromatin remodeling protein (Boerkoel et al., 2002). There are also sporadic reports of persons displaying DGI as part of a larger syndrome, but the genetic etiology remains unknown (Beighton, 1981). Opalescent teeth have been reported in a person having skeletal dysplasia with disproportionate short stature, short neck, broad chest, kyphosis, and protruding abdomen (Kantaputra, 2001), and in two siblings with microcephalic osteodysplastic primordial dwarfism (Kantaputra, 2002). The deciduous and permanent teeth were equally opalescent, and the roots were extremely short and tapered, or rootless. A family with an unusual pattern of skeletal malformations resembling OI has been reported (Moog et al., 1999). Two affected siblings had OI-like features (bone fragility, wormian bone, and DGI), but normal collagen findings. In this case, the unaffected mother also had some features of DGI, so either the DGI in this family might be independent of the skeletal dysplasia, or the syndrome has extremely variable expression and the mother is indeed affected. Recently, two brothers born of consanguineous parents had DGI, delayed tooth eruption, mild mental retardation, proportionate short stature, sensorineural hearing loss, and dysmorphic faces. No mutation in type I collagen was identified, and the mode of inheritance was proposed as autosomal-recessive (Cauwels et al., 2005).
DGI-II is one of the most common dominantly inherited disorders and affects approximately one person in every 8000. Genetic analyses linked DGI-II (Ball et al., 1982; Aplin et al., 1999), DGI-III (Boughman et al., 1986), and DD-II (Dean et al., 1997) to the chromosome 4q21 region, making the SIBLING family the prime candidate genes for these disorders. Defects in the DSPP gene can cause DGI-II (Xiao et al., 2001; Zhang et al., 2001; Kim et al., 2004; Malmgren et al., 2004), DGI-III (Dong et al., 2005; Kim et al., 2005), and DD-II (Rajpar et al., 2002). While there are other candidate genes for hereditary dentin defects (Ye et al., 2004), no disease-causing mutations outside of the DSPP gene have yet been identified (Table ). The DSPP gene structure showing the positions of known disease-causing mutations is provided in Fig. 3.
No bony defects have been reported in the kindreds with defined DSPP mutations, even though DSPP is expressed in both dentin and bone. Potential reasons for the absence of bony defects include the lower expression of DSPP in bone, tissue-specific differences in its proteolytic processing (Qin et al., 2001, 2004), and molecular redundancy (the potential for other bone extracellular matrix molecules to serve the same function as DSPP). It is also possible that bony phenotypes in DSPP mutation kindreds are subtle and go undetected. In some DGI-II kindreds with DSPP mutations (p.P17T and p.V18F), older affected members develop progressive sensorineural high-frequency hearing loss (DFNA39). Is DSPP expression especially important in the small bones of the inner ear? Defects in these bones would be expected to cause conductive, rather than neurosensory, hearing loss. Progressive hearing loss is one of the principal symptoms of OI, affecting about 50% of adult patients (Kuurila et al., 2000). The hearing loss in OI is predominantly of the conductive type. Progressive hearing loss in persons with DGI-II might be a secondary effect. Many persons with DGI-II experience significant dental attrition and a concomitant loss of vertical dimension (overclosure of the jaw). Jaw position affects the shape of the inner ear (Oliveira et al., 1992), and tooth loss, even in the absence of DGI, can lead to hearing deficits (Lawrence et al., 2001; Nagasaka et al., 2002).
The 2 main features (multiple pulp exposures and shell teeth) used to distinguish DGI-III from DGI-II are not unique to DGI-III. The pulp chambers in DGI-II are sometimes abnormally wide initially (shell teeth), but they progressively obliterate (Heimler et al., 1985; Ranta et al., 1993; Tanaka and Murakami, 1998; Sapir and Shapira, 2001). Even in the Brandywine isolate (the DGI-III prototype), the phenotype of shell teeth was described only in young children (Witkop, 1975). The similarities between DGI-II and DGI-III extend beyond the phenotype to the genotype: The same DSPP mutation (c.52G  T, p.V18F) manifested as DGI-II and DGI-III in different families (Kim et al., 2005), and the genetic defects underlying the original Brandywine phenotype are in DSPP (Dong et al., 2005). The Dspp–/– mouse teeth displayed relatively severe deficiencies in root dentin formation, similar to those in the human DGI-III phenotype (Sreenath et al., 2003).
DD-I is a rare anomaly of unknown etiology that affects approximately one patient in every 100,000. Several DD-I kindreds showed an autosomal-dominant mode of inheritance. It is not known if DD-I is another allelic disorder of the DSPP gene, or a mixed phenotype. DD-I, DD-II, and DGI-II have been observed within a single family (Graham et al., 1965) or single affected individual (Ciola et al., 1978; Tidwell and Cummingham, 1979; Diamond, 1989).
A mutation in the DSPP signal peptide codon (c.16T G, p.Y6D) was identified in a DD-II family (Rajpar et al., 2002). The effect of the mutation was a reduction (by less than 50%) of the amount of DSPP secreted into the forming dentin matrix, but the secreted protein was entirely normal. Since some of the mutations underlying DGI-II resulted in no DSPP expression from the mutant allele (a 50% reduction), the genetic data were consistent with the interpretation that the DD-II and DGI-II phenotypes are mild and severe forms of the same disease.
In humans, there are 27 different types of collagen, expressed from 42 different collagen genes (Myllyharju and Kivirikko, 2004). Type I collagen constitutes 85–90% of the dentin organic matrix (Linde et al., 1980), and is the major protein in bone. The triple-helical (3D) structure of collagen was determined by fiber diffraction over 50 years ago (Rich and Crick, 1955), and its many post-translational modifications have been characterized (Viguet-Carrin et al., 2006). The abundance of collagen in bone and dentin and the elaborate biochemistry involved in its synthesis are evident in the diverse etiology and clinical manifestations of inherited defects involving both bone and dentin. In these disorders, bone is more sensitive to collagen defects than is dentin, and the bony defects are generally a more consistent phenotypic feature than are the dentin defects. Because bone defects are the more predominant phenotype in OI and related diseases, these disorders are more properly included in classification systems other than the DGI-I designation in the Shields’ classification. The observation that bone is more sensitive to type I collagen defects than is dentin remains unexplained. Perhaps it relates to bone being a critical element of the hormonally regulated calcium and phosphate homeostasis system (Costanzo, 1998), or to the capacity of bone for regeneration and repair. Part of the reason may relate to differences between bone and dentin in the way collagen binds to, and is organized by, non-collagenous proteins. The most abundant non-collagenous proteins in dentin are the DSPP-derived proteins (MacDougall et al., 1997). Shortly after DSPP is synthesized by odontoblasts, it is cleaved into 3 structural/functional domains: dentin sialoprotein (DSP) (Ritchie et al., 1994), dentin glycoprotein (DGP) (Yamakoshi et al., 2005b), and dentin phosphoprotein (DPP) (Ritchie and Wang, 1996). In contrast to what is known about collagen structurally, the post-translational modifications of DSPP-derived proteins (excepting DGP) have been only poorly characterized (Qin et al., 2004), and their 3-D structures are completely unknown. It was only recently demonstrated that DSP is a proteoglycan capable of forming covalent dimers (Yamakoshi et al., 2005a), and that DMP1 is a proteoglycan (Qin et al., 2006). Targeted gene knockouts in mice have demonstrated that at least 5 genes encoding proteoglycans contribute to dentin formation: Dspp (Sreenath et al., 2003), Dmp1 (Ye et al., 2004), fibromodulin (Fmod) (Goldberg et al., 2006), and biglycan (Bgn) and decorin (Dcn) (Goldberg et al., 2005). All of these proteoglycans bind collagen. DSPP is the only one of these genes that is primarily dedicated to dentin formation and has been shown to be part of the etiology of isolated dentin defects. Genetic studies prove that DSPP is critical for proper dentin formation. It is apparent that DSPP-derived proteins play a role beyond biomineralization, and probably serve several important functions. Inferences about the functions of DSPP based upon the nature of DGI and DD phenotypes are limited, because of the possibility of secondary effects. The obliteration of pulp by the accelerated deposition of secondary dentin, for instance, could be the consequence of odontoblasts responding to a deficiency in the matrix or weakness of the dentin. In the Dspp knockout mice, biglycan and decorin were increased in the widened predentin zone. How much of the Dspp–/– phenotype is caused by these secondary changes? Since isolated inherited dentin defects are divided into 4 types in the Shields’ classification, it is surprising that the early results of mutational analyses have identified mutations only in DSPP, and in none of the other 4 SIBLING genes. To date, 8 different disease-causing DSPP mutations have been identified in the 5' region, up to and including the codon for Arg68 in exon 4. In our analyses of nine kindreds with inherited dentin defects, five showed disease-causing mutations in the 5' coding region of DSPP. Disease-causing mutations in the 3' coding region of DSPP might have caused the disease in some or all of the other kindreds. Currently, the DPP coding region cannot be analyzed for mutations because of its high sequence redundancy. Therefore, the initial findings of genetic studies seeking to understand the genetic causes of isolated dentin defects indicate that DSPP mutations play the predominant etiological role, but contributions by the other SIBLING genes cannot be ruled out. The human DSPP cDNA (Gu et al., 1998) and genomic (#AF163151; 9944 bp) (Gu et al., 2000) sequences have been reported. Based upon the human genomic sequence, a DSPP reference sequence was assembled (#NM_014208; 4187 bp). Also in the databases are the human chromosome 4 contig (#NT_016354.17) and alternate assembly (#NT_086651.1). These independently determined human DSPP sequences show significant variation in the DPP coding region (exon 5), so that the wild-type human DPP sequence is still unknown. (An alignment of the DSPP reference sequence against other human DSPP sequences from NCBI is provided in the APPENDIX.) These sequence differences lead us to suspect that the DPP coding region is highly polymorphic in humans. Technical advances in our ability to perform sequence analyses in the DPP coding region are needed, along with knowledge of the normal range of DSPP sequence variations, before the role played by DPP mutations in the etiology of inherited dentin defects can be elucidated. In summary, great advances are being made in our understanding of how mineralized tissues evolved over the long course of time. The growth factors and homeotic transcription factors that control tooth development and cell differentiation are being identified, and their contributions defined. The macromolecular components of mineralizing extracellular matrices have been isolated, and their structures and functions are being profitably investigated. The genetic etiologies of syndromic and isolated inherited dentin defects are being described. These exciting advances are steadily improving our understanding of normal and pathological tooth formation, and are inspiring new diagnostic and therapeutic innovations to improve our oral health.
This work was supported by a grant (A060010) from the Korea Health 21 R&D Project, Ministry of Health & Welfare, Republic of Korea, and by NIDCR grants DE12769 and DE15846.
A supplemental appendix to this article is published electronically only at http://www.dentalresearch.org. Received for publication June 2, 2006. Accepted for publication October 19, 2006.
Journal of Dental Research, Vol. 86, No. 5,
392-399 (2007) This article has been cited by other articles:
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
|
|
 |
|
Sign In to gain access to subscriptions and/or personal tools.
TOP
ABSTRACT
INTRODUCTION
