Frontiers Between Science and Clinic in Odontology Volume Title: Phosphorylated Extracellular Matrix Proteins of Bone and Dentin

In vertebrates, bone and dentin have been found in the earliest mineralized skeletal elements, which consisted of odontodes and osseous plates. For more than 450 million years, the two tissues evolved concomitantly in various skeletal elements, which were either conserved or subjected to wide changes as illustrated in extinct and extant lineages: odontodes, teeth, scales, osteoderms, fin rays, bony spines, dermal bones, and cartilage bones from the endoskeleton. Bone is cellular in most species, with a few exceptions of bone lacking osteocytes in some early vertebrates and in derived teleost fishes. Dentin has been classified into eight distinct types and some of them are highly specialized such as the elasmodin of fish scales. In this chapter, we summarize our knowledge and propose a scenario for the evolution of these two tissues. We also review hypotheses proposed for tracing back the origin of the skeletal tissues during the prechordate-vertebrate transition. Specific molecules and specialized cells were undoubtedly differentiated prior to the identification of these tissues in the fossil record. Cartilage, bone and dentin matrices were probably created when typical fibrillar collagen type II and type I, the major components of these skeletal tissues, were recruited after the genome duplications that predated the vertebrate differentiation. The secretory calcium-binding phosphoproteins (SCPPs) are crucial molecules controling crystallization of calcium phosphate in these skeletal tissues. SCPP genes are thought to have arisen from SPARCL1 (SPARC-like 1), which originated from SPARC (secreted protein, acidic, cystein-rich), at about the same period when fibrillar collagen genes used for vertebrate skeletal tissues were differentiated. Eventually, neural crest-derived specialized cells employed a fibrillar collagen and an SCPP, among other proteins, and created the first mineralized dermal skeletal elements, probably in association with sensory organs located on the body surface. Currently at least five acidic SCPPs (DSPP, DMP1, IBSP, MEPE, SPP1) are involved in the mineralization of bone and dentin tissues in mammals, and other SCPPs were identified in teleost fishes. In this chapter we analyze the evolution of these SCPPs and their relationships within the framework of vertebrate phylogeny.

Bone is composed of a mineralized matrix synthesized by osteoblastic cells and maintained by osteocytic cells embedded within this calcified environment. This review of recent literature reveals that regulation of bone formation is influenced by a complex panoply of secreted factors and cell surface receptors which control proliferation and differentiation of osteoproprogenitor cells as well as the functional activity and survival of osteoblastic cells and osteocytes. The osteocyte, the mechanosensor of bone, plays a prominent regulatory role through its control of osteoblastic lineage cells and of osteoclastic lineage cells in remodeling of bone in response to external mechanical forces. This concept is illustrated with specific examples of local, paracrine and systemic signaling mechanisms mediating the responses of cells in bone. As an indication of the inherent specificity of intercellular communication mediated by osteocytes, sclerostin and FGF23, a local and a systemic effector secreted by osteocytes, respectively, are also used as molecular and functional biomarkers. Sclerostin expression by osteocytes is inversely related to mechanical stimulation and sclerostin binds to the Lrp5/6 receptor on osteoblastic cells to inhibit the Wnt1/β-catenin signaling pathway controlling bone formation. Recent work has also shown that the viability of osteocytes is influenced by estrogen deficiency since ovariectomy in sheep increased apoptosis or programmed cell death. The mechanism whereby osteocytes regulate osteoclastogenesis was recently clarified using osteopetrotic RANKL null mice. Osteocytes were shown to be the major source for this required pro-osteoclastogenic factor displaying a greater capacity to support osteoclastogenesis than osteoblasts or bone stromal cells. Finally, while the proliferative actions of growth factors and cytokines are well documented, TGF-β appears to also play a prominent role in the prevention of apoptosis of osteoblastic cells. As illustrated above, the response of bone cells to a generic signal can lead to stimulation or suppression of proliferation of osteoprogenitor cells, to increased differentiation of bone forming osteoblastic cells, or to regulated apoptosis of osteoblasts and/or osteocytes.

The crown and root parts of mammal’s teeth associate different types of dentin: peripheral dentin (mantle dentin, hyaline’s Hopewell-Smith layer and Tome’s granular layer) and circumpulpal dentin, which include intertubular and peritubular dentin. Firstly, we describe the structural characteristics and specific composition of the different sorts of dentin. Secondly, as odontoblasts sustain dentin formation, we analyze the role of these cells in the synthesis and secretion of dentin extracellular matrix (ECM) components. Morphological changes occur during the successive stages leading the pre-odontoblasts to become post-mitotic polarized secretory odontoblasts, and ultimately senescent cells. In the third part, we report the immunohistochemical and radioautographic investigations that shed lights on the intracellular uptake and secretion of dentin components, and the self- associative properties coordinating post-secretory interactions between ECM molecules. Dentin extracellular matrix proteins include type I collagen, phosphorylated proteins, glycoproteins, proteoglycans and proteolipids. In addition, forming a minor part of ECM, a few molecules originate from intercellular diffusion of blood serum molecules. Interactions between ECM molecules, the tissue non-specific alkaline phosphatase (TNAP) and calcium-binding proteins are key factors in promoting dentin mineralization.

The Small Integrin-Binding LIgand, N-linked Glycoprotein (SIBLING) family contains five genes that have been shown to play critical functions in biomineralization as either enhancers and/or inhibitors of calcification. These genes are located in a cluster on human chromosome 4q21 and include: dentin sailophosphoprotein (DSPP), dentin matrix protein 1 (DMP-1), matrix extracellular phosphoglycoprotein (MEPE), bone sailoprotein (IBSP) and osteopontin (SPP1). Through the rapid advances in molecular genetics and studies of transgenic null mice substantial progress in determining the function of various SIBLINGs has been established. To date, two SIBLINGs have been shown to be directly involved in the pathogenesis of human diseases with altered dentin or bone mineralization phenotypes. These SIBLINGs are DSPP associated with various dentin structural diseases and DMP-1 associated with an autosomal recessive form of hypophosphatemic rickets.

Proteins produced from the DSPP gene are critical for normal dentin and tooth formation. Allelic mutations in the DSPP gene cause a variety of phenotypes that range in severity. The most common condition caused by DSPP mutations is dentinogenesis imperfecta type II (DGI – II). The DGI –II trait is highly penetrant and transmitted in an autosomal dominant manner. De novo mutations that cause DGI-II occur but appear to be rare. The clinical phenotype is characterized by amber to blue gray discoloration of both the primary and permanent dentitions and teeth that may be normal or diminished in size and can have a marked constriction at the cementumenamel junction. The pulp chambers tend to fill in with irregular dentin that has an increased water content, decreased mineral content and diminished number of dentinal tubules. The second phenotype associated with DSPP mutations is known as dentin dysplasia type II (DD-II). This trait is essentially the same as DGI-II in the primary dentition with an attenuated expression in the permanent dentition. The permanent teeth tend to have normal or only slight color changes and normal crown and root morphology. The pulp chambers are frequently abnormal in morphology with anterior teeth often having a thistle tube shaped pulp chamber. Investigators continue to search for phenotype-genotype relationships with over 30 DSPP gene mutations reported to date. Investigators suggest that certain frameshift mutations lead to DD-II while others feel there is overlap in locations of mutations and the DGI-II and DD-II phenotypes. The specific molecular basis and mechanism for the differences in clinical phenotypes and severities that result from these allelic DSPP mutations are not known at this time.

Regulation of phosphorylation of DPP and DMP1 may provide a means of fine-tuning biomineralization processes and their related signaling events. Together with the established literature, the recent studies from our group confirm that actual phosphorylation machinery is present in cells which do not normally mineralize their matrix, and is not restricted to those cells classically associated with the generation of mineral networks. Establishing how, this regulation over mineralization is achieved will constitute a fruitful area for further investigation. In particular, further experiments are necessary to characterize site-specific phosphorylation events. It will also be of interest to compare the levels and specificity of protein phosphorylation in different cell types, as well as whether these specific phosphorylation events then provide points of control over DPP and DMP1-mediated signaling. Uncovering difference in post-translational modification between cells that mineralize their matrix compared to cells that do not mineralize their matrix is also expected to provide critical information about the uniqueness of cell processes designed for mineralized tissues. These studies would not only provide valuable information on the etiology of phosphorylation in aberrant mineralization events, but also broaden basic knowledge of how cells regulate their cellular processes.

Dentin sialoprotein (DSP) and dentin phosphoprotein (DPP) play distinct roles during dentinogenesis. Both dentin extracellular matrix proteins are cleaved products of dentin sialophosphoprotein (DSPP), an important odontoblastic differentiation marker. Mutations of the DSPP gene in human and mouse cause mineralization defects in teeth. During odontogenesis, DSPP is predominantly expressed in developing teeth whereas its expression is much lower in other tissues. Therefore, DSPP serves as a unique model to study the mechanisms of spatial-temporal and tissue-specific gene regulation associated with dentinogenesis. Both in vitro and in vivo promoter analysis studies have shown that DSPP transcription is controlled by a series of growth factors and transcriptional factors as well as local factors. Some factors up-regulate DSPP expression whereas others down-regulate its transcription. Spatialtemporal DSPP expression at different stages of tooth developmental is mediated by a complex network of growth factors and transcription factors. Beyond the role of biomineralization, recent studies have demonstrated that both DSP and DPP may function as intracellular transductors. DPP is able to induce tooth- and bone-related gene expressions via the MAPK and Smad signaling pathways while DSP binds to cell membrane proteins, resulting in activation of intracellular protein kinases. This data may provide novel insights into the roles of DSP and DPP in cell proliferation and differentiation besides biomineralization.

Dentin sialophosphoprotein (DSPP) cleavage products are the most abundant non-collagenous proteins in dentin. Among the genes encoding the dentin noncollagenous proteins, only mutations in DSPP are known to cause inherited dental malformations. Characterization of DSPP-derived protein from developing porcine teeth has allowed a better understanding of DSPP structure and function, which has provided important insights into how dentin biomineralization might be controlled. Porcine DSPP is a chimera of three structural domains: dentin sialoprotein (DSP), dentin glycoprotein (DGP) and dentin phosphoprotein (DPP). DSPP is expressed by odontoblasts and is rapidly processed after its secretion by BMP1 to generate DPP. Following this initial processing, DSPP-derived proteins are further processed by matrix metalloproteinases (MMP20 and MMP2) into DSP and DGP. DSP has a heavily N-glycosylated, sialylated N-terminal domain and a C-terminal domain with two glycosaminoglycan attachments, three potential O-glycosylations, and three potential phosphoserines. DGP is a phosphorylated glycoprotein that has an affinity for hydroxyapatite, while DPP is a highly phosphorylated intrinsically disordered protein with extensive length polymorphisms resulted from genetic heterogeneity.

Dentin sialophosphoprotein (DSPP), a major non-collagenous matrix protein of odontoblasts, is cleaved mainly into dentin sialoprotein (DSP) and dentin phosphoprotein (DPP). Both DSP and DPP have been implicated to have specific roles in dentin mineralization. We had earlier reported that that DSPP null mice display a phenotype in which teeth have widened predentin and irregular dentin mineralization resulting in sporadic unmineralized areas in dentin and frequent pulp exposure. Many of the earlier in vitro studies indicated that DPP plays a significant role in initiation and completion of dentin mineralization. We have recently reported our findings on DPP conditional knockout mice that indicate distinct roles of DSP and DPP in dentin mineralization: DSP is involved in regulating initiation of dentin mineralization, and DPP being involved in the maturation of mineralized dentin. In this review, we summarized lessons learned from mouse models of DSPP, DSP, and DPP.

DMP1encodes a serine-rich acidic protein and is a member of Small Integrin- Binding Ligand, N-linked, Glycoprotein family (SIBLING) which mineralizes tissues such as bones and teeth by binding strongly to hydroxyapatite. Dual functions of the DMP1 protein have been reported based on the in vitro studies. The nonphosphorylated cytoplasmic DMP1 protein translocates to the cell nucleus and initially acts as a transcriptional factor to enhance gene transcription of the osteoblast-specific genes such as alkaline phosphatase and osteocalcin and then moves out to the extracellular matrix during the osteoblast to osteocyte transition phase to promote mineralization and phosphate homeostasis. Loss of function mutations in the human DMP1 gene have been shown to cause autosomal recessive hypophosphatemic rickets in different ethnic groups. These variant mutations in DMP1 include deletions in exon 6, nucleotide substitution in the splice acceptor sequence of intron 2, and missense mutations in exons 2 or exon 3 that introduce the premature termination codon. The DMP1-null mouse also exhibits hypophosphatemic rickets which indicates the important role of the DMP1 gene in the development of normal bone formation. Most of the functional domains identified thus far are encoded by the sixth and last exon of DMP1.

The loss- and gain-of-function animal models provide powerful approaches with which to dissect out the in vivo role of DMP1 in control of the maturation of odontoblasts and osteoblasts, in mineralization, as well as in the regulation of phosphate homeostasis via direct and indirect mechanisms. The study of Dmp1 null hypophosphatemic rickets leads to identification of DMP1 mutations in humans (a novel autosomal recessive hypophosphatemic rickets, ARHR). Characterizations of Dmp1 null mouse model reveal many novel functions of the osteocyte, the cell counts for over 90% of all bone cells. These new roles include the systemic regulations of mineral and production of FGF23, and phosphate homeostasis. Last but not the least, improvement of Dmp1 null phenotype by administration of neutralizing FGF23 antibodies represents a promising new therapeutic approach for the anabolic treatment of ARHRs associated with DMP1 mutations in the future.

Bone sialoprotein (BSP) is a highly acidic phosphorylated glycoprotein found at high levels in mineralized tissues. Similar to other members of the SIBLING protein family, BSP is a multidomain, multifunctional adhesive protein with a flexible conformation and little secondary structure. BSP has been demonstrated to be involved in cell attachment and signaling, hydroxyapatite binding and nucleation, and collagen binding. It has also been proposed that the protein has angiogenic properties, promotes bone repair, enhances osteoclast formation, and mediates attachment of certain metastatic cancer cells to bone, among other putative functions. The flexible structure of BSP is believed critical in its ability to interact with multiple binding partners. Recent studies on the BSP-null mice have provided further evidence for a significant role for BSP in bone formation and remodeling.

Osteopontin OPN] has originally been identified from various sources, but initially research focused on its role in mineralization and bone homeostasis and remodelling. This work has established a critical role for OPN in regulating the mineralization of calcium phosphate, the inorganic component of most mineralized tissues such as bones and teeth, but also in bone resorption and in pathological calcifications such as atherosclerosis or urinary and kidney stone. The role of posttranslational modifications and proteolytic processing has been studied in some detail, and OPN has been recognized as an integrin-binding protein that is involved in multiple biological processes such as arterial calcification, cardiovascular disease, inflammation and cancer. An intracellular form of OPN, which may mediate distinct biological functions of the molecule, has been discovered. OPN was found to be directly involved in regulating inflammatory and immune responses in a vast number of pathologies, as well as the progression and metastasis of many cancers. Spurred by discoveries in the last decade, therapeutic, diagnostic and prognostic approaches have been developed, and OPN is arguably the SIBLING that receives the most attention from many areas in biomedical research outside the mineralized tissues. In this chapter we attempt to provide a general overview of the many new roles of this non-collagenous protein.

Matrix Extracellular Phosphoglycoprotein (MEPE) belongs to a group of proteins called Short Integrin Binding Ligand Interacting Glycoproteins (SIBLINGs). These proteins are all localized to a defined region on chromosome 5q in mice and chromosome 4q in humans. A unifying feature of SIBLING proteins is an Acidic Serine Aspartate Rich MEPE associated motif (ASARM). The ASARM-motif and SIBLINGs appeared 300 million years ago with terrestrial vertebrates coincident with the new environmental challenge of gravity. Recent research has shown that the ASARM-motif and the released ASARM-peptide also play regulatory roles in mineralization (bone and teeth), phosphate regulation, vascularization, soft tissue calcification, osteoclastogenesis, mechanotransduction and fat energy metabolism. The MEPE ASARM-motif and peptide are physiological substrates for PHEX, a Zn metalloendopeptidase. Defects in PHEX are responsible for X-linked hypophosphatemic rickets (HYP). There is evidence that PHEX interacts with another ASARM-motif containing SIBLING protein, Dentin Matrix Protein-1 (DMP1). DMP1 mutations cause bone-renal defects that are identical with the defects caused by loss of PHEX function. This results in autosomal recessive hypophosphatemic rickets (ARHR). In both HYP and ARHR increased FGF23 expression occurs and activating mutations in FGF23 cause autosomal dominant hypophosphatemic rickets (ADHR). ASARMpeptide administration in vitro and in vivo also induces increased FGF23 expression. Recent work indicates a competitive displacement of a PHEX DMP1 interaction by ASARM-peptide and thus increased FGF23 expression.

The deposition of biominerals within living systems is a tightly controlled complex process since the mineral type, mineral location and mineral size must all be specified. Hundreds, if not thousands, of studies have been made from many perspectives. This chapter is focused on the question of the initial steps of how the mineral crystals become specifically nucleated in particular places. This requires understanding of both the physics and physical chemical concepts of crystal growth and the composition and structure of the matrix of tissue in which the crystals become positioned. Our primary interest is the mineralization of bones and teeth, and thus the bone or tooth matrix comprised of collagen fibrils and many diverse associated non-collagenous components must be considered, but this is done in a very general way looking for commonalities in different systems, such as bone and dentin. Are there differences in the type I collagen, or in the non-collagenous proteins related to mineralization? A second theme is nucleation theory. Based on classical nucleation theory, alternative non-ideal theories are examined and suggest that there is probably a multistep mechanism, yet it may be that different systems permit different pathways. One pathway that appears to be commonly considered is that random fluctuations in supersaturated solution produce clusters of the relevant ions and these must reach a critical size in liquid-like state before crystallization takes place. These varied ideas are examined in the context of recent high resolution data on the mineralization of type I collagen fibrils.

Hydroxyapatite crystals formed in various tissues throughout the body have very specific shapes and sizes. These sizes, though heterogeneous, are specific to the tissue type and the age of the animal from which the crystals were derived. A multitude of factors contribute to the regulation of crystal sizes in different tissues. This chapter reviews those factors, and presents a brief introduction to the science of nucleation and crystal growth that governs the ways these crystals form and proliferate.

As cellular components, phospholipids (PLs) are signalling molecules implicated in the regulation of stress-induced cellular responses, cell differentiation, proliferation, and cell death. The chemical comparison between demineralized and undemineralized tissues provides evidences that a specific group of PLs is present as extracellular matrix (ECM) component, and related to the mineralization process. Matrix vesicles, extensively studied in the growth plate of cartilage, are also implicated in the initial stages of bone and dentin mineralization. They play a key role in the formation of calcium-phosphate-lipoprotein complexes (CPLX) found in most mineralizing tissues. Radioautographic investigations suggest the transfer of PLs originating in the blood serum. Using an intercellular pathway they migrate directly toward the mineralization front in dentin. Genetic and pharmacological PLs alterations lead to lysosomal storage diseases, which have apparently little effects on dentin and bone mineralization. As another model to investigate, the fro-/- mice clarify the role(s) of PLs in mineralization. Positional cloning of the mutation was found to be a deletion in Smpd3, the gene encoding sphingomyelin phosphodiesterase 3 (neutral sphingomyelinase-2). The mutation was found on mouse chromosome 8. The deletion of the intron 8 - exon 9 is producing a non-functional mutant protein. As a consequence of the mutation, the fro -/- mice display defective mineralization, i.e. a non-collagenous form of osteogenesis imperfecta associated with dentin dysplasia/dentinogenesis imperfecta. This suggests that enzymatic cleavage of ECM PLs may be a pre-requisite for dentin and bone mineralization.

Matrix metalloproteinases (MMPs) are a group of mammalian enzymes that in concert can degrade practically all extracellular matrix protein. They are also acknowledged important in the regulation of cell and tissue homeostasis, as they can degrade and modify several signaling molecules, e.g. cell surface receptors and growth factors. In hard tissues several MMPs are present, and the importance of some of them in the normal skeletal and dental development is well recognized. Still, the exact mode of function is not completely clear, as their expression and function seem to vary e.g. according to the location. Even though several other MMPs are also expressed in hard tissues, the current data has not been able to demonstrate their functional significance. This may be because their role is not absolutely essential to the normal growth and development of the calcified tissues. Alternatively, in the case of absence their function may be covered by other MMPs, as the substrate specificities of these enzymes are overlapping. This review concentrates on those MMPs which are known to have functional role in hard tissue formation, mineralization and/or remodeling.