piR-368 promotes odontoblastic differentiation of dental papilla cells via the Smad1/5 signaling pathway by targeting Smurf1.

PURPOSE: The important role of non-coding RNAs in odontoblastic differentiation of dental tissue-derived stem cells has been widely demonstrated; however, whether piRNA (a subclass of non-coding RNA) involved in the course of odontoblastic differentiation is not yet available. This study aimed to investigate the expression profile of piRNA during odontogenic differentiation of mDPCs and the potential molecular mechanism in vitro. MATERIALS AND METHODS: The primary mouse dental papilla cells (mDPCs) were isolated from the first molars of 1-day postnatal Kunming mice. Then, they were cultured in odontogenic medium for 9 days. The expression profile of piRNA was detected by Small RNA sequencing. RT-qPCR was used to verify the elevation of piR-368. The mRNA and protein levels of mineralization markers were examined by qRT-PCR and Western blot analysis. Alkaline phosphatase (ALP) activity and alizarin red S staining were conducted to assess the odontoblastic differentiation ability. RESULTS: We validated piR-368 was significantly upregulated and interference with piR-368 markedly inhibited the odontogenic differentiation of mDPCs. In addition, the relationship between Smad1/5 signaling pathway and piR-368-induced odontoblastic differentiation has been discovered. Finally, we demonstrated Smurf1 as a target gene of piR-368 using dual-luciferase assays. CONCLUSION: This study was the first to illustrate the participation of piRNA in odontoblastic differentiation. We proved that piR-368 promoted odontoblastic differentiation of mouse dental papilla cells via the Smad1/5 signaling pathway by targeting Smurf1.

Collagen analysis in human tooth germ papillae.

The extracellular matrix (ECM) performs a very important role in growth regulation and tissue differentiation and organization. In view of this, the purpose of this study was to analyze the collagen, the major organic component of dental pulp ECM, in papillae of human tooth germs in different developmental phases. The maxillas and mandibles of 9 human fetuses ranging from 10 to 22 weeks of intrauterine life were removed and 16 tooth germs (1 in the cap stage, 8 in the early bell stage and 7 in the late bell stage) were obtained. The pieces were processed for histological analysis and stained with hematoxylin-eosin, Masson's Trichrome and picrosirius staining technique. Both types of collagen in the dental papilla were only detected by the picrosirius staining technique under polarized light microscopy. Type III collagen was detected in all specimens. Type I collagen was present in focal areas of the dental papilla only in some specimens. In conclusion, the findings of this study showed that type III collagen is a regular component of the papillae of human tooth germs whereas type I collagen is present in a significantly lesser amount.

Collagen analysis in human tooth germ papillae

The extracellular matrix (ECM) performs a very important role in growth regulation and tissue differentiation and organization. In view of this, the purpose of this study was to analyze the collagen, the major organic component of dental pulp ECM, in papillae of human tooth germs in different developmental phases. The maxillas and mandibles of 9 human fetuses ranging from 10 to 22 weeks of intrauterine life were removed and 16 tooth germs (1 in the cap stage, 8 in the early bell stage and 7 in the late bell stage) were obtained. The pieces were processed for histological analysis and stained with hematoxylin-eosin, Masson's Trichrome and picrosirius staining technique. Both types of collagen in the dental papilla were only detected by the picrosirius staining technique under polarized light microscopy. Type III collagen was detected in all specimens. Type I collagen was present in focal areas of the dental papilla only in some specimens. In conclusion, the findings of this study showed that type III collagen is a regular component of the papillae of human tooth germs whereas type I collagen is present in a significantly lesser amount.

A matriz extracelular (MEC) tem um papel importante na regulação do crescimento e na diferenciação e organização dos tecidos. Com base nestes aspectos o objetivo do deste estudo foi analisar o colágeno, maior componente orgânico da MEC da polpa dentária, na papila de germes dentários humanos, em diferentes fases do desenvolvimento. Foram obtidos fragmentos de maxilas e mandíbulas de 9 fetos humanos com 10 a 22 semanas de vida intra-uterina, dos quais foram analisados 16 germes dentários (1 em estágio de capuz, 8 em estágio de campânula precoce e 7 em estágio de campânula tardia). Secções histológicas seriadas foram coradas com hematoxilina e eosina, tricrômico de Masson e técnica de coloração do picrosirius. Ambos os tipos de colágeno na papila dentária foram somente detectados pela técnica de coloração do picrosirius usando microscopia de luz polarizada. Colágeno tipo III foi detectado em todas as amostras. Colágeno tipo I estava presente em áreas focais da papila dental em algumas amostras. Concluiu-se que o colágeno tipo III mostrou-se um componente regular da papila de germes dentários humanos, enquanto o colágeno tipo I esteve presente em quantidade significativamente menor.

Ultrastructure of basement membranes in monkey and shark teeth at an early stage of development.

The basement membrane, which separates the inner enamel epithelium from the dental papilla in the early stages of tooth development, is known to play a significant role in odontogenesis. In this review article, this basement membrane was described in detail based on our recent findings with the use of high-resolution electron microscopy. Tooth germs of a monkey (Macaca fuscata) and a shark (Cephaloscyllium umbratile) were processed for thin-section observations. During the early stage of development, the basement membrane of the inner enamel (dental) epithelium was composed of a lamina lucida, lamina densa, and much wider lamina fibroreticularis. At higher magnification, the lamina densa in both species was made up of a fine network of cords, which are generally the main constituents of the basement membranes. In the monkey tooth, the lamina fibroreticularis was rich in fibrils, which were now characterized as basotubules, 10-nm-wide microfibril-like structures. The space between the basotubules was filled with a cord network that extended from the lamina densa. Dental papilla cell processes were inserted into the lamina fibroreticularis, and their surface was closely associated with numerous parallel basotubules via 1.5- to 3-nm-wide filaments. In the shark tooth during its early stage of development, the basotubules were absent in the lamina fibroreticularis and only narrow extensions, 60-90 nm wide and 1-2 microm long, of the cord network of the lamina densa were present. The dental papilla cells were immobilized by means of the binding of their processes to the extensions. These results indicate that basement membranes in both monkey and shark teeth at early stage of development are specialized for functions as anchoring and firm binding, which are essential for the successful differentiation of the odontoblasts.

Development and maturation of taste buds of the palatal epithelium of the rat: histological and immunohistochemical study.

Palatal taste buds are intriguing partners in the mediation of taste behavior and their spatial distribution is functionally important for suckling behavior, especially in the neonatal life. Their prenatal development has not been previously elucidated in the rat, and the onset of their maturation remains rather controversial. We delineated the development and frequency distribution of the taste buds as well as the immunohistochemical expression of alpha-gustducin, a G protein closely related to the transduction of taste stimuli, in the nasoincisor papilla (NIP) and soft palate (SP) from the embryonic day 17 (E17) till the postnatal day 70 (PN70). The main findings in the present study were the development of a substantial number of taste pores in the SP of fetal rats (60.3 +/- 1.7 out of 122.8 +/- 5.5; mean +/- SD/animal at E19) and NIP of neonatal rats (9.8 +/- 1.0 out of 44.8 +/- 2.2 at PN4). alpha-gustducin-like immunoreactivity (-LI) was not expressed in the pored taste buds of either prenatal or newborn rats. The earliest expression of alpha-gustducin-LI was demonstrated at PN1 in the SP (1.5 +/- 0.5 cells/taste bud; mean +/- SD) and at PN4 in the NIP (1.4 +/- 0.5). By age the total counts of pored taste buds continuously increased and their morphological features became quite discernible. They became pear in shape, characterized by distinct pores, long subporal space, and longitudinally oriented cells. Around the second week, a remarkable transient decrease in the total number of taste buds was recorded in the oral epithelium of NIP and SP, which might be correlated with the changes of ingestive behaviors. The total counts of cells showing alpha-gustducin-LI per taste bud gradually increased till the end of our investigation (14.1 +/- 2.7 in NIP and 12.4 +/- 2.5 in SP at PN70). We conclude that substantial development of taste buds began prenatally in the SP, whereas most developed entirely postnatal in the NIP. The present study provides evidence that the existence of a taste pore which is considered an important criterion for the morphological maturation of taste buds is not enough for the onset of the taste transduction, which necessitates also mature taste cells. Moreover, the earlier maturation of palatal taste buds compared with the contiguous populations in the oral cavity evokes an evidence of their significant role in the transmission of gustatory information, especially in the early life of rat.

Expression and localization of connexin 43 in rat incisor odontoblasts.

We have examined the expression and localization of connexin 43 (CX43) in rat incisor odontoblasts using reverse transcriptase polymerase chain reaction, in situ hybridization and immunohistochemistry. The CX43 gene was expressed in odontoblasts, and levels of gene expression increased throughout the course of development. In contrast, CX43 was down-regulated at an incisal segment. In situ hybridization analysis showed no positive signal for CX43 RNA in the cytoplasm of differentiating dental papilla cells, but faint positive signals for CX43 RNA were observed in early pre-odontoblasts. Those signals were more intense in young and in old odontoblasts, but were less in short odontoblasts. CX43 could not be detected in differentiating dental papilla cells or in early pre-odontoblasts by immunohistochemical localization, but a positive reaction was found in the late pre-odontoblast stage where predentin had been produced. The positivity gradually increased during odontoblast maturation, and was highest in the layer of old odontoblasts. These results indicate that odontoblasts that secrete actively dentin matrix components are tightly in contact with each other by gap junctions as suggested by the intense CX43.

Ultrastructural verification of anchoring role of lamina fibroreticularis of dental basement membrane in odontogenesis.

In a previous study of the developing tooth a characteristic fibrillar layer associated with the basement membrane of the inner enamel epithelium was found to be a highly specialized lamina fibroreticularis of the basement membrane which is unusually rich in basotubules, 10 nm wide microfibril-like structures. In this study this layer was further examined in detail in the hope of ultrastructurally elucidating its role in odontogenesis. Tooth germs of the monkey (Macaca fuscata) were processed for thin section observations. Dental papilla cell processes were inserted into the lamina fibroreticularis and their surface was closely associated with numerous parallel basotubules. With high-resolution observations the space between the surface and nearest basotubules as well as the spaces between neighbouring basotubules were bridged by 1.5-3 nm wide filaments running perpendicular to the axis of basotubules. These results indicate that the processes of dental papilla cells are linked to groups of basotubules by means of 1.5-3 nm wide filaments. Immunoperoxidase staining showed the presence of fibronectin along basotubules as well as within the space between the process and basotubule. This result, together with the comparison of these filaments with microfibril-associated 1.2-3 nm wide fibronectin filaments and the reported abundance of fibronectin in the basement membrane area during odontogenesis, indicates that these 1.5-3 nm wide filaments are composed of fibronectin. After immunostaining for amyloid P component, done with the rat tissue because of the nature of an available antiserum, basotubules in the lamina fibroreticularis were positively stained, as has been shown in basotubules/microfibrils in other locations. Microfibrils function as anchoring rods by interlinking connective tissue components to one another and to the cells. Basotubules, thought to be basement membrane-incorporated microfibrils, in the lamina fibroreticularis in this study are also likely to function as an anchoring device to immobilize dental papilla cells along the basement membrane. Such an arrangement of mesenchymal cells is known to be crucial for the successful differentiation of odontoblasts in the developing tooth.

Immunohistochemistry and in situ hybridization investigation of transforming growth factor-beta: during odontoblast and ameloblast differentiation.

OBJECTIVE: The purpose of this study was to investigate the expression of transforming growth factor-beta, TGF-beta 1 during odontoblast and ameloblast differentiation. METHODS: The analysis was made on 5-micron serial sections of the glutaraldehyde-fixed and paraffin-embedded mandibular first and second molars of a neonatal mouse. A specific and affinity-purified antibody to TGF-beta 1 and a digoxigenin-labeled cRNA probe were used. RESULTS: The results showed that the TGF-beta 1 was expressed in a spatial and temporal pattern in dental tissues. The immunoreactivity of dental tissues for TGF-beta 1 was consistent with the expression of TGF-beta 1 mRNA revealed by in situ hybridization before dentin matrix formation. The TGF-beta 1 was evenly expressed in dental papilla and inner enamel epithelium. The expression of TGF-beta 1 was increased in the layer of odontoblasts and ameloblasts and in the stratum intermedium with the formation of dentin matrix, where the staining was also observed. CONCLUSIONS: These findings suggest that TGF-beta 1 may have an important role in extracellular matrix formation and cytodifferentation.

Immunolocalization of the small proteoglycan decorin in human teeth.

The immunolocalization of decorin was studied by confocal laser scanning microscopy and transmission electron microscopy. In the apical area of developing teeth, labelling for decorin was found in the dental papilla cells, prodontoblasts and also in the Hertwig's epithelial cells. Mantle dentine and the initial predentine were negative. In circumpulpal dentine, intense reactivity extended along the calcification front and dentinal tubules. Fluorescence was also evident in odontoblast cell bodies and their processes in predentine. None was perceived, however, in the predentinal matrix. Faint staining was observed on the calcified dentinal matrix. Immunoelectron microscopy revealed staining for decorin in collagen fibrils lining the predentine-dentine junction, and where arrays of labelled filaments were noted orthogonal to the collagen fibrils. Staining extending from the calcification front was observed in the matrix adjacent to the dentinal tubule. The decorin observed at the calcification front might regulate the mineralization of dentinal matrix.

Recombination of epithelial root sheath and dental papilla cells in vitro.

Hertwig's epithelial root sheath (HERS) cells were isolated and recombined with ectomesenchymal cells in vitro utilizing extracellular matrix components as substrate. After 14 days in culture, HERS cells were differentiated and exhibited a stratified organization. These features resembled those observed in vivo as epithelial rests of Malassez. A mineralization process was also present in HERS cells, in which calcium salts were deposited. This mineralization was correlated with the strong immunoexpression of osteopontin by HERS. The results obtained add support to the possible role of HERS in the secretion of hypocalcified material on the root during early cementogenesis.

In vitro effects of retinoic acid on mouse incisor development.

The developing dentition is known to express the complete set of retinoic acid (RA) nuclear receptors and cytoplasmic RA-binding proteins (CRABPI and II), and RA is required for in vitro mouse molar morphogenesis, so the role of RA during in vitro mouse incisor development was investigated. Histological procedures, immunocytochemical detection of proliferating cells, immunofluorescence detection of laminin, and in situ hybridization with RNA probes for CRABPI and II were done on the tooth-germ cultures either in the presence or in the absence of RA. RA appeared to control initial morphogenesis, particularly the asymmetrical growth of the cervical loop, and to regulate required differential mitotic activity. RA seemed also to be involved in asymmetrical laminin deposition. The distribution of the CRABP gene transcripts was similar during in vivo and in vitro incisor development. However, CRABPI gene transcript distribution in the labial part of the epithelial loop was detected in vitro only in the presence of RA. A direct role of the CRABPs during tooth development is, however, unlikely because Ch55, a synthetic RA analogue that does not bind to CRABP, had the same effects as RA on in vitro incisor development.

Immunohistochemistry of the intercellular matrix components and the epithelio-mesenchymal junction of the human tooth germ.

The immunohistochemical localization of heparan sulphate, collagen type I, III and IV, laminin, tenascin, plasma- and cellular fibronectin was studied in tooth germs from human fetuses. The lamina basalis ameloblastica or membrana preformativa, which separates the pre-ameloblasts from the pre-dentin and dentin, contained heparan sulphate, collagen type IV, laminin and fibronectin. Enamel reacted with antifibronectin, but the reaction varied depending on the type of fibronectin and the source of antibody. In early pre-dentin, collagen type I, laminin, tenascin and fibronectin were present. In late pre-dentin and dentin collagen type I was found in intertubular dentin and in the zone between enamel and dentin. The close relationship between collagen type I in dentin and fibronectin in immature enamel is interesting, as it may contribute to the stabilization of the amelodentinal interface. In dental pulp, collagen type IV and laminin were found in the endothelial basement membranes. Collagen type I and III, tenascin and fibronectin were localized to the mesenchymal intercellular matrix. The results of this study have supported the assumption that the lamina basalis ameloblastica is a basement membrane, and have lead to the suggestion that ameloblasts are producers of fibronectin or a fibronectin-like substance.

A chondroitin sulfate epitope in mammalian dental pulp and its developmental expression in mouse dental papilla.

The molecular specificity of the dental papilla of a bell-stage tooth was studied by production of dental-papilla-reactive monoclonal antibodies (Mabs). One of the Mabs, designated 7C5, recognized an epitope present in glycosaminoglycan. Several lines of evidence suggested that the 7C5-epitope consists of chondroitin 6-sulfate. The Mab did not react with mouse dental epithelium, but reacted uniformly with mesenchymal tissue in the mandibular process and accumulated in the dental sac and in the papilla of bell-stage tooth germs. The 7C5-staining was lost from the differentiating odontoblasts, while the staining in the molar tooth papilla was accumulated in the subodontoblastic layer. In the developing mouse incisor, the 7C5-epitope was restricted to the lingual-posterior area. The 7C5-epitope was also present in pulpal tissue and predentin of different types of teeth of various mammalian species, including man, sheep, swine, and rat. Collagenase pre-treatment of tissue sections abolished the bulk of the 7C5-reactivity in peridental mesenchyme during embryonic stages while leaving the staining of the dental papilla intact. In newborn and adult teeth, collagenase also impaired the reactivity in the pulp except for the subodontoblastic layer. This suggests the existence of different subpopulations of the 7C5-epitope containing proteoglycans in dental papilla and pulp. A high-molecular-weight proteoglycan, sensitive to chondroitinase ABC but not to heparinase or heparitinase, was immunoprecipitated by 7C5 from extracts of bell-stage mouse tooth germs. We suggest that the evolutionary conservation of chondroitin 6-sulfate in the dental pulp reflects its properties as non-terminally differentiated tissue and perhaps the retention of a potential to differentiate to odontoblasts.

Expression of intermediate filaments and actins in human dental pulp and embryonic dental papilla.

The localization of different cytoskeletal proteins (keratin, vimentin, desmin, actin, and alpha-smooth muscle actin) was examined by immunohistochemistry in normal human adult dental pulp and compared with dental papilla of tooth germs. Keratin and actin were localized in enamel organ. Vimentin and actin were observed in the dental papilla and in the adult dental pulp. Desmin and alpha-smooth muscle actin were present only in the vessel walls. These data are discussed paying particular attention to the origin and the peculiar functional characters of the dental papilla and pulp.

Investigation of the role of Von Korff fibers during murine dentinogenesis.

The existence and significance of Von Korff fibers during early dentinogenesis are still very controversial. The purpose of the present study was to re-examine the questions of the existence, nature and significance of Von Korff's fibers using light microscopy and immunohistochemistry. Specimens were obtained from 3 days-old CD-I mice and mandibles were carefully dissected under constant irrigation and immediately fixed in 10% neutral buffered formalin for light microscopy. Sections were treated or not with collagenase prior to silver staining. For immunohistochemistry, specimens were fixed in 95% ethanol and embedded in paraffin. Sections were reacted with goat anti-human-bovine type I or type III collagen and a rhodamine (RITC) labelled rabbit anti-goat IgG was then reacted as a secondary antibody. Slides were then examined under a Zeiss II photomicroscope equipped with epifluorescence. Our results have confirmed the presence of argyrophilic material concentrated at the periphery of the dental papilla and stretching from the subodontoblastic layer to the future dentino enamel junction. The distribution of type III collagen was very similar to the distribution of the silver staining at the cervical loop area. Type I collagen distribution was different and concentrated in areas where odontoblasts were fully differentiated. Our study showed that Von Korff fibers are not artefactual. We have established the presence of an apical compartment containing type I collagen fibers and a basal compartment containing type III collagen to explain the image of continuous silver staining crossing the entire thickness of the odontoblast layer.

Immunocytochemical localization of growth hormone receptor in rat maxillary teeth.

To address the question of what role growth hormone may have in stimulating tooth formation, the distribution of its receptor/binding protein in developing rat incisors and molars was studied immunocytochemically using well-characterized monoclonal antibodies. Ten female 45-day-old Wistar rats were perfused with 4% paraformaldehyde. Five-microns paraffin sections of the growing end of maxillary incisors and molars were cut, deparaffinized and incubated with mouse anti-growth hormone receptor antibodies or control antibodies. A three-layer streptavidin peroxidase technique was used to detect bound antibody. Immunoreaction product was associated primarily with the cytoplasm of cells at certain stages of differentiation. Dividing cells, differentiating preameloblasts and preodontoblasts, secretory ameloblasts and odontoblasts showed immunoreactivity. Undifferentiated dental epithelium cells, stellate reticulum, external dental epithelial cells, mature odontoblasts, and most of cells in the dental papilla were non-reactive. However, at certain stages of tooth development, the stratum intermedium and the external dental epithelium also stained positively. The presence of growth hormone receptor/binding protein in tooth cells at different stages of their development indicates that growth hormone may influence cell proliferation, differentiation and differentiated functions of ameloblasts, odontoblasts and cementoblasts independent of a systemic mediator, and thus may be involved in stimulating odontogenesis directly.

Colocalization of TGF-beta 1 and extracellular matrix proteins during rat tooth development.

Recent studies suggest that TGF-beta 1 controls events in embryogenesis via its modulation of the extracellular matrix (ECM). During tooth development, histomorphogenetic and cytodifferentiation events are regulated by ECM-mediated epithelial-mesenchymal interactions. In this study, we have investigated the in vivo relationship of the localization of TGF-beta 1 to various ECM proteins, including glycosaminoglycans (GAGs), fibronectin, collagens I and III and to the enzyme alkaline phosphatase, at the early and late bell stages of rat molar tooth development. We have used polyclonal antibodies specific to peptides present in the precursor and processed forms of TGF-beta 1, fibronectin and Collagens I and III and immunohistochemical methods of detection. GAGs were stained with Alcian Blue GX and alkaline phosphatase was detected with the enzyme Naphthol-AS B1. Our results indicate that both the intracellular and extracellular forms of TGF-beta 1 are expressed in a site-specific manner, and that the growth factor colocalizes with alkaline phosphatase, fibronectin, collagens I and III and GAGs during the bell stages of tooth development. These data support the hypothesis that in tooth development, the actions of TGF-beta 1 may contribute to important qualitative changes in the composition and structure of extracellular matrices.

Histochemical localization of transforming growth factor-beta 1 in developing rat molars using antibodies to different epitopes.

In this study transforming growth factor-beta 1 (TGF-beta 1) has been immunolocalized in developing rat molars using two well characterized polyclonal antibodies, Anti-CC and Anti-LC, that recognize extracellular and intracellular TGF-beta 1, respectively. With immunohistochemical methods and the ABC-peroxidase system of detection, the growth factor was immunolocalized within the ectodermally derived enamel organ and the neural crest-derived dental papilla at the early and advanced bell stages of development. With Anti-CC, widespread and abundant extracellular TGF-beta 1 was found associated with the stellate reticulum and within central and apical regions of dental papilla mesenchyme. In contrast, Anti-LC localized TGF-beta 1 intensely within the cells of the outer dental epithelium. Moderate immunostaining for TGF-beta 1 with Anti-LC was also evident within the apical cytoplasm of inner dental epithelial cells and odontoblasts. These findings support the hypothesis that TGF-beta 1 may play a paracrine role in tooth development by regulating the epithelial-mesenchymal interactions that influence growth and cytodifferentiation events.

Expression of attachment proteins during cementogenesis.

There is general agreement that during development the extracellular environment plays a critical role in controlling cell differentiation. Data generated from numerous studies support the possibility that cell attachment proteins and their corresponding cell receptors are possible candidates for this role. In particular, our studies are directed at identifying attachment proteins in mature cementum and establishing the function of these proteins during root formation. Fractionation of guanidine HCL/EDTA extracts of cementum revealed the presence of a bone-associated attachment, BSP, as well as fractions containing as of yet undetermined attachment proteins. Immunofluorescent examination of 1st molar tissues during root formation, obtained from 7 day-old mice neonates, for bone-associated attachment proteins indicated that osteopontin is expressed in the area of Hertwig's epithelial root sheath, but not in the region of the dental papillae. However, dental papillae cells, considered to have the capacity to form cementum, attached to osteopontin coated dishes, in vitro. Thus, unique attachment proteins, as well as those previously identified, were found in mature cementum and during root development. Future studies focused on identifying attachment proteins of mature cementum and determining the spatial and temporal localization of these proteins, pre- and post-cementogenesis, will provide important information necessary for establishing the function of these proteins during root development.

Changes in ultrastructural distribution of dental basement membrane heparan sulfate during early mouse tooth development.

With the invagination of the deep surface of the tooth buds of mouse embryos (15 days of gestation), heparan sulfate in the region of the lamina lucida of the dental basement membrane began to disappear. Subsequently, as the outline of the future dentino-enamel junction, which elaborates the molar cusp pattern, was formed (16-17 days of gestation), heparan sulfate disappeared not only in the lamina lucida but also at the external side of the lamina densa. However, prior to or during predentin matrix deposition (18 days of gestation), heparan sulfate reappeared in the dental basement membrane.