Expression of cytokeratins in enamel organ, junctional epithelium and epithelial cell rests of Malassez.

BACKGROUND AND OBJECTIVE: After tooth formation is complete, it is suggested that continuity exists between the epithelial cell rests of Malassez (ERM), reduced enamel epithelium (REE) and subsequently the junctional epithelium. However, the junctional epithelium was reported to differ from REE and ERM. The developmental relationships between and among them remain controversial. Therefore, in the present study we examined the expression of cytokeratins in the three types of epithelia to investigate the epithelial phenotypes. MATERIAL AND METHODS: The maxillae of Wistar rats, 1, 2, 3 and 7 wk of age, were used, and the expression of CK14, CK17, CK19, CK10/CK13 and AE1/AE3 was detected using immunoperoxidase techniques. RESULTS: There was negative staining for CK10/CK13 in all the epithelia. ERM stained strongly for AE1/AE3, CK14, CK17 and CK19. During the transformation of inner enamel epithelial (IEE) cells into reduced ameloblasts and subsequently into junctional epithelium, strong staining for CK14 was evident in IEE, REE and junctional epithelium, whereas the expression of AE1/AE3 and of CK19 were initially negative in IEE and then strong in REE and junctional epithelium, respectively. In particular, the expression of CK17 was strongly positive in ERM and REE, but was negative in IEE and junctional epithelium. CONCLUSION: ERM are of odontogenic origin and junctional epithelium has an epithelial phenotype different from REE and ERM. This is the first report to demonstrate that CK17 can be used as a marker to distinguish junctional epithelium from ERM.

[Enamel: a unique self-assembling in mineral world]. / L'émail - Un autoassemblage unique dans le monde du minéral.

Enamel is a unique tissue in vertebrates, acellular, formed on a labile scaffolding matrix and hypermineralized. The ameloblasts are epithelial cells in charge of amelogenesis. They secrete a number of matrix proteins degraded by enzymes during enamel mineralization. This ordered cellular and extracellular events imply that any genetic or environmental perturbation will produce indelible and recognizable defects. The specificity of defects will indicate the affected cellular process. Thus, depending on the specificity of alterations, the teratogenic event can be retrospectively established. Advances in the field allow to use enamel defects as diagnostic tools for molecular disorders. The multifunctionality of enamel peptides is presently identified from their chemical roles in mineralization to cell signaling, constituting a source of concrete innovations in regenerative medicine.

Dental follicle cells and cementoblasts induce apoptosis of ameloblast-lineage and Hertwig's epithelial root sheath/epithelial rests of Malassez cells through the Fas-Fas ligand pathway.

Hertwig's epithelial root sheath (HERS), epithelial rests of Malassez (ERM) cells, and reduced ameloblasts undergo apoptosis during tooth development. This study examined the effects of dental follicle cells and cementoblasts on the apoptosis of ameloblast-lineage and HERS/ERM cells derived from the enamel organ. We also elucidated the induction pathways and identified the apoptotic pathway involved in this process. Here, we showed terminal deoxynucleotidyl transferase-mediated biotin-dUTP nick-end labeling (TUNEL)-positive HERS cells and reduced ameloblasts near dental follicle cells during tooth development. Co-culturing ameloblast-lineage cell line (ALC) ameloblasts and HERS/ERM cells with either dental follicle cells or OCCM-30 cementoblasts markedly enhanced the apoptosis of ameloblasts and HERS/ERM cells compared with cells cultured alone. However, dental follicle cells and cementoblasts did not modulate the apoptotic responses of co-cultured non-odontogenic MCF10A or KB cells. When ameloblasts + HERS and cementoblasts + dental follicle cells were co-cultured, the expression of Fas ligand (FasL) increased in cementoblasts + dental follicle cells, while the expression of Fas increased in ameloblasts + HERS. Interestingly, recombinant FasL induced ameloblast apoptosis while the cementoblast-induced ameloblast apoptosis was suppressed by the Fas/FasL antagonist Kp7-6. These results suggest that during tooth development, dental follicle cells and cementoblasts induce apoptosis of ameloblast-lineage and HERS/ERM cells through the Fas-FasL pathway, but do not induce the apoptosis of non-odontogenic epithelial cells.

Enamel pits in hamster molars, formed by a single high fluoride dose, are associated with a perturbation of transitional stage ameloblasts.

Excessive intake of fluoride (F) by young children results in the formation of enamel subsurface porosities and pits, called enamel fluorosis. In this study, we used a single high dose of F administered to hamster pups to determine the stage of ameloblasts most affected by F and whether pit formation was related to F-related sub-ameloblastic cyst formation. Hamster pups received a single subcutaneous injection of either 20 mg or 40 mg NaF/kg body weight, were sacrificed 24 h later, and the number of cysts formed in the first molars were counted. Other pups were sacrificed 8 days after F injection, when the first molars had just erupted, to score for enamel defects. All F-injected pups formed enamel defects in the upper half of the cusps in a dose-dependent way. After injection of 20 mg NaF/kg, an average of 2.5 white spots per molar was found but no pits. At 40 mg NaF/kg, almost 4.5 spots per molar were counted as well as 2 pits per molar. The defects in erupted enamel were located in the upper half of the cusps, sites where cysts had formed at the transition stage of ameloblast differentiation. These results suggest that transitional ameloblasts, located between secretory- and maturation-stage ameloblasts, are most sensitive to the effects of a single high dose of F. F-induced cysts formed earlier at the pre-secretory stage were not correlated to either white spots or enamel pits, suggesting that damaged ameloblasts overlying a F-induced cyst regenerate and continue to form enamel.

Reptilian tooth development.

Dental patterns in vertebrates range from absence of teeth to multiple sets of teeth that are replaced throughout life. Despite this great variation, most of our understanding of tooth development is derived from studies on just a few model organisms. Here we introduce the reptile as an excellent model in which to study the molecular basis for early dental specification and, most importantly, for tooth replacement. We review recent snake studies that highlight the conserved role of Shh in marking the position of the odontogenic band. The distinctive molecular patterning of the dental lamina in the labial-lingual and oral-aboral axes is reviewed. We explain how these early signals help to specify the tooth-forming and non-tooth forming sides of the dental lamina as well as the presumptive successional lamina. Next, the simple architecture of the reptilian enamel organ is contrasted with the more complex, mammalian tooth bud and we discuss whether or not there is an enamel knot in reptilian teeth. The role of the successional lamina during tooth replacement in squamate reptiles is reviewed and we speculate on the possible formation of a vestigial, post-permanent dentition in mammals. In support of these ideas, we present data on agamid teeth in which development of a third generation is arrested. We suggest that in diphyodont mammals, similar mechanisms may be involved in reducing tooth replacement capacity. Finally, we review the location of label-retaining cells and suggest ways in which these putative dental epithelial stem cells contribute to continuous tooth replacement.

Wnt-RhoA signaling is involved in dental enamel development.

Transgenic mice that express dominant-negative RhoA (RhoA(DN) ) in ameloblasts have hypoplastic enamel with defects in molar cusps. ß-catenin and Wnt5a were up-regulated in enamel organs of RhoA(DN) transgenic mice, which indicated that both canonical and non-canonical Wnt pathways are implicated in the process of enamel defect formation. It was hypothesized that expression of RhoA(DN) in ameloblasts interfered with normal enamel development through the pathways that were induced by fluoride. The Wnt and RhoA pathways were further investigated in an ameloblast-lineage cell line (ALC) by treatment with sodium fluoride (NaF). The activities of RhoA and Rho-associated protein kinase (ROCK) II decreased significantly by 8-12 hours, similar to decreased activity in RhoA(DN) transgenic mice. Both canonical and non-canonical Wnt pathways were activated by treatment with NaF, which was verified by western blotting and the ß-catenin-TCF/LEF (T cell factor lymphanoid/enhancer factor) reporter gene (TOPflash) assay. ß-catenin localization to both cytoplasm and nucleus was up-regulated in NaF-treated ALC, while Gsk-3ß, the negative regulator of the Wnt pathway, showed a decreased pattern of expression. The current results indicate that both Wnt and RhoA pathways are implicated in fluoride-induced signaling transductions in the ALC as well as in the development of enamel defects in RhoA(DN) transgenic mice.

A multi-scale, discrete-cell simulation of organogenesis: Application to the effects of strain stimulus on collective cell behavior during ameloblast migration.

A multi-scale strategy is presented for simulating organogenesis that uses a single cell response function to define the behavior of individual cells in an organ-scale simulation of a large cell population. The response function summarizes detailed information about the behavior of individual cells in a sufficiently economical way that the organ-scale model can be commensurate with the entire organ. The first application demonstrates the effects of strain stimulus on the migration of ameloblasts during enamel formation. Ameloblasts are an attractive study case because mineralization preserves a complete record of their migratory paths. The response function in this case specifies the motions of cells responding to strain stimuli that propagate through the population. The strain stimuli are related to the curvature of the surface from which the ameloblasts migrate (the dentin-enamel junction or DEJ). A single unknown rate parameter is calibrated by an independent datum from the human tooth. With no remaining adjustable parameters, the theory correctly predicts aspects of the fracture-resistant, wavy microstructure of enamel in the human molar, including wavelength variations and the rate of wave amplitude damping. At a critical value of curvature of the DEJ, a transition in the ordering of cells occurs, from invariant order over the whole population to self-assembly of the population into groups or gangs. The prediction of an ordering transition and the predicted critical curvature are consistent with gnarled enamel in the cusps of the human molar. The calibration of the model using human data also predicts waves in the mouse incisor and an ordering transition at the chimpanzee cingulum. Widespread compressive strain is predicted late in the migration for both the human molar and mouse incisor, providing a possible signal for the termination of amelogenesis.

Are Hertwig's epithelial root sheath cells necessary for periodontal formation by dental follicle cells?

OBJECTIVE: The role of Hertwig's epithelial root sheath (HERS) cells in periodontal formation has been controversial. This study aimed to further clarify whether HERS cells participate in formation of the periodontium, and the necessity of HERS cells in differentiation of dental follicle cells (DFCs) for periodontal regeneration. DESIGN: HERS cells and DFCs were isolated and identified from post-natal 7-day Sprauge-Dawley rats. In vitro, direct co-culture of HERS cells and DFCs as well as the individual culture of HERS and DFCs were performed and followed by alizarin red staining and the quantitative real-time polymerase chain reaction analysis. For in vivo evaluation, the inactivated dentin matrix (iTDM) was fabricated. HERS cells and DFCs were seeded in combination or alone on iTDM and then transplanted into the rat omentum. Scanning electron microscope and further histological analysis were carried out. RESULTS: In vitro, mineral-like nodules were found in the culture of HERS cells alone or HERS + DFCs either by alizarin red staining or scanning electronic microscope. The mineralization and fiber-forming relevant mRNA expressions, such as bone sialoprotein, osteopontin, collagen I and collagen III in HERS + DFCs were significantly higher than that of the HERS or DFCs alone group. After transplantation in vivo, cementum and periodontal ligament-like tissues were formed in groups of HERS + DFCs and HERS alone, while no evident hard tissues and attached fibers were found in DFCs alone. CONCLUSIONS: Hertwig's epithelial root sheath cells directly participate in the formation of the periodontium, and they are essential for the differentiation of dental follicle cells to form periodontal structures. The combination use of Hertwig's epithelial root sheath cells and dental follicle cells is a promising approach for periodontal regeneration.

From conical to spatulate: intra- and interspecific changes in tooth shape in closely related cichlids (Teleostei; Cichlidae: Eretmodini).

The Eretmodini are closely related cichlids endemic to Lake Tanganyika with very divergent oral tooth shapes, ranging from spatulate in Eretmodus to conical in Tanganicodus. To study how closely related cichlids can generate such divergent tooth shapes, we investigated how the enamel organ directs the development of spatulate teeth in Eretmodus cf. cyanostictus (lineage A), both in ontogeny and in adults, and of conical teeth in adult Tanganicodus cf. irsacae, using 3D-reconstructions from serially sectioned tooth germs. The spatulate oral tooth shape that characterizes adult E. cf. cyanostictus (lineage A) is preceded early in ontogeny by a conical tooth shape. We propose two possible hypotheses to account for changes in the folding of the enamel organ (in particular its epithelio-mesenchymal boundary) capable of generating such distinct tooth shapes. Different arguments lead us to favor the hypothesis of an asymmetric growth and differentiation of the enamel organ, such that the tip of a conical tooth corresponds to one "corner" of a spatulate tooth. Applying current molecular models of tooth shape variation, this would imply the existence of asymmetric fields of inhibition. Whether such asymmetric growth reflects the reutilization of a simple mechanism operating in ontogeny has to be clarified.

Odontoblast ß-catenin signaling regulates fenestration of mouse Hertwig's epithelial root sheath.

The interaction between Hertwig's epithelial root sheath (HERS) and the adjacent mesenchyme is vitally important in mouse tooth root development. We previously generated odontoblast-specific Ctnnb1 (encoding ß-catenin) deletion mice, and demonstrated that odontoblast ß-catenin signaling regulates odontoblast proliferation and differentiation. However, the role of odontoblast ß-catenin signaling in regulation of HERS behavior has not been fully investigated. Here, using the same odontoblast- specific Ctnnb1 deletion mice, we found that ablation of ß-catenin signaling in odontoblasts led to aberrant HERS formation. Mechanistically, odontoblast-specific Ctnnb1 deletion resulted in elevated bone morphogenetic protein 7 (Bmp7) expression and reduced expression of noggin and follistatin, both of which encode extracellular inhibitors of BMPs. Furthermore, the levels of phosphorylated Smad1/5/8 were increased in HERS cells. In vitro tissue culture confirmed that BMP7 treatment disrupted the HERS structure. Taken together, we demonstrated that odontoblast ß-catenin signaling may act through regulation of BMP signaling to maintain the integrity of HERS cells.

Tooth enamel: current state of the art.

Important progress has been made relative to the growth, structure and function of enamel. Better understanding of the epithelial mesenchymal interactions during odontogenesis has been gained through tissue culture, and the predominant role of the dental papilla has been established. Differences between rodent and human amelogenesis have been demonstrated. With radioautography and cytochemistry, a significant amount of new information has been obtained on the metabolism of the ameloblasts, concerning the synthesis of proteins, glycoproteins and proteoglycans, as well as calcium transport. Numerous biochemical investigations have been devoted to developing and mature enamel matrix. The organic components of human adult enamel are mainly constituted of lipids and proteins, but further investigations are still needed to elucidate their precise nature. The so-called key-hole configuration of adult enamel can be questioned when amelogenesis is considered since the tissue does not develop in a prismatic head-tail fashion. The most important results have probably been obtained in the field of individual enamel apatite crystals shape and ultrastructure as well as in the description of the precise patterns of their carious dissolution which bears great similarities to the dissolution of synthetic apatites in acids.

Re-regeneration of lower jaws and the dental lamina in adult urodeles.

Transverse amputations were carried out through one-third fully regenerated jaw segments and through normal tissue of the mandible on the same and opposite sides of the jaw in adults of Notophthalmus viridescens. Collectively the results suggest that, in adult urodeles, the mandible and the dental lamina can be replaced in an identical manner more than one time. Although the major histological events are the same in jaw regeneration and re-regeneration, regrowth is more rapid in re-regeneration. It appears that recently differentiated tissues of the regenerate have a higher capacity for regeneration than normal tissues amputated for the first time. Re-regeneration of the jaw occurs by growth of the original regenerate cartilage which has undergone reorganization. In re-regeneration, the skeletal elements exhibit no polarity and regrowth occurs in both directions, while the dental lamina possesses an anterior-posterior polarity and can regrow in an anterior direction only. Information concerning the mechanisms involved in the regenerative events remain to be determined.

Immunohistochemical study of structural proteins in developing junctional epithelium.

The objective of this study was to investigate structural proteins of ameloblasts, reduced enamel epithelium and junctional epithelium using antikeratin antibodies to determine if keratin expression changes as ameloblasts differentiate to cells of junctional epithelium. Block sections of fully erupted, partially erupted and unerupted molar and incisor teeth were removed at necropsy from monkeys age 2, 5, 6 and 7 years. Tissues were fixed in Carnoy's fixative, decalcified in Kristensen's solution for 4 to 8 weeks, embedded in paraffin and sectioned at 5 mu. Representative sections were stained with hematoxylin and eosin. Remaining sections were stained with antikeratin antibodies AE1, AE3, 34 beta E12, 35 beta H11 and anti-40kD using the avidin-biotin-conjugated immunoperoxidase method. Results were analyzed qualitatively for presence or absence of staining products. Antibody 35 beta H11 reacted with ameloblasts and reduced enamel epithelium during the early maturation phase of tooth development but not with the reduced enamel epithelium of erupting teeth or with junctional epithelium. The anti-40kD antibody stained ameloblasts and reduced enamel epithelium of unerupted teeth and teeth erupting through oral epithelium; however, junctional epithelium of partially erupted and fully erupted teeth were unstained. Antibodies AE1, AE3 and 34 beta E12 stained ameloblasts and all layers of reduced enamel epithelium and junctional epithelium in all stages of development. These results suggest a relationship between expression of structural proteins and changes during development and differentiation of ameloblasts to junctional epithelial cells.

Ultrastructure of the dental epithelium during enameloid mineralization in a teleost fish, Cichlasoma cyanoguttatum.

Secretory-stage inner dental epithelial cells (IDE) of tooth buds deposited an unmineralized, ectodermally-derived, enameloid collagen matrix. Pharyngeal plates bearing tooth buds were fixed: some were demineralized, others treated with guanidine-EDTA, then fixed and post-fixed in osmium tetroxide with potassium ferricyanide. Thin Epon sections were viewed in a Jeol 100B TEM. Nascent enameloid crystals were orientated parallel to the collagen fibres and attained widths of 200 nm. Enameloid collagen was absent in demineralized mature enameloid. The outer dental epithelial plasma membrane was deeply invaginated forming extensive channels associated with elongated fuzzy-coated vesicles. Four configurations of IDE cells were characterized by cellular constituents, including elongated granules, Golgi complexes, multivesicular bodies, large electron-dense granules and extracellular amorphous material which was also adjacent to cells containing few organelles associated with protein synthesis, within infoldings of ruffled apical membranes and multivesicular bodies. This material was considered to be resorbed enameloid collagen, not a secretory product.

The effect of partial damage to the enamel-related periodontium combined with root resection on eruption of the rat incisor eruption.

Previous work has indicated that the enamel-related periodontium (ERP) has a role in the eruptive process of the rat lower incisor. By combining partial damage of this tissue with resection of the odontogenic organ, we examined the effect of the damage on subsequent incisor eruption. The connective tissue of the enamel-related periodontium was regenerated in less than 2 weeks, showing morphology close to normal. The injured part of the enamel organ was neither regenerated nor repaired, and a cement-like tissue, continuous with the true acellular cement, was formed on the denuded enamel. Before tooth exfoliation, the operated teeth erupted at a slower rate compared with root-resected and sham-operated incisors, probably because of the absence of a substantial part of the enamel organ due to surgical damage. As with the coronal dental follicle and the enamel organ in rat molars, the enamel-related periodontium and the enamel organ of rat incisors may have some control on their eruptive process.

Development of periodontal ligament and alveolar bone in homografted recombinations of enamel organs and papillary, pulpal and follicular mesenchyme in the mouse.

Mandibular first-molar tooth germs were dissected from 16-day-embryo and new-born CD1 mice. By incubation in collagenase they were separated into enamel organ, papilla and follicle. Dental pulp was obtained from mandibular first molars of 3-, 7- and 10-day-old mice. Various combinations of epithelial and mesenchymal tissues were grown for periods up to four weeks in the anterior chamber of the eye of homologous adult male mice. Recombinations of enamel organ and papilla formed teeth with regeneration of the investing layer of follicle and a root-related periodontal ligament, but no formation of alveolar bone. Bone only formed in those grafts which also included follicle. Recombinations of enamel organ and pulp produced dysplastic dentine with no enamel formation or proper tooth development. It was impossible, therefore, to assess whether the potential to regenerate an investing layer extends to the pulp later in development. At an earlier stage, however, the papillary mesenchyme has the ability to regenerate investing-layer cells which lack the capacity to form bone.

Transforming growth factor-beta 1 mRNA in neonatal ovine molars visualized by in situ hybridization: potential role for the stratum intermedium.

Human dentine contains relatively large amounts of transforming growth factor-beta (TGF-beta), which might originate from odontoblasts. The expression of the TGF-beta 1 message in developing teeth was examined by in situ hybridization. The analysis was made on 5-microns serial sections of mandibular third molars of neonatal sheep cut from tissues that had been fixed in glutaraldehyde and paraffin-embedded. A 35S-labelled cRNA probe, complementary to TGF-beta 1 mRNA, was constructed from human TGF-beta 1 cDNA. Northern analysis of total RNA from sheep placenta and neonatal third molars demonstrated hybridization to a single 2.4 kb TGF-beta 1 transcript from both tissues, indicating cross-reactivity of the human probe in the sheep. In the neonatal molars, in situ hybridization was observed in cells of the inner enamel epithelium, mature ameloblasts and mature odontoblasts, but not within preodontoblasts before dentine matrix formation. TGF-beta 1 mRNA expression was also evident in the cells of the dental papilla but scarcely so in the stellate reticulum. The most striking feature was the appearance of hybridization signal in the cells of the stratum intermedium before hybridization was evident in the inner enamel epithelium. Control sections incubated with RNAase before incubation with probe did not show evidence of hybridization. These findings suggest that TGF-beta 1 may have an important regulatory role in the differentiation of ameloblasts and odontoblasts, perhaps by modulating matrix formation during amelogenesis or odontogenesis. They also suggest a potential novel regulatory role for the cells of the stratum intermedium.

The distribution of cellular retinoic acid-binding protein I during odontogenesis in the rat incisor.

Retinoids are important molecules in various aspects of embryological development. Here the distribution of cellular retinoic acid-binding protein I (CRABPI) was studied in the continuously growing incisor of adult rats using an affinity-purified rabbit polyclonal antibody. CRABPI was present throughout the presecretory and secretory ameloblast layer. The protein disappeared from that layer during its maturation phase. The adjacent dental mesenchyme of the developing pulp stained positively for CRABPI, especially in the layer immediately beneath the fully differentiated odontoblasts. Little CRABPI was present in the odontoblast layer itself. The distribution of CRABPI, both in the undifferentiated basal region of the incisor tooth and associated with the cells during hard-tissue formation, suggests a role for this molecule during differentiation and hard-tissue genesis.

The immunohistochemical localization of Bax and Bcl-2 and their relation to apoptosis during amelogenesis in developing rat molars.

Bax and Bcl-2 are members of a family of intracellular, membrane-associated proteins that regulate programmed cell death. It has been suggested that, when Bax predominates, programmed cell death is accelerated and the apoptosis inhibitory activity of Bcl-2 is suppressed. The present study was undertaken to immunohistochemically (IHC) localize Bax and Bcl-2 in the cells of the enamel organ during amelogenesis in rat molars. Also, apoptotic cells were detected by TUNEL staining. The IHC intense localization of Bcl-2 and light staining for Bax in the pre-ameloblasts suggest that apoptosis is inhibited in the proliferating pre-ameloblasts. This is consistent with an absence of TUNEL staining for apoptosis in these cells. However, in the late secretory and transition ameloblasts, and adjacent stratum intermedium, evidence of apoptosis of the ameloblasts was observed. Bax and Bcl-2 were co-localized in the proximal ends of late secretory, transition and early maturation-stage ameloblasts, but immunoreactivity for Bax markedly increased in the proximal ends of late secretory and transition ameloblasts, while the Bcl-2 staining appeared to be lighter. This suggests that Bax antagonized Bcl-2 function, limiting the ability of Bcl-2 to prolong cell survival. In the early maturation stage, Bax staining faded while the immunoreactivity for Bcl-2 increased. Evidence of distinct apoptosis was reduced in the early maturation stage ameloblasts. When related to the occurrence of apoptosis during amelogenesis, the relative intensity of expression of Bax and Bcl-2 changed in a pattern consistent with that observed in other cell lines. This indicates that these proteins play essential roles in the process of amelogenesis, as predicted by their proposed mechanisms of action in the control of apoptosis.

Fate of horseradish peroxidase in the secretion zone of the rat incisor enamel organ.

Adult CDF albino rats were killed from 10 min to 6 hr after a single intravenous dose of HRP. Experimental and control tissues were reacted for peroxidase activity and processed for light and electron microscopy. At 10 min, all extracellular spaces of the secretion zone showed reaction product. A reaction was also seen around Tomes' processes and in a layer of enamel spaces in the region of thin enamel. At later times intervals, reactions around Tomes' processes were also seen in regions of thicker enamel. Tracer was located preferentially at the growth fronts of rod and interrod enamel, and also diffused for some distance into enamel. From 2 to 6 hr, the enamel over the transition zone became heavily labeled. The tracer penetrated for more than 90 micrometer into the enamel and was localized mainly in the interrod enamel. Droplets of dense stippled material in the extracellular spaces between Tomes' processes did not mix with tracer, but sites which contain a light stippled material in the controls (extracellular spaces, vesicles within ameloblasts) showed a reaction. It is concluded that (1) the basal terminal bars of secretory ameloblasts do not impede the flow of large molecules, (2) the apical terminal bars are permeable in early secretion, become increasingly tight as secretion progresses, and are again permeable in the transition zone, (3) ameloblasts can shuttle large extracellular molecules towards the enamel growth fronts, (4) large molecules can diffuse into enamel; rod and interrod enamel differ with regard to the diffusion of large molecules, (5) ameloblasts phagocytose significant amounts of light stippled material. The possibility is considered that extracellular enamel precursor molecules move preferentially towards the enamel growth fronts, perhaps by a mechanism involving membrane flow, and diffuse through enamel in similar fashion as HRP.