Epithelial-mesenchymal interactions in tooth morphogenesis: the roles of extracellular matrix, growth factors, and cell surface receptors.

Morphogenesis and cell differentiation in the developing tooth are controlled by a series of reciprocal interactions between the epithelial and mesenchymal tissues. The exact molecular mechanisms operating in these interactions are unknown at present, but both structural components of the extracellular matrix (ECM) and diffusible growth factors have been suggested to be involved. In this review article we summarize our findings on the distribution patterns of three ECM molecules and two cell surface receptors during tooth morphogenesis through bud, cap, and bell stages of development. The examined molecules include fibronectin, type III collagen, and tenascin, which all represent components of the mesenchymal ECM, the cell surface proteoglycan, syndecan, which functions as a receptor for interstitial matrix, and the cell surface receptor for epidermal growth factor. Based on the observed changes in distribution patterns and on experimental evidence, roles are suggested for these molecules in epithelial-mesenchymal interactions during tooth development. Fibronectin is suggested to be involved in the cell-matrix interaction that controls odontoblast differentiation. Epidermal growth factor and its receptors are suggested to be involved in a paracrine fashion in the epithelial-mesenchymal interactions regulating morphogenesis of bud- and cap-stage teeth. Tenascin and syndecan are accumulated in the dental mesenchyme during the bud stage of development, and it is suggested that they represent a couple of a cell surface receptor and its matrix ligand and that they are involved in mesenchymal cell condensation during the earliest stages of tooth morphogenesis.     

Syndecan and tenascin expression is induced by epithelial-mesenchymal interactions in embryonic tooth mesenchyme

Morphogenesis of embryonic organs is regulated by epithelial-mesenchymal interactions associating with changes in the extracellular matrix (ECM). The response of the cells to the changes in the ECM must involve integral cell surface molecules that recognize their matrix ligand and initiate transmission of signal intracellularly. We have studied the expression of the cell surface proteoglycan, syndecan, which is a matrix receptor for epithelial cells (Saunders, S., M. Jalkanen, S. O'Farrell, and M. Bernfield. J. Cell Biol. In press.), and the matrix glycoprotein, tenascin, which has been proposed to be involved in epithelial-mesenchymal interactions (Chiquet-Ehrismann, R., E. J. Mackie, C. A. Pearson, and T. Sakakura. 1986. Cell. 47:131-139) in experimental tissue recombinations of dental epithelium and mesenchyme. Our earlier studies have shown that in mouse embryos both syndecan and tenascin are intensely expressed in the condensing dental mesenchyme surrounding the epithelial bud (Thesleff, I., M. Jalkanen, S. Vainio, and M. Bernfield. 1988. Dev. Biol. 129:565-572; Thesleff, I., E. Mackie, S. Vainio, and R. Chiquet-Ehrismann. 1987. Development. 101:289-296). Analysis of rat-mouse tissue recombinants by a monoclonal antibody against the murine syndecan showed that the presumptive dental epithelium induces the expression of syndecan in the underlying mesenchyme. The expression of tenascin was induced in the dental mesenchyme in the same area as syndecan. The syndecan and tenascin positive areas increased with time of epithelial-mesenchymal contact. Other ECM molecules, laminin, type III collagen, and fibronectin, did not show a staining pattern similar to that of syndecan and tenascin. Oral epithelium from older embryos had lost its ability to induce syndecan expression but the presumptive dental epithelium induced syndecan expression even in oral mesenchyme of older embryos. Our results indicate that the expression of syndecan and tenascin in the tooth mesenchyme is regulated by epithelial-mesenchymal interactions. Because of their early appearance, syndecan and tenascin may be used to study the molecular regulation of this interaction. The similar distribution patterns of syndecan and tenascin in vivo and in vitro and their early appearance as a result of epithelial-mesenchymal interaction suggest that these molecules may be involved in the condensation and differentiation of dental mesenchymal cells.     

Syndecan from embryonic tooth mesenchyme binds tenascin.

Syndecan is a cell surface heparan sulfate-rich proteoglycan found on various epithelial cells but also in some embryonic mesenchymal tissues. We have immunoisolated syndecan from embryonic tooth mesenchyme that appeared as a 250-300-kDa molecule (Kav = 0.3 in Sepharose 4B), containing only heparan sulfate side chains (Mr = 35,000). Northern analysis of whole tooth germs and tooth mesenchymes also revealed high expression of syndecan mRNAs (2.6 and 3.4 kilobases). In the binding assay utilizing nitrocellulose as a solid phase to immobilize matrix molecules, syndecan immunoisolated from tooth mesenchyme revealed binding to tenascin, and this interaction was shown to be mediated via heparan sulfate side chains. In contrast, syndecan from mouse mammary epithelial cells showed only weak interaction with tenascin. We propose that syndecan and tenascin may represent interactions of a cell surface receptor and a matrix ligand involved in mesenchymal cell condensation and differentiation during early organogenesis.     

Sequential induction of syndecan, tenascin and cell proliferation associated with mesenchymal cell condensation during early tooth development

The cell surface proteoglycan, syndecan, and the extracellular matrix glycoprotein, tenascin, are expressed in the mesenchyme during early development of many organs. We have studied the expression patterns of syndecan and tenascin during initiation of tooth development and in association with mesenchymal cell condensation and compared these with cell proliferation. Syndecan, tenascin and bromodeoxyuridine (BrdU) incorporation were localized by triple-labelling immunohistochemistry in serial sections of molar tooth germs of mouse embryos. Prior to formation of the epithelial tooth bud, syndecan accumulated in the mesenchymal cells which underlie the presumptive dental epithelium, but tenascin was not detected at this stage. Tenascin appeared during initiation of the epithelial down-growth at the lingual aspect of the tooth germ. During subsequent formation of the epithelial bud, at the late bud stage, syndecan and tenascin became exactly colocalized in the condensed mesenchyme which was clearly demarcated from other jaw mesenchyme. The expression of syndecan and tenascin was accompanied by rapid cell proliferation as indicated by marked BrdU incorporation. When development advanced to the cap stage, syndecan staining intensity in the dental papilla mesenchyme increased further whereas tenascin became reduced. In conclusion, the results demonstrate that the expression patterns of syndecan and tenascin overlap transiently during the period of mesenchymal cell condensation and that this is accompanied by cell proliferation. Syndecan and tenascin may play a role in growth control and in compartmentalization of the dental mesenchymal cells in the condensate.     

Cell surface proteoglycan expression correlates with epithelial-mesenchymal interaction during tooth morphogenesis

Tooth morphogenesis and differentiation of the dental cells are guided by interactions between epithelial and mesenchymal tissues. Because the extracellular matrix is involved in these interactions, the expression of matrix receptors located at the cell surface may change during this developmental sequence. We have examined the distribution of an epithelial cell surface proteoglycan antigen, known to behave as a receptor for interstitial matrix, during tooth morphogenesis. Intense staining was seen around the cells of the embryonic oral epithelium as well as the dental epithelium at the early bud stage. With development, expression was greatly reduced in the enamel organ. Differentiation of these cells into ameloblasts was associated with the loss of expression, while the epithelial cells remaining in the stratum intermedium and stellate reticulum regained intense staining. The PG antigen was weakly expressed in the loose neural crest-derived jaw mesenchyme but it became strongly reactive in the condensed dental papilla mesenchyme when extensive morphogenetic movements took place. With development, the PG antigen disappeared from the advanced dental papilla mesenchyme but persisted in the dental sac mesenchyme, which gives rise to periodontal tissues. The PG antigen was not expressed by odontoblasts. Hence, the expression of the PG antigen changes during the epithelial-mesenchymal interactions of tooth development and is lost during terminal cell differentiation. The expression follows morphogenetic rather than histologic boundaries. The acquisition and loss of expression in epithelial and mesenchymal tissues during tooth development suggest that this proteoglycan has specific functions in the epithelial-mesenchymal interactions that guide morphogenesis.     

Changes in the distribution of tenascin during tooth development

Tenascin is an extracellular matrix molecule that was earlier shown to be enriched in embryonic mesenchyme surrounding the budding epithelium in various organs including the tooth. In the present study tenascin was localized by immunohistology throughout the course of tooth development in the mouse and rat using polyclonal antibodies against chick tenascin. The results indicate that tenascin is expressed by the lineage of dental mesenchymal cells throughout tooth ontogeny. The intensity of staining with tenascin antibodies in the dental papilla mesenchyme was temporarily reduced at cap stage when the tooth grows rapidly and undergoes extensive morphogenetic changes. During the bell stage of morphogenesis, the staining intensity increased and tenascin was accumulated in the dental pulp even after completion of crown development and eruption. Tenascin was present in the dental basement membrane at the time of odontoblast differentiation. The dental papilla cells ceased to express tenascin upon differentiation into odontoblasts and tenascin was completely absent from dentin. It can be speculated that the remarkable expression of tenascin in the dental mesenchymal cells as compared to other connective tissues is associated with their capacity to differentiate into hard-tissue-forming cells.     

Epidermal growth factor inhibits morphogenesis and cell differentiation in cultured mouse embryonic teeth

Although local epithelial-mesenchymal tissue interactions which are presumably mediated by extracellular matrix molecules are important regulators of tooth morphogenesis and differentiation, our studies have indicated that these developmental processes also depend on circulating molecules. The iron-carrying serum protein transferrin is necessary for the early morphogenesis of mouse tooth in organ culture (A-M. Partanen, I. Thesleff, and P. Ekblom, 1984, Differentiation 27, 59-66). In the present study we have examined the effects of other growth factors on mouse tooth germs grown in a chemically defined medium containing transferrin. Fibroblast growth factor and platelet derived growth factor had no detectable effects but epidermal growth factor (EGF) inhibited dramatically the morphogenesis of teeth, and prevented odontoblast and ameloblast cell differentiation. EGF stimulated cell proliferation in the explants measured as [3H]thymidine incorporation in DNA. However, when the distribution of dividing cells was visualized in autoradiographs, it was observed that cell proliferation was stimulated in the dental epithelium but was inhibited in the dental mesenchyme. The inhibition of cell proliferation in the dental mesenchyme apparently caused the inhibition of morphogenesis. We do not know whether the dental epithelium or mesenchyme was the primary target for the action of EGF in the inhibition of morphogenesis. It is, however, apparent that the response of the dental mesenchymal cells to EGF (inhibition of proliferation) is regulated by their local environment, since EGF enhanced proliferation when these cells were disaggregated and cultured as monolayers. This indicates that the organ culture system where the various embryonic cell lineages are maintained in their original environment corresponds better to the in vivo situation when the roles of exogenous growth factors during development are examined.     

Immunofluorescent localization of tenascin during development of the mouse urogenital sinus: Possible involvement in genital duct morphogenesis

Tenascin is a compound of the mesenchymal extracellular matrix and has been proposed as a possible mediator in epithelial-mesenchymal interactions, because of its characteristic distribution in tissues during fetal development. In the present study, we have investigated by immunofluorescence the changes in the distribution of tenascin during development of the mouse urogenital sinus, a process in which tissue interactions were found to be essential. Tenascin first appears in dorsal mesenchyme on days 13-15 of gestation, coinciding with morphological changes of the epithelium. During male development, tenascin accumulates in the dorsal mesenchyme around the junction of Wolffian ducts, but not in the ventral mesenchyme, into which prostatic buds (prostate gland anlagen) project from the sinus epithelium. During female development, the mesenchyme that participates in the downgrowth of the vagina (derived from Müllerian ducts) stains intensively for tenascin. In both of these tenascin-positive areas, the epithelium undergoes conspicuous morphogenetic changes. The results suggest that mesenchymal tenascin could be involved in the epithelial morphogenesis of the sinus, especially in the morphogenesis of the genital ducts.     

Epithelial-mesenchymal interactions regulate the stage-specific expression of a cell surface proteoglycan, syndecan, in the developing kidney

Morphogenesis of the kidney is regulated by reciprocal tissue interactions between the epithelial ureter bud and the metanephric mesenchyme. The differentiation of the kidney involves profound changes in the extracellular matrix, and therefore matrix receptors may have an important role in this process. We studied the expression of syndecan, a cell surface proteoglycan acting as a receptor for interstitial matrix materials, by using a monoclonal antibody against the core protein of the molecule. Syndecan was not detected in the uninduced metanephric mesenchyme. During the formation of the ureter bud from the Wolffian duct, syndecan appeared in the mesenchymal cells around the invaginating bud. Simultaneously with the first branching of the ureter bud, the whole nephric mesenchyme became syndecan positive, but a 3- to 10-cell-thick layer around the branching ureter bud, representing the presumptive tubular cells, was most intensely stained. During the assembly of the mesenchyme cells into pretubular aggregates, syndecan was detected in these aggregates and, to a lesser degree, in the morphologically undifferentiated mesenchyme. Thereafter syndecan was found only in the differentiating epithelium, from which it was gradually lost during maturation of the nephron. It was last detected in the periphery of the kidney, where tubulogenesis still continued. In transfilter cultures we showed that syndecan appeared in the nephric mesenchyme during the period when the mesenchyme becomes programmed to transform into epithelial structures. By using interspecies recombinations and a species-specific antibody we excluded the possibility that syndecan in the mesenchyme would originate from the inductor. We conclude that syndecan expression is regulated by epithelial-mesenchymal interactions. The findings that syndecan appeared as an early response to induction and that its distribution showed both spatial and temporal correlation with kidney morphogenesis suggest an important role for this molecule in development.     

Midkine (MK), a heparin-binding growth/differentiation factor, is regulated by retinoic acid and epithelial-mesenchymal interactions in the developing mouse tooth, and affects cell proliferation and morphogenesis

Midkine (MK) is the first cloned gene in a new family of heparin- binding growth/differentiation factors involved in the regulation of growth and differentiation. We have analyzed the expression of MK mRNA and protein during tooth development in mouse embryos and studied the regulation of MK expression and the biological effects of MK protein in organ cultures. MK expression was restricted and preferential in the tooth area as compared to the rest of the developing maxillary and mandibular processes suggesting specific functions for MK during tooth morphogenesis. MK mRNA and protein were expressed during all stages of tooth formation (initiation, morphogenesis, and cell differentiation), and shifts of expression were observed between the epithelial and mesenchymal tissue components. However, the expression of mRNA and protein showed marked differences at some stages suggesting paracrine functions for MK. Tissue recombination experiments showed that MK gene and protein expression are regulated by epithelial-mesenchymal interactions, and, moreover, that dental tissue induces the ectopic expression of MK protein in non-dental tissue. The expression of MK gene and protein in the mandibular arch mesenchyme from the tooth region were stimulated by local application of retinoic acid in beads. Cell proliferation was inhibited in dental mesenchyme around the beads releasing MK, but this effect was modulated by simultaneous application of FGF-2. Morphogenesis and cell differentiation were inhibited in tooth germs cultured in the presence of neutralizing antibodies for MK, whereas the development of other organs (e.g., salivary gland, kidney) was unaffected. These results suggest important roles for MK in the molecular cascade that regulates tooth development.     

Induction of enamel matrix protein expression in an ameloblast cell line co-cultured with a mesenchymal cell line in vitro

Interactions between epithelium and mesenchyme are important for organ and tissue development. In this study, in order to mimic interactions between epithelium and mesenchyme during native tooth development, we constructed three-dimensional culture systems in vitro using a collagen membrane. Two types of collagen membrane-based in vitro culture systems were constructed in which dental epithelial and dental follicle cell lines were cultured. One co-culture method involved inoculation of one cell line into one side of the collagen membrane, and the other cell line into the opposite side of the membrane (sandwich co-culture). As a control, the second method involved culture of one of the cell lines on a culture dish and the second cell line on a collagen membrane, facing away from the first cell line (separate co-culture). The HAT-7 cells were also grown as a monolayer culture on collagen. Ameloblast differentiation in these cultures was investigated by analysis of the mRNA and/or protein expression of ameloblastin and amelogenin. Our results suggest that interaction of epithelial and mesenchymal cells via the extracellular matrix is important for tooth differentiation in vitro. Our culture system should be a useful method for investigation of epithelial-mesenchymal interactions.     

Dorsoventral patterning of the C. elegans postembryonic mesoderm requires both LIN-12/Notch and TGFbeta signaling

The heparin binding molecules MK and HB-GAM are involved in the regulation of growth and differentiation of many tissues and organs. Here we analyzed the expression of MK and HB-GAM in the developing mouse incisors, which are continuously growing organs with a stem cell compartment. Overlapping but distinct expression patterns for MK and HB-GAM were observed during all stages of incisor development (initiation, morphogenesis, cytodifferentiation). Both proteins were detected in the enamel knot, a transient epithelial signaling structure that is important for tooth morphogenesis, and the cervical loop where the stem cell niche is located. The functions of MK and HB-GAM were studied in dental explants and organotypic cultures in vitro. In mesenchymal explants, MK stimulated HB-GAM expression and, vice-versa, HB-GAM upregulated MK expression, thus indicating a regulatory loop between these proteins. BMP and FGF molecules also activated expression of both cytokines in mesenchyme. The proliferative effects of MK and HB-GAM varied according to the mesenchymal or epithelial origin of the tissue. Growth, cytodifferentiation and mineralization were inhibited in incisor germs cultured in the presence of MK neutralizing antibodies. These results demonstrate that MK and HB-GAM are involved in stem cells maintenance, cytodifferentiation and mineralization processes during mouse incisor development.     

Effects of retinoids on tooth morphogenesis and cytodifferentiations, in vitro.

The first embryonic lower mouse molar was used as a model system to investigate the effects of two retinoids, retinoic acid (RA) and a synthetic analogue, Ch55, on morphogenesis and cytodifferentiations in vitro. Exogenous retinoids were indispensable for morphogenesis of bud, cap and bell-stage molars in serum-free, chemically-defined, culture media. Transferrin and RA or transferrin and Ch55 acted synergistically in promoting morphogenesis from bud and cap-stage explants. Transferrin, per se, had no morphogenetic effect. Epithelial histogenesis, odontoblast functional differentiation and ameloblast polarization always occurred in RA-depleted explants. Comparison of the distributions of bromodeoxyuridine (BrdU) incorporation between explants cultured in the absence or presence of RA revealed that RA could modify the patterns of cell proliferation in the inner dental epithelium and dental mesenchyme. Inner dental epithelium cell proliferation is regulated by the dental mesenchyme through basement membrane-mediated interactions, and tooth morphogenesis is controlled by the dental mesenchyme. Laminin is a target molecule of retinoid action. Using a monospecific antibody, we immunolocalized laminin and/or structurally-related molecules sharing the laminin B chain in the embryonic dental mesenchyme and in the dental basement membrane and showed that RA could promote the synthesis or secretion of these molecules. Based on previous in situ hybridization data, it was speculated that CRABPs might regulate the effects of RA on embryonic dental cell proliferation. The fact that Ch55, a retinoid which does not bind to CRABPs, is 100 times more potent than RA in promoting tooth morphogenesis in vitro seems to rule out this hypothesis. On the other hand, the stage-specific inhibition of tooth morphogenesis by excess RA is consistent with the hypothesis that CRABPs might protect embryonic tissues against potentially teratogenic concentrations of free retinoids.     

Localization and quantitation of 125I-epidermal growth factor binding in mouse embryonic tooth and other embryonic tissues at different developmental stages

We have shown earlier that epidermal growth factor (EGF) inhibits morphogenesis and cell differentiation in mouse embryonic teeth in organ culture. This inhibition depends on the stage of tooth development so that only teeth at early developmental stages respond to EGF (A-M. Partanen, P. Ekblom, and I. Thesleff (1985) Dev. Biol. 111, 84-94). We have now studied the quantity and pattern of EGF binding in teeth at various stages of development by incubating the dissected tooth germs with 125I-labeled EGF. Although the quantity of 125I-EGF binding per microgram DNA stays at the same level, localization of 125I-EGF binding by autoradiography reveals that the distribution of binding sites changes dramatically. In bud stage the epithelial tooth bud that is intruding into the underlying mesenchyme has binding sites for EGF, but the condensation of dental mesenchymal cells around the bud does not bind EGF. At the cap stage of development the dental mesenchyme binds EGF, but the dental epithelium shows no binding. This indicates that the dental mesenchyme is the primary target tissue for the inhibitory effect of EGF on tooth morphogenesis during early cap stage. During advanced morphogenesis the binding sites of EGF disappear also from the dental papilla mesenchyme, but the dental follicle which consists of condensed mesenchymal cells surrounding the tooth germ, binds EGF abundantly. We have also studied EGF binding during the development of other embryonic organs, kidney, salivary gland, lung, and skin, which are all formed by mesenchymal and epithelial components. The patterns of EGF binding in various tissues suggest that EGF may have a role in the organogenesis of epitheliomesenchymal organs as a stimulator of epithelial proliferation during initial epithelial bud formation and branching morphogenesis. The results of this study indicate that EGF stimulates or maintains proliferation of undifferentiated cells during embryonic development and that the expression of EGF receptors in different organs is not related to the age of the embryo, but is specific to the developmental stage of each organ.     

Tenascin during gut development: appearance in the mesenchyme, shift in molecular forms, and dependence on epithelial-mesenchymal interactions [published erratum appears in J Cell Biol 1989 Mar;108(3):following 1175]

Tenascin, an extracellular matrix protein, is expressed in the mesenchyme around growing epithelia in the embryo. We therefore investigated whether epithelial cells can stimulate expression of tenascin in embryonic mesenchyme. Mesenchyme from the presumptive small intestine was used because it is known that reciprocal epithelial- mesenchymal interactions are important for gut morphogenesis. Rat monoclonal antibodies against mouse tenascin were raised and were found to react specifically with mouse tenascin in ELISA. In supernatants of cultured fibroblasts, the antibodies precipitated two peptides of Mr 260 and 210 kD. One of the antibodies also reacted with these tenascin chains in immunoblots of tissue extracts. We found that tenascin was absent during early stages of gut development, at stages when the mesenchyme is already in contact with the stratified epithelium of the endoderm. Rather, it appeared in the mesenchyme when the homogenous endodermal epithelium differentiated into the heterogenous absorptive epithelium. Tenascin remained present in the stroma of the adult gut, close to the migration pathways of the continuously renewing epithelium. When first detected during intestinal differentiation, the 210-kD component was predominant but at birth the relative amount of the 260-kD component had increased. The expression data suggested that the appearance of tenascin in the mesenchyme was dependent on the presence of epithelium. To test this, isolated gut mesenchymes from 13- d-old mouse embryos were cultured for 24 h either alone or together with epithelial and nonepithelial cells. Whereas mesenchyme cultured alone or in the presence of nonepithelial B16-F1 melanoma cells produced only trace amounts of tenascin, expression was strongly stimulated by the epithelial cell line, Madin-Darby canine kidney (MDCK). We propose that growing and differentiating epithelia produce locally active factors which stimulate synthesis of tenascin in the surrounding mesenchyme.     

Fgfr2b mediated epithelial-mesenchymal interactions coordinate tooth morphogenesis and dental trigeminal axon patterning

Dental trigeminal nerve fiber growth and patterning are strictly integrated with tooth morphogenesis, but it is still unknown, how these two developmental processes are coordinated. Here we show that targeted inactivation of the dental epithelium expressed Fgfr2b results in cessation of the mouse mandibular first molar development at the degenerated cap stage and the failure of the trigeminal molar nerve to establish the lingual branch at E13.5 stage while the buccal branch develops properly. This axon patterning defect correlates to the histological absence of the mesenchymal dental follicle and adjacent Semaphorin3A-free dental follicle target field as well as appearance of ectopic Sema3A expression domain in the lingual side of the epithelial bud. Although the mesenchymal ligands for Fgfr2b, Fgf3 and -10 were present in the Fgfr2b(-/)(-) dental mesenchyme, mutant dental epithelium showed dramatically reduced proliferation and the lack of Fgf3. Tgfbeta1, which controls Sema3A was absent from the Fgfr2b(-/-) tooth germ, and Sema3A was specifically downregulated in the dental mesenchyme at the bud and cap stage. In addition, the epithelial primary enamel knot signaling center although being molecularly present neither was histologically detectable nor expressed Bmp4 and Fgf3 as well as Fgf4, which is essential for tooth morphogenesis and stimulates mesenchymal Fgf3 and Tgfbeta1. Fgf4 beads rescued Tgfbeta1 in the Fgfr2b(-/-) dental mesenchyme explants and Tgfbeta1 induced de novo Sema3A expression in the dental mesenchyme. Collectively these results demonstrate that epithelial Fgfr2b controls tooth morphogenesis and dental axon patterning, and suggests that Fgfr2b, by mediating local epithelial-mesenchymal interactions, integrates these two distinct developmental processes during odontogenesis.