Expression patterns of the homeobox gene, Hox-8, in the mouse embryo suggest a role in specifying tooth initiation and shape

A. Mackenzie 0 M. W. J. Ferguson 0 P. T. Sharpe 0 0 Molecular Embryology Laboratory, Department of Cell and Structural Biology, The University of Manchester , Stopford Building, Oxford Road, Manchester, M13 9PT , UK Expression patterns of the homeobox gene, Hox-8, in the mouse embryo suggest a role in specifying tooth initiation and shape ALASDAIR MacKENZEE, MARK W. J. FERGUSON and PAUL T. SHARPE* Corresponding author Summary - We have studied the expression patterns of the newly isolated homeobox gene, Hox-8 by in situ hybridisation to sections of the developing heads of mouse embryos between E9 and E17.5, and compared them to Hox-7 expression patterns in adjacent sections. This paper concentrates on the interesting expression patterns of Hox-8 during initiation and development of the molar and incisor teeth. Hox-8 expression domains are present in the neural crest-derived mesenchyme beneath sites of future tooth formation, in a proximo-distal gradient. Tooth development is initiated in the oral epithelium which subsequently thickens in discrete sites and invaginates to form the dental lamina. Hox-8 expression in mouse oral epithelium is first evident at the sites of the dental placodes, suggesting a role in the specification of tooth position. Subsequently, in molar teeth, this patch of Hox-8 expressing epithelium becomes incorporated within the buccal aspect of the invaginating dental lamina to form part of the external enamel epithelium of the cap stage tooth germ. This locus of Hox-8 expression becomes continuous with new sites of Hox-8 expression in the enamel navel, septum, knot and internal enamel epithelium. The transitory enamel knot, septum and navel were postulated, long ago, to be involved in specifying tooth shape, causing the inflection of the first buccal cusp, but this theory has been largely ignored. Interestingly, in the conical incisor teeth, the enamel navel, septum and knot are absent, and Hox-8 has a symmetrical expression pattern. Our demonstration of the precise expression patterns of Hox-8 in the early dental placodes and their subsequent association with the enamel knot, septum and navel provide the first molecular clues to the basis of patterning in the dentition and the association of tooth position with tooth shape: an association all the more intriguing in view of the evolutionary robustness of the patterning mechanism, and the known role of homeobox genes in Drosophila pattern formation. At the bell stage of tooth development, Hox-8 expression switches tissue layers, being absent from the differentiating epithelial ameloblasts and turned on hi the differentiating mesenchymal odontoblasts. Hox-7 is expressed in the mesenchyme of the dental papilla and follicle at all stages. This reciprocity of expression suggests an interactive role between Hox-7, Hox-8 and other genes in regulating epithelial mesenchymal interactions during dental differentiation. Hox-8 is also expressed in the distal mesenchyme and epithelia of the lateral nasal, medial nasal and maxillary processes (in a more spatially restricted domain than Hox-7), Jacobson's organs, the developing skull bones, meninges, ear, eye, whisker and hair follicles, choroid plexus, cardiac cushions and limb buds. The patterns of expression hi the facial processes resemble those of the progress zone of the limb, suggesting a similar patterning mechanism in these embryonic outgrowths. Teeth are phylogenetically ancient structures, well preserved in the fossil record where their shape and position in jaw fragments play a pivotal role in the reconstruction of the anatomy, dietary habits and lineage relationships of vertebrates (Romer, 1966). Moreover, in mammals, their shape and position in the jaws are tightly linked: there are no known mutants which, for example, develop molars at the front and incisors at the rear of the mouth (Miles and Grigson, 1990). Development of the mammalian dentition therefore involves both regional (incisors, canines, premolars and molars) and temporal (differing development times of deciduous and permanent teeth) patterning of the individual tooth anlage. Development of an individual tooth commences with an oral thickening of the jaw epithelium, its invagination into the mesenchyme to form a dental lamina, distal enlargement of the lamina to form epithelial swellings (enamel organs) associated with condensing neural crest-derived jaw mesenchyme (dental papilla and follicle) which collectively (tooth germ) progress through the well characterised morphological and differentiation stages of bud, cap and bell to form the adult tooth (Ruch, 1987; Thesleff et al., 1989; Ferguson, 1990). Despite the importance of tooth shape and position for developmental, phylogenetic and clinical dental studies, almost nothing is known about the molecular basis of this exquisitely precise patterning. Early theories (Osborn, 1978, 1984) suggested that the developmental information required for tooth initiation was carried into the jaws by clones of neural crest cells. Some credence for this patterning idea comes from neural crest transplantation studies in the chicken embryo, whereby duplicate first branchial arch structures can be induced to form by grafting future first arch premigratory neural crest cells to the region destined to form the second arch (Noden, 1983); unfortunately, birds do not develop teeth! Extensive heterotypic or heterochronic epithelial-mesenchymal recombination experiments demonstrate that initiation and regional localisation of the early dental anlage (incisors, molars) are absolutely dependent on the rostral (i.e. oral) epithelium of the mandibular arch between E9 (mouse embryonic day 9) and E10: no other epithelium can elicit tooth development from mandibular mesenchyme (Mina and Kollar, 1987; Lumsden, 1988). Equally, the E9-E10 mandibular epithelium can only specify tooth development in neural crest (either cranial or trunk) -derived mesenchyme; not, for example, in limb mesenchyme (Lumsden, 1988). One effect of this interaction between mandibular arch mesenchyme and E9/E10 mandibular epithelium is the acquisition by the former of the ability to instruct competent epithelia to participate in regional specific enamel organ morphogenesis and subsequent ameloblast differentiation and enamel matrix synthesis: an ability which persists from E l l to E16 (Kollar and Baird, 1969, 1970a, 1970b; Heritier and Deminatti, 1970; Kollar, 1972, 1981; Ruch et al., 1983; Ruch 1984; Lumsden, 1988). In vivo therefore, epithelial signalling to mandibular mesenchyme at E9/E10 is reciprocated by mesenchymal signalling to the epithelia at E13, and these numerous reciprocal interactions (summarised by Lumsden, 1988) characterised by changes in extracellular matrix molecules (Thesleff et al., 1979, 1981, 1987, 1988, 1989, 1990; Andujar et al., 1991), growth factors (Partanen and Thesleff, 1985, 1987, 1989; Cam et al., 1990; D'Souja et al., 1990; Hata et al., 1990; Kronmiller (...truncated)