Rodent incisors are covered by enamel only on their labial side. This asymmetric distribution of enamel is instrumental to making the cutting edge sharp. Enamel matrix is secreted by ameloblasts derived from dental epithelium. Here we show that overexpression of follistatin in the dental epithelium inhibits ameloblast differentiation in transgenic mouse incisors, whereas in follistatin knockout mice, ameloblasts differentiate ectopically on the lingual enamel-free surface. Consistent with this, in wild-type mice, follistatin was continuously expressed in the lingual dental epithelium but downregulated in the labial epithelium. Experiments on cultured tooth explants indicated that follistatin inhibits the ameloblast-inducing activity of BMP4 from the underlying mesenchymal odontoblasts and that follistatin expression is induced by activin from the surrounding dental follicle. Hence, ameloblast differentiation is regulated by antagonistic actions of BMP4 and activin A from two mesenchymal cell layers flanking the dental epithelium, and asymmetrically expressed follistatin regulates the labial-lingual patterning of enamel formation.

The tumor suppressor gene Apc (adenomatous polyposis coli) is a member of the Wnt signaling pathway that is involved in development and tumorigenesis. Heterozygous knockout mice for Apc have a tumor predisposition phenotype and homozygosity leads to embryonic lethality. To understand the role of Apc in development we generated a floxed allele. These mice were mated with a strain carrying Cre recombinase under the control of the human Keratin 14 (K14) promoter, which is active in basal cells of epidermis and other stratified epithelia. Mice homozygous for the floxed allele that also carry the K14-cre transgene were viable but had stunted growth and died before weaning. Histological and immunochemical examinations revealed that K14-cre–mediated Apc loss resulted in aberrant growth in many ectodermally derived squamous epithelia, including hair follicles, teeth, and oral and corneal epithelia. In addition, squamous metaplasia was observed in various epithelial-derived tissues, including the thymus. The aberrant growth of hair follicles and other appendages as well as the thymic abnormalities in K14-cre; ApcCKO/CKO mice suggest the Apc gene is crucial in embryonic cells to specify epithelial cell fates in organs that require epithelial–mesenchymal interactions for their development.

We evaluated the detailed expression patterns of Runx1 and Sox9 in various types of bone formation, and determined whether Runx1 expression was affected by Runx2 deficiency and Runx2 expression by Runx1 deficiency. Our results indicate that both Runx1 and Sox9 are intensely expressed in the future osteogenic cell compartment and in cartilage. The pattern of Runx1 and Sox9 expression suggests that both genes could potentially be involved in incipient intramembranous bone formation during craniofacial development.Introduction: Runx1, a gene essential for hematopoiesis, contains RUNX binding sites in its promoter region, suggesting possible cross-regulation with Runx2 and potential regulatory roles in bone development. On the other hand, Sox9 is essential for chondrogenesis, and haploinsufficiency of Sox9 leads to premature ossification of the skeletal system. In this study, we studied the possible roles of Runx1 and Sox9 in bone development. Materials and Methods: Runx1, Runx2/Osf2, and Sox9 expression was evaluated by in situ hybridization in the growing craniofacial bones of embryonic day (E)12-16 mice and in the endochondral bone-forming regions of embryonic and postnatal long bones. In addition, we evaluated Runx2/Osf2 expression in the growing face of Runx1 knockout mice at E12.5 and Runx1 expression in Runx2 knockout mice at E14.5. Results: Runx1 and Sox9 were expressed in cartilage, and the regions of expression expanded to the neighboring Runx2-expressing osteogenic regions. Expression of both Runx1 and Sox9 was markedly downregulated on ossification. Runx1 and Sox9 expression was absent in the regions of endochondral bone formation and in actively modeling or remodeling bone tissues in the long bones as well as in ossified craniofacial bones. Runx2 expression was not affected by gene disruption of Runx1, whereas the expression domains of Runx1 were extended in Runx2