Teeth develop from reciprocal interactions between mesenchyme cells and epithelium, where the epithelium provides the instructive information for initiation. Based on these initial tissue interactions, we have replaced the mesenchyme cells with mesenchyme created by aggregation of cultured non-dental stem cells in mice. Recombinations between non-dental cell-derived mesenchyme and embryonic oral epithelium stimulate an odontogenic response in the stem cells. Embryonic stem cells, neural stem cells, and adult bone-marrow-derived cells all responded by expressing odontogenic genes. Transfer of recombinations into adult renal capsules resulted in the development of tooth structures and associated bone. Moreover, transfer of embryonic tooth primordia into the adult jaw resulted in development of tooth structures, showing that an embryonic primordium can develop in its adult environment. These results thus provide a significant advance toward the creation of artificial embryonic tooth primordia from cultured cells that can be used to replace missing teeth following transplantation into the adult mouth.

Inductive interactions between gut endoderm and the underlying mesenchyme pattern the developing digestive tract into regions with specific morphology and functions. The molecular mechanisms behind these interactions are largely unknown. Expression of the conserved homeobox gene Barx1 is restricted to the stomach mesenchyme during gut organogenesis. Using recombinant tissue cultures, we show that Barx1 loss in the mesenchyme prevents stomach epithelial differentiation of overlying endoderm and induces intestine-specific genes instead. Additionally, Barx1 null mouse embryos show visceral homeosis, with intestinal gene expression within a highly disorganized gastric epithelium. Barx1 directs mesenchymal cell expression of two secreted Wnt antagonists, sFRP1 and sFRP2, and these factors are sufficient replacements for Barx1 function. Canonical Wnt signaling is prominent in the prospective gastric endoderm prior to epithelial differentiation, and its inhibition by Barx1-dependent signaling permits development of stomach-specific epithelium. These results define a transcriptional and signaling pathway of inductive cell interactions in vertebrate organogenesis.

The recent identification of SATB2 as a candidate gene responsible for the craniofacial dysmorphologies associated with deletions and translocations at 2q32-q33, one of only three regions of the genome for which haploinsufficiency has been significantly associated with isolated cleft palate, led us to investigate the in vivo functions of murine Satb2. We find that, similar to the way in which SATB2 is perceived to act in humans, craniofacial defects due to haploinsufficiency of Satb2, including cleft palate (in approximately 25% of cases), phenocopy those seen with 2q32-q33 deletions and translocations in humans. Full functional loss of Satb2 results in amplification of these defects and leads both to increased apoptosis in the craniofacial mesenchyme where Satb2 is usually expressed and to changes in the pattern of expression of three genes implicated in the regulation of craniofacial development in humans and mice: Pax9, Alx4, and Msx1. The Satb2-dosage sensitivity in craniofacial development is conspicuous--along with its control of cell survival, pattern of expression, and reversible functional modification by SUMOylation, it suggests that Satb2/SATB2 function in craniofacial development may prove to be more profound than has been anticipated previously. Because jaw development is Satb2-dosage sensitive, the regulators of Satb2 expression and posttranslational modification become of critical importance both ontogenetically and evolutionarily, especially since such regulators plausibly play undetected roles in jaw and palate development and in the etiology of craniofacial malformations.

Teeth are vertebrate organs that arise from complex and progressive interactions between an ectoderm, the oral epithelium and an underlying mesenchyme. During their early development, tooth germs exhibit many morphological and molecular similarities with other developing epithelial appendages, such as hair follicles, mammary and salivary glands, lungs, kidneys, etc. The developing mouse tooth germ, which is an experimentally accessible model for organogenesis, provides a powerful tool for elucidating the molecular mechanisms that control the development of these organs. Dentition patterning also provides a unique model for understanding how different shapes of teeth arise in different regions of the jaws. We review here the main signalling networks mediating the epithelial-mesenchymal interactions involved in tooth morphogenesis and patterning.

The signalling peptide encoded by the sonic hedgehog gene is restricted to localised thickenings of oral epithelium, which mark the first morphological evidence of tooth development, and is known to play a crucial role during the initiation of odontogenesis. We show that at these stages in the murine mandibular arch in the absence of epithelium, the Shh targets Ptc1and Gli1 are upregulated in diastema mesenchyme, an edentulous region between the sites of molar and incisor tooth formation. This ectopic expression is not associated with Shh transcription but with Shh protein, undetectable in the presence of epithelium. These findings suggest that, in diastema mesenchyme, restriction of Shh activity is dependent upon the overlying epithelium. This inhibitory activity was demonstrated by the ability of transplanted diastema epithelium to downregulate Ptc1 in tooth explants, and for isolated diastema mesenchyme to express Ptc1. A candidate inhibitor in diastema mesenchyme is the glycosylphosphatidylinositol-linked membrane glycoprotein Gas1. Gas1is normally expressed throughout mandibular arch mesenchyme; however, in the absence of epithelium this expression was downregulated specifically in the diastema where ectopic Shh protein was identified. Although Shh signalling has no effect upon Gas1 expression in mandibular arch mesenchyme,overexpression of Gas1 results in downregulation of ectopic Ptc1. Therefore, control of the position of tooth initiation in the mandibular arch involves a combination of Shh signalling at sites where teeth are required and antagonism in regions destined to remain edentulous.

The cranial neural crest cells, which are specialized cells of neural origin, are central to the process of mammalian tooth development. They are the only source of mesenchyme able to sustain tooth development, and give rise not only to most of the dental tissues, but also to the periodontium, the surrounding tissues that hold teeth in position. Tooth organogenesis is regulated by a series of interactions between cranial neural crest cells and the oral epithelium. In the development of a tooth, the epithelium covering the inside of the developing oral cavity provides the first instructive signals. Signaling molecules secreted by the oral epithelium 1) establish large cellular fields competent to form a specific tooth shape (mono- or multicuspid) along a proximodistal axis; 2) define an oral (capable of forming teeth) and non-oral mesenchyme along a rostrocaudal axis; and 3) position the sites of future tooth development. The critical information to model tooth shape resides later in the neural crest-derived mesenchyme. Cranial neural crest cells ultimately differentiate into highly specialized cell types to produce mature dental organs. Some cranial neural crest cells located in the dental pulp, however, maintain plasticity in their developmental potential up to postnatal life, offering new prospects for regeneration of dental tissues.

Branching morphogenesis is a molecularly conserved mechanism that is adopted by several organs, such as the lung, kidney, mammary gland and salivary gland, to maximize the surface area of a tissue within a small volume. Branching occurs through repetitive clefting and elongation of spherical epithelial structures, called endbuds, which invade the surrounding mesenchyme. In the salivary gland, lumen formation takes place alongside branching morphogenesis, but in a controlled manner, so that branching is active at the distal ends of epithelial branches while lumen formation initiates at the proximal ends, and spreads distally. We present here data showing that interaction between FGF signaling and the canonical (β-catenin dependent) and non-canonical branches of Wnt signaling coordinates these two processes. Using the Axin2(lacZ) reporter mice, we find Wnt/β-catenin signaling activity first in the mesenchyme and later, at the time of lumen formation, in the ductal epithelium. Gain and loss of function experiments reveal that this pathway exerts an inhibitory effect on salivary gland branching morphogenesis. We have found that endbuds remain devoid of Wnt/β-catenin signaling activity, a hallmark of ductal structures, through FGF-mediated inhibition of this pathway. Our data also show that FGF signaling has a major role in the control of lumen formation by preventing premature hollowing of epithelial endbuds and slowing down the canalization of presumptive ducts. Concomitantly, FGF signaling strongly represses the ductal marker Cp2l1, most likely via repression of Wnt5b and non-canonical Wnt signaling. Inhibition of canonical and non-canonical Wnt signaling in endbuds by FGF signaling occurs at least in part through sFRP1, a secreted inhibitor of Wnt signaling and downstream target of FGF signaling. Altogether, these findings point to a key function of FGF signaling in the maintenance of an undifferentiated state in endbud cells by inhibition of a ductal fate.

Homeobox genes convey positional information in embryos and their role in patterning the mammalian gut is a topic of considerable interest. Barx1 is expressed selectively in fetal stomach mesenchyme and directs differentiation of overlying endoderm. Recombinant tissue cultures and study of young mouse embryos previously suggested that Barx1 controls expression of secreted Wnt antagonists, which suppress endodermal Wnt signaling, to enable stomach epithelial differentiation. We overcame mid-gestational lethality of Barx1-/- mouse embryos and report here the spectrum of anomalies in a distinctive and unprecedented model of gastrointestinal homeotic transformation. Using various mouse models, we confirm the importance of attenuated Wnt signaling in stomach development and the role of Barx1 in suppressing endodermal Wnt activity. Absence of Barx1 also results in fully penetrant defects in positioning and expansion of the spleen, an organ that originates within the mesothelial lining of the stomach. Barx1 is absent from the spleen primordium but highly expressed in the mesogastrium, indicating an indirect effect on spleen development. However,our results argue against a role for Wnt antagonism in genesis of the spleen. Mouse spleen development relies on several homeodomain transcriptional regulators that are expressed in the spleen primordium. Loss of Barx1 does not affect expression of any of these genes but notably reduces expression of Wt1,a transcription factor implicated in spleen morphogenesis and expressed in the mesothelium. These observations place Barx1 proximally within a Wt1 pathway of spleen development and reveal how a homeotic regulator employs different molecular mechanisms to mold neighboring organs.