Introduction
Mammalian development involves the morphogenesis of complex three-dimensional structures from seemingly uniform sheets or masses of cells. A simple bud-like structure initiates the formation of many organs, including lungs, spinal cord, mammary glands, and hair follicles [1]. The multipotent, adhering epithelial cells are typically attached to an underlying basal lamina that polarizes the epithelial sheet and separates it from surrounding mesenchyme. Budding morphogenesis is guided by a reciprocal exchange of signals between epithelium and mesenchyme to specify the identity of the organ that will form and to govern its growth.
At the helm of these molecular communication pathways are Wnts, bone morphogenic proteins (BMPs), transforming growth factor βs (TGF-βs), and fibroblast growth factors (FGFs). Through activation of cell surface transmembrane receptors, these external signaling molecules trigger distinct cascades of intracellular events that culminate in changes in gene expression, growth, and differentiation [2]. How this constellation of signals collaborates in tailoring each budding process so that it executes a distinct morphogenetic program has yet to be comprehensively defined. However, the process appears to be patterned at the initial stages of bud formation, since the relative importance of these pathways and their downstream effectors differ as buds begin to develop and cell fates are specified.
The development of a bud requires a number of coordinated changes in the behavior of the targeted cells within an epithelial sheet. The process must be accompanied by alterations in the proliferation, polarity, shape, and adhesiveness of selected cells, as well as by modifications in their underlying basal lamina. Thus, extracellular epithelial-mesenchymal crosstalk must be intricately orchestrated to couple the determination of distinct cell fates with the contemporaneous remodeling of the physical and structural properties of the cell.
Among the few dispensable organs, hair follicles offer an excellent model system to study epithelial bud formation. Mammalian skin epithelium begins as a single sheet of multipotent ectodermal cells. During development, specialized mesenchymal cells populate the skin in a spatially defined pattern to initiate the complex epithelial-mesenchymal crosstalk that will specify the bud [3]. Once committed, a small cluster of epithelial cells, the placode, instructs a group of underlying mesenchymal cells to condense and form the nascent dermal papilla, which will be a permanent fixture of the hair follicle. Subsequent exchanges between the placode and nascent dermal papilla result in further growth of the follicle into the underlying dermis, or down-growth, and eventual differentiation into the six concentric layers of the mature follicle.
Previously, we delineated how two respective epithelial and mesenchymal signals, Wnts and the BMP-inhibitory factor noggin, function in concert to induce lymphoid enhancer factor-1/β-catenin (LEF-1/β-catenin)-mediated gene transcription within the follicle placode [4]. The downstream changes elicited through convergence of these two early signaling pathways include down-regulation of the gene encoding E-cadherin, the prototypical epithelial cadherin that forms the transmembrane core of intercellular adherens junctions (AJs) [5]. We subsequently showed that when E-cadherin is transgenically elevated in mouse skin, hair follicle morphogenesis is blocked, suggesting that E-cadherin down-regulation is a critical event in governing the adhesion dynamics necessary for budding morphogenesis [4]. Like LEF-1, E-cadherin also binds to β-catenin. At sites of cell-cell contact, however, E-cadherin-β-catenin complexes recruit α-catenin, which in turn coordinates the associated actin polymerization dynamics necessary to stabilize nascent AJs and integrate the cytoskeleton across an epithelial sheet [6,7,8]. α-Catenin also binds to the class III Lin-1, Isl-1, Mec-3 (LIM) protein Ajuba (a member of the zyxin family of proteins), which appears to function dually in both adhesion and in activation of the Ras-mitogen-activated protein kinase (MAPK) pathway [9,10]. Through these links, AJs appear able to couple adhesion with cytoskeletal dynamics as well as with nuclear and cytoplasmic signaling. This provides a framework for conceptualizing why E-cadherin levels appear to impact upon a plethora of developmental processes (reviewed in [11]).
As we probed more deeply into the underlying mechanisms governing E-cadherin promoter activity, we were intrigued by the close proximity of the LEF-1/β-catenin binding site to a site known to bind the Snail/Slug family of zinc finger transcriptional repressor proteins [12,13,14,15]. Both activity of Snail and down-regulation of E-cadherin play pivotal roles in epithelial to mesenchymal transitions (EMTs), typified by the transformation of polarized, adhering epithelial cells into motile mesenchymal cells [16,17]. Bud formation differs from an EMT in that E-cadherin activity needs to be down-regulated but not prevented, so that adhesive junctions are remodeled rather than quantitatively impaired. Supportive of an underlying ability to fine-tune cadherin expression at the transcriptional level, Snail seems to have an additive effect with LEF-1/β-catenin in negatively modulating E-cadherin promoter activity [4].
In the present study, we discovered that Snail is expressed briefly at an early stage of hair bud formation, when E-cadherin down-regulation and activation of proliferation take place. Thereafter, Snail disappears and remains absent during subsequent follicle down-growth and maturation. This exquisite pattern appears to be functionally relevant since altering it in vivo correspondingly affects features associated with hair bud formation, including down-regulation of E-cadherin, increased proliferation, and repressed terminal differentiation. Although the temporal spike of Snail in the hair bud is reflected at the mRNA level and seems to follow Wnt signaling and BMP inhibition, LEF-1/β-catenin activation does not appear to induce Snail gene expression in embryonic skin keratinocytes. In contrast, we provide in vitro, transgenic (Tg), and gene targeting evidence to show that TGF-β2 and small phenotype– and mothers against decapentaplegic–related protein 2 (SMAD2) signaling are upstream inducers of Snail gene expression in skin epithelium. In the absence of TGF-β2 signaling and Snail gene expression, hair placodes can form, but further follicle down-growth is blocked. Our studies point to the view that Snail likely functions downstream of cell fate specification, at a stage where the bud begins to exhibit enhanced proliferation and migration.