The skin is one of the largest organs in the body, with important functions in somato-sensation, insulation/thermoregulation, and immune defense1. Understanding the molecular and cellular basis of skin development and function has been of longstanding interest because of the fundamental importance of skin as a biological system and its relevance to dermatology. Mammalian skin contains a variety of multicellular structures, including stratified layers of keratinocytes, dermal connective tissue, several types of hair follicles, sebaceous glands, arrector pili muscles, blood vessels, and at least a dozen distinct classes of afferent (sensory) and efferent nerve fibers (Figure 1). Different regions of the body are associated with characteristically different types of skin. In most mammals, nearly the entire body surface is covered with skin that is densely packed with hair follicles. [Humans and naked mole rats constitute exceptions to this pattern.] Hair is missing from the palmar surfaces of the hands and feet, which are also associated with specialized epidermal patterns (dermatoglyphs), exocrine glands, and sensory nerve endings. The cellular and molecular events that control the growth, differentiation, and spatial arrangement of cells within the hair follicle are of special interest as each follicle exhibits, in miniature, many of the central features of organogenesis2. These features include the existence of stem cells and a stem cell niche, precisely choreographed cell migrations, and the assembly of multicellular structures from embryologically distinct components.
This article describes methods for dissecting, fixing, labeling, and imaging mouse skin as an intact two-dimensional sheet, referred to as a “whole mount” or “flat mount” preparation. Since mouse skin is relatively thin, it is possible to image through the full thickness of flattened skin using conventional confocal microscopy. The flat mount approach to imaging mammalian skin is technically advantageous because it bypasses the need for physical sectioning, thereby allowing structures to be reconstructed entirely by optical sectioning. Since nearly the entire skin is processed as a single object, the flat mount approach also facilitates the imaging of multiple regions of the body surface while preserving information about position and orientation relative to the body axes. Finally, structures within the skin are typically present in patterns that are repeated at regular intervals, thus facilitating the collection of images from multiple representatives of a given structure. These characteristics are familiar to neurobiologists who work on the retina, a two-dimensional part of the central nervous system that enjoys analogous advantages for studies of neuronal morphology3.
The flat mount approach described here is of special utility for studying structures that exhibit spatial organization on a relatively large scale within the two-dimensional plane of the skin. One example of large-scale spatial organization is the coordinated polarity of hair follicles and hair follicle-associated structures - Merkel cell clusters, arrector pili muscles, sebaceous glands, and nerve endings4. Hair follicles are oriented at an angle with respect to the plane of the skin, and the component of the follicle vector that lies within the 2-dimensional plane of the skin generally exhibits an orientation with respect to the body axes that is precisely determined for each position on the body. For example, hair follicles on the back point from rostral to caudal and hair on the dorsal surface of the feet point from proximal to distal. Hair follicle orientation is controlled by planar cell polarity signaling (PCP; also called tissue polarity5). This signaling system was discovered in Drosophila where a small set of core PCP genes was found to control the orientation of cuticular hairs and bristles. Three mammalian orthologues of core PCP genes - frizzled homolog 6 (Fzd6, also referred to as Fz6), cadherin EGF LAG seven-pass G-type receptor 1 (Celsr1), and vang-like 2 (Vangl2) - play analogous roles in mammalian skin, coordinating the orientations of hair follicles with the body axes. Studies of Fz6 knockout mice (Fzd6tm1Nat, hereafter referred to as Fz6-/-) show that the primary defect in the absence of PCP signaling is an initial randomization or disorganization of hair follicle orientation, with no effect on the intrinsic structure of the follicles6-8. A second non-PCP system acts later to promote local alignment of nearby follicles, which leads to the production of large-scale hair patterns such as whorls and tufts.
A second example of large-scale spatial organization within the skin is seen in the morphologies of sensory axon arbors. Sensory neurons that innervate the skin have their cell bodies in the dorsal root and trigeminal ganglia. These neurons detect temperature, pain, itch, and various types of mechanical deformations impinging on the skin and hair9. They can be divided into subtypes based on axon diameter and conduction velocity, terminal nerve ending structure, and the patterns of expression of receptors, channels, and other molecules. Because of the high density of innervation within the skin, analyses that involve visualizing all axons (e.g., anti-neurofilament immunostaining) or even all axons of a single class (as seen when a single cell type is marked by expression of a fluorescent reporter) generally reveals a dense superposition of axons that makes it impossible to define the morphology of an individual arbor. To circumvent this problem, we have used extremely sparse genetically-directed labeling to produce dorsal skin samples in which individual well-isolated axon arbors are visualized by expression of a histochemical reporter, human placental alkaline phosphatase10. This approach allows the unambiguous visualization of individual axon arbor morphologies and a definition of somatosensory neuron types based on morphologic criteria.