Nerve endings reveal hidden diversity in the skin

Experiments on transgenic mice have revealed that the morphologies of sensory neurons in the skin of mice are more complex and diverse than expected.

Image Nerve endings associated with hair follicles in the skin of a mouse (for sensing touch), and proprioceptors (which sense body position). Nociceptors and mechanore ceptors terminate in the skin, whereas propriocep tors termin ate in muscles and tendons.
It has long been recognized that nerve end ings in the skin display a diverse range of forms, but prior studies have generally used histological methods in tissue sections that do not reveal the complete morphology of each neuronal axon. Now, writing in eLife, Hao Wu, John Williams and Jeremy Nathans of Johns Hopkins University report the results of experiments that involve some very modern transgenic tricks, yet evoke the studies of neuronal morphology in the early 20 th century-using methods introduced by Golgi and perfected by Ramón Y Cajal-that first revealed the complex architecture of single neur ons (see Figure 1). The results of the Johns Hopkins experiment are largely descriptive in nature, and we might assume that the results reported are already buried somewhere in the lit erature, but they are not. Thus we are reminded that our knowledge of even wellstudied experi mental systems is still very fragmentary. This latest work was made possible by the CreLox system-a widelyused approach in which a Cre recombinase enzyme is used to remove chromosomal DNA flanked by two genet ically engineered loxP recognition sequences. Wu, Williams and Nathans used this method to excise a signal sequence blocking the expres sion of a histochemical marker gene which had been previously engineered into the chromosomal location of a transcription factor (Brn3a) that is important in the development of the sensory ner vous system. A key feature of the experiments was the use of a form of the Cre enzyme that is translocated to the nucleus only in the presence of an estrogenlike drug, tamoxifen, so the prob ability that the marker gene is expressed should be related to tamoxifen concentration. By trial anderror titration of the tamoxifen dose adminis tered to pregnant mice bearing transgenic litters, it was possible to label a small number of discrete sensory neurons.
The Johns Hopkins researchers attempt to systematically categorize, for the first time, the complex axons of many individual sensory neur ons with respect to their morphology, the number and density of their endings, and also their rela tionship to hair follicles, where most of the fib ers terminate. They acknowledge that this must be a preliminary system, because only a subset of sensory neurons is sampled. One class of axon ter minal may be consistent with fibers con veying itch sensation, but most of the endings conveying pain, pleasant and unpleasant tem peratures, and chemical irritants are probably not revealed by the labeling method used. One particularly interesting class of labeled neuron possesses axons with Cshaped endings that only partly encircle hair follicles, and that termi nate on a consistent side of the follicles arrayed across a large skin area. This structure may be especially suited to convey the direction of move ment of tactile stimuli across the skin. Recently it has become possible to correlate some well known markers of sensory subtypes with the terminal morph ology and electrophysiological properties of mechanoreceptors (Li et al., 2011), and the function of most of the mechanorecep tors identified here await further physiological studies.
It will be interesting to see to what extent these diverse patterns of nerve endings are pre ordained by gene regulatory programs during the developmental period when the axons are growing out from the sensory ganglia, and how much is adaptive. The overall gene regulatory cascade for the early specification of pain, touch and proprioceptive somatic sensory neurons is now fairly well understood (Liu and Ma, 2011). In mice lacking the transcription factors Islet1 and Brn3a (the gene locus used to target the reporter in this study), sensory neurons remain in a generic 'ground state' of differentiation and express few subtype specific markers (Dykes et al., 2011). Recent work has shown that the receptor tyro sine kinase cRet (Luo et al., 2009) and the tran scription factor cMaf (Wende et al., 2012) are required for the development of some classes of mechanoreceptors. However, none of these developmental mechanisms comes close to offer ing an explan ation for the diversity of sensory arbors observed by Wu, Williams and Nathans. If Figure 1. Drawings of neurons and nerve endings made more than a century apart. The drawing on the left was made by Santiago Ramón Y Cajal in 1899 and shows Purkinje cells in the cerebellum of a pigeon. The cells were stained with potassium dichromate and silver nitrate. The trace on the right shows nerve endings in the skin of a mouse. A combination of genetic and histochemical techniques were used to record the image from which the trace is taken (Wu et al., 2012).

IMAGE: INSTITUTO SANTIAGO RAMÓN Y CAJAL
a large part of the diversity observed in the present study is genetically determined, then much the regulatory program of sensory differen tiation is yet unknown.
A related question is whether the pattern of arborization will be similar in regions of skin with very different sensory properties. It is well known that the size of touch receptive fields differs widely between areas that are sparsely innervated, such as the trunk skin studied by the Johns Hopkins group, and those that are densely innervated, such as the face and fingertips. In areas with finer resolution of tactile stimuli, such as the distal limbs and face, the arbors should have smaller territories and/or have more densely packed endings. Also, since the majority of neur ons described here innervate hair follicles, the glabrous skin of the hands and feet must neces sarily have differently structured endings and arbors.
It is remarkable that in the brachial and lumbar regions, the receptive field of a single dorsal root ganglion may encompass a continuous area of skin from the midback to the tips of the digits (Takahashi et al., 2003). A single dorsal root ganglion must therefore contain at least three fundamental classes of sensory neuron, with remarkable morphological and functional diversity expressed in each class. The Johns Hopkins study is only a 'first pass' at a systematic description of this diversity, so it is possible that a truly remarkable range of neuronal form and function will ultim ately be found within a single sensory gan glion that contains just a few thousand neurons.