Regulation of neurotrophin signaling in aging sensory and motoneurons: dissipation of target support?

A hallmark of senescence is sensorimotor impairment, involving locomotion and postural control as well as fine-tuned movements. Sensory and motoneurons are not lost to any significant degree with advancing age, but do show characteristic changes in gene-expression pattern, morphology, and connectivity. This review covers recent experimental findings corroborating that alterations in trophic signaling may induce several of the phenotypic changes seen in primary sensory and motoneurons during aging. Furthermore, the data suggests that target failure, and/or breakdown of neuron-target interaction, is a critical event in the aging process of sensory and motoneurons.

Different distributions of the sensory and autonomic innervation among the microvasculature of the rat mystacial pad.

The regulation of the vasculature in the skin is a complex process involving both perivascular nerves and local endothelial-mediated control. In this study, the perivascular innervation in the mystacial pad of the rat was characterized based upon immunochemical and lectin binding characteristics and distribution. All of the innervation labeled with anti-protein gene product 9.5 (PGP 9.5), which was used in double- and triple-labeling combinations with the Griffonia simplicifolia lectin (GSA) and antibodies against a variety of neuropeptides, enzymes, and structural proteins. GSA histofluorescence revealed an intricate microvasculature within the rows of tactile vibrissae, which form a natural grid to standardize analyses. Specific features of the vascular organization were confirmed by scanning electron microscopy. Each interval between adjacent vibrissae contained a predictably organized microvascular module composed of separate arterial channels and capillary networks for each of several different structures: papillary muscles, facial muscles, the interior of vibrissal follicle-sinus complexes, vibrissal papillae, and the upper dermis of the intervibrissal fur. Each module was innervated by at least two sets of sensory, at least two sets of sympathetic, and at least one possible set of parasympathetic. These sets not only differed in their biochemical characteristics, but also in their relative position within the arterial walls and their distribution among the microvasculature to the various structures. As such, the microvasculature to each type of structure had a particular combination of innervation, suggesting that separate neuronal mechanisms may be involved in regulating the blood flow to different types of targets even within the confines of a small territory of tissue.

Innervation of nonmystacial vibrissae in the adult rat.

Vibrissal follicle-sinus complexes (F-SCs) in the mystacial pad of rodents are heavily innervated by different types of sensory nerve endings. One site in mystacial F-SCs, the inner conical body (ICB), is uniquely well innervated only in those species, such as the rat, that rhythmically whisk their mystacial vibrissae. In this study, we examined the innervation of rat nonmystacial F-SCs, which are not whisked. Supraorbital, posteroorbital, lateral cervical, median cervical, submental, and carpal forelimb F-SCs were cut on a cryostat and were either prepared for anti-human protein gene product (PGP 9.5) immunofluorescence or stained using the Winkelmann silver technique. Much of the innervation of the nonmystacial F-SCs is similar to that of mystacial F-SCs. All are innervated by a large deep vibrissal nerve (DVN) and several smaller superficial vibrissal nerves (SVNs). As in the mystacial pad, the SVNs show a distribution of Merkel and free nerve endings qualitatively similar to the rete ridge collar of all the nonmystacial F-SCs as well as provide circumferentially oriented endings to the ICBs to all but median-cervical and carpal F-SCs. Not only was the ICB innervation relatively sparse in median-cervical and carpal F-SCs, but a large portion of the carpal ICB innervation also ascended from the DVNs, which make only a small ICB contribution in other locations. Similar to mystacial pad F-SCs, the DVNs provided Merkel and lanceolate endings to the level of the ring sinus as well as reticular and irregular lanceolate-like endings to the level of the cavernous sinus. However, all but the posteroorbital F-SCs have relatively few lanceolate endings. Carpal F-SCs also have relatively few ring-sinus Merkel endings, which are diffusely distributed, are limited to the superficial portion of the outer root sheath. They also lack reticular and irregular lanceolate-like endings in the cavernous sinus. However, carpal F-SCs have a unique set of corpuscular endings in the ICB, ring sinus, and cavernous sinus that are rarely seen in other F-SCs. PGP 9.5 immunofluorescence also revealed two sets of fine-caliber profiles at the level of the ICB and ring sinus that were not previously seen in mystacial F-SCs. Although there was no correlation between ICB innervation and whisking, the regional variations in F-SC innervation suggest that functional differences may exist between vibrissae at different locations in the body.

Innervation of the hard palate in the rat studied by anterograde transport of horseradish peroxidase conjugates.

The innervation of the rat hard palate and the bordering part of the soft palate was studied after anterograde transport of horseradish peroxidase conjugated to wheat germ agglutinin (WGA-HRP) and to choleragenoid (B-HRP) in separate experiments. WGA-HRP labeling showed leakage from several types of nerve endings, whereas B-HRP did not. Both conjugates gave rise to heavy labeling of a variety of nerve endings. Intragemmal and, especially, perigemmal fibers were labeled in chemosensory corpuscles, which were most common in the medial wall of the incisive canal and in the most anterior part of the soft palate. Ruffini endings of different sizes were labeled in the incisive papilla. Other subepithelial endings forming elongated expanded profiles with medium- to large-caliber source fibers were most common in protruding parts of the palate. Labeled intraepithelial endings included Merkel endings, which were most frequent in the incisive papilla and the rugae. Other labeled profiles were medium-caliber afferents giving rise to irregular, beaded, and sometimes branched endings often located far superficially in the epithelium. Such endings were present both within and between protruding parts of the palate. Fine-caliber intraepithelial endings were labeled almost exclusively in WGA-HRP experiments.

Effects of aging and axotomy on the expression of neurotrophin receptors in primary sensory neurons.

Aging is accompanied by declined sensory perception, paralleled by widespread dystrophic and degenerative changes in both central and peripheral sensory pathways. Several lines of evidence indicate that neurotrophic interactions are of importance for a maintained plasticity in the adult and aging nervous system, and that changes in the expression of neurotrophins and/or their receptors may underpin senile neurodegeneration. We have here examined the expression of neurotrophin receptor (p75NTR, trkA, trkB, and trkC) mRNA and protein in intact and axotomized primary sensory neurons of young adult (3 months) and aged (30 months) rats. To examine possible differences among primary sensory neuron populations, we have studied trigeminal ganglia (TG) as well as cervical and lumbar dorsal root ganglia (DRG). In intact aged rats, a decrease in trk (A/B/C) mRNA labeling densities and protein-like immunoreactivities was observed. The decrease was most pronounced in lumbar DRG. In contrast, a small, not statistically significant, increase of p75NTR expression was observed in aged DRG neuron profiles. After axotomy, a down-regulation of mRNA and protein levels was observed for all neurotrophin receptors (p75NTR, trkA, trkB and trkC) in both young adult and aged rats. Consistent with the higher expression levels of neurotrophin receptors in unlesioned young adult primary sensory neurons, the relative effect of axotomy was more pronounced in the young adult than aged rats. Although a decrease in mean cell profile cross-sectional areas was found during aging and after axotomy, the characteristic distribution of neurotrophin receptor expression in different populations of NRG neurons was conserved. The present findings suggest an attenuation of neurotrophic signaling in primary sensory neurons with advancing age and that the expression of p75NTR and trks is regulated differently during aging. A similar dissociation of p75NTR and trk regulation has previously been reported in other neuronal systems during aging, suggesting that there may be a common underlying mechanism. Decreased access to ligands, disturbed axon function and systemic changes in androgen/estrogen levels are discussed as inducing and/or contributing factors.

Comprehensive immunofluorescence and lectin binding analysis of vibrissal follicle sinus complex innervation in the mystacial pad of the rat.

The innervation of the vibrissal follicle sinus complexes (FSCs) in the mystacial pad of the rat was examined by lectin binding histofluorescence with the B subunit of Griffonia simplicifolia (GSA) and by immunofluorescence with a wide variety of antibodies for neuronal related structural proteins, enzymes, and peptides. Only anti-protein gene product 9.5 labeled all sets of innervation. Several types of mechanoreceptors were distributed to specific different targets by medium to large caliber myelinated axons. All were positive for 200 kDa neurofilament subunit, peripherin, and carbonic anhydrase. Their endings expressed synaptophysin. Labeling for the 160 kDa neurofilament subunit, calbindin, and parvalbumin varied. Anti-Schwann cell protein S100 was completely co-extensive with the axons, terminal arbors, and endings of the mechanoreceptor afferents including Merkel innervation. At least 15 different sets of unmyelinated innervation were evident based upon distribution and labeling characteristics. They consisted of four basic types: 1) peptidergic; 2) GSA binding; 3) peptidergic and GSA binding; and 4) nonpeptidergic and GSA negative (peptide-/GSA-). Previous studies had not revealed that several major sets of unmyelinated innervation were peptide-/GSA-. The unmyelinated innervation had detectable peripherin but not 160 kDa or 200 kDa neurofilament subunits. GSA-positive axons uniquely lacked anti-S100 immunoreactivity. The dense circumferentially oriented unmyelinated innervation of the inner conical body contained major sets of peptide-/GSA- and GSA innervation as well as a smaller peptidergic GSA component. A small contingent of sympathetic and possibly parasympathetic innervation was affiliated with microvasculature in the FSCs. This study confirms and refutes some previous hypotheses about biochemical and morphological relationships between peripheral innervation and sensory ganglion cells.

Comprehensive immunofluorescence and lectin binding analysis of intervibrissal fur innervation in the mystacial pad of the rat.

The innervation of the intervibrissal fur in the mystacial pad of the rat and mouse was examined by immunofluorescence with a wide variety of antibodies for neuronal related structural proteins, enzymes, and peptides as well as for lectin binding histofluorescence with Griffonia simplicifolia (GSA). Anti-protein gene product 9.5 (PGP) immunofluorescence labeled all sets of axons and endings. The innervation in the upper dermis and epidermis was distributed through a four tiered dermal plexus. From deep to superficial, the second tier was the source of all apparent myelinated mechanoreceptors, the third tier of nearly all the peptidergic and GSA binding innervation, and the fourth tier of nonpeptidergic GSA negative innervation (peptide-/GSA-). Three types of mechanoreceptors-Merkel, transverse lanceolate, and longitudinal lanceolate endings-innervated guard hair follicles. All had similar labeling characteristics for 160 kDa and 200 kDa neurofilament subunits, peripherin, carbonic anhydrase, synaptophysin, and S100. Palisades of longitudinal lanceolate endings were part of piloneural complexes along circumferentially oriented sets of transverse lanceolate endings, peptidergic free nerve endings (FNEs), and peptide-/GSA- FNEs. The longitudinal lanceolate endings were the only mechanoreceptors in the mystacial pad that had detectable calcitonin gene-related peptide. The epidermis contained four types of unmyelinated endings: simple free nerve endings (FNEs), penicillate endings, cluster endings and bush endings. Only the simple FNEs were clearly peptidergic. Virtually all others were peptide-/ GSA-. Each bush ending was actually an intermingled cluster of endings formed by several unmyelinated axons and occasionally an Adelta axon. In contrast to the other unmyelinated innervation to the epidermis, bush endings labeled with an antibody against the Schwann cell protein S100. The necks and mouths of follicles, as well as superficial vasculature, were innervated by a mixture of unmyelinated peptidergic and/or GSA labeled sensory and sympathetic axons. Small presumptive sweat glands were innervated by three sets of peptidergic axons of which one was immunoreactive for somatostatin. Potential functions of the various sets of innervation are discussed.

Overexpression of nerve growth factor in skin causes preferential increases among innervation to specific sensory targets.

The impact of increased levels of skin-derived nerve growth factor (NGF) neurotrophin on sensory and sympathetic innervation to the mouse mystacial pad and postero-orbital vibrissae was determined. Consistent with an approximate doubling of neuron number in trigeminal and superior cervical ganglia, many components of the sensory and sympathetic innervation were substantially enhanced. Although the increased number of neurons raised the possibility that all types of innervation were increased, immunohistochemical analysis indicated that enhanced NGF production had a differential effect upon sensory innervation, primarily increasing unmyelinated innervation. This increased innervation occurred in specific locations known to be innervated by small, unmyelinated fibers, suggesting that NGF modulated sensory innervation density, but not targeting. In contrast, sympathetic innervation was not only increased but also was distributed to some aberrant locations. In the intervibrissal fur of the mystacial pad, both the number of sensory axons and branches appeared increased, whereas in vibrissal follicle sinus complexes, only branching increased. In some areas, sensory ending density was lower than expected based upon the size of the source nerve bundles suggesting that many axons and branches were surviving but failing to form functional endings. Furthermore, the immunochemical profile of innervation was altered in some sensory populations as demonstrated by the coexistence of RT-97 neurofilament labeling in calcitonin gene-related peptide (CGRP) positive axons, by the loss of substance P colocalization in some CGRP axons, and by an absence of neuropeptide Y labeling in tyrosine hydroxylase positive sympathetic axons. Collectively, these results indicate that the NGF mediated increase in neuron number may be selective for particular sets of innervation and that increases among some populations may result from phenotypic switching.

Impairment of peripheral sensory innervation in senescence.

Sensorimotor disturbances are common among elderly and one of the main factors depreciating life quality in senescence. Mechanistically sensory deficits during aging include not only degenerative and regressive events but also phenotypic switches among sensory neurons as well as remodeling of sensory innervation. The pattern of changes suggests that an underlying mechanism is a sustained dependence of sensory neurons on target tissues, and that this dependence, at least in part, appears to be mediated through signaling by target-derived trophic factors. This review presents and discusses evidence supporting this notion.

Alterations in mystacial pad innervation in the aged rat.

It is well established that sensory perception becomes impaired with advancing age and that, in parallel, dystrophy and degeneration of axons occur in sensory pathways. In this study, the impact of aging was examined in the mystacial pad, which receives a large variety of sensory nerve endings organized in a highly predictable pattern. Mystacial pad specimens from aged (30 months old) and young adult (2-3 months old) female Sprague-Dawley rats were processed, in parallel, for immunohistochemical analyses with antibodies against human neuronal cytoplasmic protein (protein gene product 9.5), transmitter enzymes, and several neuropeptides. Several changes in cutaneous innervation including both degenerative and regenerative processes were evident in the aged rat: (1) the Merkel endings and lanceolate endings that emanate from large-caliber afferents in the whisker follicles were reduced and showed signs of degeneration. Furthermore, a reduction of piloneural complexes at the intervibrissal hairs were evident, but only in aged rats that showed more severe behavioral sensorimotor disturbances. In contrast, Ruffini endings as well as mechanoreceptors emanating from medium-caliber axons, i.e., transverse lanceolate and reticular endings, appeared normal. (2) A reduction was evident among two sets of unmyelinated epidermal endings; however, the epidermal innervation affiliated with the intervibrissal hairs appeared normal in the aged rat. (3) A loss of sympathetic neuropeptide tyrosine (NPY) or tyrosine hydroxylase-immunoreactive (IR) and somatosensory Calcitonin gene-related peptide (CGRP)-IR perivascular axons was paralleled by an increase in presumed parasympathetic NPY/CGRP-IR axons. (4) Two "novel" networks of fine-caliber axons were observed in the outer and inner root sheaths of the whisker follicles in the aged rat. (5) NPY was present in a population of small-caliber, somatosensory CGRP-IR axons in the aged rat. This may represent a de novo synthesis, since, normally, NPY-like immunoreactivity is not observed in this set of axons. Our results suggest that the sensory impairments occurring with advancing age are part of a peripheral process instigated by changes in nerve-target interactions and/or incapacitation of the neuronal machinery to sustain the axonal integrity.

Regulation of NGF-family ligands and receptors in adulthood and senescence: correlation to degenerative and regenerative changes in cutaneous innervation.

During development, a highly differential neurotrophin dependency is reported for various types of nerve endings in the whisker follicle. To what extent these dependencies extend and play a role in adulthood is largely unresolved. We show here, using in situ hybridization and immunohistochemistry that the expression of neurotrophins and trk/p75 receptors persists in adulthood. As suggested by their expression profiles, many classes of cutaneous nerve endings disclose similar ligand-receptor dependencies in adult animals as during development, while other populations appear to switch their dependency. Furthermore, our data suggest that sensory endings that have a high turnover due to mechanical wear and tear, e. g. Merkel cell-neurite complexes at the level of ring sinus show a more complex ligand-receptor expression phenotype than do endings with a less vulnerable location, e.g. the Merkel cell-neurite complexes at the rete ridge collar. Thus, neurotrophin-3 (NT3)/trkA signalling is suggested to be important for a continuous terminal plasticity of Merkel cell-neurite complexes at the level of ring sinus in adulthood. Evidence supporting a role for neurotrophin signalling in maintaining the adult cutaneous innervation also comes from the close correlation between altered ligand-receptor expression(s) and axonal/terminal aberrations in senescence. Thus, an ageing-related decrease in target neurotrophin expression, notably NT3 and NT4, results in a site-specific loss of sensory terminals concomitant with an aberrant growth of regenerating/sprouting axons into new target fields. Ageing of the cutaneous innervation, manifested in degenerative and regenerative events, seems strongly associated with changes in neurotrophic interactions between sensory neurons and target tissues.

A rapid and dynamic regulation of GDNF-family ligands and receptors correlate with the developmental dependency of cutaneous sensory innervation.

Glial cell line-derived neurotrophic factor (GDNF) and neurturin (NTN) are members of the transforming growth factor-beta family and have been shown to elicit neurotrophic effects upon several classes of neurons including dopaminergic neurons, motoneurons, parasympathetic, sympathetic as well as primary sensory neurons. However, there is little information available on their roles in cutaneous innervation. Herein, we have studied the regulation of gdnf, ntn and the GDNF family receptors and examined their role in the development of facial cutaneous innervation in GDNF mutant mice. A dynamic spatial and temporal regulation of gdnf, ntn and their ligand binding receptors within the follicle-sinus complex correlate with development of distinct subclasses of sensory nerve endings. Furthermore, development of NGF-dependent myelinated mechanoreceptors, i.e. reticular and transverse lanceolate endings also require GDNF during ending formation and maintenance. In addition, ligand and receptor association seems to be intricately linked to a local Schwann cell-axon interaction essential for sensory terminal formation. Our results suggests that functionally specified nerve endings depend on different GDNF family members and that in contrast to neurotrophins, this family of neurotrophic factors may be acting at local sites of terminal Schwann cell-axon growth cone interactions and that they collaborate with neurotrophins by supporting the same populations of neurons but at different times in development.

The innervation of the mystacial pad in the adult rat studied by anterograde transport of HRP conjugates.

Peripheral and central terminations of mystacial pad afferents in rats were labeled by anterograde transport of wheat germ agglutinin-HRP (WGA-HRP) or choleragenoid HRP (B-HRP). Tracer was injected in the trigeminal ganglion and survival times were 6-24 h. Most of the innervation previously observed with other techniques in the mystacial pad were labeled by at least one of the tracers. This included extensive reticular endings from large-caliber afferents and a loose network of fine-caliber axons in vibrissal follicle-sinus complexes (F-SCs). Also included were individual highly branching bush-like profiles in the intervibrissal epidermis that arose from fine to medium caliber afferents. Other endings were revealed, such as beaded endings affiliated with tylotrich hairs and presumptive encapsulated lamellated endings affiliated with both vibrissae and small sinus hairs. Finally, the anterograde labeling also revealed differences in the branching pattern of Merkel afferents to the rete ridge collars and ring sinuses of F-SCs. Each tracer produced different patterns of labeling related to the survival time in the mystacial pad which corresponded to particular patterns of labeling in the trigeminal nucleus caudalis. WGA-HRP produced dense labeling of all types of afferents and peripheral endings as well as all laminae of nucleus caudalis after short survivals, but the labeling diffused as the survival times were increased. B-HRP preferentially filled the largest afferents and endings after shorter survivals, while smaller profiles became progressively labeled after longer survivals. In nucleus caudalis, profiles extending into laminae III, IV and inner part of lamina II were labeled with B-HRP after shorter survivals, but the outer part of lamina II also became labeled with longer survivals. This has not been previously observed with B-HRP. Along with other recent findings, these results reveal that the innervation of the mystacial pad especially by fine-caliber axons is far more extensive and complex than previously described. Also, depending on the survival time, the central and peripheral labeling patterns differ, which must be taken into account when interpreting results using these two tracers.

Differential dependency of cutaneous mechanoreceptors on neurotrophins, trk receptors, and P75 LNGFR.

The impact of null mutations of the genes for the NGF family of neurotrophins and their receptors was examined among the wide variety of medium to large caliber myelinated mechanoreceptors which have a highly specific predictable organization in the mystacial pad of mice. Immunofluorescence with anti-protein gene product 9.5, anti-200-kDa neurofilament protein (RT97), and anti-calcitonin gene-related product was used to label innervation in mystacial pads from mice with homozygous null mutations for nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3), neurotrophin-4 (NT-4), the three tyrosine kinase receptors (trkA, trkB, trkC), and the low-affinity nerve growth factor receptor p75. Specimens were sacrificed at birth and at 1, 2, and 4 weeks for each type of mutation as well as at 11 weeks and 1 year for p75 and trkC mutations, respectively. Our results demonstrate several major concepts about the role of neurotrophins in the development of cutaneous mechanoreceptors that are supplied by medium to large caliber myelinated afferents. First, each of the high-affinity tyrosine kinase receptors, trkA, trkB, and trkC, as well as the low-affinity p75 receptor has an impact on at least one type of mechanoreceptor. Second, consistent with the various affinities for particular trk receptors, the elimination of NGF, BDNF, and NT-3 has an impact comparable to or more complex than the absence of their most specific high-affinity receptors: trkA, trkB, and trkC, respectively. These complexities include potential NT-3 signaling through trkA and trkB to support some neuronal survival. Third, most types of afferents are dependent on a different combination of neurotrophins and receptors for their survival: reticular and transverse lanceolate afferents are dependent upon NT-3, NGF, and trkA; Ruffini afferents upon BDNF and trkB; longitudinal lanceolate afferents upon NGF, trkA, BDNF, and trkB; and Merkel afferents on NGF, trkA, NT-3, trkC, and p75. NT-4 has no obvious detrimental impact on the mechanoreceptor development in the presence of BDNF. Fourth, NT-4 and BDNF signaling through trkB may suppress Merkel innervation and NT-3 signaling through trkC may suppress Ruffini innervation. Finally, regardless of the neurotrophin/receptor dependency for afferent survival and neurite outgrowth, NT-3 has an impact on the formation of all the sensory endings. In the context of these findings, indications of competitive and suppressive interactions that appear to regulate the balance of innervation density among the various sets of innervation were evident.

Differential dependency of unmyelinated and A delta epidermal and upper dermal innervation on neurotrophins, trk receptors, and p75LNGFR.

The impact of the nerve growth factor (NGF) family of neurotrophins and their receptors was examined on the cutaneous innervation in the mystacial pads of mice. Ten sets of unmyelinated and thinly myelinated sensory and autonomic innervation were evaluated that terminated in the epidermis, upper dermis, and upper part of the intervibrissal hair follicles. Mystacial pads were analyzed from newborn to 4-week-old mice that had homozygous functional deletions of the genes for NGF, brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3), neurotrophin-4 (NT-4), tyrosine kinase (trk) A, trkB, trkC, or p75. Mystacial pads were also analyzed in adult transgenic mice that had overproduction of NGF, BDNF, or NT-3 driven by a keratin promoter gene. The innervation was revealed by using immunofluorescence and immunocytochemistry with antibodies for protein gene product (PGP) 9.5, calcitonin gene-related product (CGRP), substance P (SP), galanin (GAL), neuropeptide Y (NPY), tyrosine hydroxylase (TH), and a neurofilament protein. The cumulative results indicated that NGF/trkA signaling plays a major role in the outgrowth and proliferation of sensory axons, whereas NT-3/ trkA signaling plays a major role in the formation of sensory endings. TrkC is also essential for the development of three sets of trkA-dependent sensory innervation that coexpress CGRP, SP, and GAL. Another set of sensory innervation that only coexpressed CGRP and SP was solely dependent upon NGF and trkA. Surprisingly, most sets of trkA-dependent sensory innervation are suppressed by trkB perhaps interacting with p75. BDNF and NT-4 appear to mediate this suppressing effect in the upper dermis and NT-4 in the epidermis. In contrast to sensory innervation, sympathetic innervation to the necks of intervibrissal hair follicles depends upon NGF/trkA signaling interacting with p75 for both the axon outgrowth and ending formation. Although NT-3/trkA signaling is essential for the full complement of sympathetic neurons, NT-3 is detrimental to the formation of sympathetic terminations to the necks of hair follicles. TrkB signaling mediated by BDNF but not NT-4 also suppresses these sympathetic terminations. One sparse set of innervation, perhaps parasympathetic, terminating at the necks of hair follicles is dependent solely upon NT-3 and trkC. Taken together, our results indicate that the innervation of the epidermis, upper dermis, and the upper portion of hair follicles is regulated by a competitive balance between promoting and suppressing effects of the various neurotrophins.