Journal of Dermatological Science 65 (2012) 162–169

The dimensions and characteristics of the subepidermal nerve plexus inhuman skin – Terminal Schwann cells constitute a substantial cell populationwithin the superficial dermis

Christina M. Reinisch a, Erwin Tschachler a,b,*a Department of Dermatology, Medical University of Vienna, Waehringer Gurtel 18-20, A-1090 Vienna, Austriab Centre de Recherches et d’Investigations Epidermiques et Sensorielles (CE.R.I.E.S.), Neuilly, France

A R T I C L E I N F O

Article history:

Received 1 August 2011

Received in revised form 19 October 2011

Accepted 20 October 2011

Keywords:

Subepidermal nerve plexus

Schwann cells

Nerve growth factor receptor p75

Neurofilament

Protein-gene-product 9.5

A B S T R A C T

Background: The skin constitutes the largest sensorial organ. Its nervous system consists of different

types of afferent nerve fibers which spread out immediately beneath the skin surface to sense

temperature, touch and pain.

Objective: Our aim was to investigate the dimension and topographic relationship of the different nerve

fibers of the subepidermal nerve plexus in human hairy skin and to analyze numbers and marker

expression of terminal Schwann cells.

Methods: Nerve fibers and Schwann cells were investigated on dermal sheet preparations and thick

sections of skin from various body regions of 10 individuals.

Results: The dimension of subepidermal nerve fibers varied between different body sites with highest

values in chest skin (100 � 18 mm/mm2) and lowest in posterior forearm skin (53 � 10 mm/mm2). The

majority of fibers (85.79%) were unmyelinated, thus representing C-fibers, of which 7.84% were peptidergic.

Neurofilament-positive fibers (A-fibers) accounted for 14.21% and fibers positive for both neurofilament and

myelin (Ab-fibers) for only 0.18%. The number of Schwann cells varied in accordance with nerve fiber length

from 453 � 108 on chest skin to 184 � 58/mm2 in skin of the posterior forearm. Terminal Schwann cells

showed a marker profile comparable to Schwann cells in peripheral nerves with the notable exception of

expression of NGFr, NCAM, L1CAM and CD146 on myelinating Schwann cells in the dermis but not in

peripheral nerves.

Conclusion: Our data show that terminal Schwann cells constitute a substantial cell population within

the papillary dermis and that both nerve fiber length and Schwann cell numbers vary considerably

between different body sites.

� 2011 Japanese Society for Investigative Dermatology. Published by Elsevier Ireland Ltd. All rights

reserved.

Contents lists available at SciVerse ScienceDirect

Journal of Dermatological Science

jou r nal h o mep ag e: w ww .e lsev ier . co m / jds

1. Introduction

The subepidermal nerve plexus is the most peripheral part ofthe nervous system and represents the largest sensory organ. It isat the origin of physiological temperature and mechanicalperceptions as well as pathological sensations of pain and pruritus.Due to its location of only a few microns away from the interfacewith the environment, subepidermal nerves are prime targets fordamage due to toxic and infectious agents [1–3]. In metabolicdiseases such as diabetes mellitus, an early sign of nerve damage isthe reduction of subepidermal nerve fibers [4–6].

Schwann cells are required for the survival of sensory neuronsin the peripheral nervous system. Mice lacking ErbB3 and therefore

* Corresponding author. Tel.: +43 1 4081271; fax: +43 1 4034922.

E-mail address: Erwin.Tschachler@meduniwien.ac.at (E. Tschachler).

0923-1811/$36.00 � 2011 Japanese Society for Investigative Dermatology. Published b

doi:10.1016/j.jdermsci.2011.10.009

devoid of peripheral Schwann cells develop severe neuropathies asa consequence of neuron degeneration [7]. In addition Schwanncells play crucial roles for nerve regeneration after peripheral nerveinjury [8] and as glial support of axon function [9]. In the skinterminal Schwann cells ensheath the nerve endings entirely exceptfor the parts which enter the epidermis [10]. Apart from their rolefor nerve physiology and regeneration, terminal Schwann cellsrepresent direct targets for infectious agents such as HSV [3] andMycobacterium leprae [2] and give rise to tumors in patientssuffering from neurofibromatosis [11].

The vast majorities of studies on the skin nervous system havebeen focused on its electrophysiological properties and have led tothe definition of the different fiber qualities. According to theirtransmitted sensory modality, they can roughly be classified intonociceptors, thermo- and mechanoreceptors [12,13]. Ab-fibers inthe skin are mainly conveying innocuous mechanical stimuliwhereas Ad-fibers conduct primarily temperature and nociceptive

y Elsevier Ireland Ltd. All rights reserved.

C.M. Reinisch, E. Tschachler / Journal of Dermatological Science 65 (2012) 162–169 163

information [12]. C-fibers are also thought to be mostly tempera-ture receptors and nociceptors [12] but recently it has beenreported that distinct C-fiber subsets are able to signal pleasanttouch sensations [14] whereas others mediate itch sensations [15].

Quantification of nerve fibers in the skin by immunofluores-cence is used to determine nerve damage in patients withperipheral neuropathies [16]. Whereas this is mostly performedby analyzing thick sections [17], we have recently proposed amethod for analyzing the subepidermal nerve plexus using dermalsheet preparations allowing for the analysis of larger areas [10].

Here we use both thick sections and dermal sheet preparationsto study the constituents of the skin nervous system at differentbody sites in situ and report on the proportion and microanatomyof the different types of nerve fibers and the distribution, numberand marker expression of terminal Schwann cells in the skin.

2. Materials and methods

2.1. Preparation of dermal sheets

Split thickness skin for dermal sheet preparation was obtainedat autopsy by the use of a dermatome. In total, skin from 10individuals was analyzed. Only skin from individuals withouthistory of skin diseases nor neurological disease was used. Fromeach individual, samples were taken from the shoulder, chest,abdomen, thigh, lower leg, upper arm, anterior and posteriorforearm. The age of the 4 male and 6 female donors ranged from 42to 82 years (62.4 mean � 12.1 standard deviation). The protocol wasapproved by the Regional Committee for Medical Research Ethicsaccording to the Declaration of Helsinki. Separation of split thicknessskin into dermis and epidermis was performed as previouslydescribed [10]. In brief, samples were floated on 3.8% ammo-niumthiocyanate solution in PBS for 30 min at 37 8C. Subsequently,the epidermis was separated from the underlying dermis andepidermal and dermal sheets were washed in PBS and fixed in 4%paraformaldehyde for 3 h at 4 8C followed by incubation in 10%sucrose in PBS over night. Sheets were either directly used forimmunostaining or stored frozen at �70 8C for later use.

2.2. Preparation of thick sections

Thick sections were made by cutting 120 mm thick sections ofparaformaldehyde fixed frozen skin biopsies on a CryostatMicrotome (Thermo Scientific, Walldorf, Germany) placed in 96well plates in PBS and used for staining immediately.

2.3. Preparation of peripheral nerve sections

4 mm thick sections of paraformaldehyde fixed frozen periph-eral nerves were cut on a Cryostat Microtome (Thermo Scientific,Walldorf, Germany) and mounted on object slides.

2.4. Immunostaining procedures

Immunofluorescent and immunohistochemical staining ofdermal sheets was essentially carried out as described [10]. Thefollowing primary antibodies were used: mouse monoclonal anti-nerve growth factor receptor p75 (Neomarkers, Fremont, CA; Cat.#MS-394-P) (NGFr), chicken polyclonal anti-neurofilament heavy(Chemicon, Temecula, CA; Cat.# AB5539) (NF), rabbit polyclonalanti-protein gene product 9.5 (Biogenesis, Poole, UK; Cat.# 7863-0507) (PGP9.5), mouse monoclonal anti-calcitonin gene relatedpeptide (Genex, Helsinki, Finnland; Cat.# M-9013000) (CGRP),mouse monoclonal anti-neural cell adhesion molecule/CD56(Becton Dickinson, San Jose, CA; Cat.# 347740) (NCAM), mousemonoclonal anti-L1 cell adhesion molecule (Neomarkers, Fremont,

CA; Cat.# MS770-P) (L1CAM), mouse monoclonal anti-CD146(Chemicon, Temecula, CA, Cat.# MAB16985) (CD146), mousemonoclonal anti-S100 (Neomarkers, Fremont, CA; Cat.# MS-296-P), mouse monoclonal anti-Vimentin (Dako, Glostrup, Denmark;Cat.# M7020) and rat monoclonal anti-myelin basic protein(Chemicon, Temecula, CA; Cat.# MAB386) (MBP). Isotype controlswere diluted to the same final concentration used for therespective primary antibody. They were purchased from SantaCruz Biotechnology (chicken IgY (Cat.# sc-2718) and rabbit IgG(Cat.# sc-2027), Santa Cruz, California) and Beckman Coulter(mouse IgG1 (Cat.# PN 6602872) and mouse IgG2a (Cat.# PN6602876), Miami, Florida). The biotinylated secondary antibodieswere anti-mouse Ig (Amersham Biosciences, Little Chalfont, UK;Cat.# RPN 1001) and anti-rat IgG (Vector Laboratories, Burlingame,CA; Cat.# BA-9400). The fluorophore labeled secondary antibodieswere all Alexa Fluor1 from Molecular Probes (546 goat anti-chicken IgG (Cat.# A-11040), 488 or 546 or 633 goat anti-mouseIgG (Cat.# A-11001, A-11003, A-21052) and 488 or 546 or 633 goatanti-rabbit IgG (Cat.# A-11034, A-11010, A-21070), 488 or 546goat anti-rat IgG (Cat.# A-11006, A-11081), Eugene, OR).

On 120 mm thick tangential skin sections, immunofluorescenttriple stainings with a Schwann cell marker (NGFr, NCAM, L1CAM,CD146, S100, Vimentin) in combination with MBP and a nerve fibermarker (NF or PGP9.5) were performed. 4 mm thick peripheral nervesections were immunohistochemically stained with the followingmarkers: NGFr, NCAM, L1CAM, CD146, S100, Vimentin, MBP.

2.5. Light and laser scanning microscopy analysis

Immunohistochemically stained sheets were analyzed with aconventional light microscope (Olympus Provis, Vienna, Austria).Beginning at the basement membrane, stacks comprising thesubepidermal nerve plexus located in the papillary dermis wereacquired with a 20� objective and projected in one picture by usingMeta Morph1 Software Version 4.5 from Universal ImagingCorporation (West Chester, PA, USA). Sheets immunohistochemi-cally stained for NGFr were used for the analysis of the Schwann cellnumber and the length of the subepidermal nerve plexus. Thenumber of Schwann cells was manually counted in each of theacquired fields at the light microscope. In order to measure only thelength of the subepidermal nerve plexus and to avoid a measure-ment bias by the variation of the nerve fiber thickness, the thicknessof each nerve fiber was reduced to 1 pixel. Thereafter, the total lengthcould be displayed in the region statistics in Meta Morph1.

Fluorescently stained skin samples were analyzed on a ZeissAxiovert 200 M laser scanning microscope. Zeiss LSM 510 software(Jena, Germany) was used. Stacks of images acquired with a 40� oilobjective ranged in thickness from 39.92 mm to 84.98 mm andcomprised 40–85 slices, each spaced by approximately 1 mm.Stacks of images of the sheets featuring MBP staining wereacquired with a 20� objective, ranged in thickness from 75 mm to182.03 mm and comprised 20–46 slices, each spaced by approxi-mately 3.9 mm. These stacks were used to measure the nerve fiberlength. For illustration purposes, all images from 1 stack wereprojected in one picture. As the analyzed areas exceeded the fieldacquired in one stack by far, adjacent images were aligned in AdobePhotoshop (San Jose, California) in order to yield a comprehensivepicture from each subepidermal nerve plexus. 3D analysis ofimmunofluorescently stained sheets of chest skin was done by theuse of Imaris1 (Bitplane Ag, Zurich, Switzerland) on LSM stacks.

2.6. Data calculation and statistical analysis

By light microscopy, three non-coherent visual fields of eachdonor and body region were analyzed and the mean wascalculated. As one visual field comprised an area of

Fig. 1. Depiction of the subepidermal nerve plexus by immunohistochemical staining for NGFr. A dense and regular nerve fiber network is visible in skin of the chest (A). In skin of

the lower leg (B), the nerve plexus is regular as well, yet apparently less dense. The protrusions in the course of the nerve fibers represent Schwann cell nuclei. The arrows in panel A

and B point to some exemplary Schwann cell nuclei. (C) and (D) show the corresponding images reduced to black and white graphs on which the length of the nerve fibers was

calculated as described in Section 2. Bar = 100 mm. Shown are representative pictures of skin preparations of donor 10. In this donor the nerve fiber length was 129 � 12 mm/mm2

in chest skin (C) and 43 � 5 mm/mm2 in skin of the lower leg (D). The number of Schwann cells was 666 � 72 in chest skin (A) and 150 � 8 in skin of the lower leg (B).

Table 1Number of Schwann cells/mm2 and nerve fiber length in mm/mm2 in the

subepidermal nerve plexus of different body regions are shown. Mean number

(�standard deviation) of 10 donors for each body site and all body sites combined is

given.

Body site Number of Schwann

cells/mm2

Nerve fiber length

mm/mm2

Chest 453 (�108) 100 (�18)

Abdomen 411 (�102) 78 (�22)

Anterior forearm 269 (�70) 66 (�19)

Thigh 242 (�70) 59 (�16)

Upper arm 238 (�91) 66 (�20)

Shoulder 236 (�64) 73 (�13)

Posterior forearm 184 (�58) 53 (�10)

Lower leg 176 (�53) 42 (�16)

Mean (�SD) 276 (�95) 67 (�16)

C.M. Reinisch, E. Tschachler / Journal of Dermatological Science 65 (2012) 162–169164

0.262848 mm2, the mean was extrapolated to obtain data per1 mm2. These values of each donor and region were then used tocalculate the mean and standard deviation (SD) of all 10 donors foreach region.

By laser scanning microscopy, a square area of 9 coherent visualfields comprising 0.478 mm2 was analyzed and the data extrapo-lated to 1 mm2. In this way, the data for the length of NGFr, NF,PGP9.5 and CGRP positive fibers were collected. Because of therareness of MBP positive fibers, the method of analysis describedabove would be heavily biased by the selection of the visual fieldacquired. Therefore, all positive fibers present in the respectivesheet were analyzed and the sum was used in combination withthe total sheet size to calculate the length in mm/mm2.

Calculation of mean and SD was performed with MicrosoftExcel. Statistical analysis was done with the StatView1 software(SAS Institute, Cary, NC, USA).

3. Results

3.1. The dimension of the subepidermal nerve plexus varies

significantly according to body site

By immunohistochemical staining for NGFr, the subepidermalnerve plexus appeared as a regular network (Fig. 1A and B) as wehave reported previously [10] in all body sites examined. Theanalysis extended from the dermo-epidermal junction to about100 mm into the papillary dermis. Branching of the plexus wasobserved within 78.31 � 11.54 mm (mean � standard deviation;measured in chest skin) of the dermis. Schwann cells were readilyidentifiable since their nuclei formed protrusions along the nervefibers [10] (Fig. 1A and B, arrows). The highest total nerve fiber lengthas well as the highest number of Schwann cells per mm2 of dermis

was found in chest and abdominal skin (Table 1). By contrast, in skinfrom the posterior forearm and the lower leg both nerve fiber lengthand Schwann cell numbers were reduced to about half of that in chestskin (Table 1). Regardless of the body region nerve fiber lengthshowed no association with age or sex of the patient. By contrastSchwann cell numbers were negatively correlated with donor age inskin from the lower leg and were correlated with the sex in skin fromthe lower leg with lower numbers in male donors.

To evaluate whether the results on nerve fiber length weredependent on analysis by conventional microscopy or laser scanningmicroscopy which reconstitutes a three dimensional representation,the two methods were compared on samples from chest skin. Two-dimensional analysis revealed a nerve fiber length comparable tothat obtained by three-dimensional analysis (Table 2) in most of thepatients analyzed, showing that both methods are equally suited tostudy the extension of the skin nervous system.

Table 2Comparison of two- and three-dimensional analysis of nerve fiber length of chest skin in 10 donors.

Age Sex Immunohisto-chemical detection

and light microscopy (2-D analysis)

Immunofluorescent detection and

laser scanning microscopy (3-D analysis)

NGFr-positive fibers/mm2 NGFr-positive fibers/mm2

Donor 1 63 Male 72.15 95.87

Donor 2 75 Female 101.59 102.33

Donor 3 42 Female 125.05 150.95

Donor 4 45 Female 104.05 123.82

Donor 5 75 Male 94.16 155.34

Donor 6 64 Female 70.31 84.03

Donor 7 58 Male 100.71 85.64

Donor 8 60 Male 108.82 101.26

Donor 9 60 Female 90.8 120.61

Donor 10 82 Female 129.34 130.81

Mean (�standard error) 99.7 (�18.33) 115.07 (�25.42)

C.M. Reinisch, E. Tschachler / Journal of Dermatological Science 65 (2012) 162–169 165

3.2. Schwann cells of myelinating nerve fibers in the skin express a

marker profile different from that of myelinating fibers in peripheral

nerves

Antibodies to NGFr, NCAM, L1CAM, CD146 and MBP were usedto depict Schwann cells by immunohistochemistry and immuno-fluorescence in dermal sheet preparations, 120 mm thick skin

Fig. 2. Marker expression profile of Schwann cells of the SNP as compared to Schwann ce

immunohistochemical staining was performed on dermal sheets. In (B) 120 mm thick sk

green), for MBP (shown in red) and for a nerve fiber marker (displayed in blue). NF wa

combination with S100, Vimentin and MBP. The encircled area is shown in a higher ma

marker combinations. Again, the encircled area is shown in a higher magnification (E

Schwann cell marker (F) and the appropriate isotype control (G) are shown. Bar = 100

sections and 4 mm thick peripheral nerve sections. Staining forS100 and Vimentin were used as positive controls. The Schwanncell markers with MBP and either NF or PGP9.5 were used toperform triple stainings (Fig. 2). Nonmyelinating Schwann cells inboth the skin and peripheral nerves were positive for NGFr, NCAM,L1CAM and CD146. The same marker profile was found onSchwann cells in myelinating fibers in the skin. By contrast,

lls in peripheral nerves. In each row, the same Schwann cell marker was used. In (A)

in sections were triple stained for the respective Schwann cell marker (displayed in

s used in combination with NGFr, NCAM, L1CAM, CD146 and PGP 9.5 was used in

gnification in (C). Peripheral nerve sections were triple stained (D) with the same

). Immunohistochemical staining of peripheral nerve sections with the respective

mm in A, B, F, G. In C, E the bar = 10 mm. In D the bar = 50 mm.

Table 3Marker profiles of Schwann cells in myelinated and non-myelinated nerves in the

skin and in peripheral nerves.

Marker Schwann cells of non-mye-

linated nerves

Schwann cells of myelin-

ated nerves

Skin Peripheral nerve Skin Peripheral nerve

NGFr + + + �NCAM + + + �L1CAM + + + �CD146 + + + �S100 + + + +

Vimentin + + + +

MBP � � + +

C.M. Reinisch, E. Tschachler / Journal of Dermatological Science 65 (2012) 162–169166

myelinating Schwann cells within peripheral nerves were consis-tently negative for these markers (Table 3).

3.3. The differential expression of neuronal marker proteins allows for

the topographic and quantitative analysis of different types of nerve

fibers within the subepidermal nerve plexus

Since the nerve plexus was most dense within chest skin weused this site to analyze the expression of neuronal markers innerve fibers.

As expected each NGFr-reactive nerve fiber (Fig. 3A) was alsopositive for the pan neuronal marker PGP9.5 (Fig. 3B) [18]. Thestaining for PGP9.5 showed a discontinuous, beaded pattern asdescribed previously [19]. Since the calculation of nerve fiberlength is based on pixel analysis of stained structures, thisdiscontinuous pattern resulted in a significantly reduced calculat-ed nerve fiber length as compared to the one obtained bycalculating the length of NGFr positive structures i.e. 57 � 23 mm/mm2 as compared to 115 � 25 mm/mm2 (Table 4). In contrast to Cfibers, A fibers contain neurofilament (NF) [20] and different amountsof myelin. Staining of dermal sheets for NF showed individual nervefibers with a mostly continuous and only occasionally partialdiscontinuous staining pattern (Figs. 3D and 4B). The length of NFpositive nerve fibers was 16 � 10 mm/mm2 (14.21% of all nervefibers) (Table 4). Only a small portion of NF expressing fibers showed

Fig. 3. Differential expression of neuronal and Schwann cell markers allow for discrimina

NGFr (A) is depicting Schwann cells. Both staining for PGP9.5 (B), a pan-neuronal mark

A-fibers in the subepidermal nerve plexus compared to the total amount of nerve fibers

(C, E, F). Bar = 100 mm. Shown are representative pictures of donor 10. In this donor the

130.81 mm/mm2 for NGFr.

detectable expression of MBP, which defines Ab-fibers (Fig. 4A) [21].Morphologically, these were mainly thicker fibers in deeper layers.The length of MBP positive fibers detected in the subepidermal nerveplexus was 0.2 � 0.18 mm/mm2 (0.18% of all nerve fibers) (Table 4).In two of our donors, we could not detect any MBP positive fibers inseveral sheets covering several cm2 of skin analyzed.

Peptidergic nerve fibers were detected by staining for CGRP(Fig. 4E). The staining pattern of these fibers was discontinuous andsimilar to the staining pattern of PGP9.5. Calculated peptidergicnerve fiber length based on pixel analysis was 4 � 3 mm/mm2

(7.84% of all nerve fibers) (Table 4).

tion of different nerve fiber qualities of the subepidermal nerve plexus. Staining for

er, and NF (D), a marker for A-fibers, are demonstrating axons. There are only few

. All axons are ensheathed by Schwann cells. The respective overlays are shown in

nerve fiber length found is 67.22 mm/mm2 for PGP9.5, 11.08 mm/mm2 for NF and

Table 4Analysis of different types of nerve fibers of the subepidermal nerve plexus of chest skin in 10 donors.

Age Sex NGFr-positive

fibers/mm2

PGP9.5-positive

fibers/mm2

NF-positive

fibers/mm2

MBP-positive

fibers/mm2

CGRP-positive

fibers/mm2

Donor 1 63 Male 95.87 57.24 25.38 0.19 6.31

Donor 2 75 Female 102.33 62.03 17.09 0.06 5.67

Donor 3 42 Female 150.95 47.91 19.73 0.58 0.68

Donor 4 45 Female 123.82 72.48 24.66 0.32 11.33

Donor 5 75 Male 155.34 95.16 17.59 0.13 2.47

Donor 6 64 Female 84.03 23.34 12.89 0.02 0.42

Donor 7 58 Male 85.64 33.14 6.07 0.09 2.82

Donor 8 60 Male 101.26 40.57 20.33 None detectable 5.52

Donor 9 60 Female 120.61 67.68 8.76 0.24 4.11

Donor 10 82 Female 130.81 67.22 11.08 None detectable 5.12

Mean (�standard error) 115.07 (�25.42) 56.68 (�23.44) 16.36 (�9.78) 0.20 (�0.18) 4.45 (�3.18)

% of all nerve fibers 100% 49.25%

100%a

14.21% 0.18% 3.86%

7.84%a

a Correction factor of 2.03 used to compensate for the staining pattern.

C.M. Reinisch, E. Tschachler / Journal of Dermatological Science 65 (2012) 162–169 167

4. Discussion

The aim of this study was to analyze the extension of thedifferent types of nerve fibers within skin at different body regionsand to compare the marker profile of terminal Schwann cells withthose of the peripheral nerves. Our finding that nuclei of individualSchwann cells are readily identifiable on terminal nerve fibers after

Fig. 4. Detection of myelinated nerve fibers and peptidergic nerve fibers in the subepi

Demonstration of Ab-fibers by staining for MBP (A). All A-fibers are depicted by NF-stain

plexus is illustrated by NGFr staining, which is shown in blue in the overlay (C). Bar = 10

0.19 mm/mm2 for MBP, 25.38 mm/mm2 for NF and 95.87 mm/mm2 for NGFr. Peptidergic

for PGP9.5 (D) is shown in (F). Bar = 50 mm. Shown are representative pictures of donor 10

and 11.08 mm/mm2 for NF.

immunostaining for various Schwann cell markers [10] allowed usfor the first time to quantify these cells in the skin.

Here we provide the first comprehensive quantitative data onthe constituents of the peripheral nervous system in the skin. Ourfinding that terminal Schwann cell numbers may be as high as453 cells/mm2 establishes them as a substantial cell populationwithin the dermis. For comparison, mast cells are present at

dermal nerve plexus.

ing (B). Ad-fibers are NF positive yet MBP negative. The entire subepidermal nerve

0 mm. Shown are representative pictures of donor 1. The nerve fiber length found is

nerve fibers are demonstrated in (E) by staining for CGRP. The overlay with staining

. The nerve fiber length found is 67.22 mm/mm2 for PGP9.5, 5.12 mm/mm2 for CGRP

C.M. Reinisch, E. Tschachler / Journal of Dermatological Science 65 (2012) 162–169168

69.5 � 29.9 cells/mm2 in human skin [22]. Therefore it is conceivablethat terminal Schwann cells, apart from their function in maintainingthe integrity of peripheral nerve fibers, also play a role in skinhomeostasis through interactions with their surrounding cellularsymbionts. Since little is known about such interactions, our findingon the abundance of this cell type opens a new area for research onskin biology. The density of Schwann cells varied considerably indifferent body regions with highest numbers within trunk skin andless than half of these numbers in skin of the distal extremities. Thesame was true for the nerve fiber density. Schwann cells ensheathnerve fibers and there is no evidence that under physiologicconditions Schwann cells exist in tissues without being directlyassociated with nerve fibers. Therefore the measured length of NGFrpositive structures within dermal sheets is equivalent to total nervefiber length. One possible explanation for these regional differencescould be that while the surface areas of different dermatomes vary,the Schwann cell numbers and total amount of nerve fibers perdermatome remain constant. Support for this assumption comes fromthe fact that the trunk has approximately the same surface as bothlower extremities together and comprises 12 dermatomes (C4, TH2-Th12) whereas the lower extremities comprise only 7 dermatomes(L1-L5, S1-S2) [23]. The relative rarefaction of nerve fibers andSchwann cell numbers on the peripheral extremities might be onereason why toxic and metabolic nerve damage manifests first at thesesites [24]. The lack of any association between nerve fiber length andage in our samples might indicate that in individuals withoutneurological disease aging does not influence nerve fiber length in thesubepidermal nerve plexus of examined body sites. Likewise, in astudy determining epidermal nerve fiber length in the thigh and distalpart of the leg, no significant effect of age in age groups from 20 to 79years was seen [25]. Similarly, the data of Lauria et al. [26] suggestedlittle age-related change in intra-epidermal nerve fiber density atsites in the leg. As well, epidermal innervation of abdominal skin didnot change with age [27].

To estimate a possible bias due to measuring the nerve fiberlength on a two-dimensional projection of the reconstituted stacks,we also performed measurements of three-dimensional LSMstacks. As to be expected, three-dimensional measurementrevealed a higher nerve fiber length but the difference was notstatistically significant. Bearing in mind that the size of both the X-axis and the Y-axis of the acquired stack exceed the Z-axisapproximately 5 fold, the overall relatively low impact of the Z-axisseems comprehensible and in our opinion the measurement bias istherefore negligible.

In peripheral nerves, Schwann cells and axons are formingdensely packed nerve fiber bundles and are well separated fromthe surrounding tissue by the perineurium and epineurium. Bycontrast, in the subepidermal nerve plexus individual Schwanncells ensheathing nerve fibers are directly embedded in theextracellular matrix of the surrounding skin. We made use ofantibodies directed at diverse epitopes known to depict Schwanncells. Besides NGFr, these mainly were adhesion molecules likeNCAM, L1CAM and CD146 [28]. Additionally, S100 [29] and theintermediate filament protein Vimentin were used. To specificallylabel myelin forming Schwann cells, MBP was used. Whencomparing Schwann cell populations within peripheral nervesand in the skin with regard to marker expression we found thatnon-myelinating Schwann cells at both locations expressed readilydetectable levels of NGFr, NCAM, L1CAM, and CD146. By contrastthese markers although present on Schwann cells of themyelinating type in skin were virtually absent from myelinatingSchwann cells within peripheral nerves. This difference mightreflect their distinct integration in the surrounding dermal matrixand be suggestive also for functional differences. This hypothesisgains some support by recent reports, demonstrating that L1CAMhas opposing effects in the central and peripheral nervous system.

In the CNS, overexpression in Schwann cells accelerates functionalrecovery by enhanced myelination after spinal cord injury [30],whereas in the peripheral nervous system it negatively regulatesSchwann cell proliferation after nerve damage [31]. Whethersimilar differences also exist between myelinated fibers in largeperipheral nerves and nerve terminals remains to be determined.

With regard to the different types of nerve fibers in the skin, ithas been shown in the past that within the dermis of hairy skin,most fibers are unmyelinated and only a few myelinated fibers arepresent [16]. Using colocalization of different nerve fiber markers,we were able to differentiate and for the first time quantify thediverse fiber types.

We used the pan-axonal marker [18] PGP9.5 to depict theindividual neurons. In contrast to the staining for Schwann cellmarkers which decorated the nerve fibers continuously, thestaining pattern for PGP9.5 had a beaded appearance [19]. Sincethe algorithm of the software for measuring nerve lengthsubtracted the non-stained intervals from the total nerve length,the length of PGP9.5 positive fibers was calculated at approxi-mately 50% of the length obtained by NGFr staining. We thereforepropose a correction factor of 2 in case such a measurement ofPGP9.5 stained nerve fibers is used to describe the total nerve fiberlength.

NF is expressed by A-fibers but not by C-fiber neurons [20]. Asfast conducting nerve fibers, A-fibers are also myelinated.However, myelination depends on axon caliber and it is usuallyfound only in axons thicker >1.6 mm [32]. It has been suggested inthe past that Ad-fibers, which in contrast to Ab fibers are thinnerthan 1.6 mm, lose their myelin sheath before entering the dermis[21]. Based on expression of NF by both Ad- and Ab-fibers and thepresence of myelin only on Ab-fibers, we found that 14.21% of allnerve fibers in the subepidermal nerve plexus are Ad-fibers and0.18% Ab-fibers.

Consequently, the remaining NF negative fibers of thesubepidermal nerve plexus represent C-fibers. They comprisedaround 86%. CGRP positive fibers also referred to as peptidergic C-fibers have been implicated in neurogenic inflammation of theskin. Our results show that they represent only a minority of allPGP9.5 positive C-fibers comprising roughly 8%.

In conclusion, our study provides for the first time a detailedtopography of the subepidermal nerve plexus, which now canserve as a reference for further studies of the skin nervous systemin health and disease. Our demonstration that high numbers ofSchwann cells are present within the superficial dermis shouldprompt considerations and investigations as to their contributionto the homeostasis of healthy skin as well as to their possibleparticipation in inflammatory and neoplastic skin disorders.

Acknowledgment

This work was funded by the Medical University of Vienna.

References

[1] Nolano M, Simone DA, Wendelschafer-Crabb G, Johnson T, Hazen E, KennedyWR. Topical capsaicin in humans: parallel loss of epidermal nerve fibers andpain sensation. Pain 1999;81:135–45.

[2] Teles RM, Krutzik SR, Ochoa MT, Oliveira RB, Sarno EN, Modlin RL. Interleukin-4 regulates the expression of CD209 and subsequent uptake of Mycobacteriumleprae by Schwann cells in human leprosy. Infect Immun 2010;78:4634–43.

[3] Worrell JT, Cockerell CJ. Histopathology of peripheral nerves in cutaneousherpesvirus infection. Am J Dermatopathol 1997;19:133–7.

[4] Quattrini C, Jeziorska M, Malik RA. Small fiber neuropathy in diabetes: clinicalconsequence and assessment. Int J Low Extrem Wounds 2004;3:16–21.

[5] Reinisch CM, Traxler H, Piringer S, Tangl S, Nader A, Tschachler E. Rarefactionof the peripheral nerve network in diabetic patients is associated with apronounced reduction of terminal Schwann cells. Diabetes Care2008;31:1219–21.

C.M. Reinisch, E. Tschachler / Journal of Dermatological Science 65 (2012) 162–169 169

[6] Umapathi T, Tan WL, Loke SC, Soon PC, Tavintharan S, Chan YH. Intraepidermalnerve fiber density as a marker of early diabetic neuropathy. Muscle Nerve2007;35:591–8.

[7] Riethmacher D, Sonnenberg-Riethmacher E, Brinkmann V, Yamaai T, Lewin GR,Birchmeier C. Severe neuropathies in mice with targeted mutations in theErbB3 receptor. Nature 1997;389:725–30.

[8] Chen ZL, Yu WM, Strickland S. Peripheral regeneration. Annu Rev Neurosci2007;30:209–33.

[9] Nave KA, Trapp BD. Axon-glial signaling and the glial support of axon function.Annu Rev Neurosci 2008;31:535–61.

[10] Tschachler E, Reinisch CM, Mayer C, Paiha K, Lassmann H, Weninger W. Sheetpreparations expose the dermal nerve plexus of human skin and render thedermal nerve end organ accessible to extensive analysis. J Invest Dermatol2004;122:177–82.

[11] Zhu Y, Ghosh P, Charnay P, Burns DK, Parada LF. Neurofibromas in NF1:Schwann cell origin and role of tumor environment. Science 2002;296:920–2.

[12] Julius D, Basbaum AI. Molecular mechanisms of nociception. Nature2001;413:203–10.

[13] Lumpkin EA, Caterina MJ. Mechanisms of sensory transduction in the skin.Nature 2007;445:858–65.

[14] Loken LS, Wessberg J, Morrison I, McGlone F, Olausson H. Coding of pleasanttouch by unmyelinated afferents in humans. Nat Neurosci 2009;12:547–8.

[15] Schmelz M, Schmidt R, Bickel A, Handwerker HO, Torebjork HE. Specific C-receptors for itch in human skin. J Neurosci 1997;17:8003–8.

[16] Lauria G, Lombardi R, Camozzi F, Devigili G. Skin biopsy for the diagnosis ofperipheral neuropathy. Histopathology 2009;54:273–85.

[17] McCarthy BG, Hsieh ST, Stocks A, Hauer P, Macko C, Cornblath DR, et al.Cutaneous innervation in sensory neuropathies: evaluation by skin biopsy.Neurology 1995;45:1848–55.

[18] Wilson PO, Barber PC, Hamid QA, Power BF, Dhillon AP, Rode J, et al. Theimmunolocalization of protein gene product 9.5 using rabbit polyclonal andmouse monoclonal antibodies. Br J Exp Pathol 1988;69:91–104.

[19] Ochs S, Pourmand R, Jersild Jr RA, Friedman RN. The origin and nature ofbeading: a reversible transformation of the shape of nerve fibers. Prog Neu-robiol 1997;52:391–426.

[20] Lawson SN, Waddell PJ. Soma neurofilament immunoreactivity is related tocell size and fibre conduction velocity in rat primary sensory neurons. J Physiol1991;435:41–63.

[21] Provitera V, Nolano M, Pagano A, Caporaso G, Stancanelli A, Santoro L.Myelinated nerve endings in human skin. Muscle Nerve 2007;35:767–75.

[22] Akimoto S, Ishikawa O, Igarashi Y, Kurosawa M, Miyachi Y. Dermal mast cellsin scleroderma: their skin density, tryptase/chymase phenotypes and degran-ulation. Br J Dermatol 1998;138:399–406.

[23] Lee MW, McPhee RW, Stringer MD. An evidence-based approach to humandermatomes. Clin Anat 2008;21:363–73.

[24] Mold JW, Vesely SK, Keyl BA, Schenk JB, Roberts M. The prevalence, predictors,and consequences of peripheral sensory neuropathy in older patients. J AmBoard Fam Pract 2004;17:309–18.

[25] McArthur JC, Stocks EA, Hauer P, Cornblath DR, Griffin JW. Epidermal nervefiber density: normative reference range and diagnostic efficiency. ArchNeurol 1998;55:1513–20.

[26] Lauria G, Holland N, Hauer P, Cornblath DR, Griffin JW, McArthur JC. Epidermalinnervation: changes with aging, topographic location, and in sensory neu-ropathy. J Neurol Sci 1999;164:172–8.

[27] Besne I, Descombes C, Breton L. Effect of age and anatomical site on density ofsensory innervation in human epidermis. Arch Dermatol 2002;138:1445–50.

[28] Shih IM, Nesbit M, Herlyn M, Kurman RJ. A new Mel-CAM (CD146)-specificmonoclonal antibody, MN-4, on paraffin-embedded tissue. Mod Pathol1998;11:1098–106.

[29] Stefansson K, Wollmann RL, Moore BW. Distribution of S-100 protein outsidethe central nervous system. Brain Res 1982;234:309–17.

[30] Lavdas AA, Chen J, Papastefanaki F, Chen S, Schachner M, Matsas R, et al.Schwann cells engineered to express the cell adhesion molecule L1 acceleratemyelination and motor recovery after spinal cord injury. Exp Neurol2010;221:206–16.

[31] Guseva D, Angelov DN, Irintchev A, Schachner M. Ablation of adhesion mole-cule L1 in mice favours Schwann cell proliferation and functional recoveryafter peripheral nerve injury. Brain 2009;132:2180–95.

[32] Voyvodic JT. Target size regulates calibre and myelination of sympatheticaxons. Nature 1989;342:430–3.