Abstract Close contacts between mast cells (MC) and nerve fibers have previously been demonstrated in nor- mal and inflamed skin by light and electron micro-.
Arch Dermatol Res (1997) 289 : 292–302
© Springer-Verlag 1997
O R I G I N A L PA P E R
Vladimir A. Botchkarev · Stefan Eichmüller · Eva M. J. Peters · Peter Pietsch · Olle Johansson · Marcus Maurer · Ralf Paus
A simple immunofluorescence technique for simultaneous visualization of mast cells and nerve fibers reveals selectivity and hair cycle – dependent changes in mast cell – nerve fiber contacts in murine skin Received: 14 June 1996
Abstract Close contacts between mast cells (MC) and nerve fibers have previously been demonstrated in normal and inflamed skin by light and electron microscopy. A key step for any study in MC-nerve interactions in situ is to simultaneously visualize both communication partners, preferably with the option of double labelling the nerve fibers. For this purpose, we developed the following triple-staining technique. After paraformaldehyde-picric acid perfusion fixation, cryostat sections of back skin from C57BL/6 mice were incubated with a primary rat monoclonal antibody to substance P (SP), followed by incubation with a secondary goat-anti-rat TRITC-conjugated IgG. A rabbit antiserum to CGRP was then applied, followed by a secondary goat-anti-rabbit FITC-conjugated IgG. MCs were visualized by incubation with AMCA-labelled avidin, or (for a more convenient quantification of close MC-nerve fiber contacts) with a mixture of TRITCand FITC-labelled avidins. Using this simple, novel covisualization method, we were able to show that MCnerve associations in mouse skin are, contrary to previous suggestions, highly selective for nerve fiber types, and that these interactions are regulated in a hair cycle-dependent manner: in telogen and early anagen skin, MCs preferentially contacted CGRP-immunoreactive (IR) or SP/CGRP-IR double-labelled nerve fibers. Compared with telogen values, there was a significant increase in the number of close contacts between MCs and tyrosine hydroxylase-IR fibers during
V. A. Botchkarev · S. Eichmüller · E. M. J. Peters · M. Maurer · R. Paus (Y) Department of Dermatology, Charité, Humboldt-Universität, D-10117 Berlin, Germany P. Pietsch Department of Cardiology, Charité, Humboldt-Universität, D-10117 Berlin, Germany O. Johansson Experimental Dermatology Unit, Department of Neuroscience, Karolinska Institute, S-17177 Stockholm, Sweden
late anagen, and between MCs and peptide histidinemethionine-IR and choline acetyl transferase-IR fibers during catagen. Key words Mast cell · Nerve fibers · Avidin · Neuropeptides · Skin · Hair cycle Abbreviations AMCA 7-amino-4-methylcoumarin-3acetic acid · CGRP calcitonin gene related peptide · ChAT choline acetyl transferase · FGF2 fibroblast growth factor 2 · FITC fluorescein isothiocyanate · IgG immunoglobulin · IR immunoreative · MC mast cell · NF150 neurofilament 150 · N-CAM neuronal cell adhesion molecule · NGF nerve growth factor · PBS phosphate-buffered saline · PGP9.5 protein gene product 9.5 · PHM peptide histidine methionine · S-100 protein S-100 · SCF stem cell factor · SP substance P · TBS Tris buffer · TH tyrosine hydroxylase · TNFα tumor necrosis factor α · TRITC tetramethylrodamine isothiocyanate · VIP vasoactive intestinal polypeptide
Introduction Close contacts between mast cells (MCs) and nerve fibers have previously been demonstrated in normal and inflamed skin by light and electron microscopy [32, 44, 58, 77, 91]. The functional significance of these contacts is still unclear, but bidirectional communication between MCs and nerves is possible [4, 11, 20, 38, 52, 56, 92]. For example, neuropeptides are able to induce MC degranulation and cytokine expression [3, 25, 66], and MCs can release nerve growth factor (NGF) [47], induce the axon reflex [42], and produce neuropeptide-degrading proteases [19, 86]. A key step for any experimental study in MC-nerve interactions in situ is to achieve good and simultaneous visualization of both communication partners. A few methods for the simultaneous demonstration of nerve fibers and MCs have been described, which are based on immunoenzymatic detection of neuronal markers/neuropeptides
293
followed by immunoenzymatic detection of MC proteases or routine histochemical staining of MC glycosaminoglycans (e.g. toluidine blue or alcian blue) [34, 58, 79, 80]. However, immunoenzymatic detection of neuronal markers and neuropeptides (especially in mouse and rat skin) is notoriously plagued by background problems and, therefore, strongly depends on successful background blocking. Immunofluorescence methods have, therefore, generally been preferred by for the detection of neuronal markers [see reference 49]. Since, different types of cutaneous nerve fibers contain different combinations of neurotransmitters and/or neuropeptides [see reference 82], methods for studying MCnerve interactions ideally should allow MC visualization to be combined with the identification of two or more nerve-related antigens. However, the combination of double immunofluorescence visualization of two neuronal markers/ neuropeptides with the simultaneous demonstration of tissue MCs has proven to be difficult. For example, attempts to combine immunofluorescence methods of neuronal marker detection with MC biogenic amine determination have been successful only in pig skin and rat mesentery [2, 21]. Our attempts to reproduce these techniques in mouse or rat skin have been unsuccessful, probably owing to the necessity of distinct tissue fixation requirements for different immunocytochemical procedures. The capacity of connective tissue-type MCs to bind labelled avidin, based on the ionic interaction between the negatively charged glycosaminoglycans of MC granules and the positively charged avidin molecule was shown long ago [9, 17, 26, 83] and has been successfully employed for MC visualization in human and rodent skin [40, 51]. It has also been shown that the number of MCs detected by avidin staining is highly comparable to that detected by alcian blue-safranin or toluidine blue, independent of the different types of tissue fixation used (see, for example, references 40 and 51). FITC-labelled avidin has been used to distinguish MCs from other types of fluorescent cells in different immunofluorescence procedures [29, 88]. On this background, the first aim of this study was the application of fluoresceinated avidin for the simultaneous visualization of MCs and nerve fibers with one or two nerve-associated antigens in murine skin. MCs may be involved in the regulation of murine hair growth, and skin MC number and functions in mice fluc-
Table 1 Description of the primary antibodies used against skin nerve fiber-associated antigens
tuate in a hair cycle-dependent manner [13, 55, 60, 62]. Also, the neuropeptide substance P (SP), a potent mast cell secretagogue [25, 66], stimulates hair growth in mice, as does the pharmacological induction of neuropeptide release from skin nerve endings by capsaicin [63]. Moreover, hair cycle-dependent fluctuations of the sensory and autonomic nerve fiber density have been recently demonstrated [14, 23, 71]. Thus, MC-nerve interactions might play an important regulatory role in the control of murine hair follicle growth and cycling. Therefore, the second aim of this study was to apply the newly developed method of MC-nerve covisualization to the analysis of potential hair cycle-associated changes in MC-nerve interactions during the murine hair cycle.
Methods Syngenic C57BL/6 mice (female, 6–8 weeks old; Charles River, Hannover, Germany) with all follicles in the resting stage (telogen) were depilated on the back skin to induce a highly synchronized hair growth cycle as described previously [61–63]. At different stages of the hair cycle (telogen = day 0, anagen II = day 3, anagen VI = day 12, catagen = day 19 post depilation) mice were anesthesized by intraperitoneal injection of ketanest (0.15 ml per mouse) and perfusion-fixed with a mixture of 4% paraformaldehyde and 14% saturated picric acid injected intracardially. Skin harvesting, embedding and freezing was performed as described elsewhere [64]. For double visualization of nerve fibers and MCs, cryostat sections (15 µm) were incubated with the primary antibodies to different neuronal markers and neuropeptides (listed in Table 1) overnight at room temperature in a humid chamber, followed by incubation for 30 min at 37° C with TRITC-conjugated F(ab)2 fragments of goat antirabbit IgG (Jackson ImmunoResearch, West Grove, PA, USA) at a dilution of 1:200. All antisera contained 0.3% Triton X-100 and 2% normal goat serum. Sections then were incubated with the FITC-labelled avidin (Vector, Burlingame, CA, USA) diluted 1:1000 in 0.1 M Tris buffer (TBS), pH 7.4, for 1 h at room temperature. After every incubation, sections were rinsed in TBS three times, 15 min each time. For the triple staining of MCs and nerve fibers with two different markers (SP and CGRP), sections were incubated with monoclonal rat antibodies to SP overnight at room temperature, followed by incubation for 1 h at 37° C with TRITC-conjugated F(ab)2 fragments of goat antirat IgG, followed by incubation with the rabbit primary antisera to CGRP (overnight, room temperature), and finally with goat antirabbit FITC-conjugated IgG (Jackson ImmunoResearch, West Grove, PA, USA) diluted 1:25, for 1 h at 37° C. For the visualization of MCs, sections were then incubated with AMCA-labelled avidin (1:1000) for 1 h at room temperature. In order to quantify close contacts between MCs and nerve fibers, we also used the (more convenient) incubation of sec-
Antiserum against
Abbreviation Dilution Species
Manufacturer
Protein gene product 9.5 Neurofilament 150 Protein S-100 Substance P polyclonal Substance P monoclonal Calcitonin gene-related peptide Peptide histidine methionine (= VIP precursor) Tyrosine hydroxylase Choline acetyltransferase
PGP 9.5 NF 150 S-100 SP SP CGRP PHM
1:1000 1:500 1:1000 1:500 1:100 1:500 1:500
Rabbit Rabbit Rabbit Rabbit Rat Rabbit Rabbit
Paesel + Lorei, Frankfurt, FRG Chemicon, Temecula, CA Sigma, St. Louis, MO Amersham, Bucks, UK Serva, Heidelberg Chemicon, Temecula, CA Paesel + Lorei, Frankfurt, FRG
TH ChAT
1:500 1:100
Rabbit Rabbit
Biogenesis, Poole, UK Biogenesis, Poole, UK
294 tions with a mixture of the FITC- (1:1000) and TRITC-conjugated (1:3000) avidins for 1 h at room temperature instead of AMCAavidin. Sections were rinsed in TBS and mounted in immunomount medium (Shandon, Runcorn, UK). We used incubation with normal rabbit or rat sera (1:1000) instead of specific primary antisera, as well as incubation with primary antisera preabsorbed by 50–100 µg/ml of corresponding neuropeptide (SP, CGRP, PHM) as controls in the immunodetection of neuropeptides. Also, three different controls were performed to verify that the avidin staining was demarcating MCs. First, the avidin labelling was combined with histamine immunostaining [37]. Back skin samples of C57BL/6 female mice were fixed in 4% carbodiimide (1-ethyl-3,3-dimethylaminopropyl-carbodiimide; Sigma, St. Louis, MO), diluted in 0.1 M phosphate buffer (PBS, pH 7.4) for 2 h at 4° C, immersed in 10% sucrose (4° C, overnight), rinsed in PBS at 4° C, and sectioned at 6 µm in a cryostat. The sections were incubated with the primary rabbit antiserum to histamine (Milab AB) 1:1000 overnight at 4° C. After rinsing, the sections were incubated with a TRITC-labelled donkey antirabbit secondary antibody (Boehringer Mannheim, FRG) for 30 min at 37° C, diluted 1:200, and then incubated with FITC-avidin (Vector, Burlingame, CA, USA) 1:1000 for 1 h at room temperature. Second, the above histamine labelling was controlled by an additional histochemical staining with 0.1% toluidine blue (pH 0.5) [36], and third, sections were preincubated with 2 units heparinase (Sigma, St. Louis, MO) at 45° C for 1 h [75] to check for the capacity of MC heparin to bind fluoresceinated avidin. The sections were examined at × 400 magnification under a Zeiss fluorescence microscope, and the number of single nerve fiber profiles and MCs with or without close contacts with single nerve fibers were quantified (“close contact” defined as < 2 µm). The results are expressed per square millimeter of back skin section and calculated as number of MCs with close contacts with nerve fibers (1) relative to the total number of MCs, and (2) relative to the total number of nerve fibers. All analyses were done on at least ten microscopic fields of five different mice per hair cycle stage (i.e. a total of more than 200 microscopic fields were analyzed for every immunohistochemical reaction). For a more precise demonstration of close contacts between MCs and nerve fibers, selected sections were examined using laser confocal image systems (BioRad MRC 600, Karolinska Institute, Stockholm; Noran Instruments, Charite, Humboldt University, Berlin). The data were pooled and means and SEMs calculated. P-values were determined using independent Student’s t-test.
Results Avidin-positive cells are mast cells Multiple dermal and subcutaneous cells with bright fluorescence were visualized in C57BL/6 mouse skin after incubation with AMCA-, FITC- or TRITC-labelled avidin. The avidin-positive cells contained numerous granules, and some of these cells displayed extracellularly located avidin-positive granules in the direct vicinity of the cell. In double immunofluorescence experiments, there was a 100% overlap of histamine-positive cells and avidin-positive cells, i.e. all avidin-positive cells also showed histamine immunoreactivity and vice versa. The same held true for double staining of histamine and toluidine blue. There were no significant differences between the number of histamine-immunoreactive (IR) (78.9 ± 7.4 cells/mm2), toluidine blue-positive MCs (77.3 ± 8.8 cells/mm2) and avidin-positive MCs (79.0 ± 9.5 cells/mm2) in C57BL/6 mouse skin. Moreover, the avidin staining was completely negative after heparinase treatment of the sections. Thus,
based on the ability of heparin to bind fluoresceinated avidin, the avidin method reliably stained all differentiated murine skin MCs detectable by classical methods. Simultaneous labelling of neuronal markers and mast cells localize mast cell-nerve contacts In order to analyse spatial interactions of identified neuronal fibers and MCs, we applied a triple labelling method: nerve fibers were double-stained with two different neuronal markers (CGRP and SP), and MCs were labelled by AMCA-avidin or simultaneously with a mixture of FITCand TRITC-labelled avidin, giving a two-color visualization of MCs. This method proved to be very convenient for analyzing close contacts between MCs and nerve fibers with different neuropeptides and showed excellent reproducibility (Fig. 1). For the simultaneous visualization of MCs and nerve fibers with a single nerve fiber-associated marker, a simpler double labelling of MCs with FITC-avidin, and of nerve fibers with the specific primary antisera and TRITCor FITC-labelled secondary antibodies, was employed. Surprisingly many MCs in telogen mouse skin had microscopically visible close contacts with nerve fibers (defined as a distance of < 2 µm; Fig. 1A, B, C). MCs with such nerve contacts were located predominantly (1) in the papillary dermis, close to the dermo-epidermal basement membrane, (2) around the arrector pili muscle, (3) around the proximal part of telogen hair follicles, (4) adjacent to nerve bundles at the dermis/subcutis border and in the subcutis, distal to the panniculus carnosus, and (5) in the vicinity of subcutaneous blood vessels. Mast cells show a preference for contacts with SP-IR and/or CGRP-IR fibers in telogen skin Using this method, we determined quantitative and qualitative characteristics of MC-nerve associations in murine skin at different stages of the depilation-induced hair cycle. In unmanipulated C57BL/6 mouse skin with all hair follicles in telogen, the MC-nerve contacts apparently depended on the nature of the nerve fibers (Fig. 2). Evidently, the number of MCs showing contacts with nerve fibers strongly depended on the total number of such fibers present in the skin. In telogen skin, there was a clear predominance of PGP 9.5-IR and CGRP-IR nerve fibers over NF 150-IR, S-100IR, SP-IR, PHM-IR, TH-IR, and ChAT-IR nerve fibers (Fig. 2A). Approximately 17% of all MCs were found close to PGP 9.5-IR (pan-neuronal marker) nerve fibers, while only 5–7% of MCs had close contacts with nerve fibers expressing other structural nerve fiber markers (NF150, S-100; Fig. 2B). A comparative analysis of the number of MCs contacting nerve fibers that were immunoreactive for different neurochemical markers revealed that MCs were predominantly located near CGRP-IR fibers. Very few MCs were located near SP-IR, PHM-IR, TH-IR, or ChAT-IR nerve fibers (Fig. 2B).
295
Fig. 1A–H A, B, C One color double labelling of MC and SP-IR nerve fiber in C57BL/6 mouse skin (confocal laser scanning image). A An SP-IR nerve fiber (TRITC) shows a close contact with a TRITC-avidin-labeled MC. Another MC is also located in close proximity to an SP-IR nerve fiber, but without contact (total image from a 15-µm thick cryosection). B, C Two subsequent 1 µm step images at the 6 and 7 µm levels of the same section, demonstrating the same close contact. D, E Simultaneous two color triple labelling of nerve fibers with two neuronal markers (SP and CGRP) and MCs. D MCs and CGRP-IR nerve fibers (FITC). E MCs and SP-IR nerve fiber (TRITC). A long double-labelled CGRP- and SP-IR fiber (large arrows) is shown to demon-
strate close contact with an MC (small arrows). One SP-negative, but CGRP-IR, fiber (large arrowhead), contacting a MC (small arrowheads) can be seen. F, G, H Simultaneous three-color triple labelling on nerve fibers with two neuronal markers (CGRP and SP) and MCs. F Two AMCA-avidin-labelled MCs (large arrows). G Same section: numerous FITC-labelled CGRP-IR nerve fibers (arrowheads) are apparent; few of them are located in close proximity to MCs (small arrows indicate to the places of possible close contacts with MCs). H Same section: number of TRITC-labelled SP-IR nerve fibers (arrowheads) is substantially less than that of CGRP-IR fibers; all SP-IR nerve fibers also contain CGRP-IR, but do not show close contacts with MCs (A–H × 400)
296
A
B
C Fig. 2A–C Quantification of close contacts between MCs and nerve fibers expressing different neuronal markers and/or neuropeptides in unmanipulated back skin of C57BL/6 mice with all hair follicles in telogen. A Number of single nerve fibers, immunoreactive to different neuronal markers and neuropeptides, in telogen skin. Note the significant predominance of PGP 9.5-IR over NF150-IR and S-100-IR nerve fibers (**P < 0.01, ***P < 0.001, large asterisks) and CGRP-IR over SP-IR, PHM-IR, TH-IR, and ChAT-IR nerve fibers, (**P < 0.01, ***P < 0.001, small asterisks). B Contacts relative to the number of MCs. The percentage of MCs contacting single nerve fibers labelled with antibodies against different neuronal markers or neuropeptides in telogen skin is shown. The relative number of MCs showing contacts with PGP 9.5-IR or CGRP-IR nerve fibers is significantly higher than
the number showing contacts with the other labelled nerve fibers (**P < 0.01, large asterisks vs PGP 9.5-IR nerve fibers, small asterisks vs CGRP-IR nerve fibers). C Contacts relative to the number of single nerve fibers. The percentage of nerve fibers contacting MCs is shown, was calculated and tested statistically against values of PGP 9.5-IR fiber contacts. CGRP-IR and SP-IR nerve fibers show significantly more contacts with MCs than the average of all PGP 9.5-IR nerve fibers (*P < 0.05, **P < 0.01), whereas PHM-IR nerve fibers show significantly fewer contacts (*P < 0.05, large asterisks). The percentage of PHM-IR, TH-IR and ChAT-IR nerve fibers showing contacts with MCs is significantly lower than the percentage of CGRP-IR nerve fibers showing such contacts (*P < 0.05, **P < 0.01, small asterisks)
In addition, we calculated the number of contacts between MCs and nerve fibers with different neuronal markers or neuropeptides relative to the total number of specifically immunoreactive nerve fibers (Fig. 2C). Our results indicate that there were no significant differences in the number of nerve fibers visualized with the different structural markers (PGP 9.5, NF150, S-100) that displayed
close contacts with MCs. However, the numbers of SP-IR and/or CGRP-IR nerve fibers contacting MCs were significantly higher than the number of PHM-IR, TH-IR or ChAT-IR nerve fibers displaying close contacts with MCs.
297
A
B
C
D Fig. 3A–D Quantification of close contacts between MCs and neuropeptide-containing nerve fibers in C57BL/6 mouse back skin during various stages of the depilation-induced hair cycle. (telogen: unmanipulated skin, day 0; anagen II: 3 days after anagen induction by depilation; anagen VI and catagen: respectively, 12 and 19 days days after anagen induction). A Dynamics of the total number of MCs in the different hair cycle stages. A decrease in the number of MCs in early anagen (**P < 0.01, large asterisks), compared with telogen skin and an increase of the number in late anagen and catagen (**P < 0.01, small asterisks) compared with early anagen. B Percentage of MCs contacting different neuropeptide-containing nerve fibers in various hair cycle stages. A significant increase in the percentage of MCs contacting CGRP-IR, SP-IR, PHM-IR, and TH-IR nerve fibers (*P < 0.05, **P < 0.01, large aterisks) in early anagen, and contacting PHM-IR and ChAT-IR nerve fibers in catagen compared with telogen skin is shown. A decrease in the number of MCs contacting CGRP-IR, SP-IR nerve fibers in late anagen and catagen and contacting PHM-IR nerve fibers in late anagen (*P < 0.05, **P < 0.01, small asterisks), compared with early anagen values is also shwon. C Dynamics of the total number of different neuropeptide-containing nerve fibers in various
hair cycle stages. Significant increases (*P < 0.05, **P < 0.01, ***P < 0.001, large asterisks) in the total number of all types of neuropeptide-containing nerve fibers in early anagen, of CGRPIR, SP-IR, PHM-IR, and ChAT-IR nerve fibers in late anagen, and of SP-IR, PHM-IR, and ChAT-IR nerve fibers in catagen, compared with telogen values are shown. Decreases in the number of CGRP-IR, SP-IR and TH-IR nerve fibers during late anagen and catagen compared with early anagen values (*P < 0.05, **P < 0.01, small asterisks) are also shwon. D Percentage of different neuropeptide-containing nerve fibers contacting mast cells in various hair cycle stages. Significant decrease (*P < 0.05, **P < 0.01, large asterisks) in the percentage of the CGRP-IR, SP-IR nerve fibers displaying close contacts with mast cells in anagen and catagen compared with telogen skin, and increases in close MC contacts with of PHM-IR and ChAT-IR fibers in catagen skin (*P < 0.05, small asterisks) compared with anagen VI values are shown. Note the increase in the percentage of TH-IR nerve fibers contacting mast cells in anagen VI compared with early anagen, and the significant decrease in such contacts during catagen (*P < 0.05, **P < 0.01, small asterisks) compared with anagen VI values
298
Preferences of mast cell-nerve associations change during the hair cycle In a further step, we analyzed whether the number of close spatial contacts between MCs and nerve fibers containing different neuropeptides changes during the induced murine hair cycle. This was based on the fact that murine skin is substantially remodelled during the hair cycle [31, 60, 61] and that MC numbers and activity, as well as skin innervation change significantly during the murine hair cycle [13, 15, 23, 55, 62]. As previously reported with different histochemical staining techniques [13, 62], there was a significant decrease in avidin-positive MCs in C57BL/6 mouse skin during anagen II (3 days post depilation), while the number of MCs increased again in anagen VI (12 days post depilation) and catagen skin (19 days post depilation) compared with the early anagen values (Fig. 3A). Furthermore, the number of MCs contacting CGRPIR, SP-IR, PHM-IR, and TH-IR nerve fibers was significantly increased in early anagen skin (compared with telogen), while the number of MCs having close contacts with PHM-IR or ChAT-IR nerve fibers increased during spontaneous hair follicle regression (catagen, Fig. 3B). In contrast, there was a significant decrease in the number of MCs contacting SP-IR and CGRP-IR nerve fibers in late anagen and catagen. Furthermore, compared with telogen skin, the number of nerve fibers with various neuropeptide immunoreactivities changed substantially during the induced murine hair cycle (Fig. 3C). Compared with the telogen levels, the number of CGRP-IR nerve fibers was significantly increased in early and late anagen skin, and that of SP-IR, PHM-IR, or ChAT-IR nerve fibers increased in skin harvested 3, 12 and 19 days after anagen induction by depilation. There was a dramatic increase in the number of TH-IR nerve fibers in anagen II (day 3 post depilation), compared with telogen skin, but no significant differences between the number of TH-IR nerve fibers in telogen, anagen VI or catagen skin. However, the percentage of CGRP-IR and of SP-IR nerve fibers displaying close contacts with MCs significantly declined in anagen (days 3 and 12) and catagen (day 19) skin compared with telogen skin, while the percentage of PHM-IR, and ChAT-IR nerve fibers with close MC contacts increased in catagen skin compared with late anagen (Fig. 3D). Finally, the percentage of TH-IR nerve fibers contacting MCs was increased in late anagen (day 12), compared with early anagen (day 3), and significantly declined again during catagen (Fig. 3D).
Discussion We have demonstrated that combining the avidin method of MC visualization with the immunofluorescence detection of neuronal markers and neuropeptides offers a very simple and highly reproducible basic technique for the study of nerve-MC interactions in situ, which can be employed independently of the required fixation protocols. The de-
scribed method is perfectly suited to the use of multicolor confocal microscopy and to light microscopic investigations of MC-nerve contacts in situ. In addition, we have shown that MC-nerve contacts in murine skin are far from random, and obviously depend on the neurotransmitter and/ or neuropeptide profile of the nerve fibers, since MCs preferentially contact SP-positive or SP/CGRP-doublepositive sensory nerve fibers. Strikingly, MCs appear to change their relative preference for certain nerve fiber subtypes in a hair cycle-dependent manner. We found that, under physiological conditions, MC-nerve associations in murine skin were not at all constant, being subject to substantial reorganization, apparently dependent on the dramatic skin remodelling associated with synchronized hair follicle cycling [31, 60, 61]. The maximum frequency of close MC-sensory nerve associations can be observed in early anagen, and the minimum frequency in catagen. Finally, our study underscores previous evidence obtained in our laboratory that the number of skin MCs as well as the structural organization and the neurotransmitter/neuropeptide content of murine skin nerve fibers significantly fluctuate during the hair cycle [15, 23, 62, 68]. While the physiological functions and the control of these previously unappreciated hair cycle-dependent MCnerve interactions essentially remain obscure, the following considerations may help to interpret our findings and assist in designing meaningful follow-up studies to address these questions. In general, our results confirm previous reports demonstrating the presence of MC-nerve contacts in the skin [5, 32, 44, 58, 77, 91] and other organs [2, 34, 38, 41, 79, 80, 84, 87], and show that MCnerve contacts occur predominantly in those skin compartments known to be densely innervated [57, 82]. However, our results suggest that the frequency of MC-nerve fiber interactions in normal mammalian skin may actually be much higher than often assumed. Ultrastructural evidence indicates that, at sites of close contact with MCs, nerve fibers are deprived of a Schwann cell coating [41, 76]. Moreover, MCs that contact nerve fibers form lamellopodia, which are wrapped around and enclose the nerve fibers deeply within the cell [41]. Heparin released from MC granules at the sites of contacts with nerve fibers can downregulate N-CAM expression on the nerve fiber surface [90]. MCs, located in close proximity to nerve fibers might regulate general neuronal functions such as growth, sprouting and regression, mediated via the production and release of different neurotrophins by MCs (e.g. NGF, FGF2, leukemia inhibitory factor), thus promoting neuronal outgrowth [16, 48, 53, 75]. In telogen mouse skin, approximately 15% of the MCs were located close to sensory CGRP-IR and/or SP-IR nerve fibers (Fig. 2B), while 30-45% of these nerve fiber profiles seen in 15-µm sections were in contact with MCs (Fig. 2C). Several biologically relevant levels of MCnerve interaction in skin may be possible. For example, SP released from cutaneous sensory fibers (e.g. by antidromic stimulation) can induce MC degranulation, upregulate TNFα gene expression and may activate growth-
299
modulatory functions of MCs [3, 4, 25, 46, 66]. MC proteases can degrade SP and CGRP released from sensory nerve fibers [19, 86], and histamine released from MCs can modulate neurotransmission in peptidergic nerve fibers via neuronal histamine type 3 receptors [22, 73]. Heparin, released from MC granules, can interact with NK-1 receptors and thus modulate the action of SP [45]. These accepted bidirectional MC-nerve interactions [11, 38, 52, 56, 92] may provide a functional explanation for the surprisingly high frequency of close MC-nerve contacts and may help to understand why MCs appear to preferentially contact SP-IR and/or CGRP-IR nerve fibers. We have shown that the percentage of MCs contacting adrenergic (e.g. TH-IR) nerve fibers significantly increases in early anagen (Fig. 3B), even though the total number of skin avidin-positive or histochemically detectable MCs [13] substantially declines during this stage of the induced hair cycle (Fig.3A). Together with the fact that the percentage of TH-IR nerve fibers with MC contacts fluctuates during the hair cycle, this suggests that associations between MCs and adrenergic nerve fibers (like MC-sensory nerve contacts) in murine skin are actively regulated, apparently depending on changing hair cycle-associated functional requirements. Anatomical and functional interactions between MCs and adrenergic nerve fibers are well recognized [8, 12, 24, 30], even though there are contradictory reports of both inhibitory and enhancing effects of noradrenaline and adrenaline on mediator release from MCs [92]. Additional reasons for MC-adrenergic nerve fiber contacts may be that neuropeptide Y, which is usually co-localized with noradrenaline in some adrenergic neurons [70, 82], causes release of beta-hexosaminidase, but not prostaglandin D2 from rat peritoneal fluid MCs [6], and that histamine and serotonin released from MCs, in turn, can induce exocytotic noradrenaline release from adrenergic nerve terminals [27]. With respect to the catagen-associated increase in MC contacts with cholinergic (ChAT-IR) nerve fibers (Fig. 2C, Fig. 3B, D), which often contain VIP and PHM (27 amino acid sequence in the VIP precursor) [35, 50, 85], it is interesting to note that both acetylcholine and VIP can induce mediator release from MCs [10, 43]. Conversely, MC amines can trigger acetylcholine release from cholinergic nerve endings [27]. Furthermore, MC have been shown to contain actylcholinesterase [59] and to degrade acetylcholine released from cholinergic nerve fibers, while MC proteases are able to degrade VIP released from nerve endings [19]. Together with the hair cycle-dependent changes in MC-adrenergic nerve fiber contacts we have observed, these possible levels of bidirectional MC-cholinergic nerve fiber interactions further support the concept that autonomic nerves are indeed involved in the regulation of hair follicle morphogenesis [7] and growth [68], possibly with MC serving as a “central switchboard for tissue remodelling” [55]. Intriguingly, most changes in MC associations with specific nerve fibers occur during the telogen-anagen transformation of the hair follicle, and during spontaneous
hair follicle regression (catagen) (Figs. 2, 3). It will, therefore, be important to clarify the role of MC-nerve contacts in (a) hair cycle control, (b) hair cycle-associated remodelling of all skin compartments, and (c) in the rearrangement of skin innervation during hair follicle growth and regression [55, 60, 63]. Furthermore, the effects of MCnerve interactions on the physiological functions of both communication partners need to be elucidated. It is already known that growth factors which stimulate MC differentiation can induce neuronal outgrowth and vice versa. For example, stem cell factor (SCF) has neurotrophin-like activity, and is able to induce neuronal outgrowth and survival [18, 33]. Finally, it needs to be addressed how the apparent hair cycle dependence of MC-nerve interactions in murine skin is brought about. In this context, growth factors released from hair follicle cells and known to alter both MC and nerve functions are the most likely candidate signalling messengers between hair follicles, MCs and nerve fibers. SCF is one such candidate, since it alters both MC and nerve functions (see above), since SCF gene expression and protein content in murine skin significantly change during the murine hair cycle [67], and since it may actually be produced by follicular cells, e.g. by dermal papilla fibroblasts [74]. NGF is another candidate, as it can induce MC differentiation, maturation and survival [39, 54], and can stimulate MC secretory activity [65, 69]. Notably, murine hair follicle keratinocytes can produce NGF [1, 81], and the skin concentration of NGF is significantly increased in early anagen skin in mice, compared with telogen skin [89]. Thus, NGF may play a key role in inducing the development of sensory and autonomic “hyperinnervation” and MC degranulation in early anagen [compare references 1, 15, 62, 65, 68]. It remains to be carefully elucidated by appropriate functional studies to what extent the hair cycle-dependent changes in MC-nerve interactions uncovered in the present study are of functional significance or are epiphenomena not directly related to hair follicle remodelling. Specifically, mouse mutants in which defined MC-nerve interactions do not occur, such as mice that lack functional skin MCs (e.g. W/Wv or Sl/Sld mice [28]), or defined components of skin innervation (e.g. NGF-, neurotrophin3-, brain derived neurotrophic factor-knockouts [78]), now need to be compared with their respective wildtype littermates with respect to parameters such as hair follicle development and/or cycling or keratinocyte proliferation and differentiation. If subtle differences in these parameters should be detected, this would certainly support the concept that the MC-nerve interactions reported here are functionally important. Taken together, our results indicate that MC-nerve associations in mouse skin are much more selective, and more differentially regulated, than previously assumed, and suggest that MCs can probably change their intracutaneous location and contact with different types of nerve fibers during distinct hair cycle stages. In view of the increasing insight into the role of neural signals in hair follicle development and cycling on the one hand [7, 60, 63],
300
and into the functional importance of MCs in murine hair follicle cycling on the other [55], it deserves systematic explanation whether the hair cycle-dependent changes in MC-nerve contacts reported here are functionally important in the control of hair follicle cycling and innervation, and whether these contacts can be pharmacologically targeted for the therapeutic manipulation of hair growth and skin innervation. Acknowledgements The generous support and advice of Prof. B. M. Henz, Prof. W. Sterry and Dr. F. Noser, and the technical assistance of R. Pliet, G. Holmkvist and E.-K. Johansson are gratefully acknowledged. This study was supported in part by a grant from the Deutsche Forschungsgemeinschaft (DFG Pa 345/3-2 + 6-1) and Wella/Cosmital to R. P. and from the Medical Faculty of the Karolinska Institute to O. J., and by a visiting postdoctoral fellowship from the Swedish Cancer and Allergy Foundation to V.A.B.
References 1. Albers KM, Wright DE, Davis BM (1994) Overexpression of nerve growth factor in epidermis of transgenic mice causes the hypertrophy of the peripheral nervous system: J Neurosci 14: 1422–1433 2. Alving K, Sundstrom C, Matran R, Panula P, Hokfelt T, Lunderg JM (1991) Association between histamine-containing mast cells and sensory nerves in the skin and airways of control and capsaicin-treated pigs. Cell Tissue Res 264:529–538 3. Ansel JC, Brown JR, Payan DG, Brown MA (1993) Substance P selectively activates TNF-α gene expression in murine mast cells. J Immunol 150:4478–4485 4. Ansel JC, Kaynard AH, Armstrong CA, Olerud J, Bunnett N, Payan D (1996) Skin-nervous system interactions. J Invest Dermatol 106:198–204 5. Arizono N, Matsuda S, Hattori T, Kojima I, Maeda T, Galli SJ (1990) Anatomical variation in mast cell nerve associations in the rat small intestine, heart, lung, and skin. Similarities of distances between neural processes and mast cells, eosinophils, or plasma cells in the jejunal lamina propria. Lab Invest 62:626– 634 6. Arzubiaga C, Morrow J, Jackson Roberts L II, Biaggioni I (1991) Neuropeptide Y, a putative cotransmitter in noradrenergic neurones, induces mast cell degranulation but not prostaglandin D2 release. J Allergy Clin Immunol 87:88–93. 7. Asada-Kubota M (1995) Inhibition of hair growth by subcutaneous injection of a sympathetic neurotoxin, 6-hydroxydopamine in neonatal mice. Anat Embryol 191:407–414 8. Aubineau P, Dimitriadou V, Callabert J, Seylaz J (1987) Close contacts between nerves and serotonin-containing mast cells in cerebral vessels. Cephalagia 6:16–18 9. Bergstresser PR, Tigelaar RE, Tharp MD (1984) Conjugated avidin identifies cutaneous rodent and human mast cells. J Invest Dermatol 83:214–218 10. Blandina P, Fantozzi R, Mannaioni PF, Masini E (1980) Characteristics of histamine release evoked by acetylcholine in isolated rat mast cells. J Physiol 301:281–293 11. Blennerhassett MG (1994) Nerve and mast cell interaction: cell conflict or information exchange? Progr Clin Biol Res 390: 225–241 12. Blennerhassett MG, Tomioka M, Bienenstock J (1991) Formation of contacts between mast cells and sympathetic neurons in vitro. Cell Tissue Res 265:121–128 13. Botchkarev VA, Paus R, Czarnetzki BM, Kupriyanov VS, Gordon DS, Johansson O (1995) Hair cycle-dependent changes in mast cell histochemistry in murine skin. Arch Dermatol Res 287:683–686
14. Botchkarev VA, Eichmüller S, Johansson O, Henz BM, Paus R (1996) The innervation of murine back skin and hair follicles changes with the hair cycle (abstract). Arch Derm Res 288:310 15. Botchkarev VA, Eichmüller S, Johansson O, Paus R (1997) Hair cycle-dependent plasticity of skin and hair follicle innervation in normal murine skin. J Comp Neurol (in press) 16. Bothwell M (1995) Functional interaction of neurotrophins and neurotrophin receptors. Annu Rev Neurosci 18:223–253 17. Bussolati G, Gugliotta P(1983) Nonspecific staining of mast cells by avidin-biotin-peroxidase complexes (ABC). J Histochem Cytochem 31:1419–1421 18. Carnahan JF, Patel DR, Miller JA (1994) Stem cell factor is a neurotrophic factor for neural crest-derived chick sensory neurons. J Neurosci 14:1433–1440 19. Caughey GH, Leidig F, Viro NF, Nadel JA (1988) Substance P and vasoactive intestinal peptide degradation by mast cell tryptase and chymase. J Pharmacol Exp Ther 244:133–137. 20. Church MK, Lowman MA, Robinson C, Holgate ST, Benyon RC (1989) Interaction of neuropeptides with human mast cells. Int Arch Allergy Appl Immunol 88:70–78 21. Crivellato E, Damiani D, Mallardi F, Travan L (1991) Suggestive evidence for a microanatomical relationship between mast cells and nerve fibers containing substance P, calcitonin gene related peptide, vasoactive intestinal polypeptide, and somatostatin in the rat mesentery. Acta Anat 141:127–131. 22. Dimitriadou V, Rouleau A, Dam Trung Tuong, Newlands GJ, Miller HR, Luffau G, Schwartz JC, Garbarg M (1994) Functional relationship between mast cells and C-sensitive nerve fibers evidenced by histamine H3-receptor modulation in rat lung and spleen. Clin Science 87:151–163 23. Eichmüller S, Botchkarev VA, Johansson O, Paus R (1996) Hair cycle-dependent rearrangement of murine skin innervation (abstract). J Invest Dermatol 106:889 24. Felten SY, Felten DL, Belinger DL, Carlson SL, Ackerman KD, Madden KS, Olschowska JA, Livnat S (1988) Noradrenergic sympathetic innervation of lymphoid organs. Progr Allergy 43:14–36 25. Foreman JC (1987) Substance P and calcitonin gene-related peptide: effect on mast cells in human skin. Int Arch Allergy Appl Immunol 82:366–371 26. Fritz P, Muller J, Reiser H, Saal JG, Hadam M, Tuczek HV, Wegner J, Laschner W (1986) Avidin-peroxidase. A new mast cell staining method. Acta Histochem Suppl 32:235–239 27. Fuder H, Ries P, Schwarz P (1994) Histamine and serotonin released from the rat perfused heart by compound 48/80 or by allergen challenge influence noradrenaline or acetylcholine exocytotic release. Fundam Clin Pharm 8:477–490 28. Galli SJ, Geissler EN, Wershill B, Gordon JR, Tsai M, Hammel I (1993) Insights into mast cell development and function derived from analyses of mice carrying mutations at beige, W/c-kit, or Sl/SCF (c-kit ligand) loci. In: Kaliner M et al (ed) The mast cell in health and disease. NY. Marcel Dekker: 129–202 29. Giorno R, Choi KL, Claman HN (1987) Simultaneous in situ detection of IgE receptors and cytoplasmic granules in murine cutaneous mast cells. J Immunol Meth 99:163–166 30. Gordon DS, Sergeeva VE, Zelenova IG (1982) Neuromediators of lymphoid organs. Nauka, Leningrad (Russia):1–128 31. Hansen LS, Coggle JE, Wells J, Charles MW (1984) The influence of the hair cycle on the thickness of mouse skin. Anat Rec 210:569–573 32. Heine H, Forster FJ (1975) Histophysiology of mast cells in skin and other organs. Arch Dermatol Res 253:225–228 33. Hirata T, Morii E, Morimoto M, Kasugai T, Tsujimura T, Hirota S, Kanakura Y, Nomura S, Kitamura Y (1993) Stem cell factor induces outgrowth of c-kit-positive neurites and supports the survival of c-kit-positive neurons in dorsal root ganglia of mouse embryos. Development 119:49–56 34. Hukkanen M, Gronblad M, Rees R, Kottinen YT, Gibson SJ, Hietanen J, Polak JM, Brewerton DA (1991) Regional distribution of mast cells and peptide containing nerves in normal and adjuvant arthritic synovium. J Rheumatology 18:177–183
301 35. Itoh N, Obata K, Yanaihara N, Okamoto H (1983) Human preprovasoactive intestinal polypeptide contains a novel PHI-27like peptide, PHM-27. Nature 304:547–549 36. Johansson O, Virtanen M, Hilliges M, Yang Q (1992) Histamine immunohistochemistry: a new and highly sensitive method for studying cutaneous mast cells. Histochem J 24:283–287 37. Johansson O, Virtanen M, Hiliges M, Yang Q (1994) Histamine immunohistochemistry is superior to the conventional heparin-based routine staining methodology for investigations of human skin mast cells. Histochem J 26:424–430 38. Johnson D, Krenger W (1992) Interactions of mast cells with the nervous system – recent advances. Neurochem Res 17:939– 951 39. Kannan Y, Matsuda H, Ushio H, Kawamoto K, Shimada Y (1993) Murine granulocyte-macrophage and mast cell colony formation promoted by nerve growth factor. Int Arch Allergy Immunol 102:362–367 40. Kasper CS, Tharp MD (1987) Quantification of cutaneous mast cells using morphometric point counting and a conjugated avidin stain. J Am Acad Dermatol 16:326–331. 41. Keith IM, Jin J, Saban R (1995) Nerve mast cell interaction in normal guinea pig urinary bladder. J Comp Neurol 363:28–36 42. Kiernan JA: A pharmacological and histological investigation of the involvement of mast cells in cutaneous axon reflex vasodilatation (1975) Q J Exp Physiol 60:123–130 43. Kiernan JA (1990) Degranulation of mast cells in the trachea and bronchi of the rat following stimulation of the vagus nerve. Int Arch Allergy Appl Immunol 91:398–402 44. Kishimoto S (1984) The regeneration of substance P-containing nerve fibers in the process of burn wound healing in the guinea pig skin. J Invest Dermatol 83:219–223 45. Knaus GA, Knaus H-G, Saria A (1991) Allosteric interaction of heparin with NK-1 receptors in rat striatal membranes. Ann NY Acad Sci 632:422–425 46. Kowalski ML, Kaliner MA (1988) Neurogenic inflammation, vascular permeability, and mast cells. J Immunol 140:3905– 3911 47. Leon A, Buriani A, Dal Toso R, Fabris M, Romanello S, Aloe L, Levi-Montalcini R: Mast cell synthesize, store, and release nerve growth factor (1994) Proc Natl Acad Sci USA 91:3739– 3743 48. Levi-Montacini R, Aloe L, Alleva E (1990) A role for nerve growth factor in nervous, endocrine and immune system. Prog Neuro Endocrin Immunol 3:1–10 49. Ljungberg A, Johansson O (1993) Methodological aspects on immunohistochemistry in dermatology with special reference to neuronal markers. Histochem J 25:735–745 50. Maggi CA, Giachetti A, Dey RD, Sami I Said (1995) Neuropeptides as regulators of airway function: vasoactive intestinal peptide and the tachykinins. Physiol Rev 75:277–322 51. Markey AC, Churchill LJ, MacDonald DM (1989) Human cutaneous mast cells – a study of fixative and staining reactions in normal skin. Br J Dermatol 120:625–631 52. Marshall JS, Waserman S (1995) Mast cells and the nerves – potential interactions in the context of chronic disease. Clin Exp Allergy 25:102–110 53. Marshall J.S, Gauldie J, Nielsen L, Bienenstock J (1993) Leukemia inhibitory factor production by rat mast cells. Eur J Immunol 23:2116–2120 54. Matsuda H, Kannan Y, Ushio H, Kiso Y, Kanemoto T, Suzuki H, Kitamura Y (1991) Nerve growth factor induces development of connective tissue-type mast cells in vitro from murine bone marrow cells. J Exp Med 174:7–14 55. Maurer M, Paus R, Czarnetzki BM (1995) Mast cells as modulators of hair follicle cycling. Exp.Dermatol 4:269–277 56. McKay DM, Bienenstock J (1994) The interaction between mast cells and nerves in the gastrointestinal tract. Immunol Today 15:533–538 57. Montagna W (1974) Cutaneous innervation (1974) In: Montagna W, Parrakkal W (eds) The structure and function of skin. Academic Press. New York London, pp 157–171
58. Naukkarinen A, Harvima I, Paukkonen K, Aalto M-L, Horsmanheimo M (1993) Immunohistochemical analysis of sensory nerves and neuropeptides, and their contacts with mast cells in developing and mature psoriatic lesions. Arch Dermatol Res 285:341–346. 59. Nechushtan H, Soreg H, Kuperstein V, Tshori S, Razin E (1996) Murine and human mast cells express acetylcholinesterase. FEBS Lett 371:1–6 60. Paus R (1996) Control of the hair cycle and hair diseases as cycling disorders. Curr Opin Dermatol 3:248–258 61. Paus R, Stenn KS, Link RE (1990) Telogen skin contains an inhibitor of hair growth. Br J Dermatol 122:777–784 62. Paus R, Maurer M, Slominski A, Czarnetzki BM (1994a) Mast cell involvement in murine hair growth. Dev Biol 163:230–240 63. Paus R, Maurer M, Slominski A, Czarnetzki BM (1994b) Hair growth induction by substance P. Lab Invest 71:134–140. 64. Paus R, Hoffmann U, Czarnetzki BM, Eichmüller S (1994c) Distribution and changing density of gamma-delta T cells in murine skin during the induced hair cycle. Br J Dermatol 130:281–289 65. Paus R, Lüftl M, Czarnetzki BM (1994d) Nerve growth factor modulates keratinocyte proliferation in murine skin organ culture. Br J Dermatol 130:174–180. 66. Paus R, Heinzelmann T, Robicsek S, Czarnetzki BM, Maurer M (1995) Substance P stimulates murine epidermal keratinocyte proliferation and dermal mast cell degranulation in situ. Arch Dermatol Res 287:500–502 67. Paus R, Welker P, Jensen K, Handjiski B, Eichmüller S, Botchkarev VA, Maurer M, Scott GA (1996) Intraepithelial c-kit expression during murine hair follicle development and cycling: a role for stem cell factor in epithelial biology (abstract). J Invest Dermatol 106:834. 68. Paus R, Peters EMJ, Eichmüller S, Botchkarev VA (1997) Neural mechanisms of hair growth control. J Invest Dermatol 108 Suppl:(in press) 69. Pearce FL, Thompson HL (1986) Some characteristics of histamine secretion from rat peritoneal mast cells stimulated with nerve growth factor. J Physiol (Lond) 372:379–393 70. Pernow J, Lundberg JM, Kaijser L (1987) Vasoconstrictor effects in vivo and plasma disappearance rate of neuropeptide Y in man. Life Sci 40:47–54 71. Peters EMJ, Botchkarev VA, Eichmüller S, Paus R (1996) Hair cycle-dependence of the autonomic innervation in murine skin (abstract). J Invest Dermatol 106:890 72. Piotrowski W, Foreman JC (1985) On the actions of substance P, somatostatin, and vasoactive intestinal polypeptide on rat peritoneal mast cells and in human skin. Naunyn-Schmiedeberg’s Arch Pharmacol 331:364–368. 73. Rafferty P, Holgate ST (1989) Histamine and its antagonists in asthma.J Allergy Clin Immunol 84:144–151. 74. Randall V (1995) Beard dermal papilla cells secrete more stem cell factor in culture than non-balding scalp cells or dermal fibroblasts (abstract). In: First Tricontinental Meeting of Hair Research Societies, Brussels 75. Reed JA, Albino AP, McNutt NS (1995) Human cutaneous mast cells express basic fibroblast growth factor. Lab Invest 72:215–222 76. Schotzinger RJ, Landis SC (1990) Postnatal development of autonomic and sensory innervation of thoracic hairy skin in the rat. Cell Tissue Res 260:575–587. 77. Siguira H, Maeda T, Uehara M (1992) Mast cell invasion of peripheral nerve in skin lesions of atopic dermatitis. Acta Derm Venereol [Suppl] (Stockh) 176:74–76 78. Snider WD (1994) Function of the neurotrophins during nervous system development: what the knockouts are teaching us. Cell 77:627–638 79. Stead RH, Tomioka M, Quinonez G, Simon GT, Felten SY, Bienenstock J (1987) Intestinal mucosal mast cells in normal and nematode-infected rat intestines are in intimate contact with peptidergic nerves. Proc Natl Acad Sci USA 84:2975–2979
302 80. Stead RH, Dixon MF, Bramwell NH, Riddel RH, Bienenstock J (1989) Mast cells are closely apposed to nerves in the human gastrointestinal mucosa. Gastroenterology 97:575–585 81. Takami S, Getchell ML, Yamagishi M, Albers KM, Getchell TV (1995) Enhanced extrinsic innervation of nasal and oral chemosensory mucosa in keratin 14-NGF transgenic mice. Cell Tissue Res 282:481–491 82. Tausk F, Christian E, Johansson O, Milgram S (1993) Neurobiology of the skin. In: Fitzpatrick TB (ed) Dermatology in General Medicine. McGraw-Hill Inc, New York:396–403 83. Tharp MD, Seeling LL, Tigelaar RE, Bergstresser PR (1985) Conjugated avidin binds to mast cell granules. J Histochem Cytochem 33:27–32 84. Theoharides TS (1990) Mast cells: the immune gate to the brain. Life Sci 46:607–617 85. Vaalasti A, Tainio H, Rechardt L (1985) Vasoactive intestinal polypeptide (VIP)-like immunoreactivity in the nerves of human axillary sweat glands. J Invest Dermatol 85:246–248 86. Wallengren J, Wang ZY (1993) Interaction between tachykinins and CGRP in human skin. Acta Derm Venereol 73:259–261 87. Wang L, Stanisz AM, Wershill BK, Galli SJ, Perdue MH (1995) Substance P induces ion secretion in mouse small intestine through effects on enteric nerves and mast cells. Am J Physiol 269:G85–92
88. Weidner N, Horan RF, Austen KF (1992) Mast-cell phenotype in indolent forms of mastocytosis. Ultrastructural features, fluorescence detection of avidin binding, and immunofluorescent determination of chymase, tryptase, and carboxypeptidase. Am J Pathol 140:847–857 89. Welker P, Peters EMJ, Botchkarev VA, Pethö-Schramm A, Eichmüller S, Paus R (1996) Nerve growth factor and the murine hair cycle (abstract). J Invest Dermatol 106:910 90. Werz W, Schachner M (1988) Adhesion of neural cells to extracellular matrix constituents. Involvement of glycosaminoglycans and cell adhesion molecules. Brain Res 471:225–234 91. Wiesner-Menzel L, Schulz B, Vakilzadeh F, Czarnetzki BM (1981) Electron microscopical evidence for a direct contact between nerve fibers and mast cells. Acta Derm Venereol 61: 465–469 92. Williams RM, Bienenstock J, Stead RH (1995) Mast cells: the neuroimmune connection. In: Marone G (ed) Human basophils and mast cells: biological aspects. Karger, Basel, pp 208–235
