Archives of Dermatological Research

, Volume 298, Issue 1, pp 31–37

Increased nerve growth factor and its receptors in atopic dermatitis: an immunohistochemical study

Authors

  • Ying-Chun Dou
    • Experimental Dermatology Unit, Department of NeuroscienceKarolinska Institute
  • Lena Hagströmer
    • Department of DermatologyKarolinska University Hospital
  • Lennart Emtestam
    • Department of DermatologyKarolinska University Hospital
    • Experimental Dermatology Unit, Department of NeuroscienceKarolinska Institute
Original Paper

DOI: 10.1007/s00403-006-0657-1

Cite this article as:
Dou, Y., Hagströmer, L., Emtestam, L. et al. Arch Dermatol Res (2006) 298: 31. doi:10.1007/s00403-006-0657-1

Abstract

Evidence suggests that neurotrophins may regulate certain immune functions and inflammation. In the present study, the localization and distribution of nerve growth factor (NGF) and its receptors were explored using immunohistochemical methods, with the aim of detecting the cause of the neurohyperplasia in early lesions of atopic dermatitis (AD). In AD involved skin, strong NGF-immunoreactive (IR) cells were observed in the epidermis. In some cases, a huge number of infiltrating cells with stronger NGF immunoreactivity was seen mainly in the dermal papillae. Some trkA immunoreactivity was observed in the outer membrane of cells in the basal and spinal layers of the epidermis. In the papillary dermis, a larger number of cells demonstrated strong trkA immunoreactivity. The p75 NGFr-IR nerve fibre profiles were increased (900 per mm2; p<0.001) compared to normal [the involved skin also differed from the uninvolved skin (p<0.05)] in the dermal papillae. These nerve fibres were larger, coarser and branched, some of them terminated at p75 NGFr-IR basal cells, and also revealed a stronger fluorescence staining than the controls or the uninvolved skin. In normal healthy volunteers and AD uninvolved skin, the NGF immunoreactivity was weak in the basal layer of epidermis. Only a few trkA positive cells were seen in the basal layer of the epidermis and upper dermis. The IR epidermal basal cells revealed a striking patchy arrangement with strong p75 NGFr immunostaining in the peripheral part of the cells, and short and thick NGFr-IR nerve fibre profiles appeared as smooth endings scattered in the dermis including the cutaneous accessory organs. Using NGF and p75 NGFr double staining, both immunoreactivities showed a weak staining in the epidermis and dermis in normal and uninvolved skin. In the involved dermis of AD, the intensity of p75 NGFr-IR nerves was stronger in areas where there were also increased numbers of NGF-IR cells. These findings indicate that NGF and its receptors may contribute to the neurohyperplasia of AD.

Keywords

Atopic dermatitisSkinNerve growth factor (NGF)p75 NGFrtrkAImmunoreactivityNerve fibre

Abbreviations

AD

Atopic dermatitis

NGF

Nerve growth factor

NGFr

Nerve growth factor receptor

IR

Immunoreactive

RRX

Rhodamine red-X

FITC

Fluorescein-isothiocyanate

Introduction

Atopic dermatitis (AD) is a chronic, relapsing, inflammatory skin disease that is characterized by highly pruritic, eczematous skin lesions. Increasing evidence suggests that neurotrophins may be involved in the pathogenesis of AD and may regulate the development of AD [28]. Nerve growth factor (NGF) belongs to a family of neurotrophic proteins termed the neurotrophin family. The classic neurotrophin family also includes brain-derived neurotrophic factor (BDNF), neurotrophin-3 (NT-3), NT-4 and NT-5 [26]. The neurotrophins (including NGF) exerts their effect by binding to two classes of transmembrane receptors, a low-affinity receptor of 75 kd (p75 NGFr) [14] and a high-affinity tyrosine kinase receptor of 140 kd (trkA, trkB, trkC) [15]. The high-affinity binding site requires expression of the trkA proto-oncogene [3]. The presence of p75 increases the affinity of trkA for NGF and enhances the sensitivity of trkA-mediated neuronal response to NGF [4, 13]. Nonetheless, without p75 NGFr, the trk receptors are still capable of mediating many important functions.

Neurotrophins are synthesized and function not only in the nervous tissue, but also in some non-neuronal tissues. Neurotrophins play a crucial role as a neurotrophic molecule in the cutaneous nerve growth and functional maintenance. It is essential for the survival of certain nerves, especially the sensory ones. Its effects include normal neurite outgrowth, neuronal differentiation, survival of large numbers of neurons and support the rescue of injured and diseased nerves [24, 25]. It has been reported that the density of some nerves is dependent upon the local concentration of NGF [16]. NGF and its receptors are of great importance in the normal development of the peripheral nervous system in the early stage, including cutaneous nerves [1]. During cutaneous development, NGF is expressed at the highest in the epidermis. In particular, NGF synthesis begins in the skin with the arrival of the first axons to the epidermis (target-field), and the level of NGF mRNA increases throughout the period of skin innervation [1]. The definite evidence for NGF’s neurotrophic activity in the skin was provided by studies in transgenic mice [7, 11]. NGF transcripts have been identified in normal human keratinocytes in primary and secondary cultures, and expression of NGF mRNA was strongly down-regulated by corticosteroids [8, 23].

In order to understand NGF and its receptors’ (trkA and p75) possible roles in the pathogenesis of AD, the aim of the present study was to investigate the distribution and localization of NGF and its receptors using immunohistochemistry to visualize NGF-immunoreactive (NGF-IR) nerve fibres and immunoreactivity for NGF receptors in the skin of AD patients (involved and uninvolved) compared to the skin of healthy volunteers.

Materials and methods

Patients and samples

A total of eight AD patients with a mean age of 43 years (range 30–64) were randomly selected for this study. The AD patients were recruited from the outpatient ward of the Department of Dermatology, Karolinska University Hospital, Huddinge. Exclusion criteria were individuals below 18 years, those in the pregnancy or lactation stage, present history of any other disease other than atopy, acute infections, systemic corticosteroids for the previous 3 months, potent/semi-potent topical corticosteroids, or UV therapy/sunbeds during the previous month or systemic pharmacotherapy for 7 days prior to the study. The median age at onset of AD was 8.6 years (range 6 months–48 years). The duration of AD ranged from 10–57 years. Five of the patients had not been treated locally or systemically for the disease for the last 1 month prior to the examination, 2 patients were using a mild topical steroid since a couple of weeks and 2 patients were using a moisturizer. Five out of eight of the patients had a past or present history of respiratory atopy, but none required regular pharmacotherapy because of this. None of the patients had an allergy currently. The diagnosis was established on the basis of the morphological appearance and distribution of the skin lesions, the clinical course, and the family history of atopic disease. Thus, all the patients included fulfilled the criteria of Hanifin and Rajka for AD [5]. Since we were interested in the earliest event possible, only patients presenting, at the time of study, with relatively milder forms of AD were included, i.e. thin, slightly red and scaly lesions. Lesions presenting signs of lichenfication, nodulus and/or secondary infection did not enter the study. As controls, one biopsy specimen was taken from eight healthy volunteers. The mean age of the volunteers was 40 years (range 27–62). None had suffered from eczema or had a history of atopy (past or present). In order to minimize the risk for differences due to body location, we aimed at performing the skin biopsies from patients and control from the upper arm, in three patients the biopsies were taken from the back and in one patient from the thigh. In the controls, all biopsies were taken from the upper arm. The study protocol was approved by the local Committee of Ethics for Research in Humans and informed consent was obtained from all participants.

Single staining

The primary antibodies were mouse anti-p75 NGFr monoclonal antibody (1:25), rabbit anti-NGFβ polyclonal antibody (1:400) or rabbit anti-trkA polyclonal antibody (1:100; all from Chemicon International, Temecula, CA, USA), diluted in 0.01 M phosphate-buffered saline (PBS) containing 0.3% Triton X-100, incubated overnight at 4°C in a humid atmosphere and then incubated for 30 min at 37°C with affinity-purified donkey anti-rabbit IgG conjugated with rhodamine red-X (RRX, 1:160) or donkey anti-mouse IgG conjugated with fluorescein-isothiocyanate (FITC, 1:160; both Jackson ImmunoResearch Co., West Grove, USA) diluted in the same buffer as the primary antibodies. All rinses before and after the incubations were performed with the above buffer. The sections were viewed and photographed in a Nikon Microphot-FXA fluorescence microscope.

NGF and p75 NGFr double-staining

The steps for the immunofluorescence double-staining method were the same as those for single-staining, except as follows. The primary antibodies were a mixture of rabbit anti-NGFβ polyclonal antibody (1:400) and mouse anti-p75 NGFr monoclonal antibody (1:25; both Chemicon International, Temecula, CA, USA), and the secondary antibodies were RRX-conjugated affinity-purified donkey anti-rabbit IgG and FITC-conjugated affinity-purified donkey anti-mouse IgG (both 1:160; Jackson ImmunoResearch Co., West Grove, USA). The sections were viewed and photographed in a Nikon Microphot-FXA fluorescence microscope equipped with a special double filter.

Immunohistochemical control studies

To test for any possible non-specific binding of the primary antiserum to Fc receptors in the tissue, normal rabbit serum (1:100) was used instead of the primary antibodies. To control for any possible non-specific reactions of the secondary antisera in each independent experiment, PBS was used on certain sections instead of the primary antibodies. To test for cross-reaction of the secondary antiserum in the double-staining experiments, sections were incubated with the rabbit antibody followed by fluorophore-labelled donkey anti-mouse IgG, and other sections were incubated with mouse antibody followed by fluorophore-labelled donkey anti-rabbit IgG.

Quantification

For quantitative estimation, two fields of vision from each of two sections were investigated in each tissue biopsy. The number of fields and sections was established using a standard, stratified, random selection principle. In the dermis, under the 20× objective and using a special microscope frame, the number of p75 NGFr positive nerve fibre profiles per mm2 of projected skin area was estimated using design-based stereology. Since the lower-boundary of the dermis is not always clearly established, we only estimated one field height (0.35 mm) starting from the epidermal–dermal junction and facing downwards. In the epidermis and dermis, NGF-IR cells, p75 NGFr-IR cells and trkA-IR cells, respectively, were semiquantitatively estimated using a 5-graded scale (0=no; 1+=few; 2+=medium number; 3+=many; 4+=high number of). All sections were blind-coded.

Mann–Whitney U-test and Wilcoxon signed pairs test were used to show the difference. p<0.05 was considered to be statistically significant.

Results

NGF immunoreactivity

In normal and uninvolved skin, weak NGF-IR cells were seen in the epidermal basal layer, but there were a few stronger IR epidermal cells in the basal and spinal layers. In the middle and lower dermis, NGF-IR nerve plexuses could be seen to innervate the eccrine ducts, within and parallel to arrector pili muscle fibres (Fig. 1a). In AD involved skin, strong NGF immunoreactive cells were observed in the basal and spinal layers of the epidermis, and they were much more numerous than in controls and uninvolved AD (Fig. 1b; cf. Table 1). In some cases a huge number of infiltrating cells with stronger NGF immunoreactivity could be seen in the dermis (Table 2). These cells were small and rounded, some were spindle-like and some were dendritic in appearance. They were gathered together or occurred singly and were mainly observed in the upper part of the dermis.
https://static-content.springer.com/image/art%3A10.1007%2Fs00403-006-0657-1/MediaObjects/403_2006_657_Fig1_HTML.jpg
Fig. 1

Colour photomicrographs demonstrating NGF-like immunoreactivity. a From a normal skin sample. b Visualizes a section from atopic dermatitis involved skin. Strong NGF-IR cells are observed in the epidermis (magnification ×200)

Table 1

NGF, trkA and p75 NGFr immunoreactive cells in epidermis

 

NGF

trkA

p75 NGFr

 

IN

UN

HV

IN

UN

HV

IN

UN

HV

B1

3+

0+

1+

3+

0+

0+

2+

2+

3+

B2

2+

1+

0+

3+

1+

0+

1+

3+

4+

B3

3+

0+

1+

2+

0+

1+

2+

4+

3+

B4

2+

1+

0+

1+

1+

1+

1+

1+

2+

B5

2+

0+

1+

2+

0+

0+

1+

2+

4+

B6

2+

1+

1+

2+

0+

0+

1+

4+

3+

B7

3+

1+

0+

2+

0+

1+

2+

2+

4+

B8

3+

0+

0+

3+

1+

0+

0+

3+

3+

Semiquantitative estimation of NGF, trkA and p75 NGFr immunoreactive cells per mm2 of projected skin area of AD involved (IN), uninvolved (UN) and healthy volunteers (HV) biopsies (B), respectively

NGF Wilcoxon signed pairs test showed a significant difference between the IN and the HV groups (p<0.05) and between the IN and UN groups (p<0.05)

TrkA Wilcoxon signed pairs test showed a significant difference between the IN and the HV groups (p<0.05) and between the IN and UN groups (p<0.05)

NGFr Wilcoxon signed pairs test showed a significant difference between the IN and the HV groups (p<0.05) and between the IN and the UN groups (p<0.05)

Table 2

NGF and trkA immunoreactive cells in dermis

 

NGF

TrkA

 

IN

UN

HV

IN

UN

HV

B1

3+

1+

0+

3+

0+

1+

B2

4+

1+

1+

2+

1+

1+

B3

2+

0+

1+

3+

1+

0+

B4

3+

1+

0+

1+

0+

0+

B5

3+

0+

0+

4+

1+

0+

B6

1+

1+

0+

3+

0+

1+

B7

3+

0+

1+

3+

1+

1+

B8

3+

0+

0+

4+

1+

0+

Semiquantitative estimation of NGF and trkA immunoreactive cells per mm2 of projected skin area of AD involved (IN), uninvolved (UN) and healthy volunteers (HV) biopsies (B), respectively

NGF Wilcoxon signed pairs test showed a significant difference between the IN and the HV groups (p<0.05) and between the IN and UN groups (p<0.05)

TrkA Wilcoxon signed pairs test showed a significant difference between the IN and the HV groups (p<0.05) and between the IN and UN groups (p<0.05).

NGF receptors (trkA and p75) immunoreactivity

Staining with trkA demonstrated that trkA-IR nerve fibre profiles were never found in the epidermis in normal skin as well as in AD uninvolved skin, but a few trkA immunostained cells were seen in the basal layer of the epidermis and in the upper dermis (Fig. 2a). In the middle and lower dermis, stronger trkA-IR nerve plexuses could be seen to innervate the eccrine ducts, within and parallel to arrector pili muscle fibres. In AD involved skin, some trkA immunoreactivity was observed in the outer membrane of cells in the basal and spinal layers of the epidermis, and they were much more numerous than in controls and uninvolved AD (Table 1). In the upper dermis, especially in the papillary dermis, a larger number of cells demonstrated strong trkA immunoreactivity (Fig. 2b; cf. Table 2). In the middle and lower dermis, with regard to the intensity of trkA-IR nerves, no differences were observed between the AD involved, uninvolved and normal control skin.
https://static-content.springer.com/image/art%3A10.1007%2Fs00403-006-0657-1/MediaObjects/403_2006_657_Fig2_HTML.jpg
Fig. 2

Colour immunofluorescence micrographs of trkA-like immunoreactivity. a Shows a normal skin sample. b Demonstrates a section from atopic dermatitis involved skin. There are a larger number of trkA-IR cells in the papillary dermis (magnification ×200)

Staining with p75 NGFr showed that the p75 NGFr-IR cells revealed a striking patchy arrangement with strong immunostaining in the peripheral part of the cells, whereas the central cytoplasm and nucleus were negative (Fig. 3a), in the epidermal basal layer of normal and AD uninvolved skin. In these areas, higher numbers of p75 NGFr-IR nerve fibres were associated with higher numbers of p75 NGFr-IR basal cells. Some short and thick p75 NGFr-IR nerve fibre profiles, which appeared as smooth nerve endings, were found scattered in the dermis as well as around the cutaneous accessory organs. The median number of nerve fibre profiles per mm2 was 427 and 433, respectively, for normal control skin and uninvolved skin. A few p75 NGFr-IR free nerve endings could also be seen penetrating up into the epidermis. In the middle and lower dermis, p75 NGFr-IR nerve plexuses were seen to innervate the terminal part of small blood vessels, sweat glands, and within and parallel to the arrector pili smooth muscle fibres. Dense nerve networks were detected around the external root sheath of the lower part of the hair follicles. In AD involved skin, the intensity of p75 NGFr-IR cells tended to be decreased in the epidermal basal layer (Table 1). p75 NGFr-IR nerve fibres were never found in the epidermis. However, the fine dermal p75 NGFr-IR nerve fibre profiles were numerically increased (900 per mm2; p<0.001) as compared to normal control skin (the involved skin also differed from the uninvolved skin (p<0.05)), especially in the papillary dermis (Fig. 5). The nerve fibres were larger, coarser and branched, some of them terminated at p75 NGFr-IR basal cells, and also revealed a stronger fluorescence staining than the controls or the uninvolved skin (Fig. 3b). In the middle and lower dermis, the intensity of the p75 NGFr-IR nerve fibres around the external root sheath of hair follicles, sweat glands and blood vessels was the same as in uninvolved and control skin.
https://static-content.springer.com/image/art%3A10.1007%2Fs00403-006-0657-1/MediaObjects/403_2006_657_Fig3_HTML.jpg
Fig. 3

Colour photomicrographs showing p75 NGFr-like immunoreactivity. a Demonstrates nerves and p75 NGFr-IR cells in normal control skin. b Demonstrates dermal p75 NGFr-IR nerve fibres in involved skin of atopic dermatitis (magnification ×200)

NGF and p75 NGFr immunoreactivity

NGF and p75 NGFr double-staining showed a weak NGF immunoreactivity in the epidermal basal layer of normal and AD uninvolved skin. In the dermis, a few weak NGF-IR cells and a few short and thick p75 NGF-IR nerve fibres were also seen. Nerve bundles with an NGF immunoreactivity were mainly seen in the middle and lower dermis (Fig. 4a). The p75 NGFr-IR nerve hyperplasia appeared in the same areas where there were also increased levels of NGF-IR cells. In AD involved skin, strong NGF-IR epidermal cells and a large number of infiltrating cells mainly in the papillary dermis were seen (Fig. 4b). In the upper dermis, as was found with single staining, p75 NGFr-IR nerves were observed, and the intensity of these nerves was stronger in areas where there were also increased levels of NGF-IR cells. In the middle and lower dermis, with regard to the p75 NGFr-IR nerve hyperplasia and NGF immunoreactivity, no difference was observed between involved, uninvolved and normal skin, respectively (Fig. 5).
https://static-content.springer.com/image/art%3A10.1007%2Fs00403-006-0657-1/MediaObjects/403_2006_657_Fig4_HTML.jpg
Fig. 4

Colour photomicrographs of immunohistochemical double-staining (red NGF, green p75 NGFr). a From control skin. b From an atopic dermatitis involved skin biopsy. Note that the p75 NGFr-IR nerve fibres exist with many NGF-IR cells (magnification ×200)

https://static-content.springer.com/image/art%3A10.1007%2Fs00403-006-0657-1/MediaObjects/403_2006_657_Fig5_HTML.gif
Fig. 5

Number of dermal p75 NGFr-IR nerve fibre profiles per mm2 of projected skin area of normal human skin, atopic dermatitis uninvolved and involved skin (p<0.05 and p<0.001). Mann–Whitney U-test and Wilcoxon signed pairs test were used

Control staining

No positive staining was found in any of the control slides.

Discussion

In the present study, we have observed higher levels of NGF in the keratinocytes of the epidermal basal and spinal layers, and more in dermal infiltrating inflammatory cells in the early atopic dermatitis lesion, compared to the controls. It has been reported that NGF is synthesized and released by proliferating normal human keratinocytes [29, 33]. Normal human keratinocytes in culture express low- and high-affinity NGFr, and NGF significantly stimulates the proliferation of normal human keratinocytes in culture in a dose-dependent manner [23, 32]. The keratinocyte-derived NGF was secreted in a biologically active form as assessed by neuritis induction by sensory neurons obtained from chick embryo dorsal root ganglia [22]. Exponentially growing keratinocytes, but not confluent keratinocytes nor stratified keratinocyte cultures, are the pool of epidermal cells that synthesize NGF [8]. Consistent with the above findings, greater amounts of NGF are secreted by proliferating, pre-confluent keratinocytes than more differentiated, stratified cells [23]. This indicates that the basal proliferative cell compartment is the source of NGF in the epidermis. It has also been reported that inflammatory cells, such as mast cells [17], lymphocytes [2, 21, 27] and eosinophils [12, 26] can release NGF. In our study, in the AD dermis, there were many infiltrating inflammatory cells with different shapes, indicating that the above-mentioned cell types may be among the cells containing NGF-like immunoreactivity. However, the mechanism behind the relationship between increased NGF and such cellular proliferation in AD needs further clarification. Increased plasma levels of NGF and substance P (SP) have been observed in AD, and these NGF and SP plasma levels in AD significantly correlate with disease severity [28]. This suggested that neurogenic factors systemically modulate the allergic response in AD, probably through interactions with cells of the immune-inflammatory system. Maybe NGF-mediated nerve fibre sprouting have a functional impact on the skin immune cells involved in the regulation of the sensory nervous system. Thus, development of an allergic reaction in the skin influences the local nervous system [6]. Increasing evidence has demonstrated that neurotrophins and its receptors play an important role in both immune responses and neurogenic inflammation [10, 20].

The present study demonstrated that trkA-IR cells were observed in the epidermis and dermis, and that p75 NGFr-IR nerve fibre profiles were found increased in the dermis, especially in the dermal papillae of the involved as compared to uninvolved (p<0.05) and control (p<0.001) skin. A previous study has identified that NGF receptor trkA immunostaining could be observed in the basal keratinocytes, in the Meissner’s and Pacinian corpuscles, and around small arteries in glabrous digital skin [31]. p75 NGFr immunoreactivity has been localized in cutaneous nerves at all levels, as well as in certain basal keratinocytes [9], and in the Meissner’s and Pacinian corpuscles of the human digital skin [30]. In normal human skin, at the electron microscopic level, p75 NGFr immunoreactivity was seen mainly in Schwann cells and perineurium cell membranes [18], and the number of p75 NGFr-IR dermal nerve fibres are increased in prurigo nodularis [19]. In the presence of trkA, p75 NGFr can participate in the formation of high-affinity binding sites, resulting in enhanced NGF responsiveness [13]. Most of the biological activities elicited by NGF are mediated by ligand-dependent activation of tyrosine kinase A activity. The enhanced NGFr-IR nerve fibres demonstrated in the present study might be one of the causes of the neurohyperplasia in AD, but exactly why an increased NGFr immunoreactivity is observed in AD involved skin is still unknown. Further studies are needed to clarify the relationship between the neurohyperplasia and the disease.

In conclusion, our main findings are: (1) trkA high-affinity NGFr immmunoreactivity is increased in epidermis and upper dermis in AD involved skin; (2) strong NGF-IR cells are observed in the epidermis. In some cases, a huge number of infiltrating cells with stronger NGF immmunoreactivity are seen mainly in the dermal papillae; (3) p75 low-affinity NGFr-IR nerve fibres are more numerous and more intensely labelled in involved skin. However, the investigative technique employed did not allow us to elucidate whether the increased immunoreactivity in the lesional skin was primary or secondary, although the aim was to select as early lesions as possible. All these findings, taken together, may simply reflect common pathological features of AD. It is, of course, difficult from any data known today to reach a conclusion concerning the pathogenetic relevance of the above-studied molecules using immunohistochemistry alone. The increased neurotrophic action might be the cause of the neurohyperplasia of the diseases. Further studies are needed to investigate this hypothesis.

Acknowledgments

This work was supported by the Cancer and Allergy Foundation (Cancer- och Allergifonden). We thank Professor Tomas Hökfelt of the Department of Neuroscience, Karolinska Institute, for general support. We also thank Ms Marianne Ekman for her excellent technical assistance, and Ms Margareta Krook-Brandt of the Department of Learning, Informatics, Management and Ethics/The Medical Statistics Service Group, Karolinska Institute, for her statistical evaluation.

Copyright information

© Springer-Verlag 2006