Immunohistochemical study of nerve fibres in basal cell carcinoma

ARTICLE

Auteur(s) : Ana PANUNCIO, Raúl VIGNALE, Guislena LOPEZ

Department of Pathology, Hospital de Clínicas Dr. Manuel Quintela, School of Medicine, Montevideo, Uruguay

Article accepted on 14/2/2003

It is well known that innervation has an influence in multiple physiological events: exocrine and endocrine secretion, in maturation, multiplication and cell differentiation [1, 2] on the psycho-neuro-immune axis [3] as well as in the so-called psycho-neuro-endocrine-cutaneous-system [4]. Little is known about the relationship between neurosubstances and tumours of the skin and other locations [5-9]. We approached this subject in BCC, the most common skin cancer [10, 11]. The great difficulty in preservation and demonstration of neurotransmitters in biopsies processed for diagnosis, and the inexistence of appropriate antibodies for such conditions of fixation, probably explains the scanty knowledge in this area.
Cramer [12] stated that the integrity of the peripheral nervous system plays an important role in carcinogenesis. However, in his report we did not find experimental support for this assessment. Winkelman [13] found nerve fibres closely related to BCC cells. However, according to this author there would be little functional relationship between the fibres and the neoplastic proliferation. Ormea et al. [14] reviewed this subject and found that the information available differed and was not conclusive. In his own work using silver impregnation techniques, he evidenced a larger number of nerve fibres in the stroma of BCC compared with healthy skin. Pawlowski [15] found more numerous nervous elements in BCC than in hyperkeratotic benign lesions. He proposed that nerve fibres and Schwann cells have an active part as a defense mechanism in experimental carcinogenesis and pre-invasive conditions of skin tumours.
The aim of this work was to study by immunohistochemical methods, the existence of nerve fibres in BCC, and to establish if there is any relationship between the presence of nerve fibres and the aggressiveness of the tumour. Nerve fibres were labelled with a general neurofilament protein marker, and the sympathetic innervation with anti-TH, the key enzyme of the adrenergic route [16].
We tested many other antibodies such as anti-VIP, Substance P, NOS, Gastrin, Calciton-gene-related-peptide, but we only selected the above mentioned, because they are the ones which developed an adequate immunoreactivity in biopsies or surgical specimens fixed in formalin and paraffin embedded.

Material and methods

Forty cases of trunk and face BCC from patients with ages ranging from 49 to 82 years old were collected from the files of different health care centres in Montevideo. Sections of formalin fixed and paraffin embedded samples were obtained. Immunolabelling with polyclonal antibody against TH was performed (Chemicon International USA) (1/128 in 0.05 M Tris buffer, pH7.4); monoclonal antibody to Neurofilament, Dako MO762, peptidic subunits 70, 160 and 200 kd (1/50). Both antibodies were incubated 72 hours at 20 °C and Shi retrieval antigen method was applied [17]. Endogenous tissue peroxidase activity was blocked with hydrogen peroxide 0.03%, and non specific reactions were avoided with a Protein Block Serum Free agent (Dako, USA) when the TH antibody was used. The detection system employed was Envision (Dako, USA, with DAB as chromogen). The slides were lightly counterstained with haemalun. Negative controls were performed, omitting either the primary or secondary antibody. The periglandular and vascular innervation and the thick nerve trunks not involved in the tumour were used as positive inner controls. We discarded the 20 cases without an adequate positive internal control. Tumour samples were classified in two groups following published histological criteria of malignancy [11, 18-21]. Group 1, non-aggressive: superficial, with monomorphic cells, in clusters with more than 2 cells, and nodular and microlobular pattern; Group 2, aggressive: with any histological pattern but with invasion of all reticular dermis. The presence of nerve fibres was evaluated in the following locations: 1) in the peri-tumoral connective tissue, 2) in the close surroundings of the epithelial proliferation (periepithelial fibres) 3) in the intra-tumoral connective tissue and 4) in the peri-tumoral inflammatory infiltrate. Some samples lacked intra-tumoral connective tissue or inflammatory infiltrate, the number of cases in this category was diminished in concordance.
A semiquantitative score was established by two independent researchers as follows: 0 = without nerve fibres, 1 = focal and scarce fibres, 2 = focal and moderate number of fibres, 3 = scarce fibres disseminated throughout the tumour, 4 = large number of disseminated fibres. Statistical analysis was made with the Mann Whitney U test, by comparing the different compartments in both groups. Significance was established at the level of p£ 0.05.

Results

According to the above mentioned criteria, we classified the samples in non aggressive, Group 1 (7 cases) and aggressive, Group 2 (13 cases). A solid histological pattern predominated in group 1, whereas in the second group we found mixed histological patterns, but all of them were infiltrating. Thick trunks in the middle of the proliferation were occasionally seen, but we interpreted them as being passively trapped by the proliferation, so we did not include them in our scores (Fig. 1). Tables I and II resume the results obtained.

Table I. Non aggressive group

Table II. Aggressive group

Group 1, Non aggressive (Table I)

All of the cases of this group presented TH and NF positive fibres in the peri-tumoral connective tissue, those labelled with NF antibodies being more numerous (Fig. 2). Most of the cases had nerve fibres close to the epithelial proliferation (Fig. 3). All samples that presented intratumoral connective tissue (Fig. 4) and inflammatory infiltrate near the proliferation had positive fibres. The highest scores in this last compartment were seen with NF immunolabelling.

Group 2, Aggressive (Table II)

Focal and scarce TH-positive fibres in peri-tumoral and intra-tumoral connective tissue were observed. Few cases presented fibres close to the epithelium. Those cases with inflammatory infiltrate had lower scores of nerve fibres than group 1.

Discussion

Few studies have been performed about innervation and tumour biology [5-9, 12-15]. Most of them were made with silver impregnation methods [12-15]. In our studies, the immunolabelling of neurofilament proteins certifies that fibrilar elements were of nervous nature. Moreover, an adrenergic key enzyme was detected by TH antibody. We found a statistically significant difference when we compared similar compartments in Group 1 and 2. Nerve fibres in the peritumoral connective tissue and in contact with carcinomatous proliferation were observed. The highest scores were reached with anti-NF antibodies (Table I and II). The lower positivity for TH may be explained by the difficulty in demonstrating its existence. However, it could also indicate that other neurotransmitters or neuropeptides were in those fibres labelled by NF antibodies. Thick trunks were seen, but they might have been trapped by the proliferation and we did not include them for statistical study.
Our findings are in agreement with those obtained using silver and histochemical techniques [12-15].
It was proposed that nerves and neurosubstances could have an organoid relationship with the tumour, as a barrier to prevent tumour growth [9, 14]. In scirrous tumours of the breast, nerve fibres appeared in intimate relationship with the stroma, blood vessels and parenchyma, and it was suggested in this model, a nervous organization as a comprehensive part of cancer growth [22].
We can hypothesize that in the early stages more differentiated tumoral cells release substances like growth factors, which stimulate nerve sprouts in the neighbouring connective tissue. In this sense, it is known that nerve growth factor (NGF) is normally produced by confluent keratinocytes in culture, and a trophic activity of NGF on sensory neurons in special conditions was proposed [23]. Thus, we propose that more mutations are added in the evolution of the tumour, the production of such factors could disappear and a reduction in nerve growth stimulation would occur. Moreover, the evolved tumour could secrete inhibitory substances, causing nerve fibre degeneration.
Recent evidence showed that gastrectomy with denervation promoted the development of intestinal metaplasia, displasia, and carcinoma in the remaining body of the stomach [24]. Other studies in human liver metastases of colorectal origin, showed that autonomic perivascular nerves are absent [6]. These results are in agreement with our findings and those made with silver techniques [14, 15] that the absence of innervation is correlated with a more aggressive course.
The influence of immune cells on nerve fibre development has been already reported. There are many interactions between both systems, either by cytokines or other paracrine peptides that can act upon nerve sprouting and proliferation [24].
However further studies must be conducted in order to elucidate the exact mechanism of interactions between nerve fibres and neoplastic cells. n

References

1. Burnstock G. Autonomic neurotransmitters and trophic factors. J Auton Nerv Syst 1983; 7: 213-7.

2. Whimster IW. Nerve supply as a stimulator of the growth of tissue including skin I. Human evidence. Clin Exp Dermatol 1978; 3: 221-40.

3. Felten DL, Felten SY, Carlson SL, Olschowka JA, Livnat S. Noradrenergic and peptidergic innervation lymphoid tissue. J Immunol 1985; 135: 755s-65s.

4. Ansel JC, Kaynard AH, Armstrong CA, Olerud J, Bunnett N, Payan D. Skin-Nervous system interactions. J Invest Dermatol 1996; 106: 198-204.

5. Aso M, Hashimoto K, Iatnzavi A. Immunohistochemical studies of selected skin diseases and tumors using monoclonal antibodies to neurofilament and myelin proteins. J Am Acad Dermatol 1985; 13: 37-42.

6. Ashraf S, Crowe R, Loizidou MC, Turmaine M, Taylor I, Burnstock G. The absence of autonomic perivascular nerves in human colorectal liver metastases. Br J Can 1996; 73: 349-59.

7. Itchikawa K. Sur l’existence d’une relation entre les nerfs périfériques et le developpement du cancer. Bull Ass Franc Cancer 1928; 21: 590-6.

8. Hall GH, Turnbull LW, Richmond I, Helboe L, Atkin SL. Localization of somatostatin and somatostatin receptors in benign and malignant ovarian tumors. Br J Cancer 2002; 87: 86-90.

9. Kodama Y, Baxter RC, Martin JL. Insulin-like growth factor I inhibits cell growth in the a549 non small lung cancer cell line. Am J Respir Cell Mol Biol 2002; 27: 336-44.

10. Caro WA, Bronstein BR. Malignant tumors of epidermal appendages. Basal cell epithelioma. In: Moschella, SL; Hurley, HJ eds. Dermatology. Philadelphia. WB Saunders 1985; 1364-7.

11. Mac Kie RM. Epidermal Skin Tumors. Basal Cell Carcinoma. In: Champion RH, Burton JL, Ebling FJG. Eds. Textbook of Dermatology 5^th ed. Oxford: Blackwell Scientific Publications, 1992: 1488-95.

12. Cramer W. Innervation as a factor in the experimental production of cancer. Br J Exp Pathol 1925; 6: 71-4.

13. Winkelman RK. Cutaneous nerves in relation to epithelial tumors. J Invest Dermatol 1956; 27: 273-9.

14. Ormea F, Amerio PL, Bossi G. Contributo allo studio dell’innervazione delle neoplasie epiteliali cutanee. Minerva Dermatol 1968; 43: 658-82.

15. Pawlowski A. The role of nerve elements of the skin in the course of experimental carcinogenesis and basal cell epithelioma and squamous cell carcinoma in man. Neuropatol Pol 1970; 8: 205-39.

16. Weiner N. Regulation of norepinephrine biosynthesis. Ann Rev Pharmacol 1970; 10: 273-90.

17. Shi SR, Key ME, Kalkra KL. Antigen retrieval in formalin fixed, paraffin embedded tissues. An enhancement method for immunohistochemical staining based on microwave oven heating of tissues sections. J Histochem Cytochem 1991; 39: 741-8.

18. Jacobs GH, Rippey JJ, Altini M. Prediction of aggressive behavior in basal cell carcinoma. Cancer 1982; 49: 533-7.

19. Requena L, Sanchez E. Variantes Histológicas del epitelioma Basocelular. Piel 1991; 6: 21-30.

20. Wick MR. Malignant tumors of the epidermis. Basal cell carcinoma. In: Farmer ER, Hooh AF. Eds. Pathology of the Skin. New Jersey. Prentice-Hall International, 1990; 568-79.

21. Kirkhan N. Tumor and Cyst of the Epidermis. D Elder, R Elenitsas, Ch Jaworsky, Johnson B Eds. Lever’s istopathology of the Skin. Philadelphia. Lippincott-Raven 8^th ED. pp 685-746.

22. Oertel H. Innervation of Human Cancer. J Pathol Bact 1929; 32: 557-64.

23. Pincelli C, Servignani C, Manfredini R, Grande A, Fantini F, Bracci-Laudiero L, Aloe L, Ferrari S, Cossarizza A, Giannetti A. Expression and function of Nerve Growth Factor and Nerve Growth Factor Receptor on Cultured Keratinocytes. J Invest Dermatol 1994; 103: 13-8.

24. Kaminishi M, Shimiza N, Shimoyama S, Yamaguchi H, Tsuji E, Aoki F, Nomura S, Yoshikawa A. Denervation promotes development of cancer-related lesions in the gastric remnant. J Clin Gastroenterol 1997; 25 Suppl 1: S129-34.

25. Besedovsky HO, del Rey A. Immune-neuro-endocrine interactions: facts and hypotheses. Endocr Rev 1996 Feb; 17(1): 64-102.