We have hypothesised that melanocytes disappear in vitiligo because they are weakly attached to the epidermal basal membrane (melanocytorrhagy). In the epidermis, attachment of melanocytes to collagen IV is mediated through DDR1, which is under the control of CCN3. DDR1 genetic variants have been associated with vitiligo in patients of different ethnic origin. In vitro studies have shown that inhibition of CCN3 induces the detachment of melanocytes. We have studied in parallel the expression of CCN3 and DDR1 in lesional and perilesional skin of patients with vitiligo and the impact of the silencing of CCN3 and DDR1 in normal human melanocytes on their behaviour in epidermal reconstructs. Our in vivo study provides evidence of a dysregulation of the DDR1-CCN3 interaction in vitiligo skin as melanocytes remaining in perilesional skin did not express CCN3. Expression of DDR1 was decreased in lesional versus perilesional vitiligo skin in the majority of patients, and the expression of collagen IV was found decreased in all patients. Silencing of CCN3 in melanocytes induced a significant inhibition of cell adhesion to collagen IV whereas melanocytes transduced with shDDR1 still adhered well on collagen IV and did not increase melanocyte loss in epidermal reconstructs as compared with normal melanocytes. Melanocyte detachment was observed but not in all reconstructs using CCN3 silenced melanocytes. Overall, our study confirms that a downregulation of CCN3 is implicated in melanocyte adhesion in part through DDR1. In vitiligo skin, the interaction of CCN3 with other molecules, such as TGFβ and CCN2, needs to be addressed.

Exposure to electromagnetic radiations (EMR) produced by mobile phone concerns half the world's population and raises the problem of their impact on human health. In this study, we looked at the effects of mobile phone exposure (GSM basic, 900 MHz, SAR 2 mW g(-1) , 6 h) on a model of pigmented skin. We have analysed the expression and localization of various markers of keratinocyte and melanocyte differentiation 2, 6, 18 and 24 h after EMR exposure of reconstructed epidermis containing either only keratinocytes or a combination of keratinocytes and melanocytes grown on dead de-epidermized dermis, using histology, immunohistochemistry and Western blot. No changes were found in epidermal architecture, localization of epidermal markers, presence of apoptotic cells and the induction of p53 in both types of epidermis (with or without melanocytes) after exposure to EMR. In pigmented reconstructs, no change in the location and dendricity of melanocytes and in melanin transfer to neighbouring keratinocytes was detected after EMR exposure. Loricrin, cytokeratin 14 were significantly decreased at 6 h. The level of all markers increased at 24 h as compared to 6 h post-EMR exposure, associated with a significant decrease of the 20S proteasome activity. Our data indicate that exposure to 900 MHz frequency induces a transient alteration of epidermal homoeostasis, which may alter the protective capacity of the skin against external factors. Presence or absence of melanocytes did not modify the behaviour of reconstructs after EMR exposure.