Hair dyes have been in use for thousands of years. The market of cosmetics keeps on growing, thus hair dyes are being used by millions of consumers. Although the majority of the population reaching for hair dyes are women, there are also several male users. Hair dyes can be classified into several groups: plant dyes, metal salts, direct and oxidative dyes. Hair dyes may also be the cause of contact allergies, as some exhibit high allergenic potential in human and animal studies. For this reason, certain restrictions and prohibitions on the use of hair dyes are in effect, which are listed in the Regulation of cosmetic products no. 1223/2009 in Annex II and Annex III. In the scope of this Thesis, we studied 13 hair dyes, which most frequently appeared in cosmetic hair dye products and their autoxidation products. The latter are, to a large extent, the underlying reason for the development of allergic reactions. For skin irritation and sensitization predictions of selected hair dyes we used in silico methods, which are designed to complement existing toxicity tests on animals and humans for predicting toxicity. In silico methods predict the ability of the irritant or sensitization potential of a particular compound based on its physico-chemical properties or presence of structural alarms. The ability of the irritant and sensitization potential of selected hair dyes and their autoxidation products was examined using three softwares: Toxtree, Derek Nexus and VegaNIC. The obtained in silico predictions were compared with the already existing published data, obtained from in vivo studies in animals and humans. We used Toxtree and Derek Nexus to predict skin irritation. For most compounds, Toxtree did not provide prediction data, therefore the irritant potential of some dyes and their autoxidation products was marked as unknown. A review of the literature data revealed that Derek Nexus generated accurate predictions for most of the dyes. We used all three programs to predict the sensitization potential, which produced fairly uniform predictions. Derek Nexus once again yielded the most accurate predictions. Toxtree, VegaNIC and Derek Nexus generated accurate and uniform prediction with respect to the sensitization potential for the following dyes: m-aminophenol, 4-amino-2-hydroxytoluene, 2-methylresorcinol, 4-chlororesorcinol and HC Blue n ° 12. They also predicted that Acid Violet 43 is not a sensitiser, which was in agreement with in vivo studies. All programs incorrectly predicted sensitization potential for 2,4-diaminophenoxyethanol×HCl and 2-amino-4-hydroxyethylamino-anisole sulphate, which did not exhibit sensitization potential in in vivo studies.
|
