increasing the genetic damage produced by UV rays and contributing to the continuous raising trend of MM. This can be counterintuitive if considering a possible compensatory effect due to the sunscreen, but it is realistic to hypothesize that the compensatory effect is not complete. Due to the long human evolution pathway (8), skin melanin should have, at least in theory, a much higher efficacy (UV absorbing and protecting properties) than anthropogenic sunscreens. How to test the hypothesisThe relationship between melanoma and sunscreen use is difficult to make. This hypothesis would be far less speculative if in vivo data would be available. First of all, the real concentration of tyrosinase inhibitor sunscreens at the level of the basal lamina should be measured using microdialysis, and then, the effect of these compounds on melanoma induced by carcinogenesis protocols should be tested in mice. Relevance and perspectivesSunscreens are devised to protect us from the harmful effects of UV rays: it seems incredible and paradoxical that these products are not tested to evaluate their activity on tyrosinase, the enzyme that provides us with the natural protection afforded by melanin. It is our opinion that its use should be banned in any product that may enter in contact with human skin. Alternatively, advanced formulation strategies, which contemplate the organic UV absorbers encapsulation or inclusion with no or minimal release, should be strongly promoted. These novel formulation approaches permit to reduce or completely eliminate the contact between these molecules and the skin avoiding absorption. Encapsulated sunscreens remain efficient UV absorbers on the skin surface (9,10). Authors' contributions Conflict of interestsThe authors have declared no conflicting interests. Supporting InformationAdditional supporting data may be found in the supplementary information of this article. Appendix S1. References. Figure S1. Names, molecular formula, molecular weight, solubility and chemical structure of the 10 compounds mentioned in the study. at different concentrations. L-Tyrosine oxidation by tyrosinase was spectrophotometrically determined as previously described with minor modifications [S13]. 150 ll of 5 mM L-tyrosine in 67 mM sodium phosphate buffer (pH 6.8) was mixed with 50 ll of the same buffer with or without the compound to test in a 96-well plate. The reaction was started by further added 50 ll (150 U/ml) of mushroom tyrosinase (SigmaAldrich, St. Louis, MO, USA). Dopachrome formation from the reaction mixture was determined as the increase of absorbance at wavelength 492 nm per min (DA 492/ min) by using a Molecular Devices microplate reader every five minutes for 120 min. All compounds used in the assay (Fig. S1) were dissolved in DMSO and diluted to the finals concentrations of 50 and100 lM with exception UVINUL MS40 that was dissolved in water. The final DMSO concentration in the assay was always <3%.

The abundance of xenobiotic metabolizing enzymes (XMEs) is different in the skin and liver; therefore, it is important to differentiate between liver and skin metabolism when applying the information to safety assessment of topically applied ingredients in cosmetics. Here, we have employed EpiSkin™ S9 and human liver S9 to investigate the organ‐specific metabolic stability of 47 cosmetic‐relevant chemicals. The rank order of the metabolic rate of six chemicals in primary human hepatocytes and liver S9 matched relatively well. XME pathways in liver S9 were also present in EpiSkin S9; however, the rate of metabolism tended to be lower in the latter. It was possible to rank chemicals into low‐, medium‐ and high‐clearance chemicals and compare rates of metabolism across chemicals with similar structures. The determination of the half‐life for 21 chemicals was affected by one or more factors such as spontaneous reaction with cofactors or non‐specific binding, but these technical issues could be accounted for in most cases. There were seven chemicals that were metabolized by liver S9 but not by EpiSkin S9: 4‐amino‐3‐nitrophenol, resorcinol, cinnamyl alcohol and 2‐acetylaminofluorene (slowly metabolized); and cyclophosphamide, benzophenone, and 6‐methylcoumarin. These data support the use of human liver and EpiSkin S9 as screening assays to indicate the liver and skin metabolic stability of a chemical and to allow for comparisons across structurally similar chemicals. Moreover, these data can be used to estimate the systemic bioavailability and clearance of chemicals applied topically, which will ultimately help with the safety assessment of cosmetics ingredients.

An understanding of the bioavailability of topically applied cosmetics ingredients is key to predicting their local skin and systemic toxicity and making a safety assessment. We investigated whether short-term incubations with S9 from the reconstructed epidermal skin model, EpiSkin™, would give an indication of the rate of chemical metabolism and produce similar metabolites to those formed in incubations with human skin explants. Both have advantages: EpiSkin™ S9 is a higher-throughput assay, while the human skin explant model represents a longer incubation duration (24 hours) model integrating cutaneous distribution with metabolite formation. Here, we compared the metabolism of 10 chemicals (caffeine, vanillin, cinnamyl alcohol, propylparaben, 4-amino-3-nitrophenol, resorcinol, 4-chloroaniline, 2-amino-3-methyl-3H-imidazo[4,5-F]quinoline and 2-acetyl aminofluorene) in both models. Both models were shown to have functional Phase 1 and 2 enzymes, including cytochrome P450 activities. There was a good concordance between the models with respect to the level of metabolism (stable vs. slowly vs. extensively metabolized chemicals) and major early metabolites produced for eight chemicals. Discordant results for two chemicals were attributed to a lack of the appropriate cofactor (NADP + ) in S9 incubations (cinnamyl alcohol) and protein binding influencing chemical uptake in skin explants (4-chloroaniline). These data support the use of EpiSkin™ S9 as a screening assay to provide an initial indication of the metabolic stability of a chemical applied topically. If required, chemicals that are not metabolized by Epi-Skin™ S9 can be tested in longer-term incubations with in vitro human explant skin to determine whether it is slowly metabolized or not metabolized at all.

Absence of a functional proteasome in the suprabasal layers of the epidermis is responsible for keratosis linearis with ichthyosis congenital and sclerosing keratoderma syndrome. Patient epidermis shows hypergranulosis associated with abnormally shaped keratohyalin granules and abnormal distribution of filaggrin in the Stratum granulosum and Stratum corneum. This suggests that the proteasome is involved in the degradation of filaggrin. To test this hypothesis, the proteasome proteolytic activity was inhibited in 3D reconstructed human epidermis (RHE) with the specific clasto‐lactacystin β‐lactone inhibitor. Confirming the efficacy of inhibition, ubiquitinated proteins accumulated in treated RHEs as compared to controls. Levels of urocanic acid (UCA) and pyrrolidone carboxylic acid (PCA), the end products of filaggrin degradation, were reduced. However, neither filaggrin accumulation nor appearance of filaggrin‐derived peptides were observed. On the contrary, the amount of filaggrin was shown to decrease, and a similar tendency was observed for profilaggrin, its precursor. Accumulation of small cytoplasmic vesicles associated with a significant increase in autophagy markers indicated activation of the autophagy process upon proteasome inhibition. Taken together, these results suggest that the perturbation of UCA and PCA production after proteasome inhibition was probably due to down‐regulation of filaggrin expression rather than to blocking of filaggrin proteolysis.

No abstract