Camilla Pease

Hapten-protein binding : from theory to practical application in the in vitro prediction of skin sensitization

In view of the forthcoming European Union ban on in vivo testing of cosmetic and toiletry ingredients, following the publication of the 7th amendment to the Cosmetics Directive, the search for practical, alternative, non‐animal approaches is gathering pace. For the end‐point of skin sensitization, the ultimate goal, i.e. the development and validation of alternative in vitro/in silico assays by 2013, may be achieved through a better understanding of the skin sensitization process on the cellular and molecular levels. One of the key molecular events in skin sensitization is protein haptenation, i.e. the chemical modification of self‐skin protein(s) thus forming macromolecular immunogens. This concept is widely accepted and in theory can be used to explain the sensitizing capacity of many known skin sensitizers. Thus, the principle of protein or peptide haptenation could be used in in vitro assays to predict the sensitization potential of a new chemical entity. In this review, we consider some of the theoretical aspects of protein haptenation, how mechanisms of protein haptenation can be investigated experimentally and how we can use such knowledge in the development of novel, alternative approaches for predicting skin sensitization potential in the future.

Xenobiotic metabolism capacities of human skin in comparison with a 3D epidermis model and keratinocyte-based cell culture as in vitro alternatives for chemical testing: activating enzymes (Phase I)

Abstract:  The 7th Amendment to the EU Cosmetics Directive prohibits the use of animals in cosmetic testing for certain endpoints, such as genotoxicity. Therefore, skin in vitro models have to replace chemical testing in vivo. However, the metabolic competence neither of human skin nor of alternative in vitro models has so far been fully characterized, although skin is the first‐pass organ for accidentally or purposely (cosmetics and pharmaceuticals) applied chemicals. Thus, there is an urgent need to understand the xenobiotic‐metabolizing capacities of human skin and to compare these activities to models developed to replace animal testing. We have measured the activity of the phase II enzymes glutathione S‐transferase, UDP‐glucuronosyltransferase and N‐acetyltransferase in ex vivo human skin, the 3D epidermal model EpiDerm 200 (EPI‐200), immortalized keratinocyte‐based cell lines (HaCaT and NCTC 2544) and primary normal human epidermal keratinocytes. We show that all three phase II enzymes are present and highly active in skin as compared to phase I. Human skin, therefore, represents a more detoxifying than activating organ. This work systematically compares the activities of three important phase II enzymes in four different in vitro models directly to human skin. We conclude from our studies that 3D epidermal models, like the EPI‐200 employed here, are superior over monolayer cultures in mimicking human skin xenobiotic metabolism and thus better suited for dermatotoxicity testing.

Structure-activity relationships for selected fragrance allergens.

Fragrance substances represent a very diverse group of chemicals, a proportion of them providing not only desirable aroma characteristics, but also being associated with adverse effects, notably the ability to cause allergic reactions in the skin. However, efforts to find substitute materials are hampered by the need to undertake animal testing to evaluate both the presence and the degree of skin sensitization hazard. One potential route to avoid such testing is to understand the relationships between chemical structure and skin sensitization. In the present work we have evaluated two groups of fragrance chemicals, saturated aldehydes (aryl substituted and aliphatic aldehydes) and α,β‐unsaturated aldehydes. Data on their skin sensitization potency defined using the local lymph node assay has been evaluated in relation to their physicochemical properties. The initial outcome has been consistent with the concept that α,β‐unsaturated aldehydes react largely via Michael addition, whilst the group of saturated aldehydes form Schiff bases with proteins. Simple models of chemical reactivity based on these mechanisms suggest that it may be possible to predict allergenic potency. Accordingly, the evaluation of an additional group of similar aldehydes is now underway to assess the robustness of these models, with some emphasis being based on ensuring a wider spread of chemical reactivity.

Further evaluation of quantitative structure–activity relationship models for the prediction of the skin sensitization potency of selected fragrance allergens

Fragrance substances represent a very diverse group of chemicals; a proportion of them are associated with the ability to cause allergic reactions in the skin. Efforts to find substitute materials are hindered by the need to undertake animal testing for determining both skin sensitization hazard and potency. One strategy to avoid such testing is through an understanding of the relationships between chemical structure and skin sensitization, so‐called structure–activity relationships. In recent work, we evaluated 2 groups of fragrance chemicals – saturated aldehydes and α,β‐unsaturated aldehydes. Simple quantitative structure–activity relationship (QSAR) models relating the EC3 values [derived from the local lymph node assay (LLNA)] to physicochemical properties were developed for both sets of aldehydes. In the current study, we evaluated an additional group of carbonyl‐containing compounds to test the predictive power of the developed QSARs and to extend their scope. The QSAR models were used to predict EC3 values of 10 newly selected compounds. Local lymph node assay data generated for these compounds demonstrated that the original QSARs were fairly accurate, but still required improvement. Development of these QSAR models has provided us with a better understanding of the potential mechanisms of action for aldehydes, and hence how to avoid or limit allergy. Knowledge generated from this work is being incorporated into new/improved rules for sensitization in the expert toxicity prediction system, deductive estimation of risk from existing knowledge (DEREK).

Effect of glutathione on the covalent binding of the 13C-labeled skin sensitizer 5-chloro-2-methylisothiazol-3-one to human serum albumin: identification of adducts by nuclear magnetic resonance, matrix-assisted laser desorption/ionization mass spectrometry, and nanoelectrospray tandem mass spectrometry.

The covalent binding of 4-[(13)C]- and 5-[(13)C]-5-chloro-2-methylisothiazol-3-one (MCI) toward human serum albumin (HSA) was followed by (13)C and (1)H[(13)C] NMR spectroscopy. MCI was found to react with histidine through an addition-elimination at position 5, leading to stable substitution adducts, and with lysine to form open adducts of the thioamide or amide type. No other modification could be detected on either cysteine or tyrosine. In the presence of glutathione (GSH), we observed an increased covalent binding to lysine residues. This could be explained by the rapid reaction of GSH with MCI to form a chlorothioacyl intermediate very reactive toward primary amino groups of lysine residues. To further confirm these observations and map covalent binding sites, HSA samples modified by MCI with or without GSH were analyzed by matrix-assisted laser desorption/ionization mass spectrometry of tryptic digests and electrospray tandem mass spectrometry of modified peptides purified by reverse phase HPLC. About 80% of the HSA sequence was mapped, and several modified peptides were identified. When HSA was incubated with MCI without GSH, three peptides modified at histidine residues were characterized while when HSA was incubated in the presence of GSH, five peptides modified at histidine and lysine residues were identified. These experiments confirmed that modifications on lysine residues were of the amide and thioamide types. Observed modifications were in accordance with mass increases corresponding to structures identified by NMR, and an extra adduct corresponding to a double modification of His 338 was observed. Comparison of HSA-MCI and HSA-MCI-GSH samples confirmed that the presence of GSH increased the modification of lysine residues.

Xenobiotic metabolism in human skin and 3D human skin reconstructs: A review

In this review, we discuss and compare studies of xenobiotic metabolism in both human skin and 3D human skin reconstructs. In comparison to the liver, the skin is a less studied organ in terms of characterising metabolic capability. While the skin forms the major protective barrier to environmental chemical exposure, it is also a potential target organ for adverse health effects. Occupational, accidental or intended-use exposure to toxic chemicals could result in acute or delayed injury to the skin (e.g. inflammation, allergy, cancer). Skin metabolism may play a role in the manifestation or amelioration of adverse effects via the topical route. Today, we have robust testing strategies to assess the potential for local skin toxicity of chemical exposure. Such methods (e.g. the local lymph node assay for assessing skin sensitisation; skin painting carcinogenicity studies) incorporate skin metabolism implicitly in the in vivo model system used. In light of recent European legislation (i.e. 7(th) Amendment to the Cosmetics Directive and Registration Evaluation and Authorisation of existing Chemicals (REACH)), non-animal approaches will be required to reduce and replace animal experiments for chemical risk assessment. It is expected that new models and approaches will need to account for skin metabolism explicitly, as the mechanisms of adverse effects in the skin are deconvoluted. 3D skin models have been proposed as a tool to use in new in vitro alternative approaches. In order to be able to use 3D skin models in this context, we need to understand their metabolic competency in relation to xenobiotic biotransformation and whether functional activity is representative of that seen in human skin.

Activation of human dendritic cells by p-phenylenediamine.

Exposure to p-phenylenediamine (pPD), a primary intermediate in hair dye formulations, is often associated with the development of allergic contact dermatitis. Such reactions involve activation of the subjects immune system. The aim of these studies was to explore the relationship between pPD oxidation and functional maturation of human monocyte-derived dendritic cells in vitro. Dendritic cells were incubated with pPD and Bandrowskis base (BB) for 16 h, and expression of the costimulatory receptors CD40, CD80, CD83, CD86, and major histocompatability complex class II intracellular glutathione levels and cell viability were measured. In certain experiments, glutathione (1 mM) was added to culture medium. Liquid chromatography-mass spectrometry (LC-MS) analysis and exhaustive solvent extraction were used to monitor the rate of [14C]pPD oxidation and the extent of pPD binding to cellular and serum protein, respectively. Proliferation of allogeneic lymphocytes was determined by incorporation of [3H]thymidine. Exposure of dendritic cells to pPD (5–50 μM), but not BB, was associated with an increase in CD40 and MHC class II expression and proliferation of allogeneic lymphocytes. Dendritic cell activation with pPD was not associated with apoptotic or necrotic cell death or depletion of glutathione. Neither pPD nor BB altered dendritic cell expression of CD80, CD83, or CD86. LC-MS analysis revealed pPD was rapidly oxidized in cell culture media to BB. Addition of glutathione inhibited BB formation but did not prevent covalent binding of pPD to dendritic cell protein or dendritic cell activation. Collectively, these studies show that pPD, but not BB, selectively activates human dendritic cells in vitro.

Mass spectrometric identification of covalent adducts of the skin allergen 2,4-dinitro-1-chlorobenzene and model skin proteins.

A large proportion of allergic skin reactions are considered to be the result of skin exposure to small organic chemicals that possess the intrinsic ability to covalently modify skin proteins, either directly or following activation. In the absence of information about specific skin protein targets, studies of chemical modifications are limited to the use of model proteins. We have previously demonstrated that selected well known skin sensitizers (2,4-dinitro-1-chlorobenzene and phenyl salicylate) have the ability to covalently modify residues selectively on the model protein, human serum albumin. In the present work, we focus on the differences in covalent binding observed for two additional model proteins, human cytokeratin 14 and human cofilin, both constituent proteins of skin. Using matrix assisted laser desorption/ionization mass spectrometry (MALDI-MS) and nano LC-MS and -MS/MS strategies, the amino acid residues targeted by 2,4-dinitro-1-chlorobenzene on the two model proteins have been identified. In contrast, a structurally related non-sensitiser (2,4-dichloro-1-nitrobenzene) and a non-sensitising irritant (benzalkonium chloride) did not covalently modify the model proteins. Detailed examination of the results for the sensitizers indicate that reactive chemicals target nucleophilic amino acids residing in specific microenvironments of the 3D protein structure that are conducive to reactivity. This observation has important implications for the development of hapten-peptide binding assays. It is envisaged that the data from such assays will be integrated with outputs from other in vitro assays in the future to give a prediction of the sensitisation potential of novel chemicals.

Intermittent exposure to low-concentration paraphenylenediamine can be equivalent to single, higher-dose exposure.

Hair dye allergy is an important and increasingly common cause of allergic contact dermatitis. The role of repeated exposure in elicitation of allergy has not previously been extensively studied. We have therefore compared elicitation between single and intermittent exposure to paraphenylenediamine (PPD). 23 subjects known to be allergic to PPD from positive patch tests were exposed to 0.3% and 0.03% PPD, both in petrolatum and water, for 5 min at the same site every day for up to 8 D. In the same subjects, single exposures were also performed at different sites, from 5 to 40 min. Other experiments exposed rat skin to radiolabelled PPD as one‐off application or multiple exposures. There were 8 reactions in the cumulative exposure site using 0.3% PPD in aqueous solution. In 7 of these, there was an exact correlation with reaction to the cumulative time needed for repeat exposures to elicit a reaction and the time needed for a reaction to the single exposure. There were no reactions to 0.03% PPD in water or pet under either type of exposure condition. There was also a positive correlation between grade of original reaction in clinic (+++, ++, +) and appearance/intensity of elicitation reactions. In the animal study, cumulative time and single exposure time sites correlated with regards to retention of radiolabelled substance within the skin. This study therefore demonstrates for the first time that, over the time period tested, the allergenic component of PPD accumulates in the skin. Hence, intermittent exposure to lower concentrations of PPD may be equivalent to higher concentration, one‐off exposure.

Investigation of the skin sensitizing activity of linalool

An increasing range of chemicals appears to be capable of causing skin sensitization as a result of their capacity to undergo air oxidation (autoxidation) with the consequent formation of reactive species such as epoxides and hydroperoxides. In this small investigation, the ability of linalool, a common fragrance ingredient, to cause such effects was quantified using the local lymph node assay before and after careful purification by vacuum distillation. The commercially available grade of linalool (97% purity) was shown to be a weak skin sensitizer. Various impurities, including linalool oxide, dihydrolinalool, epoxylinalool, 3‐hexenyl butyrate and 3,7‐dimethyl‐1,7‐octadiene‐3,6‐diol were identified and were completely removed (except for the dihydrolinalool remaining at 1.4%) and the re‐purified linalool retested. Neither linalool or dihydrolinalool are protein‐reactive compounds. The sensitization potency of the re‐purified linalool sample was considerably reduced, but not entirely eliminated, suggesting either that an allergenic impurity could be very quickly reformed by mechanisms of activation or that certain potent undetectable allergens remained. Both possibilities are consistent with what is understood of the chemistry and composition of commercially available linalool.