Perspectives and strategies of alternative methods used in the risk assessment of personal care products

+Model PHARMA-413; No. of Pages 14

ARTICLE IN PRESS

Annales Pharmaceutiques Françaises (2015) xxx, xxx—xxx

Disponible en ligne sur

ScienceDirect www.sciencedirect.com

GENERAL REVIEW

Perspectives and strategies of alternative methods used in the risk assessment of personal care products Perspectives et stratégies des méthodes alternatives utilisées lors de l’évaluation du risque des produits cosmétiques P. Quantin , A. Thélu , S. Catoire , H. Ficheux ∗ Département de toxicologie, Thor Personal Care, 147, rue Irène-Joliot-Curie, 60208 Compiègne cedex, France Received 27 March 2015; accepted 11 June 2015

KEYWORDS Risk assessment; Toxicology; Skin; Cosmetics; Alternative methods

Summary Risk assessment for personal care products requires the use of alternative methods since animal testing is now totally banned. Some of these methods are effective and have been validated by the ‘‘European Union Reference Laboratory for alternatives to animal testing’’; but there is still a need for development and implementation of methods for specific endpoints. In this review, we have focused on dermal risk assessment because it is the prime route of absorption and main target organ for personal care products. Within this field, various areas must be assessed: irritation, sensitisation and toxicokinetic. Personal care product behaviour after use by the consumer and potential effects on the environment are also discussed. The purpose of this review is to show evolution and the prospects of alternative methods for safety dermal assessment. Assessment strategies must be adapted to the different chemical classes of substances studied but also to the way in which they are used. Finally, experimental and theoretical technical parameters that may impact on measured effects have been identified and discussed. © 2015 Published by Elsevier Masson SAS.

∗

Corresponding author. E-mail addresses: [email protected] (P. Quantin), [email protected] (A. Thélu), [email protected] (S. Catoire), herve.fi[email protected] (H. Ficheux). http://dx.doi.org/10.1016/j.pharma.2015.06.002 0003-4509/© 2015 Published by Elsevier Masson SAS.

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002

+Model PHARMA-413; No. of Pages 14

ARTICLE IN PRESS

2

P. Quantin et al.

MOTS CLÉS Évaluation du risque ; Toxicologie ; Peau ; Cosmétiques ; Méthodes alternatives

Résumé L’évaluation du risque en cosmétique nécessite l’utilisation de méthodes alternatives depuis que l’expérimentation sur les animaux a été totalement interdite. Certaines de ces méthodes sont efficaces et validées par l’« European Union Reference Laboratory for alternatives to animal testing » ; cependant, il existe toujours un besoin d’améliorer ces méthodes et d’en développer de nouvelles pour l’étude de certains effets toxiques. Dans cette revue générale, nous nous sommes concentrés sur l’évaluation du risque au niveau cutané car il s’agit de la voie d’absorption et de la cible principale des molécules utilisées en cosmétique. Dans ce cadre, différents aspects doivent être évalués : irritation, sensibilisation et toxicocinétique. Le devenir des produits cosmétiques après leur utilisation par le consommateur ainsi que des potentiels effets sur l’environnement sont également abordés. Le but de cette revue est de montrer l’évolution et les perspectives des méthodes alternatives utilisables pour l’évaluation du risque dans le domaine cosmétique. Les stratégies doivent être adaptées aux catégories chimiques des substances étudiées et à leur mode d’utilisation. Enfin, différents paramètres techniques, expérimentaux et théoriques, qui font varier les réponses mesurées sur les différents modèles ont été identifiés et analysés. © 2015 Publi´ e par Elsevier Masson SAS.

Introduction Risk assessment is commonly used in various industries such as pharmaceutical, chemical, personal care, food, agricultural. The first step is to identify hazard and the second to study the probability that hazard becomes a risk to the population studied. Exposure is the third component to be considered in the equation. Manufacturers must ensure that their products meet current regulations and cause no health hazards when used under normal and reasonable conditions. Personal care regulation is established at a European level and safety of use is evaluated according to specific guidelines from the Organisation for Economic Co-operation and Development (OECD), which controls substances, their chemical structures and their toxicological and exposure profiles. Numerous data and studies are required to identify hazards [1] in order to establish toxicological profiles: acute toxicity, corrosion and irritation, sensitisation, repeated toxicity, reproductive toxicity, carcinogenotoxicity, photo-induced toxicity and toxicokinetics studies. In addition to human toxicity assessment, ecotoxicity is becoming an increasingly concern and is therefore more and more investigated. The evolution of European regulations results in a reduction in the number of animals used in research to ensure the safety of chemical molecules to which humans are exposed. According to regulation No. 1223/2009 of the European Commission, experimentation on animals is prohibited for personal care products. Since September 2004, a testing ban on animals has been applied for finished personal care products, whereas the testing ban on ingredients or combinations has been effective since March 2009. For specific effects on health (repetitive dose toxicity, reproductive toxicity and toxicokinetics), the testing ban has been in force since March 2013.

Alternative approaches (in vitro, in silico and in chemico methods) have been proposed for risk assessment in cosmetology [2]. Some of these methods are effective and have been validated by EURL-ECVAM (European Union Reference Laboratory for Alternatives to Animal Testing) as tests for irritation, corrosion and phototoxicity. Concerning sensitisation, two methods have just been validated, and two others are in the last stage of validation and will be available very soon. However, doubts have been expressed by the European Scientific Committee on the possibility of replacing all animal testing by reliable alternative methods which guarantee a high level of safety for consumers. This is particularly the case for repeated dose toxicity, reproductive toxicity and toxicokinetics tests. Therefore, when dealing with the risk assessment of personal care products, there are two main items to be addressed.

Which tools/methods/models are ready to be used today for personal care risk assessment? A state of the art is proposed with the description and outlook of alternative methods available or soon to be validated by European agreements to predict local dermal toxicity of topical applied substances such as percutaneous absorption, irritation, corrosion, phototoxicity and sensitisation.

What needs to be improved in the near future and how can this be integrated into the global risk assessment strategy? Toxicological parameters and associated alternative methods, which need strong improvements and integration in global risk assessment strategies before they can be used in labs are discussed. Dermal toxicokinetic assessment,

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002

+Model PHARMA-413; No. of Pages 14

ARTICLE IN PRESS

Perspectives and strategies of alternative methods used in risk assessment targeting consumer population, use of in silico approaches and ecotoxicological monitoring of personal care products are highlighted. This work aims to inform those concerned with the evolution of toxicology techniques applied in the personal care industry. A focus is made on the area of ‘‘local dermal risk assessment’’ because it is the prime route of absorption of personal care products into the human body as noted in regulation No. 1223/2009 (annex 1, part B, section 10). Although manufacturers do not greatly address ecotoxicology assessment, it is of real importance from a consumer point of view and therefore is also covered in this review.

What is available today? Today, regulatory approved tests are available for the key area of ‘‘topical toxicity’’, which relates mainly to the skin. As a reminder, the skin is the largest and heaviest organ in the human body. This acts as a real interface with the external environment constituting a physical and vital barrier for the body [3]. However, this barrier is not completely hermetic and may be permeable to many substances. Skin consists of two layers: the epidermis and dermis. A third layer, the hypodermis, is often considered to be part of the skin. In topical toxicity assessment of personal care products, a focus is made on the epidermis because it is in this tissue that the majority of cutaneous toxic reactions take place. Epidermis (Fig. 1) is constituted of four layers accordingly to their differentiation stage. The outermost layer, the stratum corneum (SC), is considered as the major barrier to skin penetration [4]. Hence the SC is the most relevant layer in chemical absorption assessment studies. Keratinocytes are the main cell population but other cell types can also be found as immune cells (Langerhans cells), nerve cells (Merkel cells) and pigmentation cells (melanocytes). Epidermis thickness varies according to its localisation between 0.04 mm (eyelid) to 1.6 mm (sole) and is the only skin layer which is not vascularised. Stratum corneum is a structure of 10—20 ␮m thickness composed of multiple corneocyte layers, which are unsustainable enucleate cells filled with keratin (hydrophilic) and securely fastened to each other by solid junctions, corneodesmosomes. The SC structure is shown schematically as a brick and mortar wall with hydrophilic protein cells surrounded by a cement lipophilic lipid covalently linked to the proteins of the cell membrane [5]. The stratum corneum also has a reservoir function for topically applied substances [6] and is involved in skin hydration [7]. Some substances may remain stored for a few days and be released slowly in systemic circulation and therefore local and systemic exposition can continue even after the end of the application.

Percutaneous absorption Before the onset of toxic effects, potentially toxic substances have to penetrate into the epidermis. Percutaneous absorption is the transfer of a xenobiotic through the skin from an external environment into the blood circulation. Dermal penetration is the first step of absorption, which consists of the entrance of a substance into the cutaneous layer. It is widely accepted, in risk assessment, that a molecule

3

able to cross the epidermis will reach the systemic circulation.

Percutaneous absorption models Percutaneous penetration studies are mainly performed with a diffusion cell system called Franz cell [8]. The test preparation is applied to the skin for a specific duration based on appropriate human exposure. Samples of the receiver fluid are collected at different times to determine the amount (and rate) of test substance that has penetrated through the skin. At the end of the study, the total absorption of the test substance through the skin is calculated. It can be expressed as a percentage, amount or ratio of absorption per unit time, or in the case of infinite dose, permeability constant. Concerning skin models, human ex vivo skin is the best option, as recommended by the OECD [9] and the Scientific Community on Consumer Safety (SCCS, [1]) for skin penetration studies. However, it is relatively expensive and the high inter-individual variability is a drawback. Pig skin is a relevant alternative because of the similarity in skin properties to human skin but suffers the same drawbacks as human skin and in addition, ethical issues. Other relevant models available are in vitro Reconstructed Human Epidermis (RHE) but there are as yet no validated alternative methods to evaluate skin penetration [10]. There are many benefits of RHE use: good measurement repeatability (intra-model), no availability problems, no ethical issues, possibility of obtaining same magnitude of permeability as ex vivo skin when comparing several substances with different physicochemical properties. Even if good repeatability for one model can be achieved, there is a significant inter-model variability. Moreover, insufficient barrier function in current models is blamed on impairment of desquamation, existence of no keratinised microscopic foci and differences in lipid composition and organisation. Finally, skin permeability can vary by 5 to 50 times more than ex vivo skin [11,12]. Finally, another model available for percutaneous absorptions studies is artificial membrane. Various artificial systems exist; among them: Stratum Corneum Substitute [13], Parallel Artificially Membrane Permeability Assay (PAMPA) skin [14], Carbosil® membrane [15] and other miscellaneous membranes. These are used today as a screening tool to predict chemical absorption but still under assessment as potential alternatives to human skin and RHE. Results obtained on percutaneous assays with artificial membranes are still not sufficient to confirm their relevance. On one hand, several authors have shown good correlation with human skin for a limited range of chemicals (e.g. [16]), but on the other hand synthetic membranes are very sensitive to changes in experimental conditions and formulations [17].

Parameters affecting percutaneous absorption The European Food Safety Authority (EFSA) published a guidance on dermal absorption [18] in which parameters to be considered are listed with special attention to formulations. One of the main parameters is the molecular weight (MW) of the molecule studied. The diffusion of a molecule is easier when it is small. A MW of more than 500 Dalton (Da) is a limiting factor [19]; if it exceeds 1000 Da, passage is

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002

+Model

ARTICLE IN PRESS

PHARMA-413; No. of Pages 14

4

P. Quantin et al.

Stratum corneum Stratum granulosum

Langerhans cell

Epidermis Stratum spinosum

Keratinocyte

Melanocyte

Stratum basal

3 Basement membrane Fibroblast

Dermis

Blood circulation

Figure 1. Epidermis structure. Epidermis is composed of four layers distributed in a differentiation gradient (from the basement layer, the least differentiated, to cornea layer). It relies on a basement membrane that separates it from the underlying dermis, which is a supporting connective tissue. Structure de l’épiderme. L’épiderme se compose de quatre couches distribuées selon un gradient de différentiation (de la couche basale, la moins différentiée, à la couche cornée). Il repose sur une membrane basale qui le sépare du derme sous-jacent qui est un tissu conjonctif de support. Modified from ‘‘Servier Medical Art’’© , http://smart.servier.fr/servier-medical-art.

almost impossible on healthy skin. However, this empiric rule of the 500 cut-off is today under discussion and simply attributed to a low number of data available for high molecular weight molecules [20]. The arguments for the 500 Da cut-off discussed by Meinardi et al. [19] are now being questioned; there is no experimental data affirming molecules with MW > 500 Da cannot pass through the skin. ‘‘Virtually all contact allergen are under 500 Dalton, larger molecules are not known as contact sensitisers [. . .]’’, this is the main argument, which is based solely on a lack of information. In the face of such a discussion, it is clear that data on skin penetration of molecules with MW > 500 Da is needed. Beside size of chemical, other physicochemical properties can influence skin absorption, i.e. liposolubility. Substances with a higher rate of absorption are amphiphilic. Lipophilic properties are necessary to cross the stratum corneum, which is lipid rich while hydrophilic properties allow resorption into the dermis and vascular network [21]. Liposolubility of a substance is given by its octanol/water partition coefficient (Log Kow or Log P). Skin absorption of a substance is intended to be maximal when 1 < Log Kow < 2. On the contrary, substances with Log Kow <-1 or Log Kow > 4 penetrate at very low rate into the skin. It was agreed for risk assessment that in combination with a MW > 500 Da, a

default value of 10% of absorption can be applied for such substances [18]. However, as mentioned before, the empirical 500 Da cut-off rule is still questionable. Another important physical property to consider is the ionisation state. Substances, which are highly ionised are absorbed by skin at very low rate. Ionisation of the molecule depends both on its pKa and of the pH of the environment. When the molecule is non-ionised, its diffusion through the skin is optimised [22]. A final parameter to consider is the affinity of the substance to its carrier or formulation. The substance must be soluble into the vehicle but not so much that it is not retained in it and thus able to enter into the SC. In that way, high affinity between substance and vehicle is a limiting factor for skin absorption.

Dermal toxicity Skin irritation/corrosion These two endpoints are described together because of their similar mechanism of action [23]. For the majority of personal care products, corrosion is of course not expected; however a minority of personal care

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002

+Model PHARMA-413; No. of Pages 14

ARTICLE IN PRESS

Perspectives and strategies of alternative methods used in risk assessment ingredients added in hair bleaching products for example, may lead to unexpected skin corrosion reactions. There is an important difference concerning their reversibility. Skin irritation is an inflammatory reaction with reversible injuries occurring after a chemical exposition up to four hours and characterised by erythema, edema or abrasion. Cutaneous corrosion is the production of irreversible damage following the application of the test substance to skin and is typified by ulcers, bleeding, bloody scabs and by discolouration. Using validated alternative methods on reconstructed human epidermis, these two endpoints are measured by correlation with cell viability decrease as described in OECD Test Guideline (TG) 431 [24] and 439 [25] for corrosion and irritation respectively. Different criteria of viability assessment allow classification with the Globally Harmonized System (GHS) into corrosion and/or irritant chemical classes. RHE is exposed to the test substances for 2 to 60 minutes. For skin irritation, the viability assessment is carried out after a recovery time of 24 to 42 hours, whereas for skin corrosion the exposure duration is shorter and viability is measured directly after exposure. A viability threshold enables classification into corrosive or non-corrosive [24] and irritant [25]. During skin irritation reactions, injured cells express inflammatory mediators as Tumor Necrosis Factor ␣ (TNF␣) and Interleukin (Il) 1, 6 and 8 [26]. These inflammatory markers are not currently assessed in OECD TG and illustrate the limitation of the method by preventing the mild irritant classification. For skin corrosion, other validated methods exist to classify substance as corrosive or non-corrosive: transcutaneous electrical resistance (TG 430 [27]) and membrane sealing testing (TG 435 [28]). In the first, the decrease of transepidermal electric resistance can be correlated with a loss of normal stratum corneum integrity and barrier function, which is one of the major effects of skin corrosion. The guideline is developed for application on rat skin discs but could be easily applicable to RHE. For the second, there are macromolecular biological synthetic membranes such as Corrositex® which are used for membrane sealing testing. The assay is based on the detection of damage to the membrane caused by exposure to the test substance. The corrosive potential is related to the time between the application of the test substance on the membrane and the penetration of the membrane. The main issue of this assay is that aqueous substances with a pH in the range of 4.5 to 8.5 rarely qualify for assay and thus cannot be assessed.

Ocular irritation/corrosion Eye irritation is defined as the apparition of damage following test substance application to the anterior surface of the eye, reversible after 21 days (OCDE TG 405 [29]). As with skin corrosion, eye corrosion is associated with irreversible damage affecting vision. Before the animal testing ban, the most widely used method was the rabbit Draize test [30]. Another possibility is the use of ex vivo models, which are ethically acceptable as they are obtained after slaughter of the animal and particularly because the animals are not killed for research purposes but for food supply. The gold standard method to assess ocular irritation is Bovine Corneal Opacity and

5

Permeability (BCOP). This assay is used to identify ocular corrosives, severe irritants and since 2013, chemicals that do not require classification for eye irritation or serious eye damage (OECD TG 437 [31]). Induced opacity, quantified by light transmission measurement, and increased permeability related to a sodium fluorescein transepithelial assay are the two endpoints assessed after chemical exposure. Recent improvements have been shown [32] in measuring light transmission with the development of a new laser lightbased opacitometer [33], which increases the sensitivity in detecting small changes in corneal opacity. In vitro, two and three-dimensional systems exist. Short time exposure is a procedure based on a 5 minutes treatment of a rabbit corneal cell line monolayer (SIRC cells) and enables classification into irritant, moderately irritant and non-irritant with correlation of viability assessment. An intra-laboratory study of 109 chemicals [34] shown high reproducibility and an excellent predictability compared with animal based testing (Draize). A limitation is that these methods can only be used to test chemicals, which are soluble in media. Other promising new validated models are threedimensional reconstructed in vitro cornea models. The EpiOcularTM model is a human reconstructed corneal epithelium; cell viability related to chemical exposure can be correlated with identification of non-irritant and irritant substances. This test is not able to predict and distinguish mild irritant and non-irritant chemicals, providing false negative responses [35]. Takezawa et al. [36] developed a reconstructed corneal model grown on a collagen vitrigel membrane, for which an irritancy score is correlated with a transpithelial electrical resistance (TEER) kinetic profile. According to GHS classification, 30 test chemicals were categorised into irritant and non-irritant classes with a perfect sensitivity, 75% of specificity and 90% of accuracy. Three chemicals initially classed non-irritant by GHS showed disruptions of tight junctions such as shown by irritant chemicals and are predicted positive to the test. This false positive could be interesting to elucidate by adding a mild irritant class. Other models exist such as fluorescein assay (OECD TG 460 [37]), which allows the classification of substances as corrosive or strong irritant to eyes. The principle of the assay is the measurement of fluorescein diffusion through a monolayer of dog kidney cell (MDCK) growth on an inert insert, the amount of damage being proportional to fluorescein diffusion. It is only applicable to hydrosoluble substances and does not allow identification of products that would be classed as mildly irritant.

Skin phototoxicity Molecules that are able to penetrate into the skin may react and induce irritation (photo-irritation) and/or allergy (photo-allergy) when exposed to appropriate UV or visible light photons, which can cross the stratum corneum. In some cases, it has been shown that some photo-irritants are able to enhance UV induced carcinogenesis. It is known that if a substance has an absorption wave length similar to the solar spectrum, phototoxic risks can be expected. Concerning alternative methods to assess phototoxicity, there is one which is widely used since its OECD validation

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002

+Model PHARMA-413; No. of Pages 14

ARTICLE IN PRESS

6

P. Quantin et al.

in 2004 (TG 432, [38]); the 3T3-NRU in vitro phototoxicity test. In this test, phototoxicity is related and correlated to cytotoxicity of an immortalised mouse fibroblast cell line (Balb/c 3T3), measured by neutral red dye uptake, after treatment with the test substance in the presence of simulated solar light (UVA light only). Phototoxicity induces an alteration of cell membrane and thus decreases uptake and binding of neutral red compared to an untreated control and a non-light-exposed control. Dose-response curves are then obtained and compared (UV exposed vs. Non-UV exposed) and are used to measure phototoxicity prediction factors, which enable classification of the substance. Those factors, the Photo Irritancy factor and the Mean Photo Effect, are compared and used to construct statistical prediction models in Peters et al. [39]. The main limit of the method is its sole applicability to aqueous soluble substances such as many in vitro cell culture methods. Current developments and perspectives are focused on the use of 3D-reconstructed human epidermis. These models reproduce the target in vivo organ unlike 3T3 which are fibroblast cell lines and therefore allow testing of a wider panel of substances with a high pH. Finally, the use of 3D tissue allows histological observation and comparison with controls. Others methods including yeast growth inhibition phototoxicity assay [40], red blood cell photo hemolysis assay and a method based on reactive oxygen species and photo stability are in development, according to International Cooperation on Alternative Test Methods but the literature is very poor.

cells expose the hapten-protein immunogenic complex to its membrane, which is recognised by memory T-cells present in the dermis. Finally, T-cells trigger an inflammatory response, which involves various pathways and cell types such as keratinocytes, monocytes and macrophages. Over the last decade, numerous in vitro tests have been developed to replace the most widely used in vivo method, the Local Lymph Node Assay (LLNA) [42], based on the observation of lymphocyte proliferation responses in draining lymph node of mice. Two tests are currently under validation by EURL-ECVAM: MUSST (Myeloid U937 Skin Sensitization Test), h-CLAT (human Cell Line Activation Test). Two others tests have just been validated in early 2015: DPRA (Direct Peptide Reactivity Assay) and KeratinosensTM . Fig. 2 represents the three major steps of the sensitisation mechanism and the different tests developed to mimic them. To assess relevance of these tests, prediction results are usually compared with those from in vivo LLNA, which is considered as the gold standard method. To this effect, three main criteria are taken into account: • the sensitivity which is the proportion of sensitiser correctly classified by the test; • the specificity which is the proportion of non-sensitiser correctly classified by the test; • the accuracy which is the proportion of correct predictions compared to the total number of predictions.

Skin sensitisation

Correlation between epidermis protein reactivity and skin sensitisation is well known and the postulate of this test is that if a substance has the capacity to bind proteins, then it can potentially act as a contact allergen. This binding is carried out between an electrophilic chemical and nucleophilic structures (sulfur and azote rich amino acid) as lysin and cystein [43]. Cystein is the major amino acid implicated in hapten binding whereas lysin is more involved in pulmonary allergic mechanisms. After a decade of research and development [44,45], correlation between chemical reactivity of target nucleophilic amino acid and sensitisation potential has been demonstrated and DPRA procedure has been developed and implemented. This test consists in an incubation of the test substance with peptides over 24 hours at room temperature followed by High Performance Liquid Chromatography analysis to quantify the percentage of remaining free peptide. If 10% or more of the peptide is not recovered, the substance can be classified as a sensitiser. Gerberrick et al. [46] have assessed this method comparing prediction results of 81 chemicals with LLNA and found 88% sensitivity, 90% specificity and 89% accuracy. A draft TG on DPRA has just been published by OECD [47].

Physiology and mechanism Chemical exposition leads to cutaneous reactions, which often involve immunological toxicological events such as Allergic Contact Dermatitis (ACD) and can be defined by an increased sensitivity after a second contact with the substance. Briefly, ACD acts in two separated steps: induction or sensitisation and elicitation or challenge. ACD is a systemic pathology classified as a repeated dose toxicity consequence but the sensitisation phase can be seen as a topical local effect. That is why skin sensitisation is introduced in this part. The first step of the mechanism relies on the capability of a hapten to reach the epidermis and to bind to skin proteins to form an immunogenic complex. Cutaneous metabolism has the capacity to biotransform prohaptens, which are not able to bind skin proteins into competent haptens [41]. These haptenprotein complexes can then be recognized and internalised by immune epidermis cells, Langerhans cells, which mature while migrating to lymph nodes. Then, the mature Langerhans cells present the immunogenic complex exhibited on the membrane to the T-cell receptor (TCR) of a naïve CD4+ T-cells. This adaptive response is accompanied by an innate response characterised by keratinocyte pathway activation of cellular defense (antioxidant and detoxifying enzymes expression) and inflammatory (inflammatory mediators secretion) responses. This sensitisation step concludes with stimulation and proliferation of memory T-cells specific to this hapten. Unlike sensitisation, the elicitation step is associated with clinical sign expression of ACD. When a second exposure to the same hapten occurs, Langerhans

The prediction capabilities of the four tests are described in Table 1.

DPRA

KeratinosensTM Keratinocyte cellular responses to sensitising chemicals can be assessed by quantifying changes in gene expression. The activation of Nrf2-ARE pathway (Fig. 3) arises in the sensitisation phase [48] and in vitro tests have been developed to evaluate this dependent gene expression. Integration of a luciferase reporter gene under the control of an antioxidant-element response (ARE) promoter enables the

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002

+Model

ARTICLE IN PRESS

PHARMA-413; No. of Pages 14

Perspectives and strategies of alternative methods used in risk assessment

Skin protein

Hapten

1

7

HAPTEN-PROTEIN BINDING

DPRA Endpoint: Lysin & Cystein depletion

Keratinocytes

Epidermis

Langerhans cell

2

KERATINOCYTES ACTIVATION PATHWAYS

KERATINOSENSTM Endpoint: ARE /Nrf2 induction

IL; TNF-α…

DENDRITIC CELLS MATURATION

Dermis

Migration to Lymph Node

3

h-CLAT MUSST Endpoint: CD 86 / CD 54 expression

Naive T-Cell

Memory T-cell proliferation Lymph Node

Figure 2. Mechanism of skin sensitisation. This recapitulative schema represents the three principal mechanisms involved during skin sensitisation. Each mechanism is associated with an endpoint, which can be measured by a respective assessment method. Mécanisme de la sensibilisation cutanée. Ce schéma récapitulatif représente les trois principaux mécanismes impliqués durant la sensibilisation cutanée. Chaque mécanisme est associé à un critère d’évaluation qui peut être mesuré par une méthode respective.

pathway activation in keratinocytes and thus determines another endpoint in the general mechanism of sensitisation. Briefly HaCaT keratinocyte cell lines transfected with a plasmid which contains luciferase genes under ARE control are exposed to the test substance for 48 hours. Gene induction is then measured by luminescence at non-cytotoxic doses and correlated with sensitising capacities of the substance. A substance is predicted to be a sensitiser if it induces 50% luciferase activity compared to an untreated control. A minority of sensitisers is unable to activate the ARE pathway and could trigger another toxicity pathway [49]. After an inter-laboratory study [50,51], this test seems to be reproducible and highly predictive. With 102 chemicals tested, sensitivity is 81.4%, specificity 86.6% and accuracy 83% compared with LLNA data. Keratinosens is validated by EURL-ECVAM and OECD has just published a guideline recently [52]. Another similar method, LuSens, is being developed in parallel [53].

MUSST/h-CLAT These tests are based on the observation of phenotypic changes of dendritic cells occurring during maturation of Langherans cells. Originally, scientists wanted to assess this endpoint directly on human dendritic cells, however numerous drawbacks prevented this: cost, preparation, complexity in obtaining them and the inherent inter-individual variability. Many studies have shown that human myeloid cell lines U937 and THP1 have dendritic cell close phenotype. For example, THP1 cells express CD54 and CD86 markers when exposed to a sensitiser like dendritic cells in vivo response to sensitiser [54,55]. CD54 and CD86 are identified as a cutaneous sensitisation markers and its expression can vary with a dose/response profile. However, simultaneous increase of cytotoxicity due to contact with a chemical sensitiser (apoptosis) or irritant (necrosis) that could interfere was assessed and discussed by Ade et al. [56]. They finally concluded that these two factors are independent.

Table 1 Overview of evaluation criteria for each alternative method for sensitization compared to LLNA. Vue d’ensemble des critères d’évaluation pour chaque méthode alternative pour la sensibilisation comparée au LLNA. Method

Sensitivity (%)

Specificity (%)

Accuracy (%)

Total number of chemicals tested

DPRA Keratinosens MUSST h-CLAT

88 81.4 78 88

90 86.6 75 75

89 83 77 85

81 102 83 117

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002

+Model PHARMA-413; No. of Pages 14

ARTICLE IN PRESS

8

P. Quantin et al.

Nrf2

Keap 1

Nrf2 Degradation

SENSITIZER

+ Nrf2

Cytoplasm

Nrf2

ARE

Nucleus

Antioxidant and detoxifying enzymes expression

CELLULAR DEFENSE

Figure 3. Schematic of Nrf2/ARE activated pathways in keratinocytes after sensitiser exposure. Sensitisers are capable of activating a cellular defense response. Under normal conditions, the regulatory protein Keap1 binds Nrf2 in cytoplasm and targets it to its degradation. Exposure to a reactive hapten (electrophilic) can induce covalent modifications on Keap1 and enable the release of Nrf2 in the cytoplasm. Nrf2 is a nuclear transcription factor that, once free, can migrate to the nucleus, bind the ARE promotor and induce the expression of cytoprotective proteins. Représentation de la voie de signalisation Nrf2/ARE activée dans les kératinocytes après une exposition à un agent sensibilisant. Les sensibilisants sont capables d’activer une réponse de défense cellulaire. En conditions normales, la protéine régulatrice Keap1 se lie à Nrf2 dans le cytoplasme où il est dégradé. L’exposition à un haptène réactif (électrophile) peut induire des modifications covalentes sur Keap1 et permettre la libération de Nrf2 dans le cytoplasme. Nrf2 est un facteur de transcription nucléaire qui, une fois libre, peut migrer jusqu’au noyau, se fixer au promoteur ARE et induire l’expression de protéines cytoprotectrices.

MUSST The principle of this method, based on U937 cell line reactivity, is to measure CD86 overexpression of cell membrane after exposure to the test chemical at non-cytotoxic doses [57,58]. CD86 expression is quantified by flow cytometry after incubation with an anti-CD86 antibody. A substance is predicted to be a sensitiser if it induces a 50% overexpression compared to an untreated control in two independent experiences. With 83 substances tested, authors ensure 78% sensitivity, 75% specificity and 77% accuracy compared to LLNA data [59]. EURL-ECAM is currently studying the protocol for validation in the near future and reproducibility has been judged robust by Cosmetics Europe.

h-CLAT This method differs from MUSST by the cell line used (THP1, monocytes leukemia cell line) and by the measurement of one more marker: CD54 [60]. More substances have been evaluated than for MUSST, 117, and compared with LLNA sensitivity is 88%, specificity 75% and accuracy 85% [61]. Global predictive capacity is correct but low solubility chemicals seem to be predicted as false negatives and chemicals predicted as weak sensitisers according

to LLNA are not detected by this method as a sensitiser.

Other methods for skin sensitisation assessment Many other prospective methods are currently being developed and are based principally on the observation of the expression of a marker after chemical exposure. For example, in the Genomic Allergen Detection assay developed by Johansson et al. [62], sensitisation properties of a chemical are determined by its capacity to activate cell dendritic markers on the myelomonocyte cell line MUTZ-3. Here, they focus on gene expression change using real time polymerisation chain reaction. A very high accuracy has been shown (> 95%) on 38 chemicals but other validation studies are needed [61]. SENS-IS® [63] is another method based on the observation of a genomic sensitising signature; after test substance exposure on reconstructed human epidermis, gene expression of 62 biomarkers is examined and this enables classification on a potency scale. The method is currently under validation by EURL-ECVAM. In addition to the advantage of potency prediction (63% of concordance with LLNA predictions on a 40 chemicals data set), the method is applicable not only to pure chemicals but also to natural product mixtures and finished products.

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002

+Model PHARMA-413; No. of Pages 14

ARTICLE IN PRESS

Perspectives and strategies of alternative methods used in risk assessment With several disadvantages in the use of cell lines (reduced functionality, genomic instability. . .), the possibility of using peripheral blood derived cells is back under discussion. These cells have the physiological and metabolic functional capacities of dendritic cells with a low interindividual variability in CD80 expression. Many other biomarkers of sensitisation have been identified and may be interesting to assess: CD40, CD83, HLA-DR, TNF-␣. . . However, thresholds for relevant assessment need to be validated.

Integrated testing strategies Due to the complexity of the sensitisation process and the immune implication of different cell types and mechanisms, it is impossible to replace LLNA with one single in vitro test. For a relevant risk assessment, a combination of different tests targeting the main steps of the general mechanism is now well accepted. Methods integrating, compiling and weighing data from the various tests need to be developed to reproduce the in vivo situation [64,65]. As a relevant example, in various publications, Natsch et al. [66,67] tried to develop integrated strategies to fully replace animal based methods and optimise the number of studies. Two types of integrated methods are studied: majority voting testing strategies and tiered testing strategies. These strategies have been compared in Natsch et al. [67] in their capacities to identify skin sensitising chemicals. They integrated data generated for 44 chemicals for various endpoints: Quantitative Structure Activity Relationship reactivity prediction; peptide reactivity (DPRA), dendritic cell maturation (h-CLAT) and keratinocyte responses to sensitiser (KeratinoSensTM , gene signature, IL-8 expression profile). In the majority voting strategy, methods are used to test the substance and the majority response gives the overall result. If the majority of tests give a positive result, then the substance is considered as a sensitiser. Tiered strategies are described as a decision tree where a combination of tests is product dependent as the sequence of test depends on the result obtained. Tiered strategies are based on more elaborate statistical analysis (e.g. Bayesian approaches). High accuracy has been shown for the two methods (about 96%) in their ability to predict skin sensitising potential compared to LLNA. A third type of strategy exists; test battery strategy, similar to majority voting strategy, enables generation of a global predicting score after substance testing with all methods available. A very important issue related to the integrated methods available is their inability to predict sensitising potency, which is possible with animal, based testing [68]. However, studying the distribution of dose-response of the various alternative tests, Natsch et al. [69] has shown that potency information compared with LLNA classification could be obtained from this strategy. To date, correlation between distribution data in an individually alternative model and sensitiser potency can be shown but not when included in integrated testing strategies. The next step is to adapt integrated strategies so that they provide information about sensitiser potency.

9

What needs to be improved? Here are presented and discussed information on various topics which have to be improved in the general risk assessment of personal care.

Toxicokinetic Bioavailability A relevant risk assessment requires robust data on exposure thus it relies on two main factors: skin absorption and skin metabolism. As a result, a bioavailability factor could be established that can be used to extrapolate potential systemic exposure. The concept of bioavailability has been developed for drugs and is defined as the rate and extent to which the active ingredient or active moiety is absorbed from a drug product and becomes available at the site of action. Due to the non-identification of the site of action, bioavailability is redefined as the rate and extent to which the active ingredient or active moiety reaches general blood circulation. Even if personal care ingredients are not active molecules and not intended to reach systemic circulation, they can interact with internal tissues such as the dermis and therefore reach systemic circulation. In the context of the use of alternative methods, bioavailability of in vitro systems should be assessed and redefinition of this parameter, especially for in vitro risk assessment, is recommended. Reconstructed human epidermis models are becoming the main material on which personal care products are tested and bioavailability in those in vitro systems can be simply described as the amount of substance capable of crossing the layers of the tissue and reach the receiving medium.

Skin metabolism As detailed in ‘‘what is available’’, numerous tools exist to assess percutaneous absorption. However, for skin metabolism much work must be done because no alternative method has been developed or accepted to assess this parameter. Xenobiotic skin metabolism is an enzymatic based mechanism whose function is to biotransform exogenous molecules to facilitate their elimination. The aim of skin metabolism is to provide protection, but in some cases it can lead to skin diseases such as contact dermatitis caused by metabolites [70]. Proof of a skin biotransformation function is accepted by scientists but the regulatory framework make its assessment very difficult [71]. The current challenge is to characterise in vitro reconstructed human epidermis models qualitatively and quantitatively for their xenobiotic metabolic enzymes (XME) [72]. A comparison of different models, available in literature [73—76], is essential to try to correlate the data obtained from these in vitro models with in vivo data. Examination of the in vitro models of skin metabolism must help to meet various objectives during risk assessment of a personal care product: define kinetic parameters of an in vitro system by estimating clearance, evaluate important metabolic pathways (such as attempting to take into account genetic polymorphism), assess drug interactions and identify potentially active or reactive metabolites

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002

+Model PHARMA-413; No. of Pages 14

ARTICLE IN PRESS

10 (some metabolites may be present at lower levels than the parent molecule). A recent review [77] summarised various skin XME expression and activity in vivo and human skin models. After having collected and classified all the data, several gaps, which prevented comparison between studies were evidenced. Principal issues concern experimental conditions with a lack of reference points and in most cases absence of limit of detection and quantification. Similarly, critical endpoints for animal free metabolism assessments and strategic endpoints have been reported [78]. Many models appear to be not always fully representative of the in vivo specific activities. Such considerations could be explained by environmental and experimental condition modulations, which could elucidate the differences in XME expression. For example, origin of the culture medium [79] (which can be animal or vegetal) or the type of RHE [80] with only epidermis or ‘‘full thickness’’ composed of epidermis and dermis containing fibroblast capable of secreting several factors (mainly growth factors), origin of the cells (prepuce, abdomen. . .) or age of the donor. Batch by batch variability must also be considered [81]. During metabolite extraction, experimental conditions need to be carefully monitored, to avoid interference between medium and metabolite identification. As for XME expression modulation, it is essential to measure XME activities to characterise metabolism of the skin model for its use in cosmetic risk assessment. In this context, it should be noted that combination of molecules in finished products might lead to unexpected behaviour and effects such as induction, inhibition and interaction [82]. For instance, many vegetal extracts used in personal care contain enzymatic inhibitors capable of decreasing metabolism capacities; a typical inducer such as Phenobarbital has strong effects on human keratinocytes growth rate (unpublished data). Another important in vitro interacting parameter to consider is protein binding [83] because it may influence free chemical fraction and hence metabolite rate especially if there are differences in protein composition between in vitro and in vivo media. Finally, clearance is one of the most critical kinetic parameters to determine metabolising capacities of the system (in vitro intrinsic clearance). In vitro systems generally underestimate intrinsic clearance due to non-specific binding [84].

Target population Biophysical properties of skin vary in many parameters such as age, gender, body location [85] but also ethnical origin. It is very important to take these variations into consideration when a molecule is evaluated as they can interfere with many mechanisms and change behaviour. As a specific example, baby skin is a current issue. Personal care products are not only used in adult personal care but also for personal care for children. This market is expanding rapidly and requires a specific approach [86]. Recent publications have reevaluated the old notion that skin is fully matured at birth and have shown that baby skin differs in structure, function and composition from that of

P. Quantin et al. adults [87]. The most important differences between adult and child skin concern permeability, hydration and acidity parameters. Baby skin continues to develop in the two first years of life. Stratum corneum and global epidermis are thinner than in an adult [88], which could cause higher permeability. Baby skin also has a basic pH compared with adults, which can contribute to the apparition of inflammatory dermatitis [89]. Finally, low hydration capacities and insufficient hydro-lipidic film make baby skin even more sensitive. Therefore, the risk assessment of raw materials and finished products for babies cannot be established from human adult data. Adapted models should be used. In order to respect the morphology and properties of baby skin, especially the stratum corneum, an immature human reconstructed epidermis has been developed at Thor Personal Care in the In vitro Toxicology laboratory. Only characterised in an anatomical way, it is used primarily to assess skin irritation for a range of baby products. This model is derived from a previously in house developed human epidermis model (called VitroDerm). Another example concerns the genetic polymorphism of xenobiotic metabolising enzymes in ethnic origin. It has been cleared that there are interethnic differences in the metabolism of xenobiotics for many years [90] and it must be pointed out that those variations are correlated with potentially variable interethnic toxic responses. In this case, it is appropriate to mention a new reconstructed human epidermis created using keratinocytes of Asian origins [91]. This model has been developed to assess irritancy but it would be relevant to study its metabolism capacity for a potential use in target population risk assessment.

In silico approaches In silico is a rather general term that encompasses all methods performed with a computer. In silico toxicology, developed to characterise and predict toxic outcomes in humans and the environment, differs from traditional toxicology in many respects but perhaps the most important is in the scale [92]. Models mostly used are SAR (Structure Activity Relationship) and QSAR (Quantitative Structure Activity Relationship) and can be applied to predict biological mechanisms. They are constructed by bioinformatical experts based on the hypothesis that similar compounds should have similar biological activities. Those models are built with data obtained from in vitro, in vivo, and clinical studies in order to predict different types of toxic outcomes. SAR and QSAR differ in the data they generate: SAR can generate qualitative information relating to toxicological data compiled in a database with a structural alert whereas QSAR may additionally add quantitative data calculated from a mathematical equation constructed by correlation curves and regression lines representing numerical descriptors derived from structural information. Descriptors are numerical representations of molecular properties as physicochemical (e.g. logP), topological (e.g. ionisation state) and surface properties. These models can be opposed as expert rule based model (SAR) versus statistical (QSAR) [93]. Various toxicological effects can be predicted: mutagenicity, eye or skin irritation, skin sensitisation and ecotoxicity. Other interesting toxicological parameters for personal care

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002

+Model PHARMA-413; No. of Pages 14

ARTICLE IN PRESS

Perspectives and strategies of alternative methods used in risk assessment risk assessment can be anticipated by in silico modeling; this is the case for metabolism. Already highlighted is the importance of assessing skin metabolism of topical applied substances. In addition to in vitro models described previously, metabolism simulators exist and may mimic the reactivity of a chemical substrate against an enzyme pool. Metabolism in the liver (the principal organ of xenobiotic metabolism) and in the skin are simulated. To assess the relevance of those models, metabolism of a test chemical has been simulated by a computational method and compared with in vitro generation of metabolism of the same test molecule on a reconstructed human epidermis [78]. Results showed an over prediction of metabolism liver simulation compared with the in vitro study; this was of course expected due to its stronger capacity for biotransformation. The skin simulator did not generated any metabolites. This work shows the possibility of using in silico techniques but these are in need of improvement. More generally, numerous limitations can be identified [94]: quality and transparency of training set experimental data, demonstration and explanation of the model, confusing descriptors, indefinite domain of application. However, their use can be very interesting since data are quickly generated with a very low cost, the improvement and optimisation are constant with a high reproducibility. All these advantages demonstrate the usefulness of in silico modeling; such methods are becoming increasingly integrated in risk assessment strategies and could or should even become the leading methods for direct assessment of human hazard risk.

Ecotoxicology Personal care ingredients, included in finished products, can potentially reach the environment after their use, mainly after washing, at doses capable of induce harmful effects. They can be persistent, bioactive and exhibit accumulation potential and at a minimum cause endocrine disruption [95,96]. Established in 1992, the European Eco-label is the only official European label concerning ecotoxicology used in all the Member States of the European Union. This label confirms monitoring during the life cycle of the product from raw material, fabrication, distribution and use up to recycling and disposal after use. Eco-label certification must ensure various criteria of the substance of interest: prevent water pollution by limiting the quantity of potentially harmful ingredients and the total toxic load of the product; limit waste by reducing the amount of packaging and limit or prevent environmental risks related to the use of hazardous substances. For this purpose, personal care manufacturers must focus on biodegradability and ecotoxicology of their products. OECD guidelines recommend to assess these different endpoints: ready biodegradability [97], daphnia acute immobilisation test [98], fresh water algae and cyanobacteria growth inhibition test [99] and fish acute toxicity [100]. However, the ban on animal testing makes some of these methods unrealisable. For acute toxicity assessment of fish and in order to reduce animal testing, OECD TG 236 [101], performed using an embryo can replace TG 203. Indeed, between the egg stage and the embryonic stage (when able

11

to feed by mouth) fish are not considered as vertebrate organisms. There is still a way of generating the required ecotoxicological data using QSAR models, as for example EPIsuiteTM developed by the Environment Protection Agency. Knowing the physic-chemical properties of the target substance, the system is able to estimate and predict environmental fate. Such software can be very helpful when used as a screening tool in the absence of experimental data.

Conclusion The aim of this review was to study the multiple new approaches for dermal toxicity testing of personal care products. A number of laboratories have changed their methods of working by developing alternative testing due to the recent redefinition of the regulatory framework. For most of the common toxicological endpoints occurring on the skin, there are alternative methods developed to mimic and replace in vivo tests. However, these methods are mostly not yet recognised by EURL-ECVAM and OECD but should be in the near future. This paper highlights the significant interest of the scientific community in assessing the sensitising potential of topical applied chemicals and development of such methods is a real challenge. Although DPRA, MUSST, h-CLAT, KeratinoSensTM are on track to be fully validated, some false positives and false negative predictions decrease global accuracy of the tests. Other tests are being developed, at earlier stages, with varying degrees of success. Therefore, there is strong interest in the integration of other parameters, which are absorption, metabolism and QSAR data. Indeed, without transcutaneous passage, skin sensitisation cannot occur. Conversely, skin biotransformation of test substances can lead to metabolite generation, which would have to be tested instead of the parent substances. The chief difficulty with non-animal testing is to reproduce all events, which are involved in biological effects on an organism. Testing methods available today have a promising future in their combination to enable quantitative in vitro prediction of in vivo toxicity. This review also forces a step back for an overall view of this topic of general interest. It allows us to show all the difficulties in the development of new alternative methods, which have to be as close as possible to what really happens in human organisms. Data generated on alternative methods are compared to data generated on animals whereas the most logical process would be to compare to human data. However, human data are very rare due to ethical issues and we are forced to use animal comparisons. This creates a problem of interpretation because this comparison reduces the relevance of the results. Numerous false positives and false negatives appear when comparisons are made between animal and human and thus decrease relevance of in vitro assessment. Moreover, it is more and more accepted by toxicologist experts that animal models do not reflect fully human physiology and especially the kinetic and dynamic phenomena within the body, which leads to uncertainties in extrapolating animal data to human. ‘‘Traditional’’ toxicology is

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002

+Model PHARMA-413; No. of Pages 14

ARTICLE IN PRESS

12 based on animal experimentation and the observation of an endpoint. Today, approaches are based on the mechanisms involved in a toxic reaction. This approach demands the combination of in vitro, in silico and in chemico methods and helps identification of Adverse Outcome Pathways, one of the most challenging step in the evolution of non-animal based toxicology in the future.

Disclosure of interest THOR Collaborators.

References [1] Scientific Committee on Consumer Safety. The SCCS’S Notes of Guidance for the Testing of Cosmetic Substances and their Safety Evaluation 8th revision; 2012. [2] Adler S, Basketter D, Creton S, Pelkonen O, van Benthem J, Zuang V, et al. Alternative (non-animal) methods for cosmetics testing: current status and future prospects — 2010. Arch Toxicol 2011;85(5):367—485. [3] Elias PM. The skin barrier as an innate immune element. Semin Immunopathol 2007;29(1):3—14. [4] Madison KC. Barrier function of the skin: « La Raison d’Être » of the epidermis. J Invest Dermatol 2003;121(2):231—41. [5] Elias PM. Epidermal lipids, barrier function, and desquamation. J Invest Dermatol 1983;80(1s):44s—9s. [6] Vickers CH. Existence of reservoir in the stratum corneum: experimental proof. Arch Dermatol 1963;88(1):20—3. [7] Warner RR, Stone KJ, Boissy YL. Hydration disrupts human stratum corneum ultrastructure. J Invest Dermatol 2003;120(2):275—84. [8] Bartosova L, Bajgar J. Transdermal drug delivery in vitro using diffusion cells. Curr Med Chem 2012;19(27):4671—7. [9] OE.D. Test No. 428: Skin Absorption: In vitro Method [Internet]. Paris: Organisation for Economic Co-operation and Development; 2004. [10] MacGregor JT, Collins JM, Sugiyama Y, Tyson CA, Dean J, Smith L, et al. In vitro human tissue models in risk assessment: report of a consensus-building workshop. Toxicol Sci 2001;59(1):17—36. [11] Lotte C, Patouillet C, Zanini M, Messager A, Roguet R. Permeation and skin absorption: reproducibility of various industrial reconstructed human skin models. Skin Pharmacol Appl Skin Physiol 2002;15(Suppl 1):18—30. [12] Ponec M, Gibbs S, Pilgram G, Boelsma E, Koerten H, Bouwstra J, et al. Barrier function in reconstructed epidermis and its resemblance to native human skin. Skin Pharmacol Appl Skin Physiol 2001;14(Suppl 1):63—71. [13] De Jager M, Groenink W, Bielsa i Guivernau R, Andersson E, Angelova N, Ponec M, et al. A novel in vitro percutaneous penetration model: evaluation of barrier properties with paminobenzoic acid and two of its derivatives. Pharm Res 2006;23(5):951—60. [14] Ottaviani G, Martel S, Carrupt P-A. Parallel artificial membrane permeability assay: a new membrane for the fast prediction of passive human skin permeability. J Med Chem 2006;49(13):394—454. [15] Feldstein MM, Raigorodskii IM, Iordanskii AL, Hadgraft J. Modeling of percutaneous drug transport in vitro using skin-imitating Carbosil membrane. J Control Release 1998;52(1—2):25—40. [16] Ng S-F, Rouse J, Sanderson D, Eccleston G. A Comparative study of transmembrane diffusion and permeation of ibuprofen across synthetic membranes using Franz diffusion cells. Pharmaceutics 2010;2(2):209—23.

P. Quantin et al. [17] Herkenne C, Naik A, Kalia YN, Hadgraft J, Guy RH. Ibuprofen transport into and through skin from topical formulations: in vitro—in vivo comparison. J Invest Dermatol 2007;127(1):135—42. [18] European Food Safety Authority. Guidance on dermal absorption. EFSA J 2012;10(4):2665. [19] Bos JD, Meinardi MM. The 500 Dalton rule for the skin penetration of chemical compounds and drugs. Exp Dermatol 2000;9(3):165—9. [20] Roberts DW, Mekenyan OG, Dimitrov SD, Dimitrova GD. What determines skin sensitization potency-myths, maybes and realities. Part 1. The 500 molecular weight cut-off. Contact Dermatitis 2013;68(1):32—41. [21] Georgel A, Thèse de doctorat en pharmacie Pénétration transcutanée des substances actives. Application en dérmocosmétologie. Nancy: Université Henri Poincaré; 2008. p. 1—189. [22] Francis T. Transdermal Drug Delivery Systems: Revised and Expanded. 2nd Edition Hardback Taylor & Francis; 2002 [Internet]. [23] Stewart CA, Turner RB, Brown MB, Traynor MJ. Classification of Irritant/Non-Irritant and Corrosive Chemicals Using Excised Human Skin In Vitro. Advances in Dermatological Sciences; 2013 [Internet]. [24] OECD. Test No. 431: In vitro Skin Corrosion: Human Skin Model Test [Internet]. Paris: Organisation for Economic Co-operation and Development; 2004. [25] OECD. Test No. 439: In vitro Skin Irritation [Internet]. Paris: Organisation for Economic Co-operation and Development; 2010. [26] Lee HY, Stieger M, Yawalkar N, Kakeda M. Cytokines and chemokines in irritant contact dermatitis. Mediators Inflamm 2013;2013:e916497. [27] OECD. Test No. 430: In vitro Skin Corrosion: Transcutaneous Electrical Resistance Test Method (TER) [Internet]. Paris: Organisation for Economic Co-operation and Development; 2013. [28] OECD. Test No. 435: In vitro Membrane Barrier Test Method for Skin Corrosion [Internet]. Paris: Organisation for Economic Co-operation and Development; 2006. [29] OECD. Test No. 405: Acute Eye Irritation/Corrosion [Internet]. Paris: Organisation for Economic Co-operation and Development; 2012. [30] Draize JH, Woodard G, Calvery HO. Methods for the study of irritation and toxicity of substances applied topically to the skin and mucous membranes. J Pharmacol Exp Ther 1944;82(3):377—90. [31] OECD. Test No. 437: Bovine Corneal Opacity and Permeability Test Method for Identifying (i) Chemicals Inducing Serious Eye Damage and (ii) Chemicals Not Requiring Classification for Eye Irritation or Serious Eye Damage [Internet]. Paris: Organisation for Economic Co-operation and Development; 2013. [32] Verstraelen S, Jacobs A, De Wever B, Vanparys P. Improvement of the Bovine Corneal Opacity and Permeability (BCOP) assay as an in vitro alternative to the Draize rabbit eye irritation test. Toxicol In Vitro 2013;27(4):1298—311. [33] Van Goethem F, Hansen E, Sysmans M, De Smedt A, Vanparys P, Van Gompel J. Development of a new opacitometer for the bovine corneal opacity and permeability (BCOP) assay. Toxicol In Vitro 2010;24(6):1854—61. [34] Takahashi Y, Hayashi K, Abo T, Koike M, Sakaguchi H, Nishiyama N. The Short Time Exposure (STE) test for predicting eye irritation potential: intra-laboratory reproducibility and correspondence to globally harmonized system (GHS) and EU eye irritation classification for 109 chemicals. Toxicol In Vitro 2011;25(7):1425—34. [35] Cotovio J, Grandidier M-H, Lelièvre D, Bremond C, Amsellem C, Maloug S, et al. In vitro assessment of eye irritancy using

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002

+Model PHARMA-413; No. of Pages 14

ARTICLE IN PRESS

Perspectives and strategies of alternative methods used in risk assessment

[36] [37]

[38]

[39]

[40]

[41] [42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

[50]

[51]

[52]

[53]

the Reconstructed Human Corneal Epithelial Skin Ethic HCE model: application to 435 substances from consumer products industry. Toxicol In Vitro 2010;24(2):523—37. Yamaguchi H, Kojima H, Takezawa T. Vitrigel-eye irritancy test method using HCE-T cells. Toxicol Sci 2013;135(2):347—55. OECD. Test No. 460: Fluorescein Leakage Test Method for Identifying Ocular Corrosives and Severe Irritants [Internet]. Paris: Organisation for Economic Co-operation and Development; 2012. OECD. Test No. 432: In vitro 3T3 NRU Phototoxicity Test [Internet]. Paris: Organisation for Economic Co-operation and Development; 2004. Peters B, Holzhütter H-G. In vitro phototoxicity testing: development and validation of a new concentration response analysis software and biostatistical analyses related to the use of various prediction models. Altern Lab Anim 2002;30(4):415—32. Marrot L, Labarussiat A, Perez P, Meunier JR. Use of the yeast Saccharomyces cerevisiae as a pre-screening approach for assessment of chemical-induced phototoxicity. Toxicol In Vitro 2006;20(6):1040—50. Lepoittevin P. La peau : acteur majeur du métabolisme. Prog En Derm-Allergol Lille 2004 2004;10:153. Gerberick GF, Ryan CA, Dearman RJ, Kimber I. Local lymph node assay (LLNA) for detection of sensitization capacity of chemicals. Methods 2007;41(1):54—60. Ahlfors SR, Sterner O, Hansson C. Reactivity of contact allergenic haptens to amino acid residues in a model carrier peptide, and characterization of formed peptide-hapten adducts. Skin Pharmacol Appl Skin Physiol 2003;16(1):59—68. Gerberick GF, Troutman JA, Foertsch LM, Vassallo JD, Quijano M, Dobson RLM, et al. Investigation of peptide reactivity of pro-hapten skin sensitizers using a peroxidase-peroxide oxidation system. Toxicol Sci 2009;112(1):164—74. Gerberick GF, Vassallo JD, Bailey RE, Chaney JG, Morrall SW, Lepoittevin J-P. Development of a peptide reactivity assay for screening contact allergens. Toxicol Sci 2004;81(2):332—43. Gerberick GF, Vassallo JD, Foertsch LM, Price BB, Chaney JG, Lepoittevin J-P. Quantification of chemical peptide reactivity for screening contact allergens: a classification tree model approach. Toxicol Sci 2007;97(2):417—27. OECD. Test No. 442C: In Chemico Skin Sensitisation [Internet]. Paris: Organisation for Economic Co-operation and Development; 2015. El Ali Z, Gerbeix C, Hemon P, Esser PR, Martin SF, Pallardy M, et al. Allergic skin inflammation induced by chemical sensitizers is controlled by the transcription factor Nrf2. Toxicol Sci 2013;134(1):39—48. Emter R, van der Veen JW, Adamson G, Ezendam J, van Loveren H, Natsch A. Gene expression changes induced by skin sensitizers in the KeratinoSensTM cell line: discriminating Nrf2-dependent and Nrf2-independent events. Toxicol In Vitro 2013;27(8):2225—32. Natsch A. The Nrf2-Keap1-ARE toxicity pathway as a cellular sensor for skin sensitizers — functional relevance and a hypothesis on innate reactions to skin sensitizers. Toxicol Sci 2010;113(2):284—92. Andreas N, Caroline B, Leslie F, Frank G, Kimberly N, Allison H, et al. The intra- and inter-laboratory reproducibility and predictivity of the KeratinoSens assay to predict skin sensitizers in vitro: results of a ring-study in five laboratories. Toxicol In Vitro 2011;25(3):733—44. OECD. Test No. 442D: In Vitro Skin Sensitisation [Internet]. Paris: Organisation for Economic Co-operation and Development; 2015. Ramirez T, Mehling A, Kolle SN, Wruck CJ, Teubner W, Eltze T, et al. LuSens: a keratinocyte based ARE reporter gene assay

[54]

[55]

[56]

[57]

[58]

[59]

[60]

[61]

[62]

[63]

[64]

[65]

[66]

[67]

[68]

[69]

[70]

13

for use in integrated testing strategies for skin sensitization hazard identification. Toxicol In Vitro 2014;28(8):1482—97. Ashikaga T, Yoshida Y, Hirota M, Yoneyama K, Itagaki H, Sakaguchi H, et al. Development of an in vitro skin sensitization test using human cell lines: the human Cell Line Activation Test (h-CLAT): I. Optimization of the h-CLAT protocol. Toxicol In Vitro 2006;20(5):767—73. Sakaguchi H, Ashikaga T, Miyazawa M, Yoshida Y, Ito Y, Yoneyama K, et al. Development of an in vitro skin sensitization test using human cell lines; human Cell Line Activation Test (h-CLAT). II. An inter-laboratory study of the h-CLAT. Toxicol In Vitro 2006;20(5):774—84. Ade N, Martinozzi-Teissier S, Pallardy M, Rousset F. Activation of U937 cells by contact sensitizers: CD86 expression is independent of apoptosis. J Immunotoxicol 2006;3(4):189—97. Python F, Goebel C, Aeby P. Assessment of the U937 cell line for the detection of contact allergens. Toxicol Appl Pharmacol 2007;220(2):113—24. Yoshida Y, Sakaguchi H, Ito Y, Okuda M, Suzuki H. Evaluation of the skin sensitization potential of chemicals using expression of co-stimulatory molecules, CD54 and CD86, on the naive THP-1 cell line. Toxicol In Vitro 2003;17(2):221—8. Ovigne JM, Piroid C, Verda D, Rousset F, Martinozzi-Teissier S, Meunier JR. Rhemyeloid U937 skin sensitization test (MUSST) for the prediction of skin sensitization potential; 2011. Available at: http://www.episkin.com/iso album/ the myeloid u937 skin sensitization.pdf Ashikaga T, Sakaguchi H, Sono S, Kosaka N, Ishikawa M, Nukada Y, et al. A comparative evaluation of in vitro skin sensitisation tests: the human cell-line activation test (h-CLAT) versus the local lymph node assay (LLNA). Altern Lab Anim 2010;38(4):275—84. Albrekt A-S, Johansson H, Börje A, Borrebaeck C, Lindstedt M. Skin sensitizers differentially regulate signaling pathways in MUTZ-3 cells in relation to their individual potency. BMC Pharmacol Toxicol 2014;15:5. Johansson H, Albrekt A-S, Borrebaeck CAK, Lindstedt M. The GARD assay for assessment of chemical skin sensitizers. Toxicol In Vitro 2013;27(3):1163—9. Teissier S, Tourneix F, Del Bufalo A, Gomes C, Groux H, Meunier JR. Evaluation of SENS-IS® , an EPISSKIN based model for identifying chemical sensitizers. USA; San Fransisco: 51th Congress of Society Of Toxicology; 2012. Bauch C, Kolle SN, Fabian E, Pachel C, Ramirez T, Wiench B, et al. Intralaboratory validation of four in vitro assays for the prediction of the skin sensitizing potential of chemicals. Toxicol In Vitro 2011;25(6):1162—8. Jaworska J. Integrating non-animal test information into an adaptive testing strategy — skin sensitization proof of concept case. ALTEX 2011;28(3):211—25. Natsch A, Emter R, Ellis G. Filling the concept with data: integrating data from different in vitro and in silico assays on skin sensitizers to explore the battery approach for animal-free skin sensitization testing. Toxicol Sci 2009;107(1):106—21. Van der Veen JW, Rorije E, Emter R, Natsch A, van Loveren H, Ezendam J. Evaluating the performance of integrated approaches for hazard identification of skin sensitizing chemicals. Regul Toxicol Pharmacol 2014;69(3):371—9. Natsch A, Haupt T. Utility of rat liver S9 fractions to study skin-sensitizing prohaptens in a modified KeratinoSens assay. Toxicol Sci 2013;135(2):356—68. Natsch A, Ryan CA, Foertsch L, Emter R, Jaworska J, Gerberick F, et al. A dataset on 145 chemicals tested in alternative assays for skin sensitization undergoing prevalidation. J Appl Toxicol 2013;33(11):1337—52. Aeby P, Sieber T, Beck H, Gerberick GF, Goebel C. Skin sensitization to p-phenylenediamine: the diverging roles of

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002

+Model PHARMA-413; No. of Pages 14

ARTICLE IN PRESS

14

[71] [72]

[73]

[74]

[75]

[76]

[77]

[78]

[79]

[80]

[81]

[82]

[83]

P. Quantin et al. oxidation and N-acetylation for dendritic cell activation and the immune response. J Invest Dermatol 2009;129(1):99—109. Baron JM, Merk HF. Drug metabolism in the skin. Curr Opin Allergy Clin Immunol 2001;1(4):287—91. Hewitt NJ, Edwards RJ, Fritsche E, Goebel C, Aeby P, Scheel J, et al. Use of human in vitro skin models for accurate and ethical risk assessment: metabolic considerations. Toxicol Sci 2013;133(2):209—17. Hu T, Khambatta ZS, Hayden PJ, Bolmarcich J, Binder RL, Robinson MK, et al. Xenobiotic metabolism gene expression in the EpiDermin vitro3D human epidermis model compared to human skin. Toxicol In Vitro 2010;24(5):1450—63. Jäckh C, Blatz V, Fabian E, Guth K, van Ravenzwaay B, Reisinger K, et al. Characterization of enzyme activities of Cytochrome P450 enzymes, Flavin-dependent monooxygenases, N-acetyltransferases and UDP-glucuronyltransferases in human reconstructed epidermis and full-thickness skin models. Toxicol In Vitro 2011;25(6):1209—14. Götz C, Pfeiffer R, Tigges J, Blatz V, Jäckh C, Freytag EM, et al. Xenobiotic metabolism capacities of human skin in comparison with a 3D epidermis model and keratinocytebased cell culture as in vitro alternatives for chemical testing: activating enzymes (Phase I). Exp Dermatol 2012;21(5): 358—63. Götz C, Pfeiffer R, Tigges J, Ruwiedel K, Hübenthal U, Merk HF, et al. Xenobiotic metabolism capacities of human skin in comparison with a 3D-epidermis model and keratinocytebased cell culture as in vitro alternatives for chemical testing: phase II enzymes. Exp Dermatol 2012;21(5):364—9. Oesch F, Fabian E, Guth K, Landsiedel R. Xenobioticmetabolizing enzymes in the skin of rat, mouse, pig, guinea pig, man, and in human skin models. Arch Toxicol 2014;88(12):2135—90. Quantin P, Catoire S, Ficheux H. In silico-in vitro correlation in metabolism studies: points to consider. Scotland; Edimburg: Eurotox, 50th Congress of the Europeen Societies of Toxicology; 2014. Brunner D, Frank J, Appl H, Schöffl H, Pfaller W, Gstraunthaler G. Serum-free cell culture: the serum-free media interactive online database. ALTEX 2010;27(1):53—62. Wiegand C, Hewitt NJ, Merk HF, Reisinger K. Dermal xenobiotic metabolism: a comparison between native human skin, four in vitro skin test systems and a liver system. Skin Pharmacol Physiol 2014;27(5):263—75. Harris IR, Siefken W, Beck-Oldach K, Wittern K-P, Pollet D. NAD(P)H: quinone reductase activity in human epidermal keratinocytes and reconstructed epidermal models. Skin Pharmacol Appl Skin Physiol 2002;15(Suppl 1):68—73. Ahmad N, Mukhtar H. Cytochrome P450: a target for drug development for skin diseases. J Invest Dermatol 2004;123(3):417—25. Deglmann CJ, Ebner T, Ludwig E, Happich S, Schildberg FW, Koebe HG. Protein binding capacity in vitro changes metabolism of substrates and influences the predictability of metabolic pathways in vivo. Toxicol In Vitro 2004;18(6):835—40.

[84] Giuliano C, Jairaj M, Zafiu CM, Laufer R. Direct determination of unbound intrinsic drug clearance in the microsomal stability assay. Drug Metab Dispos 2005;33(9):1319—24. [85] Firooz A, Sadr B, Babakoohi S, Sarraf-Yazdy M, Fanian F, Kazerouni-Timsar A, et al. Variation of biophysical parameters of the skin with age, gender, and body region. Sci World J 2012;2012:e386936. [86] Gao X, Simpson EL. Market trends in baby skin care products and implications for clinical practice. Pediatr Dermatol 2014;31(6):734—8. [87] Stamatas GN, Nikolovski J, Luedtke MA, Kollias N, Wiegand BC. Infant skin microstructure assessed in vivo differs from adult skin in organization and at the cellular level. Pediatr Dermatol 2010;27(2):125—31. [88] Stamatas GN, Nikolovski J, Mack MC, Kollias N. Infant skin physiology and development during the first years of life: a review of recent findings based on in vivo studies. Int J Cosmet Sci 2011;33(1):17—24. [89] Ali SM, Yosipovitch G. Skin pH: from basic science to basic skin care. Acta Derm Venereol 2013;93(3):261—7. [90] Renwick AG. Inter-ethnic differences in xenobiotic metabolism. Environ Toxicol Pharmacol 1996;2(2—3):165—70. [91] Wu Y, Ding C, Li N, Qiu J, Jiang D, Zhong L, et al. Asian reconstructed epidermis model assessment for the in vitro prediction of skin irritation. USA; San Fransisco: 51st Congress of Society Of Toxicology; 2012. [92] Kavlock RJ, Ankley G, Blancato J, Breen M, Conolly R, Dix D, et al. Computational toxicology — a state of the science mini review. Toxicol Sci 2008;103(1):14—27. [93] Simon-Hettich B, Rothfuss A, Steger-Hartmann T. Use of computer-assisted prediction of toxic effects of chemical substances. Toxicology 2006;224(1—2):156—62. [94] Valerio LG. In silico toxicology for the pharmaceutical sciences. Toxicol Appl Pharmacol 2009;241(3):356—70. [95] Caliman FA, Gavrilescu M. Pharmaceuticals, personal care products and endocrine disrupting agents in the environment—a review. CLEAN — Soil Air Water 2009;37(4—5):277—303. [96] Daughton CG, Ternes TA. Pharmaceuticals and personal care products in the environment: agents of subtle change? Environ Health Perspect 1999;107(Suppl 6):907—38. [97] OECD. Test No. 301: Ready Biodegradability [Internet]. Paris: Organisation for Economic Co-operation and Development; 1992. [98] OECD. Test No. 202: Daphnia sp. Acute Immobilisation Test [Internet]. Paris: Organisation for Economic Co-operation and Development; 2004. [99] OECD. Test No. 201: Freshwater Alga and Cyanobacteria, Growth Inhibition Test [Internet]. Paris: Organisation for Economic Co-operation and Development; 2011. [100] OECD. Test No. 203: Fish, Acute Toxicity Test [Internet]. Paris: Organisation for Economic Co-operation and Development; 1992. [101] OECD. Test No. 236: Fish Embryo Acute Toxicity (FET) Test [Internet]. Paris: Organisation for Economic Co-operation and Development; 2013.

Please cite this article in press as: Quantin P, et al. Perspectives and strategies of alternative methods used in the risk assessment of personal care products. Ann Pharm Fr (2015), http://dx.doi.org/10.1016/j.pharma.2015.06.002