An alternative strategy to the use of guinea pigs for the identification of skin sensitization hazard

~

Pergamon

FdChem. Toxic. Vol. 33, No. 12, pp. 1051-1056, 1995 Copyright © 1995 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0278-6915/95 $9.50+ 0.00

0278-6915(95)00073-9

An Alternative Strategy to the Use of Guinea Pigs for the Identification of Skin Sensitization Hazard Et. A . B A S K E T T E R * ,

E. W . S C H O L E S , M . C H A M B E R L A I N and M. D. BARRATT Unilever Environmental Safety Laboratory, Colworth House, Sharnbrook, Bedford MK44 ILQ, UK (Accepted 19 April 1995) A ~ t r a c t - - F o r over half a century, guinea pig methods have dominated the field of toxicology concerned with the identification of skin sensitizers. Specific protocols, for example the guinea pig maximization test (GPMT), have been pre-eminent in the identification of skin sensitization hazard for regulatory purposes. However, there are increasingly several forces driving change, not least animal use/welfare considerations. In response to this and to address the need for a rapid screen for chemical allergens, an alternative strategy has been developed. In the first instance, a chemical is assessed by a computer-based expert system. This system is constructed from some 50 rules describing the key chemically reactive substructures of known skin sensitizers. The output from the expert system is also evaluated in the light of the understanding of the skin penetration characteristics of the chemical. In this way, and without use of animals, the likelihood that a chemical represents a skin sensitization hazard is assessed based on the two key characteristics of a skin sensitizer: (1) its direct or indirect ability to react with skin protein (i.e. does it contain a structural alert?); and (2) the ability of the chemical to partition into the appropriate epidermal compartment. When the chemical does possess a structural alert and has the capacity to penetrate skin sufficiently, then it may be regarded as a potential skin sensitizer. Subsequent to this screening phase, if necessary the chemical may be asses:~ed in the murine local lymph node assay. This assay is quicker and cheaper than traditional guinea pig assays and importantly is less stressful to the fewer animals that it requires. The assay is well validated and produces objective results which are equivalent to the GPMT in terms of identifying significant skLn sensitization hazard. In this paper, the above strategy is described in more detail, focusing on its relevance to hazard identification and its value in animal welfare terms. It is concluded that the strategy provides an important opportunity for both substantial reduction and refinement of animal use in a manner which will not compromise the existing standard of classification and labelling of skin sensitization hazard in the European Union.

Introduction It is a legal requirement in the E u r o p e a n U n i o n (EU) t h a t new substances are tested for their potential to cause skin sensitization (EC, 1993a). Generally, a substance which is found to possess a significant skin sensitization hazard has to be classified as a d a n g e r o u s substance a n d must be labelled with a St A n d r e w ' s cross and carry the R43 risk phrase ' M a y cause sensitization by skin contact'. F o r over half a century, guinea pig methods have d o m i n a t e d as the models of choice for the *Author for correspondence. Abbreviations: DEREK = Deductive Estimation of Risk from Existing Knowledge; EC = European Commission; ECETOC = European Centre for Ecotoxicology and Toxicology of Chemicals; E U = E u r o p e a n Union; FCA = Freund's complete adjuvant; GPMT = guinea pig maximization test; LLNA = local lymph node assay; OECD = Organisation for Economic Cooperation and Development; QSAR = quantitative structure--activity relationship.

identification of skin sensitization potential. Two particular protocols, the guinea pig maximization test ( G M P T ) ( M a g n u s s o n a n d Kligman, 1970) a n d the Buehler occluded patch test (Buehler, 1965) have been the preferred m e t h o d s (see B o t h a m et al., 1991). Consequently, they are described in detail in b o t h the E u r o p e a n Commission (EC, 1993b) a n d in the Organisation for Economic C o o p e r a t i o n a n d Develo p m e n t ( O E C D ) guidelines for skin sensitization ( O E C D , 1993); b o t h tests require the use of at least 30 animals. The interpretation of the data generated by these two methods in the context o f E U regulations is straightforward. If a positive incidence o f sensitization is /> 30% in the G P M T , or is i> 15% in the Buehler test, then the substance is formally classified as a skin sensitizer. However, there are now a n u m b e r o f reasons to consider a n alternative a p p r o a c h to achieving the aim o f skin sensitization hazard identification. A m a j o r factor driving change is the wish to reduce or indeed eliminate the need for animal testing. First, there is

1051

1052

D.A. Basketter et al.

currently no in vitro test for skin sensitization and although several lines of investigation are being followed, there is little probability of developing an acceptable method in the near future. Secondly, both the guinea pig procedures offer the opportunity for the development of less stressful protocols. In particular, avoidance of the use of Freund's complete adjuvant (FCA) (GPMT only) would be a welcome step forward. Lastly, alternative approaches can provide opportunities for more rapid and cost effective strategies which use the knowledge gained over recent decades to deliver an improved hazard identification process. In this paper, a strategy using a computer based expert system followed by an optional murine test, the local lymph node assay (LLNA), is described in detail and illustrated by example. Particular attention has been paid to the application of the strategy to hazard identification in EU terms and to the value of the strategy in relation to animal welfare.

The expert system (DEREK) The correlation of the protein reactivity of chemicals with their skin sensitization potential is well established (e.g. Dupuis and Benezra, 1982) so that it is now accepted that if a chemical is capable of reacting with a protein either directly or after appropriate (bio)chemical transformation, then it has the potential to be a contact allergen, assuming of course that it can locate in the appropriate epidermal compartment. Knowledge relating chemical structure to toxicity can be programmed into expert systems. An historical database (Cronin and Basketter, 1994) containing results of about 300 GPMTs carried out over a number of years according to a single protocol on defined single substances has been used to derive a set of structural alerts for skin sensitization. The approach used was to group the substances, where possible, according to their most likely mechanism of reaction with skin proteins. Where no mechanism could be clearly identified, structural alerts were derived for groups of chemicals with similar functional groups. This process initially resulted in the production of around 40 structureactivity rules (Barratt et al., 1994b), now increased to about 50. These have been incorporated into the expert system DEREK (an acronym for Deductive Estimation of Risk from Existing Knowledge) (Sanderson and Earnshaw, 1991), which is being developed by LHASA UK at the School of Chemistry, University of Leeds. DEREK embodies both a controlling programme and a chemical rulebase. The chemical rulebase consists of descriptions of molecular substructures or 'structural alerts', which correlate with specific toxicological endpoints. The user communicates with DEREK by drawing the two-dimensional chemical structure of the query molecule on the screen. The rulebase is then searched against the query structure and any structural alert is

highlighted together with a message indicating the nature of the toxicological hazard (in this case, skin sensitization). Details of the first 40 structural alerts for skin sensitization are published elsewhere I Barratt et al., 1994a).

Assessment of likely skin penetration While details of the biology of skin sensitization are only partly understood, it is now widely accepted that the ability to react with a nucleophile either directly or after appropriate metabolism is a prerequisite for the majority of skin sensitizers. However, the potential of a chemical to act as a contact allergen is further modulated by its ability to penetrate the skin; this is apparent from a number of quantitative structure-activity relationship (QSAR) studies (Basketter et al., 1992; Roberts and Basketter, 1990) in which skin sensitization potential has been found to depend crucially on physicochemical parameters such as the log(octanol/water partition coefficient) (logP). These parameters have also been found to be equally important determinants of percutaneous absorption (e.g. Flynn, 1990), with higher logP values, namely greater lipophilicity broadly leading to greater permeability. In QSAR studies of skin permeability, log (in vitro human skin permeability coefficient) has also been shown to decrease with increasing molecular weight (Flynn, 1990) or molecular volume (Barratt, 1995). The skin sensitization potential of a series of substituted phenyl benzoates was found to depend on logP and molecular volume in the same way (Barratt et al., 1994c). The logical consequence is that two chemicals may contain the same structural alert (i.e. be reactive, presumably by the same mechanism), but one will be a skin sensitizer because it can penetrate the skin while the other will not be a skin sensitizer because its skin permeability is too low. Examples of pairs of chemicals behaving in this way have been identified in the authors' laboratory (Barratt and Basketter, 1994) and elsewhere, for example N-methyl-N-nitrosourea and streptozotocin (Ashby et al., 1995). LLNA The LLNA is a sensitization test which is based on the measurement of cell proliferation in lymph nodes draining the site of chemical (allergen) application. It has been assessed extensively in terms of both interlaboratory validation (Kimber and Basketter, 1992; Scholes et al., 1992) and comparisons with existing animal and human data (Basketter and Scholes, 1992; Basketter et al., 1994; Kimber et al., 1994) The LLNA produces results with OECD recommended positive control sensitizers which are equivalent to those of the accepted guinea pig procedures (Basketter et al., 1993). Furthermore, the quantitative dose response data produced by the LLNA may offer an opportunity for definition of

Alternative skin sensitization strategy 'no observable effect' levels or similar parameters. The LLNA result,; reproduced in this paper were all generated by the standard method (Kimber and Basketter, 1992).

The strategy--in theory The scheme illustrating the strategic approach to skin sensitization hazard identification is illustrated

in Fig. 1. In the first instance, a substance of defined chemical structure which is to be investigated is entered into the D E R E K system. If no structural alert is identified, then although the chemical is unlikely to be a significant skin sensitizer, at present this should be confirmed using a standard animal assay - - the LLNA is considered most appropriate. When a skin sensitization structural alert is identified,

a skin ~ sensitisation J ~\ structuralalert ~ (DEREK)?~

No

possess

\

No

~

1053

~-- I

likely?~

and label

)

Conduct LLNA

Positive response

~_ Fig. 1.

No label required

/~ .~

1054

D . A . Basketter et al.

t h e n this s u g g e s t s t h a t the chemical has m e t one o f the t w o criteria for classification. T o b e h a v e as a skin sensitizer, h o w e v e r , a chemical m u s t n o t only be able to derivatize skin protein, it m u s t also be able to p a r t i t i o n into the relevant skin compartment. C o n s e q u e n t l y in the next p h a s e o f the strategy (Fig. 1), t h o s e chemicals identified by D E R E K as p o s s e s s i n g a skin sensitization s t r u c t u r a l alert are e v a l u a t e d in the c o n t e x t o f h o w they will traverse the s t r a t u m c o r n e u m s u c h t h a t they could react within the living e p i d e r m a l cell layers, especially with the L a n g e r h a n s ' cell. I f the skin p e n e t r a t i o n o f the chemical is j u d g e d to be high, then it m a y be a s s u m e d also to have significant skin sensitization p o t e n t i a l a n d can be classified a n d labelled accordingly. I f skin p e n e t r a t i o n is j u d g e d to be insignificant, the chemical is c o n s i d e r e d unlikely to be a skin sensitizer. I n the latter case a n d w h e r e the extent o f skin p e n e t r a t i o n is j u d g e d to be m o d e r a t e or equivocal, it is c o n s i d e r e d p r u d e n t to assess the chemical in a suitable a n i m a l m o d e l . T h e L L N A is again c o n s i d e r e d the m o s t a p p r o p r i a t e test m e t h o d .

W h a t e v e r r o u t e is t a k e n to trigger the c o n d u c t o f a L L N A , the practical o u t c o m e is the same. I f the chemical is positive in this assay, it s h o u l d be classified as a skin sensitizer a n d labelled a c c o r d i n g l y (EC, 1993a). W h e n the result clearly does n o t m e e t the criteria for classification, t h e n in the a u t h o r s ' view n o f u r t h e r w o r k s h o u l d be necessary. T h e chemical m a y be r e g a r d e d as h a v i n g insufficient sensitization potential to m e r i t classification a n d labelling as a skin sensitizer.

Animal welfare considerations I n this section, b o t h the potential for r e d u c t i o n in a n i m a l use a n d the refinement o f r e m a i n i n g a n i m a l use will be discussed. T h e c u r r e n t D E R E K r u l e b a s e h a s been developed f r o m analysis o f o v e r 300 chemicals tested in the G P M T ; a p p r o x i m a t e l y h a l f o f these w o u l d be classified a n d labelled a c c o r d i n g to c u r r e n t E U criteria ( B a r r a t t e t al., 1994b). W i t h the exception o f one o r t w o chemicals t h a t are sensitizers b e c a u s e o f the presence o f impurities, the D E R E K rulebase n o w identifies all o f these as c o n t a i n i n g s t r u c t u r a l

Table 1. Demonstration dataset Chemical* 2,4-Dinit rochlorobenzene Formaldehyde Potassium dichromate lsoeugenol 4-Ethoxymethylene-2-phenyl-2-oxazol-5-one p-Pheoylene diamine Ethylenediamine Cinnamic aldehyde Kathon CG Dowicil 200 Cobalt chloride Nickel sulfate Hexyl cinnamic aldehyde Benzocaine Mercaptobenzot hiazole Glutaraldehyde Hydroxyethylacrylate Penicillin G Toluene diamine bismaleimide Eugenol Cocoamidopropylbetaine Citral Ethylene glycol dimethacrylate Hydroxycitronellal Diphenylthiourea Methyl salicylate Sodium dodecyl sulfate p-Aminobenzoic acid Diethylphthalate 2-Hydroxypropyl methacrylate Glycerol II Zinc sulfate IJ lsopropanol II Lead acetatell Olive oil I] Tartaric acid II Dimethyl formamidell

Sensitizer#

DEREK$

+ + + + + + + + + + + + + + + + + + + + + + + + + -

+ + + + + + + + + + + + + + + + + + + + + + + + + + -

Skin penetration§ High High Low High High Moderate High High High Moderate Low Low High Moderate High High Moderate Moderate Moderate High Low High High Moderate Moderate High Low Low High Moderate Low Low Moderate Low Low Moderate High

LLNA¶I + + + + + + + + + + + + +/+ + + + + + + + + + + +/-

*The list of chemicals is from the ECETOC (Kimber et al., 1995). tClassification based on EU criteria. ~DEREK expert system assessment of the presence of a structural alert for skin sensitization. §Expert view on likelihood of skin penetration, including evaluation of logP and molecular volume by computation. ¶lResult of testing in the LLNA, Data taken from previous publications (Basketter et aL, 1994; Kimber et al., 1994) or previously unpublished resultsll.

Alternative skin sensitization strategy alerts. About 25% of the chemicals not classified as sensitizers were also found to contain structural alerts; these chemicals tended to be those containing ionic or polar groups and/or with large molecular weights and which would therefore be predicted to penetrate skin less readily (Barratt and Basketter, 1994). Of those chemicals in the database that contain a structural alert, as identified by the DEREK rulebase, the majority would be regarded as having sufficient probability of peretrating skin to be classified as sensitizers. Thus, even on the conservative estimate that 50% of these could penetrate the skin to a sufficient extent, around one-third of chemicals could be classified as skin sensitizers without any recourse to animal testing. This proportion of chemicals classified as skin sensitizers agrees well with the experience of EU competent authorities (personal communication, 1994). For the remaining substances, it is proposed that these are tested in the LLNA according to the standard protocol (Kimber and Basketter, 1992). Since the LLNA uses about half the number of animals required for OECD protocol GPMT and Buehler tests, animal use could be cut by two-thirds if the strategy described in this paper was followed. Additionally, there is a further animal welfare advantage in that the LLNA does not require the use of FCA, intradermal injections of test substance, fur removal, occlusive dressings, or the use of restraint, which are all features of the GPMT and/or the Buehler test. Thus there is an important opportunity for both subst.'mtial reduction and refinement of animal usage in a raanner which will not compromise the existing standard of classification and labelling of skin sensitization hazard in the EU. The strategy--in practice

In describing the aforementioned strategy, it is important to demonstrate that it works successfully with datasets of chemicals which either do, or do not, classify as skin sensitizers. However, it is not appropriate in this paper to repeat exhaustive lists of chemicals discussed elsewhere (Barratt and Basketter, 1994; Barratt et al., 1994a,b; Kimber et al., 1994). Consequently, an assessment of the list of positive and negative substances specifically proposed by the European Centre for Ecotoxicology and Toxicology of Chemicals (ECETOC) for the purposes of novel test evaluation has been chosen as an example (I. Kimber et al., unpublished data, 1995). Table 1 reports in brief, tl~eir categorization as skin sensitizers, assessment by the DEREK rulebase, a broad assessment of their skin penetration characteristics, and the LLNA result. Of the 25 sensitizing chemicals, all but one are identified by the DEREK rulebase as containing structural alerts for skin sensitization. The one remaining chemical, cocoamidopropylbetaine, recently shown to be a sensitizer because of the

1055

presence of a sensitizing impurity, 3-dimethylaminopropylamine, a chemical used in its manufacture (Angelini et al., 1995), is correctly identified by the LLNA. Of the 12 chemicals not classified in the EU as skin sensitizers, only two were identified by the DEREK rulebase as containing structural alerts. One of these chemicals, p-aminobenzoic acid, is unlikely to be a significant skin sensitizer because of its poor ability to penetrate skin. This is confirmed by the negative LLNA. Hydroxypropyl methacrylate met DEREK criteria for skin sensitization and was also assessed to be 'moderate' for skin penetration and thus might be judged to be a false positive. The remaining 10 chemicals did not contain structural alerts for skin sensitization and with the exception of sodium dodecyl sulfate, which at times is slightly positive in the LLNA (Basketter et al., 1994; Kimber and Basketter, 1992), were all negative in the LLNA assay. Thus the results in Table 1 serve to demonstrate that the strategic approach described in this paper will permit the correct identification of chemicals which should be classified as skin sensitizers. Furthermore, this correct identification is not achieved by over-classification of non-sensitizers. On the basis of the existing datasets, adoption of this strategy could cut animal use by two-thirds and eliminate the need for treatment of the remaining animals with FCA. Conclusion

In this paper, an alternative strategy to the currently used guinea pig methods for the identification, classification, and labelling of chemicals which present a skin sensitization hazard has been described. The approach has considerable merit on animal welfare grounds and is both rapid and highly cost effective. Furthermore, while it does not seek to change existing regulatory thresholds, it does offer a real opportunity for an improvement to the hazard identification process since the strategy is based on the fundamental mechanistic processes of skin sensitization. The evidence available strongly suggests that this strategy could be implemented now without compromising human safety. REFERENCES

Angelini G., Foti C., Rigano L. and Vena G. A. (1995) 3-Dimethylaminopropylamine:a key substance in contact allergy to cocamidopropylbetaine?Contact Dermatitis 32, 96-99. Ashby J., Hilton J., Dearman R. J. and Kimber I. (1995) Streptozotocin: inherent but not expressed skin sensitising activity. Contact Dermatitis. In press. Barratt M. D. (1995) Quantitative structure-activity relationships for skin permeability. Toxicology in Vitro 9, 27-37. Barratt M. D. and Basketter D. A. (1994) Structure-activity relationships for skin sensitization: an expert system. In In Vitro Toxicology. Edited by A. Rougier A. M.,

1056

D, A. Basketter et al.

Goldberg and H. I. Maibach. pp. 293-301. Mary Ann Liebert, Inc., New York. Barratt M. D., Basketter D. A., Chamberlain M., Admans G. D. and Langowski J. J. (1994a) An expert system rulebase for identifying contact allergens. Toxicology in Vitro 8, 1053-1060. Barratt M. D., Basketter D. A., Chamberlain M., Payne M. P., Admans G. D. and Langowski J. J. (1994b) Development of an expert system rulebase for identifying contact allergens. Toxicology in Vitro 8, 837-839. Barratt M. D., Basketter D. A. and Roberts D. W. (1994c) Skin sensitization structure-activity relationships for phenyl benzoates. Toxicology in Vitro 8, 823 826. Basketter D. A., Roberts D. W., Cronin M. and Scholes E. W. (1992) The value of the local lymph node assay in quantitative structure activity investigations. Contact Dermatitis 27, 137 142. Basketter D. A. and Scholes E. W. (1992) A comparison of the local lymph node assay with the guinea pig maximization test for the detection of a range of contact allergens. Food and Chemical Toxicology 30, 65-69. Basketter D. A., Scholes E. W. and Kimber I. 1994) Performance of the local lymph node assay with chemicals found identified as contact allergens in the human maximization test. Food and Chemical Toxicology 32, 543-547. Basketter D. A., Selbie E., Scholes E. W., Lees D., Kimber I. and Botham P. A. (1993) Results with OECD recommended positive control sensitizers in the maximization, Buehler and local lymph node assays. Food and Chemical Toxicology 31, 63-67. Botham P. A., Basketter D. A., Maurer Th., Mueller D., Potokar M. and Bontinck W. (1991) Skin sensitization - a critical review of predictive test methods in animals and man. Food and Chemical Toxicology 29, 275-286. Buehler E. V. (1965) Delayed contact hypersensitivity in the guinea pig. Archives of Dermatology 91, 171-176. Cronin M. T. D. and Basketter D. A. (1994) Multivariate

QSAR analysis of a skin sensitization database. SA R and QSAR in Environmental Research 2, 1-21. Dupuis G. and Benezra C. (1982) Contact Dermatitis to Simple Chemicals: A Molecular Approach. Marcel Dekker, New York. EC (1993a) Council Directive 92/32/EEC. 7th Amendment to Directive 67/548/EEC. Official Journal of the European Communities 35, L154. EC (1993b) B.6, Skin sensitization. O)ficial Journal of the European Communities 35, L383. Flynn G. L. (1990) Physicochemical determinants of skin absorption. In Principles of Route-to-Route Extrapolation for Risk Assessment. Edited by T. R. Gerrity and C. J. Henry. pp. 93-127. Elsevier, New York. Kimber I. and Basketter D. A. (1992) The murine local lymph node assay: a commentary on collaborative studies and new directions. Food and Chemical Toxicology 30, 165-169. Kimber I., Dearman R. J., Scholes E. W. and Basketter D. A. (1994) The local lymph node assay: developments and applications. Toxicology 93, 13-31. Magnusson B. and Kligman A. M. (1970) Allergic Contact Dermatitis in the Guinea Pig. Identification of Contact Allergens. Charles C. Thomas, Springfield, IL. OECD (1993) Organisation for Economic Cooperation and Development Guideline 406 Skin Sensitization, adopted July 1992. OECD, Paris. Roberts D. W. and Basketter D. A. (1990) A quantitative structure activity/dose response relationship for contact allergic potential of alkyl group transfer agents. Contact Dermatitis 23, 331 335. Sanderson D. M. and Earnshaw C. G. (1991) Computer prediction of possible toxic action from chemical structure; the DEREK system. Human and Experimental Toxicology 10, 261-273. Scholes E. W., Basketter D. A., Sarll A. E., Kimber I., Evans C. D., Miller K., Robbins M. C., Harrison P. T. C. and Waite S. J. (1992) The local lymph node assay: results of a final interlaboratory validation under field conditions. Journal of Applied Toxicology 12, 217-222.