- Email: [email protected]
Skin sensitisation, vehicle effects and the local lymph node assay
Food and Chemical Toxicology 39 (2001) 621±627
www.elsevier.com/locate/foodchemtox
Commentary
Skin sensitisation, vehicle eects and the local lymph node assay D.A. Basketter a,*, G.F. Gerberick b, I. Kimber c a
SEAC Toxicology Unit, Unilever Research, Colworth House, Sharnbrook, Bedford MK44 1LQ, UK b Miami Valley Laboratories, Procter & Gamble Company, Cincinnati, OH, USA c Zeneca Central Toxicology Laboratory, Alderley Park Maccles®eld, Cheshire SK10 4TJ, UK Accepted 14 October 2000
Abstract Accurate risk assessment in allergic contact dermatitis is dependent on the successful prospective identi®cation of chemicals which possess the ability to behave as skin sensitisers, followed by appropriate measurement of the relative ability to cause sensitisation; their potency. Tools for hazard identi®cation have been available for many years; more recently, a novel approach to the quantitative assessment of potency Ð the derivation of EC3 values in the local lymph node assay (LLNA) Ð has been described. It must be recognised, however, that these evaluations of chemical sensitisers also may be aected by the vehicle matrix in which skin exposure occurs. In this article, our knowledge of this area is reviewed and potential mechanisms through which vehicle eects may occur are detailed. Using the LLNA as an example, it is demonstrated that the vehicle may have little impact on the accuracy of basic hazard identi®cation; the data also therefore support the view that testing ingredients in speci®c product formulations is not warranted for hazard identi®cation purposes. However, the eect on potency estimations is of greater signi®cance. Although not all chemical allergens are aected similarly, for certain substances a greater than 10-fold vehicle-dependent change in potency is observed. Such data are vital for accurate risk assessment. Unfortunately, it does not at present appear possible to predict notionally the eect of the vehicle matrix on skin sensitising potency without recourse to direct testing, for example by estimation of LLNA EC3 data, which provides a valuable tool for this purpose. # 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction It has been recognised for many years that the vehicle matrix in which a chemical allergen is encountered on the skin can have profound eects on the response elicited and the extent to which skin sensitisation will be acquired (e.g. Kligman, 1966; Magnusson and Kligman, 1970). This, then, must be an important consideration for the interpretation of skin sensitisation testing (Robinson et al., 1989; Basketter et al., 1996; Gerberick and Robinson, 2000). Moreover, vehicle eects are not uncommonly implicated as the basis for misleading results from diagnostic human patch testing (particularly false negatives) (Ljunggren and Moller, 1972; Roeleveld and van Ketel, 1975; LideÂn and Boman, 1988). These views and experiences are supported by a relatively sparse literature on skin sensitisation. For Abbreviations: DNCB, 2,4-dinitrochlorobenzene; LC, Langerhans cells; SLS, sodium lauryl sulfate; TNF-a, tumour necrosis factor a. * Corresponding author. Tel.: +44-1234-222236; fax: +44-1234222122. E-mail address: [email protected] (D.A. Basketter).
example, it has been demonstrated in human predictive testing that the vehicle is an important modulator of the response that can be elicited in sensitised individuals (Kligman, 1966). One decade later, this was also shown to be true for the induction of sensitisation, although the eects seen in this limited study were less remarkable (Marzulli and Maibach, 1976). In the guinea pig, it has been shown that the vehicle can substantially aect the rate of sensitisation, although this could not be related to the apparent bioavailability (Andersen et al., 1985). However, the view that the vehicle matrix is likely to have an important impact on the sensitising eect of a chemical is also based on the underlying knowledge of the biology of skin penetration of chemicals (reviewed in Smith and Hotchiss, 2001) and physical chemistry (Kirchner et al., 1997). What does the above mean in relation to the prediction of skin sensitisation hazard and the subsequent hazard assessment of sensitising chemicals? Clearly, the identi®cation of skin sensitisation hazards using predictive animals models must take this into consideration (Schlede and Eppler, 1995). Furthermore, where the vehicle may have an impact on the degree of skin sensitisation observed, then it is apparently modulating the potency
0278-6915/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved. PII: S0278-6915(00)00169-1
622
D.A. Basketter et al. / Food and Chemical Toxicology 39 (2001) 621±627
of the allergen. This is of paramount importance in risk assessment (Basketter et al., 1996; Gerberick and Robinson, 2000; Gerberick et al., 2001). Matrix eects are complicated, and unfortunately there exist at present no in vitro or in silico methods to predict the eect of a vehicle on the activity of a particular skin sensitiser. Traditional animal test methods using the guinea pig are also poorly suited to this purpose, due to their relative complexity and dependence on qualitative endpoints (Kimber and Basketter, 1997; Basketter et al., 2000a). However, the local lymph node assay (LLNA) holds out more promise in this respect and has been used to provide objective and quantitative information on the impact of vehicles on skin sensitisation (Kimber and Basketter, 1997; Dearman et al., 1999; Basketter et al., 2000a). Here we review the information available on vehicle eects in skin sensitisation and consider both the relevance to hazard identi®cation and its potential utility for the purposes of risk assessment. 2. Vehicle eects: mechanisms It is necessary to consider the ways in which a vehicle or formulation matrix may impact on the skin sensitising potency of a chemical allergen. It has often been assumed that the most important in¯uence that a change in matrix may have is to alter skin penetration and the eectiveness with which a chemical allergen gains access to the viable epidermis. It is certainly the case, with speci®c examples for skin sensitisation including the common allergen nickel (Fullerton et al., 1988), the transdermal drug triprolidine, the sensitising activity of which was aected substantially by the vehicle matrix in which skin exposure occurred (Robinson and Cruze, 1996) and assessment of the alteration of skin sensitisation activity secondary to dibutyl phthalate mediated modulation of ¯uorescein isothiocyanate penentration (Dearman et al., 1996). While there is no doubt penetration is a key variable in instances where the partition coecient of the chemical allergen prevents or limits access across the stratum corneum, it is not necessarily always the relevant or predominant mechanism. For instance, it has been demonstrated that cutaneous responses induced in mice by topical exposure to suboptimal concentrations of the chemical allergen 2,4-dinitrochlorobenzene (DNCB) can be enhanced signi®cantly by co-administration of the non-sensitising surfactant sodium lauryl sulfate (SLS) (Heylings et al., 1996). Under conditions where SLS was able to augment the stimulation of lymph node cell proliferative responses to DNCB, it failed to in¯uence the eciency with which the allergen was absorbed through the skin. These and other data suggest that there exists a variety of ways in which the delivery matrix may aect the acquisition of skin sensitisation.
A further eect of the vehicle might also be to alter the protonation state of the chemical, the protonated species having a sharply reduced ability to penetrate into the skin. pKa is the pH at which 50% of the substance will be ionised. So, for example, a substance whose pKa is 5 (e.g. a weak carboxylic acid), when placed in a vehicle matrix at pH 7 is much more charged and so will penetrate much less well than if skin contact were to occur in a vehicle matrix at pH 4. The proportion of a substance which is charged can be calculated from the Henderson±Hasselbach equation: pH pKa logA=HA Another important consideration is that the induction of skin sensitisation requires the activation of Langerhans cells (LC) (reviewed in Kimber et al., 2000). Upon encounter with a chemical allergen, a proportion of LC at the site of exposure are mobilised and are stimulated to migrate, bearing antigen, to draining lymph nodes. The migration and subsequent maturation of LC, as well as other cellular interactions and processes necessary for skin sensitisation, are initiated and regulated by epidermal cytokines. The epidermis is a rich source of cytokines, some of which are expressed constitutively, while others are expressed only following receipt of an appropriate signal. It has been demonstrated that exposure to skin allergens and skin irritants is associated with the induction or upregulation of many cytokines, including those that are known to be necessary for the development of skin sensitisation (reviewed in Grabbe and Schwarz, 1998; Basketter et al., 1999a). Among the mediators known to play important roles in the induction of cutaneous immune responses are proin¯ammatory cytokines such as interleukins 1 and 6 (IL-1 and IL-6) and tumour necrosis factor a (TNF-a). The expression of some of these cytokines, such as TNF-a, will be increased following various forms of dermal trauma. Indeed, it has been postulated that some small degree of dermal trauma (and the production of epidermal cytokines associated with it) is essential for the normal development of skin sensitisation (Grabbe and Schwarz, 1998; Kimber et al., 2000). The argument is therefore that one impact that the vehicle matrix may have on skin sensitisation is to in¯uence the extent to which cytokines necessary for LC migration and maturation and other cellular processes will be produced at the time of contact with the chemical allergen. Following this reasoning, it may well be that the frequently reported ability of co-administration of SLS to augment skin sensitisation (e.g. Kligman, 1966) is at least in part attributable to the ability of this chemical to provoke the production of relevant proin¯ammatory cytokines (Grabbe and Schwarz, 1998). A requirement for some degree of dermal trauma for the eective development of skin sensitisation is of course consistent with the danger hypothesis of immune activation, which
D.A. Basketter et al. / Food and Chemical Toxicology 39 (2001) 621±627
in its simplest form argues that adaptive immune responses will be generated only if there is a certain degree of cell or tissue damage (Matzinger, 1994; McFadden and Basketter, 2000). In addition to possible eects on skin penetration and the provision of danger signals, it is possible that vehicles may in¯uence the eciency of skin sensitisation in other ways, including modulation of the metabolic activation of a prohapten, or metabolic inactivation of the hapten itself. Among other possible eects are those that will in¯uence the evaporation of a chemical allergen from the skin surface, impact a chemical phenomenon, such as allergen polymerisation (BjoÈrkner and Niklasson, 1984) or will aect in some way the formation of stable haptenprotein conjugates. 3. Vehicle eects: hazard identi®cation Although the application of predictive tests for skin sensitisation hazard identi®cation may vary to some extent globally, as a general rule the primary purpose is to discriminate between those chemicals which present a signi®cant hazard and those which do not (EC, 1996; WHO, 1997). Classi®cation as a skin sensitiser is a binary decision and as such, the key question relating to the eect of a vehicle on a predictive test is whether the matrix in which a chemical is presented will cause a change of such classi®cation. In theory, this must be possible; for example where a sensitiser gives a positive result in the guinea pig maximisation test with 35% of the test animals being sensitised, and so is just above the EC classi®cation limit (EC, 1996), then testing in a different vehicle which reduces skin penetration may give rise to a result of say 20% positive, which is below the classi®cation limit of 30%. However, there is no evidence that this is in practice a real problem, either from our data or from the following consideration. It is our experience that the distribution of responses in guinea pig tests, particularly the popular guinea pig maximisation test is not even, being rather polarised to the extremes (Fig. 1). In this analysis of 259 chemicals (taken from Cronin and Basketter, 1994), there was a preponderance of materials in the non-sensitising and strongly sensitising categories compared with the weak and moderate categories. It is in the latter group where vehicle eects are more likely to have an impact on regulatory classi®cation. This skewed distribution has the practical eect of reducing the risk of false negatives. This is con®rmed analysis of data submitted to the regulatory authority in the UK, which showed that only a very small proportion of submitted sensitisation tests fell close to the classi®cation borderline (Elliott-Minty and Evans, 1999). Published data on the impact of vehicles on hazard identi®cation in traditional guinea pig tests are very limited; to the best of our knowledge,
623
no substantial published information on this topic is available. The LLNA is a newer method which has been validated in terms of hazard identi®cation of skin sensitisation (NIH, 1999); it has been accepted as valid by the European Centre for the Validation of Alternative Methods (ECVAM) and is recommended as the preferred method (ECVAM, 2001). In contrast to guinea pig tests, the LLNA has been examined more carefully in relation to the eects of vehicles. The results indicate that hazard identi®cation is not compromised to a signi®cant extent by the choice of vehicle (Lea et al., 1999; Warbrick et al., 1999; Wright et al., 2001). The data for seven chemicals tested in the LLNA in a range of recommended vehicles revealed that the vehicle matrix had no impact on classi®cation, although it should be noted that for one substance/ vehicle combination (dimethylaminopropylamine in propylene glycol) the dataset was not conclusive (Table 1). Furthermore, these data also support the view that testing ingredients in speci®c product formulations is not warranted for hazard identi®cation purposes. It is worth mentioning at this point that advice on the choice of vehicles in relation to predictive tests for the identi®cation of skin sensitisation hazard is fairly limited. In the OECD Guideline 406 on skin sensitisation (OECD, 1992), the only mention of vehicles is in relation to the Buehler guinea pig test, where it is stated that, ``For water soluble materials, it is appropriate to use water or a dilute non-irritating solution of surfactant as a vehicle. For other test materials 80% ethanol/water is preferred for induction and acetone for challenge.'' For the LLNA, more speci®c guidance has been given (Kimber and Basketter, 1992); this has been reiterated in material related to the formal validation of the LLNA (NIH, 1999; Gerberick et al., 2001) and in the draft OECD Guideline (OECD, 2000). However, it still a dicult matter to select a vehicle system and it relies on the experimenter to balance a range of considerations, including the solubility and chemical compatiblity of the test substance, the extent to which local irritant and toxic eects are produced and the likelihood of systemic toxicity. Clearly, these considerations will have to be judged on a case by case basis but, for example, local severe irritation/necrosis at the treatment site should be avoided. An additional criterion might be to make constructive use of the recommendation in the draft OECD Guideline for the LLNA that body weights are taken at the start and the end of the experiment. Thus, to the clear advice already available for vehicle selection in the LLNA (Kimber and Basketter, 1992), it can be added that evidence of systemic toxicity, as demonstrated by a greater than 10% fall in body weight in the particular dose group, normally might invalidate the data for that group. No doubt a similar general consideration might be applied to the existing guinea pig tests.
624
D.A. Basketter et al. / Food and Chemical Toxicology 39 (2001) 621±627
Fig. 1. 259 chemicals tested in the guinea pig maximization test were placed into one of four categories, non-sensitising, weakly, moderately or strongly sensitising. The histogram shows the frequency (expressed as %) with which each category occurred.
Table 1 Classi®cation of allergens in the LLNAa,b Chemical allergen
Dimethylaminopropylamine 1,4-Dihydroquinone Cinnamic aldehyde (Chloro)methylisothiazolinone Dibromodicyanobutane Isoeugenol Benzocaine
Classi®cation as a skin sensitiser in a range of vehiclesc
Reference
AOO
MEK
DMF
PG
DMSO
+ + + + + +
+ + + + + +
+ + + + + +
? + + + NDd + ND
+ + + + ND +
Wright et al. (2001) Lea et al. (1999) Wright et al. (2001) Warbrick et al. (1999) Wright et al. (2001) Wright et al. (2001) Warbrick et al. (2000)
a Eect of a range of ®ve vehicles on the outcome of EU-based classi®cation decisions for the skin sensitisation potential of seven chemicals is presented. b ``+'' indicates that the chemical would be formally classifed as a skin sensitiser and denoted R43 according to EU regulations, `` '' means the chemical would not be classi®ed and ``?'' indicates that the data is equivocal. c AOO, acetone/olive oil (4:1, v/v), MEK, methylethylketone, DMF, dimethylformamide, PG, propylene glycol, DMSO, dimethyl sulfoxide. d ND, not done.
4. Vehicle eects: risk assessment In the section above, the impact of vehicle was considered only in the situation where a binary judgement had to be made. However, where the data indicate that the chemical possesses skin sensitisation hazard, then further use of that chemical will depend on assessment
of the risk to human health that it represents. Of course the risk will depend on the nature, extent and duration of skin exposure, but it will depend also on the potency of the allergen (Robinson et al., 2000). Both of these aspects, potency and exposure, can be modulated by the vehicle matrix in which the sensitising chemical is encountered. Whatever the mechanism(s) through which
D.A. Basketter et al. / Food and Chemical Toxicology 39 (2001) 621±627
625
Table 2 Impact of ®ve dierent vehicles on derived LLNA EC3 valuesa Chemical allergen
1,4-Dihydroquinone Cinnamic aldehyde (Chloro)methylisothiazolinone Dibromodicyanobutane Dimethylaminopropylamine Isoeugenol Ethyleneglycol dimethacrylate
EC3 values (%) for each skin sensitiser in a range of vehiclesb
Reference
AOO
MEK
DMF
PG
DMSO
0.15 1.7 0.05 5.2 2.2 1.0 36
0.09 1.1 0.007 0.4 1.8 1.0 28
0.21 0.5 0.008 6.4 1.7 1.4 32
1.5 1.4 0.05 NDc >10 2.5 15
0.35 0.9 0.008 ND 3.2 0.9 34
Lea et al. (1999) Wright et al. (2001) Warbrick et al. (1999) Wright et al. (2001) Wright et al. (2001) Wright et al. (2001) Blaikie et al. (2000)
a
For each of seven skin sensitising chemicals, the concentration (%) required to stimulate a three fold increase in cell proliferation in the draining lymph node (the EC3 value) has been estimated from the LLNA dose±response curve. EC3 values for each sensitiser in each of ®ve vehicles are presented. b AOO, acetone/olive oil (4:1, v/v); MEK, methylethylketone; DMF, dimethylformamide; PG, propylene glycol; DMSO, dimethyl sulfoxide. c ND, not done.
this modulation occurs, the important consideration for risk assessment is to gain an understanding of the eect so that it can be evaluated accurately. In guinea pig tests, quantitative assessment of the eect of vehicles is relatively dicult. This is due in part to the inherent diculties associated with the technical aspects of the conduct of guinea pig tests, which tend to conspire to ensure that when the vehicle is changed, other variables, most notably the concentration, also have to be changed. This is largely a consequence of the use of the irritation potential of the chemical as a basis for dose selection. Also, the subjective nature of the endpoint (visual assessment of erythema) does not lend itself readily to quantitation. The normal endpoint measurement in the LLNA is an immunologically relevant quantitative dose response (Kimber and Dearman, 1991). These data can be interpolated to yield a quantitative measure of relative potency, calculated as the estimated concentration of the chemical necessary to stimulate a threefold increase in lymph node cell proliferation Ð the EC3 value (Basketter et al., 1999b). This measure is highly reproducible within and between laboratories (Loveless et al., 1996; Warbrick et al., 1999) and is stable over time (Dearman et al., 1998). For the assessment of the eect of vehicles, EC3 values have been derived for sensitising chemicals tested in several vehicles in the LLNA. Such studies have been completed for a range of allergen/vehicle combinations, the data being summarised in Table 2. The results show that for some skin sensitisers, the vehicle can have a substantial impact on potency. For other allergens, changing the vehicle appears to be of little importance, with, for example, cinnamic aldehyde, or the very weak allergen, benzocaine showing virtually no vehicle-dependent changes (Warbrick et al., 2000; Wright et al., 2001). What bene®t can be derived from such quantitative estimates and how can LLNA EC3 values be applied to
risk assessment? Risk assessment for skin sensitisation requires information on potency and the EC3 value has been shown not only to be a useful measure for this purpose (Kimber and Basketter, 1997; Hilton et al., 1998) but also to have a close relationship with the apparent potency in humans (Basketter et al., 2000a; Gerberick et al., 2001). Using such an approach, the relative potency of chemicals in dierent vehicles can be compared directly using LLNA EC3 data. The vehicle may, for example, be similar in nature to the product matrix in which the skin will be exposed to the sensitising chemical. Completion of the risk assessment will require knowledge of (a) sensitising chemical(s) for which skin exposure is well documented, whose potency is understood and where the consequent (lack of) allergic contact dermatitis is known to act as points of comparison. Such chemicals are known as a benchmarks; several benchmarks may be necessary to eect accurately the risk assessment of an unknown sensitising chemical. However, the stability of results in the LLNA means that historic data, or published data from other laboratories can be utilised for this purpose. The ultimate consequence of this is that there will be a reduction in the need for animal testing, while at the same time, improved quantitative risk assessments can be made. Speci®cally, this will eliminate the need for evaluation of product formulations for the assessment of skin sensitisation potential. 5. Conclusion There has been a recognition for many years that vehicles can have important eects on the acquisition of skin sensitisation as well as on its clinical manifestation as allergic contact dermatitis (Kligman, 1966; Magnusson and Kligman, 1970; Marzulli and Maibach, 1976). It was noted in the last-mentioned paper that ``. . .few speci®c data are available on two experimental conditions
626
D.A. Basketter et al. / Food and Chemical Toxicology 39 (2001) 621±627
that may aect results obtained in skin sensitization studies: the vehicle used for evaluating the putative allergen. . .''. More than 20 years on, a similar statement might be made. The reality is that during this period there has been very little new information on the mechanisms through which vehicles might aect skin sensitisation, nor is there currently available a larger body of data qualitatively or quantitatively examining the eect of vehicles. The development of the LLNA as a new method for the identi®cation and characterisation of skin sensitisation potential has provided new possibilities (Kimber and Basketter, 1997). The quantitative nature of LLNA dose responses enabled us for the ®rst time to make meaningful measurements of relative potency (Basketter et al., 2000a; Gerberick et al., 2001). Once potency could be reliably and indeed easily quanti®ed, it has proven relatively simple to investigate the impact of vehicle. It is clear from the information presented here and elsewhere (Lea et al., 1999; Warbrick et al., 1999; Blaikie et al., 2000; Wright et al., 2001) that, while hazard identi®cation may be essentially unaected, the vehicle can have a substantial and unpredictable eect on the potency of an allergen. Indeed, from the work to date, there are no readily discernible relationships between a vehicle physicochemical characteristics and their eect of skin sensitisation. Thus, although the data indicate that the testing of in a range of vehicles or in product formulations for the speci®c identi®cation of the expression of a known sensitising hazard is unlikely to be necessary, the results have important consequences for risk assessment; because of its quantitative endpoint and high reproducibility, the LLNA provides a valuable tool through which vehicle eects on chemical sensitisers may be measured quantitatively. Appropriate use of the LLNA may thus lead to an improvement in risk assessment.
References Andersen, K.E., Carlsen, L., Egsgaard, H., Larsen, E., 1985. Contact sensitivity and bioavailability of chlorocresol. Contact Dermatitis 13, 246±251. Basketter, D.A., Gerberick, G.F., Robinson, M., 1996. Risk assessment. In: Kimber, I., Maurer, Th. (Eds.), The Toxicology of Contact Hypersensitivity. Taylor and Francis, London, pp. 152±164. Basketter, D.A., Gerberick, G.F., Kimber, I., Willis, C., 1999a. The Toxicology of Contact Dermatitis. Wiley and Sons, Chichester. Basketter, D.A., Lea, L.J., Cooper, K., Dickens, A., Briggs, D., Pate, I., Dearman, R.J., Kimber, I., 1999b. A comparison of statistical approaches to derivation of EC3 values from local lymph node assay dose responses. Journal of Applied Toxicology 19, 261±266. Basketter, D.A., Blaikie, L., Dearman, R.J., Kimber, I., Ryan, C.A., Gerberick, G.F., Harvey, P., Evans, P., White, I.R., Rycroft, R.J.G., 2000a. Use of the local lymph node assay for the estimation of relative contact allergenic potency. Contact Dermatitis 42, 344±348. Basketter, D.A., Evans, P., Gerberick, G.F., Kimber, I., 2000b. Chemical allergy: Estimation of potency, thresholds and risk assessments. Comments in Toxicology 7, 79±89.
BjoÈrkner, B., Niklasson, B., 1984. In¯uence of the vehicle on the elicitation of contact allergic reactions to acrylic compounds in the guinea pig. Contact Dermatitis 11, 268±278. Blaikie, L., Basketter, D.A., Warbrick, E.V., Dearman, R.J., Kimber, I., 2000. Utility of the LLNA for the investigation of the impact of vehicle matrix on sensitisation potency. Toxicological Sciences 54, 121. Cronin, M.T.D., Basketter, D.A., 1994. Multivariate QSAR analysis of a skin sensitization database. SAR and QSAR in Environmental Research 2, 159±179. Dearman, R.J., Cumberbatch, M., Hilton, J., Clowes, H.M., Fielding, I., Heylings, J.R., Kimber, I., 1996. In¯uence of dibutyl phthalate on dermal sensitization to ¯uorescein isothioscyanate. Fundamental and Applied Toxicology 33, 24±30. Dearman, R.J., Hilton, J., Evans, P., Harvey, P., Basketter, D.A., Kimber, I., 1998. Temporal stability of local lymph node assay responses to hexyl cinnamic aldehyde. Journal of Applied Toxicology 18, 281±284. Dearman, R.J., Basketter, D.A., Kimber, I., 1999. The local lymph node assay: use in hazard and risk assessment. Journal of Applied Toxicology 19, 299±306. EC (European Community), 1996. Dangerous Substances Directive, 22nd Adaptation to Technical Progress (96/54/EC). Ocial Journal of the European Communities 199, L248. ECVAM, in press. Statement on the validity of the Local Lymph Node Assay for skin sensitisation testing. ECVAM Joint Research Centre, European Commission, Ispra, Italy, 21 March 2000. ATLA. Elliott-Minty, C., Evans, P., 1998. Regulatory aspects of testing new industrial chemicals for skin sensitisation potential. Human and Experimental Toxicology 17, 498. Fullerton, A., Andersen, J.R., Hoelgaard, A., 1988. Permeation of nickel through human skin in vitro Ð eects of vehicles. British Journal of Dermatology 118, 509±516. Gerberick, G.F., Robinson, M.K., 2000. A skin sensitization risk assessment approach for evaluation of new ingredients and products. American Journal of Contact Dermatitis 11, 65±73. Gerberick, G.F., Ryan, C.A., Kimber, I., Dearman, R.J., Lea, L.J., Basketter, D.A., 2001. Local lymph node assay validation assessment for regulatory purposes. American Journal of Contact Dermatitis 11, 13±18. Grabbe, S., Schwarz, T., 1998. Immunoregulatory mechanisms involved in elicitation of allergic contact hypersensitivity. Immunology Today 19, 37±44. Heylings, J.R., Clowes, H.M., Cumberbatch, M., Dearman, R.J., Fielding, I., Hilton, J., Kimber, I., 1996. Sensitisation to 2,4-dinitrochlorobenzene: in¯uence of vehicle on absorption and lymph node activation. Toxicology 109, 57±65. Hilton, J., Dearman, R.J., Harvey, P., Evans, P., Basketter, D.A., Kimber, I., 1998. Estimation of relative skin sensitizing potency using the local lymph node assay: a comparison of formaldehyde and glutaraldehyde. American Journal of Contact Dermatitis 9, 29±33. Kimber, I, Basketter, D.A., 1992. The murine local lymph node assay; collaborative studies and new directions: a commentary. Food and Chemical Toxicology 30, 165±169. Kimber, I., Basketter, D.A., 1997. Contact sensitisation: a new approach to risk assessment. Human and Ecological Risk Assessment 3, 385±395. Kimber, I., Dearman, R.J., 1991. Investigation of lymph node cell proliferation as a possible immunological correlate of contact sensitizing potential. Food and Chemical Toxicology 29, 125±129. Kimber, I., Cumberbatch, M., Dearman, R.J., Bhushan, M., Griths, C.E., 2000. Cytokines and chemokines in the initiation and regulation of epidermal Langerhans cell mobilization. British Journal of Dermatology 142, 401±412. Kirchner, L.A., Moody, R.P., Doyle, E., Bose, R., Jerey, J., Chu, I., 1997. The prediction of skin permeability using physicochemical data. ATLA 25, 359±370.
D.A. Basketter et al. / Food and Chemical Toxicology 39 (2001) 621±627 Kligman, A.M., 1966. The identi®cation of contact allergens by human assay. II. Factors in¯uencing the induction and measurement of allergic contact dermatitis. Journal of Investigative Dermatology 47, 375±392. Lea, L.J., Warbrick, E.V., Dearman, R.J., Harvey, P., Kimber, I., Basketter, D.A., 1999. The impact of vehicle on assessment of relative skin sensitisation potency of 1,4-dihydroquinone in the local lymph node assay. Food and Chemical Toxicology 10, 213±218. LideÂn, C., Boman, A., 1988. Contact allergy to colour developing agents in the guinea pig. Contact Dermatitis 19, 290±295. Ljunggren, B., Moller, H., 1972. Eczematous contact allergy to chlorhexidine. Acta Dermatologica Venereologica 52, 308±310. Loveless, S.E., Ladics, G.S., Gerberick, G.F., Ryan, C.A., Basketter, D.A., Scholes, E.W., House, R.V., Hilton, J., Dearman, R.J., Kimber, I., 1996. Further evaluation of the local lymph node assay in the ®nal phase of an international collaborative trial. Toxicology 108, 141±152. Magnusson, B., Kligman, A.M., 1970. Allergic Contact Dermatitis in the Guinea Pig. Charles C. Thomas, Spring®eld, IL. Marzulli, F.N., Maibach, H.I., 1976. Eects of vehicles and elicitation concentration in contact dermatitis testing. Contact Dermatitis 2, 325±329. Matzinger, P., 1994. Tolerance, danger and the extended family. Annual Review of Immunology 12, 991±1045. McFadden, J.P., Basketter, D.A., 2000. Contact allergy, irritancy and danger. Contact Dermatitis 42, 123±127. NIH (National Institutes of Health), 1999. The Murine Local Lymph Node Assay: A Test Method for Assessing the Allergic Contact Dermatitis Potential of Chemicals/Compounds. OECD, Paris. OECD, 1992. OECD Guidelines for Testing of Chemicals. Guideline 406, Skin Sensitisation. Organisation for Economic Cooperation and Development, Paris. OECD, 2000. Draft Test Guideline for the Local Lymph Node Assay. Organisation for Economic Cooperation and Development, Paris.
627
Robinson, M.K., Cruze, C.A., 1996. Preclinical skin sensitization testing of antihistamines: guinea pig and local lymph node assay responses. Food and Chemical Toxicology 34, 495±506. Robinson, M.K., Stotts, J., Danneman, P.J., Nusair, T.L., Bay, P.H.S., 1989. A risk assessment process for allergic contact sensitization. Food and Chemical Toxicology 27, 479±489. Robinson, M.K., Gerberick, G.F., Ryan, C.A., McNamee, P., White, I.R., Basketter, D.A., 2000. The importance of exposure estimation in the assessment of skin sensitization risk. Contact Dermatitis 42, 251±259. Roeleveld, C.G., van Ketel, W.G., 1975. Allergic reaction to a hair dye elicited by an ointment containing DMSO. Contact Dermatitis 1, 331. Schlede, E., Eppler, R., 1995. Testing for skin sensitisation according to the noti®cation procedure for new chemicals: the Magnusson and Kligman test. Contact Dermatitis 32, 1±4. Smith, C.K., Hotchkiss, S.A.M., in press. Allergic Contact Dermatitis: Chemical and Metabolic Mechanisms. Taylor and Francis, London. Warbrick, E.V., Dearman, R.J., Basketter, D.A., Kimber, I., 1999. In¯uence of application vehicle on skin sensitization to the isothiazolinone CMIT: an analysis using the local lymph node assay. Contact Dermatitis 41, 325±329. Warbrick, E.V., Dearman, R.J., Basketter, D.A., Kimber, I., 2000. Failure of vehicle matrix to in¯uence local lymph node assay responses to benzocaine. Contact Dermatitis 42, 164±165. WHO, 1996. Criteria for Classi®cation of Skin and Airway Sensitizing substances in the Work and General Environment. Report of a World Health Organization Working Group, January 1996. Wright, Z., Basketter, D.A., Blaikie, L., Cooper, K.J., Warbrick, E.V., Dearman, R.J., Kimber, I., in press. Vehicle eects on skin sensitizing potency of four chemicals assessment using the local lymph node assay. International Journal of Cosmetic Science.
www.elsevier.com/locate/foodchemtox
Commentary
Skin sensitisation, vehicle eects and the local lymph node assay D.A. Basketter a,*, G.F. Gerberick b, I. Kimber c a
SEAC Toxicology Unit, Unilever Research, Colworth House, Sharnbrook, Bedford MK44 1LQ, UK b Miami Valley Laboratories, Procter & Gamble Company, Cincinnati, OH, USA c Zeneca Central Toxicology Laboratory, Alderley Park Maccles®eld, Cheshire SK10 4TJ, UK Accepted 14 October 2000
Abstract Accurate risk assessment in allergic contact dermatitis is dependent on the successful prospective identi®cation of chemicals which possess the ability to behave as skin sensitisers, followed by appropriate measurement of the relative ability to cause sensitisation; their potency. Tools for hazard identi®cation have been available for many years; more recently, a novel approach to the quantitative assessment of potency Ð the derivation of EC3 values in the local lymph node assay (LLNA) Ð has been described. It must be recognised, however, that these evaluations of chemical sensitisers also may be aected by the vehicle matrix in which skin exposure occurs. In this article, our knowledge of this area is reviewed and potential mechanisms through which vehicle eects may occur are detailed. Using the LLNA as an example, it is demonstrated that the vehicle may have little impact on the accuracy of basic hazard identi®cation; the data also therefore support the view that testing ingredients in speci®c product formulations is not warranted for hazard identi®cation purposes. However, the eect on potency estimations is of greater signi®cance. Although not all chemical allergens are aected similarly, for certain substances a greater than 10-fold vehicle-dependent change in potency is observed. Such data are vital for accurate risk assessment. Unfortunately, it does not at present appear possible to predict notionally the eect of the vehicle matrix on skin sensitising potency without recourse to direct testing, for example by estimation of LLNA EC3 data, which provides a valuable tool for this purpose. # 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction It has been recognised for many years that the vehicle matrix in which a chemical allergen is encountered on the skin can have profound eects on the response elicited and the extent to which skin sensitisation will be acquired (e.g. Kligman, 1966; Magnusson and Kligman, 1970). This, then, must be an important consideration for the interpretation of skin sensitisation testing (Robinson et al., 1989; Basketter et al., 1996; Gerberick and Robinson, 2000). Moreover, vehicle eects are not uncommonly implicated as the basis for misleading results from diagnostic human patch testing (particularly false negatives) (Ljunggren and Moller, 1972; Roeleveld and van Ketel, 1975; LideÂn and Boman, 1988). These views and experiences are supported by a relatively sparse literature on skin sensitisation. For Abbreviations: DNCB, 2,4-dinitrochlorobenzene; LC, Langerhans cells; SLS, sodium lauryl sulfate; TNF-a, tumour necrosis factor a. * Corresponding author. Tel.: +44-1234-222236; fax: +44-1234222122. E-mail address: [email protected] (D.A. Basketter).
example, it has been demonstrated in human predictive testing that the vehicle is an important modulator of the response that can be elicited in sensitised individuals (Kligman, 1966). One decade later, this was also shown to be true for the induction of sensitisation, although the eects seen in this limited study were less remarkable (Marzulli and Maibach, 1976). In the guinea pig, it has been shown that the vehicle can substantially aect the rate of sensitisation, although this could not be related to the apparent bioavailability (Andersen et al., 1985). However, the view that the vehicle matrix is likely to have an important impact on the sensitising eect of a chemical is also based on the underlying knowledge of the biology of skin penetration of chemicals (reviewed in Smith and Hotchiss, 2001) and physical chemistry (Kirchner et al., 1997). What does the above mean in relation to the prediction of skin sensitisation hazard and the subsequent hazard assessment of sensitising chemicals? Clearly, the identi®cation of skin sensitisation hazards using predictive animals models must take this into consideration (Schlede and Eppler, 1995). Furthermore, where the vehicle may have an impact on the degree of skin sensitisation observed, then it is apparently modulating the potency
0278-6915/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved. PII: S0278-6915(00)00169-1
622
D.A. Basketter et al. / Food and Chemical Toxicology 39 (2001) 621±627
of the allergen. This is of paramount importance in risk assessment (Basketter et al., 1996; Gerberick and Robinson, 2000; Gerberick et al., 2001). Matrix eects are complicated, and unfortunately there exist at present no in vitro or in silico methods to predict the eect of a vehicle on the activity of a particular skin sensitiser. Traditional animal test methods using the guinea pig are also poorly suited to this purpose, due to their relative complexity and dependence on qualitative endpoints (Kimber and Basketter, 1997; Basketter et al., 2000a). However, the local lymph node assay (LLNA) holds out more promise in this respect and has been used to provide objective and quantitative information on the impact of vehicles on skin sensitisation (Kimber and Basketter, 1997; Dearman et al., 1999; Basketter et al., 2000a). Here we review the information available on vehicle eects in skin sensitisation and consider both the relevance to hazard identi®cation and its potential utility for the purposes of risk assessment. 2. Vehicle eects: mechanisms It is necessary to consider the ways in which a vehicle or formulation matrix may impact on the skin sensitising potency of a chemical allergen. It has often been assumed that the most important in¯uence that a change in matrix may have is to alter skin penetration and the eectiveness with which a chemical allergen gains access to the viable epidermis. It is certainly the case, with speci®c examples for skin sensitisation including the common allergen nickel (Fullerton et al., 1988), the transdermal drug triprolidine, the sensitising activity of which was aected substantially by the vehicle matrix in which skin exposure occurred (Robinson and Cruze, 1996) and assessment of the alteration of skin sensitisation activity secondary to dibutyl phthalate mediated modulation of ¯uorescein isothiocyanate penentration (Dearman et al., 1996). While there is no doubt penetration is a key variable in instances where the partition coecient of the chemical allergen prevents or limits access across the stratum corneum, it is not necessarily always the relevant or predominant mechanism. For instance, it has been demonstrated that cutaneous responses induced in mice by topical exposure to suboptimal concentrations of the chemical allergen 2,4-dinitrochlorobenzene (DNCB) can be enhanced signi®cantly by co-administration of the non-sensitising surfactant sodium lauryl sulfate (SLS) (Heylings et al., 1996). Under conditions where SLS was able to augment the stimulation of lymph node cell proliferative responses to DNCB, it failed to in¯uence the eciency with which the allergen was absorbed through the skin. These and other data suggest that there exists a variety of ways in which the delivery matrix may aect the acquisition of skin sensitisation.
A further eect of the vehicle might also be to alter the protonation state of the chemical, the protonated species having a sharply reduced ability to penetrate into the skin. pKa is the pH at which 50% of the substance will be ionised. So, for example, a substance whose pKa is 5 (e.g. a weak carboxylic acid), when placed in a vehicle matrix at pH 7 is much more charged and so will penetrate much less well than if skin contact were to occur in a vehicle matrix at pH 4. The proportion of a substance which is charged can be calculated from the Henderson±Hasselbach equation: pH pKa logA=HA Another important consideration is that the induction of skin sensitisation requires the activation of Langerhans cells (LC) (reviewed in Kimber et al., 2000). Upon encounter with a chemical allergen, a proportion of LC at the site of exposure are mobilised and are stimulated to migrate, bearing antigen, to draining lymph nodes. The migration and subsequent maturation of LC, as well as other cellular interactions and processes necessary for skin sensitisation, are initiated and regulated by epidermal cytokines. The epidermis is a rich source of cytokines, some of which are expressed constitutively, while others are expressed only following receipt of an appropriate signal. It has been demonstrated that exposure to skin allergens and skin irritants is associated with the induction or upregulation of many cytokines, including those that are known to be necessary for the development of skin sensitisation (reviewed in Grabbe and Schwarz, 1998; Basketter et al., 1999a). Among the mediators known to play important roles in the induction of cutaneous immune responses are proin¯ammatory cytokines such as interleukins 1 and 6 (IL-1 and IL-6) and tumour necrosis factor a (TNF-a). The expression of some of these cytokines, such as TNF-a, will be increased following various forms of dermal trauma. Indeed, it has been postulated that some small degree of dermal trauma (and the production of epidermal cytokines associated with it) is essential for the normal development of skin sensitisation (Grabbe and Schwarz, 1998; Kimber et al., 2000). The argument is therefore that one impact that the vehicle matrix may have on skin sensitisation is to in¯uence the extent to which cytokines necessary for LC migration and maturation and other cellular processes will be produced at the time of contact with the chemical allergen. Following this reasoning, it may well be that the frequently reported ability of co-administration of SLS to augment skin sensitisation (e.g. Kligman, 1966) is at least in part attributable to the ability of this chemical to provoke the production of relevant proin¯ammatory cytokines (Grabbe and Schwarz, 1998). A requirement for some degree of dermal trauma for the eective development of skin sensitisation is of course consistent with the danger hypothesis of immune activation, which
D.A. Basketter et al. / Food and Chemical Toxicology 39 (2001) 621±627
in its simplest form argues that adaptive immune responses will be generated only if there is a certain degree of cell or tissue damage (Matzinger, 1994; McFadden and Basketter, 2000). In addition to possible eects on skin penetration and the provision of danger signals, it is possible that vehicles may in¯uence the eciency of skin sensitisation in other ways, including modulation of the metabolic activation of a prohapten, or metabolic inactivation of the hapten itself. Among other possible eects are those that will in¯uence the evaporation of a chemical allergen from the skin surface, impact a chemical phenomenon, such as allergen polymerisation (BjoÈrkner and Niklasson, 1984) or will aect in some way the formation of stable haptenprotein conjugates. 3. Vehicle eects: hazard identi®cation Although the application of predictive tests for skin sensitisation hazard identi®cation may vary to some extent globally, as a general rule the primary purpose is to discriminate between those chemicals which present a signi®cant hazard and those which do not (EC, 1996; WHO, 1997). Classi®cation as a skin sensitiser is a binary decision and as such, the key question relating to the eect of a vehicle on a predictive test is whether the matrix in which a chemical is presented will cause a change of such classi®cation. In theory, this must be possible; for example where a sensitiser gives a positive result in the guinea pig maximisation test with 35% of the test animals being sensitised, and so is just above the EC classi®cation limit (EC, 1996), then testing in a different vehicle which reduces skin penetration may give rise to a result of say 20% positive, which is below the classi®cation limit of 30%. However, there is no evidence that this is in practice a real problem, either from our data or from the following consideration. It is our experience that the distribution of responses in guinea pig tests, particularly the popular guinea pig maximisation test is not even, being rather polarised to the extremes (Fig. 1). In this analysis of 259 chemicals (taken from Cronin and Basketter, 1994), there was a preponderance of materials in the non-sensitising and strongly sensitising categories compared with the weak and moderate categories. It is in the latter group where vehicle eects are more likely to have an impact on regulatory classi®cation. This skewed distribution has the practical eect of reducing the risk of false negatives. This is con®rmed analysis of data submitted to the regulatory authority in the UK, which showed that only a very small proportion of submitted sensitisation tests fell close to the classi®cation borderline (Elliott-Minty and Evans, 1999). Published data on the impact of vehicles on hazard identi®cation in traditional guinea pig tests are very limited; to the best of our knowledge,
623
no substantial published information on this topic is available. The LLNA is a newer method which has been validated in terms of hazard identi®cation of skin sensitisation (NIH, 1999); it has been accepted as valid by the European Centre for the Validation of Alternative Methods (ECVAM) and is recommended as the preferred method (ECVAM, 2001). In contrast to guinea pig tests, the LLNA has been examined more carefully in relation to the eects of vehicles. The results indicate that hazard identi®cation is not compromised to a signi®cant extent by the choice of vehicle (Lea et al., 1999; Warbrick et al., 1999; Wright et al., 2001). The data for seven chemicals tested in the LLNA in a range of recommended vehicles revealed that the vehicle matrix had no impact on classi®cation, although it should be noted that for one substance/ vehicle combination (dimethylaminopropylamine in propylene glycol) the dataset was not conclusive (Table 1). Furthermore, these data also support the view that testing ingredients in speci®c product formulations is not warranted for hazard identi®cation purposes. It is worth mentioning at this point that advice on the choice of vehicles in relation to predictive tests for the identi®cation of skin sensitisation hazard is fairly limited. In the OECD Guideline 406 on skin sensitisation (OECD, 1992), the only mention of vehicles is in relation to the Buehler guinea pig test, where it is stated that, ``For water soluble materials, it is appropriate to use water or a dilute non-irritating solution of surfactant as a vehicle. For other test materials 80% ethanol/water is preferred for induction and acetone for challenge.'' For the LLNA, more speci®c guidance has been given (Kimber and Basketter, 1992); this has been reiterated in material related to the formal validation of the LLNA (NIH, 1999; Gerberick et al., 2001) and in the draft OECD Guideline (OECD, 2000). However, it still a dicult matter to select a vehicle system and it relies on the experimenter to balance a range of considerations, including the solubility and chemical compatiblity of the test substance, the extent to which local irritant and toxic eects are produced and the likelihood of systemic toxicity. Clearly, these considerations will have to be judged on a case by case basis but, for example, local severe irritation/necrosis at the treatment site should be avoided. An additional criterion might be to make constructive use of the recommendation in the draft OECD Guideline for the LLNA that body weights are taken at the start and the end of the experiment. Thus, to the clear advice already available for vehicle selection in the LLNA (Kimber and Basketter, 1992), it can be added that evidence of systemic toxicity, as demonstrated by a greater than 10% fall in body weight in the particular dose group, normally might invalidate the data for that group. No doubt a similar general consideration might be applied to the existing guinea pig tests.
624
D.A. Basketter et al. / Food and Chemical Toxicology 39 (2001) 621±627
Fig. 1. 259 chemicals tested in the guinea pig maximization test were placed into one of four categories, non-sensitising, weakly, moderately or strongly sensitising. The histogram shows the frequency (expressed as %) with which each category occurred.
Table 1 Classi®cation of allergens in the LLNAa,b Chemical allergen
Dimethylaminopropylamine 1,4-Dihydroquinone Cinnamic aldehyde (Chloro)methylisothiazolinone Dibromodicyanobutane Isoeugenol Benzocaine
Classi®cation as a skin sensitiser in a range of vehiclesc
Reference
AOO
MEK
DMF
PG
DMSO
+ + + + + +
+ + + + + +
+ + + + + +
? + + + NDd + ND
+ + + + ND +
Wright et al. (2001) Lea et al. (1999) Wright et al. (2001) Warbrick et al. (1999) Wright et al. (2001) Wright et al. (2001) Warbrick et al. (2000)
a Eect of a range of ®ve vehicles on the outcome of EU-based classi®cation decisions for the skin sensitisation potential of seven chemicals is presented. b ``+'' indicates that the chemical would be formally classifed as a skin sensitiser and denoted R43 according to EU regulations, `` '' means the chemical would not be classi®ed and ``?'' indicates that the data is equivocal. c AOO, acetone/olive oil (4:1, v/v), MEK, methylethylketone, DMF, dimethylformamide, PG, propylene glycol, DMSO, dimethyl sulfoxide. d ND, not done.
4. Vehicle eects: risk assessment In the section above, the impact of vehicle was considered only in the situation where a binary judgement had to be made. However, where the data indicate that the chemical possesses skin sensitisation hazard, then further use of that chemical will depend on assessment
of the risk to human health that it represents. Of course the risk will depend on the nature, extent and duration of skin exposure, but it will depend also on the potency of the allergen (Robinson et al., 2000). Both of these aspects, potency and exposure, can be modulated by the vehicle matrix in which the sensitising chemical is encountered. Whatever the mechanism(s) through which
D.A. Basketter et al. / Food and Chemical Toxicology 39 (2001) 621±627
625
Table 2 Impact of ®ve dierent vehicles on derived LLNA EC3 valuesa Chemical allergen
1,4-Dihydroquinone Cinnamic aldehyde (Chloro)methylisothiazolinone Dibromodicyanobutane Dimethylaminopropylamine Isoeugenol Ethyleneglycol dimethacrylate
EC3 values (%) for each skin sensitiser in a range of vehiclesb
Reference
AOO
MEK
DMF
PG
DMSO
0.15 1.7 0.05 5.2 2.2 1.0 36
0.09 1.1 0.007 0.4 1.8 1.0 28
0.21 0.5 0.008 6.4 1.7 1.4 32
1.5 1.4 0.05 NDc >10 2.5 15
0.35 0.9 0.008 ND 3.2 0.9 34
Lea et al. (1999) Wright et al. (2001) Warbrick et al. (1999) Wright et al. (2001) Wright et al. (2001) Wright et al. (2001) Blaikie et al. (2000)
a
For each of seven skin sensitising chemicals, the concentration (%) required to stimulate a three fold increase in cell proliferation in the draining lymph node (the EC3 value) has been estimated from the LLNA dose±response curve. EC3 values for each sensitiser in each of ®ve vehicles are presented. b AOO, acetone/olive oil (4:1, v/v); MEK, methylethylketone; DMF, dimethylformamide; PG, propylene glycol; DMSO, dimethyl sulfoxide. c ND, not done.
this modulation occurs, the important consideration for risk assessment is to gain an understanding of the eect so that it can be evaluated accurately. In guinea pig tests, quantitative assessment of the eect of vehicles is relatively dicult. This is due in part to the inherent diculties associated with the technical aspects of the conduct of guinea pig tests, which tend to conspire to ensure that when the vehicle is changed, other variables, most notably the concentration, also have to be changed. This is largely a consequence of the use of the irritation potential of the chemical as a basis for dose selection. Also, the subjective nature of the endpoint (visual assessment of erythema) does not lend itself readily to quantitation. The normal endpoint measurement in the LLNA is an immunologically relevant quantitative dose response (Kimber and Dearman, 1991). These data can be interpolated to yield a quantitative measure of relative potency, calculated as the estimated concentration of the chemical necessary to stimulate a threefold increase in lymph node cell proliferation Ð the EC3 value (Basketter et al., 1999b). This measure is highly reproducible within and between laboratories (Loveless et al., 1996; Warbrick et al., 1999) and is stable over time (Dearman et al., 1998). For the assessment of the eect of vehicles, EC3 values have been derived for sensitising chemicals tested in several vehicles in the LLNA. Such studies have been completed for a range of allergen/vehicle combinations, the data being summarised in Table 2. The results show that for some skin sensitisers, the vehicle can have a substantial impact on potency. For other allergens, changing the vehicle appears to be of little importance, with, for example, cinnamic aldehyde, or the very weak allergen, benzocaine showing virtually no vehicle-dependent changes (Warbrick et al., 2000; Wright et al., 2001). What bene®t can be derived from such quantitative estimates and how can LLNA EC3 values be applied to
risk assessment? Risk assessment for skin sensitisation requires information on potency and the EC3 value has been shown not only to be a useful measure for this purpose (Kimber and Basketter, 1997; Hilton et al., 1998) but also to have a close relationship with the apparent potency in humans (Basketter et al., 2000a; Gerberick et al., 2001). Using such an approach, the relative potency of chemicals in dierent vehicles can be compared directly using LLNA EC3 data. The vehicle may, for example, be similar in nature to the product matrix in which the skin will be exposed to the sensitising chemical. Completion of the risk assessment will require knowledge of (a) sensitising chemical(s) for which skin exposure is well documented, whose potency is understood and where the consequent (lack of) allergic contact dermatitis is known to act as points of comparison. Such chemicals are known as a benchmarks; several benchmarks may be necessary to eect accurately the risk assessment of an unknown sensitising chemical. However, the stability of results in the LLNA means that historic data, or published data from other laboratories can be utilised for this purpose. The ultimate consequence of this is that there will be a reduction in the need for animal testing, while at the same time, improved quantitative risk assessments can be made. Speci®cally, this will eliminate the need for evaluation of product formulations for the assessment of skin sensitisation potential. 5. Conclusion There has been a recognition for many years that vehicles can have important eects on the acquisition of skin sensitisation as well as on its clinical manifestation as allergic contact dermatitis (Kligman, 1966; Magnusson and Kligman, 1970; Marzulli and Maibach, 1976). It was noted in the last-mentioned paper that ``. . .few speci®c data are available on two experimental conditions
626
D.A. Basketter et al. / Food and Chemical Toxicology 39 (2001) 621±627
that may aect results obtained in skin sensitization studies: the vehicle used for evaluating the putative allergen. . .''. More than 20 years on, a similar statement might be made. The reality is that during this period there has been very little new information on the mechanisms through which vehicles might aect skin sensitisation, nor is there currently available a larger body of data qualitatively or quantitatively examining the eect of vehicles. The development of the LLNA as a new method for the identi®cation and characterisation of skin sensitisation potential has provided new possibilities (Kimber and Basketter, 1997). The quantitative nature of LLNA dose responses enabled us for the ®rst time to make meaningful measurements of relative potency (Basketter et al., 2000a; Gerberick et al., 2001). Once potency could be reliably and indeed easily quanti®ed, it has proven relatively simple to investigate the impact of vehicle. It is clear from the information presented here and elsewhere (Lea et al., 1999; Warbrick et al., 1999; Blaikie et al., 2000; Wright et al., 2001) that, while hazard identi®cation may be essentially unaected, the vehicle can have a substantial and unpredictable eect on the potency of an allergen. Indeed, from the work to date, there are no readily discernible relationships between a vehicle physicochemical characteristics and their eect of skin sensitisation. Thus, although the data indicate that the testing of in a range of vehicles or in product formulations for the speci®c identi®cation of the expression of a known sensitising hazard is unlikely to be necessary, the results have important consequences for risk assessment; because of its quantitative endpoint and high reproducibility, the LLNA provides a valuable tool through which vehicle eects on chemical sensitisers may be measured quantitatively. Appropriate use of the LLNA may thus lead to an improvement in risk assessment.
References Andersen, K.E., Carlsen, L., Egsgaard, H., Larsen, E., 1985. Contact sensitivity and bioavailability of chlorocresol. Contact Dermatitis 13, 246±251. Basketter, D.A., Gerberick, G.F., Robinson, M., 1996. Risk assessment. In: Kimber, I., Maurer, Th. (Eds.), The Toxicology of Contact Hypersensitivity. Taylor and Francis, London, pp. 152±164. Basketter, D.A., Gerberick, G.F., Kimber, I., Willis, C., 1999a. The Toxicology of Contact Dermatitis. Wiley and Sons, Chichester. Basketter, D.A., Lea, L.J., Cooper, K., Dickens, A., Briggs, D., Pate, I., Dearman, R.J., Kimber, I., 1999b. A comparison of statistical approaches to derivation of EC3 values from local lymph node assay dose responses. Journal of Applied Toxicology 19, 261±266. Basketter, D.A., Blaikie, L., Dearman, R.J., Kimber, I., Ryan, C.A., Gerberick, G.F., Harvey, P., Evans, P., White, I.R., Rycroft, R.J.G., 2000a. Use of the local lymph node assay for the estimation of relative contact allergenic potency. Contact Dermatitis 42, 344±348. Basketter, D.A., Evans, P., Gerberick, G.F., Kimber, I., 2000b. Chemical allergy: Estimation of potency, thresholds and risk assessments. Comments in Toxicology 7, 79±89.
BjoÈrkner, B., Niklasson, B., 1984. In¯uence of the vehicle on the elicitation of contact allergic reactions to acrylic compounds in the guinea pig. Contact Dermatitis 11, 268±278. Blaikie, L., Basketter, D.A., Warbrick, E.V., Dearman, R.J., Kimber, I., 2000. Utility of the LLNA for the investigation of the impact of vehicle matrix on sensitisation potency. Toxicological Sciences 54, 121. Cronin, M.T.D., Basketter, D.A., 1994. Multivariate QSAR analysis of a skin sensitization database. SAR and QSAR in Environmental Research 2, 159±179. Dearman, R.J., Cumberbatch, M., Hilton, J., Clowes, H.M., Fielding, I., Heylings, J.R., Kimber, I., 1996. In¯uence of dibutyl phthalate on dermal sensitization to ¯uorescein isothioscyanate. Fundamental and Applied Toxicology 33, 24±30. Dearman, R.J., Hilton, J., Evans, P., Harvey, P., Basketter, D.A., Kimber, I., 1998. Temporal stability of local lymph node assay responses to hexyl cinnamic aldehyde. Journal of Applied Toxicology 18, 281±284. Dearman, R.J., Basketter, D.A., Kimber, I., 1999. The local lymph node assay: use in hazard and risk assessment. Journal of Applied Toxicology 19, 299±306. EC (European Community), 1996. Dangerous Substances Directive, 22nd Adaptation to Technical Progress (96/54/EC). Ocial Journal of the European Communities 199, L248. ECVAM, in press. Statement on the validity of the Local Lymph Node Assay for skin sensitisation testing. ECVAM Joint Research Centre, European Commission, Ispra, Italy, 21 March 2000. ATLA. Elliott-Minty, C., Evans, P., 1998. Regulatory aspects of testing new industrial chemicals for skin sensitisation potential. Human and Experimental Toxicology 17, 498. Fullerton, A., Andersen, J.R., Hoelgaard, A., 1988. Permeation of nickel through human skin in vitro Ð eects of vehicles. British Journal of Dermatology 118, 509±516. Gerberick, G.F., Robinson, M.K., 2000. A skin sensitization risk assessment approach for evaluation of new ingredients and products. American Journal of Contact Dermatitis 11, 65±73. Gerberick, G.F., Ryan, C.A., Kimber, I., Dearman, R.J., Lea, L.J., Basketter, D.A., 2001. Local lymph node assay validation assessment for regulatory purposes. American Journal of Contact Dermatitis 11, 13±18. Grabbe, S., Schwarz, T., 1998. Immunoregulatory mechanisms involved in elicitation of allergic contact hypersensitivity. Immunology Today 19, 37±44. Heylings, J.R., Clowes, H.M., Cumberbatch, M., Dearman, R.J., Fielding, I., Hilton, J., Kimber, I., 1996. Sensitisation to 2,4-dinitrochlorobenzene: in¯uence of vehicle on absorption and lymph node activation. Toxicology 109, 57±65. Hilton, J., Dearman, R.J., Harvey, P., Evans, P., Basketter, D.A., Kimber, I., 1998. Estimation of relative skin sensitizing potency using the local lymph node assay: a comparison of formaldehyde and glutaraldehyde. American Journal of Contact Dermatitis 9, 29±33. Kimber, I, Basketter, D.A., 1992. The murine local lymph node assay; collaborative studies and new directions: a commentary. Food and Chemical Toxicology 30, 165±169. Kimber, I., Basketter, D.A., 1997. Contact sensitisation: a new approach to risk assessment. Human and Ecological Risk Assessment 3, 385±395. Kimber, I., Dearman, R.J., 1991. Investigation of lymph node cell proliferation as a possible immunological correlate of contact sensitizing potential. Food and Chemical Toxicology 29, 125±129. Kimber, I., Cumberbatch, M., Dearman, R.J., Bhushan, M., Griths, C.E., 2000. Cytokines and chemokines in the initiation and regulation of epidermal Langerhans cell mobilization. British Journal of Dermatology 142, 401±412. Kirchner, L.A., Moody, R.P., Doyle, E., Bose, R., Jerey, J., Chu, I., 1997. The prediction of skin permeability using physicochemical data. ATLA 25, 359±370.
D.A. Basketter et al. / Food and Chemical Toxicology 39 (2001) 621±627 Kligman, A.M., 1966. The identi®cation of contact allergens by human assay. II. Factors in¯uencing the induction and measurement of allergic contact dermatitis. Journal of Investigative Dermatology 47, 375±392. Lea, L.J., Warbrick, E.V., Dearman, R.J., Harvey, P., Kimber, I., Basketter, D.A., 1999. The impact of vehicle on assessment of relative skin sensitisation potency of 1,4-dihydroquinone in the local lymph node assay. Food and Chemical Toxicology 10, 213±218. LideÂn, C., Boman, A., 1988. Contact allergy to colour developing agents in the guinea pig. Contact Dermatitis 19, 290±295. Ljunggren, B., Moller, H., 1972. Eczematous contact allergy to chlorhexidine. Acta Dermatologica Venereologica 52, 308±310. Loveless, S.E., Ladics, G.S., Gerberick, G.F., Ryan, C.A., Basketter, D.A., Scholes, E.W., House, R.V., Hilton, J., Dearman, R.J., Kimber, I., 1996. Further evaluation of the local lymph node assay in the ®nal phase of an international collaborative trial. Toxicology 108, 141±152. Magnusson, B., Kligman, A.M., 1970. Allergic Contact Dermatitis in the Guinea Pig. Charles C. Thomas, Spring®eld, IL. Marzulli, F.N., Maibach, H.I., 1976. Eects of vehicles and elicitation concentration in contact dermatitis testing. Contact Dermatitis 2, 325±329. Matzinger, P., 1994. Tolerance, danger and the extended family. Annual Review of Immunology 12, 991±1045. McFadden, J.P., Basketter, D.A., 2000. Contact allergy, irritancy and danger. Contact Dermatitis 42, 123±127. NIH (National Institutes of Health), 1999. The Murine Local Lymph Node Assay: A Test Method for Assessing the Allergic Contact Dermatitis Potential of Chemicals/Compounds. OECD, Paris. OECD, 1992. OECD Guidelines for Testing of Chemicals. Guideline 406, Skin Sensitisation. Organisation for Economic Cooperation and Development, Paris. OECD, 2000. Draft Test Guideline for the Local Lymph Node Assay. Organisation for Economic Cooperation and Development, Paris.
627
Robinson, M.K., Cruze, C.A., 1996. Preclinical skin sensitization testing of antihistamines: guinea pig and local lymph node assay responses. Food and Chemical Toxicology 34, 495±506. Robinson, M.K., Stotts, J., Danneman, P.J., Nusair, T.L., Bay, P.H.S., 1989. A risk assessment process for allergic contact sensitization. Food and Chemical Toxicology 27, 479±489. Robinson, M.K., Gerberick, G.F., Ryan, C.A., McNamee, P., White, I.R., Basketter, D.A., 2000. The importance of exposure estimation in the assessment of skin sensitization risk. Contact Dermatitis 42, 251±259. Roeleveld, C.G., van Ketel, W.G., 1975. Allergic reaction to a hair dye elicited by an ointment containing DMSO. Contact Dermatitis 1, 331. Schlede, E., Eppler, R., 1995. Testing for skin sensitisation according to the noti®cation procedure for new chemicals: the Magnusson and Kligman test. Contact Dermatitis 32, 1±4. Smith, C.K., Hotchkiss, S.A.M., in press. Allergic Contact Dermatitis: Chemical and Metabolic Mechanisms. Taylor and Francis, London. Warbrick, E.V., Dearman, R.J., Basketter, D.A., Kimber, I., 1999. In¯uence of application vehicle on skin sensitization to the isothiazolinone CMIT: an analysis using the local lymph node assay. Contact Dermatitis 41, 325±329. Warbrick, E.V., Dearman, R.J., Basketter, D.A., Kimber, I., 2000. Failure of vehicle matrix to in¯uence local lymph node assay responses to benzocaine. Contact Dermatitis 42, 164±165. WHO, 1996. Criteria for Classi®cation of Skin and Airway Sensitizing substances in the Work and General Environment. Report of a World Health Organization Working Group, January 1996. Wright, Z., Basketter, D.A., Blaikie, L., Cooper, K.J., Warbrick, E.V., Dearman, R.J., Kimber, I., in press. Vehicle eects on skin sensitizing potency of four chemicals assessment using the local lymph node assay. International Journal of Cosmetic Science.