- Email: [email protected]
New approaches to the assessment of eye and skin irritation
Toxicology Letters, 641’65(1992) 157-164 0 1992 Elsevier Science Publishers B.V., All rights reserved 03784274/‘92/$5.00
157
New approaches to the assessment of eye and skin irritation George Calvin Procter 8 Gamble European Technical Center, Strombeek-Beuer
(Belgium)
Key words: Eye irritation; Skin irritation
SUMMARY Assessment of eye and skin irritation potential is an important part of any comprehensive toxicology programme for new chemicals and consumer products. The original skin and eye irritation assessment methods described by Draize et al. are still widely used for regulatory purposes with only slight modification. These methods have been reevaluated in the light of the modern approach to toxicological testing requiring refinement of protocols, reduction in animal numbers, and replacement of in viva tests by in vitro assays. A refinement in the original Draize skin irritation test accepted by regulatory authorities is reduction of test substance exposure time from 24 to 4 h. This makes the test less stressful. The Low Volume Eye Irritation Test is also a less stressful, more predictive refinement of the Draize eye irritation procedure. Many in vitro models are being developed for use in the assessment of eye and skin irritation potential in the hope that the animal tests may some day be replaced. In order for alternative tests to be incorporated into safety assessment, they must undergo a defined evaluation that allows toxicologists to determine whether the new procedures provide relevant information. As the validity of the methods is determined, the new procedures can be incorporated into safety assessment using a tier assessment process. The first step in tier testing is a thorough analysis of existing in-house and published data on the test substance or similar material. A review of physical/chemical characteristics can also provide valuable information. If more data are needed in order to complete the safety assessment, the next step is to evaluate the test substance in in vitro assays. The in vitro data may provide sufficient information to make a safety assessment. If additional information is required, then in viuo testing may be conducted using a limited number of animals or human clinical studies. If alternative procedures are to be ultimately accepted for use in safety assessment, regulatory authorities must become active participants in programmes designed to assess and validate the new tests.
INTRODUCTION
The assessment of acute eye and skin irritation is a part of any comprehensive toxicology programme on new and established chemicals and proCorrespondence to: G. Calvin, Procter & Gamble European Technical Center, Temselaan 100, 1853 Strombeek-Bever, Belgium
158
ducts. This is understandable. The skin is the major contact organ with the outside world. It is our primary defence mechanism against the external environment, and its response to chemical insult is a key factor to be determined in making both hazard and risk assessments. The eyes are important, vulnerable sense organs and it is essential that their response to chemical insult be understood. Thus, there is a need for chemical manufacturing companies and user companies to understand the hazard presented by their respective product categories. EEC regulations focus on specific test methods for both skin and eye irritation assessment, with passing reference to the use of human data. With the continued entry of new chemicals and products into commerce, there is clearly a need to conduct animal studies to assess their irritation potential. The established test methods are useful but have several limitations which have led to considerable effort being directed towards the development of alternative methods, which allow for Reduction in animal numbers, Refinement in the methods used and Replacement of current methods with in vitro alternatives.
PRESENT METHODOLOGY
The present regulations governing the evaluation of the eye and skin irritation potential of chemicals and preparations for classification purposes within the EEC require the Draize methods to be used. The regulators who developed the protocols in Annex V of the 6th Amendment to the Directive on Dangerous Substances 111opted for the Draize methods, which were the best methods then available. When the Directive was being developed in the 196Os, there were essentially no viable alternatives. It should also be kept in mind that the protocols in Annex V of the 6th Amendment were largely adaptations of the OECD Test Guidelines 121.Today we have an appreciation of the limitations of these methods. I see three fundamental problems in the application of these protocols, and the other requirements of the 6th Amendment, in today’s world. Firstly, the science of toxicology has advanced considerably. There are now alternative methods that can and are being used routinely in the assessment of eye and skin irritation. Secondly, although the Directive on Dangerous Substances allows for the use of human data for regulatory purposes there is a lack of guidance as to how data generated in human tests can be used. Thirdly, the test methods used to assess eye and skin irritation overpredict human irritation responses. I am not going to dwell at length on present methodologies. These have been fully reviewed in the ECETOC monographs [3,41 and I would much rather spend my time on reviewing the “New Approaches to the Assessment of the Eye and Skin Irritation”.
159 NEW APPROACHES TO THE ASSESSMENT OF EYE AND SKIN IRRITATION
The science of toxicology has made considerable progress since the 1960s when the present protocols for the assessment of eye and skin irritation were first incorporated into worldwide regulatory schemes. This, combined with a changing attitude towards animal testing, has led to: 1. Refinement in the present methodology in which more meaningful data are obtained in less stressful in uiuo protocols. 2. Reduction in animal numbers, in generating data in an efficient manner. 3. Replacement whereby animal tests are replaced with in vitro systems. A. Eye Irritation (i) In uiuo methods: The most promising alternative method for the assessment of eye irritation potential is the Low Volume Eye Test (LVET) [5]. The value of the Low Volume Eye Test as a predictor of a human response to an eye insult has been reported by Freeberg et al. [61.In this study the irritation response in the eyes of human volunteers exposed to various concentrations of detergent, soap, shampoo and fabric softening products was compared to the response observed in the rabbit in the Draize test and the LVET. The LVET gave a better correlation with actual human responses than did the Draize test. This work, from actual testing in human volunteers, has been substantiated by monitoring the eye irritation responses in consumers accidentally exposed to household products 171.The human exposure data presented in this study were reported consumer eye incidents with Procter & Gamble products, and the reported eye accidents with employees during manufacture. The comparative animal data were derived from the rabbit Draize Test and the LVET. The main conclusion reached was that each of the animal methods overestimates the human response to accidental eye exposure. The LVET is more closely related to human experience with consumer household cleaning products. We believe that the LVET is a better test than Draize for assessing the eye irritation potential of soaps, detergents and household cleaning products, and personal care products. Our rationale for this is that the LVET more accurately predicts the human response and that the test is less stressful to animals than the Draize test. The LVET has proven to be an acceptable model for assessing the eye irritancy potential of soap and detergent products 1141. It would seem to be important that products and chemicals in other industrial categories be evaluated in order to check the wider applicability of the LVET. Clearly its application across a broad range of product types could lead to better prediction of human risk and a clear advance in refinement of in uiuo methodology. (ii) In vitro methods: Dr. Leon Bruner and his group at Procter & Gamble
160
have carried out an evaluation of seven in vitro test methods that may be useful for assessing eye irritancy potential [81. In their evaluation the group examined 17 test materials covering chemicals, household cleaners, hand soaps, dish-washing liquids, shampoos and liquid laundry products. In uiuo ocular irritation scores for the seventeen test materials were obtained from historical LVET data. In a review of this nature there is simply not the time to go into the test methods used in any detail. This can be obtained from the reference quoted. This initial work, and subsequent studies indicate that the silicon microphysiometer, the luminescent bacteria toxicity test, and the neutral red assay hold out some promise as useful in vitro alternatives. These methods all rely on the presentation of the test material in the assay system in an aqueous buffered medium. More recent work by Osborne et al. [9] has led to the development of an in vitro assay for eye and skin irritation assessment, which allows for solids and materials not suitable for testing in an aqueous medium to be evaluated. These methods are now being used for safety assessment of test substances where available data supports their use. In order to gain broader acceptance, the assays next need to be evaluated in well-designed, independent, peer-reviewed validation programmes. These studies will provide the expanded range of data that will be necessary better to define the predictive capacity of the assays. B. Skin Irritation (i) In uiuo tests: As was discussed in the ECETOC Monograph 141there are a number of alternative in uiuo models to the Draize Test in rabbits. Although the Annex V test methods allow data to be generated for classification purposes in other species, there has been no universal move away from the rabbit. Other species evaluated have included the guinea pig, mouse, rat, monkey and beagle dog. Alternative skin grading parameters to erythema and oedema have been developed. These have included cutaneous blood flow by Laser Doppler Flowmetry, infra-red detection of skin temperature and skin thickness assessment. Such technologies are innovative but no consensus exists on their general applicability. Alternatives to the Draize occluded and semi-occluded patch systems have included an open application cup, an occlusive chamber, skin painting and intradermal injection. Again no single method has uniform applicability and the closed patch remains the method of choice for many in uiuo tests. The species of ultimate concern is, of course, man. There is no reason why, with appropriate attention being paid to ethical considerations, skin irritation tests could not be carried out in man. Importantly, no extrapolations are necessary and, also, no animal species are involved. This is the perfect example of the replacement principle.
161
My colleague, Dr. Rosemarie Osborne, has recently carried out a review [IO] of the comparative animal (rabbit) and human data. These data clearly show that rabbit skin is much more reactive than human skin, and that responses in the rabbit cannot accurately be used to predict human response. Put another way, a test in rabbit skin will inevitably yield a number that can be used for classification purposes. Such a classification will most probably have no meaning in terms of human risk assessment. It should be noted that the text of the 6th Amendment makes reference to the use of data generated in humans, but no guidance is given to the quantitative and qualitative nature of what would constitute acceptable human data. It is suggested that appropriately designed human skin patch tests and productuse tests are an acceptable approach for assessing human response. (ii) In vitro tests: The ECETOC Monograph on “Skin Irritation” 141gives an adequate overview of the methods that are available for the in vitro assessment of skin irritation. I do not think that it would be helpful to repeat that information by listing test systems. Rather, I would like to illustrate the potential of one system which is the “Human Skin Keratinocyte Cytotoxicity/Neutral Red (KC) Assay” 111-131 and I am grateful to my colleague, Dr. Rosemarie Osborne, for allowing me to use her data. The basis of the assay system is the ability of cultured human keratinocytes to take up the vital dye neutral red. Viable cells will incorporate the dye: damaged cells will not. Thus, the test system measures the response of human epidermis derived epithelial basal cells (keratinocytes) to test materials applied in culture. From dose-inhibition curves the concentration of test material that inhibits neutral red (NR) uptake by 50% is calculated. This is the NR50 value. The NR50 value for different test materials is a measure of the in vitro cytotoxicity potential of those materials, which is then taken to be an indicator of ultimate skin irritation potential. In validating the system a standard curve was constructed for the keratinocyte assay using human skin patch test data for a range of chemical irritants covering a wide range of irritation potential and cytotoxicity. The method showed a promising correlation with human skin irritation for selected chemicals and ingredients that cover a wide range of irritation potential. For other chemicals, such as acids and alkalis, the KC assay underestimated in uiuo irritation. This discrepancy is likely due to the buffering capacity of the cell culture medium, that reduces the effect of extremely high or low pH on the cells. The major use of preclinical skin irritation testing at P&G is to provide safety support data prior to human skin exposures. For example, we would like to ensure that panelists’ skin is not unduly irritated in a clinical patch test (an irritation score of I 2 out of a possible total of 5). Thus, one valuable use of KC assay data is to confirm the selection of doses of test substances
162
before a clinical patch test. Based on the KC assay score for a new material, potential patch test doses can be positioned relative to a standard curve. For example, for a test material with a cytotoxicity score [log (l/NR50)] of 3, the log (score/dose) value is approximately 1.5. To select a dose of material that would produce a patch test score of no greater than 2, the highest dose in the human patch test should be 0.002%. The effects of various product formulations which contained surfactants were tested in the KC assay. A poor correlation between the cytotoxicity data and human skin irritation was obtained for these products. For example, the bar soaps were similar in human skin irritation, but differed by 2 orders of magnitude in cytotoxicity in vitro. In contrast, the shampoos differed by nearly 3 orders of magnitude in skin irritation, but were similarly cytotoxic. One explanation for the lack of correlation is that, in the in vitro assay the presentation of the test products differs fundamentally from that which occurs to skin in Go. In a patch test, epidermal basal cells are likely exposed to, at most, a dilution of only a few of the formulation components. In the KC assay, the keratinocytes are exposed to dilutions of the entire product formulation. Nonetheless, KC assay data may be useful for determination of whether a new formulation is far more cytotoxic than previous formulations of a similar type. These data illustrate two points made in the ECETOC monograph on skin irritation. Firstly, it is essential that any in vitro model be validated for the class of chemical being evaluated and in the laboratory performing the test. Secondly, and this is the converse, if careful thought is not given to the choice and use of specific assays one can get meaningless and confusing data. It is preferable to rely on systems that model important aspects of human skin response to the materials tested. In this way, data from the in vitro system can be interpreted based on the physiology of the skin, and applied directly to safety assessment of potential human skin exposure. APPLICATION IN PRACTICE
The new in uiuo and in vitro methods for eye and skin irritation assessment should become part of the review process which takes place before any exposure of humans to test materials. A first step is a review of existing data from all available sources. Information on physicochemical characteristics, such as pH and buffering capacity, may indicate the potential for eye or skin irritancy. A structure-activity assessment against chemicals of known irritancy potential may provide additional perspective. In some cases, skin irritancy data may become available from other tests which are designed to assess the dermal safety of topically applied chemicals. Irritancy data might come available from guinea pig skin sensitization
163
studies, or from rabbit percutaneous toxicity studies. .As the design of such studies is not primarily directed towards assessment of skin irritancy, care must be taken in using data from such studies. Concentrations of test materials tend to be low in sensitization studies but set at the highest tolerable dose in percutaneous toxicity studies. An overall assessment of this existing skin irritation data may be sufficient for the selection of a dose or exposure conditions in a clinical skin irritation test or a product use study. For many products and chemicals, an adequate skin irritation assessment can be made without additional testing. In some cases, however, additional in vitro analysis may be required to position irritancy potential relative to chemicals and products of known potency. A stepwise procedure is recommended in the ECETOC monograph on “Skin Irritation” [41. IRRITATION ASSESSMENT AND REGULATION
As has been explained earlier, regulations worldwide call for Draize Tests for skin and eye irritation. Regulators are interested in progress in this area, and need to provide support for the development and validation of alternative methods in order that a scheme can be developed to give industry the option of using the new methods such as 1 have described. We cannot stay forever with the Draize Tests. The omens are, I believe, good. The rational arguments over the LD50 test have led to its virtual elimination with no loss in the thoroughness of the hazard evaluation process. The up-and-down method, the limit test and the non-lethality method provide an adequate basis for assessing acute lethality. This provides a good model and hopefully the scientific-regulatory dialogue will see Draize Tests eventually replaced. SUMMARY AND CONCLUSIONS
1. Essentially all regulatory authorities in the OECD countries require a Draize Test to be carried out for the assessment of eye and skin irritation, and subsequent classification. 2. These methods are out-dated. The regulatory process has not allowed the easy incorporation of new methodology into regulation. 3. Modern approaches to eye and skin irritation assessment require that attention be paid to the three “R”s of Refinement, Reduction and Replacement in the design of protocols developed to assess toxicity. 4. Alternative in uiuo methods exist for both eye and skin irritation assessment. Excellent progress has been made in developing in vitro
164
methods that may significantly reduce our dependence on animal based tests. 5. Toxicologists should arrange for peer review and publication of the results of research into in uiuo and in vitro alternative methods in order to make their methods available for broad evaluation. 6. Regulators should encourage such work and actively embrace the principle that valid data for classification purposes can be obtained from protocols other than those mandated in Annex V. ACKNOWLEDGMENTS
I would like to acknowledge the help provided by Dr. Rosemarie Osborne, Dr. Leon Bruner and Dr. K.A. Kohrman in the preparation of this manuscript and Mrs. Petra Bogaerts for typing it. REFERENCES 1 Legislation on Dangerous Substances, Classification and Labelling in the European Communities, Volume 1. Office for Official Publications of the European Communities, Graham & Trotman Ltd., Luxembourg and London, 1987. OECD Guidelines for Testing of Chemicals. The Organisation for Economic Co-operation and Development, 1981. Eye Irritation Testing, Monograph 11, European Chemical Industry Ecology and Toxicology Centre, Brussels, June 1988. Skin Irritation, Monograph 15, European Chemical Industry Ecology and Toxicology Centre, Brussels, July 1990. Griffith, J.F., Nixon, G.A., Bruce, R.D., Reer, P.J. and Barman, E.A. (1980) Dose-response studies with chemical irritants in the albino rabbit eye as a basis for selecting optimum testing conditions for predicting hazard to the human eye. Toxicol. Appl. Pharmacol. 55, 501313. 6 Freeberg, F.E., Nixon, G.A., Reer, P.J., Weaver, J.E., Bruce, R.D., Griffith, J.F. and Sanders, L.W. (1986) Human and rabbit eye responses to chemical insult. Fund. Appl. Toxicol. 7,626-234.1 7 Freeberg, F.E., Griffith, J.F., Bruce, R.D. and Bay, P.H.S. (1984) Correlation of animal test methods with human experience for household products. J. Toxicol. Cut. Ocular Toxicol. 3 (l), 53-64. 8 Bruner, L.H., Kain, D.J., Roberts, D.A. and Parker, R.D. (1991) Evaluation of seven in-vitro alternatives for ocular testing. Fund. Appl. Toxicol. 17,136149. 9 Perkins, M.A., Roberts, D.A. and Osborne, R. (1992) Human skin cell cultures for in vitro skin and eye irritancy assessments of neat test materials. The Toxicologist 12: 296. 10 Osborne, R., Comparison of rabbit and human skin responses to primary dermal irritants, Manuscript in preparation. 11 Osborne, R. and Perkins, M.A. (1991) In vitro skin irritation testing with human skin cell cultures, Toxicol. In Vitro 5 (5/6), 563-567. 12 Osborne, R. and Perkins, M.A. (1991) Evaluation of Human Skin Cell CulturesforIn Vitro Skin Irritancy Testing. In: A.M. Goldberg (Ed.)ln Vitro Toxicology: Mechanisms and New Toxicology. Alternative Methods in Toxicology, 8. Mary Ann Liebert, Publisher, New York. 13 Osborne R. and Perkins M.A. (1992) An approach for development of alternative test methods based on mechanisms of skin irritation, Food Chem. Toxicol. In press. 14 Inter Regulatory Agency Group Workshop on updating Eye Irritation Tests, September 26 and 27,1991, Washington D.C.
157
New approaches to the assessment of eye and skin irritation George Calvin Procter 8 Gamble European Technical Center, Strombeek-Beuer
(Belgium)
Key words: Eye irritation; Skin irritation
SUMMARY Assessment of eye and skin irritation potential is an important part of any comprehensive toxicology programme for new chemicals and consumer products. The original skin and eye irritation assessment methods described by Draize et al. are still widely used for regulatory purposes with only slight modification. These methods have been reevaluated in the light of the modern approach to toxicological testing requiring refinement of protocols, reduction in animal numbers, and replacement of in viva tests by in vitro assays. A refinement in the original Draize skin irritation test accepted by regulatory authorities is reduction of test substance exposure time from 24 to 4 h. This makes the test less stressful. The Low Volume Eye Irritation Test is also a less stressful, more predictive refinement of the Draize eye irritation procedure. Many in vitro models are being developed for use in the assessment of eye and skin irritation potential in the hope that the animal tests may some day be replaced. In order for alternative tests to be incorporated into safety assessment, they must undergo a defined evaluation that allows toxicologists to determine whether the new procedures provide relevant information. As the validity of the methods is determined, the new procedures can be incorporated into safety assessment using a tier assessment process. The first step in tier testing is a thorough analysis of existing in-house and published data on the test substance or similar material. A review of physical/chemical characteristics can also provide valuable information. If more data are needed in order to complete the safety assessment, the next step is to evaluate the test substance in in vitro assays. The in vitro data may provide sufficient information to make a safety assessment. If additional information is required, then in viuo testing may be conducted using a limited number of animals or human clinical studies. If alternative procedures are to be ultimately accepted for use in safety assessment, regulatory authorities must become active participants in programmes designed to assess and validate the new tests.
INTRODUCTION
The assessment of acute eye and skin irritation is a part of any comprehensive toxicology programme on new and established chemicals and proCorrespondence to: G. Calvin, Procter & Gamble European Technical Center, Temselaan 100, 1853 Strombeek-Bever, Belgium
158
ducts. This is understandable. The skin is the major contact organ with the outside world. It is our primary defence mechanism against the external environment, and its response to chemical insult is a key factor to be determined in making both hazard and risk assessments. The eyes are important, vulnerable sense organs and it is essential that their response to chemical insult be understood. Thus, there is a need for chemical manufacturing companies and user companies to understand the hazard presented by their respective product categories. EEC regulations focus on specific test methods for both skin and eye irritation assessment, with passing reference to the use of human data. With the continued entry of new chemicals and products into commerce, there is clearly a need to conduct animal studies to assess their irritation potential. The established test methods are useful but have several limitations which have led to considerable effort being directed towards the development of alternative methods, which allow for Reduction in animal numbers, Refinement in the methods used and Replacement of current methods with in vitro alternatives.
PRESENT METHODOLOGY
The present regulations governing the evaluation of the eye and skin irritation potential of chemicals and preparations for classification purposes within the EEC require the Draize methods to be used. The regulators who developed the protocols in Annex V of the 6th Amendment to the Directive on Dangerous Substances 111opted for the Draize methods, which were the best methods then available. When the Directive was being developed in the 196Os, there were essentially no viable alternatives. It should also be kept in mind that the protocols in Annex V of the 6th Amendment were largely adaptations of the OECD Test Guidelines 121.Today we have an appreciation of the limitations of these methods. I see three fundamental problems in the application of these protocols, and the other requirements of the 6th Amendment, in today’s world. Firstly, the science of toxicology has advanced considerably. There are now alternative methods that can and are being used routinely in the assessment of eye and skin irritation. Secondly, although the Directive on Dangerous Substances allows for the use of human data for regulatory purposes there is a lack of guidance as to how data generated in human tests can be used. Thirdly, the test methods used to assess eye and skin irritation overpredict human irritation responses. I am not going to dwell at length on present methodologies. These have been fully reviewed in the ECETOC monographs [3,41 and I would much rather spend my time on reviewing the “New Approaches to the Assessment of the Eye and Skin Irritation”.
159 NEW APPROACHES TO THE ASSESSMENT OF EYE AND SKIN IRRITATION
The science of toxicology has made considerable progress since the 1960s when the present protocols for the assessment of eye and skin irritation were first incorporated into worldwide regulatory schemes. This, combined with a changing attitude towards animal testing, has led to: 1. Refinement in the present methodology in which more meaningful data are obtained in less stressful in uiuo protocols. 2. Reduction in animal numbers, in generating data in an efficient manner. 3. Replacement whereby animal tests are replaced with in vitro systems. A. Eye Irritation (i) In uiuo methods: The most promising alternative method for the assessment of eye irritation potential is the Low Volume Eye Test (LVET) [5]. The value of the Low Volume Eye Test as a predictor of a human response to an eye insult has been reported by Freeberg et al. [61.In this study the irritation response in the eyes of human volunteers exposed to various concentrations of detergent, soap, shampoo and fabric softening products was compared to the response observed in the rabbit in the Draize test and the LVET. The LVET gave a better correlation with actual human responses than did the Draize test. This work, from actual testing in human volunteers, has been substantiated by monitoring the eye irritation responses in consumers accidentally exposed to household products 171.The human exposure data presented in this study were reported consumer eye incidents with Procter & Gamble products, and the reported eye accidents with employees during manufacture. The comparative animal data were derived from the rabbit Draize Test and the LVET. The main conclusion reached was that each of the animal methods overestimates the human response to accidental eye exposure. The LVET is more closely related to human experience with consumer household cleaning products. We believe that the LVET is a better test than Draize for assessing the eye irritation potential of soaps, detergents and household cleaning products, and personal care products. Our rationale for this is that the LVET more accurately predicts the human response and that the test is less stressful to animals than the Draize test. The LVET has proven to be an acceptable model for assessing the eye irritancy potential of soap and detergent products 1141. It would seem to be important that products and chemicals in other industrial categories be evaluated in order to check the wider applicability of the LVET. Clearly its application across a broad range of product types could lead to better prediction of human risk and a clear advance in refinement of in uiuo methodology. (ii) In vitro methods: Dr. Leon Bruner and his group at Procter & Gamble
160
have carried out an evaluation of seven in vitro test methods that may be useful for assessing eye irritancy potential [81. In their evaluation the group examined 17 test materials covering chemicals, household cleaners, hand soaps, dish-washing liquids, shampoos and liquid laundry products. In uiuo ocular irritation scores for the seventeen test materials were obtained from historical LVET data. In a review of this nature there is simply not the time to go into the test methods used in any detail. This can be obtained from the reference quoted. This initial work, and subsequent studies indicate that the silicon microphysiometer, the luminescent bacteria toxicity test, and the neutral red assay hold out some promise as useful in vitro alternatives. These methods all rely on the presentation of the test material in the assay system in an aqueous buffered medium. More recent work by Osborne et al. [9] has led to the development of an in vitro assay for eye and skin irritation assessment, which allows for solids and materials not suitable for testing in an aqueous medium to be evaluated. These methods are now being used for safety assessment of test substances where available data supports their use. In order to gain broader acceptance, the assays next need to be evaluated in well-designed, independent, peer-reviewed validation programmes. These studies will provide the expanded range of data that will be necessary better to define the predictive capacity of the assays. B. Skin Irritation (i) In uiuo tests: As was discussed in the ECETOC Monograph 141there are a number of alternative in uiuo models to the Draize Test in rabbits. Although the Annex V test methods allow data to be generated for classification purposes in other species, there has been no universal move away from the rabbit. Other species evaluated have included the guinea pig, mouse, rat, monkey and beagle dog. Alternative skin grading parameters to erythema and oedema have been developed. These have included cutaneous blood flow by Laser Doppler Flowmetry, infra-red detection of skin temperature and skin thickness assessment. Such technologies are innovative but no consensus exists on their general applicability. Alternatives to the Draize occluded and semi-occluded patch systems have included an open application cup, an occlusive chamber, skin painting and intradermal injection. Again no single method has uniform applicability and the closed patch remains the method of choice for many in uiuo tests. The species of ultimate concern is, of course, man. There is no reason why, with appropriate attention being paid to ethical considerations, skin irritation tests could not be carried out in man. Importantly, no extrapolations are necessary and, also, no animal species are involved. This is the perfect example of the replacement principle.
161
My colleague, Dr. Rosemarie Osborne, has recently carried out a review [IO] of the comparative animal (rabbit) and human data. These data clearly show that rabbit skin is much more reactive than human skin, and that responses in the rabbit cannot accurately be used to predict human response. Put another way, a test in rabbit skin will inevitably yield a number that can be used for classification purposes. Such a classification will most probably have no meaning in terms of human risk assessment. It should be noted that the text of the 6th Amendment makes reference to the use of data generated in humans, but no guidance is given to the quantitative and qualitative nature of what would constitute acceptable human data. It is suggested that appropriately designed human skin patch tests and productuse tests are an acceptable approach for assessing human response. (ii) In vitro tests: The ECETOC Monograph on “Skin Irritation” 141gives an adequate overview of the methods that are available for the in vitro assessment of skin irritation. I do not think that it would be helpful to repeat that information by listing test systems. Rather, I would like to illustrate the potential of one system which is the “Human Skin Keratinocyte Cytotoxicity/Neutral Red (KC) Assay” 111-131 and I am grateful to my colleague, Dr. Rosemarie Osborne, for allowing me to use her data. The basis of the assay system is the ability of cultured human keratinocytes to take up the vital dye neutral red. Viable cells will incorporate the dye: damaged cells will not. Thus, the test system measures the response of human epidermis derived epithelial basal cells (keratinocytes) to test materials applied in culture. From dose-inhibition curves the concentration of test material that inhibits neutral red (NR) uptake by 50% is calculated. This is the NR50 value. The NR50 value for different test materials is a measure of the in vitro cytotoxicity potential of those materials, which is then taken to be an indicator of ultimate skin irritation potential. In validating the system a standard curve was constructed for the keratinocyte assay using human skin patch test data for a range of chemical irritants covering a wide range of irritation potential and cytotoxicity. The method showed a promising correlation with human skin irritation for selected chemicals and ingredients that cover a wide range of irritation potential. For other chemicals, such as acids and alkalis, the KC assay underestimated in uiuo irritation. This discrepancy is likely due to the buffering capacity of the cell culture medium, that reduces the effect of extremely high or low pH on the cells. The major use of preclinical skin irritation testing at P&G is to provide safety support data prior to human skin exposures. For example, we would like to ensure that panelists’ skin is not unduly irritated in a clinical patch test (an irritation score of I 2 out of a possible total of 5). Thus, one valuable use of KC assay data is to confirm the selection of doses of test substances
162
before a clinical patch test. Based on the KC assay score for a new material, potential patch test doses can be positioned relative to a standard curve. For example, for a test material with a cytotoxicity score [log (l/NR50)] of 3, the log (score/dose) value is approximately 1.5. To select a dose of material that would produce a patch test score of no greater than 2, the highest dose in the human patch test should be 0.002%. The effects of various product formulations which contained surfactants were tested in the KC assay. A poor correlation between the cytotoxicity data and human skin irritation was obtained for these products. For example, the bar soaps were similar in human skin irritation, but differed by 2 orders of magnitude in cytotoxicity in vitro. In contrast, the shampoos differed by nearly 3 orders of magnitude in skin irritation, but were similarly cytotoxic. One explanation for the lack of correlation is that, in the in vitro assay the presentation of the test products differs fundamentally from that which occurs to skin in Go. In a patch test, epidermal basal cells are likely exposed to, at most, a dilution of only a few of the formulation components. In the KC assay, the keratinocytes are exposed to dilutions of the entire product formulation. Nonetheless, KC assay data may be useful for determination of whether a new formulation is far more cytotoxic than previous formulations of a similar type. These data illustrate two points made in the ECETOC monograph on skin irritation. Firstly, it is essential that any in vitro model be validated for the class of chemical being evaluated and in the laboratory performing the test. Secondly, and this is the converse, if careful thought is not given to the choice and use of specific assays one can get meaningless and confusing data. It is preferable to rely on systems that model important aspects of human skin response to the materials tested. In this way, data from the in vitro system can be interpreted based on the physiology of the skin, and applied directly to safety assessment of potential human skin exposure. APPLICATION IN PRACTICE
The new in uiuo and in vitro methods for eye and skin irritation assessment should become part of the review process which takes place before any exposure of humans to test materials. A first step is a review of existing data from all available sources. Information on physicochemical characteristics, such as pH and buffering capacity, may indicate the potential for eye or skin irritancy. A structure-activity assessment against chemicals of known irritancy potential may provide additional perspective. In some cases, skin irritancy data may become available from other tests which are designed to assess the dermal safety of topically applied chemicals. Irritancy data might come available from guinea pig skin sensitization
163
studies, or from rabbit percutaneous toxicity studies. .As the design of such studies is not primarily directed towards assessment of skin irritancy, care must be taken in using data from such studies. Concentrations of test materials tend to be low in sensitization studies but set at the highest tolerable dose in percutaneous toxicity studies. An overall assessment of this existing skin irritation data may be sufficient for the selection of a dose or exposure conditions in a clinical skin irritation test or a product use study. For many products and chemicals, an adequate skin irritation assessment can be made without additional testing. In some cases, however, additional in vitro analysis may be required to position irritancy potential relative to chemicals and products of known potency. A stepwise procedure is recommended in the ECETOC monograph on “Skin Irritation” [41. IRRITATION ASSESSMENT AND REGULATION
As has been explained earlier, regulations worldwide call for Draize Tests for skin and eye irritation. Regulators are interested in progress in this area, and need to provide support for the development and validation of alternative methods in order that a scheme can be developed to give industry the option of using the new methods such as 1 have described. We cannot stay forever with the Draize Tests. The omens are, I believe, good. The rational arguments over the LD50 test have led to its virtual elimination with no loss in the thoroughness of the hazard evaluation process. The up-and-down method, the limit test and the non-lethality method provide an adequate basis for assessing acute lethality. This provides a good model and hopefully the scientific-regulatory dialogue will see Draize Tests eventually replaced. SUMMARY AND CONCLUSIONS
1. Essentially all regulatory authorities in the OECD countries require a Draize Test to be carried out for the assessment of eye and skin irritation, and subsequent classification. 2. These methods are out-dated. The regulatory process has not allowed the easy incorporation of new methodology into regulation. 3. Modern approaches to eye and skin irritation assessment require that attention be paid to the three “R”s of Refinement, Reduction and Replacement in the design of protocols developed to assess toxicity. 4. Alternative in uiuo methods exist for both eye and skin irritation assessment. Excellent progress has been made in developing in vitro
164
methods that may significantly reduce our dependence on animal based tests. 5. Toxicologists should arrange for peer review and publication of the results of research into in uiuo and in vitro alternative methods in order to make their methods available for broad evaluation. 6. Regulators should encourage such work and actively embrace the principle that valid data for classification purposes can be obtained from protocols other than those mandated in Annex V. ACKNOWLEDGMENTS
I would like to acknowledge the help provided by Dr. Rosemarie Osborne, Dr. Leon Bruner and Dr. K.A. Kohrman in the preparation of this manuscript and Mrs. Petra Bogaerts for typing it. REFERENCES 1 Legislation on Dangerous Substances, Classification and Labelling in the European Communities, Volume 1. Office for Official Publications of the European Communities, Graham & Trotman Ltd., Luxembourg and London, 1987. OECD Guidelines for Testing of Chemicals. The Organisation for Economic Co-operation and Development, 1981. Eye Irritation Testing, Monograph 11, European Chemical Industry Ecology and Toxicology Centre, Brussels, June 1988. Skin Irritation, Monograph 15, European Chemical Industry Ecology and Toxicology Centre, Brussels, July 1990. Griffith, J.F., Nixon, G.A., Bruce, R.D., Reer, P.J. and Barman, E.A. (1980) Dose-response studies with chemical irritants in the albino rabbit eye as a basis for selecting optimum testing conditions for predicting hazard to the human eye. Toxicol. Appl. Pharmacol. 55, 501313. 6 Freeberg, F.E., Nixon, G.A., Reer, P.J., Weaver, J.E., Bruce, R.D., Griffith, J.F. and Sanders, L.W. (1986) Human and rabbit eye responses to chemical insult. Fund. Appl. Toxicol. 7,626-234.1 7 Freeberg, F.E., Griffith, J.F., Bruce, R.D. and Bay, P.H.S. (1984) Correlation of animal test methods with human experience for household products. J. Toxicol. Cut. Ocular Toxicol. 3 (l), 53-64. 8 Bruner, L.H., Kain, D.J., Roberts, D.A. and Parker, R.D. (1991) Evaluation of seven in-vitro alternatives for ocular testing. Fund. Appl. Toxicol. 17,136149. 9 Perkins, M.A., Roberts, D.A. and Osborne, R. (1992) Human skin cell cultures for in vitro skin and eye irritancy assessments of neat test materials. The Toxicologist 12: 296. 10 Osborne, R., Comparison of rabbit and human skin responses to primary dermal irritants, Manuscript in preparation. 11 Osborne, R. and Perkins, M.A. (1991) In vitro skin irritation testing with human skin cell cultures, Toxicol. In Vitro 5 (5/6), 563-567. 12 Osborne, R. and Perkins, M.A. (1991) Evaluation of Human Skin Cell CulturesforIn Vitro Skin Irritancy Testing. In: A.M. Goldberg (Ed.)ln Vitro Toxicology: Mechanisms and New Toxicology. Alternative Methods in Toxicology, 8. Mary Ann Liebert, Publisher, New York. 13 Osborne R. and Perkins M.A. (1992) An approach for development of alternative test methods based on mechanisms of skin irritation, Food Chem. Toxicol. In press. 14 Inter Regulatory Agency Group Workshop on updating Eye Irritation Tests, September 26 and 27,1991, Washington D.C.