In vitro methods for estimating eye irritancy of cleaning products phase i: preliminary assessment

The Development and Evaluation of In Vitro Tests by the FRAME Alternatives Laboratory

This review outlines the work which has been conducted in the FRAME Alternatives Laboratory during the first ten years of the FRAME Research Programme. A number of in vitro tests, including the kenacid blue, neutral red release and fluorescein leakage assay methods, have been evaluated and have subsequently been included in validation schemes organised by the US Soap and Detergent Association, the US Cosmetic, Toiletry and Fragrance Association, the European Commission and the European Cosmetic, Toiletry and Perfumery Association, as well as in the Scandinavian multicentre evaluation of in vitro cytotoxicity testing scheme. More recently, research has been undertaken in the areas of phototoxicity, immunotoxicity, dermal toxicity and intercellular communication, in addition to investigations into fundamental mechanisms of toxicity.

Validation of In Vitro Cytotoxicity Tests — Past and Present Strategies

In recent years, conventional toxicity testing in animals has been reinforced by in vitro methods. As a result, toxicity testing in some sectors has become more effective and at the same time more ethical. This trend is probably only at its beginning, as many of the newly-developed methods have not yet won general acceptance as a basis for the large-scale replacement and reduction of animal experimentation. What limits the wider use of these methods is validation, i.e. the evaluation of their reliability and relevance. The present paper is a short review of the validation efforts made hitherto, including projects being planned and under discussion. Our own MEIC approach is compared with other strategies. Finally, our opinion on the effectiveness of one large consensus project relative to several different smaller validation programmes is expressed — we advocate the latter strategy, because it will save time and reduce costs.

Soaps and Detergents: Alternatives to Animal Eye Irritation Tests

An extensive computer search was conducted, and a comprehensive overview of the current status of alternatives to animal eye irritation tests was obtained. A search of Medline and Toxline databases (1988 to present) was supplemented with references from sources regarding in vitro eye irritation. Particular attention was paid to soap and detergent products and related ingredients. Eighty-five references are included in the review; the in vitro assays are categorized, and their predictive values for assessing acute ocular irritation are evaluated and compared with the Draize rabbit eye irritation assay and with each other. The present review shows that the increased activity of scientists from academia, industry, and regulatory agencies has resulted in substantial progress in developing alternative in vitro procedures and that a number of large, interlaboratory evaluations and international workshops have assisted in the selection process. However, none of these methodologies has obtained acceptance for regulatory classification purposes. Conclusions drawn from this review include that (a) no single in vitro assay is considered capable of replacing the Draize eye irritation test; (b) the chorioallantoic membrane vascular assay (CAMVA) or the hen egg test-chorio-allantoic membrane test (HET-CAM), the chicken or bovine enucleated eye test, the neutral red and plasminogen activation assays for cytotoxicity, and the silicon microphysiometer appear to have the greatest potential as screening tools for eye irritation; and (c) choosing a specific assay or series of assays will depend on the type of agent tested and the impact of false-negative or false-positive results. New assays will continue to be developed and should be included in future evaluations, when sufficient data are available.

Evaluation of surfactant cytotoxicity potential by primary cultures of ocular tissues: I. Characterization of rabbit corneal epithelial cells and initial injury and delayed toxicity studies

This investigation was undertaken to develop cytotoxicity assay systems using primary cultures of rabbit corneal epithelial cells as an experimental model to evaluate oculotoxic agents and the ability of these in vitro assay systems to predict irritancy potential and delayed toxicity. We have characterized the epithelial nature of the cultures by identifying keratins with antikeratin antibodies (AE1/AE3) and by demonstrating metabolic enzymes important to the integrity of the cells: lactate dehydrogenase, glucose 6-phosphate dehydrogenase and aldolase. Eight surfactants were compared and ranked according to their cytotoxic potential. We evaluated cytotoxicity by measuring leakage of the cytosolic enzyme, lactate dehydrogenase, into the medium, by making morphological observations and by assessing lysosomal neutral red uptake and mitochondrial 3-(4,5-dimethythiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) reduction. The cells were treated for 1 h with the surfactants and the possibility of delayed toxicity was evaluated 24 h after removal of the surfactant. The cytotoxicity of the different types of surfactants as shown by all the tests was cationic > anionic = amphoteric > non-ionic. Triton X-100, a non-ionic surfactant but a severe irritant, had a ranking similar to anionic surfactants. The in vitro rankings corresponded well to reported in vivo Draize rabbit eye test data. The 24-h test for lactate dehydrogenase leakage showed that mild and non-irritating surfactants did not demonstrate any subsequent damage after a 1-h exposure, but the extreme and severe surfactants continued to show further damage after the 1-h exposure. These in vitro findings were similar to reported in vivo results. The neutral red and MTT tests did not adequately predict the prolonged toxicity of the more irritating surfactants, as was demonstrated by the lactate dehydrogenase leakage test. We conclude that in vitro cytotoxicity assays using primary cultures of rabbit corneal epithelial cells may be used to rank the cytotoxic potential of surfactants, but only the lactate dehydrogenase leakage test was able to assess prolonged cell injury.

In vitro toxicity of various classes of test agents using the neutral red assay on a human three-dimensional physiologic skin model

A new three-dimensional human skin model consisting of several layers of actively dividing and metabolically active human neonatal foreskin-derived fibroblasts and epidermal keratinocytes grown on nylon mesh has been used to assess the in vitro toxicity of test agents from various classes. Utilizing a slight modification of the published neutral red viability assay for endpoint determination, we have assayed and obtained dose-dependent toxicity curves for test agents from the following classes: detergents (n = 15), alcohols (n = 5), metal chlorides (n = 10), perfumes and colognes (n = 5), shampoos (n = 4), conditioners (n = 3), moisturizers (n = 3), pesticides (n = 3), and antimicrobial preservatives (n = 4). Limited comparisons to in vivo ocular irritancy data with alcohols and detergents are encouraging. We have demonstrated the utility of this metabolically active dermal substrate containing naturally secreted collagen and other extracellular matrix proteins along with the neutral red viability assay for assessing the toxicity of a number of test agents from a variety of different classes with broad industrial applications.

The EYTEX™ System and the Neutral Red Uptake Inhibition Assay in Cultured Hep G2 Cells as Alternative Methods for In Vivo Eye Irritation Following the EEC Protocol

The neutral red uptake inhibition assay and the EYTEX™ system were investigated as alternative methods for the assessment of eye irritation, determined according to the EEC protocol. The 17 test chemicals used were mainly organic solvents. The xenobiotics were applied to Hep G2 cells for 24 hours at different concentrations. Neutral red uptake inhibition was then measured. The results are expressed as the NI50 value, which is the concentration of test compound required to induce a 50% reduction in neutral red uptake. The same chemicals were also tested as coded samples by the EYTEX™ test according to the manufacturer's directions. A nearly identical quantitative correlation was found for both in vitro methods with corneal opacity scores: r = 0.84 for EYTEX™ scores and r = 0.83 for log NI50, expressed in μg/ml. Whilst these correlations are certainly not perfect, it is clear that both in vitro methods can be used as valuable prescreening methods.

In vitro predictive tests for eye irritants

The developments and attitudes towards in vitro testing since the first major workshop on irritation testing five years ago (Reinhardt et al., 1985) are summarized. Many test systems have been described and an increasing number of compounds tested. However, the in vivo data basis used for comparison is still heterogeneous and a proper analysis of most in vivo/in vitro correlations is difficult. Some progress has been achieved with various controlled validation programmes on a national and international level, all of which have not been completed. In order to improve this slow progress the following measures are proposed: (1) expectations for a uniform test need to be discouraged when heterogeneous chemical actions may occur, and rigid test guidelines have to be replaced by flexible ones (differentiated approach); (2) many practical small steps need to be taken rather than one theoretical large step, that is by accepting simple in vitro tests for restricted groups of chemicals such as severe irritants (adaptive approach); and (3) the threshold for acceptance of in vitro tests needs to be lowered by regulatory bodies and routine laboratories (optimistic approach).