Introduction
Contact dermatitis in metal workers is considered to be the most relevant occupational illness in the metal working industry. Different studies report a 20-25% prevalence of hand eczema in metal workers (1,2). Irritant contact dermatitis is more often diagnosed than the allergic contact dermatitis (3). Several different chemicals with which metal workers have skin contact contribute to the high prevalence of contact eczema, including cleaning detergents, solvents, degreasers and metalworking fluids. Though rarely harmful on short-term exposure, metalworking fluids may evolve their harmful attributes after long term and repeated contact with the skin. Since several hours of contact on each working day is typical in many working environments, even a low irritant capacity contributes to the high incidence of contact eczema in this work group. The use of cutting fluids with good skin compatibility is therefore an important factor in preventing occupational hand dermatitis in metal workers. In recent studies, both in vitro and in vivo, the irritancy of different cutting fluids was compared. Both strategies have advantages, but also limitations.
In vivo studies on human volunteers are of the highest relevance for predicting skin irritancy. However, when looking at chronic skin irritancy induced by mild irritants, established in vivo studies often fail to predict the irritant potential. For example in a study by de Boer et. al. (4) different metal working fluids induced only minimal skin irritation, even after stripping of the stratum corneum and repeated application to the forearm skin over 5 days. Differentiation between the tested fluids and water was only possible for one of three cutting fluids tested. In addition, in in vivo tests the safety of volunteers has the highest priority. Hence, some products may not be tested. A study director should take into account the higher penetration rate of components and a higher risk of sensitisation under the conditions of a repeated patch test design. This holds especially true for all used cutting fluids for which the exact composition is not known.
For in vitro tests correlation to the in vivo situation has to be verified before they can be used as a predictive tool. On the other hand, in vitro test procedures can be used with any sample and at any concentration. The analysis of the irritancy potential is based on the measurement of inflammation mediators or on the quantification of cell damage. This allows a relatively sensitive analysis of the irritant potential of test substances, since the biochemical parameters increase before clinical symptoms are manifest. However, careful examination of the test results is necessary due to the model nature of in vitro tests. Correlation with the in vivo situation can only be assumed when validated test designs are used. In previous studies, the BUS model was accepted as an in vitro model for testing the skin compatibility of cutting fluids (5). Although it served for some time as a suitable test system, it also has some disadvantages such as the lack of standardisation and insufficient data for correlation with the in vivo situation.
Today, 3-dimensional cell cultures from human skin cells, so called skin models, have gained increasing importance as alternative to in vivo test systems (6). Due to great efforts of the cosmetic and raw material industry to find alternative methods to replace animal testing, the development of skin models has made fast improvements in the last years. The reconstituted human skin model, cultivated from human epidermal skin cells on a collagen matrix, was validated for the toxicological endpoint skin corrosion by ECVAM (European Centre for the Validation of Alternative Methods) (7). The validation study was performed in three independent laboratories and proved that this human skin model was able to correctly predict the corrosion potential of 12 OECD reference chemicals. The validation procedure led to a new OECD Guideline for the testing of chemicals for skin corrosion based on skin models (8). So far, three different skin models have been validated according to the OECD TG 431 as suitable test systems for skin corrosion: EPISKIN™, EpiDerm™ and SkinEthic™. Skin corrosion, defined as the production of irreversible tissue damage of the skin following the application of a test material, is certainly not induced by the application of metal working fluids. The damage to skin by long term contact with this product category is a mild irritation, with irritation being defined as the production of reversible tissue damage of the skin. However, to date no validated in vitro assay for testing and classification of skin irritation exists. However, in a prevalidation study by ECVAM for the testing of skin irritation and skin compatibility, human skin models turned out to be the most promising of several different in vitro methods (9). In addition to the investigations on alternative skin irritation models for classification purposes, several positive investigations on the testing of consumer products with low irritation potential on skin models also exist (10). In the experiments presented here we used an established in vivo study, the 24h Patch Test and an accepted in vitro test procedure on human skin models to test the skin compatibility of three cutting fluids. The aim was to find out whether these test systems are able to differentiate between the three test samples.
Experimental
Materials
Cutting Fluids
Table 1 lists the cutting fluids referred to in this paper. Product A, B and C are based on formulations of the MULTAN® series (Henkel technologies). General information on fields of application, components, pH, and use concentrations are listed.
Skin Models
SkinEthic human epidermis model (RHE) and SkinEthic human cornea model (HCE) were purchased from SkinEthic Laboratories (Nice, France). On arrival, cultures were transferred to fresh culture medium and incubated overnight at 37°C before application of the test samples.
Methods
Treatment of skin models
A 30µl aliquot of test sample was applied directly onto the surface of the epithelial culture. Each product was tested in triplicates (minimum) against negative (water) and positive (0.5% or 1% SDS in water) controls.
Membrane Integrity Assay
Media were sampled at the specified time intervals and tested for lactate dehydrogenase (LDH) activity. (Cytotoxicity Detection Kit LDH, Roche Diagnostic Corporation, Mannheim, Germany).
24 h Patch Test
The test was carried out on 20 healthy female and male volunteers in a mixed panel representing the average spectrum of the normal skin type. Volunteers were informed about the aim of the study, test procedure, test substances, and any possible health risks or discomforts before the start of the study. Written consent was signed by all volunteers before participating in the study. The test substances (5% dilution in demin. water) were applied side by side on the back of the volunteers over a period of 24 hours in a volume of 70ml under occlusive plasters (Finn Chamber on Scanpor, 12 mm). Any irritant effects observed were recorded and documented 6, 24, 48 and 72 hours after removal of the plasters, divided into erythema, edema, squamation and fissuration parameters according to the scale of Frosch (11). The resulting score values are evaluated numbers reflecting the strength of the individual reactions. Individual score values were summed for all recorded time points and divided by the number of volunteers. The resulting total irritation score was calculated for each test substance, for erythema and a parameter combination of erythema + edema + squamation + fissuration (shown in tables). SDS 0.5% and demin. water were used reference substances for this test. For statistical comparison of the visual scores the “area under curve” (AUC) was calculated. A Many to One Comparison of product A and product C versus product B was performed. The Friedman test was used to clarify whether there were significant differences between the groups. The significance level was p<0.05.
Results and Discussion
Three water soluble cutting fluids, A, B and C were analysed with regard to their skin compatibility in a 24h Patch Test and in skin models respectively. Differences in product formulas are listed in Table 1.
In the 24h Patch Test the positive control (0,5% SDS) led to a total irritation score of 9.16. The negative control demin. water induced a total irritation score of 0.37. Products A, B and C induced irritation scores between 1.05 and 1.47. The irritation score for erythema was between 0.68 and 1.05 for products A, B and C. All irritation scores and related standard deviations are shown in Table 2.
The 5% product dilution used in the patch test resembles the recommended usage concentration for the products. At this concentration all products were well tolerated with irritation scores low above that of demin. water. Under these test conditions differences in skin compatibility between product A, B and C were not detectable. The Friedman test yielded no statistical significant differences in the score values of the three products. This finding corresponds with that of other investigators, who found patch tests on human volunteers insufficient to detect compatibility differences between well tolerated formulations (5).
In the in vitro tests, human epidermal skin models and human cornea models respectively, the products were tested with 5% and 10% product dilutions. Table 3 and Figures 1 and 2 show the LDH release in epidermis models after treatment with products A, B and C and with the positive control 1% SDS. An increase in LDH release indicates damage to the cell membrane leading to the leakage of enzymes from the cell. Not surprisingly, the 10% product dilution induced higher LDH-release compared to the 5% product dilution for all products tested. Clear differences between the three products in the amount of LDH release they induced were observed at both concentrations. At both concentrations product B induced the highest cell damage with product C producing the lowest.
Figure 1: Course of LDH-release from epidermis models after treatment with products A, B and C in 5% product dilution.
Values are shown as % of negative control (n=4).
Figure 2: Course of LDH-release from epidermis models after treatment with products A, B and C in 10% product dilution.
Values are shown as % of negative control (n=4).
A similar picture was seen after application of the products to cornea models. Since cornea models are more sensitive to chemical noxes, a shorter application time was chosen before measuring LDH in this test system. The data generated with cornea models is summarized in Table 4 and Figures 3 and 4. Again, as expected the 10% product dilution induced a higher LDH-release as a measure of cell damage compared to the 5% product dilution for all products tested. Clear differentiation between product A, B and C on cornea models was only seen after application of the products at 10% dilution. At 5% product dilution no significant difference was detected in the LDH released by the three products.
Figure 3: Course of LDH-release from cornea models after treatment with products A, B and C in 5% product dilution.
Values are shown as % of negative control (n=4).
Figure 4: Course of LDH-release from cornea models after treatment with products A, B and C in 10% product dilution.
Values are shown as % of negative control (n=4).
In practice skin problems with working fluids often appear to be due to contamination or chemical changes in the product during use. Testing used cutting fluids on human volunteers is untenable for ethical reasons since the precise composition of a used working fluid is not known. With in vitro systems, where such concerns do not exist, tests were performed in order to differentiate between used and unused cutting fluids. Product C was tested as a 5% product dilution on cornea models. In parallel, used samples of product C taken from different tanks in an Italian factory were tested. The results are summarised in Table 5 and Figure 5. Product from tank 1 and tank 2 induced a LDH release comparable or only slightly higher than that induced by the unused product C. In contrast product C from tank 3 induced a higher LDH release, heralding upcoming problems in skin compatibility.
Figure 5: Course of LDH-release from cornea models after treatment with used and unused products in 5% product
dilution. Values are shown as % of negative control (n=4).
Comments