Dermal penetration testing, also known as percutaneous penetration, measures the absorption or penetration of a substance "through the skin barrier and into the skin" (OECD, 2004a). Dermal penetration studies are conducted to determine how much of a chemical penetrates the skin, and thereby whether it has the potential to be absorbed into the systemic circulation. Dermal penetration is considered to occur by passive diffusion; however, biotransformation of the test substance within the skin (metabolism) prior to systemic absorption can occur (OECD, 2004a).
Dermal toxicity testing, on the other hand, is conducted to assess the local and/or systemic effects of a chemical following dermal exposure. It may provide an indication that the substance penetrates the skin if it produces systemic toxicity, but the amount of chemical absorbed is not quantified by dermal toxicity testing (ECVAM, 2002).
One of the primary roles of the skin is to form a barrier to protect humans from substances contacted in the environment. Permeation of a substance through the skin depends upon a number of factors, including: area of contact; duration of exposure; lipophilicity (fat solubility), molecular weight, and concentration of the test substance; and integrity of the stratum corneum and thickness of the epidermis (OECD, 2004b). The qualities of the outermost layers of the skin, the stratum corneum, typically determine the rate of dermal penetration. [For a description of the structure of the skin, see the AltTox section Skin Irritation/Corrosion.]
Both in vivo and in vitro methods are available for determining the percutaneous penetration of a substance. Selection of the method used may result in obtaining different types of information, variations in accuracy of the results, and relevance of the test results to humans. Another factor in the selection of an in vivo versus in vitro test system may be the preference of the national regulatory authority. The in vitro dermal penetration test methods are considered advantageous to the in vivo tests in the EU. Some US agencies have accepted in vitro dermal penetration data, however, they have not formally accepted the Organisation for Economic Co-operation and Development's (OECD) test guideline (TG) for in vitro dermal absorption (OECD TG 428, Skin Absorption: In Vitro Method).
The OECD recommends consulting their guidance document, the OECD Guidance Document for the Conduct of Skin Absorption Studies (No. 28, March 5, 2004), to determine the most suitable method for a particular study. Whatever test system is used, its performance in the testing laboratory should be confirmed with reference chemicals showing comparable results to the published values (OECD, 2004a).
The Animal Test(s)
The rat is the most common species used for in vivo dermal penetration testing. The in vivo method for determining the penetration of a substance through the skin of an animal and into the systemic compartment is described in the OECD TG 427 (revised version adopted April 13, 2004). One or more doses of the test material, usually a radioactive-labeled sample, are applied to the shaved skin of an animal for a specified time. The time and dose are based on expected human exposure. The animals are then observed at regular intervals for signs of toxicity, and daily excreta (and sometimes expired air) are measured for the test substance. Blood is collected periodically and when the animal is euthanized, and distribution of the test substance is measured in tissue samples from the application site and the body. The results can be reported as the rate, amount, or percentage of skin absorption. The animal test method is considered advantageous for the "generation of systemic kinetic and metabolic information" (OECD, 2004a).
Non-animal Alternative Methods
A variety of in vitro methods have been developed for dermal penetration testing. Most of these methods use full or partial thickness human or animal skin mounted in a diffusion chamber. Pig skin is commonly used. Human keratinocyte cell-culture based models called reconstituted, 3-dimensional (3D), or organotypic skin cell models, have also been used, but are generally more permeable than skin preparations. Quantitative structure-activity relationship ((Q)SAR) and other computational models have been developed for the prediction of dermal penetration, but are not widely used for hazard identification.
Advantages beyond the fact that live animals are not used include: human skin more closely estimates human exposure; the early phase of absorption can be determined; replicate measurements can be made using the same individual's skin; exposure conditions can be varied; and a wider range of physical forms of test substances can be evaluated (OECD, 2004a). The primary limitation described for in vitro percutaneous penetration testing "is that sink conditions of the peripheral blood flow may not be fully reproduced." Metabolism of the test substance within the skin may vary among the different in vivo and in vitro (viable skin) test systems.
In vitro methods to determine skin absorption are described in OECD TG 428, Skin Absorption: In Vitro Method. The in vitro test is based on the permeability of a test substance across various types of human or animal skin preparations. The skin sample is mounted in a static or flow-through diffusion chamber and the test substance that permeates the skin is collected in a fluid reservoir. The amount and rate of test substance accumulating in the fluid chamber is measured. If the test substance is metabolized in the skin, its metabolites in the fluid chamber should also be determined. The amounts of test substance rinsed off, remaining on, and within the skin sample are also measured to provide an estimate of total recovery of the applied substance.
Both fresh and frozen human and animal skin can be used to assess in vitro percutaneous penetration (OECD, 2004c). The use of viable skin is preferred in certain circumstances. Non-viable human or animal skin can be used; however, the ability to assess skin metabolism of the test substance will be lost. Regardless of the type of skin preparation used, its integrity must be demonstrated. Several methods, including transepidermal water loss and tritiated water permeability, are used for assessing the integrity of skin samples prior to their use (OECD, 2004a).
The measurement or estimation of the dermal penetration of a substance is an important component of estimating its blood levels for use in physiologically based pharmacokinetic (PBPK) models. A PBPK model to predict plasma levels of benzoic acid following dermal exposure in the hairless guinea pig was developed using in vitro dermal penetration data from a guinea pig skin flow-through cell (Macpherson, et al., 1996). Predicted values were compared to in vivo measurements with one conclusion being that "varying the transdermal input parameters produced closer agreement between predicted and measured values."
Isolated skin flaps have been used to show the relationship between in vivo and in vitro dermal penetration. Isolated perfused porcine skin flaps were used as a pharmacokinetic model to study the flux of compounds through pig skin, and 8-hour skin flap results were described as highly predictive of 6-day in vivo absorption totals (Carver, et al., 1989). Another example was the use of an in vitro flow-through cell with porcine skin to estimate the rate and extent of dermal absorption of organophosphate pesticides (Van der Merwe, et al., 2006).
Experimental variables that may affect the results of in vitro dermal penetration studies include: the selection of the in vitro test system, the preparation and storage of the skin samples, the preparation of the radio-labeled test substance, volatility of the test substance, test substance degradation, the dose tested, exposure time, occlusion/non-occlusion of the test sample on the skin, temperature, removal of test substances at the end of the experiment, sampling and data analysis techniques, and analytical method problems (OECD, 2004a). Many protocol variables have been shown to influence the rate and extent of skin absorption of a test substance; these variables include the use of infinite versus finite dosing, test substance volatility, vehicle of application, exposure time, and leave on versus rinse off of test substance (Brain, et al., 1995; Walters, et al., 1997). The test substance retained within the skin (except for the stratum corneum) is usually included in the calculated absorbed dose, although in the live animal it may eventually be lost by shedding and renewal of the skin.
The OECD guidance and TG for in vitro percutaneous penetration testing were issued in 2004, but the guidelines are very broad and do not endorse specific protocols or protocol components. An example of some of the experimental variables to be considered with in vitro dermal penetration studies are described in the report assessing the percutaneous penetration of diethanolamine in cosmetic formulations using in vitro human skin (Brain, et al., 2005), including the testing of rinse off versus leave in formulation, varying the formulation ingredients, using replicate skin samples from different donors, and comparing permeability in fresh versus frozen skin tissues.
In vitro dermal penetration methods are widely used; picking the correct protocol is what is confusing. To promote the development of standardized protocols and better agreement of in vitro dermal permeation methods, the Institute for In Vitro Sciences (IIVS) hosted a workshop for a small group of international stakeholders in 2005 to discuss the OECD guidance and to make recommendations on implementing specific aspects of the guidance. Recommendations from this meeting are summarized in a poster available on the IIVS website, and include items such as selection of the skin model and receptor fluid, best practices for storing frozen skin, barrier integrity assessment methods, and more. Additional suggestions by the workshop participants were that human skin be considered the "gold standard," that in vitro data be submitted to regulatory agencies to promote its acceptance, and that some mechanism be developed for sharing among stakeholders the responses of regulatory agencies to submitted in vitro data.
Human cell-based or reconstituted human skin cell models such as EpiDerm, EpiSkin, and SKINETHIC (Asbill, et al., 2000; Schäfer-Korting, et al., 2006; Schreiber, et al., 2005; Zghoul, et al., 2001), and a rat keratinocyte culture model (ROC) (Marjukka Suhonen, et al., 2003; Pappinen, et al., 2007) are also being used to evaluate dermal penetration. These models are considered to be metabolically active, but in most cases are more permeable than in vitro human and animal skin preparations.
(Q)SAR models, which are predictive computational models based on molecular structure, have been developed to assess the skin permeability of drugs and chemicals. Many models have been developed, and a reassessment of the statistical relevance of existing models suggested that heterogeneity in the data used to develop the models (i.e., skin origin and experimental conditions) contributes to their variance (Fujiwara, et al., 2003; Geinoz, et al., 2004).
Geinoz, et al. (2004) concluded that some models are "based on nonsignificant parameters," and that structure-based skin permeability prediction, like for other (Q)SAR models, is more reliable when the model is developed and used within similar chemical classes. Fujiwara, et al., (2003), on the other hand, extracted information from 10 published data sets on diverse species and compounds and concluded that skin permeability can be reasonably well explained by two descriptors, the octanol/water partition coefficient (log P) and the molecular weight (MW). A later study suggests that a linear relationship exists for log P and MW when water is the vehicle (solvent), but not for MW when ethanol or ethanol/water is the vehicle (Van der Merwe & Riviere, 2005). Because of therapeutic relevance as well as toxicological studies, considerable activity continues in the development and evaluation of predictive computational models for skin permeation based on chemical structure (Katritzky, et al., 2006; Neumann, et al., 2006).
Validation and Acceptance of Non-animal Alternative Methods
Dermal penetration studies using in vitro skin preparations are broadly used without formal validation.
The OECD Guidance Document for the Conduct of Skin Absorption Studies (2004a) states that "...skin absorption is primarily a passive process and studies undertaken using appropriate in vitro experimental conditions have produced data for a wide range of chemicals that demonstrate the usefulness of this method. Such methods have found use in, for example, comparing delivery of chemicals into and through skin from different formulations and can also provide useful models for the assessment of risk due to percutaneous absorption in humans."
Non-animal methods for dermal penetration have not been formally validated, but the in vitro methods described by OECD TG 428 have been accepted by EU authorities. US agencies have not formally acknowledged acceptance of the in vitro percutaneous penetration methods described by OECD 428, but there are anecdotal reports of in vitro methods being accepted by some US agencies.
Although not a formal validation study, an inter-laboratory assessment of in vitro skin absorption methods was conducted in which three compounds were tested in 10 laboratories using standardized protocols (Van de Sandt, et al., 2004). Some variability in the results was reported, but the in vitro methods were considered to be robust. Nine labs used human skin, and it was concluded that "variation observed may be largely attributed to human variability in dermal absorption and the skin source." One lab used rat skin, which was more permeable than human skin to caffeine, but similar for benzoic acid and testosterone. The type of diffusion cell used did not appear to affect the results. Skin thickness did alter the results - only slightly for benzoic acid and caffeine, but significantly for testosterone, where absorption was higher with thin, dermatomed human skin.
The OECD Guidance Document states that "reconstituted human skin models can be used if data from reference chemicals are consistent with those in the published literature" (OECD, 2004a). Schäfer-Korting, et al. (2006) conducted an interlaboratory prevalidation study to assess the performance of Reconstructed Human Epidermal (RHE) models. The skin permeability of caffeine and testosterone was compared for three RHE models (SkinEthic, EpiDerm, and EPISKIN), human epidermis, and animal skin. Reproducibility was good, but the RHE models overestimated dermal penetration for these chemicals. Prevalidation of the RHE models was considered successful, and a validation study is being conducted.