A Novel Explant Model for Skin Delivery Assessment

The efficacy and safety of cosmetic products depends on the chemical structure and toxicological profile of their ingredients, as well as the levels of these ingredients that the user will encounter. These parameters are not only linked to the ingredients in a formulation, but also to what they become on the skin. Studies have reported that skin absorption of the same ingredient in different topical formulations can differ as much as 10 to 50 times.1 Moreover, determining absorption itself is challenging due to the complexity of structures and mechanisms constituting the skin delivery pathway. Delivery depends on various parameters, including the physicochemical properties of the active ingredient, skin condition, formulation type, etc.

For reasons of safety and efficacy, there is a clear need to evaluate the skin permeation of cosmetic actives, and in vitro methods are preferred from both reproducibility and regulatory standpoints. Therefore, to obtain near in vivo conditions, a rigorous protocol and an appropriate skin model are key. The skin model should have good barrier function and metabolic activity. It also must be practical and easy to use, and provide reproducible results. This paper describes the evolution of an experimental explant modela to meet these needs and provide the desired test conditions.

In vitro Skin Delivery Tests

The pharmaceutical industry2 inspired cosmetic industry test guidelines for in vitro assessments of the dermal absorption and percutaneous penetration of ingredients.3 In 2004, the Organization for Economic Co-operation and Development (OECD) adopted4 Test Guideline 428 and a corresponding5 Guidance Document 2, which have been incorporated into in vitro dermal absorption studies in toxicological dossiers of cosmetic ingredients. The OECD documents present the use of static or dynamic diffusion cells and describe the main steps to conduct studies, i.e., the dose to apply, application time, etc., to harmonize test protocols. They also give an overview of pertinent skin models—an important parameter for experimentation.

In Europe, the 7th Amendment to the Cosmetic Directive bans the use of animal tests to develop new cosmetics and cosmetic ingredients, which has led the cosmetics industry on a pursuit for reconstructed human epidermis (RHE) models as test substrates. Initially, these models were used for purely basic research interests, to obtain a better understanding of skin physiology. Over time, however, it became clear that cosmetic ingredients could be tested with RHEs to validate their efficacy and safety.

Some RHE models have been validated for skin corrosion6 and skin irritation7 testing. However, for most of them, the level of documentation and quality of the barrier function are quite variable, and often inadequately characterized.

One study developed according to OECD principles8 evaluated the skin diffusion of nine compounds with widely varying physicochemical characteristics into three commercially available types of RHEsb-d: human epidermis, pig skin and bovine udder skin. Results demonstrated the quality of the barrier properties in existing RHE models was limited, as their over-predictions in the rate of metabolism9, 10 of topically applied xenobiotics confirmed.11

These results are also supported by the work of Garcia et al.,12 who showed RHEs present a higher TEWL than human skin, and that the penetration rate of caffeine can be estimated as 20 to 25 times higher in RHE models. These inaccuracies are likely caused by the fundamental challenge of producing a stratum corneum (SC) barrier as efficient as natural human skin barrier. Currently, the use of RHEs as models for skin delivery evaluations is limited, and the Scientific Committee on Consumer Safety (SCCS) in Europe has stated it considers their use as being “under development,” noting that such systems “are not yet advised for in vitro testing on the basis of their insufficient barrier function.”13

Skin Explant Models

The pertinence of porcine to human skin from both histological and physiological viewpoints has been established, thus excised specimens have been widely used to test the permeation of ingredients into skin; in contrast, rat skin is considered two to ten times more permeable than human skin14 and is therefore not recommended as a substitute. Pig skin is collected as a by-product from slaughterhouses and in this particular context, is not in contradiction with the Cosmetic Directive.15 Using this approach to collect the test substrate, it is also possible to control the age of the animal skin as well as the body site from which the sample was taken. Suitable membranes, i.e., full dermatomed skin, can thus be prepared to set up reproducible experiments with a large enough sample size.

Obviously, studies with human skin are the most accurate way to determine the capacity of a cosmetic preparation to penetrate. However, its availability, variability and quality can cause practical problems. For example, parameters such as the patient’s age, the sampling site, size of the sample area, the subject’s enlightened consent, etc., can be difficult to obtain. In addition, the skin may be damaged, distended or even injured, as is the case with stretch marks. Due to these difficulties, along with the challenge of maintaining viable skin, only a limited number of studies have been made to determine the delivery of actives through human skin. Indeed, the majority of such experiments are conducted with frozen skin having inactivated biotransformation enzymes, so the influence of skin metabolism on penetration is not considered.

Short-term skin explant cultures have recently been highlighted for their ability to study human skin permeation of an active compound while accounting for cutaneous metabolism, which has gained interest. The major pathways of biotransformation,16 such as via skin enzymes, are now well-described in literature.17 If active cosmetic ingredients are metabolized by cutaneous enzymes, this impacts their overall diffusion rate.18 Although the role of cutaneous biotransformation is well-admitted, its investigation is still limited. Therefore, to study the human skin permeation of an active compound and take potential cutaneous metabolism into account, the development of short-term skin explant culture could represent an interesting alternative.

Living human skin explants were first developed 20 years ago and have been optimized throughout the years. Obtained from fresh skin biopsies, they can be kept for a few hours, or up to 10-14 days on supports or in culture medium under the appropriate conditions; i.e., preferably in serum-free medium in temperature- and CO2-controlled conditions. They offer practical advantages in that they are easy to prepare and use, and they have good barrier function and metabolic activity for skin delivery testing. This paper describes a newly developed skin explant modela,19 and presents studies investigating its skin delivery and cutaneous metabolism assessment capabilities.

Materials and Methods

The tested human skin explant model consists of a 0.5-cm2 full-thickness human skin biopsy embedded in a proprietary nourishing matrix and disposed in a cell culture insert. The system is loaded into multi-well companion plates with a lid. It can be cultured up to seven days with a proprietary defined culture medium, free from serum, hydrocortisone and growth factors. This model retains the features of normal skin in terms of cell populations, viability, structure, metabolism, lipid synthesis and barrier function. To explore percutaneous penetration, the culture insert acts as a diffusion cell, where the insert and well act as the donor and receptor compartments, respectively (see Figure 1).

The first set of experiments was conducted on the test models (n = 6) under the following protocol. A 5-mg/cm2 hydroalcoholic gel sample with 1% caffeine (log P = -0.07), as a reference compound, and 0.3% propylparaben (log P = 2.93), a lipophilic compound of interest, was applied to each skin surface for 24 hr. The defined medium, acting as the receptor medium, was evaluated prior to testing to ensure good solubility of both compounds. To compare the barrier properties of the test model with a non-viable skin model, frozen human skin samples also were tested (n = 3) using Franz diffusion cells (see Figure 2) under the same conditions. The donor was a Caucasian, 66-year-old female; the anatomical site was the abdomen.

After 24 hr, the skin surface, the skin, the matrix, the medium from the new test skin model, and the receptor fluid from the diffusion cells were collected for both studies and analyzed by high-performance liquid chromatography with diode-array detection (HPLC-DAD) to quantify the amount of caffeine and propylparaben in each compartment.

A second set of experiments was conducted by topically applying a blend of four parabens (differing alkyl moieties)—methylparaben, ethylparaben, propylparaben and butylparaben—to the new test skin model (n = 6) to further investigate the metabolism activity of the model. Again, a 5-mg/cm2 sample of gel containing caffeine and the parabens was topically applied for 24 hr. The amounts of parent paraben and 4-hydroxybenzoic acid at 24 hr in the medium, skin and skin surface were collected and analyzed by HPLC.

Results and Discussion

All results are expressed as the percentage of the applied dose recovered in the three compartments: the skin surface, the skin and the medium (see Figure 3). The distribution of caffeine was similar for the new skin model and the frozen human skin samples. After 24 hr diffusion, 30% of the caffeine was still present on the skin surface, meaning the skin barrier function was maintained in both cases. Less than 35% of the applied dose was recovered in the medium fraction, which is relevant for this hydrophilic molecule. Therefore, the test skin model was found to be a suitable in vitro model to follow the skin diffusion of hydrophilic molecules, as the integrity of the barrier function was maintained.

For the propylparaben, however, the distribution profile in the two skin models was completely different. The recovery was approximately 2% in the new skin model but nearly 98% in frozen skin. This shows that propylparaben may be metabolized in viable skin, compared with frozen skin. As an explanation, in viable skin, carboxylesterases may be hydrolyzing the propylparaben. This is in agreement with previous studies reporting that parabens are hydrolyzed in human and animal skin by carboxylesterases into 4-hydroxybenzoic acid.20 In relation, the recovery of propylparaben and its metabolite in the two skin models after 24 hr application of the gel with only propylparaben confirmed the impact of metabolism on skin delivery (see Figure 4).

Using the new skin model as a short-term culture, propylparaben was found to readily absorb into skin, since less than 1% of the applied dose was recovered at the skin surface. The quantitative analysis of 4-hydroxybenzoic acid confirmed this model is metabolically active, compared with the frozen human skin sample, which is not. Significant amounts of this metabolite were detected mainly in the culture media after diffusion through the new skin model.

In the second set of experiments on the test skin model (n = 6) with the four-paraben blend (see Figure 5), the distribution profile of caffeine was used as a reference for a non-metabolized molecule.

After 24 hr skin application, the recovery of each paraben was low, and different distribution profiles were observed. These results matched the different partition coefficients (log P). For example, methylparaben (log P = 1.96) had optimal properties to diffuse quickly through the different layers of the skin, both through the lipophilic SC and the hydrophilic viable epidermis and dermis. It was only recovered in the medium fraction, with a higher fraction not metabolized due to the quick rate of penetration.

On the contrary, butylparaben (log P = 3.57) diffused slowly through the skin due to its higher affinity with the SC and its lower affinity for the viable skin. Higher amounts were therefore recovered on the skin surface and not metabolized. Ethylparaben (log P = 2.51) and propylparaben (log P = 3.04) showed the lowest total recovery, with only traces of their non-metabolized forms recovered.

Consistent with previous studies, these four parabens were hydrolyzed by carboxylesterases mainly into 4-hydroxybenzoic acid after skin absorption. This main metabolite was again analyzed with HPLC-DAD to confirm its detection in the different compartments. Results indicated that nearly 70% of the applied doses of the four parabens were recovered as the metabolite 4-hydroxybenzoic acid—mainly in the medium compartment and in low amounts in the skin. These results indicate the new model is suitable to test the absorption of test compounds and predict their metabolism in skin.


The bioavailability of cosmetic ingredients in the skin is important to efficacy and is formulation-dependent. To ensure that an active ingredient is suitably formulated in a functional cosmetic product, it must follow the 4 Rs rule—i.e., Right chemical, Right location, Right concentration and for the corRect period of time. In relation, its in vitro evaluation for skin penetration constitutes an important step.

While the cosmetics industry has guidelines for the development of such studies, the challenge is to find an alternative model exhibiting both good barrier function and effective metabolic activities. In these respects, the relevance of a new test skin model was evaluated as described here for the delivery of caffeine and propylparaben. The penetration profile of caffeine was found to be the same as in frozen skin, indicating this model provides good barrier function. The data obtained for propylparaben and other parabens also confirmed the new model exhibits good enzymatic activity, which plays a major role in penetration.

Further, as demonstrated by these simple evaluations, the new skin model is easy to use and handle, thus it could be considered a good alternative to test the skin delivery of cosmetic ingredients. Its use represents a bridge between in vitro testing and in vivo clinical studies by achieving a close approximation to reality, enabling much broader studies around percutaneous penetration, e.g., enhancement, reduction, new vehicles, etc., for the cosmetic area.


  1. NA Megrab, AC Williams and BW Barry, Oestradiol permeation through human skin and silastic membrane: Effects of propylene glycol and supersaturation, J Contr Rel 36 277–294 (1995)
  2. WJ Addicks, G Flynn, L Weiner and N Weiner, Validation of a flow-through diffusion cell for use in transdermal research, Pharm Res 4 337–341 (1987)
  3. W Diembeck et al, Test guidelines for in vitro assessment of dermal absorption and percutaneous penetration of cosmetic ingredients, Food Chem Toxicol 37 191–205 (1999)
  4. OECD, Guidelines for the testing of chemicals no. 428, Skin absorption: In vitro Method, Organization for Economic Cooperation and Development, Paris (2004)
  5. OECD, Guidance document no. 28 for the conduct of skin absorption studies, Organization for Economic Cooperation and Development, Paris (2003)
  6. OECD, Guidelines for the testing of chemicals no. 430, In vitro skin corrosion: Transcutaneous electrical resistance test (TER), Organization for Economic Cooperation and Development, Paris 12 (2004)
  7. OECD, Guidelines for the testing of chemicals no. 439, In vitro skin irritation: Reconstructed human epidermis test method, Organization for Economic Cooperation and Development, Paris18 (2010)
  8. T Hartung et al, A modular approach to the ECVAM principles on test validity, Altern Lab Anim 32 467–472 (2004)
  9. M Schäfer-Korting et al, The use of reconstructed human epidermis for skin absorption testing: Results of the validation study, Altern Lab Anim 36 161-187 (2008)
  10. M Van Gele, B Geusens, L Brochez, R Speeckeart and J Lambert, Three-dimensional skin models as tools for transdermal drug delivery: Challenges and limitations, Expert Opin Drug Deliv 8(6) 705–720 (2011)
  11. A Mavon, V Raufast and D Redoules, Skin absorption and metabolism of a new vitamin E prodrug, delta-tocopherol-glucoside: In vitro evaluation in human skin models, J Control Release 100 221–231 (2004)
  12. N Garcia, O Doucet, M Bayer, D Fouchard, L Zastrow and JP Marty, Characterization of the barrier function in a reconstituted human epidermis cultivated in chemically defined medium, Int J Cosmet Sci 24(1) 25–34 (2002)
  13. Scientific Committee on Consumer Safety (SCCS), Basic criteria for the in vitro assessment of dermal absorption of cosmetic ingredients, SCCS/1358/10 (2010)
  14. JH Ross, MH Dong and MI Krieger, Conservatism in pesticide exposure assessment, Regulatory Toxicol and Pharmacol 31 53–58 (2000)
  15. S Richert, A Schrader and K Schrader, Transdermal delivery of two antioxidants from different cosmetic formulations, Int J Cosmet Sci 25 5–13 (2003)
  16. S van Eijl et al, Elucidation of xenobiotic metabolism pathways in human skin and human skin models by proteomic profiling, PLoS One 7(7) 1–13 (2012)
  17. NJ Hewitt et al, Use of human in vitro skin models for accurate and ethical risk assessment: Metabolic considerations, Toxicol Sci 133(2) 209–217 (2013)
  18. C Jacques et al, Effect of skin metabolism on dermal delivery of testosterone: Qualitative assessment using a new short-term skin model, Skin Pharmacol Physiol 27(4) 188 (2014)
  19. www.genoskin.com/en/skin-models/ex-vivo-human-skin-model (Accessed Jun 22, 2014)
  20. C Jewel et al, Hydrolysis of a series of parabens by skin microsomes and cytosol from human and minipigs and in whole skin in short-term culture, Tox Appl Pharm 225 221–228 (2007)
More in Method/Process