- Email: [email protected]
Skin Culture Model: a Possible Alternative to the Use of Excised Human Skin for Assessing In Vitro Percutaneous Absorption
Toxicology in Vitro 12 (1998) 423±430
Skin Culture Model: a Possible Alternative to the Use of Excised Human Skin for Assessing In Vitro Percutaneous Absorption O. DOUCET, N. GARCIA and L. ZASTROW Lancaster Group-Benckiser Cosmetics, International Research Center, BP 669, MC 98014, Monaco (Accepted 27 February 1998) AbstractÐThe interest of a Reconstituted Epidermis (REp) model for screening tests in the ®eld of percutaneous absorption has been investigated. A simpli®ed test procedure was elaborated to screen the percutaneous absorption of drugs through human skin. Using compounds with very dierent physicochemical properties, we compared this new test procedure realized with REp exposed for 14 days at the air±liquid interface to a traditional skin penetration protocol utilized with human skin biopsies and ¯ow-through diusion cells. Whatever the hydrophobicity of the drugs applied, the screening test protocol elaborated allowed determination, by HPLC measurements, of the ability of each test drug to diffuse through the skin. Interestingly, the magnitude of the drug diusion linearly correlated between the two test protocols. For compounds applied under a saturated aqueous solution, we observed that their permeability coecient obtained through REp with the screening procedure was 10 times greater than that found through human skin and a traditional diusion protocol. From these preliminary results, the test procedure elaborated for screening tests using REp cultures seemed to be a possible alternative to the use of traditional percutaneous absorption protocols carried out with human skin samples. # 1998 Elsevier Science Ltd. All rights reserved Abbreviations: ASA = acetyl salicylic acid; BA = benzoic acid; BSA = bovine serum albumin; CAF = caeine; NA = nicotinic acid; OMC = octyl methoxy cinnamate; PBS = phosphate buered saline; pvB5 = pro-vitamin B5; REp = reconstituted epidermis; SA = salicylic acid; SCO = scopolamine; TES = testosterone; THE = theophylline. Keywords: human reconstituted epidermis; percutaneous absorption; human skin.
INTRODUCTION
Traditionally, in vitro percutaneous absorption studies are carried out using animal skin biopsies (pigs, rats) or excised human skin from cosmetic surgery or autopsy. Until now, these kinds of membranes have been reported as useful models for predicting human, in vivo, percutaneous penetration. They may be considered of particular interest for ranking the ability of drug to diuse (Anjo et al., 1980; Bronaugh, 1993), keeping in mind that there are still insucient data concerning the real quantitative relationship existing between in vivo and in vitro percutaneous absorption (Howes et al., 1996). The increasing diculties in obtaining human skin samples urge the pharmaceutical and cosmetic industries to develop new reliable models of membranes which can be successfully used to assess the drug transport. Nowadays, various two-dimensional skin models are commercially available (EpidermTM, Recon-
stituted EpidermisTM, etc.). These skin cultures are widely used in the ®eld of cell toxicology and cell pharmacology where they are recognized as useful tools for screening tests involving cytotoxic and in¯ammatory phenomena (Cohen et al., 1994; Doucet et al., 1996; Ponec and Kempenaar, 1995). The use of these three-dimensional cultures as a membrane for percutaneous absorption studies can therefore be considered one of the most interesting alternatives to excised human skin. In humans, the barrier function of the skin is located in its outermost layer, the stratum corneum (SC) (Scheuplein, 1976). This speci®c epidermis layer is composed of terminally dierentiated keratinocytes (corneocytes) closely stacked upon each other. Intercellular spaces are ®lled with lipids secreted by living keratinocytes during their maturation process (Elias, 1995). To ensure a barrier function against topically applied compounds, skin culture models have to undergo a morphologic dierentiation similar to that seen in vivo. This
0887-2333/98/$19.00+0.00 # 1998 Elsevier Science Ltd. All rights reserved. Printed in Great Britain PII: S0887-2333(98)00023-X
424
O. Doucet et al.
includes the formation of lamellar bodies able to extrude their contents between the cells, creating regular intercellular lipid bilayers (Madison et al., 1988; Ponec et al., 1988). The aim of this study was to develop a very simpli®ed test procedure using a commercial skin culture model which could be easily used to perform screening tests in the ®eld of percutaneous absorption. Using a simpli®ed test protocol, we assessed the permeability of Reconstituted Epidermis (REp) cultivated in chemically-de®ned medium to several drugs characterized by dierent physicochemical properties. In order to determine whether this new test procedure may be of interest for predicting human in vitro percutaneous absorption, the permeation data obtained through REp were compared with data obtained through full-thickness human skin and under a traditional experimental protocol for in vitro percutaneous absorption.
MATERIALS AND METHODS
Compounds Various drugs with dierent chemical structures and physicochemical properties were selected: octyl methoxy cinnamate (OMC) was provided by Haarmann and ReimerTM (France), testosterone (TES) and benzoic acid (BA) by SigmaTM (USA). Caeine (CAF), salicylic acid (SA) and nicotinic acid (NA) were obtained from MerckTM (Germany), acetyl salicylic acid (ASA) from AldrichTM (Germany). Pro-vitamin B5 (pvB5) and theophylline (THE) were supplied by BASFTM (France), and scopolamine (SCO) by BoehringerTM (Germany). Reconstituted epidermis (REp) Reconstituted Epidermis (REp) were supplied by SkinethicTM Laboratory (Nice, France). The main characteristics of this model have been described elsewhere (Rosdy and Clauss, 1990; Rosdy et al., 1993). In brief, a fully dierentiated epithelium having the features of the human epidermis was obtained by culturing normal human keratinocytes in a chemically-de®ned medium, on an inert microporous polycarbonate ®lter Millicell-PCF MilliporeTM (USA) at the air±liquid interface for 14 days. At this time point the horny layer presented more than 10 compact cell layers and expressed the main dierentiation markers (Fartasch et al., 1996). Human skin biopsies Human skin biopsies were obtained from cosmetic surgery on women. Fresh abdominal biopsies were sampled and immediately frozen at ÿ188C. Just before each experiment the skin was thawed at
room temperature. The subcutaneous fat was gently removed using scalpel and forceps. Permeation experiment The purpose of this study was to develop a simple and fast test protocol which could be used to compare the ability of drugs to diuse through the skin. Therefore, a simpli®ed diusion procedure (static diusion procedure) was elaborated in order to avoid the use of a sophisticated diusion system and fraction collector. These studies using REp as membrane were conducted in parallel with whole thickness human skin and a traditional diusing procedure (dynamic diusion procedure) close to those described by the FDA and AAPS report (Skelly et al., 1987), or more recently by COLIPA (1995) in its guideline for percutaneous absorption studies. Simpli®ed diusion procedure used for REp REp (0.6 cm2) were placed in a 12-well culture plate (Becton DickinsonTM, USA), each well containing 0.6 ml of receiver medium. Culture plates were installed on a heating block allowing the receptor medium and the epidermis surface to reach a temperature close to 378C and 328C, respectively. Phosphate buered saline (PBS) (Dulbecco's buffer without calcium magnesium and sodium bicarbonate (Life TechnologiesTM, France) was used as a receptor medium. For lipophilic compounds (TES, OMC, PvB5), 1.5% bovine serum albumin (BSA) (SigmaTM, USA) was added to the receiver medium. Before product application REp set in their diffusing plates were exposed to standard environmental conditions (room temperature 22 2 28C, relative humidity 552 5%) for 2 hr in order to equilibrate with the ambient atmosphere. After a de®ned period of time following the application of drugs solutions (0.25 hr, 0.5 hr, 0.75 hr, 1 hr, 2 hr, 4 hr and 6 hr), REp were gently transferred using forceps to a new well containing fresh medium. At the end of the experiment (6 hr), the receptor medium of each plate well was evaporated for one night under a gentle nitrogen gas ¯ow. Exactly 1 ml of HPLC mobile phase was then added to each well for analysing. Standard diusion procedure used for human skin Full-thickness biopsies were placed in LGATM ¯ow-through diusion cells with 1 cm2 area (LGATM, USA). As in the REp experiment, PBS Dulbecco's buer without calcium magnesium and sodium bicarbonate was used as receptor medium. For lipophilic compounds (TES, OMC, pvB5), 1.5% BSA was added to the medium. The ¯ow rate of receptor ¯uid (4 ml/hr) circulating under the skin was regulated with a peristaltic pump (Ismatec IPCTM, Switzerland) to ensure a complete replacement of
Skin culture model for percutaneous absorption
the receptor with fresh medium between two sampling intervals. The receptor medium was continuously stirred using a te¯on-coated magnet (500 rpm). It was maintained at 378C, allowing the skin surface to reach a temperature of 32 2 18C. Before starting each experiment, human skin biopsies set onto their diusing cells were exposed for 18 hr to standard environmental conditions (room temperature 22 2 28C, relative humidity 55 2 5%) to equilibrate with the ambient atmosphere. Following the application of drugs, the receptor compartment of each cell was sampled hourly for 24 hr using a fraction collector Retriever IVTM (ISCOTM, USA), and analysed by HPLC.
425
aliquots of the digestat were ®ltered through a 0.45mm ®lter (MilliporeTM, USA) and dissolved in mobile phase in order to be analysed. Only experiments in which the total mass balance recovery was 90 2 10% were considered satisfactory. Drugs analysis Drugs were quanti®ed using reverse phase liquid chromatography HPLC (Merck-HitachiTM, Germany) with an UV detector. The nature of the stationary and mobile phases were chosen according to the nature of the diusing compounds and are presented in Table 2.
RESULTS
Product application
Preliminary experiment
pvB5 was applied under a 50% propylene glycol solution while OMC was dissolved in dioctylcyclohexane (Cetiol-STM; Sidobre SinnovaTM, France) under a 5% concentration. All the other drugs were applied under saturated solutions using water as the vehicle, except TES, which was applied under a saturated solution in propylene glycol (OrchidisTM, France). Concentrations of the various drugs applied and their respective vehicles are listed in Table 1. Solutions were applied to the human skin biopsies and REp under a de®ned volume of 2602 11 ml/cm2, creating a large excess of drug on the skin surface during the whole experiment (in®nite dose model). At the end of each experiment the total mass balance of each experiment was determined as follows: the skin surface was washed three times with 200 ml of a 1% soap solution (CetavlonTM, France) and rinsed three times with the same volume of water. The skin was then gently wiped with a cotton tip. The drug contained in the washing liquids and tips was extracted in HPLC mobile phase and measured in order to determine the amount of drug remaining on the skin surface. Skin membranes were cut in small pieces and digested for 48 hr by Trypsine (Life technologiesTM, USA) at 408C. After homogenization and centrifugation (1000 g) for 10 min,
In a previous experiment, saturated solutions of the test compounds were applied to the polycarbonate microporous ®lters used for the REp culture covered or not covered by the skin culture. The diffusion was then recorded for 24 hr to determine the ability of the skin culture to create a barrier to the diusion. The results obtained clearly demonstrated that a 14-day air-exposed culture could form a consistent barrier to the diusion of some topically applied compounds. Under the experimental conditions used, a 6-hr study was sucient to reach a steady-state ¯ux even for slow diusing compounds. Typical kinetic pro®les obtained for some of the test drugs are presented in Fig. 1. Drug screening Taking into account the cumulative amount of drug absorbed at the end of the two protocols, we observed that, under the experimental procedure used, most of the drugs diused much faster through REp than through full-thickness human skin (Fig. 2). However, the simpli®ed test procedure using REp was able to separate fast diusing compounds such as BA, SA, ASA, from molecules, such as SCO, NA, THE, TES, pvB5 and OMC which diused at a slower rate. Owing to the dierence of permeability between the two membranes, signi®cant dierences in the drug diusion could be
Table 1. Characteristics of the compounds applied to human skin and Reconstituted Epidermis Compounds OMC TES BA CAF SA ASA pvB5 NA THE SCO
MW 290 288 112 194 138 180 205 123 180 303
Purity
K oct
98% 99% 99% 99% 99% 99% 98% 99% 99% 99%
/ 2089 (a) 67 (a) 1.02 (a) 181.97 (b) 18.19 (a) / 1.41 (c) / 17.38 (b)
(a) From Roskos et al., 1989; (b) from Barratt, 1995; (c) from Guy et al., 1986.
Vehicle used Dioctylcyclohexane Propylene glycol Distilled water Distilled water Distilled water Distilled water Propylene glycol Distilled water Distilled water Distilled water
Concentration of drug applied (mg/ml) 52 238 3 23 2 3 510 18 6 107
426
O. Doucet et al.
observed even from the ®rst hours following the application, for example, after 2 or 4 hr (Fig. 3). A signi®cant correlation (R2 =0.94, P < 0.0001) was found between the two protocols by considering the cumulative amount of drug absorbed at the end of each type of experiment, for example, after
6 hr through the REp and after 24 hr through the human skin (Fig. 4). Focusing on compounds applied as saturated aqueous solutions (CAF, BA, SA, SCO, NA, THE, ASA) we observed that BA and SA diused very quickly through both human skin and REp, indu-
Fig. 1. Diusion of (a) caeine, (b) testosterone and (c) acetyl salicylic acid through polycarbonate ®lters covered or uncovered by Reconstituted Epidermis (REp). REp results are expressed as the mean 2SEM of a representative experiment with three replicate cultures derived from the same pool of cells. Polycarbonate ®lters results are expressed as the mean2SEM of a representative experiment with three samples.
Skin culture model for percutaneous absorption
427
Table 2. Nature of the HPLC stationary and mobile phases used to quantify the drug diusion Compounds (nm)
Flow rat Mobile phase
Stationary phase
Column
Wavelength (ml/min)
OMC
Lichrosphere 60 5 mm 25 mm
RP select B
ÿ35% Distilled water ÿ65% Methanol
1
308
TES
Lichrosphere 60 7 mm 125 mm
RP18
ÿ30% Distilled water ÿ70% Methanol
1
245
BA
Lichrosphere 60 7 mm 125 mm
RP18
ÿ60% Buer: pH 3 Citrate/hydrochloric acid MERCKTM ÿ40% Methanol
1
230
CAF
Lichrosphere 60 7 mm 125 mm
RP18
ÿ65% Distilled water ÿ35% Methanol
1
272
SA
Lichrosphere 60 5 mm 25 mm
RP select B
ÿ80% Buer: pH 3 KH2PO4 (6.58 g/litre) H3PO4 85% (4.8 ml/litre) ÿ20% Acetonitrile
1
220
ASA
Lichrosphere 60 5 mm 25 mm
RP select B
ÿ80% Buer: pH 3 KH2PO4 (6.58 g/litre) H3PO4 85% (4.8 ml/litre) ÿ20% Acetonitrile
1
220
pvB5
Lichrosphere 60 7 mm 125 mm
RP18
ÿ85% Distilled water ÿ15% Methanol
1
205
NA
Lichrosphere 60 5 mm 25 mm
RP select B
ÿ95% Buer: pH 3 KH2PO4 (6.58 g/litre) H3PO4 85% (4.8 ml/litre) ÿ5% Methanol
1
265
THE
Lichrosphere 60 7 mm 125 mm
RP18
ÿ65% Distilled water ÿ35% Methanol
1
270
SCO
Lichrosphere 60 7 mm 125 mm
RP18
ÿ80% Buer: pH 3 KH2PO4 (6.8 g/litre) H3PO4 85% (400 ml/litre) ÿ20% Acetonitrile
1
254
cing a prompt depletion of the donor compartment. As it was not possible to de®ne a real steady-state ¯ux for these two latter compounds, we used the maximal ¯ux as pseudo steady-state value to calculate the permeability coecient (Bronaugh and Franz, 1986). Under these conditions, a linear correlation (R2 =0.95, P < 0.0002) was found between the permeability coecients obtained for each drug through the two types of membranes (Fig. 5). This suggested that, for these test compounds and this kind of vehicle, the simpli®ed test procedure using REp allowed a 10 times faster diusion of the test compounds than the standard procedure conducted with full-thickness human skin. However, slight dierences between the two experimental protocols used (e.g. static diusion procedure for the screening test with REp and dynamic diusion procedure for the human skin) did not allow conclusions about any dierences between the two membranes in term of permeability. Indeed, the aim of this study was to globally compare two kinds of test protocols and not to directly assess the magnitude of the barrier oered by the REp culture at 14 days.
DISCUSSION
The ultimate advantage of the three-dimensional culture system is to allow direct topical application of insoluble material onto a skin surface in order to reproduce as closely as possible the topical application achieved in vivo. For that purpose, these models have to include in their structure a SC able to ensure its barrier function against topically applied compounds. Former experiments conducted using dierent skin culture models have shown that, despite recognizable barrier properties, skin culture models were more permeable than human skin (Gay et al., 1992; Ponec et al., 1990; Regnier et al., 1992). However, whatever the permeability of these type of culture models we thought that they could be of interest for screenning the drug diusion or for comparing the drug delivery process. In this study we tried to elaborate a simpli®ed experimental protocol which can be easily and quickly used as screening test for in vitro percutaneous absorption studies and which avoids the use of human skin samples.
428
O. Doucet et al.
Fig. 2. Percutaneous absorption of chemical compounds through Reconstituted Epidermis (for 6 hr) and human skin (for 24 hr). REp results are expressed as the mean 2SEM of a representative experiment with three replicate cultures derived from the same pool of cells. Human skin results are expressed as the mean 2SEM of a representative experiment with ®ve skin biopsies derived from the same donor.
The aim of our work was to determine whether the simpli®ed test procedure elaborated using REp culture could be of interest for reducing or replacing the traditional in vitro protocol with human skin and ¯ow-through diusion cells. In order to fully simplify the experimental procedure, we deliberately selected a static diusion protocol rather than a dynamic diusion procedure which is generally presented as reproducing diusion conditions closer to the in vivo situation (Bronaugh and Stewart, 1985). Similarly, REp were used as supplied, without reducing with a glass disc
or a seal the size of the area to which the drugs were applied. This latter point probably increased the global permeability of the REp model since it is generally admitted that the morphological structure of cultured epidermis is impaired in the surrounding of the junction between the skin culture and the plastic insert. Therefore, the permeability characteristics of the skin culture model used could not be considered a representative statement about its real and optimal permeability. In our study, for the comparison of the two test protocols, the molecules were selected on the basis
Fig. 3. Early events in the percutaneous absorption of chemical compounds through Reconstituted Epidermis (Rep). Results are expressed as the mean2 SEM of a representative experiment with three replicate cultures derived from the same pool of cells.
Skin culture model for percutaneous absorption
429
Fig. 4. Correlation between the cumulative amount of drug absorbed through human skin and Reconstituted Epidermis (REp). REp results are expressed as the mean of a representative experiment with three replicate cultures derived from the same pool of cells. Human skin results are expressed as the mean of a representative experiment with ®ve skin biopsies derived from the same donor.
of their respective skin pharmacokinetic behaviour. Previous in vivo studies in humans, had clearly underlined the high diusing potential of BA, SA and ASA (Franz, 1975; Roskos et al., 1989) compared with chemicals such as NA (Franz, 1975). For these compounds, and using acetone as the vehicle, these authors observed a quantitative correlation between in vivo and in vitro data. For the test compounds used in this study and under the experimental conditions used, the REp model seemed more permeable than full-thickness human skin. However, the dierent test protocols used for REp and human skin did not allow the comparison of the permeability of the two mem-
branes. Moreover, it is reasonable to expect that the use of skin culture samples exposed for a longer period to the air±liquid interface could signi®cantly decrease the permeability of this REp model. Recent experiments conducted in that ®eld with the same skin culture model (Fartasch et al., 1996) suggested that the duration of the air exposure period dramatically in¯uenced the ultrastructural intercellular lipid arrangement. This latter point is often presented of major importance in the control of the drug transport (Fartasch and Ponec, 1994; Hou et al, 1991; Potts and Francoeur, 1991). Nevertheless, the simpli®ed test procedure elaborated, which did not require a sterile environment
Fig. 5. Correlation between permeability coecients obtained through Reconstituted Epidermis (REp) and human skin. REp results are expressed as the mean of a representative experiment with three replicate cultures derived from the same pool of cells. Human skin results are expressed as the mean of a representative experiment with ®ve skin biopsies derived from the same donor.
430
O. Doucet et al.
(laminar ¯ow), sophisticated diusing chamber systems, automated samplers or human skin samples, was able to predict the drug diusion obtained in vitro through human skin. Whatever the respective hydrophobicity of the 10 drugs tested, the permeability results obtained linearly correlated between the two protocols. From these preliminary results it seemed that the new screening test procedure using REp culture could be a possible alternative to the traditional percutaneous absorption protocol utilized in vitro with human skin samples. Complementary, in vitro, studies will have to ®rst include additional molecules and dierent kinds of vehicles. The ultimate step will remain the validation of this new, in vitro, protocol by assessing in vivo, in humans the percutaneous penetration of the studied compounds.
REFERENCES
Anjo D. M., Feldman R. J. and Maibach H. I. (1980) Methods for predicting percutaneous penetration in man. In Percutaneous Penetration of Steroids. Edited by P. Mauvais-Jarvais. pp. 31±51. Academic Press, New York. Barratt M. D. (1995) Quantitative structure-activity relationships for skin permeability. Toxicology in Vitro 9, 27±37. Bronaugh R. L. (1993) Diusion cell design. In Topical Drug Bioavailability, Bioequivalence and Penetration. Edited by V. P. Shaw and H. I. Maibach. pp. 117±125. Plenum Press, New York. Bronaugh R. L. and Franz T. J. (1986) Vehicle eects on percutaneous absorption: in vivo and in vitro comparison with human skin. British Journal of Dermatology 115, 1±11. Bronaugh R. L. and Stewart R. F. (1985) Methods for in vitro percutaneous absorption studies. IV: The ¯ow through diusion cell. Journal of Pharmaceutical Sciences 74, 474±480. Cohen C., Dossou G., Rougier A. and Roguet R. (1994) Episkin: an in vitro model for the evaluation of the phototoxicity and sunscreen photoprotective properties. Toxicology in Vitro 8, 669±671. COLIPA (1995) Cosmetic ingredients, guidelines for percutaneous absorption/penetration. The European Cosmetic, Toiletry and Perfumery Association, Brussels. Doucet O., Robert C. and Zastrow L. (1996) Use of a serum-free reconstituted epidermis as a skin pharmacological model. Toxicology in Vitro 3, 305±313. Elias P. M. (1995) Role of lipids in barrier function of the skin. In Pharmacology of the Skin. Edited by H. Mukhtar. pp. 29±38. CRC Press Boca Raton, Ann Arbor. Fartasch M. and Ponec M. (1994) Improved barrier structure formation in air exposed human keratinocyte culture. Journal of Investigative Dermatology 102, 366±374. Fartasch M., Ponec M. and Rosdy M. (1996) The development of a structurally competent epidermal barrier in air-exposed keratinocytes cultures: a time course study. Abstract, 23rd Annual Meeting of the Society for Cutaneous Ultrastructure Research (April 11±13). Journal of Investigative Dermatology 107, 656.
Franz T. J. (1975) Percutaneous absorption. On the relevance of in vitro data. Journal of Investigative Dermatology 64, 190±195. Gay R., Swiderek M., Nelson D. and Ernesti A. (1992) The living skin equivalent as a model in vitro for ranking the toxic potential of dermal irritants. Toxicology in Vitro 6, 303±315. Guy R. H., Carlstom E. M., Bucks D. A. W., Hinz R. S. and Maibach H. I. (1986) Percutaneous penetration of nicotinates: in vivo and in vitro measurements. Journal of Pharmaceutical Sciences 75, 968±972. Hou S. Y. E., Mitra A. K., White S. H., Menon G. K., Ghadially R. and Elias P. M. (1991) Membrane structure in normal and essential fatty acid de®cient stratum corneum: characterized by ruthenium tetroxide staining and X ray diraction. Journal of Investigative Dermatology 96, 215±223. Howes D., Guy R., Hadgraft J., Heylings J., Hoeck U., Kemper F., Maibach H. I., Marty J. P., Merck H., Parra J., Rekkas D., Rondelli I., Schaefer H., Tauber U. and Verbiese N. (1996) Methods for assessing percutaneous absorption. The report and recommendations of ECVAM workshop report 13. ATLA 24, 81±106. Madison K. C., Swartzendruber D. C., Wertz P. W. and Downing D. T. (1988) Lamellar granule extrusion and stratum corneum intercellular lamellae in murine keratinocyte cultures. Journal of Investigative Dermatology 90, 110±116. Ponec M. and Kempenaar J. (1995) Use of human skin recombinants as an in vitro model for testing the irritation potential of cutaneous irritant. Skin Pharmacology 8, 49±59. Ponec M., Wauben-Penris P. J. J., Burger A., Kempenaar J. and Bodde H. E. (1990) Nitroglycerin and sucrose permeability as quality markers for reconstructed human epidermis. Skin Pharmacology 3, 126±135. Ponec M., Weerheim A., Kempenaar J., Mommaas A. M. and Nugteren D. H. (1988) Lipids composition of cultured human keratinocytes in relation to their dierentiation. Journal of Lipid Research 29, 949±961. Potts R. O. and Francoeur M. L. (1991) The in¯uence of stratum corneum morphology on water permeability. Journal of Investigative Dermatology 96, 495±499. Regnier M., Caron D., Reichert U. and Schaefer H. (1992) Reconstructed human epidermis: a model to study in vitro the barrier function of the skin. Skin Pharmacology 5, 49±56. Rosdy M. and Clauss L. C. (1990) Terminal epidermal dierentiation of human keratinocytes grown in chemically-de®ned medium on inert ®lter substrates at the air liquid interface. Journal of Investigative Dermatology 95, 409±414. Rosdy M., Pisani A. and Ortonne J. P. (1993) Production of basement membrane components by a reconstructed epidermis cultured in the absence of serum and dermal factors. British Journal of Dermatology 129, 227±234. Roskos K. V., Maibach H. I. and Guy R. H. (1989) The eect of aging on percutaneous absorption in man. Journal of Pharmacokinetics and Biopharmaceutics 17, 617±630. Scheuplein R. J. (1976) Permeability of the skin : a review of major concepts and some new developments. Journal of Investigative Dermatology 67, 672. Skelly J. P., Shah P., Maibach H. I., Guy R. H., Wester R. C., Flynn G. and Yacobi A. (1987) FDA and AAPS report of the workshop on principles and practices of in vitro percutaneous penetration studies: Relevance to bioavailability and bioequivalence. Pharmaceutical Research 4, 265±267.
Skin Culture Model: a Possible Alternative to the Use of Excised Human Skin for Assessing In Vitro Percutaneous Absorption O. DOUCET, N. GARCIA and L. ZASTROW Lancaster Group-Benckiser Cosmetics, International Research Center, BP 669, MC 98014, Monaco (Accepted 27 February 1998) AbstractÐThe interest of a Reconstituted Epidermis (REp) model for screening tests in the ®eld of percutaneous absorption has been investigated. A simpli®ed test procedure was elaborated to screen the percutaneous absorption of drugs through human skin. Using compounds with very dierent physicochemical properties, we compared this new test procedure realized with REp exposed for 14 days at the air±liquid interface to a traditional skin penetration protocol utilized with human skin biopsies and ¯ow-through diusion cells. Whatever the hydrophobicity of the drugs applied, the screening test protocol elaborated allowed determination, by HPLC measurements, of the ability of each test drug to diffuse through the skin. Interestingly, the magnitude of the drug diusion linearly correlated between the two test protocols. For compounds applied under a saturated aqueous solution, we observed that their permeability coecient obtained through REp with the screening procedure was 10 times greater than that found through human skin and a traditional diusion protocol. From these preliminary results, the test procedure elaborated for screening tests using REp cultures seemed to be a possible alternative to the use of traditional percutaneous absorption protocols carried out with human skin samples. # 1998 Elsevier Science Ltd. All rights reserved Abbreviations: ASA = acetyl salicylic acid; BA = benzoic acid; BSA = bovine serum albumin; CAF = caeine; NA = nicotinic acid; OMC = octyl methoxy cinnamate; PBS = phosphate buered saline; pvB5 = pro-vitamin B5; REp = reconstituted epidermis; SA = salicylic acid; SCO = scopolamine; TES = testosterone; THE = theophylline. Keywords: human reconstituted epidermis; percutaneous absorption; human skin.
INTRODUCTION
Traditionally, in vitro percutaneous absorption studies are carried out using animal skin biopsies (pigs, rats) or excised human skin from cosmetic surgery or autopsy. Until now, these kinds of membranes have been reported as useful models for predicting human, in vivo, percutaneous penetration. They may be considered of particular interest for ranking the ability of drug to diuse (Anjo et al., 1980; Bronaugh, 1993), keeping in mind that there are still insucient data concerning the real quantitative relationship existing between in vivo and in vitro percutaneous absorption (Howes et al., 1996). The increasing diculties in obtaining human skin samples urge the pharmaceutical and cosmetic industries to develop new reliable models of membranes which can be successfully used to assess the drug transport. Nowadays, various two-dimensional skin models are commercially available (EpidermTM, Recon-
stituted EpidermisTM, etc.). These skin cultures are widely used in the ®eld of cell toxicology and cell pharmacology where they are recognized as useful tools for screening tests involving cytotoxic and in¯ammatory phenomena (Cohen et al., 1994; Doucet et al., 1996; Ponec and Kempenaar, 1995). The use of these three-dimensional cultures as a membrane for percutaneous absorption studies can therefore be considered one of the most interesting alternatives to excised human skin. In humans, the barrier function of the skin is located in its outermost layer, the stratum corneum (SC) (Scheuplein, 1976). This speci®c epidermis layer is composed of terminally dierentiated keratinocytes (corneocytes) closely stacked upon each other. Intercellular spaces are ®lled with lipids secreted by living keratinocytes during their maturation process (Elias, 1995). To ensure a barrier function against topically applied compounds, skin culture models have to undergo a morphologic dierentiation similar to that seen in vivo. This
0887-2333/98/$19.00+0.00 # 1998 Elsevier Science Ltd. All rights reserved. Printed in Great Britain PII: S0887-2333(98)00023-X
424
O. Doucet et al.
includes the formation of lamellar bodies able to extrude their contents between the cells, creating regular intercellular lipid bilayers (Madison et al., 1988; Ponec et al., 1988). The aim of this study was to develop a very simpli®ed test procedure using a commercial skin culture model which could be easily used to perform screening tests in the ®eld of percutaneous absorption. Using a simpli®ed test protocol, we assessed the permeability of Reconstituted Epidermis (REp) cultivated in chemically-de®ned medium to several drugs characterized by dierent physicochemical properties. In order to determine whether this new test procedure may be of interest for predicting human in vitro percutaneous absorption, the permeation data obtained through REp were compared with data obtained through full-thickness human skin and under a traditional experimental protocol for in vitro percutaneous absorption.
MATERIALS AND METHODS
Compounds Various drugs with dierent chemical structures and physicochemical properties were selected: octyl methoxy cinnamate (OMC) was provided by Haarmann and ReimerTM (France), testosterone (TES) and benzoic acid (BA) by SigmaTM (USA). Caeine (CAF), salicylic acid (SA) and nicotinic acid (NA) were obtained from MerckTM (Germany), acetyl salicylic acid (ASA) from AldrichTM (Germany). Pro-vitamin B5 (pvB5) and theophylline (THE) were supplied by BASFTM (France), and scopolamine (SCO) by BoehringerTM (Germany). Reconstituted epidermis (REp) Reconstituted Epidermis (REp) were supplied by SkinethicTM Laboratory (Nice, France). The main characteristics of this model have been described elsewhere (Rosdy and Clauss, 1990; Rosdy et al., 1993). In brief, a fully dierentiated epithelium having the features of the human epidermis was obtained by culturing normal human keratinocytes in a chemically-de®ned medium, on an inert microporous polycarbonate ®lter Millicell-PCF MilliporeTM (USA) at the air±liquid interface for 14 days. At this time point the horny layer presented more than 10 compact cell layers and expressed the main dierentiation markers (Fartasch et al., 1996). Human skin biopsies Human skin biopsies were obtained from cosmetic surgery on women. Fresh abdominal biopsies were sampled and immediately frozen at ÿ188C. Just before each experiment the skin was thawed at
room temperature. The subcutaneous fat was gently removed using scalpel and forceps. Permeation experiment The purpose of this study was to develop a simple and fast test protocol which could be used to compare the ability of drugs to diuse through the skin. Therefore, a simpli®ed diusion procedure (static diusion procedure) was elaborated in order to avoid the use of a sophisticated diusion system and fraction collector. These studies using REp as membrane were conducted in parallel with whole thickness human skin and a traditional diusing procedure (dynamic diusion procedure) close to those described by the FDA and AAPS report (Skelly et al., 1987), or more recently by COLIPA (1995) in its guideline for percutaneous absorption studies. Simpli®ed diusion procedure used for REp REp (0.6 cm2) were placed in a 12-well culture plate (Becton DickinsonTM, USA), each well containing 0.6 ml of receiver medium. Culture plates were installed on a heating block allowing the receptor medium and the epidermis surface to reach a temperature close to 378C and 328C, respectively. Phosphate buered saline (PBS) (Dulbecco's buffer without calcium magnesium and sodium bicarbonate (Life TechnologiesTM, France) was used as a receptor medium. For lipophilic compounds (TES, OMC, PvB5), 1.5% bovine serum albumin (BSA) (SigmaTM, USA) was added to the receiver medium. Before product application REp set in their diffusing plates were exposed to standard environmental conditions (room temperature 22 2 28C, relative humidity 552 5%) for 2 hr in order to equilibrate with the ambient atmosphere. After a de®ned period of time following the application of drugs solutions (0.25 hr, 0.5 hr, 0.75 hr, 1 hr, 2 hr, 4 hr and 6 hr), REp were gently transferred using forceps to a new well containing fresh medium. At the end of the experiment (6 hr), the receptor medium of each plate well was evaporated for one night under a gentle nitrogen gas ¯ow. Exactly 1 ml of HPLC mobile phase was then added to each well for analysing. Standard diusion procedure used for human skin Full-thickness biopsies were placed in LGATM ¯ow-through diusion cells with 1 cm2 area (LGATM, USA). As in the REp experiment, PBS Dulbecco's buer without calcium magnesium and sodium bicarbonate was used as receptor medium. For lipophilic compounds (TES, OMC, pvB5), 1.5% BSA was added to the medium. The ¯ow rate of receptor ¯uid (4 ml/hr) circulating under the skin was regulated with a peristaltic pump (Ismatec IPCTM, Switzerland) to ensure a complete replacement of
Skin culture model for percutaneous absorption
the receptor with fresh medium between two sampling intervals. The receptor medium was continuously stirred using a te¯on-coated magnet (500 rpm). It was maintained at 378C, allowing the skin surface to reach a temperature of 32 2 18C. Before starting each experiment, human skin biopsies set onto their diusing cells were exposed for 18 hr to standard environmental conditions (room temperature 22 2 28C, relative humidity 55 2 5%) to equilibrate with the ambient atmosphere. Following the application of drugs, the receptor compartment of each cell was sampled hourly for 24 hr using a fraction collector Retriever IVTM (ISCOTM, USA), and analysed by HPLC.
425
aliquots of the digestat were ®ltered through a 0.45mm ®lter (MilliporeTM, USA) and dissolved in mobile phase in order to be analysed. Only experiments in which the total mass balance recovery was 90 2 10% were considered satisfactory. Drugs analysis Drugs were quanti®ed using reverse phase liquid chromatography HPLC (Merck-HitachiTM, Germany) with an UV detector. The nature of the stationary and mobile phases were chosen according to the nature of the diusing compounds and are presented in Table 2.
RESULTS
Product application
Preliminary experiment
pvB5 was applied under a 50% propylene glycol solution while OMC was dissolved in dioctylcyclohexane (Cetiol-STM; Sidobre SinnovaTM, France) under a 5% concentration. All the other drugs were applied under saturated solutions using water as the vehicle, except TES, which was applied under a saturated solution in propylene glycol (OrchidisTM, France). Concentrations of the various drugs applied and their respective vehicles are listed in Table 1. Solutions were applied to the human skin biopsies and REp under a de®ned volume of 2602 11 ml/cm2, creating a large excess of drug on the skin surface during the whole experiment (in®nite dose model). At the end of each experiment the total mass balance of each experiment was determined as follows: the skin surface was washed three times with 200 ml of a 1% soap solution (CetavlonTM, France) and rinsed three times with the same volume of water. The skin was then gently wiped with a cotton tip. The drug contained in the washing liquids and tips was extracted in HPLC mobile phase and measured in order to determine the amount of drug remaining on the skin surface. Skin membranes were cut in small pieces and digested for 48 hr by Trypsine (Life technologiesTM, USA) at 408C. After homogenization and centrifugation (1000 g) for 10 min,
In a previous experiment, saturated solutions of the test compounds were applied to the polycarbonate microporous ®lters used for the REp culture covered or not covered by the skin culture. The diffusion was then recorded for 24 hr to determine the ability of the skin culture to create a barrier to the diusion. The results obtained clearly demonstrated that a 14-day air-exposed culture could form a consistent barrier to the diusion of some topically applied compounds. Under the experimental conditions used, a 6-hr study was sucient to reach a steady-state ¯ux even for slow diusing compounds. Typical kinetic pro®les obtained for some of the test drugs are presented in Fig. 1. Drug screening Taking into account the cumulative amount of drug absorbed at the end of the two protocols, we observed that, under the experimental procedure used, most of the drugs diused much faster through REp than through full-thickness human skin (Fig. 2). However, the simpli®ed test procedure using REp was able to separate fast diusing compounds such as BA, SA, ASA, from molecules, such as SCO, NA, THE, TES, pvB5 and OMC which diused at a slower rate. Owing to the dierence of permeability between the two membranes, signi®cant dierences in the drug diusion could be
Table 1. Characteristics of the compounds applied to human skin and Reconstituted Epidermis Compounds OMC TES BA CAF SA ASA pvB5 NA THE SCO
MW 290 288 112 194 138 180 205 123 180 303
Purity
K oct
98% 99% 99% 99% 99% 99% 98% 99% 99% 99%
/ 2089 (a) 67 (a) 1.02 (a) 181.97 (b) 18.19 (a) / 1.41 (c) / 17.38 (b)
(a) From Roskos et al., 1989; (b) from Barratt, 1995; (c) from Guy et al., 1986.
Vehicle used Dioctylcyclohexane Propylene glycol Distilled water Distilled water Distilled water Distilled water Propylene glycol Distilled water Distilled water Distilled water
Concentration of drug applied (mg/ml) 52 238 3 23 2 3 510 18 6 107
426
O. Doucet et al.
observed even from the ®rst hours following the application, for example, after 2 or 4 hr (Fig. 3). A signi®cant correlation (R2 =0.94, P < 0.0001) was found between the two protocols by considering the cumulative amount of drug absorbed at the end of each type of experiment, for example, after
6 hr through the REp and after 24 hr through the human skin (Fig. 4). Focusing on compounds applied as saturated aqueous solutions (CAF, BA, SA, SCO, NA, THE, ASA) we observed that BA and SA diused very quickly through both human skin and REp, indu-
Fig. 1. Diusion of (a) caeine, (b) testosterone and (c) acetyl salicylic acid through polycarbonate ®lters covered or uncovered by Reconstituted Epidermis (REp). REp results are expressed as the mean 2SEM of a representative experiment with three replicate cultures derived from the same pool of cells. Polycarbonate ®lters results are expressed as the mean2SEM of a representative experiment with three samples.
Skin culture model for percutaneous absorption
427
Table 2. Nature of the HPLC stationary and mobile phases used to quantify the drug diusion Compounds (nm)
Flow rat Mobile phase
Stationary phase
Column
Wavelength (ml/min)
OMC
Lichrosphere 60 5 mm 25 mm
RP select B
ÿ35% Distilled water ÿ65% Methanol
1
308
TES
Lichrosphere 60 7 mm 125 mm
RP18
ÿ30% Distilled water ÿ70% Methanol
1
245
BA
Lichrosphere 60 7 mm 125 mm
RP18
ÿ60% Buer: pH 3 Citrate/hydrochloric acid MERCKTM ÿ40% Methanol
1
230
CAF
Lichrosphere 60 7 mm 125 mm
RP18
ÿ65% Distilled water ÿ35% Methanol
1
272
SA
Lichrosphere 60 5 mm 25 mm
RP select B
ÿ80% Buer: pH 3 KH2PO4 (6.58 g/litre) H3PO4 85% (4.8 ml/litre) ÿ20% Acetonitrile
1
220
ASA
Lichrosphere 60 5 mm 25 mm
RP select B
ÿ80% Buer: pH 3 KH2PO4 (6.58 g/litre) H3PO4 85% (4.8 ml/litre) ÿ20% Acetonitrile
1
220
pvB5
Lichrosphere 60 7 mm 125 mm
RP18
ÿ85% Distilled water ÿ15% Methanol
1
205
NA
Lichrosphere 60 5 mm 25 mm
RP select B
ÿ95% Buer: pH 3 KH2PO4 (6.58 g/litre) H3PO4 85% (4.8 ml/litre) ÿ5% Methanol
1
265
THE
Lichrosphere 60 7 mm 125 mm
RP18
ÿ65% Distilled water ÿ35% Methanol
1
270
SCO
Lichrosphere 60 7 mm 125 mm
RP18
ÿ80% Buer: pH 3 KH2PO4 (6.8 g/litre) H3PO4 85% (400 ml/litre) ÿ20% Acetonitrile
1
254
cing a prompt depletion of the donor compartment. As it was not possible to de®ne a real steady-state ¯ux for these two latter compounds, we used the maximal ¯ux as pseudo steady-state value to calculate the permeability coecient (Bronaugh and Franz, 1986). Under these conditions, a linear correlation (R2 =0.95, P < 0.0002) was found between the permeability coecients obtained for each drug through the two types of membranes (Fig. 5). This suggested that, for these test compounds and this kind of vehicle, the simpli®ed test procedure using REp allowed a 10 times faster diusion of the test compounds than the standard procedure conducted with full-thickness human skin. However, slight dierences between the two experimental protocols used (e.g. static diusion procedure for the screening test with REp and dynamic diusion procedure for the human skin) did not allow conclusions about any dierences between the two membranes in term of permeability. Indeed, the aim of this study was to globally compare two kinds of test protocols and not to directly assess the magnitude of the barrier oered by the REp culture at 14 days.
DISCUSSION
The ultimate advantage of the three-dimensional culture system is to allow direct topical application of insoluble material onto a skin surface in order to reproduce as closely as possible the topical application achieved in vivo. For that purpose, these models have to include in their structure a SC able to ensure its barrier function against topically applied compounds. Former experiments conducted using dierent skin culture models have shown that, despite recognizable barrier properties, skin culture models were more permeable than human skin (Gay et al., 1992; Ponec et al., 1990; Regnier et al., 1992). However, whatever the permeability of these type of culture models we thought that they could be of interest for screenning the drug diusion or for comparing the drug delivery process. In this study we tried to elaborate a simpli®ed experimental protocol which can be easily and quickly used as screening test for in vitro percutaneous absorption studies and which avoids the use of human skin samples.
428
O. Doucet et al.
Fig. 2. Percutaneous absorption of chemical compounds through Reconstituted Epidermis (for 6 hr) and human skin (for 24 hr). REp results are expressed as the mean 2SEM of a representative experiment with three replicate cultures derived from the same pool of cells. Human skin results are expressed as the mean 2SEM of a representative experiment with ®ve skin biopsies derived from the same donor.
The aim of our work was to determine whether the simpli®ed test procedure elaborated using REp culture could be of interest for reducing or replacing the traditional in vitro protocol with human skin and ¯ow-through diusion cells. In order to fully simplify the experimental procedure, we deliberately selected a static diusion protocol rather than a dynamic diusion procedure which is generally presented as reproducing diusion conditions closer to the in vivo situation (Bronaugh and Stewart, 1985). Similarly, REp were used as supplied, without reducing with a glass disc
or a seal the size of the area to which the drugs were applied. This latter point probably increased the global permeability of the REp model since it is generally admitted that the morphological structure of cultured epidermis is impaired in the surrounding of the junction between the skin culture and the plastic insert. Therefore, the permeability characteristics of the skin culture model used could not be considered a representative statement about its real and optimal permeability. In our study, for the comparison of the two test protocols, the molecules were selected on the basis
Fig. 3. Early events in the percutaneous absorption of chemical compounds through Reconstituted Epidermis (Rep). Results are expressed as the mean2 SEM of a representative experiment with three replicate cultures derived from the same pool of cells.
Skin culture model for percutaneous absorption
429
Fig. 4. Correlation between the cumulative amount of drug absorbed through human skin and Reconstituted Epidermis (REp). REp results are expressed as the mean of a representative experiment with three replicate cultures derived from the same pool of cells. Human skin results are expressed as the mean of a representative experiment with ®ve skin biopsies derived from the same donor.
of their respective skin pharmacokinetic behaviour. Previous in vivo studies in humans, had clearly underlined the high diusing potential of BA, SA and ASA (Franz, 1975; Roskos et al., 1989) compared with chemicals such as NA (Franz, 1975). For these compounds, and using acetone as the vehicle, these authors observed a quantitative correlation between in vivo and in vitro data. For the test compounds used in this study and under the experimental conditions used, the REp model seemed more permeable than full-thickness human skin. However, the dierent test protocols used for REp and human skin did not allow the comparison of the permeability of the two mem-
branes. Moreover, it is reasonable to expect that the use of skin culture samples exposed for a longer period to the air±liquid interface could signi®cantly decrease the permeability of this REp model. Recent experiments conducted in that ®eld with the same skin culture model (Fartasch et al., 1996) suggested that the duration of the air exposure period dramatically in¯uenced the ultrastructural intercellular lipid arrangement. This latter point is often presented of major importance in the control of the drug transport (Fartasch and Ponec, 1994; Hou et al, 1991; Potts and Francoeur, 1991). Nevertheless, the simpli®ed test procedure elaborated, which did not require a sterile environment
Fig. 5. Correlation between permeability coecients obtained through Reconstituted Epidermis (REp) and human skin. REp results are expressed as the mean of a representative experiment with three replicate cultures derived from the same pool of cells. Human skin results are expressed as the mean of a representative experiment with ®ve skin biopsies derived from the same donor.
430
O. Doucet et al.
(laminar ¯ow), sophisticated diusing chamber systems, automated samplers or human skin samples, was able to predict the drug diusion obtained in vitro through human skin. Whatever the respective hydrophobicity of the 10 drugs tested, the permeability results obtained linearly correlated between the two protocols. From these preliminary results it seemed that the new screening test procedure using REp culture could be a possible alternative to the traditional percutaneous absorption protocol utilized in vitro with human skin samples. Complementary, in vitro, studies will have to ®rst include additional molecules and dierent kinds of vehicles. The ultimate step will remain the validation of this new, in vitro, protocol by assessing in vivo, in humans the percutaneous penetration of the studied compounds.
REFERENCES
Anjo D. M., Feldman R. J. and Maibach H. I. (1980) Methods for predicting percutaneous penetration in man. In Percutaneous Penetration of Steroids. Edited by P. Mauvais-Jarvais. pp. 31±51. Academic Press, New York. Barratt M. D. (1995) Quantitative structure-activity relationships for skin permeability. Toxicology in Vitro 9, 27±37. Bronaugh R. L. (1993) Diusion cell design. In Topical Drug Bioavailability, Bioequivalence and Penetration. Edited by V. P. Shaw and H. I. Maibach. pp. 117±125. Plenum Press, New York. Bronaugh R. L. and Franz T. J. (1986) Vehicle eects on percutaneous absorption: in vivo and in vitro comparison with human skin. British Journal of Dermatology 115, 1±11. Bronaugh R. L. and Stewart R. F. (1985) Methods for in vitro percutaneous absorption studies. IV: The ¯ow through diusion cell. Journal of Pharmaceutical Sciences 74, 474±480. Cohen C., Dossou G., Rougier A. and Roguet R. (1994) Episkin: an in vitro model for the evaluation of the phototoxicity and sunscreen photoprotective properties. Toxicology in Vitro 8, 669±671. COLIPA (1995) Cosmetic ingredients, guidelines for percutaneous absorption/penetration. The European Cosmetic, Toiletry and Perfumery Association, Brussels. Doucet O., Robert C. and Zastrow L. (1996) Use of a serum-free reconstituted epidermis as a skin pharmacological model. Toxicology in Vitro 3, 305±313. Elias P. M. (1995) Role of lipids in barrier function of the skin. In Pharmacology of the Skin. Edited by H. Mukhtar. pp. 29±38. CRC Press Boca Raton, Ann Arbor. Fartasch M. and Ponec M. (1994) Improved barrier structure formation in air exposed human keratinocyte culture. Journal of Investigative Dermatology 102, 366±374. Fartasch M., Ponec M. and Rosdy M. (1996) The development of a structurally competent epidermal barrier in air-exposed keratinocytes cultures: a time course study. Abstract, 23rd Annual Meeting of the Society for Cutaneous Ultrastructure Research (April 11±13). Journal of Investigative Dermatology 107, 656.
Franz T. J. (1975) Percutaneous absorption. On the relevance of in vitro data. Journal of Investigative Dermatology 64, 190±195. Gay R., Swiderek M., Nelson D. and Ernesti A. (1992) The living skin equivalent as a model in vitro for ranking the toxic potential of dermal irritants. Toxicology in Vitro 6, 303±315. Guy R. H., Carlstom E. M., Bucks D. A. W., Hinz R. S. and Maibach H. I. (1986) Percutaneous penetration of nicotinates: in vivo and in vitro measurements. Journal of Pharmaceutical Sciences 75, 968±972. Hou S. Y. E., Mitra A. K., White S. H., Menon G. K., Ghadially R. and Elias P. M. (1991) Membrane structure in normal and essential fatty acid de®cient stratum corneum: characterized by ruthenium tetroxide staining and X ray diraction. Journal of Investigative Dermatology 96, 215±223. Howes D., Guy R., Hadgraft J., Heylings J., Hoeck U., Kemper F., Maibach H. I., Marty J. P., Merck H., Parra J., Rekkas D., Rondelli I., Schaefer H., Tauber U. and Verbiese N. (1996) Methods for assessing percutaneous absorption. The report and recommendations of ECVAM workshop report 13. ATLA 24, 81±106. Madison K. C., Swartzendruber D. C., Wertz P. W. and Downing D. T. (1988) Lamellar granule extrusion and stratum corneum intercellular lamellae in murine keratinocyte cultures. Journal of Investigative Dermatology 90, 110±116. Ponec M. and Kempenaar J. (1995) Use of human skin recombinants as an in vitro model for testing the irritation potential of cutaneous irritant. Skin Pharmacology 8, 49±59. Ponec M., Wauben-Penris P. J. J., Burger A., Kempenaar J. and Bodde H. E. (1990) Nitroglycerin and sucrose permeability as quality markers for reconstructed human epidermis. Skin Pharmacology 3, 126±135. Ponec M., Weerheim A., Kempenaar J., Mommaas A. M. and Nugteren D. H. (1988) Lipids composition of cultured human keratinocytes in relation to their dierentiation. Journal of Lipid Research 29, 949±961. Potts R. O. and Francoeur M. L. (1991) The in¯uence of stratum corneum morphology on water permeability. Journal of Investigative Dermatology 96, 495±499. Regnier M., Caron D., Reichert U. and Schaefer H. (1992) Reconstructed human epidermis: a model to study in vitro the barrier function of the skin. Skin Pharmacology 5, 49±56. Rosdy M. and Clauss L. C. (1990) Terminal epidermal dierentiation of human keratinocytes grown in chemically-de®ned medium on inert ®lter substrates at the air liquid interface. Journal of Investigative Dermatology 95, 409±414. Rosdy M., Pisani A. and Ortonne J. P. (1993) Production of basement membrane components by a reconstructed epidermis cultured in the absence of serum and dermal factors. British Journal of Dermatology 129, 227±234. Roskos K. V., Maibach H. I. and Guy R. H. (1989) The eect of aging on percutaneous absorption in man. Journal of Pharmacokinetics and Biopharmaceutics 17, 617±630. Scheuplein R. J. (1976) Permeability of the skin : a review of major concepts and some new developments. Journal of Investigative Dermatology 67, 672. Skelly J. P., Shah P., Maibach H. I., Guy R. H., Wester R. C., Flynn G. and Yacobi A. (1987) FDA and AAPS report of the workshop on principles and practices of in vitro percutaneous penetration studies: Relevance to bioavailability and bioequivalence. Pharmaceutical Research 4, 265±267.