Content uploaded by Ulrich F Schaefer
Author content
All content in this area was uploaded by Ulrich F Schaefer on Mar 14, 2014
Content may be subject to copyright.
Introduction
To protect human health and the environment,
substances and products are evaluated by toxicolog-
ical testing, hazard analysis and risk assessment.
During the next few decades, the REACH
(Registration, Evaluation and Authorisation of
Chemicals) initiative of the European Union (EU)
will lead to an increase in the toxicological testing of
chemicals. Yet animal welfare is of ethical concern
to Western society, so alternatives to animal
experiments are necessary. Non-animal testing pro-
cedures for regulatory purposes must be independ-
ently validated, to ensure that they provide relevant
and reliable data for hazard prediction and risk
assessment in humans. The principles to ensure
test validity and the prerequisites for experimental
validation have been published by the European
Centre for the Validation of Alternative Methods
(ECVAM; 1) and the Organisation for Economic
Cooperation and Development (OECD; 2).
Prevalidation involves performing a small-scale
inter-laboratory study with a few test samples. In
the first steps, an experimental procedure is
established, transferred to another laboratory,
and refined. The refined protocol is subsequently
evaluated with respect to reliability and relevance
by testing a limited number of compounds in at
least three laboratories. If the procedure is
deemed suitable with respect to transferability
and performance, it is then tested in a larger-scale
validation study. In addition, a prediction model is
developed, which is then tested to assess the pre-
dictive capacity of the test system. Indeed, the EU
(2000) and the OECD have already accepted the
scientifically validated in vitro tests for phototox-
icity, skin corrosion and embryotoxicity testing
(for reviews see 3, 4).
Reconstructed Human Epidermis for Skin Absorption
Testing: Results of the German Prevalidation Study
Monika Schäfer-Korting,
1
Udo Bock,
2
Armin Gamer,
3
Annekathrin Haberland,
1
Eleonore
Haltner-Ukomadu,
2
Monika Kaca,
2
Hennicke Kamp,
3
Manfred Kietzmann,
4
Hans Christian
Korting,
5
Hans-Udo Krächter,
6
Claus-Michael Lehr,
7
Manfred Liebsch,
8
Annette Mehling,
6
Frank
Netzlaff,
7
Frank Niedorf,
4
Maria K. Rübbelke,
5
Ulrich Schäfer,
7
Elisabeth Schmidt,
8
Sylvia
Schreiber,
1
Klaus-Rudolf Schröder,
9
Horst Spielmann
8
and Alexander Vuia
1
1
Freie Universität Berlin, Institut für Pharmazie, Berlin, Germany;
2
Across Barriers GmbH, Saarbrücken,
Germany;
3
BASF Aktiengesellschaft, Experimental Toxicology and Ecology, Ludwigshafen, Germany;
4
Stiftung Tierärztliche Hochschule Hannover, Institut für Pharmakologie, Hannover, Germany;
5
Ludwig-
Maximilians-Universität München, Klinik und Poliklinik für Dermatologie und Allergologie, München,
Germany;
6
Cognis Deutschland GmbH & Co. KG, Düsseldorf, Germany;
7
Universität des Saarlandes,
Biopharmazie u. Pharmazeutische Technologie, Saarbrücken, Germany;
8
ZEBET, Berlin, Germany;
9
Henkel
KGaA, Düsseldorf, Germany
Summary — Exposure to chemicals absorbed by the skin can threaten human health. In order to stan-
dardise the predictive testing of percutaneous absorption for regulatory purposes, the OECD adopted
guideline 428, which describes methods for assessing absorption by using human and animal skin. In this
study, a protocol based on the OECD principles was developed and prevalidated by using reconstructed
human epidermis (RHE). The permeation of the OECD standard compounds, caffeine and testosterone,
through commercially available RHE models was compared to that of human epidermis and animal skin. In
comparison to human epidermis, the permeation of the chemicals was overestimated when using RHE. The
following ranking of the permeation coefficients for testosterone was obtained: SkinEthic > EpiDerm,
EPISKIN > human epidermis, bovine udder skin, pig skin. The ranking for caffeine was: SkinEthic, EPISKIN
> bovine udder skin, EpiDerm, pig skin, human epidermis. The inter-laboratory and intra-laboratory repro-
ducibility was good. Long and variable lag times, which are a matter of concern when using human and
pig skin, did not occur with RHE. Due to the successful transfer of the protocol, it is now in the validation
process.
Key words: bovine udder skin, human skin, human skin models, pig skin, percutaneous absorption,
prevalidation, reconstructed human epidermis, skin absorption.
Address for correspondence: M. Schäfer-Korting, Freie Universität Berlin, Institut für Pharmazie,
Königin-Luise-Straße 2–4, 14195 Berlin, Germany.
E-mail: msk@zedat.fu-berlin.de
ATLA 34, 283–294, 2006 283
In 2003–2004, the OECD released Test
Guidelines (TGs) 427 and 428, which are used to
study the percutaneous uptake of chemicals in vivo
(5) and in vitro (6), and are accompanied by
Technical Guidance Document (TGD) 28 (7). This
is of key importance, since the skin is the third
major absorption organ — after the gastrointestinal
tract and the lung — and is the primary exposure
route for chemicals such as pesticides (8, 9).
Moreover, several groups of professionals are
exposed to dangerous chemicals in an occupational
setting via dermal absorption; for example, the poi-
soning of tobacco farmers due to the percutaneous
absorption of nicotine when handling wet tobacco
leaves (10).
Since in vitro studies are aimed at predicting skin
absorption in man, viable human skin should be
used for testing in preference to animal skin, which
is generally more permeable — except for porcine
skin (11–14). Nevertheless, due to their ready avail-
ability, in vitro studies are also conducted with pig
and rat skin. Rat skin is often used for in vitro test-
ing, since this species is also used for the estimation
of toxicity following single and repeated dose
administration, as well as to study the percuta-
neous absorption of chemicals and formulations in
vivo (5). In the rat, an in vitro–in vivo comparison
of the skin absorption of eight pesticides showed
that absorption in vitro was 2–3-fold higher than in
vivo (15). While the OECD proposals will accelerate
the spread of the diffusion (Franz) cell technique
for penetration studies, other in vitro approaches,
such as the isolated perfused porcine skin flap (16,
17), perfused pig forelimb (18, 19) and perfused
bovine udder (20, 21), can be used to study specific
problems.
Due to the limited availability of excised human
skin for experimental purposes, human-based alter-
natives for uptake testing are being investigated. In
recent years, organotypic models have become more
useful for investigators, and today, reconstructed
human epidermis (RHE) models are commercially
available. They are well described with respect to
tissue architecture and lipid composition (22, 23),
and have already proved to be of value for corrosiv-
ity testing (24, 25). In a project funded by the 5th
Framework Programme of the EU and the German
Research Foundation, the uptake of reference stan-
dard compounds by animal skin was compared to
that by human skin (26, 27). These studies also
included investigations on formulations (28–31). In
another investigation, the penetration properties of
four dermatological drugs with human, pig and rat
skin were compared with results with the
SkinEthic
®
RHE and the Graftskin™ LSE models
(32). In fact, the OECD TGD 28 states that recon-
structed human skin models can be used for hazard
assessment, if the data obtained with reference
chemicals are consistent with those in the pub-
lished literature (7). Therefore, RHE may become
another favoured test matrix. Scientists in the field
of skin absorption testing have decided to define
what relevant procedures are, and to evaluate them
as the basis for validation studies.
The results of initial experiments, in which the
skin permeation of caffeine, testosterone and a dye
was compared, permitted the identification of
highly relevant parameters for ensuring repro-
ducible permeation. Aspects related to protocol
development will be, or have been, published sepa-
rately (33–35). Based on the outcome of an initial
comparison of a small number of comparative
experiments, the authors agreed on a test protocol
which was then tested thoroughly by five toxicolog-
ical laboratories in Germany (both in academia and
in industry) and at ZEBET. This prevalidation
study, funded by the German Ministry of Education
and Research (BMBF), was carried out to qualify
RHE for uptake studies. The test protocols focused
on skin permeation and the results of the inter-lab-
oratory comparison are reported here.
Materials and Methods
The study was performed according to the modular
approach for validation, a stepwise procedure pro-
posed by ECVAM (1). Uptake tests using human,
pig and rat skin can be regarded as retrospectively
validated due to the approved OECD in vitro TG
428 (6), and data generated with RHE can therefore
be compared to those for uptake with human and
animal (pig) skin. The experiments were conducted
from March to August 2004, in general accordance
with the principles of Good Laboratory Practice.
Study management and organisation
An initial test protocol was developed, based on the
principles of TGD 28 (7). Specific questions con-
cerning test substance preparation, analytics (tech-
niques used for quantification), and composition
and histopathological structure of the skin models
used, as well as the handling of the different mod-
els, were each evaluated by one of the participating
laboratories. Based on the outcome of these experi-
ments, refined and detailed standard operation pro-
cedures (SOPs) were defined for conducting the
skin penetration tests.
Test chemicals and preparation of solutions
For cutaneous uptake studies, the OECD proposes
the use of caffeine (logP = 0.01; MW = 194; 58-08-
2) and testosterone (logP = 3.32; MW = 288; 58-22-
0) as reference standard substances with a low or
high lipophilicity, respectively. Therefore, these
agents were included in this study. The chemicals
284 M. Schäfer-Korting et al.
were obtained from Sigma (St. Louis, MO, USA), as
were ethanol (64-17-5), Igepal
®
CA-630 ([octylphe-
noxy]polyethoxyethanol, 9043-52-1) and phosphate
buffered saline (PBS), pH 7.4. All the laboratories
used donor solutions of caffeine (0.1%, 284.1µg/cm²)
and testosterone (0.004%, 11.36µg/cm²), which were
freshly prepared by dissolving caffeine in PBS and
diluting the stock solution of testosterone
(10mg/ml 96% ethanol) in PBS containing 2% (v/v)
of the solubiliser, Igepal
®
CA-630. The stock solu-
tions were stable for at least four weeks when
stored at 4°C. Two laboratories assessed penetra-
tion by using radiochemical detection and therefore
spiked the samples with 1-methyl-
14
C-caffeine
(51.2mCi/mmol; 77196-81-7; Perkin Elmer Life
Sciences, Boston, MA, USA) and 1,2,6,7-³H-testo-
sterone (100Ci/mmol; 6384-79-8; Amersham, Frei-
burg, Germany), both at purities of higher than
97%, to achieve a total radioactivity of 1µCi per
Franz cell.
Reconstructed human epidermis models and
skin
RHE was purchased from three manufacturers:
EpiDerm™ Skin Model (EPI-606X) from MatTek
Corporation (Ashland, MA, USA); Reconstructed
Human Epidermis Kit EPISKIN
®
(J13, 1.07cm²)
from L’Oréal (Paris, France); and SkinEthic
®
Skin
Model (RHE/L/17: Reconstituted Human Epider-
mis, large, age day 17, 4.00cm²) from Laboratoire
SkinEthic (Nice, France). All the RHEs were
shipped for delivery on a Tuesday or Wednesday
morning, and were used for the experiments within
24 hours after delivery, according to the recommen-
dation of the manufacturers. The storage period
was documented for each experiment. All the han-
dling before incubation was performed under a ster-
ile air flow. Before opening the EPISKIN kit, the
integrity of the kit was verified by the colour of the
agar medium and the temperature indicator. The
EPISKIN and SkinEthic tissues were removed from
nutrient agar immediately after delivery, transferred
into six-well plates (SkinEthic) or twelve-well plates
(EPISKIN), filled with the manufacturer’s mainte-
nance media, then kept overnight in an incubator at
37°C and 5% CO
2
. The EpiDerm™ tissue was stored
overnight at 4°C. The next morning, the tissues were
transferred into six-well plates and kept in the incu-
bator at 37°C and 5% CO
2
, for at least 1 hour.
Human skin (abdomen or breast) was obtained
from females aged 20–75 years, who had been sub-
jected to cosmetic surgery. Pig (Deutsche Landrasse
breed; no soaking of the cadaver in boiling water)
and bovine udder skin (Schwarzbunte breed) were
obtained from local abattoirs. The skin was placed
in an ice-cold cloth and immediately transferred to
the laboratories. Great care was taken to avoid con-
tamination of the skin surface by subcutaneous
lipids. In the laboratories, subcutaneous fat and
connective tissue were removed from the skin, and
the tissue then subjected to cryopreservation at
–25°C for at least 1 day, up to a maximum of 6
months. The skin was thawed immediately before
performing the experiments. Human epidermis
sheets (HES) were prepared from human skin by
heat separation (34, 36), whereas skin with a thick-
ness of 1000 ± 100µm was prepared from pig and
udder skin by using a Dermatome™ (Aesculap,
Tuttlingen, Germany). Alternatively, the upper side
of the skin was frozen to obtain split skin of identi-
cal thickness by a microtome (Leica 1325CM,
Nussloch, Germany). A detailed comparison based
on pig skin from the same donor did not indicate
any differences in the permeation of testosterone
when these different approaches were used for skin
preparation.
Refined test protocol and SOPs
The refined test protocols were transformed into
SOPs and used for this study. Except for the inspec-
tion of tissue integrity, the protocols used were
more detailed than TGD 28 (7) but followed the pro-
cedures described therein. Briefly, following a
visual check of tissue integrity by using a magnify-
ing glass, human epidermis sheets, animal skin
(both rehydrated in PBS for 30 minutes) and RHE
were mounted in Franz cells (15mm in diameter,
with a receptor chamber of 12cm³; PermeGear,
Bethlehem, PA, USA). Tissues were discarded if
there was a wet skin surface due to the appearance
of receptor medium. Since the size of the available
EPISKIN models was too small to mount directly
into the Franz cells used, a special EPISKIN insert
was made, in order to accommodate the reduced
surface area of 0.357cm
2
. The stratum corneum was
placed facing the air and the dermis was in contact
with the supporting membrane and the receptor
medium PBS. The receptor medium was kept at a
constant temperature of 33.5 ± 0.5°C by using an
incubator or a water bath, and stirred with a mag-
netic bar at 500rpm. The absence of air bubbles was
monitored throughout the experiment.
After equilibration for 30 minutes, 500µl of the
donor solution was applied to the skin surface of the
tissues. When using EPISKIN, 110µl was used to
adapt the applied amount to the smaller surface
area. The opening of the Franz cell was kept cov-
ered by Parafilm
®
. In all the experiments,
281.4µg/cm² caffeine or 11.36µg/cm² testosterone
was applied and left in place for the entire experi-
ment. The concentrations of the donor solutions
chosen ensured that the concentration in the recep-
tor fluid was clearly below the solubility limits of
the compounds. The uptake of caffeine and testos-
terone was measured by high performance liquid
chromatography (HPLC; 3 laboratories) or radio-
Reconstructed human epidermis for skin absorption testing 285
chemical detection (2 laboratories) of the amounts
that had permeated into the receptor medium. The
receptor medium was sampled at 6 hours and 24
hours, as well as at additional time points selected
by the individual centres, to estimate skin perme-
ation from a regression line based on six valid sam-
ples. Thus, the individual sampling times varied
between the types of test skin and the laboratories.
To assure the free diffusibility of test compounds,
the saturation of solubility was determined by
preparing saturated solutions of caffeine and testos-
terone in PBS (pH 7.4), which were kept at 37°C.
After 24 hours, an aliquot was taken and filtered
immediately. Then the solution was diluted and the
concentrations of the compounds were quantified
by HPLC.
Analysis
Laboratories using HPLC analysis to quantitate the
permeated substances in the receptor medium used
a validated method. Briefly, a Waters Alliance HT
2695 equipped with a Waters 996 Photo diode array
detector (Waters, Milford, MA, USA) was used for
detection. The HPLC column used (XTerra MS
C18; 50 × 2.1mm i.d. 5µm; Waters) was maintained
at 40°C. The mobile phase consisted of a 10mM
phosphate buffer, adjusted to pH 3.5 with o-phos-
phoric acid; the organic modifier was acetonitrile
(MeCN). Chromatography was performed by using
a gradient (5–95% MeCN from 0–5 min, 5% MeCN
from 5.1–7 min). The flow rate was set to
0.6ml/minute, and 10µl of the solution was injected.
For caffeine, concentrations ranging from 0.25 to
10µg/ml, corresponding to the linear portion of the
regression line, with a correlation coefficient of
0.999, were used for calibration; the limit of detec-
tion (LOD) was 0.25µg/ml. For the calibration of
testosterone, concentrations from the linear por-
tion of the curve ranging from 0.025 to 10µg/ml and
exhibiting a correlation coefficient of 0.999 were
used; the LOD was 0.025µg/ml. In addition, an iso-
cratic HPLC procedure was used after testing for
validity. Radiolabelled caffeine and testosterone
were quantified in the receptor medium by scintil-
lation counting (Microbeta Plus, Wallac, Turku,
Finland) employing a scintillation cocktail
(Optiphase Supermix, Wallac), as previously
described (34). The results were comparable, inde-
pendently of the method chosen — HPLC or radio-
chemical quantification.
Data analysis procedure/biostatistical
methods
Human epidermis sheets were tested in three labo-
ratories, and reconstructed epidermis and animal
skin in two laboratories. Animal skin was obtained
from at least two donors, RHE from at least two
batches was used, and the human epidermis sheets
originated from three donors. In general, experi-
ments with human epidermis sheets were per-
formed in triplicate for each donor, and
experiments involving the other matrices were con-
ducted in quadruplicate for each donor or batch.
Therefore, in general, eight tests for each test agent
and skin type (nine for human epidermis sheets)
were run by the laboratory in charge, resulting in
the performance of about 25 parallel experiments
for each substance (Table 1).
Permeation values were given as caffeine or
testosterone amounts in the receptor medium at 6
hours (µg/cm²), normalised to the exposed skin sur-
face, as well as by the apparent permeability coeffi-
cient P
app
(= [V/A*C
i
]*dC
A
/dt), which takes the
exposed surface area (A) into account (EpiDerm,
SkinEthic: 1.768cm²; EPISKIN: 0.385cm²). The vol-
ume (V) was 12cm
3
in all the experiments, with the
exception of experiments with EPISKIN due to the
different size of the EPISKIN insert (11.4cm
3
). C
i
gives the initial concentration of the applied sub-
stance in µg/cm³, and dC
A
/dt is the increasing con-
centration of the substances in the receptor
medium with increasing time. The P
app
value and
lag time (intersection of the linear part of the
regression line with the x-axis) are derived from the
skin of each donor or batch. Data are presented as
either the arithmetic mean value, standard devia-
tion (sd) and the coefficient of variation (CV) of the
experiments performed with the skin of each
donor/batch, or as the mean of the respective labo-
ratory, as indicated. For each test agent, differences
of P
app
were analysed by a two-factor analysis of
variance (ANOVA) with the factor “skin” (6 types)
and the factor “laboratory” (5 types). Intra-labora-
tory and inter-laboratory variations were calculated
according to DIN ISO 5725-2. The parameter s
r
%
estimates the intra-laboratory CV, and s
L
%
estimates the inter-laboratory CV, to describe the
repeatability (37).
Results
The experimental procedure was first tested to
ensure that the concentrations of the test com-
pounds in the receptor medium did not exceed the
permitted range of 10% saturation solubility. The
maximum caffeine concentration was 10.8µg/ml, cor-
responding to 0.04% saturation solubility (c
s
:
31.08mg/ml, pH 7.4). With respect to testosterone,
0.6556µg/ml was found, corresponding to 2.09% sat-
uration solubility (c
s
: 31.31µg/ml, pH 7.4). Thus, dif-
fusion was not restricted.
The partner laboratories performed the experi-
ments according to the refined SOPs. There were
only very few variations of the protocol procedures.
Since testosterone concentrations in the receptor
286 M. Schäfer-Korting et al.
fluid were below the detection limit when using pig
skin with 1000µm thickness, one laboratory used
pig skin with a thickness of 700µm. These experi-
ments were repeated with skin of the correct
dimension. These latter data were included in the
further processing.
Figure 1 depicts mean caffeine concentrations
and Figure 2 mean testosterone concentrations in
the receptor media, as determined by the partner
laboratories. When evaluating the concentrations of
both substances, permeation of human epidermis
sheets showed pronounced variability of the P
app
value. The CVs for caffeine and testosterone were
62% and 93%, respectively. The CV of the P
app
val-
ues was less with RHE (20–55%). Significant differ-
ences between the two laboratories studying the
same tissue were not found, except for caffeine per-
meation when using the SkinEthic model (Figure
Table 1: Donors of skin and batches of reconstructed human epidermis, as used in the
partner laboratories
Laboratory
Skin type FU US TiHo LMU ACB
Human epidermal sheets/ breast, 39
c
abdomen, 39
c,t
breast, 30
c,t
experiments in triplicate breast, 50
t
abdomen, 49
c
breast, 55
c,t
per donor breast, 55
c
abdomen, 51
c,t
breast, 52
c,t
breast, 58
c
breast, 72
t
Pig skin S030
c
S027
c
S032
c
S028
c
S033
t
S034
c
S037
t
S040
t
S041
t
Bovine udder skin R023
c
126-0503
c
R031
c
178-0904
c
R034
c
179-0904
c
R044
t
180-0904
c
R045
t
181-0904
t
R047
t
183-1104
t
184-1104
t
SkinEthic (04/05 022A) 0414
c,t
0402
c
0501
c,t
0405
t
0406
t
0501
t
0503
c
0702
c
0704
c
EpiDerm 5025
c
5025
c,t
5034
t
5412
c,t
5036
t
5619
c,t
5412
c
EPISKIN 021
c,t
014
c,t
(04 EPIS-J13) 022
c,t
015
c,t
031
c,t
018
c,t
031
c,t
The tissues used were from at least two donors or batches; experiments were run in quadruplicate for each batch or
donor.
FU = Freie Universität Berlin; US = Universität des Saarlandes; TiHo = Stiftung Tierärztliche Hochschule
Hannover; LMU = Ludwig-Maximilians-Universität München; ACB = Across Barriers GmbH;
c
= caffeine;
t
= testosterone.
Reconstructed human epidermis for skin absorption testing 287
1e; p ≤ 0.05). A more detailed analysis of differences
between RHE models is based on P
app
values, total
permeation within 6 hours, and lag times, as sum-
marised in Table 2. In fact, the amounts of testo-
sterone and caffeine permeated after 6 hours are in
good accordance with the P
app
values.
Testosterone permeation
Testosterone permeation through the EpiDerm and
EpiSkin model (P
app
: 2.89 ± 1.09 × 10
–6
cm/s and 2.11
± 0.63 × 10
–6
cm/s, respectively) was very similar,
while the SkinEthic model appeared to be more per-
a = EpiDerm; b = human epidermis sheets; c = EPISKIN; d = pig skin; e = SkinEthic; f = bovine udder skin.
Permeation was measured following the application of 500µL or 110µl (EPISKIN) of a 0.1% caffeine solution in PBS
for at least 6 hours to reconstructed epidermis (a, c, e), or for 26–30 hours to human epidermis sheets or split skin (b, d,
f). Each line shows the results obtained within a single laboratory.
= Across Barriers GmbH; = Freie Universität Berlin; = Ludwig-Maximilians-Universität München;
= Stiftung Tierärztliche Hochschule Hannover; = Universität des Saarlandes.
Figure 1: Caffeine permeation into phosphate buffered saline in different laboratories
0
time (hours)
permeation (µg/cm
2
)
5
f)
10 15 20 25 30
100
80
60
40
20
0
0
time (hours)
permeation (µg/cm
2
)
5
d)
10 15 20 25 30
100
80
60
40
20
0
0
time (hours)
permeation (µg/cm
2
)
5
b)
10 15 20 25 30
100
80
60
40
20
0
0
time (hours)
permeation (µg/cm
2
)
2
e)
4 6 8 10 12
100
80
60
40
20
0
0
time (hours)
permeation (µg/cm
2
)
2
a)
4 6 8 10 12
100
80
60
40
20
0
0
time (hours)
permeation (µg/cm
2
)
2
c)
4 6 8 10 12
100
80
60
40
20
0
288 M. Schäfer-Korting et al.
meable (6.00 ± 1.17 × 10
–6
cm/s; p ≤ 0.05). Compared
to human epidermis (0.42 ± 0.39 × 10
–6
cm/s), pig
skin (0.08 ± 0.01 × 10
–6
cm/s) and bovine udder skin
(0.32 ± 0.28 × 10
-6
cm/s), the P
app
values suggested
that the barrier functions of the reconstructed tissues
were clearly less developed (p ≤ 0.05). The ANOVA
grouped skin permeability, as measured by P
app
, for
testosterone as follows: SkinEthic > EpiDerm,
EPISKIN > human epidermis, bovine udder skin,
pig skin. Total permeation within 6 hours ranged
from 0.6% of the applied testosterone when using
pig skin, to 39.5% when using the SkinEthic model.
The grouping of the tissues was the same as that of
the P
app
values except for an additional significant
a = EpiDerm; b = human epidermis sheets; c = EPISKIN; d = pig skin; e = SkinEthic; f = bovine udder skin.
Permeation was measured following the application of 500µL or 110µl (EPISKIN) of 0.004% testosterone in PBS for
6–8 hours to reconstructed epidermis (a, c, e), or for 26–36 hours to human epidermis sheets or split skin (b, d, f). Each
line shows the results obtained within a single laboratory.
= Across Barriers GmbH; = Freie Universität Berlin; = Ludwig-Maximilians-Universität München;
= Stiftung Tierärztliche Hochschule Hannover; = Universität des Saarlandes.
Figure 2: Testosterone permeation into phosphate buffered saline in different laboratories
0
time (hours)
permeation (µg/cm
2
)
5
f)
10 15 20 25 30
6
5
4
3
2
1
0
0
time (hours)
permeation (µg/cm
2
)
5
d)
10 15 20 25 30
6
5
4
3
2
1
0
0
time (hours)
permeation (µg/cm
2
)
2
e)
4 6 8 10 12
6
5
4
3
2
1
0
0
time (hours)
permeation (µg/cm
2
)
2
a)
4 6 8 10 12
6
5
4
3
2
1
0
0
time (hours)
permeation (µg/cm
2
)
2
c)
4 6 8 10 12
6
5
4
3
2
1
0
0
time (hours)
permeation (µg/cm
2
)
5
b)
10 15 20 25 30
6
5
4
3
2
1
0
Reconstructed human epidermis for skin absorption testing 289
difference for EpiDerm and EPISKIN (Table 2;
p ≤ 0.05).
Caffeine permeation
The EpiDerm model appeared to be less permeable
(P
app
: 0.24 ± 0.14 × 10
–6
cm/s; p ≤ 0.05) with respect
to caffeine, compared to the EPISKIN and
SkinEthic models (P
app
: 2.77 ± 0.78 × 10
–6
cm/s and
3.63 ± 1.91 × 10
–6
cm/s, respectively). In fact,
according to the ANOVA, the P
app
value of caffeine
permeating EpiDerm was in the range of the P
app
values of human epidermis (0.06 ± 0.04 ×
10
–6
cm/s), pig skin (0.07 ± 0.05 × 10
–6
cm/s), and
bovine udder skin (0.63 ± 0.23 × 10
–6
cm/s). The
following rank order of the preparations was
found: SkinEthic, EPISKIN > bovine udder skin,
EpiDerm, pig skin, human epidermis. The order
changed slightly (pig skin and human epidermis)
when comparing the amounts of permeated caffeine
at 6 hours, which was 0.17–26% of the amount
applied — the extreme values in this case were
obtained with human skin and SkinEthic. ANOVA
again discriminated between SkinEthic and
EPISKIN.
Lag times
Another difference between RHE and human epi-
dermis sheets or animal skin is the delay of perme-
ation (Table 2). The lag time of both test substances
was low to absent, depending on the RHE model
used. For example, with caffeine the lag time for
EPISKIN was approximately 1 hour, and that for
EpiDerm less than 0.5 hours, while there was no lag
time with the SkinEthic model. In contrast, when
testing caffeine, a mean lag time of 4 hours was
observed with pig skin, and enormous fluctuations
of lag time were found with human epidermis,
sometimes making test periods of 24 hours or more
necessary. The lag times found when using bovine
udder skin were in between those of the RHE and
human epidermis or pig skin.
Inter-laboratory variability
Importantly, while the ANOVA revealed clear dif-
ferences between the test skins, inter-laboratory
variability (s
L
%) between the five laboratories was
low. Even when studying human epidermis sheets,
the s
L
% of the P
app
values varied by 9.3 for caffeine
and was almost zero when testosterone was tested
(inhomogenities excluded a precise quantification).
Thus, the protocol was successfully transferred
between the partner laboratories.
Discussion
During the last decade, major progress was made in
the replacement of animal experiments, including
the approved in vitro approaches for regulatory
phototoxicity (38) and skin corrosivity testing (24,
39). In addition, non-animal tests for skin irritation
(40–42) and skin sensitisation (43) are being devel-
Table 2: Summary of permeation data
P
app
(10
–6
cm/s)
Permeation (
µµ
g/cm
2
) Lag time (hours)
Skin type mean ± sd mean ± sd CV (%) mean ± sd n
Caffeine
Human epidermis sheets 1.12 ± 1.18 0.06 ± 0.04 62.29 1.73 ± 1.48 8
Pig skin 0.48 ± 0.41 0.07 ± 0.05 74.82 3.92 ± 0.87 6
Bovine udder skin 8.24 ± 3.86 0.63 ± 0.23 37.23 1.88 ± 0.42 7
EpiDerm 4.87 ± 2.67 0.24 ± 0.14 55.59 0.33 ± 0.06 5
SkinEthic 73.65 ± 36.58 3.63 ± 1.91 52.74 0.14 ± 0.05 6
EPISKIN 51.25 ± 9.84 2.77 ± 0.78 24.37 1.04 ± 0.26 7
Testosterone
Human epidermis sheets 0.32 ± 0.27 0.42 ± 0.39 93.18 1.03 ± 2.52 8
Pig skin 0.07 ± 0.15 0.08 ± 0.01 14.91 –0.13 ± 11.92 4
Bovine udder skin 0.14 ± 0.15 0.32 ± 0.28 89.89 1.19 ± 1.30 6
EpiDerm 2.36 ± 0.90 2.89 ± 1.09 37.82 0.00 ± 0.09 5
SkinEthic 4.47 ± 0.57 6.00 ± 1.17 19.55 0.14 ± 0.09 5
EPISKIN 1.53 ± 0.47 2.11 ± 0.63 29.89 0.93 ± 0.33 7
P
app
(apparent permeation coefficient) values, lag time and drug permeated into the acceptor medium after 6 hours for
caffeine (0.1%) and testosterone (0.004%) applied to human epidermis sheets, reconstructed epidermis and pig and
bovine udder skin. CV = coefficient of variation; sd = standard deviation; n = number of independent experiments.
290 M. Schäfer-Korting et al.
oped, but their final validation and regulatory
acceptance have not yet been achieved. These in
vitro tests are either based on monolayer cultures of
skin cells (phototoxicity, sensitisation) or make use
of RHE (corrosivity, irritation) or human skin
explant cultures (44). RHE is often favoured for
skin toxicity testing, because of its close resem-
blance to human skin (for a review, see 42). The
cutaneous uptake of a chemical is the first step in
the induction of skin damage in contact irritant and
allergic dermatitis, as well as for systemic toxicity.
Therefore, a validated in vitro approach involving
the use of RHE for skin uptake testing would be
welcome. The method should permit the determi-
nation of the amounts penetrating (e.g. for sensiti-
sation) and permeating (e.g. for systemic toxicity)
the skin. In safety assessment, the sum of the
amounts of a substance permeating and penetrat-
ing the skin is used to determine absorption. If no
data are available, the worst-case-scenario absorp-
tion of 100% is normally used. Approval of methods
by the OECD, in this case the use of human and ani-
mal skin in diffusion cells in skin absorption studies
(6, 7), results in an almost world-wide acceptance of
the data generated according to the appropriate
OECD TGs.
There are, however, two major drawbacks. First
of all, the enormous number of tests expected as a
result of the REACH initiative will result in a
shortage of human skin for in vitro experiments.
Testing might then be delayed, or animal skin will
have to be substituted for human skin. Moreover,
the rather generally defined test procedure laid
down in OECD TGD 28 (7) can result in a large
variability in the data generated when relying only
on this document. In fact, the results of a recent
ring trial, in which the OECD reference sub-
stances, benzoic acid, caffeine, and testosterone,
were tested, varied widely between the ten labora-
tories of the EDETOX project. The published data
(45) suggested a maximum flux variation of 50%
with caffeine and 109% with testosterone, between
the nine EDETOX laboratories using human skin
as a test matrix. These differences may result from
the variations in the thickness of the human skin
obtained from surgical intervention and from
human cadavers, which ranged from 300 to
1800µm. Additionally, the Franz cell surface areas
employed varied by ten-fold and the Franz cell vol-
umes by even more (45). In order to be able to com-
pare the results of permeation experiments
performed in one laboratory with those of another
laboratory, a higher degree of standardisation of
testing procedures needs to be achieved than would
follow from the use of the OECD TGD alone.
In this study, five laboratories used human and
animal skin (generally originating from different
donors), in addition to different batches of various
RHE models. The highly standardised test proce-
dure clearly reduced the variability of the perme-
ation profile. This applied to both the medium and
highly lipophilic agents (caffeine, Figure 1; testos-
terone, Figure 2), as well as to the permeation
parameters, P
app
value, permeated amount and lag
time (Table 2). When studying human epidermis
sheets, the inter-laboratory variability (s
L
%) of the
P
app
values was only 9.3 for caffeine and almost
zero for testosterone, although inhomogeneities did
not allow precise quantification. Taking into
account the fluctuation seen when using human
skin with a thickness of 1000µm (34), heat isolation
of the epidermis does not increase variability to a
relevant extent. Therefore, any variability is proba-
bly caused by the individual skin structure of the
donors. The variability observed in this study was
in the upper range of scatter found in the perme-
ation of 18 compounds (mostly beta-blockers) tested
on human cadaver epidermis (variation coefficients
27–92%) in addition to a rat keratinocyte model
(6–83%; 46). However, the data from the latter
model were generated only in a single laboratory.
Recent studies comparing the permeability of
chemicals tested on reconstructed epidermis of
human and rat origin with that of human epidermal
sheets (32, 34, 46), human skin (18, 27, 30, 32, 34),
pig skin (18, 19, 32, 34) or rat skin (32), have already
indicated a higher degree of permeability for the
reconstructed tissues. The detailed study protocol
adhered to by the five independent laboratories dur-
ing testing in our study allows a more precise analy-
sis of the differences, as well as a comparison of the
three commercially available RHE models. The P
app
values for caffeine and testosterone permeation
exceeded those calculated for human epidermis
sheets by 4.0-fold and 6.9-fold with EpiDerm, 46.0-
fold and 5.0-fold with EPISKIN, and 60.6-fold and
14.3-fold with the SkinEthic model. Thus,
hydrophilic agents seem to permeate RHE even bet-
ter than lipophilic ones, and the lack of hair follicles
that facilitate permeation of hydrophilic compounds
is not relevant for caffeine. Whether, this holds true
for more hydrophilic substances, e.g. mannitol, will
be evaluated during the validation study. Suhonen
and coworkers (46) reported that, on average, there
was a two-fold to three-fold enhancement of the per-
meability coefficients of 18 test compounds, ranging
from 0.3 (hydrophilic compounds) to 5.2 (lipophilic
compounds), in the rat epidermis model compared to
HES. In particular with testosterone, which was
tested in both studies, the permeability of recon-
structed rat epidermis (46) and the RHE models,
EpiDerm and EPISKIN, exceeded HES permeabil-
ity by about 5-fold. Low lag times were also seen
with reconstructed epidermis built up from a rat
keratinocyte cell line (46), while long and variable
lag times of several hours had already been
reported in previous studies when using HES (11,
12).
The higher permeation and thus overestimation
of skin absorption when using reconstructed epi-
Reconstructed human epidermis for skin absorption testing 291
dermis models is in accordance with the incomplete
barrier found in these models. The deficiencies in
the barrier are caused by lower concentrations of
free fatty acids and hydrophilic ceramide fractions,
and also by the expression of cytokines and growth
factors leading to hyper-proliferation of the epider-
mal cells (23). In our study, bovine udder skin also
tended to be oversensitive for predicting the skin
absorption of chemicals in comparison with human
epidermis (P
app
ratio: 10.5 for caffeine and 0.8 for
testosterone). Pig skin (P
app
ratio: 0.2 for testos-
terone and 1.2 for caffeine) tended to be less per-
meable. A higher uptake by bovine skin as
compared to pig skin was also observed when study-
ing the veterinary drug, abamectin (47). To permit
correlation of the permeability of RHE to the per-
meability of human skin, a prediction model will be
developed after finishing the next step of the formal
validation process, which is described below.
In the limited number of experiments conducted in
this study, a tendency for lower variation of perme-
ation data was observed when using RHE. Should
this also be demonstrated in the validation phase, a
reduction in the number of individual experiments
necessary to be performed may be possible. From a
practical point of view, the constant and short lag
times found in permeation studies when using RHE
(Table 2) also have various advantages. Uptake
experiments can be performed within 6–8 hours, the
sampling of receptor medium needed for valid exper-
iments can be kept to a minimum, and, more impor-
tantly, a smaller number of batches are needed, in
comparison with the numbers of donors needed for
studies on human and animal skin. This will also
facilitate the evaluation of the influences of donor
vehicles, which is relevant for pesticides (48), actives
of cosmetics, and drugs in human medicine (29–31)
and veterinary medicine (47). In addition, studies on
interferences from the environment (14, 16, 17), the
pre-treatment of skin, and the influence of dermal
metabolism (13, 29, 49), which have only recently
become a matter of concern, can also be simplified.
The important effects of these factors have been
reported; for example, the pre-treatment of skin with
various ingredients of sunscreens and also N,N-
diethyl-m-toluamide (DEET) has been found to
increase the absorption of 2,4-dichlorophenoxy-
acetic acid. This is relevant when agricultural work-
ers use sunscreens and/or repellents, which may
increase the risk of pesticide uptake (14).
OECD TG 428 (6) calls for integrity testing before
permeation experiments are performed, and the
TGD 28 (7) recommends the measurement of
transepidermal water loss (TEWL) or transepider-
mal electrical resistance (TER). Alternatively, triti-
ated water can be used as a permeation marker.
When agreeing on the test protocol, however, the
study group decided against the use of these meth-
ods. This decision was made because of the poor pre-
dictability of TEWL measurements for skin barrier
integrity, as was recently observed by one partner
laboratory when comparing intact, stripped and nee-
dle-punctured human epidermis sheets. TEWL
measurements detected only very severe damage to
the stratum corneum (35). Furthermore, the appli-
cation of solutions prior to each experiment limited
the use of the TER, since a complete removal of the
solution, especially of lipophilic preparations, could
not be ensured. Finally, the use of tritiated water
was not regarded as suitable, since its application for
5 hours prior to each experiment might affect the
quality of the skin samples (50). Moreover, the ³H-
label might result in an overestimation of the quan-
tity of permeated test compound, due to radioactive
contamination. Furthermore, the study group aimed
to define and characterise methods that could easily
be established by each laboratory. Based on these
considerations, the study group decided to check for
integrity by visual inspection only.
According to the ECVAM principles for validation
(1), the acceptance of a non-animal test does not
only require transferability and acceptable variabil-
ity both within and between laboratories, but also
relevance based on predictive capacity, as deter-
mined experimentally, and on mechanistic rele-
vance. The mechanistic principles for the uptake
into and through human skin and RHE are quite
similar. In both systems, the stratum corneum is
the major barrier to the uptake of chemicals. In con-
trast to the skin of furry animals, in human skin
the shunt pathway of absorption via the hair folli-
cles is of little relevance. Since the prevalidation
experiments reported here were run in five labora-
tories, each of them performing an overlapping sub-
set of experiments, the results are in line with the
ECVAM principles for demonstrating the predictive
capacity of new testing procedures. The validation
study will further improve the predictive capacity of
the protocol, since the number of participating lab-
oratories will be greater and the set of test chemi-
cals will be expanded to cover a broader spectrum of
physicochemical parameters. Substances with a
molecular weight over 500 Daltons will also be
included, since the skin penetration of these sub-
stances may be very low (51). Moreover, a finite-
dose approach which reflects accidental intoxication
will be included, to comply with the needs of regu-
latory toxicology.
At the time when the project was designed, all the
known commercially available RHE models were
included. However, the situation, changed, when in
2002 to 2004, new commercially available reconsti-
tuted skin models were introduced into the market
(for example: the full-thickness model AST-2000,
CellSystems, St. Katharinen, Germany; the full-
thickness models EpiDerm FT, Mattek, Ashland,
MA, USA; SkinEthic RFT, Laboratoire Skinethic,
Nice, France; and Phenion FT, Phenion, Frankfurt,
Germany; as well as the epidermal model, EST-
1000, CellSystems). In addition to the commercially
292 M. Schäfer-Korting et al.
available organotypic models, there are several in-
house models, including a rat keratinocyte cell line
(46), which may also be suitable for uptake studies.
According to the principles of the modular valida-
tion approach, new methods may be evaluated
experimentally, without conducting a full valida-
tion study (1). In this case, all the skin models avail-
able for testing may be used for skin absorption
testing in the near future, without any restrictions.
Acknowledgements
Financial support of the German Ministry of
Education and Research (0312881-0312886) is
gratefully acknowledged.
Received 07.11.05; received in final form 20.03.06;
accepted for publication 05.04.06.
References
1. Hartung, T., Bremer, S., Casati, S., Coecke, S.,
Corvi, R., Fortaner, S., Gribaldo, L., Halder, M.,
Hoffmann, S., Janusch Roi, A., Prieto, P., Sabbioni,
E., Scott, L., Worth, A. & Zuang, V. (2004). A modu-
lar approach to the ECVAM principles on test valid-
ity. ATLA 32, 467–472.
2. OECD (2004). Principles for the Validation, for
Regulatory Purposes, of (Quantitative) Structure-
Activity Relationship Models. Website http://www.
oecd.org/document/23/0,2340,en_2649_34373_33957
015_1_1_1_1,00.html (Accessed 19.4.06).
3. Liebsch, M. & Spielmann, H. (2002). Currently avail-
able in vitro methods used in regulatory toxicology.
Toxicology Letters 127, 127–134.
4. Spielmann, H. (2003). Validation and regulatory
acceptance of new carcinogenicity tests. Toxicologic
Pathology 31, 54–59.
5. OECD (2004). OECD Guidelines for the Testing of
Chemicals, No. 427: Skin absorption: in vivo method,
8pp. Paris, France: OECD.
6. OECD (2004). OECD Guidelines for the Testing of
Chemicals, No. 428: Skin absorption: in vitro method,
8pp. Paris, France: OECD.
7. OECD (2003). Guidance Document for the Conduct
of Skin Absorption Studies. Series on testing assess-
ment No. 28. Paris, France: OECD.
8. Lundehn, J. (1992). Uniform principles for safeguard-
ing the health of applicators of plant protection prod-
ucts (uniform principles for operator protection).
Mitteilungen aus der Biologischen Bundesanstalt 277,
Berlin, Germany.
9. van Ravenzwaay, B. & Leibold, E. (2004). The signifi-
cance of in vitro rat skin absorption studies to human
risk assessment. Toxicology in Vitro 18, 219–225.
10. Onuki, M., Yokoyama, K., Kimura, K., Sato, H.,
Nordin, R.B., Naing, L., Morita, Y., Sakai, T., Kobay-
ashi, Y. & Araki, S. (2003). Assessment of urinary coti-
nine as a marker of nicotine absorption from tobacco
leaves: a study on tobacco farmers in Malaysia.
Journal of Occupational Health 45, 140–145.
11. Ramsey, J.D., Woollen, B.H., Auton, T.R. & Scott, R.C.
(1994). The predictive accuracy of in vitro measure-
ments for the dermal absorption of a lipophilic pene-
trant (fluazifop-butyl) through rat and human skin.
Fundamental and Applied Toxicology 23, 230–236.
12. Frantz, S.W., Ballantyne, B., Beskitt, J.L., Tallant,
M.J. & Greco, R.J. (1995). Pharmacokinetics of 2-
ethyl-1,3-hexanediol. III. In vitro skin penetration
comparisons using the excised skin of humans, rats
and rabbits. Fundamental and Applied Toxicology 28,
1–8.
13. Hewitt, P.G., Perkins, J. & Hotchkiss, S.A. (2000).
Metabolism of fluroxypyr, fluroxypyr methyl ester,
and the herbicide fluroxypyr methylheptyl ester. I:
during percutaneous absorption through fresh rat
and human skin in vitro. Drug Metabolism and
Disposition 28, 748–754.
14. Pont, A.R., Charron, A.R. & Brand, R.M. (2004).
Active ingredients in sunscreens act as topical
penetration enhancers for the herbicide 2,4-
dichlorophenoxyacetic acid. Toxicology and Applied
Pharmacology 195, 348–354.
15. Scott, R.C., Batten, P.L., Clowes, H.M., Jones, B.K.
& Ramsey, J.D. (1992). Further validation of an in
vitro method to reduce the need for in vivo studies
for measuring the absorption of chemicals through
rat skin. Fundamental and Applied Toxicology 19,
484–492.
16. Qiao, G.L. & Riviere, J.E. (2002). Systemic uptake
and cutaneous disposition of pentachlorophenol in a
sequential exposure scenario: effects of skin preex-
posure to benzo[a]pyrene. Journal of Toxicology and
Environmental Health: Part A 65, 1307–1331.
17. Riviere, J.E., Baynes, R.E., Brooks, J.D., Yeatts, J.L.
& Monteiro-Riviere, N.A. (2003). Percutaneous
absorption of topical N,N-diethyl-m-toluamide
(DEET): effects of exposure variables and coadmin-
istered toxicants. Journal of Toxicology and Envir-
onmental Health: Part A 66, 133–151.
18. Mahmoud, A., Haberland, A., Dürrfeld, M., Heydeck,
D., Wagner, S. & Schäfer-Korting, M. (2005).
Cutaneous estradiol permeation, penetration and
metabolism in pig and man. Skin Pharmacology and
Physiology 18, 27–35.
19. Wagner, S.M., Nogueira, A.C., Paul, M., Heydeck,
D., Klug, S. & Christ, B. (2003). The isolated nor-
mothermic hemoperfused porcine forelimb as a test
system for transdermal absorption studies. Journal
of Artificial Organs 6, 183–191.
20. Kietzmann, M., Loscher, W., Arens, D., Maass, P. &
Lubach, D. (1993). The isolated perfused bovine
udder as an in vitro model of percutaneous drug
absorption. Skin viability and percutaneous absorp-
tion of dexamethasome, benzoyl peroxide, and
etofenamate. Journal of Pharmacological and Tox-
icological Methods 30, 75–84.
21. Pitterman, W., Kietzmann, M. & Jackwerth, B.
(1995). The isolated perfused bovine udder skin.
ALTEX 12, 196–200.
22. Netzlaff, F., Lehr, C.M., Wertz, P.W. & Schaefer,
U.F. (2005). The human epidermis models EpiSkin,
SkinEthic and EpiDerm: an evaluation of morphol-
ogy and their suitability for testing phototoxicity,
irritancy, corrosivity, and substance transport.
European Journal of Pharmaceutics and Biopharm-
aceutics 60, 167–178.
23. Ponec, M., Boelsma, E., Gibbs, S. & Mommaas, M.
(2002). Characterization of reconstructed skin mod-
els. Skin Pharmacology and Applied Skin Phys-
iology 15, Suppl. 1, 4–17.
24. Liebsch, M., Traue, D., Barrabas, C., Spielmann, H.,
Reconstructed human epidermis for skin absorption testing 293
Uphill, P., Wilkins, S., McPherson, J.P., Wiemann, C.,
Kaufmann, T., Remmele, M. & Holzhütter, H-G.
(2000). The ECVAM prevalidation study on the use of
EpiDerm for skin corrosivity testing. ATLA 28,
371–401.
25. Spielmann, H., Muller, L., Averback, D., Balls, M.,
Brendler-Schwaab, S., Castell, J.V., Curren, R., de
Silva, O., Gibbs, N.K., Liebsch, M., Lovell, W.W.,
Merk, H.F., Nash, J.F., Neumann, N.J., Pape,
W.J.W., Ulrich, P. & Vohr, H-W. (2000). The second
ECVAM workshop on phototoxicity testing. The
report and recommendations of ECVAM workshop
42. ATLA 28, 777–814.
26. Lotte, C., Patouillet, C., Zanini, M., Messager, A. &
Roguet, R. (2002). Permeation and skin absorption:
reproducibility of various industrial reconstructed
human skin models. Skin Pharmacology and App-
lied Skin Physiology 15, Suppl. 1, 18–30.
27. Dreher, F., Fouchard, F., Patouillet, C., Andrian, M.,
Simonet, J.T. & Benech-Kieffer, F. (2002). Com-
parison of cutaneous bioavailability of cosmetic
preparations containing caffeine or alpha-tocopherol
applied on human skin models or human skin ex vivo
at finite doses. Skin Pharmacology and Applied Skin
Physiology 12, Suppl. 1, 40–58.
28. Dreher, F., Patouillet, C., Fouchard, F., Zanini, M.,
Messager, A., Rouget, R., Cottin, M., Leclaire, J. &
Beech-Kieffer, F. (2002). Improvement of the experi-
mental setup to assess cutaneous bioavailability on
human skin models: dynamic protocol. Skin Pharm-
acology and Applied Skin Physiology 15, Suppl. 1,
31–39.
29. Gysler, A., Kleuser, B., Sippl, W., Lange, K., Korting,
H.C., Höltje, H.D. & Schäfer-Korting, M. (1999). Skin
penetration and metabolism of topical glucocorticoids
in reconstructed epidermis and in excised human
skin. Pharmacological Research 16, 1386–1391.
30. Santos Maia, C., Mehnert, W., Schaller, M., Korting,
H., Gysler, A., Haberland, A. & Schäfer-Korting, M.
(2002). Drug targeting by solid lipid nanoparticles for
dermal use. Journal of Drug Targeting 10, 489–495.
31. Sivaramakrishnan, R., Nakamura, C., Mehnert, W.,
Korting, H.C., Kramer, K.D. & Schäfer-Korting, M.
(2004). Glucocorticoid entrapment into lipid carriers
— characterisation by parelectric spectroscopy and
influence on dermal uptake. Journal of Controlled
Release 97, 493–502.
32. Schmook, F.P., Meingassner, J.G. & Billich, A.
(2001). Comparison of human skin or epidermis
models with human and animal skin in in vitro per-
cutaneous absorption. International Journal of
Pharmaceutics 215, 51–56.
33. Haberland, A., Schreiber, S., Maia, C.S., Rübbelke,
M.K., Schaller, M., Korting, H.C., Kleuser, B.,
Schimke, I. & Schäfer-Korting, M. (2006). The
impact of skin viability on drug metabolism and per-
meation-BSA toxicity on primary keratinocytes.
Toxicology in Vitro 20, 347–354.
34. Schreiber, S., Mahmoud, A., Vuia, A., Rübbelke, M.K.,
Schmidt, E., Schaller, M., Kandarova, H., Haberland,
A., Schaefer, U.F., Bock, U., Korting, H.C., Liebsch, M.
& Schäfer-Korting, M. (2005). Reconstructed epider-
mis versus human and animal skin in skin absorption
studies. Toxicology in Vitro 19, 813–822.
35. Netzlaff, F., Kostka, K.H., Lehr, C.M. & Schaefer,
U.F. (2006). TEWL measurements as a routine
method for evaluating the integrity of epidermis
sheets in static Franz type diffusion cells in vitro.
Limitations shown by transport data testing.
European Journal of Pharmaceutics and Biopharm-
aceutics 63, 44–50.
36. Kligman, A.M. & Christophers, E. (1963). Prep-
aration of isolated sheets of human stratum
corneum. Archives of Dermatology 88, 702–705.
37. DIN (1991). DIN ISO 5725-2: Genauigkeit (Rich-
tigkeit und Präzision) von der Ermittlung der
Wiederhole und Vergleichspräzision von festgelegten
Messverfahren. Entwurf. Berlin, Germany: Beuth
Verlag GmbH.
38. Spielmann, H., Liebsch, M., Döring, B. & Molden-
hauer, F. (1994). First results of an EC/COLIPA val-
idation project of in vitro phototoxicity testing
methods. ALTEX 11, 22–31.
39. Fentem, J.H. (1999). Validation of in vitro tests for
skin corrosivity. ALTEX 16, 150–153.
40. Fentem, J.H. & Botham, P.A. (2002). ECVAM’s
activities in validating alternative tests for skin cor-
rosion and irritation. ATLA 30, Suppl. 2, 61–67.
41. Kandarova, H., Liebsch, M., Genschow, E., Gerner,
I., Traue, D., Slawik, B. & Spielmann, H. (2004).
Optimisation of the EpiDerm test protocol for the
upcoming ECVAM validation study on in vitro skin
irritation tests. ALTEX 21, 107–114.
42. Welss, T., Basketter, D.A. & Schröder, K.R. (2004).
In vitro skin irritation: facts and future. State of the
art review of mechanisms and models. Toxicology in
Vitro 18, 231–243.
43. Aeby, P., Wyss, C., Beck, H., Griem, P., Scheffler, H. &
Goebel, C. (2004). Characterization of the sensitizing
potential of chemicals by in vitro analysis of dendritic
cell activation and skin penetration. Journal of
Investigative Dermatology 122, 1154–1164.
44. Jacobs, J.J., Lehe, C.L., Cammans, K.D., Das, P.K. &
Elliott, G.R. (2004). Assessment of contact allergens
by dissociation of irritant and sensitizing properties.
Toxicology in Vitro 18, 681–690.
45. van der Sandt, J.J.M., Maas, W.J.M., van Burgsteden,
J.A., Sartorelli, P., Montomoli, L., Larese, F., Payan, J-
P., Limaset, J.C., Carmichael, P., Kenyon, S., Robin-
son, E., Dick, I., Nilsen, J., Schaller, K-H., Korinth, G.,
Cage, S., Wilkinson, S.C. & Williams Faith, M. (2004).
In vitro predictions of skin absorption of caffeine,
testosterone and benzoic acid: a multi-centre compar-
ison study. Regulatory Toxicology and Pharmacology
39, 271–281.
46. Suhonen, T., Pasonen-Seppänen, S., Kirjavainen, M.,
Tammi, M., Tammi, R. & Urtti, A. (2003). Epidermal
cell culture model derived from rat keratinocytes with
permeability characteristics comparable to human
cadaver skin. European Journal of Pharmaceutical
Sciences 20, 107–113.
47. Baynes, R.E. (2004). In vitro dermal disposition of
abamectin (avermectin B[1]) in livestock. Research in
Veterinary Science 76, 235–242.
48. Navidi, W.C. & Bunge, A.L. (2002). Uncertainty in
measurements of dermal absorption of pesticides. Risk
Analysis 22, 1175–1182.
49. Gysler, A., Lange, K., Korting, H.C. & Schäfer-Kor-
ting, M. (1997). Prednicarbate biotransformation in
human foreskin keratinocytes and fibroblasts. Pharm-
acological Research Communications 14, 793–797.
50. COLIPA (1995). Cosmetic Ingredients: Guidelines for
Percutaneous Absorption/penetration. 6pp. Brussels,
Belgium: COLIPA.
51. Bos, J.D. & Meinardi, M.M. (2000). The 500 Dalton
rule for the skin penetration of chemical compounds
and drugs. Experimental Dermatology 9, 165–169.
294 M. Schäfer-Korting et al.