`
`&
`ELSEVIE
`
`Available online at www.sciencedirect.com
`
`science @)onecr:
`
`Regulatory Toxicology and Pharmacology 39 (2004) 271-281
`
`Regulatory
`
`Toxicology and
`Pharmacology
`
`www.elsevier.com/locate/yrtph
`
`In vitro predictions of skin absorption of caffeine, testosterone,
`and benzoic acid: a multi-centre comparison study
`
`J.J.M. van de Sandt,** J.A. van Burgsteden,’ S. Cage,' P.L. Carmichael,”! I. Dick,!
`S. Kenyon,* G. Korinth," F. Larese,° J.C. Limasset,? W.J.M. Maas,* L. Montomoli,”
`J.B. Nielsen,® J.-P. Payan,‘ E. Robinson,’ P.. Sartorelli,” K.H. Schaller,"
`S.C. Wilkinson; and F.M. Williams!
`
`“TNO Nutrition and Food Research, Zeist, The Netherlanels
`® Istinuto di Medicina del Lavoro, Siena, Italy
`“ Universita di Trieste, Italy
`" Institut National de Recherche et de Sécurité, Vandoeuvre Cedex, France
`* Biological Chemistry, Faculty of Medicine, Imperial College London, London, UK
`" Health and Safety Laboratory, Sheffield, UK
`® University of Southern Denmark, Odense, Denmark
`X University of Erlangen-Nuremberg, Erlangen, Germany
`' Huntingdon Life Science Ltd, Eye, UK
`i) The Medical School, University of Newcastle, Newcastle upon Tyne, UK
`Received 18 November 2003
`Available online 22 April 2004
`
`Abstract
`
`To obtain better insight into the robustness ofin vitro percutaneous absorption methodology, the intra- and inter-laboratory
`variation in this type of study was investigated in 10 European laboratories. To this purpose, the in vitro absorption of three
`compounds through humanskin (9 laboratories) and rat skin (1 laboratory) was determined. The test materials were benzoic acid,
`caffeine, and testosterone, representing a range of different physico-chemical properties. All laboratories performed their studies
`according to a detailed protocol in which all experimental details were described and each laboratory performed at least three
`independent experiments for each test chemical. All laboratories assigned the absorption of benzoic acid through human skin, the
`highest ranking of the three compounds(overall mean flux of 16.54 +4 11.87 pg/cm?/h). The absorption ofcaffeine and testosterone
`
`through humanskin wassimilar, having overall mean maximumabsorptionrates of 2.24 + 1.43 ug/em?/h and 1.63 + 1.94 g/em*/h,
`respectively. In 7 out of 9 laboratories, the maximum absorption rates of caffeine were ranked higher than testosterone. No dif-
`ferences were observed between the mean absorption through human skin and the onerat study for benzoic acid and testosterone.
`For caffeine the maximum absorption rate and the total penetration through rat skin were clearly higher than the mean value for
`human skin. When evaluating all data, it appeared that no consistent relation existed between the diffusion cell type and the ab-
`sorption of the test compounds. Skin thickness only slightly influenced the absorption of benzoic acid and caffeine. In contrast, the
`maximum absorption rate of testosterone wasclearly higher in the laboratories using thin, dermatomed skin membranes. Testos-
`terone is the most lipophilic compound and showed also a higher presence in the skin membrane after 24h than the two other
`compounds. The results of this study indicate that the in vitro methodology for assessing skin absorption is relatively robust. A
`major effort was made to standardize the study performance, but, unlike in a formal validation study, not all variables were
`controlled. The variation observed may be largely attributed to human variability in dermal absorption and the skin source. Forthe
`most lipophilic compound, testosterone, skin thickness proved to be a critical variable.
`© 2004 Elsevier Inc. All rights reserved.
`
`
`* Corresponding author. Fax; +31-30-6960264.
`E-mail adéress: vandesandt@yoeding.tno.nl (J.J.M. van de Sandt),
`' Present address: Unilever Colworth, Sharnbrook, UK
`’
`0273-2300/$ - see front matter © 2004 Elsevier Inc. All rights reserved.
`doi:10.1016/j.yrtph.2004.02.004
`0001
`
`Noven Pharmaceuticals, Inc.
`EX2023
`Mylan Tech., Inc. v. Noven Pharma., Inc.
`IPR2018-00173
`
`

`

`272
`
`JM. van de Sandt et al. | Regulatory Toxicology and Pharmacology 39 (2004) 271-281
`
`1. Introduction
`
`Reproducible data on percutaneous absorption in
`humansare required to predict the systemic risk from
`dermal exposure to chemicals, such as hazardous sub-
`stances at the workplace, agrochemicals, and cosmetic
`ingredients (EC 2002; EEC 1991; SCCNFP 2003). In
`this context, there is a need for reliable in vitro models
`since the European Union advocates this approach and
`national legislation stipulates that animal experiments
`should be avoided wheneverscientifically feasible. Fur-
`thermore, owing to the difference in skin structure, an-
`imal studies do not always reflect the humansituation.
`Absorption through the skin is the primary route
`of exposure for most pesticides both occupationally
`(Benford et al. 1999) and in residential settings (Ross
`et al. 1992). Despite the often relatively high dermal
`(and inhalation) exposure in occupational settings, reg-
`ulations for pesticides and other chemical exposure have
`evolved from concern about the oral route of exposure.
`In the absenceofreliable dermal absorption data, route-
`to-route extrapolation has been used to assess dermal
`risk. It should be noted that this extrapolation is not
`always straightforward in cases when differences in
`biotransformation exist between the oral and dermal
`route, excessive first pass effects occur and/orlarge dif-
`ferences in rate of absorption exist between the various
`routes of exposure. When no information is available on
`percutaneous absorption, risk assessments may assume
`an absorption percentage of 100%, a worst case scenario
`(EC 2002). This is a very conservative approach and a
`more accurate measure of absorption would have a
`major impact on risk assessments for many chemicals in
`regulatory toxicology. The specific need for a valid
`method of assessing human dermal absorption has led
`the OECD (2000a,b,c) and EPA (1996, 1999) to produce
`guidelines for in vitro and in vivo assessment of percu-
`taneous absorption.
`A review of available data from published literature
`on in vitro dermal absorption was performed underthe
`auspices of the OECD in order to evaluate the perfor-
`mance of in vitro and in vivo percutaneous absorption
`measurements. It was concluded that evaluation of in
`vitro test methods from published literature wasdifficult
`(OECD 2000d) because studies containing direct com-
`parisons of in vitro and in vivo measurements were
`very limited. There were too many variables, such as
`different species, thickness and types of the skin, expo-
`
`sure duration, and vehicles. Also, very few multi-centre
`studies have been performed (Becket al. 1994) and these
`studies were limited in their approach (e.g., with respect
`to the number of laboratories involved). Therefore, no
`proper data on the intra- and inter-laboratory repro-
`ducibility of the in vitro methodology are available.
`The purpose of the present research was therefore to
`assess intra- and inter-laboratory variability in deter-
`mination of percutaneous penetration by in vitro
`methods on a larger scale than done previously. This
`report contains data generated by 10 independent lab-
`oratories from within the European Union, each testing
`the percutaneous absorption of three chemicals that are
`recommended by the OECDassuitable reference com-
`pounds for regulatory studies (OECD 2000c). The ex-
`perimental conditions (amount applied, exposure time,
`vehicle, receptor fluid, preparation of membranes, and
`analysis) were standardized according to a detailed
`protocol that adopted many ofthe guidelines proposed
`by the OECD.
`
`2. Materials and methods
`
`2.1. Test substances and preparation of dose solutions
`
`The test substances were chosen on the basis of their
`range in physico-chemical properties (Table 1) and their
`recommendationas reference compounds by the OECD
`(OECD 2000c). All participating laboratories used the
`same batches of test substances. Non-radiolabelled tes-
`tosterone, caffeine, and benzoic acid were purchased
`from Steraloids
`(Newport, RI, USA) and Sigma
`Chemical Company by the study coordinator and were
`then supplied to the participants.
`[4-'*C]testosterone
`(53.6mCi/mmol) and [l-methyl-'*C]caffeine (51.2 mCi/
`mmol) were purchased from Perkin-Elmer Life Sci-
`ences, while [ring-UL-'*C]benzoic acid (6.2 mCi/mmol)
`wasobtained from Sigma Chemical Company. The dose
`solutions were prepared freshly by each laboratory in
`ethanol/water (1:1, v/v), yielding a concentration of
`4.0mg/mL for each compound. Participants with a li-
`cense to handle radiochemicals prepared the dose solu-
`tions by mixing appropriate amounts of radiolabelled
`and non-radiolabelled test substances. The dose solu-
`tions were measured for exact total radioactivity prior to
`and directly after the application to the skin membranes.
`The
`radioactive
`concentration was
`approximately
`
`Table |
`Test substances
`
`Test substance
`
`Benzoie acid (benzenecarboxylic acid)
`Testosterone (4-androsten-17$-ol-3-one)
`Caffeine (3.7-dihydro-1.3,7-trimethyl-1 H-purine-2.6-dione)
`
`MW
`122.1
`288.4
`194.2
`
`log Po/w
`1.83
`3.32
`0.01
`
`CAS No.
`65-85-(0)
`58-22-(0)
`58-(8-2
`
`0002
`
`

`

`J.J.M. van de Sandt et al.
`
`| Regulatory Toxicology and Pharmacology 39 (2004) 271-281
`
`273
`
`1 MBq/mL for testosterone and caffeine and approxi-
`mately 4 MBq/mL for benzoic acid.
`
`2.4. Experimental design
`
`2.2. Preparation ofskin membranes
`
`Both human and rat skin membranes were prepared
`from frozen skin. Whole skin was cleaned of subcuta-
`neous fat and the skin was stored at approximately
`—20°C (participants 1 and 2 at approximately —70 °C)
`for a maximum period of one year. The supply and use
`of human and animal tissue was in full accordance with
`national ethical guidelines. Detailed information on the
`human skin source was recorded (Table 2). Most par-
`ticipants used human skin with a thickness between 0.7—
`1.1mm, while one participant used skin that was 0.8-
`1.8mm. Three laboratories used dermatomed skin with
`a thickness of 0.5-0.7 mm (participants | and 7) or 0.3—
`0.4mm (participant 10). The range of skin thickness
`used by the various participants allowed for the assess-
`mentofthe influence ofskin thickness on the absorption
`characteristics of the test compounds. Skin from more
`than one donor was used in each experiment and each
`experimental group consisted of 5—7 skin membranes
`form different individuals. Rat full-thickness skin was
`used by participant 5 and was collected from the back
`(clipped carefully) of four weeks old male Sprague
`Dawley rats.
`
`2.3. Diffusion cells and receptor fluid
`
`Each participant used the diffusion cell that was es-
`tablished in their laboratory (details are shown in Table
`3). For experiments with caffeine and benzoic acid, the
`receptor fluid consisted of saline (0.9%NaCl), while for
`experiments with
`testosterone,
`the
`receptor
`fluid
`consisted of saline (0.9% NaCl+5% Bovine Serum
`Albumin (BSA), adjusted to pH 7.4. For systems using
`flow-through diffusion cells, the flow of receptor fluid
`was approximately 1.5 mL/h.
`
`Table 2
`Details of source of human skin
`
`All participating laboratories performedtheir studies
`according to a detailed study protocol in which the ex-
`perimental design and parameters such as the dose of
`the test chemical, vehicle, duration of the experiment,
`preparation of the skin membranes, receptor fluid type,
`occlusion, temperature, sampling times, and number of
`replicates were defined. Skin membranes were thawed,
`mounted in the diffusion cell and the skin integrity was
`assessed by either visual assessment, permeation oftri-
`tiated water (cut-off K, > 3.5 x 10-4cm/h) or capaci-
`tance (cut-off: 55nF), depending on the participant.
`Subsequently,
`the test substances were applied at a
`concentration of 4.0mg/mL ethanol/water (1:1, v/v).
`The application volume was 25uL/em? which is con-
`sidered the minimum volume necessary to produce a
`homogeneous distribution on the skin surface. This
`represented a finite dose (100 ug/cm?), in order to mimic
`occupationally relevant situations. The exposure time
`was 24h, during which the donor compartment
`re-
`mained occluded. Aliquots of the receptor fluid were
`collected at various time points (minimally at 1, 2, 4, 8,
`and 24h post-dosing). For static cells, the original vol-
`ume of the receptor fluid was restored by adding fresh
`receptor fluid to the receptor compartmentdirectly after
`each sampling. In case of non-radiolabelled test com-
`pounds, the receptor fluid samples were stored at ap-
`proximately —20°C until analysis. At
`the end of the
`experiment, the test compound remaining at the appli-
`cation site was
`removed, using five cotton swabs
`dampened with ethanol/water (1:1, v/v), followed by one
`dry cotton swab. When a radioactive test compound was
`used, the cotton swabs, donor compartmentrinse, re-
`ceptor compartment rinse, and skin membranes [after
`digestion with 1.5M KOHin water/ethanol (1:4)] were
`analysed for presence of the test compound by f-
`counting. Each laboratory performed 3—5 independent
`experiments for each test chemical.
`
`Participant Number of—Post-mortem/ Sex and age donor Bodysite Skin thickness
`
`
`
`
`
`donors
`surgical waste
`(mm)
`
`
`
`1. University of Newcastle, UK
`2. Instituto di Medicina del Lavoro, Italy
`3. Universita di Trieste, Italy
`4, TNO Nutrition and Food Research,
`The Netherlands
`6. Imperial College London, UK
`7. Health and Safety Laboratory, UK
`8. University of Southern Denmark,
`Denmark
`9. University of Erlangen-Nuremberg,
`Germany
`
`10. Huntingdon Life Sciences, UK 0.3-0.4 5 Post-mortem Male, female (40-72 y) Abdomen, leg
`Participant No. 5 used rat skin,
`
`Surgical waste
`Post-mortem
`Post-mortem
`Surgical waste
`
`Female (20-59 y)
`Male (67-90 y)
`Male, female (67-89 y)
`Female (28-69 y)
`
`Breast
`Leg
`Abdomen
`Abdomen
`
`Surgical waste
`Surgical waste
`Surgical waste
`
`Female (29-50 y)
`Female (26-60 y)
`Female (16-68 y)
`
`Abdomen
`Abdomen
`Breast, abdomen
`
`0.5
`0.7-0.9
`0.8-1.8
`0.7
`
`0.9
`0.5-0.7
`0.7-1L.1
`
`Surgical waste
`
`Male, female (40-79 y)
`
`Breast, leg
`
`0.9
`
`
`
`
`
`
`
`0003
`
`17
`6
`7
`6
`
`3
`3
`22
`
`3
`
`
`
`

`

`274
`
`JEM. van de Sandt et al. | Regulatory Toxicology and Pharmacology 39 (2004) 271-28]
`
`Table 3
`Details of diffusion cell systems
`Diffusion
`Reference
`Participant
`Exposed skin
`Receptor
`cell type
`compartment
`area (cm*)
`0.64
`Volume: 0.25 mL:
`stirrer bar: yes
`Volume: 3.5 mL;
`stirrer bar: yes
`Volume: 15mL;
`stirrer bar: yes
`Volume: 0.2 mL;
`stirrer bar: no
`Volume: 5.15mL;
`stirrer bar: yes
`Volume: 0.4mL;
`stirrer bar: no
`Volume: 0.35 mL:
`stirrer bar: yes
`Volume: 17.7 mL:
`stirrer bar: yes
`Volume: 5.0mL;
`stirrer bar: yes
`Volume: 0.25 mL;
`stirrer bar: yes
`
`0.95
`
`3.14
`
`0.64
`
`1.76
`
`0.32
`
`2.12
`
`0.64
`
`0.64
`
`Cloweset al. (1994)
`
`Reifenrath et al. (1994)
`
`Larese Filon et al. (1999)
`
`Bronaugh and Stewart (1985)
`
`=
`
`Bronaugh and Stewart (1985)
`
`Nielsen and Nielsen (2000)
`
`Franz (1975)
`
`Cloweset al. (1994)
`
`1. University of Newcastle, UK
`
`Flow-through
`
`2. Instituto di Medicina del Lavoro, Italy
`
`Flow-through
`
`3. Universita di Trieste. Italy
`
`Stalie
`
`4. TNO Nutrition and Food Research,
`The Netherlands
`5. Institut National de Recherche et de
`Sécurité, France
`6. Imperial College London, UK
`
`Flow-through
`
`Static
`
`Flow-through
`
`7. Health and Safety Laboratory, UK
`
`Flow-through
`
`8. University of Southern Denmark,
`Denmark
`9. University of Erlangen-Nuremberg,
`Germany
`10. Huntingdon Life Sciences Ltd., UK
`
`Static
`
`Static
`
`Flow-through
`
`2.5. Analysis of non-radiolabelled test substances
`
`The analysis of non-radiolabelled test substances in
`the dose solutions and receptor fluid samples was per-
`formed centrally: benzoic acid by the Health and Safety
`Laboratory (UK), caffeine by the University of Trieste
`(Italy), and testosterone by TNO Nutrition and Food
`Research (The Netherlands). Established protocols were
`used for the HPLC-UV analysis of benzoic acid (Phe-
`nomenex column, SphereClone ODS(2), eluent:metha-
`nol:phosphate buffer
`(pH 6)
`(4:6),
`flow 1 mL/min,
`A =229nm), caffeine (Hypersil ODS column, eluent:
`methanol:water (1:3), flow 1 mL/min, 4 = 276 nm), and
`testosterone (according to Bogaards et al. 1995). The
`amount of non-radiolabelled test substance was not
`determined in the skin tissue and therefore total recov-
`ery values were not calculated.
`
`2.6. Analysis of radiolabelled test substances
`
`Radioactivity measurements were madeby individual
`participating laboratories. Radioactivity in the various
`samples (receptor
`fluid,
`skin,
`skin swabs, and cell
`washings) was determined byliquid scintillation count-
`ing. Receptor fluid samples were added directly to an
`appropriate scintillation fluid. For analysis of the skin
`membranes,an aliquot of the tissue digest (1.5 M KOH
`in 20% aqueous ethanol) was used.
`
`time course was constructed from the amount oftest
`substance in the receptor fluid and the maximum ab-
`sorption rate was determined fromthe steepest, linear
`portion of the curve. The time to maximum rate,
`the
`percentage of the dose recovered in the receptorfluid in
`24h,
`the percentage in the skin membrane, and the
`percentage total recovery (for radiolabelled studies) was
`also calculated. The data of each laboratory were pre-
`sented as mean +standard deviation, together with the
`coefficient of variation (CV). The presence of the test
`compoundin the skin membrane after washing the ap-
`plication area at 24h was expressed by the ratio between
`the percentage of the dose in skin and receptor fluid
`[total penetration (TP)] and the percentage ofthe dose in
`receptor fluid (RF).
`
`3. Results
`
`The absorption of caffeine, benzoic acid, and testos-
`terone through the skin was defined on the basis of
`maximum absorption rate, time to maximum rate, per-
`centage dose recovered in the skin membrane (at 24h
`post-dosing), and percentage dose recovered in the
`receptor
`fluid (at 24h post-dosing). The results of
`individual
`laboratory measurements
`are
`shown in
`Tables 4-6 and overviews of the mean values are given
`in Figs. 1-4.
`
`2.7. Calculation ofresults
`
`3.1. Benzoic acid
`
`The calculations were performed using a standardized
`Excel spreadsheet prepared by the study coordinator. A
`cumulative amount absorbed per unit skin area versus
`
`The mean maximum absorption rate of benzoic acid
`through human skin membranes was 16.54 + 11.87 pg/
`cm?/h, while the amountin the receptorfluid after 24h
`
`0004
`
`

`

`JM. van de Sandt et al. | Regulatory Toxicology and Pharmacology 39 (2004) 271-28]
`
`275
`
`Abd
`
`ory
`
`(asopJo%)
`
`AIaAooar
`
`
`
`(dL)wonrnauad
`
`(3980pJO%)
`
`[Pol
`
`[rol
`
`onl
`
`aN
`
`rot
`
`sol
`
`dN
`
`aL
`
`aN
`
`col
`
`ero
`
`eee
`
`STL
`
`css
`
`ss
`
`$48
`
`Sr
`
`ssol
`
`Lie
`
`fLFccg
`
`Los
`
`rL9
`
`on
`
`ch
`
`Use
`
`£96
`
`waeLP
`
`POLERD
`
`LRCFLOS
`
`OCF086
`
`Yee
`
`PLES86
`
`SolTE
`
`£'L6
`
`C66
`
`166
`
`606
`
`6r6
`
`ct6
`
`Or
`
`tor
`
`POL
`
`BIECO6
`
`“aT
`
`TFL6£96
`
`fod€
`
`ager
`
`L16
`
`LS6
`
`£06
`
`cl
`
`ese
`
`TOP
`
`SL9
`
`CSIFoes
`
`PLIFLLF
`
`PinkRE
`
`an
`
`LoL
`
`888
`
`GEL
`
`Fre
`
`dN
`
`808
`
`Vie
`
`PHFER
`
`fale
`
`Pad'LOSFOEL
`
`Sak'DOCFRLE
`
`c'B8
`
`oes
`
`ct
`
`OCFCb
`
`vad€
`
`£08
`
`OSL
`
`S18
`
`TSE
`
`YadOF
`
`GFE
`
`O38
`
`YahOE
`
`FCL
`
`SL9
`
`(4)ping
`
`soidaaay
`
`(380JO%)
`
`BEFOEo
`
`“ORL
`
`Ves
`
`6s9
`
`669
`
`haceels
`
`PEFEES
`
`ue
`
`css
`
`rLe
`
`6L8
`
`PEFROR
`
`ST
`
`68
`
`716
`
`£6
`
`‘“s
`
`998
`
`678
`
`$08
`
`oak
`
`vl
`
`ce
`
`Lt6
`
`TE
`
`(as0pJO%)
`
`Ye
`
`ceFzy
`
`Pas19
`
`re
`
`tt
`
`ee
`
`Ir9
`
`Be
`
`FFFR68
`
`L1F6F
`
`LeiFlce
`
`tal
`
`Pog
`
`CSE
`
`6c9
`
`TP
`
`aN
`
`roo
`
`Ctl
`
`eso
`
`rao
`
`RTFRRO
`
`if
`
`618
`
`F98
`
`scol
`
`FOFE06
`
`MoT
`
`PEFRLE
`
`“ae
`
`O6L
`
`CPL
`
`cos
`
`F0F6r
`
`one
`
`ts
`
`os
`
`oF
`
`aN
`
`of
`
`o's
`
`Te
`
`Le
`
`OfFIF
`
`haeEL
`
`COFRT
`
`YakCE
`
`el
`
`v1
`
`el
`
`jusjuesUlyS
`
`wadTfoFeE
`
`aPE
`
`o7auth,
`
`(q)287WInNWIeW
`aeuondiosge
`
`POFFO
`
`YakFG
`
`s0
`
`50
`
`oo
`
`se
`
`st
`
`LZ
`
`cE
`
`6T
`
`FOFOE
`
`takEF
`
`SoFLT
`
`Oe
`
`OT
`
`oc
`
`Tl
`
`coFoD
`
`aBE
`
`£0FT70
`
`eaeELT
`
`oo
`
`oo
`
`so
`
`60Fc1ol
`
`PahSL
`
`coFreu
`
`MadFE
`
`foFs0
`
`69
`
`s0
`
`s0
`
`60
`
`so
`
`e0
`
`s0
`
`dN
`
`oo
`
`0c
`
`S000
`
`(Wew/a)
`
`EEOELLIE
`
`SOO
`
`OSTE
`
`OCSE
`
`PSST
`
`POTFOOS
`
`PieeS
`
`“ae8CVOFSEE
`
`£99
`
`OL
`
`cL
`
`ERTFEPEC
`
`OPC
`
`9L1C
`
`BOT
`
`HeTEI
`
`bcar
`
`fat’
`
`oFGL
`
`eS'CZ
`
`sole
`
`EL
`
`rot
`
`OEt
`
`O'S
`
`we
`
`OFTFL99
`
`aN
`
`Lg
`
`COOFLEE
`
`PSEFOLFT
`
`wotFE
`
`6E1E
`
`OP'ST
`
`LSFe
`
`CherPoe
`
`ale
`
`és'lt
`
`60TE
`
`89LE
`
`60
`
`iv9
`
`$96
`
`SobTP
`
`to'P
`
`OL'F
`
`Slt
`
`BaeCE
`
`uwFY
`
`urwn
`
`unumFy
`
`URL
`
`uewny
`
`urwnpy
`
`ueuinyy
`
`URW}
`
`uRUMpy
`
`unum
`
`unumyy
`
`ueWny}
`
`ueuny
`
`ueum}y
`
`wy
`
`1ey
`
`1ey
`
`urwny
`
`uEWn}
`
`uEWn,
`
`Uru}
`
`URW
`
`URW}
`
`ueuinyy
`
`aN
`
`uFun}
`
`unumFy
`
`unumyy
`
`urumyy
`
`urn
`
`uRuunyy
`
`rane
`
`oom em oy
`
`oy
`
`a ~
`
`co]
`
`reer
`
`~
`
`eo
`
`GS upapy
`
`ean
`
`CsBe
`
`aS=uvapy
`
`AD
`
`GS=weapy££!
`
`AD
`
`GS=upapy
`
`AD
`
`OSFweapy£ra!
`
`AQ
`
`asFuvayy
`
`=m
`
`AD
`
`aN
`
`=—a en +
`
`dSFuvapy
`
`AD
`
`OSFueapy£cI
`
`AD
`
`aS=unayy£t!
`
`AD
`
`OSTI
`
`OSTc
`
`AQN-O1dH£
`
`OSTr
`
`OST§
`
`OST9
`
`aNL
`
`OST8
`
`AN-O1dH6
`
`OSTOr
`
`salaadg
`
`[dasJo“ON,
`‘ONTuawLadxy
`
`sisd[euyjundiiurg
`
`
`
`“ON,
`
`
`
`plorsiozuag
`
`PAGEL
`
`
`
`
`
`
`
`
`
`
`
`

`

`
`
`
`
`276
`
`aulayt>©BL
`
`
`
`
`
`oneyUtzO7out,[eMIxeyysaladsJo‘ony‘ONluauULedya,sisdjRuymedionieg
`
`
`
`
`
`A9a09a1(dL)wonenauad(44)png(asopJo%,)
`
`
`
`[royprio,sordacayquaqUOSULYS(y/m9/8n/)aes
`
`
`
`
`
`
`
`
`
`
`
`
`
`AW(q)aleswinuixeuuondiosqesaynoydas“ON
`
`{asopJO%)(asopJO%)(asopJO%)
`
`
`
`61F90orLETFICEPCFlolCIFSESOFTELFLsdS+upayy
`
`MoT209“dOPAVEEYak£PfobTGAD
`
`
`
`
`9°66VolPLeta0oc'TueumHs£F'66Foevolofotcot
`
`beamyLcailbaalELELiveoF6oF'oluEumMpy9IOSTl
`
`
`
`
`
`fFroo£EF0EFEFOOOFLTcrFeycOFRLOdSFpay
`
`SalCEitET268OEPil£e2o8EEPalOFAD
`
`L6LcsiLvlstcs£80UenIneL£sVosUslesloTroIPTueuny£+POsx01VsscLs6s'0MeO9£leLsFLelesFeOuruinHec0c'TirsLTolaIgc3'0Wem£IISTc
`
`
`
`
`
`=~rgFerr~COFRTEro=oroCSFpay
`
`ca=CatPP=PacytPaREAD
`
`aN=~Fol~LtIs'0Geum9£.=so=VeLEOueunHLcaN==vor=Lc9s"0BeUnETsIAN*O1dH£
`
`
`
`
`
`VERBCLSFFOOForeLtF96POFLEOFOFPETdSFunayy
`
`haeSabLE2HTEvalREaeOtvobSCAD
`
`£16SreeLee96cr991ueuinyé£6f6CrPSE69SeLLCwenLc81686SlsoSeCstBoCGeniLIOSTF
`
`wetSCokSsMeSTfabblCAELAD
`
`COFTTés0FE89aSFanaF66£'6soeseo£1989weyé£F'66B85clsOLoléV'9weécFil866aco$oga)olawy8lOSTs
`TOFS66OFFETS6TFLESRIFLE
`
`
`
`6sOeVolVeSf6eGeumL€
`
`
`
`S6LceostLeVc90°CeumLcrelBLClrLoc90£solgeuiny£IOST9
`
`
`
`
`
`COFOCSVLESEETEFLIC9LFRTTeTFceOSOF087GSFlnay
`
`
`
`
`
`JM. van de Sandt et al. | Regulatory Toxicology and Pharmacology 39 (2004) 271-281
`
`
`
`
`
`OCFTHESFOSLOroror9OFOTHOTcoreCSFupayy
`
`FobcePalGEMdOCeaeTTYadTEPalEEAD
`
`£96cle9cVTroSIFDean9£Lt6“sivelrlPCEFTuBuinHscsotOc6SLyor60elOFTueuMsIO81a
`
`“aOFPALE“aBEYoTE2O'0SalGEAD
`
`aeumyL£~gsc~eltreuewinHLcCIN=~Lt=Vl9STWeuMH£IAN-O1TdH6
`==cer=rcOst
`
`—waBC=fae£E2aOFAD
`
`POFOTE20FLTSOFESTCSFtvoy
`
`9000
`
`
`
`
`
`VLF869SOFLSEPOFSESLOFOTEoFoeOFOasFupayy
`
`beamLcOnlacoTEETleocoztrlueunyLIOSTg
`LoL|relLla39'0ueunyL£SOL68oscotoccel
`
`Ba£SadECwaePLBOEDSabLE"OLEAD
`
`
`
`=~ETFSOFA£0oFtTesOFee’eGS+upapy
`
`=~Yad6~cotaaMalCiAD
`
`healt¥caN=~cor~elotueumnH§IAQ-O'1dHb
`“lecl69CGeum££_=SBS7olLol
`
`

`

`
`
`
`
`AUOINSOISA|,9SEL
`
`
`
`
`
`
`
`oneyUtoO}awit,PeUUEKePysaizadgJo“ony“ON,WeuadxysIsk[ruyqwedionEg
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`[IO][RIOLl,pingsoidasayjUa}UODULYS(4)wa/art)
`
`
`
`
`(dL)wonenausd(au)(as0pJo%)(a80pJo%(asopJO%)(3850pJO%Aaan0991
`
`
`
`JIM. van de Sandt et al. | Regulatory
`
`
`
`
`
`PissearORFPRTLOFItPoFL0SCFERTdSFupapy
`
`PadOFPasOEPoeOFMaeFEadELAD
`
`£Or9oFx0sromean}9£w'L6Lrets0SepceumqL€
`
`
`
`6t'lVL6aslfF£0SOEmeumnHLIOSTI
`
`LOT916eel6'66'EtrqoUPL}£lOST,
`
`0c6SLiOLs0l69srouewny9serso3issST6ceoMITES£¥P68o0rtrostvso£0mewn},££oroLlSr666rleoPenney£zc
`
`
`
`OMFEROTF69!cCFEDSRESOCTFRY£00F£0aSFupapy
`
`
`
`
`
`SoETSebCTLOREllTPwilTEwieGEAD
`
`
`
`==“AOSoeFSweGEAD
`
`
`
`OCFOECUFFL00F£10dSFunayy
`
`ot9910ueuinH8==gs=oesv0EE8gcCIN=~61=99troDRUMLlArO1dH£
`
`PILES96OGIFGLEPOFEETMELERPE9OFFScoFos'odSFupapy
`
`CaSTTieOFVk69walEPeaeTESakSOAD
`
`66oo68SLIUtecouewnyL££98besFLgorgstroueumnyHLcFeeFs01etaOotgsOFTuewnHLlOST¥
`
`sology and Pharmacology 39 (2004) 271-281
`
`
`
`£OF£66SEFCLEEPERICBEFTTCOFLTSOFFTGS+unayy
`
`we0wae92aOE8aTE268OFSolAD
`
`Toxic
`
`166O'6EVigollotcS'Twy$£666PreoeolrT£0wey8cre686fe90Cfuloc81wy8lOST§
`
`
`
`LF£0689FT6CLTFS8oSFL0C6TFST800=cF'0aS+anayy
`
`See'DMabEEeadTTBabSEMaPLMasOTAD
`
`VesSOEL6897o'r0uewnyi£clsVeeo8stFO6f0uewnyLcPretL6StflVorreseoweumnH£tOST9
`
`
`
`
`
`=SGPFOSE:SOFTTf0'0=£6GS+upapy
`
`==SLE=aT$9'EuRWn}y§£sre$c60"ueunyrcCINttrLtLOueunH5lANPOTdH£
`
`
`
`LOEFECSVLFEL8SF7sPIFETPiFReOFOFLEOdS+upapy
`
`28OEadPOWH2609MohSE20TETAD
`
`oroLoea1tOLsooueWnyyL-9Sveel61o'sai)urWnyyL£oSvser60ortroue}LctrlssogifelerOFoLoueuinyLiIOST8—_MCL=YadSEwe'9AD
`
`-~lr;Lt970neuL£==ge=67eousuLeCIN-o's6T0uBWnyyllAQrO'1dH6
`
`
`
`
`
`
`
`==£TFES=POFRTICOF0F0dSFupapy
`
`=.=icPP—PalwetTSAD
`
`6'S6O87a9rltelMEUL£16Llfel09e0969uenmpyL€
`
`selS16osteuGEelOOFuewnH9iOsTol
`
`
`
`&TFTS6OOFETESSERSTPIFSSEOFITEVEF6ESdSFupapy
`
`
`
`
`
`
`
`277
`
`"aCBalTE266FEMadSEYGETtolOFAOD
`
`
`
`2000
`
`
`
`Awd(y)ayescunuxedayesuondiosqesayroydas‘ON,
`
`
`

`

`278
`
`JJM. van de Sandt et al. | Regulatory Toxicology and Pharmacology 39 (2004) 271-28]
`
`7
`
`2
`
`?
`
`¥
`
`ND)
`
`°
`>
`Participant no.
`
`¢
`
`2
`
`@
`
`*
`%,
`%
`
`40.0
`
`0 0 3
`
`.0
`20.0
`15.0
`10.0
`5.0
`
`00
`
`
`
`Flux(ug/em’h)
`
`Fig.
`
`|. Overview of the maximum absorption rates of benzoic acid (grey), caffeine (white) and testosterone (black). ND is not determined.
`
`mo
`90.0
`oo
`nO
`ao
`D0
`40.0
`
`0a
`
`0
`10.0
`oo
`
`%ofdose
`
`Fig. 2. In vitro skin absorption of benzoic acid, expressed as percentage of the dose present in the receptor fluid (grey) or present in the receptor
`fluid + skin membrane(total penetration—white). ND is not determined.
`
`Participant no,
`
`rs
`
`?
`
`?
`
`IND
`?
`
`”
`
`IND
`>
`o
`Participantno.
`
`IND
`2
`
`“
`
`?
`
`“,
`%
`
`7.0
`a0
`
`0 0 0 2
`
`0
`10.0
`
`a0
`
`%ofdose
`
`Fig. 3. In vitro skin absorption of caffeine, expressed as percentage of the dose present in the receptor fluid (grey) or present in the receptor
`fluid+skin membrane(total penetration—white). ND is not determined.
`
`<
`
`-
`
`ND
`-
`
`vr
`
`r
`
`IND
`*
`
`¢
`
`ND
`2
`
`“D
`
`Participant no.
`
`a
`%
`%
`
`m0
`
`a0
`
`0 2
`
`00
`x0
`
`20
`10.0
`
`no
`
`%ofdose
`
`Fig. 4. In vitro skin absorption of testosterone, expressed as percentage of the dose present in the receptor fluid (grey) or present in the receptor
`fluid + skin membrane(total penetration—white). ND is not determined.
`
`0008
`
`

`

`JIM, van de Sandt et al. | Regulatory Toxicology and Pharmacology 39 (2004) 271-281
`
`279
`
`was 70.64£17.2% of the dose applied (8 laboratories).
`The mean maximum absorption rate of benzoic acid
`through rat skin (1 laboratory) was 21.21 g/cm?/h and
`the amount in the receptor fluid after 24h was 89.8%.
`For both human and rat skin,
`the ratio TP:RF was
`approximately 1.0,
`indicating that almost no benzoic
`acid remained in the skin membrane after washing the
`application area. The total recovery of the radioactivity
`ranged between 53.6 and 98.5%(7 laboratories).
`Each laboratory performed 3—5 independent experi-
`ments. The coefficient of variation (CV) of the maxi-
`mum absorption rate varied from 6.3%(lab 4) to 52.2%
`(lab 2). For the percentage in the receptorfluid (at 24h),
`the CV values ranged between 1.6%(lab 4) and 57.1%
`(lab 2).
`
`3.2. Caffeine
`
`The mean maximum absorption rate of caffeine
`
`through human skin membranes was2.24 +1.43 p1g/cm?/
`h, while the amount in the receptor fluid after 24h was
`24.5+11.6%of the dose applied (9 laboratories). The
`mean maximum absorption rate of caffeine throughrat
`skin (1 laboratory) was 6.82 .g/cm?/h and the amountin
`the receptor fluid after 24h was 53.7%. For both human
`and rat skin, the ratio TP:RF wasonly slightly higher
`than 1.0, indicating that only a small amount caffeine
`remained in the skin membrane after washing the ap-
`plication area. The total recovery of the radioactivity
`ranged between 66.4 and 100.6%(7 laboratories).
`Each laboratory performed 3—5 independent experi-
`ments. The CV value of the maximum absorption rate
`varied from 12.0%(lab 5) to 91.4%(lab 1). For the
`percentage in the receptor fluid (at 24h), the CV values
`ranged between 5.4% (lab 5) and 66.0%(lab 1).
`
`3.3. Testosterone
`
`The mean maximum absorption rate of testosterone
`through human skin was 1.63+1.94ug/em*/h, while
`the amount
`in the receptor
`fluid after 24h was
`11.8+10.9% of the dose applied (9 laboratories). The
`mean maximum absorption
`rate of
`testosterone
`through rat skin (1 laboratory) was 1.84 ng/cm7/h and
`the amount in the receptor fluid after 24h was 21.4%.
`For both human andrat skin, the ratio TP:RF ranged
`between 1.35 and 3.54, indicating that a considerable
`amount testosterone remained in the skin membrane
`after washing the application area. The total recovery
`of the radioactivity ranged between 52.3 and 103.5%
`(7 laboratories).
`Each laboratory performed 3—5 independent experi-
`ments. The CV value of the maximum absorption rate
`ranged from 6.3%(lab 7)
`to 111.0%(lab 8). For the
`percentage in the receptor fluid (at 24h), the CV values
`ranged between 12.6%(lab 7) and 111.7%(lab 8).
`
`4. Discussion
`
`The presence of international guidelines has led to a
`partial standardization of in vitro skin absorption
`studies for regulatory purposes. On the other hand, the
`guidelines allow for certain flexibility in order to study
`compounds with widely
`differing physicochemical
`properties and undercircumstances which are the most
`relevant for its use, resulting in e.g., different exposure
`times, dose levels, and vehicle/formulations.
`In the
`OECD guidance document (OECD 2000c), useful
`in-
`formation is provided on how to properly design in vitro
`and in vivo skin absorption studies. Both static and
`flow-through diffusion cell types are considered suitable.
`In order to prevent underestimation of skin absorption,
`the test compound should be soluble in the receptor
`fluid, but the receptor fluid should notalter the barrier
`properties of the skin membrane. Skin membranes can
`be prepared in various ways, but the use of skin mem-
`branes with a thickness of more than 1.0mm (epidermis
`and dermis) is not recommended and mustbe justified
`by the researcher, since the absorption of lipophilic
`compounds may be impeded by a thick dermis. This
`guidance has been proved useful for both investigators
`in the laboratory and for regulatory agencies which
`evaluate this type of data for risk assessment purposes.
`Only very limited data exist on the intra-laboratory
`and inter-laboratory variation of in vitro skin absorp-
`tion studies. In 1994, Beck et al. reported a good cor-
`relation of in vitro absorption of hair dyes through full-
`thickness pig skin in 2 laboratories. Recently, using an
`artificial (silicone rubber) membrane, the intra-labora-
`tory and inter-laboratory variation of methyl paraben
`absorption was assessed in 18 laboratories (Chilcott
`et al. submitted). In their study, the CV values between
`laboratories were approximately 35%, while the intra-
`laboratory variation averaged 10%.
`In the study presented here, the in vitro absorption of
`three compounds through human skin (9 laboratories)
`and rat skin (1 laboratory) was investigated. The com-
`pounds (testosterone, caffeine, and benzoic acid) have a
`wide spread in their physico-chemical properties and
`have been recommended as reference compoundsbythe
`OECD (2000c). The studies were performed according
`to a very detailed protocol. Two participants were GLP-
`compliant while the other laboratories adhered to this
`quality system as muchaspossible. Analysis of samples
`from studies using non-radiolabelled test compounds
`was performed centrally in order to limit analytical
`variation and data analysis ofall laboratories was car-
`ried out according to a study-specific Excel spreadsheet.
`The total recovery of the radioactivity at the end ofthe
`experiment was not always as high as required by the
`
`guidelines (100+10% for OECD and 100415%for
`SCCNFP). Of the 7 laboratories that determined mass
`balance, 3 (benzoic acid), 4 (caffeine), and 5 (testoster-
`
`0009
`
`

`

`280
`
`J.LM. van de Sandt et al. | Regulatory Toxicology and Pharmacology 39 (2004) 271-281
`
`one) obtained a mass balance larger than 85". The most
`probable cause of the low recovery observed in some
`cases is the technical difficulty of evenly spreading the
`small volume of the dose solution on the skin surface
`(25uL/cm). It may be that part of the dose solution
`may have adhered to the pipet tip and therefore was not
`applied to the skin. It is important to mention that for
`regulatory studies most often very small volumes should
`be applied which are relevant for the in-use situation: 2-
`5mg/cm? for solid and semi-solid preparations and
`10 uL/em? for liquids (OECD 2000a; SCCNFP 2003).
`Although information on the mass balance is often
`lacking in the open literature, we present these data in
`order to give a complete overview of the sources of
`variation when performing in vitro skin absorption
`studies. For caffeine and testosterone, no clear difference
`was observed between laboratories with high and low
`recovery with respect
`to inter-experimental variation.
`Interestingly, the laboratories with good recovery ten-
`ded to have a lower inter-experimental variation for
`benzoic acid, which is probably related to the very high
`relative absorption of this compound.
`Each laboratory performed three to five independent
`experiments with each compound. The reproducibility
`of the maximum absorption rate determinations within
`each laboratory was generally higher for benzoic acid
`than for caffeine and testosterone. The observation that
`the intra-laboratory variation in this study was generally
`higher than that observed for absorption of methyl
`paraben througha silicone rubber membrane (Chilcott
`et al. submitte