Human skin levels of retinoic acid and
cytochrome P-450-derived 4-hydroxyretinoic
acid after topical application of retinoic acid in
vivo compared to concentrations required to
stimulate retinoic acid receptor-mediated
transcription in vitro.
E A Duell, … , P Chambon, J J Voorhees
J Clin Invest.
1992;
90(4)
:1269-1274.
https://doi.org/10.1172/JCI115990
.
Metabolism of retinoic acid to a less active metabolite, 4-hydroxyretinoic acid, occurs via
cytochrome P-450 isozyme(s). Effect of a pharmacological dose of retinoic acid on the level
of retinoic acid in skin and on cytochrome P-450 activity was investigated. A cream
containing 0.1% retinoic acid or cream alone was applied topically to adult human skin for
four days under occlusion. Treated areas were removed by a keratome and a microsomal
fraction was isolated from each biopsy. In vitro incubation of 3H-retinoic acid with
microsomes from in vivo retinoic acid treated sites resulted in a 4.5-fold increase (P =
0.0001, n = 13) in its transformation to 4-hydroxyretinoic acid in comparison to in vitro
incubations with microsomes from in vivo cream alone treated sites. This cytochrome P-450
mediated activity was oxygen- and NADPH-dependent and was inhibited 68% by 5 microM
ketoconazole (P = 0.0035, n = 8) and 51% by carbon monoxide (P = 0.02, n = 6).
Cotransfection of individual retinoic acid receptors (RARs) or retinoid X receptor-alpha
(RXR-alpha) and a chloramphenicol acetyl transferase (CAT) reporter plasmid containing a
retinoic acid responsive element into CV-1 cells was used to determine the ED50 values for
stimulation of CAT activity by retinoic acid and its metabolites. Levels of all trans and 13-cis
RA in RA-treated tissues were greater than the ED50 values […]
Research Article
Find the latest version:
Human Skin
Levels of Retinoic
Acid
and
Cytochrome P-450-derived 4-Hydroxyretinoic Acid after Topical
Application
of
Retinoic Acid In Vivo Compared
to Concentrations
Required
to
Stimulate
Retinoic
Acid Receptor-mediated Transcription In Vitro
ElizabethA.Duell, Anders Astr6m, ChristopherE.M.Griffiths, PierreChambon,* and John J.Voorhees
Department ofDermatology, University ofMichigan,AnnArbor, Michigan 48109-0528;and*Laboratorie deGenetiqueMoleculaire des Eucaryotesdu Centre National de la RechercheScientifiqueU.184Institut National de la Sante et de la RechercheMedicale,
Faculte de Medecine, Strasbourg,France
Introduction
Metabolism of retinoic acidto a less active
metabolite,
4-hy-droxyretinoicacid,
occurs viacytochrome
P450isozyme(s).
Effect ofapharmacologicaldoseof retinoic acidonthe level of retinoic acid in skin andoncytochrome P-450
activity
was in-vestigated. A creamcontaining
0.1% retinoic acid or cream alone wasappliedtopically
toadulthuman skin forfourdays
under occlusion.Treated areas wereremoved
by
a keratome anda microsomal fractionwas isolatedfrom eachbiopsy.
In vitro incubation of 3H-retinoic acid with microsomes from in vivo retinoic acid treatedsites resulted ina4.5-fold increase(P
=
0.0001,
n =13)
in itstransformation
to4-hydroxyretinoic
acid in
comparison
to in vitro incubations with microsomes from invivocreamalonetreatedsites. Thiscytochrome
P.450 mediated activity was oxygen- and NADPH-dependent andwasinhibited68% by 5
,M
ketoconazole(P=0.0035,n =8)
and 51% bycarbonmonoxide
(P
=0.02,
n =6).
Cotransfection of individual retinoic acidreceptors
(RARs)
orretinoidXreceptor-a(RXR-a)
andachloramphenicol
ace-tyl transferase
(CAT)
reporterplasmid containing
aretinoic acid responsive element into CV-1 cellswasusedtodeterminethe
ED50
valuesfor stimulation of CATactivity
by retinoic acid andits metabolites. Levels ofal
transand13-cis
RAin RA-treated tissues were greaterthan theED5,
valuesdetermined for all three RARs with these compounds.Furthermore,
thelevel ofall trans RA was greater than the
ED5o
for RXR-a whereas the 4-OH RA level was greater than theED5.
forRAR-,Band
RAR-'y
but lessthanforRAR-aandRXR-a. Thesedatasuggestthatthereare
sufficient
amountsof retinoic acid in treatedskintoactivategenetranscriptionoverbothRARsandRXR-a. (J. Clin. Invest. 1992. 90:1269-1274.) Key words:
high performanceliquid chromatography*
P450 inhibitors.
reporter assay.ED50
valuesThis workwaspresented inpart atthe annualmeeting oftheSociety for Investigative Dermatology, Seattle,WA, 1-4 May 1991.
Addressreprintrequests to Dr.ElizabethA.Duell,Departmentof Dermatology,University ofMichigan Medical Center, R6558KresgeI, Box0528,AnnArbor,MI48109-0528.
Receivedforpublication 21 October 1991 and in revisedform 11 February1992.
Retinoidssuchasalltransretinol
(ROL)'
and alltransretinoic acid(RA)areimportantinepithelial proliferationand differ-entiation( 1-3)andrepairofphotodamagedskin(4, 5).ROL is stored in tissuesasretinyl
esters(6-8)
orboundtocellularretinol binding protein (CRBP) (3). LittleornoRA is stored in tissues. Retinoidspresent in human skin areROL,
retinyl
esters,3,4-didehydroretinylesters, and RA(9).RA and3,4-di-dehydroRA(ddRA)act asimportant morphogensin the chick
wing budsystemeither directly (2)orindirectly (10, 11). The majorretinoic acid receptor(RAR)form expressed in the
epi-dermis is
RAR-'y
( 12,
13). Expressionofretinoid Xreceptor-a (RXR-a)has also been demonstrated in human skin( 14).Metabolism of RA occurs via a cytochrome P-450(P-450)
mediated enzyme system in tissues such as liver ( 15-17), tra-chea( 15),intestine(16), and rodent skin ( 18 ). Theprimary product formedis4-OH RA which is further transformedto 4-oxo RA. Topical application of RA to skin of 4-d-old rats
increasesRA metabolism fourfold. In vitro ketoconazole in-hibits conversion of RA to 4-OH RA bymicrosomesfromrat
epidermis ( 18 ).Inliver the formation of 4-OH RA via P-450
isozyme(s)isinhibitedboth in vivo and in vitro by ketocona-zole ( 19).
Inthis study a single topical dose of 0.1% RA cream or creamvehicle was applied to adult human skin andkeptunder constantocclusion. After 96 h vehicle- and RA-treated sites weretested for their capacity to metabolize RA and were
ana-lyzedfor RA and RA metabolite content. RA and its
metabo-lites were also tested for theircapacitytoinducea, RAR-(,RAR-'y,orRXR-a-mediatedtranscription ofa reporter gene inacotransfectionsystem.
Methods
Isocitrate, isocitratedehydrogenase, NADPH,NAD, 13-cisRA,and RAwerepurchased fromSigmaChemical Co. (St. Louis, MO).CV-l cells, African Green Monkey kidney cells,were purchasedfrom Ameri-canType Culture Collection (Rockville, MD).3H-RA was obtained from DuPont NEN (Boston, MA)and ['4C]chloramphenicol from AmershamCorp. (Arlington Heights,IL).RA metabolites 4-OH RA, 4-oxo RA, 13-cis 4-oxo RA,and5,6-epoxyRAweregifts fromDrs.
1. Abbreviations used in this paper: CAT, chloramphenicol acetyl transferase; CRBP, cellular retinol binding protein; ddRA, 3,4-didehy-dro retinoic acid; RA, alltransretinoic acid; RAR,RA receptor;ROL, alltransretinol; RXR-a, retinoidX receptor-a.
Abstract
J.Clin. Invest.
C©The American Societyfor Clinical Investigation, Inc. 0021-9738/92/10/1269/06 $2.00
Michael Rosenberger and P. F. Sorter of Hoffmann-LaRoche (Nutley, NJ). Metabolite 3,4-ddRA wasagift from Drs. I. Kompis and M. Klaus ofHoffmann-LaRocheAG (Basel, Switzerland). The 0.1 % RA (Retin-AO)andvehiclecream weresupplied by Ortho Pharmaceutical Corp.(Raritan,NJ). HPLC gradeacetonitrileand methanol were pur-chased fromBurdick& JacksonLaboratoriesInc., adivision ofBaxter Products (Romulus, MI). Spherisorb ODSl column was purchased fromPhaseSeparations Inc. (Norwalk, CT).
RAapplication and biopsy procedure. 0.5 g of RA (0.1% Retin-Ae cream) and 0.5 g of vehicle cream were applied to separately marked 3 x 2 inch areas on the buttock region of adult volunteers and kept oc-cludedfor four dayswithSaran Wrap and secured with surgical tape (20). Atfourdays test areas were evaluated for erythema, washed, and keratomed under 1%lidocainelocal anesthesia (20). In experiments to determine RA and RA metabolite levels, all three RA-treated areas were washed with chlorhexadine soap solution. In addition, two sites were washed with ethanol and one of the ethanol plus soap-washed siteswastape-strippedto"glistening", i.e., forremovalofstratum cor-neum,before keratoming. Epidermalsamples were immediately fro-zen inliquid nitrogen and stored at -80'C until used. Skin punch biopsies (2 mm)were takenfromtape-strippedsites and placed in 10% neutralbuffered formalin.Thesebiopsieswereprocessedfor hematox-ylinandeosin histological analysis ofadequacyofstratum corneum removal bystripping.All subjects gave informed, written consent. The study wasapproved bythe University of Michigan Medical Center Institutional ReviewBoard.
RAmetabolism assay. Microsomal fractions were prepared asprevi-ously described (21 ).RAand RAmetaboliteswerekeptin amber glass in darkenedroomswith yellowlightingwheneverpossible.The rateof RAmetabolismwasdeterminedby incubating 100 gg microsomal pro-tein in 0.01 Mphosphate buffer (pH 7.4) containinganNADPH re-generatingsystem(21),and 3H-RA(1 uM,spact 1 mCi/0.I
timol).
Sampleswereincubated for 15-45 minat350C.Theaddition of100M1 methanolcontaining 100,g/ml
ofbutylatedhydroxytoluene cooledto -20°Cterminatedthereaction.Aftercentrifugationat1,000gfor10 min, the supernatantfractionswereanalyzed forRAmetabolitesby reverse-phaseHPLC andliquid scintillationspectrometry.Ketoconazole,aninhibitorof cytochrome P-450,wasincubatedat 35°C for10 min with all componentsofthe assay present except for 3H-RA whichwasaddedto startthereaction.Forincubationswith CO asinhibitor,all componentsofthe assaywerepresent but tubeswere keptoniceduring 15 minflushingwith CO. Tubesweresealed and transferred to a 35°Cwater bath forformation of RAmetabolites. Control tubeswerehandled in thesame mannerexcept theatmosphere
wasnotequilibrated withCO.
HPLCseparation ofretinoids.AnHPLCsystem(1090M;Hewlett PackardCo.,PaloAlto,CA)containingaSpherisorbODSl column (PhaseSeparations Inc.,25cm X4.6mm),adiode arraydetector,and chemworkstationwasusedtoseparate retinoids. Mobilephasesused forgradientelutionofretinoidswerethoseof Buggeetal.(22).Mobile phase A was acetonitrile:0.02 M ammonium acetate:acetic acid (1:1:0.01) and mobilephaseBwasacetonitrile:0.2 M ammonium ace-tate:aceticacid( 19:1:0.008).At thestartoftherunsolventAwas100% followedbya lineargradientto 30% solvent Bat3 min, ashallow gradientto81%solventB at38 min, and 100% solventB at40 min.The flowratewas0.5ml/minand the total time for theseparationwas45 min. Effluent from the HPLC column floweddirectly into a flow-through scintillation spectrometer (Beta-One; Radiomatic Instru-mentsandChemicalCo., Inc.,Tampa,FL)withacomputerizeddata capture systemtodetermineradioactivityin eachmetabolitepeak.The amountofmetabolite formedwascalculated basedonthespecific activ-ity of added retinoic acid.
Extraction ofretinoidsfrom biopsy samples. Frozen keratome biopsies,eachweighing - 120 mg,werepowderedandimmediately
transferred into 20 vol of chloroform:methanol(2:1 ) containingtracer amounts of3H-RA.Retinoids were extractedwithgentlemixingfor 15 min followed bycentrifugation at 1,000g for 10 min. Supernatant fractionswereremoved and saved. Pelletswerereextracted with 5 vol
of the same solvent. After centrifugation supernatant fractions were savedand pellets were extracted with methanol:water acidified to pH 3.2 (95:5). After centrifugation the three supernatant fractions were combined and were evaporated to dryness under a stream of nitrogen. Samples were resuspended in methanol for HPLC injection. Recovery ofretinoids and assessment of nonenzymaticisomerizationof alltrans RA to 13-cis RA during extraction procedure were based on the amount of3H-RArecovered as3H-alltrans or 3H- 1 3-cis RA in compar-ison to the amount initially added as all trans3H-RA.
Cotransfection ofCV-J cells and reporter assay. Procedures re-ported by Astrom et al. (23) were used to determine
ED50
values for chloramphenicolacetyltransferase (CAT) activity by RA and RA me-tabolites.CV-l cells were cotransfected with RAR-a,RAR-f3,
RAR-y,or RXR-a expression vectors and the reporter plasmid TRE3-tk-CAT. Human RXR-a cDNA (24) was cloned from human skin as described ( 14)and subcloned into the expression vectorpSVL(Pharmacia LKB Biotechnology, Inc., Piscataway, NJ) for use in the transfection assay ( 14). The effect of variousconcentrationsof RA and RA metab olites (10-"-5 x 104 M) on CAT activity was determined as de-scribed (23).
Other methods. Statistical analyses for all assays were performed using a two-tailed paired t test. Protein content was determined by the method of Lowry et al. (25) with BSA as standard.
The concentrations of retinoids in skin were calculated as follows. Thebiopsies were weighed before pulverization. The recovery of unla-beled material was equated to the recovery of added3H-RA. Calibra-tion curves for the retinoids were stored in the HP-1090 HPLC unit. Theabsorption (peak area) from each sample was converted to nano-grams of retinoid based on the calibration curves. The nanonano-grams of retinoiddivided by the percent recovery gave the value for total reti-noid. This value was divided by the wet weight in grams to give micro-grams per gram skin. 1 g wet wt of tissue was equivalent to 1 ml ofwater for purposes of calculating molarity.
Results
Retention times for nonradioactive retinoid standards are
shown in chromatogram depicted in Fig. 1A.RAmetabolites separated by HPLC after in vitro incubation of 3H-RA with microsomal fractions obtained from biopsies treated in vivo withvehiclecream or0. 1% RAcream areshown in chromato-gramsdepicted in Fig. 1 B(vehicle)and Fig. 1 C (0.1% RA).
Incubation of 3H-RA with microsomes fromRA-treated sites showed
increased
formation of4-OH RA, 4-oxo-RA, andmorepolarmetabolites in comparison to metabolites generated by microsomes from vehicle-treated sites.
Formation ofRAmetabolites wastime (10-60min) and
protein concentration (100-400 ,ug) dependent (data not
shown).Datashowing increased P-450mediated RA metabo-lism four days aftera single topical
application
of 0.1% RA creamin comparison to vehiclecreamfor 13 normalvolun-teers aregiven in
Fig.
2. Therewas a4.5-fold (P= 0.0001) increase in formation of4-OH RA (124.4±17.4pg/min per mgmicrosomal protein)bymicrosomes from RA-treated sites in comparison to formation of metabolites by microsomesfrom vehicle treated sites (27.7±5.5 pg/min per mg micro-somal
protein).
There was a 2.6-fold increase in 4-oxo RAformation (P=0.0003) anda4.8-fold increase in formation of
morepolar metabolites (P=0.007) in RA-treatedversus
vehi-cle-treatedareas.
Datashowing inhibition ofP-450mediated
activity
byke-toconazolearegiveninFig.3.Withbiopsies fromRA-treated
sites assourceof microsomestherewas a68% decrease in4-OH
RAformation
(P
=0.0035,
n=8)
inincubationscontaining
5micro-( 61
E
41
21
% 6,00
a
CM
a4,00
U
a
.2
2,006,00 U IL
a
>.' 4,00
.U
.2
2,00la
0-O\
-!o
-0O
0-0'
TIME (MIN)
I I
10 20
TIME (MIN)
z W 200
I-0
2150
2 1
*t5
30
I0
-C
10
10
0-,.
A
10 20
TIME (MIN)
30
Figure 1. Reverse-phaseHPLCseparation of retinoidsasdetailed in Methods.(A)Chromatogramshowsretention times for nonradioac-tiveretinoidstandards.(B)Depictsthechromatogramobtained after invitro incubationof3H-RA with microsomal fractions from vehi-cle-treatedsites. (C) Depictschromatogramobtained after in vitro incubation of3H-RAwithmicrosomal fraction fromRA-treatedsites.
T
T
IL-w
UR
umE
Figure3. Ketoconazoleinhibits retinoic acid-induced cytochrome P-450-mediated transformation of retinoic acid into 4-OHretinoic acid. Conditionsfor in vitroassayofthemicrosomal fraction from RA-treatedsitesandanalytical proceduresaredetailed inMethods. Comparisonof3H-RAversus3H-RA+ 5 uMketoconazolegave the followingpercent decreasesand P values:4-OHRA(68%; 0.0035), 4-oxo RA(82%; 0.003),andpolarmetabolites (53%; 0.029).n =8. *,3H-RA- 1
AM;
o, 3H-RA+KETO-5MM.somal protein) in comparison to 3H-RA alone (154.1±26.2 pg/min per mgmicrosomal protein). There weresignificant
decreases in formation of 4-oxo RA (82%, P = 0.003, n = 8)as well as more polar metabolites (53%, P = 0.029, n = 7) in
incubations containing ketoconazoleand 3H-RA in compari-son to3H-RA alone. As shown inFig. 4,COsignificantly de-creased by 51%(P=0.02,n = 6)formation of 4-OH RA.
Concentrations of retinoids required to obtain half-maxi-mal stimulation
(ED50)
of CAT. activity by RAR-a,RAR-fl,
RAR-y,
or RXR-a in cotransfected CV-1 cells aregiven in Table I. Data for RAR-,y, the major receptorexpressedin epi-dermal cells, show that RA and 3,4 ddRA have anEDm
of 1-2 nM while 4-oxo RA requires a 10-fold higher concentration (13 nM) and 4-OH RA a 20-fold higher concentration (30nM). The 5,6epoxy RA ED50for
RAR-Ey
was49nM. To determinewhetherconcentrations of RA or its metabo-litesreachedsufficient levelstopotentially activate transcrip-tionby RARs or RXR-a after topicalapplicationofRA,levels of RA and its metabolites present in viableepidermisat fourdaysweredetermined. Fig. 5 shows that washing the sites with
soap and ethanol was notsufficienttoremove RAremaining
onthesurface of skin. After tapestripping theamountofRA
measured decreased 80% whereas 4-OH RA levels remained
z W 125
I. 100'
75
8
25'
E 0'
4-OHRA 4-OXO-RA POLAR METABOLITES
Figure2.Retinoic acidtreatmentof human skin in vivocauses
cy-tochromeP-450 mediated increased retinoic acid metabolism.In vi-troassayconditions and analytical proceduresaredetailed in
Meth--ods.Comparison of RA-treatedversusvehicle-treated sitesgavethe
followingpercentincreases and P values:4-OH RA(349%; 0.000 ), 4-oxo RA(162%; 0.0003),andpolarmetabolites (381%;0.007).n
= 13.m,VEH;o,RA.
7r
4-OH-RA 4-OXO-RA
Figure 4. COinhibits cytochrome P-450-mediated retinoic acid
me-tabolism into 4-OH retinoic acid. Microsomal fractions from sites treatedinvivo with 0.1%RAcream wereusedin in vitroincubations.
Assayconditions and analytical proceduresaredetailed in Methods.
Acomparison of incubation withRA+COversusRAalonegavethe
followingpercentdecreases and Pvalues:4-OH RA(51 %;0.02)and 4-oxoRA(53%; 0.07).n=6. .,3H-RA- I MM;o,3H-RA+CO
- MM.
z i 150 E 125'
IL -i
R 100' 0 75'
I 50
25 o5
TableI. Concentrations of Retinoids RequiredtoObtain Half-Maximal Stimulation
(ED5,)
ofTRE3-tk-CAT
Transcription by Retinoic Acid Receptorsand RXR-aCotransfectedintoCV-1 CellsED50
Compounds RAR-a RAR-,B RAR-Dy RXR-a nM
RetinoicAcid 24* 4* 1.3* 236
dd-Retinoic Acid 34 13 1.8 245
4-oxo-Retinoic Acid 79 28 13 377
4-OHRetinoic Acid >500 75 30 >500 5,6-EpoxyRetinoic Acid 237 56 49 >500
Data representtheaverage valuefortwo tothreetitrations from101 to5X
l-4
Mretinoid concentrations. * These datapoints forRA werepreviously published (23)andareincludedtofacilitate compar-isontometabolitedata.the sameregardless ofthe method of tissue preparation. All transand 13-cisRA were present in almost equal amounts.
The 13- cisRA present intissuewasnotduetononenzymatic
isomerization
during extraction since 91.2%±1.7% of the 'H-RAaddedastracer todeterminerecovery of materialremainedasalltransand 8.8%±1.1%was 13-cis RA. As shown in Table
II,
all
trans RA was present as824 ng/gmwetwt or- 2.75,uM
concentration
(threelog units higher
than theED50
forRAR-y),
13 cis-RAas745 ng/gm wet wt or - 2.5tiM
(60X
ED"o
value
for
RAR-y),
and4-OHRA as93 ng/gmwet wt or - 0.3,M
concentration
(10XED50
value forRAR-'y).
4-Oxo RA wasnotdetected in RA-treatedbiopsies
eventhough
epidermal
microsomes from similar
biopsies
showed conversiontoboth 4-OH and 4-oxo RA.Discussion
The
effects of
daily application
ofRA without occlusion tophotodamaged skin
havebeeninvestigated
in clinicalstudies (4,5). Theimprovement
seenis
variable and dosedependent
(4,5). Pilot studies with 0.1%RAcreamwithorwithout
occlu-sion
usedtopically
for variousperiods
of time showed that RAunder occlusion for four
days
gaveahighly
consistenterythem-atous
reaction
as well as characteristichistological
(20)
re-7
5
0
WASH WASH+EtOH WASH
EtH
+STRIPPED
Figure 5. Retinoidcontentinareastreated fourdayswith0.1%RA creamfollowedbyremovalof surface RAbytheindicated methods. Extractionproceduredetailed in Methods. Concentrations instripped
tissuesampleswerealltransRA( 2.7
AM),
13 cis-RA(nu
2.5MAM),
and 4-OH RA(-- 0.3
MM).
n= 5foreachmethod ofsurfaceRA removal.o, all trans RA; m, 13-cisRA;.,
4-OHRA.Table I. Comparison ofTissueLevelsafterTopical Application
ofAllTrans RetinoicAcidto HumanSkinIn Vivowith
ED50
ValuesforTranscriptionalActivation ofTRE3-tk-CA
T in aCotransfectionAssay
alltransRA 13-cis RA 4-OH RA
ng/g wet wt 824±281 745±24 93±36
nM 2750 n~2480 310
Retinoid(nM)
ED50RAR-a 24* 190* > 500
EDORAR-#
4* 43* 75ED50RAR-y
1.3* 140* 30ED,1RXR-a 236 ND > 500
*These datapointswerepreviously published (23) and are included
forfacilitatingcomparisons.
sponsein 9
of
10patients. Becauseof thisconsistent retinoid response,thefourdayoccludedpatch was used in the currentstudy.
Routine
nonoccluded clinical use ofRA creampro-ducedaninconsistent retinoidresponse too variable for
experi-mentation.
Inthepresent study asingletopical application of RA with
occlusioncontinuously for four daystoadult human skin
re-sulted in
sufficient
retinoic acid penetration into viablelayersof the
epidermis
toinduceRAmetabolism.Our data and thatof
others (18)
suggestthatthis
increased metabolism isproba-bly dueto
increased
P-450activity.
Furthermore, inductionofP-450
activity
in response to RAhasbeen reported in othertissues ( 15-19).Inaddition, 5,6epoxy RA can beformed
ei-therbya
lipid peroxidase
(26) or aP-450-mediated activity (27). However, the P-450 isozyme(s) induced inhumanskin inresponse totopical
RAis unknown.P-450enzymespurified from rabbit liverhave been
evalu-ated for
their
capacity
to metabolize RA. IA2 and IIB4iso-zymes were mosteffective in metabolizing RA to 4-OH RA
while
IIC3,IIE 1,IG2, and IIE2wererelatively ineffective,andisozymes
IA 1 andIIIA6wereineffective (28).Cultured human
keratinocytes
have beenshowntocontain the mRNA for P-450isozyme
IA 1 (29). However, the IAIisozyme
was oneoftwoformsfrom
rabbit liver that didnotmetabolize
RA. Animmunohistochemical
study of cultured humanepidermal
foreskin keratinocytes showed binding of monoclonal antibodiestoisozymes
belongingtofamilies
IIandIII
(30).
Inneonatalratstopical application of
3-methylcholan-threnewas moreeffective
thantopical application of
RA in inducing RA metabolism by P-450 isozyme(s)(18).Pheno-barbitolwastheleasteffective inducer ofRAmetabolism ( 18).
In hamsters intraperitoneal injection of methylcholanthrene increased RA metabolism in liver twofold while orallygiven
RA increased RA metabolism eightfold in liver ( 16). Thus
metabolism ofRA in response to inducers of P-450 activity
may vary as a result ofroute of administration, species in-volved, ortissuestudied.
Cytochrome
P-450mediated metabolism ofRA may beasignificant mechanism for
regulation
ofpharmacological
RA levels invivo.Infacttheconversion ofRAtolessactiveattenuatesphysiologically relevant, endogenous RA levelscan notbe addressed with the data presented here. However, our
results do demonstrate that topical application ofpharmacolog-ical dosesof RA in humans induce a 4.5-fold increase in
metab-olism
ofRAto4-OH RA andmorepolar metabolites. Keto-conazole,atopical antifungal agent currentlyin routineclini-cal use, is an inhibitor ofsome P-450 mediated activities.
Ketoconazoleatconcentrations of 7X
10-7
Mto5 X 10-6Minhibits
invitroRAmetabolismby
ratepidermal
microsomes. Similarconcentrationsarerequiredtosuppresscertain P-450 dependent reactions in steroid synthesis (19). One possible side effect of ketoconazoleuse is increased RAlevels duetodecreasedRAmetabolism. Thismaybe of clinical importance inas muchas simultaneoususe oftopical ketoconazole and
RAcould decreasetheamountof RA requiredtoobtain clini-calefficacy.
Todetermine thepotency
of
ddRA and RAmetabolites
incomparison
to RA and 13-cis RA in RAR- and RXR-me-diatedtranscriptional activation,
a reporter gene (TRE3-tk-CAT)wascotransfected together withreceptorexpression vec-torsintoCV-l
cellsandconcentrationsrequired
toobtain half-maximal inductionof
CATactivity
(ED50)weredetermined.Similar
reporter assay systemshave beenestablished with cell lines other than CV-1cells. TheED50s
observedforRAin thesesystems varywith RARreceptor typerather than the cell line used
for
cotransfectionprocedure (23, 24).
The CV-l cells used in thereporter assay areepithelial
cells but theappropriate-ness
of
anyinvitro
system to in vivo occurrences is alwaysopen to
question.
Of
the three knownRARs,
RAR-'y
is themajor expressed
form in human skin (
12, 13)
and inthereporter assay systemused in these
experiments
RAR-,y has the lowest ED50for all
trans RA(1.3 nM)(23),
alltransddRA(1.8 nM),all
trans4-oxo-RA (13 nM), and
all
trans4-OH RA (30 nM). TheED50valuesare two- tofourfold higher forRAR-,B than those
determined for
RAR-y
except forall
trans ddRA which isseventimes
higher.
The ED50 valuesforRAR-a werethehigh-est
obtained with
anyof
the three RARs. TheED50
valuesforall
trans RA(23) andall
transddRAwith
RAR-awassimilartothe valueobtained with 4-OHRAand
RAR-'y
(30 nM). TheED50of 4-OH RA for RAR-a wassimilarto thatofRXR-a
(> 500nM).RXR-ahas beenshowntobeexpressed inskin in vivo (14)but demonstrates higher ED50 values for all
com-pounds tested comparedtoRARs.TheED50sinthe presenceof
RXR-a are20times higher foreach compound in comparison tovalues
for
RAR-y.
The lowestED_0
for 13-cis RA is 43 nMwithRAR-,B receptor (2
3).
The value for 1 3-cis RA with RAR--yis 140nM(23) (TableII).Todetermine whether metabolites mightplay a significant
role inretinoid effectsobservedfollowingtopical application of
RA, treated siteswere analyzedfor presence of RA and RA
metabolites.
Washing withsoap and ethanol was not sufficient to remove RAstillpresent on the surface and within stratum corneumof skinasevidencedby
4jig
RA per gwet wt(
12,uM)
in thespecimen.
Removal ofstratumcorneum andsur-faceRAwithtapestripping reduced these values to 1.5
,tg
RA per g wet wt (5 ,M). The method of preparing the site forbiopsy didnotchange the level of 4-OH RA (93 ng/g wet
wt,
0.3 ,uM).
Thissuggests that the presence of4-OH
RA indicatesincreasedRAcontent and increased metabolism of RA within
living
layers
ofepidermis
withoutnecessitating
tape
stripping
ofskintodetermineRAmetabolism.Intape-stripped biopsy samples, 53% ofRA extracted was
alltransRA(2.7
,qM)
and 47%was 13-cis RA(2.5,M).
This concentration of alltransRA is2,000
times higher than theED50
forRAR-y
and50-foldhigher thantheED50 forRXR-ainthe cotransfectionassay.Thehigh levels of extracted13-cis RA aftertopical application of alltransRAandapparentlack ofsignificant nonenzymatic isomerization during retinoid
ex-traction from tissue suggests that
epidermis
may contain anactive RA isomerase. Theamountof 13-cisRAin these
sam-ples
is10-fold
higher
than theEDOs
in thereporter assayof
140nMwith
RAR-y
and 190nMwithRAR-aand50-fold higher than theED,0
of43nMfor RAR-3(23).
The4-OHRAlevelswere 10times
higher
than the ED50determined for RAR-,yandfour times
higher
than the ED50 for RAR-gi but less than theED50
of> 500 nM with RAR-a and RXR-a in the cotransfec-tionassay.The above
comparisons
aremadeontheassumption
that concentrationsofretinoids
in skin havefunctionalactivity
sim-ilartothatobserved when thesameretinoids
areaddedtocell culturescontaining genetically engineered
retinoid response systems.However,
theremaybeunrecognized
differencesbe-tween invivo and in vitro cell
penetration, distribution,
and metabolism.In
conclusion,
the effects observedclinically
aftertopicalapplication
ofpharmacological
doses ofalltrans RA under occlusionfor fourdays
may be due in parttoall
trans RAand 13-cis RAaswellas4-OH RA. The largequantity of all
trans RA as well as lesser butsubstantial
amountsof 13-cis
and 4-OH RA in relation totheconcentrationrequired
for half-maximalactivation of transcription
in thecotransfection
assay suggeststhat thebiological
responseof skin(20)
inresponseto RA under short-term occlusion maybemediated
atleast inpart
by
both RARs and RXR-a. Whether molecularchanges
inskin inresponsetoRA
application
canbemediated byRAR-and
RXR-a-independent
mechanisms remains unknown.Acknowledgments
WewishtothankUlrikaPetterssonforexcellenttechnical assistancein the reporter gene assay, Robin Gardnerfor providing biopsy materials, and TedHamiltonfor statisticalanalyses.
Thisstudywassupportedin partbyagrantfromR. W. Johnson PharmaceuticalResearchInstituteand theBabcock Endowment.
References
1.Roberts,A.B.,and M. B.Sporn.1984.Cellularbiologyandbiochemistry of
theretinoids.InTheRetinoids.M. B.Sporn,A.B.Roberts,and D.S.Goodman, editors. AcademicPress,NewYork. Vol.2,209-286.
2.Thaller, C.,and G.Eichele.1990.Isolation of3,4-didehydroretinoicacid, a novelmorphogenetic signalin thechick wingbud. Nature(Lond.).345:815-819.
3.Dolle, P.,E.Ruberte,P.Leroy,G.Morriss-Kay,and P.Chambon.1990.
Retinoicacid receptors and cellularretinoid binding proteins. 1.Asystematic studyof their differential pattern oftranscription duringmouseorganogenesis. Development(Camb.). 110:1133-1151.
4.Weiss,J.S.,C. N.Ellis,J. T.Headington,T.Tincoff, T. A. Hamilton, and J. J. Voorhees. 1988.Topicaltretinoinimprovesphotodamagedskin: a double blind vehicle-controlledstudy.JAMA(J.Am.Med. Assoc.). 259:527-532.
5.Weinstein,G.D.,T.P.Nigra,P.E.Pochi,R.C. Savin, A. Allan, K. Benik,
E.Jeffes,L. Lufrano,and G. Thorne. 1991. Topical tretinoin for treatment of
photodamagedskin.Arch.Dermatol. 127:659-665.
6.Shingleton,J.L.,M.K.Skinner,andD. E.Ong. 1989. Retinol esterification in Sertoli cellsbylecithin-retinolacyltransferase. Biochemistry.28:9647-9653.
7.Torma, H.,andA.Vahlquist.1987. Retinol esterification bymouse epider-malmicrosomes: evidence foracyl-CoA:retinol acyltransferase activity.J. Invest.
8.Torma, H., and A.Vahlquist. 1990. Vitamin A esterification in human
epidermis:arelation to keratinocyte differentiation. J. Invest. Dermatol. 94:132-138.
9.Vahlquist, A. 1982. Vitamin A in human skin. I. Detection and
identifica-tion of retinoids innormalepidermis.J. Invest. Dermatol.79:89-93.
10.Wanek, N., D. M.Gardiner,K.Muneoka, and S. V. Bryant. 1991. Con-version by retinoic acidofanterior cells into ZPA cells in the chick wing bud. Nature(Lond.).350:81-83.
11.Noji, S., T. Nohno, E Koyama, K. Muto, K. Ohyama, Y. Aoki, K. Tamura, K. Ohsugi, H.Ide, S. Taniguchi, and T. Saito. 1991. Retinoic acid inducespolarizingactivity but is unlikely to be a morphogen in the chick limb bud. Nature(Lond.). 350:83-86.
12. Krust, A., P. H. Kastner, M.Petkovich,A.Zelent, and P. Chambon. 1989. Athirdhumanretinoic acidreceptor, hRAR-'y.Proc. Natl. Acad. Sci. USA.
86:5310-5314.
13. Elder, J. T., G. J. Fisher, Q. Y. Zhang, D. Eisen, A. Krust, P. Kastner, P.
Chambon,and J. J.Voorhees. 1991. Retinoic acidreceptorgene expression in human skin. J. Invest.Dermatol. 96:425-433.
14.Elder, J. T., A. Astrom, U. Pettersson, A. Krust, P. Kastner, P.Chambon, and J. J. Voorhees. 1992.Differentialregulation of retinoic acid receptors and
binding proteinsin human skin. J. Invest.Dermatol.98:673-679.
15.Frolik, C. A., A. B. Roberts, T. E. Tavela, P. P. Roller, D. L. Newton, and M.B.Sporn. 1979. Isolation and identification of 4-hydroxy- and 4-oxoretinoic
acid.Invitro metabolites of all-trans-retinoic acidin hamster trachea and liver.
Biochemistry. 18:2092-2097.
16. Roberts, A. B., M. D. Nichols, D. L. Newton, and M. B. Sporn. 1979. In vitrometabolism of retinoicacid in hamster intestine and liver. J.Biol.Chem. 254:6296-6302.
17.Leo, M. A.,S.lida,and C. S.Lieber.1984.Retinoic acidmetabolism by a systemreconstitutedwith cytochrome P-450. Arch. Biochem. Biophys. 234:305-312.
18. Vanden Bossche,H.,G.Willemsens,and P.A.J. Janssen. 1988. Cy-tochromeP450-dependent metabolism of retinoic acidin ratskin microsomes: inhibitionby ketoconazole.Skin Pharmacol.1:176-185.
19. Van Wauwe, J. P., M. C. Coene, J. Goossens, G. VanNijen,W.Cools, and W.Lauwers. 1988. Ketoconazole inhibits the in vitro and in vivo metabolism of
all-trans-retinoicacid. J.Pharm.Exp. Ther.245:718-722.
20. Fisher, G. J., J. Esmann, C. E. M.Griffiths,H.S.Talwar, E.A.Duell,C. Hammerberg, J. T. Elder, L. J.Finkel,G. D.Karabin,B.J.Nickoloff,etal.1991. Cellular,immunologicandbiochemical characterization oftopicalretinoic acid-treatedhumanskin.J. Invest.Dermatol. 96:699-707.
21. Duell, E., V.Fischer, G. Fisher,A.Astrom, and J. Voorhees. 1991. Cyclo-sporineAmetabolism by cytochrome P-450111occursinmicrosomes fromrat
liver but notfromnormalepidermisorpsoriaticlesions. J. Invest.Dermatol. 96:827-831.
22. Bugge, C. J. L., L. C.Rodriguez,and F. M. Vane.1985.Determination of isotretinoinoretretinateandtheirmajor metabolitesinhuman bloodby re-versed-phase high-performance liquid chromatography.J.Pharm.Biomed. Anal. 3:269-277.
23. Astrom, A., U. Pettersson, A. Krust, P. Chambon, and J. J.Voorhees.
1990.Retinoic acidand syntheticanalogs differentially activate retinoic acid
receptordependent transcription. Biochem. Biophys.Res.Commun. 173:339-345.
24.Mangelsdorf,D.J., E. S.Ong,J. A.Dyck,and R. M. Evans.1990.Nuclear receptorthat identifiesanovel retinoic acidresponsepathway.Nature(Lond.). 345:224-229.
25. Lowry,0. H., N. J. Rosebrough, A. L.Farr,and R. J. Randall. 1951. ProteinmeasurementswiththeFolin phenolreagent. J.Biol. Chem. 193:265-275.
26. Sietsema,W.K., and H. F. DeLuca. 1982.Retinoic acid5,6-epoxidase.J.
Biol. Chem. 257:4265-4270.
27.Watabe, T., and T. Sawahata. 1979.Biotransformation of cholesterolto
cholestane-3,S5a,6jf-triol
via cholesterol a-epoxide (5a,6a-epoxycholestan-3fl-ol)inbovine adrenalcortex.J. Biol.Chem. 254:3854-3860.28. Roberts,E.S., A. D. N. Vaz, and M. J. Coon. 1991.Metabolism of retinoidsbyrabbitP450cytochromes.FASEB(Fed.Am.Soc.Exp.Biol.)J. 5:15 14a.(Abstr.).
29. Berghard,A., K.Gradin,and R.Toftgard. 1990. Serum and extracellular calcium modulate induction ofcytochromeP-4501A1inhumankeratinocytes.J. Biol.Chem. 265:21086-21090.