Journal of Cell Science 102, 113-121 (1992)Printed in Great Britain © The Company of Biologists Limited 1992

113

Retinoic acid receptor expression in human skin keratinocytes and dermal

fibroblasts in vitro

CHRISTOPHER P. F. REDFERN* and CAROLE TODD

Medical Molecular Biology Group, Department of Dermatology, 4th Floor Cookson Building, Medical School, University of Newcastle,Framlington Place, Newcastle upon Tyne NE2 4HH, UK

•Author for correspondence

Summary

Retinoic acid is essential for the normal differentiation ofepithelia but its cellular function is obscure. Theexpression patterns of retinoic acid receptors (RARs) inskin cell types may give an insight into the role of retinoicacid in skin. We have compared the patterns of RARexpression in human keratinocytes and dermal fibro-blasts in vitro, and studied the effects of retinoic acid onRAR expression. RAR-a and RAR-y were expressed hikeratinocytes and fibroblasts: RAR-y was expressed atsimilar levels in both cell types but RAR-a was moreabundant hi fibroblasts. There were no differences hiexpression of either RAR-a or RAR-y between stratify-ing (high-calcium medium) and proliferating (low-calcium medium) keratinocytes and expression of theseRARs was unaffected by retinoic acid. RAR-/J wasundetectable in keratinocytes. In the majority offibroblast cell lines, RAR-/J transcripts were either

undetectable or expressed at a low level. Retinoic acid atlow concentrations (10~10 to 10~9 M) rapidly inducedthe expression of RAR-/?. Cyclic adenosine monophos-phate (cAMP) analogues inhibit RAR-/} induction interatocarcinoma cells. However, dibutyryl-cAMP didnot affect RAR-/? induction in fibroblasts. Forskolin, anadenylate cyclase activator, and the phosphodiesteraseinhibitor 3-isobutyl-l-methylxanthine (TBMX) de-creased constitutive RAR-/J mRNA levels but did notblock induction of RAR-/3 by retinoic acid. Sinceintracellular cAMP levels were only increased detectablyhi response to forskolin, the reduction in constitutivelevels of RAR-/J mRNA may be mediated by mechanismsother than via cAMP.

Key words: retinoic acid, retinoic acid receptors,keratinocytes, fibroblasts, human.

Introduction

Retinoic acid has marked effects on skin and epithelia,and in excess inhibits normal keratinocyte differen-tiation and may induce mucous metaplasia; conversely,retinoic acid deficiency results in hyperkeratosis orsquamous metaplasia (Sengel, 1976). Many studieshave been directed at elucidating the effects of retinoicacid on keratinocytes, the major epidermal cell type,and it is clear that retinoic acid directly influencesproliferation and alters the differentiation pathway ofthese cells in vitro (Green and Watt, 1982; Marcelo andTomich, 1983; Redfern and Todd, 1988). Althoughthere is little doubt that retinoic acid has direct effectson epidermis, some of the observed effects of retinoicacid on epithelial tissues and their derivatives in vivomay, in fact, be mediated by underlying mesenchymalor stromal cells (Tickle et al. 1989; Covant and Hardy,1990; Hardy et al. 1990).

Specific cellular binding proteins (cellular retinoicacid binding protein or CRABP) and nuclear receptorsfor retinoic acid have been identified (Ong and Chytil,

1978; Daly and Redfern, 1987) and cloned (Petkovich etal. 1987; Giguere et al. 1987; Brand et al. 1988; Zelentet al. 1989; Krust et al. 1989); these presumably mediatethe diverse effects of retinoic acid on different tissues.Retinoic acid receptors (RARs) are ligand-activatedtranscriptional regulators closely related in structure tosteroid and thyroid hormone receptors (review: Red-fern, 1992). Three main classes of RAR have beendescribed and, for RAR-/S and RAR-y at least,alternative splicing and transcription from differentpromoters generates transcripts coding for RAR pro-teins differing at their amino-terminal ends (Zelent etal. 1991; Lehmann et al. 1991). RARs and the cytosolicretinoid binding proteins, CRABP and cellular retinolbinding protein (CRBP), are expressed in precisespatiotemporal patterns during embryological develop-ment (Doll6 et al. 1990; Ruberte et al. 1991). Thisdevelopmentally regulated expression of specific reti-noic acid receptors, coupled with the dramatic disrup-tions of normal pattern formation caused by excessretinoic acid (Tickle et al. 1982; Eichele, 1989), arguesthat retinoic acid functions as a regulatory molecule

114 C. P. F. Redfern and C. Todd

involved in the control of morphogenesis. Cell andtissue specificity in the biological effects of retinoic acidwill be determined, at least in part, by the patterns ofexpression of different retinoic acid receptors.

With respect to skin, although retinoic acid can, invitro at least, apparently determine whether keratino-cytes differentiate to a squamous or mucus-secretingcell phenotype (Sengel, 1976), its role in the normaldevelopment and maintenance of epidermis in vivo isuncertain. Since interactions between epithelial andmesenchymal elements are important in both epidermaldifferentiation and the development of epidermalstructures (Sengel, 1976), it is important to considerdermis as a potential mediator of the effects of retinoicacid on epidermis in vivo. Cultured cells can berelatively readily derived from normal human skintissue and represent potentially useful models forstudying the roles of retinoic acid in skin biology. Wehave addressed the question of whether or not retinoicacid is likely to have different developmental roles indifferent skin cell types by studying the patterns ofexpression of RARs in human keratinocytes andfibroblasts cultured in vitro.

Materials and methods

Cell cultureKeratinocyte cultures were established from foreskin,retroauricular skin or foetal skin (abdominal), grown underserum-free conditions in medium MCDB 153 (Sharpe et al.1989) and used after three to six passages. To inducestratification, the calcium ion concentration was increasedfrom 0.07 mM to 1.2 mM, and the cells were used forexperiments after 4 days at this higher Ca2+ concentration.Dermal fibroblast cultures were established by explantoutgrowth from small blocks (~8 mm3) of dermal tissueplaced on a scratched surface of plastic tissue-culture dishes.Cells that grew from the explant were subsequently passagedand cultured in Dulbecco's minimal essential medium(DMEM) containing 10% foetal bovine serum (FBS). Theorigin of each fibroblast primary cell line is given in Table 1.

For experiments with retinoic acid, dermal fibroblasts were

Table 1. Dermal fibroblast cell lines and their sites oforigin

fs2£s3fs4£s5bx4bx7bx8Fetal*scdmflO.ll

Site

ForeskinForeskinForeskinForeskinFaceBackForearmFetal backBreastChestChest

Age/sex

2 yChild5 y11 y- F12y/F89y/F24 wk -- F64y/M64y/M

In the case of non-foreskin samples, cell lines were establishedby explant outgrowth from 4 mm punch biopsies, or, for cell lineflO.ll, from a strip of chest skin obtained during surgery forcardiac bypass graft (Latham et al. 1989).

•Aborted after diagnosis of polycystic kidney disease.

used when approximately 70% confluent and the medium wasreplaced 12-24 hours before adding all trans retinoic acid(Sigma) in ethanol to final concentrations within the range 0.1to 1000 nM. An equal volume (<5 /A per 10 ml medium) ofethanol was added to control cultures. Concentrations ofretinoic acid stock solutions were estimated using an extinc-tion coefficient of 36,500 at 343 nm. The adenosine 3':5'-cyclicmonophosphate (cAMP) analogues, A^,2'-O-dibutyryladeno-sine 3 :5'-cyclic monophosphate (dibutyryl-cAMP) and 8-bromoadenosine 3':5'-cyclic monophosphate (8-bromo-cAMP) (Sigma), were dissolved in culture medium and addedto the cell cultures to give a final concentration of 1 mM. Thephosphodiesterase inhibitor, 3-isobutyl-l-methylxanthine(IBMX) (Aldrich), was dissolved in 1 M NaOH and used at afinal concentration of 1 mM and the adenylate cyclaseactivator, forskolin (Sigma), was dissolved in ethanol andused at a final concentration of 10~5 M. The appropriatevehicle was added to control cultures for each experiment.

The effects of dibutyryl-cAMP on fibroblast proliferationwere studied by seeding 50,000 cells into 35 mm diameterculture dishes; the cells were allowed to attach for two hoursand dibutyryl-cAMP was added to a final concentration of 1mM. Cells were refed with fresh medium containing 1 mMdibutyryl-cAMP each day, and after three days the cells weredetached and counted using a haemocytometer. Controldishes set up in parallel were treated in the same way exceptthat dibutyryl-cAMP was omitted from the medium.

cAMP assayFibroblasts, seeded into 25 cm2 tissue-culture flasks (0.5 x 106

cells/flask) and used the following day, were treated for sixhours with retinoic acid, forskolin or IB MX (as above). Thecell monolayers were then washed twice with phosphate-buffered saline (Flow Laboratories, Dulbecco's formula,without calcium and magnesium), extracted with 1 ml of 70%(v/v) ethanol in water and finally with 1 ml of 65% (v/v)ethanol. Ethanolic extracts were combined and a sample waslyophilized for assay with a cAMP radioimmunoassay kit(Amersham International, Amersham, UK).

RNA extractions and Northern blottingTotal cellular RNA was prepared by the guanidiniumisothiocyanate/caesium chloride method (Chirgwin et al.1979). RNA samples (15 ng per track for 0.5 cm wide slots andcontaining ethidium bromide as a visual check on RNAloading) were size-fractionated on 1.2% agarose/formalde-hyde gels and transferred by vacuum blotting with 1.8 MNaCl, 0.01 M EDTA, 0.1 M sodium phosphate, pH 7.4(IOXSSPE), to nylon membranes (Amersham). Membraneswere hybridized at 42°C with 32P-labelled probe using 50%formamide, 6xSSPE, 0.2% (w/v) Ficoll 400, 0.2% (w/v)polyvinylpyrrolidone, 0.2% bovine serum albumin (fractionV), 0.5% SDS, 5% dextran sulphate, 200 ̂ g ml"1 tRNA and100 fig ml"1 single-stranded carrier DNA as the prehybridiz-ation and hybridization buffer. After hybridization, mem-branes were washed 3-4 times in O.lxSSPE, 0.1% SDS for>15 minutes each at 68°C and exposed to X-ray film withintensifying screens at — 70°C. For quantitative autoradiogra-phy, X-ray film was preflashed and the autoradiographsscanned using an LKB laser scanning densitometer or animage analysis system.

ProbesThe RAR-ar probe was a Kpnl/Sacl fragment (503 bp) fromthe 5' end of the human RAR-Q-1 CDNA (Petkovich et al.1987). The human RAR-/3 probe consisted of the completeRAR-/S2 cDNA insert (1400 bp) of the plasmid pCOD20

(Brand et al. 1988). The human RAR-y probe was the full-length, 1500 bp yl cDNA insert (Krust et al. 1989). As afurther check on RNA loading, the membranes for someexperiments were reprobed with either a mouse 18 Sribosomal RNA cDNA probe (Edwards et al. 1985) or a rat fi-actin probe consisting of a 1200 bp BgH fragment of thepRp"A-l cDNA clone isolated by P. Gunning. A mouse a-actin probe consisting of a 1150 bp Pstl fragment of the cDNAclone pAM (Minty et al. 1981) was used as a test of fibroblastphenotype. Probes were labelled with [32P]dCTP (AmershamInternational, 3000 Ci mmol"1) to a specific activity ofapproximately 109 disints min"1 ng~l (Feinberg and Vogel-stein, 1983).

With the high-stringency post-hybridization washing con-ditions used for these experiments there is no detectablecross-hybridization of the RAR probes. The human RAR-aand RAR-ycDNA probes were provided by Martin Petkovichand Pierre Chambon, Strasbourg, France, and the RAR-/Sprobe by Anne Dejean, Paris, France.

Results

RAR expression in cultured keratinocytesKeratinocytes proliferate to form a monolayer of cellswhen cultured in media with a low (0.07 mM) Ca2+

concentration. At higher Ca2+ levels (>0.1 mM), thecells stratify and grow in tight colonies (Hennings et al.1980; Boyce and Ham, 1983). To ask whether or notRAR expression varies in relation to these Ca2+-induced changes in vitro, RNA was extracted fromkeratinocytes cultured continuously in low-calciummedium, and from keratinocytes cultured for four daysin high-calcium medium. Both RAR-a and RAR-ywere expressed in keratinocytes and there was noconsistent difference in RAR expression between cellsgrown in high-calcium or low-calcium medium (Fig. 1),and no differences between keratinocytes obtainedfrom different sites or ages of donor. The RAR-ff probedetected two transcripts, approximately 3.6 and 2.8 kb,as has been described in other tissues (Rees et al. 1989),and the shorter of the two transcripts was moreabundant. RAR-y transcripts were detectable as a bandat approximately 3.2-3.3 kb. RAR-p1 mRNA wasundetectable in all keratinocyte RNA samples analysed(Fig- 1).

Since RAR- a and RAR-/J are inducible in responseto retinoic acid in some cell types (de The' et al. 1989;Redfern et al. 1990; Clifford et al. 1990; Leroy et al.1991; Kastner et al. 1990), the effect of retinoic acid onRAR expression in keratinocytes was investigated.However, for cells grown in either low-calcium or high-calcium medium, there were no significant changes inRAR expression in response to treatment with 10~7 Mretinoic acid for up to 24 hours (Fig. 1).

Dermal fibroblastsRAR expression patterns

We have reported previously that of two human dermalfibroblast lines isolated, a cell line (fl0.ll) isolated fromthe chest skin of a 60-year-old male expressed RAR-/3at a high level whereas a breast-skin fibroblast line (lineSC) did not (Rees and Redfern, 1989). In view of this

Retinoic acid receptors in cultured skin cells 1151 2 3 4 5 6 7 8 9 1 0 1 1 1 2

RAR-a

RAR-/S

rr^M n ' T , BMUB.I

III

L L L H"^ 18S rRNA

H L L H L

Fig. 1. RAR expression in human keratinocytes from fetalskin (24 weeks gestation; tracks 1 to 5), foreskin (tracks 6to 8, and 10) and retroauricular skin (tracks 9, 11 and 12),cultured in either low-calcium (L) or for 4 days in high-calcium (H) medium. In tracks 3 and 4, RNA wasextracted from keratinocytes grown in low-calcium mediumand treated with 10~7 M retinoic acid (RA) for 24 hours;the control for this experiment is track 2. Similar resultswere obtained using keratinocytes grown for 4 days inhigh-calcium medium and treated for 8 h with 10~8 Mretinoic acid (not shown). The blot was probed successivelywith all three RAR-probes, and finally with a 18 S rRNAcDNA probe as a check on RNA loading.

potential heterogeneity of dermal fibroblasts, we stud-ied RAR expression in fibroblast cultures establishedfrom biopsies taken from different body sites and fromindividuals ranging in age from 24 weeks gestation to 89years (Table 1). RAR- a was expressed in all fibroblastcultures tested and at a 1.5- to 2-fold higher level thankeratinocytes, relative to RAR-y (Figs 2 and 3). RAR-ywas also expressed in all samples, and at a levelcomparable to keratinocytes (Figs 2 and 3). Althoughthere was some variation in RAR-y signal intensitybetween the cell lines, this correlated with slightvariations in RAR-a signal intensity and RNA loading.RNA was extracted from only one sample of the fl0.llcell line but for other fibroblast cell lines the RAR-expression patterns were consistent between differentbatches of cells.

With the exception of the dermal fibroblast cell linefl0.ll, RAR-/3 transcripts were either undetectable orpresent at a low level in cultured fibroblasts. The fl0.llfibroblast cell line was unique with respect to the veryhigh level of expression of RAR-p\ These cells are nolonger in existence and to investigate the possibility thatthe individual from which the fl0.ll cells were derivedhad consistently high levels of RAR-/3 expression in hisdermal fibroblasts, new dermal fibroblast cultures wereestablished from the same body site of the originaldonor. However, in these subsequent cultures theexpression of RAR-p' was undetectable (Fig. 4).

116 C. P. F. Redfern and C. Todd

All fibroblast cell lines used for these experimentswere obtained by explant outgrowth and no attemptwas made to derive clonal cell populations; the variablelevel of RAR-p1 expression between different cell linesmay reflect heterogeneity of dermal fibroblasts in vivo.One of the biopsies used to establish a fibroblast culture(bx4) was obtained from skin adjacent to a basal cell

F F(RA) Fdm dm

F(RA)

f 1 c . i : f s _ t s i

3 6

31

RAR-/3

RAR--/

actin

Fig. 2. RAR expression in dermal fibroblast lines. Theorigin of each primary cell line is given in Table 1. TheNorthern blot was probed successively with the three RARprobes and finally with an a-actin probe. Fibroblast line SCwas treated with (sc+) or without (sc—) 1CT7 M retinoicacid for 24 hours. The positions and approximate lengthsof RAR transcripts are given on the right of the figure (in

K- K+ F- F+

• sRAR-a

RAR-B

RAR-y

Fig. 3. RAR expression incultured keratinocytes (K)and dermal fibroblasts (F)treated with (+) orwithout ( - ) 10~7 Mretinoic acid for 24 hours.Samples (nominally 15 ngtotal RNA per gel track)were analysed together onthe same gel and the filterwas probed successivelywith probes for RAR-or,-/S and -y The highersignal intensity for RAR-aand RAR-y in the RNAsample from retinoic acid-treated keratinocytes is aloading artefact.

dm

1*1RAR-a RAR-jS RAR-7

Fig. 4. RAR expression in a dermal fibroblast line derivedfrom the chest skin of donor dm (Fa™), the original donorof line flO.ll, compared to fs5 fibroblasts treated withretinoic acid [F(RA)] as a positive control.

carcinoma. Myofibroblasts, as defined by their ex-pression of a-actin (Oda et al. 1988), can appear in skinas part of a stromal reaction to tumour, and to assess thecontribution of myofibroblasts to culture heterogeneity,Northern blots were reprobed with an ar-actin probeunder conditions of low stringency, allowing thedetection of both a- and /J-actin. The fibroblast line bx4expressed ar-actin at the highest level (Fig. 2); RNAfrom this cell line also gave a relatively strong signal forRAR-/S. Expression of a^actin was also detectable inRNA from bx7 cells, derived from skin adjacent to amole, but the signal intensity was low and this samplegave no detectable signal for the RAR-p1 probe at theexposures used. If it is true that myofibroblasts expressRAR-p1 at a higher level than fibroblasts, the proportionof myofibroblasts present in the bx7 sample may havebeen too low to allow detection of RAR-p". The flO.llfibroblasts expressed a^actin only at a low level bycomparison with the bx4 and bx7 cells (Fig. 2).

Changes in RAR expression in response to retinoicacid

To investigate the effects of retinoic acid on RARexpression in dermal fibroblasts, the breast-skin cellline (sc) was treated with 10~7 M retinoic acid for 24hours and this resulted in a marked induction of RAR-p1

mRNA, relative to ethanol-treated control cells (Figs 2and 3). This response was studied in more detail usingthe fetal fibroblast cells and the foreskin-derivedfibroblasts fs5. In both fibroblast lines, the retinoic acid-mediated induction of RAR-p1 was rapid, detectablewithin two hours of adding retinoic acid to the cells andgave a 6- to 30-fold induction of RAR-p1 within 16 hours(Fig. 5). Induction of RAR-p1 was detectable with aretinoic acid concentration of 10~10 M and a response50% of maximal was produced with retinoic acidconcentrations in the range 10~10 to 10~9 M (Fig. 6).There was a slight reduction in response at a highretinoic acid concentration of 10~7 M. No changes inthe abundance of RAR-or or RAR-y mRNA wereobserved in response to retinoic acid (Figs 2 and 3).

It has been reported for embryonal carcinoma cellsthat the induction of RAR-p1 in response to retinoic acidis inhibited or attenuated by cAMP analogues (Hu andGudas, 1990; Martin et al. 1990). cAMP is an importantelement of intracellular signal transduction pathways,controlling, for example, the activities of particular

Retinoic acid receptors in cultured skin cells 117

A Time (hours)0 z 4 .-. 16

IffRAR-B

DC # 4 8 12Time (hours)

16

Fig. 5. Time-course of RAR-/3 induction by retinoic acid infoetal dermal fibroblasts. Cells were exposed to 10~8 Mretinoic acid for the times indicated. (A) Northern blotprobed for RAR-/S and then for /3-actin (lower panel) tocorrect for RNA loading. (B) Time course of RAR-/Sinduction in foetal ( • , two separate experiments) andforeskin (fs5, • ) dermal fibroblasts in response to 10~8 Mretinoic acid. The RAR-/3 signal intensity is expressedrelative to control, ethanol-treated cells and is correctedfor RNA loading by reference to the signal intensity for/S-actin.

transcription factors by regulating protein kinases(Karin, 1989), and therefore could be important in theregulation of RAR-p1 expression in fibroblasts. Dermalfibroblasts were exposed to retinoic acid in the presenceor absence of 1 mM dibutyryl-cAMP but this did notinhibit the retinoic acid-mediated induction of RAR-p*(Fig. 7). The fact that others have obtained biologicaleffects in response to the treatment of fibroblasts withdibutyryl-cAMP (Lin et al. 1988; Yoneda et al. 1988)suggests that penetration of the analogue into the cellsshould not be a problem. To verify that dibutyryl-cAMP had some biological effects on human dermalfibroblasts, the cells were grown for three days in thepresence or absence of 1 mM dibutyryl-cAMP. Com-pared to the control, untreated cells, dibutyryl-cAMPsignificantly inhibited proliferation: mean cell numberswere 186,000 (n=18) and 101,000 (n=Yl) per 35 mmdiameter dish for the control and dibutyryl-cAMP-

RAR-fi

•ff«f0 0.1 1 10 100

Retinoic acid (nM)a*oco

.ao

£

CD

I

CC 0 0.1 1 10 100

Retinoic acid (nM)

Fig. 6. Dose-response curve (B) for the induction of RAR-/3 in fs5 fibroblasts in response to treatment with 10~8 Mretinoic acid for 7.5 hours. (A) Northern blot probed forRAR-/S and then /S-actin. In B, RAR-/3 signal intensity isexpressed relative to control, ethanol-treated cells and iscorrected for RNA loading by reference to the signalintensity for /3-actin.

mRA cAMP RA

cAMP

Fig. 7. Changes in RAR-/3 mRNA abundance in foreskin(fs5) fibroblasts treated with retinoic acid (10~8 M) in thepresence or absence of dibutyryl-cAMP (1 mM). Theslightly greater RAR-/3 signal intensity for cells treatedwith both retinoic acid and dibutyryl-cAMP, compared toretinoic acid alone, is a loading artefact and was not seenin repeat experiments, c, control.

treated cells, respectively (Student's *28=8.765,P<0.001).

Intracellular cAMP levels may be increased usingphosphodiesterase inhibitors or activators of adenylatecyclase. Both IB MX, a phosphodiesterase inhibitor,and forskolin, an activator of adenylate cyclase,attenuated the retinoic acid-mediated induction of

118 C. P. F. Redfern and C. Todd

c I RA RA + I

RAR-jS

BRA RA+F

IIRAR-/S

Fig. 8. (A) RNA fromforeskin fibroblasts treatedwith 10~8 M retinoic acid(RA) in the presence orabsence of IBMX (I) andprobed successively withRAR-a, RAR-P, IL1-/3and /3-actin probes, c,control, untreated cells.(B) RNA from cellstreated with 10~8 Mretinoic acid in thepresence or absence of1(T5 M forskolin (F) andprobed for RAR-0.

RAR-/3 relative to untreated control cells (Fig. 8).However, the basal level of RAR-/J expression inunstimulated cells was reduced in response to IBMXand forskolin (Fig. 8). Relative to this reduced level ofRAR-/J expression, RAR-/3 mRNA levels wereincreased by adding retinoic acid to IBMX- or forsko-lin-treated cells (Fig. 8). To see if the effects of IBMXand forskolin on RAR-̂ 3 expression were specific toRAR-/J, blots were reprobed for RAR-a-. The ex-pression of a gene unrelated to RARs was investigatedby also probing the RNA samples from IBMX-treatedcells with an interleukin-1/3 (ILl-fJ) probe. RAR-a"transcript abundance increased approximately 2-fold inresponse to IBMX alone whereas IL1-/J expressiondecreased (Fig. 8). Forskolin had no detectable effecton RAR-a expression. These results show that theeffects of IBMX in reducing RAR-jS expression werenot specific to RAR-/3 and, since RAR-or mRNAabundance was increased in response to IBMX, werenot due to a general reduction in cellular mRNA levels.In addition to the difference between IBMX andforskolin in their effects on RAR-ar expression, therewas a marked difference in their effects on intracellularcAMP concentrations: forskolin increased intracellularcAMP levels 7-fold after six hours incubation (Fig. 9)whereas there was no detectable change in intracellularcAMP in response to IBMX. While we cannot rule outthe possibility that cAMP does affect RAR-/3 ex-pression in forskolin-treated fibroblasts, the lack ofeffect of dibutyryl-cAMP and the effects of IBMX inthe absence of measurable increases in intracellularcAMP suggest that other pathways mediate the changesin RAR-/3 and RAR- a expression in response to IBMXand forskolin.

Fig. 9. Intracellular cAMP concentrations in foreskindermal fibroblasts treated for 6 hours with 1 mM IBMX (I)or 10 fjM forskolin (F) in the presence or absence of 10nM retinoic acid (RA). c, control, untreated cells. Datagiven are the mean and 95% confidence limits for 5replicates at each treatment, and are expressed as pmolescAMP per 25 cm2 flask.

Discussion

RAR expression patterns in cultured skin cellsIt has been suggested that RAR-y is the predominantRAR form expressed in skin (Zelent et al. 1989).However, the present results clearly show that RAR-ais also expressed in cultured keratinocytes and dermalfibroblasts, in agreement with previous studies onwhole rodent and human skin (Rees and Redfern, 1989;Leroy et al. 1991). Overall, keratinocytes and dermalfibroblasts had similar patterns of RAR expression invitro, expressing both RAR-ar and RAR-y but differingin the relative abundance of RAR-ar mRNA. A recentreport in which RAR-a- was described as undetectablein human skin and cultured keratinocytes and fibro-blasts (Elder et al. 1991) is likely to be an artefact due tothe use of hybridization probes too short for thewashing stringency employed.

Although one dermal fibroblast cell line, flO.ll, hada high level of expression of RAR-/J, this appears to bean anomaly and was not a feature of other dermalfibroblasts derived from the same body site of theoriginal donor, or of other dermal fibroblast linesanalysed. Contamination of the original flO.ll primaryculture with an established cell line is unlikely as thepattern of RAR expression in these cells was differentfrom that in other cell lines cultured in the samelaboratory. Thus, the high RAR-̂ S expression in flO.llcells either represents an altered phenotype that hasarisen during culture or other mesenchyme-derivedcells within the dermis may have distinctive patterns ofRAR expression. This raises the important question ofwhether RAR expression patterns of the remaining celllines accurately reflect those of keratinocytes andfibroblasts in vivo. In whole human skin, expression ofboth RAR-a-and RAR-yis detectable, whereas RAR-/3is not, or only at a low level (Rees and Redfern, 1989;

Retinoic acid receptors in cultured skin cells 119

Leroy et al. 1991). Furthermore, recent in situ hybridiz-ation studies (Viallet et al. 1991) show that theexpression of RAR-a- and RAR-y, but not RAR-/5, isdetectable in the dermis and epidermis of mouse skin.We would therefore argue that RAR-expression pat-terns in human dermal fibroblasts and keratinocytesremain at least qualitatively stable in vitro.

Induction of RAR-/3 expression by retinoic acidA major difference between keratinocytes and fibro-blasts is their differential responsiveness to retinoicacid. The mechanism of RAR-/J induction in fibroblastshas not been fully characterized but is likely to resultfrom an RAR-mediated increase in the rate of RAR-Btranscription, as has been shown for hepatocellularcarcinoma cells (de Th6 et al. 1989) and embryonalcarcinoma cells (Martin et al. 1990). A retinoic acidresponse element (RARE) has been defined within theRAR-B promoter (de Th6 et al. 1990; Sucov et al.1990). Similar RAREs are present within the promotersof other retinoic acid-responsive genes (review: Red-fern, 1992). The finding that RAR-B induction infibroblasts occurred at nanomolar retinoic acid concen-trations suggests that the response was indeed receptor-mediated: a response 50% of the maximum within a 0.1to 1 nM range is indicative of a Ku comparable to valuesof 0.2-0.4 nM reported for RAR-a-by Nervi et al. (1989)and Yang et al. (1991).

The effects of cAMP and cAMP-elevating drugs onRAR-/? expression are difficult to interpret. Dibutyryl-cAMP itself had no effect, either on the basal level ofRAR-/J expression or on RAR-B induction by retinoicacid. This implies that the effects of IBMX andforskolin in reducing the constitutive level of RAR-/Jexpression may be mediated by a mechanism other thandirectly via intracellular cAMP. This conclusion issupported by the finding that IBMX, unlike forskolin,did not increase cAMP levels detectably after a six-hourexposure, yet both IBMX and forskolin decreasedconstitutive expression of RAR-B but did not inhibitRAR-/3induction by retinoic acid. Furthermore, IBMXalone increased the abundance of RAR-a, an effect notobserved with forskolin. Although a putative cAMPresponse element (CRE) has been described within theRAR-B promoter (Zelent et al. 1991), there is noevidence that this CRE is functional; in embryonalcarcinoma cells, Martin et al. (1990) have shown thatthe attenuation of RAR-B expression in response todibutyryl-cAMP is mediated at a post-transcriptionallevel.

Retinoic acid in skin developmentThe finding that fibroblasts and keratinocytes differwith respect to changes in RAR-gene expression inresponse to retinoic acid raises two important ques-tions: (1) does RAR-B expression vary in fibroblasts invivo during normal development as a result of regu-lation of intracellular retinoic acid concentration; (2)what are the developmental and phenotypic conse-quences of RAR-/3 expression in fibroblasts in vivo andin vitro!

Detailed in situ hybridization studies of RARexpression during mouse embryogenesis have shownthat the expression of all three receptors is spatiotem-porally regulated (Doll6 et al. 1990; Ruberte et al.1991). For example, RAR-/J shows spatially restrictedpatterns of expression in the mesenchyme of limb budsand facial structures (Dolle' et al. 1990). However,whether or not these spatiotemporal changes in RAR-/Jexpression are brought about through differentialregulation of intracellular retinoic acid concentrations isunknown.

Studies on F9 teratocarcinoma cells suggest thatRAR-B could determine whether the cells differentiateinto parietal or visceral endoderm in response toretinoic acid. In the presence of retinoic acid alone,aggregated F9 cells differentiate into visceral endo-derm, whereas in the presence of retinoic acid andcAMP, conditions in which RAR-/3 expression isattenuated (Martin et al. 1990), F9 cells differentiateinto parietal endoderm (Strickland et al. 1980; Darrowet al. 1990). Unlike F9 cells, fibroblasts do not show aclear differentiation response to retinoic acid. Retinoicacid stimulates fibroblast proliferation and theincreased expression of extracellular matrix proteins,but this may vary with growth conditions in vitro(Varani et al. 1990). Otherwise, the longer-termphenotypic consequences of retinoic acid treatment offibroblasts are unknown.

In vivo studies have recently shown that RAR-/Jexpression is increased in the limb-bud mesenchymeand embryonic mouse-lip dermis in response to retinoicacid (Tickle et al. 1989; Viallet et al. 1991; Noji et al.1991). In the embryonic mouse lip, retinoid treatmentresults in the development of vibrissae follicle buds intoexocrine glands (Hardy, 1968). Such morphologicalchanges appear to be mediated by the dermis (Hardy etal. 1990) and are associated with increased dermalRAR-£ expression (Viallet et al. 1991). Since theinduction of RAR-/3 by retinoic acid in dermalfibroblasts is a rapid event, RAR-/? is likely to beinvolved in the transcriptional regulation of genesspecifying the phenotype of associated epithelial cells.The characterization of long-term phenotypic changesin fibroblasts in response to retinoic acid is offundamental importance for understanding the role ofretinoic acid in dermal-epidermal interactions.

We thank Pierre Chambon, Martin Petkovich and AnneDejean for gifts of RAR probes, J. L. Rees and colleagues forthe skin biopsies, J. L. Rees for preparing RNA from theoriginal fibroblast line, Val Randall for her valuable com-ments on the manuscript and Jane Taylor for the cAMPmeasurements and help with the fibroblast growth exper-iments. This research was supported in part by grants toC.P.F.R. from the Wellcome Trust and the North of EnglandCancer Research Campaign.

References

Boyce, S. T. and Ham, R. G. (1983). Calcium-regulateddifferentiation of normal human epidermal keratinocytes in

120 C. P. F. Redfern and C. Todd

chemically-defined clonal culture and serum-free serial culture. J.Invest. Derm. 81 (Suppl.), 33s^lOs.

Brand, M., Petkovich, M., Krust, A., Chambon, P., de The\ H.,Marchio, A., TloUais, P. and Dejean, A. (1988). Identification of asecond human retinoic acid receptor. Nature 332, 850-853.

Chlrgwln, J. M., Przybyla, A. E., McDonald, R. and Rotter, W. J.(1979). Isolation of biologically active ribonucleic acid from sourcesenriched in ribonuclease. Biochemistry 18, 5294-5299.

Clifford, J. L., Petkovich, M., Chambon, P. and Lotan, R. (1990).Modulation by retinoids of mRNA levels for nuclear retinoic acidreceptors in murine melanoma cells. Mol. Endocrinol. 4, 1546-1555.

Covant, H. A. and Hardy, M. H. (1990). Excess retinoic acid actsthrough the stroma to produce mucous glands from newbornhamster cheek pouch in vitro. J. Exp. Zool. 253, 271-279.

Daly, A. K. and Redfern, C. P. F. (1987). Characterisation of aretinoic acid-binding component from F9 embryonal carcinoma cellnuclei. Eur. J. Biochem. 168, 133-139.

Darrow, A. L., Rkkles, R. J. and Strickland, S. (1990). Maintenanceand use of F9 teratocarcinoma cells. Meth. Enzymol. 190, 110-117.

de The\ H., Marchio, A., TioUais, P. and Dejean, A. (1989).Differential expression and ligand regulation of the retinoic acidreceptor a-and 0 genes. EMBO J. 8, 429-433.

de The", H., Vivanco-Rulz, M., TioUais, P. et al. (1990). Identificationof a retinoic acid response element in the retinoic acid receptor-/?gene. Nature 343, 177-180.

Dolled P., Ruberte, E., Leroy, P., Morriss-Kay, G. and Chambon, P.(1990). Retinoic acid receptors and cellular retinoid bindingproteins I. A systematic study of their differential pattern oftranscription during mouse organogenesis. Development 110,1133-1151.

Edwards, D. R., Parfltt, C. L. J. and Denhardt, D. T. (1985).Transcriptional regulation of two serum-induced RNAs in mousefibroblasts: equivalence of one species to B2 repetitive elements.Mol. Cell. Biol. 5, 3280-3288.

Eichele, G. (1989). Retinoids and vertebrate limb pattern formation.Trends Genet. 5, 246-251.

Elder, J. T., Fisher, G. J., Zhang, Q. Y., Elsen, D., Krust, A.,Kastner, P., Chambon, P. and Voorhees, J. J. (1991). Retinoic acidreceptor gene expression in human skin. / . Invest. Dermatol. 96,425-433.

Feinberg, A. P. and Vogelstein, B. (1983). A technique forradiolabelling DNA restriction endonuclease fragments to highspecific activity. Anal. Biochem. 132, 6-13.

Glguere, V., Ong, E. S., Segui, P. and Evans, R. M. (1987).Identification of a receptor for the morphogen retinoic acid. Nature330, 624-629.

Green, H. and Watt, F. (1982). Regulation by vitamin A of envelopecross-linking in cultured keratinocytes derived from differenthuman epithelia. Mol. Cell. Biol. 2, 1115-1117.

Hardy, M. H. (1968). Glandular metaplasia of hair follicles and otherresponses to vitamin A excess in cultures of rodent skin. J.Embryol. Exp. Morph. 19, 157-180.

Hardy, M. H., Dhouially, D., Torma, H. and Vahlquist, A. (1990).Either chick embryo dermis or retinoid-treated mouse dermis caninitiate glandular morphogenesis from mammalian epidermaltissue. J. Exp. Zool. 256, 279-289.

Hennlngs, H., Michael, D., Cheng, C , Steward, S., Holbrook, K. andYuspa, S. H. (1980). Calcium regulation of growth anddifferentiation of mouse epidermal cells in culture. Cell 19, 245-254.

Hu, L. and Gudas, L. J. (1990). cAMP analogs and retinoic acidinfluence the expression of retinoic acid receptor a, f}, and ymRNAs in F9 teratocarcinoma cells. Mol. Cell. Biol. 10, 391-396.

Karin, M. (1989). Complexities of gene regulation by cAMP. TrendsGenet. 5, 65-67.

Kastner, P., Krust, A., Mendelsohn, C , Garnier, J. M., Zelent, A.,Leroy, P., Stanb, A. and Chambon, P. (1990). Murine isoforms ofretinoic acid receptor y with specific patterns of expression. Proc.Nat. Acad. Sci. U.S.A. 87, 2700-2704.

Krust, A., Kastner, P., Petkovich, M., Zelent, A. and Chambon, P.(1989). A third human retinoic acid receptor, hRAR-y. Proc. Nat.Acad. Sci. U.S.A. 86, 5310-5314.

Latham, J. A. E., Redfern, C. P. F., Thody, A. J. and De Kretser, T.

A. (1989). Immunohistochemical markers of human sebaceousgland differentiation. J. Histochem. Cytochem. 37, 729-734.

Lehmann, J. M., Hoffmann, B. and Pfahl, M. (1991). Genomicorganization of the retinoic acid receptor gamma gene. Nucl. AcidsRes. 19, 573-578.

Leroy, P., Krust, A., Zelent, A., Mendelsohn, C , Garnier, J.-M.,Kastner, P., Dierich, A. and Chambon, P. (1991). Multipleisoforms of the mouse retinoic acid receptor a are generated byalternative splicing and differential induction by retinoic acid.EMBO J. 10, 59-69.

Lin, P., Ahluwalia, M. and Gruenstein, E. (1988). Regulation ofconductive Cl" transport in human fibroblasts. Amer. J. Physiol.255, 552-558.

Marcelo, C. L. and Tomich, J. (1983). cyclic AMP, glucocorticoid andretinoid modulation of in vitro keratinocyte growth. J. Invest.Dermatol. 81 (Suppl.), 64s-68s.

Martin, C. A., Ziegler, L. M. and Napoli, J. L. (1990). Retinoic acid,dibutyryl-cAMP, and differentiation affect the expression ofretinoic acid receptors in F9 cells. Proc. Nat. Acad. Sci. U.S.A. 87,4804-4808.

Minty, A., Caravatti, M., Robert, B., Cohen, A., Daubas, P.,Weydert, A., Gros, F. and Buckingham, M. E. (1981). Mouse actinmessenger RNAs: construction and characterisation of arecombinant plasmid molecule containing a complementary DNAtranscript of mouse a^actin mRNA. J. Biol. Chem. 256, 1008-1014.

Nervi, C , Grippo, J. F., Sherman, M. I., George, M. D. and Jetten,A. M. (1989). Identification and characterisation of nuclear retinoicacid binding activity in human myeloblastic leukaemia HL-60 cells.Proc. Nat. Acad. Sci. U.S.A. 86, 5854-5858.

Noji, S., Nohno, T., Muto, K., Ohyama, K., Aoki, Y., Tamura, K.,Ohsugi, K., Ide, H., Taniguchi, S. et al. (1991). Retinoic acidinduces polarizing activity but is unlikely to be a morphogen in thechick limb bud. Nature 350, 83-86.

Oda, D., Gown, A. M., Vande Berg, J. S. and Stern, R. (1988). Thefibroblast-nature of myofibroblasts. Exp. Mol. Pathol. 49, 316-329.

Ong, D. E. and Chytil, F. (1978). Cellular retinoic acid bindingprotein from rat testis: purification and characterization. J. Biol.Chem. 253, 4551-4554.

Petkovich, M., Brand, N., Krust, A. and Chambon, P. (1987). Ahuman retinoic acid receptor which belongs to the family of nuclearreceptors. Nature 330, 444-450.

Redfern, C. P. F. (1992). Retinoic acid receptors. Pathobtology (inpress).

Redfern, C. P. F., Daly, A. K., Latham, J. A. E. and Todd, C. (1990).The biological activity of retinoids in melanoma cells. Induction ofexpression of retinoic acid receptor-/? by retinoic acid in S91melanoma cells. FEBS Lett. 273, 19-22.

Redfern, C. P. F. and Todd, C. (1988). The effects of retinoic acid onrat epidermal cells in vitro: changes in patterns of proteinphosphorylation in relation to growth and differentiation. Exp. CellRes. 174, 367-377.

Rees, J. L., Daly, A. K. and Redfern, C. P. F. (1989). Differentialexpression of the a-and /3 retinoic acid receptors in tissues of the rat.Biochem. J. 259, 917-919.

Rees, J. L. and Redfem, C. P. F. (1989). Expression of the a-and /Sretinoic acid receptors in skin. J. Invest. Derm. 93, 818-820.

Ruberte, E., Dolle\ P., Chambon, P. and Morriss-Kay, G. (1991).Retinoic acid receptors and cellular retinoid binding proteins: theirdifferential pattern of transcription during early morphogenesis inmouse embryos. Development 111, 45-60.

Sengel, P. (1976). Morphogenesis of Skin. Cambridge UniversityPress, Cambridge.

Sharpe, G. R., Gillesple, J. I. and Greenwell, J. R. (1989). Anincrease in intracellular free calcium is an early event duringdifferentiation of cultured human keratinocytes. FEBS Leu. 254,25-28.

Strickland, S., Smith, K. K. and Marotti, K. R. (1980). Hormonalinduction of differentiation in teratocarcinoma stem cells:generation of parietal endoderm by retinoic acid and dibutyrylcAMP. Cell 21, 347-355.

Sucov, H. M., Murakami, K. K. and Evans, R. M. (1990).Characterization of an autoregulated response element in themouse retinoic acid receptor type /5 gene. Proc. Nat. Acad. Sci.U.S.A. 87, 5392-53%.

Retinoic acid receptors in cultured skin cells 121

Tickle, C , Alberts, B. M., Woipert, L. and Lee, J. (1982). Localapplication of retinoic acid to the limb bud mimics the action of thepolarising region. Nature 296, 564-566.

Tickle, C , Crawley, A. and Farrar, J. (1989). Retinoic acidapplication to chick wing buds leads to a dose-dependentreorganization of the apical ectodermal ridge that is mediated bythe mesenchyme. Development 106, 691-705.

Varani, J., Mitra, R. S., Gibbs, D., Phan, S. H., Dixit, V. M., Mitra,R., Jr, Wang, T., Siebert, K. J., NickolotT, B. J. and Voorhees, J. J.(1990). All-trans retinoic acid stimulates growth and extracellularmatrix production in growth-inhibited cultured human skinfibroblasts. J. Invest. Derm. 94, 717-723.

Viallet, J. P., Ruberte, E., Du Manior, S., Krnst, A., Zelent, A. andDhouallly, D. (1991). Retinoic acid-induced glandular metaplasia inmouse skin is linked to the dermal expression of retinoic acidreceptor-/? mRNA. Develop. Biol. 144, 424^28.

Yang, N., Schule, R., Mangelsdorf, D. J. and Evans, R. M. (1991).Characterization of DNA binding and retinoic acid binding

properties of retinoic acid receptor. Proc. Nat. Acad. Sci. U.S.A.88, 3559-3563.

Yoneda, M., Yamagata, M., Susuki, S. and Kimata, K. (1988).Hyaluronic acid modulates proliferation of mouse dermalfibroblasts in culture. J. Cell Sci. 90, 265-273.

Zelent, A., Krust, A., Petkovich, M., Kastner, P. and Chambon, P.(1989). Cloning of murine a and fi retinoic acid receptors and anovel receptor /predominately expressed in skin. Nature 339, 714-717.

Zelent, A., Mendelsohn, C , Kastner, P., Krust, A., Gamier, J-M.,Ruffeach, H., Leroy, P. and Chambon, P. (1991). Differentiallyexpressed isoforms of mouse retinoic acid receptor /S are generatedby usage of two promoters and alternative splicing. EMBO J. 10,71-81.

(Received 20 November 1991 - Accepted 23 January 1992)