All-Trans Retinoic Acid Induces Cellular Retinol-Binding Protein in Human Skin In Vivo

All-Tvans Retinoic Acid Induces Cellular Retinol-Binding Protein in Human Skin In Vivo Gary J. Fisher, Ambati P. Reddy, Subhash C. Datta, Sewon Kang, Jong Y. Yi,t Pierre Chambon,* and John J. Voorhees Department ot" Dermatology, Utiivetsity of Micliigan Medical Center, Ann Arbor, Michigan, U.S.A.; and *Iiistitut de Chimie Biologique, Facultc de Medecine, Strasbourg Cedex, France

We examined the regulation of cellular retinol-binding protein (CRBP) mRNA and protein expression in human skin in vivo hy All-trans retinoic acid and eill-trans retinol. Treatment of human skin for 24 h with all-trans retinoic acid (0.1%) or all-(rans retinol (1.6%) induced CRBP mRNA 5.5-fold (p < 0.01, n = 10) and 5.7-fold (p < 0.01, n = 5), respectively, compared with skin treated with vehicle or sodium lauryl sulfate (used as an irritant control). In vitro translation of poly A+ RNA from all-trans retinoic acid, all-trans retinol, sodium lauryl sulfate, and vehicle-treated human skin demonstrated that the ohserved increased CRBP mRNA in aU-trans retinoic acid- and aU-trans retinol-treated skin was ahle to direct increased (2.3-2.9-fold) CRBP protein synthesis. Rihoprobe in situ hyhridization revealed that CRBP mRNA was uniformly elevated throughout the epidermis and in dermal cells after aU-trans retinoic acid treatment of human skin. Western analysis revealed that CRBP protein was elevated 3.2-fold (p < 0.01, n = 6) and 3.0-fold (p < 0.01, n = 6) after

all-trans retinoic acid treatment of human skin in vivo for 24 and 96 h, respectively, compared with vehicleand sodium lauryl sulfate-treated skin. In addition, functional CRBP levels measured hy [''H]all-(i'a»ii retinol binding were elevated 1.9-fold (p < 0.01, n = 6) and 3.5-fold (p < 0.01, n = 6) at 24 and 94 h, respectively, after all-trans retinoic acid treatment, compared with vehicle- or sodium lauryl sulfatetreated skin. Gel mobility shift analysis revealed that retinoid receptors in nuclear extracts from human skin formed a specific complex with a DNA probe containing the retinoic acid response element in the mouse CRBP gene. Monoclonal antibodies to nuclear retinoid receptors demonstrated that predominantly retinoic acid receptor-a/retinoid X receptor-a heterodimers bound to the CRBP retinoic acid response element. These data demonstrate that CRBP expres-: sion in human skin in vivo is regulated by exogenous all-(raMs retinoic acid and all-trans retinol. Key tvords: Epidermis/gene regtilationlretinoic acid receptors Iretinol me-, tabolism. J Invest Dermatol 105:80-86, 1995

C

believed to be the major storage form of alUtratts retinol, whereas! retinoic acid is a ligand for nuclear retinoic acid receptors, wbich regulate transcription of specific genes. CIVBP is positively regulated by all-ri-rt(/.s retinoic acid and negatively regulated by glucocorticoids [4-8]. Evidence indicates tbat regulation by aXUlratts retinoic acid is primarily transcriptional and is mediated by nuclear retinoid receptors [9,10]. Tbese receptors are members ofthe steroid/thyroid hormone supergene family of nuclear receptors. Members of this family characteristically bind both specific DNA sequences (response elements) in target gene promoters, and hormones. Hormone binding results in receptor|l activation which, in conjunction with other transcription factors,?! results in enhancement or reduction in tlie rate of gene transcrip-| tion. Functional retinoic acid response elements (RAREs) havefj been identified in the rat [9] and mouse [10] CRBP gene proinot-! ers. These RAREs are composed of two 6—base-pair (bp) repeating motifs separated by 2 bp. There are two families of nuclear retinoid receptors: nuclear retinoic acid receptors (RARs) and nuclear retitioid X receptors (RXRs), each encoded by tliree genes, a, |8, and 7 [11-14]. Accumulating evidence indicates that aW-ttatis retinoic acid binds to and activates RARs, which bind to DNA and funcdon primarily as heterodimers with RJCR [9,15-20]. It has beeti demonstrated in transient transfection studies that RAR/ii

ellular retinol-binding protein (CRBP) is a ubiquitously expressed 15.8-kDa cytosolic protein that binds all-tratts retinol (vitamin A) with high specificity and affinity. A highly homologous protein, CRBPII, which is encoded by a distinct gene, is expressed in intestine [1,2]. CRBP is a member of a large family of fatty acid binding proteins and is believed to be critical for cellular uptake and metabolism of all-d-aiii retinol. AU-fraiis retinol is carried in the circulation bound to retinol-binding protein [3]. AU-ttatts retinol dissociates from retinol-binding protein and presumably, because of its lipophilic nature, enters cells by diffusion through the plasma membrane. In the cytoplasm, ai\-traiis retinol is bound by C l ^ P . In the cell, CKJBP-bound al\-tratts retinol can undergo at least two different metabolic fates: it may be either esterified to retinyl esters or oxidized to aU-tt'cttis retinoic acid. Retinyl esters are Manuscript received September 22, 1994; final revision received February 8, 199.S; accepted for publication March 7, 1995. Reprint requests to: Or. Gary J. Fisher, Department of Dermatology, University of Michigan Medical School, Ann Arbor, MI 48109. Abbreviation: SLS, sodium lauryl sulfate. t Present address: Department of Dermatology, Kangiiam St. Mary's Hospital, Catholic University Medical College, 505 Banpo-dong, Seochogu, Seoul, 137-04(1, Korea.

0022-202X/95/S09.50 • SSDI0022-202X(95)00206-2 • Copyrigbt © 1995 by The Society for Investigative Dermatology, luc. 80

if-

VOL. 105. NO. I JULY 1995

R X R heterodimers bind to and transactivate heterologous reporter constructs containing the CIU3P R A l ^ [9,10]. Retinoic acid is vital to the maintenance of cellular bomeostasis in adult mammalian skin. Dietary all-ttatts retinol deficiency results in altered epidennal differentiation (byperkeratosis), in whicb the skin beconie.s thin and scaly because of reduced cellular growth and increased terminal differentiation [21], Tliis condition can be reversed by administration of either all-trcttts retinol or al\-tratts retinoic acid. Because aW-tratts retinoic acid is formed from all-frrt;;.? retinol, it has been concluded tbat retinoic acid is essential for normal skin function. W e have previously reported that topical treatment of adult human skin with aU-ttattx retinoic acid results in a variety of cellular and biochemical alterations [22—25]. In addition, we and others have previously demonstrated that adult human skin expresses RAR and IOCR transcripts and proteins [14,26-29], Based on these data, we hypothesize that CRBP expression in human skin is regulated by a\l-tratts retinoic acid. To test this hypothesis, we determined C l ^ P mRNA and protein levels in human epidermis after topical al\-ttatts retinoic acid treatment. MATERIALS AND METHODS Materials Reagents for ['^^Pllabeling DNA probes for Northern analysis and gel retardation assays, and for labeling and detecting digoxigeninlabeling cRNA probes for iu situ hybridization, were obtained from Boehringer Mannheim. 01igo(dT)2s-coupled magnetic beads (Dynabeads) were from Dynal, Inc. Wheat germ extract for itt I'itro translation was from Promega. Oligonucieotides were synthesized by tbe DNA synthesis facility. University of Micliigau. ['Hjall-traiis retinol and (a-'"P]-dCTP (3000 Ci/mmol) were purchased from DtiPont New England Nuclear. Human multiple tissue poly A-I- RNA blot was obtained from Clontecb. Htiman CRBP cDNA and antibodies were generously provided by Dr. William S. Blaner (Institute of Human Nutrition, Columbia University, New York). Horseradish peroxidase-conjugated goat anti-rabbit IgG was from Cappel Research Products. Enhanced chemilnminescent detection reagent was from Amersham Corporation. Ammonyx LO (N,N-dimethyl-dodecylamine-N-oxide) was obtained from Fluka Ciiemika. Zeta Probe nylon membrane for Northern analysis was from Bio-Rad. Procurement of Human Skin Biopsy Specimens A\\-traus retinoic acid cream (0.1%), the skin irritant sodium lanryl sulfate (SLS; 2% in aW-tratts retinoic acid cream), and vebicle cream were applied to the btittocks of healthy adult volunteers [22]. In separate experiments, subjects were treated topically with a\\-ttans retinol (1.6% dissolved in 95% ethanol/propylene glycol, 7:3, containing 0.1 mg/ml butylated hydroxytoluene as antioxidant) and vehicle. Treated areas were covered with light-tight plastic wrap to prevent surface loss and oxidation and to enhance penetration of compounds into the skin. After 6, 1 6, 24, or 96 h of treatment, keratonie biopsy specimens were obtained from treated sites as described previously [22]. In addition, full-tbickness 4-mm punch biopsy specimens from all-ttatts retinoic acid aud vebicle-treated skin were obtained for itt sitn hybridization studies. All procedures involving human subjects were approved by the University of Michigan Institutiotial Review Hoard, and ali subjects provided written informed consent. Preparation of Nuclear Extracts From Human Skin Nuclear extracts containing retinoid receptors were prepared from keratonie biopsy specimens as descrihed previously [28,30]. Extracts were aliquoted and stored at —70°C before use. Electrophoretic Mobility Shift Assay Gel mobility shift assays were performed as described by Garner and Revzin |31] and Smidi et al [10]. Double-stranded oiigodeoxynucleotides containing the 1?J\RE (shown in bold) in the mouse CRBP promoter (5'-GATCCAGGTCAaaAAGTCAGACAC-3') [10] were used as probe. For antibody supersbifts. we used monoclonal antibodies to RAR-a, AB9a(F) (9ff-9A()) [32]; RAR-y, AB4-)'hF (4y-7All) [33]; and RXR-a, 4RX-1F0. Radioactive proteinDNA complexes in tbe gels were visualized by Pbospliorlmager (Molecular Dynamics). Northern Analysis Total RNA was isolated from keratonie biopsy specimens by guanidinium hydrochloride lysis and ultracentrifugatiou, as described previously [34]. Equal quantities (20 /ag per lane) of total RNA were separated electrophoretically in 2.2 M formaldehyde, 1% agarose gels, transferred to nylon membranes, aud liybridized with randomly primed [^^P]labeled cDNA probes for CRBV or 36B4 (36B4 encodes a ribosomal protein and was used as an internal control). Visualization and quantitation of mRNA levels were performed using a Phosphorlmager.

CRBP INDUCTION IN HUMAN SKIN

81

In Viti'o Translation Oligo(dT)25-labeled Dyiiabeads were used to purify poly A+ RNA from total RNA derived from liunian skin treated with aW-trtins retinoic acid (0.1%), a\\-tratts retinol (1.6%), SLS (2%), or vehicle before biopsy. Poly A+ RNA (2 /ig) from each treatment group was translated tising the wheat germ extract system according to the manufacturer's instructions. The reaction mixtures containing the translated protein products (100 fig protein) were separated by sodium dodecylsulfatepolyacrylamide gel electrophoresis (SDS-PAGE), transferred to nitrocellulose membrane, and subjected to Western blot analysis, as described below. Western Analysis CRBP protein levels were detemiined in bigh-speed stipernatants, prepared as described below, from vehicle-treated. .SLStreated, and all-frdii.? retinoic acid-treated skin. Samples (100 fig) were separated by 15'M) SDS-PAGE, transferred to nitrocellulose membrane, and incubated for 16 h with polyclonal rabbit anti-C!U3P antibodie.s [35]. Immunoreactive C l ^ P was visualized by enhanced cheiiiiluminescence detection. Band intensities were determined by laser densitometry. Measurement of CRBP Levels by Ligand Binding CRBP levels were determined in high-speed supernatants from vehicle-treated, SLS-treated, and all-r(vin.« retinoic acid-treated skin samples by measurement of ['Hlallttans retinol binding. Keratome biopsy specimens were ground to a fine powder in liquid nitrogeti using a mortar and pestle; homotrenized in 20 lnM Tris, pH 7.4, containing 1 niM dithiothreitol and 1 niM phenylmethylsulfonyl fluoride, using a glass homogenizer; and centrifuged at 100,000 X g at 4°C for 1 h. The supernatant was assayed for ['^H]all-rrrt».s" retinol binding. [•^H[all-fivjiK retinol (50 pmol, 200,000 cpm) in ethanol was transferred to microcentrifuge tubes, and tbe solvent was evaporated witb a stream of nitrogen gas. Dried a\\-traiis retinol was dissolved in 2 f^l Ammonyx LO, followed by addition of 848 /nl assay buffer (20 niM Tris-HCl, 10% glycerol, 10 inM thioglycerol, pH 7.5) and 150 /xl sample (100 /Ltg protein). After incubation ofthe mixtures in the dark at 4°C for 16 ll, 0.5 ml 3% dextran-coated charcoal [36] was added for 30 min on ice to adsorb free a\\-traiis retinol. Charcoal was removed hy centrifugation for 5 min at 10,000 X g, and aliquots ofthe supernatants, containing ['Hlall-rrrtiK retinol bound to C1U3P, were counted in a liquid scintillation counter. Ill Situ Hybridization The bluescript plasmid containing human CRBP cDNA was linearized with Hindlll or Pstl for sense and anti-sense probes, respectively. Digoxigenin-containiug sense and anti-sense cW^A probes were synthesized by iu vitto transcription using T3 and T7 RNA polymerases. Frozen sections (5 /nm tliick) were tbaw-mounted onto slides coated with 3-aiiiinopropyl-trietlio.xysilaiie, fixed with 4% (w/v) paraformaldehyde, treated with 0.2 N HCl for 5 min at room temperature, and subjected to proteinase K digestion (5 /ng/ml) at 37''C for S min, followed by a second fixation witb 4"/a paraformaldehyde. After rinsing in phosphatebuffered saline, the slides were acetylated with 0.25% (v/v) acetic anhydride in 0.1 M triethanolamine, pH 8.0, rinsed in nuclease-free water, and equilibrated in 10 niM Tris-HCl (pH 7.5), 50% formamide, 0.6 M NaCl, 1 mM ethyleuediamiuetetraacetic acid, 10 niM ditbiothreitol, and 50 /xg/ml heparin. Prehybridization was performed for 1 h at 50°C iu equilibration buffer, to which 10% polyethylene glycol 7500 and 1 X Denhardt's solution were added. Hybridization was performed for 18 li at 50°C in prehybridization buffer, to wbich 0.5 mg/ml denatured salmon sperm DNA, 0.5 mg/nil yea.st tl^NA, and 1 /xg/ml sen.se or anti-sense cRNA probes were added. Sections were washed once after hybridization in 2 X sodium citrate/sodium chloride buffer (SSC) for 30 min at 50°C, digested with RNase A (20 /ng/ml) for 30 miu at 37°C, and washed for 30 min in 2 X SSC/50% formamide/10 mM ditliiotbreitol at 37°C and twice in 1 X SSC/14 niM didiiodireitol/0.067% sodium pyrophosphate at 50°C for 30 min. Hybridization signals were detected immunoliistocliemically using alkaline-phospliatase—conjugated anti-digoxigenin antibody. Protein Assay Protein concentration was measured by the Pierce/BCA metbod using bovine serum albumin as standard [37]. Statistical Analysis Comparisons of CRBP induction between treatment groups were perfonned with eitber the paired t test or repeatedmeasures analysis of variance witb the Tuke)- procedure for multiple group comparisons. Summary statistics are expressed as mean ± SEM. Data analysis was performed using the Michigan Data Analysis System (MIDAS), a statistical software package developed at the Center for Statistical Consultation and Research at the University of Michigan.

INSULTS Induction of CRBP Gene Expression by All-Trans Retinoic Acid and All-Traiii Retinol in Human Skin To investigate the regulation of CIVBP gene expression in human skin, adult subjects were treated topically with ail-ttatts retinoic acid cream

82

FISHER ET AL

THE JOUK.NAL OF INVESTIGATIVn DERMATOLOGY

24 Hours Figure 1. All-fraHS retinoic acid induces CRBP mRNA in human skin. Adult human subjects were treated topically with vciiicle, SLS (2%), and sW-lraiis retinoic acid (RA, 0.1%) (each subject was treated on separate areas with each of tbe tbree agents). Keratome biopsy specimens were obtained after 24 h {open hiit's) or 96 h (shaded bars) of treatment, and steady-state levels of CRBP mRNA and 36B4 mRNA (used as an internal control) were analyzed by Northern blot. Top, representative Northern blot (20 /LLg total RNA per lane) results from three subjects; V, vehicle; S, SLS; R, aW-fraus retinoic acid. Bars indicate mean ± SEM (n = 10) of ClUJP mlOsIA band intensities normalized to those of 36B4 niRNA, determined by Phosphorlmager. For 24 h treatment (n = 10): 1*^ versus vehicle, p < 0.01; RA versus SLS, p < 0.01; vehicle versus SLS, not significant. For 96 h treatment (n = 10): RA versus vehicle, p < 0.05; RA versus SLS, p < 0.05; vehicle versus SLS, not significant.

96 Hours

Subj 1 Subj 2 Subj 3 I

II

II

Subj 1 Subj 2 Subj 3 I

VSRVSRVSR 0.7 kb -

Treatment

I

II

II

1

V S R V S R V S R

il

CRBP

(Control Gene) <

6

1' m 4 CO

3 "i 2 1 Vehicle

(0.1%), SLS (2%), or vehicle for 6, 24, and 96 h and determined steady-state CRBP mlOvIA levels (normalized to the internal control 36B4 gene transcript) by Northern hlot. Because all-/ra(;i retinoic acid treatment causes some skin "irritation" (i.e., the skin becomes pink), treatment with the irritant SLS, which causes a similar degree of pinkness [22], served as a control for the efFects of irritation. In all skin samples, a single 0.7-kb CRBP transcript was detected. After 6 h of treatment, the level of C l ^ P mRNA was similar in al\-traus retinoic acid-treated, SLS-treated, and vehicletreated skin (data not shown), hi contrast, after 24 h of a\l-trans retinoic acid treatment, Cm3P mliJSIA was increased 3.2-fold (p < 0.01, n = 10) compared with SLS-treated skin and 5.5-fold (p < 0.01, n = 10) compared with vehicle-treated skin (Fig 1). There was no difference between SLS-treated and vehicle-treated skin. Induction of C1U3P by a\]-traus retinoic acid observed at 24 h persisted for at least 96 h. After 96 h of treatment, C1U3P mRNA was elevated 3.5-fold (p < 0.05, n = 10) compared with SLStreated skin and 2.6-fold (p < 0.05, n = 1(1) compared with vehicle-treated skin (Fig 1). We next examined whether al\-tiaus retinol, the ligand for CRBP, could induce C l ^ P gene expression in htiman skin. Adult subjects were treated topically with ail-lraiis retinol (1.6%) or vehicle for 24 h, and Cm3P mRNA was quantified hy Northern blot as described above. AW-lraiis retinol induced CRBP mRNA 5.7-fold (p < 0.01, n = 5), compared with vehicle (Fig 2). The magnitude of this induction was similar to that observed with an-traiis retinoic acid treatment (Fig 1). CRBP In Situ Hybridization in Human Skin Having demonstrated induction of CRBP mRNA by all-/raiw retinoic acid, we next performed iii situ hybridization to localize CRBP transcripts in human skin treated for 24 h with all-traiis retinoic acid (0.1%) or vehicle. In vehicle-treated skin, CRBP mRNA expression was detectable at a low level throughout the epidermis, with somewhat more intense, patchy expression in the lower layers of epidermis (Fig 3.4). No significant C I ^ P expression was detected in the dermis. In aW-traiis retinoic acid-treated skin, C l ^ P mlVNA expression was elevated throughout the epidermis in a difftise manner (Fig 3B). CIUJP transcripts were also present in perivascular inflammatory cells, endothelial cells, spindle-shaped cells in the

SLS

RA

Subj1 Subj2Subj3Subj4 V R V R V R V R 18S-

Mil (Control Gene) i

Vehicle

Retinol

Figure 2. All-trans retinol induces CRBP mRNA in human skin. Adult human subjects were treated topically with all-r/rt((.« retinol (1 .b"/a) and vehicle for 24 h. Keratome biopsy specimens were obtained, and steadystate levels of CRBP mRNA and 36B4 mRNA (used as an internal control) were analyzed by Northern blot. Top, representative Northern blots of| lysIA (20 /Ltg total RNA per lane) from three subjects; V, vehicle; R, all-(ra/i.s retinol. liars indicate mean ± SEM fold change relative to veliicle treatment of CIU3P mRNA band intensities, normalized to those of 36B4 mRNA, determined by Phosphorlmager. Vehicle versus all-(raii,« retinol: p < 0.01, n = 5.

CIU3P INDUCTION IN HUMAN SKIN

VOL. 105. NO. 1 JULY 1995

83

not vary under the conditions of treatment (data not shown). In contrast to CRBP, as expected, the level of nuclear l^LAR-yl protein synthesized itt t'itro did not differ among the four treatment groups (Fig 4). These data indicate that increased CRBP nil^JMA levels in aW-tratts retinoic acid—treated and all-fivi/i.« retinol—treated skin can direct increased CRBP protein synthesis. Measurement of CRBP Protein Levels and Ail-Trans Retinol Binding Activities in All-Trans Retinoic Acid-Treated, SLS-Treated, and Vehicle-Treated Human Skin The above data demonstrate that all-fra;;,'; retinoic acid and a\\-tratts retinol induce Cm3P mlVNA in human skin in uit'o and that the increased CRBP mRNA is capable of directuig increased CRBP protein synthesis itt vitro. We next examined whether the observed increase in CRBP miySIA after a\\-ttatts rednoic acid treatment of human skin itt vivo resulted in increased CRBP protein content and increased all-fra».? retinol binding activity. Western analysis revealed that after 24 h of a\\-tratts retiiioic acid treatment, CRBP protein was elevated 3.0-fold (p < 0.01, n = 6) compared with SLS-treated skin and 3.2-fold (p < 0.01, n = 6) compared with vehicle-treated skin (Fig 5). Similar results were obtained after 96 h of a\\-tratts retinoic acid treatment. Cl^JBP protein was increased 2.8-fold (p < 0.01, n = 6) compared with SLS-treated skin and 3.0-fold (p < 0.01, n = 6) compared with vehicle-treated skin (Fig 5). There were no differences in CRBP protein levels, at either 24 or 96 h of treatment, between SLS- and vehicle-treated skin (Fig 5). We next determined all-f)V7);.? retinol binding, a measure of functional CIVBP protein, in supernatants from a\\-tratts retinoic

24 Hours Treatment

c

•

I

Veh SLS RA ROL

Figure 3. Localization of CRBP mRNA in vehicle-treated and alltrans retinoic acid—treated human skin by in situ hybridization. Adult

CRBP

human subjects were treated with veliiclc and wW-traits retinoic acid (0.1%) for 24 h. Full-thickness 4-mm specimens were ohtained, enihcdded in OCT, and frozen in liquid nitrogen. Ci'yostat sections (5 /xm) were hyhridized widi sense or anti-sense CRBP rihoprohes, as descrihed in Matciials and Methods. A, vehicle-treated skin with anti-sense prohe; B, all-(ra(« retinoic acid-treated skin with and-sense prohe; C, aH-trans retinoic acid—treated skin with sense prohe. Results are representadve of four experiments that yielded similar results.

QC

langeJ C R B P ;

In Vitro Translation of CRBP mRNA From All-TraHi Retinoic Acid—Treated, All-Trans Retinol-Treated, SLSTreated, and Vehicle-Treated Human Skin To determine whether increased CRBP mRNA observed in all-fivi/;.« retinoic acid—treated and a\\-ttatts retinol—treated skin could direct increased CRBP protein synthesis, we performed itt vitro translation of equal amounts of poly A+ PJMA isolated from human skin treated with all-tratts retinoic acid (0.1%), aW-ttatts retinol (1.6%), SLS (2%), and vehicles for 24 h. Levels of CRBP protein and RAR-yi protein (used an internal control) synthesized itt i'itro were determined by Western blot. The level of Cm3P protein synthesized with poly A+ l^UMA obtained from aW-ttatts retinoic acid-treated and aW-ttatts retinol-treated skin was elevated 2.3-fold and 2.9-fold, respectively, coinpared with that obtained with vehicle-treated skin (Fig 4). There was no difference in the level of C l ^ P protein synthesized with poly A+ IUMA isolated from vehicle-treated versus SLS-treated skin (Fig 4). To examine the specificity ofthe observed increase in C l ^ P protein synthesis with all-Ir
3.02.52.01.51.0-

o 0.5old

dermis, cells in the hair follicles, sweat glands, and ducts. Hybridization of a\\-trittts retinoic acid—treated skin with a sense CRBP cRNA probe (Fig 3 Q yielded no detectable signal, indicating that hybridization of the anti-sense probe was specific.

RAR-7

0.0Veh

SLS

ROL

Figure 4. Increased 111 vitro translation of CRBP mRNA in all-trans retinoic acid-treated and all-fraiw retinol-treated human skin. Poly A+ RNA was isolated from human skin treated for 24 h witii vehicle (Veh), SLS (2%), all-rro».< retinoic acid (RA, 0.1%), and all-frniK retinol (ROL, 1.6%), and transl.ited in vitro (2 ^Lg poly A+ RNA) using wheat gemi extract. Translation products (100 ;xg total protein) were resolved by SDS-PAGE and immunoblotted with andhodies specific to CIU3P and l^U\ft-7. Bar graph displays relative CR13P/RAR-7 hand intensities, determined hy laser densitometry. Data are representative of two experiments that yielded similar results.

84

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

FISHER l-r AL

1 2 3 A

4 5 6 7 8 9 10

•— — — —

^ ^

O)

c c

^ CRBP

ffl

B

^ — . —^^^^

I a>

c

3.5o 3.02.5OL CO CC 2.0O (U 1.5O) CO 1.0O 0.52 o u- 0.0

3-

o

T

D> (0

6o u.

Vehicle

SLS

RA

Figure 6. All-trans retinoic acid induces CRBP binding activity in

Vehicle

SLS

RA

Figure 5. All-traiis retinoic acid induces CRBP protein in human skin in vivo. Adult human subjects were treated topically with vehicle, SLS (2'yii), and all-rraiis retinoic acid (RA, O.V/n) (each subject was treated with eacli of tlie three agents) for 24 h (open bars) or 96 h (shaded bars). CRUP protein levels in high-speed supernatants were determined hy Western blot. Top, Western blot results from three individuals treated with vehicle (lanes 1,4,7), SLS (hm:( 2,5,8), and retinoic acid {hues 3,6,9) for 24 h (A) and 96 h (B). Lane 10 is purified rat testes CRBP standard. Bars indicate mean ± SEM fold change relative to vehicle treatment of CRBP protein hand intensities, determined by laser densitometry. For 24 h treatment (n = 6): RA versus vehicle, p < 0.01; WA versus SLS, p < 0.01; vehicle versus SLS, not significant. For 96 h treatment (n = 6); RA versus vehicle, p < 0.01; RA versus SLS, p < 0.01; vehicle versus SLS, not significant.

acid-treated (0.1%), SLS-treated (0.2%), and vehicle-treated skin. After 24 h of aW-trans retinoic acid treatment, aW-trans retinol binding activity was elevated 2.0-fold (p < 0.01, n = 6) compared with SLS-treated skin and 1.9-fold (p < 0.01, n = 6) compared with vehicle-treated skin (Fig 6). Similar results were obtained after 96 h of aW-tiims retinoic acid treatment. AW-trans retinol binding activity was increased 2.5-fold (p < 0.01, n = 6) compared with SLS-treated skin and 3.5-fold (p < O.Ol, n = 6) compared with vehicle-treated skin (Fig 6). There were no difFerences in a\{~trans retinol binding activity, at either 24 or 96 h of treatment, between SLS- and vehicle-treated skin (Fig 6). Binding of Retinoid Receptors in Human Skin to the RARE in the CRBP Gene The above data demonstrate that a\\-trans retinoic acid induces CK-UP tnRNA and protein expression in human skin. To delineate further the mechanism of CRBP induction, we investigated whether retinoid receptors in human skin can bind to the I^JVRE in the Cl^BP gene. The liJVI^ element in the CIVBP gene in mouse and rat is composed of direct repeats of two hexameric half sites spaced by 2 bp (DR-2). This motif binds nuclear retinoid receptors and is responsible for stimulation of CRBP gene transcription by aW-trans retinoic acid [9,1 Oj. Because the human CRUP gene promoter has not been seqnenced, we examined the interaction of human skin retinoid receptors with the RAI?^ in the mouse C1U3P gene. Electrophoretic mobility sliift assays revealed that nuclear extracts from human skin formed a retarded complex with a synthetic double-stranded [' PJdeoxynucleotide probe containing the mouse CRBP gene RARE (Fig 7, tane 1). Addition of 100-fold excess unlabeled probe effectively reduced the intensity of the complex, indicating that it was specific (Fig 7, laiic 2). To identify which retinoid receptors in skin bind to the CRBP RARE, we performed

human skin in vivo. Adult human subjects were treated topically with vehicle, SLS (2%), and all-rrnii-v retinoic acid (RA, 0.1%) (each subject was treated with each of the three agents) for 24 or 96 h. ['HJretinol hinding by C[^J3P was determined in high-speed supernatants as described in Mntcrhis and Methods. Bars indicate mean ± SEM fold change relative to vehicle treatment of CIU3P hinding activity. Values for vehicle-treated skin were 0.17 ± 0.01 and 0.09 ± 0.02 pmol retinol bound/mg protein at 24 and 96 h, respectively. For 24 h treatment (n = 6); RA versus veliicle, p < O.OI; RA versus SLS, p < 0.01; vehicle versus SLS, not significant. For 96 h treatment (n = 6): RA versus vehicle, p < 0.01; IVA versus SLS, p < 0.01; veliicle versus SLS, not significant.

supershift gel retardation assays with monoclonal antibodies specific for KAR.-a, RAR-)3, RAR-7, and RXR-cj. A faint but detectable supershifted complex was obtained with antibody to RAR-a (Fig 7, lane 3). Intense supershifted complexes were obtained with antibodies to l^J^R-7 (Fig 7, lane 4) and RXR-a (Fig 7, lane 5). Antibodies to RAR-y and R X R - a together double supershifted essentially all of the retarded complex (Fig 7; compare lanes. I, S). No shifted complex was observed with antibody to RAR-j3 (data not shown). These data demonstrate that RAR-7 and RXR-a are the predominant nuclear retinoid receptors in human skin that bind to the CIU3P RARE. DISCUSSION This study demonstrates that topical application of aW-trniis retinoic acid or all-trans retinol induces CRBP expression in adult human skin. These results are consistent with previous studies demonstrating induction of CRBP mRNA by aW-traiis retinoic acid and/or a\[-trans retinol in the lung of retinol-sufBcient [7] and -deficient [5,38] rats, cultured Sertoli cells [39], and P19 embryonal carcinoma cells (40]. Although the mechanism of CRBP induction in human skin by retinoids remains to be elucidated, one likely possibility is that it involves nuclear retinoid receptor-mediated transcriptional activation of the CRBP gene. Although the promoter of the human CI?J3P gene has not been sequenced, both the mouse [10] and rat [9] ClUiP genes contain RAREs that specifically bind nuclear retinoid receptors and confer a\\-tratis retinoic acid inducibility on CRBP promoter reporter gene constructs. In gel mobility supershift assays with nuclear extracts from human skin, we found that WJVR-y/lOCR-a heterodiniers, which are the major retinoid receptors in human skin [28], bound to the mouse CRBP R A l ^ . These data are compatible with the involvement of RAR7/lOCR-a heterodimers in mediating CBJiP gene induction by all-fra»i- retinoic acid in human skin. More direct evidence for tliis mechanism must await sequencing of the human C l ^ P gene. We observed that all-trans retinol, in addition to aU-tfatts retinoic acid, induced CRBP expression in human skin. Induction of CRBP by a\\-trans retinol presumably reflects its conversion to all-trniis retinoic acid (reviewed in [41]). We [42] and others [43,44] have demonstrated the capacity of cultured human keratinocytes to synthesize a\\-trans retinoic acid from all-^r
CRBP INDUCTION IN HUMAN SKIN

VOL. 105, NO. 1 JUEY 1995

ct d •

Antibody ^ I

^

d

I

I

dOC I

oc oc QC OC + < < X X

85

induction of Cl-UiP by exogenous all-trans retinoic acid probably reflects, at least in part, a physiologic mechanism to promote storage of all-trtins retinol as retinyl esters, under conditions of perceived all-trans retinol excess. Induction of CRBP levels by exogenous all-frrt(« retinoic acid may therefore have the potential to modulate metabolism of endogenous aIl-(ra/).« retinol. This issue is currently under investigation in our laboratory.

OC OC OC OC

This study was supported in pan hy the Babcock Fund for Dermatological Research. We acknowledge Mr. William Dnnkle and Ms. Sujatha Venkatapnrain for expen technical assistance and Mr. Ted Hamilton for e.xpert statistical analysis.

2 •

INFERENCES

1

J lib*

12

3 4 5 6

Figure 7. Nuclear retinoid receptors in huinan skin bind to the CRBP RARE. Binding of retinoid receptors in nuclear extracts from adult human skin to the R A l ^ in the mouse CI?J3P gene was assessed hy the gel sliift assay, as descrihed in Materials and Methods. Triangle 1, lane 1, specific

complex of CRJ3P-RAI^LE with human skin nuclear retinoid receptors; lane 2, competition of specific DNA binding with lOO-fold excess unlaheled probe; triangle 2, lane 2, supershifted complex with monoclonal antihody to RJ\R-a; triangle 2, lane 3, supershifted complex with monoclonal antihody to IVAR-7; triangle 4, lane 4, supershifted complex with monoclonal antibody to RXR-a; triangle 3, lane 5, double-supershifted complex with antibodies to I^JVR-y and RXR-a. Restiits are representative of three experiments that yielded similar results.

deinonstrated in cultured human keratinocytes that transcriptional activation by all-(ra//.s retinol of a K>AR-dependent reporter gene, under the control of a RARE, is inhibited by citral, which blocks the metabolic conversion ofall-tratis retinol to all-trans retinoic acid [42]. Thus, we hypothesize that all-trans retinol-induced CRBP expression, and likely the majority of other biologic responses of the skin to all-trans retinol, are mediated through nuclear retinoid receptors, subsecjuent to the metabolic conversion of all-fraii.s retinol to all-trans retinoic acid. Accumulating evidence indicates that C l ^ P plays a critical role in cellular uptake and metabolism of all-/ra/w retinol (reviewed in [45,46]). The ratio of apo-CRBP to holo-CIU3P appears to be a key determinant in regulating the relative rates of retinyl ester synthesis and hydrolysis. Thus, under conditions of relative retinol insufficiency (i.e., a p o - C l ^ P greater than holo-CRBP), net all-frfliw retinol esterification is reduced in favor of all-trans retinol oxidation to all-trans retinoic acid, which also uses holo-CRBP as substrate. Because aU-ttans retinol is the precursor of all-trans retinoic acid.

1.

Li E, Denimer LA, Sweetser DA, Ong DE, Gordon JI: Rat cellular retinolbinding protein 11; use ot';i cloned cDN A to define its primary' structure, tissue specific expression and developmental regulation. Proc Natl Acad Sci USA 83:.S779-5783, 1986 2. Levin MS, Li E, Ong DE, Gordon JI: Comparison of the tissue specific expression and developmental regulation of two closely linked parent genes encoding cytosolic retinol-hinding proteins. J Biol Chem 262:7118-7124, 1987 3. BlomhoffR, Green MH. Berg T. Nonim KR: Transport and storage of vitamin A. Sriciia'250:399-404, 1990 4. Blaner WS, Das K, Mertz JS, Das SR, Goodman DS: EfFects of dietary- retinoic acid on cellular retinol und retinoic acid-binding protein levels in various rat tissues. / Lipiil Res 27:1084-1088. 1986 5. Haq RU, Chytil F: Retinoic acid rapidly indnces king cellular retinol-binding protein inRNA levels in retinol deficient rats. liiocltem liiiipbys Res Conimtitl 1.56:712-716, 1988 6. Kato M, Blaner WS, Mertz JR, Das SR, Goodman DS: Influence of retinoid nutritional status on cellular retinol and retinoic acid binding protein concentrations in various rat tissues../ Biol Cbem 260:4832-4838, 1985 7. Rush GM, Haq RU, Chytil F: Opposing effects of retinoic acid and dexamethasone on cellular retinol binding protein ribonucleic acid levels in tlie rat. Ijiilocriiiohxy 129:7(15-709, 1991 8. Wei LN. Mertz JR, Goodman DS, Nguyen-Huu C: Regulation of the cellular retinoid binding proteins and their messenger ribonucleic acids during pl9 embryonal carcinoma cell differentiation indnced by retinoic acid. Mol Endocrinol 1:526-534, 1987 9. Husmann M, Hoffmann B, Stnnip GD, Cbytil F, Pfahl M: A retinoic acid response element from the rat CRBIM promoter is activated by an RAR/RXR heterodinier. Biocbem Biophys Res Commun 187:1558-1564, 1992 10. Smith WC, Naksliatri H, Leroy I', ReesJ, Cbambon 1': A retinoic acid response element is present in the mouse cellular-retinol binding protein 1 (CRBPI) promoter. EMliOJ 10:2223-2230, 1991 1 1. Leid M, Kastiier P, Chambon P: Multiplicity generates diversity in the rednoic acid signalling patbways. Trends Biochem Sci 17:427-433, 1992 12. Leid M: Role ot nuclear retinoic acid receptors in the regulation ot gene expression. In: 151onihoff R (ed.). Vitamin .4 (')( Heahh mid Disease. Marcel Dekker, New York. 1994. pp 189-238 13. Chambon P: The retinoid signaling pathway: molecular and genetic analysis. Semili Cell Biol 5:115-125, 1994 14. Mangelsdorl" DJ, Borgmeyer U, Hayman KA, Zhou JY. Ong ES, Oro AE, Kakizuka A, Evans RM: Characterization of three RXR genes that mediate the action of 9-cis retinoic acid. Genes Dev 6:329-344, 1992 15. Durand B, Saunders M. Leroy P, Leid M. Chambon P: All trans and 9-cis retinoic acid induction of CRABIMI transcription is mediated by l ^ ^ R / R X R heterodimers bound to DRl aud DR2 repeated motifs. Cell 71:73-85, 1992 16. Hermann T, Hoffmann B, Zhang X-K, Tran P, Ptahl M: Heterodimeric receptor complexes determine 3,5,3'-triiodothyrouine and retinoid signaling specificities. Mol Bidocriiiol 6:1 153-1 162, 1992 17. Kliewer SA, Umesono K, Mongelsdorf DJ, Evans RM: Retinoid X receptor is an auxiliary protein for thyroid hormone and retinoic acid receptors. Nanirc 355:446-449. 1992 18. Leid M, Kastner P, Lyons R. Nakshatri, Saunders M, Zacharewski T, CbenJ-Y, Staub A, Gamier J-M, Mader S, Chambon P: Purification, cloning and lOCR identity of the HeLa cell factor with which RAR, TR heterodimerizes to bind target sequences efficiently. Cell 68:377-392. 1992 19. Yu VC, Delsert C, Anderson B, HollowayJM, Devary OV, Naar AM, Kim SY, Boutin JM, Glass CK, Rosenfeld MG: lOCRbeta: a coregulator that enhances binding of retinoic acid, thyroid hormone and vitamin D receptors to tlieir cognate response elements. Cell 67:1251-1266, 1991 20. Zhang X-K. Hoffinann B, Tran P, Graupner G, Pfahl M: Retinoid X receptor interacts with nuelear receptors in retinoic acid, thyroid hormone and vitamm D signalling. Nrtrrm-355:441-448, 1992 21. Bollag W: Vitamin A and retinoids: from nutrition to pharniacotlierapy in dermatology and oncology. Lancet 1:860-863. 1983 22. Fisher GJ, Esmann J, Griffiths CE, Talwar HS, Duell EA, Hammerberg C, Elder IT, Finkel LJ. Karabin GD, Nickoloff BJ, Cooper KD, Voorhees JJ: Celltllar, immunologic and biochemical characterization of topical retinoic acid-treated human skin. / hm-st Dennatol 96:699-707, 1991 23. Fisher GJ, Tavakkol A, Griffiths C, ElderJT, Zhang Q-Y, Finkel LJ, Danielpour

86

24.

25. 26. 27.

28.

29. 30.

31.

32. 33.

34.

FISHER ET AL

D, Glick AB, Higlcy H, Ellingsworth L, VoorheesJJ: DifFerential modulation of transforming growtli factor-/31 expression and muciii deposition by retinoic acid and sodium lauryl sulfate in human skin. / Iiircst Dermatol 98:102-108, 1992 GrifBtks CEM, Rosenthal D, Leach K, Wang TS, Finkcl LJ, Yuspa SH, Voorhees JJ, Fisher GJ: Short-term retinoic acid treatment increases in vivo, but not in vitro, epidermal transglutaniinase-K enzyme activity and immunoreaetivity. _/ hwfst Dermatol 99:283-288, 1992 Astrom A, Tavakkol A, Pettersson U, Cromie M, Elder JT, Voorhees JJ: Molecular cloning of two human cellular retinoic acid-binding proteins (CRABP). J Biol Chem 266:17662-17666, 1991 ReesJL, Redfcm CP: Expression ofthe alpha and beta retinoic acid receptors in skin.J luvesi Dermatol 93:818-820, 1989 Tavakkol A, Griffiths CEM, Keane KM, Palmer RD, Voorhees JJ: Cellular localization of mRNA for cellular retinoic acid binding protein-H and nuclear retinoic acid receptor-Ti in retinoic acid treated human skin, / Invest Dennatol 99:146-150, 1992 Fisher GJ, Talwar HS, Xiao J-H, Datta SC, Reddy AP, Gaub M-P, RochetteEgly C, Chambon C, Voorbees JJ: Immunoiogical identification and functional quantitation of retinoic acid and retinoid X receptor proteins in human skin.J Biol Chem 269:21)629-20635, 1994 Elder JT, Astrom A, Pettersson U, Tavakkol A, Griffiths CE, Krust A, Kastner P, Chambon P, VoorheesJJ: Differential regulation of retinoic acid receptors and binding proteins in buman skin.J Invest Dermatot 98:673—679, 1192 Dignam JD, Lebovitz RM. Roeder RG: Accurate transcription initiation of RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res 11:1475-1489, 1983 Gamer MM, Revzin A: A gel retardation method for quantifying tbe binding of proteins to specific DNA regions: application to components ofthe Uscherichia coli lactose regulatory system. Nucleic Acids Res 9:3047-3060, 1981 Gaub MP, Rochette-Egly C, Lutz Y, Ali S, Matthes H, Scheuer I, Chambon P: Immunodetection of multiple species of retinoic acid receptor alpha: evidence for pbosphorylatioh. lixp Celt Res 201:335-346, 1992 Rocbette-Egly C, Lutz Y, Saunders M, Scheuer I, Gaub MP, Chambon P: Retinoic acid receptor gamma: specific immunodetection and phosphorylation. J Cell liiol 115:535-545, 1991 Chirgwin JM, Przyhyla AE, MacDonald RJ, Rutter WJ: Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 18:5296-5300, 1979

THE JOURNAL OF INVESTIGATIVE DEIUVIATOLOGY

35.

36.

37.

38.

39.

40.

41.

42.

43. 44.

45.

46.

Kato M, Kato K, Goodman DS: bnnumochemical studies ofthe localization of plasma and cellular retinol binding proteins and of transthyretin (prealbuniin) in rat liver and kidney. J Cell Biol 98:1696-1704, 1984 Reddy AP, Cben JY, Zacharewski T, Gronenieyer H, VoorheesJJ, Fisher GJ: Cbaracterization and purification of buman rednoic acid receptor- gamma 1 overexpressed in tbe baculovirus-insect cell system. Biochem J 833-840, 1992 Smith PK, Krohn RI, Hernianson GT, Mallia AK, Gartner TH, Provenzano MD, Fujimoto EK, Goeke NM, Olson BJ, KJenk DC: Measurement of protein using bicinchoninic acid. Aval Biochem 150:76-85, 1985 Rajan N, Blaner WS, Soprano DR, Suliara A, Goodman DS: Cellular retinolbinding protein messenger RNA levels in normal and retinoid-deficient rats. J Lipid Res 31:821-829. 1990 Eskild W. Oyen O. Beebe S.JahnsenT, Hansson V: Regulation of mRNA levels for cellular retinol binding protein in rat Sertoli cells by cyclic AMP and retinol. Biochem Biophys Res Commun 152:1504-151(1, 1988 Wei NL, Blaner WS, Goodman DS, Nguyen-Huu MC: Cellular retinoic acid and cellular retinol binding proteins: complementary deoxyiibonucleic acid cloning, chromosomal assignment and tissue specific expression. Mol lindocriiml 3:454-463, 1989 IJlaner WS, Olson JA. Retinol and retinoic acid metabolism. In: Spom MB, Roberts AB, Goodman DS (eds.). The Retinoids: Iiioh_i;y, Chcmistiy, aud Mediciuc, 2nd ed. Raven Press, New York, 1994, pp 229-255 Kurlandsky SB, Xiao J-H, Duell EA, Durand B, Chambon P, VoorheesJJ, ' Fisher GJ: Biological activity of all-trans retinol requires metabolic conversion to all-trans retinoic acid and is (nediated through activation of nuclear retinoid receptors in human keratinocytes. J Biol Chem 269:32821-32827, 1994 Randolph RK, Simon M: Characterization of retinol metabolism in cultured human epidermal keratinocytes. J Biol Chem 268:9198-9205, 1993 Siegenthaler G, Saurat JH, Ponec M: Retinol and retinal metabolism. Relationsbip to tbe state of differenriation of cultured buman keratinocytes. Biochem J 268:371-378, 1990 Ong DE, Newcomer ME, Cbytil F: Cellular retinoid-binding proteins. In: Sporn MB, Roberts AB, Goodman DS {eds.). The Retinoids: Bioloi^y, Cheufistry, iwd Mediciue, 2ud ed. Raven Press. New York. 1994. pp 283-317 Napoli JL: Biosynthesis and metabolism of retinoic acid: roles of CRBP and in retinoic acid homeostasis. J Nuir 123:362-366. 1993