Retinoic Acid Receptors as Regulators of Human Epidermal Keratinocyte Differentiation

Thomas M. Vollberg, Sr., Clara Nervi, Margaret D. George, Wataru Fujimoto, Andrbe Krust, and Anton M. Jetten Cell Biology Section Laboratory of Pulmonary Pathobiology National Institute of Environmental Health Sciences National Institutes of Health Research Triangle Park, North Carolina 27709 Department of Dermatology Okayama University Medical School (W.F.) 2-5-l Shikata-cho, Okayama City 700, Japan lnstitut de Chimie Biologique Faculte de Medecine LGME-CNRS (A.K.) 63085 Strasbourg Cedex, France

transglutaminase, and SQ37, and to activate transcription of the RARB response element-CAT reporter. These results demonstrate that the control of NHEK differentiation by RA is consistent with the interaction of the retinoid with RAR and the regulation of transcription by that ligand-receptor complex. (Molecular Endocrinology 6: 667-676, 1992)

To examine the role of nuclear retinoic acid (RA) receptors (RARs) in the regulation of squamous differentiation in normal human epidermal keratinocytes (NHEK), we analyzed binding activity, mRNA expression, and transcriptional activity of the endogenously expressed RARs. Specific RA-binding activity eluted from size-exclusion HPLC with an apparent mol wt of 50 kilodaltons and was predominantly (>95%) associated with the NHEK nuclear cell fraction. This RAR-binding activity represented in part the expression of RARa and RARy genes, whose transcripts were expressed in similar abundance in undifferentiated NHEK. Differentiation resulted in lower mRNA expression of RARa relative to the mRNA expression of RARr. Treatment of NHEK cells with lo-’ M RA did not induce expression of RARB mRNA. Similarly, three squamous cell carcinoma cell lines derived from human skin and oral cavity expressed RARa and RARr transcripts, but not RARB transcripts. Transfection of NHEK with chloramphenicol acetyltransferase (CAT) reporter plasmids indicated that the endogenously expressed RARs could activate transcription through the RARB response element in a concentration-dependent manner with doses of lo-’ M RA and higher. CAT expression was not activated through TRE, a palindromic thyroid hormone response element with purported RA responsiveness. The competitive binding of benzoic acid derivatives of RA to RAR correlated with the ability of each analog to suppress mRNA expression of the squamous cell markers, involucrin, type I O&36-6609/92/0667-0676$03.00/0 Molecular Endocrinology Copyright 0 1992 by The Endcmne

INTRODUCTION

Retinoidsprofoundly influencethe proliferationand differentiation of many epithelial tissues, including the epidermis (1, 2). In viva, both the excess and the deficiency of vitamin A, retinol, are capableof producing pathological changes in the epidermis(1, 3). In vitro, treatment with retinoic acid (RA), the active metabolite of retinol, inhibitsthe stratification and cornification of epidermalkeratinocytes (4, 5). Although stromal influences may mediate some retinoid effects in vivo, in vitro effects indicate that keratinocytes intrinsicallyrespond to retinoids. In culture, retinoids inhibit the expression of squamouscell markers, including keratins 1 and 10 (6-l 0), epidermaltype I transglutaminase (TGase) (1l-l 4) cholesterolsulfotransferase(13) and squamousspecific mRNAs detected by various cDNA clones selectedfor their differentialexpression(14, 15). The down-regulationof severalkeratins(16) and TGase (Saunders,N., and A. M. Jetten, in preparation)appears to occur at the level of gene transcription. Expression of keratin K19 is induced by RA at least in part by an increasein the transcriptionof this keratin gene(8-l 0). RA exerts these effects on keratinocytes at nanomolar

Smety

667 The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 26 May 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

MOL ENDO. 1992 668

Vol6 No. 5

concentrations (9, 13) and this action requires a particular retinoid structure, as exhibited by the activities of specific RA analogs (13, 17). These observations suggest that keratinocytes possess specific, high affinity receptors that mediate retinoid action. The identification of nuclear RA receptor @AR) genes has led to the assumption that many retinoid effects are mediated by these receptors (18-22). The RAR genes RARLu, RARP, and RARr are members of the steroid/thyroid hormone receptor family (23). Variation in the expression of RAR genes and their isoforms, the products of alternative splicing, has been hypothesized to account at least in part for the tissue specificity of RA responses (24-26). RAR genes encode 50-kilodalton (kDa) nuclear proteins that bind RA with high specific affinity (27, 28). These receptors possess domains for DNA and ligand binding and are thought to act as hormone (retinoid)-dependent transcriptional factors (23). RARP mRNA expression is induced in many cell types as a direct response to RA (29-33). The RA response element (@RARE), a specific DNA sequence in the RARP promoter, binds RARs and is responsible for RA-RAR-activated transcription of the RARP gene (34-36). In addition to activating transcription, a recent study suggests that one RAR isoform, RARy, is capable of inhibiting PRARE-mediated transcription (37). Thus, RARs can regulate gene expression through transcriptional activation and suppression. In the epidermis of embryonic mouse skin and in human skin, RARcv and RARy expression has been demonstrated by in situ hybridization (38-40). In recent reports, RARy and, to a lesser extent, RARa transcripts were detected in RNA from human epidermis and from keratinocytes cultured on fibroblast feeder layers (16, 41, 42). These results imply that RARs may mediate some RA regulation of gene expression in keratinocytes. Using cultured normal human epidermal keratinocytes (NHEK cells), we have examined the expression of RAR genes in order to define their role in the regulation of squamous differentiation. In this report we describe the RA-binding activity of NHEK RARs, establish the pattern of RAR gene expression with respect to RA treatment and cell differentiation, demonstrate that endogeneous retinoic acid receptors, presumably RARs, function as transcriptional activators, and relate RA-RAR interaction to RA regulation of squamous differentiation.

RESULTS HPLC Analysis

of RA-Specific

Binding

in NHEK

The sensitivity of cultured NHEK cells to RA inhibition of cell differentiation is consistent with the modulation of these effects by high affinity binding proteins. Both cytosolic and nuclear extracts prepared from proliferating undifferentiated keratinocyte cultures were incubated with 5 nM [3H]RA in the presence or absence of 1 go unlabeled RA or the RA analog Ch55 and analyzed

for the expression of RA-binding activity by size-exclusion HPLC analysis (Fig. 1). The RA analog Ch55 is an active retinoid which does not bind to cytoplasmic RAbinding protein (17). A single peak of specific binding was detected in the nuclear extracts. This RA-binding activity was eluted with a retention time (27 min) that corresponded to a mol wt (M,) of 50,000. The specificity of the binding was demonstrated by the competition of [3H]RA binding with unlabeled RA or Ch55. Radiolabeled RA in three other peaks eluting at 16, 39, and 43 min was not competed by the presence of cold RA or Ch55 and, thus, represented nonspecific binding. By HPLC analysis, cytosolic extracts of the cells did not contain RAR-binding activity, nor was there any RAspecific binding detected that would be consistent with the expression of the 15-kDa cytoplasmic RA-binding protein (results not shown). Thus, the predominant RAbinding activity in undifferentiated keratinocytes was RA specific, 50 kDa in size, and nuclear in localization, in agreement with the characteristics of recombinantly expressed RAR genes (27,28). Expression

of RAR mRNA in Epidermal

Cells

To establish which RAR genes contributed to the expression of the nuclear RA-binding activity, specific probes for the coding regions of each RAR mRNA were hybridized to Northern blots of NHEK RNA (Fig. 2A). RARol [3.8 and 2.8 kilobases (kb)] and RARy (3.3 and 3.1 kb) transcripts were detected in the cultured NHEK cells. The undifferentiated cells (Fig. 2A, lane 1) expressed similar levels of RARol and RAR-r, as noted from the specific activity of the probes (3 x 10’ and 6 X 10’ dpm/pg, respectively) and the length of exposure of the autoradiograms (14 and 10 days, respectively). Treatment of the cells with 0.1 PM RA for 24 h had little effect on RARcv and RARy expression and failed to induce any detectable levels of RARP mRNA (Fig. 2A, lane 3). RAR@ mRNA expression and its induction by RA were detectable in the lung carcinoma NCI-H647 cell line (Fig. 2A, lanes 5 and 6). In confluent cultures of the keratinocytes that have undergone phenotypic differentiation (Fig. 2A, lane 2) expression of RARol transcripts was diminished, but the level of RARy expression was unchanged, leading to the predominance of RARy transcripts in the differentiated cells. Additionally, there was an apparent qualitative change, which resulted in augmented expression of the 3.1-kb RARr transcript relative to expression of the 3.3-kb RARy transcript. Growth arrest of the proliferating NHEK cultures by treatment with 100 PM transforming growth factor-p, for 24 h did not alter the expression of RARa and RAR-y transcripts qualitatively, but did result in a modest decrease in the relative expression of transcripts for both RAR genes (Fig. 2A, lane 4). To determine if altered RAR expression may play a role in epidermal carcinogenesis, RAR gene expression was examined in several squamous cell carcinoma cell lines. RAR mRNA expression in three carcinoma cell lines, derived from stratified squamous epithelia, was

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 26 May 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

RA Receptor Regulation of Keralnocyte Differentiation

669

20

Retention

30

40

50

time (min)

Fig. 1. RA-Binding Activities in NHEK by HPLC Analysis Nuclear extracts were labeled with 5 nM [3H]RA in the absence (0) or presence of a 200-fold excess of RA (0) or retinoid analog Ch55 (A) and analyzed by HPLC size-exclusion chromatography, as described in Maferials and Methods.

comparable to the expression seen in NHEK cells (Fig. 28). Additionally, nuclear extracts of these cells contained RA-binding activity similar to that measured in NHEK (results not shown). Thus, these carcinoma cells express RARs that are present in the normal epithelial cells, and gross alteration of the expression of RARL~ and RARy does not appear to be associated with the transformed phenotype. The SCC 15 and SqCC/Yl cell lines are derived from epithelia within the oral cavity. Although normal oral epithelial cells express RARP transcripts (42), RARP expression was not observed in SCC 15 and SqCC/Yl cells.

to the PRARE. Induction of CAT activity above the basal transcription of the thymidine kinase promoter was detectable at a concentration of 1 nM and increased in a dose-dependent fashion to 39.5-fold above the basal level in the presence of 1 FM RA, the highest dose. In contrast, the TRE was used poorly by endogenous RA receptor. Background levels of CAT activity were detected in each of the cell lysates from TRES-tkCAT cells; even the highest dose (1 PM RA) induced CAT activity to less than 1.3-fold of the basal expression produced by the thymidine kinase promoter. RAR and Retinoid

Transcriptional NHEK Cells

Regulation

by Endogenous

Inhibition

of Squamous

Genes

RAR in

The functional role of the RA-RAR complex in regulating gene transcription in NHEK was examined by transient transfection with two different chloramphenicol acetyltransferase (CAT) reporter plasmids. The thyroid hormone response element (TRE) is a palindromic sequence that mediates transactivation by the thyroid hormone and other members of the steroid hormone family. @RARE is the direct repeat sequence in the RAR p2 promoter that mediates RA induction of that gene. In cotransfection experiments, recombinantly expressed RARs are capable of activating transcription through either of two response elements, @RARE and TRE (34-37, 43-45). Since it was not known whether either of these promoter sequences would mediate transactivation at endogenous levels of RAR expression, constructs containing each were transfected into NHEK cells. The cells were treated with dimethylsulfoxide (DMSO) or the indicated dose of RA for 48 h and then assayed for CAT activity (Fig. 3). The endogenous RA receptors, in response to RA treatment, were able to induce expression of the CAT reporter when linked

Functional benzoic acid derivatives of RA, analogs of the Ch series, induce retinoid effects on cells without binding to cytoplasmic retinoid-binding proteins (17). We reasoned that structural differences in these analogs might affect their ability to bind RARs and transcriptionally activate PRARE. Thus, if RARs were involved in the regulation of squamous differentiation, we expected that those Ch analogs that are less effective in suppressing squamous gene expression would be the same analogs that bind RAR less well and are less active in activating PRARE transcription. Suppression of mRNA expression of squamous cell markers was assessed by Northern analysis of confluent cultures of NHEK cells that had been treated with DMSO or 1 Om7M retinoid (RA or one of five Ch analogs (Ch20, Ch30, Ch40, Ch55, and Ch80) for 5 days (Fig. 4). The control (DMSO) NHEK cells expressed mRNAs corresponding to the cDNA clone SQ37, TGase, and involucrin at levels characteristic of squamous differentiated keratinocytes. In agreement with earlier studies (6-13, 15-17), treatment of the cells with RA inhibited their phenotypic differentiation and down-regulated the expression of the squamous cell-specific genes.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 26 May 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

Vol6 No. 5

MOL ENDO. 670

A NHEK I1

2

3

H647 4115 -

kb 4.4

-

2.4

RAR a - 2.4 - 4.4 RAR p

RARp -

2.4

-

4.4

RARy

RAR y -

2.4

- 2.4

- 2.4 - 2.4

GPDH - 1.4 GPDH

-

1.4

Fig. 2. Northern Analysis of RAR mRNA Expression in NHEK and Human Carcinoma Cell Lines Total RNA (30 pg/lane) was separated by denaturing electrophoresis in 0.66 M formaldehyde-l .2% agarose gel, transfered to a 0.45~pm Nytran nylon membrane, and successively probed for RAR gene expression, as described in Materials and Methods. GPDH is a constitutively expressed mRNA probed to control for loading of the RNA samples to the gel. The migration of RNA ladder markers (BRL Life Technologies) in an adjacent lane of the gel or that of ribosomal RNAs in each sample are indicated in the right margin. A, RAR genes expressed in NHEK cells. Expression of RAR genes in: lane 1, proliferating undifferentiated NHEK; lane 2, confluent, differentiated NHEK; lane 3, undifferentiated NHEK treated with 1 PM RA for 24 h; and lane 4, undifferentiated NHEK treated with 100 PM transforming growth factor-p, for 24 h. RNA from the lung carcinoma cell line NCI-H647 (from Dr. A. Gazdar, NCI, Bethesda, MD) cultured in the absence or presence of 1 PM RA for 24 h (lanes 5 and 6, respectively), as previously described (33) served as a control for the ability to measure expression of all three RAR genes. B, Comparison of RAR gene expression in NHEK cells and squamous carcinoma cell lines. NHEK cells and carcinoma cell lines cultured in the absence or presence of 1 NM RA for 24 h before collection, as described in Materials and Methods, were loaded onto the gel as indicated. LC, RNA from the lung carcinoma cell line SK-LU-1 (American Type Culture Collection).

The analogsCh30, Ch40, Ch55, and Ch80 were each effective in suppressingexpression of the squamous cell-specific mRNAs. The analog Ch20 was relatively ineffective in suppressingexpressionof squamousdifferentiation genes. Expression of glyceraldehyde 3phosphatedehydrogenase(GPDH) mRNA, a constitutively expressed gene, was unaffected by any of the treatments. Next, the ability of each analog to compete with RA for bindingto RARs in nuclear extracts of NHEK cells was measuredby HPLC analysis(Fig. 5A). The reduction of [3H]RA binding to RARs by the 200-fold molar excess of a competitor was plotted as a percentageof [3H]RA bound in the absence of any competitor. The active analogs Ch30, Ch55, and Ch80 were effective competitorsto RA for bindingto RAR. The Ch20 analog, which was relatively inactive in suppressingexpressionof squamousdifferentiationmarkers,failedto compete [3H]RA binding. The Ch40 analog, which was

effective in suppressingthe squamousmarker mRNAs, failed to compete for bindingto RAR. Since the activity in suppressing squamous gene expression was assessedby incubatingintact cells with the retinoid analogs and since binding activity was determinedin vitro with cell-free extracts, it may be possible that intact cells metabolize Ch40 to a form capableof bindingand activating RAR. The ability of each analog to effect transcriptional activation through endogenousRAR was measuredby transient transfection of NHEK cells with the PRARECAT vector (Fig. 58). The transfected cellswere treated with lo-’ M of each analog for 48 h before collection and measurementof CAT enzyme activity. The activation of CAT reporter expression closely paralleledthe suppressionof squamous cell-specific mRNAs. The analogsCh30, Ch40, Ch55, and Ch80 causeda greater than 5-fold increasein CAT enzyme activity. The analog Ch20 failed to significantly increase CAT activity over

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 26 May 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

RA Receptor Regulation of Keratinocyte Differentiation

671

kb

0

10

-12

,o-ll

,o-lo RA

,o-9

,o-*

&

SQ37

- 1.2

TGase

- 3.6

GPDH

- 1.4

INVOLUCRIN

- 2.1

,o.6

WV

Fig. 3. Transcriptional Regulation through RAREs by Endogenous RARs in NHEK Ceils were cultured and transfected with TRE3-tk-CAT (0) or @RARE-tk-CAT (D) plasmid and treated with the indicated dose of RA for 48 h. Collection and assay of CAT activity were as noted in Materials and Methods. Activation of the CAT reporter is indicate.d as the percentage of [‘4C]chloramphenicol converted to the acetylated form. Each determination was based on the activity in triplicate culture dishes. Error bars represent the SEM.

the level in control cells.The activity of the Ch40 analog in effecting transcriptionalactivation through the RARE confirmed that in intact cells Ch40 was capable of interacting with RAR, perhaps through a metabolite. Thus, the ability to inhibit the expression of squamous differentiationwas closely related to the transcriptional activation of geneexpressionby that retinoid-RARcomplex.

DISCUSSION

The identification of genes encoding RARs had suggestedthat RA-responsivecells, such as NHEK, would express the products of one or more of these genes. Similar to the characteristics of RARs expressed in other cells (27, 28, 32, 33) the RA-binding activity in NHEK cells was predominantly associated with the nucleus(>95%), eluted as a 50-kDa protein, and demonstrated RA-specific binding. This RAR-binding activity representedat least in part the expressionof RARa and RARr. Antibodies specific to each RAR protein (gifts from Dr. P. Chambon,CNRS, Strasbourg, France; and Dr. J. Grippo, Hoffman LaRoche, Nutley, NJ) were not sensitiveenough to determineif the relative mRNA expressionof each RAR gene is directly proportionalto the expressionof their respective RAR proteins(results not shown). Undifferentiated NHEK cells expressed RARcI and RARr transcripts in similar amounts. Previous reports have described the predominance of

Fig. 4. Suppression of Squamous Marker Gene Expression by Retinoid Analogs NHEK cells were cultured to within 80-90% of confluent density and treated with 1 Oe7M retinoid (from 1 OW4M stock) or an equal volume of DMSO (NA), as indicated. Benzoic acid derivatives of RA, Ch series analogs, were a gift from Dr. K. Shudo, University of Tokyo (Tokyo, Japan). Total RNA (20 fig) from treated cultures was analyzed by Northern blot for expression of the squamous marker genes SQ37, TGase, or involucrin. GPDH was probed as a control.

RARy expression in human skin and in keratinocytes derived from epidermis (16, 41, 42). We found that differentiationof NHEK cellscauseda shift in the mRNA expression of RAR genes, which resulted in the predominant expression of RARy transcripts. In situ hybridization of human skin was in agreement with the expression of RARa and RARr mRNA by the keratinocyte cells (39, 40). Thus, nuclear RA-binding activity, presumablythe expressionof RARa!and RARy genes, was the primary RA receptor activity detected in undifferentiated NHEK cells. The detection of this RARbinding activity by analysis using physiological(nanomolar) RA concentrations further implicatesRARs as effecters of retinoid responsein NHEK cells. Because of the role of RA in epithelialgrowth and differentiation, we suspectedthat changesin RAR activity or expression might contribute to pathological conditionsin which epidermalgrowth and differentiation are aberrant, such as neoplasiaor psoriasis.The distribution and expression of RAR genes, as detected by in situ hybridization, was similarin psoriaticskin to that in normal skin (results not shown). In squamouscell carcinomacell linesthe expressionof RA-bindingactiv-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 26 May 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

MOL

ENDO.

1992

Vol6

Retinoid

Treatment Fig. 5. Retinoid Analog Effectiveness in Binding to RAR and in Activating RARE-Mediated Transcription A, Competition of Ch series retinoids with RA for binding to NHEK RAR. Nuclear extract was incubated with 5 nM [3H]RA in the absence or presence of a 200-fold molar excess of the indicated retinoid. Binding of each retinoid was graphed as percentage of [3H]RA in the RAR peak fractions competed by the analog. B, Transcriptional activation of BRARE-tk-CAT by RA and RA analogs of the Ch series. Cells transfected with the @RARE-tk-CAT reporter were treated with DMSO or 1 O-’ M retinoid, as indicated. After 48 h of treatment, CAT activity was measured, as outlined in Materials and Methods. This experiment was carried out concomitantly with the experiment in A. The value for CAT activity in the DMSO control and the 1 O-’ M RA are derived directly from that figure. CAT activity is plotted relative to expression in the DMSO control and in each case was determined from triplicate culture dishes. Error bars indicate the SEM.

No. 5

ity and the mRNA expression of RARol and RARy genes were similar to those in the nontransformed NHEK cells, suggesting that carcinoma cells maintain the mechanism for RAR control of gene expression. Specifically, SCC 13 carcinoma cells have been shown to down-regulate some keratin genes in response to RA, although the overall expression of keratins and the magnitude of the response differed from those of normal keratinocytes (16). Thus, the putative role of RARs in the development of psoriasis or neoplasia cannot be defined by the mRNA expression of RAR genes alone, but will require the understanding of RAR target genes and their role in regulating cell growth and differentiation. The expression of RAR@ may be restricted to epithelia, which are programmed to undergo nonkeratinizing differentiation. RAR@ expression, which is modulated by RA in tracheobronchial epithelial cells (33), was not noted in NHEK cells under any culture conditions and was not detected after treatment of the NHEK cells with 1 FM RA for 24 h (Fig. 2A) or with 0.1 PM RA for 5 days (results not shown). Expression of RARP was also undetectable in the epidermis of normal or psoriatic skin by in situ hybridization (39, 40) (results not shown). Since RAR/3 is expressed in tracheobronchial epithelia (33), but is not required for the inhibition of squamous differentiation genes in NHEK cells, RARB induction must have a distinct function with respect to differentiation of mucosal epithelia. In the orally derived carcinoma cells, SCC 15 and SqCC/Yl , which undergo abnormal keratinization (42,46), we failed to detect RARP expression in the absence or presence of RA. Nonkeratinizing normal keratinocytes from the oral cavity express RARP mRNA, and this expression appears to be lost during the neoplastic progression of these cells (42) (Fig. 28). The loss of RARp expression in the neoplastic cells apparently results from gene inactivation, rather than gene deletion or rearrangement (42). Potentially, this inactivation could result from the inactivation or inhibition of RAR function. Husmann et al. (37) have shown that the RAR?, isoform is a potential inhibitor of @RARE-mediated Vansactivation by other RARs. In that cotransfection study, the sequence context of the ORARE appeared to influence the RAR7, antagonism of transactivation (37). Since our study did not examine the effect of sequence context on BRARE usage by the endogenous RARs, we cannot conclude whether RARr, antagonism occurs in NHEK cells. However, our results would seem to indicate that inactivation of the RARP gene occurs through some mechanism other than a modification of the ability of NHEK RARs to function as transcriptional activators. In addition to the modification of RAR function, the inactivation of the RARP gene could occur through a number of other known mechanisms of gene inactivation, including a modification of the RAR@ promoter (e.g. methylation) or of an additional factor that regulates RARP promoter, but not the thymidine kinase promoter construct. Alternatively, a transcriptional block to RARP mRNA elongation or a drastic decrease in RARP mRNA half-life

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 26 May 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

RA Receptor Regulation of Keratinocyte Differentiation

673

could prevent detectable expression of RAW in NHEK and squamous cell carcinoma cells. In addition to demonstrating RA-specific binding, we established that endogenous receptors in NHEK cells function as RA-dependent transcriptional regulators. This regulation was specific to the reporter construct containing @RARE sequence and was responsive to RA at concentrations consistent with the binding of RA to RARs. RXRs, also members of the steroid hormone superfamily, can mediate transcriptional regulation in response to RA treatment (47, 48). However, the dose response observed in our experiments would indicate that ligand-dependent @RARE-tyrosine kinase (tk)-CAT transactivation in the NHEK cells was more closely related to the interaction of the retinoid with RARs rather than RXRs. Recently, RXRs were shown to act in a ligand-independent manner as RAR coregulators by forming heterodimers with RARs and thereby enhancing the DNA binding and transcriptional regulation by RAR through RAREs (49). NHEK cells express RXRtv mRNA (Bernacki, S., and A. Jetten, unpublished observation). Hence, the activation of the reporter in NHEK could involve RXR coregulation of RARs. RA analogs differentially suppressed expression of squamous cell markers, competed with RA for binding to RAR, and activated PRARE-mediated gene transcription. Collectively, these results appear to indicate that interaction of a retinoid with RARs correlates with the retinoid regulation of the expression of the squamous phenotype. For some keratin genes (10, 16) and for TGase (Saunders, N., and A. M. Jetten, in preparation), RA suppression appears to occur at the transcriptional level. Furthermore, RA-dependent suppression can be conferred on reporter genes linked to the promoter region of the keratin or TGase genes (16) (Saunders, N., and A. M. Jetten, in preparation). The results in this manuscript suggest that RARs, either alone or in concert with other factors, such as RXRs, are an intimate part of this regulation. RA suppression of squamous cell-specific gene expression could be mediated by RAR binding to a RARE whose sequence context in the promoter leads directly to the ligand-dependent inactivation of transcription, or perhaps by a mechanism similar to the cross-talk between AP-1 factor and the vitamin D and RA receptors in the promoter of the osteocalcin gene (50, 51), RAR could interfere with the transcriptional activity of another transcription factor within a squamous cell gene promoter. Alternatively, RARs could indirectly suppress squamous cell genes by regulating positively or negatively the expression of reaulatorv factors that control transcription of squamous cell genes. Our full understanding of this suppression awaits identification of the critical promoter sequences that regulate expression of squamous cell genes, including the presence or absence of RAREs. In this report, using cultured keratinocytes that are devoid of other cell types or the remnants of other cells, such as fibroblast feeder layers, we have demonstrated the

intrinsic

expression

gene expression,

of RAR-binding

and RARE-dependent

activity,

RAR

transcriptional

activation. The expression and transcriptional activity of endogenously expressed RARs were consistent with their role as mediators of RA effects in keratinocytes. Furthermore, these effects are observed with characteristics that are in agreement with a physiological role of RA in hormonally regulating the maintainence of the epidermis through the product of RAR genes. Further studies will be required to address the mechanism by which RA regulates transcriptionally the suppression of squamous genes and to define how the metabolism of retinoids, including retinol gradients and the expression of cytoplasmic retinoid-binding proteins, relates to the physiological control of keratinocyte differentiation.

MATERIALS

AND METHODS

Cell Culture NHEK cells isolated from human foreskin were obtained as cryopreserved primary culture cells from Clonetics Corp. (San Diego, CA) and cultured in serum-free KGM medium [modified MCDB153 medium containing 0.15 mM CaC12 and supplemented with 0.1 rig/ml recombinant epidermal growth factor, 5 @g/ml insulin, 0.5 pg/ml hydrocortisone, and 0.4% (vol/vol) bovine pituitary extract] supplied by Clonetics. Undifferentiated or differentiated keratinocytes were obtained as previously described (14). All experiments were carried out with cells passaged no more than three ttmes after receipt. The human epidermal squamous cell carcinoma cell lines SCC 13 (52) and SqCC/Yl (46) were obtained from Dr. J. G. Rheinwald (Harvard University, Boston, MA) and Dr. J. McLane (Hoffman LaRoche), respectively. The carcinoma cell line SCC 15 (52) was obtained from the American Type Culture Collection (Rockville, MD). Carcinoma cells were cultured in the serumfree KGM medium. RA (all-trans.retinoic acid, Hoffman LaRoche) and RA analogs were prepared as 1 OOO-fold concentrated stocks in DMSO. Benzoic acid derivatives of RA, analogs of the Ch series, were provided by Dr. K. Shudo (University of Tokyo, Tokyo, Japan). Analysis

of RA-Binding

Activity

Undifferentiated NHEK cells (0.5-l .5 x 10’) were trypsinized and washed twice with PBS (pH 7.4) containing 2 mM EDTA and collected by centrifugation. Nuclear and cytosolic cellular fractions were prepared as previously described (28). To assess RA-binding activity, nuclear or cytosolic extract was incubated with 5 nM [3H]RA (55.7 Ci/mmol; DuPont-New England Nuclear, Boston, MA) in the absence or presence of 1 PM (200.fold excess) RA, CH55, or other retinoids for 18 h at 4 C and analyzed by HPLC using a Superose 12 HR lo/30 sizeexclusion column (LKB Pharmacia, Uppsala, Sweden), as previously described (28). Northern

Analysis

Total RNA was isolated and separated by electrophoresis in 0.66 M formaldehyde-l% agarose gel in 1 x MOPS buffer (3[N-morpholino]propanesulfonlc acid), as described (14). After transfer to Nytran nylon membrane (0.45 pm; Schleicher and Schuell, Keene, NH) in a Vacublot apparatus using the manufacturer’s protocol (American Bionetics, Inc., Hayward, CA), nucleic acid was permanently fixed to the membrane by UV cross-linking (53). Ribosomal RNA in the samples and a 0.24to 9.5.kb RNA ladder (BRL Life Technologies, Inc., Gaitherburq, MD) in an adiacent lane were used as size markers. Hybridization conditions were described previously (14). Blots

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 26 May 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

MOL 674

ENDO.

1992

Vol6

were washed to a stringency of 0.5 x SSC (15 mM sodium citrate, 150 mM NaCI, pH 7.0)-0.1% sodium dodecyl sulfate (SDS) at 55 C. For RAR genes, Northern blots were also washed in 0.2 x SSC-0.1% SDS at 60 C. Radiolabeled DNA probes were detected by exposure to Kodak XAR-5 film (Eastman Kodak, Rochester, NY) with enhancing screens (Liqhtninq Plus, DuPont, Wilminqton, DE) at -70 C. Previouslv deFected-probes were removed by washing at 75 C in 5 rni Tris-HCI IDH 8.0). 0.2 mM EDTA. 0.1 x Denhardt’s solution, and 0.05% (wt/vol) sodium pyrophosphate. Radioactively labeled DNA probes were synthesized from purified cDNA inserts using a random prime labelling kit (BRL Life Technologies) according to the manufacturer’s protocol with [ol-32P]deoxy-CTP (3000 Ci/mmol; Amersham Corp., Arlington Heights, IL). The specific activity of the radioactivity in the DNA probes was 3-8 x 10’ dpm/pg DNA. The RARa and RARr probes (from Dr. P. Chambon, Strasbourg, France) were synthesized from a 1.7-kb EcoRl restriction fragment of pHRARa0 (18, 20) and a 1.9-kb EcoRl restriction fragment of pHRARy0 (21). The RARp probe was synthesized from the 1.4-kb Psrl-Xhol restriction fragment from the plasmid pCOD20 (54) (a gift from Dr. H. deThe, Paris, France). TGase DNA probe was synthesized from the 2.6-kb EcoRl insert of clone DTG~ (12). The 0.8-kb cDNA oortion of clone SQ37 was used to probe expression of a squamous-specific mRNA that encodes an unknown gene product (15). lnvolucrin probe was synthesized from the 0.8-kb insert of clone pBRI-3 (55) (provided by Dr. R. L. Eckert, Case Western Reserve University, Cleveland, OH). GPDH, a 1 .12-kb Pstl cDNA fragement of pGAD 28, was used to probe for this constitutively expressed 1.45kb mRNA (56). Transient

Transfections

and CAT Assay

The plasmid CAT reporter constructs pBLCAT8+C25 and pBLCAT2-ORARE were gifts from Dr. P. Chambon (Strasbourg, France) and Dr. M. Pfahl (La Jolla Cancer Research Foundation, La Jolla, CA), respectively. The p(TRE3)-tk-CAT (45) consisted of the double stranded oligonucleotide

supernatant (50 ~1) was combined with 20 ~1 4 mM acetyl coenzyme-A (in 0.25 M Tris-HCI, pH 7.5) 0.25 &i [‘“Cl chloramphenicol (55 mCi/mmol; Amersham), and 0.25 M TrisHCI, pH 7.5, to a final volume of 150 ~1. After incubation at 37 C for 1 h, each reaction was extracted with 1 ml ethyl acetate. The organic phase was evaporated to dryness. The acetylated and unacetylated chloramphenicol in the dry tube were dissolved in 30 ~1 ethyl acetate and chromatographed on a silica gel TLC plate (Eastman Kodak), using chloroform-methanol (19:l) as solvent. The chromatograph was exposed directly to Kodak XAR-5 film for 7 days at room temperature. The CAT activity was quantitated by liquid scintillation spectroscopy of the radioactive acetylated and unacetylated chloramphenicol. Acknowledgments The authors express their appreciation to Drs. K. Shudo, P. Chambon, M. Petkovich, H. de The, J. Grippo, R. L. Eckert, and M. Pfahl for providing the reagents and plasmids used in this study. We thank Drs. J. Arata, S. Taniguchi, and S. Noji for sharing their findings from in situ hybridization of human skin. We gratefully acknowledge Drs. N. Saunders, S. Randall, and V. Davis for their suggestions during the preparation of the manuscript.

Received October 31, 1991. Revision received January 23, 1992. Accepted February 10, 1992. Address requests for reprints to: Dr. Anton M. Jetten, National Institute of Environmental Health Sciences, National Institutes of Health, P.O. Box 12233, Research Triangle Park, North Carolina 27709.

REFERENCES

5’-AGCTTAGGTCAGGGACGTGACCTT-3’, 3’-ATCCAGTCCCTGCACTGGAAGATC-5’ containing the palindromic TRE inserted between the HindIll and Xbal sites adjacent to position -105 of the thymidine kinase promoter in pBL-CAT8+. The pBLCAT2-@RARE (37) BRARE-tk-CAT, consisted of the double stranded oligonucleotide,

5.

5’-GATCTGTAGGGTTCACCGAAAGTTCACTCA-3’ 3’-ACATCCCAAGTGGCTTTCAAGTGAGTCTAG-5’, containing the direct repeat RARE of the RARP2 promoter inserted in the Bglll site adjacent to position -105 of the thymidine kinase promoter in pBLCAT2. Plasmid DNAs were obtained from cesium chloride-banded preparations. Transient transfections were accomplished by lipofection, using reagent and protocols from BRL Life Technologies. Briefly, NHEK cells (l-l .5 x 1 O5 cells) were cultured in 3 ml KGM in 60-mm tissue culture dishes (Falcon, Becton Dickinson, Lincoln Park, NJ) to a density of 0.8-l x 10’ cells/dish. For each culture dish, CAT plasmid DNA (2.5 pg) and pSVBGal (2.5 pg) were diluted with 20 +g lipofectin in sterile H20. The lipofectin-DNA mixture was then added dropwise to the culture dish. After 16-18 h at 37 C, the cells were fed fresh KGM (4 ml); treated by adding 4 ~1 DMSO, RA, or retinoid stock; and returned to the incubator. Each reporter plasmid and treatment was performed in triplicate culture dishes. After 48 h, the cells were collected and resuspended in 150 ~1 0.25 M Tris-HCI, pH 8.0. The cells were lysed by four consecutive cycles of freezing in dry ice-ethanol and thawing at 37 C. After the cell debris was microcentrifuged, the cleared lysate was transfered to a fresh tube and stored at -70 C. CAT activity was assayed as described by Gorman et al. (57). Cell lysate

No. 5

6.

7.

8.

9.

10.

11.

12.

Sporn MB, Roberts AB, Goodman DS (eds) 1984 The Retinoids, vol 1 and 2. Academic Press, New’York Eckert RL. Rorke EA 1989 Molecular bioloav of keratinocyte differentiation. Environ Health Perspeci80:109-116 Elias PM 1987 Retinoid effects on the epidermis. Dermatologica 175:S28-S36 Yuspa SH, Harris CC 1974 Altered differentiation of mouse epidermal cells treated with retinyl acetate in vitro. Exp Cell Res 86:95-l 05 Green H, Watt FM 1982 Regulation by vitamin A of envvelope cross-linking in cultured keratinocytes derived from different human epitheilia. Mol Cell Biol2:1115-1117 Connor MJ, Ashton RE, Lowe NJ 1986 A comparitive study of the induction of epidermal hyperplasia by natural and synthetic retinoids. J Pharmacol Exp Ther 237:3135 Kopan R, Traska G, Fuchs E 1987 Retinoids as important regulators of terminal differentiation: examining keratin expression in individual cells at various staqes of keratinization. J Cell Biol 105:427-440 Eckert RL. Green H 1984 Clonina of cDNA’s soecifvina vitamin A responsive keratins. Proc Natl Acad ‘Sci USZ 81:4321-4325 Gilfix BM, Eckert RL 1985 Coordinate regulation by vitamin A of keratin gene expression in human keratinocytes. J Biol Chem 260:14026-l 4029 Stellmach VM, Fuchs E 1989 Exploring the mechanisms underlying cell type specific and retinoid-mediated expression of keratins. New Biol 1:305-317 Thacher SM, Coe EL, Rice RH 1985 Retinoid suppression of transglutaminase activity and envelope competence in cultured human epidermal carcinoma cells. Differentiation 29:82-87 Floyd EE, Jetten AM 1989 Regulation of type I (epidermal) transglutaminase mRNA levelss during squamous differ-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 26 May 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

RA Receptor

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23. 24.

25.

26.

27.

28.

29.

30.

31.

Regulation

of Keratinocyte

Differentiation

entiation: down regulation by retinoids. Mol Cell Biol 9:4846-4851 Jetten AM, George MA, Nervi C, Boone LR, Rearick JI 1989 Increased cholesterol sulfate and cholesterol sulfotransferase activity in relation to the multistep process of differentiation in human epidermal keratinocytes. J Invest Dermatol92:203-209 Vollberg TM, George MA, Jetten AM 1991 Induction of extracellular matrix gene expression in normal human keratinocytes by transforming growth factor beta is altered by cellular differentiation. Exp Cell Res 193:93-l 00 Smits HK, Floyd EE, Jetten AM 1987 Molecular cloning of gene sequences regulated during squamous differentiation of tracheal epithelial cells and controlled by retinoic acid. Mol Cell Biol 7:4017-4023 Stellmach V. Leask A. Fuchs E 1991 Retinoid mediated transcriptional regulation of keratin genes in human epidermal and squamous cell carcinoma cells. Proc Natl Acad Sci USA 88:4582-4586 Jetten AM, Anderson K, Deas MA, Kagechika H, Lotan R, Rearick JI, Shudo K 1987 New benzoic acid derivatives with retinoid activity: lack of direct correlation between biologic activity and binding to cellular retinoic acid binding protein. Cancer Res 47:3523-3527 Petkovich M, Brand NJ, Krust A, Chambon P 1987 A human retinoic acid receptor which belongs to the family of nuclear receptors. Nature 330:444-450 Giguere V, Ong ES, Segui P, Evans R 1988 Identification of a new class of steroid hormone receptors. Nature 331:91-94 Brand N, Petkovich M, Krust A, Chambon P, deThe H, Marchio A, Tiollais P, Dejean A 1988 Identification of a second human retinoic acid receptor. Nature 332:850853 Krust A, Kastner P, Petkovich M, Zelent A, Chambon P 1989 A third human retinoic acid receptor, hRAR-gamma. Proc Natl Acad Sci USA 86531 O-531 4 Zelent A, Krust A, Petkovich M, Kastner P, Chambon P 1989 Cloning of murine LY and p retinoic acid receptors and a novel receptor gamma predominantly expressed in skin. Nature 339:714-717 Evans RM 1988 The steroid and thyroid hormone reeptor superfamily. Science 240:889-895 Kastner P, Krust A, Mendelsohn C, Garnier J-M, Zelent A, Leroy P, Staub A, Chambon P 1990 Murine isoforms of retinoic acid receptor y with specific patterns of expression Proc Natl Acad Sci USA 87:2700-2704 Leroy P, Krust A, Zelent A, Mendelsohn C, Garnier J-M, Kastner P, Dierich A, Chambon P 1991 Multiple isoforms of the mouse retinoic acid receptor (Y are generated by alternative splicing and differential induction by retinoic acid. EMBO J 10:71-81 Zelent A, Mendelsohn C, Kastner P, Krust A, Garnier JM, Ruffenach F, Leroy P, Chambon P 1991 Differentially expressed isoforms of the mouse retinoic acid receptor p are generated by usage of two promoters and alternative splicing. EMBO J lo:71 -81 Nervi C, Grippo JF, Sherman MI, George MD, Jetten AM 1989 Identification and characterization of nuclear retinoic acid-binding activity in human myeloblastic leukemia HL60 cells. Proc Natl Acad Sci USA 86:8854-8858 Jetten AM, Grippo JF, Nervi C 1990 Isolation and binding characteristics of nuclear retinoic acid receptors. Methods Enzymoll89:248-255 de The H, Marchio A, Tiollais P, Dejean A 1989 Differential expression and ligand regulation of the retinoic acid receptor 01 and ,R genes. EMBO J 8:429-433 Hu L, Gudas LJ 1990 Cyclic-AMP analogs and retinoic acid influence the expression of retinoic acid receptor OL, p and y mRNA’s in F9 teratocarcinoma cells. Mol Cell Biol lo:391 -396 Martin CA, Ziegler LM, Napoli JL 1990 Retinoic acid, dibutyryl-c-AMP, and differentiation affect expression of

675

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

retinoic acid receptors in F9 cells. Proc Natl Acad Sci USA 87:4804-4808 Nervi C, Vollberg TM, Grippo JF, Lucas DA, George MD, Sherman MI, Shudo K, Jetten AM 1990 Expression of nuclear retinoic acid receptors in wild-type and mutant embryonal carcinoma PCC4.azal R cells. Cell Growth Differ 1:535-542 Nervi C, Vollberg TM, George MD, Zelent A, Chambon P, Jetten AM 1991 Expression of nuclear retinoic acid receptors in normal tracheobronchial cells and in lung carcinoma cells. Exp Cell Res 195:163-l 70 de The H, delMar Vivanco-Ruiz M, Tiollais P, Stunnenberg H, Dejean A 1990 Identification of a retinoic acid response element in the retinoic acid receptor p gene. Nature 343:177-l 80 Sucov HM, Murakami KK, Evans RM 1990 Characterization of an autoregulated response element in the mouse retinoic acid receptor type p gene. Proc Natl Acad Sci USA 87:5392-5396 Hoffmann B, Lehmannn JM, Zhang X-K, Hermann T, Husmann M, Graupner G, Pfahl M 1990 A retinoic acid receptor-specific element controls the retinoic acid receptor p promoter. Mol Endocrinol 4:1727-l 736 Husmann M, Lehmann J, Hoffmann B, Hermann T, Tzukerman M, Pfahl M 1991 Antagonism between retinoic acid receptors. Mol Cell Biol 11:4097-4103 Dolle P. Ruberte E, Kastner P, Petkovich M, Stoner CM, Gudas LJ, Chambon P 1989 Differentiatl expression of genes encoding 01, (I’ and y retinoic acid receptors and CRABP in the developing limbs of the mouse. Nature 342:702-705 Noji S, Yamaai T, Koyama E, Nohno T, Fujimoto W, Arata J. Taniguchi S 1989 Expression of retinoic acid receptors in keratinizing front of skin. FEBS Lett 259:86-90 Fujimoto W 1990 Expression pattern of retinoic acid receptor genes in normal human skin. Jpn J Dermatol 100:1227-1233 Elder JT, Fisher GJ, Zhang Q-Y, Eisen D, Krust A, Kastner P, Chambon P 1991 Retinoic acid receptor gene expression in human skin. J Invest Dermatol 96:425-433 Hu L, Crowe DL, Rheinwald JG, Chambon P, Gudas LJ 1991 Abnormal expression of retinoic acid receptors and keratin 19 by human oral and epidermal squamous cell carcinoma cell lines. Cancer Res 51:3972-3981 Umesono K, Giguere V, Glass CK, Rosenfeld MG, Evans RM 1988 Retinoic acid and thyroid hormone induce gene expression through a common responsive element. Nature 336:262-265 Astrom A, Pettersson U, Krust A, Chambon P, Voorhees JJ 1990 Retinoic acid and svnthetic analoos differentiallv activated retinoic acid receptor dependeni transcription. Biochem Bioohvs Res Commun 173:339-345 de Verneuil H, Metzger D 1990 The lack of transcriptional activation of the v-erbA oncogene is part due to a mutation present in the binding domain of the protein. Nucleic Acids Res 1814489-4497 Reiss M, Pitman SW, Sartorelli AC 1985 Modulation of the terminal differentiation of human squamous carcinoma cells in vitro by all-trans-retinoic acid. J Natl Cancer lnst 74:1015-l 023 Mangelsdorf DJ, Ong ES, Dyck JA, Evans RM 1990 Nuclear receptor that identifies a novel retinoic acid response pathway. Nature 345:224-229 Mangelsdorf DJ, Umesono K, Kilewer SA, Borgmeyer U, Onq ES, Evans RM 1991 A direct repeat in the cellular retinal-binding protein type II gene confers differential reaulation bv RXR and RAR. Cell 66:555-561 Y;VC, Delsert C, Anderson B, Holloway JM, Devary OV, Naar AM, Kim SY, Boutin J-M, Glass CK, Rosenfeld MG 1991 RXR/3: a coregulator that enhances binding of retinoic acid, thyroid hormone and vitamin D receptors to their cognate response elements. Cell 67:1251-l 266 Schule R, Umesono K, Mangelsdorf DJ, Bolado J, Pike JW, Evans RM 1990 Jun-Fos and receptors for vitamins

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 26 May 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.

MOL 676

51

52.

53.

ENDO.

Vol6

1992

A and D recocoonize a common resoonse element in the human osteoca6in gene. Cell 61:497-504 Owen TA, Bortell R, Yocum SA, Smock SL, Zhang M, Abate C, Shalhoub V, Aronin N, Wright KL, van Wijnen AJ. Stein JL. Curran T. Lian JB. Stein GS 1990 Coordinate occupancy of AP-1 sites in the vitamin D-responsive and CCAAT box elements bv FosJun in the osteocalcin aene: model for phenotype suppression of transcription.-Proc Natl Acad Sci USA 87:9990-9994 Rheinwald JG, Beckett MA 1981 Tumorigenic keratinocyte lines requiring anchorage and fibroblast support cultured from human squamous cell carcinomas. Cancer Res 41 :1657-l 663 Church GM, Gilbert W 1984 Genomic sequencing. Proc Natl Acad Sci USA 81 :1991-l 995

No. 5

54.

de The H, Marchio A, Tiollais P, Dejean A 1987 A novel steroid thyroid hormone receptor-related gene inappropriately expressed in human hepatocellular carcinoma. Nature 330:667-670 55. Eckert RL, Green H 1986 Structure and evolution of the human involucrin gene. Cell 46:583-589 56. Dugaiczyk A, Haron JA, Stone EM, Dennison OE, Rothblum KN, Schwartz RJ 1983 Cloning and sequencing of a deoxyribonucleic acid copy of glyceraldehyde-3-phosphate sehydrogenase messenger ribonucleic acid isolated from chicken muscle. Biochemistry 22:1605-l 613 57. Gorman CM, Moffat LF, Howard BH 1982 Recombinant genomes which express chloramphenicol acetyltransferase in mammalian cells. Mol Cell Biol 2:1041-l 051

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 26 May 2015. at 12:09 For personal use only. No other uses without permission. . All rights reserved.