Retinoic Acid Regulates Oral Epithelial Differentiation by Two Mechanisms

Retinoic Acid Regulates Oral Epithelial Differentiation by Two Mechanisms

Retinoic Acid Regulates Oral Epithelial Difierentiation by Tw^o Mechanisms Mikael B. Kautsky,* Philip Fleckman,t and Beverly A. Dale*t *Dcpartment of ...

9MB Sizes 156 Downloads 55 Views

Retinoic Acid Regulates Oral Epithelial Difierentiation by Tw^o Mechanisms Mikael B. Kautsky,* Philip Fleckman,t and Beverly A. Dale*t *Dcpartment of Oral Biology. University of Washington; and tl^cpartment of Medicine (Dermatology), University of Wasliington. Seattle. Washington. U.S.A.

The effect of retinoic acid (RA) concentration on differentiation of oral keratinocytes and the influence of fibroblasts on RA-dependent regulation were investigated in a lifted culture system. Keratinocyte differentiation was assessed by morphology, immunohistochemistry and immunohlotting. Filaggrin/ profilaggrin and keratin 1 were used as biochemical markers for cornified epithelium and keratins 13 and 19 as markers for noncornified epithelium. Cultured oral keratinocytes in R_A-free conditions differentiated in a manner that closely resemhled the differentiation pattern of gingival epithelia in vivo. Increasing RA concentrations altered the in fiVo—like terminal differentiation of oral keratinocytes hy disruption of organized stratification, inhihition of filaggrin/profilaggrin and Kl expression, and stimulation of K13 and K19 expression. Differentiation of keratinocytes from both cornified and noncornified regions of the oral cavity varied in a similar manner in response to

added RA, with the exception of K19 expression. K19 was consistently expressed at higher levels in keratinocytes originating from noncornified epithelia as compared to those from cornified epithelia. The level of RA regulation was ultimately dependent on the type of fihrohlasts underlying the epithelial cells. Homologous fibroblasts rendered the oral keratinocytes less sensitive to the effects of RA than skin fibroblasts. In addition, at a given RA concentration, fihrohlasts from cornified oral mueosa potentiated keratinocyte expression of RA sensitive markers of keratinization as compared to the influence exerted by fihrohlasts originating from noncornified oral mueosa. These results indicate that the RA regulation of oral epithelial differentiation is mediated hy two separate mechanisms: a direct, RA concentrationdependent effect, and an indirect, fihrohlast-mediated effect. Key words: oval tnucosalvitamin Aldevmal equivalent. J Invest Demtatol 104:546-553, 1995

T

[7-10]. In epidermal keratinocytes cultured at the air-liquid interface, which undergo a high degree of differentiation, RA inhibits morphologic structure and e.\pression of several markers of coriiiOed epithelia, such as Kl and profilaggrin [11,12]. In contrast, keratins 13 and 19, normally found in noncornified epithelia, are upregulated by \KA in these experimental conditions. These observations suggest a possible role of RA in region-specific differentiation of oral epithelia. RA affects gene transcription via interaction with specific nuclear receptor proteins that function as retinoid-activated transcription factors [13—15]. To date, three different retinoic acid receptor subtypes (liJ^Ra, RAR/3, and RAR7) have been cloned and sequenced. Strong interspecies sequence homology suggests that each RAR subtype has distinct cellular functions ([15,16], review). It has recently been shown that the transcription of a number of difFerentiation markers in human epidermal keratinocytes is controlled by I^U^/RAR complexes [17,18]. In addition, a second family of retinoid responsive nuclear receptors (RXRs), which preferentially bind the RA metabolite 9-cis-RA, has been characterized [19]. rOCRs form heterodimers with RARs and function as co-regulators of gene transcription [20]. Another sot of proteins thought to be involved in mediating RA actions in cells are the cellular retinoic acid-binding proteins 1 and II (CRABP-I and CRABP-II). CI
he oral cavity exhibits extensive regional variation in epithelial differentiation. This characteristic makes oral epithelia an e.xcellent model for studies of differentiation processes in epithelia. The two major differentiation patterns of oral epithelia, the noncornified epithelium of lining mueosa and the cornified epithelium of masticatory mueosa, can be distinguished hy e.xpression of a number of structural proteins [1-3]. For example, the noncornified epithelia express keratins 13 and 4 in suprabasal cell layers and keratin 19 in basal cells, whereas the cornified epithelia express only small amounts of K13 and the major keratin products in suprabasal cells are keratins 1 and 10 along with keratins 6 and 16. Profilaggrin, a keratin-associated protein of epidermis, is also expressed in oral cornified epithelia [4]. Vitamin A exerts profound effects on differentiation of epithelial tissues; hypervitaminosis causes mucous metaplasia, whereas deficiency leads to squamous metaplasia (|5J, review). The response of epidermal keratinocytes to retinoic acid (RA), the active form of vitamin A in epithelia [6], has been shown by numerous studies Manuscript received July 28, 1994; revised October 20, 1994; accepted for publication October 2.S, 1994. Reprint request.s to: Dr. M. B. Kautsky, Department of Oral Biology SB-22. University of Washington, Seattle, WA 98195. Abbreviations: K, keratin; HOFs, human oral fibroblasts.

0022-202X/95/S09.50 • SSDI0022-2()2X(94)00382-H

• Copyright © 199.S by The Society for Investigative Dermatology, hic.

546

RA, FIBROBLASTS, AND ORAL KERATINOCYTE DIFFERENTIATION

VOL. t()4, NO. 4 APRIL 1995

Table I.

Antibody

Keratin Kl Keratin K13

KS19,1

Keratin K19

AKH-1 8959 (blots)

Profilaggrin/filaggriii Profilaggrin/filaggrin

5C10

KS19,1 (K19)

Source

Marker for

AE8

AEg (K13)

H&E

Antibodies Used for Identification of Markers of Epithelial Differentiation

547

T,-T, Sun" A, Schemier and T.-T. Sun" Unpublished results ICN, Immunobiologicals, Lisle, IL Date et al, 1987 [48] Fleckman et al, 1985 [49]

Gingival epithelium

' These antibodies were generously provided by T,-T, Sun, Cultured epithelium

intracellular RA concentrations for differential regulation of gene transcription [22,23], Connective tissue is also important in determining epithelial differentiation [24], Studies by recombination of oral and dermal tissues have shown that the connective tissues are able to influence differentiation and proliferation of keratinocytes [25-27], However, the molecular mechanisms of epithelial-niesenchymal interactions in the oral cavity are still largely unknown and the role of retinoids not understood, Sanquer at al [28] showed that dennal fibroblasts modulate the effects of retinoids on epidermal growth. It is therefore conceivable that subepithelial fibroblasts may also modulate the retinoid effect on adult skin and oral keratinocyte differentiation. The aims of this study were to achieve iu I'lVo—like differentiation of oral keratinocytes in culture to investigate the effects of RA and the influence of subepithelial fibroblasts on oral keratinocyte differentiation in vitro. This work is a first step in understanding a possible role of retinoids in region-specific differentiation of oral epithelia. Our results demonstrate two levels of the RA effect on differentiation of oral keratinocytes: a direct RA concentrationdependent regulation, which is similar for keratinocytes from both comified and noncornified oral mucosa, and an indirect regulation mediated by the underlying fibroblast type. Both mechanisms are likely to be involved in the region-specific differentiation of oral epithelia, MATERIALS AND METHODS Cell Culturc Oral tissues were obtained from patients undergoing routine oral surgical procedures. Epithelial cells were separated from the underlying tissue by dispase treatment [29], and primai-y cultures were grown on plastic in keratinocyte growth medium (KGM) (Clonetics Co., San Diego, CA). Keratinocytes were expanded through one passage in KGM, Lifted cultures were grown essentially as described by Asselineau and Prunieras [30], Briefly, bovine type I collagen (Bioetica, France) was mixed with a suspension based on minimum essential medium (Gibco, Life Technologies, Inc., Grand Island, NY), 10% fetal bovine serum (FBS) (Hyclone Laboratories I n c , Logan, UT) and 2 X 10'"^ fibroblasts per lattice, Fibroblasts tested included GMlt) fibroblasts, a cell line from human fetal dermal tissue (N,LG,M,S,, Camden, NJ), and oral and dermal fibroblasts, obtained from appropriate tissues according to conventional explant technique [31]- Tbi^ collagen lattices were allowed to contract at 37°C for 3-7 d, depending on fibroblast type. Second-passage keratinocytes originating from either coinified or noncornified oral mucosa were then seeded on the contracted collagen lattices at a density of 2 X lO'' cells/cm". The cultures were grown submerged in Dulbecco's modified Eagle's medium (Gibco) for 1 week, then raised to the air-liquid interface and grown for two more weeks. The medium contained 1 ()% of either nonnal (control) or delipidized FBS. Delipidization was carried out according to the method of Rothblat et al [32] and the delipidized product was ledissolved at an equivalent protein concentration in Hepes buffered saline. During the first week, all culture media were supplemented with normal FBS, penicillin/ streptomycin (Gibco), hyckocortisone (0.04 fj,g/nil) (Gibco), cholera toxin (10"'" mol/l) (Calbioehem Behring Diagnostics, Lajolla, CA), insulin (.S ^g/ml) (Sigma Chemical Co., St, Louis, MO), and transferrin (,=> /xg/ml) (Sigma), At the time of lattice elevation, all media, except the control medium, were switched to delipidized FBS and lt)()t)X stock solutions of all-trans RA in isopropyl alcohol (Sigma) were added together with the

^

Figure 1. Gingival keratinocytes form a tissue morphologically and iminunohistochemically similar to gingiva. Comparison of morphology (hematoxylin and eosin) and expression of profilaggrin (AKH-1), keratin 1 (5Clt)), keratin 13 (AE8), and keratin 19 (KS19,1) between gingival tissue (io\i iiirr) and cultured gingival keratinocytes {hottom row). The keratinocytes were cultured as described in Materials and Methods in media containing 10% delipidized seinm without addition of 1
above supplements under yellow ligbt to a final RA concentration of 0, 1 0 ' " ' , 1 0 " ' , 10"", or 1 0 " ' mol/l. Cultures were kept in the dark and media changed twice per week. Retinoid concentration in control serum was 5 X 1 0 " ' mol/l retinol as detennincd by the Clinical Nutrition Research Laboratory at the University of Washington using the technique of Bieri et al [33], Retinol was not detectable in tbe delipidized serum by this technique. Epithelial Differentiation Analysis Two lattices were used for each experimental group. One specimen was fixed in methyl Carnoy's, sectioned, and used for light microscopy of hematoxylin and eosin—stained sections and for immunohistochemical analysis by the avidin biotin peroxidase complex method [34], The other lattice was used for protein analysis using 7,5—12.5% gradient sodium dodecylsultate—polyacrylamide gel elctrophoresis and immunoblotting. The cultured epithelium was mechanically separated from the collagen lattice and epithelial proteins extracted by homogenization in Urea/Tris buffer (8 mol/l and 0.05 mol/l, pH 7,6), as described previously [35], The extracts were frozen at —70°C until analysis. Discontinuous sodium dodecylsulfate gels were performed by the technique of Laemmli [36] and stained with Coomassie brilliant blue. Proteins from duplicate gels were electrophoretically transferred to nitrocellulose filters (Schleicher and Schuell, Keene, NH) and protein markers of interest were identified with appropriate antibodies (Table 1) employing the avidin biotin peroxidase method, RESULTS Oral Culture Parameters The dermal equivalent culture system [30] was modified as described above to allow for in I'li'ii-like oral keratinocyte differentiation, A comparison of morphology between oral coniified epithelitiin and cultured oral keratinocytes grown in medium containing 10% delipidized serum at the airliqtiid interface is shown in Fig 1, The cultured epithelium was more compact and lacked the prominent epithelial ridges seen in the gingival tissue. However, the cultured epithelium did exhibit the four characteristic layers of a normal gingival epithelium, i,e,, basal, spinous, granular, and comified cell layers, A partially dissociated zone of cells was seen above the well differentiated cell layers in most cultures. These cells were most likely formed early in the culture period prior to exposure to differentiating conditions at the air-liquid interface and remained associated with the differen-

548

KAUTSKY ET AL

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

H&E

AKH-1 (ProFG/FG)

KS19.1 (K19)

RA (mol/L)

0

10-10

10-8

Figure 2. RA inhibits terminal dificrentiation of gingival keratinocytes in lifted cultures—immunohistocheinistry. Morphology (hematoxylin and eosin) and expression of profilaggrin (AKH-1), keratin 1 (.SCIO). keratin 1.^ (AE8). and keratin 19 (K,SI9.1) by gingival keratinocytes is regulated by l
tiated culture due to lack of desquamation under these conditions [37]. Immunohistologic staining for four representative difFerenriation markers is also shown in Fig 1. In the tissue as well as in the cultured epithelium, the granular layer was well delineated by reaction with antibody AKH-1, showing a batid of profilaggrin expression

immediately below the cornified cell layer. Both epithelia exhibited suprabasal expression of Kt as detected by iintibody 5C1(). In contrast, Kl.3 was expressed only in occasional, individual cells in the tissue and was also very poorly expressed in the differentiating portion of tlie cultured epithelium. K19 was not expressed in either epithelium.

VOL. 104. NO. 4 APRIL 1995

RA. FIDRODLASTS. AND ORAL KERATINOCYTE DIFEERENTIATION

DifFerentiation of oral keratinocytes is extremely sensitive to culture conditions. Epithelial cells of higher passage number (greater than three) or those that grew slowly in secondary cultures were less well differentiated when transferred to lattices (data not shown), suggesting the necessity of u.sing standardized conditions for growth and expansion of these cells. Second-passage keratinocytes (tertiary) expanded tor approximately two and a half weeks were used in subsequent experiments. Effect of Retinoic Acid on Oral Epithelial Differentiation The effect of RA on comified oral keratinocytes in dermal equivalent cultures is shown in Fig 2. In medium containing no RA or low concentrations of RA (10 '"mol/1) a squamous stratified epithelium was formed and markers of cornified differentiation, profilaggrin and K l , were expressed. High-IVA conditions (10 ** mol/1) prevented terminal differentiation of the keratinocytes, resulting in a disorganized epithelium with increased expression of markers of noncornified epithelium, K13, and K19, and decreased expression of markers of cornified epithelium, profilaggrin, and K l . The immunoblots in Fig 3 further support the findings of the immunohistochemistry. An increase in K19 expression (KS19.1) was observed with increasing RA concentrations, and a decrease in profilaggrin/filaggrin and Kl expression (8959 and 5C10) occurred under these conditions. Keratinocytes from oral noncornified mucosa responded similarly to RA in a parallel series of experiments (data not shown). To control for inter-experimental variation, keratinocytes from cornified and noncornified oral mucosa taken from one patient were compared for lVA response in a single experiment. Differentiation of these two cell types in response to l^A. was similar (Fig 4). In BjV-free conditions, both cell types formed a parakeratinized, stratified epithelium with similar expression of profilaggrin, Kl, and K13. In 10 " mol/1 1\A, both keratinocyte types formed an epithelium that was less well keratinized, but again morphology and expression of profilaggrin, Kl, and K13 were comparable. In contrast, expression of K19 differed consistently between these two epithelia in various growth conditions (Fig 4 and data not shown). K19 expression was consistently higher in the cultured epithelia originating from noncornified oral mucosa, mimicking the pattern of expression seen in the original tissues [38].

2 2 2 2

549

Influence of Fibroblasts on Oral Epithelial DiflFerentiation The above re.sults indicated that although there was some degree of intrinsic diversity (K19 expression) between the basal cells of cornified and noncornified oral epithelia, this diversity was not sufficient to explain tlie variation in regional differentiation iit sitti. This suggested that extrinsic signals influence the pathway of terminal keratinocyte differentiation. One source of extrinsic signals is the subepithelial fibroblast population. Therefore, we tested the ahility of different fibroblast types to influence the RA-sensitive differentiation of oral keratinocytes. Fibroblasts were introduced as the variable in three separate comparisons of oral keratinocyte differentiation. In medium containing 10% control serum (comparable to approximately 10"** mol/1 1^^), oral keratinocytes influenced by fetal dermal fibroblasts (GMlOs) expressed markers of noncornified oral epithelia, K13, and K19, and, conversely, when influenced by oral fihroblasts (HOFs 91-14) expressed markers of coniified epithelia, profilaggrin, and Kl (Fig 5). Results at three different l^JK concentrations (0, 10 '. and 10 ^ mol/1, data not shown) support this finding and suggest that in comparison with oral fibroblasts, htiman fetal fibroblasts potentiated the RA response ofthe keratinocytes. When adult oral fibroblasts (HOFs 91-14) and adult dermal fibroblasts (J.G.) were compared, a similar effect was ohserved. The epithelial cells cultured with dermal fihrohlasts differentiated as if they were exposed to higher amounts of retinoid than when cultured witli oral fibroblasts. This difference was most prominent at high l^A concentrations (Fig 6), where 10 ' mol/1 RA resulted in disruption of orderly stratification in the cultures with dennal fibroblasts (J.G.). hut not with oral fibrohlasts (HOFs 91-14). The final comparison was done between fibroblasts originating from oral cornified mucosa (HOFs 92-21) and oral noncornified mucosa (HOFs 92-22). Of these two oral fibrohlast subtypes, those from cornified epithelia (HOFs 92-21) induced a lower RA sensitivity in the keratinocytes, promoting expression of markers of coniification, profilaggrin, and Kl, and inhibiting markers of noncomified epithelia, K13, and K19 (Fig 7). The differences in VJK sensitivity can also be seen in the morphology of these cultured epithelia (Fig 8). This difference was most prominent in the high RA concentration, as the effect of seemingly higher RA concentration (compare to Fig 6) emerged in the cultures containing fihroblasts from noncornified oral epithelia (HOFs 92-22). Although the epithelial cells displayed differential l^A sensitivity when influenced by fibroblasts from either cornified or noncomified oral mucosa, the difference was not sufficient to implicate subepithelial fibroblasts as the sole regulators of regional variation in oral epithelial differentiation. DISCUSSION

KS19.1

5C10

8959

Figure 3. RA inhibits terminal diflerentiation of gingival keratinocytes in lifted cultures—immunoblots. Epithelial proteins were separated on a 7.5'Mi to 12.5'Mi gradient sodium dodecylsulfate gel, transferred to nitrocellulose, and reacted with antibodies as indicated (KS19.I is specific for keratin I'), 5C1() for keratin 1. and S95') detects filaggrin and profilaggrin). The first panel shows a duplicate gel stained with Coomassie brilliant blue. Molecular weight markers (size in kl)a) lire shown on the left and proteins of interest on the right (P. profilaggrin: K l , keratin 1; K19, keratin 19; and FG, filaggrin). l^A concentnitions are indicated above each lane. The control lanes contain protein extracts from keratinocytes grown in medium with 1 (1% normal calf serum. The retinoid concentration of this medium approximates the 10"" mol/1 1<,A culture eonditions (.see Matetials attd Methods).

The aim of this work was to investigate R ^ effects on oral epithelial differentiation. An air-liquid interface culture system was adapted for grov\'th and differentiation of oral keratinocytes. This culture system was used to show that I^LA inhihits morjihologic differentiation and expression of biochemical markers of keratinization of oral epithelial cells irrespective of the original iit sittt keratinocyte differentiation. In addition, we showed that subepithelial fihrohlasts play an iniportant role in the regulation ofthe RA responsiveness of oral keratinocytes and subsequent differentiation. The observation that RA acts as an inhibitor of oral comified keratinocyte differentiation parallels the findings of Asselineiiu et al [12], who demonstrated a similar RA effect on epidermal differentiation. They showed that a specific range of RA concentrations (10 ''-10 " mol/1) was required to ohtain in vitro epithelial architecture closest to that found in vivo. A lower optimal range was found in the present study. In fact the oral keratinocytes seemed to differentiate equally well in delipidized media with no added KA as in delipidized media with low RA concentrations (10 '" moI/1). This difference in RA levels for optimal differentiation between the two studies may be due to differences in culture conditions or to inherent differences between epidermal and oral keratinocytes. In the present study, we have found inherent differences between two

550

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

KAUT,SKY r-r AL

+RA

-RA Cornified

Noncornified

Cornified

Noncornified

H&E

AKH-1

5C10

AE8

KS19.1 t

.

Figure 4. Keratinocytes from cornified and noncornified oral epithelia di£E°erentiate similarly in response to RA in lifted cultures. Keratinocytes from the same donor from the two major differentiation p;itterus of oral epithelium were grown at the air-liquid interface on lattices containing gingival fibroblasts. The growth media were supplemented with delipidized serum either without RA or with RA added back to a final concentration of 10 " mol/l. Markers of differentiation were analyzed by ininiunoliistocheinistry (same as in Fig 2), Note the similar expression of markers iu the two epithelial types, except for K19 (KS19,1), which was more strongly expressed in noncornified epithelia under all conditions tested. Also note that in the high IVA concentrations, the use of oral fibroblasts here yields more satisfactory morphology and differentiation than with GMIO fibroblasts used in Fig 2, Interrupted lines mark the epithelial border. Bar, 100 /J-ni,

oral keratinocyte types in tbeir keratin 19 expression. Evidence of inherent differences between various keratinocyte populations is also presented by Lindbcrg and Rlieinwald [39|, wlio identified three distinct subtypes of human oral keratinocytes based on differentiation patterns in xenografts subsequent to primary cultures. Our results, however, indicate tbat regional variation in oral epithelial differentiation cannot be entirely explained by intrinsic

differences in the two keratinocyte types and their response to RA, W e suggest an important role of tbe subepithelial fibroblast, Tbe fibroblasts were able to influence differentiation of tbe oral keratinocytes in accordance with their tissue of origin via an apparent modulation of keratinocyte RA sensitivity, Consec]uently, fibroblasts from cornified oral mucosa inhibited the RA response of oral keratinocytes, thus leading to more abundant expression of markers of comification in the epithelial cells, whereas fibroblasts from

RA. FIBROBLASTS. AND ORAL KERATINOCYTE DIFFERENTIATION

VOL. 104. NO. 4 APRIL 1995

IP

GMIO

551

HOFs 91-14

HOFs 91-14

H&E H&E

AKH-1

....V

5C10

AE8

Figure 6. Adult dermal fibroblasts (J.G.) and gingival fibroblasts (HOFs 91-14) induce different responses in RA-sensitive differentiation of gingival keratinocytes. Epithelial morphology of gingival keratinocytes grown on lattices containing either J.G. fibroblasts (dennal) or HOFs 91-14 (gingival) in medium with 10 mol/1 l^JK was assessed by hematoxylin and eosiii staining. Note that gingival keratinocytes form a nonconiified type epitlielium in the presence of gingival fibroblasts. but a disorganized epithelium is formed in the presence of dennal fibroblasts at this RA concentration. Bar, 100 ^Jim.

fibrohlasts (see Fig 8) are hy themselves capable of inducing the large variation in terminal differentiation seen in the oral cavity. Consequently, this suggests that there must be a continuous cross talk between the oral keratinocytes and the oral fibroblasts during the induction process leading to terminal differentiation of the epithelial cells. Furthermore, we have shown in the present study that evaluation of the epithelial response to signals fi-om the underlying connective tissues may depend on the differentiation marker measured. Thus, the adult oral keratinoeyte phenotype may be intrinsically predetermined in some regards (K19 expression), but receptive for extrinsic influences in others (Kl, K13, and profilaggrin expression).

RA lO-'' 10-' 10-7 10-9 10-7 10-9 10-7 10-9 10-7 10-9 (niol/L) II 1 1 1 1 1 1 1 1 1

KS19.1

FIBROS

C

N

C

N

C

N

C

N

C

N

C

N

C

N

C

N

C

N

C

h 65-

Figure 5. Human fetal dermal fibroblasts (GMIO) induce greater RA sensitivity in gingival keratinocytes than gingival fibroblasts /jjOFs 91-14). Gingival kcratinocytcs were grown on lattices containing either GMIO fibroblasts or HOFs 91-14 in medium with 1 {)% control serum. These conditions approximate 10"" mol/1 RA (see Materials and Methods). Markers of difterentiation were assessed by innnnnohistochemistry (same a.s in Fig 2). Note prominent expression of K1.1 (AE8) and K19 (KSI9.1) in epithelia cultured with GMIO fibroblasts. but expression of Kl (5C10) and filaggrin (AKH-1) in cultures with oral fibroblasts. Interrupted lines mark the epithelial border. Bar. 100 fini.

nonconiified epithelia potentiated the RA response of oral keratinocytes. Some previous studies suggest that the adult oral epithelial phenotype is an intrinsic property of the epithelium [39-41]. Other oh.servations indicate that connective tissues are responsible for the final pattern of oral epithelial differentiation [42—44]. The present study suggests that neither the intrinsically predetermined keratinoeyte phenotype (see Fig 4), nor the influence of regional

43-

N

I -Kl -K13 -K19

=

KS19.1

AE8

5C10

8959

Figure 7. Fibroblasts from different regions of oral cavity differentially influence expression of RA-sensitive markers of keratinoeyte differentiation—immunoblots. Keratinocytes were cultured on lattices containing either fibrohlasts from coniified (C) or nonconiified (N) oral mucosa in media with delipidized serum supplemented with 10 ' mol/1 or 10 '' mol/I RA. Proteins were separated on a 7.5% to 12.5% gradient sodium dodecylsnltate gel. transferred to nitrocellnlose. and probed with antibodies as indicated (KS19.1 is specific for keratin 19. AE8 for keratin 13. 5C10 for keratin 1. and 8959 detects profilaggrin and filaggrin). The first panel shows a duplicate gel stained with Coomassie brilliant blue. Molecular weight markers (kDa) are indicated on the left and proteins of interest on the right.

552

KAUTSKY ET AL

HOFS 92-21 (C)

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

HOFs 92-22 (N)

RA (mol/L)

10-9

10-7

Figure 8. Fibroblasts from different regions of oral cavity differentially aflfect morphology of gingival keratinocytes grown at the air-liquid interface. Keratinocytes were cultured on lattices containing fibroblasts from either comified (C) or noncornified (N) oral mucosa in media with delipidized serum supplemented with 1 0 ^ niol/1 or 10~ ^ mol/l RA and morphology was assessed by hematoxylin and eosin staining. Note the greater sensitivity of keratinocytes to to,xic efFects of RA, i,e,, disruption of orderly stratification, when cultured in the presence of fibroblasts from a noncornified region. Bar, 100 /xin.

Certain variability between consecutive experiments in the present study can be attributed to the use of primary cultures for each experiment. However, this variation does not alter the interpretation and conclusions ofthe results reported. The efFects of RA on oral keratinocyte differentiation and the influences of the various fibroblasts were consistently observed, Furtherniore, to ensure reliability, most of the results shown in this report were derived from comparisons within a single experiment and any direct comparison between experiments was made with caution. This study indicates that the dermal equivalent is a useful model system for studies of oral keratinocyte differentiation. It will complement the nude-mouse-xenograft model [39] as a system for study of highly differentiated oral keratinocytes. One advantage of the dermal equivalent is the ease of administering various agents influencing oral epithelial differentiation. Our results suggest that at least two different RA-associated regulatory mechanisins influence terminal differentiation of oral epithelia: one is via a direct, RA-concentration-dependent mechanism. In this regulation low RA concentrations decreased expression of the two markers of noncornified oral epithelia, K13 and K19, whereas expression of the two markers of cornification, profilaggrin and K l , were increased and high RA concentrations had the opposite effect, Keratinocytes from comified and noncornified oral regions showed no difference in the sensitivity for this type of RA regulation for three ofthe four markers examined in the present study. Only K19 expression was consistently different in these two keratinocyte types at various RA concentrations. This phenomenon may be explained by differential expression of RAR subtypes in the two epithelial cell types. Thus, for example, RAR7, which is thought to be equally expressed in the two cell types, may control K13, Kl, and profilaggrin expression, whereas RARj8, which is differentially expressed in these two cell types, may control K19 expression [45,46], We show here that a second level of regulation, affecting RAsensitive markers of differentiation, is mediated by subepithelial fibroblasts. The nature of this regulation is more indirect, because it seems

to influence the apparent RA exposure of the keratinocytes independently of the actual KA concentration in the ctilture medium. One possible mechanism for this kind of regulation may be exerted via the CRABPs through their proposed ability to sequester KA in the cytoplasm of epithelial cells [21], preventing it from acting as a ligand for the nuclear RARs, This effect may in turn be mediated by tissue-specific expression of one or more soluble fibroblast products that influence epithelial cells. Indeed, Boukamp cl al have demonstrated that dennis regulates epidermal keratinocyte growth and differentiation via diffusable factors [47], The dermal equivalent system will serve as a basis for fliture experiments aimed at clarification of the,se hypotheses.

IVe thatik Dr. Rtitger Perssoti from the Departtnettt of Peiiodotitics atid Ms. Char Bell-Yotiiigerfrom the Department of Oral and MaxiDofacial Siirgeq', botli at lhe Uttiuersity of IVasltiitgloti, for providing ns with hiiniatt oral ti.tsne for our experiments. Tlii.t research is siipporled hy National Institute of Dental Research grattt 5-Ki5-DE003n.

REFERENCES 1.

2. 3. 4. 5. 6.

7. 8.

9. 10.

11.

12. 13. 14. 15.

16.

17.

18.

19. 20.

21.

Ouhayoun JP, Gosscliii E, Forest N, Winter S, Eranke WW: Cylokor,itin patterns of hutnan oral epithelia: differences in cytokeratin synthesis in piiigival epitiielium and the adjacent alveolar mucosa, DiJI'creiitJiitiott 30:123-129, 1985 Clausen H, Vedtofte I>, Moe D, Dahelsteen E, Sun TT, Dale B: DilFerentiationdepeiident expression of keratins in human oral epithelia. J Invest Dermatol 86:249-2,54, 1986 Dale BA, Salonen J, [ones AH: New approaches and concepts in the study of differentiation of oral epithelia. Crit Rci> Oral Hiol Med 1:167-190, 1990 Smith SA, Dale BA: Imuiunologic localization of fllaggrin in human oial epithelia and correlation with keratinization. J hivrst Dernuitol 86:168-172, 1986 Dannon M: Retinoic acid in skin and epiUielia. Snn Dev Uhl 2:219-228, 1991 Siegeiuhaler G, Saura( JH, I'ouec M: Retinol and retinal nietaboli.sni. Relationship to the state of differentiation of cultured human keratinocytes. liwclifin J 268:371-378, 1990 Fuchs E, Green H: Regulation of terminal differentiation of cultured iuinian keratinocytes hy vitamin A. Cell 2.S:617-625, 1981 Green H, Watt FM: Regulation by vitamin A of envelope cross-linking in cultured keratinocytes derived froni different human epithelia. Mol Cell Biol 2:111.5-1117, 1982 Choi Y, Fuchs E: TGF-heta and retinoic acid: regulators of growth and modifiers of differentiation in luniian epidermal cells. Cell Regiil 1:791-809, 1990 Asselineau D, Dale LiA, Bernard BA: Filaggrin production hy cultured human epidermal keratinocytes and its regulation by retinoic acid. Differentiation 45:221-229, 1990 Kopan R, Traska G, Fuchs E: Retinoids as important regulators of terminal differentiation: examining keratin expression in individual epidermal eells at various stages of keratiuization, J Cell Biol 105:427-440, 1987 Asselineau D, Bernard BA, Bailly C, Darmon M: Retinoic acid improves epidermal morphogenesis. Dev Hiol 133:322-335, 1989 I'etkovich M, Brand NJ, Krust A, C^liamhon P: A human retinoic acid receptor which belongs to tbe family ofnuclcvir receptors. JVi»»rr 330:444-450, 1987 Evans RM: Tbe steroid and tbyroid hormone receptor superfamily. Scienee 24«:«a9-89S, 1988 Zelent A, Krust A, Petkovicil M, Ka.stner J', Cliamhon P: C^loning of murine alpha and beta retinoic acid receptors and a novel receptor gamma predominantly expressed in skin. Natnre 339:714-717, 1989 Lobnes D, Dierich A, Ghyselinck N, Kastner 1', Lampron C, LeMeur M, Lufkin T, Mendelsobn C, Naksbatri H, Cbambon P: Retinoid receptors and binding proteins.i Cell Sci (suppl 16):69-76, 1992 Vollberg TM Sr, Nervi C, George MD, Fvijimolo W, Krust A, Jetten AM: Retinoic acid receptors as regulators of buman epidermal keratinocyte differentiation, Mol Entiocriiiol 6:667-676, 1992 Dlumenberg M, Connolly DM, Freedherg IM: Regulation of keratin gene expression: tbe role of tbe nuclear receptors for retinoic acid, tbyroid hormone, and vitamin D3.J liii'est Dernuitol 98(suppl 6):42S-49S, 1992 Maugelsdorf DJ, Ong ES, Dyck JA, Ev,nns RM: Nucle,ir receptor tbat identifies a novel retinoic acid response pathway. Nature 345:224-229, 1990 Yu VC, Delsert C, Andersen B, Holloway JM, Devary OV, Naar AM, Kim SY, Boutin J-M, Glass CK, Rosenfeld MG: lOCf^ beta: a coregulator that enhances binding of retinoic acid, tbyroid hormone, and vitamin D receptors to tbeir cognate response elements. Cell 67:1251-1266, 1991 Astrom A, Tavakkol A, Pettersson U, Cromie M, Elder J'i', Voorhees JJ: Molecular cloning of two human cellular retinoic acid-binding proteins (CRABP). Retinoic acid-induced expression of CRABP-ll but not CRABP-1 in adult btiman skin in vivo and in skin fibrohlasts in vitro. J Hiol Chem 266:17662-17666, 1991

VOL. 104. NO. 4

22.

23.

24. 25.

26. 27.

28.

29.

30. 31. 32. 33.

34.

35.

36.

APRIL

RA. FIBROBLASTS. AND OI!J\L KERATINOCYTE DIFFERENTIATION

Madeii M, Oiig DE. SiinmiLMbell D, Chyril F: Spatial distribution of ccllulaiprotein binding to retinoic acid in tbe chick limb bnd. Niitme 335:73.1-735. 1988 Boylan JF. Gndas LJ: Overcxpression of tbe cellular retinoic acid binding protein-I (CRABP-I) results in a reduction in differentiation-specific gene expression in F9 teratocarcinonia cells. J CeU liiol 112:965-979, 1991 Briggaman WJK: Epidermal-dermal interactions in adult skin.y 7i/i'rs( Detviatol 79 (suppl 1):21S-24S. 1982 Billingliani RE, Silvers WK: Stndies on tbe conservation of epidermal specificities of skin and certain niucosas in adnit mammals. /fc.v;)Med 125:429-446. 1967 Hill MW, Mackenzie IC: Tbe inlluence of subepitlielial connective tissues on epitbelial proliferation in adult mouse. CV// T/.«iii' Res 255:179-182. 1989 Wilson jL, DoUard SC, Cbow LT, Broker TR: Epitbelial-specitic gene expression during diiferentiation of stratified primary human keratinocyte cultures. Cf//C;riiM'(/i D/Z/h 3:471-483, 1992 Sanqner S, Coulomb B, Lebreton C, Dubertret L: Human dernial fibroblasts modnlate tbe effects of retinoids on epidermal growth. J Invest Dctitiatol 95:7(1(1-7114, 199(1 Oda D, Watson E: Human oral epitbelial ciiltnre I: Improved conditions for reprodncible culture in serum-free medinin. In Vitw CeU Dev liiot 26:589-595. 1 990 Asselineau D. Prunieras M: Reconstruction of 'simplified' skin: control of fabrication. BrJ Denmitot lll(snppl 27):219-222, 1984 Cooper JT. Goldstein S: Skin biopsy and successful fibroblast culture (letter). LdHfrt 2:673, 1973 Rotbblat GH. Arbogast LY. Ouellette L. Howard BV: Preparation of delipidized serum protein for use in cell culture systems, hi Vitro 12:554-557. 1976 Bieri JG. Tolliver TJ. Catignani GL: Simultaneons determination of alpbatocopherol and retinol in plasma or red cells by bigh pressure liquid chromatography. Atn J Ctitt Ni(» 32:214.3-2149. 1979 Hsu SM. Raine L, Fanger H: Use of avidin-hiotin-peroxidase complex (ABC) in innnnnoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures. J H/.t/iii/zfi/i Cytochetti 29:577-580, 1981 Dale BA, Holbrook KA. Kimball JR, Hoff M, Sun TT: Expression of epidennal keratins and filaggriu dnring human fetal skin development. / Cetl liiot 101:1257-1269, 1985 " Laeminli UK: Cleavage of structural proteins during the assemhly ofthe head of bacteriophage T4. Natme 227:680-685, 1970

37. 38. 39. 40.

41.

42. 43.

44.

45.

46.

47.

48.

49.

553

Lillie JH. MacCallum DK, Jepsen A: Growth of stratified squamous epithelium on reconstituted extracellular matrices. J hivest Dermatot 90:100-109. 1988 Sawaf MH. Ouhayonn JP, Forest N: Cytokeratin profiles in oral epitheha: a review and a new classification. / Biol Hitccale 19:187-198, 1991 Lindberg K, Rliehiwald JG: Tbree distinct keratinocyte subtypes identified in buman oral epithelium by their patterns of keratin expression in culture and in xenografts. Diffcretttidtioii 45:230-241. 1990 OuhayounJP. Sawaf MH. GofflauxJC. Etienne D. Forest N: Re-epithehalization of a palatal connective tissue graft trimsplanted iu a non-keratinized alveolar mucosa: a histological and biocbemical study in linmans. / Periodotttal Res 23:127-133. 1988 De Luca M, Alhanese E. Megua M, Cancedda R, Mangiante PE. Cadoni A. Franzi A l \ Evidence that lumian oral epithelium reconstituted in vitro and transplanted onto patients with defects in the oral mucosa retains properties of the original donor site. Trausplatttatioii 50:454-459. 1990 Karring T, Lang NP, Loe H: The role of gingival connective tissue in determining epithelial difl'erentiation. / l'eiioiloutal Res 10:1-11. 1975 Mackenzie IC. Hill MW: Connective tissue iiiHueiices on patterns of epithelial architecture and keratinization in skin and oral mucosa ofthe adult mouse. CeU Tisstie Res 235:551-559. 1984 Boisnic S, Branchet M-C. Beranger J-Y. Szpirglas H, Frances C. Le Charpender Y: Corinin fihrohlasts influence the development of oral leukoplakia in an in vitro leconstitution model. EiirJ Dermatol 3:124-131. 1993 Hu L, Crowe DL, lUleiuwald JG. Chambon P, Gudas LJ: Abnormal expression of retinoic acid receptors and keratin 19 by human oral and epidennal sqnainous cell carcinoma cell lines. Cancer Res 51:3972-3981. 1991 Crowe DL. Hu L. Gudas LJ. RheinwaldJG: Variahle expression of retinoic acid receptor (RAR heta) mRNA in human oral and epidermal keratinocytes; relation to keratin 19 expression and keratinization potential. Dijferctttiatioii 48:199-208, 1991 Boukainp 1'. Bieitkrentz D. Stark H-J. Fusenig NE: Mesenclnine-mediated and endogenous regulation of growth and differentiation of human skin keratinocytes derived from different hody sites. Dijfereittiatiott 44:150-161, 1990 Dale BA, Gown AM. Fleckman P. Kimball JR, Resing KA: Characterization of two mom.clonal antibodies to hiiinan epidermal keratohyalin: reactivity with filaggriu and related proteins. J Itivest Deivuttot 88:306-313. 1987 Fleckman P. Dale BA, Holbrook KA: Profilaggrin, a bigh-molecular-weigbt precursor of filaggrin in human epideniiis and cultured keratinocytes. J Ittvest Dcmirjro/85:507-512, 1985