Differentiation-Dependent Expression of Keratins Human Oral Epithelia

. tn

Henrik C lausen, D.D.S ., Poul Vedtofte, D.D.S., Dennis M oe, Ph.D., Erik Dabclstcen, D.D.S., Tuna-Tien Sun , Ph .D., and Beverly Dale, Ph .D.

0 Departments of Ora l Di3gnosis, Surgery, and Biochemistry. Hoyal Denta l College (HC. PV. DM, ED), Copenh agen. Dc111mrk; Departments of Dermatology and Pharmacology, New York University (T-TS), New York, New York; and Depart ments of Periodonti cs, Ora l Biology, and MedJcmc/ Dermato logy, UmverS!ty of Washin gton (BD). Seattle, Washington, U . . A.

The polypeptide composition of epith elial keratins va ries with the state of differentiation. T he epithelia lining the human oral cavity show regional variations in their his­tology . In the present study, paired sa mples of nonker~­tinized buccal epithehum and keratlmzed hard palate epi­thelium were analyzed by sodium dodecyl sul fate­polyacrylamide gel electrophoresis and by immunoblots wi th monoclonal antibodies AEl, AE2, and AE3, and re­sults were correlated with immunofluorescence staining of tissue sections of the same sa mples . Keratins from hard palate (M, 67K, 63-65K, SSK, 56 .5K, 56K, SOK, 48K) and epidermis (M,. 67K, 63- 6SK, SSK, 56.5K, SOK) were sun­ilar to each other but distinctly different from those of buccal epithelium (major bands of M, 52K and 59K, minor bands of SOK and SSK). The immunoblot analysis further indicated the similarity of hard palate and epidermal ker­atins, in contrast to those of buccal epithelium. Each ora l

The differentiation of the epitheliLlll1 lining the ora l cavity shows regio nal vanat1on w 1th respect to the extent and type ofkeratinization (1 ,2). T he epithelium of hard palate is mainl y o rthokeratini zed similar to that of epidermis, while the epitheliu m of the buccal

mucosa and floor of the mouth are nonkeratinized. The differ­entiation pattern is defined histologicall y by th e presence o r ab­sence of a nucleated (parakeratinization) or anucleated stratum corn eum w ith a stratum granu losum (o rthokeratiniza tion), and is clearly reflected in th e synthesis of keratin and fi laggrin (3-10]. Because the bucca l and hard pal ate epithelia are derived from th e same germ layer, and represent the extremes of differentiation of stratified squamous epithelia, they prov ide a unique m atenal for study of epithelial diffe ren tiation. .

It has been established that all epithelial cells contam m em bers of the keratin fami ly of protein s that form intermediate fi lam ents

Manuscript received March I 2, 1985; accepted for pub lica tion O ctober 11 ' 1985.

Supported by grants from Ingeborg og Leo Dannin Fondcn and Vera og Carl Johan Michaelsens Lcga t, Denmark (HC and ED), by NIDH grant D£04660 (BAD), and by National Institutes of Health grant AM3541 I (T-TS).

Reprint requests to: Dr. Henrik Clausen, Department of Oral Diag­nosis, Jagtvej 160, Copenhagen DK-2200, Denmark.

Abbreviations: ABC: avidin-biotin-peroxidase complex PAGE: polyacrylamide gel electrophoresis PBS: phosphate-buffered saline SDS : sodium dodccyl sulfide

tissue exp ressed ke ratins of the type I (AE 1, acidic) subfam­il y and type II (AE3, bas ic) sub fami ly. In tissue sections, the predominant stainin g patte rn for nonkeratinized buccal epithelium was: AE1, positive in the basal layer; AE2, neg­ative; AE3, positive in all layers. In contra st, the staining pattern for keratinized palatal epithelium was: AE1 and AE2, positive in the suprabasa l laye rs; AE3, positive in all layers. Strong suprabasal AEl stain ing in pa late may be related to the presence of the 48K keratin. Some buccal sa mples showed an alternate staining pattern of spotty su­pra basa l stainin g with AEl and AE2 which was co rrelated with the expression of the 56. SK and 63-67K keratins , as well as filaggrin . These results suggest differentiation-spe­cifi c expression of the keratins and show immunologica ll y detectable variation in the apparently normal differentiation pattern of nonkeratinized bu ccal epithelium. J hwcst Der­matol 86:249-254, 1986

of the cy toskeleto n 111 - 14] . Two subfamilies of keratins arc de­fined by immunoreactivity, isoelect ri c poin t, and eD NA h yb rid­ization group j1 5-17]. The ex pression of an acidic (type I) keratin is freq uentl y paired with that of a specific basic (type II) keratin [ 16, 17]. The expression of specifi c keratin s appears to depend on the type of tiss ue , as well as on the state of differentiation or development, certain extrinsic factors , and pathologic conditions [3, 13, 14, 18-26].

The monoclonal antibodies A£1, A£2, and AE3 [5] have been extrem ely useful in esta blishi ng the correlation of keratin poly­peptide exp ression and ti ssue immuno reactiv ity , as recentl y re­viewed by Su n et a! [1 7].

Prev iousl y, it was found that the keratins expressed in kera­tinized hard pa late epi thelium arc similar to those of epidermis , and that the non keratinized bucca l epith elium expressed a dis­tinctly different set of keratins [6, 7 ]. In the present stud y, we used the monoclona l antibodies AE 1, AE2, and AE3 to de fmc kerat in subfamily members present in normal palate and buccal mucosa , and to correlate the subunit specificity with the im­muno histochemical staining pattern on the same tiss ue samples. This should serve as a normal baselin e with which changes in patho logic conditions can be co1npared .

MATERIALS AND METHODS

Ten pa ired biopsies of clinica ll y norm al hum an hard palate and bucca l mu cosa were obtain ed from patients undergoing surgery. Three- or fo ur-millimeter punch biopsies were taken from the fo llowi ng reg ions : (l ) hard pa late, posteri or to the rugae, and latera l of the midline; (2) buccal mu cosa, below the occlusa l line and above the sul cus in the regio n of the first premolar; and (3)

0022-202)(/86/$03.50 Copyright © 1986 by T he Society for Investigative Dermatology, Inc.

249

250 CLAUSEN ET AL

epidermi s, fl ank reg io n (fo r immunoch emistr y) o r neonatal fo re­skin (fo r keratin ex trac tion). All biopsies were immed iatel y fro zen in isopentane, precoo led to -70°C, and sto red at - 80°C. In an attempt to correlate the immunohi stoc hemica l stainin g pattern s of th e monoclo na l antibodies w ith th eir subunit specifi cit y, 6 o f the o ral tissue sa mples were f1rst cut as frozen sections; the residu al materi al was w ashed in phosph ate-buffered sa line (PBS) and then processed as described fo r keratin extra cti on (sec be low). T wo additi onal sa mpl es of each tissue were Ca rn oy's-fi xed and used as co ntro l fo r the immun o histochemi cal stainin g o f frozen sec­tions.

Antibodies Three mouse mon oclo nal antibodies (AE1. AE2, AE3) we re used [SJ. The specificity to wa rd keratin polypeptides from a number of ti ssues except hum an o ral epitheli a have been described earlier [3-5[. Po lyclon:d antifilaggrin was prepa red as previo usly described [271.

Imm.unohistochentical Technique Frozen sectio ns of 8 fLm we re rin sed in phos phate-buffered sa line (PBS) fo r 5 min , air­dried, and in cubated w ith the mo nocl o nal antibod ies diluted 1 :20 and '1: I 00 for 18 h at 4°C, washed 3 times in PBS for 15 min , and stained with Auon:scein isothiocy:mate-conjuga ted rabbit anti­mouse lgG (Dako, Copenhagen) or swine antirab bit , and mounted in g lycerol and p-phenylenediamine (pH 8.0). The sections were exa mined in a Leitz O rth olu x 12 phoco mi croscope with epiflu­orescence and w ith a Xenon 75-W lamp as lig ht source. Nei gh­bo rin g sectio ns we re stained w ith hem atoxylin-eos in in o rder to correlate the histo logic features with th e im mun oAu o rescent stainin g . Ca rn oy's-fi xed ti ssues were e mbedded in paraffin, sec­tioned, and stained by the avidin-bio tin-peroxidase complex (ABC) meth od using the m onoclo nal antik erat in antibodi es and biotin­conju gated horse antim ouse lgG and lg M [28 J. The m onoclonal antibodies AE 1 and AE3 sta ined identi ca ll y, irres pective of fi x­ati on technique. AE2 showed un exp lained differen ces on o ral ti ss ues. Sin ce AB C immuno peroxid ase stainin g of fi xed sections of ora l epithelia correlated w ith immuno blo t data, res ults o btained w ith this m ethod onl y arc considered in the present report.

Keratin Extraction E pitheli um was sepa rated from th e con­nective tissue by 2-h in cubatio n in 20 m M EDTA in PBS at 37°C [29]. T he separated epithelia were ho m ogenized in buffer A (I 50 mM N aCI, 5 m M EDTA , 0.4 rnM phen ylmeth ylsulfon yl flu o ride, 10 mM T ris-HC I, pH 7.4) in a Dounce hom ogenize r, and sub­sequently extracted in buffer B (0 .6 M Kl , 1% [v/v·J Triton X-100, 5 mM EDTA, 0.4 m M phenylmeth ylsulfon yl flu o ride, 10 mM Tris-HC I, pH 7.4), and buffer C (1.5 M KC I in stead of 0 .6 M Kl, other co mpo nents as buffer B) J7 , 13 [. Rem ainin g m aterial was solubilized either in sa m ple buffer containin g 3% sodium dodecy l sul fate (S l S) o r the sa m e buffer w ith the additio n of 8 M urea. after the addition of 10 m M di thiothrcitol and 25 mM iodoacctamidc , by bo ilin g for 2-3 min . Contro ls extracted in a similar manner but w ith out additi on of iodoacetamid e we re in­cluded. These were redu ced by bo ilin g w ith SDS and 2-mercap­toethanol just pri o r to electro pho res is .

Gel Electrophoresis and lmmunoblot Samples were ana­lyzed b y SDS-po lyacry lam id e gel elect ropho resis (PA GE) in a discon tinuous buffer system acco rdin g to Lae mmli [30 J on a 7. 5-12% linea r acrylamide g radient, th en stain ed w ith Coomassie Bri lli ant B lu e. Pro teins on dupli cate un stained gels we re tran s­ferred in Tris- g lycine buffer plus 20% m ethano l onto nitrocel­lulose paper (Schlei cher and Schull ASS, po re size 0.45 fLm ) J31l . The nitrocellulose paper was then in cubated seq uentiall y in 3% bovine se rum albumin for 1 h, antik e rat in antibody for 2 h, and peroxidase-antiperox id ase co mplex (PAP, Sternberger-Mcycr, Jarretsv ill e, M aryland) for I h. All diluti ons and rin ses were done in 50 m M Tri s-HC I, pH 7.6, 150 m M N aC I. The membrane was rinsed well with bu ffer between each in cubat ion . After a final buffer rinse, the blo t w as in cubated w ith freshl y prepared sub­strate soluti on containin g 3,3-diamin obenzidinc-HC I (0.6 m g/ ml) and 0.005% hydrogen perox ide.

TH E JOURNA L OF IN VESTIGATIVE DERMATOLOGY

RES ULTS

Specificity of the Monoclonal Antibodies with Oral Keratins The reactivity of th e m onoclo nal an tib od ies w ith ker­atin polypeptides from hum an ora l epithelia was determined b th e immunoblo t technique f3ll . Six of the paired samples of h ard palate and buccal mu cosa we re tested; 3 of each type a re shown in Fi g 1. Two samples of hum an ne w bo rn foreskin were included as controls. The Coom assie Blue-stained gels (SDS-PA GE) showed that bucca l mu cosa consistentl y ex pressed 3 m ain co mpo nents of M , 52K, 57-58 1<. and 59K (F ig I , lanes 1-3) . The h ard palate mu cosa exp ressed 6 m aj or components of M , 481<, SOK, 56.5K 58 1<. , 63-65 1<. , and 67 1<. (Fig I , lanes 4- 6) . co nsistent w ith preli m­in ary res ults [6,7 1. The controls from epidermis ex pressed 4 maj o r co mponents of M,. SO K, 56.5 1< , SS K, and 67 1<. (also 63-65K, wea k) , in accord w ith ea rlier findin gs [4,5, 14,32,33 ]. Controls of the extra ction and carboxymethylation procedure showed no clunge in the ke ratins detectable b y stainin g o r immuno reactivity in an y of th e in ves ti ga ted epitheli a (not shown).

Antibody AEl lmmuno blo t anal ys is n.: vea led th at A E 1 reacted w ith th e low-M,. keratin s from all tissues investiga ted . A s d e­scribed ea rli er, th e SOK and 56.5 1< kera tin s fro m epidermis were stained (Fi g 1, lanes 7-8) [SJ . In bucca l mucosa, the 52K kera tin, as w ell as a fa int band o f SOK wh ich was just barel y v isible on the stain ed gel, i. e., lane 3 in Fig 1 ). rea cted w ith AE I (Fig 1.

Buccal Palate Epid 67

~65 =::;V 63 =:::::.58 ~56.5

50 48

___/ 56.5 52- -.....,50

48

_....--67 -63 ---...... 56.5

AE2 67

59....._ ~65 -63

58.r- -......58

AE3 1 2 3 4 5 6 7 8

Figure 1. Ana lys is of oral kerat ins by SDS gradient gel dccrrophoresi and immunoblots. La11e.< 1-3, bu cca l epithelium; lm1 cs 4-6, palatal epi­thelium ; lanes 7 and If, epidermis. Note that all samples contain both AEI­reactivc (ac idi c. type I) and AE3-rcactivc (basic, type II) keratins. M o­lecular weights ( X 10 - 3) are indica ted by numbers. The trace staining of the 63-67 1< ke rat ins by AE I and 48K keratin by AE2 is due ro use of asci tes Auid and was abscJH when hybridoma culture supernatants were used.

VOL. 86. NO. 3 MARCH 1986 KERATIN IN HUMAN ORAL EPITHELIA 251

Table I. Keratins Expressed in Oral Epithelia

Basic, Type II Subfami ly"

Tissue 63-67K' 59K 58K

(1,2)" (4) (5)

Buccal + + Hard palate + + Epide rmis + +

' React with AE3. 'React with AEI. 'Also react wi th AE2. "Numbers in parentheses, catalog numbers, according to Moll et all14] .

Janes 1-3). In hard palate, th e 48K, SOK, and 56.5K keratins (Fig 1 lanes 4-6) were stained with AE1. These 3 keratins of palate a~e present in similar am ounts in the Coomassie Blue-stained gel, but the 48K keratin has a much m o re intense reaction with the antibody than either th e SOK or 56.5K keratins [5, 16]. AEJ stain­ing of a 48K band in bucca l ex trac ts was o bserved in 3 samples (e. g. , Fig 1, lane 2), and the 56. SK band was seen 111 1 sample (also see Fig 3£, lane b , below). The 48K keratm was not seen in epi derm al ex tracts; howeve r, th1s kcratm has prev iously. been observed in variab le amounts in heel epidermis, hypcrprolifcra­tive epidermis, and cultured cells (5 , 13,26.34].

Antibody AE2 AE2 reacted predominantly with the high-M, keratins 63-65K and 67K found in both hard palate and ep idermis (Fig 1, lanes 4-6) and with the 56.5K keratin also stained by AE1. Jn general, the stai nin g of buccal extracts was nega tive or very weak; however, AE2-positive bands were detectable in a few of the buccal samples (Fig 1, lane 2, and also see Fig 3£, lane b).

Antibody AE3 AE3 stained 58K and the hig h-M, keratins 63-651< and 67K from both palate and epidermis (Fig 1, lanes 4-8). AE3 also reacted w ith the 56K keratin in palate. The 57-581< and 59K bands from buccal mucosa were consistentl y stained vvith AE3, w hereas sta ining of a 63-65K band varied in this tissue (Fig .I , lanes 1-3). Althoug h the 57-58K and 59K bands found in bucca l extracts did not ap pea r to co mig rate with the SSK band fro m epidermis, both 57-58K and 59K bands sta ined wi th AE3. The immunoblot data are summarized in Table I. Based on lllllllU­

noreactivity and M,., the o ral keratins were cata logued tentatively acco rdin g to Mo ll ct al (Table I) [14] .

Tissue Staining AE I consistentl y stained onl y supra basal cell s in keratinized ha rd palate ep ithelia (Fig 2A) in contrast to ep i­dermis [3,5] in w hi ch it sta ins on ly the basa l layer. T he staining of bucca l epithelium was generall y localized to the basal layer (Fig 3A). However, considera ble var iation was found ; some sam­ples s h owed a few isolated positive cell s in stratum sp mosum , o r a more patchy distribution of positive cell s in the entire la yer. One sa n1ple ac tu all y showed stronger sta ining of the spino us layer than that of th e b::tsal layer (Fig 3D) .

Acidic, Type I Subfamilyh

56K 56.5K' 52K SOK 48K

(6) (10) (13) (14) (16)

+ + + + + +

+ +

AE2 sta ined all suprabasal cel l layers in hard palate (Fig 2B), consistent with previous results for epidermis [3,5], whereas buc­cal epithelium showed a weak to negative reaction (Fig 3B).

AE3 sta ined throughout the en tire epithelium of both types of ora l tissues (Figs 2C, 3C).

The va riation in immunohistochemi cal stainin g of buccal sam­ples suggested an altered differentiation pattern in so me of the sa m ples. Therefore, keratins from buccal samples representative of the 2 AE1 staining patterns, basal (Fig 3A) vs suprabasal (Fig 3D), were directly compared by immuno blot analysis (Fig 3£, lanes a and b, respectively). Samples with the most frequent stain­ing pattern (Fig 3A; Fig 3£, lane a) sho wed the 52K, 57-58K, and 59K keratins, w hile sa mples w ith the less frequent pattern (Fig 3D; Fig 3£, lane b) also contained 56.5K and 63- 67K keratins.

Because the occurrence of the keratin-associated protein, fi lag­grin , is strongly correlated with keratinization [1 0], it was used to independentl y test for individual va riation within the samples of buccal mu cosa . Most buccal samples showed nega tive to very weak, granular staining (Fig 4A). One exception showed areas w ith strong staining, especially in connection with epithelial ridges (Fig 4B). This sample also showed suprabasa l AE1 stai nin g.

D ISCUSSION

In the present study we ha ve confirm ed that the pattern of keratin synthesis in human oral orthokerat inized epithelium trom hard palate is very si milar to that produced in epidermis, but the pattern of keratin synthesis in oral no nkeratinized buccal epithelium is distin ctly different from the keratinized epithelia (Fig 1) [6, 7]. The immunoreactivity of the hard palate keratins was almos t identi ca l to that from epidermi s, further indicating the similarity of these 2 sets of keratins (Fig 1) [3,5 , 14,32].

The concept of expression of specific keratins as markers for specific type of differentiation is supported by the results of this study of oral mu cosa. TheM, 65-67K and 56.5K keratins, mark­ers for keratinized epitheli a, are found in palate (ortho kcrat inizcd) but not buccal (nonkeratinizcd) epithelia. In contrast, the major keratins of bucca l epithelia arc 52K and 59K. The 50K and 58K keratins , markers for stratified epithelia, are found in both palate

Figure 2. Immunohistochemical staining of hard palate epithelia. A, Immunofl uorescence staining of stratum spinosu m and granulosum by AE1; B, immunoperoxidase staining of all cell layers except basal cells by AE2; C, im­munofl uorescence staining of the enti re epi­thelium by AE3.

252 C LAUSEN ET A L

Figure 3. Immuno histochemi cal s t ~tinin ~ of bucca l epithe lia. A. Typica l intmu ll oftu o rcs­cencc stainin g o f basal cells by AE I; B. wea k to m:g:Hive i1n 1nunopcroxicbsc st:t ining by AE2; C, immunoflu o rescence sta inin g of the entire epi thelium by AE3; D, va riation ol AE I ilnmuno Au orcsccncc st::l inin g showi ng in­tense reaction with spin ous cel ls :111d alm ost negative reaction w ith basa l ce ll s: lower right paud, immun oblo t ana lys is o l tissue ex tracts from sampics shown in (A) (laue a) and (D) (laue b) w ith AE I , A E2. and A E3.

and bucca l epithe li a; however, they a rc prese nt in ve r y s m all quamitics in bu cca l epith e lium. A lth o u gh th e bu cca l 57-58K band docs no t com ig rate ex act ly w ith t he 581<. in t he p:d ate , bo th sta in w ith AE3, indi ca tin g the 2 bands a rc closeiy re lated . The im­mun o b lo t with AE 1 (specific fo r m ost o f the t ype I acid ic ke ra tin s) and AE3 (spccif1c for the t ype II b as ic ke ratin s) showed th at t he o ra l ep ithel ia express members of bo th t he ac idi c and bas ic subfam ili es (Fig 1), in ag reement w ith th e current con ce pt th at m embers of both sub fa mili es arc ex pressed in an y epithe li a l ce ll [4, 17, 19.35,361.

Severa l studies have shown that the express io n o f kerat in s w ithin a g iven cell chan ges durin g epi th elia l d ifferentiation 13.5,2 '1 ,32,

Figure 4. lmmu no Au oresccncc staining with antibody to hum an fi lag­grin . A , IJucca l epithelium with :t n csscnti :t ll y nega ti ve reaction; B, buccal epithelium show ing a pos iti ve suprabasal reaction associated with ti ssue ridges.

THE JOU HNA L OF INV ESTIGATIV E DERMATOLOGY

56.5........_ 52-48/

-AE1 AE2 AE3

65 :f-63 ~59

58

.37,38 1. In epid erm is , the SO K and 581<. ke ratin s arc expressed in th e basa l la ye r , w h ereas th e 56.5K and 65-67K keratins arc mainl y sy n thes ized in the sp in o us and g ranular hyers [3,5,32 1. T he sim­ilarity o f th e kerat in s fou n d in epide rmis and lnrd palate epith e­li um sugges ts that a s im il a r o rg ani za tion exis ts in the kera tinized o ra i epithe li a . The sim ple pattern found in bu cca l epith elium , in para ll el , co u ld indi ca te tint the SOK and 58K ke ratin s arc syn­thesized in the basa l layer , w hereas the 521<. and S':I K ma y b e the pair ex pres sed supraba sa ll y in thi s ti ss ue .

T he m aj o r kc r:ttin s found in the no n kerati nized bucca l epithelia, 5':11<. (Moll ct a l [1 4 l no . 4) and 52 1<. (pres umabl y M o ll ct al [1 4], no . 13 , 54K), ha ve been fo un d to also be the m;J_jor keratins of no n kcra tin izcd st ratified sq uamous epithelia of interna l orga n . such as esopha g us, epi g lo tti s. and ce rvix , w hich a rc derived from endode r m 11 3, 14 , 17,33 1. Si nce the bucca l epithe lium is ecroder­m all y deriv ed , this indi ca tes th at ke rat in exp ress io n of linin g ep­ithe lia is co rrcbted w ith t he type of differen tiation, ra ther th an germ- laye r o rig in .

As ex pected from cl:!ta p rev iously obtained abo u t t he ep idermis lS I, s tri ct co rre lat io n of the kerati n patte rn and the immunohis­to logic loca lizat io n detected by th e m o noclo n al ai)tibodies wa no t poss ible. Positive immu nofl uo rescen ce s tainin g could alw ay be accou n ted fo r, w hereas nega ti ve stainin g in som e cases has been shown to be d ue to nusk in g 15 ,39 1.

I3a sa l s tainin g in bu cca l epithelium and epiderm is with AEl i consistent w ith the presence of th e SOK keratin (Table I) [3,5]. but the absence of suprabasa l A E l s ta inin g o f the 52K in bu ccal ep ithe lium co uld be due to m ask in g , as su gges ted for the 56.5K in ep ide rmi s 151. P alate epithel ium was, howeve r, s ta ined only in hi g he r cell layers (spi no us and g ranular) by AEI . T h e striking diffe rence in sta inin g patte rn w ith AE 1 in the 2 keratini zed epi­thelia (pa late and epidermi s) is m os t pro bably due to the presence o f th e 481<. keratin in h ard pa late, sin ce t he kerat ins oth crvvise are idcmica l. T he 481< k eratin h as been sh own to b e associated with

V OL. 86. N O. 3 MARC H 1986

h y perplasia in epidermis [26,40,41] . The presence of the 48K keratin in epidermal disorders changes the AE J immunoreactivity from the norm al basal stainin g to a suprabasal staining pattern [26] as we see in normal palate epithelium . The presence of the 48 K k eratin apparently facilitates AEl immunoreactivity in spi­nous and g ranular cells, either by its own presence and strong AEl reactiv ity (see Figs 1, 2) or by dem asking AEJ epitopes o n the 56. SK keratin abundant in these cell layers . The 48K keratin is also found in norm al heel epidermis, which also show supra­basa l s taining with AEl (Woodcock-Mitchell , Sun , unpublished observations). Altho ug h palate epithelium is not considered h y­perproliferative as in epidermal disorders studied by Weiss et al [26, 42], the epithelium of both palate and heel are much thick er than normal epidermis.

Differences in the immunolog ic stainin g pattern of the o ral epithelia w ere also o bserved w ith AE2. The generally w ea k to n egative staining o f buccal epitheli a was consistent with the im­munoblot d ata and with the nonkeratinized differentiation pattern of this tissue. AE2-staincd spinous and g ranular cells in hard pal ate are similar to staining of epidermis. Both tissues are keratinized, genera te a stratum corneum, and contain the 56. 5K and 67K keratins.

AE3 stainin g o f the entire epithelium of all investi gated epithelia is consistent with the immunoblo t staining results of a strong 59K k eratin in buccal mucosa and the 58K and 67K keratins in pala t e. AE3 also reacts with the 56K keratin previousl y identified in h y perproli fe rative epithelia. However, this does not result in a unique AE3 tissue-staining reaction fo r palate in contras t to AE1 (see a bove).

The immunohistologic sta ining patterns were consistent in all individuals for th e palate epithelium , but va riation was noted in b ucc al mu cosa . The variable suprabasal staining found in the bu cc a l sa mples (Fi g 3A,D) tended to correlate with the presence of the 56.5K and 63-65K keratin s and slight increase in the 48K keratin (Fig 18, lane 2; Fig 3E, lane b). This could indicate that th es e keratins are expressed in some bucca l epithelial cells, i. e., cells s tartin g to keratinize at a molecular level. The presence o f 63- 6 5K keratins could be due to contamination with stratum corne um ; however, another marker for keratinization, fil aggrin [10], identified b y immunofluorescence in some tissue samples, is consistent with thi s interpretation o f variation in the differ­entia tion o f cells in buccal mu cosa.

Buccal epithelium is more exposed to fri ction and subsequent subclinical chan ges to ward keratinization th an , for example, the fl o or of the mouth . The latter may be a m o re invariant source of o ra l nonkeratinized epithelia [7) . It sho uld be no ted that traditi onal his tolo gic evaluation o f these buccal samples show ed that all w ere normall y nonkeratinized epithelia , and no ch ara cteristics o f ker­atiniza tio n w ere found . The samples w ere also analyzed fo r blood­g roup cell surface antigens, previo usl y shown to express different staining patterns correlated with the sta ge of histologic differ­enti a tion [43]. These buccal sa mples sho w ed the staining pattern normally fo und in nonkeratinized epithelia (Vedto fte, C lausen, Dab e lsteen, unpub lished observations) . Thus it appea rs that cyro­skel e t al m aturation , i.e. , keratin and fil aggrin synthesis, ma y pre­ced e the morphologic changes as w ell as cell surface changes associated with change of differentiation pattern from nonkera­tinize d to keratinized .

The regional variation of th e human o ral mu cosa provides a unique material for the study o f fea tures related to epithelial dif­fe re ntiation . In th e present report we show that distin ct patterns of k e ratin synthesis and immunoreactivity with keratin subset­specifi c antibodies is correlated with different keratiniza tion pat­terns. The results obtained will be of value as a baseline fo r furth er studies of keratin synthesis and immuno reactivity in various be­nign and m alig nant keratiniza tion disorders in the o ral mucosa.

The excellerll tee/mien/ nssistnr1ce of } nllct Kimbn/1 n11d ]11/ie Scofield is grntt:Jidly

acknowledged.

KERATIN IN H U MA N O RAL EPITHELI A 253

RE FERENCES

1. Squier CA, Johnson NW, Hopps RM: Hum an O ral Mucosa: De­velopment, Structure and Function. Blackwell Scientific Publi­ca tions, Oxford, 1976, pp 7- 44

2. Schroeder HE: Difte rentiation of Hum an Oral Stratified Epithelia. Basel, Karger, 1981, pp 35- 119

3. Sun T-T, Eichner R, Nelson WG , Tseng SCG, Weiss RA,Jarvinen M, Woodcock-Mitchell J: Keratin classes: molecular markers for different types of epithelial d iffc rcnti :~ tion . J In ves t Derm a to! 81 (suppl): 109S-11 5S, 1983

4. Tseng SCG, Jarvinen MJ , Nelson WG, Huang J-W, Woodcock­Mitchell J , Sun T-T: Co rrelation of specific keratins with different types of epithelial differentiation: monoclonal antibody studies. Cell 30:361-372, 1982

5. Woodcock-Mi tchell J, Eichner R, Nelson WG, Sun T -T: Immuno­localiza tion of keratin polypeptides in human epidermis using monoclonal antibodies. J Cell Bioi 95:580- 588, 1982

6. Clausen H, Vedtofte P, Moe D, Dabelsteen E: Keratin pattern in human oral buccal and hard palate mucosa . Scand J Dent Res 91:411-41 3, 1983

7. C lausen H, Moe D, Buschard K. Dabelsteen E: Keratin proteins in hum an oral mucosa. J O ral Pathol. in press

8. Sibrack LA , Gray RH , Bernstein lA: Localiza tion of the histidine­rich protein in keratohyalin: a morphological and macromolecular marker in epiderm al di fferentiation. J In ves t Dermatol 62:394-405, 1974

9. Dale BA, Holbrook KA, Steinert PM: Assembly of stratum corneum bas ic protein and keratin fil aments in macrofibril s. Nature 276:223- 227, 1978

10. Smith SA, Dale BA: Immunohistochemical localiza tion of fila ggrin in hum an oral epithelium and correl ation with keratinization. J Inves t Dcrm atol 86:168-1 72, 1986

11. Franke WW, Schmid E, Osborn M, Weber K: Different interme­diate-sized filan)ents distinguished by immunofluorescence mi­croscopy. Proc Nat! Acad Sci USA 75:5034-5038, 1978

12. Sun T-T, Green H: Immunoflu orescent sta ining of keratin fibers in cultured ce lls. Cell 14:469- 476, 1978

13. Franke WW, Schiller DL, Moll R, Winter S, Schmid E, Engelbrecht 1: Diversity of cytokeratins. Di fferentiation specific express ion of cytokeratin polypeptides in epithelial cells and tissues. J Mol Bioi 153:933-959, 1981

14. Moll R, Franke WW, Schiller DL: The ca talog of hum an cytoker­atins: patterns of expression in normal epithel ia, tumors and cul­tured cells. Cell 31 :11-24, 1982

IS. Fuchs E, Coopock SM, Green H, Cleveland DW: Two distinct classes of keratin genes and their evolutionary signif1cance. Cell 27: 1033-1042, 198 1

16. Eichner R, Bonitz P, Sun T-T : Classification of epidermal keratins according to their immunoreactivity, isoelectric point, and mode of expression. J Cell Bioi 98: 1388-1396, 1984

17. Sun T-T, Eichner R, Schermer A. Cooper D, Nelson WG. Weiss RA : Class ifica tion, ex pression, and possible mechanisms of evo­lution of mammalian epithelial keratins: a uni fy ing model, Cancer Cell , vol 1, The transfo rmed phenotype. Edited by A Levine, W Torp , G Vandewoude, JD Watson. New York, Cold Spring Har­bor Laboratory, pp 169- 176, 1984

18. Banks-Schlegel SP: Keratin altera tions during embryonic epidermal differentiation: a presage of adult epidermal maturation. J Cell Bioi 93:551-559, 1982

19. Moll R, Moll I, Wies t W: Changes in the pattern of cytokeratin polypeptides in epidermis and hai r follicles during skin develop­ment in human fetuses. Differentiation 23 :1 70-1 78, 1982

20. Doran TI, Vidrich A, Sun T-T: Intrinsic and extrinsic regulation of the differentiation of skin , corneal and esophageal cells. Cell 22: 17- 25, 1980

21. Breitkreutz D, Bohnert A, Herzmann E, Bowden PE, Boukamp P, Fusenig NE: Differentiation specific fun ctions in cultured and transplanted mouse keratinocytes: environmental influences on ul­tras tructure and keratin expression. Differentiation 26: 154- 169, 1984

22. Schweizer J , Winter H. Hill MW, Mackenzie TC: The keratin poly-

254 C LAUSEN ET AL

peptide patterns in hetero typicall y reco m bi ned epithel ia of skin :1nd mucosa of adult mouse. Differentiation 21'i: 144- 153, I \184

23. Bo wden PE, Wood Ej. C unliffe WJ : Co mpari son of prckcratin and k eratin polypeptides in no rmal and psoriati c human epidermis. 13iochim 13i o phys Acta 743: 172- 17\1, 1983

24. Steinert PM , Peck GL, Idler WW : Stru ctural changes o f hum an epidermal kerat in in diso rders o f kerati ni zation , Biochemistr y o f No rma l and Abno rmal Ep ide rmal Diffl:rentiatio n. Edi ted by lA. Uernstein, M Sciji . T okyo. Univ ofTokyo Press, 198 1, pp 39 1-406

25. Lb ning T . Staquct M-J . Thi vo let J , Seifert G: Keratin po lypeptide di stributi on in no rm al and diseased hum an epide rmis and o ral mucosa. Virchows Arch jCe ll l'atho lj 388:273-288, 1980

26. Weiss RA, Eichner R, Sun T-T: Monocl onal an tibody analys is o f keratin expression in ep idermal di seases : a 48- and 56-K Dalton ke rat in as mo lecular markers for h ype rprolifc rative keratin ocy tcs . J Cell Bio l \18 :1397-1406, 1984

27. Sybert VP , Oak BA , Holbroo k KA : Ichth yos is vu lga ris: identifi­cation o f a defect in synthesis o f fila gg rin co rre lated with an ab­sence of kcratohya lin c g ranules. J In vest Dcrmatol 84:191 - 194, 1\185

28. Hsu SM. !{ai nc L, Fan ge r H : Usc of av idin-bio tin-perox idase co nl­p lcx (ABC) in in11nuno pcrox idasc techniques : a co mparison be­tween ABC and un labeled antibody (PAP) procedures. J 1-listo ­che tn Cytochem 2\1:577-580, 198 1

29. Ma ckenzie I C. Dabelstecn E. !{oed-Petersen B: A method fo r stud y­in g cpithe l i~ l -m cse n ch y ma l imeractions in human o ra l mu cosal lesions. Sca nd J Dent l~cs 1)7 :234-243, I \179

30. Lacmmii UK : C leavage of stru ctural proteins durin g the assemb ly of the head o f bacte riophage T4. Nature 227:680- 685 , 1970

3 1. Towbin H , Staehclin T. Gordon J: Elect ro phoreti c tra nsfe r o f pro­teins fro m polyac rylamide gels to nitrocellulose sheets: procedure and some appli ca tions. Proc N at! Acad Sci USA 76:4350-4354, 1979

32. Fuchs E. G reen H : C hanges in keratin gene express ion during ter­mina l differentiation of the kcratinocytc. Cell 19: I 033- 1042. I \180

33. !3anks-Schlegd Si' . H am is OC: Tissue-specifiC ex press ion o f keratin

TI-l E JOU I\NAL OF INVEST IGATIVE DEHMATOLOGY

proteins in hunun esophagc:li and epidermal epithelium and their cultured keratinocytes. Exp C ell Res 146:27 1-280. 1983

34. Nel son WG, Sun T-T: The 50- and Sti-Kda lto n kerati n classes as molecular markers fo r stratifi ed squamous epithelia: cell cul ture studies. J C ell l:lio l \17:244-25 1. 1983

35. Kim Kl-1. f~ heinwa ld J, Fuchs EV: T issue specifi city of epithelial keratin s: differential express io n of m RN As fro m two multigene fa milies. M ol C d l Ui o l 3: 495-502, 19H3

36. Sch iller DL. Franke WW. Geiger B: A subfa mily of relati vd y la rge and basic cyrokcratin po lypeptides as ddined by peptide mapping is represented by one o r seved po lypeptides in epithe li a l ce ll s. EMl:lO J 1:76 1-71'i9, 1982

37. Skerrow D . Skerro w CJ: T onofilamcnt differentiatio n in hum an epidermis: iso lation and polypeptide chain composition of kcra­tin ocyte subpopcd:Hion. Exp Cell Res 143:27-35 , 1983

38. Sch weizer J, Kinjo M. Furstc nbcrge r G. Winter H : Sequential express io n of mRNA-en coded keratin se ts in neonatal mouse ep­ide rmis: basa l ce ll s with propert ies of tcnnina ll y differenti a ting ce ll s. Cell 37 :159- 170, 1984

3\1 . Franke WW. Sch mid E. Wd lstced J, G rund C, G ig i 0, Geiger B: C hange of cy tokeratin fi lam en t o rg:mi z:Hion during the ce ll cycle: sel ective mask in g o f ~ n immun olog ical d~tcrminant in interph ase l'tK 2 ce ll s. J Cell Bioi 'J7 : 1255- 12(>(). I 9tl3

40. M oll H, Moll I, Franke WW: Difference o f exp ression ofcytokcrat in polypeptides in vario us ~pithc lial skin tum ors. Arch Derm atol R c 276:349-364, 1984

-! I . McGuire J, Osbor M, Li ghtfoot L: Two knati ns M W 50,000 and 56, 000 are sy nthes ized by psori at ic epidermis. l:lr J Dcrm ato l Il l (supp l 27) :27- 37, 19R4

42. Weiss HA , Gui llet G Y A, Frcedbcrg IM, Fa rm er EH. Small EA , Weiss MM , Sun T-T: Th~ usc of mo nocl onal an tibod y to ke rarin in human epidermal di seas~: alterati ons in immunohis tochemi cal stainin g patt<:rn . J In ves t Derma to! 8 1:224-230, 1983

43. v,· dtoft ~ P, Dabclstecn E. 1-bkom ori S. Young WW: Regional vari­ations o f cd l surface carbohyd rates in hum an o ral s tratifi ed epi­thdium . Diff~rcn ti arion 25:22 1-228. 198-l