Inhibition of the Transforming Growth Factor-b/Smad Signaling ...

ORIGINAL ARTICLE

Inhibition of the Transforming Growth Factor-b/Smad SignalingPathway in the Epithelium of Oral Lichen

Andreas Karatsaidis, Olav Schreurs,TonyAxeŁ ll,� Kristen Helgeland, and Karl Schenck

Department of Oral Biology and �Department of Clinical Dentistry, Dental Faculty, University of Oslo, Oslo, Norway

The basal cells in epithelium of the erythematous formof oral lichen display hyperproliferation compared withnormal oral mucosa. In this study we examinedwhether this is associated with disrupted production,activation, or signal transduction of the epithelialgrowth inhibitor transforming growth factor (TGF)b1. In situ immunostaining showed that most epithelialcells in normal oral mucosa had nuclear and cytoplas-mic Smad4 and phosphorylated Smad2/3, but expressedlittle or no Smad7. Expression of latency-associatedpeptide TGF-b1, latent TGF-b binding protein 1,TGF-btype I receptor, and TGF-b type II receptor was readilyseen, but only very little TGF-b1 was activated. In er-ythematous oral lichen, basal and lower spinous epithe-lial layers showed staining for latency-associatedpeptide TGF-b1,TGF-b type I receptor, and TGF-b typeII receptor. A band with scanty staining for these mole-

cules, but with marked staining for active TGF-b1, wasseen in the upper spinous and granular layers. Numbersof epithelial cell nuclei with Smad4 and phosphorylatedSmad2/3 staining were signi¢cantly reduced in erythema-tous oral lichen compared with normal oral mucosa. Ba-sal and suprabasal cell layers in erythematous oral lichenshowed strong cytoplasmic Smad7 protein staining, butin spinous and granular layers Smad7 was localized tothe cell membrane. In situ hybridization showed strongSmad7 mRNA expression in almost all basal keratino-cytes in erythematous oral lichen; by contrast, no or oc-casionally very weak Smad7 mRNA expression was seenin these cells in normal oral mucosa.The observations in-dicate that inhibition of the TGF-b/Smad pathway mayaccount for the hyperproliferation of keratinocytes in er-ythematous oral lichen. Key words: epithelial proliferation/keratinocytes/oral mucosa. J Invest Dermatol 121:00 ^00, 2003

Oral lichen (OL), which includes oral lichenoid re-actions and oral lichen planus (OLP), is a com-mon oral disease a¡ecting about 2% of thepopulations examined (Axell and Rundquist,1987; Salonen et al, 1990). Oral lichenoid reactions

are de¢ned as adverse reactions to dental materials, whereaschanges with unknown etiology are termed OLP (Bolewskaet al, 1990). Histologically, oral lichenoid reactions and OLP areindistinguishable from each other (Bolewska and Reibel, 1989).The lesions are characterized by a typical band-like mononuclearin£ammatory in¢ltrate in the connective tissue, dominated byTlymphocytes (Hedberg et al, 1986). Pathologic changes in theepithelium include hyperorthokeratosis or parakeratosis, acantho-sis, atrophy, and liquefaction degeneration of basal cells (Andrea-sen, 1968; Kramer et al, 1978). Clinically, OL shows white papular,reticular, erythematous, plaque, and ulcerative forms (Andreasen,1968; Kramer et al, 1978). Erythematous OL (ERY OL) is one ofthe common forms of the disease and is often associated with oraldiscomfort in the a¡ected persons (Andreasen, 1968).

In order to maintain the strati¢ed epithelium’s normal anatomyand thereby its functions, there is a ¢ne-tuned balance betweenthe continuous proliferation of basal keratinocytes and matura-tion and cell death of terminally di¡erentiating keratinocytes(Gandarillas, 2000). Epithelial proliferation and di¡erentiationare regulated by an intricate signaling network of di¡erent pep-tides (cytokines/growth factors) produced by keratinocytes them-selves, stromal cells, and in¢ltrating in£ammatory cells and theirrespective receptors (Dotto, 1999; Freedberg et al, 2001). Trans-forming growth factor b (TGF-b) has been shown to have im-portant regulatory functions in epithelial proliferation anddi¡erentiation (Roberts, 1998; Ten Dijke et al, 2002). Among thethree TGF-b isoforms, TGF-b1 is known as the most prominentregulator. TGF-b1 inhibits epithelial cell growth through tran-scriptional repression of the growth promoting gene c-myc (Pie-tenpol et al, 1990), and upregulates the cyclin-dependent kinaseinhibitors p15 and p21 (Hannon and Beach, 1994; Datto et al,1995). This leaves the retinoblastoma protein in a hypo-phos-phorylated state and thereby retains the cells in the G1 phase(Weinberg, 1995). TGF-b1 can also induce apoptosis in keratino-cytes (Min et al, 1999), and a¡ect keratinocyte di¡erentiation byupregulating the expression of integrins and keratins (Mans-bridge and Hanawalt, 1988; Jiang et al, 1995; Zambruno et al,1995). The crucial role of TGF-b1 for epithelial development ishighlighted in studies on transgenic mice: the epidermis ofTGF-b1-null mice displays marked hyperproliferation (Glicket al, 1993), whereas mice with targeted overexpression of activeTGF-b1 exhibit severely disrupted skin development, leading toneonatal death (Sellheyer et al, 1993).

Reprint requests to: Andreas Karatsaidis, Department of Oral Biology,PO Box 1052, Blindern, N-0316 Oslo, Norway. Email: [email protected]: ERY OL, erythematous oral lichen; LAP-TGF-b1, la-

tency-associated peptide TGF-b1; LTBP-1, latent TGF-b binding protein 1;NOM, normal oral mucosa; OL, oral lichen; OLP, oral lichen planus;TbRI,TGF-b type I receptor; TbRII,TGF-b type II receptor.

Manuscript received January 31, 2003; revised June 20, 2003; accepted forpublication July 20, 2003

0022-202X/03/$15.00 . Copyright r 2003 by The Society for Investigative Dermatology, Inc.

1

TGF-b1 is predominantly secreted as a large latent proteincomplex, consisting of the mature (active) TGF-b1 homodimer,the latency-associated peptide (LAP), and the high molecularweight latent TGF-b binding protein (LTBP) (Wake¢eld et al,1988; Miyazono et al, 1991). Extracellularly, latent TGF-b1 is pro-teolytically cleaved with the release of activeTGF-b1. On the tar-get cell, active TGF-b1 binds to a hetero-tetrameric complex oftwo transmembrane signaling receptors: TGF-b type I (TbRI)and TGF-b type II (TbRII) receptors (Ebner et al, 1993). TbRIIphosphorylates TbRI upon binding of TGF-b1 (Wrana et al,1994) and TbRI subsequently phosphorylates either of two intra-cellular protein homologs, Smad2 or Smad3 (Nakao et al, 1997b).Phosphorylated Smad2 and Smad3 associate in the cytoplasmwith Smad4 and the resulting complexes translocate to the nu-cleus, where they modulate transcription in collaboration withother coactivators and corepressors (Massague, 2000). This signal-ing pathway, however, can be blocked by binding of Smad7 tothe intracellular domain of the TbRI, which inhibits the phos-phorylation of Smad2 and Smad3 (Nakao et al, 1997a).We have previously reported that ERY OL lesions show

epithelial atrophy and hyperproliferation of basal keratinocytescompared with reticular OL and normal oral mucosa (NOM)epithelium (Karatsaidis et al, 2003). From studies on transgenicmice with null expression of Smad3, or active TGF-b1 constitu-tively targeted to the keratin 10 promoter, or overexpression ofSmad7 (Cui et al, 1995; Ashcroft et al, 1999; He et al, 2002), it isknown that interference with theTGF-b/Smad pathway can causesuch changes. In this study, we show that this pathway is inhib-ited in ERYOL because, concomitantly with an increased activa-tion of TGF-b1, there is an increased Smad7 expression andmembrane translocation, and signi¢cantly reduced nuclear accu-mulation of Smad2, Smad3, and Smad4 in the epithelium ofERY OL. Parallel to the ¢ndings in transgenic mice, this thusmay be the cause for the epithelial hyperproliferation in ERYOL.

MATERIALS AND METHODS

Specimens Biopsies were taken, after informed consent, fromvolunteers with ERY OL (n¼12) or with NOM (n¼11). No distinctionwas made between OLP and oral lichenoid reactions. Clinical diagnosiswas made by an experienced clinician (TA), and pathologic diagnosis wascon¢rmed on hematoxylin and eosin sections of the biopsies. The clinicaland histopathologic criteria used were those described by Kramer et al(1978). Buccal mucosal biopsies were taken from sites that typi¢ed theclinical diagnosis and character of the lesion. The volunteers were testedfor oral Candida infection using Dentocult dip slides (Orion Diagnostica,Espoo, Finland) and the biopsies were examined by PAS staining. Noneof the patients included in this study showed Candida infection at the timeof biopsy. Biopsies were snap-frozen on dry ice (�701C), oriented,embedded in OCT compound (Sakura Finetek, Tokyo, Japan), and storedat �801C. Biopsies for in situ hybridization were ¢xed in 4% bu¡erformaldehyde. Five micron thick sections were cut at �201C in a cryostatand slides were stored at �201C until used. The study was carried outaccording to the Helsinki Declaration’s principles for biomedical researchand approved by the Ethical Committee of Health, Oslo, Norway.

Immunohistology Antibodies against the following proteins were usedin this study: the mature and active TGF-b1 (mouse IgG, 1.25 mg per ml,BioSource International, Camarillo, CA), LAP (TGF-b1) (goat IgG, 1.3 mgper ml, R&D, Oxfordshire, UK), LTBP-1 (rabbit IgG, 0.3 mg per ml, akind gift from Dr C.-H. Heldin, Ludwig Institute for Cancer Research,Uppsala, Sweden), keratinocyte transglutaminase (mouse IgG, 0.45 mg perml, Biomedical Technologies, Stoughton, MA). The following antibodiesfrom Santa Cruz Biotechnology (Santa Cruz, CA) were used: TGF-bRI(rabbit IgG, 0.33 mg per ml), TGF-bRII (rabbit IgG, 0.25 mg per ml),phosphorylated Smad2/3 (rabbit IgG, 0.25 mg per ml), Smad4 (mouse IgG,2 mg per ml), Smad7 (goat IgG, 4 mg per ml), plasminogen activatorinhibitor 1 (PAI-1) (mouse IgG, 1 mg per ml).Single stainings were performed after ¢xation of cryosections in 4%

paraformaldehyde for 20 min, followed by 2� 5 min washes inphosphate-bu¡ered saline (PBS) and preincubation with 0.3% H2O2 for

30 min. Sections were rewashed and incubated in 5% of appropriatenormal serum for 30 min (the species of the normal serum used wasequivalent to the species of the secondary antibody used; see below).Primary antibodies were diluted in PBS with 1% bovine serum albumin,applied on the sections, and incubated overnight at 41C. Subsequently, thesections were washed in PBS and incubated 30 min with the appropriatesecondary biotinylated antibody preparation (horse antimouse IgG, goatantirabbit IgG, or rabbit antigoat IgG, all diluted 1:200 and all fromVectorLaboratories, Burlingame, CA). After washing in PBS, sections wereincubated with avidin�biotin complex conjugated with horseradishperoxidase (Vector Laboratories) for 30 min, washed, and ¢nallydeveloped with 3,30 -diaminobenzidine as substrate (Sigma-Aldrich, StLouis, MO).In double stainings, the sections were ¢xed as above, washed in PBS, and

incubated in 10% normal goat serum and 10% normal horse serum for 30min. Primary antibodies for keratinocyte transglutaminase and LTBP-1were diluted and incubated for 1 h. Sections were washed again in PBSfollowed by incubation of a secondary goat antirabbit biotinylatedantibody preparation (1:200, Vector Laboratories) for 30 min. Afterwashing, the sections were incubated in Cy3-labeled streptavidin (1:1000,Amersham Pharmacia Biotech, Uppsala, Sweden) for 30 min. Sectionswere again washed, followed by sequential incubation with avidin (10 mgper ml) and biotin (1 mg per ml), both from Sigma-Aldrich, for 10 mineach, and an incubation of a secondary horse antimouse biotinylatedantibody preparation (1:200, Vector Laboratories) for 30 min. This wasfollowed by incubation with Cy2-labeled streptavidin (1:1000, AmershamPharmacia) for 30 min. Sections were ¢nally washed and counterstainedwith DAPI (Molecular Probes, Eugene, OR). Isotype-matched controlantibodies showed no staining with the immunohistologic detectiontechniques used.

In situ hybridization After ¢xation in 4% bu¡ered formaldehydesolution for 24 h, biopsies were processed through graded alcohols,oriented, and embedded in para⁄n. Four micron thick para⁄n sectionswere cut and mounted on polylysine-coated glass slides. Prior to staining,sections were dewaxed and rehydrated according to standard procedures.Endogenous peroxidase activity was quenched with 3% H2O2 inmethanol. The sections were then rehydrated with distilled water andequilibrated in TE-bu¡er (10 mM Tris�HCl, 1 mM ethylenediaminetetraacetic acid, pH 7.6). Permeabilization was performed with proteinaseK (Sigma-Aldrich, 20 mg per ml in TE-bu¡er, 30 min, 371C). Proteolyticactivity was stopped by incubation with glycine in PBS (2 mg per ml).Sections were post¢xed in 4% paraformaldehyde in PBS for 15 min.Acetylation was done with freshly prepared triethanolamine bu¡er (100mM, pH 8) with 0.5% acetic anhydride (2� 5 min). Equilibration andprehybridization were carried out with 5� sodium citrate/chloride bu¡er(SSC) (10 min, room temperature) and hybridization bu¡er (Dako; 2 h,52 1C), (DAKO, Glostrup, Denmark) respectively. For hybridization,sections were incubated for 18 h at 52 1C with biotinylated antisense orsense probe (GreenStar Biotin10 oligonucleotide probe; antisense, 48 bphybridizing to nucleotides 440^487 in the coding sequence of humanSmad7; GeneDetect, Auckland, New Zealand) diluted in hybridizationbu¡er (1 mg per ml). Stringency washes were 2� SSC (1 h, 521C),2� SSC diluted with 50% formamide (20 min, 521C), 0.1�SSC (20 min,521C). Blocking was done withTris-bu¡ered saline (TBS; 0.1M Tris�HCl,pH 7.5, 0.15 M NaCl) with 1% casein (TBS-C). Further incubations andwashings were then peroxidase-conjugated rabbit antibiotin antibodies(Dako P5106, 1:100 in TBS-C, 30 min), washing with TBS with 0.05%Tween 20 (TBS-T), biotinyl-tyramide (Dako GenPoint kit K0620; 8 min),washing with PBS-T, alkaline phosphatase-conjugated rabbit antibiotinantibodies (Dako D5107; 1:75 in TBS-C, 30 min). The sections were ¢nallydeveloped with BCIP/NBT substrate (20 min) and counterstained withNuclear Fast Red. All aqueous solutions for in situ hybridizations wereprepared with DEPC-distilled water.

Cell countings Images of sections were captured with a digital camerausing objectives with 20� , 40� , 60� , or 100�magni¢cation andmonitored on a 17-inch computer screen by means of an imaging soft-ware program (Soft Imaging System, Mˇnster, Germany). Cell countingswere made using objectives with 20�magni¢cation. For ERY OL speci-mens, the ¢eld chosen for counting was located over the in£ammatorycell in¢ltrate. Total numbers of keratinocytes and numbers of immuno-histochemically stained nuclei were counted in the entire part of theepithelium visible on the computer screen.

Statistics t tests were used to compare di¡erences between groups.Di¡erences were considered to be statistically signi¢cant at po0.001.

2 KARATSAIDIS ETAL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

RESULTS

An overview of the stainings is presented inTable I.In NOM, basal, spinous, and granular cell layers showed peri-

nuclear, nuclear, and cytoplasmic staining for LAP-TGF-b1 thatgradually decreased in intensity from the spinous to the outerepithelial surface (Fig 1a). Cytoplasmic staining for LTBP-1 wasseen in all epithelial cell layers in NOM, and this was strongest inspinous, granular, and super¢cial cell areas (Fig 1c). Speckledstaining for active TGF-b1 was observed in upper spinous andgranular layers (Fig 1ewith enlargement, showing arrow-markedspots). Strong con£uent staining for active TGF-b1 was seen inthe outermost super¢cial epithelial cell layer in NOM (Fig 1e).TbRI andTbRII expression was present as a nuclear, a cell mem-brane, and a cytoplasmic staining in all NOM epithelial celllayers (Fig 1g, i). In NOM epithelium, the staining was most in-tense in basal cell layers for both TbRI and TbRII. The majorityof the cells throughout the epithelium showed nuclear and cyto-plasmic staining for both phosphorylated Smad2/3 and Smad4(Fig 1k, m, respectively). NOM epithelium showed areas withpatched weak or no cytoplasmic staining for Smad7 (Fig 1o withenlargement). In situ hybridization showed only very weak andscattered nuclear and perinuclear Smad7 mRNA in basal kerati-nocytes of NOM (Fig 1q; antisense). Control hybridization(sense) did not show any staining (Fig 1r).In ERY OL, basal and lower spinous layers showed cytoplas-

mic staining for LAP-TGF-b1, which was especially intensearound cell nuclei (Fig 1b with enlargement). Typically, a band-like area with only scattered or absent LAP-TGF-b1 stainingwas observed in the uppermost spinous and granular cell layers(Fig 1bwith enlargement).There was strong cytoplasmic stainingfor LTBP-1 in basal and lower spinous cell layers (Fig 1d). Inupper spinous and granular epithelial layers, this staining had amarked pericellular and intercellular location (Fig 1d with enlar-gement). Active TGF-b1 consistently showed a clear cell mem-brane/pericellular staining in the granular and upper spinousepithelial layers, but not in the basal cell layers in ERY OLepithelium (Fig 1f with enlargement). This band of active TGF-b1 staining was complementary to the cell layers that displayeddecreased LAP-TGF-b1 staining (compare Fig 1b, f with respec-tive enlargements). The reduced stores of LAP-TGF-b1 maytherefore be a direct result of depletion of latent TGF-b1 in these

layers as a consequence of increased activation. Cell membrane,cytoplasmic, and intense perinuclear staining for TbRI andTbRII was seen (Fig 1h, j, respectively). This staining wasstrongly reduced or absent in granular and in upper parts of thespinous epithelial cell layers, often in a band-like pattern, similarto that described above for LAP-TGF-b1. Phosphorylated Smad2/3(Fig 1l with enlargement) and Smad4 (Fig 1nwith enlargement)stainings revealed signi¢cantly reduced numbers of cells with nu-clear and/or cytoplasmic staining among keratinocytes in allepithelial layers in OL compared with epithelial cells in NOM(Table II, pp0.001, t test). The Smad7 staining pattern in ERYOL was profoundly altered compared with NOM (Fig 1p withenlargement). In contrast to the nearly absent Smad7 expressionin NOM, basal and suprabasal cell layers in ERY OL showedstrong cytoplasmic Smad7 staining. In spinous and granularlayers, the Smad7 staining was particularly localized to the cellmembrane. Almost all basal keratinocytes showed strong nuclearand perinuclear expression of Smad7 mRNA in ERYOL epithe-lium (Fig 1; antisense). Control hybridization (sense) did notreveal any staining (Fig 1t).LTBP-1 has been demonstrated to be targeted to elements of

the extracellular matrix through cross-linking by the enzymetransglutaminase (Nunes et al, 1997; Verderio et al, 1999). Doublestaining for LTBP-1 and keratinocyte transglutaminase in NOMepithelium showed that LTBP-1 was localized preferentially inthe cytoplasm and keratinocyte transglutaminase in cell mem-brane/pericellular epithelial areas (Fig 1u). In contrast, doublestaining for LTBP-1 and keratinocyte transglutaminase in ERYepithelium showed a cell membrane/pericellular codistributionof these molecules in spinous and granular cell layers (Fig 1(v1^v3)).Finally, we carried out staining for PAI-1, which is upregulated

byTGF-b through Smad (Dennler et al, 1998). NOM epitheliumshowed ample cytoplasmic and pericellular staining in spinousand granular cell layers (Fig 1w). In contrast, there was very littlestaining for PAI-1 in ERYOL epithelium, and this was restrictedto only scattered epithelial cells (Fig 1x).In sum,TGF-b1was strongly activated in parts of the ERYOL

epithelium, but the signal transduction through Smad appearedto be blocked by Smad7.This was accompanied by a concomitantdecreased expression of PAI-1 in ERYOL epithelium.

DISCUSSION

The integrity of normal strati¢ed epithelium is maintained by abalance between proliferation of basal epithelial cells and contin-uous shedding of terminally di¡erentiated epithelial cells (Jettenand Harvat, 1997; Gandarillas, 2000). Epithelial and stromal cellsproduce an array of cytokines and growth factors that bind tocorresponding receptors on epithelial cells, which upon activationcan have a wide variety of antagonistic and synergistic e¡ects onkeratinocyte proliferation and di¡erentiation (Dotto, 1999; Freed-berg et al, 2001). In this in situ study, we examined the activationstatus of TGF-b1 and its intracellular signaling pathway mediatedthrough Smad molecules.In NOM, we disclosed evidence for active nuclear Smad sig-

naling in the epithelium: numerous Smad4 and phosphorylatedSmad2/3 positive nuclei were present, with only sparse signs of

Table I. Summary of speci¢c stainings in NOM and ERYOL

NOM OL

LAP-TGF-b Basal/parabasal layers þ þ þ a þ þ þSpinous/granular þ þ þ 0/þ

LTBP-1 Basal/parabasal layers þ þ þ þ þSpinous/granular þ þ þ þ þ þ

Active TGF-b Basal/parabasal layers 0 0Spinous/granular Spots þ þ þ

TbRI Basal/parabasal layers þ þ þ þ þ þSpinous/granular þ þ 0/þ

TbRII Basal/parabasal layers þ þ þ þ þ þSpinous/granular þ þ 0/þ

pSmad2/3 Basal/parabasal layers þ þ þ 0/þSpinous/granular þ þ þ 0/þ

Smad4 Basal/parabasal layers þ þ þ 0/þSpinous/granular þ þ þ 0/þ

Smad7 Basal/parabasal layers 7 þ þ þSpinous/granular 7 Membrane þ þ þ

Smad7 mRNA Basal/parabasal layers 7 þ þ þSpinous/granular 0 0

PAI-1 Basal/parabasal layers þ þ 7Spinous/granular þ þ þ 7

a0, no staining; 7, very weak staining; þ, weak staining; þ þ, moderatestaining; þ þ þ, strong staining.

Table II. Percentages of cells with nuclei stained for Smad2/3and Smad4 in epithelium of NOM and ERYOL

(mean7standard deviation)

NOM (n¼11) OL (n¼12)

pSmad2/3 60718 19711�

Smad4 56717 20712�

�Statistical di¡erence between NOM and OL groups (po0.001, t test).

Smad SIGNALING BLOCKAGE IN ORAL LICHEN 3VOL. 121, NO. 6 DECEMBER 2003

Smad7 expression. This accords with observations in normal skinepidermis, where Smad2, Smad3, and Smad4 nuclear stainingalso was seen in the epidermal basal, spinous, and granular kera-tinocytes (He et al, 2001). The Smad activation indicates that this,by inference, has a physiologic regulatory function of keratino-cyte proliferation and di¡erentiation. Several factors can triggersuch Smad signaling. (1) Active TGF-b1 could have been a likelyactivator, but this was only seen at low levels in the uppermostspinous and granular epithelial cell layers and not among basaland suprabasal keratinocytes where it should have been to exertits growth inhibitory function. This accords with observationsfrom normal epidermis where it is also absent (Kane et al, 1991).(2) Growth factors that in£uence epithelial proliferation and dif-ferentiation by signaling through MAPK (e.g., epidermal growthfactor and hepatocyte growth factor) can mediate activation andnuclear translocation of Smad2 (de Caestecker et al, 1998). Thistype of cross-talk between Smad and MAPK pathways may alsoexist in NOM epithelium. (3) Signaling could be induced byTGF-b2 and/or TGF-b3 (Paterson et al, 2001). This is not likelybecause TGF-b2 and TGF-b3 are only expressed in the upper spi-nous and the granular layers in NOM epithelium (own observa-tions, data not shown) where any action on growth regulation ofbasal cells is unlikely. (4) Signaling could occur through activin,as activin can regulate keratinocyte di¡erentiation (Beer et al,2000) and mediate downstream Smad2/3-induced transcription(Lebrun et al, 1999).In OL, the normal epithelial structure is severely disrupted,

with changes such as hyperkeratosis, atrophy, and degenerationof basal cells (Andreasen, 1968; Kramer et al, 1978). In addition,the epithelium shows an abnormally high proliferation rate inthe basal cell layer in the ERY form of the disease (Karatsaidiset al, 2003). Such changes could be caused by inhibition of thegrowth regulatory activity of TGF-b1. In this study we thereforeexamined TGF-b1 signaling in ERY OL at the level of produc-tion and activation of the ligand, of TbRI and TbRII expression,and of Smad signal transduction, and we conclude the following.(1) Production and activation of TGF-b1 are not de¢cient in ERYOL because there was ample staining for latent TGF-b1(as shown by the presence of LAP-TGF-b1) and a strong stainingfor active TGF-b1 in the upper parts of the spinous layer and inthe granular cell layer. The activation markedly exceeded thelevels seen in NOM, but TGF-b1 activation was con¢ned to themore super¢cial di¡erentiating cell layers and was not seenamong basal cells where it could have had an antiproliferativee¡ect. (2) TGF-b signaling is probably not obstructed by de¢-cient expression of TbRI and TbRII because the receptors werenormally expressed in basal and spinous layers of the ERY OLepithelium. Incidentally, the upper spinous and granular layersshowed a reduced TbRI/TbRII expression. This could be adesensitizing/downregulating response to the increased TGF-b1activity�a mechanism that has been documented in vitro (Primeet al, 1994; Zwaagstra et al, 1999). It has also recently been shownthat Smad7 can be involved in degradation of TGF-b receptors atthe protein level (Kavsak et al, 2000). As Smad7 was located at thelevel of the cell membrane in the uppermost spinous and granularlayers in ERY OL, this also could contribute to the reduced ex-pression of TbRI and TbRII at this location. (3) There appears tobe blockage of the TGF-b1 signal through Smad2 and Smad3 inERY OL, because there were signi¢cantly decreased numbers ofcells with nuclear accumulation of the signaling moleculesSmad2, Smad3, and Smad4 in all epithelial cell layers, concomi-tant with a marked cell membrane localization of the inhibitorysignaling molecule Smad7. Moreover, Smad7 transcription wasshown to be signi¢cantly upregulated in basal keratinocytes ofERYOL compared with NOM. Blockage of Smad2 and Smad3phosphorylation by Smad7 may thus be the cause for the in-creased proliferation rate in ERY OL epithelium. This notion isstrongly supported by studies in transgenic mice that show thatboth Smad3-null mice and mice with overexpression of Smad7display increased keratinocyte proliferation (Ashcroft et al, 1999;

He et al, 2002). Moreover, transgenic mice with constitutive activeTGF-b1 coupled to the keratin 10 promoter (which targets TGF-b1 to suprabasal epithelial cell layers) also show epidermal hyper-proliferation, suprabasal TGF-b1 activation, and epithelial atrophy(Cui et al, 1995). The ¢ndings in the latter model strongly resem-ble our observations in ERY OL epithelium. It is likely that asignaling blockage by Smad7 was in part responsible for the(then unexpected) increased epithelial proliferation seen in thekeratin-10-TGF-b1 transgenic mice.The site (i.e., the epithelial layers) where TGF-b1 exerts its ac-

tivity is known to be crucial for the morphology of the epithe-lium. TGF-b1 activation within the basal cell layer in£icts majordamage to the epithelium: e.g., transgenic mice with constitutiveexpression of active TGF-b1 targeted to basal keratinocytes donot survive because of a total inhibition of basal cell proliferation(Sellheyer et al, 1993). Other transgenic strains, where latent TGF-b1 expression was coupled to keratin 14 promoter (which is con-ditionally activated in basal keratinocytes), showed marked de-fects in re-epithelialization during wound healing (Yang et al,2001). This may be the reason whyTGF-b1 during normal woundhealing is activated in a con¢ned compartment of the moresuper¢cial epithelial cell layers, at a distance from the proliferativebasal cells (Kane et al, 1991). The observed restricted activation ofTGF-b1 in ERY OL epithelium resembles this pattern. Ourstainings for LTBP-1 and keratinocyte transglutaminase expres-sion might give a clue to why this activation shows this con¢ne-ment. In connective tissue, LTBP-1 has a function in targetinglatent TGF-b to the extracellular matrix and transglutaminasecontributes to this process because it cross-links LTBP-1 to extra-cellular matrix elements (Nunes et al, 1997). Transglutaminase isnot expressed in basal keratinocytes in oral mucosa (Ta et al,1990;present observations) and if transglutaminase and LTBP-1have a similar TGF-b targeting function in epithelium as in con-nective tissue, this might explain the con¢nement of the TGF-b1activation. Therefore, as long as the TGF-b1 activity remainscompartmentalized as in the present ERYOL biopsies, the lesioncould be regarded as a wound that is trying to heal normally butis kept in an unresolved state by the chronic in£ammatory in¢l-trate (see below). If TGF-b1, however, would become verystrongly activated and a¡ect the basal cell layers, epithelial prolif-eration could become so severely inhibited that the lesion thenperhaps might turn into its ulcerative form.The epithelial di¡erentiation pattern and cell death mode are

disturbed in ERYOL (Karatsaidis et al, 2003), which is the under-lying cause for the morpho-pathologic changes typical for thedisease. Besides inhibiting epithelial growth, TGF-b1 can a¡ectepithelial di¡erentiation by regulating, for example, keratin andintegrin expression and apoptosis (Mansbridge and Hanawalt,1988; Jiang et al, 1995; Zambruno et al, 1995; Min et al, 1999).Altered epidermal di¡erentiation resembling that in OL can beseen in transgenic mice in which the TGF-b signaling pathwaywas manipulated; mice with overexpressed epidermal Smad7show acanthosis and hyperkeratosis (He et al, 2002) and keratin-10-TGF-b1 transgenic mice in which the epidermis was treatedwith 12-O-tetradecanoylphorbol-13-acetate (a hyperplasia indu-cing agent) acquire pronounced epidermal atrophy (Cui et al,1995). It is therefore possible that the detected inhibition of theTGF-b signaling pathway plays a role in the epithelial atrophyin ERYOL (Karatsaidis et al, 2003).Based on prospective follow-up studies of large groups of OLP

patients, it has been suggested that OLP can be a premalignantcondition because a small fraction of the initial lesions diagnosedas plaque or ERY OLP over time developed into cancer(Holmstrup et al, 1988). Loss of Smad and TGF-b receptor signal-ing has been observed in cancer cells, and this is associated withdecreased growth inhibitory control of TGF-b (Pasche, 2001).Even though the TGF-b/Smad signaling blockade was not com-plete in ERY OL, alterations in this growth inhibitory pathwaymight perhaps, in concert with other predisposing factors, play arole in dysplastic transformation of ERYOL.


Figure1. Micrographs of immunohistologically stained tissue sections from biopsies obtained from persons with NOM (a, c, e, g, i, k, m, o,u, w) and patients with OL (b, d, f, h, j, l, n, p, v, x). After immunohistochemical staining the labeled molecules are seen in brown (counterstain withhematoxylin) and for immuno£uorescence (u, v) the speci¢c molecules are displayed in red and green (overlapping red and green appear yellow; blue nuclearstain with DAPI). Antibodies used were against (a), (b) LAP-TGF-b1; (c), (d ) LTBP-1; (e), (f ) active TGF-b1 (arrows in (e) indicate spotted staining); (g), (h)TbRI; (i), (j ) TbRII; (k), (l ) phoshorylated Smad2/3; (m), (n) Smad4; (o), (p) Smad7; (q), (r) LTPB-1 in red, keratinocyte transglutaminase in green (position ofthe basement membrane indicated with white broken lines); (w), (x) mature PAI-1; in situ hybridization for Smad7 (q, r, s, t; labeled molecules are in blue;counterstain with red). Low magni¢cation pictures (lettered a to x, except for q, r, s, t) were obtained with a 20� objective (scale bar: 100 mm). High magni-¢cation pictures (without lettering) were taken with a 60� objective, except for enlargements (e) and (k), which were taken with 100� and 40� objectives,respectively. Micrographs of in situ hybridization (q, r, s, t) were obtained with a 40� objective. Frames of high magni¢cation are connected to their respec-tive low magni¢cation areas with arrows. Scale bar for £uorescence ¢gures (u, v) is 10 mm.


Figure1. Continued


Upregulation of the expression of Smad7 can be caused bypro-in£ammatory cytokines like interferon-g and tumor necrosisfactor a (Ulloa et al, 1999; Bitzer et al, 2000). These are believed tobe central regulatory cytokines in OL, and are synthesized andsecreted by activated keratinocytes and in£ammatory cells in thesubepithelial in¢ltrate (Yamamoto and Osaki, 1995). Typically,Smad7 mRNA was observed in the basal keratinocyte layer inERY OL. The increased basal cell proliferation in ERY OLepithelium (Karatsaidis et al, 2003) can thus be due to inhibitionof TGF-b/Smad signaling by Smad7, induced by interferon-gand tumor necrosis factor a.TGF-b is a strong upregulator of the transcription of PAI-1 in

human keratinocytes (Keski-Oja and Koli, 1992). Elements in thePAI-1 promoter that are binding sites for the Smad3�Smad4 het-erodimer have indeed been identi¢ed (Dennler et al, 1998). PAI-1expression in ERY OL epithelium was strongly reduced, whichindicates a blockage at the functional level of the TGF-b/Smadsignaling pathway.In sum, these results indicate that there is a functional blockage

of the TGF-b signaling pathway by increased Smad7 expressionand translocation in ERY OL. This may explain the increasedhyperproliferation and the disturbed terminal di¡erentiation ofkeratinocytes observed in this disease.

We thank Tannlegesenteret-Bergr�dveien 13 in Oslo for help with collection of thebiopsies and Dr. Guttorm Haraldson and Dr. MarjanVeuger for technical advice within situ hybridization.

REFERENCES

Andreasen JO: Oral lichen planus. 1. A clinical evaluation of 115 cases. Oral Surg OralMed Oral Pathol 25:31^42, 1968

Ashcroft GS,Yang X, Glick AB, et al: Mice lacking Smad3 show accelerated woundhealing and an impaired local in£ammatory response. Nat Cell Biol 1:260^266,1999

Axell T, Rundquist L: Oral lichen planus ^ A demographic study. Community DentOral Epidemiol 15:52^56, 1987

Beer HD, Gassmann MG, Munz B, Steiling H, Engelhardt F, Bleuel K,Werner S:Expression and function of keratinocyte growth factor and activin in skinmorphogenesis and cutaneous wound repair. J Investig Dermatol Symp Proc5:34^39, 2000

Bitzer M, von Gersdor¡ G, Liang D, Dominguez-Rosales A, Beg AA, Rojkind M,Bottinger EP: A mechanism of suppression of TGF-b/SMAD signaling byNF-kB/Rela. Genes Dev 14:187^197, 2000

Bolewska J, Reibel J: T lymphocytes, Langerhans cells and HLA-DR expression onkeratinocytes in oral lesions associated with amalgam restorations. J Oral PatholMed 18:525^528, 1989

Bolewska J, Holmstrup P, Moller-Madsen B, Kenrad B, Danscher G: Amalgam asso-ciated mercury accumulations in normal oral mucosa, oral mucosal lesions oflichen planus and contact lesions associated with amalgam. J Oral Pathol Med19:39^42, 1990

de Caestecker MP, Parks WT, Frank CJ, Castagnino P, Bottaro DP, Roberts AB,Lechleider RJ: Smad2 transduces common signals from receptor serine-threo-nine and tyrosine kinases. Genes Dev 12:1587^1592, 1998

Cui W, Fowlis DJ, Cousins FM, Du⁄e E, Bryson S, Balmain A, Akhurst RJ:Concerted action of TGF-b1 and its type II receptor in control of epidermalhomeostasis in transgenic mice. Genes Dev 9:945^955, 1995

Datto MB, Li Y, Panus JF, Howe DJ, Xiong Y, Wang XF: Transforming growthfactor-b induces the cyclin-dependent kinase inhibitor p21 through a p53-independent mechanism. Proc Natl Acad Sci USA 92:5545^5549, 1995

Dennler S, Itoh S, Vivien D, ten Dijke P, Huet S, Gauthier JM: Direct binding ofSmad3 and Smad4 to critical TGF b-inducible elements in the promoter ofhuman plasminogen activator inhibitor-type 1 gene. EMBO J 17:3091^3100,1998

Dotto GP: Signal transduction pathways controlling the switch between keratino-cyte growth and di¡erentiation. Crit Rev Oral Biol Med 10:442^457, 1999

Ebner R, Chen RH, Shum L, et al: Cloning of a type I TGF-b receptor and its e¡ecton TGF-b binding to the type II receptor. Science 260:1344^1348, 1993

Freedberg IM, Tomic-Canic M, Komine M, Blumenberg M: Keratins and thekeratinocyte activation cycle. J Invest Dermatol 116:633^640, 2001

Gandarillas A: Epidermal di¡erentiation, apoptosis, and senescenc: Commonpathways? Exp Gerontol 35:53^62, 2000

Glick AB, Kulkarni AB, Tennenbaum T, et al: Loss of expression of transforminggrowth factor-b in skin and skin tumors is associated with hyperproliferationand a high risk for malignant conversion. Proc Natl Acad Sci USA 90:6076^6080, 1993

Hannon GJ, Beach D: P15ink4b is a potential e¡ector of TGF-b-induced cell cyclearrest. Nature 371:257^261, 1994

HeW, CaoT, Smith DA, Myers TE,Wang XJ: Smads mediate signaling of the TGFbsuperfamily in normal keratinocytes but are lost during skin chemical carcino-genesis. Oncogene 20:471^483, 2001

HeW, Li AG,Wang D, et al: Overexpression of Smad7 results in severe pathologicalalterations in multiple epithelial tissues. EMBO J 21:2580^2590, 2002

Hedberg N, Ng A, Hunter N: A semi-quantitative assessment of the histopathologyof oral lichen planus. J Oral Pathol 15:268^272, 1986

Holmstrup P, Thorn JJ, Rindum J, Pindborg JJ: Malignant development of lichenplanus-a¡ected oral mucosa. J Oral Pathol 17:219^225, 1988

Jetten AM, Harvat BL: Epidermal di¡erentiation and squamous metaplasia: Fromstem cell to cell death. J Dermatol 24:711^725, 1997

Jiang CK, Tomic-Canic M, Lucas DJ, Simon M, Blumenberg M: TGF b promotesthe basal phenotype of epidermal keratinocytes: Transcriptional induction ofk#5 and k#14 keratin genes. Growth Factors 12:87^97, 1995

Kane CJ, Hebda PA, Mansbridge JN, Hanawalt PC: Direct evidence for spatial andtemporal regulation of transforming growth factor-b 1 expression during cu-taneous wound healing. J Cell Physiol 148:157^173, 1991

Karatsaidis A, Schreurs O, Helgeland K, Axell T, Schenck K: Erythematous andreticular forms of oral lichen planus and oral lichenoid reactions di¡er inpathological features related to disease activity. J Oral Pathol Med 32:275^281,2003

Kavsak P, Rasmussen RK, Causing CG, Bonni S, Zhu H,Thomsen GH,Wrana JL:Smad7 binds to Smurf2 to form an E3 ubiquitin ligase that targets the TGF breceptor for degradation. Mol Cell 6:1365^1375, 2000

Keski-Oja J, Koli K: Enhanced production of plasminogen activator activity in hu-man and murine keratinocytes by transforming growth factor-b1. J Invest Der-matol 99:193^200, 1992

Kramer IR, Lucas RB, Pindborg JJ, Sobin LH: De¢nition of leukoplakia and relatedlesions: An aid to studies on oral precancer. Oral Surg Oral Med Oral Pathol46:518^539, 1978

Lebrun JJ, Takabe K, Chen Y, Vale W: Roles of pathway-speci¢c and inhibitorySmads in activin receptor signaling. Mol Endocrinol 13:15^23, 1999

Mansbridge JN, Hanawalt PC: Role of transforming growth factor-b in the matura-tion of human epidermal keratinocytes. J Invest Dermatol 90:336^341, 1988

Massague J: How cells read TGF-b signals. Nat Rev Mol Cell Biol 1:169^178, 2000Min BM,Woo KM, Lee G, Park NH:Terminal di¡erentiation of normal human oral

keratinocytes is associated with enhanced cellular TGF-b and phospholipasec-g1 levels and apoptotic cell death. Exp Cell Res 249:377^385, 1999

Miyazono K, Olofsson A, Colosetti P, Heldin CH: A role of the latent TGF-b1-binding protein in the assembly and secretion of TGF-b1. EMBO J 10:1091^1101, 1991

Nakao A, Afrakhte M, Moren A, et al: Identi¢cation of Smad7, a TGFb-inducibleantagonist of TGF-b signalling. Nature 389:631^635, 1997a

Nakao A, Imamura T, Souchelnytskyi S, et al: TGF-b receptor-mediated signallingthrough Smad2, Smad3 and Smad4. EMBO J 16:5353^5362, 1997b

Nunes I, Gleizes PE, Metz CN, Rifkin DB: Latent transforming growth factor-bbinding protein domains involved in activation and transglutaminase-depen-dent cross-linking of latent transforming growth factor-b. J Cell Biol 136:1151^1163, 1997

Pasche B: Role of transforming growth factor-b in cancer. J Cell Physiol 186:153^168,2001

Paterson IC, Matthews JB, Huntley S, Robinson CM, Fahey M, Parkinson EK,Prime SS: Decreased expression of TGF-b cell surface receptors duringprogression of human oral squamous cell carcinoma. J Pathol 193:458^467,2001

Pietenpol JA, Holt JT, Stein RW, Moses HL:Transforming growth factor b1 suppres-sion of c-myc gene transcription: Role in inhibition of keratinocyte prolifera-tion. Proc Natl Acad Sci USA 87:3758^3762, 1990

Prime SS, Matthews JB, Patel V, et al: TGF-b receptor regulation mediates the re-sponse to exogenous ligand but is independent of the degree of cellular di¡er-entiation in human oral keratinocytes. Int J Cancer 56:406^412, 1994

Roberts AB: Molecular and cell biology of TGF-b. Miner Electrolyte Metab 24:111^119,1998

Salonen L, Axell T, Hellden L: Occurrence of oral mucosal lesions, the in£uence oftobacco habits and an estimate of treatment time in an adult Swedish popula-tion. J Oral Pathol Med 19:170^176, 1990

Sellheyer K, Bickenbach JR, Rothnagel JA, et al: Inhibition of skin development byoverexpression of transforming growth factor b1 in the epidermis of transgenicmice. Proc Natl Acad Sci USA 90:5237^5241, 1993

Ta BM, Gallagher GT, Chakravarty R, Rice RH: Keratinocyte transglutaminase inhuman skin and oral mucosa: Cytoplasmic localization and uncoupling of dif-ferentiation markers. J Cell Sci 95:631^638, 1990

Ten Dijke P, Goumans MJ, Itoh F, Itoh S: Regulation of cell proliferation by Smadproteins. J Cell Physiol 191:1^16, 2002

Ulloa L, Doody J, Massague J: Inhibition of transforming growth factor-b/SMADsignalling by the interferon-g/STAT pathway. Nature 397:710^713, 1999


Verderio E, Gaudry C, Gross S, Smith C, Downes S, Gri⁄n M: Regulation of cellsurface tissue transglutaminase: E¡ects on matrix storage of latent transform-ing growth factor-b binding protein-1. J Histochem Cytochem 47:1417^1432, 1999

Wake¢eld LM, Smith DM, Flanders KC, Sporn MB: Latent transforming growthfactor-b from human platelets. J Biol Chem 263:7646^7654, 1988

Weinberg RA: The retinoblastoma protein and cell cycle control. Cell 81:323^330,1995

Wrana JL, Attisano L,Wieser R,Ventura F, Massague J: Mechanism of activation ofthe TGF-b receptor. Nature 370:341^347, 1994

Yamamoto T, Osaki T: Characteristic cytokines generated by keratinocytes andmononuclear in¢ltrates in oral lichen planus. J Invest Dermatol 104:784^788, 1995

Yang L, ChanT, Demare J, IwashinaT, GhaharyA, Scott PG,Tredget EE: Healing ofburn wounds in transgenic mice overexpressing transforming growth factor-b1 in the epidermis. AmJ Pathol 159:2147^2157, 2001

Zambruno G, Marchisio PC, Marconi A,Vaschieri C, Melchiori A, De Giannetti A,Luca M: Transforming growth factor-b1 modulates b1 and b5 integrin recep-tors and induces the de novo expression of the avb6 heterodimer in normal hu-man keratinocytes: Implications for wound healing. J Cell Biol 129:853^865,1995

Zwaagstra JC, Kassam Z, O’Connor-McCourt MD: Down-regulation of transform-ing growth factor-b receptors: Cooperativity between the types I, II, and IIIreceptors and modulation at the cell surface. Exp Cell Res 252:352^362, 1999


Inhibition of the Transforming Growth Factor-b/Smad Signaling ...

Documents

Transcript of Inhibition of the Transforming Growth Factor-b/Smad Signaling ...