expression of a retinoid-inducible tumor suppressor...

12
Expression of a Retinoid-inducible Tumor Suppressor, Tazarotene- inducible Gene-3, Is Decreased in Psoriasis and Skin Cancer 1 Madeleine Duvic, 2 Bharati Helekar, Claudia Schulz, Mimi Cho, Dan DiSepio, Carina Hager, Dominique DiMao, Parul Hazarika, Brooke Jackson, Joan Breuer-McHam, John Young, Gary Clayman, Scott M. Lippman, Roshandra A. S. Chandraratna, Nancy A. Robinson, Anne Deucher, Richard L. Eckert, and Sunil Nagpal Section of Dermatology, Department of Internal Medical Specialties [M. D., B. H., C. S., M. C., C. H., D. DiM., P. H., B. J., J. B-M., J. Y.], Department of Head and Neck Surgery [G. C.], Division of Cancer Prevention [S. M. L.], M. D. Anderson Cancer Center, Houston, Texas 77030; Allergan Research, Irvine, California 92612 [D. DiS., R. A. S. C., S. N.]; and Department of Physiology/ Biophysics, Case Western Reserve, Cleveland, Ohio 44106 [N. A. R., A. D., R. L. E.] ABSTRACT Tazarotene-induced gene-3 (TIG-3), isolated from hu- man keratinocytes treated with the retinoic acid receptor- selective retinoid Tazarotene, is homologous to H-rev, a class II tumor suppressor. TIG-3 gene localized to chromosome 11q23, a site of loss of heterozygosity in several malignan- cies. Retinoids influence epidermal differentiation and are used to treat and prevent skin cancer. Therefore, we studied TIG-3 mRNA expression in psoriasis and in basal and SCCs by in situ hybridization and a quantitative QT-RT-PCR assay. Psoriasis lesions had significantly lower staining (me- dian, 3) than paired normal control skin (median, 4; P 5 0.012). TIG-3 mRNA was significantly higher in normal control skin (P 5 0.001), in paired adjacent skin (median, 3; P 5 0.007), and in overlying epidermis (median, 3.0; P 5 0.0001) than in 21 SCC specimens as a group (median, 1.5). Aggressive SCCs (median, 1.0) were lower in TIG-3 mRNA staining than nonaggressive SCCs (median, 1.5; P 5 0.07). Three aggressive tumors had no TIG-3 mRNA staining. TIG-3 protein as shown by immunohistochemistry was high- est in the suprabasal epidermis of normal skin, just under the stratum corneum, and was decreased in basal and squa- mous cell carcinomas, similar to the mRNA staining. Reduc- tion in TIG-3 mRNA expression in psoriasis and basal car- cinomas and loss in some aggressive SCCs support the hypothesis that TIG-3 may function as a tumor suppressor in both normal and malignant epidermal differentiation. INTRODUCTION The protective skin barrier arises from a tightly controlled genetic differentiation program resulting in keratinocyte apo- ptosis and stratum corneum formation (1, 2). Despite the fact that most keratinocytes undergo apoptosis, malignancies of ke- ratinocytes (basal, SCC, 3 or basosquamous cell carcinomas) commonly arise and are the most common malignancies in humans (3). The molecular basis of some squamous carcinomas is related to mutations in p53 and in H-ras, induced by UV light (4, 5). Mutations in genes of the sonic-hedgehog or patched pathway have been implicated recently in the formation of BCCs, although phenotypic variation is complex (6 –9). Many other genes may contribute to the pathogenesis of skin cancers because of the complexity of the epidermal differentiation path- way and the need for chronic repair of DNA damage (10 –12). Benign epidermal hyperplasia may also result in the pres- ence of a cutaneous inflammatory reaction, such as psoriasis, in genetically susceptible individuals (13). The genetic programs controlling epidermal differentiation, although only partially understood, are known to be regulated in part by steroid hor- mones, including vitamin A (retinoids) and vitamin D (deltoids; Ref. 14). Retinoids exert their effects in part through binding to differentially expressed RAR or retinoid X receptors (15–17). Retinoids and their receptors also interact with specific DNA transcription factors, including AP1, AP2, CBP300, and nuclear factor-interleukin 6, thought to regulate epidermal differentia- tion (18 –21). The cytokine-inducible nuclear transcription fac- tor, nuclear factor-kB, has also been implicated in control of epidermal proliferation (22). Topical retinoid treatment is known to alter the expression of epidermal keratin and other differentiation markers (23). Retinoids are active clinically in psoriasis (24 –26) and for both chemoprevention and therapy of squamous carcinomas and oral leukoplakia (27).). Tazarotene, a new synthetic RAR b,g- Received 9/3/99; revised 4/25/00; accepted 5/4/00. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported in part by a clinical research grant from Allergan, by the Dermatology Foundation as a Fellowship sponsored by Ortho Dermatological (to C. S.), and an American Society for Derma- tologic Surgery Dermatology Foundation Grant (to B. J.); by the NIH Grants PO1-CA16672-22 (to M. D., G. C., S. L.), R21-CA74117 (to B. H. and M. D.), the M. D. Anderson Cancer Center Core Grant CA16672, and The M. D. Anderson Skin Cancer Research Fund. This work was presented in part at the SID Meeting, Chicago, Illinois, May 5– 8, 1999. 2 To whom requests for reprints should be addressed, at Section of Dermatology, Box 28, M. D. Anderson Cancer Center, Houston, Texas 77030. Phone: (713) 745-1113; Fax: (713) 745-3597; E-mail: mduvic @mdanderson.org. 3 The abbreviations used are: SCC, squamous cell carcinoma; RAR, retinoic acid receptor; BCC, basal cell carcinoma; ISH, in situ hybrid- ization; TIG, Tazarotene-induced gene; QT-RT-PCR, quantitative real- time PCR; BCE, basal cell epithelioma; IHC, immunohistochemistry. 3249 Vol. 6, 3249 –3259, August 2000 Clinical Cancer Research Research. on June 16, 2018. © 2000 American Association for Cancer clincancerres.aacrjournals.org Downloaded from

Upload: tranduong

Post on 05-May-2018

231 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Expression of a Retinoid-inducible Tumor Suppressor ...clincancerres.aacrjournals.org/content/clincanres/6/8/3249.full.pdfExpression of a Retinoid-inducible Tumor Suppressor,Tazarotene-inducible

Expression of a Retinoid-inducible Tumor Suppressor,Tazarotene-inducible Gene-3,Is Decreased in Psoriasis and Skin Cancer1

Madeleine Duvic,2 Bharati Helekar,Claudia Schulz, Mimi Cho, Dan DiSepio,Carina Hager, Dominique DiMao,Parul Hazarika, Brooke Jackson,Joan Breuer-McHam, John Young,Gary Clayman, Scott M. Lippman,Roshandra A. S. Chandraratna,Nancy A. Robinson, Anne Deucher,Richard L. Eckert, and Sunil NagpalSection of Dermatology, Department of Internal Medical Specialties[M. D., B. H., C. S., M. C., C. H., D. DiM., P. H., B. J., J. B-M.,J. Y.], Department of Head and Neck Surgery [G. C.], Division ofCancer Prevention [S. M. L.], M. D. Anderson Cancer Center,Houston, Texas 77030; Allergan Research, Irvine, California 92612[D. DiS., R. A. S. C., S. N.]; and Department of Physiology/Biophysics, Case Western Reserve, Cleveland, Ohio 44106[N. A. R., A. D., R. L. E.]

ABSTRACTTazarotene-induced gene-3(TIG-3), isolated from hu-

man keratinocytes treated with the retinoic acid receptor-selective retinoid Tazarotene, is homologous to H-rev, a classII tumor suppressor. TIG-3 gene localized to chromosome11q23, a site of loss of heterozygosity in several malignan-cies. Retinoids influence epidermal differentiation and areused to treat and prevent skin cancer. Therefore, we studiedTIG-3 mRNA expression in psoriasis and in basal and SCCsby in situ hybridization and a quantitative QT-RT-PCRassay. Psoriasis lesions had significantly lower staining (me-dian, 3) than paired normal control skin (median, 4; P 50.012). TIG-3 mRNA was significantly higher in normalcontrol skin (P 5 0.001), in paired adjacent skin (median, 3;P 5 0.007), and in overlying epidermis (median, 3.0;P 50.0001) than in 21 SCC specimens as a group (median, 1.5).

Aggressive SCCs (median, 1.0) were lower inTIG-3 mRNAstaining than nonaggressive SCCs (median, 1.5;P 5 0.07).Three aggressive tumors had noTIG-3 mRNA staining.TIG-3 protein as shown by immunohistochemistry was high-est in the suprabasal epidermis of normal skin, just underthe stratum corneum, and was decreased in basal and squa-mous cell carcinomas, similar to the mRNA staining. Reduc-tion in TIG-3 mRNA expression in psoriasis and basal car-cinomas and loss in some aggressive SCCs support thehypothesis that TIG-3 may function as a tumor suppressorin both normal and malignant epidermal differentiation.

INTRODUCTIONThe protective skin barrier arises from a tightly controlled

genetic differentiation program resulting in keratinocyte apo-ptosis and stratum corneum formation (1, 2). Despite the factthat most keratinocytes undergo apoptosis, malignancies of ke-ratinocytes (basal, SCC,3 or basosquamous cell carcinomas)commonly arise and are the most common malignancies inhumans (3). The molecular basis of some squamous carcinomasis related to mutations inp53and in H-ras,induced by UV light(4, 5). Mutations in genes of thesonic-hedgehogor patchedpathway have been implicated recently in the formation ofBCCs, although phenotypic variation is complex (6–9). Manyother genes may contribute to the pathogenesis of skin cancersbecause of the complexity of the epidermal differentiation path-way and the need for chronic repair of DNA damage (10–12).

Benign epidermal hyperplasia may also result in the pres-ence of a cutaneous inflammatory reaction, such as psoriasis, ingenetically susceptible individuals (13). The genetic programscontrolling epidermal differentiation, although only partiallyunderstood, are known to be regulated in part by steroid hor-mones, including vitamin A (retinoids) and vitamin D (deltoids;Ref. 14). Retinoids exert their effects in part through binding todifferentially expressed RAR or retinoid X receptors (15–17).Retinoids and their receptors also interact with specific DNAtranscription factors, including AP1, AP2, CBP300, and nuclearfactor-interleukin 6, thought to regulate epidermal differentia-tion (18–21). The cytokine-inducible nuclear transcription fac-tor, nuclear factor-kB, has also been implicated in control ofepidermal proliferation (22). Topical retinoid treatment isknown to alter the expression of epidermal keratin and otherdifferentiation markers (23).

Retinoids are active clinically in psoriasis (24–26) and forboth chemoprevention and therapy of squamous carcinomas andoral leukoplakia (27).). Tazarotene, a new synthetic RARb,g-

Received 9/3/99; revised 4/25/00; accepted 5/4/00.The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisementin accordance with 18 U.S.C. Section 1734 solely toindicate this fact.1 This work was supported in part by a clinical research grant fromAllergan, by the Dermatology Foundation as a Fellowship sponsored byOrtho Dermatological (to C. S.), and an American Society for Derma-tologic Surgery Dermatology Foundation Grant (to B. J.); by the NIHGrants PO1-CA16672-22 (to M. D., G. C., S. L.), R21-CA74117 (toB. H. and M. D.), the M. D. Anderson Cancer Center Core GrantCA16672, and The M. D. Anderson Skin Cancer Research Fund. Thiswork was presented in part at the SID Meeting, Chicago, Illinois, May5–8, 1999.2 To whom requests for reprints should be addressed, at Section ofDermatology, Box 28, M. D. Anderson Cancer Center, Houston, Texas77030. Phone: (713) 745-1113; Fax: (713) 745-3597; E-mail: [email protected].

3 The abbreviations used are: SCC, squamous cell carcinoma; RAR,retinoic acid receptor; BCC, basal cell carcinoma; ISH,in situ hybrid-ization; TIG, Tazarotene-induced gene; QT-RT-PCR, quantitative real-time PCR; BCE, basal cell epithelioma; IHC, immunohistochemistry.

3249Vol. 6, 3249–3259, August 2000 Clinical Cancer Research

Research. on June 16, 2018. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Page 2: Expression of a Retinoid-inducible Tumor Suppressor ...clincancerres.aacrjournals.org/content/clincanres/6/8/3249.full.pdfExpression of a Retinoid-inducible Tumor Suppressor,Tazarotene-inducible

selective retinoid, is effective topically for treatment of psoriasisand acne (28–30). Tazarotene treatment is associated with de-creased expression of epidermal genes overexpressed in psori-asis, including type I keratinocyte transglutaminase, epidermalgrowth factor receptor, MRP-8, and scalp/elafin (29, 31, 32). Inaddition, Tazarotene reduces the expression of inflammatorymarkers including intercellular adhesion molecule-1 andHLA-DR on cells in psoriasis lesions and increases the reex-pression of profilaggrin, a marker of terminal differentiation(31, 33). Recently, a clinical study has suggested that topicalTazarotene may induce remissions of basal cell carcinomas (34).

To understand the mechanism of action of Tazarotene andits effect on epidermal gene expression, novelTIG cDNAs wereisolated by differential display PCR from mRNA collected fromtreated or mock-treated human keratinocytes (35–37).TIG-3(RANTES-3; Ref. 38), a cDNA of 736 bp, was induced 4-foldin treated human keratinocyte cultures. The cDNA isolatedpredicted anMr 18,000 protein of 164 amino acids with 52%homology to H-rev 107, a known class II tumor suppressor (39).H-rev 107 is down-regulated by ras transfection into rat fibro-blasts, up-regulated in fibroblasts that spontaneously revert afterras transfection, and is induced by IFN-g in astrocytes (40, 41).TIG-3 has been expressed with an inducible promoter causinggrowth retardation in EcR-293 kidney cell lines (38). Similarstudies done in a SCC line show growth inhibition byTIG-3sense constructsin vitro.4 These studies suggest that the func-tion of TIG-3 is as a tumor suppressor (38). Of interest,TIG-3maps to chromosome 11q23 that has been reported previously toshow loss of heterozygosity in a number of different tumors andis suspected to contain a tumor suppressor (42, 43).

Cell culture studies have shown that TIG-3 is up-regulatedby all RAR-selective retinoids including Tazarotene and was in-ducible in keratinocytes and in the spontaneously transformedkeratinocytes, the HaCat cell line (38). Psoriasis is a disease char-acterized by rapid epidermal proliferation in response to T-cellinflammation, which is inhibited by topical Tazarotene (28). Thus,we hypothesized that: (a) TIG-3expression may also be associatedwith growth inhibition in normal as well as inflamed epidermis;and (b) loss ofTIG-3 expression could result in epidermal malig-nancies. We examined the level and pattern of expression of TIG-3mRNA and protein as a putative tumor suppressor in normal skin,compared with psoriasis and BCCs and SCCs.

MATERIALS AND METHODSPatient Specimens. Skin biopsy specimens were taken

from unaffected, normal skin and from 29 psoriasis patients’lesions at baseline prior to their use of an investigational 0.1%Tazarotene gel and at 3 and 14 days of treatment. Aggressive(n 5 11) and nonaggressive (n5 10) SCCs and BCCs (n5 6)of the skin and paired normal skin specimens from overlyingand/or adjacent normal skin, as available, were obtained frompatients undergoing surgery in Dermatology Section and theDepartment of Head and Neck Surgery, M. D. Anderson CancerCenter (Table 1). Additional specimens from BCCs, SCCs, and

normal adjacent skin for mRNA isolation were taken fromMoh’s surgery cases and snap frozen at270°C (Table 2). Allpatients signed informed consent under approved protocols.

Specimens were fixed in 4% paraformaldehyde at 4°C for16–24 h, processed in alcohol, and embedded in paraffin asdescribed previously (1, 11). Sections of 4mm were mounted onsialinated, nuclease-free glass slides (CEL Associates, Houston,TX) for in situ hybridization.

Preparation of Riboprobes. Plasmid (pAGN-TIG 3)containing a 600-bp 39cDNA from TIG-3 was inserted inreverse orientation inEcoRI in a PCRII vector. Onemg ofcDNA template was linearized withNotI or HindIII and tran-scribed using either Sp6 or T7 polymerase to yield antisense andsense riboprobes, respectively. The riboprobes were transcribedin the presence of UTP-digoxigenin using a Genius 4 kit, ac-cording to the manufacturer’s instructions (Boehringer-Mann-heim, Indianapolis, IN) as described previously (44). The em-pirical concentration of the probes was estimated using a dotblot method involving serial dilutions with a Dig DNA labelingand detection kit (Boehringer Mannheim).

ISH and Detection of Riboprobes. ISH was performedaccording to a modified procedure of Xuet al. (44). Deparaf-finized and rehydrated tissue sections were washed in PBS (pH7.2), followed by 0.2N HCl at room temperature for 10 min.They were treated with proteinase K in 10 mM Tris and 2 mM

CaCl2 (pH 8.0) for 20 min at 37°C. After three washes in PBS,sections were postfixed with 4% paraformaldehyde for 5 min atroom temperature, washed three times in PBS, and acetylated atpH 8.0. After a final PBS wash, sections were dehydrated inethanol and air-dried.

Prehybridization was carried out in a humidification cham-ber in 50% deionized formamide, 23 SSC solution, 23Den-hardt’s solution (0.02% Ficoll 400, 0.02% polyvinylpyrrolidone,and 0.2% BSA), 10% dextran sulfate, 400mg/ml yeast tRNA,250 mg/ml salmon sperm DNA, and 20 mM DTT in diethylpyrocarbonate-treated water. Sense or antisense riboprobesadded at 400 ng/ml were hybridized for 4 h at 42°C. Posthy-bridization washes were in 23SSC and 0.1% SDS at roomtemperature for 1 h and 40 min, followed by another wash with0.13SSC and 0.1% SDS for 20 min at room temperature. Probebinding was detected using a Fab fragment of anti-digoxigeninantibody and chromogen substrate solution (45ml of nitrobluetetrazolium and 35ml of 5-bromo-4-chloro-3-indolyl-phosphatesolution in 10 ml of buffer 3 containing 10 mM Tris, 0.1M NaCl,and 50 mM MgCl2 pH 9.5). The color reaction was for 1 h.

Analysis of in Situ Hybridization. The sections weregraded for intensity by three blinded observers using a semi-quantitative scale (11, 31). Staining intensity was graded as 0(no staining), 1 (light blue, faint staining), 2 (blue staining), 3(moderate staining, purple color, 4 (strong, very deep purple),and 5 (dark black staining obscuring the architecture). The datadid not fit a Poisson distribution and therefore were analyzedusing nonparametric comparisons. Median scores were com-pared using Wilcoxon signed ranks or Mann-Whitney with anSSPS software package.

IHC. IHC was performed on paraffin-embedded tissuesections from all tumors studied by ISH by methods describedpreviously (11, 31). TIGd C rabbit polyclonal antibody raised torecombinant TIG-his tag protein was used as the primary antibody4 C. Schulzet al., manuscript in preparation.

3250Retinoid-inducible Tumor Suppressor in Skin Cancer

Research. on June 16, 2018. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Page 3: Expression of a Retinoid-inducible Tumor Suppressor ...clincancerres.aacrjournals.org/content/clincanres/6/8/3249.full.pdfExpression of a Retinoid-inducible Tumor Suppressor,Tazarotene-inducible

at a dilution of 1:1000 and was detected with horse antirabbitsecondary antibody at 1:100 using a Vector Stain ABC kit (VectorLabs, Burlingame, California). Preimmune rabbit serum (1:1000)was used as a negative control for TIG-3 antibody staining.5

QT-RT-PCR mRNA Analysis. Total RNA was ex-tracted from normal and tumor specimens (Table 2) using Trizolreagent (Life Technologies, Inc., Gaithersburg, MD). Contam-inating DNA was removed by DNase I treatment in 4 mM MgCl2at 37°C for 30 min and then terminated at 72°C for 5 min. ThePCR primers (45F-CAAGAGCCCAAACCTGGAG and 111RTATACAGGGCCCAGTGCTCAT) and fluorescent oligonu-cleotide probe (65T-CCTGATTGAGATTTTCCGCCTTGGC)for the TIG-3 QT-RT-PCR assay were designed from the pub-lished sequence, accession no. AR060228 (38), using the PrimerExpress software (Perkin-Elmer). Primers were purchased fromIntegrative DNA Technologies (Coralville, IA). RNA samples(100 ng) were reverse transcribed with SuperScript reversetranscriptase (1 unit/ml; Boehringer Mannheim) in a 20-ml totalvolume reaction that included either Tig3 or 36B4 reverseprimer (750 nM), 13 PCR buffer (Perkin-Elmer), 4 mM MgCl2,and 500mM deoxynucleotide triphosphates. The 36B4 forwardprimer-AGATGCAGCAGATCCGCAT and reverse primer-ATATGAGGCAGCAGTTTCACCAG were used to amplifycDNA for 36B4 standard and detected with the probe-AGGCT-GTGGTGCTGATGGGCAAGAAC. The reaction was incu-

bated at 50°C for 30 min and then terminated by incubation at72°C for 5 min. After reverse transcription, 13 PCR buffer, 4 mM

MgCl2, 500 mM deoxynucleotide triphosphates, 300 nM forwardand reverse primer, 100 nM fluorescent oligonucleotide probe, andTaq polymerase were added to make a final volume of 50ml.

PCR was performed in an ABI 7700 Prizm SequenceDetector (Perkin-Elmer) with preheating at 95°C for 1 min,followed by 40 cycles of melting (95°C for 12 s) and annealingand extension (60°C for 60 s). Dequenching of the fluorescenthybridization probe was continuously monitored, and theCt(PCR cycles to threshold) were calculated using Sequence De-tector 1.6 software (Perkin-Elmer). Unknown RNA sampleswere analyzed in triplicate with a control that lacked reversetranscriptase (2RT control). Samples were run in parallel with10-fold dilutions of mRNA from tumor line 22B HNSCC(Ref. 45; 600 ng to 600 pg) and assigned arbitrary number ofmolecules to generate a standard curve (Ct versusarbitrary RNAtemplate molecules). The number of template molecules in thesamples were calculated by interpolation of theCt against the22B HNSCC RNA standard curve. 36B4 mRNA was alsomeasured by QT-RT-PCR and used to normalize as a standard,using the primers described previously (46).

RESULTSTIG-3 mRNA and Protein Is Expressed Highly in the

Suprabasalar Epidermis and Adnexa of Normal Skin. IfTIG-3 acts as a tumor suppressor and controls normal epithelialdifferentiation, then it would be expressed in normal epidermis.The intensity of the antisense riboprobe staining was strongest

5 R. L. Eckert, N. A. Robinson, and A. Deucher, manuscript in prepa-ration.

Table 1 TIG-3mRNA hybridization in SCC tumors compared with normal overlying or adjacent skin specimens

Aggressive SCCs

Site/Metastasis Tumor no.

Differentiation TIG-3 levels compared with tumor

Well Moderate Poor Overlying skin Adjacent skin

? 96-36569 1 2 2 Marked patchyincrease

NAa

Ear 96-34065A24 2 1 2 2-fold increase NAParotid, LN, PNUpper lip, Cheek 96-38303A4 2 1 2 2.66-fold increase NAEar 96-54304 2 1 2 2.33-fold increase NA? 97-4570C13 2 1 2 Marked increase NANose 96-39845A8 2 1 2 1.2-fold increase No changeScalp/skull, multiple tumors 96-41288A7 2 1 2 2-fold increase 1.5-fold increaseFace, multiple tumors 96-51845A11 2 1 2 No change NAInfraorbital/PN 96-38787A1 2 2 1 2-fold increase NADermis/parotid, L, V 96-35269A21 2 2 1 2.33-fold increase NAParotid/LN 97-38951E16 2 2 1 NA NA

Non-aggressive squamous cell carcinomas

Elbow 96-05316A1 1 2 2 1.2-fold increase 1.4-fold increaseRetroauricular 96-51337A2 1 2 2 No change No changeForehead 97-00664D2 1 2 2 No change 1.5-fold increaseEar 96-46239A6 2 1 2 1.25-fold increase 1.5-fold increaseForehead 97-09809A6 2 1 2 2-fold increase 2.66-fold increaseForehead 97-20557A3 2 1 2 1.2-fold increase 1.6-fold increaseNose 96-35773A6 2 2 1 1.5-fold increase 1.5-fold increaseForehead 97-53342A2 2 2 1 1.5-fold increase 2-fold increaseForehead 97-02450A1 2 2 1 1.5-fold increase 1.5-fold increasea NA, not available; LN, lymph node; PN, perineural invasion; L, lymphatics; V, vasculature.

3251Clinical Cancer Research

Research. on June 16, 2018. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Page 4: Expression of a Retinoid-inducible Tumor Suppressor ...clincancerres.aacrjournals.org/content/clincanres/6/8/3249.full.pdfExpression of a Retinoid-inducible Tumor Suppressor,Tazarotene-inducible

by in situ hybridization throughout the epidermis, hair follicles,and sebaceous glands of normal control skin (Fig. 1,A andB).In thinner epidermis of normal skin (Fig. 1,A andG), all layersstained strongly with the antisense probe. The sense probeshowed only faint background staining intensity (Fig. 1C).Where the epidermis was thicker, staining intensity was morepronounced in the suprabasal layers than in the basal layers innormal skin (Fig. 1B) and in hyperproliferative psoriatic epider-mis (Fig. 1,H and I).

Immunoreactive TIG-3 protein staining in the upper layersof normal skin (Fig. 1,D andE) was similar to the pattern of themRNA seen byin situ hybridization (Fig. 1,A and B). TIG-3protein was more restricted than the mRNA because it was mostintense in the most well-differentiated layers of the normalsuprabasal epidermis, just under the stratum corneum (Fig. 1,Dand E). The hair follicles, sebaceous glands, and small bloodvessels of the dermis also stained prominently with TIG-3antibody. The immunoreactivity was cytoplasmic in appearancein keratinocytes. Preimmune rabbit serum gave only backgroundreactivity in normal skin (Fig. 1F).

TIG-3 mRNA Expression Is Significantly Decreased inPsoriatic Epidermis Compared with Paired Normal, Unin-volved Skin. Normal, uninvolved skin from psoriasis pa-tients was similar in intensity and pattern of mRNA stainingto normal controls without skin cancer in mRNAin situhybridization intensity (Fig. 1G). Psoriasis lesions prior totreatment (Fig. 1H) were severalfold thicker than normal skinbecause of epidermal acanthosis and increased numbers ofsuprabasalar keratinocytes. The intensity of staining forTIG-3 antisense probe was reduced in the basal layers ofpsoriasis lesions (Fig. 1H) compared with normal controlskin (Fig. 1G). As shown in Fig. 2, normal, uninvolved pairedskin from 30 psoriasis patients (median, 4.0) had signifi-cantly more intense TIG-3 antisense staining than the basaland suprabasal layers of psoriasis lesions (median, 3.0;P 50.012, Wilcoxon signed ranks test). Suprabasal TIG-3 stain-ing was more intense in paired, uninvolved normal skinspecimens compared with baseline psoriasis lesions in 16 of30 patients’ lesions examined, was less intense in 4, and didnot differ by more than 1 degree in 10 patients.

Pooled mRNA had been analyzed previously from lesionsin this clinical study from the 15 responding patients who had.50% decrease in plaque elevation by 8 weeks (38). In theresponders, topical Tazarotene gel application was associatedwith increased levels of totalTIG-3 mRNA transcripts as com-

pared with untreated lesions by a quantitative assay (38). In theanalysis done by considering responding and nonrespondingpatients together, the median staining intensities after Tazaro-tene treatment for 3 or 14 days, as judged by ISH, were notsignificantly different from baseline (Fig. 2).

TIG-3 Expression Is Significantly Decreased in BothBasal and SCCs Compared with Adjacent and OverlyingNormal Skin. BCCs and SCCs studied by ISH were derivedfrom sun-exposed areas (except one case) from subjects be-tween 50 and 80 years of age (Table 1). All except 2 patientswere males. SCCs of the skin were defined as aggressive if theywere.2 cm in size, had neural invasion, and/or were metastaticto regional lymph nodes. Aggressive tumors studied extendedinto deeper structures including auricle, periauricular soft tis-sues, parotid glands, intraorbital nerves, scalp, nose, and intrapa-rotid lymph nodes. As indicated in Table 1, some tumors did nothave paired overlying or adjacent skin available to study. Non-aggressive tumors were restricted to scalp and superficial facialtissues and were,2 cm in size.

We examined the median intensity ofTIG-3 mRNA insitu staining comparing normal skin from psoriasis patientsas well as overlying and adjacent paired normal skin fromskin cancer patients to their tumors (Fig. 2). As shown inFigs. 2– 4, the median staining intensity of the TIG-3 anti-sense probe was decreased in BCCs, basosquamous carcino-mas, and SCCss when compared with normal skin from psoriasispatients and from paired tumor patients. Psoriasis control normalskin and paired adjacent normal suprabasalar layers of skin hadsimilar staining (median, 4.0). Both were significantly higher inTIG-3 mRNA staining than basal cell carcinomas (median, 1.33;P 5 0.045, Mann-Whitney;n 5 6). BCCs studied by ISH showedreduction inTIG-3mRNA staining in the tumor that was similar tothe normal overlying basal cell layer (Fig. 3A). IHC stainingintensity in the suprabasal layer of adjacent skin (Fig. 3B) washigher than in the basal layer of normal skin and in basal cellcarcinomas shown in Fig. 3,D and E. Normal skin showed nospecific reactivity with preimmune serum and secondary antibody(Fig. 3C).

The staining intensity of the TIG-3 antisense riboprobe byISH was also significantly higher in normal skin specimensfrom all sources, compared with all SCCs as a group. As shownin Fig. 2, TIG-3 mRNA staining was stronger in normal skin frompsoriasis controls (median, 4.0;P 5 0.001), in normal skin over-lying SCCs (median, 3.0;P 5 0.0001), and in the adjacent skin toSCCs (median, 3.0;P 5 0.007) compared with staining intensity in

Table 2 Specimens from adjacent normal skin and tumors for QT-RT-PCR RNA measurements

Patient no. Age/sex Diagnosis Sites of biopsies TIG-3 levels: normalvs. tumor

1 65 WM BCE Left back Similar to NLa skin2 62 WM Well-differentiated SCC Nasal root Tumor. NL3 70 WF BCE Left forehead Both low4 77 WF Benign lichenoid keratosis Right leg Similar to NL control5 61 WF BCC Right upper cheek Normal. tumor6 66 WM SCCin situ Right upper cheek Both low7 71 WM BCE Right preauricular Nl skin low

BCE Left preauricular N1 skin lowBCE Posterior neck N1 skin low

a NL, normal; W, white; M, male; F, female.

3252Retinoid-inducible Tumor Suppressor in Skin Cancer

Research. on June 16, 2018. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Page 5: Expression of a Retinoid-inducible Tumor Suppressor ...clincancerres.aacrjournals.org/content/clincanres/6/8/3249.full.pdfExpression of a Retinoid-inducible Tumor Suppressor,Tazarotene-inducible

21 aggressive and nonaggressive SCC specimens as a group (me-dian, 1.5). Examples of aggressive SCCs shown in Fig. 4 had eitheroverlying or adjacent normal skin. Aggressive SCCs were.2 cmand had neural invasion or metastasis to parotid gland or regionallymph nodes. For this reason, they were not always available withnormal adjacent or overlying skin (Table 1).

All SCC specimens were studied both by ISH and by IHC.There is reducedTIG-3 mRNA signal in SCC tumors comparedwith overlying normal skin with higher expression in adjacentskin. All tumors showed correlation in the levels of mRNAstaining and TIG-3 protein reactivity. (Examples of specimensstudied by both techniques are found in Fig. 4,B andC, E andF, I andJ, M andN). We did not quantitate the IHC staining butfound it to be at least 2-fold or greater in the suprabasal epider-mis than in tumor specimens. The most highly differentiatedareas of keratin pearls were highest in bothTIG-3 mRNA andprotein (Fig. 4,H, M, andN).

The variation in staining intensities found in normal skinoverlying both aggressive and nonaggressive SCC tumors, fromwhich the tumors arose, was greater than the variation seen innormal control skin as well as in adjacent skin (Fig. 2). Fig. 4shows examples of overlying normal skin with highTIG-3 (Band D), moderateTIG-3 (A), and low TIG-3 (E, L) mRNAstaining. Pseudoepitheliomatous hyperplasia found in two tu-mors (DandE) could be either high or low in staining intensity.Adjacent skin at a distance from tumors (Fig. 4,G, I, and K)showed very strongTIG-3mRNA and protein staining as a rule,but the intensity was decreased at the site where tumors arose(Fig. 4,G–JandK–N). The mRNA and protein staining patternswere similar, as shown.

TIG-3 Expression Is Less in Aggressive SCCs Com-pared with Nonaggressive SCC Tumors. The median stain-ing intensity in 11 aggressive SCCs (median, 1.0) was alsodecreased compared with 10 nonaggressive SCCs (median,

Fig. 1 TIG-3mRNA and protein levels are high in normal skin and reduced in psoriasis lesions.TIG-3 mRNA was detected in paraffin-embeddedspecimens of skin by ISH, using digoxigenin-labeled cRNA probes (Boehringer Mannheim) and antibody to digoxigenin (35, 44). TIG-3immunoreactivity was assayed using a rabbit polyclonal antibody raised to recombinant TIG-3 protein at a dilution of 1:1000 and was detected byhorse antirabbit secondary antibody with a Vector Stain ABC kit (Vector Laboratories, Burlingame, CA).A, normal skin specimen shows dense purplesignal in epidermis, hair follicles, and sebaceous glands with antisense probe (340).B, high power view of epidermis from normal skin specimenadjacent to tumor with ISH signal to antisense probe (3400).C, normal skin specimen byin situ hybridization to sense riboprobe as negative control(3100).D, normal skin from patient without skin cancer stained with antibody to TIG-3 protein (1:1000) detected with horse antirabbit IgG (1:100;340). E, specimen inD at high magnification (3100).F, specimen inD stained with preimmune rabbit serum (1:1000) and detected with horseantirabbit secondary antibody (1:100) as negative control (3100).G, ISH of normal unaffected skin from a psoriasis patient before treatment withantisense riboprobe probe (3100).H, same patient’s psoriasis lesion at baseline, antisense probe (3100).I, psoriasis lesion fromG at day 3 ofTazarotene application, antisense probe (3100).

3253Clinical Cancer Research

Research. on June 16, 2018. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Page 6: Expression of a Retinoid-inducible Tumor Suppressor ...clincancerres.aacrjournals.org/content/clincanres/6/8/3249.full.pdfExpression of a Retinoid-inducible Tumor Suppressor,Tazarotene-inducible

1.5;P 5 0.038, Mann-Whitney; Fig. 2). Ten of 11 aggressiveSCC versus7 of 10 nonaggressive SCCs showed decreasedTIG-3 mRNA ISH signals compared with paired normal skin.All tumors studied showed reducedTIG-3 mRNA and proteinstaining compared with normal control donor skin specimens.Three aggressive SCCs and a metastatic SCC lesion to thelymph node had no staining present forTIG-3 mRNA, sug-gesting that complete loss of gene expression may have occurred.In three nonaggressive SCCs, overlying skin and tumor were sim-ilarly both reduced (compared with normal skin). Two of thesethree tumors were interpreted by the pathologist as well-differen-tiated, and the third was poorly differentiated.

Development of QT-RT-PCR Assay to MeasureTIG-3mRNA Transcripts. A quantitative fluorescent real time QT-RT-PCR assay (47) was developed to measureTIG-3 mRNAtranscripts in paired patient specimens (Table 2). The cDNA for36B4, encoding a ribosomal protein, was used to normalizelevels of TIG-3 mRNA, expressed as a ratio. Levels ofTIG-3mRNA transcripts were determined from two samples of nor-mal, non-sun-exposed abdominal skin from patients withoutcancer and from 7 skin tumors with paired adjacent skin spec-

imens taken at the time of Moh’s surgery (Table 2; Fig. 5).TIG-3 mRNA levels from triplicate samples of reverse-tran-scribed mRNA were normalized to 36B4. Two specimens fromnormal abdomen skin of patients without skin cancer (Fig. 5,AandB) were similar and showed highestTIG-3 mRNA levels,similar to the levels measured in HaCat keratinocytes (notshown). High levels ofTIG-3 mRNA were also found in onevery well-differentiated SCC (patient 2) and in a lichenoidactinic keratosis with epidermal acanthosis (patient 4). A small,well-localized BCE on the infraorbital cheek had reducedTIG-3transcripts compared with the normal adjacent skin specimen(patient 5). Similar to thein situ findings, three patients had lowTIG-3 levels in both tumor and adjacent normal skin. Patient 7,with lowest TIG-3 in normal adjacent skin, also had multiplefacial basal cell carcinoma occurring in the setting of severe,extensive photodamage over the whole area.

DISCUSSION

TIG-3 (RANTES-3) is a newly described, retinoid-inducibleputative tumor suppressor that was isolated from Tazarotene-

Fig. 2 Median values of staining intensity of ISH antisense TIG-3 riborobes; normal skin is higher than psoriasis lesions and SCCs. Paired normalskin, psoriasis lesions, and tumors mRNA staining intensities were graded in the suprabasal epidermis or in the tumor using a semiquantitative scaleof 0–5 with 0.5 increments: 0, no staining; 1, faint; 2, light; 3, moderate; 4, dark; and 5, most intense, black dense staining. Wilcoxon signed ranksof related or independent variables were performed on nonparametric data from each group. Comparisons are shown as standard box plots of eachgroup: psoriasis patients’ normal skin (black hatch), and lesions at baseline (PsL0), after 3 days of Tazarotene (PsL3), and at day 14 of therapy (PsL14,white boxes) are at theleft side. Skin cancer specimens of normal overlying skin (nl_over) and nonaggressive SCC (nl_nonaggr) are shown bygrayhatched boxesin middle. All SCCs (Scc_all), aggressive SCCs (SCC_agg), and nonaggressive SCCs (SCC_nagg) are indicated at theright sidewithgray boxes. Theblack horizontal barswithin the boxesrepresent the median values for ISH staining intensities. TheX axisshows the number ofspecimens analyzed in each group. TheY-axisshows the intensity of mRNA staining intensity from 0 to 5. Theboxessurrounding themediansindicatethe values falling between the 25th and 75th percentiles. Ranges are shown by thevertical bars. Outliers, shown bycircles,are included in calculationof the medians.

3254Retinoid-inducible Tumor Suppressor in Skin Cancer

Research. on June 16, 2018. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Page 7: Expression of a Retinoid-inducible Tumor Suppressor ...clincancerres.aacrjournals.org/content/clincanres/6/8/3249.full.pdfExpression of a Retinoid-inducible Tumor Suppressor,Tazarotene-inducible

treated keratinocytes.TIG-3 has been shown to inhibit cellularproliferation when expressed and is associated with retinoidresponsiveness in malignant cell linesin vitro (38). This studyshows for the first time thatTIG-3 mRNA and protein expres-sion are significantly reduced in basal and SCCs arising fromsun-exposed skin, compared with normal skin. LowerTIG-3mRNA and loss of expression are seen in aggressive SCCscompared with nonaggressive SCCs in this study.

The highest levels ofTIG-3 mRNA by ISH and TIG-3protein by IHC are found within the normal suprabasal epider-mis, hair follicles, and sebaceous glands. TIG-3 in normal skinmay be, therefore, associated with terminal keratinocyte differ-entiation and growth arrest occurring in the suprabasal layers. Inhyperproliferative epidermis, as in psoriasis lesions and overly-ing SCCs,TIG-3mRNA by ISH and protein by IHC are reducedrelative to skin of normal thickness.

Treatment of hyperproliferative psoriasis lesions with top-ical Tazarotene gel induces 4-fold increased expression ofTIG-3 mRNA in epidermis of responding patients comparedwith pretreatment levels (38). In psoriasis lesions, inflammationis present that affects epidermal differentiation, and the skin is ableto normalize after topical retinoid therapy (28, 33). Up-regulationof TIG-3 expression may be part of the normalization process, andlow TIG-3 expression may contribute to the increased rate ofepidermal proliferation that characterizes psoriasis (48).

Recently, topical Tazarotene gel treatment has also beenshown to induce clinical remission of 47% of basal cell carci-nomas treated over a period of months (34). Whereas loss ofTIG-3 may be an important event leading to skin cancer, ex-pression of TIG-3 in suprabasalar epidermis may help to regu-late normal terminal differentiation.TIG-3 mRNA levels arehighest in normal skin and differed significantly between non-

Fig. 3 TIG-3 staining is higher in suprabasalar normal skinthan in BCCs.A, BCC with overlying skin studied by ISH, asdescribed in Fig. 1, using antisense riboprobe (3100).B, normalskin adjacent to BCE showing TIG-3 protein by IHC, as inA, inepidermis and dermal vessels (340).C, same section asB, withpreimmune serum as negative control (340).D, diffuse BCC indermis stained with antibody to TIG-3 (3100).E, well-demar-cated nodule of BCE with adjacent normal skin stained withTIG-3 antibody (3100).

3255Clinical Cancer Research

Research. on June 16, 2018. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Page 8: Expression of a Retinoid-inducible Tumor Suppressor ...clincancerres.aacrjournals.org/content/clincanres/6/8/3249.full.pdfExpression of a Retinoid-inducible Tumor Suppressor,Tazarotene-inducible

Fig. 4 Expression ofTIG-3 mRNA by ISH and protein by IHC is reduced in paired adjacent and overlying normal skin compared with SCCs.Specimens were studied as in Fig. 1 by either ISH or IHC.A, moderately well-differentiated, aggressive SCC in dermis with normal overlying skinby ISH (340).B, moderately well-differentiated, aggressive SCC (34065) in dermis with normal overlying skin by ISH (340). C, same tumorB byIHC with normal epidermis toleft, perpendicular to the base of photo and tumor nodules to theleft (340).D, moderately well-differentiated aggressiveSCC (38303) from face studied by ISH. Tumor is throughout the dermis with pseudoepithelomatous hyperplasia in overlying normal skin (340).E,

3256Retinoid-inducible Tumor Suppressor in Skin Cancer

Research. on June 16, 2018. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Page 9: Expression of a Retinoid-inducible Tumor Suppressor ...clincancerres.aacrjournals.org/content/clincanres/6/8/3249.full.pdfExpression of a Retinoid-inducible Tumor Suppressor,Tazarotene-inducible

aggressive and aggressive SCCs. Therefore, with further inves-tigation, TIG-3 may become the first molecular marker of ag-gressiveness and of retinoid action during treatment orchemoprevention studies. These new data support the rational ofusing retinoids for both prevention and treatment of skin cancer.

TIG-3 is homologous to a class II tumor suppressor foundin H-ras revertant lines, H-rev 107 (40, 49). TIG-3 mapped tochromosome 11q23, which has long been hypothesized to con-tain a tumor suppressor (38). This region has shown loss ofheterozygosity and poor prognosis in cancers of the head andneck, melanoma, cervix, breast, lung, and ovarian malignanciesand also leukemias, lymphomas, and rhabdomyosarcomas (38,42, 43, 50–54). With respect to head and neck tumors, 25%showed loss of heterozygosity at 11q23 with a significant asso-ciation with persistent or recurrent disease after radiation (55).The expression ofTIG-3mRNA in most aggressive SCC tumorsis reduced by at least 50% compared with paired specimens ofadjacent or overlying normal skin byin situ hybridization stud-ies, suggesting that loss of heterozygosity may be present. All11 aggressive SCC tumors have markedly reduced TIG-3 ex-pression levels compared with overlying or adjacent normalskin. In three aggressive skin tumors, one a metastasis to node,TIG-3 mRNA expression is absent. Loss of TIG-3 protein in

aggressive SCC is also confirmed by IHC. Analysis of theTIG-3 promoter, coding sequences, and gene sequences will berequired to understand the basis of this finding.

Loss of heterozygosity in tumors reflects deletion of a largeportion of the genomic DNA containing one allele and should beassociated with a 50% reduction of the mRNA transcripts.Complete inactivation of a tumor suppressor requires that thesecond allele also acquire a second mutation within the tumortissue. It is possible that inactivation of both copies of TIG-3may be associated with the development of more aggressiveskin cancers, and this would agree with studies demonstratingloss of chromosome 11q23 in head and neck SCCs (55). Incontrast, mutations may be subtle and may arise without con-sistently altering mRNA levels (56). Loss of TIG-3 expressionmay also result from mutations or small deletions in the retinoidresponse elements (38) or hypermethylation of the gene’s pro-moter. More direct measures of mRNA levels, such as theQT-RT-PCR assay we report, and sequence information areultimately required to understand the importance of tumor sup-pressors in carcinogenesis. ISH localized mRNA expressionwithin tumor specimens, but it is only semiquantitative.

Basal cell carcinomas, generally nonaggressive, localizedtumors arising from differentiated hair follicle keratinocytes,express levels of TIG-3 similar to the basal layer of normalepidermis. In keeping with the hypothesis that TIG-3 is a tumorsuppressor, basal layers have lessTIG-3mRNA and protein thanmore differentiated epidermal layers. A quantitative, sensitiveQT-RT-PCR assay was developed to measure TIG-3 transcriptswithin specimens. By this assay, normalizedTIG-3 mRNAlevels are 2–4-fold higher in normal non-sun-exposed skin fromcontrols without cancer than in normal skin adjacent to threebasal cell carcinomas and one SCC. One BCE and one SCCshowed only slightly lower levels, and one very well-differen-tiated, nonaggressive SCC was 3-fold increased in tumor com-pared with normal skin. In these studies, there was no opportu-nity to study mRNA transcripts in aggressive SCCs and nomechanism to control for the relative amount of tumor andnormal skin contributing material to total mRNA transcriptpools. Microdissection may be required to dissociate tumorspecimen from normal skin. IncreasedTIG-3mRNA levels maybe measured from a large tumor mass, even if the relative levelsare low, in comparison with normal skin.

Overlying normal skin has more variable expression ofTIG-3 compared with adjacent skin specimens; therefore, wealso used normal skin specimens from psoriasis patients ascontrols. Skin overlying tumors may have lowTIG-3 mRNA byISH and low protein by IHC. In some SCCs where the point oftumor origin was present, the normal skin at the tumor origin haslow TIG-3 staining. These observations, although preliminary,

poorly differentiated aggressive SCC (38787) of the infraorbital area studied by ISH. Overlying skin shows decreased TIG-3 and pseudoepithelo-matous hyperplasia (340).F, same tumor as inE by IHC (340).G, moderately well-differentiated aggressive SCC (412288) by ISH shows normaladjacent skin toleft, take off point to tumor (middle) and tumor mass (left;340). H, same patient, SCC in dermis by ISH. Keratin pearls are presentin dermis (340).I, same patient with BCC next to normal epidermis and hair follicle (left) by ISH (340).J, same section by IHC with TIG-3 antibody(340). K, ISH of normal adjacent skin with hair follicle in patient with tumor (39845) that is shown inL–N (340). L, ISH of normal overlying skinand moderately well-differentiated SCC (39845;340). M, same tumor by ISH (340) showing normal adjacent skin (left) and point of tumor originmiddle (340).N, same tumor section asM, by IHC (340).

Fig. 5 QT-RT-PCR measurement of TIG-3 mRNA transcripts in nor-mal skin and tumor specimens. Total RNA was prepared from fresh skinsamples from tumor and paired normal skin taken from surgery cases(Table 2). One hundred ng of mRNA were reverse transcribed withreverse transcriptase. Triplicates were subjected to PCR using forwardand reverse primers with a fluorogenic primer. The release of thefluorescence by 59nuclease activity was measured in real time using anABI 7700 prism sequence detector. TIG-3 signals were normalizedagainst the standard 36B4, and the graphs represent the ratios.ColumnsA andB, TIG-3mRNA levels from normal abdominal skin taken fromsurgery in two older patients without skin cancer. Paired normal andtumor samples were compared from seven patients undergoing surgeryin Table 2. Normalized values for TIG-3:f, in normal skin; and , intumors.

3257Clinical Cancer Research

Research. on June 16, 2018. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Page 10: Expression of a Retinoid-inducible Tumor Suppressor ...clincancerres.aacrjournals.org/content/clincanres/6/8/3249.full.pdfExpression of a Retinoid-inducible Tumor Suppressor,Tazarotene-inducible

do suggest that loss of TIG-3 in normal skin may be associatedwith the development of SCC tumors. The QT-RT-PCR assayshowed thatTIG-3 mRNA transcripts are low in both the per-ilesional photodamaged skin and in several basal cell carcino-mas. One patient with SCC had lowTIG-3 mRNA both in normalskin and tumor by QT-RT-PCR. If TIG-3 is lost or reduced inoverlying normal skin, there could be a field effect, as reported forloss of p53 both in sun-damaged skin and resulting tumors (57, 58).Mutations in p53 are reported in 90% of SCCs and 50% of basalcell carcinomas and result from UV light-induced pyrimidine-cytosine photoproducts (57). Whether these types of changes andthis kind of frequency will also be found in skin cancers withrespect to TIG-3 is not yet determined.

Heterogeneity of mRNA expression is present in sometumors, especially when there was mixed differentiation. This wasconfirmed using IHC showing high TIG-3 within areas of kerati-nization (keratin pearl formation). Heterogeneity in TIG-3 expres-sion and ability to be induced by retinoid treatment might help toexplain the clinical heterogeneity in response to retinoids.

It is well known that retinoids may prevent or retard thegrowth of some SCCs, although the mechanisms are not wellestablished (27, 59, 60). Whereas loss of heterozygosity at thechromosome 11q23 region has been found in 25% of recurrentSCCs of the head and neck, no specific gene has been implicateduntil now (55). On the basis of these new data, TIG-3 is a leadcandidate tumor suppressor explaining loss of heterozygosity atthis locus. Irreversible loss or decrease inTIG-3 mRNA andprotein in SCCs may be associated with more aggressive tu-mors, loss of retinoid response, and ultimately tumor progres-sion or invasion. Mutations in TIG-3 that alter its function couldalso occur without influencing the mRNA levels. It should benoted that the expression of TIG-3 is significantly lower inaggressive compared with nonaggressive SCCs, unlike DNArepair or p53 mutations (58, 61). TIG-3 is thus a possiblebiomarker for distinguishing between nonaggressive and ag-gressive biological behavior. Further studies involving the reg-ulation of TIG-3 by retinoids and elucidating the structure of theTIG-3 protein and the identification of key functional mutationswill be important to understand the role of TIG-3 in squamouscarcinogenesis and epidermal differentiation.

ACKNOWLEDGMENTSWe acknowledge the participation of Dr. Margaret Kripke, Dr.

Sarah Strom Lori Buschman, Dr. Adan El-Nagaar, Dr. Stephen Mays,Dr. Kendra Woods, and Annette Basey in obtaining and/or processingthe clinic specimens and lines for analysis. Dr. Peter Davies, Depart-ment of Integrative Biology, University of Texas Medical School, wasresponsible for the creation of the QT-RT-PCR assay.

REFERENCES1. Stoler, A., Kopan, R., Duvic, M., and Fuchs, E. The use of mono-specific antisera and cRNA probes to localize the major changes inkeratin expression during normal and abnormal epidermal differentia-tion. J. Cell Biol.,107: 427–446, 1988.2. Eckert, R. L., Crish, J. F., Banks, E. B., and Welter, J. F. The epidermis:genes on, genes off. J. Investig. Dermatol.,109: 501–509, 1997.3. Kanjali, S., and Duvic, M. Cutaneous carcinogenesis: molecularmodel of cancer development.In: S. Miller (ed.), Cutaneous Oncology.Cambridge: Blackwell Scientific Publications, Inc., 1998.4. Ziegler, A., Leffel, D. F., Kunala, S., Sharma, H., Gailani, M.,Simon, J. A., Halperin, A. J., Baden, H. P., Shapiro, P. E., Bale, A. E.,and Brash, D. E. Mutation hotspots due to sunlight in thep53 gene ofnon-melanoma skin cancer. Proc. Natl. Acad. Sci. USA,90: 4216–4220, 1993.

5. Glick, A. B., Sporn, M. B., and Yuspa, S. H. Altered regulation ofTGF-B1 and TGF-ain primary keratinocytes and papillomas expressingv-Ha-ras. Mol. Carcinog.,4: 210–219, 1991.6. Gailini, M. R., Bale, S. J., Leffell, D. J., DiGiovanna, J. J., Peck, G.L., Poliak, S., Drum, M. A., Pastakia, B., McBride, O. W., Kase, R.,etal. Developmental defects in Gorlin syndrome related to a putativetumor suppressor gene on chromosome 9. Cell,69: 111–117, 1992.7. Ananthaswamy, H. N., Loughlin, S. M., Cox, P., Evans, R. L.,Ullrich, S. E., and Kripke, M. L. Sunlight and skin cancer: inhibition ofp53 mutations in UV-irradiated mouse skin by sunscreens. Nat. Med.,3:510–514, 1997.8. Quinn, A. G., Campbell, C., Healy, E., and Rees, J. L. Chromosome9 allele loss occurs in both basal and squamous cell carcinomas of theskin. J. Investig. Dermatol.,102: 300–303, 1994.9. Wiching, C., Shanley, S., Smyth, I., Gillies, S., Negus, K., Graham,S., Suthers, G., Haites, N., Edwards, M., Wainwright, B., and Chenevix-Trench, G. Most germ-line mutations in the nevoid basal cell carcinomasyndrome lead to a premature termination of the PATCHED protein,and no genotype-phenotype correlations are evident. Am. J. Hum.Genet.,60: 21–26, 1997.10. Scott, R. E., Wilke, M. S., Wille, J. J., Pittelkow, M. R., Hsu, B. M.,and Kasperbauer, J. L. Human squamous carcinoma cells express com-plex defects in the control of proliferation and differentiation. Am. J.Pathol.,133: 374–380, 1988.11. Duvic, M., Chema, D., Annarella, M., Cho, M., Esgleyes-Ribot, T.,Remenyik, E., Ulmer, R., Rapini, R., Sacks, P. G., Clayman, G. L.,Davies, P. J. A., and Thacher, S. Keratinocyte transglutaminase expres-sion varies in squamous cell carcinomas. J. Investig. Dermatol.,102:462–469, 1994.12. Alkemade, H. A. C., Molhuizen, H. O. F., Vlijmen-Wilems, M. J. J.,Haelst, U. J. G. M., and Schalkwijk, J. Differential expression ofSKALP/elafin in human epidermal tumors. Am. J. Pathol.,143: 1679–1687, 1993.13. Baker, B. S., and Fry, L. The immunology of psoriasis. Br. J.Dermatol.,126: 1–9, 1992.14. Aneskievich, B. J., and Fuchs, E. Terminal differentiation in kera-tinocytes involves positive and negative regulation by retinoic acidreceptors and retinoid X receptors at retinoid response elements. Mol.Cell. Biol., 12: 4862–4871, 1992.15. Kliewer, S. A., Umesono, K., Mangelsdorf, D. J., and Evans, R. M.Retinoid receptor X interacts with nuclear receptors in retinoic acid,thyroid hormone, and vitamin D3 signalling. Nature (Lond.),355:446–449, 1992.16. Chambon, P. The molecular and genetic dissection of the retinoidsignaling pathway. Recent Prog. Horm. Res.,50: 317–332, 1995.17. Elder, J. T., Astrom, A., Pettersson, U., Tavakkol, A., Griffiths,C. E. M., Krust, A., Kastner, P., Chambon, P., and Voorhees, J. J.Differential regulation of retinoic acid receptors and binding proteins inhuman skin. J. Investig. Dermatol.,98: 673–679, 1992.18. Chakravarti, D., LaMorte, V., J., Nelson, M. C., Nakajima, T., Schul-man, I. G., Juguilon, H., Montminy, M., and Evans, R. M. Role of CBP/P300 in nuclear receptor signalling. Nature (Lond.),383: 99–102, 1996.19. Nagpal, S., Athanikar, J., and Chandraratna, R. A. S. Separation oftransactivation and AP1 antagonism functions of retinoic acid receptora. J. Biol. Chem.,270: 923–927, 1995.20. Saiki, R. K., Scharf, S. , Faloona, F., Mullis, K. B., Horn, G. T.,Erlich, H. A., and Arnheim, N. Enzymatic amplification ofb-globingenomic sequences and restriction site analysis for diagnosis of sicklecell anemia. Science (Washington DC),230: 1350–1354, 1985.21. DiSepio, D., Malhotra, M., Chandraratna, R. A. S., and Nagpal, S.Retinoic acid receptor nuclear-factor interleukin 6 antagonism: a novelmechanism of retinoid dependent inhibition of a keratinocyte hyperprolif-erative differentiation marker. J. Biol. Chem.,272: 25555–25559, 1997.22. Seitz, C. S., Lin, Q., Deng, H., and Khavari, P. A. Alterations in NF-kBfunction in transgenic epithelial tissue demonstrate a growth inhibitory rolefor NF-kB. Proc. Natl. Acad. Sci. USA,95: 2307–2313, 1998.23. Eichner, R., Kahn, M., Capetola, R. J., Gendimenico, G. J., andMezick, J. A. Effects of topical retinoids on cytoskeletal proteins:implications for retinoid effects on epidermal differentiation. J. Investig.Dermatol.,98: 154–161, 1992.

3258Retinoid-inducible Tumor Suppressor in Skin Cancer

Research. on June 16, 2018. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Page 11: Expression of a Retinoid-inducible Tumor Suppressor ...clincancerres.aacrjournals.org/content/clincanres/6/8/3249.full.pdfExpression of a Retinoid-inducible Tumor Suppressor,Tazarotene-inducible

24. Gunther, S. The therapeutic value of retinoic acid in chronic, acute,guttate, and erythrodermic psoriasis. Dermatology,89: 515–517, 1973.25. Ellis, C. N., and Voorhees, J. J. Etretinate therapy. J. Am. Acad.Dermatol.,16: 267–291, 1987.26. Ranki, A., Lauharanta, J., and Keanerva, L. Effect of etretinate onthe distribution of Langerhans cells and T lymphocytes in psoriatic skin.Arch. Dermatol. Res.,276: 102–104, 1984.27. Lippman, S. M., Lotan, R., and Clayman, G. L. Systemic preventionand treatment of skin cancer.In: R. S. Weber, M. J. Miller, and H.Goepfert (eds.), Basal and Squamous Skin Cancers of the Head andNeck, pp. 351–355. Baltimore: Williams and Wilkins, 1996.28. Duvic, M., Nagpal, S., Asano, A., and Chandraratna, R. A. S.Molecular mechanisms of Tazarotene action. J. Am. Acad. Dermatol.,37: 85–92, 1997.29. Weinstein, G., Krueger, G. G., Lowe, N. J., Duvic, M., Friedman,D. J., Jegasothy, B. V., Shmunes, E., Tschen, E. H., Lew-Kaya, D., Lue,J. C., Sefton, J., Gibson, J. R., and Chandraratna, R. A. S. Tazarotenegel, a new retinoid, for topical therapy of psoriasis. Vehicle-controlledstudy of safety, efficacy, and duration of therapeutic effect. J. Am. Acad.Dermatol.,37: 85–92, 1997.30. Duvic, M. The pharmacologic profile of Tazarotene. Cutis,61:22–26, 1998.31. Esgleyes-Ribot, T., Chandraratna, R. A., Lew-Kaya, D. A., Sefton,J., and Duvic, M. Response of psoriasis to a new topical retinoid, AGN190168. J. Am. Acad. Dermatol.,30: 581–590, 1994.32. Thacher, S. M., Krasinski, G., Malhotra, M., Duvic, M., Nagpal, S.,and Chandraratna, R. A. S. Coordinate regulation of keratinocyte dif-ferentiation markers by retinoids in cell culture and in psoriasis. J. In-vestig. Dermatol.,106: 819, 82, 1996.33. Duvic, M., Asano, A. T., Hager, C., and Mays, S. The pathogenesisof psoriasis and mechanism of action of Tazarotene. J. Am. Acad.Dermatol.,39: S126–S129, 1998.34. Peris, K., Fargnoli, M. C., and Chimenti, S. Preliminary observa-tions on the use of topical Tazarotene to treat basal-cell carcinoma.N. Engl. J. Med.,341: 1767–1768, 1999.35. Nagpal, S., Patel, S., Asano, A. T., Johnson, A. T., Duvic, M., andChandraratna, R. A. S. Tazarotene-induced gene 1 (TIG1), a novelretinoic acid receptor-responsive gene in skin. J. Investig. Dermatol.,106: 269–274, 1996.36. Nagpal, S., Patel, S., Jacobe, H., DiSepio, D., Ghosn, C., Malhotra,M., Teng, M., Duvic, M., and Chandraratna, R. A. S. Tazarotene-induced gene 2 (TIG-2), a novel retinoid responsive gene in skin.J. Investig. Dermatol.,109: 91–95, 1997.37. Nagpal, S., Thacher, S. M., Patel, S., Friant, S., Malhotra, M.,Shafer, J., Krasinski, G., Asano, R. T., Teng, M., Duvic, M., andChandraratna R. A. S. Negative regulation of two hyperproliferativekeratinocyte differentiation markers by an RAR-selective retinoid: in-sight into the mechanism of retinoid action in psoriasis. Cell GrowthDifferentiation,7: 1783–1791, 1996.38. DiSepio, D., Ghosn, C., Eckert, R. L., Deucher, A., Robinson, N.,Duvic, M., Chandraratna, R. A. S., and Nagpal, S. Identification andcharacterization of a retinoid-inducible class II tumor suppressor/anti-proliferative gene. Proc. Natl. Acad. Sci. USA,95: 14811–14815, 1998.39. Hajnal, A., Klemenz, R., and Schafer, R. Subtraction cloning ofH-rev107,a gene specifically expressed in H-ras resistant fibroblasts.Oncogene,9: 419–490, 1994.40. Szpiere, C., Szpiere, J., Riviere, M., Hajnal, A., Kiess, M., Scharm,B., and Schafer, R. Chromosomal assignment of three rat and humanH-revgenes, putative tumor suppressors, down-regulated in malignantlyHRAS-transformed cells. Mamm. Genome,7: 701–793, 1996.41. Kuchinke, W., Hart, R. P., and Jonakait, G. M. Identification ofmRNAs regulated by interferon-gin cultured rat astrocytes by PCRdifferential display. Neuroimmunomodulation,2: 347–355, 1995.42. di Iasio, M. G., Calin, G., Tibiltti, M. G., Vorchovsky, I., Benedikts-son, K. P., Taramelli, R., Barbanti-Brodano, G., and Negrini, M. Re-finement of the LOH region 1 at 11q23.1 deleted in human breastcarcinomas and sublocalization of 11 expressed sequence tags within therefined region. Oncogene,18: 1635–1638, 1999.

43. Murakami, Y., Nobukuni, T., Tamura, K., Maruyama, T., Sekiya,T., Arai, Y., Gomyou, H., Tanigami, A., Ohki, M., Cabin, D., Frisch-myer, P., Hunt, P., and Reeves, R. H. Localization of tumor suppressoractivity important in non-small cell lung carcinoma on chromosome11q. Proc. Natl. Acad. Sci. USA,95: 8153–8157, 1998.44. Xu, X. C., Clifford, J. L., Hong, W. K., and Lotan, J. R. Detectionof nuclear retinoic acid receptor mRNAs in histological tissue sectionsusing non-radioactivein situ hybridization histochemistry. Diagn. Mol.Pathol.,3: 122–131, 1994.45. Carey, T. Head and neck tumor cell lines.In: Atlas of HumanTumor Cell Lines, pp. 79–120. New York: Academic Press, Inc., 1994.46. Ritter, S. J., and Davies, P. J. A. Identification of a TGF-B1/BMP4response element within the mouse tissue transglutaminase gene pro-moter. J. Biol. Chem.,273: 12798–12806, 1998.47. Gibson, U. E. M., Heid, C. A., and Williams, P. M. A novel methodfor real time quantitative RT-PCR. Genome Res.,6: 995–1001, 1996.48. Greaves, M. W., and Weinstein, G. D. Treatment of psoriasis.N. Engl. J. Med.,332: 581–588, 1995.49. Husman, K., Sers, C., Fietze, E., Mincheva, A., Lichter, P., andSchafer, R. Transcription and translational down-regulation of H-Rev-107,a class II tumour suppressor gene located on human chromosome11q11–12. Oncogene,17: 1305–1312, 1998.50. Herbst, R. A., Gutzmer, R., Matiaske, F., Mommert, S., Casper, U.,Kapp, A., and Weiss, J. Identification of two distinct deletion targets at11q23 in cutaneous malignant melanoma. Int. J. Cancer,80: 205–209,1999.51. Hampton, G. M., Mannerman, A., Winquist, R., Alavaikko, M.,Blanco, G., Taskinen, P. J., Kiviniemi, H., Newsham, I., Cavenee,W. K., and Evans, G. A. Loss of heterozygosity in sporadic humanbreast carcinoma: a common region between 11q22 and 11q23.3. Can-cer Res.,54: 4586–4589, 1994.52. Gabra, H., Watson, J. E., Taylor, K. J., Mackay, J., Leonard, R. C.,Steel, C. M., Porteous, D. J., and Smyth, J. F. Definition and refinement ofa region of loss of heterozygosity at 11q23.3-q24.3 in epithelial ovariancancer associated with poor prognosis. Cancer Res.,56: 950–954, 1996.53. Bethwaite, P. B., Koreth, J., Herrington, C. S., and McGee, J. O. Lossof heterozygosity occurs at theD11S29locus on chromosome 11q23 ininvasive cervical carcinoma. Br. J. Cancer,71: 814–818, 1995.54. Launonen, V., Stenboack, F., Pusitola, U., Bloigu, R., Huusko, P.,Kytola, S., Kauppila, A., and Winquist, R. Chromosome 11q22.3–q25LOH in ovarian cancer: association with a more aggressive diseasecourse and involved subregions. Gynecol. Oncol.,71: 299–304, 1998.55. Lazar, A. D., Winter, M. R., Nogueira, C. P., Larson, P. S.,Finnemore, E. N., Dolan, R. W., Fulcihan, N., Chakravarti, A., Zietman,A., and Rosenberg, C. L. Loss of heterogeneity at 11q23 in squamouscell carcinoma of the head and neck is associated with recurrent disease.Clin. Cancer Res.,4: 2787–2793, 1998.56. Weinberg, R. A. Tumor suppressor genes. Science (WashingtonDC), 254: 1138–1146, 1991.57. Brash, D. E., Ziegler, A., Jonason, A. S., Simon, J. A., Kunala, S.,and Leffell, D. J. Sunlight and sunburn in human skin cancer: p53,apoptosis, and tumor promotion. J. Investig. Dermatol. Symp. Proc.,1:136–142, 1996.58. Kanjilal, S., Strom, S., Clayman, G. L., Weber, R. S., El-Naggar,A. K., Kapur, V., Cummings, K. K., Hill, L. A., Spitz, M. R., Kripke,M. L., and Ananthaswamy, H. N. p53 mutations in non-melanoma skincancer of the head and neck: molecular evidence for field cancerization.Cancer Res.,55: 3604–3609, 1995.59. Kraemer, K. H., DiGiovanna, J. J., Moshell, A. N., Tarone, R. E., andPeck, G. L. Prevention of skin cancer in xeroderma pigmentosum with theuse of oral isotretinoin. N. Engl. J. Med.,318: 1633–1637, 1988.60. Lippman, S. M., and Meyskens, F. L. Results of the use of vitaminA and retinoids in cutaneous malignancies. Pharmacol. Ther.,40: 107–122, 1989.61. Wei, A., Cheng, L., Wang, L., Guo, Z., Strom, S. S., Spitz, M. R.,Lippman, S. M., Duvic, M., and Kripke, M. L. Role of DNA repair forUV damage in non-melanoma skin cancer progression. Proc. Am. As-soc. Cancer Res.,39: 477, 1998.

3259Clinical Cancer Research

Research. on June 16, 2018. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Page 12: Expression of a Retinoid-inducible Tumor Suppressor ...clincancerres.aacrjournals.org/content/clincanres/6/8/3249.full.pdfExpression of a Retinoid-inducible Tumor Suppressor,Tazarotene-inducible

2000;6:3249-3259. Clin Cancer Res   Madeleine Duvic, Bharati Helekar, Claudia Schulz, et al.   Skin Cancer

Is Decreased in Psoriasis and-3,Tazarotene-inducible GeneExpression of a Retinoid-inducible Tumor Suppressor,

  Updated version

  http://clincancerres.aacrjournals.org/content/6/8/3249

Access the most recent version of this article at:

   

   

  Cited articles

  http://clincancerres.aacrjournals.org/content/6/8/3249.full#ref-list-1

This article cites 56 articles, 17 of which you can access for free at:

  Citing articles

  http://clincancerres.aacrjournals.org/content/6/8/3249.full#related-urls

This article has been cited by 10 HighWire-hosted articles. Access the articles at:

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  [email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://clincancerres.aacrjournals.org/content/6/8/3249To request permission to re-use all or part of this article, use this link

Research. on June 16, 2018. © 2000 American Association for Cancerclincancerres.aacrjournals.org Downloaded from