GENETIC DETERMINANTS OF MALIGNANCY

Role of and Tumor Suppressor Genes in Multistage

Stuart H. Yuspa, Andrzej A. Dlugosz, Christina K. Cheng, Mitchell F. Denning, Tamar Tennenbaum, Adam B. Glick, and Wendy C. Weinberg Laboratory of Cellular Carcinogenesis and Tumor Promotion, National Cancer Institute, Bethesda, Maryland U.S.A.

The introduction of the techniques of molecular biol­ tiation and enhanced growth rate of initiated cells im­ ogy as tools to study skin carcinogenesis has provided parta growth advantage when the epidermis isexposed more precise localization of biochemical pathways to promoters. The frequencyof premalignant progres­ that regulate the tumor phenotype. This approach has sion varies among papillomas, and subpopulations at identified genetic changes that are characteristic of high risk for progression have been identified. These each of the specific stages of squamous cancer patho­ high-risk papillomas overexpress the a6p4 integrin genesis: initiation, exogenous promotion, premalig­ and are deficient in transforming growthfactor P1 and nant progression, and malignant conversion. Initiation fJ2 peptides, two changes associated with a very high can result from mutations in a single gene, and the proliferation rate in this subset of tumors. The intro­ Harvey allele of the rasgene family has been identified duction of an oncogenic rasH- gene into epidermal cells as a frequent site for initiating mutations. Heterozy­ derived from transgenic mice with a null mutation in gous activating mutations in c-rasH- are dominant, and the TGFp1 gene have an accelerated rate of malignant affected keratinocytes hyperproliferate and are resist­ progression when examined in .,ivo. Thus members of ant to signals for terminal differentiation. An impor­ the TGFp gene family contribute a tumor-suppressor tant pathway impacted by c-rasH- activation is the pro­ function in carcinogenesis. Accelerated malignant tein kinase C (PKC) pathway, a major regulator of progression is also found with v-rasH- transduced kera­ keratinocyte differentiation. Increased activity of tinocytesfrom skinof mice with a null mutation inthe PKCa and suppression ofPKC6by tyrosine phospho­ gene. The similarities in risk for malignant con­ rylation contribute to the phenotypic consequences of version by initiated keratinocytes from TGp1 and p53 rasH- gene activation in keratinocytes. Tumor promot­ null geneotypes suggest that a common, growth-re­ ers disturb epidermal homeostasis and cause selective lated pathway may underly the tumor-suppressive clonal expansion of initiated cellsto produce multiple functions of these proteins in the skin carcinogenesis benign squamous papillomas. Resistance to differen- mode!.] Invest Dermatol103:90S-95S, 1994

he introduction of the techniques of molecular biol­ context of the intact epidermis. In cell culture, initiated keratino­ ogy as tools to study skin carcinogenesis has provided cytes display an altered response to signals for terminal differentia­ more precise localization of pathways that regulate tion, a characteristic that provides a selective growth advantage T the tumor phenotype. However, the application of under culture conditions favoring differentiation [2-4]. Exploita­ these tools to answer relevant questions has depended tion of this difference in vitro has been particularly helpful in isolat­ on the establishment of a conceptual framework that developed ing initiated keratinocytes of mouse and human origin [5-8]. over four decades of research on the biology and cellular physiology Exogenous tumor promotion causes the selective clonal expan­ of skin cancer induction by chemical exposures. These studies indi­ sion of initiated cells to produce multiple benign squamous cell cated that predictable and progressive stages occurred during the papillomas, each representing a clone of thousands of initiated cells. clonal evolution of a normal keratinocyte into a squamouscell car­ Potent exogenous promoters of the phorbol ester class activate, cinoma, and both genetic and epigenetic events contributed to these protein kinase C (PKC), and this enzyme activity accelerates epi­ changes [1]. Progress in understanding carcinogenesis wasenhanced dermalterminal differentiation [9-11]. Because initiated cells resist when certain properties associated with a particular stage of skin the induction of terminal differentiation by activators ofPKC [12], cancer pathogenesis were characterized [1]. Operational analyses, the differential response to phorbol esters favors the growth of the which defined the experimental requirements to produce a specific initiated subpopulation. This process recapitulates in vivo, the clonal neoplastic phenotype, have also been important for understanding selection of initiated cells by differentiation inducing agents in ke­ the process of tumor development (Fig 1). ratinocyte culture [13-15]. Squamous papillomas demonstrate a The earliest event documentedin skin carcinogenesis , initiation, �ign proliferation rate and delayed maturation, properties that are is carcinogen induced and mutational in nature. Initiated keratino­ analogous to the phenotype of individual initiated cells in vitro cytes express a subtle change in phenotype, unrecognizable in the [16,17]. Because most exogenous promoting agents are not muta­ genic [1], papillomas are diploid [18,19], and a single genetic change Reprint requests to: Dr. StuartYuspa, Building 37, Room 3B25, National in normal keratinocytes is sufficient to produce a papilloma pheno­ Cancer Institute, National Institutes of Health, Bethesda, MD 20892. type [20-22], the mechanism of exogenous promotion is likely to

0022-202Xj94jS07.00 Copyright © 1994 by The Society for Investigative Dermatology, Inc.

90S VOL. 103, NO.5, SUPPLEMENT, NOVEMBER 1994 ONCOGENES AND TUMOR SUPPRESSORS IN CARCINOGENESIS 91S

Initiation Promotion Premalignant Progression Malignant Conversion pression of mRNA for K1 and K10 is blocked, loricrin and fiJaggrin (Mutation) (Epigenetic) (Chromosomal Deletion, (Mutation) Duplication) expression areenhanced, and TGK is expressed constitutively (Fig • 2). Similar changes in expression of these epidermal markers had been documented in normal keratinocytes treated withPKC activa­ tors such asphorbol esters and diacylglycerol(Fig 2) [11,12,47,50]. Furthermore,diacylglycerol levels are markedlyelevated in v-rasH. Promoter Promoter keratinocytes and in keratinocytes expressing a mutated c-rasHa, Dependent Independent when comparedto normalepidermal cells [51,52]. Diacylglycerol is the endogenous activatorofPKC [53,54]. WhenPKC activitywas H. Figure 1. Operational stages in experimental skin carcinogenesis. measured in v-ras keratinocytes and compared to activity in nor­ Each stage is ed fined by the biologic consequences of a specific experimental mal epidermal cells, PKC activity that is Ca++-dependentwas in­ protocol.This scheme serves as a framework for molecular analysis of multi- creased whereas PKC activity that is Ca++-independent was de­ s. i stage carcinogenesis. See text and [1] for detail creasedin the v-rasHainitiated cells (not shown). To examine if the PKC pathway could be responsible for the phenotypic alterations in v-rasHa keratinocytes, inhibitors of PKC were used to determine be epigenetic inmost cases. In the absence of exposure to exogenous their influence on the expression of epidermal markersin cells initi­ H promoters, initiated skin rarely develops tumors. Thus, exogenous ated by v-ras a transduction. Both constitutive and Ca++-induced promotion is a rate-limiting early event in carcinogenesis. expression ofTGK mRNA in v-rasHakeratinocytes was prevented by Premalignant progression is most often a spontaneous process, inactivating PKC with bryostatin or GF 109203X, selective PKC independentof exogenous promoters [23,24]. Genetic studies indi­ inhibitors. Bryostatin treatment also restored the Ca++-inducedex­ catethat non-random sequentialchromosomal aberrations are asso­ pression ofK1 andK10 to v-rasH. cells (not shown). These findings ciated with premalignant progression [19,25-28]. This stage of support the hypothesis that rasHa-induced alterations in keratino­ cancer pathogenesis constitutes the major time-dependent compo­ cyte differentiationare the result of chronic activation of a Ca++-de­ nent of carcinogenesis and must involve repeated episodes of cell pendent isoform of PKC and reduced activity of a Ca++-indepen­ selection since modal dominance of a specificchromosomal aberra­ dent isoform of PKC. Thus, pharmacologic modulation of this tionindicates clonal outgrowth. Thus, at least one functionof the pathway may provide an approach to reverse the neoplastic pheno­ relevant genetic events occurringduring premalignant progression type. must imparta growth advantage on the affectedcell. Since PKC activation results int erminaldifferentiation of normal Malignantconversion of benign tumorsis a relatively rare occur­ keratinocytes [9-11], the chronic activation ofPKC measured in rence; less than 5% of papillomas spontaneously convert to cancers v-rasHa keratinocytes that did not spontaneously undergo terminal [23,24]. The risk for malignant conversion is variable among benign differentiation presented a paradox. Neoplastic keratinocyte cell tumors,and subpopulations of papillomaswith a much higher risk lines expressing an activated c-rasH• gene also have elevated diacyl­ for malignant conversion have been identified[2 9 -31]. Because the glycerol levels [52], but are resistant to terminal differentiationin­ conversion frequency can be significantly increased by exposing duced by either Ca++or phorbol esters. This suggested that qualita- animals bearing benign tumors to mutagens, the crucial events in the conversion process appear to be genetic in nature [23,24]. How­ ever, even mutagen-induced malignant conversion varies in fre­ Ca2+(mM) 0.05 0. 12 0.05 0.12 quency, and the subclass of papillomas at high risk for spontaneous r-,. r----JI '" conversion is also at higher risk for mutagen-induced conversion "0... "0... "0... :r ...... e... CI) [30]. Thus, the factors that determine risk for conversion to malig­ c c t!) c c 0 0 0 0 � nancy susceptibility to spontaneous or induced ge­ <{ I must increase U U 0 U U > netic changes. The identification of these factors would be an im­ portantgoal in understanding cancer pathogenesis. K1 THE METABOLIC BASIS OF INITIATION INDUCED BY A ras The mutational basisfor initiation of mouse epidermal carcinogen­ Ha esis was supported by the discovery that the c-ras gene was fre­ quently mutated in chemically-induced squamous papillomas, and Loricrin thesite of the mutation was determined by the initiating carcinogen [32-37]. A causal relationship for a rasmutation and initiation of skin carcinogenesis was established when an activatedv-rasHa onco­ gene was introduced into normal epidermis or culturedkeratino­ cytesand recipient cells developed into squamous papillomas in vivo Filaggrin 120,38]' Subsequent studies indicated that a significant number of human skin cancers also contained mutated c-ras genes [39-44]. We have used cultured keratinocytes into which the v-rasHagene was introduced via retroviral transduction to further characterize Figure 2. Protein kinaseC activation and the v-rasH- oncogene block the initiatedphenotype and to identify pathways affected by the expression of mRNA for spinous cell markers and enhance expres­ expressionof the v-rasHa p21 oncoprotein [45 -47]. sion of mRNA for granular cell markers. Mouse keratinocytes were Primary keratinocytes expressing the v-rasHa oncogene, intro­ cultured in Eagles medium containing 8% fetal bovine serum and a Ca++ duced by a defective retroviral vector, exhibit an altered response to concentration of 0.05 mM. Some cultures were transducedwith the v_rasH. signalsfor terminaldifferentiation in cell cultureand produce squa­ oncogene nsing a defective retroviral vector as described previously (20). a m u mous papillomas when grafted onto nude mice [20,45-47]. Thus, After 8 d in vitro subset of cultures was switched to edi m containing this single gene change can accomplish all the requirements for 0.12 mM Ca2+ for 24 h. In one group of cells the 0.12 mM Ca++ medium was supplemented with 125 pM 1-oleoyl-2-acetylglycerol (OAG), a syn­ initiation.In cultures of normal keratinocytes, increasingextracel­ thetic diacylglycerol that activates PKC [11]. Total RNAwas isolated, sepa­ mM lularCa++ from 0.05 to 0.12 induces both mRNA transcripts rated on a 1 % agarose gel containing0.6 M formaldehyde and Northernblot andprotein for severaldiff erentiationmarkers, includingkeratin 1 analysisperformed as describedpreviously (\ \, . A.tthe timepointexamined (Kl), KI0, loricrin, filaggrin, and keratinocyte transglutarninase (24 h), transcription of the later markers, loricrin and filaggrin, is low in (TGJ[48,49] (Fig 2). In v-rasHakeratinocytes, Ca++-mediated ex- normal keratinocytes (48). 9ZS YUSPA ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

tive changes in PKC isoforms may also contribute to the initiated phenotype and could account for differences in Ca++-dependent and independent PKC activity in v-rasfh keratinocytes. PKC is a familyof enzymes, members of which are differentiallyexpressed in particular tissues and cell types [55]. Five PKC isozymes are ex­ pressed in keratinocytes (a, �, E, C, ,,) [56]. Of these, PKCa is the only Ca++-dependent isoform. In an analysis of modification of these isoforms, we found tyrosine phosphorylation of PKC� in v-rasfh keratinocytes [57]. PKC� was also tyrosine phosphorylated in neoplastic keratinocytes containing an activated allele of the c­ H. ras gene. The activity of tyrosine phosphorylated PKC� was con­ stitutively low and could not be stimulated significantly by phorbol ester treatment, indicating that tyrosine phosphorylationblocks ac­ tivation of PKC� [57]. This interaction between tyrosine kinases Figure 3. Immunohistochemicalmarkers distinguish high riskand and PKC permits signaling through specific PKC isozyme in cells low riskpapillomas. Papillomaswere generated on the skinof SEN CAR expressing an oncogenic rasfh gene and may represent a molecular mice by initiation with 10 pg of7,12-dimethylbenz[a]anthracene and pro­ block to differentiation in neoplastic keratinocytes. motion with 12-0-tetradecanoylphorbol-13-acetate (TPA, 2 JAg, once weekly) to produce papillomas at low risk for malignant progression or CHARACTERISTICS OF TUMORS AT HIGH RISK FOR mezerein (4 JAg,twice weekly) to produce papillomasat high riskfor malig­ MALIGNANT CONVERSION nant progression [31]. High risk papillomas (a,b,e) were excised at 8 weeks when they were first detected clinically and only about 1-2 mm in diame­ Variations in protocols used for inducing skin tumors yield sub­ ter. Lowrisk papillomas (d,ef) were excised at 11 weeks, also at firstdetec­ groups of papillomas with either low or high riskfor progression to tion. Frozen sections were processed for immunochemistry using specific malignancy [32]. High risk tumors represent a relatively small co­ antibodies for the 0:6 integrin subunit (a,d), keratin 1 (b,e) and keratin 13 hort of papillomas. They are selectively induced by limited expo­ (ef). Reagents and procedures for frozen sectious and immunochemistry sure to strong promoters or by more prolonged exposure to weak have been describedpreviously [59]. Each panel represents a 5-p.m frozen promoters [32] . Thus, they appear to be more sensitive to exogenous section through the entiretumor. Bar, 200 JAm. tumor promotion. In addition, high risk papillomas erupt early, grow larger, and do not regress when promoter exposure is discon­ tinued. Thus, high risk papillomas have unique growth properties. We have performed comparative phenotypic analyses of high and strata [17]. In low-risk tumors, 90% of the proliferating cells are low risk papillomas to identify differences that may suggest a mo­ confinedto the basal compartment. In contrast, the majorityof high lecular basis for the altered biologic potential of this class of benign risk papillomas are devoid of both TGF-p1 and TGF-fJ2 as soon as tumors. Several tumor induction protocols were utilized that pref­ they arise; these tumors have up to 40% of the proliferating cells in erentially produce each papilloma subtype. The tumors were evalu­ the suprabasal layers [17]. Squamous cell carcinomas are also devoid ated at the time they first erupted (8 -11 weeks) and sequentially of TGF-p [17,64], suggesting that they arise from the TGF-p­ during the course of tumor progression. Particular attention was deficient high-risk papillomas. In some high-risk papillomas, concentrated on expression of specific integrins and cytokeratins as TGF-p loss can occur first and correlates with basal cell hyperpro­ changes in these markers previously were identified with particular liferation, whereas TGF-fJ2 loss correlates with suprabasal hyper­ stages of skin tumor progression [58,59]. High-risk papillomas proliferation. In tumors, loss of TGF-p is controlled at the post­ demonstrated basal and suprabasal expression of the a6p4 integrin, transcriptional level [17,64]. These results show that TGF-p ex­ loss of keratin 1, and aberrant expression of keratin 13 when they pression and function are compartmentalized in epidermis and firsterupted at 8 weeks after promotion started (Fig 3a,b.c). In these epidermal tumors, and that loss ofTGF-p is an early, biologically tumors, a6p4 integrin expression coincided with an expansion of relevant risk factor for malignant progression. the proliferating compartment as indicated by suprabasal bromo­ The importance ofTGF-p loss in progression was demonstrated deoxyuridine labeling [60]. In contrast, a6p4 immunostaining was directly using transgenic mice with null mutations introduced into confined to basal cells in low risk tumors, keratin 1 was abundant, the TGF-p110cus [65]. Keratinocyte cultures were established from and keratin 13 was absent or focal in the majority of this group, epidermisof newbornmice either homozygous or heterozygous for whereas proliferating cells were largely in the basal compartment the TGF-p1 null mutation or from wildtype littermates. The v­ H. (Fig 3d,e!). By 33 weeks after promotion started, a6p4 suprabasal ras oncogene was introduced by a defective retrovirus into cells of expression continued to distinguish papillomas at higher risk for each genotype, and recipient cells were grafted onto nude mouse malignant conversion, but keratin 13 was expressed in all groups hosts. As expected, the v-rasH'-transduced wild type keratinocytes (Fig 4). At this time, high risk papillomas displayed focal expres­ produced benign papillomas in grafts. In contrast, the tumors that sion of keratin 8 [46,61,62] and y-glutamyltranspeptidase, markers developed from v-rasH'-transduced TGF-p1 null keratinocytes were also found in chemically-induced carcinomas but not low-risk either very dysplastic papillomas or squamous cell carcinomas. tumors (Fig 4). Keratin 13 was lost in most carcinomas. Keratin 8 These results provide strong genetic proof to indicatethat the loss of and y-glutamyltranspeptidase were only expressed in a6p4 positive TGF-p1 is a critical step in the progression to malignancy in the . cells [60]. These results indicate that expression of a6p4 integrinin e�idermis. suprabasal strata serves as an early predictive marker to identify THE P53 AND benign squamous tumors at high risk for malignant progression and MALIGNANT CONVERSION that this marker is associated with a higher proliferation rate and suprabasal proliferation in high-risk tumors. The p53 gene is frequently mutated in human squamous cell and basal cell carcinomas of the skin [66-69]. This is likely to result GROWTH FACTORS AND THE RISK FOR from the direct mutational action of ultraviolet light since the most MALIGNANT CONVERSION frequent mutagenic change is a C -+ T transition commonly found The transforming growth factor-p's are expressed in the epidermis at C-C dinucleotides [66-69]. The E6 transforming gene of onco­ and are growth inhibitors for both normal and initiated mouse kera­ genic human papilloma virusestargets p53 protein to fonna rapidly tinocytes in vitro [63]; altered TGF-p expression could influencethe degraded complex. This action of E6 is essential for transforming growth properties of high risk papillomas. Normal epidermis, and activity of this virus for cultured keratinocytes [70-73]. Together skin papillomas at low risk for malignant conversion express TGF­ these results suggest an important role for p53 inepiderm al carcino­ P1 in the basal cell compartment and TGF-fJ2 in the suprabasal genesis. In experimental mouse skin carcinogenesis, p53 mutations VOL. 103, NO.5, SUPPLEMENT, NOVEMBER 1994 ONCOGENES AND TUMOR SUPPRESSORS IN CARCINOGENESIS 93S

,;-' .� . ., "

Figure S. Tumors induced on nude mice from grafts of keratino­ cytes expressing the v-rasHo oncogene. Skin keratinocytes were cultured from p53 null transgenic mice (e) or their wildtype (a) or heterozygous (b) Figure 4. Low-risk tumors are phenotypically distinct from carci­ littermates. In culture, cells of each genotype were transduced with the nomas after 33 weeks of tumor promotion. Tumors were induced by v-rasH> oncogene by a defective retroviral vector and removed fromculture the low-risk protocol described in the legend to Fig 3. After 33 weeks after 3 d for skin grafting as described [20]. Within 3-5 weeks, tumors papillomas were excised for immunohistochemical staining of frozen sec­ developed on recipient mice that had either the benign (a,b) or malignant (e) tions (a,b,c). Simultaneous staining was erformed on the few carcinomas p phenotype. See text for quantitative evaluation. Bar, 1 cm. that also develop fromthis protocol (d,eJ). The specificantibodies used were to recognize (a,d) keratin 13; (b,e) a6 integrin subunit; (en keratin 8. Re­ agents and procedures have been reported previously [60]. Bar, 100 Jlm.

tigation. In previous studies we have shown that a v-fos oncogene can cause malignant conversion of benign keratinocytes expressing are extremely rare in papillomas, but mutations are detected in car­ a ras oncogene [87,88]. Because thefos gene product is a transcrip­ cinomas [74-77]. This suggests that loss of p53 function may be tion factor known to regulate the expression of a number of genes relevant to malignant conversion [78]. To test this question, kerati­ that containAP-1 consensus sequences [89,90], the converting ac­ nocytes from skin of newborn p53 null transgenic mice [79] were tion of v-fos is likely to result from altered expression of normal H> cultured and transduced with the v-ras gene in experiments simi­ cellular genes. We are currently attempting to identify those cellu­ lar to those performed on TGF-fJ1 null mice. In all groups large lar genes that may be contributing to the malignant phenotype of tumors were apparent two weeks after grafting. Mice grafted with keratinocytes. p53(+j+) keratinocytes developed papillomas (eight of eight), whereas all mice grafted with p53(-j-) keratinocytes developed REFERENCES squamouscell carcinomas (eight of eight). p53(+j-) cells produced 1. Yuspa SH , D)' ugosz AA:C utaneouscarcinogenesis: natural and experimental.In: ra idly growing 10 of 19 of these papillo p papillomas; mas progressed Goldsmith L (ed.). PhYSiology, Biochonistry andMolecular Biology ofThe Skin. to carcinomas within five weeks (Fig 5). Thus, loss of the p53 Oxford University Press, New York, 1991, pp 1365-1402 H> protein in combination with a mutant ras gene is associated with 2. Yuspa SH, Morgan DL: Mouse skin cells resistant to terminal differentiation with i i car is. 293:72-74,1981 the malignant conversion stage of epidermal carcinogenesis. associated n tiation of cinogenes Nature 3. Kulesz-Martin MF, Koehler B, Hennings H, Yuspa SH: Quantitative assay for carcinogen altered differentiation in mouse epidermal cells. Carcinogenesis CONCLUSIONS 1:995-1006, 1980 4. Kilkenny AB,Morgan D, SpanglerEF, Yuspa SH: Correlation of initiating po­ The use of molecular approaches to analyze established biologic tencyof skincarcinogens with potency to induce resistance to terminal differ­ conceptshas provided insights into functionalc hanges in cell physi­ entiation incultured mouse keratinocytes. Cancer Res 45:2219-2225, 1985 ology during carcinogenesis. We now recognize that the initiated 5. YuspaSH , MorganD, LichtiU, Spangler EF, MichaelD,Kilkenn y A, Hennings H: Cultivationand characterization of cells derived fromm ouse skinpapillo­ phenotype results from intrinsic changes in intracellular signaling mas induced by an initiation-promotion protocol. Carcinogenesis 7:949 -958, pathways, particularly those related to terminal differentiation. 1986 When activation of the c-rasH> gene is responsible for initiation of 6. Strickland]E, GreenhalghDA, Koceva-Chyla A, Hennings H, Restrepo C, Ba­ keratinocyte neoplasia, a critical target is the PKC pathway. The laschak M, Yuspa SH: Development of murine epidermal cell lines which contain an activated rasH. oncogene and form papillomas in skin grafts on five PKC isoforms in keratinocytes contribute to specific compo­ athymic nude mouse hosts. Cancer Res 48:165-169,1988 nents of the differentiation response. Presently,the analysis suggests 7. Hennings H, Michael D, Lichti U, Yuspa SH: Response of carcinogen-altered that chronic activation of PKCa through increased cellular diacyl­ mOuse epidermal cells to phorbol ester tumor promoters and calcium.]Invest glycerol and functionalinh ibition ofPKCa due to phosphorylation Derm4toI88:60-65, 1987 8. Kulesz-MartinM, Kilkenn AE, Holbrook KA. Digernes V, Yuspa SH: Proper­ on tyrosine residue(s) contribute to the altered differentiation re­ y ties of carcinogen altered mouse epidermal cells resistant to calcium-induced sponse characteristic of initiated epidennal cells. terminal diKerentiation. CarcilUlgmesis 4:1367-1377,1983 Extrinsic tumor promotion and premalignant progression are as­ 9. Yuspa SH, Ben T, Hennings H. LichtiU: Phorbol ester rumorpromoters induce sociated withchanges in extrinsic signaling pathways. The control epidermal tTansglutaminase activity. Biochon Bwphys Res Co",,,,"n 97:700- 708, 1980 of proliferation of neo lastic epidermal cells seems central to these 10. Yuspa SH, BenT, Hennings H, Lichti U: Divergent responses inepidermal basal stages. Selective clona rexpansion of the neoplastic population is a cells exposed to the tumor promoter 12-0-tetradecanoylphorbol-13-acetate. requirement. The accumulation of genetic and chromosomal CancerRes 42:2344-2349, 1982 changes associated with rogression may be a consequence of al­ 11. D)' ugosz AA,Yuspa SH: Coordinatechanges in gene expression which mark the spinous to granular cell tTansition inepidermis are regulated by rot in tered controf . We now know that loss of expres ion or p e kinase proliferation s C.] Cell BioI 120:217-225, 1993 a localization of the TGF-fJ class of growth factors is common 12. YuspaSH, Ben T, HenningsH: The inductionof epidermaltransglutaminase and pathway through which premalignant progression proceeds in skin terminal differentiationby tumorpromoters in cultured epi dermal cells. Carci­ carcinogenesis. Loss of response by neoplastic cells to negative nogenesis 4:1413-1418, 1983 13. HenningsH, Robinson VA, Michael DM, Pernt GR,Jung R, YuspaSH : Devel­ growth factors such as TGFfJ be an alternative pathway to opment of an in vitro analogue of initiated mouse epidermis to study tumor achieve the same endpoint [80-84mar. Loss ofTGF-fJresponsiveness promotersand antipromoters. CancerRes 50:4794-4800.1990 has been associated with mutation and functional inactivation in 14. Hennings H, LowryDT. Robinson VA. Morgan DL, Fujiki H. YuspaSH: Activ­ p53 [85,86] and could account for the accelerated skin tumor pro­ ity of diverse tumorpromoters in a keratinocyteco-culture model of initiated epidermis. Carcinogenesis 13:2145-2151,1992 gression observed for both p53 and TGF-fJ null mice transduced 15. S trickland JE. Ueda M, Hennings H, Yuspa SH: A model for initiatedm ouse skin: with the v-rasH> gene. suppression of papilloma but not carcinoma formation by normal epidermal The molecular,bas is for malignant conversion is still under inves- cells in grafts on athymic nude mice. Cancer Res 52:1439-1444, 1992 94S YUSPA ET AL THEJOURNAL OF INVESTIGATIVE DERMATOLOGY

16. Roop DR, KriegTM, Mehrel T, Cheng CK, Yuspa SH: Transcriptional control 44. Daya-Grosjean L, Robert C, Drougard C, Suarez H, Sarasin A: High mutation of high molecular weight keratin gene expression in multistage mouse skin frequency in ras genes of skin tumors isolated from DNA repair deficient carcinogenesis. Cancer Res48:32 45-3252, 1988 xeroderma pigmentosum patients. Cancer Res53:1 625 - 1629, 1993 17. GlickAB, KulkarniAB, T ennenbaum T, Hennings H, Flanders KC, O'Reilly M, 45. Yuspa SH, Kilkenny AE, Stauley J, Lichti U: Keratinocytesb locked inp horbol SpornMB, Karlsson S, Yuspa SH: Lossof expression of transforming growth ester-responsive early stage of terminal differentiation by sarcomaviruses. Na­ factorP in skin and skin tumors is associated with hyperproliferation and a high ture 314:459-462, 1985 riskfor malignant conversion. Pro, Natl Acod Sci USA 90:6076-6080, 1993 46. Cheng C, Kilkenny AE. Roop D, Yuspa SH: The v-ras oncogene inhibits the 18. Aldaz CM, Conti CJ, Yuspa SH, Slaga TJ: Cytogenetic profile of mouse skin expression of differentiation markersand fa cilitates expression of cytokeratius8 tumors induced by the viral Harvey-ras gene. Carcinogenesis 9:1503-1505, and 18 inmouse keratinocytes. Mol Carcinog 3:363-373, 1990 1988 47. Roop DR, Huitfeldt H, Kilkenny A, Yuspa SH: Regulated expression of dilfe ren­ 19. Aldaz CM, Conti CJ,Larcher F, Trono D, Roop DR, ChesnerJ, Whitehead T, tiation-associated keratinsin cul tured epidermal cells detected by monospecific Slaga TJ : Sequential development of aneuploidy, keratin modification and antibodies to unique peptides of mouse epidermal keratins. D!fferentiation garnma-glutamyltransferase expression in mouse skin papillomas. CancerRes 35:143-150, 1987 48:3253-3257, 1988 48. Yuspa SH, Kilkenny AE, Steinert PM, Roop DR: Expression ofmurine epidermal 20. Roop DR, Lowy DR, Tambourin PE, Strickland J, Harper JR, Balaschak M, dilferentiation markers is tighdy regulated by restricted extracellular calcium Spangler EF, Yuspa SH: An activated Harvey rasoncogene produces benign concentrations invitro.J Cell Bioi 109:1207 -1217, 1989 tumourson mouse epidermal tissue.Nature 323:822-824, 1986 49. Lichti U, Yus pa SH: Modulation of tissue and epidermal transglutaminases in 21. Greenhalgh DA, Rothnagel JA, Quintanilla MI,Ore ngo CC, Gagne TA, Bund­ mouse epidermal cells aftertreatment with 12-0-tetradecanoylphorbol-13-ac­ man DS, LongleyMA, Roop DR: Induction of epidermal hyperplasia,hyper­ etate and/or retinoic acid in vivo and in culture. Cancer Res 48:74-81, 1988 keratosis. andpapillomas intransgenic mice by a targeted v-Ha-rasoncogene. 50. Tofrgard R, Yuspa SH, Roop DR: Keratin gene expression in mouse skin tumors Mol Carcinog 7:99-110, 1993 . and in mouse skin treated with 12-0-tetradecanoylphorbol-13-acetate. Cancer 22. Bailleul B, SuraniMA. White S, BartonSC, Brown K, Blessing M. Jorcano J, Res 45:5845 -5850, 1985 Balmain A: Skin hyperkeratosis and papilloma formation in transgenic mice 51. Lee E, Punnonen K, Cheng C, Glick A, D'ugosz A, Yuspa SH: Analysis of expressing a ras oncogene from a suprabasal keratin promoter. Cell 62:697- phospholipid metabolism in murine keratinocytes transformed by the v-ras 708, 1990 oncogene: relationship of phosphatidylinositol turnover and cytokinestimula­ 23. Hennings H. ShoresR, Wenk ML, SpanglerEF, Tarone R, Yuspa SH: Malignant tion to the transformed phenotype. Carcinogenesis 13:2367 -2373, 1992 conversionof mouse skin tumorsis increased b y tumor initiators and unaffected 52. Lee E. Yuspa SH: Changes in inositol phosphate metabolism are associatedwith by tumorpromoters. Nature 304:67 -69, 1983 terminalerentiation dilf and neoplasia in mouse keratinocytes. Carcinogenesis 24. Hennings H, Shores RA, Poirier MC, Reed E, TaroneRE. Yuspa SH: Enhanced 12:1651-1658, 1991 malignant conversion of benign mouse skin tumors by cisplatin.J Natl Cancer 53. Blumberg PM, Pettit GR. Warren BS, Szallasi A, Schuman LD, Sharkey NA, Irut 82:836-840, 1990 NakakumaH, Dell'Aquila ML, deVries DJ: The protein kinase C pathway in 25. Aldaz CM. Conti CJ, Klein-Szanto AJ P, Slaga TJ: Progressive dysplasia and tumor promotion. Prog Clin Bioi res 298:201-212, 1989 aneuploidy are hallmarks of mouseskin papillomas:relevance to malignancy. 54. Blumberg PM: Complexities of the protein kinase C pathway. Mol Carcinog Proc Natl Acad Sci USA 84:2029- 2032, 1987 4:339-344, 1991 26. AldazC M, Trono D, Larcher F, Slaga TJ, ContiCJ: Sequential trisomization of 55. Stabel S, Parker PJ: Protein kinase C. Pharmacol Ther 51:71-95, 1991 chromosomes6 and7 in mouse skin premalignant lesions. Mol Carcinog 2:22- 56. Dlu gosz AA,Mischak H, Mushinski JF, Yuspa SH: Transcripts encoding protein 26, 1989 kinase C alpha. delta, epsilon, zeta, and eta are expressed in basal and dilferen­ 27. Bianchi AB,Aldaz CM, Conti CJ:Non-random duplication of the chromosome tiating mouse keratinocytes in vitro and exhibit quantitative changes in neo­ bearing a mutated Ha-ras-1 allele in mouse skin tumors. Proc Natl A,ad Sci USA plastic cells. Mol Carcinog 5:286-292, 1992 87:6902-6906, 1990 57. Denning MF,D 'ugosz AA,Howett MK, Yuspa SH: Expressionof an oncogenic 28. Conti CJ,Aldaz C M, O'Connell J. Klein-Szanto AJP, Slaga TJ: Aneuploidy, an raslla gene in murine keratinocytes induces tyrosine phosphorylation and re­ early event in mouse skin tumor development. Carcinogenesis 7:1845-1848, duced activity of protein kinase C 6.J Bioi Chem 268:26079-26081, 1993 1986 58. Nischt R, Roop DR, Mehrel T, Yuspa SH, Rentrop M,W inter H, Schweizer J: 29. Hennings H, Robinson VA, Yuspa SH. Welty DJ, GreenhalghDG: Activation of Aberrantexpression during two-stagemouse skin carcinogenesis of type I 47- the rasllagene in urethane-initiated papillomas induced by protocols with high kDa Keratin,K 13. normally associatedwith terminaldifferentiation of internal and low frequencies of malignant conversion. Cancer Commun 3:277 -281. stratifiede pithelia. Mol Carcinog 1:96- 108, 1988 1991 59. Tennenbaum T, Yuspa SH, Grover A, Castronovo V, Sobel ME, Yamada Y, De 30. Hennings H, Shores R, Balaschak M. Yuspa SH: Sensitivity ofsub populations of Luca LM: Extracellularmatrix receptors and mouse skin carcinogenesis: altered mouse skin papillomas to malignant conversion by urethane or 4-nitroquino­ expression linked to appearance of early markers of tumorprogression. Cancer line N-oxide. CancerRes 50:653-657, 1990 Res 52:2966-2976, 1992 31. Hennings H, Shores R, Mitehell P, Spangler EF,Y uspa SH: Induction of papillo­ 60. Tennenbaum T, Weiner AK,Belanger AJ, Glick AB, HenningsH, Yuspa SH: mas with a high probability of conversion to malignancy. Carcinogenesis The suprabasal expression of o.6P4 integtin is associated with a high riskfor 6:1607-1610, 1985 malignant progression in mouse skin carcinogenesis. Cancer Res 53:4803- 32. Balmain A, RamsdenM, Bowden GT, SmithJ: Activation of the mouse cellular 4810, 1993 Harvey-ras gene in chemically induced benign skin papillomas. Nature 61. Larcher F, Bauluz C, Diaz-Guerra M. Quintanilla M, Conti CJ, Ballestin C. 307:658-660, 1984 Jorcano JL: Aberrant expression of the simple epithelial typeII keratin 8 by 33. Bizub D, Wood AW, SkalkaAM:Mutagenesis of the Ha-ras oncogene in mouse mouse skin carcinomas but not papillomas. Mol Carcinog 6:112-121, 1992 skin tumors induced by polycyclic aromatichydrocarbons. Proc Natl Acod Sci 62. Diaz-GuerraM, Haddow S, Bauluz C,JorcanoJL, Cano A, Balmain A, Quintan­ USA 83:6048 -6052, 1986 illa M:Ex pression of simple epithelial cytokeratius in mouse epidermal kerati­ 34. Balmain A, Pragneil m: Mouse skin carcinomas induced in vivo by chemical nocytes harboring Harvey ras gene alterations. Cancer Res 52:680 - 687, 1992 carcinogenshave a transforming harveyras oncogene. Nature 303:72 -74, 1983 63. Glick AB, Sporn MB, Yuspa SH: Altered regulationof TGF-p 1 and TGF-o.in 35. Brown K,Buchmann A, BalmainA: Carcinogen-induced mutations in the mouse primary keratinocytes and papillomas expressing v-Ha-ras. Mol Carcinog c-Ha-ras gene provide evidence of multiple pathways for tumor progression. 4:210-219, 1991 Proc Natl Acod Sci USA 87:538-542, 1990 64. Fowlis DJ, Flanders KC, Dullie E, Balmain A, AlthurstRJ: Discordant trans­ 36. QuintanillaM, Brown K, Ramsden M, Balmain A: Carcinogen-specificmu tation forming growth factorP1 RNA and protein localizationduring chemicalc arci­ and amplification ofHa-rasduring mouse skincarcinogenesis. Nature 322:78- nogenesis of the skin. Cell Growth Differ 3:81 -91, 1992 80, 1986 65. Kulkarni AB.Huh GG, Becker D, Geiser A, Lyght M, Flanders KC, RobertsAB. 37. Harper JR, Reynolds SH, Greenhalgh DA, Strickiand JE, Lacal JC, Yuspa SH: SpornMB, Ward JM. Karlsson S: Transforming growth bctor-p1 null muta­ Analysis of the rasH oncogene and itsp21 product in chemicallyinduced skin tion inmice c auses excessive inflammatory response and early death. Pro,Nail tumors and tumor-derived cell lines. Carcinogenesis 8: 1821- 1825, 1987 Acad Sci USA 90:770-774, 1992 38. Brown K, QuintanillaM, RamsdenM. Kerrm , Young S. Balmain A: v-ras genes BrashDE, RudolphJA, SimonJA, LinA, McKenua GJ ,Baden HP, Halperin AJ, fromHarvey and BALB murine sarcoma virusescan act asinitiators of two stage Ponten J: A role for sunlight in skin cancer: UV-induced p53 mutations in mouse skin carcinogenesis. Cell 46:447 -456, 1986 squamous cell carcinoma. Pro, Natl Acad Sci USA 88:10124-10128, 1991 39. LeonJ, Kamino H, Steinberg JJ, Pellicer A: H-ras activation in benign and 67. Rady p. Scinicariello F, Wagner RFJr, TyringSK: p53 mutations in basal cell self-regressing skin tumors (keratoacanthomas) in both humansand an animal carcinomas. Cancer Res 52:3804-3806, 1992 model system. Mol Cell Bioi 8:786-793, 1988 68. Ziegler A, Keffell DJ, Kunaia S. SharmaHW.Gailani M. SimonJA, Halperin AJ, 40. Corominas M. Karnino H, Leon J, Pellicer A: Oncogene activation in human Baden HP, Shapiro PE, Bale AE:Mutation hotspots due to sunlight in thep53 benign tumorsof the skin(keratoacanthomas): is HRAS involved in dilferen­ gene of nonmelanoma skin cancers. Proc Natl Acad Sci USA 90:4216-4220, tiation as well as proliferation? Proc Natl Acod Sci USA 86:6372-6376, 1989 1993 41. Ananthaswamy HN, Price JE, ldGo berg LH, BalesES: Detection andidentifica­ 69. Kanjilal S, Pierceall WE, Cummings K, Kripke ML, Ananthaswamy HN:Hig h tion of activated oncogenes in human skin cancers occurring onsun-exposed frequency of p53 mutations in ultraviolet radiation-induced murine skin body sites. CancerRes 48:3341-3346. 1988 tumors:evidence for strand bias and tumor heterogeneity. Cancer Res53:2961 - 42. Ananthaswamy HN, Applegate LA, Goldberg LH, Bales ES: Deletion of the 2964, 1993 c-Ha-ras-1 allele in human skin cancers. Mol Carcinog 2:298-301, 1989 70. WernessBA, LevineAJ, Howley PM: Association of humanpapillomavirus types 43. Suarez HG, Daya-Grosjean L, Schlaifer D, Nardeux P, Renault H, BosjL, Sarasin 16 and 18 E6 proteins with p53. Science 248:76-79, 1990 A: Activated oncogenes in human skin tumors from a repair-deficient syn­ 71. Schdfner M. Werness BA, Huibregtse JM, Levine AJ, Howley PM: The B6 drome, xeroderma pigmentosum. CancerRes 49:1223-1228. 1989 oncoprotein encoded by human papillomavirus types 16 and 18 promotes the degradation of p53. Cell 63:1129-1136, 1990 VOL. 103 , NO. 5, SUPPLBMBNT, NOVEMBER 1994 ONCOGENES AND TUMOR SUPPRESSORS IN CARCINOGENESIS 958

72. Kessis TD, Slebos RJ , Nelson WG, KastanMB , PlunkettBS, Han SM, Lorincz 81. Missero C, Filvaro/f E, Dotto GP: Induction of transforming growth factor P, AT, Hedrick L, Coo KR:Human papillomavirus 16 E6 exprcssiondismpts the resistance bythe E1A oncogene requires binding to a specific set of cellular p53-mediated cellular response to DNA damage. p"" NIIll ACtUl Sci USA proteins.Proc Na tl AuulSci USA 88:3489-3493, 1991 90:3988-3992, 1993 82. Krieg P, Schnapke R, Furstenbergcr G, Vogt I, MarksF: TGF-P-1 and skin 73. Woodworth CD, Wang H, Simpson S, Alvarez-Salas LM,N otarioV: Overex­ carcinogenesis: antiproliferative efFect inTG virroand F-{Il mRNA expression pression of wild-type p53 alters growth and dilferentiation of normal human during epidermal hyperproliferation and multistage tumorigenesis. Mol Car­ keratinocytes but not human papillomavirus-expressing cell lincs. Cell Growth cinog 4:129 -137, 1991 D!ffer 4:367 -376, 1993 83. Missero C, Ramon C, Dotto GP: Escape from transforming growth fac tor P 74. Bums PA, Kemp C}, Gannon }V, Lane DP, Bremner R, Balmain A: Loss of conrrol andoncogene cooperationin skintornor development. p"" NIIllACtUl heterozygosity and mutational alterations of thep53 gene in skintumours of Sd USA 88:9613-9617, 1991 interspecific hybrid mice. OnrogetU!6:23 63-2369, 1991 84. Reiss M, SartomliAC: Regulationof growth and differentiationofhuman kera­ 75. Ruggcri B, Caamano J, GoocIrow T, DiRado M, Bianchi A, Trcmo D, Conti C}, tinocytesby Type B transfo rming growthfactor and epidermalgrowth factor. Klein-Szanto AJ: Alterations of the p53 tumor suppressor gene during mouse Cancer Res 47:6705 -6709, 1987 skin tumor progression. Cancer Res 51:6615-6621, 1991 85. GerwinBI, Spil1areE, ForresterK, LehmanTA, KispertJ, WelshJA, Pfeifer AM, 76. Kress S, Sutter C, stricklandPT, M ukbtar H, Schweizer J, Schwarz M: Can;ino­ Lechner JF, Baker SJ. Vogelstein B: Mutant p53 can induce tumorigeniccon­ gen-specific mutational pattern in the p53 gene in ultraviolet B radiation-in­ version of humanbronchial epithelial cells and reduce their responsiveness to a duced squamous cell carcinomas of mouse skin. CtJnarRes 52:6400-6403, negative growth factor, transforming growth factorb eta,.Proc Natl ACtUl Sci 1992 USA 89:2759-2763, 1992 77. Ruggeri B, DiRado M, Zhang SY, Haner B, Goodrow T, Klein-SZanto AJ: 86. ReissM, vellucciVF, ZhOll ZL: Mutant p53 tornor suppressorcauses gene resist­ Benzo[a]pyrene-indua:d murine skintumon exhibit frequentand characteris­ ance to transforming growthfactor P, in murine keratinoeytcs. CtJnar .Ra tic G to T mutations in the p53 gene. ProcNatl Acad Sci USA 90:1013-1017, 53:899-904, 1993 1993 87. Greenhalgh DA, Yuspa SH: Malignant conversion of murine squamous papil­ 78. KempC], Donehower LA, Bradley A,B almaiD.A: Reduction of p53 gene dosage loma celllines by transfection with the fos oncogene. Mol Carclnog 1: 134 - 143, docsnot increaseinitiation or promotion but enhances malignant progression 1988 or chemically inducedskin rumors. CJl74:813-822, 1993 88. Greenhalgh DA, Welty DJ, Player A, Yuspa SH: Two oncogenes,v-fos andv-ras, 79. DonehowerM, LA,Harvey Slagle BL, McArthurMJ, Montgomery CAJr, Butel cooperate to convert normal keratinocytes to squamous cell carcinomas. Proc Js, BradleyA: Mice cldicicntp53 for are developmentally normalbut suscepti­ Natl Acad Sd USA87:643 -647, 1990 ble to spontaneous rumours. Nature 356:215-221, 1992 89. CurranT, FranzaBR:Fos andjun: The AP-1 connection. CtlI 55:395-397, 1988 80. Sipes NJ, Lyons RM,Moses HL: Isolation and characterizationof Kirsten murine 90. Rauscher FJ, Sambucerti LC, Curran T, Distel RJ , SpiegelmanBM: Common sarcoma virus-transformed mouse keratinocytes resistant to transforming DNA binding site forfos protein complexes and transcriptionfactor AP t. Cell growth factor beta. Mol Carcinog 3:12-19, 1990 52:471- 480, 1988