Ink4a/Arf Deficiency Promotes Ultraviolet Radiation-induced Melanomagenesis1

[CANCER RESEARCH 62, 6724 – 6730, November 15, 2002]

Ink4a/Arf Deficiency Promotes Ultraviolet Radiation-induced Melanomagenesis1 Juan A. Recio, Frances P. Noonan, Hisashi Takayama,2 Miriam R. Anver, Paul Duray, Walter L. Rush, Gerd Lindner, Edward C. De Fabo, Ronald A. DePinho, and Glenn Merlino3 Laboratory of Molecular Biology, National Cancer Institute, Bethesda, Maryland 20892-4264 [J. A. R., H. T., G. M.]; Laboratory of Photobiology and Photoimmunology, Departments of Immunology and Dermatology, George Washington University Medical School, Washington, DC 20037 [F. P. N., E. C. D. F.]; Pathology/Histotechnology Laboratory, Science Applications International Corporation, National Cancer Institute at Frederick, Maryland 21702-1201 [M. R. A.]; Laboratory of Pathology, National Cancer Institute, Bethesda, Maryland 20892 [P. D.]; Department of Dermatopathology, Armed Forces Institute of Pathology, Washington, DC 20306-6000 [W. L. R.]; Department of Dermatology, University Hospital Eppendorf, University of Hamburg, Hamburg, Germany D-20246 [G. L.]; and Department of Adult Oncology, Medicine and Genetics, DanaFarber Cancer Institute, Harvard Medical School, Boston, Massachusetts 02115 [R. A. D.]

ABSTRACT Cutaneous malignant melanoma (CMM), already known for its highly aggressive behavior and resistance to conventional therapy, has evolved into a health crisis by virtue of a dramatic elevation in incidence. The underlying genetic basis for CMM, as well as the fundamental role for UV radiation in its etiology, is now widely accepted. However, the only bona fide genetic locus to emerge from extensive analysis of CMM suppressor candidates is INK4a/ARF at 9p21, which is lost frequently in familial and occasionally in somatic CMM. The functional relationship between INK4a/ARF and UV radiation in the pathogenesis of CMM is largely unknown. Recently, we reported that hepatocyte growth factor/scatter factor (HGF/SF)-transgenic mice develop melanomas after a single erythemal dose of neonatal UV radiation, supporting epidemiological data implicating childhood sunburn in CMM. Here we show that neonatal UV irradiation induces a full spectrum of melanocyte pathology from early premalignant lesions through distant metastases. Cutaneous melanomas arise with histopathological and molecular pathogenetic features remarkably similar to CMM, including loss of ink4a/arf. A role for ink4a/arf in UV-induced melanomagenesis was directly assessed by placing the HGF/SF transgene on a genetic background devoid of ink4a/arf. Median time to melanoma development induced by UV radiation was only 50 days in HGF/SF ink4a/arfⴚ/ⴚ mice, compared with 152 and 238 days in HGF/SF ink4a/arfⴙ/ⴚ and HGF/SF ink4a/arfⴙ/ⴙ mice, respectively. These studies provide experimental evidence that ink4a/arf plays a critical role in UV-induced melanomagenesis and strongly suggest that sunburn is a highly significant risk factor, particularly in families harboring germ-line mutations in INK4a/ARF.

INTRODUCTION CMM4 is a highly aggressive, potentially fatal malignancy often resistant to currently available therapy. The recent alarming elevation in incidence of CMM (1, 2) has emphasized the need to understand the molecular mechanisms by which melanoma develops in humans and the risks that influence that process. It is well appreciated that such risks include both genetic and environmental factors (3, 4). Documentation of germ-line and somatic mutations in familial and sporadic CMM has unambiguously identified INK4a/ARF at 9p21 as a melanoma susceptibility locus, whereas a variety of other candidates remain to be discovered or confirmed (reviewed in Refs. 4 – 6). The INK4a/ARF locus consists of two overlapping tumor suppressor genes, p16INK4a and p14ARF (p19arf in mice), encoding two unreReceived 6/17/02; accepted 10/3/02. 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 NIH Award CA-92258 (to F. P. N.) and the National Cancer Institute under Contract NOI-CO-56000. 2 Present address: Department of Internal Medicine, Fujioka General Hospital, 942-1 Fujioka, Gunma 375-8503, Japan. 3 To whom requests for reprints should be addressed, at Laboratory of Molecular Biology, National Cancer Institute, Building 37, Room 5002, Bethesda, MD 20892-4264. Phone: (301) 496-4270; Fax: (301) 480-7618; E-mail: [email protected]. 4 The abbreviations used are: CMM, cutaneous malignant melanoma; HGF/SF, hepatocyte growth factor/scatter factor; RB, retinoblastoma; SED, standard erythemal dose; TRP, tyrosinase-related protein.

lated proteins in alternative reading frames (7). Acting through pRB and p53, respectively, these factors help regulate transit through the cell cycle, as well as cellular senescence and apoptosis (reviewed in Ref. 8). Data obtained from human tumors have implicated loss of p16INK4a, and with it pRB function, as the most significant mutational event at this locus in melanomagenesis (reviewed in Ref. 4). In sporadic CMM, where INK4a/ARF mutations are less frequent, functional loss of p16INK4a can also occur through epigenetic mechanisms such as promoter hypermethylation (9, 10). Germ-line mutations in cyclin-dependent kinase 4 that prevent binding to p16INK4a also abolish normal pRB-mediated cell cycle control, which could explain susceptibility to CMM in those families (11). Transgenic mouse models have been used to test these hypotheses. Interestingly, although ink4a/arf⫺/⫺ mice did not exhibit melanoma susceptibility (12), expression of activated H-ras on an Ink4a/Arf-deficient background produced a high incidence of melanoma with a relatively short latency (13). p16ink4a-specific gene targeting can also facilitate melanomagenesis, particularly after exposure to 7,12-dimethylbenz(a)anthracene (14, 15). Mice expressing the cyclin-dependent kinase 4 R24C allele identified in melanoma-prone kindreds are susceptible to spontaneous and 7,12-dimethylbenz(a)anthracene-initiated melanoma (16, 17). Receptor tyrosine kinases play an important role in melanocyte function and melanoma development, including the HGF/SF receptor c-Met (reviewed in Refs. 4, 18 –20). HGF/SF can serve as a multifunctional regulator of proliferation, motility, and morphogenesis in cells expressing c-MET (reviewed in Refs. 21, 22) and is required for the development of liver, placenta, and skeletal muscle (23–25). HGF/SF has been shown to stimulate growth and invasiveness and associated angiogenesis in tumor cells (reviewed in Refs. 26, 27). HGF/SF-Met autocrine loops have been detected in numerous human tumor types, including melanoma (21, 27, 28). The c-MET protooncogene resides at chromosome 7q33, a region amplified in human melanoma (29, 30), where it has been implicated in metastatic progression (31, 32). Transgenic mice broadly expressing HGF/SF develop a number of tumor types, demonstrating its potential as an oncogenic agent (33). At a mean onset age of 21 months, HGF/SFtransgenic mice develop melanoma, ⬃15% of which acquire a metastatic phenotype (33, 34). With respect to environmental factors, exposure to the UV spectrum of solar radiation is thought to be a causal agent in at least 80% of CMM (3, 35, 36). Retrospective epidemiological data currently suggest that unlike other skin cancers that are associated with cumulative lifetime UV exposure, CMM is provoked by intense intermittent exposure to UV, particularly during childhood (37, 38). However, the functional relationship between genes and the environment in melanoma pathogenesis is not well understood. These circumstances are fueled, at least in part, by the lack of a suitable genetically tractable, UV-dependent mouse melanoma model. Melanoma-prone transgenic mouse models have been previously described (reviewed in Ref. 39; see above), and some have demonstrated sensitivity to UV irradiation (40, 41); however, responses have been relatively inefficient. More-

6724

Downloaded from cancerres.aacrjournals.org on September 5, 2015. © 2002 American Association for Cancer Research.

INK4A/ARF DEFICIENCY PROMOTES UV-INDUCED MELANOMA

over, unlike human skin, melanocytes in murine skin are typically confined to hair follicles and resistant to melanoma induction by UV irradiation alone. An exception is the skin of albino mice overexpressing the HGF/SF transgene, which exhibits ectopic presence of melanocytes outside the confines of the hair follicle, in or near the epidermis. The development of this more humanized skin is likely the result of documented effects of HGF/SF on melanocytic cell survival and adhesion molecule expression, especially E-cadherin (28, 42, 43). Chronic UV irradiation using suberythemal doses in adult HGF/SF-transgenic mice induced nonmelanocytic cutaneous neoplasms but not melanoma (44). In contrast, we recently reported that a single dose of erythemal UV radiation in neonatal HGF/SF-transgenic mice was necessary and sufficient to induce melanoma at a relatively low latency and high cumulative incidence (45). These data provided the first experimental validation of epidemiological evidence that childhood sunburn poses a significant risk for CMM. Here we show that melanomas arising in these UV-irradiated, HGF/SF-transgenic mice are reminiscent of human CMM at the histopathological and molecular levels and are distinguishable from melanomas that develop in other murine model systems, which arise within the dermis and lack the epidermal component characteristic of conventional human CMM. Furthermore, we provide genetic evidence that disruption of ink4a/arf pathways promotes UV-induced melanomagenesis.

Therefore, the UV doses used in these experiments appear to be physiologically relevant. Histological and Molecular Analysis of Lesions. c-Met Western blotting, histopathology, and immunohistochemistry of mouse tissues were performed as described elsewhere (34). Metastases were grossly assessed at necropsy and confirmed by histopathology and immunohistochemistry. HGF/SF, c-Met, and Ki67 immunofluorescence studies were according to Lindner et al. (48). Analysis of c-Met phosphorylation by immunoprecipitation and Western blotting was according to Otsuka et al. (34). Antibodies: rabbit anti-c-Met (Santa Cruz Biotechnology); sheep anti-HGF/SF (48); rabbit anti-TRP1 and antiTRP2 (Ref. 34; generous gifts from Vince Hearing); rabbit anti-Ki67 (Dianova, Hamburg, Germany); mouse monoclonal anti-BrdUrd (Dako); mouse monoclonal antiphosphotyrosine (Upstate Biotechnology); mouse monoclonal anti-p16ink4a (F-12; Santa Cruz Biotechnology); sheep anti-p53 (Oncogene Research Products); and rabbit anti-S100 (Sigma). In some experiments, nuclei were visualized using Hoechst 33342 (Sigma). For c-met in situ hybridization, SP6 polymerase and plasmid pMMETS, carrying 4.15 kbp of mouse c-met cDNA (49), were used to generate a 33P-labeled RNA probe; a randomsequence riboprobe was used as control. PCR analysis of mouse exons 1 and 2 of p16ink4a was according to Zhuang et al. (50).

MATERIALS AND METHODS Transgenic Mice. Transgenic mice, in which expression of a murine HGF/SF cDNA was driven by the metallothionein promoter and locus control regions, were generated on the inbred albino FVB/NCr genetic background (hereafter referred to as FVB) as described previously (46). Ink4a/arf⫺/⫺ mice were originally made on a mixed genetic background as reported previously (12). In the UV-irradiation studies described here, the mutant ink4a/arf allele was placed on the FVB background through backcrossing for 10 generations. In earlier studies (unpublished observations), spontaneous tumorigenesis was observed in untreated HGF/SF-transgenic Ink4a/arf-deficient mice on a background consisting of about half FVB and half C57BL/6. All animal work was performed in accordance with guidelines established by the NIH. UV Irradiation. A bank of six Phillips F40 UV lamps was used. The Spectral Power Distribution and UV monitoring regimens have been described previously (47). Neonatal mice were irradiated in single wells of a 6-well Falcon plastic tissue culture plate (No. 3046; Becton-Dickinson, NJ) without the lid, placed 20 cm beneath the bank of sunlamps. The exposure time was 15 min for a total dose of 6.24 kJ/m2 UVB (280 –320 nm), 3.31 kJ/m2 UVA (320 – 400 nm), 0.03 kJ/m2 UVC (⬍280 nm), and 5.04 kJ/m2 of visible radiation (400 – 800 nm). Exposed mice exhibited skin reddening and occasional superficial desquamation; no UV-associated neonatal mortality was noted. In some experiments, young adult mice (6 weeks of age) were shaved and exposed to a second dose twice that delivered to neonates (45). Data for time to tumor development was analyzed by Kaplan-Meier survival analysis using Statview (SAS, Cary, NC). Calculations of the SED. To get a better quantitative measure of erythemogenic UV radiation, the Commission Internationale de l’Eclairage5 has recommended the use of the SED. The SED is calculated by multiplying the irradiance (W/unit area/nm) of a UV-emitting source (e.g., an artificial lamp or natural sunlight) by the human reference action spectrum for erythema. Integration of the area under the product curve gives the erythemal-weighted irradiance. Multiplying this weighted irradiance by the exposure time in seconds gives the erythemal-effective dose in Joules/unit area. The Commission Internationale de l’Eclairage has defined 1 SED as 100 J/m2 erythemalweighted irradiation. For comparative purposes, the number of SEDs given to neonatal mice in these experiments was calculated as 23. We determined previously that 23 SEDs could have been received in ⬃2 h and 40 min of sunlight exposure at northern mid-latitudes (Ref. 45, supplementary data). 5

Internet address: www.cie.co.at/cie/framepublications.html.

Fig. 1. Induction of cutaneous malignant melanoma in HGF/SF-transgenic mice as a consequence of neonatal erythemal UV irradiation. Shown is the gross appearance of early-stage “nevoid” lesions (a and b; arrows), a subset of which progressed to melanoma at sacrifice and frank melanomas (c and d; arrows). Effective regimens included erythemal UV irradiation at 3.5 days of age alone (b– d) and at both 3.5 days and 6 weeks of age (a). The UV doses used at 3.5 days and 6 weeks of age were 9.58 and 19.16 kJ/m2, respectively (45).

6725

Downloaded from cancerres.aacrjournals.org on September 5, 2015. © 2002 American Association for Cancer Research.

INK4A/ARF DEFICIENCY PROMOTES UV-INDUCED MELANOMA

RESULTS Exposure of neonatal HGF/SF-transgenic mice (3.5 days of age) to erythemal doses of UV radiation, with or without a subsequent erythemal dose at 6 weeks of age, was sufficient to induce the development of melanocytic tumors (45). Grossly, early lesions appeared on the albino FVB background as persistent, discolored spots between 2 and 3 months of age (Fig. 1, a and b), giving rise to frank amelanotic tumors with a median onset age of 238 days (Fig. 1, c and d). Upon histopathological examination, it was observed that unlike other animal models, most cutaneous melanomas (16 of 19) arising in HGF/SF mice as a consequence of neonatal UV irradiation possessed a junctional as well as a dermal component, demonstrating a remarkable similarity to lesions found in human patients at various stages of progression. The diagnosis of melanoma was confirmed through immunohistochemical assessment of a panel of melanocytic markers, including TRP1 and S100 protein (Fig. 2b, d, f, and h). Fig. 2 displays a representative human early melanoma in situ (Fig. 2a) and a more advanced radial growth phase or microinvasive malignant melanoma

exhibiting so-called “pagetoid” spreading (Fig. 2c). Melanocytic mouse lesions with strikingly similar features are shown in Fig. 2, b and d, respectively. Fig. 2 also displays more progressive human lesions, vertical growth phase or invasive malignant melanoma (Fig. 2e), and lymph node metastasis (Fig. 2g) and comparable lesions in the mouse (Fig. 2, f and h, respectively). One highly pleomorphic melanoma was identified in UV-irradiated HGF/SF transgenic mice (Fig. 3h), resembling the tumor type found rarely in patients. Not all mouse melanomas conformed to the classic human pattern of melanoma development. Examples were noted of nodular melanomas in vertical growth phase without a radial growth phase characterized by pagetoid spreading. The acquisition of receptor tyrosine kinase autocrine signaling loops, and consequent autonomous cellular growth control, occurs early in melanomagenesis and is considered a hallmark of human melanoma (4, 19, 51). Recently, Li et al. (28) reported that a majority of human melanoma cell lines produce both HGF/SF and c-MET. Within the skin of HGF/SF transgenic mice, ectopic HGF/SF expres-

Fig. 2. Histopathological comparison between human melanomas (left) and melanocytic lesions arising in UVirradiated HGF/SF-transgenic mice (right). a, human early malignant melanoma in situ in which solitary atypical melanocytes (arrow) are scattered into the upper reaches of the epidermis; H&E stain. b, lesion arising in HGF/SF-transgenic mouse with same features shown in (a; arrow); H&E stain. Inset, identification of scattered epidermal cells as melanocytes (brown TRP1 staining). c, two panels showing more advanced human microinvasive malignant melanoma exhibiting proliferation of atypical melanocytes disposed as solitary units (arrowhead) and variably sized nests (arrow), distributed along and above the dermal-epithelial junction; H&E stain. d, two lesions arising in HGF/SF-transgenic mice highly similar to (c), showing large nests of atypical melanocytes (arrow) and pagetoid spread melanoma cells even into the keratinized layer (arrowhead); H&E stain. Inset, identification of nested cells as melanocytes (brown TRP1 staining). e, representative human invasive malignant melanoma exhibiting broad growth, asymmetry, and irregular distribution of melanocytes along the junction; lower magnification, H&E stain. f, invasive lesion from transgenic mice comparable with (e), exhibiting broad growth, asymmetry, and chaotic melanocyte distribution in the epidermis; lower magnification, H&E stain. Inset displays S100 stain of same tumor. g, lymph node metastasis from melanoma patient; H&E stain. Inset, lower magnification of S100positive (brown) metastatic cells in subcapsular region. h, lymph node metastasis from UV-irradiated transgenic mouse; H&E stain. Inset, lower magnification of TRP1positive (brown) metastatic cells in subcapsular region.

6726

Downloaded from cancerres.aacrjournals.org on September 5, 2015. © 2002 American Association for Cancer Research.

INK4A/ARF DEFICIENCY PROMOTES UV-INDUCED MELANOMA

Fig. 3. UV-induced melanomagenesis in HGF/ SF-transgenic albino mice. a, double immunofluorescence of untreated adult transgenic truncal skin (E, epidermis; D, dermis); c-Met (red) and HGF/SF (green). Note solitary green HGF/SF-expressing melanocytes (white arrows). Inset, high magnification of double immunofluorescence identifying HGF/SF-expressing cells as melanocytes; HGF/SF (green), TRP1 (red), and nuclei (blue, Hoechst). b, bright field appearance of untreated transgenic adult skin showing melanocytes (brown stain) located in dermis (D), epidermis (E) and at dermalepidermal junction. c, in situ hybridization analysis of expression of c-met RNA transcripts in early lesion of UV-initiated transgenic skin. Note heavy silver grains over atypical melanocytes in dermis (D) and at junction; grains are infrequent in normal cells of dermis and epidermis (E). d, early melanocytic lesion (brown TRP1 staining) showing junctional components. e, Expression of p16ink4a (brown nuclei, arrowheads) in early melanocytic lesion. f, absence of p16ink4a expression in malignant melanoma; note positive staining in normal follicular cells (bottom of panel). g, visualization of melanocyte proliferation in an early lesion by Ki67 immunofluorescence (red); nuclei are Hoechst stained (blue). h, robust expression of p53 (brown nuclei) in a highly progressive, pleomorphic melanoma (shown as H&E stain at higher magnification in inset). Confirmation that this tumor was carrying an exon 8 mutation in p53 was by single-strand conformational polymorphism (data not shown).

sion was readily detectable in solitary melanocytes by immunofluorescence (Fig. 3a). Expression of the HGF/SF transgene resulted in the presence of melanocytes outside the hair follicle, in the dermis, epidermis, and dermal-epidermal junction (Fig. 3b). Early hyperplastic melanocytic lesions (Fig. 3, d and g) were found to strongly express c-met by in situ hybridization (Fig. 3c). The importance of the resulting intensified autocrine loops in melanomagenesis was substantiated by the frequent detection of c-Met autophosphorylation in melanomas arising in these UV-treated transgenic mice (Fig. 4a). Molecular analysis of frank tumors from UV-irradiated HGF/SFtransgenic mice demonstrated frequent homozygous loss of ink4a/arf sequences, including the second exon (Fig. 4b), resembling genetic lesions commonly found in human melanoma and creating a deficiency of both p16ink4a and p19arf. The absence of p16ink4a expression in these tumors was confirmed immunohistochemically (Fig. 3f and Fig. 4b). In contrast, p16ink4a expression could be unambiguously observed in several hyperplastic lesions (Fig. 2e), suggesting that ink4a/arf loss may not be required to initiate melanomagenesis in this UV-induced model. To confirm a role for ink4a/arf loss in UV-induced melanomagenesis, the HGF/SF transgene was placed onto an Ink4a/arf-deficient background. Ink4a/arf wild-type, heterozygote, and homozygote null HGF/SF-transgenic neonatal mice were UV-irradiated and monitored for the development of melanomas. Fig. 5a shows that the absence of one ink4a/arf allele significantly accelerated UV-induced melanomagenesis, with a median time of tumor onset of 152 days. This behavior was more pronounced in the ink4a/arf null cohort in which melanomas arose with a median onset of just 50 days (Fig. 5a). Melanomas originating in HGF/SF-transgenic, Ink4a/arf-deficient

mice (Fig. 5b) were highly malignant, demonstrating both locally invasive and distantly metastatic phenotypes (Fig. 5, c and d). Favored metastatic sites were lymph node and liver. Neonatal UV irradiation failed to induce melanoma in Ink4a/arf-deficient mice in the absence of the HGF/SF transgene (Fig. 5a). Notably, the specific combination of broad HGF/SF overexpression and Ink4a/arf deficiency also induced development of the highly malignant skeletal muscle tumor, rhabdomyosarcoma, to which all HGF/SF ink4a/arf⫺/⫺ mice succumbed by an average of ⬃100 days of age (52). In fact, untreated HGF/SF ink4a/arf⫺/⫺ mice did not survive long enough to assess melanoma development. However, dermal melanoma was the second most prevalent tumor in untreated HGF/SF ink4a/arf⫹/⫺ mice, arising with a median time of 308 days. Compared with UV-irradiated HGF/SF ink4a/arf⫹/⫺ mice (Fig. 5a), tumor-free survival for unirradiated HGF/SF ink4a/arf⫹/⫺ mice was significantly greater (P ⬍ 0.001) with time to 50% melanoma-free survival ⬎ 8 months longer, confirming a critical role for ink4a/arf in melanoma induction by UV irradiation. DISCUSSION UV exposure is thought to be a causal agent in at least 80% of CMM (3, 35, 36). In this report, we describe an animal model, the HGF/SF transgenic mouse, in which melanocytic tumors arise as a consequence of a single neonatal dose of erythemal UV radiation with histopathological and molecular pathogenetic profiles reminiscent of human CMM. This was achieved at a total UV dose that was 30-fold lower than the suberythemal intensities administered chronically in adult HGF/SF transgenics, which failed to induce melanoma (44),

6727

Downloaded from cancerres.aacrjournals.org on September 5, 2015. © 2002 American Association for Cancer Research.

INK4A/ARF DEFICIENCY PROMOTES UV-INDUCED MELANOMA

Fig. 4. Molecular and biochemical characterization of UV-induced melanomas. a, direct Western analysis of c-Met expression and autophosphorylation in six primary melanomas. c-Met was visualized by anti-c-Met antibody, revealing two bands representing unprocessed p170 and processed p140 c-Met (arrows). c-Met activity was assessed through analysis of phosphotyrosine residues using three different probes: an antibody specific for Y1365; an antibody specific for Y1230, Y1234, and Y1235; and a general antiphosphotyrosine antibody. It is assumed that some tumors exhibiting relatively low levels of c-Met protein, but high kinase activity, are constitutively activated through the creation of strong HGF/SF-Met autocrine loops and consequently down-regulated. b, PCR analysis of the structure of p16ink4a exons 1 and 2 in seven different primary melanomas arising in HGF/SF-transgenic mice, performed as described previously (50). Note homozygous loss of exon 2 sequences in three of seven tested tumors. Results of immunohistochemical (IHC) analysis of p16ink4a expression in these tumors is shown at the bottom (⫹, detectable expression; ⫺, no staining).

corroborating epidemiological evidence that childhood sunburn poses a highly significant risk for CMM (45). Although the reason for the preferential UV sensitivity exhibited by neonatal relative to adult skin is presently unclear, this mouse model provides a unique opportunity to experimentally address this question; the answer to which may be fundamental to our understanding of the human disease. Compared with adult skin, neonatal wild-type mouse skin has been reported to harbor a considerably higher proportion of melanocyte progenitors (53), a finding consistent with our own immunohistochemical analysis of melanocyte differentiation in HGF/ SF-transgenic mice (data not shown). Moreover, neonatal murine melanocytic cells demonstrate an enhanced proliferative capacity, realized as a consequence of UV irradiation or wounding (54, 55). We propose that melanocyte progenitors in neonatal skin constitute critical cellular targets for sunburning levels of UV radiation, which directly damages DNA while facilitating its fixation by encouraging cell cycle entry. An intriguing additional possibility is that early exposure to intense UV radiation irreversibly compromises the developing immune system, promoting future tolerance to subsequent melanocytic lesions. It has been well described that UV radiation is a modulator of immune function, and there is considerable evidence that UV immunosuppression plays a critical role in photocarcinogenesis (56). Many experimental studies of UV-induced nonmelanoma skin cancer have indicated a critical role for UV immunosuppression in the outgrowth of these cancers (reviewed in Ref. 57); however, comparable studies

have not been possible for melanoma because of the lack of a suitable mouse model for UV-induced melanoma. Point mutations with UV signatures in skin cancers have been reported previously, and the identity of relevant genetic targets of UV radiation is currently the focus of intense investigation. Some reports have implicated the RAS genes in human CMM (reviewed in Ref. 4); however, c-H-ras and c-N-ras mutations were not detected in the seven UV-induced melanomas examined (data not shown). Immunohistochemical analysis showed that of eight tumors tested, only the most pleomorphic melanoma was strongly positive for nuclear p53 (Fig. 3h), typically indicative of a stabilizing mutation. These data are consistent with reports of infrequent p53 mutagenesis in human CMM (reviewed in Ref. 4) and with the notion that p53 plays a relatively rare role in progression to a more genetically unstable phenotype. An obvious candidate mutational target is the INK4a/ARF locus, widely regarded as a tumor suppressor in human melanoma (reviewed in Refs. 4, 6, 8). Data reported here, generated from UV-irradiated HGF/SF-transgenic Ink4a/arf-deficient mice, indicate an important role for ink4a/arf in UV-induced melanomagenesis. The significant acceleration of melanoma development observed in UV-irradiated, Ink4a/arf-deficient, HGF/SF-transgenic mice is consistent with epidemiological data, suggesting that individuals living in areas of intense sunlight exposure and bearing germ-line INK4a/ARF mutations are at the highest risk for CMM (58) and additionally supports this transgenic mouse as a valid model for human CMM. What is the mechanistic relationship between ink4a/arf and UV radiation? Reports of UV signature mutations in melanoma and other skin tumors, but not in internal tumors, have suggested that the INK4a/ARF locus is a direct target of UV radiation (59, 60). However, ink4a/arf mutations in melanomas arising in HGF/SFtransgenic mice carrying wild-type ink4a/arf alleles and UV irradiated as neonates were characterized by DNA deletions, not typically associated with UV irradiation. Although these data are consistent with a later role for ink4a/arf in melanomagenesis, p16INK4a expression has been reported to be up-regulated epigenetically by UV exposure in human skin cultures and melanoma cell lines, causing a delay at the G2-M junction of the cell cycle (61). Early loss of p16ink4a would eliminate this block, facilitating fixation of UV-induced mutations. Determining the relationship between UV and ink4a/arf is additionally complicated by the fact that genetic mutations at this locus can affect p16ink4a, p19arf, or both. Although UV mutations affecting exon 1␤ of p14ARF alone are rare, reported p16INK4a exon 2 mutations with UV signatures can generate altered p14ARF as well as p16INK4a (60). Moreover, p14ARF has been reported to be inactivated as a consequence of a germ-line mutation (62). Loss of p53 function through mutation of p19arf could encourage melanocyte survival after significant UVinduced damage. A better understanding of the distinct roles of p16ink4a and p19arf in UV-induced melanomagenesis must await the outcome of UV irradiation of HGF/SF-transgenic mice in the context of p16ink4a-specific and p19arf-specific deficiencies. Here we show that melanocytic neoplasms arising in this experimental mouse model uniquely resemble human CMM with respect to etiology, histopathology and molecular derangements. The HGF/SF-transgenic mouse described here provides a badly needed platform for the rigorous assessment of genetic and environmental melanoma risk factors, acting both individually and in concert. It is anticipated that exploitation of this mouse model will facilitate derivation of both effective sun protection strategies and antimelanoma therapies.

6728

Downloaded from cancerres.aacrjournals.org on September 5, 2015. © 2002 American Association for Cancer Research.

INK4A/ARF DEFICIENCY PROMOTES UV-INDUCED MELANOMA

Fig. 5. Ink4a/Arf deficiency cooperates with neonatal UV irradiation in the promotion of melanoma in HGF/SF transgenic mice. a, Kaplan-Meier survival analysis of melanomagenesis in UV-irradiated HGF/SF transgenic (Tg) and nontransgenic (nTg) mice in the presence of two (⫹/⫹), one (⫹/⫺), or no (⫺/⫺) wild-type ink4a/arf alleles. All mice in this study were UV-irradiated once at 3.5 days of age (45). The following number of total mice and number of mice with melanomas (in parentheses) was used to generate these curves: Tg,⫹/⫹ 43 (10); Tg,⫹/⫺ 6 (4); Tg,⫺/⫺ 4 (3); nTg,⫹/⫺ 6 (0); and nTg,⫺/⫺ 3 (0). Some mice had more than one melanoma. Both the UV-irradiated Tg,⫹/⫺ and Tg,⫺/⫺ cohorts developed an equivalent number of rhabdomyosarcomas and a smaller number of squamous cell carcinomas as well. The curves for both the HGF/SF ink4a/arf⫺/⫺ and HGF/SF ink4a/arf⫹/⫺ mice were found to be significantly different from the HGF/SF ink4a/arf⫹/⫹ mice (P ⬍ 0.001 in both cases) and from each other (P ⬍ 0.01). Note that in the absence of the HGF/SF transgene, ink4a/arf mutant mice failed to develop melanoma as a consequence of UV irradiation. b, histopathological profile of highly invasive, nodular melanoma arising in a 50-day-old HGF/SF ink4a/arf⫺/⫺ mouse; H&E stain. c, vascular invasion of melanoma shown in (b). Brown represents TRP1 expression; arrows indicate location of endothelial cell layer; L, vessel lumen. d, distant metastasis to liver exhibited in a 194-day-old HGF/SF ink4a/arf⫹/⫺ mouse. Tumor is in brown (TRP2 immunohistochemical stain); P, liver parenchyma.

ACKNOWLEDGMENTS We thank Drs. Ruth Halaban, Dot Bennett, Dorothea Becker, Meenhard Herlyn, Peggy Tucker, Pamela Pollock, Marcus Bosenberg, Stuart Yuspa, Lalage Wakefield, Raymond Barnhill, Aizen Marrogi, Nick Hayward, Lynda Chin, Yanlin Yu, and Sandy Kazakis for useful and stimulating discussions. We also thank Drs. Vince Hearing for the TRP1 antibody, Carmen Birchmeier for mouse c-met cDNA, Aurora Medina and Lee Helman for p53 single-strand conformational polymorphism, William LaRochelle for mouse HGF/SF cDNA, Peter Soderkvist and Svetlana Bolshakov for mutant c-N-ras samples, Greg Buzard for mutant c-H-ras samples, and Cecil Fox for in situ hybridization. We thank Vicki Joe, Michael Amesquita, and Richard Sharp for DNA analysis and Heather Wimbrow, Cari Cherry, Dee Green and Barbara Kasprzak for technical assistance.

REFERENCES 1. Howe, H. L., Wingo, P. A., Thun, M. J., Ries, L. A., Rosenberg, H. M., Feigal, E. G., and Edwards, B. K. Annual report to the nation on the status of cancer (1973 Through 1998). Featuring cancers with recent increasing trends. J. Natl. Cancer Inst. (Bethesda), 93: 824 – 842, 2001. 2. Jemal, A., Devesa, S. S., Hartge, P., and Tucker, M. A. Recent trends in cutaneous melanoma incidence among whites in the United States. J. Natl. Cancer Inst. (Bethesda), 93: 678 – 683, 2001. 3. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Solar and ultraviolet radiation, Vol. 55. Lyon, France: IARC, 1992.

4. Chin, L., Merlino, G., and DePinho, R. A. Malignant melanoma: modern black plague and genetic black box. Genes Dev., 12: 3467–3481, 1998. 5. Herlyn, M. Molecular and cellular biology of melanoma, pp. 1–102, Austin, TX: R. G. Landes Co., 1993. 6. Pollock, P. M., and Trent, J. M. The genetics of cutaneous melanoma. Clin. Lab. Med., 20: 667– 690, 2000. 7. Quelle, D. E., Zindy, F., Ashmun, R. A., and Sherr, C. J. Alternative reading frames of the INK4a tumor suppressor gene encode two unrelated proteins capable of inducing cell cycle arrest. Cell, 83: 993–1000, 1995. 8. Chin, L., Pomerantz, J., and DePinho, R. A. The INK4a/ARF tumor suppressor: one gene-two products-two pathways. Trends Biochem. Sci., 23: 291–296, 1998. 9. Gonzalgo, M. L., Bender, C. M., You, E. H., Glendening, J. M., Flores, J. F., Walker, G. J., Hayward, N. K., Jones, P. A., and Fountain, J. W. Low frequency of p16/ CDKN2A methylation in sporadic melanoma: comparative approaches for methylation analysis of primary tumors. Cancer Res., 57: 5336 –5347, 1997. 10. Funk, J. O., Schiller, P. I., Barrett, M. T., Wong, D. J., Kind, P., and Sander, C. A. p16INK4a expression is frequently decreased and associated with 9p21 loss of heterozygosity in sporadic melanoma. J. Cutan. Pathol., 25: 291–296, 1998. 11. Zuo, L., Weger, J., Yang, B., Goldstein, A. M., Tucker, M. A., Walker, G. J., Hayward, N., and Dracopoli, N. C. Germline mutations in the p16INK4a binding domain of CDK4 in familial melanoma. Nat. Genet., 12: 97–99, 1996. 12. Serrano, M., Lee, H., Chin, L., Cordon-Cardo, C., Beach, D., and DePinho, R. A. Role of the INK4a locus in tumor suppression and cell mortality. Cell, 85: 27–37, 1996. 13. Chin, L., Pomerantz, J., Polsky, D., Jacobson, M., Cohen, C., Cordon-Cardo, C., Horner, J. W., and DePinho, R. A. Cooperative effects of INK4a and ras in melanoma susceptibility in vivo. Genes Dev., 11: 2822–2834, 1997. 14. Sharpless, N. E., Bardeesy, N., Lee, K. H., Carrasco, D., Castrillon, D. H., Aguirre, A. J., Wu, E. A., Horner, J. W., and DePinho, R. A. Loss of p16Ink4a with retention of p19Arf predisposes mice to tumorigenesis. Nature (Lond.), 413: 86 –91, 2001.

6729

Downloaded from cancerres.aacrjournals.org on September 5, 2015. © 2002 American Association for Cancer Research.

INK4A/ARF DEFICIENCY PROMOTES UV-INDUCED MELANOMA

15. Krimpenfort, P., Quon, K. C., Mooi, W. J., Loonstra, A., and Berns, A. Loss of p16Ink4a confers susceptibility to metastatic melanoma in mice. Nature (Lond.), 413: 83– 86, 2001. 16. Sotillo, R., Garcia, J. F., Ortega, S., Martin, J., Dubus, P., Barbacid, M., and Malumbres M. Invasive melanoma in Cdk4-targeted mice. Proc. Natl. Acad. Sci. USA, 98: 13312–13317, 2001. 17. Rane, S. G., Cosenza, S. C., Mettus, R. V., and Reddy, E. P. Germ line transmission of the Cdk4(R24C) mutation facilitates tumorigenesis and escape from cellular senescence. Mol. Cell. Biol., 22: 644 – 656, 2002. 18. Bennett, D. C. Genetics, development, and malignancy of melanocytes. Int. Rev. Cytol., 146: 191–260, 1993. 19. Shih, I. M., and Herlyn, M. Autocrine and paracrine roles for growth factors in melanoma. In Vivo, 8: 113–123, 1994. 20. Halaban, R. Growth factors and melanomas. Semin. Oncol., 23: 673– 681, 1996. 21. Jeffers, M., Rong, S., and Vande Woude, G. F. Hepatocyte growth factor/scatter factor-Met signaling in tumorigenicity and invasion/metastasis. J. Mol. Med., 74: 505–513, 1996. 22. Matsumoto, K., and Nakamura, T. Emerging multipotent aspects of hepatocyte growth factor. J. Biochem. (Tokyo), 119: 591– 600, 1996. 23. Schmidt, C., Bladt, F., Goedecke, S., Brinkman, V., Zschiesche, W., Sharpe, M., Gherardi, E., and Birchmeier, C. Scatter factor/hepatocyte growth factor is essential for liver development. Nature (Lond.), 373: 699 –702, 1995. 24. Uehara, Y., Minowa, O., Mori, C., Shiota, K., Kuno, J., Noda, T., and Kitamura, N. Placental defect and embryonic lethality in mice lacking hepatocyte growth factor/ scatter factor. Nature (Lond.), 373: 702–705, 1995. 25. Bladt, F., Rlethmacher, D., Isenmann, S., Aguzzi, A., and Birchmeier, C. Essential role for the c-met receptor in the migration of myogenic precursor cells into the limb bud. Nature (Lond.), 376: 768 –771, 1995. 26. Rosen, E. M., Lamszus, K., Laterra, J., Polverini, P. J., Rubin, J. S., and Goldberg, I. D. HGF/SF in angiogenesis. Ciba Found. Symp., 212: 215–226, 1997. 27. Comoglio, P. M., and Boccaccio, C. Scatter factors and invasive growth. Semin. Cancer Biol., 11: 153–165, 2001. 28. Li, G., Schaider, H., Satyamoorthy, K., Hanakawa, Y., Hashimoto, K., and Herlyn, M. Down-regulation of E-cadherin and desmoglein 1 by autocrine hepatocyte growth factor during melanoma development. Oncogene, 20: 8125– 8135, 2001. 29. Wiltshire, R. N., Duray, P., Bittner, M. L., Visakorpi, T., Meltzer, P. S., Tuthill, R. J., Liotta, L. A., and Trent, J. M. Direct visualization of the clonal progression of primary cutaneous melanoma: application of tissue microdissection and comparative genomic hybridization. Cancer Res., 55: 3954 –3957, 1995. 30. Bastian, B. C., LeBoit, P. E., Hamm, H., Brocker, E. B., and Pinkel, D. Chromosomal gains and losses in primary cutaneous melanomas detected by comparative genomic hybridization. Cancer Res., 58: 2170 –2175, 1998. 31. Natali, P. G., Nicotra, M. R., Di Renzo, M. F., Prat, M., Bigotti, A., Cavaliere, R., and Comoglio, P. M. Expression of the c-Met/HGF receptor in human melanocytic neoplasms: demonstration of the relationship to malignant melanoma tumour progression. Br. J. Cancer, 68: 746 –750, 1993. 32. Hendrix, M. J., Seftor, E. A., Seftor, R. E., Kirschmann, D. A., Gardner, L. M., Boldt, H. C., Meyer, M., Pe’er, J., and Folberg, R. Regulation of uveal melanoma interconverted phenotype by hepatocyte growth factor/scatter factor (HGF/SF). Am. J. Pathol., 152: 855– 863, 1998. 33. Takayama, H., LaRochelle, W. J., Sharp, R., Otsuka, T., Kriebel, P., Anver, M., Aaronson, S. A., and Merlino, G. Diverse tumorigenesis associated with aberrant development in mice overexpressing hepatocyte growth factor/scatter factor. Proc. Natl. Acad. Sci. USA, 94: 701–706, 1997. 34. Otsuka, T., Takayama, H., Sharp, R., Celli, G., LaRochelle, W. J., Bottaro, D. P., Ellmore, N., Vieira, W., Owens, J. W., Anver, M., and Merlino, G. c-Met autocrine activation induces development of malignant melanoma and acquisition of the metastatic phenotype. Cancer Res., 58: 5157–5167, 1998. 35. Armstrong, B. K., Kricker, A., and English, D. R. Sun exposure and skin cancer. Australas. J. Dermatol., 38 (Suppl.) 1: S1–S6, 1997. 36. MacKie, R. M. Incidence, risk factors and prevention of melanoma. Eur. J. Cancer, 34 (Suppl.) 3: S3–S6, 1998. 37. Holman, C. D. J., Armstrong, B. K., and Heenan, P. J. A theory of the etiology and pathogenesis of human cutaneous malignant melanoma. J. Natl. Cancer Inst. (Bethesda), 71: 651– 656, 1983. 38. Whiteman, D. C., Whiteman, C. A., and Green, A. C. Childhood sun exposure as a risk factor for melanoma: a systematic review of epidemiologic studies. Cancer Causes Control, 12: 69 – 82, 2001. 39. Tietze, M. K., and Chin, L. Murine models of malignant melanoma. Mol. Med. Today, 6: 408 – 410, 2000.

40. Kelsall, S. R., and Mintz, B. Metastatic cutaneous melanoma promoted by ultraviolet radiation in mice with transgene-initiated low melanoma susceptibility. Cancer Res., 58: 4061– 4065, 1998. 41. Broome Powell, M., Gause, P. R., Hyman, P., Gregus, J., Lluria-Prevatt, M., Nagle, R., and Bowden, G. T. Induction of melanoma in TPras transgenic mice. Carcinogenesis (Lond.), 20: 1747–1753, 1999. 42. Kos, L., Aronzon, A., Takayama, H., Maina, F., Ponzetto, C., Merlino, G., and Pavan, W. Hepatocyte growth factor/scatter factor-Met signaling in neural crest-derived melanocyte development. Pigm. Cell Res., 12: 13–21, 1999. 43. Kunisada, T., Yamazaki, H., Hirobe, T., Kamei, S., Omoteno, M., Tagaya, H., Hemmi, H., Koshimizu, U., Nakamura, T., and Hayashi, S-I. Keratinocyte expression of transgenic hepatocyte growth factor affects melanocyte development, leading to dermal melanocytosis. Mech. Dev., 94: 67–78, 2000. 44. Noonan, F. P., Otsuka, T., Bang, S., Anver, M. R., and Merlino, G. Accelerated ultraviolet radiation-induced carcinogenesis in hepatocyte growth factor/scatter factor transgenic mice. Cancer Res., 60: 3738 –3743, 2000. 45. Noonan, F. P., Recio, J. A., Takayama, H., Duray, P., Anver, M. R., Rush, W. L., De Fabo, E. C., and Merlino, G. Neonatal sunburn and melanoma in mice. Nature (Lond.), 413: 271–272, 2001. 46. Takayama, H., LaRochelle, W. J., Anver, M., Bockman, D. E., and Merlino, G. Scatter factor/hepatocyte growth factor as a regulator of skeletal muscle and neural crest development. Proc. Natl. Acad. Sci. USA, 93: 5866 –5871, 1996. 47. Noonan, F. P., and De Fabo, E. C. Ultraviolet-B dose response curves for local and systemic immunosuppression are identical. Photochem. Photobiol., 52: 801– 810, 1990. 48. Lindner, G., Menrad, A., Gherardi, E., Merlino, G., Welker, P., Handjiski, B., Roloff, B., and Paus, R. Involvement of hepatocyte growth factor/scatter factor and Met receptor signaling in hair follicle morphogenesis and cycling. FASEB J., 14: 319 – 332, 2000. 49. Iyer, A., Kmiecik, T. E., Park, M., Daar, I., Blair, D., Dunn, K. J., Sutrave, P., Ihle, J. N., Bodescot, M., and Vande Woude, G. F. Structure, tissue-specific expression, and transforming activity of the mouse met proto-oncogene. Cell Growth Diff., 1: 87–95, 1990. 50. Zhuang, S. M., Schippert, A., Haugen-Strano, A., Wiseman, R. W., and Soderkvist, P. Inactivations of p16INK4a-␣, p16INK4a-␤, and p15INK4b genes in 2⬘, 3⬘-dideoxycytidine- and 1,3-butadiene-induced murine lymphomas. Oncogene, 16: 803– 808, 1998. 51. Scott, G., Stoler, M., Sarkar, S., and Halaban, R. Localization of basic fibroblast growth factor mRNA in melanocytic lesions by in situ hybridization. J. Investig. Dermatol., 96: 318 –322, 1991. 52. Sharp, R., Recio, J. A., Jhappan, C., Otsuka, T., Liu, S., Yu, Y., Liu, W., Anver, M., Navid, F., Helman, L. J., DePinho, R. A., and Merlino, G. Synergism between INK4a/ARF inactivation and aberrant HGF/SF signaling in rhabdomyosarcomagenesis. Nature Medicine, 8: in press, 2002. 53. Hirobe, T. Histochemical survey of the distribution of the epidermal melanoblasts and melanocytes in the mouse during fetal and postnatal periods. Anat. Rec., 208: 589 –594, 1984. 54. Rosdahl, I. K., and Szabo, G. Mitotic activity of epidermal melanocytes in UVirradiated mouse skin. J. Investig. Dermatol., 70: 143–148, 1978. 55. Hirobe, T. Developmental changes of the proliferative response of mouse epidermal melanocytes to skin wounding. Development (Camb.), 102: 567–574, 1988. 56. Noonan, F. P., and De Fabo, E. C. Immunosuppression by ultraviolet B radiation: initiation by urocanic acid. Immunol. Today, 13: 250 –254, 1992. 57. Kripke, M. L. Effects of UV radiation on tumor immunity. J. Natl. Cancer Inst. (Bethesda), 82: 1392–1396, 1990. 58. Bishop, D. T., Demenais, F., Goldstein, A. M., Bergman, W., Newton Bishop, J., Bressac-de Paillerets, B., Chompret, A., Ghiorzo, P., Gruis, N., Hansson, J., Harland, M., Hayward, N., Holland, E. A., Mann, G. J., Mantelli, M., Nancarrow, D., Platz, A., Tucker, M. A., and the Melanoma Genetics Consortium. Geographical Variation in the penetrance of CDKN2A mutations for melanoma. J. Natl. Cancer Inst. (Bethesda), 94: 894 –903, 2002. 59. Maestro, R., and Boiocchi, M. Sunlight and melanoma: an answer from MTS1 (p16). Science (Wash. DC), 267: 15–16, 1995. 60. Pollock, P. M., Yu, F., Qiu, L., Parsons, P. G., and Hayward, N. K. Evidence for u. v. induction of CDKN2 mutations in melanoma cell lines. Oncogene, 11: 663– 668, 1995. 61. Pavey, S., Conroy, S., Russell, T., and Gabrielli, B. Ultraviolet radiation induces p16CDKN2A expression in human skin. Cancer Res., 59: 4185– 4189, 1999. 62. Rizos, H., Puig, S., Badenas, C., Malvehy, J., Darmanian, A. P., Jimenez, L., Mila, M., and Kefford, R. F. A melanoma-associated germline mutation in exon 1␤ inactivates p14ARF. Oncogene, 20: 5543–5547, 2001.

6730

Downloaded from cancerres.aacrjournals.org on September 5, 2015. © 2002 American Association for Cancer Research.

Ink4a/Arf Deficiency Promotes Ultraviolet Radiation-induced Melanomagenesis Juan A. Recio, Frances P. Noonan, Hisashi Takayama, et al. Cancer Res 2002;62:6724-6730.

Updated version

Cited articles Citing articles

E-mail alerts Reprints and Subscriptions Permissions

Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/62/22/6724

This article cites 56 articles, 23 of which you can access for free at: http://cancerres.aacrjournals.org/content/62/22/6724.full.html#ref-list-1 This article has been cited by 11 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/62/22/6724.full.html#related-urls

Sign up to receive free email-alerts related to this article or journal. To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at [email protected]. To request permission to re-use all or part of this article, contact the AACR Publications Department at [email protected].

Downloaded from cancerres.aacrjournals.org on September 5, 2015. © 2002 American Association for Cancer Research.