Keratinocyte-Specific Pten Deficiency Results in Epidermal Hyperplasia, Accelerated Hair Follicle Morphogenesis and Tumor Formation1

[CANCER RESEARCH 63, 674–681, February 1, 2003] Keratinocyte-specific Pten Deficiency Results in Epidermal Hyperplasia, Accelerated Hair Follicle Morphogenesis and Tumor Formation1

Akira Suzuki, Satoshi Itami,2 Minako Ohishi,2 Koichi Hamada, Tae Inoue, Nobuyasu Komazawa, Haruki Senoo, Takehiko Sasaki, Junji Takeda, Motomu Manabe, Tak Wah Mak,3, 4 and Toru Nakano3, 5 Departments of Biochemistry [A. S., K. H.], Dermatology [T. I., M. M.], and Anatomy [H. S.], Akita University School of Medicine, Akita 010-8543, Japan; Department of Molecular Cell Biology, Research Institute for Microbial Disease [M. O., T. N.], Department of Dermatology [S. I.], and Department of Social and Environmental Medicine [N. K., J. T.], Graduate School of Medicine, Osaka University, Osaka 565-0871, Japan; Department of Pharmacology, Tokyo Metropolitan Institute of Medical Science, Tokyo 113-8613, Japan [T. S.]; and Advanced Medical Discovery Institute, University of Toronto, Toronto, Ontario, M5G 2C1 Canada [T. W. M.]

ABSTRACT The epidermis of the skin is a keratinized and stratified squamous epithelium composed mainly of keratinocytes, cells whose prolifera- PTEN is a tumor suppressor gene mutated in many human cancers. We tion and differentiation must be tightly regulated and coordinated used the Cre-loxP system to generate a keratinocyte-specific null mutation (13–15). Keratinocytes first attach to the basal membrane of the of Pten in mice (k5Ptenflox/flox mice). k5Ptenflox/flox mice exhibit wrinkled skin because of epidermal hyperplasia and hyperkeratosis and ruffled, epidermis as undifferentiated precursor cells. These precursors mi- shaggy, and curly hair. Histological examination revealed that skin mor- grate toward the surface of the epidermis and subsequently form its phogenesis is accelerated in k5Ptenflox/flox mice. Within 3 weeks of birth, outermost layer (16). Both the normal development of keratinocytes 90% of k5Ptenflox/flox mice die of malnutrition possibly caused by hyper- and the onset of tumors involving these cells require complex prolif- keratosis of the esophagus. All k5Ptenflox/flox mice develop spontaneous erative/apoptotic events. The biochemical pathways underlying many tumors within 8.5 months of birth, and chemical treatment accelerates the of these events have been delineated through the use of transgenic and onset of tumors. k5Ptenflox/flox keratinocytes are hyperproliferative and knockout mice (17). Hair follicle morphogenesis and hair remodeling resistant to apoptosis and show increased activation of the Pten down- also depend on a balance of proliferative and apoptotic events. How- stream signaling mediators Akt/protein kinase B (PKB) and extracellular ever, the molecular mechanisms underlying hair morphogenesis are signal-regulated kinase. Pten is thus an important regulator of normal not well understood. A normal hair follicle cyclically traverses three development and oncogenesis in the skin. phases of organ growth: hair shaft formation (anagen); organ involu- tion (catagen); and relative quiescence (telogen). Hair follicle mor- INTRODUCTION phogenesis continues until the first hair cycle after birth (18), after PTEN (also called MMAC1 or TEP1) is a tumor suppressor gene which periodical hair remodeling occurs throughout the life of the mutated in both human sporadic cancers (1) and in hereditary cancer animal. syndromes such as Cowden disease and Bannayan-Zonana syndrome The importance of EGFR, IGF-1, and Ha-ras signaling in morpho- (2, 3). Among the characteristic symptoms of Cowden disease are genesis and carcinogenesis is clearly evident in skin (19–21). Acti- ␣ those involving the skin such as trichilemmomas in the face papules, vation of the transforming growth factor /EGFR pathway (22–24), papillomatosis in the mucosal and cutaneous tissues, and hyperkera- IGF-1 (20, 25), or v-Ha-ras (26) in the epidermis of transgenic mice tosis in the acral region of the skin (4). The association of cutaneous leads to skin hyperplasia, hyperkeratosis, and tumor formation. Trans- squamous cell carcinomas with Cowden disease has also been re- genic animals overexpressing IGF-1 also show accelerated hair ported (5, 6). growth (27) and mice lacking IGF-1R exhibit hypoplastic skin (28). Ј PTEN is a dual protein and lipid phosphatase (7, 8). PTENЈs major Because each of EGFR, IGF-R, and Ras triggers PI3 K signaling and substrate is PIP36, a second messenger molecule generated by PI3ЈK PTEN regulates this pathway, we felt it likely that the PTEN gene activated in response to numerous growth factors such as EGF (9), would play an important role in skin development and oncogenesis. In previous work, we used gene targeting to create a complete null hepatocyte growth factor (10), fibroblast growth factors (11), and IGF-1 Ϫ/Ϫ (12). PIP3 in turn activates the serine-threonine kinase Akt/PKB, which mutation of Pten in mice (Pten mice; Ref. 29). However, total is involved in antiapoptosis, proliferation and oncogenesis. Thus, by deficiency for Pten proved to be embryonic lethal in the very early dephosphorylating PIP3, PTEN negatively regulates cell survival. stages of development, precluding the functional analysis of Pten in various organs. We subsequently showed in PtenϪ/Ϫ MEFs that Akt/PKB is a key player downstream of Pten and that Akt/PKB is Received 8/5/02; accepted 12/16/02. The costs of publication of this article were defrayed in part by the payment of page hyperactivated in the absence of Pten (30). Mice heterozygous for the ϩ Ϫ charges. This article must therefore be hereby marked advertisement in accordance with Pten null mutation (Pten / mice) were frequently found to develop 18 U.S.C. Section 1734 solely to indicate this fact. lymphoid hyperplasia and endometrial, prostatic, and breast cancers. 1 This work was supported by grants from the Ministry of Education, Science, Sports and Culture, Japan; Japan Medical Association; the Naito Foundation; the ONO Medical With respect to skin anomalies, focal hyperkeratosis was reported in Research Foundation; the Japanese Society for Promotion of Science (JSPS- a proportion of Ptenϩ/Ϫ mice (29, 31, 32). Because Pten is expressed RAT98L01101); the Sumitomo Foundation; and the Public Trust Haraguchi Memorial Cancer Research Foundation. in keratinocytes and decreased Pten activity contributes to the malig- 2 These individuals contributed equally as second authors. nant conversion stage in chemically induced mouse skin cancers (33), 3 These individuals contributed equally as last authors. we decided to investigate the role of Pten in skin development and 4 To whom requests for reprints should be addressed, at Advanced Medical Discovery Institute, 620 University, Suite 706, Toronto, Ontario, M5G 2C1 Canada. Phone: oncogenesis using the Cre-loxP system. (416) 204-2236; Fax: (416) 204-5300; E-mail: [email protected]. 5 To whom requests for reprints should be addressed, at Department of Molecular Cell Biology, Research Institute for Microbial Disease, Osaka University, 3-1 Yamadaoka, Suita, Osaka 565-0871, Japan. Phone: 81-6-6879-8361; Fax: 81-6-6879-8362; E-mail: MATERIALS AND METHODS [email protected]. 6 The abbreviations used are: PIP3, phosphatidylinositol 3,4,5-triphosphate; PI3ЈK, Generation of k5Ptenflox/flox Mice. Ptenflox/flox mice (129 Ola x C57BL6/J phosphoinositide-3-kinase; EGF, epidermal growth factor; EGFR, epidermal growth F2) generated as previously described (29) were mated to Keratin5Cre trans- factor receptor; IGF, insulin-like growth factor; MEF, murine embryonic fibroblast; SE, genic mice (C57BL6/J background; Ref. 34) in which expression of Cre is scanning electron microscopy; DMBA, 7,12-dimethylbenz(a)anthracene; TPA, 12-O- tetradecanoylphorbol-13-acetate; MAPK, mitogen-activated protein kinase; LOH, loss of controlled by the endogenous promoter of the keratinocyte-specific gene heterozygosity; PKB, protein kinase B. Keratin5. Offspring carrying Keratin5Cre and two copies of the floxed Pten 674

allele (k5CrePtenflox/flox), Keratin5Cre plus one copy of the floxed Pten allele medium without supplements and cultured in the presence of EGF (30 ␮g/ml) (k5CrePtenflox/ϩ), and Keratin5Cre plus two copies of the wild- for 72 h. Cells were pulsed with 1␮Ci [3H]thymidine (Amersham) for another type Pten allele (k5CrePtenϩ/ϩ) were used in the analyses as homozygous 16 h before harvesting by trypsinization. Incorporated radioactivity was meas- ϩ mutant (k5Ptenflox/flox), heterozygous mutant (k5Ptenflox/ ) and wild-type ured using a ␤-scintillation counter. ϩ ϩ (k5Pten / ) mice, respectively. Apoptosis Assay. Keratinocytes cultured in 6-well plates coated with type PCR Analysis of Pten Genotypes. Genomic DNA from mouse tails was I collagen were treated with either UV- or ␥-irradiation at the doses indicated isolated and amplified by PCR following a published protocol (29). Sense in Fig. 4. One day after treatment, cell viability was determined by staining primer [5Ј-GTCACCAGGATGCTTCTGAC-3Ј] and antisense primer [5Ј- with 7-amino-actinomycin D (Sigma) as described previously (35). GAAACGGCCTTAACGACGTAG-3Ј] were used to detect the floxed Pten SE. Hairs plucked from the anterior backs of mice were attached to alu- allele; sense primer [5Ј-GTCACCAGGATGCTTCTGAC-3Ј] and antisense minum mounts and coated with 350 angstrom gold using EIKO IB-5 sputter primer [5Ј-GTGACATCAACATGCAACACTG-3Ј] were used to detect the coater. Samples were observed under a JEOL JSM-T200 scanning electron wild-type Pten allele; and sense primer [5Ј-ATGCCAATGCCCCCTCAGT- microscope operated at 15-kV accelerating voltage. Photographs were taken Ј Ј Ј TCCT-3 ] and antisense primer [5 -TGCCCCTTTTTATCCCTTCCAGA-3 ] using Fuji Neopan SS film. were used to detect the Keratin5Cre transgene. Amplified fragments of 512 bp, Tumor Induction. Mice (6–7 weeks old) were shaved on their backs 2 ϳ 413 bp and 300 bp, respectively, were obtained. days before tumor induction. To induce tumors, the shaved dorsal skin of mice Preparation and in Vitro Culture of Keratinocytes. Full-thickness skin was treated topically with 25 nmol of DMBA (Sigma) in acetone. After 1 taken from newborn mice was treated with 250 units/ml dispase (Godoshusei, week, each animal received subsequent topical treatments of 10 nmol of TPA Tokyo, Japan) overnight at 4°C. The epidermis was peeled off from the dermis (Sigma) in acetone twice weekly for 4 weeks. Control mice were treated with and trypsinized to prepare single cells that were suspended in Defined Kera- acetone only. tinocyte-SFM medium (Life Technologies, Inc.) with supplements and 2% FCS. Cells were seeded at 1 ϫ 107 cells/10-cm dish (for proliferation assays) or 5 ϫ 105 cells/well of a 6-well plate (for apoptosis assays), which had been RESULTS precoated with collagen type I (Iwaki Glass, Tokyo, Japan). Cells were cultured at 37°Cin5%CO for 5 h until cells had attached and spread. flox/flox 2 Generation of Keratin5-CrePten Mice. Keratinocyte- Unattached cells were removed by washing with PBS, and the attached cells specific Pten-deficient mice were generated using the strategy described were additionally cultured in fresh medium without FCS for 24 h before proliferation and apoptosis assays (see below). in “Materials and Methods” and Fig. 1A. Briefly, Keratin5Cre transgenic Southern and Western Blots. Genomic Southern blots of DNA obtained (k5CreTg) mice (34) were crossed to mice homozygous for the floxed flox/flox flox/flox from keratinocytes were performed using a previously described probe and Pten allele (Pten mice) to generate Keratin5CrePten mice protocol (29). For Western blots, keratinocytes (5 ϫ 106/10 ml) were either left (k5Ptenflox/flox mice). The Keratin5 promotor directs gene expression untreated or stimulated for 10 min or 16 h with 30 ␮g/ml EGF (Biomedical from day 13.5 postcoitum (d.p.c.) in the basal layer of epidermal and Technologies). Total cell lysates were prepared, and 15 ␮g of lysate aliquots follicular keratinocytes so that expression of a floxed gene under the were analyzed by Western blotting as described previously (30). Antibodies control of this promoter is disrupted throughout the epidermis and the directed against the NH terminus of Pten and antiactin were from Santa Cruz, 2 outer root sheaths of hair follicles (36, 37). In mice bearing a floxed allele whereas antiphospho-Akt/PKB (Ser473), antitotal Akt/PKB, antiphospho- MAPK (p42/p44), and antitotal MAPK(p42/p44; Thr202/Tyr204) antibodies of the Pig-a gene, expression of Cre using the Keratin5 promotor induced were from New England Biolabs. the disruption of Pig-a not only in the vast majority of keratinocytes but Histological Analysis and Immunohistochemistry. For histological anal- also in the esophagus and stomach (34). These results are consistent with ysis, dorsal skin samples and tumors were fixed in formalin and embedded in a previous report in which the Keratin5 gene was found to be expressed paraffin before sectioning according to standard protocols. Sections of 5 ␮m in esophagus and stomach as well as in epidermis (36, 38). were cut and stained with H&E. For immunohistochemical staining, freshly k5Ptenflox/flox mice were born alive and initially appeared healthy. dissected skin samples were covered with Tissue-Tek OCT compound (Miles, Genomic Southern blotting of DNA from the keratinocytes of new- ␮ Inc.) and quickly frozen in liquid nitrogen. Frozen sections (5- m thick) were borns showed that Cre-mediated recombination of loxP sites resulted fixed in ice-cold acetone. Immunofluorescent staining was performed using in the deletion of much of the 6.0-kb Ptenflox allele in almost all anti-Keratin5, anti-Keratin6, anti-Keratin10 (all from BabCo) and anti-Ki-67 ⌬ (Novocastra) as primary antibodies, and fluorescein-conjugated donkey anti- keratinocytes, leaving the 2.3-kb Pten allele (Fig. 1B). The deletion rabbit or antimouse IgG (Jackson ImmunoResearch) as secondary antibodies. of Pten was confirmed at the protein level by Western blotting using Proliferation Assay. Keratinocytes cultured in a 10 cm dish were replated antibody recognizing the NH2 terminus of Pten (Fig. 1C). Extracts of in round-bottomed 96-well plates (1 ϫ 104/well) in Defined Keratinocyte-SFM keratinocytes from k5Ptenflox/flox mice showed a major reduction in

Fig. 1. Generation of keratinocyte-specific Pten- deficient (k5Ptenflox/flox) mice. A, targeting strategy. Exons of the murine Pten gene are represented by f, loxP sites by black arrowheads. The probe used to analyze Southern blots is indicated and has been described previously (35). The floxed (Ptenflox) and deleted (Pten⌬) alleles are shown. B, genomic Southern blot. DNA (20 ␮g) extracted from keratinocytes of the indicated genotypes was digested with HindIII. The vast majority of keratinocytes from k5CrePtenflox/flox mice (k5Ptenflox/flox) showed deletion of the Pten gene. C, Western blot analysis of Pten protein expression by keratinocytes of the indicated genotype. Actin, loading control.

675

Pten protein, whereas keratinocytes of k5Ptenflox/ϩ mice showed only giving the mutants the appearance of a wrinkled bear (Fig. 2, D a minor decrease. and E). Lethality in the Lactation Period and Gross Skin Abnormalities Surprisingly, Ͼ90% of k5Ptenflox/flox mice died ϳ6 days after birth in k5Ptenflox/flox Mice. We monitored the growth and health of because of malnutrition during the lactation period (Fig. 2F). How- k5Ptenϩ/ϩ, k5Ptenflox/ϩ, and k5Ptenflox/flox mice for several months. ever, animals that survived this period and made it to the 2-month All k5Ptenflox/flox mice could be identified without genotyping be- mark had normal life spans without severe malnutrition. Overall, these cause the skin of mutant animals 3 days of age or older was wrinkled mutant mice looked similar to k5Ptenϩ/ϩ and k5Ptenflox/ϩ mice, because of hyperplasia (Fig. 2A). In addition, most k5Ptenflox/flox mice although their skin and hair abnormalities were still obvious. Gross were significantly smaller than their k5Ptenϩ/ϩ littermates from 3–5 histological examination of k5Ptenflox/flox pups revealed marked hy- days of age on (Fig. 2B), and their hair coats were abnormally ruffled perkeratosis in the esophagus (Fig. 2, G and H). Milk was easily and shaggy (Fig. 2C). k5Ptenflox/flox mice had hyperplastic noses and identified in the stomachs of wild-type mice but only faintly visible in lips (Fig. 2, D and E) and slightly enlarged ears (data not shown), the stomachs of k5Ptenflox/flox pups. It is highly likely that esophageal dysfunction because of hyperkeratosis caused the dysphagia, perhaps accounting for the small body size of the mutant mice and their deaths because of malnutrition. Examination of the hair of k5Ptenflox/flox mice showed that the cuticles, which normally cover the hair shaft, were almost completely detached from the hair shafts (Fig. 2, I and J). In contrast, k5Ptenflox/ϩ mice were indistinguishable from k5Ptenϩ/ϩ mice in both gross appearance and histological findings (data not shown). Precocious Hair Follicle Morphogenesis in Pten-deficient Epidermis. To determine the cause of the gross abnormalities of k5Ptenflox/flox skin, we analyzed the microscopic architecture of skin in tissue specimens taken at various time points after birth. Histolog- ical comparison with wild-type mice revealed that hyperkeratosis, hypergranulosis, and epidermal hyperplasia were present in Pten- deficient newborns (Fig. 3; day 0, top and middle panels). Moreover, the high density of the hair follicles in the mutant caused a reduction of interfollicular epidermis (Fig. 3; day 0, top panel and arrow in bottom panel; day 7, top panel), and sebaceous glands showed advanced development in k5Ptenflox/flox skin (Fig. 3; day 7, arrows in top and bottom panels). At ϳ10 days after birth, the epithelia of k5Ptenflox/flox mice became papillomatous. Immunostaining using antibodies specific for Keratin5 and Keratin10 showed no obvious difference between k5Ptenϩ/ϩ and k5Ptenflox/flox mice, except that ectopic staining by anti-Keratin10 was observed in the infundibular epidermis of the mutants (Fig. 3; day 10, middle panel). In wild-type mice, skin morphogenesis is completed by the end of the first hair follicle cycle. Hair remodeling then begins its lifelong cycle of regression, resting, and spontaneous regrowth (39). In this study, the first anagen phase of hair development in wild-type mice continued until day10, the day on which the thickest skin was present (Fig. 3; day10, top left panel). In contrast, the skin of k5Ptenflox/flox mice was much thicker than that of wild-type animals on day 7 but significantly thinner on day 10, suggesting that the mutants had already regressed into the catagen phase (Fig. 3; day 7, top panel and day 10, top panel). In wild-type mice, relative quiescence (telogen) was observed at day 19 (Fig. 3; day 19, top and middle left panels). However, on day 19 in k5Ptenflox/flox mice, the epithelia were papillomatous and thicker than in the wild-type, and well-developed sebaceous glands were present (Fig. 3; day 19, top and middle right panels). Immunostaining using anti-Ki-67 antibody indicated that on day 19, basal cells in the mutant were proliferating more than those of wild-type littermates (Fig. 3, day 19, bottom panels). Moreover, the proliferation of basal cells in the mutant on day 19 was greater than ϩ ϩ Fig. 2. Gross abnormalities of k5Ptenflox/flox mice. A–C, reduced size. k5CrePten / that of mutant basal cells on day 15 (data not shown), consistent with (left) and k5CrePtenflox/flox (right) pups at (A) 4 days, (B) 6 days, and (C) 12 days after birth. The skin of k5CrePtenflox/flox mice exhibits hyperproliferation and the hair is ruffled accelerated entrance into the next anagen phase. Although follicular and shaggy. D and E, wrinkled bear appearance. Lower face regions of k5CrePtenϩ/ϩ (D) remodeling was not impaired in the absence of Pten (data not shown), flox/flox flox/flox and k5CrePten mice (E) at 1 month of age. F, lethality of the k5CrePten most epithelia in the mutant mice continued to show hyperkeratosis, mutation. Numbers of dead mice (f) on the indicated day after birth and the total number of the surviving mice ( ) are plotted. G and H, severe hyperkeratosis (arrow) in the lower hyperplasticity (especially the granular layer), and well-developed esophagus. Comparative histology of lower esophagus of k5CrePtenϩ/ϩ (G) and sebaceous glands even at 3 months of age (Fig. 3, 3 months, right k5CrePtenflox/flox (H) mice. I and J, cuticle detachment. SE analysis of cuticles (arrow)of k5CrePtenϩ/ϩ (I) and k5CrePtenflox/flox (J) mice. The cuticle of the k5CrePtenflox/flox hair panels). The hair was consistently abnormal (data not shown). The has almost completely separated from the hair shaft. presence of hyperplastic epithelia and the acceleration of skin mor- 676

Fig. 3. Hyperkeratosis, epidermal hyperplasia, and accelerated morphogenesis in k5Ptenflox/flox mice. Histological skin sections of k5CrePtenϩ/ϩ (left panels) and k5CrePtenflox/flox (right panels) mice sacrificed on the indicated day after birth. The second rows of the day 0, 7, 19, and 3 month panels show a higher magnification of the respective first rows. Immunohistochemical analyses using anti-Keratin10 (day 10, middle row), anti-Keratin5 (day 10, bottom row), and anti-Ki-67 (day 19, bottom row) antibodies are also shown. The arrow in the bottom right panel day 0 shows the reduction in interfollicular epidermis in k5CrePtenflox/flox mice, and the arrows in the top and bottom right panels of day 7 indicate a well-developed sebaceous gland. phogenesis in k5Ptenflox/flox mice indicate that Pten is indispensable homozygous mutant mice that survived over 2 months after birth for the normal development of skin. were monitored for spontaneous tumorigenesis. As shown in Skin Tumor Formation in k5Ptenflox/؉ and k5Ptenflox/flox Mice. Fig. 4A, no skin tumors were detected on k5Ptenϩ/ϩ mice during To further characterize the effects of Pten gene disruption on the 8.5 month observation period. In contrast, spontaneous tumors mouse epidermis, 32 wild-type, 31 heterozygous mutant, and 11 developed in 23% of k5Ptenflox/ϩ mice and, most surprisingly, in 677

Fig. 4. Spontaneous and chemically induced tumors in k5CrePtenflox/flox mice. A, spontaneous tumor incidence. The health of 11 k5CrePtenflox/flox, 31 k5CrePtenflox/ϩ, and 32 k5CrePtenϩ/ϩ mice was monitored for 8.5 months, and the percentage of animals in each group that survived without developing tumors was plotted. B–I, occurrence of specific tumor types: (B and F) spontaneous papillomas; (C and G) low-grade squamous cell carcinoma; (D and H) sebaceous carcinoma; (E and I) adenocarcinoma of the eccrine gland. F–I are H&E-stained histological sections of the tumors shown in situ in B–E. J, PCR analysis demonstrating LOH in low-grade squamous cell carcinomas (scc) observed in 3 different k5CrePtenflox/ϩ mice. Analysis of liver DNA from a k5CrePtenflox/ϩ mouse is shown as a control. K, multiple papillomas induced in a k5CrePtenflox/flox mouse by DMBA initiation and TPA promotion. L and M, histological H&E staining of a section of the papillomas shown in (K).

100% of k5Ptenflox/flox mice. Most of these spontaneous tumors mice treated with DMBA followed by TPA developed 5–15 skin were squamous papillomas that occurred on the face and palms of papillomas in the treated area (only) within 5 weeks of the initial the front paws (Fig. 4, B and F). However, many of these papil- DMBA treatment (Fig. 4, K–M). In contrast, no tumors were present lomas went on to develop into squamous cell carcinomas with on the skin of either k5Ptenflox/ϩ or wild-type mice treated in the same nuclear atypia and increased mitosis. These latter tumors were fashion for the same duration. TPA treatment alone failed to induce capable of invading the dermis (Fig. 4, C and G). In addition to tumorigenesis in any of the three groups during the 6-week observa- papillomas and squamous cell carcinomas, we observed sebaceous tion period (data not shown). This result was not unexpected because carcinomas characterized by an obvious sebaceous gland-like spontaneous tumors did not appear in k5Ptenflox/flox mice until 3.5 structure (in 11% of tumor-bearing mice; Fig. 4, D and H) and months of age. These observations indicate that Pten functions as a adenocarcinomas of the sweat gland (in 11% of tumor-bearing tumor suppressor for both spontaneous and induced skin tumors in mice; Fig. 4, E and I). To confirm that the tumors arising in mice. k5Ptenflox/ϩ mice arose because of a LOH, PCR assays on tumor Hyperproliferation and Resistance to Apoptosis in k5Ptenflox/flox DNA were carried out. The loss of the wild-type Pten allele was Keratinocytes. We next investigated whether the accelerated skin observed in 3 of 3 squamous carcinomas obtained from flox/flox ϩ morphogenesis and oncogenesis in k5Pten mice were associ- k5Ptenflox/ mice (Fig. 4J). To examine induced carcinogenesis in Pten-deficient skin, ated with a defect in keratinocyte proliferation or apoptosis. Kerati- k5Ptenϩ/ϩ, k5Ptenflox/ϩ and k5Ptenflox/flox mice of 6–7 weeks of age nocytes were stimulated in vitro with EGF to induce proliferation or (n ϭ 8/group) were treated with either DMBA plus TPA or TPA subjected to UV- or ␥-irradiation to induce apoptosis. As shown in alone. DMBA initiates skin tumorigenesis, and TPA promotes growth Fig. 5A, Pten-deficient keratinocytes showed enhanced proliferation of an established skin tumor. Surprisingly, 100% of k5Ptenflox/flox in response to EGF. In addition, the mutant keratinocytes were more 678

characteristic clinical features of Cowden’s disease include skin abnormalities such as trichilemmomas in the face papules, papillomatosis in the mucosal and cutaneous tissues, and hyperkeratosis in the acral region of the skin (4). At the molecular level, skin lesions in Cowden’s disease patients have shown evidence of LOH for PTEN (42). In this study, we observed significant similarities to Cowden’s disease symptoms in k5Ptenflox/flox mice. The mutant animals exhib- ited hyperkeratosis and spontaneous tumors such as papillomas and cutaneous squamous cell carcinomas. Although trichilemmomas, a type of tumor of the epidermal appendages, did not occur in k5Ptenflox/flox mice, the mutants frequently displayed other types of epidermal appendage tumors. We therefore believe our k5Ptenflox/flox mice represent not only a reasonable model of the skin lesions characteristic of Cowden’s disease but also a suitable model of skin carcinogenesis in general. The phenotypes observed in k5Ptenflox/flox mice are reminiscent of those of Ha-ras (26), transforming growth factor ␣ (22), sons of sevenless (SOS) (24), and IGF-1 (25, 27) transgenic mice. The similarities between k5Ptenflox/flox mice and transgenic mice expressing IGF-1 under the control of the Keratin5 promotor (k5IGF-1 Tg) are Fig. 5. Enhanced proliferation and resistance to apoptosis of Pten-deficient keratinocytes. A, hyperproliferation. Purified keratinocytes from newborn k5Ptenϩ/ϩ and particularly striking. k5IGF-1 Tg mice are small in size as neonates k5Ptenflox/flox mice were incubated with EGF (10 ␮g/ml), and proliferation was measured and have wrinkled, thick skin because of hyperkeratosis and epithelial by thymidine uptake. Keratinocytes from k5Ptenflox/flox mice showed enhanced prolifer- hyperplasia. The hair of these mutants is ruffled and shaggy, and P Ͻ 0.05. Results ,ء .ation. Mean thymidine uptake Ϯ SE for 3 mice/group is shown shown are representative of four independent trials. B, resistance to apoptosis. Purified spontaneous tumors such as papillomas and squamous cell carcinomas keratinocytes from newborn k5Ptenϩ/ϩ and k5Ptenflox/flox mice were prepared as de- develop with age. Cells of these mutants show evidence of apoptotic 2 scribed in “Materials and Methods” and stimulated with UV- (5 or 15J/m )or␥-irradi- resistance and activation of PI3ЈK and its downstream mediators ation (5 or 20 Gy). Cell viability was determined by 7-amino-actinomycin D staining after 16-h stimulation. The percentage of viable cells remaining after treatment as compared Akt/PKB and ERK (25). These observations, taken together with the with the viability of untreated cells cultured for the same length of time is shown. results of our study, suggest that signal transduction downstream of Keratinocytes from k5Ptenflox/flox mice were significantly more resistant to apoptosis than the wild type. Results are expressed as the mean Ϯ SE for 3 mice/group and are many molecules affecting cell growth/death involves common Akt/ representative of three trials. PKB- and ERK-mediated pathways. Disruption of these pathways may account for the phenotypes of k5Ptenflox/flox mice. One of the most striking phenotypes in k5Ptenflox/flox mice is the resistant than wild-type keratinocytes to apoptosis induced by either precocious morphogenesis of skin. The wnt/␤-catenin/Lef-1 pathway high dose UV- or ␥-irradiation (Fig. 5B). has been reported to be very important for the acceleration of devel- Activation of Akt/PKB and MAPK in Pten-deficient Keratino- opmental morphogenesis in the skin (43). In the embryonic skin of cytes. We have previously reported that regulation of Akt/PKB acti- 14.5 d.p.c mice, pilosebaceous units develop from epidermal down- vation by Pten is critical for normal apoptosis in MEFs and for growths under the influence of specific mesenchymal cell condensa- proliferation/apoptosis in T cells (30, 35). MAPK, a major signaling tions. These condensations supply permissive and instructive signals molecule downstream of Ras, is also activated downstream of PIP3 that govern the position and type of hairs and other appendages (40). Our previous demonstration that both MAPK and Akt/PKB are developed (reviewed by Refs. 38, 44). The expression of patterning activated in Pten-deficient T cells (35) prompted us to analyze the genes such as those in the wnt/␤-catenin/Lef-1 signaling pathway are phosphorylation of Akt/PKB and MAPK in keratinocytes from thought to regulate these signals (43). Perhaps significantly, Pten has k5Ptenϩ/ϩ, k5Ptenflox/ϩ, and k5Ptenflox/flox newborn mice. After stim- been shown to negatively regulate the ␤-catenin/Lef-1 pathway by ulation with EGF, the phosphorylation of Akt/PKB (Fig. 6A) and MAPK (Fig. 6B) was significantly elevated in k5Ptenflox/flox keratinocytes compared with k5Ptenϩ/ϩ or k5Ptenflox/ϩ keratinocytes. Thus, in keratinocytes, as in T cells, Akt/PKB and MAPK activation is subject to negative regulation by Pten.

DISCUSSION In previous studies, we showed in MEFs, T cells, and neuronal cells that Pten deficiency causes accumulation of PIP3 and constitutive activation of Akt/PKB. These events result in cellular hyperproliferation and resistance to apoptosis that can lead to abnormal development and malignancy (29, 30, 41). We now report similar findings in murine skin in vivo and in keratinocytes in vitro from studies of mice in which the Cre-loxP system was used to selectively disrupt the Pten gene in the epidermis. Fig. 6. Activation of Akt/PKB and MAPK in Pten-deficient keratinocytes. Immunoblot analysis of the expression and activation of Akt/PKB (A) and MAPK/ERK (B) proteins. Hereditary heterozygous mutation of PTEN in humans is associated Keratinocytes of the indicated genotypes were stimulated with EGF (10 ␮g/ml) for 15 min with Cowden’s disease (2), a disorder characterized by the onset of (for pAkt/PKB) or 10 min and 16 h (for pMAPK), and phosphorylation of Akt/PKB and ERKs was determined by Western blotting. Levels of phosphorylated Akt/PKB, ERK1 multiple hamartoma in various tissues. These hamartoma frequently (p44) and ERK2 (p42) were significantly elevated in k5Ptenflox/flox keratinocytes com- develop into malignancies such as breast and thyroid cancers (5). The pared with k5Ptenϩ/ϩ cells. 679

Downloaded from cancerres.aacrjournals.org on September 23, 2021. © 2003 American Association for Cancer Research. ROLE OF PTEN IN SKIN MORPHOGENESIS AND ONCOGENESIS inhibiting the nuclear accumulation of ␤-catenin and activation of REFERENCES Lef-1 in a prostatic cell line (45). However, in our hands, no definite 1. Li, J., Yen, C., Liaw, D., Podsypanina, K., Bose, S., Wang, S. I., Puc, J., Miliaresis, flox/flox difference in the subcellular distribution of ␤-catenin in k5Pten C., Rodgers, L., McCombie, R., Bigner, S. H., Giovanella, B. C., Ittmann, M., Tycko, cells was observed (data not shown). Both k5Ptenflox/flox mice and B., Hibshoosh, H., Wigler, M. H., and Parsons, R. PTEN, a putative protein tyrosine phosphatase gene mutated in human brain, breast, and prostate cancer. Science (Wash k5IGF-1 Tg mice show accelerated hair growth at day 5 (27), indi- DC), 275: 1943–1947, 1997. cating that common molecules downstream of PKB/Akt in addition to 2. Liaw, D., Marsh, D. J., Li, J., Dahia, P. L., Wang, S. I., Zheng, Z., Bose, S., Call, ␤-catenin, or molecules downstream of ERK, may account for the K. M., Tsou, H. C., Peacocke, M., Eng, C., and Parsons, R. Germline mutations of the PTEN gene in Cowden disease, an inherited breast and thyroid cancer syndrome. Nat. accelerated skin morphogenesis in these mice. Genet., 16: 64–67, 1997. Several lines of evidence suggest that Akt/PKB may be a key 3. Marsh, D. J., Dahia, P. L., Zheng, Z., Liaw, D., Parsons, R., Gorlin, R. J., and Eng, molecule regulating the onset of skin carcinogenesis in mice. The C. Germline mutations in PTEN are present in Bannayan-Zonana syndrome. Nat. Genet., 16: 333–334, 1997. transplantation of keratinocytes overexpressing Akt/PKB results in 4. Hildenbrand, C., Burgdorf, W. H., and Lautenschlager, S. Cowden syndrome-diag- highly aggressive skin tumors characterized by increased invasiveness nostic skin signs. Dermatology (Basel), 202: 362–366, 2001. and altered differentiation (33). In addition, Akt/PKB activation is one 5. Nuss, D. D., Aeling, J. L., Clemons, D. E., and Weber, W. N. Multiple hamartoma syndrome (Cowden’s disease). Arch. Dermatol., 114: 743–746, 1978. of the first events in the chemical induction of skin tumors (33). 6. Camisa, C., Bikowski, J. B., and McDonald, S. G. Cowden’s disease. Association Finally, the onset of skin tumor formation in mice requires EGFR and with squamous cell carcinoma of the tongue and perianal basal cell carcinoma. Arch. Akt/PKB signaling in addition to SOS/Ras/ERK signaling (24). In- Dermatol., 120: 677–678, 1984. Ј 7. Myers, M. P., Stolarov, J. P., Eng, C., Li, J., Wang, S. I., Wigler, M. H., Parsons, R., deed, MAPK/ERK has been reported to act in synergy with the PI3 K and Tonks, N. K. P-TEN, the tumor suppressor from human chromosome 10q23, is pathway to stimulate CycD1 transcription in NIH3T3 cells (46). Thus, a dual-specificity phosphatase. Proc. Natl. Acad. Sci., USA 94: 9052–9057, 1997. the onset of tumors in k5Ptenflox/flox mice may be caused primarily by 8. Maehama, T., and Dixon, J. E. The tumor suppressor, PTEN/MMAC1, dephospho- rylates the lipid second messenger, phosphatidylinositol 3,4,5-trisphosphate. J. Biol. cellular hyperproliferation and/or apoptotic resistance induced by Chem., 273: 13375–13378, 1998. PI3ЈK and Akt/PKB hyperactivation, with a contribution by deregu- 9. Hu, P., Margolis, B., Skolnik, E. Y., Lammers, R., Ullrich, A., and Schlessinger, lated ERK activation. J. Interaction of phosphatidylinositol 3-kinase-associated p85 with epidermal growth factor and platelet-derived growth factor receptors. Mol. Cell. Biol., 12: 981–990, k5IGF-1 Tg mice are reportedly hypersensitive to both DMBA and 1992. TPA. However, analysis of papillomas observed in k5IGF-1 Tg mice 10. Ponzetto, C., Bardelli, A., Maina, F., Longati, P., Panayotou, G., Dhand, R., treated with TPA alone revealed that essentially all of the tumors had Waterfield, M. D., and Comoglio, P. M. A novel recognition motif for phosphatidylinositol 3-kinase binding mediates its association with the hepatocyte growth factor/ Ha-ras mutations (25). This finding suggests that enhanced IGF-1 scatter factor receptor. Mol. Cell. Biol., 13: 4600–4608, 1993. signaling does not substitute for initiation but rather enhances the 11. Raffioni, S., and Bradshaw, R. A. Activation of phosphatidylinositol 3-kinase by ability of TPA to select for already existing transformed cells (25). epidermal growth factor, basic fibroblast growth factor, and nerve growth factor in PC12 pheochromocytoma cells. Proc. Natl. Acad. Sci. USA, 89: 9121–9125, 1992. Because both Pten deficiency and IGF-1 overexpression result in 12. Yamamoto, K., Lapetina, E. G., and Moxham, C. P. Insulin like growth factor-I Akt/PKB and subsequent MAPK activation, it is conceivable that the induces limited association of phosphatidylinositol 3-kinase to its receptor. Endocri- same mechanism underlies the phenotypes in k5Ptenflox/flox mice. nology, 130: 1490–1498, 1992. flox/flox 13. Fuchs, E. Epidermal differentiation. Curr. Opin. Cell Biol., 2: 1028–1035, 1990. Prolonged observation of k5Pten mice subjected to treatment 14. Fuchs, E. Epidermal differentiation: the bare essentials. J. Cell Biol., 111: 2807–2814, with either DMBA or TPA alone may help to define the role of Pten 1990. in chemically induced carcinogenesis. 15. Fuchs, E. Keratin genes, epidermal differentiation and animal models for the study of flox/flox flox/ϩ human skin diseases. Biochem. Soc. Trans., 19: 1112–1115, 1991. In this study, both k5Pten and k5Pten mice developed 16. Jones, P. H., and Watt, F. M. Separation of human epidermal stem cells from transit flox/flox spontaneous skin tumors at high frequency. Fully 100% of k5Pten amplifying cells on the basis of differences in integrin function and expression. Cell, mice and 23% of k5Ptenflox/ϩ mice acquired papillomas and/or squamous 73: 713–724, 1993. 17. Wu, X., and Pandolfi, P. P. Mouse models for multistep tumorigenesis. Trends Cell cell carcinomas during the 8.5-month observation period. The result for Biol., 11: S2–S9, 2001. flox/ϩ the k5Pten mice is particularly surprising in light of the fact that skin 18. Sano, S., Itami, S., Takeda, K., Tarutani, M., Yamaguchi, Y., Miura, H., Yoshikawa, tumors are observed in Ptenϩ/Ϫ mice at a frequency of Ͻ5%. We K., Akira, S., and Takeda, J. Keratinocyte-specific ablation of Stat3 exhibits impaired skin remodeling, but does not affect skin morphogenesis. EMBO J., 18: 4657–4668, speculate that the tissue-specificity of the Pten mutation examined in this 1999. study may in some way account for this discrepancy. For example, an 19. Yamamoto, T., Kamata, N., Kawano, H., Shimizu, S., Kuroki, T., Toyoshima, K., immunosurveillance mechanism able to block skin tumor formation may Rikimaru, K., Nomura, N., Ishizaki, R., Pastan, I., et al. High incidence of amplifi- cation of the epidermal growth factor receptor gene in human squamous carcinoma be triggered in response to global (only) inactivation of Pten in heterozy- cell lines. Cancer Res., 46: 414–416, 1986. gotes. Such a mechanism would be in line with our previous observations 20. Rho, O., Bol, D. K., You, J., Beltran, L., Rupp, T., and DiGiovanni, J. Altered of mice with a T-cell-specific Pten deficiency. These mutants experience expression of insulin-like growth factor I and its receptor during multistage carcinogenesis in mouse skin. Mol. Carcinog., 17: 62–69, 1996. an accumulation of T cells that produce increased levels of Th1 and Th2 21. Vassar, R., Hutton, M. E., and Fuchs, E. Transgenic overexpression of transforming cytokines (35). growth factor alpha bypasses the need for c-Ha-ras mutations in mouse skin tumor- Our study is the first reported in vivo analysis of Pten function in igenesis. Mol. Cell. Biol., 12: 4643–4653, 1992. 22. Vassar, R., and Fuchs, E. Transgenic mice provide new insights into the role of mouse skin. We clearly demonstrate that Pten is an essential regulator TGF-␣ during epidermal development and differentiation. Genes Dev., 5: 714–727, of normal homeostasis and oncogenesis in the organ. Our results 1991. suggest that inhibition of the PIP3-Akt/PKB pathway may be an 23. Dominey, A. M., Wang, X. J., King, L. E., Jr., Nanney, L. B., Gagne, T. A., Sellheyer, K., Bundman, D. S., Longley, M. A., Rothnagel, J. A., Greenhalgh, D. A., et al. attractive therapeutic target for the treatment of skin malignancies. Targeted overexpression of transforming growth factor ␣ in the epidermis of trans- Studies to examine PTEN expression in human keratinocyte malig- genic mice elicits hyperplasia, hyperkeratosis, and spontaneous, squamous papillo- nancies are ongoing. mas. Cell Growth Differ., 4: 1071–1082, 1993. 24. Sibilia, M., Fleischmann, A., Behrens, A., Stingl, L., Carroll, J., Watt, F. M., Schlessinger, J., and Wagner, E. F. The EGF receptor provides an essential survival signal for SOS-dependent skin tumor development. Cell, 102: 211–220, 2000. ACKNOWLEDGMENTS 25. DiGiovanni, J., Bol, D. K., Wilker, E., Beltran, L., Carbajal, S., Moats, S., Ramirez, A., Jorcano, J., and Kiguchi, K. Constitutive expression of insulin-like growth We thank Dr. Tetsuo Noda (Tohoku University), Dr. Shinichi Ansai (Akita factor-1 in epidermal basal cells of transgenic mice leads to spontaneous tumor University), Miki Nagai (Osaka University), and Noriko Asada (Osaka Uni- promotion. Cancer Res., 60: 1561–1570, 2000. 26. Greenhalgh, D. A., Rothnagel, J. A., Quintanilla, M. I., Orengo, C. C., Gagne, T. A., versity) for helpful discussions and technical assistance; Dr. Mary Saunders Bundman, D. S., Longley, M. A., and Roop, D. R. Induction of epidermal hyperplasia, (Advanced Medical Discovery Institute) for scientific editing; and Miki Sato hyperkeratosis, and papillomas in transgenic mice by a targeted v-Ha-ras oncogene. Suzuki for her valuable assistance. Mol. Carcinog., 7: 99–110, 1993. 680

27. Bol, D. K., Kiguchi, K., Gimenez-Conti, I., Rupp, T., and DiGiovanni, J. Overexpression 36. Byrne, C., and Fuchs, E. Probing keratinocyte and differentiation specificity of the of insulin-like growth factor-1 induces hyperplasia, dermal abnormalities, and spontane- human K5 promoter in vitro and in transgenic mice. Mol. Cell. Biol., 13: 3176–3190, ous tumor formation in transgenic mice. Oncogene, 14: 1725–1734, 1997. 1993. 28. Liu, J. P., Baker, J., Perkins, A. S., Robertson, E. J., and Efstratiadis, A. Mice carrying 37. Murillas, R., Larcher, F., Conti, C. J., Santos, M., Ullrich, A., and Jorcano, J. L. null mutations of the genes encoding insulin-like growth factor I (Igf-1) and type 1 Expression of a dominant negative mutant of epidermal growth factor receptor in the IGF receptor (Igf1r). Cell, 75: 59–72, 1993. epidermis of transgenic mice elicits striking alterations in hair follicle development 29. Suzuki, A., de la Pompa, J. L., Stambolic, V., Elia, A. J., Sasaki, T., del Barco and skin structure. EMBO J., 14: 5216–5223, 1995. Barrantes, I., Ho, A., Wakeham, A., Itie, A., Khoo, W., Fukumoto, M., and Mak, 38. Ramirez, A., Bravo, A., Jorcano, J. L., and Vidal, M. Sequences 5Ј of the bovine T. W. High cancer susceptibility and embryonic lethality associated with mutation of keratin 5 gene direct tissue- and cell-type-specific expression of a lacZ gene in the the PTEN tumor suppressor gene in mice. Curr. Biol., 8: 1169–1178, 1998. adult and during development. Differentiation, 58: 53–64, 1994. 30. Stambolic, V., Suzuki, A., de la Pompa, J. L., Brothers, G. M., Mirtsos, C., Sasaki, 39. Hardy, M. H. The secret life of the hair follicle. Trends Genet., 8: 55–61, 1992. T., Ruland, J., Penninger, J. M., Siderovski, D. P., and Mak, T. W. Negative 40. Shan, X., Czar, M. J., Bunnell, S. C., Liu, P., Liu, Y., Schwartzberg, P. L., and regulation of PKB/Akt-dependent cell survival by the tumor suppressor PTEN. Cell., Wange, R. L. Deficiency of PTEN in Jurkat T cells causes constitutive localization of 95: 29–39, 1998. Itk to the plasma membrane and hyperresponsiveness to CD3 stimulation. Mol. Cell. 31. Stambolic, V., Tsao, M. S., Macpherson, D., Suzuki, A., Chapman, W. B., and Mak, Biol., 20: 6945–6957, 2000. 41. Backman, S. A., Stambolic, V., Suzuki, A., Haight, J., Elia, A., Pretorius, J., Tsao, T. W. High incidence of breast and endometrial neoplasia resembling human Cowden M. S., Shannon, P., Bolon, B., Ivy, G. O., and Mak, T. W. Deletion of Pten in mouse syndrome in ptenϩ/Ϫ mice. Cancer Res., 60: 3605–3611, 2000. brain causes seizures, ataxia and defects in soma size resembling Lhermitte-Duclos 32. Podsypanina, K., Ellenson, L. H., Nemes, A., Gu, J., Tamura, M., Yamada, K. M., disease. Nat. Genet., 29: 396–403, 2001. Cordon-Cardo, C., Catoretti, G., Fisher, P. E., and Parsons, R. Mutation of Pten/ 42. Trojan, J., Plotz, G., Brieger, A., Raedle, J., Meltzer, S. J., Wolter, M., and Zeuzem, Mmac1 in mice causes neoplasia in multiple organ systems. Proc. Natl. Acad. Sci. S. Activation of a cryptic splice site of PTEN and loss of heterozygosity in benign USA, 96: 1563–1568, 1999. skin lesions in Cowden disease. J. Investig. Dermatol., 117: 1650–1653, 2001. 33. Segrelles, C., Ruiz, S., Perez, P., Murga, C., Santos, M., Budunova, I. V., Martinez, 43. Gat, U., DasGupta, R., Degenstein, L., and Fuchs, E. De novo hair follicle morpho- J., Larcher, F., Slaga, T. J., Gutkind, J. S., Jorcano, J. L., and Paramio, J. M. genesis and hair tumors in mice expressing a truncated ␤-catenin in skin. Cell., 95: Functional roles of Akt signaling in mouse skin tumorigenesis. Oncogene, 21: 53–64, 605–614, 1998. 2002. 44. Stenn, K. S., Combates, N. J., Eilertsen, K. J., Gordon, J. S., Pardinas, J. R., Parimoo, 34. Tarutani, M., Itami, S., Okabe, M., Ikawa, M., Tezuka, T., Yoshikawa, K., Kinoshita, S., and Prouty, S. M. Hair follicle growth controls. Dermatol. Clin., 14: 543–558, T., and Takeda, J. Tissue-specific knockout of the mouse Pig-a gene reveals important 1996. roles for GPI-anchored proteins in skin development. Proc. Natl. Acad. Sci. USA, 94: 45. Persad, S., Troussard, A. A., McPhee, T. R., Mulholland, D. J., and Dedhar, S. Tumor 7400–7405, 1997. suppressor PTEN inhibits nuclear accumulation of ␤-catenin and T cell/lymphoid 35. Suzuki, A., Yamaguchi, M. T., Ohteki, T., Sasaki, T., Kaisho, T., Kimura, Y., enhancer factor 1-mediated transcriptional activation. J. Cell Biol., 153: 1161–1174, Yoshida, R., Wakeham, A., Higuchi, T., Fukumoto, M., Tsubata, T., Ohashi, P. S., 2001. Koyasu, S., Penninger, J. M., Nakano, T., and Mak, T. W. T cell-specific loss of Pten 46. Gille, H., and Downward, J. Multiple ras effector pathways contribute to G1 cell cycle leads to defects in central and peripheral tolerance. Immunity, 14: 523–534, 2001. progression. J. Biol. Chem., 274: 22033–22040, 1999.

681

Downloaded from cancerres.aacrjournals.org on September 23, 2021. © 2003 American Association for Cancer Research. Keratinocyte-specific Pten Deficiency Results in Epidermal Hyperplasia, Accelerated Hair Follicle Morphogenesis and Tumor Formation

Akira Suzuki, Satoshi Itami, Minako Ohishi, et al.

Cancer Res 2003;63:674-681.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/63/3/674

Cited articles This article cites 46 articles, 21 of which you can access for free at: http://cancerres.aacrjournals.org/content/63/3/674.full#ref-list-1

Citing articles This article has been cited by 44 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/63/3/674.full#related-urls

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/63/3/674. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.