Human Cripto-1 overexpression in the mouse mammary ... - Nature

Oncogene (2005) 24, 4094–4105

& 2005 Nature Publishing Group All rights reserved 0950-9232/05 $30.00 www.nature.com/onc

Human Cripto-1 overexpression in the mouse mammary gland results in the development of hyperplasia and adenocarcinoma Christian Wechselberger1, Luigi Strizzi1, Nicholas Kenney 1,2, Morihisa Hirota1, Youping Sun1, Andreas Ebert3, Olivia Orozco4, Caterina Bianco1, Nadia I Khan1, Brenda Wallace-Jones1, Nicola Normanno5, Heather Adkins4, Michele Sanicola4 and David S Salomon*,1 1

Tumor Growth Factor Section, Mammary Biology and Tumorigenesis Laboratory, National Cancer Institute, Bethesda, MD 20892, USA; 2Department of Biological Science, Hampton University, Hampton, VA 23668, USA; 3Department of Obstetrics and Gynecology, Medical Center Benjamin Franklin, Free University Berlin, 12200 Berlin, Germany; 4Biogen Idec Inc., 14 Cambridge Center, Cambridge MA 02142, USA; 5Division of Haematological Oncology and Department of Experimental Oncology, IT Fondazione Pascale, 80131 Naples, Italy

Human Cripto-1 (CR-1) is overexpressed in approximately 80% of human breast, colon and lung carcinomas. Mouse Cr-1 upregulation is also observed in a number of transgenic (Tg) mouse mammary tumors. To determine whether CR-1 can alter mammary gland development and/or may contribute to tumorigenesis in vivo, we have generated Tg mouse lines that express human CR-1 under the transcriptional control of the mouse mammary tumor virus (MMTV). Stable Tg MMTV/CR-1 FVB/N lines expressing different levels of CR-1 were analysed. Virgin female MMTV/CR-1 Tg mice exhibited enhanced ductal branching, dilated ducts, intraductal hyperplasia, hyperplastic alveolar nodules and condensation of the mammary stroma. Virgin aged MMTV/CR-1 Tg mice also possessed persistent end buds. In aged multiparous MMTV/ CR-1 mice, the hyperplastic phenotype was most pronounced with multifocal hyperplasias. In the highest CR1-expressing subline, G4, 38% (12/31) of the multiparous animals aged 12–20 months developed hyperplasias and approximately 33% (11/31) developed papillary adenocarcinomas. The long latency period suggests that additional genetic alterations are required to facilitate mammary tumor formation in conjunction with CR-1. This is the first in vivo study that shows hyperplasia and tumor growth in CR-1-overexpressing animals. Oncogene (2005) 24, 4094–4105. doi:10.1038/sj.onc.1208417 Published online 16 May 2005 Keywords: Cripto-1; transgenic mice; mammary development; mammary tumorigenesis; MMTV-LTR

*Correspondence: DS Salomon, Tumor Growth Factor Section, Mammary Biology & Tumorigenesis Laboratory, National Cancer Institute, Building 10, Room 5B39, 10 Center Drive, Bethesda, MD 20892, USA; E-mail: [email protected] Received 13 April 2004; revised 8 November 2004; accepted 8 November 2004; published online 16 May 2005

Introduction Human Cripto-1 (CR-1) is a cell surface glycosylphosphatidylinositol (GPI)-linked protein that is overexpressed in a majority of solid human carcinomas, including breast, ovary, colon, gastric, cervix, uterus, lung and pancreas (Salomon et al., 1999, 2000). CR-1 is a member of the EGF-CFC protein family and structurally contains an epidermal growth factor (EGF)-like domain and a cysteine-rich region called the Cripto-1/FRL-1/Cryptic (CFC) domain (Ciccodicola et al., 1989; Salomon et al., 1999, 2000). The EGF-CFC family includes mouse cripto-1 (Cr-1) (Dono et al., 1993), chicken Cripto-1 (Colas and Schoenwolf, 2000), zebrafish one-eyed pinhead (oep) (Zhang et al., 1998), Xenopus Frl-1 (Kinoshita et al., 1995; Shen and Schier, 2000) and mouse and human cryptic (Shen et al., 1997). Cripto-1’s most well-defined role is as a coreceptor for the transforming growth factor b (TGF-b) ligands, Nodal and Vg1/GDF-1 (Schier and Shen, 2000; Shen and Schier, 2000; Yan et al., 2002; Cheng et al., 2003). Genetic studies in zebrafish and mice have defined an essential role for Cripto-1 in the specification of mesoderm, endoderm and patterning of the anterior– posterior axis during embryogenesis (Ding et al., 1998; Xu et al., 1999). These genetic studies were also the first to implicate Cr-1 and oep as coreceptors for Nodal, a TGF-b family member. Subsequent biochemical studies have demonstrated that Cripto-1 binds directly to both Nodal and to the Activin type I receptor, Alk 4 (ActRIB), to recruit the Activin type II receptor and to facilitate Nodal signaling (Shen and Schier, 2000; Yeo and Whitman, 2001; Bianco et al., 2002; Adkins et al., 2003). Site-directed mutagenesis experiments have shown the EGF-like domain to be critical for the Cripto-1-Nodal interaction (Yeo and Whitman, 2001; Yan et al., 2002), while the CFC domain has been identified as the domain that interacts with Alk 4 (Yeo and Whitman, 2001; Bianco et al., 2002; Adkins et al., 2003). Recent work has expanded Cripto-1’s role as a negative modulator of TGF-b ligand activity. CR-1 is able to block Activin signaling in tumor cell lines via Alk

Mammary tumors in Cripto-1 transgenic mice C Wechselberger et al

4095

4 receptor binding and by its capacity to directly bind to Activin B and/or the Activin type II receptor (Adkins et al., 2003; Gray et al., 2003). In addition, studies in Xenopus have shown that Frl-1, Cr-1 or CR-1 can bind to Alk4 and facilitate signaling of Vg1/GDF ligands in embryos (Cheng et al., 2003). CR-1 has also been shown in mammalian cells to bind directly to the GPI-linked heparin sulfate-containing proteoglycan, glypican-1, which can then activate c-src (Bianco et al., 2003). Bianco et al. (2003) demonstrated that the CR-1– glypican-1 interaction, which results in c-src activation, can lead to downstream activation of the ras/raf/MAPK pathway and a PI-3 kinase/Akt pathway. Cripto-1’s ability to stimulate cell migration and to block cell differentiation and apoptosis have been associated with these src-dependent responses that are independent of any Nodal, Alk4 and/or TGF-b ligand interactions. Finally, Xenopus Frl-1 has been recently implicated as a coreceptor for Xnr3, a distant member of the Nodal family (Yokota et al., 2003). Unlike most Nodal proteins, Xnr3 does not signal via Alk 4 and Activin type II receptors. Xnr3 is a direct target gene of the wnt/b-catenin pathway and signals indirectly via the FGFR1 receptor. This probably occurs by the ability of Frl-1 to directly regulate through its interaction with Xnr3 the expression in Xenopus of N-CAM, which is a FGF-independent FGFR activator (Kinoshita et al., 1995; Yokota et al., 2003). Therefore, the EGF-CFC proteins can bridge different signaling pathways as comodulators/coreceptors, ligands or indirect effectors (Adamson et al., 2002). Perturbations of the EGF, FGF, TGF-b and Wnt growth factor signaling pathways that control cell proliferation and differentiation are often associated with cell transformation and malignant tumors in vivo (Salomon et al., 2000; Howe and Brown, 2004). CR-1 is overexpressed in a number of different types of human carcinomas and has been demonstrated to modulate the signaling of these pathways in human tumor cell lines in vitro (Adamson et al., 2002). In vivo, experiments have shown that tumor growth can be suppressed by downregulation of CR-1 using an antisense strategy (Ciardiello et al., 1994; Normanno et al., 1996, 1999; De Luca et al., 1997, 1999) or inhibition of CR-1’s CFC signaling domain with an anti-CR-1 mouse monoclonal antibody or with an anti-EGF-like domain CR-1 rat monoclonal antibody (Adkins et al., 2003; Xing et al., 2004). These experiments validate that blockade of CR1 can modulate tumor growth and points to CR-1 as a potential therapeutic target. To ascertain the biological effects and oncogenic potential of CR-1, CR-1 was overexpressed in the mouse mammary gland and its ability to modify mammary gland development and/or contribute to mammary tumorigenesis in vivo was examined. CR-1 is known to be overexpressed in human breast carcinomas (Salomon et al., 2000). Specifically, greater than 80% of infiltrating ductal carcinomas and 50% of ductal carcinomas in situ were found to be positive for CR-1 expression. Only 15% of noninvolved breast tissue stained weakly for CR-1 in these studies (Qi et al., 1994; Normanno et al.,

1995; Panico et al., 1996; Srinivasan et al., 2000). Expression of endogenous mouse Cr-1 mRNA and protein in the developing mouse mammary gland has already been described and assessed by in situ hybridization, Northern blot analysis, Western blot analysis and immunohistochemistry (Kenney et al., 1995; Herrington et al., 1997; Niemeyer et al., 1998). These studies show that Cr-1 expression in the prepubescent and pubescent mammary gland can first be detected in the cap stem cells of the growing terminal end buds (TEBs) and later in the ductal epithelial cells after puberty but not in the surrounding stroma. Also, during pregnancy, expression of the endogenous Cr-1 in the lobuloalveolar cells is increased approximately three- to five fold and decreases again later during lactation. Overexpression of Cr-1 can be detected in preneoplastic hyperplastic lesions and in tumors of the mammary gland in several different transgenic (Tg) mouse lines that overexpress, in the mammary gland, oncogenes such as int-3, c-neu (rat ortholog of human erbB-2), TGFa, polyoma middle t and SV40 Large T (Kenney et al., 1996; Niemeyer et al., 1999). Whether these specific oncogenes are involved in the regulation of CR-1 expression is not known. In vitro, CR-1 overexpression can induce anchorage-independent growth of nontransformed mouse mammary epithelial cells (Ciardiello et al., 1991; Niemeyer et al., 1998; Wechselberger et al., 2001). In addition, stable overexpression of Cr-1 or CR-1 in several nontransformed mouse mammary epithelial cell lines stimulates chemotaxis, chemokinesis and branching morphogenesis, suggesting a possible role for these proteins during the process of epithelial–mesenchymal transition (EMT) (Ebert et al., 2000; Wechselberger et al., 2001; Strizzi et al., 2004). Finally, primary mouse mammary epithelial cells which retrovirally overexpress Cr-1 when introduced into the cleared mammary fat pad of syngeneic, ovariectomized virgin mice give rise to outgrowths that exhibit an increase in secondary and tertiary lateral side branching (Kenney et al., 1997). Ductal hyperplasias were also observed in these mammary outgrowths. However, no mammary carcinomas were observed after several in vivo transplantations, but the level of Cr-1 expression was not quantifiable and the nature of the transplantation experiment did not allow for the analysis of the transduced Cr-1 gene in multiparous animals. Thus while Cr-1 expression has been correlated with mouse and human mammary tumors and Cripto-1 can transform normal mammary epithelial cells in vitro, no studies have directly assessed whether CR-1 can modify mammary gland development and/or may contribute to mammary tumorigenesis in vivo. To address this question, we generated several transgenic mouse lines in which mammary gland expression of the human CR-1 protein was selected by use of the mouse mammary tumor virus long terminal repeat (MMTV-LTR) as a promoter. Three sublines were established that express the CR-1 transgene during early stages of lactation at low, intermediate and high levels, respectively. These data provide the first evidence that mammary-specific overexpression of CR-1 results in alterations of the Oncogene


4096

development of the mouse mammary gland, including induction of ductal hyperplasias in virgin animals and the development of carcinomas in aged, multiparous animals.

Results Generation of MMTV/CR-1 transgenic mice and expression analysis of the transgene To create stable mouse Tg lines overexpressing CR-1 in mammary epithelial cells in vivo, single cell FVB/N strain mouse zygotes where microinjected with a construct containing the full-length cDNA of human CR-1 under control of the mouse mammary tumor virus (MMTV) long terminal repeat promoter (Figure 1a; Donehower et al., 1981). A total of 18 founder transgenic FVB/N mice (11 males and seven females) were generated. Integration of the transgene construct

2B3-7 into the genome of potential founders was assessed by Southern blot analysis. Founders exhibiting germline transmission of the transgene were identified by PCR analysis of genomic DNA using human CR-1specific primers. Protein extracts prepared from various tissues were used to determine the sites and levels of CR1 expression in these animals with a mouse monoclonal antibody that only detects the human transgene-derived protein. Transgenic CR-1 protein expression could be detected in the mammary gland 3 months after weaning. Three founders were identified which genetically passed the transgene through the germline into sublines and which also expressed the human CR-1 transgene protein in the lactating mammary gland at low (G2-line), intermediate (K3-line) and high levels (G4-line) (Figure 1b). Expression levels of the CR-1 protein and mRNA were found to correlate with the transgene copy number differences (data not shown). In addition, several tissues that are known to be permissive for the MMTV-promoter expressed the human CR-1 protein. In male animals, strong expression of CR-1 protein was observed in brain, lung, testis, epididymis, seminal vesicles and lacrimal gland (Figure 1c), consistent with the tissue specific-expression pattern of transgene expression noted in other MMTV-based Tg strains (Webster et al., 1998). In females, expression levels in brain, lung, uterus and lacrimal gland were comparable to those found in male mice (data not shown). No protein expression was detected in the ovaries. Mice from all three heterozygous MMTV/CR-1 transgenic lines lactated and bred normally. Mammary gland development in virgin MMTV/CR-1 transgenic mice

Figure 1 (a) EcoR1 fragment (1.22 kb) containing the human CR1 cDNA was placed downstream of the MMTV-LTR promoter/ enhancer to direct expression on the transgene in mammary epithelium. (b) CR-1 protein expression as detected by Western blot analysis using a mouse anti-CR-1 monoclonal antibody that detects human CR-1 in day 3 lactating mammary gland from three independent founder lines of MMTV/CR-1 transgenic mice. Lanes 1–3, 5 and 8 correspond to lysates prepared from wild-type FVB/N mammary glands or mammary glands from transgenic lines (5 and 8) not expressing the human CR-1 protein. Lanes 6, 4 and 7 correspond to the level of CR-1 protein expression detected in mammary lysates prepared from lactating G2, K3 and G4 mice, respectively. (c) Expression of CR-1 protein in 6-month-old male MMTV/CR-1 G4 brain (B), colon (C), testis (T), epididymis (E), seminal vesicles (S), mammary gland rudiment (M), lacrimal gland (L) and wild-type, 6-month-old FVB/N female mammary gland (Ctl) Oncogene

Since CR-1 or Cr-1 overexpression in several different mouse mammary epithelial cell lines leads to enhanced branching in type I collagen gels in vitro (Wechselberger et al., 2001), we assessed whether mammary epithelialspecific overexpresssion of CR-1 could perturb ductal morphogenesis in virgin Tg mice by performing whole mount and histological analyses on MMTV/CR-1 animals. TEB morphology was indistinguishable between Tg and wild-type animals at 4 weeks of age at this stage (Figure 2a and b). In contrast, examination at 12 weeks of age revealed that 8/10 (80%) of the virgin G4 MMTV/CR-1 Tg mice displayed an increase in the number of TEBs compared to 12-week-old FVB/N control mice (Figure 2c and d). By 12 weeks, all of the G4 MMTV/CR-1 transgenic mice exhibited a greater number of small, speculated side buds compared to 12-week-old FVB/N control mice (Figure 2c and d). This phenotype was still evident in nulliparous 6-monthold G4 MMTV/CR-1 transgenic mice (Figure 2e and f). This phenotype was observed in several different MMTV/CR-1 G4 mice both at 4 and 6 weeks of age as compared to wild-type FVB/N virgin mice (Figure 3A). Identification of dilated lateral ducts having a cross-sectional diameter greater than 100 mm were significantly more frequent in the MMTV/CR-1 G4 transgenic mice relative to wild-type FVB/N mice at


4097

4 and 6 weeks (Figure 3A, B). When compared to 13-month-old wild-type FVB/N mammary gland (Figure 3Ba), K3 MMTV-CR-1 Tg mammary gland showed significant dilation of ducts and focal areas of intraductal epithelial proliferation (Figures 3Bb). This phenotype was most pronounced in the G4 subline. Likewise, in contrast to wild-type 13-monthold FVB/N mice, there is pronounced condensation of the stroma, particularly the periductal mesenchyme, and compression of the adipocytes and adipose tissue immediately adjacent to the ducts (Figure 3Bb). This dilatory phenotype of the ducts and stromal condensation occurs in 3/10 (30%) of the K3 and 8/10 (80%) of the G4 MMTV-CR-1 mammary glands analysed. Compared to 13-month-old FVB/N control mice (Figure 4a), there was an increase in lateral side branching of the secondary and tertiary ducts observed in all three of the MMTV/CR-1 Tg sublines that was even more readily apparent at 13 months (Figure 4b, c and d). Quantification of the lateral side branching showed a significant increase in lateral side branching in

Figure 2 Whole-mount histology of mammary glands from virgin wild-type FVB/N (a, c and e) used as control or MMTV/CR-1 G4 female mice (b, d and f) at 4 weeks (a and b), 12 weeks (c and d) or 6 months (e and f) showing increased number of TEBs in the MMTV/CR-1 mammary glands (arrows) compared to control

Figure 3 (A) Number of dilated mammary ducts (black), TEBs (red), end buds (green) and lateral secondary or tertiary side branches (blue) observed in 4-week-old and 6-week-old FVB/N (WT) and G4 (CR-1) mammary glands. Data are expressed as numbers per mm27standard deviation. (B) Hematoxylin and eosin-stained sections from 13-month-old virgin mammary glands from wild-type FVB/N (a) or MMTV/CR-1 K3 transgenic mice (b). Note dilation of ducts, intraductal epithelial proliferation (red arrows) and increased eosin staining due to increase in stroma content (black arrows) in MMTV-CR-1 K3 mammary gland versus wild-type mammary gland

Figure 4 Whole mounts of mammary glands from nulliparous 13month-old (a) virgin wild-type FVB/N, (b) MMTV/CR-1 G2, (c) MMTV/CR-1 K3 or (d) MMTV/CR-1 G4 mice exhibiting increased number of secondary and tertiary ductal side branching and end buds in MMTV/CR-1 virgin mammary glands. In (e), quantification shows significantly increased lateral secondary or tertiary side branches in 13-month-old mammary glands from MMTV-CR-1 G2, MMTV-CR-1 K3 and MMTV-CR G4 female mice compared to wild-type FVB/N. Values represent the mean of branch points counted in at least four different fields from 10 separate animals. Error bars represent standard deviation. Epithelial content in 6-month-old founder MMTV/CR-1 transgenic mice expressed as percent relative to the content in whole mounts prepared from age-matched wild-type FVB/N is represented in (f) Oncogene


4098

the three different Tg lines compared to FVB/N control mice and also suggested a dose-dependent effect between the different levels of expression of CR-1 protein among the three Tg lines and number of branch points (Figure 4e). A similar correlation was also observed for the increase of the epithelial content in the mammary gland of these three MMTV-CR-1 Tg sublines relative to FVB/N mice as assessed at 6 months of age by morphometric image scanning (Figure 4f). Mammary gland whole mounts of 6–15-month old virgin mice exhibit signs of focal proliferation (Figure 5a) in 5/10 (50%) of the K3 and 8/10 (80%) of the G4 MMTV-CR-1 mice analysed. Histological analysis of these areas showed hyperplastic alveolar nodules

Figure 5 A representative whole mount (a) shows a hyperplastic region in the mammary gland from a 13-month-old G4 transgenic. Corresponding histology shows hyperplastic alveolar nodular (HAN) lesion (b) and focal hyperplasia (c) Oncogene

(HANs) with areas of hypersecretion and cystic alteration (Figure 5b) and intraductal epithelial hyperplasia (Figure 5c). Mammary gland development in MMTV/CR-1 transgenic mice during pregnancy, lactation and involution During early pregnancy and lactation, there were no significant phenotypic differences either in the timing, morphology or histology of lobuloalveolar development or in the level of expression of different milk proteins such as b-casein or whey acidic protein (WAP) between the wild-type FVB/N or MMTV-CR-1 transgenic mice (data not shown). In addition, no significant differences were observed in estrogen or progesterone receptor expression in the mammary gland between wild-type FVB/N and MMTV-CR-1 transgenic mice as assessed by immunohistochemistry either in the virgin or pregnant mammary glands (data not shown). Examination of the mammary gland of Tg animals during early involution also revealed that there were no significant differences in the timing or rates of involution in the MMTV/CR-1 mice as compared to FVB/N female mice when analysed at day 2 of involution (Figure 6a versus b). This was also found when the percentage of

Figure 6 Histological sections of mammary glands from MMTV/ CR-1 transgenic (a) and FVB/N (b) obtained at day 2 of involution show no significant differences. At day 30 of involution, whole mounts and histological sections of mammary glands from MMTV/CR-1 G4 (respectively, c and e) and wild-type FVB/N (respectively, d and f) mice show incomplete involution with less atrophic ducts (c versus d) and increased epithelial content (e versus f) at 30 days in MMTV/CR-1 G4 mammary gland compared to the fully regressed age-matched FVB/N wild-type mammary gland


4099

apoptotic cells was assessed by tunnel assay at day 2 of involution in wild-type FVB/N and MMTV/CR-1 G4 mice (data not shown). However, whole mount (Figure 6c versus d) and histological analysis at 30 days postinvolution (Figure 6e versus f) showed fewer atrophic ducts and increased retention of a epithelial cells in the G4 MMTV-CR-1 as compared to the wildtype FVB/N mice of comparable age. To assess whether the hyperplasias that were found in the MMTV/CR-1 G4 nulliparous animals represent a functionally competent transplantable population of cells that can give rise to hyperplasias, we transplanted a hyperplastic lesion obtained from a nulliparous MMTV/ CR-1 G4 transgenic mouse into the cleared mammary fat pad of the fourth inguinal mammary gland of a syngenic 4-week-old FVB/N virgin mouse. After 11.5 weeks, we observed signs of proliferation and increased ductal outgrowth in areas of the mammary tree seen in whole mounts (Figure 7a) that suggested the presence of hyperplastic lesions. Histological analysis of sections from these hyperproliferative areas showed intraductal epithelial hyperplasias (Figure 7b). In contrast, a wildtype FVB/N mammary gland explant transplanted into the contralateral cleared mammary fat pad of the same

Figure 7 (a) A representative whole mount shows hyperplasias from MMTV/CR-1 G4 transplanted into the cleared fourth inguinal mammary fat pad of 4-week-old virgin FVB/N mouse and allowed to grow for 11.5 weeks. Note increase in density of epithelial structures from the MMTV/CR-1 G4 outgrowth, suggesting hyperplasia (arrows). Histological analysis of sections from the transplanted tissue shows intraductal epithelial hyperplasia (b)

syngenic FVB/N virgin mouse did not give rise to any hyperproliferative tissue (data not shown). This transplantable hyperplastic phenotype was observed in the transplants from four additional hyperplasias of G4 MMTV-CR-1 transgenic mice. Finally, the hyperplastic phenotype was consistently observed after four to five serial transplantations of each of the five original lesions without senescence that is normally observed in nonmalignant mammary epithelial cells. Mammary gland development in multiparous MMTV/ CR-1 transgenic animals Multiparous, aged MMTV/CR-1 animals from all sublines that had undergone at least 4–5 pregnancies were found to exhibit an increase in density of the mammary gland (Figure 8a) as compared to wild-type multiparous FVB/N mice. Multiple focal hyperplasias (Figure 8b) were observed in the G2, K3 and G4 MMTV/CR-1 mice at this stage and papillary adenocarcinomas (Figure 8c) were observed in the inguinal fourth mammary gland and in other mammary glands such as the thoracic glands in the K3 and G4 MMTV/ CR-1 mice. Depending upon the subline, 28/70–56/70 (40–80%) of multiparous transgenic animals develop multifocal hyperplasias within 12–20 months (Figure 8d). A number of these hyperplastic lesions that resembled HANs, particularly in the G4 MMTV/CR-1 subline, exhibited regions of secretory vacuolization (Figure 8b) and showed areas of intraductal epithelial

Figure 8 (a) A representative whole mount shows increased epithelial density in 12-month-old multiparous MMTV/CR-1 G4 mammary gland. Hematoxylin and eosin stained sections of MMTV/CR-1 G4 mammary gland in (b) shows secretory epithelium, some lining cystic structures and dysplastic alveolar nodules with epithelial hyperplasia (inset). In (c), Hematoxylin and eosin-stained histological section of a representative papillary adenocarcinoma from a mammary gland of a 12-month multiparous MMTV/CR-1 G4 transgenic mouse shows characteristic frond-like structures (inset). The histogram in (d) represents frequency of formation of hyperplastic lesions and carcinomas in multiparous (4–5 pregnancies) MMTV/CR-1 G2 (n ¼ 19), MMTV/ CR-1 K3 (n ¼ 20) and MMTV/CR-1 G4 (n ¼ 31) lines after 12–20 months Oncogene


4100

cell proliferation (Figure 8b inset). The lower CR-1 expressing sublines (G2 and K3) have a predominance of hyperplasias, approximately 50–80% (15/19 to 10/20, respectively) with less than 10% (2/20) of the K3 subline developing tumors (Figure 8d). In the highest CR-1expressing subline, G4, approximately 38% of the animals (12/31) that are at least 20 months and that have gone through at least four pregnancies develop hyperplasias and 33% (10/31) develop focal tumors presumably due to hyperplasias that had progressed to tumors. The tumors that arise in the MMTV/CR-1 sublines were all mixed papillary adenocarcinomas (Figure 8c). Occasionally, several tumors could be found in the same mammary gland. All of the tumors that were found had papillary lesions and were multiloculated and cystic. The epithelial cells appeared hyperchromatic and multilayered along the frod-like papillary structures (Figure 8c inset). In some tumors, areas with undifferentiated, oval to spindle-shaped tumor cells resembling mesenchyme were also observed and expressed mesenchymal markers such as vimentin, smooth muscle actin, fibronectin and N-cadherin and lacked expression of epithelial markers such as E-cadherin (Strizzi et al., 2004). No other evident tumor subtype was found in the multiple animals from the different transgenic sublines. Data for hyperplasia and tumor development was collected over a 12–20 month period from approximately 70 MMTV/CR-1 mice Tg mice from the three sublines and compared to an equal number of age-matched FVB/N multiparous animals. None of the wild-type FVB/N mice were observed to have hyperplasias or tumors at 20 months. In the three sublines, there was a decrease in the frequency of hyperplasias comparing the aged, multiparous G2, K3 and G4 mice, which was coincident with an increase in mammary tumor incidence. Mammary tumors in Tg mice appear to arise from areas of hyperplasia (Cardiff and Munn, 1995) and this suggests that the decrease in hyperplasias in the sublines K3 and G4 that develop tumors may reflect this conversion process. To ascertain if the mammary papillary tumors overexpress cripto-1, immunohistochemical analysis was performed on the MMTV-CR-1 tumors using a rabbit polyclonal anti-cripto-1 antibody capable of detecting both mouse and the transgenic human CR-1 proteins. Figure 9b shows staining for total cripto-1 expression (i.e. both mouse and human) in both the papillary component of the tumor and in the thin surrounding stroma. No staining for cripto-1 was observed in the tumor sections when only the secondary goat anti-rabbit antibody was used (Figure 8a) or when the primary anticripto-1 antibody was preabsorbed with recombinant cripto-1 protein (data not shown). Finally, Cr-1 overexpression has been observed in several different histological types of mammary tumors that arise from the activation of different oncogenes in the mammary gland of other transgenic mouse lines (Kenney et al., 1996; Niemeyer et al., 1999). Additionaly, mouse and human cripto-1 have recently been demonstrated to be a target gene in a canonical wnt/b-catenin/Lef-1 signaling pathway (Morkel et al., 2003). Since Wnt-1 is also Oncogene

Figure 9 Immunohistochemistry shows Cripto-1 expression in MMTV/CR-1 G4 (MMTV/CR-1) and MMTV/Wnt-1 (Wnt) mammary tumors

oncogenic in the mouse mammary gland (Rosner et al, 2002; Li et al., 2003; Liu et al., 2004), we examined the level of Cr-1 expression in several different mammary tumors that arise in MMTV/Wnt-1 mice (Li et al., 2003). Most Wnt-1 mammary tumors are differentiated adenocarcinomas that emerege in a solitary fashion in mammary glands that are extensively hyperplastic (Rosner et al., 2002). However, approximately 15% of MMTV-Wnt-1 tumors are papillary adenocarcinomas with areas of undifferentiated carcinoma cells (Liu et al., 2004). Intense Cr-1 staining was observed in the more anaplastic areas of the Wnt-1 tumors (Figure 9c).

Discussion The generation of MMTV/CR-1 stable Tg lines directing CR-1 expression in the mammary gland epithelium


4101

has allowed us to directly assess if CR-1 overexpression can alter normal mammary gland developmental program and/or contribute to a tumorigenic phenotype in vivo. MMTV/CR-1 Tg mammary glands from several different sublines that express different levels of the transgene versus wild-type control mammary glands from FVB/N mice were analysed at all stages of mammary gland development. The most striking phenotypes that were observed in all three MMTV-CR-1 Tg lines occurred at puberty, in aged nulliparous and in aged multiparous female mice. In virgin CR-1 transgenic mice, there was an increase in lateral ductal side branching of the secondary and tertiary ducts which agrees with the ex vivo data using primary mammary cells that overexpress a transduced Cr-1 transgene and that had been transplanted into the cleared mammary fat pad of syngeneic virgin mice (Kenney et al., 1997). In the virgin mammary gland, allometric ductal growth normally occurs from the TEBs. Signaling cues that can positively or negatively regulate mammary ductal growth and lateral side branching and that are initiated at puberty around 4 weeks in the postnatal mammary gland and during early pregnancy include systemic hormones such as estrogen, progesterone and prolactin and locally derived growth factors such amphiregulin, hepatocyte growth factor (HGF), Wnt-4, TGF-b1 and activin B. These growth factors are expressed in either the mammary epithelium (amphiregulin, TGF-b1 and Wnt-4) or stroma (HGF and activin B) (Robinson and Henninghausen, 1997; Robinson et al., 2000; Pollard, 2001; Silberstein, 2001; Naylor and Ormandy, 2002). Interestingly, while TGF-b1 generally inhibits TEB growth and lateral side branching, presumably by attenuating HGF expression in the stroma (Pollard, 2001; Silberstein, 2001), activin B is required for ductal elongation and alveolar development and branching (Robinson and Henninghausen, 1997). Likewise, during pregnancy, progesterone stimulates branching morphogenesis by its ability to stimulate the expression of wnt-4 (Brisken et al., 2000). Therefore, cripto-1 similar to HGF and wnt-4 can positively regulate lateral side branching both in vitro and in vivo (Kenney et al., 1997; Soriano et al., 1998; Uyttendale et al., 1998; Robinson et al., 2000; Wechselberger et al., 2001). Also, like the phenotype observed in the virgin mammary gland of our MMTV-CR-1 sublines, overexpression of Wnt-1 can lead to extensive lateral side branching in the mammary gland of virgin MMTV-Wnt-1 mice and to the development of multifocal hyperplastic mammary glands (Rosner et al., 2002; Li et al., 2003; Liu et al., 2004). In this respect, it is worth noting that Cr-1 and CR-1 have recently been shown to be target genes that are upregulated in a canonical wnt/b-catenin/Lef-1 signaling pathway during mouse embryogenesis and in mouse colon adenomas in Apcmin mice and in human colon carcinoma cells that possess activated b-catenin (Morkel et al., 2003). These data suggest that there may be a functional connection between Wnt-1 and cripto-1 that may lead to a partial mimicry of phenotype in the mammary gland when CR-1 is overexpressed. The increased expression of cripto-1 observed in the Wnt-1

mammary tumors indirectly would support such a possible connection in mammary tumors that are induced by an activated Wnt-1 signaling pathway involving b-catenin (Imbert et al., 2001). This is also supported by the recent observations that the MMTVCR-1 mammary tumors show elevated levels of dephosphorylated (active) b-catenin and phosphorylated (inactive) GSK-3b, two downstream effectors in the canonical wnt signaling pathway (Strizzi et al., 2004). Finally, mouse strain-specific variations in lateral ductal side branching in virgin mice have been reported and these variations are directly related to the susceptibility of different mouse strains to spontaneous mammary cancer development (Naylor and Ormandy, 2002). The enhanced lateral side branching, dilation of ducts, some of which show an increase in intraductal epithelial proliferation, and hyperplasias that are found in the mammary gland of nulliparous and multiparous MMTV-CR-1 mice along with the similarity, in part, to the phenotype in the mammary gland of MMTVWnt-1 mice, suggests that both genes when inappropriately overexpressed may contribute to the development of a premalignant state in mammary epithelial cells. Further evidence for this possibility stems from the observation that Cr-1 is frequently overexpressed in mammary hyperplasias that are found in other transgenic mice that overexpress different oncogenes (Niemeyer et al., 1999). In the case of Wnt-1 overexpression, this has been formally shown to contribute to the clonal expansion of a population of progenitor stem cells that can give rise to hyperplasias that can eventually develop into adenocarcinomas, which contain both luminal epithelial and myoepithelial cells (Li et al., 2003; Liu et al., 2004). The tumors that develop in the MMTVCR-1 mice are composed almost entirely of epithelial cells forming papillary structures but also show rare foci containing mesenchymal-like cells arranged in an undifferentiated pattern. These areas express high levels of CR1 as determined by immunohistochemistry and express markers and activated signaling molecules suggesting that EMT is occurring in these CR-1-overexpressing mammary gland tumors EMT (Strizzi et al., 2004). Further studies are in progress to clearly establish whether these mesenchymal-like cells are myoepithelial cells or truly arise via EMT from pre-existing epithelial cells. Likewise, whether Cr-1 is expressed in a population of progenitor stem cells in the mammary gland is also unknown. However, this may be possible since Cr-1 and CR-1 are expressed in pluripotential mouse and human embryonic stem (ES) cells, respectively, and are markers for stemness that are downregulated in expression following differentiation (Bhattacharya et al., 2004; Carpenter et al., 2004). Wnt-1 tumors develop in nulliparous mice within 6–12 months at a high frequency and that are predominantly well-differentiated adenocarcinomas, a small fraction of which are represented by papillary adenocarcinomas (Liu et al., 2004). The long latency period and low penetrance level observed for the formation of the MMTV-CR-1 papillary adenocarcinomas that occurs only in multiparous MMTV-CR-1 transgenic mice suggests that the tumorigenic phenotype Oncogene


4102

in the MMTV-CR-1 mice is less severe than in the MMTV/Wnt-1 mice and that additional genetic alterations may be required to compliment the effects of CR-1 and to complete the tumorigenic process. When a more rigorous comparison is made between the histology, latency and frequency of the MMTV/CR1 mammary tumors, the CR-1 mammary tumors most closely resemble mammary tumors found in the MMTV/integrin-linked kinase (ILK) Tg mouse (White et al., 2001). Approximately 55% of the MMTV/ILK Tg virgin female mice exhibit in the mammary gland an increase in ductal side branching with a large number of secondary and tertiary branches and small speculated side buds. We have found a similar phenotype in a majority of the MMTV-CR-1 G4 Tg virgin animals. In both Tg mouse models, hyperplasias are seen in virgin and multiparous animals, with a hyperplastic phenotype becoming more pronounced after multiple pregnancies. Mammary papillary adenocarcinomas in MMTV/ILK Tg mice arise in approximately 34% of the multiparous animals around 18 months, while in MMTV-CR-1 G4 Tg mice 35% of the multiparous animals develop mammary tumors of similar histology at >14 months of age. ILK is a cytoplasmic serine–threonine kinase effector of integrin receptors known to interact with different b intergrin cytoplasmic subunits. Interestingly, in the MMTV-CR-1 papillary tumors, there is an increase in expression of ILK (unpublished results), b1, b3 and b4 integrins (Strizzi et al., 2004). ILK has also been shown to induce the phosphorylation and inhibition of GSK-3b, a negative regulator of the wnt signaling pathway (Delcommenne et al., 1998; Oloumi et al., 2004). The inhibition of GSK-3b by ILK by phosphorylation results in the activation of b-catenin by dephosphorylation and the subsequent activation of the b-catenin/LEF-1 transcription factor complex. The similarity of the phenotypes observed in the MMTV/ ILK and MMTV/CR-1 Tg mice argues that cripto-1 upregulation noted in human and mouse mammary tumors may be a direct result of inappropriate b-catenin activation through an ILK and/or Wnt-1 signaling pathway. ILK can regulate downstream components in the Wnt-1 signaling pathway that can lead to tumors that exhibit evidence of EMT similar to an EMT phenotype, which is also noted in the MMTV-CR-1 tumors (White et al., 2001; Oloumi et al., 2004; Strizzi et al., 2004). A similar phenotype of papillary adenocarcinomas that develop at a low penetrance (B30%) and with a long latency period (24 months) in multiparous female mice has also been described in MMTV/ protein casein kinase 2 alpha (CK2a) transgenic mice (Landesman-Bollag et al., 2001). Similar to ILK, CK2a is also a positive regulator of the wnt signaling pathway and MMTV/CK2a transgenic mice that develop tumors also exhibit evidence of EMT. It will therefore be important to determine the pathways by which CR-1 promotes mammary epithelial cell transformation. For example, it has recently been demonstrated that CR-1 overexpression in human T47D breast cancer cells can block the growth inhibitory effects of activin B (Adkins et al., 2003). This may be Oncogene

one mechanism by which CR-1 promotes a loss of growth control. Additionally, the recent identification of the ability of CR-1 to interact with other cell surface molecules like glypican-1, which can activate c-src, MAPK and PI-3 kinase and the ability of cripto-1 to function as a target gene in the Wnt-1 signaling pathway (Bianco et al., 2003; Morkel et al., 2003; Strizzi et al., 2004) or the ability of cripto-1 to regulate the FGFR1 pathway (Yokota et al., 2003) or tomoregulin signaling (Harms and Chang, 2003), suggests that overexpression of Cripto-1 may result in perturbations of more than one signaling pathway in cancer cells. Taken together, both recent signaling data and the work presented here demonstrating that CR-1 overexpression in the mammary gland leads to the development of mammary hyperplasias and tumors highlights CR-1 as an important oncogene and as a potential target for cancer therapy. Materials and methods Generation of MMTV - CR-1 transgenic mice In order to create the expression vector 2B3(7), a 890 bp fragment containing the entire open reading frame of the human CR-1 cDNA flanked by 220 bp of 50 - and 110 bp of 30 untranslated regions was cloned into the EcoRI site of the MMTV/SV40/Bssk vector (Huang et al., 1981). The biologic activity of this construct was tested by transient transfection into mouse NMuMG cells. The production of CR-1 mRNA after induction by 1 mM dexamethasone was verified by PCR and Northern Blot analysis (data not shown). Transgenic FVB/N mice expressing CR-1 under the control of the complete MMTV long terminal repeat (LTR) were generated by pronuclear injection of the 4990 bp SpeI/SalI-digested DNA fragment from construct 2B3(7) (Laboratory Animal Sciences Program, SAIC Frederick). Analysis of the transgene integration The integration of the CR-1 construct was verified by PCR and restriction mapping of genomic DNA extracted from tail biopsies (Laird et al., 1991). PCR was performed using primers specific for the human CR-1 cDNA sequence (HCR-F1: 50 -CAG GAA TTT GCT CGT CCA TCT CGG-30 and HCRR1: 50 -TAG TAC GTG CAG ACG GTG GTA GTT-30 ). The amplification protocol for reaction conditions and cycling parameters were performed as described by the Taq polymerase manufacturer (Promega, Madison, WI, USA). Southern Blot analysis was used to determine the copy number of the integrated transgene. Briefly, genomic tail DNA was digested with EcoRI and fragments were separated on a 0.8% agarose gel, denatured with 0.5 M NaOH, 1.0 M NaCl for 30 min and blotted on nylon membrane by capillary transfer overnight in denaturation solution (Botchan et al., 1976). Transgene integration, integrity and copy numbers were determined using the EcoRI fragment of clone 2B3(7) as a probe labeled by the random priming procedure (Feinberg and Vogelstein, 1984). Unincorporated nucleotides were removed by Quick Spin Columns as described by the manufacturer (Promega, Madison, WI, USA). Hybridization was performed at 651C for 20 h as previously described (Wechselberger et al., 1992). After hybridization, membranes were washed two times for 20 min at 651C with 0.5 SSC/0.1% SDS and once for 20 min with 0.1 SSC/ 0.1% SDS. The membranes were then exposed to BioMax film (Kodak, Rochester, NY, USA) at 701C with an intensifying


4103 screen. Transgene copy numbers were estimated by comparing the hybridization signal to the signal obtained with a known amount of purified plasmid DNA of the transgene mixed with genomic tail DNA from nontransgenic mice. Tissue specificity of transgene expression and Western blot analysis For Western blotting, tissues were grounded on dry ice and lysed in 20 mM Tris-HCl (pH 7.5) containing 150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 5 mM MgCl2, 2 mg/ml aprotinin, 2 mg/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate and 20 mM sodium fluoride. Crude protein lysates (50 mg) were seperated by SDS– PAGE on 5–20% polyacrylamide gels (Novex, San Diego, CA, USA), transferred to PVDF membranes (Millipore, Bedford, MA, USA) and blocked for 1 h with 5% dry milk in Trisbuffered saline with 0.05% Tween 20. For the detection of CR1 protein, a monoclonal mouse antibody (A10.B2.18; Biogen Inc., Cambridge, MA, USA) that is directed against the N terminal domain of human CR-1 and that does not crossreact with mouse Cr-1 was used at a dilution of 1 : 2000 overnight at 41C. As a secondary antibody, a horseradish peroxidaseconjugated rabbit-anti-mouse IgG (Amersham, Piscataway, NJ, USA) was used at a concentration of 1 : 5000. Immunoreactive bands were detected by enhanced chemiluminescence (Amersham, Piscataway, NJ, USA).

acid : ethanol, 1 : 3) for 60 min at room temperature. The glands were rehydrated prior to overnight staining in aluminium carmine (1 g carmine, 2.5 g AlK(SO4)2 boiled for 20 min in distilled H2O, filtered and brought to a final volume of 500 ml). The glands were then dehydrated, cleared with xylene and mounted. Photographs were taken using a Polaroid DMC-1 digital camera (Polaroid, Cambridge, MA, USA) mounted on a Leica MZ125 microscope (Leica, Wetzlar, Germany). Areas of interest and suspect lesions were cut out of the whole mount, embedded in paraffin and processed for histology. Epithelial content of selected whole mount areas were determined by the use of NIH image analysis system software. Histomorphometric analysis of wild-type FVB/N and MMTV-CR-1 transgenic mammary gland whole mounts or histological sections and serial transplants were carried out as previously described (Kenney et al., 1996, 1997; Johnson et al., 2003). TEBs, end buds (EB), lateral side branches (LSB) and lateral ducts were assessed in whole mounts or histological sections in quadrants of the mammary gland which were not adjacent to the nipples and which measured 1 mm2. Lateral ducts with cross-sections >100 mm were considered dilated and quantification of LSB was performed by counting the number of side branches per 10 field. A total of 10 mammary inguinal whole mounts or tissue sections per group7standard deviation from individual animals were scored. Data were analysed for statistical significance by Student t test and P-valueo0.05 were considered statistically significant and identified throughout by an asterisk (*).

Histology and Immunohistochemistry Macroscopically identifiable masses were isolated and fixed with 4% phosphate-buffered paraformaldehyde overnight at 41C. Paraffin sections were prepared by American HistoLabs (Gaithersburg, MD, USA). Hematoxylin and eosin staining was performed according to standard procedures. Immunocytochemical staining with a rabbit anti-Cripto-1 antibody 1579 (1 : 100) that reacts with both mouse and human Cripto-1 protein was utilized on paraffin-embedded, formalin-fixed tissue sections using the Vectastain Kit and following the manufacturer’s suggestions (Vector Laboratories, Burlingame, CA, USA). The specificity of the antibody was checked by substitution of the primary antibody with isotype nonspecific Ig or following preabsorbtion of the primary antibody with recombinant CR-1 (10 mg/ml). Whole mount fixation Mammary glands from selected mice were dissected out and fixed on glass slides with Carnoy’s solution (glacial acetic

Characterization of tumor samples Tumor samples were either frozen on dry ice or fixed with paraformaldehyde as described in Materials and methods. Histopathology and immunohistochemistry were assessed by examination of hematoxylin and eosin-stained tumor sections. Acknowledgements We thank Dr Lionel Feigenbaum (SAIC Frederick) for generating the MMTV-CR-1 transgenic FVB/N mice and Amanda Kalmanowicz and Heidi Olesch from the Building 28 mouse facility at NIH for excellent animal care. We also thank Dr Humphrey Gardner of the Biogen Idec Department of Research Pathology for review of the MMTV/CR-1 histology slides. We gratefully acknowledge the generosity of Dr Yi Li (Baylor College of Medicine) for providing slides of the MMTV/Wnt-1 mammary tumors. This work was supported in part by funding from the Associazione Italiana per la Ricerca sul Cancro (AIRC) to Nicola Normanno.

References Adamson ED, Minchiotti G and Salomon DS. (2002). J. Cell. Physiol., 190, 267–278. Adkins HB, Bianco C, Schiffer SG, Rayhorn P, Zafari M, Cheung AE, Orozco O, Olson D, De Luca A, Chen LL, Miatkowski K, Benjamin C, Normanno N, Williams KP, Jarpe M, LePage D, Salomon D and Sanicola M. (2003). J. Clin. Invest., 112, 575–587. Bhattacharya B, Miura, Brandenberger R, Mejido, Luo Y, Yang AX, Joshi BH, Ginis I, Thies RS, Amit M, Lyons, Condie BG, Itskovitz-Eldor J, Rao MS and Puri RK. (2004). Blood, 103, 2956–2964. Bianco C, Adkins HB, Wechselberger C, Seno M, Normanno N, De Luca A, Sun Y, Khan N, Kenney N, Ebert A, Williams KP, Sanicola M and Salomon DS. (2002). Mol. Cell. Biol., 22, 2586–2597.

Bianco C, Strizzi L, Rehman A, Normanno N, Wechselberger C, Sun Y, Khan N, Hirota M, Adkins H, Williams K, Margolis RU, Sanicola M and Salomon DS. (2003). Cancer Res., 63, 1192–1197. Botchan M, Topp W and Sambrook J. (1976). Cell, 9, 269–287. Brisken C, Heineman A, Chavarria T, Elenbaas B, Tan J, Dey SK, McMahon JA, McMahon AP and Weinberg RA. (2000). Genes Dev., 14, 650–654. Cardiff RD and Munn RJ. (1995). Cancer Lett., 90, 13–19. Carpenter MK, Rosler ES, Fisk GJ, Brandenberger R, Ares X, Miura T, Lucero T and Rao MS. (2004). Develop. Dyn., 229, 243–258. Cheng SK, Olale F, Bennett JT, Brivanlou AH and Schier AF. (2003). Genes Dev., 17, 31–36. Oncogene


4104 Ciardiello F, Dono R, Kim N, Persico MG and Salomon DS. (1991). Cancer Res., 51, 1051–1054. Ciardiello F, Tortora G, Bianco C, Selvam MP, Basolo F, Fontanini G, Pacifico F, Normanno N, Brandt R and Persico MG. (1994). Oncogene, 9, 291–298. Ciccodicola A, Dono R, Obici S, Simeone A, Zollo M and Persico MG. (1989). EMBO J., 8, 1987–1991. Colas JF and Schoenwolf GC. (2000). Gene, 255, 205–217. Delcommenne M, Tan C, Gray V, Rue L, Woodgett J and Dedhar S. (1998). Proc. Natl. Acad. Sci. USA, 95, 11211–11216. De Luca A, Casamassimi A, Selvam MP, Losito S, Ciardiello F, Agrawal S, Salomon DS and Normanno N. (1999). Int. J. Cancer, 80, 589–594. De Luca A, Selvam MP, Sandomenico C, Pepe S, Bianco AR, Ciardiello F, Salomon DS and Normanno N. (1997). Int. J. Cancer, 73, 277–282. Ding J, Yang L, Yan YT, Chen A, Desai N, Wynshaw-Boris A and Shen MM. (1998). Nature, 395, 702–707. Donehower LA, Huang AL and Hager GL. (1981). J. Virol., 37, 226–238. Dono R, Scalera L, Pacifico F, Acampora D, Persico MG and Simeone A. (1993). Development, 118, 1157–1168. Ebert AD, Wechselberger C, Nees M, Clair T, Schaller G, Martinez-Lacaci I, Wallace-Jones B, Bianco C, Weitzel HK and Salomon DS. (2000). Exp. Cell Res., 257, 223–229. Feinberg AP and Vogelstein B. (1984). Anal. Biochem., 137, 266–267. Gray PC, Harrison and Vale V. (2003). Proc. Natl. Acad. Sci. USA, 100, 5193–5198. Harms PW and Chang C. (2003). Genes Dev., 17, 2624–2629. Herrington EE, Ram TG, Salomon DS, Johnson GR, Gullick WJ, Kenney N and Hosick HL. (1997). J. Cell Physiol., 170, 47–56. Howe LR and Brown AM. (2004). Cancer Biol. Ther., 3, 36–41. Huang AL, Ostrowski MC, Berard D and Hager GL. (1981). Cell, 27, 245–255. Imbert A, Eelkema R, Jordan S, Feiner H and Cowin P. (2001). J. Cell Biol., 153, 555–568. Johnson MD, Kenney NJ, Stoica A, Hilakivi-Clarke L, Singh B, Chepko G, Clarke R, Sholler PF, Lirio AA, Foss C, Reiter R, Trock B, Paik S and Martin MB. (2003). Nature Med., 9, 1081–1084. Kenney NJ, Huang RP, Johnson GR, Wu JX, Okamura D, Matheny W, Kordon E, Gullick WJ, Plowman G and Smith GH. (1995). Mol. Reprod. Dev., 41, 277–286. Kenney NJ, Smith GH, Johnson MD, Rosenberg K, Salomon DS and Dickson RB. (1997). Pathogenesis, 1, 57–71. Kenney NJ, Smith GH, Maroulakou IG, Green JH, Muller WJ, Callahan R, Salomon DS and Dickson RB. (1996). Mol. Carcinog., 15, 44–56. Kinoshita N, Minshull J and Kirschner MW. (1995). Cell, 83, 621–630. Laird PW, Zijderveld A, Linders K, Rudnicki MA, Jaenisch R and Berns A. (1991). Nucleic Acids Res., 19, 4293. Landesman-Bollag E, Romieu-Mourez R, Song DH, Sonenshein GE, Cardiff RD and Seldin DC. (2001). Oncogene, 20, 3247–3257. Li Y, Welm B, Podsypanina K, Huang S, Chamorro M, Zhang X, Rowlands T, Egeblad M, Cowin P, Werb Z, Tan LK, Rosen JM and Varmus HE. (2003). Proc. Natl. Acad. Sci. USA, 100, 15853–15858. Liu BY, McDermott SP, Khwaja SS and Alexander CM. (2004). Proc. Natl. Acad. Sci. USA, 101, 4158–4163. Oncogene

Morkel M, Huelsken J, Wakamiya M, Ding J, van de Wetering M, Clevers H, Taketo MM, Behringer RR, Shen MM and Birchmeier W. (2003). Development, 130, 6283–6294. Naylor MJ and Ormandy CJ. (2002). Dev. Dynam., 225, 100–105. Niemeyer CC, Persico MG and Adamson ED. (1998). Cell Death Differ., 5, 440–449. Niemeyer CC, Spencer-Dene B, Wu JX and Adamson ED. (1999). Int. J. Cancer, 81, 588–591. Normanno N, Bianco C, Damiano V, de Angelis E, Selvam MP, Grassi M, Magliulo G, Tortora G, Bianco AR, Mendelsohn J, Salomon DS and Ciardiello F. (1996). Clin. Cancer Res., 2, 601–609. Normanno N, Kim N, Wen D, Smith K, Harris AL, Plowman G, Colletta G, Ciardiello F and Salomon DS. (1995). Breast Cancer Res. Treat., 35, 293–297. Normanno N, Tortora G, De Luca A, Pomatico G, Casamassimi A, Agrawal S, Mendelsohn J, Bianco AR and Ciardiello F. (1999). Oncol. Rep., 6, 1105–1109. Oloumi A, McPhee and Dedhar S. (2004). Biochim. Biophys. Acta., 1691, 1–15. Panico L, D’Antonio A, Salvatore G, Mezza E, Tortora G, De Laurentiis M, De Placido S, Giordano T, Merino M, Salomon DS, Mullick WJ, Pettinato G, Schnitt SJ, Bianco AR and Ciardiello F. (1996). Int. J. Cancer, 65, 51–56. Pollard JW. (2001). Breast Cancer Res., 3, 320–337. Qi CF, Liscia DS, Normanno N, Merlo G, Johnson GR, Gullick WJ, Ciardiello F, Saeki T, Brandt R and Kim N. (1994). Br. J. Cancer, 69, 903–910. Robinson GW and Henninghausen L. (1997). Development, 124, 2701–2708. Robinson GW, Henninghausen L and Johnson PJ. (2000). Genes Dev., 14, 889–894. Rosner A, Miyoshi K, Landesman-Bollag E, Xu X, Seldin DC, Moser AR, MacLeod CL, Shyamala G, Gillgrass AE and Cardiff RD. (2002). Am. J. Path., 161, 1087–1097. Salomon DS, Bianco C and De Santis M. (1999). BioEssays, 21, 61–70. Salomon DS, Bianco C, Ebert AD, Khan NI, De Santis M, Normanno N, Wechselberger C, Seno M, Williams K, Sanicola M, Foley S, Gullick WJ and Persico G. (2000). Endocr. Relat. Cancer, 7, 199–226. Schier AF and Shen MM. (2000). Nature, 403, 385–389. Shen MM and Schier AF. (2000). Trends Genet., 16, 303–309. Shen MM, Wang H and Leder P. (1997). Development, 124, 429–442. Silberstein GB. (2001). Breast Cancer Res., 3, 218–223. Soriano, Pepper MS, Orci L and Montesano R. (1998). J. Mammary Gland Biol. Neoplasia, 3, 133–150. Srinivasan R, Gillett CE, Barnes DM and Gullick WJ. (2000). Cancer Res, 60, 1483–1487. Strizzi L, Bianco C, Normanno N, Seno M, Wechselberger C, Wallace-Jones B, Khan NI, Hirota M, Sun Y, Sanicola M and Salomon DS. (2004). J. Cell. Physiol., 201, 266–276. Uyttendale H, Sorian JV, Montesano R and Kitajewski J. (1998). Dev. Biol., 196, 204–217. Webster MA, Martin-Soudant N, Nepveu A, Cardiff RD and Muller WJ. (1998). Oncogene, 16, 1963–1972. Wechselberger C, Ebert AD, Bianco C, Khan NI, Sun Y, Wallace-Jones B, Montesano R and Salomon DS. (2001). Exp. Cell Res., 266, 95–105. Wechselberger C, Kreil G and Richter K. (1992). Proc. Natl. Acad. Sci. USA, 89, 9819–9822. White DE, Cardiff RD, Dedhar S and Muller WJ. (2001). Oncogene, 20, 7064–7072.


4105 Xing PX, Hu XF, Hosick HL and McKenzie IFC. (2004). Cancer Res., 64, 4018–4023. Xu C, Liguori G, Persico MG and Adamson ED. (1999). Development, 126, 483–494. Yan YT, Liu JJ, Luo Y, E C, Haltiwanger RS, Abate-Shen C and Shen MM. (2002). Mol. Cell. Biol., 22, 4439–4449.

Yeo C and Whitman M. (2001). Mol. Cell, 7, 949–957. Yokota C, Kofron M, Zuck M, Houston DW, Isaacs H, Asashima M, Wylie CC and Heasman J. (2003). Development, 130, 2199–2212. Zhang J, Talbot WS and Schier AF. (1998). Cell, 92, 241–251.

Oncogene