Increased susceptibility to carcinogen-induced mammary tumors inMMTV-Cdc25B transgenic mice

Yao Yao1,2,7, Eric D Slosberg1,3,7, Lei Wang2, Hanina Hibshoosh4, Yu-Jing Zhang5,Wang-Qiu Xing1, Regina M Santella1,5 and I Bernard Weinstein*,1,3,5,6

1Herbert Irving Comprehensive Cancer Center, College of Physicians and Surgeons, Columbia University, 701 West 168th Street,New York, NY 10032, USA; 2Integrated Program in CMBS, College of Physicians and Surgeons, Columbia University, 701 West168th Street, New York, NY 10032, USA; 3Department of Genetics, College of Physicians and Surgeons, Columbia University, 701West 168th Street, New York, NY 10032, USA; 4Department of Pathology, College of Physicians and Surgeons, ColumbiaUniversity, 701 West 168th Street, New York, NY 10032, USA; 5School of Public Health, Division of Environmental HealthSciences, College of Physicians and Surgeons, Columbia University, 701 West 168th Street, New York, NY 10032, USA;6Department of Medicine, College of Physicians and Surgeons, Columbia University, 701 West 168th Street, New York,NY 10032, USA

Cdc25 phosphatases activate cyclin-dependent kinases(Cdks) by dephosphorylating critical phospho-tyrosineand phospho-threonine residues on these proteins. Severaltypes of studies indicate that Cdc25s can enhance cellproliferation and oncogenesis. Furthermore, overexpres-sion of Cdc25A and/or B have been detected in severaltypes of primary human cancers, including breastcancers. To further assess the oncogenic capacity ofCdc25B in vivo, we have generated transgenic mice thatoverexpress Cdc25B in the mammary epithelium, drivenby the MMTV±LTR promoter. Although these mice aregrossly normal for up to 18 months, the ectopicexpression of Cdc25B in their mammary glands increasesthe susceptibility of these mice to induction of mammarytumors by the carcinogen 9,10-dimethyl-1,2-benzanthra-cene (DMBA).

Keywords: Cdc25B; transgenic mice; DMBA; synergy;mammary tumor; MMTV-LTR

Introduction

In eukaryotic cells, cell-cycle progression is controlledby the ordered activation of a family of cyclin-dependent kinases (Cdks). These proteins are sub-jected to both positive and negative regulations, atmultiple levels. Cdks are activated by the binding ofregulatory subunits known as cyclins, and phosphor-ylation by the Cdk-activating kinase (CAK). Cyclin/Cdk complexes are also negatively regulated by at leasttwo di�erent mechanisms: binding of Cdk inhibitorsbelonging to the INK family (p15, p16, p18 and p19)or the CIP family (p21, p27 and p57), and phosphor-ylation of conserved threonine and tyrosine residues,usually at positions 14 and 15 in the Cdks (Sherr,1993). Cdc25s are a family of related dual-speci®cphosphatases that catalyze the dephosphorylation of

both phospho-Thr-14 and phospho-Tyr-15 in the Cdks(Morgan, 1995).

Cdc25 was originally discovered as a lethal allelethat arrests Schizosaccharomyces pombe in the G2phase and was then cloned by functional complementa-tion (Russell and Nurse, 1986). It was shown to be adose-dependent mitotic inducer (Russell and Nurse,1986). In humans, there are at least three Cdc25 familymembers (Cdc25A, B and C) which share 40 ± 50%homology at the amino acid level, and are thought toplay unique roles in cell cycle progression (Draetta andEckstein, 1997; Ho�mann and Karsenti, 1994).Cdc25A is expressed at late G1 and the microinjectionof antibodies to Cdc25A into mammalian cells preventsentry into the S phase (Ho�mann et al., 1994; Jinno etal., 1994). Cdc25B and Cdc25C are expressed at lowlevels throughout the cell cycle with a sharp increase inexpression and activity at the G2 and M phases(Gabrielli et al., 1996; Nagata et al., 1991; Sadhu etal., 1990). The function of Cdc25C is most closelyrelated to that of the S. pombe homologue in regulatingthe G2-M transition. Cdc25B expression overlaps thatof Cdc25C and it is thought to play an important rolein regulating microtubule nucleation at the centro-somes, which is essential for the formation of themitotic spindle in prophase (Gabrielli et al., 1996).Consistent with these functions is the ®nding thatoverexpression of Cdc25B results in accumulation ofcells in the G2/M phase and the formation ofminispindles in the cytoplasm (Gabrielli et al., 1996).In addition, Cdc25 may also provide a link betweenmitogenic signal transduction (Clarke et al., 1993;Galaktionov et al., 1995a; Grieco et al., 1994; Kumagaiand Dunphy, 1996) and DNA damage pathways to thecell cycle machinery (Furnari et al., 1997; Ogg et al.,1994; Peng et al., 1997; 1998; Sanchez et al., 1997).

Several lines of evidence suggest that Cdc25s canenhance growth and oncogenesis in mammalian cells.Cdc25A and Cdc25B can interact and be activated bythe Raf1 kinase (Galaktionov et al., 1995a), and theycan transform rodent ®broblasts in cooperation withan activated Ha-ras gene or loss of the Rb gene(Galaktionov et al., 1995b). The transforming virusesSV40 and adenovirus resulted in increased Cdc25B andCdc25A mRNA levels as well as their phosphataseactivities (Nagata et al., 1991; Spitkovsky et al., 1996).Down-regulation of Cdc25A appears to mediate TGF-

*Correspondence: IB Weinstein, Department of Medicine, College ofPhysicians and Surgeons, Columbia University, 701 West 168thStreet, New York, NY 10032, USA7The ®rst two authors have contributed equally to the present workReceived 29 September 1998; revised 14 April 1999; accepted 14 April1999

Oncogene (1999) 18, 5159 ± 5166ã 1999 Stockton Press All rights reserved 0950 ± 9232/99 $15.00

http://www.stockton-press.co.uk/onc

b-induced cell cycle arrest in mammary epithelial cells(Iavarone and Massague, 1997). Furthermore, over-expression of Cdc25A and/or Cdc25B have beendetected in some cancer cell lines (Nagata et al.,1991) and in several types of human cancers, includinggastric cancer, squamous cell carcinomas of the headand neck (Gasparotto et al., 1997), and non-small celllung carcinomas (NSCLC) (Wu et al., 1998). Inaddition, Cdc25B is overexpressed in over 32% ofprimary human breast cancers, with a strong correla-tion between its overexpression and negative prognosticfeatures in these cases (Galaktionov et al., 1995b).

To assess the in vivo oncogenic potential of Cdc25Bin the mammary gland, we generated transgenic micethat express the human Cdc25B gene under the controlof a Mouse Mammary Tumor Virus (MMTV) LTR.These mice do not develop spontaneous mammarytumors but they are highly susceptible to mammarytumor formation induced by the carcinogen DMBA.These ®ndings provide evidence for an oncogenic roleof Cdc25B in the mammary epithelium.

Results

Generation of MMTV-Cdc25B founder mice

To assess the potential role of Cdc25B as an oncogenein mammary epithelium, we examined the conse-quences of overexpression of Cdc25B in the mammarygland of transgenic mice. The human Cdc25B cDNAwas cloned into an expression plasmid containing themouse mammary tumor virus long terminal repeat(MMTV±LTR), a 5'-arti®cial intron, a 3'-intron fromthe SV40 large T sequence and polyadenylation (polyA) signals (Figure 1). The S®I-linearized fragment ofthe MMTV-Cdc25B construct was microinjected intofertilized mouse oocytes obtained from B6/CBA F1

donor mice and eight transgenic founder mice wereidenti®ed by Southern blot analyses of tail DNAsamples (Figure 2). Founder mice 554, 558, 561, 562,565 and 566 were back-crossed with C57BL6 mice (TheJackson Laboratory) for successive generations andtheir o�spring were used for further studies.

Expression of ectopic Cd25B in mouse mammary glands

To examine the expression of ectopic Cdc25B in mousetissues, Northern and Western blot analyses wereperformed (Figure 3). Northern blot analysis revealedthe expected *2.5 kb transgenic mRNA in both themammary glands of non-lactating transgenic andlactating transgenic mice from ®ve di�erent lines(Figure 3a), although the levels of Cdc25B mRNAvaried from line to line. Thus, lines 554 and 558expressed much higher levels than lines 561, 565 and566. However, none of the mammary glands from the

Figure 1 A schematic diagram of the pMMTV-Cdc25B plasmid. The human Cdc25B2 cDNA was cloned downstream of theMMTV±LTR and was ¯anked by a 5'-arti®cial intron (intron 1) and a 3' SV40 large T intron (intron 2) plus the polyA sequences.The MMTV-intron 1-Cdc25B2-intron 2-polyA cassette was released from the plasmid DNA by S®I digestion. The diagnosticBamHI restriction enzyme site is shown

Figure 2 Screen of MMTV-Cdc25B transgenic founder mice bySouthern blot analyses of tail DNA. DNA was prepared from thetails of non-transgenic (C) and transgenic mice from six founderlines (554, 558, 561, 562, 565 and 566) and examined by Southernblot analyses. A 32P-labelled SV40 intron-polyA fragment wasused as the probe to identify mice carrying the transgene

Synergy between Cdc25B and DMBA in inducing mammary tumorsY Yao et al

5160

non-lactating or lactating control mice expresseddetectable levels of Cdc25B mRNA (lanes 1 and 7).Endogenous murine Cdc25B transcripts were notdetected in the mRNAs isolated from the mammaryglands of either the transgenic (Tg) or control mice,perhaps due to sequence variations between the humanCd25B2 cDNA probe and the endogenous mouseCdc25B mRNA sequences (only 78% identity) andthe high stringency washing conditions used. We alsoobserved that the ectopic Cdc25B mRNA wasexpressed at higher levels in the lactating than thenon-lactating mammary glands. In addition, thetransgene was expressed in the salivary glands andtestes, but not the liver, of Tg mice (data not shown).These were expected since the MMTV promoter ishighly-inducible by lactogenic hormones (Henrard andRoss, 1988; Rollini et al., 1992) and is active in somenon-mammary tissues, including various lymphoidtissues, the testes, salivary gland, prostate gland,Harderian gland and the skin (Amundadottir et al.,1995; Cardi� and Muller, 1993; Halter et al., 1992;Hennighausen et al., 1995; Henrard and Ross, 1988;Rollini et al., 1992; Ross et al., 1990).

To con®rm that the transgenic Cdc25B transcriptsresulted in elevated levels of the Cdc25B protein, weperformed Western blot analyses using proteinsprepared from lactating and non-lactating mammaryglands of transgenic and non-transgenic mice (Figure3b). A strong Cdc25B protein band (*70 kDa) wasfound in the mammary glands of the lactating Tg micefrom all the ®ve lines examined (lanes 8 ± 12), but thisprotein was not detected in either the mammary glands

of non-lactating (lane 1) or lactating control mice (lane7). Therefore, we were successful in obtaining a highlevel of expression of the Cdc25B protein in themammary glands of the lactating Tg mice. However,we did not detect the exogenous Cdc25B protein (lanes2 ± 6) in the mammary glands of the non-lactating Tgmice, even though the corresponding exogenousmRNA was readily detected by Northern blot analysis(Figure 3a, lanes 2, 3 and 6). This may re¯ect thegreater sensitivity of the Northern blots, or expressionof the exogenous Cdc25B protein might be stimulatedby a post-translational mechanism during lactation.The high expression of the transgenic Cdc25B mRNAduring lactation is consistent with the hormone-responsive nature of the MMTV promoter (Henrardand Ross, 1988; Rollini et al., 1992).

Immunohistochemical staining of mammary glandswith an antibody speci®c for Cdc25B was alsoperformed (Figure 4). Very weak and sparse Cdc25Bexpression can be seen in both non-lactating (Figure4a) and lactating (Figure 4c) mammary glands of wild-type mice. This highly sensitive technique was able todetect levels of Cdc25B that were not detectable byWestern blot analysis (Figure 3b). Mammary glandsfrom Tg mice show strong staining for Cdc25B (Figure4b and d), particularly upon lactation (Figure 4d). Asexpected Cdc25B is localized within the nucleus ofexpressing cells.

Lack of spontaneous tumors in Cdc25B transgenic mice

The MMTV-Cdc25B transgenic mice were viable andfertile and displayed no obvious abnormalities. Femalemice were examined by palpatation every month for

a

b

Figure 3 (a) Expression of exogenous Cdc25B mRNA intransgenic mice. RNAs were prepared from non-lactating (NL,lanes 1 ± 6) and lactating (L, lanes 7 ± 12) mammary glands of ®vetransgenic mouse lines (554, 558, 561, 565 and 566, lanes 2 ± 6 and8 ± 12) and control mice (lanes 1 and 7) by the GTC method (seeMaterials and methods). Ten mg of each RNA sample wasresolved on a 1% agarose-formaldehyde gel. 32P-labeled humanCdc25B2 cDNA was used as the probe in the Northern blothybridization (upper panel). As a loading control, the ethidiumbromide staining signal of the 28S ribosomal RNA was alsoincluded (lower panel). (b) Expression of the Cdc25B protein inthe mammary glands of transgenic (554, 558, 561, 565 and 566)and control (c) mice. Protein lysates from mammary glands ofnon-lactating (NL, lanes 1 ± 6) and lactating (L, lanes 7 ± 12) micewere prepared and examined by Western blot using an antibodyspeci®c for Cdc25B (upper panel) or actin (as a loading control,lower panel). Molecular masses are indicated on the left

Figure 4 Representative histological sections demonstratingCdc25B immunostaining. Mammary gland sections from non-lactating (a and b) and lactating (c and d) wild-type (a and c) andtransgenic (b and d) mice were stained with an antibody speci®cfor Cdc25B. The immunostaining used the avidin-biotin complextechnique and 3,3'-diaminobenzidine development (brown). Cellswere lightly counterstained with harris hematoxylin and photo-graphed at a magni®cation of 4006. Note the strong Cdc25Bstaining in lactating transgenic mice. The arrows point to cellsshowing clear nuclear expression of Cdc25B. Also note theabsence of hyperplasia or other abnormalities in the mammaryglands of the transgenic mice

Synergy between Cdc25B and DMBA in inducing mammary tumorsY Yao et al

5161

tumor development. Among the 98 female Tg miceobtained from the ®ve original founder lines that werefollowed until they were over 1 year of age (medianage=14 months), no mammary tumors were detected.A thorough histologic analysis of both lactating andnon-lactating mammary glands from the Tg mice(n=24) did not show any evidence of hyperplasia orother abnormalities (Figure 4 and data not shown).Since the expression of the exogenous Cdc25B mRNAwas also detected in the testis and salivary gland of Tgmice (data not shown), these organs and the othermajor organs and tissues (including spleen, skin, heart,liver, kidney and brain) from both male and female Tgmice were examined by gross pathology and byhistology. However, no abnormalities were found inthese tissues (data not shown). Among the 180 Tg micethat were over 1 year of age, ®ve developed lymphoma.Since C57BL6 mice can develop spontaneous lympho-mas (Stutman, 1975; Turusov, 1994), we are not certainthat the lymphomas that arose in the transgenic micewere related to the Cdc25B transgene.

Synergy between Cdc25B and DMBA in mammarytumor development

The above studies indicated that overexpression ofCdc25B in the mammary gland of these mice was notby itself su�cient to cause the development ofmammary tumors. To investigate the potential role ofCdc25B in enhancing mammary tumor development in

combination with other carcinogenic factors weadministered the chemical carcinogen DMBA (7,12-dimethyl-1,2-benzanthracene) to these mice, during thetime of ductal development in the mammary gland(Medina et al., 1980). Virgin female Tg mice from fourfounder lines (n=43) and control mice (n=34) fromfour lines were given 1 mg doses of DMBA by oralgavage for ®ve consecutive weeks beginning at 8 weeksof age. The mice were then examined biweekly fortumors, by palpatation. Within 250 days afteradministration of the DMBA mammary tumors arosein six of the 29 (20.7%) surviving non-transgeniccontrol mice (Figure 5). These data are consistent withthose reported by others who used a similar DMBAtreatment protocol in wild-type mice (Aldaz et al.,1996; Pierce et al., 1995; Witty et al., 1995). It is ofinterest that the number of Cdc25B Tg mice thatdeveloped mammary tumors during the same timeperiod was 17 of the 39 (43.6%) survivors. Thisfrequency was signi®cantly higher than that in thecontrol animals (P=0.0493, t-test). Although themedian latency period for tumor development of bothgroups was not signi®cantly di�erent (136 vs 155 days,P=0.4149), mammary tumors were ®rst detected in theDMBA-treated Tg mice within 50 ± 60 days but in theDMBA-treated control mice they were not detecteduntil 120 days. The increased incidence of tumorformation occurred in all the four transgenic linestested, providing evidence that this phenotype is notdue to random mutations in endogenous genes arising

Figure 5 Susceptibility of the MMTV-Cdc25B transgenic mice to DMBA-induced mammary tumor formation. This ®gure displaysa Kaplan ±Meier analysis of the time to development of palpable mammary tumors in virgin MMTV-Cdc25B transgenic femalesand their wild-type littermates, following DMBA treatment. DMBA was administered for ®ve consecutive weeks, beginning at 8weeks of age. The solid blue curve indicates the transgenic group and the solid red curve indicates the wild-type control group.Individual transgene sublines (554Tg, 558Tg, 561Tg and 565Tg) are plotted as dashed blue lines. The number of mice in each groupare indicated on the right

Synergy between Cdc25B and DMBA in inducing mammary tumorsY Yao et al

5162

from insertion of the transgene into the mouse genome.There were nine deaths amongst all of the 77 miceinitially treated with DMBA, four in the Tg group and®ve in the control group. These mice had noidenti®able tumors and their death was probablyrelated to the DMBA treatment, since this compoundhas been previously shown to cause death due totoxicity to various organs (Aldaz et al., 1996;Archuleta et al., 1992).

The tumors that were found in the control and Tgmice were remarkably similar, since almost all of theanimals developed malignant adenoacanthomas of themammary glands. This type of tumor is characterizedby the presence of squamous di�erentiation with focalareas of keratinization (horny pearls) (Figure 6a andb). Most of these adenoacanthomas were well-di�erentiated and only two cases were classi®ed as`mid-stage' since they displayed invasion of theadjacent skin tissue (Figure 6c). Furthermore, nearlyall of the cases of tumor formation involved only asingle tumor per animal, with no noticeable hyperplasiain the non-a�ected mammary glands (Figure 6d),suggesting a weak oncogenic role for Cdc25B in themouse mammary epithelium (see Discussion).

Discussion

In this paper, we have described the generation andanalysis of mice containing a MMTV-Cdc25B trans-gene. Five independent founder lines of Tg miceexpressed the exogenous Cdc25B mRNA in the non-lactating and lactating mammary glands (Figure 3a).Production of the exogenous Cdc25B protein was also

increased in the lactating mammary glands of the Tgmice (Figures 3b and 4d). The increased levels ofCdc25B mRNA and protein in the mammary glandsduring lactation is consistent with the known respon-siveness of the MMTV promoter to lactogenichormones (Henrard and Ross, 1988; Rollini et al.,1992), and also to results obtained in mice carryingtransgenes controlled by similar promoters (Bortnerand Rosenberg, 1995; 1997; Santarelli et al., 1996). Asmentioned in the Introduction, several types ofevidence support an oncogenic role for Cdc25B,including its ability to transform ®broblasts incooperation with an activated oncogene or the loss ofa tumor suppressor gene, and its overexpression inseveral types of human cancer, including breast cancer(Galaktionov et al., 1995b). Surprisingly our transgenicmice did not develop spontaneous mammary tumors,nor did they show any evidence of hyperplasia ordysplasia in the mammary gland (Figure 4) or in othertissues that overexpress Cdc25B. However, afteradministering the chemical carcinogen DMBA thetransgenic mice displayed an increased frequency indeveloping mammary tumors, when compared inparallel to control mice that were also given DMBA(Figure 5). Therefore, increased expression of Cdc25Bcan markedly enhance the development of DMBA-induced mammary tumors.

There are several possible explanations for theabsence of spontaneous mammary tumors in theuntreated transgenic mice that overexpress Cdc25B.(1) It is widely believed that oncogenes need tocooperate with each other and/or the loss of tumorsuppressors to achieve a fully transformed phenotype.This is consistent with the theory of multi-stagecarcinogenesis (Weinberg, 1989). Some oncogenes,however, are highly potent and appear to be capableof inducing tumor formation on their own (i.e.MMTV-PyV middle T and MMTV-activated-neutransgenic mice (Guy et al., 1992; Muller et al,1988)). No or very few secondary mutations or `hits'may be required in these systems to induce tumorformation. Alternatively, these oncogenes may causegenomic instability, thus enhancing tumor progression.However, the majority of oncogenes that are over-expressed as transgenes in the mouse mammary glandinduce mammary tumors in a stochastic fashion andwith lengthy latency periods (Cardi� and Muller,1993). Thus, in MMTV-c-Myc (Stewart et al., 1984)and MMTV-TGF-a (Matsui et al., 1990) transgenicmice, development of tumors requires extensive cellproliferation and multiple pregnancies. Thus Cdc25Bmay be a weak oncogene that requires one, or more,secondary `hits', which do not occur during the normalrelatively short lifespan of a mouse, to contribute totumorigenesis. (2) The MMTV promoter is active inmammary glands of mice in a hormone-dependentmanner (Beato, 1991; Bruggemeier et al., 1991; Truss etal., 1992). Thus it is active at basal levels duringnormal growth, but is abruptly and markedly activatedduring lactation. This is also true in our MMTV-Cdc25B mice, since only during lactation did we detecthigh levels of Cdc25B protein and mRNA in themammary gland (see Figure 3). We mated the femaletransgenic mice as often as possible to provide thehighest levels of Cdc25B overexpression, but evenmultiparous mice have active MMTV expression for

Figure 6 Histological analysis of DMBA-induced tumors inMMTV-Cdc25B mice. Photomicrographs of hematoxylin andeosin stained normal mammary glands or DMBA-inducedmammary tumors (6200). Adenoacanthomas of the molluscoidtype were observed in the mammary tumors from transgenic lines554 (a) and 565 (b). A tumor invasion into adjacent skeletalmuscle (d) from transgenic line 554 was also shown (c). Themammary tumors obtained from the non-transgenic mice had asimilar histology (not shown). A normal virgin mammary gland(d) is shown for comparison

Synergy between Cdc25B and DMBA in inducing mammary tumorsY Yao et al

5163

only a fraction of their lifespan. Perhaps the absence ofsustained high levels of Cdc25B limits the oncogenice�ect of Cdc25B in our transgenic mice. In futurestudies it will be of interest to use ectopic hormonaltreatments that mimics a lactating state (Li et al.,1995), and thus constitutively activate the MMTVpromoter, to see if under these conditions the Cdc25Btransgene can induce mammary tumors. (3) TheMMTV-Cdc25B transgenic mice were maintained in aC57BL6 genetic background. This background ishighly resistant to spontaneous mammary tumorformation (Turusov, 1994) and was used in an attemptto reduce background levels of tumor formation innon-transgenic control mice. In view of the absence ofspontaneous mammary tumor formation in ourtransgenic mice, we are in the process of transferringthe MMTV-Cdc25B transgene into a FVB geneticbackground, one that is less resistant to tumorformation and is also commonly used in manytransgenic models of breast cancer (Cardi� andMuller, 1993; Turusov, 1994).

During the course of our studies other investigatorsdiscovered that Cdc25B is alternatively spliced toproduce three di�erent isoforms (Baldin et al.,1997a,b; Gabrielli et al., 1997). The most abundantlyexpressed isoform, Cdc25B3, produces a protein with 14residues (aa 68 ± 81) that are not found in the originalCdc25B cDNA (Cdc25B1). A second isoform, Cdc25B2,which was used in the construction of our transgenicmice, is expressed at lower levels and contains the 42base pair insert but has a 123 b.p. (41 residue) in-framedeletion further downstream (aa 154 ± 194). Thesevariants are expressed at di�erent levels in di�erentcell lines, but Cdc25B3 is the most abundantly expressedform, while Cdc25B2 and Cdc25B1 are expressed at lowlevels or not at all, respectively (Baldin et al., 1997b;Gabrielli et al, 1997). Functionally these splice variantsdi�er, having di�erent in vitro a�nities for di�erentcyclin-Cdk complexes and di�erent e�ciencies asmitotic inducers in S. pombe (Baldin et al., 1997b).The di�erences in molecular weights and transcript sizesbetween the three variants are small, making it di�cultto determine which endogenous Cdc25B variants wereexamined in previous studies. In particular, althoughseveral studies have found elevated levels of Cdc25BmRNAs in cancers (Galaktionov et al., 1995b;Gasparotto et al., 1997; Kudo et al., 1997; Wu et al.,1998), these studies do not distinguish among splicevariants, leaving open the possibility that the di�erentsplice variants may have markedly di�erent abilities totransform tissues. To construct MMTV-Cdc25B micewe used the Cdc25B2 cDNA splice variant, the ®rstvariant cloned, and one that has been previously used toprovide evidence for the transforming ability of Cdc25B(Galaktionov et al., 1995a,b; 1996). Thus, it is possiblethat transgenic mice made with the other splice variantsmight have a di�erent e�ect on the development ofspontaneous or DMBA-induced mammary tumors.

The DMBA-induced mammary tumors, in all of theMMTV-Cdc25B mice and the majority of the tumorsin the wild-type mice, were adenoacanthomas. Humanbreast cancers are rarely classi®ed as adenoacanthomas(Hanina Hibshoosh, personal communication), how-ever this type is common in DMBA-induced mammarytumors in mice (Medina et al., 1980; Witty et al., 1995).In addition, the same type of mammary tumors form

spontaneously in transgenic mice bearing a MMTV-cyclin D1 transgene (Nielsen et al., 1991; Wang et al.,1994). It is di�cult to know how much signi®canceshould be given to the speci®c type of tumorpathology. It is, however, of interest that DMBA-induced tumors are frequently associated with muta-tional activation of the Ha-ras oncogene (Aldaz et al.,1996; Dandekar et al., 1986), and Ha-ras transgenicmice also develop spontaneous mammary adenoa-canthomas (Nielsen et al., 1991). Furthermore, it hasbeen shown that the Ras-Raf pathway can activateCdc25B activity (Galaktionov et al., 1995a), suggestinga possible scenario where Cdc25B lays downstream ofmutant Ras on a pathway that favors the developmentof adenoacanthoma-type tumors.

The ability of cells to arrest cell cycle progression inresponse to DNA damage is essential for themaintenance of genomic integrity, allowing for therepair of damaged DNA or, in the case of severedamage, for the activation of programmed cell death(Hartwell and Weinert, 1989). Following DNAdamage, Cdc25 is phosphorylated on Ser-216 by thecheckpoint kinase CHK1 (Furnari et al., 1997; Ogg etal., 1994; Peng et al., 1997; 1998; Sanchez et al., 1997),and Cdc25 (S216A) mutants, which cannot bephosphorylated in response to DNA damage nor bindthe 14-3-3 protein, perturb mitotic timing and allowcells to escape the G2 checkpoint arrest induced byunreplicated DNA or radiation-induced damage (Penget al., 1997). These data suggest that Cdc25 is a centralcomponent in the cellular response to DNA damage.Cdc25B was also found to have a role in formation ofthe mitotic spindle (Gabrielli et al., 1996). Therefore,the increased susceptibility to DMBA-induced tumorformation in our MMTV-Cdc25B mice may be due toeither chromosomal loss or gain, or inappropriatemitosis in the presence of unrepaired DNA damage,since DMBA forms DNA adducts, thereby inducingDNA damage. We are currently exploring thishypothesis.

Taken together, the results described in this paperprovide the ®rst in vivo evidence that Cdc25B can beoncogenic. Although mice that harbor the MMTV-Cdc25B transgene do not display detectable histologicabnormalities or spontaneous mammary tumors, theyare more susceptible to carcinogen-induced mammarytumor formation. A signi®cant proportion of humanbreast cancers express elevated Cdc25B levels. Our datasuggest that the increased expression of Cdc25B inhuman breast cancers may play a causal role and is notsimply a secondary e�ect during tumor development.Likewise, these studies highlight similarities betweenmechanism(s) involved in the development of bothhuman and mouse mammary cancers. These MMTV-Cdc25B mice may, therefore, provide a useful modelsystem for further studies on mammary tumordevelopment.

Materials and methods

Transgene construction

The poly-linker between KpnI and NotI of the pGEMplasmid was substituted with a poly-linker containing S®I ±AscI ±HincII ±NotI ±AscI ±S®I sites to produce the plasmid

Synergy between Cdc25B and DMBA in inducing mammary tumorsY Yao et al

5164

pHM. A 1.6 kb NdeI-XbaI fragment of the MMTV±LTRpromoter was inserted into the HincII site of pHM after®lling in the ends with Klenow, to produce pMMTV. A2.1 kb EcoRV/HpaI human Cdc25B2 cDNA fragmentcontaining the entire ORF was inserted into the EcoRV siteof pNN265, which is ¯anked by two introns (Mayford et al.,1996). The resulting plasmid pNN265-Cdc25B2 was digestedwith NotI and the 5' intron-Cdc25B2-3' intron cassette wasinserted into the NotI site of pMMTV to produce the ®nalplasmid pMMTV-Cdc25B.

Generation of the transgenic mice

Transgenic mice were generated by the microinjection of theS®I linearized MMTV-Cdc25B cassette into the pronucleusof single-cell embryos isolated from super-ovulated B6/CBAmice, according to standard procedures (Hogan et al., 1994).Embryos that survived microinjection were implanted intopseudopregnant females and allowed to develop to term.Genotyping of the transgenic MMTV-Cdc25B mice wasperformed by Southern blot analysis of tail DNA. ForSouthern blots, 10 ml of each tail DNA sample was digestedwith BamHI and resolved in a 1% agarose gel byelectrophoresis. Hybridization was then performed using a32P-labeled (Prime-It II, Stratagene) SV40 intron-poly Afragment (3' intron-poly A). This probe is speci®c for theinjected DNA construct.

Isolation of total RNA, electrophoresis and Northern blot

Mammary tissue was isolated from the mammary glands ofnon-lactating or 10-day post-partum lactating transgenic micefrom ®ve di�erent lines and two non-transgenic control mice.Total cellular RNA from these tissues was prepared aspreviously described (Klein et al., 1998). Ten mg of totalRNA was resolved in a 1% agarose-6% formaldehyde geland were then blotted onto a Hybond N nylon membrane(Amersham). The blots were hybridized to the randomprimed human Cdc25B2 cDNA probe, and washed threetimes with 0.56SSC/0.1% SDS at 658C for 15 min eachtime. The blots were autoradiographed and the relativeabundance of RNA per lane was compared with the ethidiumbromide staining signal of the 28S ribosomal RNA bands.

DMBA treatment

The carcinogen 7,12-dimethyl-1,2-benzanthracene (DMBA,Sigma) was dissolved in corn oil at a concentration of 5 mg/ml. This mixture was heated at 378C and shaken vigorouslyto fully dissolve the DMBA. Eight-week old virgin femalemice were treated with 5 weekly doses of 1 mg DMBA viaoral gavage. Oral gavage consisted of inserting a curved blunt

tipped needle attached to a 1 ml syringe into the mid-throatof a ®rmly grasped mouse and injecting 200 ml of a 5 mg/mlsolution of DMBA. The mice were maintained in the absenceof males and were checked weekly by palpation for mammarytumor formation. Mice were sacri®ced by CO2 inhalationeither when tumors were ®rst noted or if the mice becamemoribund.

Histo-pathological examination

Mice were sacri®ced by CO2 inhalation, and tissue or tumorsamples were immediately dissected and either frozen or ®xedin 10% bu�ered formalin. Following ®xation in formalin thetissues were embedded in para�n, sectioned and stained withhematoxylin and eosin (H&E), as described elsewhere(Quezado et al., 1998). Tumor classi®cations were performedby at least two pathologists. For immunohistochemicalstaining depara�nized sections were incubated with anantibody speci®c for mouse and human Cdc25B (Transduc-tion Laboratories), followed by incubation with biotinylatedsecondary antiserum and visualized with the avidin-biotincomplex technique and 3,3'-diaminobenzidine development(ABC reagent; Vector Laboratories).

Protein preparation, antibodies and Western blots

For Western blotting assays, tissues were homogenized inK4IP bu�er (50 mM HEPES pH 7.5, 150 mM NaCl, 1.0 mM

EDTA, 2.5 mM EGTA, 1.0 mM DTT, 0.1% Tween-20, 10%glycerol, 1.0 mM PMSF, 10 mg/ml aprotinin, and 10 mg/mlleupeptin) and sonicated for three 10 s intervals, followed bycentrifugation at 14 000 r.p.m. at 48C for 20 min. Fifty mg ofeach protein sample was resolved in a 10% SDS-polyacrylamide gel. For Western blot analyses, antibodiesspeci®c for Cdc25B (Santa Cruz) and actin (UBI) were usedon the same membrane, as described by the manufacturers.Western blots were performed as previously described (Jianget al., 1993). The Cdc25B antibody recognizes both humanand murine Cdc25B.

AcknowledgementsWe are grateful for the excellent administrative assistanceprovided by Barbara Castro. We thank Dr David Beachfor the Cdc25B2 cDNA and Dr Giorgio Cattoretti fortechnical assistance and helpful discussions. This work wassupported by National Institutes of Health Grant RO1CA/ES 63467, and an award from the National Founda-tion for Cancer Research (to Dr I Bernard Weinstein). EDSlosberg was supported by NCI Cancer Training GrantT32CA09503 and a NIH Pre-Doctoral Training GrantT32GM07088.

References

Aldaz CM, Liao QY, LaBate M and Johnston DA. (1996).Carcinogenesis, 17, 2069 ± 2072.

Amundadottir L, Johnson M, Merlino G, Smith G andDickson R. (1995). Cell Growth Di�er., 6, 737 ± 748.

Archuleta M, Born J, Montano R and Burchiel S. (1992).Toxicol. Appl. Pharmacol., 113, 133 ± 137.

Baldin V, Cans C, Knibiehler M and Ducommun B. (1997a).J. Biol. Chem., 272, 32731 ± 32734.

Baldin V, Cans C, Superti-Furga G and Ducommun B.(1997b). Oncogene, 14, 2485 ± 2495.

Beato M. (1991). FASEB J., 5, 2044 ± 2051.Bortner DM and Rosenberg MP. (1995). Cell Growth Di�er.,

6, 1579 ± 1589.Bortner DM and Rosenberg MP. (1997). Mol. Cell. Biol., 17,

453 ± 459.

Bruggemeier U, Kal� M, Franke S, Scheidereit C, Beato M.(1991). Cell, 64, 565 ± 572.

Cardi� RD and Muller WJ. (1993). Transgenic mousemodels of mammary tumorigenesis. In: Cancer Surveys:The Molecular Pathology of Cancer. Imperial CancerResearch Fund, Vol. 16, pp. 97 ± 113.

Clarke PR, Ho�mann I, Draetta G and Karsenti E. (1993).Mol. Cell. Biol., 4, 397 ± 411.

Dandekar S, Sukumar S, Zarbl H, Young L and Cardi� RD.(1986). Mol. Cell. Biol., 6, 4104 ± 4108.

Draetta G and Eckstein J. (1997). Biochim. Biophys. Acta.,1332, M53 ±M63.

Furnari B, Rhind N and Russell P. (1997). Science, 277,1495 ± 1497.

Synergy between Cdc25B and DMBA in inducing mammary tumorsY Yao et al

5165

Gabrielli BG, Clark JM, McCormack AK and Ellem KA.(1997). J. Biol. Chem., 272, 28607 ± 28614.

Gabrielli BG, DeSouza C, Tonks I, Clark J, Hayward N andEllem K. (1996). J. Cell. Sci., 109, 1081 ± 1093.

Galaktionov K, Chen X and Beach D. (1996). Nature, 382,511 ± 517.

Galaktionov K, Jessus C and Beach D. (1995a). Genes Dev.,9, 1046 ± 1058.

Galaktionov K, Lee AK, Eckstein J, Draetta G, Meckler J,Loda M and Beach D. (1995b). Science, 269, 1575 ± 1577.

Gasparotto D, Maestro R, Piccinin S, Vukosavljevic T,Barzan L, Sulfaro S and Boiocchi M. (1997). Cancer Res.,57, 2366 ± 2368.

Grieco D, Avvedimento EV and Gottesman ME. (1994).Proc. Natl. Acad. Sci. USA, 91, 9896 ± 9900.

Guy CT, Cardi� RD and Muller WJ. (1992).Mol. Cell. Biol.,12, 954 ± 961.

Halter SA, Dempsey P, Matsui Y, Stokes M, Graves-Deal R,Hogan B and Co�ey R. (1992). Am. J. Path., 140, 1131 ±1146.

Hartwell LH and Weinert TA. (1989). Science, 246, 629 ±634.

Hennighausen L, Wall RJ, Tillmann U, Li M and Furth P.(1995). J. Cell. Biochem., 59, 463 ± 472.

Henrard D and Ross SR. (1988). J. Virol., 62, 3046 ± 3049.Ho�mann I, Draetta G and Karsenti E. (1994). EMBO J.,

13, 4302 ± 4310.Ho�mann I and Karsenti E. (1994). J. Cell. Sci. Suppl., 18,

75 ± 79.Hogan B, Beddington R, Costantini F and Lacy E. (1994).

Manipulating the Mouse Embryo: A Laboratory Manual.CSHL Press.

Iavarone A and Massague J. (1997). Nature, 387, 417 ± 422.Jiang W, Kahn SM, Zhou P, Zhang YJ, Cacaoe AM, Infante

AS, Doi S, Santella RM and Weinstein IB. (1993).Oncogene, 8, 3447 ± 3457.

Jinno S, Suto K, Nagata A, Igarashi M, Kanaoka Y, NojimaH and Okayama H. (1994). EMBO J., 13, 1549 ± 1556.

Klein MG, Yao Y, Slosberg ED, Lima CD, Doki Y andWeinstein IB. (1998). Exp. Cell Res., 244, 26 ± 32.

Kudo Y, Yasui W, Ue T, Yamamoto S, Yokozaki H, NikaiH and Tahara E. (1997). Jpn. J. Cancer Res., 88, 947 ± 952.

Kumagai A and Dunphy WG. (1996). Science, 273, 1377 ±1380.

Li B, Kittrell FS, Median D and Rosen JM. (1995). Mol.Carcinog., 14, 75 ± 83.

Matsui Y, Halter SA, Holt JT, Hogan BL and Co�ey RJ.(1990). Cell, 61, 1147 ± 1155.

Mayford M, Bach ME, Huang YY, Wang L, Hawkins RD,Kandel ER. (1996). Science, 274, 1678 ± 1683.

Medina D, Butel JS, Socher SH and Miller FL. (1980).Cancer Res., 40, 368 ± 373.

Millar JB, McGowan CH, Lenaers G, Jones R and Russell P.(1991). EMBO J., 10, 4301 ± 4309.

Morgan DO. (1995). Nature, 374, 131 ± 134.Muller W, Sinn E, Pattengale P, Wallace R and Leder P.

(1988). Cell, 54, 105 ± 115.Nagata A, Igarashi M, Jinno S, Suto K and Okayama H.

(1991). New Biol., 3, 959 ± 968.Nielson LL, Discafani CM, Gurnani M and Tyler RD.

(1991). Cancer Res., 51, 3762 ± 3767.Ogg S, Gabrielli B and Piwnica-Worms H. (1994). J. Biol.

Chem., 269, 30461 ± 30469.Peng C, Graves P, Thoma R, Wu Z, Shaw A and Piwnica-

Worms H. (1997). Science, 277, 1501 ± 1505.Peng C-Y, Graves PR, Ogg S, Thoma RS, Byrnes MJ, Wu Z,

Stephenson MT and Piwnica-Worms H. (1998). CellGrowth Di�er., 9, 197 ± 208.

Pierce DF, Gorska AE, Chytil A, Meise KS, Page DL, Co�eyRJ and Moses HL. (1995). Proc. Natl. Acad. Sci. USA, 92,4254 ± 4258.

Quezado MM, Middleton LP, Bryant B, Lane K, Weiss SW,Merino MJ. (1998). Hum. Pathol., 29, 604 ± 608.

Rollini P, Billotte J, Kolb E and Diggelmann H. (1992). J.Virol., 66, 4580 ± 4586.

Ross SR, Hsu CL, Choi Y, Mok E and Dudley JP. (1990).Mol. Cell. Biol., 10, 5822 ± 5829.

Russell P and Nurse P. (1986). Cell, 45, 145 ± 153.Sadhu K, Reed SI, Richardson H and Russell P. (1990).

Proc. Natl. Acad. Sci. USA, 87, 5139 ± 5143.Sanchez Y, Wong C, Thoma RS, Richman R, Wu Z,

Piwnica-Worms H and Elledge SJ. (1997). Science, 277,1497 ± 1501.

Santarelli R, Tzeng YJ, Zimmermann C, Guhl E, Graess-mann A. (1996). Oncogene, 12, 496 ± 505.

Sherr CJ. (1993). Cell, 73, 1059 ± 1065.Spitkovsky D, Jansen-Durr P, Karsenti E and Ho�man I.

(1996). Oncogene, 2549 ± 2554.Stewart TA, Pattengale PK and Leder P. (1984). Cell, 38,

627 ± 637.Stutman O. (1975). Adv. Cancer Res., 22, 261 ± 422.Truss M, Chalepakis G, Pina B, Barettino D, Bruggemeier

U, Kal� M, Slater EP, Beato M. (1992). J. SteroidBiochem. Mol. Biol., 41, 241 ± 248.

Turusov VS. (1994). Tumors of the mouse. Pathology oftumours in laboratory animals. 2, 776, InternationalAgency for Research on Cancer, Lyon.

Wang TC, Cardi� RD, Zukerberg L, Lees E, Arnold A andSchmidt EV. (1994). Nature, 369, 669 ± 671.

Weinberg RA. (1989). Cancer Res., 49, 3713 ± 3721.Witty JP, Lempka T, Co�ey RJ and Matrisian LM. (1995).

Cancer Res., 55, 1401 ± 1406.Wu W, Fan YH, Kemp BL, Walsh G, Mao L. (1998). Proc.

Am. Assoc. Cancer Res., 39, 544.

Synergy between Cdc25B and DMBA in inducing mammary tumorsY Yao et al

5166