Angiogenic Acceleration of Neu Induced Mammary Tumor Progression and Metastasis

[CANCER RESEARCH 64, 169–179, January 1, 2004]

Angiogenic Acceleration of Neu Induced Mammary Tumor Progressionand Metastasis

Robert G. Oshima,1,2 Jacqueline Lesperance,1 Varinia Munoz,1 Lionel Hebbard,1 Barbara Ranscht,1 Niki Sharan,2

William J. Muller,3 Craig A. Hauser,1 and Robert D. Cardiff4

1Oncodevelopmental Biology Program, The Burnham Institute, La Jolla, California; 2The Institute for Molecular Biology and Biotechnology, McMaster University, Hamilton,Ontario, Canada; 3Molecular Oncology Group, Departments of Medicine and Biochemistry, McGill University, Montreal, Quebec, Canada; and 4Pathology Department,University of California, Davis, California

ABSTRACT

The Neu (ErbB2, HER2) member of the epidermal growth factorreceptor family is implicated in many human breast cancers. We havetested the importance of increased angiogenic signaling in the NeuYD[mouse mammary tumor virus (MMTV)-Neundl-YD5] mammary tumormodel. Transgenic mice expressing vascular endothelial growth factor(VEGF)164 from the MMTV promoter were generated. These mice ex-pressed VEGF164 RNA and protein at 20- to 40-fold higher levels through-out mammary gland development but exhibited normal mammary glanddevelopment and function. However, in combination with the NeuYDoncogene, VEGF164 expression resulted in increased vascularization ofhyperplastic mammary epithelium and dramatic acceleration of tumorappearance from 111 to 51 days. Gene expression profiling also indicatedthat the VEGF-accelerated tumors were substantially more vascularizedand less hypoxic. The preferential vascularization of early hyperplasticportions of mammary epithelia in NeuYD;MMTV-VEGF animals wasassociated with NeuYD RNA expression, disorganization of the tightjunctions, and overlapping transgenic VEGF expression. NeuYD;MMTV-VEGF164 bigenic, tumor-bearing animals resulted in an average of 10tumor cell colonies/lung lodged within vascular spaces. No similar lungcolonies were found in control NeuYD mice with similar tumor burdens.Overall, these results demonstrate the angiogenic restriction of earlyhyperplastic mammary lesions. They also reinforce in vivo the importanceof activated Neu in causing disorganization of mammary luminal epithe-lial cell junctions and provide support for an invasion-independent mech-anism of metastasis.

INTRODUCTION

Activation of Neu (HER2, ErbB2), a member of the epidermalgrowth factor receptor family, has been implicated in the progressionof a subset of human breast cancers. The evidence includes theamplification and elevated expression of ErbB2 in human breastcancers, the clinical efficacy of anti-ErbB2 antibodies for patientswith elevated ErbB2 levels, and mammary tumors arising in trans-genic mice as a consequence of mammary epithelial expression ofoncogenic forms of Neu (1, 2). Systematic investigations of themechanisms by which activated Neu induces tumor formation andmetastasis have revealed distinct and overlapping transforming sig-nals emanating from four of five tyrosine autophosphorylation sites(3). Use of add back mutants, in which only one of the five autophos-phorylation sites is capable of signaling, revealed that recruitment ofShc (Src homology 2 domain-containing transforming protein 1) toNeu tyrosine 1227 within the context of an activating extracellular

deletion was sufficient to generate mammary tumors similar to thosecaused by the fully competent oncogene and resembling human come-docarcinoma (4). This form of Neu was also shown capable of causingdisruption of epithelial cell interaction but was limited in its ability toinduce invasive behavior in vitro (5). Tumors arising in mice expressingNeuYD driven by the MMTV promoter did not efficiently metastasize tothe lung.

Vascular endothelial growth factor (VEGF) A is a critical angio-genic growth factor necessary for the development, remodeling, andcontrol of permeability of blood vessels (6, 7). Of the three majorforms of VEGF (VEGF121, VEGF164, VEGF189), VEGF164 rescuesthe vasculature of VEGF-deficient tumor cells most efficiently (8). Intransgenic mice, VEGF164 is a potent angiogenic stimulus (9). Thetransition of hyperplasia to neoplasia has been correlated with theinduction of angiogenesis (10). The acquisition of angiogenic activityduring pancreatic tumor formation in transgenic mice involves VEGFand its mobilization by the action of matrix metalloproteinase-9 (11,12). Transgenic expression of VEGF in pancreatic tumors acceleratesthe onset of tumors but does not increase metastasis (13). During skintumor progression, inflammatory cells contribute metalloproteinaseto mobilize VEGF from extracellular matrix (14). In breast cancercells, ErbB2 signaling may increase VEGF expression (15, 16)through phosphatidylinositol 3�-kinase and AKT-dependent increasedhypoxia-inducible factor-1� synthesis (17). The coactivation of epi-dermal growth factor receptor (EGFR, ErbB1) or ErbB3 signaling inNeu-initiated mammary tumors (18) and the demonstrated importanceof phosphatidylinositol 3�-kinase in polyomavirus middle T antigen-induced tumors (19) may reflect requirements for increasing VEGFexpression in mammary tumor cells.

Increased tumor vessel density is correlated with metastasis andpoor outcome. Currently, tumor metastasis is thought to involveincreased tumor cell motility and invasion of vascular or lymph vesselspaces, movement to a distal site, and infiltration across cellular andbasement membrane barrier (20, 21). However, a noninvasive mech-anism of metastasis has been suggested on the basis of analysis of ahighly angiogenic and metastatic experimental tumor (22). For ahighly angiogenic tumor, tumor emboli may bud off into tumorvascular spaces and lodge in the capillary beds of distal organs, wherethey may grow noninvasively.

Here we have tested the importance of angiogenesis on the pro-gression of Neu initiated mammary tumors by generating transgenicmice that express both NeuYD and VEGF164 in the luminal epithelialcells of the mammary gland. Our results demonstrate a key role ofangiogenesis on the very early progression of mammary hyperplasticdisease, reinforce in vivo the role of Neu in causing disorganization ofthe polarized mammary epithelium, and show that increased vascu-larity of tumors leads to increased metastasis, apparently by an inva-sion-independent mechanism.

MATERIALS AND METHODS

Vector Construction. The murine cDNA coding for the 164-residue formof VEGF was cloned from mouse kidney RNA by reverse transcription-PCR

Received 7/1/03; revised 10/20/03; accepted 10/21/03.Grant support: Supported by grants from the California Breast Cancer Research

Program (BCRP 6JB-0073 to R. G. O.), in part by grants from Department of DefenseBreast Cancer Research Program (DAMD-17-00-0175 to R. G. O.), the National CancerInstitute (CA 74597 to R. G. O and C. A. H.), the Canadian Breast Cancer Initiative (toW. J. M.), the National Institute of Child Health and Human Development (HD 25938 toB. R.), and a grant from Cancer Center Support (CA 30199).

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

Requests for reprints: Robert G. Oshima, The Burnham Institute, 10901 North TorreyPines Road, La Jolla, California 92037. Phone: (858) 646-3147; Fax: (858) 713-6268;E-mail: [email protected].

169

Research. on November 22, 2015. © 2004 American Association for Cancercancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

with the primers shown in Table 1. This amplified a 707 bp region of theVEGF mRNA (GenBank accession no. NM_009505) including 56 nucleo-tide(s) (nt) of the 5� noncoding region and 77 nt of 3� noncoding region coding.Cleavage of the peptide leader of the cloned form would generate VEGF164.After digestion with HinDIII and EcoRI, the amplified fragment was clonedinto the MMTV vector (1) and sequenced.

Transgenic Mice. MMTV-VEGF transgenic mice were generated by astandard pronuclear DNA injection and screened by PCR of tail DNA for thepresence of the MMTV vector or specifically for the MMTV-VEGF transgene(Table 1). Of 21 transgenic founders detected by PCR, 16 transmitted thetransgene, and 14 were screened for transgene expression in virgin mammarygland. Two moderately expressing lines, MMTV-VEGF-25 (VEGF-25) andMMTV-VEGF-64 (VEGF-64), and one very high level expressing line,MMTV-VEGF-89 (VEGF-89), were examined in more detail.

Mammary Tumor Analysis. VEGF-25 and VEGF-64 mice were mated tothe MMTV-Neundl-YD5 (NeuYD) line that develops mammary tumors with100% penetrance (4). The PCR primers for genotyping these animals areshown in Table 1. A subset of animals was also mated to the FVB/N-TgN(TIE2-lacZ) 182Sato (Tie2-LacZ) transgenic animals obtained from the Jack-son Laboratory for visualization of endothelial cells. Histochemical stainingfor �-galactosidase of ear punch pieces was routinely used for the identifica-tion of these transgenic mice. All three mouse lines were in the FVB/N geneticbackground. Female animals with the appropriate genotypes were examinedfor tumors by palpation three times weekly starting at 30 days of age. Afterdiscovery, the length and width of tumors were measured with calipersminimally three times weekly until the tumors reached approximately 1 cm indiameter. Tumor volume was calculated as (length � width2)/2. Tumor growthwas measured for the single largest tumor of 11 NeuYD;VEGF-25 mice and 9NeuYD mice. Average values represent between 3 and 11 individual values foreach time point. Linear regression analysis of all the data points and compar-isons of the slopes were performed by Prism software (GraphPad).

Protein Analysis. VEGF protein was measured with an ELISA assay(R&D Systems). Frozen tissues were homogenized in the provided lysis buffer,and tissues lysates were cleared by centrifugation at 15,000 � g at 4°C. Ifpresent, the fat layer was discarded, and the cleared protein lysate was used.Protein content of tissue lysates was determined with the Bio-Rad detergent-compatible protein assay.

For the analysis of Neu, ErbB2, Grb2, and Akt, tissue samples wereprepared and analyzed by electrophoresis and immunoblotting as describedpreviously (4).

Histology. Mammary fat pads were excised, mounted on glass slides, fixedin acidic ethanol (Carnoy’s fixative or buffered formaldehyde), and stainedwith carmine alum overnight (23). For visualization of bacterial �-galactosid-ase, mammary tissues were fixed and stained as described previously (24).Apoptotic cells were identified with a commercial terminal deoxynucleotidyl-transferase-mediated nick end labeling staining kit (Apotag kit; Intergen)according to the instructions of the manufacturer. For vascular perfusion of the

mammary gland, animals under deep anesthesia were perfused through the leftventricle with the use of a peristaltic pump and a volume equal to 50% of thebody weight of the animal. Colloidal carbon particles from Higgins India inkwere sized by differential centrifugation (25). After resuspension in a startingvolume of PBS, material was filtered through 4.5-�m nitrocellulose filters andused after diluting 4-fold.

Immunochemistry. Antibody staining of CD31 antigen was performed onformaldehyde-fixed, paraffin embedded tumors. For increased sensitivity, nor-mal and hyperplastic mammary tissue was fixed and extracted overnight inacetone at �20°C. The tissue was then incubated with 30% sucrose in PBS forseveral hours at 4°C and frozen in OCT medium for frozen sections. Air driedfrozen sections were pretreated with hydrogen peroxide, blocked, washed, andvisualized using the Vector Laboratories avidin-biotin complex method kitexcept mouse-absorbed, biotin-labeled rabbit antirat IgG secondary antibody(Vector Laboratories) was used. The primary antibody was rat monoclonalantibody to mouse CD31 (catalogue no. 557355; PharMingen). For immuno-fluorescence, mammary glands and tumors were snap frozen in liquid nitrogen.Ten-�m cryosections were prepared and fixed in acetone for 10 min. Sectionswere blocked in 10% FCS in Tris-HCl buffered saline for 30 min andincubated overnight with rat anti-E-cadherin (catalogue no. 13-1900; Zymed)and zonula occludens-1 (ZO-1) antibodies (catalogue no. R26.4C; Develop-mental Studies Hybridoma Bank, Iowa). Rabbit anti-mouse keratin 18 (26),rabbit antibody against mouse T-cadherin5, rat anti-CD31 (PharMingen) anti-bodies, and FITC-labeled Bandeiraea simplicifolia lectin B4 (catalogue no.B-1205; Vector Laboratories) were applied the next day for an hour. Antibodybinding was detected with Alexa 488 donkey antirabbit and Alexa 594 goatantirat secondary antibodies (Molecular Probes), observed on a Bio-Rad 1024confocal microscope, and images were analyzed with Photoshop software(Adobe).

In Situ Hybridization. Freshly excised tumors and tissues were fixed in4% paraformaldehyde-PBS overnight, dehydrated, and embedded in paraffin.VEGF and NeuYD probes were generated from PCR-amplified cDNA frag-ments using primers incorporating the T7 promoter (Table 1). In vitro tran-scription was performed in the presence of digoxigenin UTP (Roche). In situhybridization was performed as described previously (27) with minor modifi-cations. Briefly, tissue sections were deparaffinated and rehydrated, washedtwice in PBS, and postfixed in 4% paraformaldehyde-PBS for 15 min at roomtemperature. Sections were subsequently washed in PBS before treatment in0.2 M HCl for 8 min to block endogenous alkaline phosphatases. Sections werewashed again and treated with proteinase K (Dako) for 3 min at roomtemperature. After washing, sections were incubated twice for 15 min in PBScontaining 0.1% diethylpyrocarbonate and equilibrated for 15 min in 5� SSC(0.75 M NaCl , 0.075 M Na-citrate) before prehybridization for 2 h at 65°C inhybridization solution (Dako). After denaturing, probes were added to thehybridization mix (200 ng/�l), and hybridization was carried out at 65°C for16 h. Sections were then washed (27) and incubated for 2 h at room temper-ature with alkaline phosphatase-coupled antidigoxigenin antibody (BoehringerMannheim) diluted 1:500. Color development was performed with nitrobluetetrazolium at room temperature for 24 h. Sections were counterstained innuclear red before dehydration and mounting.

RNA Analysis. Total RNA was isolated using Tri-reagent (MolecularResearch Center, Inc; Ref. 28). RNase protection was performed with acommercial kit (RPAIII; Ambion) as recommended by the manufacturer withthe angiogenesis probe set (mAngio-1; BD PharMingen) or with the SPAprobe for the SV40 splice and polyadenylation region of MMTV206 transgenicvector mRNA probe (19)

For quantitative real-time PCR (Q-PCR), cDNA was prepared from 4 �g ofwhole RNA by reverse transcription using oligo(dT)18 primer and the Super-Script II First-Strand Synthesis System kit (Invitrogen). The PCR primers weredesigned using Primer 3 software (29) to amplify sequences within 1200 bp ofthe 3� end of the mRNA and span introns when possible. The specific primersare shown in Table 1. Q-PCR reactions were performed using a LightCyclerinstrument and the LightCycler SYBR green DNA master mix (Roche, Mann-heim Germany). The 10-�l Q-PCR reactions were performed essentially asrecommended by the manufacturer, containing cDNA derived from 10 ng of

5 M. Garlatti, B. Marturano, M. W. Porter, and B. Ranscht. Expression of T-cadherinin subdomains of the neuroepithelium and initial axon tracts of the embryo, manuscript inpreparation.

Table 1 PCR primers

Target Primer sequence

VEGFa cDNA CGCGAAGCTTAGTCCGAGCCGGAGAGAGGGCCGAATTCGTGACATGGTTAATCGGTCTTTC

VEGF ISH TTTCTGCTCTCTTGGGTGTAATACGACTCACTATAGGGCCGCCTTGGCTTGTCACA

Neu ISH CGCTCACTGCGGGAGCTGTAATACGACTCACTATAGGGGATGAATGTCACCGGGCT

MMTV vector CGTGTTTGAATTTGGACTGACGACATCACTGAGCTAAATCCCCAACCC

MMTV-VEGF GAAAGACCGATTAACCATGTCACTCAGCAGTAGCCTCATCATCA

MMTV-NeundlYD5 TTTCCTGCAGCAGCCTAACGGAACCCACATCAGGC

CD31/Pecam CCAAGGCCAAACAGAAACGATTTATGGGTTTTACTGCATC

Flk1/Kdr CCGAAGAATTGTGAGAACAGGAGCAAACCAACCAATTAAG

VE-Cad/Cdh5 CATCTCAGGGAATGAACCTCGCCTCTTTGTGTCTGTATGC

a VEGF, vascular endothelial growth factor; ISH, in situ hybridization; MMTV, mousemammary tumor virus; Pecam, platelet/endothelial cell adhesion molecule; Flk1, fetalliver kinase 1; Kdr, kinase insert domain-containing receptor; VE-Cad, vascular endothe-lial-cadherin. Cdh5, H-cadherin.

170

ANGIOGENESIS-INDUCED METASTASIS



whole RNA, a final concentration of 4 mM MgCl2, and 0.5 �M of each primer.Amplification was initiated at 95°C for 70 s, followed by 42 amplificationcycles of 95°C for 0 s, 56°C for 7 s, and 72°C for 20 s. The cyclophilin A(CPH/peptidylprolyl isomerase A) housekeeping gene was used as constantinternal reference for each cDNA preparation as described previously (30).Specificity of the SYBR green Q-PCR signal was monitored by melting curveanalysis and by agarose gel electrophoresis analysis, confirming that eachQ-PCR product was the anticipated size. Relative gene expression was ana-lyzed using the LightCycler software (version 3.5), which calculated therelative concentration of each specific cDNA to that of CPH cDNA from thesame cDNA preparation. This normalized analysis used a standard curvegenerated for each experiment, relating differences in Q-PCR-crossing thresh-olds (�CT) to the cDNA concentrations of CPH.

RESULTS

Tolerance of VEGF164 Expression During Normal MammaryGland Development. To test the importance of increased angiogen-esis on the progression of Neu/ErbB2 tumors, transgenic mice wereconstructed which express VEGF164 in mammary luminal epithelia. A707 bp form of the VEGF cDNA coding for the 190 residue unproc-essed form of VEGF was placed under the control of the MMTVlong-terminal repeat promoter (Fig. 1A). Cleavage of the precursor isexpected to generate the 164-residue form of secreted mouse VEGFthat retains the proteoglycan-binding domain. Two of the 14 lineswere male sterile and discontinued. Eight lines expressed the trans-gene in female virgin mammary gland. Three lines of mice withhigher expression were expanded and examined in greater detail. Oneof these lines (VEGF-89) required foster care of maternally transmit-ted progeny. Males of the VEGF-89 line and the two other discarded

male sterile lines had testes abnormalities (data not shown) as de-scribed by other investigators (31).

The expression of transgene-encoded RNA in different organs ofthe VEGF-25 line is shown in Fig. 1B. Expression was high in virginmammary gland and low but detectable in salivary gland and lung.VEGF mRNA was elevated in the virgin, pregnant, lactating, andinvoluting mammary tissue of VEGF-25 mice (Fig. 1C, Lanes 6–10).VEGF protein expression paralleled RNA expression in transgenicmammary tissue (Fig. 1D). In nontransgenic mammary fat pads,VEGF production by nonepithelial components was confirmed inmammary fat pads cleared of epithelium (Fig. 1D) and fat pad withoutlymph node (Fig. 1D). Elevated levels of VEGF were found in milk.VEGF protein levels in virgin VEGF-25 transgenic mammary tissuewere 28-fold higher than control animals and remained 6- to 12-foldhigher than controls at different stages of gland development. VEGFlevels were normalized to total protein and were not corrected for thelarge amounts of secretory proteins generated during pregnancy andlactation. VEGF was very high in the milk of VEGF-25 animals butwas not elevated in serum or tissue lysates of organs other thanmammary gland (data not shown). VEGF levels in the VEGF-89 linewere 2.8-fold higher than VEGF-25 values and 80 times the value fornontransgenic animals. These results are consistent with the reliablemammary epithelial tissue specificity of the MMTV vector and showthat the VEGF164 transgene is expressed in all major stages of mam-mary gland biology. Transgenic VEGF164 appears to remain largelylocalized in mammary tissue and milk.

Normal Mammary Gland Development in VEGF-25 Mice. Al-though VEGF164 was overexpressed in the mammary tissues ofVEGF-25 mice (and the VEGF-64 line, data not shown), mammary

Fig. 1. Elevated vascular endothelial growth factor (VEGF) expression in VEGF164 transgenic mice. A, schematic map of mouse mammary tumor virus (MMTV)-VEGF transgeneand RNase protection probes and protected fragments. The location of the MMTV long terminal repeat promoter (LTR), VEGF cDNA and SV40-derived processing regions areindicated by the black, white, and hatched portions of the horizontal bar. The gray area indicates the residual 5� noncoding region of the Ras cDNA (1). The second small white regionindicates an SV40 intron. The site of polyadenylation of the SV40 major transcripts is indicated by pA. The mRNAs, RNase protection probes, and protected fragments are indicatedbelow the map with sizes indicated in nucleotides (nt). Double stranded complexes of the 847 nt SPA probe with the two mRNAs (one spliced, one not) are both cleaved by RNasein an A/U rich region (vertical arrow) generating two 5� fragments. The internal 174 nt VEGF probe is trimmed of vector sequences to 148 nts by RNase. B, RNase protection analysisof transgenic RNAs from different organs of a VEGF-25 virgin female mouse. The three protected probe fragments are indicated at the left. Organ abbreviations are as follows:brn, brain; hrt, heart; liv, liver; lng, lung; spl, spleen; sg, salivary gland; kid, kidney; mg, mammary gland; M, size markers. The signals for the L32 ribosomal protein RNA, analyzedsimultaneously, are shown in the bottom portion of the gel. C, RNase protection analysis of VEGF, L32, and glyceraldehydes 3-phosphate dehydrogenase (GAPDH) RNAs. RNAs frommammary tissue dissected from the following: V, virgin; P, 15 days pregnant; L, 16 days lactating; In, 14 days involution; 25 and T1, tumor from NeuYD;VEGF-25 (mouse mammarytumor virus-Neundl-YD5 mouse strain) bigenic animal; wild-type (wt); and T2, NeuYD tumor without transgenic VEGF. Note that the transgenic VEGF mRNA generates the samesignal as endogenous VEGF with this probe. Identities of the protected fragments are indicated at the left. D, VEGF protein expression in tissue homogenates of mammary glands andtumors. Open bars indicate wt animals, and filled bars indicate tissues from VEGF-25 animals. Gray bars indicate values for VEGF-89 animals. Abbreviations are as follows:FP, cleared fat pad; FP-LN, fat pad without lymph node; E, pregnant with indicated days; Lac, lactating with indicated days after litter birth; In, involution with the indicated numberof days after weaning. VEGF was measured by ELISA assay. E, rescue of VEGF-89 pups by foster mother feeding. Six newborn pups from the VEGF-89 line were fed by a wt ICRstrain foster mother (VEGF-89/wt mom, Œ) and six wt ICR strain pups were fed by the VEGF-89 mother (wt/VEGF-89, F). The growth of a typical litter of six FVB/N pups is alsoshown (w/w, f). Individual pups were weighed each day. Each point represents the average and SD.

171




gland development was not distinguishable from control animalsduring virgin development, pregnancy, lactation, and involution (Fig.2, A and B; additional data not shown). Furthermore, the interactionbetween mammary fat pad capillaries and mammary epithelial tubuleswas not dramatically altered in VEGF-25 female mice (Fig. 2, C–F).However, mothers of the VEGF-89, highly expressing line, weregenerally unable to support litters because of a lactation defect. Thisline required propagation of pups by foster mothers because maleswere sterile. Male sterility was associated with testis and epididymidisabnormalities as described previously (31; data not shown). VEGF-89lactating mothers revealed a dramatic deficiency of differentiatedalveoli (Fig. 2, G and H). The deficient growth of VEGF-89 pups was

confirmed to be attributable to lactation based upon normal growth ofVEGF-89 pups (transgenic and nontransgenic progeny) that had fostermothers and the deficient growth of wild-type (wt) pups that hadVEGF-89 mothers (Fig. 1E). Thus, high expression of VEGF164 cancause female mammary gland and male testis abnormalities. How-ever, more modest levels of the VEGF164 expression are surprisinglywell tolerated by both males and females in the VEGF-25 andVEGF-64 lines.

VEGF164 Acceleration of Tumor Appearance and Growth. Totest the effects of mammary epithelial expression of VEGF on mam-mary tumor development, the VEGF-25 transgene was combined withthe NeuYD transgenic mouse mammary tumor model (4). Tumors

Fig. 2. Mammary epithelia and vascular patternsof VEGF-25 mice and lactation defect of VEGF-89mice. Whole mount preparations of mammary fatpads of wild-type (wt; A, C, E) and VEGF-25 mice(B, D, F) with additional Tie2-LacZ transgene (Aand B). Major vessels are revealed by blue-bacterial�-galactosidase staining (A and B). Epithelium isstained red with carmine. Panels C-F are fromanimals perfused with ink before sacrifice. Ink con-taining vessels are black. Note the extensive capil-lary structure and the close juxtaposition of epithe-lium and capillaries. The size bars are indicated inmicrons. G and H represent H&E-stained sectionsof mammary tissue from wt and VEGF-89 mothers,which had been lactating for 3.5 days. Note theexpanded structure of the milk-containing lobulesin G and the paucity of expanded lobules in H.

172




were detected in NeuYD;VEGF-25 animals after an average of 51days compared with an average of 111 days for NeuYD mice.(Logrank test, P � 0.001; Fig. 3A). This extreme acceleration oftumor appearance was also associated with significantly faster growthof the tumors after appearance (Fig. 3, B and C). For purposes ofclarity, average data for each time point are shown. However, linearregression analysis of all data points of the two types of tumorsindicates that the difference in slopes between the two data sets issignificant. (P � 0.001). To investigate the dramatic differences intumor appearance associated with VEGF164 expression, the earlystages of tumor development were examined.

Increased Vascular Association with Hyperplastic MammaryEpithelium. In NeuYD animals during mammary epithelium inva-sion of the fat pad, hyperplastic epithelia in some but not all epithelialbranches were evident in both NeuYD and NeuYD;VEGF-25 mice(Fig. 4, A and B). VEGF expression did not change these first signs ofaltered mammary epithelial growth control. However, the appearanceof the hyperplastic epithelium was accompanied by preferential asso-ciation of the vasculature with the hyperplastic epithelium in NeuYD;VEGF-25 animals (Fig. 4, C and F). The hyperplastic epitheliumattracts blood vessels in both NeuYD and NeuYD;VEGF-25 animals,but the bigenic animals have much more extensive and closer asso-ciation (Fig. 4F).

The preferential association of new vessels with hyperplastic epi-thelium was confirmed in sectioned material stained for the endothe-lial marker CD31 (Fig. 5, A and B). Endothelial cells were foundinterdigitating and surrounding hyperplastic and neoplastic epithelialcells (Fig. 5B). The increased vascularity of the early lesions persistedin solid tumors and generated a distinctive, highly vascularized mor-phology. (Fig. 5, C and D). Large diameter vascular cysts werecommon in NeuYD tumors (Fig. 5C, black arrows). Additional areasof degeneration (Fig. 5C, red arrow) were present in NeuYD tumorsbut were not observed in NeuYD;VEGF-25 tumors. Tumors acceler-ated by VEGF-25 had characteristic small fingers or nests of tumorcells surrounded by blood vessels. (Fig. 5D). All of the NeuYD;VEGF-25 tumors appeared to have well defined boundaries definedby endothelial cells.

Mosaic Expression of Transgenic NeuYD and VEGF. The rea-son for the mosaic patterns of hyperplastic growth in developingNeuYD mammary epithelial trees was investigated by in situ hybrid-ization. Elevated Neu RNA was detected in hyperplastic cellularmasses but generally not in the epithelium with normally organizedepithelial structures (Fig. 6, A, arrow and C). The degree of cellularhybridization increased with decreasing epithelial organization andincreasing neoplastic morphology, culminating in frank tumors thatappeared nearly uniform in the cellular pattern of hybridization (datanot shown). The VEGF transgene RNA was also detected in a non-

uniform epithelial pattern although single-expressing cells in normalepithelial structures were common (Fig. 6B, arrows). EndogenousVEGF RNA levels were generally at or below the threshold ofdetection of this method except in NeuYD tumors. Detection oftransgenic Neu and VEGF164 RNAs on sequential portions of thesame bigenic mammary tissue revealed that the two transgenes hadoverlapping patterns of expression. Neither transgene appeared to beuniformly expressed in the early epithelial tubules. Small dysplasticnests of cells were positive for both Neu and VEGF. However, thehybridization pattern for both probes appeared nearly uniform in franktumors (Fig. 6E; additional data not shown). Apoptotic areas ofNeuYD tumors were identified by labeling of DNA breaks by theterminal deoxynucleotidyltransferase-mediated nick end labelingmethod (Fig. 6F). Only a few scattered single apoptotic cells wereobserved in NeuYD;VEGF-25 tumor sections (data not shown). Insitu hybridization revealed increased endogenous VEGF RNA sur-rounding apoptotic areas presumably caused by hypoxia, whereastransgenic VEGF was uniformly expressed in NeuYD;VEGF-25 tu-mors (Fig. 6E).

These results indicate that the expression of the MMTV-drivenNeuYD and VEGF transgenes are not uniformly expressed coinciden-tally with the hormonally driven mammary epithelial tree morpho-genesis but rather are expressed in a mosaic pattern within mammaryepithelial cells. Both transgenes appear more highly expressed inhyperplastic tissues and neoplasia. Differential NeuYD expressionappears to be the cause of the nonuniform hyperplastic epithelialpattern.

Disruption of Tight Junction Organization in Hyperplasia andNeoplasia. The dramatic differences in vessel organization betweencontrol NeuYD mammary tissue and NeuYD;VEGF-25 bigenic mam-mary epithelia, and the transgenic VEGF expression in milk and theabsence of increased VEGF serum levels, suggested that transgenicVEGF164 may not be available to normal fat pad capillaries becauseof the organization of polarized epithelial junctions and the vectorialsecretion of mammary luminal epithelial cells into the lumen. Theinfluence of NeuYD-induced hyperplasia and neoplastic growthon epithelial junctional complexes was examined by immunohisto-chemical detection of the ZO-1 tight junction protein, E-cadherin, andcytoplasmic mouse keratin 18 (EndoB) to identify epithelial cells.Epithelial ducts of both NeuYD and NeuYD;VEGF-25 mammarytissues had the expected polarized organization of ZO-1 near theapical surface in appropriate cutting planes (Fig. 7A). The ZO-1distribution of organized ducts was not distinguishable from control,normal virgin ductal structure (data not shown). In addition, bloodvessels stained with ZO-1 antibody were distinguished from epithe-lium by their lack of mouse keratin 18 reaction (Fig. 7B). Hyperplastictissues commonly had disorganized ZO-1 even in close proximity

Fig. 3. Acceleration of NeuYD tumor appearance and growth by VEGF-25. A, time of detection of tumors in NeuYD female mice (F) with the additional VEGF-25 transgene(�VEGF, Œ) are shown as a function of age. The average and SD of the time of first detection is indicated. Logrank tests of survival plots of the data indicated a statistically significantdifference between NeuYD;VEGF-25 and NeuYD (P � 0.001). B and C, tumor growth rates. The mean tumor volume (length � width2 and SD is shown as a function of elapsedtime after first detection for the largest tumors of 9 NeuYD and 11 NeuYD;VEGF-25 mice. Trend lines were determined by linear regression analysis of the means. B, NeuYD tumors;C, NeuYD;VEGF-25 tumors. Linear regression analysis of all individual data points indicates that the difference between the slopes is significant (P � 0.0001).

173




with relatively normal-appearing ductal structures (Fig. 7, A and B).Because the dysplastic growth pattern of the epithelium is associatedwith NeuYD expression, the disruption of the tight junction organi-zation appears to be an early consequence of NeuYD expression.ZO-1 was persistently expressed in tumors arising from the hyper-plastic growths but was not found in an organized pattern (Fig. 7C).In NeuYD;VEGF-25 tumors, ZO-1 was more organized around singleor groups of tumor cells (Fig. 7D). However, this organization colo-calized with the increased density of blood vessels as revealed by theBandeiraea simplicifolia lectin and CD31 reaction (Fig. 7, H and K).The ZO-1 not associated with blood vessels in NeuYD;VEGF-25tumors appeared as a punctate pattern similar to NeuYD tumors (Fig.7H). These results show that hyperplastic growth of the NeuYD;

VEGF-25 epithelium is associated with disruption of the organizationof tight junctions.

E-cadherin was found in the expected basolateral pattern of normalductal structures of both NeuYD and NeuYD;VEGF-25 tissues (Fig.7, E and F; additional data not shown). E-cadherin continued to beexpressed at intercellular surfaces of hyperplastic and neoplastic ep-ithelial cells (Fig. 7, E–G). The E-cadherin pattern was similar in bothNeuYD and NeuYD;VEGF-25 tumors (Fig. 7G; data not shown).T-cadherin, a glycosylphosphatidyl inositol-linked cadherin foundpreviously on mouse tumor blood vessels, was colocalized with CD31in the normal mammary fat pad (data not shown) and in the vessels ofNeuYD and NeuYD;VEGF-25 hyperplastic epithelium and in tumors(Fig. 7, I–L).

Fig. 4. Mosaic development of mammary hy-perplasia and preferential vascularization in NeuYD;VEGF-25 animals. A and B, whole mount mam-mary tissues were stained with carmine to revealepithelium in red. Arrows point to branches dis-playing abnormal dysplastic morphology. C, anunfixed, 45-day-old NeuYD;VEGF-25 bigenic fe-male mammary fat pad was photographed beforedissection from the skin, revealing a swollen lymphnode (white arrow) and increased diffuse vascula-ture (arrows). D, after dissection, mounting, andstaining, the pattern of the diffuse vessels corre-sponds to the hyperplastic mammary epithelium(black arrows). Ink perfused, carmine stained,whole mounted NeuYD (E) and bigenic NeuYD;VEGF-25 (F) mammary fat pads are shown. Notethe close juxtaposition of the vessels with the hy-perplastic growths in the bigenic animals (F).

174




Fig. 5. Increased endothelial cell association withVEGF-25 epithelium and tumors. Acetone-fixed,frozen-sectioned mammary tissues from female micewith the indicated genotypes and ages were stainedwith CD31 antibody (brown color) and hematoxylin.Hyperplastic epithelium of 79 d NeuYD (A) and 45 dNeuYD;VEGF-25 bigenic mammary gland (B) areshown. C, 108 d NeuYD tumor. Black arrow pointsto large vessels with unstained erythrocytes. Redarrow points to area of degeneration. D, 72 dNeuYD;VEGF-25 mammary tumor. Note extensivevasculature surrounding fingers of tumor cells.

Fig. 6. Overlapping vascular endothelial growthfactor (VEGF) and Neu RNA expression patternsin hyperplastic and tumor tissue. In situ hybridiza-tion (ISH) with Neu (A and C) or VEGF (B, D, E)probes is revealed by blue color. Nuclei were coun-terstained with neutral red. A and B, ISH wasperformed on sequential sections of NeuYD;VEGF-25 (mouse mammary tumor virus-Neundl-YD5 mouse strain;mouse mammary tumor virusvascular endothelial growth factor-25 mouse strain)bigenic, hyperplastic tissue. Control sense probesdid not develop signals (not shown). C, Neu RNAexpression in hyperplastic tissue of NeuYD ani-mals. D, endogenous VEGF signal of section of aNeuYD tumor containing an apoptotic area. Apop-totic cells of an adjacent section of the tumorshown in D were detected by end labeling of DNAbreaks (terminal deoxynucleotidyltransferase-me-diated nick end labeling; brown color) in F. E,VEGF RNA detected in NeuYD;VEGF-25 tumorcells.

175




Accelerated Tumor Formation Does Not Reflect Increased NeuSignaling. Because the early appearance of dysplastic mammarylesions in NeuYD mice is correlated with expression of the oncogene,the accelerated growth of NeuYD;VEGF tumors might reflect anindirect effect of increased NeuYD expression or signaling. However,immunoblot analysis of the levels of Neu, ErbB3, Akt, activated Akt,and Grb2 proteins in NeuYD tumors was not distinguishable fromNeuYD;VEGF-25 tumors (data not shown). Thus, neither major al-terations in expression of NeuYD and ErbB3 nor downstream signal-ing to Akt account for the more rapid growth of NeuYD;VEGF-25tumors.

Decreased Hypoxia-Responsive Gene Expression in Well-Vascularized Tumors. To evaluate whether the increased vascularityof NeuYD;VEGF-25 tumors impacted tumor physiology, the mRNAsfor a number of endothelial cell and metabolic cell markers in NeuYDand NeuYD;VEGF-25 tumors were measured by Q-PCR (Fig. 8). ThemRNAs for the endothelial cell markers vascular endothelial-cadherin(VE-cad), fetal liver kinase-1 (FIK-1), and CD31 were elevatedapproximately 5-fold in NeuYD;VEGF-25 tumors. T-cadherin mRNAexpression was also increased, consistent with the histology andimmunostaining results (Figs. 5 and 7). Expression of severalhypoxia-responsive genes was also analyzed. Endogenous VEGFmRNA expression (distinguished from transgenic VEGF164 by itsunique 3� noncoding region) was decreased 4-fold in NeuYD;VEGF-25 tumors. Expressions of glycolytic pathway componentsglyceraldehyde-3-phosphate dehydrogenase and phosphoglyceratekinase-1 were also significantly reduced in these tumors, whereasglucose transporter 1 and hypoxia-inducible factor-1� expressionwere unchanged (Fig. 8). Increased hypoxia and glycolysis are char-acteristics of tumors, and reduced expression of several hypoxia-responsive genes in NeuYD;VEGF-25 likely reflects their increasedperfusion.

Increased Vascularity Results in Tumor Cell Colonization ofthe Lung. Lung tissues of tumor-bearing NeuYD and NeuYD;VEGF-25 mice were examined for evidence of metastasis. No meta-static lesions were found in the lungs of NeuYD animals (Table 2).However, the lungs of NeuYD;VEGF-25 animals contained an aver-age of 9.6 easily recognizable nests of tumor cells bounded by cellularborders and commonly located within vascular spaces (Table 2; Fig.9, C and D). The tumor cells of some of these emboli were identifiedby the expression of Neu RNA (Fig. 9, A and B). The tumor cells ofthese emboli continued to grow to small tumors (Fig. 9D), but nonehad clearly invaded surrounding tissues. Additional metastases in liver

Fig. 7. Loss of tight junction organization in NeuYD mammary tumors. Hyperplasia and tumors were immunostained with antibodies against mouse keratin 18 (mK18; green), zonulaoccludens-1 (ZO-1; red), E-cadherin (red), CD31 (red), T-cadherin (green), and Bandeiraea simplicifolia lectin B4 (BSL-1, green) and analyzed with a confocal microscope. Keratinantibody reacts with the mammary epithelium and tumor cells (A-G, green). ZO-1 staining is organized near the lumen surface of ductal structures but in a punctate pattern withinNeuYD and NeuYD;VEGF-25 (mouse mammary tumor virus-Neundl-YD5 mouse strain;mouse mammary tumor virus vascular endothelial growth factor-25 mouse strain) hyperplasia(A and B, red arrows). In NeuYD tumors, ZO-1 staining remains punctated (C, red arrow) and is increased in NeuYD;VEGF-25 tumors (D, red) but in some instances does notcolocalize with mK18 (see red arrow). Staining with Bandeiraea simplicifolia confirmed that although there is some punctate staining (H, red arrow), the majority of ZO-1 expressionin NeuYD;VEGF-25 tumors are associated with the endothelium (H, yellow arrow). Basolateral distribution of E-cadherin (F, red arrow) is found over all surfaces of hyperplasticcells (E and F, red) and in tumors (G, red). Vascularity increases in NeuYD;VEGF-25 tumors as seen by T-cadherin on the endothelium (I and J, green) and colocalized with CD31(K, red and L, yellow). Scale bars, 25 �m.

Fig. 8. Increased endothelial and decreased hypoxia-responsive gene expression inNeuYD;VEGF-25 tumors. RNA from NeuYD and NeuYD;VEGF-25 mammary tumorswas transcribed into cDNA and amplified by real-time PCR with primers specific for theindicated gene products. Signals for each product were normalized to that of cyclophilinA (CPH/peptidylprolyl isomerase A) and then to NeuYD tumor values. The values are theaverage and SD from three different tumors of each genotype, each measured in at leasttwo separate Q-PCR assays. NeuYD;VEGF tumors are indicated (�) relative to NeuYDtumors (f).

176




were not found in several of the VEGF;NeuYD animals, but a sys-tematic examination of additional organs was not performed. The lungemboli were distinctive because some colonies were surrounded byCD31-positive endothelial cells within CD31-positive vascular spaces

(Fig. 9, E and F). Increased vascularization caused by transgenicVEGF164 expression in tumor cells results in increased colonization ofthe lungs despite the decreased elapsed time between detection andsacrifice (13 � 4.6 days for NeuYD;VEGF-25 and 21 � 3.5 days forNeuYD). However, many tumor cell colonies were confined within avascular boundary and failed to invade surrounding lung tissues at thetime of animal sacrifice.

DISCUSSION

The expression of VEGF164 led to increased vascularity and dra-matically accelerated growth of mammary tumors. These findingsconfirm the importance of angiogenic signaling in mammary tumordevelopment. However, moderate VEGF164 expression in normalmammary luminal epithelial cells did not dramatically alter normalmammary development or the association of vessels with developingepithelial tubules. This moderation of the effect of transgenicVEGF164 is likely because of the polarized vectorial secretion byluminal epithelial cells. The elevated expression of VEGF in the milkof lactating VEGF164 transgenic mice and the absence of elevatedVEGF in serum is supportive of this conclusion. The absence ofelevated VEGF in the serum of tumor-bearing NeuYD;VEGF-25transgenic mice may reflect the local deposition of VEGF164, whichretains a heparin-binding domain and is commonly cell associated.

Table 2 Lung metastasis

Genotype Mouse no. Age (days) Mets/sectiona Methodb

Neu 84 111 0 H&ENeu, Tc 15 123 0 H&ENeu 83 129 0 H&ENeu 56 139 0 H&E, ISHNeu 85 135 0 H&E, CD31Neu 70 139 0 H&ENeu, VEGF 43 56 21 H&ENeu, VEGF 79 65 2 H&E, ISHNeu, VEGF, T 42 66 12 H&E, ISH, CD31Neu, VEGF 112 66 12 H&ENeu, VEGF, T 72 67 4 H&ENeu, VEGF 14 68 21 H&ENeu, VEGF 23 72 3 H&ENeu, VEGF 25 72 4 H&E, CD31Neu, VEGF, T 91 72 13 H&E, CD31Neu, VEGF, T 89 74 7 H&ENeu, VEGF, T 25 89 7 H&E

9.6 � 6.8a Tumor cell emboli per lung section. Mets, metastases/section.b Method of detection; H&E, hematoxylin and eosin stain; CD31, immunohistochem-

ical stain; ISH, in situ hybridization.c T, Tie2-LacZ; Neu, MMTV-NeuYD; VEGF, MMTV-VEGF-25.

Fig. 9. Lung metastasis in NeuYD;VEGF-25(mouse mammary tumor virus-Neundl-YD5 mousestrain;mouse mammary tumor virus vascular endo-thelial growth factor-25 mouse strain) mice. Allpanels show histological sections of lung from aNeuYD;VEGF-25 tumor-bearing animal. Sectionsshown in A, C, and D were stained with H&E. Bwas a section of the same specimen shown in A butsubjected to Neu in situ hybridization (ISH). Insetsin A and B show increased magnifications of em-boli. Arrow in A points to cell-limited boundary. Cand D show additional examples of tumor emboli.Note erythrocytes adjacent to the tumor emboluswithin a vascular space in C and large tumor colonyin D. E and F show the appearance of antibodystaining of CD31. Both panels represent examplesof small tumor cell colonies surrounded by endo-thelial cells located within a larger vessel. E and Fshow CD31 immunohistochemical staining inbrown.

177




VEGF immunostaining of NeuYD;VEGF-25 animals supports thisview (data not shown).

The disruption of tight junction organization of luminal epithelialcells by oncogenic Neu correlates well with the increased vascular-ization of early hyperplasia. This early effect of activating Neu con-firms in vivo the disruption of epithelial cell polarity observed incultured epithelial cells (5) and mammary cell spheroid culture (32).Disruption of epithelial polarity may be a key transforming event (33,34). Tight junction protein expression and organization has beenimplicated previously in human breast cancer progression (35, 36) andmay be regulated at least in part by protein phosphatase 2A (37),which interacts with Src, which in turn is activated by Neu. However,evidence for this possible pathway in mammary epithelial cells in vivorequires experimental confirmation.

The E-cadherin tumor suppressor protein continued to be expressedat tumor cell surfaces of NeuYD tumors. The persistent expression ofE-cadherin may contribute to the compact growth behavior of NeuYDtumor cells. The increased metastasis of NeuYD;VEGF-25 tumorcells occurs despite persistent E-cadherin expression. However, it isnot known whether the expression of E-cadherin also reflects contin-ued signaling because E-cadherin signaling is also dependent oncytoplasmic catenin proteins.

In situ hybridization revealed that both the NeuYD oncogene andthe VEGF164 transgene are expressed in a mosaic pattern at thehyperplastic stage of tumor development. The differential expressionof the NeuYD transgene in some mammary epithelial cells is thelikely cause for the differential hyperplastic growth of some branchesof the epithelial tree observed in young NeuYD females. However, thecellular expression patterns of the NeuYD and VEGF overlap inhyperplastic regions of the mammary epithelium. These results indi-cate that although the mammary epithelial specificity of the MMTVpromoter is retained, the penetrance of cellular expression may de-pend on the particular integration site of the transgene. The disruptionof the epithelial junctional organization by NeuYD likely permitstransgenic VEGF to stimulate neighboring endothelial cells in the fatpad, resulting in close juxtaposition of vessels with the dysplasticepithelium. This greatly accelerates the development of tumors. Therelatively uniform expression of both VEGF and NeuYD RNA intumors may reflect either a selection for cells expressing both trans-genes or a preferential activation of the MMTV-driven transgenes. Wefavor the former because we have found no evidence of stimulation ofthe MMTV promoter by activated Neu or the polyomavirus middle Tantigen oncogenes (data not shown). However, the details of epige-netic restriction of integrated MMTV promoter transgenes remain tobe elucidated.

The increased vascularization of NeuYD tumors by transgenicVEGF was confirmed with CD31 and T-cadherin staining as well asRNA analysis. T-cadherin was first implicated in neuronal guidance(38) but was also identified as a differentially expressed gene asso-ciated with tumor vasculature (39). More recent data implicate T-cadherin in pathological angiogenesis.6 The up-regulation of T-cadherin in the tumors parallels the vascular endothelial-cadherin,CD31, and fetal liver kinase-1 endothelial marker genes and raises thepossibility that T-cadherin may be functionally important in tumorangiogenesis.

The increased tumor vasculature of NeuYD;VEGF-25 tumors hasat least three major consequences. First, NeuYD tumor formation isgreatly accelerated. NeuYD;VEGF-25 transgenic tumors grow with adistinctive morphology of nests of tumor cells essentially surrounded

by endothelial cells. The distinctive morphology of VEGF-acceleratedtumors resembles that of mammary tumor cell lines selected forincreased metastasis (22). This tumor morphology is derived from anearly hyperplastic state in which vessels interdigitate small groups oftumor cells. The tumor expands in association with the vasculature.The increased association of endothelial cells with the tumor cells andthe generally increased vascularity likely provide the environmentnecessary for budding of tumor cell emboli into the vasculaturespaces.

The second distinctive characteristic of NeuYD;VEGF-25 tumors isthe RNA expression profile, which reflects an increased perfusion ofthe tumors. Increased vascular endothelial-cadherin, T-cadherin, andCD31 RNAs reflect the increased vessel density. Phosphoglyceratekinase, glyceraldehydes 3-phosphate dehydrogenase, and endogenousVEGF RNAs are known to be regulated at least in part by hypoxiathrough the hypoxia-inducible factor-1� transcription factor. Induc-tion of endogenous VEGF RNA in control NeuYD tumors was foundin close proximity with regions undergoing apoptosis, presumablybecause of hypoxia and/or nutrient restriction. The relative decreasesin phosphoglycerate kinase, glyceraldehydes 3-phosphate dehydro-genase, and endogenous VEGF RNA expression likely reflect in-creased oxygen and/or nutrient availability in NeuYD;VEGF-25 tu-mors. However, the glucose transporter 1 mRNA appeared to beexpressed similarly in NeuYD and NeuYD;VEGF-25 tumors althoughglucose transporter 1 RNA is regulated in part by hypoxia-induciblefactor-1� (40). This may reflect a combination of increased perfusionand a dominant effect of oncogene transcriptional activation (41, 42).The hypoxic response of phosphoglycerate kinase, glyceraldehydes3-phosphate dehydrogenase, and endogenous VEGF may be lesssensitive to oncogenic stimulation.

Although the increased progression, growth rate, and gene expres-sion pattern of NeuYD;VEGF-25 tumors is consistent with an angio-genic stimulation of these tumors, VEGF164 may have additionaleffects on nonendothelial cells. For example, high-level VEGF ex-pression from the MMTV promoter results in male sterility associatedwith VEGFR receptor (VEGFR) 1 expression in certain spermato-genic cells and VEGFR1 and VEGFR2 in Leydig cells of the testis(31). Males of the VEGF-89 line had the same phenotype (data notshown). Furthermore, VEGF165 binds to neuropilin-1 that cooperateswith VEGFR2 (43) and can directly effect the survival of some humanbreast cancer cell lines (44). It is possible that VEGF164 could pro-mote NeuYD tumor growth directly. However, immunohistochemicalstaining for VEGFR2 was found in endothelial cells not tumor cells(data not shown), and we have not found a stimulatory effect of VEGFon the growth of NeuYD tumor cells in cell culture (data not shown).These observations and the decreased expression of hypoxic-respon-sive genes support increased vessel density as the primary mecha-nisms of NeuYD tumor acceleration.

The third consequence of increased vascularization of NeuYD;VEGF-25 tumors is the increase in metastasis of tumor cell emboli tothe lung. This process closely resembles the invasion-independentmetastasis described previously for highly vascularized and metastatictumors (22). Under a heavy tumor burden at 30–60 days after tumordetection, NeuYD tumors do not metastasize efficiently to the lung(4). In this study when mice were sacrificed earlier, lung metastaseswere not detected in NeuYD animals. However, many small emboli oftumor cells were found in VEGF;NeuYD-25 lungs. Mitotic figuresand some larger colonies indicate tumor cells continue to grow ex-pansively. However, it remains to be determined whether such tumorswill be invasive because their continued expression of VEGF164

would be expected to continue to attract endothelial cells. Althoughincreased vascularity clearly facilitates metastasis in NeuYD animals,forced VEGF expression did not cause pancreatic tumors to metasta-

6 L. W. Hebbard, M. Garlatti, L. J. T. Young, R. D. Cardiff, R. G. Oshima, andB. Ranscht. T-cadherin supports mouse mammary tumor growth through an angiogenicmechanism, manuscript in preparation.

178




size (13). However, the vessel density of pancreatic tumors withsupplementary transgenic VEGF was not increased, perhaps becauseof the retention of epithelial polarity in SV40 T antigen-transformedcells or lower levels of VEGF expression. Without increased vascu-larity, increased metastasis might not be expected.

Increased vascularization of NeuYD-initiated tumors is sufficientfor accelerated colonization of distal organs. This metastatic processappears to differ from the current paradigm of invasion of the vascu-lature by tumor cells and extravasation at distal sites. The invasion-independent model (22) is consistent with the persistent expression ofthe E-cadherin adhesive protein in NeuYD;VEGF-25 tumors. Inva-sion-independent metastasis may represent an important contributionto the well-documented correlation between vascular density andmetastasis.

ACKNOWLEDGMENTS

We thank Jacqueline Avis for generating transgenic mice by pronuclearinjection and Abraham Gomez for expert care of animals.

REFERENCES

1. Muller, W. J., Sinn, E., Pattengale, P. K., Wallace, R., and Leder, P. Single-stepinduction of mammary adenocarcinoma in transgenic mice bearing the activatedc-neu oncogene. Cell, 54: 105–115, 1988.

2. Siegel, P. M., Hardy, W. R., and Muller, W. J. Mammary gland neoplasia: insightsfrom transgenic mouse models. Bioessays, 22: 554–563, 2000.

3. Dankort, D., Jeyabalan, N., Jones, N., Dumont, D. J., and Muller, W. J. MultipleErbB2/Neu phosphorylation sites mediate transformation through distinct effectorproteins. J. Biol. Chem., 276: 38921–38928, 2001.

4. Dankort, D., Maslikowski, B., Warner, N., Kanno, N., Kim, H., Wang, Z., Moran,M. F., Oshima, R. G., Cardiff, R. D., and Muller, W. J. Grb2 and Shc adapter proteinsplay distinct roles in Neu (ErbB2)- induced mammary tumorigenesis: implications forhuman breast cancer. Mol. Cell. Biol., 21: 1540–1551, 2001.

5. Khoury, H., Dankort, D. L., Sadekova, S., Naujokas, M. A., Muller, W. J., and Park,M. Distinct tyrosine autophosphorylation sites mediate induction of epithelial mes-enchymal like transition by an activated ErbB2/Neu receptor. Oncogene, 20: 788–799, 2001.

6. Gale, N. W., and Yancopoulos, G. D. Growth factors acting via endothelial cell-specific receptor tyrosine kinases: VEGFs, angiopoietins, and ephrins in vasculardevelopment. Genes Dev., 13: 1055–1066, 1999.

7. Yancopoulos, G. D., Davis, S., Gale, N. W., Rudge, J. S., Wiegand, S. J., and Holash,J. Vascular-specific growth factors and blood vessel formation. Nature (Lond.), 407:242–248, 2000.

8. Grunstein, J., Masbad, J. J., Hickey, R., Giordano, F., and Johnson, R. S. Isoforms ofvascular endothelial growth factor act in a coordinate fashion to recruit and expandtumor vasculature. Mol. Cell. Biol., 20: 7282–7291, 2000.

9. Thurston, G., Suri, C., Smith, K., McClain, J., Sato, T. N., Yancopoulos, G. D., andMcDonald, D. M. Leakage-resistant blood vessels in mice transgenically overexpress-ing angiopoietin-1. Science (Wash. DC), 286: 2511–2514, 1999.

10. Folkman, J., Watson, K., Ingber, D., and Hanahan, D. Induction of angiogenesisduring the transition from hyperplasia to neoplasia. Nature (Lond.), 339: 58–61,1989.

11. Inoue, M., Hager, J. H., Ferrara, N., Gerber, H. P., and Hanahan, D. VEGF-A has acritical, nonredundant role in angiogenic switching and pancreatic � cell carcinogen-esis. Cancer Cell, 1: 193–202, 2002.

12. Bergers, G., Brekken, R., McMahon, G., Vu, T. H., Itoh, T., Tamaki, K., Tanzawa,K., Thorpe, P., Itohara, S., Werb, Z., and Hanahan, D. Matrix metalloproteinase-9triggers the angiogenic switch during carcinogenesis. Nat. Cell Biol., 2: 737–744,2000.

13. Gannon, G., Mandriota, S. J., Cui, L., Baetens, D., Pepper, M. S., and Christofori, G.Overexpression of vascular endothelial growth factor-A165 enhances tumor angio-genesis but not metastasis during �-cell carcinogenesis. Cancer Res., 62: 603–608,2002.

14. Coussens, L. M., Raymond, W. W., Berger, G., Laig-Webster, M., Behrendtsen, O.,Werb, Z., Caughey, G. H., and Hanahan, D. Inflammatory mast cells up-regulateangiogenesis during squamous epithelial carcinogenesis. Genes Dev., 13: 1382–1397,1999.

15. Yen, L., You, X. L., Al Moustafa, A. E., Batist, G., Hynes, N. E., Mader, S., Meloche,S., and Alaoui-Jamali, M. A. Heregulin selectively upregulates vascular endothelialgrowth factor secretion in cancer cells and stimulates angiogenesis. Oncogene, 19:3460–3469, 2000.

16. Rak, J., Mitsuhashi, Y., Sheehan, C., Tamir, A., Viloria-Petit, A., Filmus, J., Mansour,S. J., Ahn, N. G., and Kerbel, R. S. Oncogenes and tumor angiogenesis: differentialmodes of vascular endothelial growth factor up-regulation in ras-transformed epithe-lial cells and fibroblasts. Cancer Res., 60: 490–498, 2000.

17. Laughner, E., Taghavi, P., Chiles, K., Mahon, P. C., and Semenza, G. L. HER2 (neu)signaling increases the rate of hypoxia-inducible factor 1� (HIF-1�) synthesis: novel

mechanism for HIF-1-mediated vascular endothelial growth factor expression. Mol.Cell. Biol., 21: 3995–4004, 2001.

18. Siegel, P. M., Ryan, E. D., Cardiff, R. D., and Muller, W. J. Elevated expression ofactivated forms of Neu/ErbB2 and ErbB-3 are involved in the induction of mammarytumors in transgenic mice: implications for human breast cancer. EMBO J., 18:2149–2164, 1999.

19. Webster, M. A., Hutchinson, J. N., Rauh, M. J., Muthuswamy, S. K., Anton, M.,Tortorice, C. G., Cardiff, R. D., Graham, F. L., Hassell, J. A., and Muller, W. J.Requirement for both Shc and phosphatidylinositol 3� kinase signaling pathways inpolyomavirus middle T-mediated mammary tumorigenesis. Mol. Cell. Biol., 18:2344–2359, 1998.

20. Saaristo, A., Karpanen, T., and Alitalo, K. Mechanisms of angiogenesis and their usein the inhibition of tumor growth and metastasis. Oncogene, 19: 6122–6129, 2000.

21. Liotta, L. A., and Kohn, E. C. The microenvironment of the tumour-host interface.Nature (Lond.), 411: 375–379, 2001.

22. Sugino, T., Kusakabe, T., Hoshi, N., Yamaguchi, T., Kawaguchi, T., Goodison, S.,Sekimata, M., Homma, Y., and Suzuki, T. An invasion-independent pathway ofblood-borne metastasis: a new murine mammary tumor model. Am. J. Pathol., 160:1973–1980, 2002.

23. Sympson, C. J., Talhouk, R. S., Alexander, C. M., Chin, J. R., Clift, S. M., Bissell,M. J., and Werb, Z. Targeted expression of stromelysin-1 in mammary gland providesevidence for a role of proteinases in branching morphogenesis and the requirement foran intact basement membrane for tissue-specific gene expression. J. Cell Biol., 125:681–693, 1994.

24. Wen, F., Cecena, G., Munoz-Ritchie, V., Fuchs, E., Chambon, P., and Oshima, R. G.Expression of conditional cre recombinase in epithelial tissues of transgenic mice.Genesis, 35: 100–106, 2003.

25. Soemarwoto, I. N., and Bern, H. A. The effects of hormones on the vascular patternof the mouse mammary gland. Am. J. Anat., 103: 403–435, 1958.

26. Trevor, K., and Oshima, R. G. Preimplantation mouse embryos and liver express thesame type I keratin gene product. J. Biol. Chem., 260: 15885–15891, 1985.

27. Braissant, O., and Wahli, W. Differential expression of peroxisome proliferator-activated receptor-�, -�, and -� during rat embryonic development. Endocrinology,139: 2748–2754, 1998.

28. Chomcznski, P., and Sacchi, N. Single-step method of RNA isolation by acidguanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem., 162: 156–159, 1987.

29. Rozen, S., and Skaletsky, H. Primer3 on the WWW for general users and for biologistprogrammers,. In: S. Krawetz and S. Misener (eds.), Bioinformatics Methods andProtocols: Methods in Molecular Biology, pp. 365–386. Totowa, New Jersey: Hu-mana Press, 2000.

30. Feroze-Merzoug, F., Berquin, I. M., Dey, J., and Chen, Y. Q. Peptidylprolyl isomer-ase A (PPIA) as a preferred internal control over GAPDH and �-actin in quantitativeRNA analyses. Biotechniques. 32: 776–778, 780, 782, 2002.

31. Korpelainen, E. I., Karkkainen, M. J., Tenhunen, A., Lakso, M., Rauvala, H., Vierula,M., Parvinen, M., and Alitalo, K. Overexpression of VEGF in testis and epididymiscauses infertility in transgenic mice: evidence for nonendothelial targets for VEGF.J. Cell Biol., 143: 1705–1712, 1998.

32. Muthuswamy, S. K., Li, D., Lelievre, S., Bissell, M. J., and Brugge, J. S. ErbB2, butnot ErbB1, reinitiates proliferation and induces luminal repopulation in epithelialacini. Nat. Cell Biol., 3: 785–792, 2001.

33. Bissell, M. J., and Bilder, D. Polarity determination in breast tissue: desmosomaladhesion, myoepithelial cells, and laminin 1. Breast Cancer Res., 5: 117–119, 2003.

34. Bilder, D., Li, M., and Perrimon, N. Cooperative regulation of cell polarity andgrowth by Drosophila tumor suppressors. Science (Wash. DC), 289: 113–116, 2000.

35. Hoover, K. B., Liao, S. Y., and Bryant, P. J. Loss of the tight junction MAGUK ZO-1in breast cancer: relationship to glandular differentiation and loss of heterozygosity.Am. J. Pathol., 153: 1767–1773, 1998.

36. Kominsky, S. L., Argani, P., Korz, D., Evron, E., Raman, V., Garrett, E., Rein, A.,Sauter, G., Kallioniemi, O. P., and Sukumar, S. Loss of the tight junction proteinclaudin-7 correlates with histological grade in both ductal carcinoma in situ andinvasive ductal carcinoma of the breast. Oncogene, 22: 2021–2033, 2003.

37. Nunbhakdi-Craig, V., Machleidt, T., Ogris, E., Bellotto, D., White, C. L. III, andSontag, E. Protein phosphatase 2A associates with and regulates atypical PKC and theepithelial tight junction complex. J. Cell Biol., 158: 967–978, 2002.

38. Ranscht, B., and Dours-Zimmermann, M. T. T-cadherin, a novel cadherin celladhesion molecule in the nervous system lacks the conserved cytoplasmic region.Neuron, 7: 391–402, 1991.

39. Wyder, L., Vitaliti, A., Schneider, H., Hebbard, L. W., Moritz, D. R., Wittmer,M., Ajmo, M., and Klemenz, R. Increased expression of H/T-cadherin in tumor-penetrating blood vessels. Cancer Res., 60: 4682–4688, 2000.

40. Ryan, H. E., Lo, J., and Johnson, R. S. HIF-1 � is required for solid tumor formationand embryonic vascularization. EMBO J., 17: 3005–3015, 1998.

41. Murakami, T., Nishiyama, T., Shirotani, T., Shinohara, Y., Kan, M., Ishii, K., Kanai,F., Nakazuru, S., and Ebina, Y. Identification of two enhancer elements in the geneencoding the type 1 glucose transporter from the mouse which are responsive toserum, growth factor, and oncogenes. J. Biol. Chem., 267: 9300–9306, 1992.

42. Dang, C. V., and Semenza, G. L. Oncogenic alterations of metabolism. TrendsBiochem. Sci., 24: 68–72, 1999.

43. Mamluk, R., Gechtman, Z., Kutcher, M. E., Gasiunas, N., Gallagher, J., andKlagsbrun, M. Neuropilin-1 binds vascular endothelial growth factor 165, placentagrowth factor-2, and heparin via its b1b2 domain. J. Biol. Chem., 277: 24818–24825,2002.

44. Bachelder, R. E., Crago, A., Chung, J., Wendt, M. A., Shaw, L. M., Robinson, G., andMercurio, A. M. Vascular endothelial growth factor is an autocrine survival factor forneuropilin-expressing breast carcinoma cells. Cancer Res., 61: 5736–5740, 2001.

179




2004;64:169-179. Cancer Res Robert G. Oshima, Jacqueline Lesperance, Varinia Munoz, et al. Progression and MetastasisAngiogenic Acceleration of Neu Induced Mammary Tumor

Updated version

http://cancerres.aacrjournals.org/content/64/1/169

Access the most recent version of this article at:

Cited articles

http://cancerres.aacrjournals.org/content/64/1/169.full.html#ref-list-1

This article cites 43 articles, 21 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/64/1/169.full.html#related-urls

This article has been cited by 12 HighWire-hosted articles. Access the articles at:

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

SubscriptionsReprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

[email protected] at

To request permission to re-use all or part of this article, contact the AACR Publications


http://cancerres.aacrjournals.org/content/64/1/169

http://cancerres.aacrjournals.org/content/64/1/169.full.html#ref-list-1

http://cancerres.aacrjournals.org/content/64/1/169.full.html#related-urls

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

mailto:[email protected]


Documents

Angiogenic Acceleration of Neu Induced Mammary Tumor Progression and Metastasis