2 BOL ET AL. MOLECULAR CARCINOGENESIS 21:2–12 (1998)

© 1998 WILEY-LISS, INC.

Severe Follicular Hyperplasia and Spontaneous PapillomaFormation in Transgenic Mice Expressing the NeuOncogene Under the Control of the Bovine Keratin5 PromoterDavid Bol,1 Kaoru Kiguchi,1 Linda Beltrán,1 Tim Rupp,1 Samantha Moats,1 Irma Gimenez-Conti,1 José Jorcano,2 andJohn DiGiovanni1*1Department of Carcinogenesis, The University of Texas M. D. Anderson Cancer Center, Science Park–Research Division,Smithville, Texas

2Department of Cell and Molecular Biology, Ciemat Instituto, Madrid, Spain

Transgenic mice were developed to explore the role of the erbB2 during epithelial homeostasis and tum-origenesis, through targeted expression of the neu oncogene (neu*). Expression of a neu* cDNA was tar-geted to the basal layer of skin epidermis as well as other epithelial tissues of transgenic mice via the bovinekeratin 5 promoter. Two transgenic founders were obtained that were morphologically distinguishable fromnon-transgenic littermates by their visibly thickened skin and patchy hair growth by day 3 after birth. Thepresence of the transgene was confirmed by polymerase chain reaction analysis of tail DNA and immunofluo-rescence analysis of neu* protein in skin sections. Histological evaluation revealed significant hyperplasia ofthe follicular and interfollicular epidermis, the abnormal presence of horny material in the dermis and hypo-dermis, and a dramatic increase in epidermal proliferation. Many areas of the dermis involving this abnormalepithelial proliferation exhibited a squamous cell carcinoma–like appearance. In addition, there was unusualproliferation of the sebaceous glands. One founder died at day 14 and the other at day 20. The latter founderhad two papillomas at the time of death. Additional phenotypic changes resulting from the expression ofneu* in other tissues included hyperkeratosis in the forestomach and esophagus. In addition, there was alack of distinction of the cortical-medullary boundaries and an increased rate of cell death in lymphocytes inthe thymus. The phenotypic changes in these other tissues correlated with transgene expression. The datasuggest that erbB2 signaling has an important role in epidermal proliferation. In addition, the data providestrong support for a role for erbB2 signaling during epidermal carcinogenesis in mouse skin. Mol. Carcinog.21:2–12, 1998. © 1998 Wiley-Liss, Inc.

Key words: epithelial homeostasis; erbB2; neu; skin tumorigenesis

INTRODUCTION

Receptor tyrosine kinases are involved in growthcontrol, and alterations in several of these ligand-receptor systems may result in uncontrolled growthand oncogenic transformation [1,2]. One of the best-studied receptor tyrosine kinase families is the type1 growth factor receptors, also called the erbB fam-ily. The epidermal growth factor (EGF) receptor (EGFr,or erbB1) was the first member to be cloned andshows considerable homology to the avian erythro-blastosis virus transforming protein, v-erbB [3]. erbB2is the human homologue of the neu oncogene (neu*),which was initially isolated from a chemically in-duced rat neuroblastoma [4,5] and shares close ho-mology with EGFr [6]. The erbB family also includestwo more recently identified members, erbB3 [7,8]and erbB4 [9]. All erbB family members have a simi-lar primary structure with a ligand-binding extracel-lular domain, a transmembrane domain, and anintracellular domain that contains a highly conserved

tyrosine kinase domain and a carboxy terminus[1,10]. The binding of EGF to the EGFr leads to acti-vation of the intrinsic tyrosine kinase and phospho-rylation of multiple tyrosine residues on its carboxyterminus. Phosphorylation on the carboxy-terminal

The current address of David Bol is Department of Oncology,Bristol-Myers Squibb Pharmaceutical Research Institute, P.O. Box4000, Princeton, NJ 08543.

*Correspondence to: Department of Carcinogenesis, The Uni-versity of Texas M. D. Anderson Cancer Center, Science Park–Re-search Division, P.O. Box 389, Smithville, TX 78957.

Received 1 October 1997; Revised 30 October 1997; Accepted 3November 1997

Abbreviations: neu*, neu oncogene; EGF, epidermal growth fac-tor; EGFr, epidermal growth factor receptor; TGFα, transforminggrowth factor α; AR, amphiregulin; HB-EGF, heparin-binding EGF-like growth factor; TPA; 12-O-tetradecanoylphorbol-13-acetate;DMBA, 7,12-dimethylbenz[a]anthracene; BK5, bovine keratin 5,PCR, polymerase chain reaction; PBS, phosphate-buffered saline;BSA/PBS, PBS, pH 7.5, containing 1% bovine serum albumin; H&E,hematoxylin and eosin; BrdU, bromodeoxyuridine; K6, keratin 6;LI, labeling index; TUNEL, terminal deoxytransferase–mediated dUTPnick end labeling.

FAST TRACK

EXPRESSION OF NEU ONCOGENE 3

tyrosine residues is believed to occur through recep-tor dimerization (homodimerization, heterodi-merization, or both) and transphosphorylation [11].The phosphotyrosines then serve as recognition sitesfor src homology-2 domain–containing proteins, thebinding of which in turn triggers signal-transductionpathways from the membrane to the nucleus andother cellular locations [12]. Although all erbB fam-ily members have similar primary structures, recep-tor activation mechanisms, and signal transductionpatterns, as described above, they bind to differentligands. The EGFr binds to a family of peptides in-cluding EGF, transforming growth factor-α (TGFα),amphiregulin (AR), heparin-binding EGF-like growthfactor (HB-EGF), and betacellulin [reviewed in 10].erbB3 and erbB4 bind to neu differentiation factorand other members of the family of peptides calledheregulins, which consists of at least 10 distinctisoforms [13,14]. Recently, betacellulin was alsoshown to bind to erbB4 [15]. In contrast, despiteextensive research, a specific ligand for erbB2 has notyet been identified [16,17]. The absence of a specificligand for erbB2 has led to the hypothesis that thiserbB family member functions primarily by formingheterodimers with other erbB family members,thereby expanding the array of downstream signal-ing pathways activated by ligand binding [17]. In fact,it has been proposed that erbB2-EGFr heterodimersare the primary receptor dimer involved in mediat-ing EGF-induced signaling in certain cell types[11,17–20].

A number of recent studies, including work in ourlaboratory, provide considerable support for an im-portant role for signaling through the EGFr duringmultistage skin carcinogenesis. Initially, work in ourlaboratory led to the finding that topical applicationof 12-O-tetradecanoylphorbol-13-acetate (TPA) andthe non–phorbol ester tumor promoter chrysarobininduces the loss of epidermal protein kinase C, in-creases EGF binding to its receptor, and induces epi-dermal TGFa mRNA and protein synthesis in mouseepidermis [21]. We also found that diverse classes oftumor promoters (including okadaic acid andthapsigargin) increase epidermal TGFa mRNA andprotein levels after topical application [22] and alsoincrease the levels of other EGFr ligands such as ARand HB-EGF [23]. In addition, we found that the EGFris activated in the epidermis of tumor promoter–treated mice and that tyrosine kinase inhibitors blockthe EGFr activation and epidermal proliferation in-duced by TPA [24]. Furthermore, the mRNA and pro-tein expression of both EGFr ligands (including TGFα,AR, and HB-EGF) and the EGFr is constitutively in-creased in primary papillomas and squamous cellcarcinomas induced by an initiation-promotion regi-men [23,25]. Several other lines of evidence also pro-vide support for a critical role for the EGFr inmultistage skin carcinogenesis. First, several labora-tories have created transgenic mice that overexpress

TGFα in mouse epidermis [26–28]. Studies with thesemice have shown that TGFα overexpression can sub-stitute for Ha-ras activation and, more importantly,that TGFα overexpression in mouse epidermis leadsto epidermal hyperplasia [27–30]. Furthermore, oneof these studies showed that TGFa transgenic miceinitiated with 7,12-dimethylbenz[a]anthracene(DMBA) develop papillomas, whereas non-transgenicmice initiated with DMBA do not [28]. These latterdata, in combination with the observation that TGFαoverexpression leads to epidermal hyperplasia, indi-cate an important role for TGFα during the tumor-promotion stage of multistage skin carcinogenesis.In addition to these studies, Yuspa and colleagues[31] recently showed that EGFr-null keratinocytes arerelatively resistant to transformation by v-Ha-ras.Finally, Jorcano and colleagues [32] recently reportedthe creation of transgenic mice that overexpress adominant negative mutant of the EGFr in epidermisunder the control of the bovine keratin 5 (BK5) pro-moter. Signaling through the EGFr is significantlyreduced in skin keratinocytes, and these mice arerelatively resistant to two-stage carcinogenesis usingDMBA initiation and TPA promotion (Jorcano JL,unpublished studies). Collectively, this evidence in-dicates that signaling through the EGFr is importantin the tumor-promotion stage of multistage skin car-cinogenesis and that constitutive signaling throughthe EGFr contributes to the development of autono-mous growth in premalignant papillomas inducedin this carcinogenesis model system.

Recently, we discovered that erbB2 (but not erbB3)also becomes activated in EGF-stimulated keratin-ocytes, in TPA-treated epidermis, and in theepidermis of K14.TGFα transgenic mice [33]. In EGF-stimulated keratinocytes, we found an increased as-sociation of erbB2 with EGFr, suggesting that erbB2(but again not erbB3) becomes activated throughheterodimerization with the EGFr. These data raisedthe possibility that erbB2 activation is an integralcomponent of signaling through the EGFr in mousekeratinocytes and that erbB2 plays an important rolein EGFr signaling during tumor promotion. In thestudy reported here, we expressed the neu oncogene(referred to as neu*) in the epidermis of mice undercontrol of the BK5 promoter [32]. The goal of thisstudy was to determine the role of erbB2 signalingin epidermal homeostasis and multistage skin car-cinogenesis in mice.

MATERIALS AND METHODS

Preparation of the DNA Construct

The DNA construct used to generate the transgenicmice is shown in Figure 1A. The neu* cDNA was ex-cised from the parent vector pSV-neu* (a gift fromDr. Mien-Chie Hung, The University of Texas M. D.Anderson Cancer Center) [34] with the restrictionenzymes HindIII and SaPI. The termini of this frag-

4 BOL ET AL.

ment were filled in with the Klenow fragment of DNApolymerase, and the blunt-ended fragment was li-gated into the SmaI restriction enzyme site of thekeratinocyte expression vector pBK5 [32]. The ori-entation and integrity of the cDNA insert was de-termined by a series of diagnostic restrictiondigests. The resulting construct contained the BK5promoter for transcription of the rabbit β-globinintron, the neu* cDNA, and the simian virus 40polyadenylation signal.

Preparation of DNA for Microinjection

Plasmid DNA was prepared by standard techniques[35]. pBK5.neu* was digested with EcoRI to releasethe vector sequences from the expression construct,and the two fragments were separated by electro-phoresis through a 0.8% agarose gel. The larger frag-ment containing the expression construct wasisolated from a gel slice by electrophoresis into a di-alysis bag, and the DNA was purified by using thePrepagene DNA cleanup kit according to manu-facturer’s recommendations (Bio-Rad Labs., Hercules,CA). The DNA was eluted from the purification resinwith 10 mM Tris-HCl, pH 7.6, and 0.1 mM EDTAand quantitated. For injection, the DNA was dilutedto 5 ng/µL in sterile water and centrifuged repeat-edly to remove residual resin.

Generation and Identification of Transgenic Mice

Donor embryos for injection were generated bymating of FVB male mice (National Cancer Institute,Research Facility, Frederick, MD) to superovulatedICR female mice (Harlan Sprague Dawley, Indianapo-lis, IN) and were isolated as previously described [36].DNA was injected into the pronuclei of these F1 em-bryos, and survivors were transferred to pseudopreg-nant ICR females. Transgenic animals were identifiedby polymerase chain reaction (PCR) of the DNA fromthe tails of weanlings [37]. The oligonucleotides spe-cific for the rabbit β-globin intron cDNA that we usedwere 5´-TTC AGG GTG TTG TTT AGA ATG G-3´ (up-stream) and 5´-CAA TAA GAA TAT TTC CAC GCC A-3´ (downstream), which yielded a 450-bp fragment ina PCR with the rabbit β-globin cDNA as a template.

Analysis of Transgene Expression

Transgene expression was determined by indirectimmunofluorescence analysis of histologic sectionsfrom various tissues. The expression and localizationof endogenous mouse erbB2 and neu* was determinedby immunofluorescence analysis of sections of dorsalskins and skin tumors. The tissues were fixed in for-malin and embedded in paraffin, and 4-µm sectionswere adhered to slides. After deparaffinization, theslides were microwaved twice for 5 min each time toenhance the staining of neu*. The sections were incu-bated with 10% non-immunized goat serum for 30min to block the nonspecific Fc receptor in the tissueand then washed three times with phosphate-bufferedsaline (PBS), pH 7.5, containing 1% bovine serum al-bumin (BSA/PBS). The sections were then incubatedeither with a 1:300 dilution of the primary rabbitpolyclonal antibody against the epitope correspond-ing to amino acids 1169–1186, which map to thecarboxy terminus of the precursor forms of humanneu gp 185 (Santa Cruz Biotechnology Inc., Santa Cruz,CA) in BSA/PBS or with preimmune rabbit serum asnegative controls for 40 min. After three washes withBSA/PBS, the sections were incubated with the sec-

Figure 1. (A) DNA construct used for generating K5.neu*transgenic mice. The construct contains the BK5 promoter, therabbit β-globin intron, the neu* cDNA insert, and the simianvirus 40 polyadenylation signal. (B) Transgenic founder(K5.neu*A) and corresponding non-transgenic littermate ob-tained from injection of the BK5.neu* construct at day 5 (up-per panel) and day 20 (lower panel).

EXPRESSION OF NEU ONCOGENE 5

ondary fluorescence (fluorescein isothiocyanate)–con-jugated affinity pure F(ab´)2 fragment goat anti–rabbitimmunoglobulin G (Jackson Immuno Research Lab,West Grove, PA; diluted 1:200) for 40 min. The sec-tions were then covered with Vectashield mountingsolution (Vector Labs. Inc., Burlingame, CA) before thecover slips were attached. The immunospecificity ofthe neu* immunoglobulin G was confirmed bypreabsorption with the antigen.

Histological Analyses

Dorsal skin samples, internal organs, and tumorswere fixed in formalin or ethanol and embedded inparaffin before sectioning. Sections of 4 µm were cutand stained with hematoxylin and eosin (H&E). Micewere injected intraperitoneally with bromodeoxy-uridine (BrdU) in PBS (100 µg/gm body weight) 30min before they were killed. For the analysis of epi-dermal labeling index (LI), paraffin sections werestained with anti-BrdU antibody as previously de-scribed [38]. Analysis of keratin 6 (K6) expression wasalso performed as previously described [39]. Themeasurements of epidermal thickness and LI wereperformed as previously described [40].

Terminal Deoxytransferase–Mediated dUTP NickEnd Labeling Assay

The individual apoptotic cells in formalin-fixedtissue sections were identified by the terminaldeoxytransferase–mediated dUTP nick end labeling(TUNEL) technique (In Situ Cell Death Detection Kit,Fluorescein; Boehringer Mannheim, Indianapolis, IN).

RESULTSGeneration of K5.neu* Transgenic Mice

To target expression of neu* to the epidermis oftransgenic mice, we made a construct by insertingthe neu* cDNA into the BK5 expression vector asshown in Figure 1A. This vector uses rabbit β-globingene intron sequences as well as the simian virus 40polyadenylation signals to direct optimal expressionof the inserted cDNA specifically to the epidermalbasal cells of the skin [41]. A 12-kb KpnI recombi-nant fragment was purified and injected into one-cell embryos from (FVB × ICR)F1 mice. Two founderswere shown to carry the transgene by PCR analysisof tail DNA. Both of these founders presented mor-phological differences that distinguished them fromnon-transgenic animals by day 3. The gross physicalcharacteristics of both founders were similar, as wasthe appearance of their skins at the microscopic level(see below). One of the founders, designatedK5.neu*B, died at 14 d of age, and the other founder,designated K5.neu*A, died at 20 d of age.

Physical Characteristics Resulting fromTransgene Expression

To confirm expression of the K5.neu* transgene,indirect immunofluorescence analysis of neu* pro-

tein was performed (see below). The presence of thetransgene in genomic DNA (detected by PCR; datanot shown) and the expression of the transgene cor-related with obvious physical characteristics in thetwo founders. Specifically, within 5 d after birth, theskins of both founders exhibited areas thicker thanthe skins of the controls (Figure 1B, top panel). Inaddition, hair growth was patchy on the transgenicmice but not on non-transgenic littermates at thistime. Both founders eventually developed relativelythin-looking hair coats, and by approximately 14 dbegan to exhibit significant alopecia. In addition,both founders exhibited weight loss starting at about10 d of age. Figure 1B (bottom panel) shows thephysical appearance of founder K5.neu*A at day 20.This mouse also had two papillomas, one on thesnout and one on the right hind leg (both tumorsare visible in the bottom panel of Figure 1B). Thegross appearance of these tumors resembled that oftypical exophytic papillomas generated by standardinitiation-promotion regimens [42].

Histological Evaluation of Skin from K5.neu*Transgenic Mice

Histologic evaluation of skin from both transgenicfounders at the times of their deaths showed dra-matic alterations compared with non-transgenic lit-termates. Figure 2 shows the histologic features ofskin from founder K5.neu*A. There was marked fol-licular hyperplasia and a more moderate hyperpla-sia of the interfollicular epidermis (Figure 2B). Inaddition, there was a significant amount of hornymaterial in the dermis and hypodermis, althoughthere was little evidence of hyperkeratosis in the stra-tum corneum of the interfollicular epidermis. Insome areas, epidermal proliferation was observeddown to and within the muscle layer. Many areas ofthe dermis involving this abnormal epithelial prolif-eration exhibited a squamous cell carcinoma–likeappearance. Another striking feature of the skin fromK5.neu* transgenic mice was a marked enlargementof the sebaceous glands. The hair on both transgenicfounders was somewhat more oily and matted, alsosuggesting overproduction of sebaceous secretionsin these mice.

Figure 2C and D shows the BrdU-stained sectionsfrom an age-matched non-transgenic littermate andfounder K5.neu*A, respectively. The LI was signifi-cantly higher in the interfollicular epidermis of thisfounder than in its non-transgenic littermate (21.2± 3.4% vs 2.8 ± 0.7%, respectively). The LI of epider-mal cells lining the sebaceous glands was also el-evated. Furthermore, there was significantlyincreased proliferation of epidermal cells invadingthe hypodermis. Overall, the skin was markedlythicker in the transgenic mice than in the non-transgenic mice. A particularly striking feature oftransgenic skin observed during this analysis was thelarge number of proliferating epidermal cells in the

6 BOL ET AL.

Figure 2. Histological analysis and immunostaining for erbB2,neu*, and K6 in the skin of transgenic founder K5.neu*A and anage-matched non-transgenic littermate (control) at 20 d of age.H&E stain (150×) of a dorsal skin section from the control (A) andfounder K5.neu*A (B); BrdU-labeled cells (150×) of control (C)

and founder K5.neu*A (D); staining for K6 (150×) in the control(E) and founder K5.neu*A (F); staining for endogenous erbB2(300×) in the control (G) and endogenous erbB2 + neu* infounder K5.neu*A (H).

EXPRESSION OF NEU ONCOGENE 7

hypodermis. As a result, the fat cell layer could notbe readily discerned. To further evaluate the extentof epidermal proliferation induced by expression ofneu* in epidermal basal cells, we examined the ex-pression of K6 in skin sections as previously described[39]. Although K6 is normally expressed in the outerroot sheath of hair follicles of adult mice, as shownin Figure 2E, the expression of K6 can be induced inthe interfollicular epidermis by wounding or treat-ment with proliferative stimuli such as TPA [43–45].Figure 2F shows that expression of neu* in epider-mal basal cells induced expression of K6 both in theinterfollicular epidermis and in those epidermal cellsundergoing proliferation in the dermis, especiallythose cells invading the hypodermis. These resultsconfirm that expression of neu* in epidermal basalcells led to increased proliferation.

Transgene expression was verified in sections ofdorsal skin by indirect immunofluorescence analy-sis (Figure 2G and H). Figure 2G shows the patternof expression of endogenous mouse erbB2 in the skinof a 20-d-old non-transgenic littermate. EndogenouserbB2 was expressed in all epidermal layers of theinterfollicular epidermis, in epidermal cells lining theouter root sheath of hair follicles, and in sebaceousglands. In founder K5.neu*A, transgene expression

was clearly observed in the epidermal cells invadingthe dermis and hypodermis. Transgene expressionin the interfollicular epidermis and hair follicles wassomewhat more difficult to ascertain because of thestaining of endogenous erbB2; however, the fluores-cence intensity of cells in the basal layer was stron-ger, consistent with transgene expression in thisepidermal compartment. In founder K5.neu*B, simi-lar results were obtained (data not shown). Thus,transgene expression correlated with skin phenotypein both founders.

As noted above, the founder that died at age 20 d(K5.neu*A) had two papillomas at the time of death.Figure 3 shows H&E sections from both of these tu-mors. Both tumors exhibited the typical features ofsquamous papillomas induced by standard initiation-promotion regimens in mice. In addition, transgeneexpression in the leg tumor (Figure 3C) was com-pared to endogenous erbB2 expression in a skin pap-illoma induced by an initiation-promotion regimenin SENCAR mice (Figure 3D). Expression of endog-enous erbB2 was confined to the epithelial compo-nent (mostly the basal layer) of the tumor inducedby initiation of SENCAR mice with DMBA (25 nmol)followed by promotion with TPA (1.7 nmol) for 25wk. Intense staining for neu* was observed in the

Figure 3. Histological analysis and immunostaining forerbB2 in tumors from the transgenic founder K5.neu*A. (A)H&E stain of facial papilloma from the K5.neu*A transgenicfounder (75×); (B) H&E stain of the leg papilloma from the

K5.neu*A transgenic founder (75×); (C) staining for neu* in theleg papilloma (300×); (D) staining for erbB2 in papilloma froma SENCAR mouse initiated with DMBA (25 nmol) followed bypromotion with TPA (1.7 nmol twice weekly) for 25 wk.

8 BOL ET AL.

epithelial component of the leg tumor from founderK5.neu*A. Based on the fluorescence intensity in tis-sue sections at the same magnification, the level oftransgene expression was at least fivefold higher thanthe endogenous erbB2 staining.

Histologic Features of Forestomach and Thymus ofK5.neu* Transgenic Mice

The BK5 promoter drives expression in a varietyof mouse epithelial tissues in addition to skin, in-cluding thymus, esophagus, and forestomach [41].

These tissues were also examined for possible effectsof neu* expression, as shown in Figures 4 and 5.Figure 4 shows cross sections of the forestomach offounder K5.neu*B (panel B) and a non-transgeniclittermate (panel A). The forestomach exhibited sig-nificant hyperkeratosis and a thinning of the epi-thelial layer. In some sections, the hyperkeratosiswas particularly pronounced. The thinning of theepithelial layer appeared to be accompanied bypicnotic nuclei in most of the cells along the base-ment membrane, which suggests increased apop-

Figure 4. Histological analysis and immunostaining forerbB2 and neu* in the forestomach of transgenic founderK5.neu*B at day 14 and in an age-matched non-transgenic lit-termate (control). H&E stain (150×) of the control (A) and

founder K5.neu*B (B); staining for endogenous erbB2 in thecontrol (C, 300×) and for endogenous erbB2 + neu* in founderK5.neu*B (D, 300×); and staining of the section in panel D witherbB2 antibody with neutralizing peptide (E, 300×).

EXPRESSION OF NEU ONCOGENE 9

tosis. This effect of transgenic expression may havecontributed to the rapid wasting observed in theseanimals. Similar histologic changes were also ob-served in the esophagus (data not shown). Transgeneexpression (Figure 4D) correlated with these phe-notypic changes in the forestomach and esophagus.

Figure 5 shows H&E sections of the thymuses of anon-transgenic littermate and transgenic founderK5.neu*A (panels A and B, respectively). The mostpronounced feature of the thymus from founderK5.neu*A was the lack of distinction between thecortical and medullary regions. In addition, there

appeared to be a reduced number of lymphocytes.Analysis of transgene expression by indirect immu-nofluorescence analysis revealed weak staining forerbB2 both in the cortical region and to a slightlyhigher degree in the medullary region of the thy-mus of the non-transgenic mouse (Figure 5C). TheBK5 promoter also drives expression of neu*throughout the thymus, as shown in Figure 5D. TheTUNEL assay for apoptotic cells confirmed a signifi-cant increase in the number of lymphocytes under-going apoptosis as a result of transgene expression(Figure 5F). Thus, transgene expression produced a

Figure 5. Histological analysis, immunostaining for erbB2and neu*, and TUNEL assay of the thymus of the transgenicfounder K5.neu*A and of an age-matched non-transgenic lit-termate (control). H&E stain (75×) of the control (A) and

founder K5.neu*A (B), staining for endogenous erbB2 in thecontrol (C, 150×) and for endogenous erbB2 + neu* in founderK5.neu*A (D, 150×); TUNEL assay of the control (E, 150×) andfounder K5.neu*A (F, 150×).

10 BOL ET AL.

significant phenotype in the forestomach, esopha-gus, and thymus.

DISCUSSION

This study was designed to evaluate the role oferbB2 signaling in epidermal homeostasis and itspotential role during epidermal carcinogenesis. Toachieve these goals, neu* was expressed in skin epi-dermis under the control of the BK5 promoter. Themajor findings of this study are as follows: (i) ex-pression of neu* under the control of BK5 produceda dramatic phenotype characterized by severe folli-cular epidermal proliferation and a marked overpro-duction of keratin in the dermal compartment, withmany areas of the skin exhibiting a squamous cellcarcinoma–like appearance; (ii) expression of neu*also led to hyperplasia of the interfollicular epider-mis characterized by an increased LI but little or nohyperkeratosis; (iii) expression of neu* under thecontrol of BK5 led to spontaneous papilloma for-mation in founder K5.neu*A; (iv) expression of neu*under the control of BK5 led to hyperkeratosis anda thinning of the epithelial layer in the forestom-ach and esophagus; and (v) expression of neu* un-der the control of BK5 produced a phenotype in thethymus characterized by a lack of delineation of thecortical-medullary boundaries and an increased rateof cell death in lymphocytes. Overall, the data dem-onstrate that abnormal erbB2 signaling in a varietyof tissues produces significant phenotypic changes.In the skin, these phenotypic changes were strikingand provide strong support for an important role forerbB2 in skin carcinogenesis.

Alterations in erbB2 signaling have been impli-cated in neoplastic transformation in vitro and inneoplasia in both experimental animals and in hu-man cancer. Overexpression of erbB2 occurs in 20–30% of breast and ovarian cancers and correlateswith poor prognosis in both types of cancer [46,47].The transforming or oncogenic activity of neu hasbeen demonstrated in cell lines transfected with theproto-oncogene and in other transgenic mice over-expressing the neu oncogene. For example, overex-pression of the normal coding sequence of the rat(i.e., wild-type erbB2) neu gene by using long termi-nal repeat–based expression vectors in NIH/3T3 cellscauses cell transformation [48]. In addition,overexpression of neu (i.e., wild-type erbB2) underthe control of the mouse mammary tumor virus pro-moter in transgenic mice resulted in the appearanceof focal mammary tumors after long latency [49].Expression of the neu oncogene under control ofthe mouse mammary tumor virus promoter intransgenic mice resulted in multifocal mammary ad-enocarcinomas that involved the entire epitheliumand both mammary glands [50,51]. It was suggestedthat an appropriate level of oncogenic neu is suffi-cient for rapid production of mammary tumors in

transgenic mice [51]. Our results are consistent withthe latter observation in that the skin of bothtransgenic founders exhibited characteristics ofsquamous cell carcinomas. In addition, founderK5.neu*A developed skin papillomas before it diedat 20 d of age. Thus, in mouse epidermis an appro-priate level of neu* expression appears to be suffi-cient for tumor development.

A particularly interesting observation in our studywas the dramatic phenotype observed in the hairfollicles and the subsequent proliferation of thesecells in the dermis and hypodermis as a result oftransgene expression. Previous work, primarily inhuman skin, has shown expression of erbB2 in themore differentiated layers and in the outer rootsheath of hair follicles [52,53]. In our study, erbB2expression was observed in all epidermal layers ofnormal mouse epidermis (at age 20 d), includingthe outer root sheath of hair follicles and in seba-ceous glands. Consistent with this finding, previ-ous work in our laboratory demonstrated that erbB2protein is expressed in cultured epidermal basal cellsof adult mouse epidermis [33]. The strong prolifera-tive phenotype in both the hair follicles and theinterfollicular epidermis is consistent with the pat-tern of endogenous erbB2 protein expression weobserved. As noted in the Introduction, the EGFrappears to be an important partner in the functionof erbB2. EGFr expression was previously reportedin the basal layer, the outer root sheath cells of hairfollicles, the ducts of the eccrine sweat glands, andthe outer layer of sebaceous glands as well as sev-eral other structures in human skin [54,55]. Re-cently, we postulated that the interaction betweenthe EGFr and erbB is important in mediating theeffects of EGFr ligands in mouse keratinocytes andduring skin tumor promotion [33]. The presence ofEGFr in the same cells expressing endogenous erbB2further supports this hypothesis. In addition, the pres-ence of both endogenous erbB2 and EGFr in the samecells expressing high levels of neu* from the BK5 pro-moter may also play an important role in producingthe phenotype in these transgenic founders.

Transgenic models with increased signalingthrough the EGFr have been developed. In thesemodels, overexpression of TGFα driven by the hu-man keratin 14 promoter, which has an expressionpattern similar to BK5, produces dramatic hyper-plasia and hyperkeratosis in the interfollicular epi-dermis [26]. In addition, the dermis of these mice isthinner than those of control animals, but onlymodest phenotypic changes in hair follicles wereobserved. As noted above, in our recent studies wefound that both the EGFr and erbB2 were activatedin the epidermis of K14.TGFα transgenic mice [33],suggesting a role for erbB2 signaling in producingthe phenotype in these mice. While both theK14.TGFα transgenic mice and the K5.neu*transgenic mice exhibit epidermal hyperplasia, there

EXPRESSION OF NEU ONCOGENE 11

are major phenotypic differences in their skins, par-ticularly in the dermis. The dramatic changes seenin hair follicles and the accompanying downwardproliferation of follicular epidermal cells in K5.neu*founders were not observed in either of theK14.TGFα transgenic models produced [26,33].There are three possible explanations for these data:(i) the activation of erbB2 in K14.TGFα transgenicmice has no significant biologic consequences; (ii)overexpression of neu* leads to the activation of sig-naling pathways that are not activated when TGFαbinds to the EGFr and initiates EGFr-erbB2heterodimer formation; (iii) overexpression of neu*leads to activation of a completely abnormal set ofsignaling pathways not activated by endogenouserbB2. Future studies will address these possibilitiesby using a variety of approaches, including tissue-specific erbB2 knockout mice.

Another interesting finding from our study wasthe phenotypic changes observed in several otherepithelial tissues to which BK5 targets transgene ex-pression [41]. In particular, the esophagus, forestom-ach, and thymus all exhibited abnormal phenotypiccharacteristics, and the changes in the thymus werequite striking. Analysis of erbB2 expression in thethymus of a non-transgenic littermate by indirectimmunofluorescence showed very weak erbB2 stain-ing in both the cortical and medullary epithelial cells(Figure 5C). Thus, the transgene was expressed inthe same cells expressing endogenous erbB2. Otherstudies found no expression of erbB2 in either hu-man thymus [56] or rat thymus [57]. erbB2 was alsoexpressed in epithelial cells of the forestomach (Fig-ure 4C) and esophagus (data not shown). Transgeneexpression in the epithelial component of these tis-sues produced hyperkeratosis and a thinning of theepithelial layer. However, while expression of neu*in all three of these epithelial tissues produced phe-notypic changes, no evidence of neoplastic changeswas observed. Nevertheless, our data suggest an im-portant role for erbB2 in epithelial homeostasis inall three of these tissues that warrants further in-vestigation.

In conclusion, our data provide strong supportfor an important role of erbB2 signaling in the con-trol of cell proliferation in skin epidermis. In addi-tion, the data also support a role for erbB2 signalingin the development of tumors in this tissue. Tofurther explore the role of erbB2 signaling path-ways in multistage skin carcinogenesis, we are plac-ing both the wild-type erbB2 and oncogenic erbB2(neu*) under the control of different keratin pro-moters to obtain viable transgenic lines for moredetailed studies.

ACKNOWLEDGMENTS

We wish to thank Cheryl Carpenter for her excel-lent secretarial assistance in the preparation of thismanuscript. This research was supported by United

States Public Health Service Grant CA 57596, NIEHSCenter Grant ES07784, and M. D. Anderson CoreGrant CA 16672.

REFERENCES

1. Ullrich A, Schlessinger J. Signal transduction by receptors withtyrosine kinase activity. Cell 61:203–212, 1990.

2. Aaronson SA. Growth factors and cancer. Science 254:1146–1153, 1991.

3. Ullrich A, Coussens L, Hayflick JS, et al. Human epidermal growthfactor receptor cDNA sequence and aberrant expression of theamplified gene in A431 epidermoid carcinoma cells. Nature309:418–425, 1984.

4. Shih C, Shilo B-Z, Goldfarb MP, Dannenberg A, Weinberg RA.Passage of phenotypes of chemically transformed cells via trans-fection of DNA and chromatin. Proc Natl Acad Sci USA 76:5714–5718, 1979.

5. Shih C, Padhy LH, Murray M, Weinberg RA. Transforming genesof carcinomas and neuroblastomas introduced into mouse fi-broblasts. Nature 290:261–264, 1981.

6. Bargmann CI, Hung MC, Weinberg RA. The neu oncogene en-codes an epidermal growth factor receptor-related protein. Na-ture 319:226–230, 1986.

7. Kraus MH, Issing W, Miki T, Popescu NC, Aaronson SA. Isolationand characterization of ERBB3, a third member of the ERBB/epi-dermal growth factor receptor family: Evidence for overexpressionin a subset of human mammary tumors. Proc Natl Acad Sci USA86:9193–9197, 1989.

8. Plowman GD, Whitney GS, Neubauer MG, et al. Molecularcloning and expression of an additional epidermal growthfactor receptor-related gene. Proc Natl Acad Sci USA 87:4905–4909, 1990.

9. Plowman GD, Culouscou J-M, Whitney GS, et al. Ligand-specificactivation of HER4/p180erbB4, a fourth member of the epider-mal growth factor receptor family. Proc Natl Acad Sci USA90:1746–1750, 1993.

10. Mason S, Gullick WJ. Type 1 growth factor receptors: An over-view of recent developments. The Breast 4:11–18, 1995.

11. Lemmon MA, Schlessinger J. Regulation of signal transductiondiversity by receptor oligomerization. Trends Biochem Sci 19:459–463, 1994.

12. Pawson T. SH2 and SH3 domains in signal transduction. AdvCancer Res 64:87–110, 1994.

13. Holmes WE, Sliwkowski MX, Akita RW, et al. Identification ofheregulin, a specific activator of p185erbB2. Science 256:1205–1210, 1992.

14. Ben-Baruch N, Yarden Y. Neu differentiation factors: A family ofalternatively spliced neuronal and mesenchymal factors. Proc SocExp Biol Med 206:221–227, 1994.

15. Riese DJ 2nd, Bermingham Y, Van Raaij TM, Buckley S, PlowmanGD, Stern DF. Betacellulin activates the epidermal growth factorreceptor and erbB-4 and induces cellular response patterns dis-tinct from those stimulated by epidermal growth factor orneuregulin-beta. Oncogene 12(2):345–353, 1996.

16. Dougall WC, Qian X, Peterson NC, Miller MJ, Samanta A, GreeneMI. The neu-oncogene: Signal transduction pathways, transfor-mation mechanisms and evolving therapies. Oncogene 9:2109–2123, 1994.

17. Karunagaran D, Tzahar E, Beerli RR, et al. ErbB-2 is a commonauxiliary subunit of NDF and EGF receptors: Implications for breastcancer. EMBO J 15:254–264, 1996.

18. Goldman R, Levy R, Peles E, Yarden Y. Heterodimerization of theerbB-1 and erbB-2 receptors in human breast carcinomas cells:A mechanism for receptor transregulation. Biochemistry29:11024–11028, 1990.

19. Wada T, Qian X, Greene MI. Intermolecular association of thep185neu protein and EGF receptor modulates EGF receptor func-tion. Cell 61:1339–1347, 1990.

20. Earp HS, Dawson TL, Li X, Yu H. Heterodimerization and func-tional interaction between EGF receptor family members: A newsignaling paradigm with implications for breast cancer research.Breast Cancer Res Treat 35:115–132, 1995.

21. Imamoto A, Beltrán LM, DiGiovanni J. Evidence for autocrine/paracrine growth stimulation by transforming growth factor-αduring the process of skin tumor promotion. Mol Carcinog 4:52–60, 1991.

22. Kiguchi K, Beltrán LM, You J, Rho O, DiGiovanni J. Diverse tu-mor promoters elevate expression of transforming growth fac-

12 BOL ET AL.

tor α mRNA and protein in SENCAR mouse epidermis. MolCarcinog 12:225–235, 1995.

23. Kiguchi K, Beltrán LM, DiGiovanni J. Altered expression of epi-dermal growth factor receptor ligands during multistage skincarcinogenesis. Abstract. Proceedings of the American Associa-tion for Cancer Research 36:149, 1996.

24. Xian W, Kiguchi K, Imamoto A, Rupp T, Zilberstein A, DiGiovanniJ. Activation of the epidermal growth factor receptor by skintumor promoters and in skin tumors from SENCAR mice. CellGrowth Differ 6:1447–1455, 1995.

25. Rho O, Beltrán LM, Gimenez-Conti IB, DiGiovanni J. Alteredexpression of the epidermal growth factor receptor and trans-forming growth factor-α during multistage skin carcinogenesisin SENCAR mice. Mol Carcinog 11:19–28, 1994.

26. Vassar R, Fuchs E. Transgenic mice provide new insights into therole of TGFα during epidermal development and differentiation.Genes Dev 5:714–727, 1991.

27. Dominey AM, Wang X-J, King LE, et al. Targeted over-expressionof transforming growth factor α in the epidermis of transgenicmice elicits hyperplasia, hyperkeratosis, and spontaneous squa-mous papillomas. Cell Growth Differ 4:1071–1082, 1993.

28. Jhappan C, Takayama H, Dickson RB, Merlino G. Transgenic miceprovide genetic evidence that transforming growth factor α pro-motes skin tumorigenesis via H-Ras-dependent and H-Ras-inde-pendent pathways. Cell Growth Differ 5:385–394, 1994.

29. Vassar R, Hutton ME, Fuchs E. Transgenic overexpression of trans-forming growth factor α bypasses the need for c-Ha-ras muta-tions in mouse skin tumorigenesis. Mol Cell Biol 12:4643–4653,1992.

30. Wang X-J, Greenhalgh DA, Eckhardt JN, Rothnagel JA, RoopDR. Epidermal expression of transforming growth factor-α intransgenic mice: Induction of spontaneous and 12-O-tetradecanoylphorbol-13-acetate-induced papillomas via amechanism independent of Ha-ras activation or overexpression.Mol Carcinog 10:15–22, 1994.

31. Dlugosz AA, Hansen L, Cheng C, et al. Targeted disruption of EGFRimpairs keratinocyte response to v-rasHa. Abstract. Proceedings ofthe American Association for Cancer Research 36:149, 1996.

32. Murillas R, Larcher F, Conti CJ, Santos M, Ullrich A, Jorcano JL.Expression of a dominant negative mutant of epidermal growthfactor receptor in the epidermis of transgenic mice elicits strik-ing alterations in hair follicle development and skin structure.EMBO J 14(21):5216–5223, 1995.

33. Xian W, Rosenberg MP, DiGiovanni J. Activation of erbB2 and c-src in phorbol ester-treated mouse epidermis: Possible role inmouse skin tumor promotion. Oncogene 14:1435–1444, 1997.

34. Yu DH, Hung MC. Expression of activated rat neu oncogene issufficient to induce experimental metastasis in 3T3 cells.Oncogene 6(11):1991–1996, 1991.

35. Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Labora-tory Manual, Second Edition. Cold Spring Harbor LaboratoryPress, Cold Spring Harbor, NY 1994.

36. Hogan B, Beddington R, Costantini F, Lacy E. Manipulating theMouse Embryo, A Laboratory Manual, Second Edition. ColdSpring Harbor Laboratory Press, Cold Spring Harbor, NY, 1994.

37. Miller SA, Dykes DD, Polesky HF. A simple salting out procedurefor extracting DNA from human nucleated cells. Nucleic AcidsRes 16(3):1215, 1988.

38. Eldridge SR, Tilbury LF, Goldsworthy TL, Butterworth BE. Mea-surement of chemically induced cell proliferation in rodent liverand kidney: A comparison of 5-bromo-2´-deoxyuridine and[3H]thymidine administered by injection or osmotic pump. Car-cinogenesis 11(12):2245–2251, 1990.

39. Bol DK, Kiguchi K, Gimenez-Conti I, Rupp T, DiGiovanni J.Overexpression of insulin-like growth factor-1 induces hyperpla-

sia, dermal abnormalities, and spontaneous tumor formation intransgenic mice. Oncogene 14:1725–1734, 1997.

40. Naito M, Naito Y, DiGiovanni J. Comparison of the histologicalchanges in the skin of DBA/2 and C57BL/6 mice following expo-sure to various promoting agents. Carcinogenesis 8(12):1807–1815, 1987.

41. Ramírez A, Bravo A, Jorcano JL, Vidal M. Sequences 5´ of thebovine keratin 5 gene direct tissue- and cell-type-specific expres-sion of a lacZ gene in the adult and during development. Differ-entiation 58:53–64, 1994.

42. Aldaz CM, Conti CJ. The premalignant nature of mouse skinpapilloma: Histopathologic, cytogenetic, and biochemical evi-dence. In: Conti CJ, Klein-Szanto AJP, Slaga TJ (eds), Skin Tu-mors: Experimental and Clinical Aspects. Raven Press, NY, 1989,pp. 227–242.

43. Mansbridge JN, Knapp AM. Changes in keratinocyte maturationduring wound healing. J Invest Dermatol 89:253–262, 1987.

44. Toftgard R, Yuspa SH, Roop DR. Keratin gene expression in mouseskin tumors and in mouse skin treated with 12-O-tetra-decanoylphorbol-13-acetate. Cancer Res 45:5845–5850, 1985.

45. Molloy CJ, Laskin JD. Specific alterations in keratin biosynthesisin mouse epidermis in vivo and in explant culture following asingle exposure to the tumor promoter 12-O-tetradecanoyl-phorbol-13-acetate. Cancer Res 47:4674–4680, 1987.

46. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuireWL. Human breast cancer: Correlation of relapse and survivalwith amplification of Her-2/neu oncogene. Science 235:177–182, 1987.

47. Slamon DJ, Godolphin W, Jones LA, et al. Studies of the HER-2/neu protooncogene in human breast and ovarian cancer. Sci-ence 244:707–712, 1989.

48. DiMarco E, Pierce JH, Kingsley CL, DiFiore PP. Transformation ofNIH 3T3 cells by overexpression of the normal coding sequenceof the rat neu gene. Mol Cell Biol 10:3247–3252, 1990.

49. Guy CT, Webster MA, Schaller M, Parson TJ, Cardiff RD, MullerWJ. Expression of the neu protooncogene in the mammary epi-thelium of transgenic mice induces metastatic disease. Proc NatlAcad Sci USA 89:10578–10582, 1992.

50. Muller WJ, Sinn E, Wallace R, Pattengale PK, Leder P. Single-step induction of mammary adenocarcinoma in transgenicmice bearing the activated c-neu oncogene. Cell 54:105–115, 1988.

51. Guy CT, Cardiff RD, Muller WJ. Activated neu induces rapid tu-mor progression. J Biol Chem 271(13):7673–7678, 1996.

52. Maguire HC Jr, Jaworsky C, Cohen JA, Hellman M, Weiner DB,Greene MI. Distribution of neu (c-erbB-2) protein in human skin.J Invest Dermatol 89:786–790, 1989.

53. Dougall WC, Qian X, Peterson NC, Miller MJ, Samanta A, GreeneMI. The neu-oncogene: Signal transduction pathways, transfor-mation mechanisms and evolving therapies. Oncogene 9:2109–2123, 1994.

54. Green MR, Mycock C, Smith CG, Couchman JR. Biochemicaland ultrastructural processing of [125I] epidermal growth factorin rat epidermis and hair follicles: Accumulation of nuclear label.J Invest Dermatol 88(3):259–265, 1987.

55. Nanney LB, Magid M, Stoscheck CM, King LE Jr. Comparison ofepidermal growth factor binding and receptor distribution innormal human epidermis and epidermal appendages. J InvestDermatol 83(5):385–393, 1984.

56. Natali PG, Nicotra MR, Bigotti A, et al. Expression of the p185encoded by HER2 oncogene in normal and transformed humantissues. Int J Cancer 45:457–461, 1990.

57. Kokai Y, Cohen J, Drebin JA, Greene MI. Stage- and tissue-spe-cific expression of the neu oncogene in rat development. ProcNatl Acad Sci USA 84:8498–8501, 1987.