[CANCER RESEARCH 63, 3945–3954, July 15, 2003]

Antizyme Overexpression in Transgenic Mice Reduces Cell Proliferation,Increases Apoptosis, and Reduces N-Nitrosomethylbenzylamine-inducedForestomach Carcinogenesis1

Louise Y. Y. Fong, David J. Feith, and Anthony E. Pegg2

Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, Pennsylvania 19107 [L. Y. Y. F.], and Department of Cellular and Molecular Physiology, Pennsylvania StateUniversity College of Medicine, Hershey, Pennsylvania 17033 [D. J. F., A. E. P.]

ABSTRACT

Antizyme (AZ) is known to be a regulator of polyamine metabolism thatinhibits ornithine decarboxylase activity and polyamine transport, thus re-stricting polyamine levels. Transgenic mice with AZ expression targeted tothe basal cell layer of the forestomach epithelium by the keratin 5 promoterwere used to investigate whether AZ overexpression inhibited uncontrolledcell proliferation in zinc-deficient (ZD) mice and reduced their susceptibilityto forestomach carcinogenesis by N-nitrosomethylbenzylamine (NMBA).Four-week-old keratin 5/AZ and wild-type (Wt) littermates were placed onZD or zinc-sufficient (ZS) diets to form four groups: ZD:AZ, ZD:Wt, ZS:AZ,and ZS:Wt. After 5 weeks, 27–45 mice in each group were treated twice withNMBA and sacrificed 14 weeks later. Independent of zinc intake, AZ micehad significantly lower forestomach tumor incidence and tumor multiplicitythan respective Wt littermates (P < 0.001): 21% of ZD:AZ versus 76% ofZD:Wt mice and 3% of ZS:AZ versus 33% of ZS:Wt mice developed tumors.Spermidine content was reduced in NMBA-treated ZD:AZ forestomachs.Zinc deficiency increased the forestomach cell proliferation in Wt mice, butthis effect was blocked by AZ. Conversely, apoptosis was substantially higherin control and NMBA-treated ZD:AZ than respective ZD:Wt forestomachs.The restored ZD:AZ forestomach epithelium displayed strong expression ofBax, a proapoptotic protein, and weak staining of cyclin D1 and its catalyticpartner Cdk4, key regulatory proteins controlling G1 to S progression. Incontrast, proliferative ZD:Wt forestomach showed strong expression of Bcl-2,an antiapoptotic protein, and overexpression of cyclin D1/Cdk4. Treatment ofZD:Wt mice with �-difluoromethylornithine, an inhibitor of ornithine decar-boxylase, had similar results to AZ in reducing tumor incidence, spermidinecontent, decreasing cell proliferation, and increasing apoptosis. These resultsdemonstrate that AZ may act as a tumor suppressor gene stimulatingapoptosis and restraining cell proliferation, thereby inhibiting forestomachtumor development. Although effects of AZ on functions other than poly-amine metabolism are possible, alterations in polyamines are the most likelyexplanation for the reduction in tumors, supporting the use of strategies tomodulate polyamine levels for cancer chemoprevention in individuals at highrisk of developing malignancies of the gastrointestinal tract.

INTRODUCTION

Polyamines are ubiquitous small basic molecules that play roles inmany aspects of cellular physiology. Polyamines are essential formammalian cell growth and development (1, 2), and gene disruptionsof ODC3 or S-adenosylmethionine decarboxylase, two key enzymesin polyamine biosynthesis are lethal at early stages of embryonic

development (3, 4). Many studies have indicated a correlation be-tween polyamine content and neoplastic growth (5), and the poly-amine biosynthetic pathway may be a useful target for the design ofcancer chemopreventive and chemotherapeutic agents (6–8). Theirreversible ODC inhibitor DFMO is currently in clinical trials forchemoprevention of numerous types of cancers (9, 10).

ODC (EC 4.1.1.17), which catalyzes the formation of putrescinefrom ornithine, the first step of cellular polyamine biosynthesis, hasbeen associated with neoplastic growth in many experimental studies.ODC overexpression and increased levels of polyamines alone or incooperation with other oncogenes can transform cells in vitro (11–14).Increased ODC activity is associated with transformation caused byseveral oncogenes (15–18), and ODC transcription is negatively reg-ulated by the Wilms’ tumor suppressor (19). Transgenic mice over-expressing ODC in the skin are more sensitive to tumor developmentin response to carcinogens (20) or UV radiation (21) and develop skintumors without need for tumor promoters (22).

ODC activity is very highly regulated at multiple levels includingtranscription, translation and protein degradation (reviewed in Refs.23–25). A major factor in the regulation of ODC and polyaminehomeostasis is AZ (26, 27). AZ mRNA contains an internal stopcodon preventing the synthesis of the active protein, but a polyamine-dependent �1 frameshifting event allows the translation of AZmRNA into full-length AZ protein. AZ then binds to the ODCmonomer, which prevents formation of the enzymatically active ho-modimer and, therefore, reduces ODC activity. More critically, thebinding of AZ stimulates the degradation of ODC by the 26S protea-some in an ATP-dependent but ubiquitin-independent manner. Mam-malian cells also possess a highly inducible polyamine transportsystem, which is activated when internal polyamine content falls. Thisuptake system is also inhibited by AZ, and there is some evidence thatAZ may also stimulate polyamine excretion (27). Thus, AZ reducespolyamine content and contributes to polyamine homeostasis throughmultiple mechanisms. Recent studies have indicated that there may bea family of AZ molecules in addition to that first discovered, which isnow termed AZ-1 (27, 28). The function of these multiple forms ofAZ is not yet known.

Transgenic mice in which AZ-1 is expressed from the bovine K5and K6 promoter elements have been derived (29). The AZ-1 cDNAconstruct used had a single nucleotide deletion (T205) to remove therequirement for polyamine-stimulated frameshifting in the translationof the mRNA. There are two potential start codons in the AZ-1 cDNA,and both are present in the transgene, but Western blots of epidermalextracts indicated that the second site was used preferentially in vivo(29). Both K5/AZ and K6/AZ transgenic mice developed normallyand were phenotypically indistinguishable from Wt littermates. How-ever, the transgenic AZ expression blocked the increase in skin ODCinduced by tumor promoters, and reduced epidermal and dermalpolyamine content, particularly spermidine. Carcinogenesis studiesusing a two stage protocol with initiation with 7,12-dimethylben-z(a)anthracene and treatment with the tumor promoter 12-O-tetradec-anoylphorbol-13-acetate showed that two founder lines of K6/AZ

Received 12/16/02; accepted 5/8/03.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1 Supported by grants CA-18138 from the National Cancer Institute (to A. E. P.) and99B045-REN from the American Institute for Cancer Research (to L. Y. Y. F.).

2 To whom requests for reprints should be addressed, at Department of Cellular andMolecular Physiology, H166, Pennsylvania State University College of Medicine, 500University Drive, Hershey, PA 17033. Phone: (717) 531-8152; Fax: (717) 531-5157;E-mail: aep1@psu.edu.

3 The abbreviations used are: ODC, ornithine decarboxylase; AZ, antizyme; ZD,zinc-deficient; NMBA, N-nitrosomethylbenzylamine; K5, keratin 5; Wt, wild-type; ZS,zinc-sufficient; DFMO, �-difluoromethylornithine; K6, keratin 6; SCJ, squamocolumnarjunction between fore- and hindstomach; PCNA, proliferating cell nuclear antigen;TUNEL, terminal deoxynucleotidyltransferase-mediated nick end labeling; LI, labelingindex; AI, apoptotic index; DAB, 3,3�-diaminobenzidine tetrahydrochloride.

3945

Research. on January 13, 2015. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

mice had a delay in tumor onset and a substantial reduction in tumormultiplicity compared with normal littermates. K5/AZ mice alsodeveloped fewer papillomas than littermate controls, and combinationof these lines to produce K5/AZ-K6/AZ double transgenic animalsyielded an additive decrease in tumor multiplicity (29). These studieswere carried out on a mixed B6D2 genetic background, but the resultswere confirmed in a defined genetic background after the transgeniclines were backcrossed onto the carcinogenesis-resistant C57BL/6Jinbred strain as well as the sensitive DBA/2J strain (30). There wasa greater effect with the K6/AZ lines, which is consistent with themore complete inhibition of ODC activity in K6/AZ comparedwith K5/AZ mice in response to 12-O-tetradecanoylphorbol-13-acetate application.

These results suggest that AZ acts as a tumor suppressor and that thetransgenic mice expressing AZ can be used to investigate whichpathways of carcinogenesis have an essential increase in polyaminemetabolism as a contributory factor. AZ expression from the K5promoter is not limited to the skin but may occur in other epithelialcells. Numerous studies have demonstrated that the K5 promoter candrive transgene expression in other stratified epithelia in addition toskin, such as the esophagus and forestomach (31–34), thymus (31,32), and prostate (35).

Therefore, we have studied the incidence of tumors in K5/AZ miceafter treatment with NMBA. This carcinogen is known to produceesophageal cancer in rats, and tumor development is accelerated byexposure to a ZD diet that induces proliferation in target cells (31, 37).Treatment of rats with NMBA and a ZD diet provides an importantmodel system, which reproduces many aspects of the development ofesophageal squamous cell carcinoma in humans in high risk areassuch as northern China and Iran (38, 39). Mice also develop tumorsreadily in response to NMBA and a ZD diet (40, 41). However, inmice there is a higher incidence of tumors in the forestomach, whichis considered to be a dilation of the lower esophagus (42), than inesophagus. A loss of p53 increased the sensitivity of ZD mice toNMBA-induced esophageal/forestomach carcinogenesis (41). Therapid rate of tumor induction/progression in ZD:p53�/� mice wasaccompanied by an increase in the rate of cell proliferation and adecrease in apoptosis.

The studies reported here show that AZ expression has the oppositeeffect to zinc deficiency and lack of p53 in that it stimulates apoptosis,restrains cell proliferation, and inhibits forestomach tumor develop-ment. These results are supported by studies in which DFMO wastested and found to also reduce tumor development after NMBAtreatment. These findings emphasize the potential value of using drugstargeting the polyamine metabolic pathway as cancer chemopreven-tive agents.

MATERIALS AND METHODS

Transgenic Mice. This study was approved by the Thomas JeffersonUniversity Institutional Animal Care and Use Committee and conducted underNIH guidelines. The generation of K5-AZ transgenic mice was describedpreviously (29). All of the animals used for the experiments were from the fifthor sixth backcross generation to C57BL/6J mice; therefore, �98% of theirgenes are derived from this inbred strain. Breeding of transgenic males from asingle founder line to C57BL/6J females (The Jackson Laboratory, Bar Harbor,ME) generated a total of 130 AZ transgenic mice and 136 Wt littermates (withabout equal number of males and females) for the forestomach tumorigenesisstudy. The AZ and Wt offspring were differentiated by genotyping of tail DNAusing a PCR-based method (29).

Chemicals and Diets. NMBA was purchased from Ash Stevens, Inc.(Detroit, MI). Custom-formulated, egg white-based ZD and ZS diets contain-ing 1.5 and 75 ppm zinc, respectively, were prepared by Teklad (Madison,

WI). The ZD diet is nutritionally complete and is identical to ZS diet except forthe concentration of elemental zinc (36).

Zinc Determination. The testes were removed from male and hair fromfemale animals at necropsy. Samples of testis or hair were dried to constantweight at 90°C and ashed in a furnace. Ashed samples were dissolved in 0.1N HCl and the zinc content determined by atomic spectrometry as described(36). Zinc content was expressed as �g/g dry weight of testis or hair. Overtsigns of zinc deficiency that are well-described for ZD rats (36), including fociof alopecia, skin lesions, and retarded growth, were not evident in ZD:Wt orZD:AZ mice, which had similar body weights at end point (Table 1). However,zinc content in the testis (male) and hair (female) was significantly lower in ZDthan ZS mice, regardless of genotype and other treatment (P � 0.001; Table 1).

Forestomach Tumorigenesis Experiment in AZ Transgenic Mice. Theexperiment was conducted in batches when the mice became available fromour breeding colonies. Four-week-old mice (AZ male, 13.1 � 1.5 g andfemale, 12.3 � 1.4 g; Wt male, 13.0 � 1.9 g and female, 12.6 � 1.7 g) werehoused 3–5 to a polycarbonate cage with a wire stainless steel floor. They weregiven free access to deionized drinking water. The mice were randomized intotwo dietary groups and were fed ad libitum a ZD or control ZS diet, formingfour experimental groups: ZD:Wt, ZD:AZ, ZS:Wt, and ZS:AZ. After 5 weeks,10 mice from each group (control mice) were sacrificed to determine the extentof cell proliferation and apoptosis in the forestomach. The remaining animals(27–45 mice/group and 15–29 mice/group in 2� NMBA and 4� NMBAstudies, respectively) were treated with two or four intragastric doses ofNMBA at 2 mg/kg body weight, twice weekly. NMBA-treated animals con-tinued on their respective diet and were sacrificed 14 weeks after the firstcarcinogen treatment for end point tumor incidence analysis.

Effect of DFMO on Forestomach Tumorigenesis in C57BL/6 Mice. Toevaluate the effect of DFMO in ZD mice, 75 weanling male C57BL/6 mice(11.7 � 1.3 g) were purchased from Taconic Laboratory (Germantown, NY).The animals were randomly divided into two dietary groups as describedabove. To determine the extent of cell proliferation before NMBA dosing (0 h),5 ZD and 5 ZS mice were killed after 5 weeks of experimental diet. Theremaining mice received two intragastric doses of NMBA at 2 mg/kg bodyweight per week. After the first dose, the animals were divided into four groups(14–20 mice/group): ZD/DFMO�; ZD/DFMO�; ZS/DFMO�, and ZS/DFMO�. DFMO� mice remained on deionized water, whereas DFMO�

animals were given deionized water containing 1% DFMO. All of the micecontinued on their respective diet. After 2 weeks, 5 mice from each group werekilled to determine the effect of DFMO on forestomach cell proliferation. Theremaining 45 mice were sacrificed 12 weeks after the first NMBA dose fortumor incidence analysis.

Tumor Analysis. After anesthetization with isoflurane (Ohmeda Inc.,Madison, WI), the mice were sacrificed and subjected to complete necropsies.Whole stomachs were excised and opened longitudinally. Tumors �0.5 mm indiameter in the forestomach were mapped and counted. Whole forestomachswere fixed in buffered formalin and embedded in paraffin. The forestomachwas cut into 4–6 strips, across the SCJ. Then, 4-�m thick cross-sections werecut. Typically, there were 4 to 6 sections per slide, representing the entireforestomach/SCJ. Sections were either stained with H&E for histopathology orleft unstained for immunohistochemical studies.

Table 1 Effect of a zinc-deficient diet on zinc content and body weight

Four-week-old mice were fed ZD or ZS diet for 5 weeks, and were then givenintragastrically, two or four NMBA doses, twice per week. The animals were sacrificed14 weeks later. Values are shown as the mean � SD. Zinc content (2� NMBA,4� NMBA): testis (n � 5–16), hair (n � 5–18), ZD:Wt versus ZS:Wt, ZD:AZ versusZS:AZ, P � 0.001. All statistical tests were two-sided.

Parameter

Diet:Genotype

ZD:Wt ZD:AZ ZS:Wt ZS:AZ

2� NMBAZinc content (�g/g), testis 135 � 4 133 � 5 168 � 6 163 � 8Zinc content (�g/g), hair 129 � 4 130 � 7 187 � 14 197 � 12Body weight (g), male 25 � 2 26 � 3 27 � 2 28 � 2Body weight (g), female 21 � 3 20 � 2 22 � 1 22 � 1

4� NMBAZinc content (�g/g), testis 128 � 5 128 � 6 165 � 13 161 � 11Zinc content (�g/g), hair 124 � 7 131 � 9 187 � 122 184 � 10Body weight (g), male 25 � 3 25 � 4 26 � 2 25 � 2Body weight (g), female 21 � 2 21 � 2 22 � 1 22 � 1

3946

AZ-MEDIATED INHIBITION OF FORESTOMACH CARCINOGENESIS

Research. on January 13, 2015. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

Polyamine Analysis. At necropsy, a small strip of forestomach epitheliumwas cut and snap-frozen in liquid nitrogen. The samples were stored at �80°Cuntil polyamine analysis. Each sample represented three pooled strips frommice of the same treatment group for the 2�NMBA experiment or a singletissue sample for the DFMO experiment. Polyamine content in forestomachwas determined after separation by ion-pair reversed-phase high-pressureliquid chromatography using fluorescence detection after postcolumn derivat-ization with o-phthalaldehyde as described previously (43) and normalized totissue wet weight (in grams). Normalization to mg protein/sample yieldedsimilar results.

Cell Proliferation Determination by PCNA Immunohistochemistry.Monoclonal mouse anti-PCNA (Santa Cruz Biotech., Santa Cruz, CA) wasused at 1:500 dilution, followed by incubations with biotinylated goat anti-mouse antibody and streptavidin horseradish peroxidase. PCNA was localizedby a final incubation with 3-amino-9-ethylcarbazole-substrate-chromogen sys-tem (Dako Corp., Carpinteria, CA), and a light hematoxylin counterstain. Cellswith red reaction product in the nucleus were considered positive for thepresence of PCNA.

Apoptosis Analysis. Apoptosis was assessed by the TUNEL method andby morphological characterization of cells in H&E stained sections.

TUNEL Assay. The 3�-OH end labeling of DNA in tissue sections wasperformed with an ApopTaq in situ peroxidase detection kit (SerologicalsCorp., Norcross, GA). Sections were deparaffinized, rehydrated in a gradedalcohol series, and incubated with proteinase K (20 �g/ml, 37°C for 10 min).Endogenous peroxidase in the sections was inhibited with 3% hydrogenperoxide, and slides were incubated (37°C for l h) with terminal deoxynucle-otidyl transferase to catalyze the addition of digoxigenin-labeled nucleotides tothe 3�-OH ends of fragmented DNA. Next, slides were incubated with horse-radish peroxidase-conjugated antidigoxigenin antibodies, and DNA fragmen-tation was detected by staining with DAB. Finally, sections were counter-stained with methyl green. Sections from rat mammary gland (SerologicalsCorp.), in which extensive apoptosis occurs, served as a positive control.Negative controls omitted terminal deoxynucleotidyl transferase.

Morphological Criteria. The incidence of apoptosis was measured ongood-quality sections stained with H&E. The morphology of apoptotic cells

depends on their stage in the process. Apoptotic morphologies include diffusecytoplasmic staining with only minimal nuclear condensation, distinct apo-ptotic bodies resulting from nuclear disintegration, or dense staining nucleiwith normal nuclear structure (44). All three of the forms were considered to beequivalent, and the AI, expressed as a percentage, was calculated by dividing thenumber of apoptotic cells by the total number of cells in the tissue section.

K5 and AZ Immunohistochemistry. Antiserum to AZ was produced inrabbits immunized with a purified recombinant polyhistidine-tagged AZ fusionprotein and then purified using an AZ-affinity column as described previously(29). After deparaffinization and rehydration in graded alcohols, forestomachsections were heated in citrate buffer [0.01 M (pH 6.0)] in a microwave oven(88–95°C; 3 � 5 min) before nonspecific binding sites were blocked with goatserum. Sections were incubated overnight at 37°C in a humidified chamberwith rabbit anti-AZ antiserum at 1:200 dilution followed by incubation withbiotinylated goat antirabbit antibody serum (1:500 dilution). Slides were thenincubated with streptavidin horseradish peroxidase (1:1000 dilution). Cyto-plasmic expression of AZ was localized by a final incubation with DAB anda light hematoxylin counterstain.

K5 was detected using a polyclonal antibody against mouse K5 (CovanceResearch Products, Berkeley, CA) at a 1:500 dilution and visualized with theVectastain Elite ABC kit with DAB chromagen (Vector Laboratories, Burl-ingame, CA).

Bax and Bcl-2 Immunohistochemistry. To detect Bax and Bcl-2, fores-tomach sections were incubated overnight at 37°C in a humidified chamberwith a rabbit anti-Bcl-2 polyclonal antiserum (Santa Cruz) at a 1:400 dilution,or with a rabbit anti-Bax polyclonal antiserum (Santa Cruz) at a 1:200 dilution,followed by incubation with a biotinylated goat antirabbit antibody serum.Cytoplasmic Bcl-2 and Bax expression was visualized with DAB.

Cyclin D1 and Cdk4 Immunohistochemistry. To detect cyclin D1 andCdk4, forestomach sections were incubated as described above with a rabbitanticyclin D1 polyclonal antiserum (Lab Vision Corp., Fremont, CA) at a1:100 dilution or with a rabbit anti-Cdk4 polyclonal antiserum (Santa Cruz) ata 1:100 dilution, followed by incubation with a biotinylated goat antirabbitantibody serum. Nuclear cyclin D1 and Cdk4 expression was visualizedwith DAB.

Fig. 1. Reduction of NMBA-induced forestom-ach carcinogenesis in AZ transgenic mice. Four-week-old mice were fed ZD or ZS diet for 5 weeks,and were then given two (A and C) or four (B andD) NMBA doses, intragastrically twice per week.The animals were sacrificed 14 weeks later. A andB, show the percentage of tumor incidence in theforestomach and SCJ with the glandular stomach.The number of animals with tumors and the totalnumber of mice is shown above each bar. C and Dshow the tumor multiplicity of forestomach tumorsand the SE. Statistical analysis of significance ofpercentage of tumor incidence: 2� NMBA,ZD:AZ versus ZD:Wt, forestomach and SCJ,P � 0.001; ZS:AZ versus ZS:Wt, forestomach,P � 0.003 and SCJ, P � 0.001; ZD:Wt versusZS:Wt, forestomach, P � 0.001 and SCJ,P � 0.05; ZD:AZ versus ZS:AZ, forestomach,P � 0.032 and SCJ, P � 0.013; 4� NMBA,ZD:AZ versus ZD:Wt, forestomach, P � 0.001and SCJ, P � 0.004; ZS:AZ versus ZS:Wt, fore-stomach, P � 0.048; ZD:Wt versus ZS:Wt, fore-stomach, P � 0.025 and SCJ, P � 0.009. Statisti-cal analysis of significance of tumors/forestomach:2� NMBA, ZD:AZ versus ZD:Wt and ZD:Wtversus ZS:Wt, P � 0.001; 4� NMBA, ZD:AZversus ZD:Wt; and ZD:Wt versus ZS:Wt,P � 0.001. All statistical tests were two-sided.

3947

AZ-MEDIATED INHIBITION OF FORESTOMACH CARCINOGENESIS

Research. on January 13, 2015. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

Statistical Analysis. Tumor incidence differences were analyzed by two-tailed Fisher’s exact test, and data on cell proliferation and testis zinc levelwere analyzed by one-way ANOVA with the SAS statistical computer pro-gram as described (41). All of the statistical tests were two-sided and wereconsidered statistically significant at P � 0.05.

RESULTS

Effect of AZ Expression on Forestomach Carcinogenesis. Four-teen weeks after two NMBA doses, ZD:Wt mice invariably showed athickened and shrunken forestomach, with a 76% and 56% tumorincidence in the forestomach and SCJ, respectively, and a forestomachtumor multiplicity of 2.6 (Fig. 1, A and C). There was a large andhighly significant reduction in tumor incidence in ZD:AZ mice (21%forestomach and 18% SCJ) and in tumor multiplicity (0.4). Althoughthe overall tumor incidence was lower in ZS:Wt mice (which isconsistent with previous results in C57BL/6 mice; Refs. 40, 41)ZS:AZ mice, relative to their ZS:Wt counterparts, also had signifi-cantly lower tumor incidence in the forestomach (3% versus 33%) andSCJ (0% versus 30%; Fig. 1A). Similar results were obtained in asecond experiment in which the mice were given four doses of NMBAand sacrificed 14 weeks later for tumor incidence analysis (Fig. 1, Band D). Notably, only 15% of ZD:AZ mice had forestomach tumorsversus 80% of their ZD:Wt counterparts, and tumor multiplicity was0.5 versus 3.8. In the ZS mice, AZ expression caused a drop inforestomach tumors from 41% to 12%. Tumors at the SCJ were alsoreduced by AZ in both ZD and ZS mice. Esophageal tumors were notdetected in ZD:Wt mice probably because the animals in the presentstudy were exposed to fewer NMBA doses than in the previous study(40).

Tumor-bearing ZD:Wt mice regularly displayed large, fused fore-stomach tumors (Fig. 2A) or large, solitary tumors (Fig. 2B). On thecontrary, 80% of ZD:AZ mice showed a tumor-free forestomachthat was mostly large and thin (Fig. 2C) or slightly thickened at theSCJ (Fig. 2D). The results shown in Fig. 2 were obtained with micegiven two doses of NMBA but similar results (data not shown) wereobtained after four doses of the carcinogen.

The expression pattern of AZ in the forestomach epithelium ofK5/AZ transgenic mice was determined both indirectly by K5 immu-nostaining and directly by AZ immunostaining. Endogenous K5 wasreadily visualized in the forestomach of both AZ and control mice byimmunohistochemistry. K5 staining was strong and uniform through-

Fig. 3. Immunohistochemical detection of AZ in ZS:AZ forestomach. A forestomachsection from an NMBA-treated ZS:AZ mouse displayed moderately strong cytoplasmicstaining of AZ in many cells (representative cells shown by arrows) in the proliferativeforestomach epithelium (B) versus forestomach from an NMBA-treated ZS:wt mouseshowing lack of AZ expression (A). The bars represent 50 �m.

Fig. 2. Macroscopic appearance of forestomachs ofZD:Wt and ZD:AZ mice 14 weeks after two doses ofNMBA. Representative pictures of individual mice areshown. ZD:Wt forestomachs display large, fused tu-mors (A), and large solitary tumors (B), whereasZD:AZ forestomachs show a tumor-free, thin fores-tomach and SCJ (C), or exhibit a tumor-free forestom-ach with thickened SCJ (D). Gross photographs, �5.3.

3948

AZ-MEDIATED INHIBITION OF FORESTOMACH CARCINOGENESIS

Research. on January 13, 2015. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

out the basal cell layer of the forestomach and within areas ofhyperproliferation. The staining was weak and patchy within thesuprabasal layer. These patterns of K5 staining were consistent inde-pendent of diet, genotype, or carcinogen treatment (data not shown).These results are consistent with previous studies showing that the K5promoter is active in the forestomach (31–34). AZ expression wasvisualized using immunohistochemical staining with an antiserum topurified recombinant rat AZ. This revealed higher levels of AZexpression in the cytoplasm of forestomach epithelial cells fromK5/AZ mice relative to control animals (Fig. 3).

Reduced Cell Proliferation in Control ZD:AZ Mice. Visualinspection of control mice after 5 weeks of the ZD diet typicallyshowed a thickened forestomach and SCJ in ZD:Wt mice but a largeand thin forestomach in ZD:AZ animals, suggesting nutritional zincdeprivation did not have an effect on forestomach cell proliferation inthe latter. Histopathologic examination of control forestomach sec-tions showed that the level of cell proliferation was highest in ZD:Wtmice, followed by ZS:Wt, and then ZD:AZ � ZS:AZ. Although AZoverexpression reduced cell proliferation in both ZD and ZS fores-tomach, its effect was more dramatic in ZD mice, and photomicro-graphs are presented of ZD tissue sections (Fig. 4). Relative to ZS:Wtforestomach that exhibited occasional and mild proliferation (data notshown), ZD:Wt forestomach regularly displayed a proliferative epi-

thelium with small upward and downward focal hyperplastic lesions(Fig. 4A), showing many PCNA-positive nuclei (Fig. 4E), and over-expression of Cdk4 (Fig. 4I) and cyclin D1 (Fig. 4M) in these lesions.PCNA is an endogenous cell proliferation marker; cyclin D1 and itscatalytic partner, Cdk4, are major G1-S regulatory proteins (45).These results are consistent with previous data that demonstrated thatproliferative esophagi from untreated ZD rats had altered expressionprofiles for genes that control G1-S progression (43). In contrast, AZmice on a ZD diet typically showed a thin forestomach epithelium of2–5 cells thick (Fig. 4B), with PCNA-positive nuclei mainly in thebasal cell layer (Fig. 4F), and weak and sporadic expression of Cdk4(Fig. 4J) and cyclin D1 (Fig. 4N). Thus, AZ overexpression in ZDmice restrains Cdk4/cyclin D1 expression and inhibits cell prolifera-tion in the forestomach.

Quantitative PCNA-immunohistochemistry showed that dietary zincdeficiency induced a high rate of cell proliferation in the forestomach ofWt mice (Fig. 5A, Expt. 1), a result consistent with previous studies in rats(46) and mice (40). The PCNA-LI, a measure of cellular proliferation,was significantly higher in ZD:Wt than ZS:Wt forestomach. Overexpres-sion of AZ in ZD mice counteracted the effect of zinc deficiency on cellproliferation and brought about a significantly reduced LI compared withZD:Wt forestomach. Likewise, LI was significantly lower in ZS:AZ thanZS:Wt forestomach (Fig. 5A, Expt. 1).

Fig. 4. Cell proliferation in control and NMBA-treated ZD:AZ and ZD:Wt forestomachs. H&E staining (A–D), immunohistochemistry for PCNA (E–H), Cdk4 (I–L), and cyclin D1(M–P) were used. Representative pictures of individual mice are shown. A, B, E, F, I, J, M, and N show results from mice fed ZD diet for 5 weeks with no NMBA treatment. ZD:Wtmice showed: a hyperplastic epithelium with focal hyperplastic lesions (A) displaying numerous PCNA-positive nuclei (3-amino-9-ethylcarbazole, red) in S phase, and G1-S/G2 phase(E); strong nuclear expression of Cdk4 (I, DAB, brown); and strong nuclear staining of cyclin D1 (DAB, brown) in basal cells and focal hyperplastic lesions (M). Conversely, ZD:AZmice showed: a thin epithelium (B) with PCNA-positive cells mainly in the basal cell layer (F); moderate expression of Cdk4 in basal cells (J); and sporadic expression of cyclin D1in basal cells (N). C, D, G, H, K, L, O, and P show results from mice fed ZD diet, treated with two NMBA doses, and sacrificed 14 weeks later. Sections from ZD:Wt mice showed:a dysplastic epithelium with deep focal hyperplastic lesions (C), abundant PCNA-positive nuclei (G), overexpression of cyclin D1 (O) in these lesions, and overexpression of Cdk4in a dysplastic epithelium (K). Forestomach sections from ZD:AZ mice showed: a thin epithelium displaying many apoptotic cells (D and inset); infrequent PCNA-positive nuclei mainlyin the basal cell layer (H); a few strongly stained Cdk4-positive (L); and cyclin D1-positive (P) nuclei, mainly in basal cells. Bars: 100 �m for A–C and G; 50 �m for D–F and H–P;25 �m for the inset.

3949

AZ-MEDIATED INHIBITION OF FORESTOMACH CARCINOGENESIS

Research. on January 13, 2015. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

Increased Apoptosis in Control ZD:AZ Mice. AIs were deter-mined in high quality H&E-stained forestomach sections of controlZD and ZS mice of both AZ and Wt genotypes (Fig. 5B, Expt. 1). Zincdeficiency produced a significant decrease in the AI of Wt mice.Expression of AZ increased the AI, which was significantly higher inZD:AZ compared with ZD:Wt forestomach and in ZS:AZ comparedwith ZS:Wt forestomach. In addition, TUNEL analysis (Fig. 6)showed frequent occurrence of darkly stained TUNEL-positive nucleiin the restored and sometimes still proliferative forestomach epithe-

lium of ZD:AZ mice compared with ZD:Wt forestomach, which hadonly occasional incidences of apoptotic cells (Fig. 6, B versus A).

Fig. 6 also shows immunohistochemical staining for Bax, a pro-apoptotic protein, and Bcl-2, an antiapoptotic protein in forestomachsections. This revealed diffuse and weak staining of Bax in ZD:Wtsections (Fig. 6E) but strong staining in ZD:AZ epithelium (Fig. 6F).On the other hand, Bcl-2 expression was typically strong in the basaland proliferative areas of ZD:Wt (Fig. 6I) forestomach but weak andinfrequent in ZD:AZ epithelium (Fig. 6J).

Cell Proliferation and Apoptosis in NMBA-treated ZD:AZMice. Detailed cell proliferation and apoptosis results are presentedin Figs. 4–6 for treated ZD:AZ and ZD:Wt mice of the two NMBAdose experiment. At 14 weeks, the ZD:AZ forestomach demonstratedsignificantly lower LI than their ZD:Wt counterparts (Fig. 5A, Expt. 2)a result consistent with those from respective control groups (Fig. 5A,Expt. 1). Typically, ZD:Wt forestomach displayed numerous PCNA-positive nuclei in areas of hyperplasia, dysplasia, and papilloma,whereas ZD:AZ forestomach showed a few PCNA-positive nucleimainly in the basal cell layer (Fig. 4, G versus H). In addition, ZD:Wtforestomach lesions exhibited overexpression of Cdk4 (Fig. 4K) andcyclin D1 (Fig. 4O) in these lesions, whereas ZD:AZ forestomachshowed moderate expression of these two G1 to S regulatory proteins,mostly in basal cells (Fig. 4, L and P).

On the other hand, the AI was significantly higher in ZD:AZ thanZD:Wt forestomachs (Fig. 5B, Expt. 2). These data are in line withthose from respective control groups (Fig. 5B, Expt. 1). H&E-stainedand TUNEL-stained forestomach sections from ZD:AZ mice dis-played frequent occurrence of apoptotic cells (Fig. 4D; Fig. 6D),whereas those from ZD:Wt animals showed only isolated occurrences(Fig. 4C; Fig. 6C). Bax expression was intense and abundant inZD:AZ forestomach (Fig. 6H) but moderate and sparse in ZD:Wtforestomach (Fig. 6G). Conversely, Bcl-2 expression was moderateand mostly in the basal cells of ZD:AZ forestomach (Fig. 6L) butstrong and abundant in ZD:Wt forestomach lesions (Fig. 6K).

Effect of AZ Expression on Polyamine Levels in the MouseForestomach. The most probable explanation for the striking effectof transgenic AZ expression on tumor incidence, cell proliferation,and apoptosis described above is that polyamine levels influence theseprocesses, and that the inhibitory effects of AZ on ODC and poly-amine transport alter polyamine levels. Direct assessment of alter-ations in ODC levels were not possible, because ODC activity meas-urements could not be made reliably on tissue samples because of thesmall amount of tissue available and the limited sensitivity of theseassays. Except for putrescine, which was at the limit of detection,polyamine levels could be measured accurately in the forestomachtissue. In samples collected at 14 weeks after treatment with two dosesof NMBA, there was a statistically significant reduction in the sper-midine content of the ZD:AZ group compared with ZD:Wt (Fig. 7).However, the differences were small, and spermidine values were notstatistically significantly decreased in the comparison of the ZS:AZand ZS:Wt groups. A probable explanation for this is that the relevantepithelial cell population makes up only a small proportion of thetissue used for this analysis because the underlying connective tissueand muscle layers are also retained in this analysis. Therefore, addi-tional study of the possible role of polyamines in cell proliferation andforestomach carcinogenesis in ZD mice was carried out using theODC inhibitor DFMO.

Effect of DFMO on Polyamine Levels, Tumor Incidence, andCell Proliferation in Forestomach of ZD Mice. ZD C57BL/6 micewere treated with the two-dose NMBA protocol, and half of theanimals in each group received 1% DFMO in the drinking waterstarting after the first NMBA dose. DFMO treatment did lead to asmall reduction in body weight but did not affect the zinc deficiency

Fig. 5. Rates of cell proliferation (A) and apoptosis (B) in forestomach. Results areshown from three experiments. Experiment 1 shows data from control AZ transgenic andWt littermates fed a ZD or ZS diet for 5 weeks. Experiment 2 shows data from ZD micetreated with 2 doses of NMBA as described in the legend to Fig. 1. Experiment 3 showsresults from ZD mice treated with two doses of NMBA with and without DFMO asdescribed in the legend to Fig. 8. The PCNA-LI (%; A) is calculated by dividing thenumber of respective labeled cells in S phase by the total number of cells/cross-section oftissue, and the result is expressed as a percentage. The AI (%; B) is calculated by dividingthe number of apoptotic cells by the total number of cells/cross-section of tissue, and theresult is expressed as a percentage. Statistical analysis was two sided with 7–11 animalsin each group; bars, �SD. Statistical analysis for LI in experiment 1: ZD:Wt versusZS:Wt, P � 0.001; ZD:AZ versus ZD:Wt, P � 0.001; ZS:AZ versus ZS:Wt, P � 0.001;ZD:AZ versus ZS:AZ, NS. Statistical analysis for AI in experiment 1: ZD:Wt versusZS:Wt, P � 0.001; ZD:AZ versus ZD:Wt, P � 0.001; ZS:AZ versus ZS:Wt. P � 0.05;ZD:AZ versus ZS:AZ, P � 0.001. Statistical analysis for experiment 2: ZD:AZ versusZD:Wt, P � 0.001 (n � 11) for both LI and AI. Statistical analysis for experiment3: ZD:DFMO� versus ZD:DFMO�, P � 0.001 (n � 11) for both LI and AI.

3950

AZ-MEDIATED INHIBITION OF FORESTOMACH CARCINOGENESIS

Research. on January 13, 2015. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

(Compare Table 2 with Table 1). Table 2 shows that 87% of maleC57BL/6 mice developed forestomach tumors at week 12 after twodoses of NMBA, with a tumor multiplicity of 3.7. ZD mice that wereswitched to drinking water containing 1% DFMO after 5 weeks of adeficient diet exhibited a greatly reduced tumor incidence of 9%, witha lower tumor multiplicity of 0.2 (Table 2). The spermidine content ofthe forestomach was reduced significantly by DFMO to about thesame extent as that by AZ (compare Table 2 and Fig. 7). DFMOtreatment also reduced the tumor incidence and the spermidine con-

tent in ZS mice treated with the two-dose NMBA protocol, althoughthe tumor incidence in the ZS mice not receiving DFMO was con-siderably lower than in ZD mice (results not shown).

DFMO, administered after the establishment of increased cell pro-liferation by nutritional zinc deficiency, reversed the increased cellproliferation in the mouse forestomach (Fig. 8). After 5 weeks of theZD diet, the forestomach typically displayed a proliferative epitheliumwith many PCNA-positive nuclei in small focal hyperplastic up- anddown-growths (Fig. 8A). At week 2 after NMBA treatment, there wasan expansion in focal hyperplastic lesions that was accompanied by anincrease in PCNA-positive S phase cells (Fig. 8B). At week 12,PCNA-positive cells were found in areas of hyperplasia, dysplasia,focal hyperplastic lesions, and papilloma (Fig. 8C). In contrast, whenDFMO was given starting at the time of the first NMBA treatment,within 2 weeks after NMBA and DFMO administration, a restoredforestomach epithelium with a thickness of 2–6 cells was in place,with PCNA-positive nuclei mainly in the basal cell layer (Fig. 8D). Atweek 12, the forestomach in the DFMO-treated group was mostly 2–5cells thick with a few PCNA-positive nuclei, mainly in basal cells(Fig. 8E). Quantitative PCNA immunohistochemistry demonstrated asignificantly lower PCNA-LI in ZD/DFMO� than ZD/DFMO� fore-stomachs at week 12 (Fig. 5A, Expt. 3). In addition, these ZD/DFMO�

forestomachs showed a significantly higher AI (%) than the corre-sponding ZD/DFMO� forestomach (Fig. 5B, Expt. 3).

DISCUSSION

The moderate expression of AZ from the K5 promoter has noobvious deleterious effects in these transgenic mice. Detailed exam-

Fig. 6. Apoptosis in control and NMBA-treated ZD:AZ and ZD:Wt forestomachs. TUNEL analysis (A–D), immunohistochemistry for Bax (E–H), and Bcl-2 (I–L) were used.Representative pictures of individual mice are shown. A, B, E, F, I, and J show results from mice fed ZD diet for 5 weeks with no NMBA treatment. ZD:Wt mice forestomachs showed:lack of strongly stained TUNEL-positive apoptotic nuclei (A); very weak and diffuse cytoplasmic expression of Bax (E); and strong cytoplasmic staining of Bcl-2 in basal cells andfocal hyperplastic lesion (I). In contrast, ZD:AZ forestomachs showed: strongly stained TUNEL-positive nuclei in the basal cell layer of restored epithelium (B); with overexpressionof Bax in a proliferative location (F); and weak expression of Bcl-2 in basal cells (J). C, D, G, H, K, and L show results from mice fed ZD diet, treated with two NMBA doses, andsacrificed 14 weeks later. Sections of ZD:Wt forestomachs showed: the occurrence of a single TUNEL-positive nucleus in deep focal hyperplastic lesions; scattered expression of Bax(G); and intense expression of Bcl-2 in these lesions (K). In contrast, ZD:AZ forestomachs showed: numerous darkly stained TUNEL-positive nuclei in a restored epithelium (D); intensecytoplasmic staining of Bax in a mildly proliferative epithelium (H and inset); and moderate expression of Bcl-2 in basal cells (L). Bars, 50 �m for A–L, and 25 �m for the inset.

Fig. 7. Polyamine content of forestomach of NMBA-treated ZD and ZS mice. Four-week-old mice (AZ or Wt) were fed ZD or ZS diet for 5 weeks, and were then given twointragastric doses of NMBA and sacrificed 14 weeks later. Forestomach samples weretaken and polyamine content analyzed. Results are shown for putrescine, spermidine, andspermine. The number of samples analyzed was: ZD:Wt, 11; ZD:AZ, 11; ZS:Wt, 5; andZS:AZ, 10. Because of the small amount of tissue available, each sample contained tissuespooled from 3 mice; bars, � SD. Statistical analysis was two sided, and the only valuesthat were significantly different were spermidine content for ZD:Wt versus ZD:AZ;P � 0.01.

3951

AZ-MEDIATED INHIBITION OF FORESTOMACH CARCINOGENESIS

Research. on January 13, 2015. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

ination has only been carried out in the skin (29, 30) and in theesophagus/forestomach, but no significant effects were seen in theabsence of treatment with carcinogens or tumor promoters. The trans-genic mice have no obvious phenotype and live a normal life span.However, the results provided here and those in which skin tumordevelopment was examined (29, 30) show that they are highly resist-ant to tumor development after treatment with regimes that initiateand promote carcinogenesis.

It is firmly established that AZ is a regulator of polyamine content.There have been suggestions that AZ may cause a reduction in the

content of other proteins including cyclin D1, cdk4, Smad1, andSNIP1 (27, 47). These reports are highly preliminary and have not yetbeen confirmed. However, we certainly cannot at present rule out thepossibility that the tumor-suppressive effects of AZ are mediated orenhanced via effects on these pathways or others.

It is most probable that the cancer preventative effect is attributableto a reduction in cellular polyamine content. Although there was onlya small decline in spermidine in the treated AZ mice compared withtreated controls (Fig. 7), this reduction was similar to that produced bythe administration of DFMO in doses that had a similar effect ontumor incidence. The close similarity between the effects on cellproliferation and apoptosis of DFMO treatment or transgenic AZexpression is strong evidence that AZ is modulating carcinogenesisthrough effects on polyamines. The importance of polyamines andODC in esophageal and stomach carcinogenesis is also supportedby studies showing that ODC increases in these tissues of ratsresponding to regimes that cause tumors (48, 49). Similarly, trans-genic rats with human c-Ha-ras are highly susceptible to NMBAinduction of esophageal tumors (50), and ODC is a downstreamtarget of ras (17).

Two factors contribute to the cancer protective effect of AZ. Ourfindings clearly demonstrate that overexpression of AZ both stimu-lates apoptosis and inhibits ZD-induced cell proliferation, a conditionknown to promote forestomach carcinogenesis in ZD mice (40, 41).The antiproliferative effects of blocking increases in polyamine con-tent have been documented extensively in cell culture and in vivo (2,51). The ability of cancer chemopreventive agents including DFMO tocause reductions in the proliferative index is a key factor in the currentevaluations of these agents in Phase II/III trials (7, 52–54).

The stimulation of apoptosis by AZ or by DFMO, which is shownclearly in the results of Figs. 5 and 6, has been less widely recognized

Fig. 8. Reversal of pre-established cell prolifer-ation in forestomach of NMBA-treated ZD mice byDFMO. Mice were treated for 5 weeks with the ZDdiet and then with two doses of NMBA as in thelegend to Fig. 1. Half of the mice received DFMOin the drinking water (DFMO�) and the others didnot (DFMO�). PCNA immunohistochemistry withcounterstaining with hematoxylin was then carriedout on forestomach samples. A–C show represent-ative results for ZD; DFMO� mice: at 0 h (5 weeksafter ZD diet), at 2 weeks (B), and 12 weeks (C)after two NMBA doses. Numerous PNCA-positiveregions in small upward and downward focal hy-perplastic lesions are seen at 0 h (A) with abundantPCNA-positive nuclei in expanded focal hyper-plastic lesions at 2 weeks (B) and papillomas at 12weeks (C). D and E show results for DFMO� miceat 2 weeks (D) and 12 weeks (E). These show arestored epithelium of 2–3 cells thick with PCNA-positive nuclei, mainly in the basal cell layer(D and E). Bars, 100 �m.

Table 2 Reduction of NMBA-induced forestomach carcinogenesis by DFMO inC57BL/6 mice on a ZD diet

Male weanling C57BL/6 mice were fed ZD diet for 5 weeks. The mice were then giventwo doses of NMBA. After the first dose, DFMO� groups received 1% DFMO in thedrinking water. All mice were sacrificed at week 12. Statistical analysis of significance of% tumor incidence for both forestomach and squamocolumnar junction and tumormultiplicity, DFMO� versus DFMO�, P � 0.001. Statistical analysis of significance ofpolyamine content, spermidine: DFMO� versus DFMO�, P � 0.003. Statistical analysisof significance of body weight: DFMO� versus DFMO�, P � 0.01. All statistical testswere two-sided.

Parameter

Diet:DFMO

ZD:DFMO� ZD:DFMO�

Forestomach tumor incidencea

(%)13/15 (87) 1/11 (9)

Squamocolumnar junction tumorincidencea (%)

14/15 (93) 1/11 (9)

Tumor multiplicityb 3.7 � 3.0 0.2 � 0.6Polyamine content (nmol/g tissue)Putrescine 7 � 6 6 � 7Spermidine 328 � 142 249 � 104Spermine 216 � 112 217 � 58Zinc content (�g/g), testis 135 � 4 133 � 5Body weight (g) 31 � 2 26 � 1

a Number of mice with tumors/total number of mice.b Number of tumors per forestomach.

3952

AZ-MEDIATED INHIBITION OF FORESTOMACH CARCINOGENESIS

Research. on January 13, 2015. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

than the effects on proliferation, because polyamine levels can havebiphasic effects both inhibiting and increasing apoptosis (see Refs. 55,56). However, it is consistent with results of DFMO treatment in theZD rat esophagus (43) and of human gastric tumors growing in nudemice (57). Recent studies have shown that polyamine depletion trig-gers the mitochondrial-mediated cell death pathway (58).

More than 50% of human cancers including many esophagealsquamous cell carcinomas exhibit alterations that cause a loss ofactivity of the p53 tumor suppressor protein. Mutations in p53 alsooccur frequently in rodent esophageal tumors induced by NMBA(59). Recent studies using p53-deficient mice have shown that thecombination of ZD with loss of p53 leads to a large increase in therate of cell proliferation, a decrease in apoptosis, and the very rapiddevelopment of forestomach and esophageal tumors (41). Ourstudies on forestomach reported here and in the skin (29, 30) showthat AZ has opposite effects to the loss of p53 and suggest that AZmay be a tumor suppressor gene. Several other findings supportthis possibility. The content of AZ mRNA is reduced in malignantoral keratinocytes, and AZ expression reversed their malignantphenotype (60). Expression of AZ blocked tumor formation byH-ras-transfected 3T3 cells in nude mice (61). AZ-mediated ef-fects on cell proliferation have been documented by several groups(60 – 63). Therefore, additional study of the potential role of AZ asa tumor suppressor gene is warranted. Because of the uniqueregulation of AZ synthesis in which translational frameshifting isneeded for production of the protein, the levels of AZ may notcorrelate with the mRNA content, and measurement of the activeprotein content will be required for such studies.

DFMO is currently in clinical trials to prevent cancer in high-riskgroups (9, 64, 65). The success of the transgenic AZ approach may bebecause of the ability of AZ to augment a modest reduction of ODCalong with a blockage of the uptake of exogenous polyamines. There-fore, an improved approach might be to combine low doses of DFMO,which are tolerated well, with treatment with an inhibitor of poly-amine transport (66, 67). Another possibility would be to test inducersof AZ as chemopreventive agents. Several polyamine analogues in-cluding some that are in clinical trials as antitumor agents are able toinduce AZ (63). Combination therapy has much promise for cancerprevention, and significant effects in reducing esophageal and stom-ach cancer has been observed in rodent models with phenethyliso-thiocyanate, piroxicam, curcumin, and resveratrol (68–70). The com-binations of these agents with induction of AZ may also be a usefulstrategy.

ACKNOWLEDGMENTS

We thank David Super (Medical Media Services, Thomas Jefferson Uni-versity) for mouse forestomach photography, Vu T. Nguyen for animal main-tenance, and Suzie Sass-Kuhn for the polyamine analysis.

REFERENCES

1. Heby, O., and Persson, L. Molecular genetics of polyamine synthesis in eukaryoticcells. TIBS, 15: 153–158, 1990.

2. Pegg, A. E. Polyamine metabolism and its importance in neoplastic growth and as atarget for chemotherapy. Cancer Res., 48: 759–774, 1988.

3. Pendeville, H., Carpino, N., Marine, J. C., Takahashi, Y., Muller, M., Martial, J. A.,and Cleveland, J. L. The ornithine decarboxylase gene is essential for cell survivalduring early murine development. Mol. Cell. Biol., 21: 6459–6558, 2001.

4. Nishimura, K., Nakatsu, F., Kashiwagi, K., Ohno, H., Saito, H., Saito, T., andIgarashi, K. Essential role of S-adenosylmethionine decarboxylase in mouse embry-onic development. Genes Cells, 7: 41–47, 2002.

5. Bachrach, U., Wang, Y. C., and Tabib, A. Polyamines: New cues in cellular signaltransduction. News Physiol. Sci., 16: 106–110, 2001.

6. Marton, L. J., and Pegg, A. E. Polyamines as targets for therapeutic intervention.Annu. Rev. Pharm., 35: 55–91, 1995.

7. Kelloff, G. J., Boone, C. W., Steele, V. E., Lubet, R. A., Doddy, L. A., Malone, W. F.,Hawk, E. T., and Sigman, C. C. New agents for cancer chemoprevention. J. Cell.Biochem., 26(Suppl.): 1–28, 1996.

8. Casero, R. A., Jr., and Woster. P. M. Terminally alkylated polyamine analogues aschemotherapeutic agents. J. Med. Chem., 44: 1–26, 2001.

9. Simoneau, A. R., Gerner, E. W., Phung, M., McLaren, C. E., and Meyskens, F. L.�-Difluoromethylornithine and polyamine levels in the human prostate: results of aphase II trial. J. Natl. Cancer Inst., 93: 57–59, 2001.

10. Doyle, K. J., McLaren, C. E., Shanks, J. E., Galus, C. M., and Meyskens, F. L. Effectsof difluoromethylornithine chemoprevention on audiometry thresholds and otoacous-tic emissions. Arch. Otolaryngol. Head Neck Surg., 127: 553–558, 2001.

11. Clifford, A., Morgan, D., Yuspa, S. H., Peralta Soler, A., and Gilmour, S. Role ofornithine decarboxylase in epidermal tumorigenesis. Cancer Res., 55: 1680–1686,1995.

12. Tabib, A., and Bachrach, U. Role of polyamines in mediating malignant transforma-tion and oncogene expression. Int. J. Biochem. Cell Biol., 31: 1289–1295, 1999.

13. Ravanko, K., Jarvinen, K., Paasinen-Sohns, A., and Holtta, E. Loss of p27Kip1 fromcyclin E/cyclin-dependent kinase (CDK) but not from cyclin D1/CDK4 complexes incells transformed by polyamine biosynthetic enzymes. Cancer Res., 60: 5244–5253,2000.

14. Geng, H., Naylor, P. H., Dosescu, J., Skunca, M., Majumdar, A. P. N., and Moshier,J. A. TGF� is required for full expression of the transformed growth pheneotype ofNIH 3T3 cells overexpressing ornithine decarboxylase. Carcinogenesis (Lond.), 21:567–572, 2000.

15. Packham, G., and Cleveland, J. L. Ornithine decarboxylase is a mediator of c-Myc-induced apoptosis. Mol. Cell. Biol., 14: 5741–5747, 1994.

16. Holtta, E., Auvinen, M., and Andersson, L. C. Polyamines are essential for celltransformation for pp60v-src: Delination of molecular events relevant for the trans-formed phenotype. J. Cell Biol., 122: 903–914, 1993.

17. Shantz, L. M., and Pegg, A. E. Ornithine decarboxylase induction in transformationby the Raf/MAPK and Rac/Rho signal transduction pathways. Cancer Res., 58:2748–2753, 1998.

18. Hurta, R. A. R., Lee, J., and Voskas, D. Transformation by H-ras can result in aberrantregulation of ornithine decarboxylase gene expression by transforming growth factor-�1. J. Cell. Biochem., 81: 39–55, 2001.

19. Li, R-S., Law, G. L., Seifert, R. A., Romaniuk, P. J., and Morris, D. R. Ornithinedecarboxylase is a transcriptional target of tumor suppressor WT1. Exp. Cell Res.,247: 257–266, 1999.

20. Chen, Y., Megosh, L. C., Gilmour, S. K., Sawicki, J. A., and O’Brien. T. G. K6/ODCtransgenic mice as a sensitive model for carincogen identification. Toxicol. Lett., 116:27–35, 2000.

21. Ahmad, N., Gilliam, A. C., Katiyar, S. K., O’Brien, T. G., and Mukhtar, H. Adefinitive role of ornithine decarboxylase in photocarcinogenesis. Am. J. Pathol., 159:885–892, 2001.

22. O’Brien, T. G., Megosh, L. C., Gilliard, G., and Peralta Soler, A. Ornithine decar-boxylase overexpression is a sufficient condition for tumor promotion. Cancer Res.,57: 2630–2637, 1997.

23. Morris, D. R., Davis, R., and Coffino, P. A new perspective on ornithine decarbox-ylase regulation: prevention of polyamine toxicity is the overriding theme. J. Cell.Biochem., 46: 102–105, 1991.

24. Pegg, A. E., Shantz, L. M., and Coleman, C. S. Ornithine decarboxylase: structure,function and translational regulation. Biochem. Soc. Transact., 22: 846–852, 1995.

25. Shantz, L. M., and Pegg, A. E. Translational regulation of ornithine decarboxylaseand other enzymes of the polyamine pathway. Int. J. Biochem. Cell Biol., 31:107–122, 1999.

26. Ivanov, I. P., Matsufuji, S., Murakami, Y., Gesteland, R. F., and Atkins, J. F.Conservation of polyamine regulation by translational frameshifting from yeast tomammals. EMBO J., 19: 1907–1917, 2000.

27. Coffino, P. Regulation of cellular polyamines by antizyme. Nat. Rev. Mol. Cell. Biol.,2: 188–194, 2001.

28. Ivanov, I. P., Rohrwasser, A., Terreros, D. A., Gesteland, R. F., and Atkins, J. F.Discovery of a spermatogenesis stage-specific ornithine decarboxylase antizyme:antizyme 3. Proc. Natl. Acad. Sci. USA, 97: 4808–4813, 2000.

29. Feith, D. J., Shantz, L. M., and Pegg, A. E. Targeted antizyme expression in the skinof transgenic mice reduces tumor promoter induction of ornithine decarboxylase anddecreases sensitivity to chemical carcinogenesis. Cancer Res., 61: 6073–6081, 2001.

30. Pegg, A. E., Feith, D. J., Fong, L. Y. Y., Coleman, C. S., O’Brien, T. G., and Shantz,L. M. Transgenic mouse models for studies of the role of polyamines in normal,hypertophic and neoplastic growth. Biochem. Soc. Transact., 31: 356–360, 2003.

31. Bol, D., Kiguchi, K., Beltran, L., Rupp, T., Moats, S., Gimenez-Conti, I., Jorcano, J.,and DiGiovanni, J. Severe follicular hyperplasia and spontaneous papilloma forma-tion in transgenic mice expressing the neu oncogene under the control of the bovinekeratin 5 promoter. Mol. Carcinog., 21: 2–12, 1998.

32. He, W., Li, A. G., Wang, D., Han, S., Zheng, B., Goumans, M. J., Ten Dijke, P., andWang, X. J. Overexpression of Smad7 results in severe pathological alterations inmultiple epithelial tissues. EMBO J., 21: 2580–2590, 2002.

33. Pierce, A. M., Schneider-Broussard, R., Gimenez-Conti, I. B., Russell, J. L., Conti,C. J., and Johnson, D. G. E2F1 has both oncogenic and tumor-suppressive propertiesin a transgenic model. Mol. Cell. Biol., 19: 6408–6414, 1999.

34. Ramirez, A., Bravo, A., Jorcano, J. L., and Vidal, M. Sequences 5� of the bovinekeratin 5 gene direct tissue- and cell-type-specific expression of a lacZ gene in theadult and during development. Differentiation, 58: 53–64, 1994.

35. DiGiovanni, J., Kiguchi, K., Frijhoff, A., Wilker, E., Bol, D. K., Beltran, L., Moats,S., Ramirez, A., Jorcano, J., and Conti, C. Deregulated expression of insulin-likegrowth factor 1 in prostate epithelium leads to neoplasia in transgenic mice. Proc.Natl. Acad. Sci. USA, 97: 3455–3460, 2000.

3953

AZ-MEDIATED INHIBITION OF FORESTOMACH CARCINOGENESIS

Research. on January 13, 2015. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

36. Fong, L. Y. Y., Li, J. X., Farber, J., and Magee, P. N. Cell Proliferation andesophageal tumorigenesis in the zinc deficient rat. Carcinogenesis (Lond.), 17:1841–1848, 1996.

37. Fong, L. Y. Y., Sivak, A., and Newberne, P. M. Zinc deficiency and methylbenzylni-trosamine induced esophageal cancer in rats. J. Natl. Cancer Inst., 61: 145–150, 1978.

38. Stinson, S. F., Squire, R. A., and Sporn, M. B. Pathology of esophageal neoplasmsand associated proliferative lesions induced in rats by N-methyl-N-benzylnitro-samine. J. Natl. Cancer Inst., 61: 1471–1475, 1978.

39. Lu, S. H., Montesano, R., Zhang, M. S., Feng, L., Luo, F. J., Chui, S. X.,Umbenhauer, D., Saffhill, R., and Rajewsky, M. F. Relevance of N-nitrosaminesto esophageal cancer in China. J. Cell. Physiol., 4(Suppl.): 51–58, 1986.

40. Fong, L. Y. Y., and Magee, P. N. Dietary zinc deficiency enhances esophageal cellproliferation and N-nitrosomethylbenzylamine (NMBA)-induced esophageal tumorincidence in C57BL/6 mouse. Cancer Lett., 143: 63–69, 1999.

41. Fong, L. Y. Y., Ishii, H., Nguyen, V. T., Vecchione, A., Farber, J. L., Croce, C. M.,and Huebner, K. p53 deficiency accelerates induction and progression of esophagealand forestomach tumors in zinc-deficient mice. Cancer Res., 63: 186–195, 2003.

42. Leininger, J. R., and Jokinen, M. P. Tumours of the oral cavity, pharynx, oesophagusand stomach. In: V. S. Turusov and U. Mohr (eds.), Tumours of the Mouse, Ed. 2, pp.167–193. Lyon, France: IARC, 1994.

43. Fong, L. Y. Y., Nguyen, V. T., Pegg, A. E., and Magee, P. N. �-Difluoromethylor-nithine induction of apoptosis: a mechanism which reverses pre-established cellproliferation and cancer initiation in esophageal carcinogenesis in zinc-deficient rats.Cancer Epidemiol. Biomark. Prev., 10: 191–199, 2001.

44. Kerr, J. F. R. Apoptosis, its significance in cancer and cancer therapy. Cancer (Phila.),73: 2013–2026, 1994.

45. Ortega, S., Malumbres, M., and Barbacid, M. Cyclin D-dependent kinases. INK4inhibitors and cancer. Biochim. Biophys. Acta, 1602: 73–87, 2002.

46. Fong, L. Y. Y., Lau, K-M., Huebner, K., and Magee, P. N. Induction of esophagealtumors in zinc deficient rats by single low doses of N-nitrosomethylbenzylamine(NMBA): analysis of cell proliferation, and mutations in H-ras and p53 genes.Carcinogenesis (Lond.), 18: 1477–1484, 1997.

47. Lin, Y., Martin, J., Gruendler, C., Farley, J., Meng, X., Li, B. Y., Lechleider, R., Huff,C., Kim, R. H., Grasser, W. A., Paralkar, V., and Wang, T. A novel link between theproteasome pathway and the signal transduction pathway of the bone morphogeneticproteins (BMPs). BMC Cell Biol, 3: 15, 2002.

48. Daliam, A., Savoure, N., Ramee, M. P., Desrues, B., Dazord, L., and Nicol, M.Ornithine decarboxylase basal activity in liver, oesophagus and lung of vitamin Adeficient rats, and the effect of retinoic acid. Carcinogenesis (Lond.), 9: 2161–2164,1988.

49. Otani, K., Yano, Y., Hasuma, T., Arakawa, T., Kobayashi, K., Matsui-Yuasa, I., andOtani, S. Polyamine metabolism of rat gastric mucosa after oral administration ofhypertonic sodium chloride solution. Am. J. Physiol., 274: G299–G305, 1998.

50. Asamoto, M., Toriyama-Baba, H., Ohnishi, T., Naito, A., Ota, T., Ando, A., Ochiya,T., and Tsuda, H. Transgenic rats carrying human c-Ha-ras proto-oncogene are highlysusceptible to N-nitrosomethylbenzylamine induction of esophageal tumorigenesis.Jpn. J. Cancer Res., 93: 744–751, 2002.

51. McCann, P. P., and Pegg, A. E. Ornithine decarboxylase as an enzyme target fortherapy. Pharmacol. Ther., 54: 195–215, 1992.

52. Steele, V. E., Sharma, S., Mehta, R., Elmore, E., Redpath, J. L., Rudd, C., Bagheri,D., Sigman, C. C., and Kelloff, G. J. The use of in vitro assays to predict the efficacyof chemopreventive agents in whole animals. J. Cell. Biochem., 265: 29–53, 1996.

53. D’Ambrosio, S. M., Gibson-D’Ambrosio, R., Milo, G. E., Casto, B., Kelloff, G. J.,and Steele, V. E. Differential response of normal, premaligant, and malignant humanoral epithelial cells to growth inhibition by chemopreventive agents. Anticancer Res.,20: 2273–2380, 2000.

54. Elmore, E., Stringer, D. E., Steele, V. E., Gerner, E. W., and Redpath, J. L.Chemoprevention by difluoromethylornithine: correlation of an in vitro human cell

assay with human clinical data for biomarker modulation. Anticancer Res., 21:1163–1165, 2001.

55. Schipper, R. G., Penning, L. C., and Verhofstad, A. A. J. Involvement of polyaminesin apoptosis. Facts and controversies: effectors or protectors? Cancer Biol., 10:55–68, 2000.

56. Stefanelli, C., Tantini, B., Fattori, M., Stanic, I., Pignatti, C., Clo, C., Guarnieri, C.,Caldarera, C. C., Mackintosh, C. A., Pegg, A. E., and Flamigni, F. Caspase activationin etoposide-treated fibroblasts is correlated to ERK phosphorylation and both eventsare blocked by polyamine depletion. FEBS Lett., 527: 223–228, 2002.

57. Takahashi, Y., Mai, M., and Nishioka, K. �-Difluoromethylornithine induces apo-ptosis as well as anti-angiogenesis in the inhibition of tumor growth and metastasis ina human gastric cancer model. Int. J. Cancer, 85: 243–247, 2000.

58. Nitta, T., Igarashi, K., and Yamamoto, N. Polyamine depletion induces apoptosisthrough mitochondria-mediated pathway. Exp. Cell Res., 276: 120–128, 2002.

59. Wang, D., Weghorst, C. M., Calvert, R. J., and Stoner, G. D. Mutations in the p53tumor suppressor gene in rat esophageal papillomas induced by N-nitrosomethylben-zylamine. Carcinogenesis (Lond.), 17: 625–630, 1996.

60. Tsuji, T., Usui, S., Aida, T., Tachikawa, T., Hu, G. F., Sasaki, A., Matsumura, T.,Todd, R., and Wong, D. T. W. Induction of epithelial differentiation and DNAdemethylation in hamster malignant oral keratinocyte by ornithine decarboxylaseantizyme. Oncogene, 20: 24–33, 2001.

61. Iwata, S., Sato, Y., Asada, M., Takagi, M., Tsujimoto, A., Inaba, T., and Yamada, T.Anti-tumor activity of antizyme which targets the ornithine decarboxylase requiredfor cell growth and transformation. Oncogene, 18: 165–172, 1999.

62. Koike, C., Chao, D. T., and Zetter, B. R. Sensitivity to polyamine-induced growtharrest correlated with antizyme induction in prostate carcinoma cells. Cancer Res., 59:6109–6112, 1999.

63. Mitchell, J. L., Leyser, A., Holtorff, M. S., Bates, J. S., Frydman, B., Valasinas, A.,Reddy, V. K., and Marton, L. J. Antizyme induction by polyamine analogues as afactor of cell growth inhibition. Biochem. J., 366: 663–671, 2002.

64. Alberts, D. S., Dorr, R. T., Einspahr, J. G., Aickin, M., Saboda, K., Xu, M. J., Peng,Y-M., Goldman, R., Foote, J. A., Warneke, J. A., Salasche, S., Roe, D. J., andBowden, G. T. Chemoprevention of human actinic keratoses by topical 2-(difluoro-methyl)-dl-ornithine. Cancer Epidemiol. Biomark. Prev., 9: 1281–1286, 2000.

65. Carbone, P. P., Pirsch, J. D., Thomas, J. P., Douglas, J. A., Verma, A. K., Larson,P. O., Snow, S., Tutsch, K. D., and Pauk, D. Phase I Chemoprevention study ofdifluoromethylornithine in subjects with organ transplants. Cancer Epidemiol.Biomark. Prev., 10: 657–661, 2001.

66. Burns, M. R., Carlson, C. L., Vanderwerf, S. M., Ziemer, J. R., Weeks, R. S., Cai, F.,Webb, H. W., and Graminski, G. F. Amino acid/spermine conjugates: polyamineamides as potent spermidine uptake inhibitors. J. Med. Chem., 44: 3632–3644, 2001.

67. Graminski, G. F., Carlson, C. L., Ziemer, J. R., Cai, F., Vermeulen, N. M.,Vanderwerf, S. M., and Burns, M. R. Synthesis of bis-spermine dimers that are potentpolyamine transport inhibitors. Bioorg. Chem., 12: 35–40, 2002.

68. Carlton, P. S., Gopalakrishnan, R., Gupta, R., Liston, b. W., Habib, S., Morse, M. A.,and Stoner, G. D. Piroxicam is an ineffective inhibitor of N-nitrosomethylben-zylamine-induced tumorigenesis in the rat esophagus. Cancer Res., 62: 4376–4382,2002.

69. Ushida, J., Sugie, S., Kawabata, K., Pham, Q. V., Tanaka, T., Fujii, K., Takeuchi, T.,Ito, Y., and Mori, H. Chemopreventive effect of curcumin on N-nitrosomethylben-zylamine-induced esophageal carcinogenesis in rats. Jpn. J. Cancer Res., 91: 893–898, 2002.

70. Li, Z. G., Hong, T., Shimada, Y., Komoto, I., Kawabe, A., Ding, Y., Kaganoi, J.,Hashimoto, Y., and Imamura, M. Suppression of N-nitrosomethylbenzylamine(NMBA)-induced esophageal tumorigenesis in F344 rats by resveratrol. Carcinogen-esis (Lond.), 23: 1531–1536, 2002.

3954

AZ-MEDIATED INHIBITION OF FORESTOMACH CARCINOGENESIS

Research. on January 13, 2015. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from

2003;63:3945-3954. Cancer Res   Louise Y. Y. Fong, David J. Feith and Anthony E. Pegg  Carcinogenesis -Nitrosomethylbenzylamine-induced Forestomach

NProliferation, Increases Apoptosis, and Reduces Antizyme Overexpression in Transgenic Mice Reduces Cell

  Updated version

  http://cancerres.aacrjournals.org/content/63/14/3945

Access the most recent version of this article at:

   

   

  Cited Articles

  http://cancerres.aacrjournals.org/content/63/14/3945.full.html#ref-list-1

This article cites by 62 articles, 27 of which you can access for free at:

  Citing articles

  http://cancerres.aacrjournals.org/content/63/14/3945.full.html#related-urls

This article has been cited by 20 HighWire-hosted articles. Access the articles at:

   

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

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

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

  Permissions

  .permissions@aacr.orgDepartment at

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

Research. on January 13, 2015. © 2003 American Association for Cancercancerres.aacrjournals.org Downloaded from