Enhanced Susceptibility to Chemical Induction of Ovarian Tumors in Mice with a Germ Line P53 Mutation

Enhanced Susceptibility to Chemical Induction of Ovarian Tumors in Mice with a Germ Line p53 Mutation

Yian Wang,1 Zhongqiu Zhang,1 Yan Lu,1 Ruisheng Yao,1 Dongmei Jia,1 Weidong Wen,1 Marie LaRegina,2 Keith Crist,3 Ronald Lubet,4 and Ming You1

1Department of Surgery and 2Division of Comparative Medicine, Washington University School of Medicine, St. Louis, Missouri; 3Medical College of Ohio, Toledo, Ohio; and 4Chemoprevention Branch, National Cancer Institute, Bethesda, Maryland

Abstract common ovarian cancers are epithelial cancers, which originate Mice with a germ line p53 mutation (p53Ala135Val/wt) from the ovarian surface lining. This type accounts for f90% display increased susceptibility to lung, skin, and of all ovarian cancers. The second type of ovarian cancers are colon carcinogenesis. Here, we show that p53Ala135Val/wt germ cell tumors, which account for f5% of all ovarian mice developed ovarian tumors significantly cancers. The third type of ovarian cancers are the specialized more rapidly than their wild-type littermates after stromal cell cancers. Although it accounts for only <5% of all 7,12-dimethylbenz(a)anthracene (DMBA) treatment. ovarian cancers, it is a highly aggressive form of cancer. Approximately 50% of the ovarian tumors in Ovarian cancer is idiopathic, although it is associated with p53wt/wt mice and 23% in p53Ala135Val/wt mice are multiple genetic alterations. Dysfunction of p53 is one of the adenocarcinomas and the remaining tumors were most common genetic alterations found in cancer, occurring in adenocarcinoma mixed with sarcoma or ovarian up to 50% of all human malignancies (2). The p53 tumor sarcomas. All of the p53Ala135Val/wt mice had died of suppressor protein is a central regulator of cell growth, DNA ovarian tumors 25 weeks after the initial DMBA damage repair, and apoptosis (3). It directly activates the treatment, whereas >50% of p53wt/wt mice were still expression of a substantial number of genes important for alive. These mice not only have a shortened tumor cell cycle regulation and apoptosis (4). Although, p53 gene latency but also closely resemble a subset of human alterations are rare events in benign ovarian tumors, previous ovarian tumors containing the p53 mutation. studies have shown that p53 is mutated in f50% of late stage Microarray and GenMAPP analyses revealed that (stage III/IV) ovarian cancers (5). Furthermore, mutated p53 the mutant p53 (Ala135Val) affected several cellular genes with resultant overexpression of p53 proteins occurs processes, including the cell cycle, apoptosis, frequently in ovarian sarcomas (6). Ala135Val/wt and Wnt pathways. These findings indicate that a We reported that p53 mice are highly susceptible germ line p53 mutation significantly enhanced to carcinogen-induced tumors including lung (7) and skin DMBA-induced ovarian tumor development and tumors (8). However, the role of germ line p53 mutation in progression. (Mol Cancer Res 2008;6(1):99–109) mouse ovarian carcinogenesis has not been examined previously in this rodent model. For this reason, we have carried out an ovarian carcinogenesis study with 7,12-dimethylben- Introduction z(a)anthracene (DMBA). Here, we report that p53Ala135Val/wt Ovarian cancer is the fifth leading cause of cancer death in mice display significantly increased susceptibility to DMBA- women, and is the leading cause of death from gynecologic induced ovarian tumor carcinogenesis. Alterations of several malignancy. It has been estimated that there were 20,180 new cellular processes, including cell cycle arrest, apoptosis, and cases and 15,310 ovary cancer deaths in the United States Wnt pathways, may mediate enhanced DMBA-induced ovarian during 2006 (1). Ovarian cancer is classified into three major tumor carcinogenesis in p53Ala135Val/wt mice. groups based on their distinct histologic origins. The most Results and Discussion Three months post-DMBA treatment, p53Ala135Val/wt mice Received 5/10/07; revised 8/28/07; accepted 9/4/07. had developed significantly more advanced tumor phenotypes Grant support: National Cancer Institute, NIH (R01CA113793 and (larger ovarian tumors) compared with their wild-type litter- N01CN43308). mates. The survival rate of p53wt/wt mice was significantly The costs of publication of this article were defrayed in part by the payment of Ala135Val/wt page charges. This article must therefore be hereby marked advertisement in higher than p53 mice (Fig. 1). Most of the accordance with 18 U.S.C. Section 1734 solely to indicate this fact. p53Ala135Val/wt mice were dead or moribund before the end of Note: Supplementary data for this article are available at Molecular Cancer the experiment due to advanced ovarian tumors. On an average Research Online (http://mcr.aacrjournals.org/). Ala135Val/wt Requests for reprints: Ming You, Department of Surgery, The Washington of 17 weeks post-DMBA treatment, 100% of the p53 University School of Medicine, 660 South Euclid Avenue, St. Louis, MO 63110. mice (40 of 40) had developed tumors. Whereas in the p53wt/wt Phone: 314-462-9294; Fax: 314-362-9366. E-mail: [email protected] Copyright D 2008 American Association for Cancer Research. group, 40 weeks posttreatment, >80% (28 of 35) of the mice doi:10.1158/1541-7786.MCR-07-0216 had developed tumors. No mice were found dead or sick in

FIGURE 1. DMBA-induced ovarian sarcomas in p53Ala135Val/wt transgenic mice. Both p53Ala135Val/wt transgenic mice and their wild-type littermates were monitored for the development of ovarian tumors. Moribund animals were terminated, and the tissues were collected for histopathologic and molecular analysis. Genotypes: E, wild-type; n,p53Ala135Val/wt. Total number of animals in each group: 35 mice in the p53wt/wt group; 40 mice in the p53Ala135Val/wt group.

either of the p53wt/wt or p53Ala135Val/wt DMSO control groups. including the pancreas, spleen, uterus, and kidney. For Morphologically, mice with tumors had an enlarged abdomen. comparison, the normal structure of the ovary was stained DMBA-treated ovaries were grossly enlarged and hyperemic, and are shown in Fig. 2 (A, B, and C), and the ovarian tumors occupying the pelvis and part of the abdominal cavity. Tumor in (D, E, and F) at 40Â, 100Â, and 400Â magnifications, size varied from 1.0 to 3.5 cm in diameter. The size of the respectively. Adenocarcinoma cells (Fig. 2F, a) are arranged p53Ala135Val/wt tumors was twice as large as that of the p53wt/wt with glandular structures in back-to-back nests which are tumors. The tumors presented as large masses with reddish separated by lumina. The nuclear/cytoplasmic ratio is high in color from the ovary on the left side, and aggressively invaded these cells. The nuclei show obvious pleomorphisms, granular surrounding organs including the spleen, intestine, kidney, chromatin, and large plump nuclei with prominent nucleoli. and uterus. The control ovaries (right side) looked normal at Sarcoma cells (Fig. 2F, b) are arranged in bundles and streams f2 mm in diameter. Histologically, 50% (4 of 8) of the ovarian and are in an overall spindle shape. Their nuclei are elongated tumors in p53wt/wt mice and f23% (3 of 13) of the ovarian or oval and variable in size. Although the nuclei contained tumors in p53Ala135Val/wt mice are adenocarcinomas (Fig. 2F, a). coarse to clumped chromatin and the occasional prominent Some of the adenocarcinomas consisted of areas with some nucleoli, the cytoplasm was eosinophilic and fibrillar. There is degree of sarcoma features. The remaining tumors displayed also no significant difference in the mitotic index between typical features of ovarian sarcomas (Fig. 2F, b). Approximately mutant p53 mice and wild-type p53 mice. The average mitotic 40% of the tumors had invaded into the surrounding organs, index was 2 F 0.875 per field under a 400Â magnification.

FIGURE 2. Histopathology of ovarian tumors. Histologically, 50% of the ovarian tumors in p53wt/wt mice and f23% of the ovarian tumors in p53Ala135Val/wt mice are adenocarcinomas. Some of the adenocarcinomas consisted of areas with some degree of sarcoma features. The remaining tumors displayed typical features of ovarian sarcomas. Adenocarcino- ma cells are arranged with glandular structures in back-to-back nests which are separated by lumina. The nuclear/cytoplasmic ratio is high in these cells. The nucleus shows obvious pleomorphism, granular chromatin, and large plump nuclei with prominent nucleoli. Sarcoma cells are arranged in bundles and streams and are in an overall spindle shape. Their nuclei are elongated or oval and variable in size. Although the nuclei contained coarse to clumped chromatin and the occasional prominent nucleoli, the cytoplasm was eosinophilic and fibrillar. A normal mouse ovary [magnification, Â40 (A), Â100 (B), and Â400 (C)]. DMBA-induced ovarian tumors [magnification, Â40 (D), Â100 (E), and Â400 (F; a, adenocarcinoma; b, sarcoma)].

Mol Cancer Res 2008;6(1). January 2008 Downloaded from mcr.aacrjournals.org on September 29, 2021. © 2008 American Association for Cancer Research. Induction of Ovarian Tumors in p53 Transgenic Mice 101

No tumors were found in either p53wt/wt or p53Ala135Val/wt Tcfec, Gata4, etc.), as well as the oncogene Myb and tumor DMSO control mice. There were no statistically significant suppressor gene Wt1. In Table 2, genes altered only in p53wt/wt differences between the body weights of any of the groups tumors include genes related to cancer (Ris2, Srcasm, and treated with DMBA as compared with the controls. However, Sdccag33l), cell cycle regulation (cdc6, Bub1, Bub1b, Ccnb2, the majority of p53Ala135Val/wt mice were terminated before the Nusap1, and Trp63), cell growth and differentiation (Edg7 and end of the experiment due to the abundance of tumors. The Mal), oncogenesis (Rin1, Src, Brca1, Ect2, and Erbb3), signal survival data was shown in Fig. 1. Our data clearly showed that transduction (Rad23b, Rgs2, Stard5, Gpr56, Rab3d, Rerg, p53Ala135Val/wt mice have an increased susceptibility to DMBA- Mark1, Rab25,andFrzb), and transcriptional regulation induced ovarian tumors. Because spontaneous ovarian tumors (Rcor1). In Table 2, genes altered only in p53Ala135Val/wt are rare in the mouse, DMBA-induced ovarian tumors in tumors include genes involved in apoptosis (Perp and Bcl2l10), p53Ala135Val/wt mice has become a potential model for human a cancer-related gene (Gfra1), cell growth and differentiation ovarian tumors to aid our studies on ovarian cancer chemo- (Csf1 and FGFr3), signal transduction (Gpr85, Arhgdib, prevention, early detection, and therapy. Rad51ap1, Gpr155, and Gnai1), and transcriptional regulation The RASmitogen-activated protein kinase signal transduc- (Gata6 and Tcfcp2l2). tion pathway has been shown to interact with the ARF-p53 Genes that were altered only in p53Ala135Val/wt tumors as pathway in the processes of cell transformation and carcino- compared with p53wt/wt tumors and only in p53val135/wt normal genesis. We did K-ras2 mutation analyses on a total of 20 ovary as compared with p53wt/wt normal ovary (Table 1A; tumors (10 from p53Ala135Val/wt mice and 10 from p53wt/wt Fig. 3A) are listed in Table 3. The tumor-related genes are mice). The direct sequencing analysis revealed that 30% of the involved in apoptosis (Bcl7a and Biklk), tumorigenesis (Xpa, tumors (6 out of 20) harbored mutations in codons 13 and 61 Gfra1, Bcas3, Mas1, Hrasls, and Vav2), cell cycle regulation (Supplementary Table S1). There was no mutation identified in (Ccna1, Cwf19l1, and Cdkn2a), cell growth and differentiation codon 12. Also, there was no significant difference in the (IGFals, Neurod1, and FGF6), signal transduction (Gpr74, incidence and types of mutations between tumors from p53wt/wt Rab3c, Wnt8a, Mark1, Gpr73, Fzd5,andRhpn2), and and p53Ala135Val/wt mice (Supplementary Table S1). transcriptional regulation (Pou6f1, Pou4f3, Pou2f1,and Microarray analyses were then done on normal and tumor Tcfcp2l2). We used GenMAPP to illustrate all the signaling tissues from ovaries treated with DMBA. Differential gene pathways containing differentially expressed genes associated expression was based on p53 genotype status and ovarian tumor with p53 status in ovarian tumorigenesis. Several cellular phenotype (z2.0-fold change, P V 0.05). A total of 20 samples pathways showed alterations in ovarian tumors (yellow repre- (five from each group, normal and tumor samples from both sents the changes in wild-type tumors and blue represents p53wt/wt and p53Ala135Val/wt mice) were examined with micro- the changes in p53Ala135Val/wt tumors with their correspond- array analysis. In order to see the genetic effect of mutated p53 ing normal as control). The altered pathways include ones on gene expression in mouse ovary tissues, we compared the that regulate apoptosis (Supplementary Fig. S1), cell cycle expression profiles of normal ovaries of p53Ala135Val/wt mice to regulation (Supplementary Fig. S2), and Wnt (Supplementary those of p53wt/wt mice. We identified 30 genes with at least Fig. S3). These indicated that reduction of apoptotic activity 2.5-fold change in expression (P V 0.05; Table 1A). Of these, and altered cell cycle regulations contribute, at least in 15 were underexpressed and 15 were overexpressed in part, to the observed increase in ovarian malignancy p53Ala135Val/wt mice. The overexpressed genes include transfor- predisposition in mice treated with DMBA. It is important mation-related protein 53, early growth response 2, peroxisome to note that the majority of cells in the normal ovary are of proliferator-activated receptor binding protein, arginine vaso- the stromal subtype, and that some of the differences in the pressin receptor 2, defensin-related cryptdin 6, and phosphoi- expression profiles between normal ovaries and adenocarci- nositide-3-kinase adaptor protein 1. Underexpressed genes nomas might be due to the enrichment of the epithelial cells include ELAV-like 3, carbonic anhydrase 1, islet amyloid in adenocarcinomas. polypeptide, integrin h7, prolactin-induced protein ubiquitin C, Among the genes with altered expression, we selected 12 and forkhead box N1. for real-time PCR validation using qPCR analysis (Bio-Rad Interestingly, the elevated expression levels of 12 genes seen Laboratories) on the Bio-Rad iCycler iQ Real Time PCR in the p53Ala135Val/wt normal samples were also maintained at Instrument. Mouse glyceraldehyde-3-phosphate dehydroge- the same or even higher levels in the p53Ala135Val/wt tumors nase was used as the reference control for normalization (Table 1B). Similarly, the same pattern was seen with two of template. Ten of the 12 genes were correctly confirmed down-regulated genes, IFN-activated gene 205 and ELAV (Fig. 3B); the validating rate was 86%. Figure 3B shows 10 (embryonic lethal, abnormal vision, Drosophila)-like 3 (Table of the confirmed genes (two no-change genes were not 1B). We have identified a large set of genes which are shown). In a comparison of p53wt/wt tumor to p53wt/wt normal, modulated during tumorigenesis regardless of p53 status. Only Cdh1, Bak1, Apaf1, Cdc20, CyclinB2, cdc25C, and Bub1 were the genes with >5-fold change (P < 0.05) are listed in Table 2. confirmed. Casp7 was not confirmed. In a comparison of These genes play important roles in several critical biological p53Ala135Val/wt tumor to p53Ala135Val/wt normal, Cdh1, CyclinA2, functions, including apoptosis (Bak1, Pawe, and Tnfrsf19), and cdc25c were confirmed. Bad was not confirmed. tumorigenesis (Fosl1, Glipr1, Plk1, etc.), cell cycle regulation The ovary is an organ formed of many different cell types. (Cdkn2a, Cdc25c, Ccnb1, etc.), cell growth and differentiation Tumors that arise from this organ include epithelial ovarian (TGFa, EGFl6, FGFr2, etc.), signal transduction (Rgs16, carcinomas, germ cell tumors, and stroma tumors. There are Racgap1, Prkca, etc.), and transcriptional regulation (Etv4, several transgenic mouse models established for the ovarian

Table 1. Differentially Expressed Genes in Normal Ovaries of p53Ala135Val/wt Mice

(A) Differentially expressed genes identified by microarray analysis in normal ovaries of p53Ala135Val/wt mice with normal ovaries of p53wt/wt mice as control (12 unknown sequences were not listed)

Accession no. Unigene name Unigene Fold

BB828014 Transformation-related protein 53 Trp53 10.8 BG061809 MutShomologue 5 ( Escherichia coli) Msh5 4.4 NM_019690 GNAScomplex locus Gnas 3.7 NM_010502 IFN a family, gene 1 Ifna1 3.6 X06746 Early growth response 2 Egr2 3.6 NM_138648 Oxidized low density lipoprotein (lectin-like) receptor 1 Olr1 3.0 NM_007784 Casein a Csna 2.9 BB760479 Peroxisome proliferator-activated receptor binding protein Pparbp 2.7 NM_019404 Arginine vasopressin receptor 2 Avpr2 2.7 NM_007845 Defensin-related cryptdin 6 Defcr6 2.7 NM_015766 EBV-induced gene 3 Ebi3 2.7 NM_013631 Pyruvate kinase liver and RBC Pklr 2.7 BF020871 BTB (POZ) domain containing 1 Btbd1 2.7 NM_031376 Phosphoinositide-3-kinase adaptor protein 1 Pik3ap1 2.5 BC002065 Serine/cysteine proteinase inhibitor, clade A, member 3G Serpina3g 2.5 AU067745 ELAV-like 3 Elavl3 À6.1 BC011223 Carbonic anhydrase 1 Car1 À3.5 BB434117 Islet amyloid polypeptide Iapp À3.4 AF217002 Two pore channel 1 Tpcn1 À3.3 NM_010775 Mannose binding lectin, liver (A) Mbl1 À3.1 BC003725 ADP-ribosylation factor-like 4 Arl4 À3.1 AF374470 CD209a antigen Cd209a À3.0 NM_013566 Integrin h7 Itgb7 À2.9 NM_008018 SH3 multiple domain 1 Sh3md1 À2.8 NM_020516 Solute carrier family 16, member 8 Slc16a8 À2.7 U76759 NF of activated T cells, calcineurin-dependent 2 interacting Nfatc2ip À2.6 BC018264 Prolactin-induced protein Pip À2.6 D50527 Ubiquitin C Ubc À2.5 NM_008238 Forkhead box N1 Foxn1 À2.5 NM_027052 Solute carrier family 38, member 4 Slc38a4 À2.5

(B) Altered expression of p53 mutation-related genes

Accession no. Unigene gene name Unigene gene symbol Mean Fold

+/+ N +/À N+/À T+/ÀN vs. +/+N +/ÀT vs. +/+ N

BB828014 Transformation-related protein 53 Trp53 555 5,993 11,934 10.8 21.5 BG061809 MutShomologue 5 ( E. coli) Msh5 42 184 240 4.4 5.7 X06746 Early growth response 2 Egr2 258 796 681 3.1 2.6 NM_138648 Oxidized low-density lipoprotein (lectin-like) receptor 1 Olr1 52 157 255 3.0 4.9 NM_007784 Casein a Csna 28 81 215 2.9 7.7 BF020871 BTB (POZ) domain containing 1 Btbd1 26 70 161 2.7 6.2 NM_013631 Pyruvate kinase (liver and RBC) Pklr 52 137 160 2.7 3.1 AV026617 FBJ osteosarcoma oncogene Fos 1,498 3,208 2,335 2.1 1.6 NM_010115 Kallikrein 26 Klk26 65 137 225 2.1 3.5 NM_007913 Early growth response 1 Egr1 3,513 7,372 6,081 2.1 1.7 AI481797 IFN activated gene 205 Ifi205 3,776 1,690 527 À2.2 À7.2 AU067745 Embryonic lethal, abnormal vision, Drosophila, like 3 Elavl3 164 27 78 À6.1 À2.1

NOTE: +/+, p53wt/wt;+/À, p53Ala135Val/wt; N, normal ovarian tissue; T, ovarian tumor.

tumor. In one mouse model, re-introduction of ovarian cells transplantation of ovaries from p53 homozygous knockout exhibiting loss of p53 in combination with two activated mice into histocompatible wild-type female hosts induces oncogenes (c-Myc, K-Ras, or Akt) to the ovarian bursa results angiosarcoma rather than epithelial malignancy 1 year post- in the development of ovarian carcinomas (9). In another transplantation (13). model, 97% of the animals developed spontaneous ovarian Here, we report the induction of ovarian tumors by DMBA carcinomas in conditionally inactivated p53 and Rb mice (10). in p53Ala135Val/wt transgenic mice and the identification of the Also, Connolly et al. reported that TgMISIIR-TAg transgenic potential molecular mechanisms involved. The ovarian tumors mice developed bilateral ovarian carcinomas accompanied induced by DMBA treatment in p53Ala135Val/wt mice were by ascites and peritoneal implants (11). These tumors have diagnosed as either adenocarcinomas or sarcomas. Nearly half all been characterized histopathologically as ovarian adeno- of the tumors invaded into its surrounding organs. Although we carcinoma derived from the ovarian surface epithelial cells. observed 100% and 80% tumor frequencies in p53Ala135Val/wt There have been reports that radiation induces ovarian and p53wt/wt mice, respectively (data not shown), the granulosum cell tumors (12). Chen et al. reported that surgical p53Ala135Val/wt mice developed more aggressive tumors with a

Table 2. Differentially Expressed Genes Identified by Microarray Analysis in the Comparison of Tumor to Normal p53wt/wt and p53Ala135Val/wt Mice and in Human Ovarian Tumors

Accession no. Unigene name Unigene symbol Mouse Human

p53wt/wt T/N p53Ala135Val/wt T/N p53wt/wt T/N p53mut/wt T/N

Apoptosis AF402617 BCL2-antagonist/killer 1 Bak1 5.6 6.0 2.4 2.5 NM_009684 Apoptotic protease-activating factor 1 Apaf1 2.5 2.2 2.8 2.3 BB398886 PRKC, apoptosis, WT1, regulator Pawr À8.1 À10.8 À1.7 À1.8 NM_013869 TNF receptor superfamily, member 19 Tnfrsf19 À19.5 À20.5 NM_022032 PERP, TP53 apoptosis effector Perp À6.9 NM_013479 Bcl2-like 10 Bcl2l10 À8.7 Cancer-related genes NM_010235 Fos-like antigen 1 Fosl1 49.0 31.1 3.6 2.1 BC025083 GLI pathogenesis-related 1 (glioma) Glipr1 11.5 17.6 2.0 1.5 NM_011121 Polo-like kinase 1 (Drosophila) Plk1 7.0 7.4 1.7 2.5 AF477481 Retroviral integration site 2 Cdt1 6.3 2.1 2.9 AF335325 DNA-damage-inducible transcript 4-like Ddit4l À5.8 À11.7 BC004710 Src activating and signaling molecule Tom1l1 À6.2 BC028776 Colon cancer antigen 33 like Tshz2 À7.0 NM_013864 N-myc downstream regulated gene 2 Ndrg2 À13.2 À6.3 À4.9 À6.2 NM_015764 Regulated by estrogen in breast Greb1 À19.3 À18.6 cancer protein AF055573 Fragile histidine triad gene Fhit À20.9 À10.2 AF015172 Neurotrophic factor family receptor a1 Gfra1 7.3 Cell cycle regulation NM_009877 Cyclin-dependent kinase inhibitor 2A (p16) Cdkn2a 39.2 8.2 5.7 17.3 NM_009860 Cdc25 homologue C (Saccharomyces Cdc25c 9.5 8.7 4.3 6.6 cerevisiae) X58708 Cyclin B1 Ccnb1 9.4 7.5 2.7 3.5 NM_023223 Cdc20 homologue (S. cerevisiae) Cdc20 7.9 5.5 X75483 Cyclin A2 Ccna2 7.7 6.3 5.2 5.5 NM_025415 Cdc28 protein kinase regulatory subunit 2 Cks2 7.3 5.5 3.0 4.8 M64403 Cyclin D1 Ccnd1 6.9 5.9 37.0 20.7 NM_007659 Cdc2 homologue A (Schizosaccharomyces Cdc2a 6.6 5.7 5.9 9.1 pombe) NM_011799 Cdc6 homologue (S. cerevisiae) Cdc6 6.5 3.4 4.5 AK017841 E2F transcription factor 1 E2f1 6.5 6.6 6.8 9.4 AF002823 Budding uninhibited by benzimidazoles 1 Bub1 6.4 2.6 3.9 NM_009773 Budding uninhibited by benzimidazoles 1h Bub1b 5.9 4.5 5.6 AK013312 Cyclin B2 Ccnb2 5.4 4.7 7.4 NM_133851 Nucleolar and spindle-associated protein 1 Nusap1 5.3 2.3 3.9 NM_007633 Cyclin E1 Ccne1 3.5 5.6 8.6 16.9 NM_011641 Transformation-related protein 63 Trp63 À7.0 NM_009864 Cadherin 1 Cdh1 À20.9 À13.8 Growth/differentiation U65016 Transforming growth factor a Tgfa À6.4 À6.2 NM_019397 EGF-like-domain, multiple 6 Egfl6 À6.5 À12.8 NM_022983 Endothelial differentiation, G protein – Edg7 À6.8 coupled receptor 7 NM_010207 Fibroblast growth factor receptor 2 Fgfr2 À6.8 À7.6 NM_007836 Growth arrest, DNA damage – Gadd45a À8.0 À6.6 inducible 45a NM_009886 Cadherin EGF LAG seven-pass G-type Celsr1 À9.1 À14.1 receptor 1 AK011784 Insulin-like growth factor binding protein 2 Igfbp2 À11.1 À12.2 BM932451 Fibroblast growth factor 1 Fgf1 À16.2 À5.8 NM_010762 T-cell differentiation protein Mal À16.9 NM_016719 Growth factor receptor bound protein 14 Grb14 À20.3 À12.0 À2.7 À4.1 M21149 Colony-stimulating factor 1 (macrophage) Csf1 5.4 2.6 3.6 BB732903 Fibroblast growth factor receptor 3 Fgfr3 À7.0 Oncogene and tumor suppressor gene BC011277 Ras and Rab interactor 1 Rin1 6.4 1.6 BG868120 Rous sarcoma oncogene Src 6.3 U36475 Breast cancer 1 Brca1 5.5 2.3 NM_007900 Ect2 oncogene Ect2 5.1 9.1 15.4 BF140685 v-Erb-b2 oncogene homologue 3 Erbb3 À5.3 NM_033597 Myeloblastosis oncogene Myb À6.4 À8.6 M55512 Wilms tumor homologue Wt1 À23.2 À17.7 Signal transduction U72881 Regulator of G protein signaling 16 Rgs16 9.9 23.9 1.8 NM_012025 Rac GTPase-activating protein 1 Racgap1 8.2 7.3 4.3 5.4 BF138887 RAD23b homologue (S. cerevisiae) Rad23b 5.9 2.3 BM246539 Protein kinase C, a Prkca 5.4 5.1 3.8 3.5 BI658627 Secreted frizzled-related sequence protein 1 Sfrp1 À5.3 À9.6 À3.5 À4.9

(continued on the following page)

Table 2. Differentially Expressed Genes Identified by Microarray Analysis in the Comparison of Tumor to Normal p53wt/wt and p53Ala135Val/wt Mice and in Human Ovarian Tumors (Cont’d)

Accession no. Unigene name Unigene symbol Mouse Human

p53wt/wt T/N p53Ala135Val/wt T/N p53wt/wt T/N p53mut/wt T/N

AK004849 Rhophilin, Rho GTPase binding protein 2 Rhpn2 À5.4 À10.2 BG067321 Regulator of G protein signaling 2 Rgs2 À5.4 À15.4 À25.1 BI076697 START domain containing 5 Stard5 À5.4 NM_018882 G protein – coupled receptor 56 Gpr56 À5.8 AF263365 RAB3D, member RASoncogene family Rab3d À6.1 BC026463 RAS-like, estrogen-regulated, Rerg À6.2 growth-inhibitor BM213279 MAP/microtubule affinity-regulating Mark1 À6.3 kinase 1 NM_016899 RAB25, member RASoncogene family Rab25 À7.2 BC019741 Regulator of G protein signaling 11 Rgs11 À8.2 À6.1 AV101562 Guanine nucleotide binding protein, a 14 Gna14 À8.7 À5.7 À2.7 À4.1 NM_011356 Frizzled-related protein Frzb À15.3 À2.8 BC005618 Tumor-associated calcium signal Tacstd1 À24.9 À16.4 transducer 1 BC026975 G protein – coupled receptor 85 Gpr85 5.7 2.5 AK002516 Rho, GDP dissociation inhibitor h Arhgdib 5.7 BC003738 RAD51-associated protein 1 Rad51ap1 5.6 3.9 14.8 BB762731 G protein – coupled receptor 155 Gpr155 À5.4 BQ174580 G protein, a inhibiting 1 Gnai1 À7.4 À1.8 Transcription factor X63190 E1A enhancer binding protein Etv4 30.7 23.8 8.9 9.4 NM_054048 RE1-silencing transcription factor Rcor1 6.0 1.6 2.1 corepressor 1 NM_031198 Transcription factor EC Tcfec 5.6 5.5 AB075549 GATA binding protein 4 Gata4 À7.5 À11.7 À9.0 À11.8 BC004672 Liver-specific bHLH-Zip transcription Lisch7 À8.7 À7.3 factor NM_053085 Transcription factor 23 Tcf23 À34.3 À6.5 NM_011545 Transcription factor 21 Tcf21 À37.1 À21.2 À11.8 À25.8 BC006789 Ets homologous factor Ehf À41.6 À21.2 BM214048 GATA-binding protein 6 Gata6 À6.3 BC013080 Transcription factor CP2-like 2 Tcfcp2l2 À5.6

NOTE: Genes were differentially expressed >5.0-fold and P < 0.05; real-time PCR – confirmed (values in boldface, P < 0.05). N, normal ovarian tissue; T, ovarian tumor.

shorter survival time. At 25 weeks post-DMBA treatment, Through microarray analysis, we identified clusters of genes 100% of the p53Ala135Val/wt mice had died, whereas only 50% of which might hold the keys to explaining the mechanisms the p53wt/wt mice had died. Our study has established a unique underlying (a) DMBA-induced ovarian carcinogenesis and (b) animal model for further investigation of this disease. aggressive behavior of the ovarian tumors in p53Ala135Val/wt We previously reported that similar treatment with DMBA mice. One cluster of genes was altered in the normal ovary of sutures in Wistar Furth rats resulted in the development of p53Ala135Val/wt mice as compared with the levels in the normal primarily ovarian adenocarcinomas and ovarian granulosum ovary of p53wt/wt mice (Supplementary Table S1). Because cell tumors (14). Similarly, we observed that 50% of the ovarian the only difference that existed in the normal tissues was the tumors were adenocarcinomas and the remaining 50% were p53 germ line mutation, we can predict that altered expression ovarian sarcomas in p53wt/wt mice. Approximately 23% of could be the consequence of this mutation. Interestingly, the the ovarian tumors were adenocarcinomas and 77% were expression levels of some altered expressions (including Trp53, sarcomas in p53Ala135Val/wt mice. Thus, p53Ala135Val/wt mice Msh5, Egr2, Olr1, Ifi205, and Elavl3) were maintained at a seem to be a model for human ovarian tumors because alte- similar level in ovarian tumors of p53Ala135Val/wt mice as well ration of the p53 gene is the most frequently identified genetic (Table 1). One such gene is MSH5 (MutS homologue 5), which change in ovarian carcinomas, affecting >50% of advanced is a member of the DNA mismatch repair protein family which and early stage carcinomas (5, 6). The role of the p53 tumor play important roles in maintaining accurate genetic transmis- suppressor gene in malignancy has been intensely investi- sion in both mitotic and meiotic processes (15, 16). In our case, gated in the past. It has been shown that mutation of the p53 Msh5 overexpression might not contribute to ovarian tumori- is the most frequently identified genetic alteration in ovarian genesis. Egr1 is a member of the immediate-early gene family. tumors as it affects >50% of advanced carcinomas and ovarian It is involved in the regulation of cell growth and differentia- sarcomas (5). In the current studies, we show that p53 is an tion in response to signals, such as mitogens, growth factors, important factor in ovarian carcinogenesis induced by DMBA: and stress stimuli (17, 18). Egr1 could act as both a tumor the germ line mutation of p53 leads to an increased response suppressor and a tumor promoter in a tissue-specific manner. of ovarian cells to DMBA treatment, possibly, through im- In most human tumors, such as breast cancer, fibrosarcoma, and pairment of the ability to control cell cycle regulation and glioblastoma, Egr1 is described as a tumor suppressor gene apoptosis. (19). One of the mechanisms of suppression has been

FIGURE 3. Differential gene expression in DMBA-induced ovarian sarcomas in p53Ala135Val/wt transgenic mice. A. Gene clusters. Approximately 2,700 genes and expressed sequence tags were differentially expressed in at least one of the groups (wt/wt T versus wt/wt N, Ala135Val/wt T versus Ala135Val/wt N, Ala135Val/wt N versus wt/wt N, and Ala135Val/wt T versus wt/wt T; P < 0.05, t test). Green, expression below the mean value for the gene; black, near the mean; and red, above the mean. Two clusters (1 and 2) of genes had expression levels which were altered in p53Ala135Val/wt tumors only. The only known genes are listed. B. Confirmation of gene expression changes of selected genes from microarray analyses by real-time PCR. Solid columns, real-time PCR data; shaded columns, microarray expression data. The fold change is indicated in the Y-axis (above 0, increased expression; below 0, decreased expression in tumors as compared with normals).

Table 3. Differentially Expressed Genes (>2.0-Fold, P < 0.05) only in p53Ala135Val/wt Tumor (T+/À)

À Accession no. Unigene gene name Unigene symbol T+/ vs. T+/+ Function

NM_008340 Insulin-like GF binding protein Igfals 5.1 Growth/differentiation BC017640 B cell CLL/lymphoma 7A Bcl7a 4.6 Apoptosis NM_008552 MAS1 oncogene Mas1 4.4 Oncogene NM_010127 POU domain class 6 transcription factor 1 Pou6f1 3.8 Transcription factor BM117916 XP, complementation A Xpa 3.5 Cancer-related gene AF236084 G protein – coupled receptor 74 Gpr74 3.5 Signal transduction AK014050 RAB3C, member RASoncogene family Rab3c 3.5 Signaltransduction NM_007628 Cyclin A1 Ccna1 3.2 Cell cycle regulation AK005073 Neurogenic differentiation 1 Neurod1 2.9 Growth/differentiation NM_013751 HRAS-like suppressor Hrasls 2.8 Oncogene NM_009290 Wingless-related MMTV integration 8A Wnt8a 2.8 Signal transduction M92416 Fibroblast growth factor 6 Fgf6 2.6 Growth/differentiation BM213279 MAP/microtubule affinity-regulating kinase 1 Mark1 2.6 Signal transduction NM_021381 G protein – coupled receptor 73 Gpr73 2.5 Signal transduction NM_138945 POU domain, class 4, transcription factor 3 Pou4f3 2.5 Transcription factor NM_007546 Bcl2-interacting killer-like Biklk 2.4 Apoptosis AF015172 Glial cell – derived neurotrophic factor receptor a1 Gfra1 2.4 Cancer-related gene NM_022721 Frizzled homologue 5 (Drosophila) Fzd5 2.3 Signal transduction BB749215 CWF19-like 1, cell cycle control (S. pombe) Cwf19l1 2.2 Cell cycle regulation AW701903 Breast carcinoma – amplified sequence 3 Bcas3 À2.0 Cancer-related gene AK004849 Rhophilin, Rho GTPase binding protein 2 Rhpn2 À2.0 Signal transduction NM_009500 Vav2 oncogene Vav2 À2.1 Oncogene X56230 POU domain, class 2, transcription factor 1 Pou2f1 À2.3 Transcription factor NM_009877 Cyclin-dependent kinase inhibitor 2A Cdkn2a À2.9 Cell cycle regulation BC013080 Transcription factor CP2-like 2 Tcfcp2l2 À3.8 Transcription factor

implicated in the regulation of p53 leading to apoptosis (20). many important processes within eukaryotic cells during However, the tumor-suppressing role seems to be tissue- interphase (26). During cell division, dynein functions in the specific because recent studies implicated the tumor growth– formation, alignment, and maintenance of mitotic spindles, as promoting role of Egr1 in prostate cancer progression (21). well as the movement and positioning of chromosomes (27). Higher levels of Egr1 were found in prostate cancer (22). Egr1 The cytoplasmic dynein complex is composed of four subunit is growth-promoting for vascular smooth muscle cells and for classes: the heavy, intermediate, light-intermediate, and light rat kidney tumor cells (23, 24). Our results suggest that the chains. Dynein light chain 1 (DLC1) interacts with BimL, a Egr1 gene acts as tumor promoters in DMBA-induced ovarian Bcl-2 family member (28) and Pak1 (29). Pak1 is the p21- tumorigenesis as their expression levels were increased in activated kinase 1, which could form a complex with DLC1 and tumors. However, it is more likely due to the increased BimL to phosphorylate both proteins. Phosphorylated BimL is p53Ala135Val/wt expression in the tissues. This hypothesis is then prevented from interacting with and inactivating Bcl-2. supported by recent observations that p53 is a direct target of The interaction of DLC1-Pak1-BimL promotes cell survival by Egr1 as the p53 promoter contains two Egr1-binding sites (25). blocking BimL’s apoptotic functions. Overexpression of DLC1 We identified two clusters (A and B) of genes with altered is detected in human breast tumors and breast cancer cells, expression only in p53Ala135Val/wt ovarian tumors as compared and up-regulation of DLC1 exists with overexpressed Pak1 in with p53Ala135Val/wt normal ovary. No expression changes were 50% of breast cancers (29). Deregulation of Pak1 expression identified in these genes in p53wt/wt ovarian tumors compared is reported in breast and ovarian cancers (30). Although up- with p53wt/wt normal ovary (Fig. 3). Cluster A includes regulation of dynein is p53 dependent, we saw overexpressed overexpressed genes, and cluster B contained underexpressed Pak1 in ovarian tumors regardless of p53 status. BimL is not genes. From the past, cumulative studies have shown that p53 found on the array. The p53 performs its tumor suppressor mutations are the most common genetic alterations in advanced function by regulating transcription (3). In order for it to act, ovarian cancer, implicating that p53 mutations play important translocation of p53 from the cytoplasm into the nucleus is roles in human ovarian cancer progression. The fact that the required, and this transportation is facilitated by cytoplasmic expression levels of these genes were altered in p53Ala135Val/wt dynein along microtubule tracks (31). Mutant p53 protein in tumors but not in p53wt/wt tumors suggests the possibility that p53Ala135Val/wt mice is increased dramatically. We can speculate these genes have a role in the more aggressive behavior of that the nuclear transportation of mutant p53 is facilitated by p53Ala135Val/wt ovarian tumors. One of the overexpressed genes the elevated dynein light intermediate chain 1. is microtubule-associated cytoplasmic dynein light intermediate It is a very well established fact that p53 mutation is a chain 1. Microtubules play a central role in coordinating various common event in human tumors. Approximately 50% of cellular functions, including cell shape, motility, signal trans- ovarian adenocarcinomas (1, 5) have mutants in the p53 gene. duction and intracellular organelle transport during interphase, We can safely predict that the remaining f50% of the tumors and the formation of mitotic spindles and the cytokinesis bridge must retain the wild-type p53 gene. The question that needs to during cell division. Evidence also indicates that microtubule- be answered is why the wild-type p53 retained in these tumors associated molecules may have a role in cancer cell proli- is unable to perform its function to suppress tumor formation. feration and cell survival. In general, dynein is implicated in Of course, p53 is not the only tumor suppressor gene to have

been identified, and other tumor suppressors might well be tumors containing a p53 mutation. Although the general idea involved in the development of cancers whose p53 remains that p53val135 would enhance carcinogen-induced cancer is itself unaltered. Previous studies have shown that p53 is mutated in not new, the observation that direct application of a chemical ovarian malignancies with f50% in stage III/IV, f15% in carcinogen to ovaries consistently induced non–granulosa cell– stage I/II ovarian adenocarcinomas (1, 5), and in f50% of related ovarian carcinomas in mice as well as the finding of ovarian sarcomas (7). So how do the remaining ovarian tumors synergistic interactions of the carcinogen and p53 in ovarian tolerate wild-type p53 and overcome the tumor suppression that tumor progression are new and have not been reported wild-type p53 provides? Activation of oncogenes and/or previously. Although this may be expected, the end results are inactivation of other tumor suppressor genes are ways to by no means certain because we have seen that the presence of bypass the requirement for p53 mutation. In Table 2, there is a p53val135 in several organ sites failed to enhance carcinogen- set of ‘‘p53-bypassing’’ genes as their expression levels were induced tumorigenesis.5 We believe that we have succeeded only altered in p53wt/wt tumors, not in p53Ala135Val/wt tumors. in generating a potential mouse model for the development Altered expression of these genes could contribute to ovarian of ovarian carcinogenesis in mice. tumorigenesis in p53wt/wt mice. These genes primarily fall into Finally, we begin to address the mechanism of ovarian three functional groups, overexpressed oncogene/tumor sup- carcinogenesis by conducting a series of microarray experi- pressor genes (Rin1, Src, Brca1, and Ect2), overexpressed cell ments. We have presented altered gene expression profiles that cycle regulatory genes (Cdc6, Bub1, Bub1b, Ccnb2, and showed the gene expression differences between normal Nusap1), and underexpressed signal transduction regulatory ovarian tissues from p53Ala135Val/wt mice and those from genes (Rgs2, Stard5, Gpr56, Rab3d, Rerg, Mark1, Rab25, and p53wt/wt mice (Tables 1 and 2). Furthermore, we showed the Frzb). Although Brca1 is defined as a tumor suppressor gene, it gene expression differences of ovarian tumors versus normal is overexpressed in DMBA-induced mouse ovarian tumors. ovarian tissues, dependent or independent of p53 mutant status This overexpressed Brca1 could be a mutant form. Interest- (Table 2), as well as the gene expression differences between ingly, a member of the p53 family, Trp63, is significantly ovarian tumors of p53Ala135Val/wt mice and those from p53wt/wt down-regulated in the p53wt/wt tumors. This down-regulation mice (Table 3). We also attempted to make a biological con- of Trp63 was not observed in p53Ala135Val/wt tumors. On the nection in terms of the genes altered in this model system other hand, there are some genes which are only altered in compared with the genes that are typically known to be p53Ala135Val/wt tumors, indicating that these genes function involved in human ovarian cancer. Specifically, we compared independently of p53. One set of these genes is signal gene expression profiles from the National Center for transduction–related, including the genes Gpr85, Arhgdib, Biotechnology Information data set GSE6008 with 99 human Rad51ap1, Gpr155, and Gnai1. The second set of genes, which individual ovarian tumors (37 endometrioid, 41 serous, 13 included Perp and Bcl 2110, are apoptosis-related. mucinous, and 8 clear cell carcinomas) and 4 individual normal Another common genetic alteration in ovarian cancer are ovary samples with our mouse array data. We found 42 genes point mutations in the K-ras2 gene. Mutated K-ras2 product is in human ovarian tumors that had similar changes seen in involved in the control of growth-stimulatory and cell death mouse ovarian tumors (Table 2). The p53 is a central regulator pathways. Previous studies have shown that activation of of cell growth, DNA damage repair, and apoptosis (3). The p53 K-ras2 by a point mutation in codon 12/13 was frequently directly activates the expression of a substantial number of detected in ovarian tumors (32), with a mutation frequency of genes important for cell cycle regulation and apoptosis (4). 41% to 60%, suggesting the involvement of K-ras2 in this type These two pathways, apoptosis and cell cycle, are heavily of tumor. DMBA-induced mammary tumors in rodents harbored involved in DMBA-induced mouse ovarian tumorigenesis, mutations. The majority of the mutations were purine (AG) to regardless of p53 mutational status. These two mechanisms pyrimidine (CT) transversions (12/19) on the nontranscribed also branch out to the Wnt pathway. We believe that the results strand, which is likely to be due to depurination created by represent a starting point for follow-up in-depth molecular DMBA adduct formation on the nontranscribed strand (33). In analyses. Some of the genes are currently further investigated in the DMBA-induced ovarian tumors, base substitutions at A or G our laboratory. In light of the recent emphasis on the problems accounted for 95% of all the point mutations. The predominance of ovarian cancer in humans, we believe that some aspects of of purine (A or G) mutations is consistent with the fact that this article, especially those related to the roles of both chemical DMBA adduct formation preferentially occurs on dA and dG, carcinogen and the p53 mutation in the ovarian tumorigenic leading to depurination (34). In our study, we analyzed point process, are novel and important. mutations in K-ras2 and found that mutation occurred in codons 13 and 61 with 40% (4 of 10) and 20% (2 of 10) of mutation frequency in ovarian sarcomas of p53wt/wt and p53Ala135Val/wt Materials and Methods mice, respectively (Supplementary Table S1). The pattern of Animal Bioassays mutations are demonstrating the important role of K-ras2 during To conduct the animal bioassays, 6-week-old female mice on DMBA-induced ovarian carcinogenesis, and it is independent of A/J background were grouped based on their p53 genotype, histopathologic types. p53Ala135Val/wt and their wild-type littermates (p53wt/wt). These In this study, we have successfully used the classical carcinogenesis approach of ovarian tumor induction (35) in p53 transgenic mice. These mice not only have shortened tumor latencies but also closely resemble a subset of human ovarian 5 You et al. unpublished results.

mice were produced by backcrossing UL53-3 mice to A/J strain Z-transform. These values were input to the GeneCluster mice for more than 10 times. Based on an early publication program of Eisen et al. (37) and genes were clustered using (35), DMBA was heated to its melting point (f124jC) in a average linkage and correlation dissimilarities. Pathways of chemical fume hood. Sterile 6-0 silks (2 cm in length beyond interest, such as cell cycle, apoptosis, mitogen-activated protein the attached needle) were immersed in the melted DMBA or in kinase, and G protein, were evaluated for differential regulation DMSO (as control), and dried before use. Surgical implantation using the visualization tool GenMAPP at the University of on the ovary was done on 6-week-old mice. Anesthesia was California, San Francisco (38).6 We imported our results into induced with ketamine (90 mg/kg)/acepromazine (2 mg/kg) the program and used GenMAPP to illustrate the pathways intraperitoneally. Supplemental buprenex (0.01 mg/kg) was containing differentially expressed genes. Differential gene given subcutaneously to every mouse before surgery. The lower expression was based on p53 genotype status and ovarian back was shaved and swabbed with Nolvasan followed by 70% tumor phenotype (P V 0.05, z2.0-fold change). A total of ethanol. A 2 cm dorsal incision was made to open the retro- 20 samples (5 samples from each group, normal and tumor peritoneal cavity and to expose the left ovary. Using a sterile from both the p53wt/wt and p53Ala135Val/wt) were examined with needle, a 1-cm length of 6-0 silk, impregnated with either microarray analysis. DMBA or DMSO, was implanted into the left ovary and secured with one knot. The dorsum was closed in two layers Real-time PCR with sutures and wound clips. All animals were observed daily Selected genes were validated by real-time PCR using for clinical signs of illness for the first 2 weeks postsurgery and qPCR analysis (Bio-Rad Laboratories) on the Bio-Rad iCycler twice a week thereafter. Three months after surgery, 50% of iQ Real Time PCR Instrument. Mouse glyceraldehyde-3- the mice from the p53Ala135Val/wt group and a few mice from the phosphate dehydrogenase was used as the reference control p53wt/wt group showed obvious tumors which could be easily for normalization of template. RNAs from four individual palpated through the enlarged abdominal or dorsal wall. Mice, ovarian tumors of p53wt/wt and p53Ala135Val/wt were tested with either moribund or with dramatically enlarged abdomens, were four individual normal ovary RNAs as normal controls. Each euthanized by CO asphyxiation. Gross necropsy was done: all 2 sample was run in duplicate for both tested genes and for mouse organs including the ovary, uterus, colon, intestine, lung, liver, glyceraldehyde-3-phosphate dehydrogenase. The amplifications and kidney were examined for tumors. During the tumor were carried out by a hot-start at 95jC for 2 min, followed dissection, first of all, the implanted suture was located. Then, j j the surrounding tissue was collected, a representative portion of by 40 cycles of 95 C for 15 s and 58 C for 45 s. The primers ¶ ¶ the ovarian tumors was snap-frozen in liquid nitrogen for (5 to 3 ) used for real-time PCR confirmation on gene molecular biological study and part was fixed in buffered expression in ovarian tumors are available upon request. formalin for histopathologic examination. The right side normal ovary was also removed. Tissue sections (4 Am) were stained Detection of Point Mutations in the K-ras2 Gene with H&E for histopathologic evaluation according to existing The RASmitogen-activated protein kinase signal transduc- human and mouse ovarian cancer classifications (36). Fishers tion pathway has been shown to interact with the ARF- exact test was used to determine the difference in incidence p53 pathway in the processes of cell transformation and between p53Ala135Val/wt (n = 40) and p53wt/wt mice (n = 35). carcinogenesis. Activation of K-ras2 by point mutation in codon 12 or 13 was frequently detected in ovarian mucinous Gene Expression Analysis tumors (32). In our study, we did K-ras2 mutation analysis on a Microarray analysis was then done on normal and tumor total of 20 tumors (10 from p53Ala135Val/wt mice and 10 from tissues from ovaries treated with DMBA. Total RNAs from four p53wt/wt mice). Tumors used were ovarian adenocarcinomas paired adenocarcinomas [ovarian tumor (left side) and normal either from p53wt/wt (5 of 10) or p53Ala135Val/wt mice (4 of 20), ovary (right side)], were isolated by Trizol (Invitrogen) and respectively. The rest of the tumors (5 of 10 from p53wt/wt mice; purified with RNeasy Mini Kit and RNase-free DNase Set 15 of 20 from p53Ala135Val/wt mice) were adenocarcinomas (QIAGEN) according to the manufacturer’s protocols. In vitro mixed with sarcoma or sarcomas. PCR amplification and direct transcription-based RNA amplification was then done and DNA sequencing analysis were used to identify the muta- cDNA was synthesized using a Superscript cDNA Synthesis tions in the K-ras2 gene. Briefly, DNA was isolated from Kit (Invitrogen). The cDNA was cleaned using phase-lock ovarian tumors using TRIzol reagent (Invitrogen). Sequences of gels (Fisher Scientific). Biotin-labeled cRNA was transcribed PCR primers for K-ras2 exons 1 (contains codons 12 and 13) in vitro from cDNA using a BioArray HighYield RNA Tran- and 2 (contains codon 61) were described previously (39). script Labeling Kit (ENZO Biochem). The labeled cRNA was A50AL reaction mixture containing 100 ng of genomic applied to the Affymetrix Moe430Av2 GeneChips (Affymetrix) DNA, 10 mmol/L of Tris-HCl (pH 8.5), 50 mmol/L of KCl, according to the manufacturer’s recommendations. Array nor- 2.5 mmol/L of MgCl2, 100 Amol/L of each deoxyribonucleo- malization and gene expression estimates were obtained using side triphosphate (dATP, dCTP, dGTP, dTTP), 1.0 unit of Affymetrix Microarray Suite 5.0 software. Differential expres- Taq DNA polymerase (Promega), and 40 pmol of each sion was determined on the combined basis of statistical testing primer were overlaid with sterile mineral oil and subjected using t tests and fold changes. A cutoff of P V 0.05 and a fold change of z2 is called positive for differential expression. For the selected genes, expression indices were transformed across samples to an N (0, 1) distribution using a standard statistical 6 http://www.genmapp.org/

to 35 cycles of PCR amplification. Each cycle consisted of 20. Nair P, Muthukkumar S, Sells SF, Han SS, Sukhatme VP, Rangnekar VM. Early growth response-1-dependent apoptosis is mediated by p53. J Biol Chem 1 min each at 94jC, 52jC, and 72jC. The mutations were 1997;272:20131 – 8. analyzed with an ABI PRISM 3700 DNA analyzer (PE Applied 21. Abdulkadir SA, Carbone JM, Naughton CK, Humphrey PA, Catalona WJ, Biosystems). Milbrandt J. Frequent and early loss of the EGR1 corepressor NAB2 in human prostate carcinoma. Hum Pathol 2001;32:935 – 9. 22. Eid MA, Kumar MV, Iczkowski KA, Bostwick DG, Tindall DJ. Expression Acknowledgments of early growth response genes in human prostate cancer. Cancer Res 1998;58: We thank Dr. Daolong Wang for Fisher exact test on the bioassay data. 2461 – 8. 23. Scharnhorst V, Menke AL, Attema J, et al. EGR-1 enhances tumor growth References and modulates the effect of the Wilms’ tumor 1 gene products on tumorigenicity. 1. Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2006. CA Cancer J Clin Oncogene 2000;19:791 – 800. 2006;56:106 – 30. 24. Fahmy RG, Khachigian LM. Antisense Egr-1 RNA driven by the CMV 2. Greenblatt MS, Bennett WP, Hollstein M, Harris CC. Mutations in the p53 promoter is an inhibitor of vascular smooth muscle cell proliferation and regrowth tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. after injury. J Cell Biochem 2002;84:575 – 82. Cancer Res 1994;54:4855 – 78. 25. Krones-Herzig A, Adamson E, Mercola D. Early growth response 1 protein, 3. Vogelstein B, Lane D, Levine AJ. Surfing the p53 network. Nature 2000;408: an upstream gatekeeper of the p53 tumor suppressor, controls replicative 307 – 10. senescence. Proc Natl Acad Sci U S A 2003;100:3233 – 8. 4. el-Deiry WS, Tokino T, Velculescu VE, et al. WAF1, a potential mediator of 26. Lane J, Allan V. Microtubule-based membrane movement. Biochim Biophys p53 tumor suppression. Cell 1993;75:817 – 25. Acta 1998;1376:27 – 55. 5. Aunoble B, Sanches R, Didier E, Bignon YJ. Major oncogenes and tumor 27. Suelmann R, Fischer R. Nuclear migration in fungi—different motors at suppressor genes involved in epithelial ovarian cancer (review). Int J Oncol 2000; work. Res Microbiol 2000;151:247 – 54. 16:567 – 76. 28. Puthalakath H, Huang DC, O’Reilly LA, King SM, Strasser A. The 6. Liu FS, Kohler MF, Marks JR, Bast RC, Jr., Boyd J, Berchuck A. Mutation proapoptotic activity of the Bcl-2 family member Bim is regulated by interaction and overexpression of the p53 tumor suppressor gene frequently occurs in uterine with the dynein motor complex. Mol Cell 1999;3:287 – 96. and ovarian sarcomas. Obstet Gynecol 1994;83:118 – 24. 29. Vadlamudi RK, Bagheri-Yarmand R, Yang Z, et al. Dynein light chain 1, a 7. Zhang Z, Liu Q, Lantry LE, et al. A germ-line p53 mutation accelerates p21-activated kinase 1-interacting substrate, promotes cancerous phenotypes pulmonary tumorigenesis: p53-independent efficacy of chemopreventive agents [Erratum in: Cancer Cell. 2004;6:101]. Cancer Cell 2004;5:575 – 85. green tea or dexamethasone/myo-inositol and chemotherapeutic agents taxol or 30. Vadlamudi RK, Adam L, Wang RA, et al. Regulatable expression of adriamycin. Cancer Res 2000;60:901 – 7. p21-activated kinase-1 promotes anchorage-independent growth and abnormal 8. Zhang Z, Yao R, Li J, et al. Induction of invasive mouse skin carcinomas in organization of mitotic spindles in human epithelial breast cancer cells. J Biol transgenic mice with mutations in both H-ras and p53. Mol Cancer Res 2005;3: Chem 2000;275:36238 – 44. 563 – 74. 31. Giannakakou P, Sackett DL, Ward Y, Webster KR, Blagosklonny MV, Fojo 9. Orsulic S, Li Y, Soslow RA, Vitale-Cross LA, Gutkind JS, Varmus HE. T. p53 is associated with cellular microtubules and is transported to the nucleus by Induction of ovarian cancer by defined multiple genetic changes in a mouse dynein. Nat Cell Biol 2000;2:709 – 17. model system. Cancer Cell 2002;1:53 – 62. 32. Enomoto T, Inoue M, Perantoni AO, Terakawa N, Tanizawa O, Rice JM. 10. Flesken-Nikitin A, Choi KC, Eng JP, Shmidt EN, Nikitin AY. Induction of K-ras activation in neoplasms of the human female reproductive tract. Cancer Res carcinogenesis by concurrent inactivation of p53 and Rb1 in the mouse ovarian 1990;50:6139 – 45. surface epithelium. Cancer Res 2003;63:3459 – 63. 33. Yamamoto K, Nakata D, Tada M, et al. A functional and quantitative 11. Connolly DC, Bao R, Nikitin AY, et al. Female mice chimeric for expression mutational analysis of p53 mutations in yeast indicates strand biases and different of the simian virus 40 TAg under control of the MISIIR promoter develop roles of mutations in DMBA- and BBN-induced tumors in rats [Erratum in: Int J epithelial ovarian cancer. Cancer Res 2003;63:1389 – 97. Cancer 2000;85:898]. Int J Cancer 1999;83:700 – 5. 12. Ullrich RL. Tumor induction in BALB/c female mice after fission neutron or 34. Manjanatha MG, Chen JB, Shaddock JG, Jr., Harris AJ, Shelton SD, gamma irradiation. Radiat Res 1983;93:506 – 15. Casciano DA. Molecular analysis of lacI mutations in Rat2 cells exposed to 7,12- 13. Chen CM, Chang JL, Behringer RR. Tumor formation in p53 mutant ovaries dimethylbenz[a]anthracene: evidence for DNA sequence and DNA strand biases transplanted into wild-type female hosts. Oncogene 2004;23:7722 – 5. for mutation. Mutat Res 1996;372:53 – 64. 14. Crist KA, Zhang Z, You M, et al. Characterization of rat ovarian 35. Jacobs AJ, Curtis GL, Newland JR, Wilson RB, Ryan WL. Chemical adenocarcinomas developed in response to direct instillation of 7,12-dimethyl- induction of ovarian epithelial carcinoma in mice. Gynecol Oncol 1984;18: benz[a]anthracene (DMBA) coated suture. Carcinogenesis 2005;26:951 – 7. 177 – 80. 15. Harfe BD, Jinks-Robertson S. DNA mismatch repair and genetic instability. 36. Davis B, Harleman JH, Heinrichs M, et al. Female genital system. In: Mohr Annu Rev Genet 2000;34:359 – 99. U, editor. International classification of rodent tumors. The Mouse. Springer 16. Modrich P, Lahue R. Mismatch repair in replication fidelity, genetic (Berlin); 2001. p. 211 – 68. recombination, and cancer biology. Annu Rev Biochem 1996;65:101 – 33. 37. Eisen MB, Spellman PT, Brown PO, Botstein D. Cluster analysis and dis- 17. Milbrandt J. A nerve growth factor-induced gene encodes a possible play of genome-wide expression patterns. Proc Natl Acad Sci U S A 1998;95: transcriptional regulatory factor. Science 1987;238:797 – 9. 14863 – 8. 18. Sukhatme VP, Cao XM, Chang LC, et al. A zinc finger-encoding gene 38. Dahlquist KD, Salomonis N, Vranizan K, Lawlor SC, Conklin BR. coregulated with c-fos during growth and differentiation, and after cellular GenMAPP, a new tool for viewing and analyzing microarray data on biological depolarization. Cell 1988;53:37 – 43. pathways. Nat Genet 2002;31:19 – 20. 19. Huang RP, Fan Y, de Belle I, et al. Decreased Egr-1 expression in human, 39. Wang Y, Zhang Z, Yan Y, et al. A chemically induced model for squamous mouse and rat mammary cells and tissues correlates with tumor formation. Int J cell carcinoma of the lung in mice: histopathology and strain susceptibility. Cancer 1997;72:102 – 9. Cancer Res 2004;64:1647 – 54.

Mol Cancer Res 2008;6(1). January 2008 Downloaded from mcr.aacrjournals.org on September 29, 2021. © 2008 American Association for Cancer Research. Enhanced Susceptibility to Chemical Induction of Ovarian Tumors in Mice with a Germ Line p53 Mutation

Yian Wang, Zhongqiu Zhang, Yan Lu, et al.

Mol Cancer Res 2008;6:99-109.

Updated version Access the most recent version of this article at: http://mcr.aacrjournals.org/content/6/1/99

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2008/02/20/6.1.99.DC1

Cited articles This article cites 38 articles, 13 of which you can access for free at: http://mcr.aacrjournals.org/content/6/1/99.full#ref-list-1

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

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/6/1/99. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.