J Toxicol Pathol 2005; 18: 121–134

Review Susceptibility of Heterozygous and Nullizygous Knockout Mice to Chemical Carcinogens: Tissue Dependence and Role of p53 Mutations

Tetsuya Tsukamoto1, Akihiro Hirata1, and Masae Tatematsu1

1Division of Oncological Pathology, Aichi Cancer Center Research Institute, 1–1 Kanokoden, Chikusa-ku, Nagoya 464–8681, Japan

Abstract: Mutations of the p53 tumor suppressor gene constitute one of the most frequent molecular changes in a wide variety of human cancers and mice deficient in p53 have recently attracted attention for their potential to identify chemical genotoxins. In this article, we review the data on the susceptibility of p53 nullizygote (–/–), heterozygote (+/ –), and wild type (+/+) mice to various carcinogens. Induction of esophageal and tongue squamous cell carcinomas (SCCs) by methyl-n-amylnitrosamine was shown to be increased in p53 (+/–) mice, in addition to the high sensitivity shown by p53 (–/–) littermates. N,N-dibutylnitrosamine (DBN) treatment also caused more tumor development in p53 (+/–) than wild-type mice, as with N-butyl-N-(4-hydroxybutyl)nitrosamine (BBN) in the urinary bladder. In addition, p53 (+/–) heterozygotes proved more sensitive than wild type littermates to the induction of stromal cell tumors like hemangiomas/hemangiosarcomas by N-bis(2-hydroxypropyl)nitrosamine (BHP) or lymphomas and fibrosarcomas with other carcinogens. Analysis of exons 5–8 of the p53 gene demonstrated mutations in approximately one half of the lesions. With N-methyl-N-nitrosourea, preneoplastic lesions of the stomach, pepsinogen altered pyloric glands (PAPG), and a gastric adenocarcinoma, were found after only 15 weeks in p53 (–/–) mice, although there was no significant difference in the incidence of gastric tumors between p53 (+/+) and (+/–) mice in the longer-term. Regarding colon carcinogenicity, adenocarcinomas were observed limited to 1, 2-dimethylhydrazine treated p53 (–/–) mice in the short term, but again, no significant difference was evident between the p53 (+/+) and (+/–) cases at the end of the study. Furthermore, diethylnitrosamine or aminophenylnorharman treated p53 (+/–) mice did not demonstrate elevated susceptibility to tumors in the liver. With BHP, which induces tumors in multiple organs, p53 (+/–) mice were not more statistically sensitive with regard to lung tumor development than p53 (+/+). Of the malignant tumors examined in p53 (+/+) and (+/–) mice, as many as 10% demonstrated mutations in the p53 gene. These results suggest that p53 may not be a direct target for mouse adenomas/adenocarcinomas, but rather plays an important role as a gatekeeper in their genesis. In contrast p53 itself is frequently mutated in squamous, urothelial, or stromal tumors with a clear order of susceptibility: p53 (–/–) > p53 (+/–) > p53 (+/+) mice. p53 (–/–) mice are versatile animals for carcinogenicity testing, despite their disadvantage of a high background of spontaneous tumor development, and tissue dependence must be taken into account when exposing p53 (+/–) mice to chemical carcinogens. (J Toxicol Pathol 2005; 18: 121–134) Key words: p53 knockout mouse, susceptibility, tissue dependence

Introduction development of a variety of tumors2. As an animal model for this disorder, mice deficient in p53 alleles have been Mutations of the p53 gene constitute one of the most established3–6. These animals show normal development, frequent molecular changes been in a wide variety of human but homozygote knockouts (–/–) show extreme sensitivity to cancers1. In addition to somatic mutations in sporadic chemical carcinogens, almost irrespective of the organ3. cancers, germline mutations in the p53 gene are associated However, the frequent spontaneous development of tumors with the Li-Fraumeni syndrome, a familial autosomal in p53 knockout mice, including thymic lymphomas and dominant disease characterized by a predisposition to sarcomas at young adult age, hinders their application for risk assessment purposes. p53 (+/–) mice, on the other hand, have only a low incidence of background spontaneous Received: 1 June 2005, Accepted: 3 August 2005 tumors until late in life7,8 and thus have attracted interest as Mailing address: Tetsuya Tsukamoto, Division of Oncological experimental animals in tests for the carcinogenic potential Pathology, Aichi Cancer Center Research Institute, 1–1 Kanokoden, of chemicals. The same mouse model has also been widely Chikusa, Nagoya 464–8681, Japan utilized for functional analysis of the p53 gene in TEL: 81-52-762-6111 (ext. 7062) FAX: 81-52-764-2972 carcinogenesis studies. There have been several reports that E-mail: [email protected] 122 Susceptibility of the p53 Knockout Mice p53 (+/–) mice are highly sensitive to genotoxic carcinogens, damage checkpoints employ damage sensor , such as such as N-butyl-N-(4-hydroxybutyl)nitrosamine (BBN) ataxia telangiectasia mutated (ATM), ATM- and Rad3- targeting the urinary bladder9, N-ethyl-N-nitrosourea (ENU) related (ATR), the Rad17-replication factor C (Rad17-RFC) inducing uterine endometrial stromal sarcomas10, N,N- complex, and the toroidal Rad9-Rad1-Hus1 checkpoint dibutylnitrosamine (DBN) causing esophagus and urinary complex (9-1-1 complex), to detect DNA damage and to bladder tumors11, methyl-n-amylnitrosamine (MNAN) initiate signal transduction cascades involving checkpoint active in the esophagus12 and tongue13 and urethane inducing kinase (Chk) 1 and Chk2 Ser/Thr kinases and Cdc25 vascular tumors14. However, p53 (+/–) mice have not been phosphatases. The signal transducers activate p53 and shown to have increased susceptibility to induction of inactivate cyclin-dependent kinases to inhibit cell cycle hepatocellular tumors by diethylnitrosamine (DEN)15 or progression33. The ATM gene, when mutated in the genetic breast tumors by 7,12-dimethyl[a]benzanthracene instability syndrome ataxia telangiectasia, is linked to (DMBA)16. We have also shown only a minimal increment increased cancer risk. ATM and its related kinase, ATR, of sensitivity in p53 (+/–) mice to stomach tumors induced phosphorylate a number of down stream proteins, including by N-methyl-N-nitrosourea (MNU)17, colon tumors after p53, to effect stabilization34. Then p53 in turn plays a critical treatment with 1,2-dimethylhydrazine (DMH)18, and liver role in maintaining genome integrity by activating a tumors induced with a heterocyclic amine, biochemical chain reaction that leads to cell cycle aminophenylnorharman (APNH)19. Furthermore, utilization checkpoint activation from G1 to S (the G1/S checkpoint)35 of N-bis(2-hydroxypropyl)nitrosamine (BHP) for the or from G2 to mitosis (the G2/M checkpoint)36. In the G1/S induction of tumors in multiple organs, pointed to tissue checkpoint case, cyclin E-Cyclin-dependent kinase 2 (Cdk2) specificity regarding susceptibility (Hirata et al., manuscript is inactivated by a p53-inducible Cdk inhibitor, p21CIP1/WAF1 submitted). In the present report, we review our data and 37. The G2/M checkpoint involves inhibition of Cdc2, the discuss the mechanisms of the sensitivity of p53 knockout cyclin-dependent kinase required to enter mitosis. Cdc2 is mice. inhibited simultaneously by three transcriptional targets of p53, Gadd45, p21, and 14-3-3σ. Binding of Cdc2 to Cyclin p53 as a Tumor Suppressor B1 is required for its activity, and repression of the cyclin B1 gene by p53 also contributes to blocking entry into mitosis36. Discovery of the p53 Cdc25A phosphatase38 and Cdk2 kinase39 appear central in p53 is a tumor suppressor gene which is frequently the transient intra-S-phase response for repair of DNA altered in human malignancies20. Located on the short arm damage during DNA replication (the intra-S checkpoint). of human 17p1321, where loss of Cells try to repair their DNA during checkpoint delay. heterozygosity (LOH) is frequently observed in human However, if the repair processes fail, p53 stimulates multiple cancers22, the gene yields a 2.8 kb mRNA transcript and apoptotic mechanisms including both nuclear and encodes a 53 kD nuclear phosphoprotein composed of 393 mitochondrial pathways40. p53 can directly induce amino acids23. The p53 protein was initially identified in permeabilization of the outer mitochondrial membrane by SV40-transformed cells, where it was thought to be a forming complexes with the protective BclXL and Bcl2 transformation-specific protein or tumor antigen24. proteins, resulting in cytochrome c release with binding to Transfection assays in NIH3T3 fibroblast cells initially BclXL via its DNA binding domain, in the intrinsic suggested p53 to be an oncogene25, but subsequently it was pathway41. p53 also functions to activate the pro-apoptotic demonstrated that only mutant forms of p53 possess the proteins BAX and BAK on the mitochondrial membranes capacity to immortalize cells and the wild-type protein through direct physical interaction42. Besides these actually suppresses transformation26. Furthermore, a tumor mitochondrial pathways, p53 also induces the expression of suppressor role was suggested by the fact that LOH and point proteins in extrinsic death- pathways mediated by mutations in the remaining allele were frequently identified elements such as FAS and DR5/KILLER43. p53 induces Fas in colorectal27 and lung28 cancers. There is now mRNA expression by binding to its promoter region. DR5 is overwhelming evidence that the normal function of p53 is in also induced by p53 in response to DNA damage and in turn fact that of a tumor suppressor29. promotes cell death through caspase-844. p53 stimulates expression of caspase-645, which cleaves the nuclear Responses of p53 to DNA damaging agents envelope protein lamin A and several transcription factors46. Damage to DNA induces several cellular responses that The pro-apoptotic Bid connects activation of the extrinsic enable the cell either to eliminate or cope with the damage or death receptor pathway to activation of the mitochondrial to activate a programmed cell death process, so that disruption processes associated with the intrinsic pathway. potentially catastrophic mutations are not passed on to the p53 therefore appears to promote the convergence of next generation. There are many DNA damaging agents, intrinsic and extrinsic pathways through Bid regulation47. including ionizing irradiation30, genotoxic chemical carcinogens31, and oxidative radicals32, which can cause Development of p53 Knockout Mice genotoxic insult, including DNA strand breaks, and stimulate activation of checkpoint proteins. The DNA To generate animals featuring loss of p53, homologous Tsukamoto, Hirata, Tatematsu 123 recombination has been utilized with mouse embryonic stem with 5 ppm MNAN, the incidence of squamous cell (ES) cells to make null alleles. Two models have been carcinomas (SCCs) was significantly higher (P<0.001) in developed by replacement of exons 2 through 6 with a p53 (–/–) than in p53 (+/–) and (+/+) mice at 15 weeks after neomycin resistant cassette (neoR)5,6. Tsukada et al. starting treatment. At 25 weeks, the SCC incidence in p53 introduced neoR into the second exon to disrupt the (+/–) was significantly greater (P<0.05) than in p53 (+/+) translation initiation codon generating a null mutation of the mice. Only one SCC developed, in a p53 (+/+) mouse at 25 p53 gene using ES cells derived from an F1 embryo with a weeks. The total number of tumors per mouse (mean ± SD) C57BL/6 × CBA background4. Donehower and co-workers was also higher in the p53 (–/–) (4.8 ± 1.4 /mouse) than in the inserted a PolII promoter-driven neoR into exon 5, and also p53 (+/–) (1.5 ± 1.2) and p53 (+/+) (0.8 ± 0.9) groups at 15 deleted 350 nucleotides of intron 4 and 106 nucleotides of weeks (P<0.001) and it was greater in p53 (+/–) (4.4 ± 2.6) exon 5 in 129Sv derived ES cells3. p53 knockout (–/–) mice than in p53 (+/+) (2.7 ± 1.5) mice at 25 weeks (P<0.05). In develop normally but show extreme sensitivity to chemical the mice receiving 15 ppm MNAN, there was a trend of carcinogens, mostly irrespective of the organ3. The problem higher SCC incidence in p53 (+/–) as compared to p53 (+/+) is that they frequently develop spontaneous tumors, mice at 15 and 25 weeks. PCR-SSCP and sequencing including thymic lymphomas or sarcomas at young adult analyses for exons 5–8 of the p53 gene were performed and age, hindering their practical use for test purposes. In mutations were identified in 6 out of 12 esophageal SCCs contrast, p53 (+/–) mice show relatively late spontaneous (50%) and 14 out of 23 SCCs (61%) in p53 (+/+) and p53 (+/ tumor development with lower incidences than nullizygous –) mice, respectively. Only one of 19 papillomas examined p53 (–/–) knockout mice7,8 and thus have attracted interest as had a mutation. LA-PCR-amplified DNAs from frozen experimental animals for assessing the carcinogenic samples exhibited missense mutations not in the mutant but potential of chemicals, as well as for the functional analysis rather in the wild type allele, indicating loss of functional of the roles of p53 in carcinogenesis. In fact, the p53 protein (Fig. 1 and Table 1). Alternatives to Carcinogenicity Testing (ACT) project of the Nishikawa et al.11 administered 0.025% or 0.05% N,N- International Life Sciences Institute (ILSI) has already dibutylnitrosamine (DBN) in drinking water to p53 (+/–) shown positive results48,49. mice as well as to the wild type controls for 20 weeks. This For our experiments, p53 knockout mice produced by chemical targets different organs, including the esophagus in Donehower et al.3 were backcrossed to C57BL/6J females to which papillary hyperplasia, papillomas, and SCCs were obtain N4 mice, which are maintained at the Animal Facility induced. The incidence of papilloma was significantly of Aichi Cancer Center Research Institute in a collaborative higher in p53 (+/–) than in p53 (+/+) mice (P<0.05) in the effort. The genetic background of our p53 knockout mouse 0.05% dose group. Furthermore, p53 (+/–) mice had more is estimated to be more than 15/16 C57BL/6J and less than 1/ SCCs than their wild type counterparts with both doses 16 129Sv. To analyze susceptibility, three different (P<0.05). Analysis of the p53 gene status in esophageal experimental periods were basically selected to examine (1) mucosa in DBN treated animals revealed mutations at high preneoplastic lesions in the short term, (2) differences in frequencies in both p53 (+/+) and p53 (+/–) mice without tumor yield between p53 heterozygous (+/–) and nullizygous any differences between the p53 genotypes (Table 1). (–/–) mice using small amounts of carcinogens, in the mid- term, usually 15 weeks, and (3) lastly to show differences Tongue carcinogenesis between p53 wild-type (+/+) and heterozygous (+/–) mice MNAN induces papillomas and SCCs of the tongue as using standard carcinogen doses in longer term experiments well as the esophagus13. At week 15, although no for up to one year. microscopic changes were observed in either p53 (+/+) or (+/ –) mice, SCCs were found in 5/12 (41.7%) p53 (–/–) mice, Heterozygous p53 Knockout Mice are Susceptible the incidence being significant (P<0.05 and P<0.01 vs. p53 to Carcinogens (+/+) and p53 (+/–), respectively). Additionally, one of five p53 (–/–) mice exhibited SCC development and another had Esophageal carcinogenesis a papilloma. At week 25, carcinoma in situ (CIS) was To analyze esophageal squamous cell carcinogenesis12, observed in one each of the p53 (+/+) and (+/–) mice, drinking water containing 5 or 15 ppm of MNAN was together with a papilloma in the latter animal. The rare provided to p53 (+/+), (+/–), and (–/–) mice for 8 weeks. occurrence of tongue neoplasms in wild type and The animals were then maintained without further treatment heterozygous p53 knockout mice in that study might have for an additional 7 or 17 weeks. Sacrifices of mice of the been a reflection of murine resistance to tongue three genotypes took place at week 15 and at week 25 p53 carcinogenesis by MNAN as compared to 4-nitroquinoline- (+/+) and (+/–) mice were sacrificed. Three groups of p53 1-oxide (4NQO)50. p53 mutations were identified in the CIS (+/+), (+/–), or (–/–) control animals received in the p53 (+/+) mouse, and in the papilloma of the p53 (+/– unsupplemented drinking water. Administration of MNAN ) mouse (Table 1). induced a 100% incidence of esophageal diffuse hyperplasia in p53 (–/–), (+/–), and (+/+) mice, often accompanied by Carcinogenesis in the urinary bladder subepithelial inflammatory infiltration. In the mice treated Ozaki et al.9 used BBN to induce tumors in the urinary 124 Susceptibility of the p53 Knockout Mice

Fig. 1. Esophageal tumors in p53 knockout mice. (A) Papilloma induced in an MNAN treated p53 (+/–) mouse. (B) Squamous cell carcinoma in an MNAN treated p53 (+/–) mouse. (A and B) H&E staining. Original magnification: (A) 100×; (B) 200×. bladder. p53 (+/+) and p53 (+/–) mice were administered 0.004, 0.0075, or 0.025% BBN in drinking water for 20 weeks, resulting in development of dysplasia, transitional cell carcinoma (TCCs or urothelial carcinomas) and SCCs. In the 0.0075% treated groups, the incidence of dysplasia was significantly higher in p53 (+/–) than in wild type (P<0.05) mice. The frequency of p53 mutations in dysplasia and TCC was about 20% for both genotypes. In contrast, 50% (4/8) of SCCs had mutations in p53 (+/–) mice. The p53 (+/+) mice did not demonstrate any SCCs (Table 1). In Fig. 2. Malignant stromal tumors in p53 knockout mice. (A) the same experiment, hemangiomas and hemangiosarcomas Hemangiosarcoma in a BHP treated p53 (–/–) mouse. (B) A were also observed in the urinary bladder. The incidence of sarcoma developing in a BHP treated p53 (–/–) mouse. (C) these stromal tumors in p53 (+/–) was significantly higher Thymic lymphoma spontaneously developed in a p53 (–/–) than in their wild type counterparts (P<0.005) (Table 1). mouse. (A–C) H&E staining. Original magnification: (A and B) 200×; (C) 400×. In another study, DBN induced neoplastic lesions in the urinary bladder as well as in the esophagus11 and the incidence of TCCs was significantly higher in p53 (+/–) than in p53 (+/+) mice, with both 0.025% and 0.05% doses Tsukamoto, Hirata, Tatematsu 125 11 11 11 10 10 9 9 12 13 , , 2003 , , 2003 , 2003 , , 2000 , , 2000 , , 1998 , 1998 , 2002 , , 2002 , et al. et al. et al. et al. et al. et et al. et et al. et al. et al. et al. Hirata Shirai, Shirai, manuscript submitted Mitsumori, Mitsumori, -nitrosourea. -nitrosourea. N Esophageal 43.8% (7/16) 43.8% p53 mutations -ethyl- N Sex Frequency of References Male ND Nishikawa, Male (1/19) 5.3% Male mucosa: Nishikawa, Male 50% (1/2) Male ND Nishikawa, Male (20/35) 57% Male (1/1) 100% D Female ND Mitsumori, in each p53 genotype in each p53 -bis(2-hydroxypropyl)nitrosamine; ENU, N (+/+) (+/–) (–/–) Incidence of tumors Incidence of 21% (3/14)21% (6/14) 43% ND 0% (0/14) 36% (5/14)† ND 90.9% (10/11)90.9% (13/14) 92.9% ND 0% (0/13) 0% (0/15) (5/12)* 41.7% 14% (2/14)14% (8/14)† 57% ND 9.1% (1/11)9.1% (2/14) 14.3% ND (9/14)64% (14/14)† 100% 0% (0/13) ND 0% (0/15) (2/12) 16.7% 7.1% (2/28)7.1% (15/37)† 40.5% ND Male ND Ozaki, a 0% (0/19) 31% (5/16)† N ndometrial (5/19) 26% (10/16)† 63% ND Female ND SCC SCC TCC ysplasia (5/10) 50.0% % (10/10)† 100 ND (6/28) 21.4% -(4-hydroxybutyl)nitrosamine; BHP, SCC+CIS Papilloma Papilloma Papilloma N hyperplasia Fibrosarcoma -butyl- N nd/or Nullizygous p53 Knockout Mice p53 Knockout nd/or Nullizygous & 0.025% Hemangiosarcoma icantly higher than p53 (+/+). ; TCC, transitional cell carcinoma. ND, not determined. in situ -dibutylnitrosamine; BBN, -dibutylnitrosamine; BBN, periods 25 25 weeks 25 weeks 5 ppm ppm 15 (13/13) 100% (10/10) 100% (15/16) 93.8% (13/14) 92.9% ND ND 20 weeks20 0.05% 25 weeks 5 ppm (3/13) 23% (6/10)† 60% ND (1/13) 7.7% (1/16) 6.3% ND 15 weeks15 ppm 15 15 15 weeks weeks15 5 ppmppm 15 20 weeks 0.025% weeks15 ppm 5 0% (0/13) 20 weeks (1/15) 6.7% 0.0075% weeks20 (10/12)* 83.3% 0.05% weeks20 SCC 15 weeks 0.05% weeks15 ppm 20 weeks40 ppm 200 0% (0/10)ppm 20 weeks15 weeks40 % (2/10) 20.0 ppm 200 Hemangiomappm 20 (5/13) 39% ND Hemangiosarcoma (9/20) 45.0% (7/10)† 70% (16/16)† 100% (5/20) 25.0% (1/14) 7% (16/16)† 100% ND 0% (0/32) ND (3/10) 30% (4/8) 50.0% (7/19) 36.8% ND 0% (0/32)(14/17)† 82.4% Male (31.6%)* 6/19 Male ND (2/19) 10.5% ND ND (12/20) 60.0% (14/17)† 82.4% ND 25 25 weeks weeks25 5 ppmppm 15 weeks20 (1/13) 7.7% 0.0075% (2/10) 20.0% (7/16)† 43.8% (8/14) 57.1% ND ND TCC (2/10) 20.0% 15 weeks(6/10) 60.0% ppm 20 ND Male (7/32) 21.9% Ozaki, 0% (0/32) 0% (0/32) (21.1%)* 4/19 20 weeks20 0.05% (7/13) 54% (10/10)† 100% ND 25 25 weeks 5 ppm 0% (0/13) (1/16) 6.3% ND N,N sceptibility of Heterozygous a Heterozygous of sceptibility DBN 20 weeks 0.025% DBN 20 weeks 0.025% DBN 20 weeks 0.025% 3 (+/+) and (+/–). †: Signif (+/–). 3 (+/+) and Uterus ENU 26 weeks 120 mg/kg BW Endometrialstromal sarcoma 37% (7/19) 94% (15/16)† ND Female 100% (10/10) OthersUterus ENU weeks 26 BW mg/kg 120 ENU 26 weeks mg/kg BW 120 Sarcoma Atypicale (0/19) 0% (4/16) 25% ND Female ND Tongue MNAN weeks 15 ppm 5 Hematopoietic ENU weeks 26 BW mg/kg 120 Lymphom Liver vascular cells BHP List of Experiments Showing Su Showing Experiments List of SCC, squamous cellSCC, carcinoma; CIS, carcinoma *: Significantly higher than p5 MNAN, methyl-n-amylnitrosamine; DBN, methyl-n-amylnitrosamine; MNAN, Urothelial cellUrothelial bladder Urinary BBN weeks 20 0.0075% D Columnar cell Columnar Lung ENU weeks 26 BW mg/kg 120 Adenoma (8/19) 42% (13/16)† 81% ND Female ND Table 1. Table typesCell organs Target Treatments Expeimental Dose types Tumor Squamous cellSquamous Esophagus MNAN weeks 15 ppm 5 (7/13) 53.8% (12/15) 80.0% (10/12) 83.3% Stromal cell Stromal bladder Urinary BBN weeks 20 0.0075, 0.004, Hemangioma/ 126 Susceptibility of the p53 Knockout Mice

(P<0.05). Stromal tumors were also observed, p53 (+/–) partial reduction of p53 protein amount in (+/–) mice animals demonstrating more fibrosarcomas in the 0.025% compared with (+/+) mice for cancer promotion and group than p53 (+/+) mice (P<0.05) (Table 1). progression. In the latter case, the progression rate was also greater in (+/–) than (+/+) mice and was associated with loss Development of vascular tumors in the liver of the remaining wild type allele. For simultaneous tumor induction in various organs, Ozaki et al.9 reported a high susceptibility of p53 (+/–) BHP was chosen to induce liver vascular tumors as well as mice to BBN urinary bladder carcinogenesis without lung tumors (Hirata et al., manuscript submitted). In a mid- frequent p53 gene mutations. Therefore, a lack of increased term experiment, male p53 (+/+), (+/–), and (–/–) mice were susceptibility may reflect organ/tissue specificity for the given drinking water containing 20 ppm BHP, or 200 ppm levels of p53 protein required for tumorigenesis, in addition for homozygotes and the wild type. Liver vascular tumors to involvement of p53 gene alterations. were found and histologically diagnosed as hemangiomas and hemangiosarcomas; some of them were confirmed to be Heterozygous p53 Knockout Mice are not immunohistochemically positive for von Willebrand factor. Sufficiently Susceptible to Carcinogens for In the 15-week experiment group treated with 20 ppm BHP, Effective Short-Term Testing vascular tumors were found in only p53 (–/–) mice, at incidences showing statistical significance compared with Carcinogenesis in the stomach p53 (+/+) and (+/–) mice (P<0.005 and P<0.05, An experiment was conducted to determine the respectively). Hepatic vascular tumors also developed in all sensitivity to induction of preneoplastic lesions, pepsinogen surviving p53 (+/–) mice, whereas p53 (+/+) tumor yields altered pyloric glands (PAPGs), within 5 weeks, and to were significantly lower with 200 ppm (hemangiomas clarify any differences in tumor yields between p53 (+/–) P<0.001; hemangiosarcomas P<0.001). For longer term and (–/–) mice administered 30 ppm MNU for 15 weeks17. A observation, male p53 (+/+) and (+/–) mice were given 20 longer study was also conducted to show the variation in ppm BHP for 40 weeks. Incidences of liver vascular tumors responses between p53 (+/+) and (+/–) mice administered 30 in BHP-treated p53 (+/–) mice were significantly higher than or 120 ppm MNU. Serial sections including pyloric glands in the wild type counterparts (hemangioma P<0.01; were stained with anti-pepsinogen 1 (Pg 1) antibody57 and hemangiosarcoma P<0.0001). Twenty liver were evaluated for PAPG. PAPG were stained weakly or hemangiosarcomas were subjected to microdissection, PCR- negative for Pg 1, and it was established that they are SSCP, and sequencing analysis of p53 exons 5–8 and p53 putative precancerous lesions in the mouse57, and also in the mutations were identified in 12 of 20 (60.0%) lesions. The rat58,59. In the 5-week experiment, histological sections frequency was significantly higher than that in lung tumors showed slightly irregular glands and weak hyperplasia or (described below) in which the mutation rate was only as low atrophy, especially in the p53 (–/–) mice. Pg 1 staining as 10%. Mutations in all five hemangiosarcomas (5/5, revealed PAPG even in normal looking mucosa in MNU- 100%) were found to be located in the wild-type allele of p53 treated p53 (–/–) mice, in contrast to the control mice in (+/–) mice, indicating complete loss of functional p53 which Pg 1 was consistently present in pyloric gland cells. protein (Fig. 2A and Table 1). The frequency of PAPGs was 22.6 ± 14.9% (mean ± SD) in MNU-treated p53 (–/–) mice, significantly elevated Target cells decide sensitivity of p53 (+/–) mice (P<0.001) compared with the values for MNU-treated p53 Not only nullizygous but also heterozygous p53 (+/+) and p53 (+/–) mice (1.8 ± 1.0 and 1.7 ± 0.8%, knockout mice are more susceptible to esophageal respectively) and the control groups (1.0 ± 0.4, 1.0 ± 0.3, and tumorigenesis induced by MNAN12, than their wild-type 1.2 ± 0.6 % in p53 (+/+), (+/–), and (–/–) mice, respectively). counterparts, as indicated by an increased incidence of We were thus convinced that nullizygous (–/–) p53 knockout SCCs. In addition, accelerated tumor development with mice are much more susceptible to stomach carcinogenesis chemical exposure in p53 (+/–) mice has been reported with induced with MNU than their heterozygous or wild type regard to lymphomas51, mesotheliomas52, skin tumors53, counterparts. In the 15-week experiment, MNU strongly vascular tumors14, urinary bladder tumors53, and lung affected p53 (–/–) mice, resulting in high mortality tumors54 (representative photos of a sarcoma and a compared to the other two groups, with development of lymphoma are shown in Fig. 2, B and C). Transitional cell lymphomas or sarcomas3,7,8. Stomach lesions including carcinoma development in the urinary bladder was also hyperplasia, adenomas, and an adenocarcinoma were found found to be enhanced in BBN treated p53 (+/–) mice as only in MNU-treated groups. Adenomas were found only in compared with the C57BL/6 parental strain9. The data thus p53 (+/–) and (–/–) mice, along with a single well suggest that squamous cell and transitional cell epithelia may differentiated adenocarcinoma in a homozygote. The trend be more influenced by decrease of p53 protein amount or of these lesions was revealed to be significantly shifted that the p53 gene of the wild type allele could be a direct toward malignancy in p53 (–/–) mice compared with p53 (+/ target in these tissues. In contrast, spontaneous sarcomas55 +) and (+/–) mice. In the long term experiment, there were and skin SCC induced with dimethylbenzanthracene plus no significant differences in the incidences of glandular 12-O-tetradecanoyl-phorbol-13-acetate (TPA)56 need only stomach lesions at week 40 between p53 (+/+) and (+/–) Tsukamoto, Hirata, Tatematsu 127 mice. Fifty-one animals were observed to have found in p53 (+/–) and p53 (+/+) female mice receiving 30 adenocarcinomas, which included 44 well differentiated, 4 ppm of APNH, relative to controls. However, there were no poorly differentiated, and 3 signet-ring cell carcinomas. The significant differences between the p53 (+/+) and p53 (+/–) results for PAPG coincided well with the fact that tumors in cases. Analysis of exons 5-8 of the p53 gene was performed p53 (–/–) mice developed within 15 weeks. A total of 68 on 13 and 10 HCC samples from p53 (+/–) and p53 (+/+) gastric tumors were subjected to PCR-SSCP analysis. mice, respectively, and 2 hepatocellular adenoma samples Although simultaneously developing lymphomas and each from p53 (+/–) and (+/+) mice. No evidence of p53 sarcomas in MNU-treated groups showed mutations in the gene mutations, however, was obtained (Table 2). wild type allele of the p53 gene, none of those gastric tumors proved to have any mutations17 (Fig. 3, A–C and Table 2). Lung carcinogenesis We have recently performed a multiorgan Colon carcinogenesis carcinogenesis study to assess the mechanisms that may p53 (+/+), (+/–), and (–/–) mice were treated with determine the susceptibility of p53 knockout mice to subcutaneous injections of DMH at a dose of 20 mg/kg body tumorigenesis (Hirata et al., manuscript submitted). The weight, once a week for 5 or 15/16 weeks in short- and mid- three genotypes were given drinking water containing 20 or term experiments, then sacrificed two weeks after the final 200 ppm BHP for 15 weeks in a mid-term experiment as injection18. Colonic lesions were classified into focal atypias, described above under “Development of vascular tumors in adenomas, and adenocarcinomas. In the DMH-treated p53 (– the liver”. Lung adenocarcinomas were observed after a /–) mice, three mice developed adenomas (3/14 = 21.4%) and period as short as 15 weeks and the incidence was the mean number of focal atypia per unit length of colon was significantly higher in p53 (–/–) than the other genotypes. 0.059 ± 0.13 /cm in contrast to a complete lack of them in (+/ p53 (+/+) and (+/–) mice were found to have only adenomas +) and (+/–) animals in the short-term experiment. In the at 15 weeks. In contrast, there were no statistically mid-term experiment, focal atypias and adenomas appeared significant differences in the incidences of lung lesions in DMH-treated p53 (+/+) and p53 (+/–) mice, along with between p53 (+/+) and (+/–) mice. A longer-term carcinomas in the p53 (–/–) case. The incidences of tumors of observation of male and female p53 (+/+) and (+/–) mice DMH-treated p53 (–/–) mice were significantly higher than given 20 ppm BHP water for 40 weeks also did not reveal values for the other DMH-treated genotypes and control any statistically significant differences in specific lung groups. In the long-term experiment (30 weeks), there were lesions, although the incidences of lung tumors in no significant differences between values for tumors in p53 heterozygotes of both sexes tended to be higher. In addition, (+/+) and (+/–) mice (P>0.999 for adenomas and P=0.315 for bronchial tumors were found in the long-term observation carcinomas). A total of 56 colonic proliferative lesions were and again the incidences were higher in p53 (+/–) than in p53 subjected to PCR-SSCP analysis of p53 exons 5–8. PCR (+/+) mice, but without statistical significance. Twenty lung products in most cases appeared to migrate at the same speed adenomas and 17 adenocarcinomas were subjected to as normal controls, but bands in four cases for exons 7 and 8 analysis of the p53 gene and mutations were identified in were shifted. Mutations were observed in one carcinoma in a two (10.0%) and two (11.8%), respectively. The frequency p53 (+/–) mouse and one each in a focal atypia, an adenoma of p53 mutations in lung tumors was significantly lower than and a carcinoma in three p53 (+/+) mice. No LOH was found in hemangiosarcomas (P<0.01) (Fig. 3, F and G and Table in any tumor or focal atypia in p53 (+/–) mice (Fig. 3, D and 2). E and Table 2). Intraperitoneal injection of ENU has also been used to induce tumors in multiple organs by Mitsumori et al.10. Liver carcinogenesis Regarding lung tumors, the incidence of adenomas was APNH, a novel heterocyclic amine, was synthesized by significantly higher in p53 (+/–) than that in p53 (+/+) mice Totsuka et al.60 and mixed into CA-1 (Clea Japan, Tokyo, (P<0.05). Furthermore, the incidence of atypical Japan) diet at 3, 10, and 30 ppm for ad libitum administration endometrial hyperplasia was also higher in the heterozygotes to p53 (+/+), (+/–), and (–/–) male mice for 15 weeks, and to (P<0.05). The p53 deficient mice used in that study had both sexes of p53 (+/+) and (+/–) mice for 40 weeks19. Liver been established by Tsukada et al.4 and backcrossed with lesions were diagnosed as oval cell hyperplasia, foci, female CBA mice and this might explain the discrepancy adenoma, and adenocarcinoma. In the 15-week experiment, with our data (Table 1). preneoplastic foci were limited to p53 (–/–) mice treated with 10 ppm or 30 ppm of APNH, albeit without statistical p53 (–/–) Mice are most sensitive significance as compared to the other genotypes. Various Since stomach cancer developed in wild type mice in a hepatic lesions including hepatocellular adenomas and previous study57, loss of p53 is not indispensable for hepatocellular carcinomas (HCCs) developed by week 40, stomach carcinogenesis. The results of PCR-SSCP analysis especially in females. Altered cell foci were observed at all reported above support the idea of no essential role for p53 dose levels in p53 (+/–) males and females, 30 ppm p53 (+/ by revealing no mutations in exons 5–8 of the p53 gene in 68 +) males, and 10 and 30 ppm dosed p53 (+/+) females. stomach tumors, analyzed in BALB/c mice as reported by Significant induction of adenomas and carcinomas was Furihata et al.61. However, complete loss of wild type p53 128 Susceptibility of the p53 Knockout Mice

Fig. 3. Adenocarcinomas in p53 knockout mice. (A) Normal pyloric gland preserving Pg1 in a control p53 (+/+) mouse. (B) PAPG showing pyloric glands with loss of Pg1 in an MNU treated p53 (–/–) mouse (red arrow). (C) Stomach adenocarcinoma in a p53 (+/ –) mouse. (D) Colonic neoplastic lesion (focal atypia) in a DMH treated p53 (+/+) mouse (blue arrow). (E) Colonic adenocarcinoma in a p53 (+/–) mouse. (F) Lung adenoma in a BHP treated p53 (+/–) mouse. (G) Lung adenocarcinoma in a BHP treated p53 (+/–) mouse. (A and B) Pg1 immunohistochemistry. (C–G) H&E staining. Original magnification: (A–D) 200×. (G) 100×. Tsukamoto, Hirata, Tatematsu 129 17 19 18 16 , 15 , 2000 , 2005 , , 2003 , , 1994 et al. et al. et al. et al. et al. References Hirata manuscript submitted manuscript mutations Frequency of p53 Frequency Sex Male (4/56) 7.1% Sakai, Male (2/17) 11.8% Male 0% (0/68) Yamamoto, Male (2/20) 10.0% Female 0% (0/27) Iidaka, ND Male LOH=0% (0/26) 1995 Kemp, oxypropyl)nitrosamine; DMBA, 7,12-dimethylbenz[a]anthracene. DMBA, oxypropyl)nitrosamine; # (n=21) # (+/+) (+/–) (–/–) Incidence of tumors in each p53 genotype (n=16) 0% (0/10)0% 0% (0/21)(1/13) 7.7% 0% (0/10)0% 0% (0/15)(14/20)* 70% 0% (0/19)0% 0% (0/21)(1/10) 10% 0% (0/19)0% (2/21) 9.5% (6/10)* 60% 0% (0/32)0% 0% (0/32)(5/21)* 23.8% 10% (1/10) 20% (3/15) 90% (18/20)* 3.1% (1/32)3.1% (1/32) 3.1% (0/21) 0% 8.3% (2/24)8.3% (2/19) 10.5% ND 37.5% (9/24)37.5% (4/19) 21.1% ND 63 tumors/mouse63 tumors/mouse 48 Foci HCC Adenoma Adenoma Adenoma Adenocarcinoma Adenocarcinoma nylnorharman; DEN, diethylnitrosamine; BHP, N-bis(2-hydr BHP, diethylnitrosamine; DEN, nylnorharman; 6 wk 6 Mammary tumor(7/20) 35% (12/30) 40% ND Female (1/7) 14.3% Jerry, 15 times 15 Adenoma times 15 Adenocarcinoma × × × Dose types Tumor mol/g BW tumor Liver 30 ppm30 (12/27) 44.4% (16/32) 34.8% ND 30 ppm30 ppm 10 (0/10) 0% 0% (0/23)(1/14) 7.1% 10 ppm 10 10 ppm 10 30 ppm30 µ (10/27) 37.0% (14/46) 30.4% ND Nullizygous p53 Knockout Mice Knockout p53 Nullizygous 20 mg/kg BW mg/kg 20 BW mg/kg 20 thylhydrazine; APNH, aminophe APNH, thylhydrazine; periods 30 30 weeks weeks 15 (5/22) 22.7% (7/18) 38.9% ND 30 30 weeks (3/22) 13.6% (2/18) 11.1% ND 15 weeks15 40 weeks weeks 15 ppm 30 120 ppm weeks 15 (7/12) 58.3% weeks15 56.5% (13/23) ND ppm 20 40 weeks40 ppm 120 (11/12) 91.7% (20/23) 87.0% ND 40 weeks40 ppm 20 (11/19) 57.9% (14/17) 82.4% ND 40 weeks40 ppm 20 (2/19) 10.5% (6/17) 35.3% ND 15 weeks15 ppm 20 Expeimental DEN 32 weeks 0.2 APNH Showing Susceptibility of Only of Susceptibility Showing than p53 (+/+) and (+/–). organs Treatments Lung BHP Liver weeks 40 Colon DMH Stomach MNU Mammary gland DMBA months 7 mg/wk 1.0 List of Experiments Experiments List of : Estimated from a figure. *: Significantly higher*: Significantly # MNU, N-methyl-N-nitrosourea; DMH, 1,2-dime DMH, N-methyl-N-nitrosourea; MNU, HCC, hepatocellular carcinoma. determined. ND, not Table 2. Table Cell typesCell Target Columnar cell weeks 15 ppm 30 130 Susceptibility of the p53 Knockout Mice expression greatly enhanced stomach carcinogenesis. Thus experiments73. Similarly, Okamoto et al.74,75 conducted a the p53 gene might not be a direct target of MNU but the p53 PCR-SSCP analysis of exons 5–8 in p53 in DMH- protein might act as a gatekeeper62,63 for initiation of induced colonic neoplasms and revealed 33 mutations in 22 neoplasia, as indicated by the PAPG data, and for tumor of 182 (12.1%) colon tumors (including multiple mutations), promotion and progression, as reflected in the more but again allelic loss was detected in only 2 of 163 lesions advanced stomach lesion development in p53 (–/–) mice. (1.2%). Since p53 (+/–) mice have one inactive allele, they Regarding colon carcinogenesis, the C57BL/6J strain, the might have been predicted to be more susceptible to background to the p53-deficient mice reviewed here used in carcinogenesis compared with their wild type counterparts. the study mentioned above3, is moderately resistant to Analysis of the p53 gene demonstrated no mutations in the DMH, with 43–48% tumor incidences reported64,65. liver tumors developing in either p53 (+/+) or p53 (+/–) However, colon carcinomas were already observed in 70% mice, consistent with the lack of structural aberrations in the of p53 (–/–) mice after only 16 weeks of DMH treatment, p53 gene in carcinogen-induced liver tumors in C57BL/6J with considerable reduction in the latency. The high strain mice76. In contrast, PCR-SSCP analysis revealed a susceptibility of p53 (–/–) to DMH-induction of colorectal high frequency of missense mutations in p53 with evidence carcinogenesis indicates that loss of p53 function has an of loss of functional p53 protein in esophageal malignancies. accelerating effect. In vitro experiments have shown that Taken together, these results support the hypothesis that non-tumorigenic colonic epithelial cells from p53 (–/–) mice mutational inactivation of the retained wild-type allele or acquire tumorigenic potential upon ectopic expression of an loss of p53 heterozygosity, with consequent loss of p53 activated Ki-ras gene66. Therefore, complete lack of p53 function, eventually results in development of neoplasias as may facilitate tumorigenesis on activation of oncogenes or occurs in human Li-Fraumeni syndrome77. Our results point inactivation of other tumor suppressor genes, and degrees of to a high frequency of p53 mutations in esophageal gene alteration may be reflected in grades of malignancy. In malignancies, the most common being missense, as reported particular, activating mutations in the β-catenin gene are for human cancer78. In contrast to the frequent p53 thought to be responsible for the excess β-catenin signaling mutations in SCCs, even in small carcinomas, only one featured in the majority of carcinogen-induced colonic mutation was detected in 19 papillomas analyzed. Similar lesions67, including small foci68. findings have been reported in murine skin tumors induced Mammary carcinogenesis induced by DMBA16 and with benzo[a]pyrene in which the majority of p53 mutations hepatocarcinogenesis initiated with the genotoxic were identified in SCCs; they were rare in papillomas79. carcinogens dimethylnitrosamine (DMN), 2-amino-1- Moreover, abnormal p53 protein is only infrequently methyl-6-phenylimidazo [4,5-b]pyridine (PhIP), 6- identified in human esophageal squamous cell papillomas80 nitrochrysene (6-NC)69, 2-amino-3,8-dimethylimidazo[4,5- while dysplastic lesions exhibit p53 mutations81,82. f]quinoxaline (MeIQx)70, DEN15, and other chemicals71 Regarding the location of p53 mutations, they were including APNH, a novel heterocyclic amine19, may reflect randomly distributed through exons 5 to 8, with more than organ/tissue specificity in the threshold for the p53 gene one mutation detected at codons 164 (8%, 2/25), 219 (4/25, product. While the mechanism may involve an additional hit 16%), 232 (2/25, 8%) and 250 (4/25, 16%). Approximately to inactivate the second normal allele, Venkatachalam et half of the mutations observed in our study were G:C to A:T al.55 have proposed that reduction of the p53 gene products transitions at non-CpG sites. Retrospective analyses of p53 may be sufficient to promote tumorigenesis. gene mutations in human esophageal cancers have also shown a predominance of G:C to A:T transitions83,84, Conclusions indicating that the MNAN-induced esophageal carcinogenesis model mimics the human disease. In Mutations in the p53 gene commonly occur at hot spots dimethylbenzanthracene / 12-O-tetradecanoyl-phorbol-13- in human cancers1, but the mutation database for laboratory acetate induced skin carcinogenesis in p53 deficient mice, animals is limited. Our previous study revealed no LOH of the p53 gene in (+/–) mice enhanced malignant mutations in 68 MNU-induced gastric tumors using PCR- progression56. From these results, p53 alteration may be SSCP analysis in p53 deficient mice17, while the mutation organ-specific (Fig. 4). rate observed for the p53 gene was 3 in 46 (6.5%) colonic In summary, our studies indicate that p53 (–/–) mice tumors in p53 (+/–) and (+/+) mice, much lower than that may be universally susceptible to carcinogens, whereas observed for human colon cancers. Moreover, allelic loss of sensitivity in p53 (+/–) mice is organ-specific and tumors in heterozygous mice was not detected in that study, dramatically increased by involvement of p53 gene while LOH is frequent (75%) in colon carcinomas in alterations in the carcinogenic process, in addition to humans. In humans, inactivation of p53 is most often due to mechanisms related to haploinsufficiency. Though p53 a missense mutation combined with loss of the other allele72. knockout mice are powerful tools for identification of In colon adenocarcinomas of rats treated with DMH or carcinogens, organ/tissue and carcinogen specific azoxymethane, p53 gene alteration was only rarely detected susceptibility must be taken into consideration for with PCR-SSCP analysis in an earlier series of appropriate evaluation. Tsukamoto, Hirata, Tatematsu 131

Fig. 4. p53 gene mutations and susceptibility of p53 knockout mice. The p53 gene may not be a direct target of carcinogens in columnar cells like stomach, colon, and liver cells, as evidenced by susceptibility limited to nullizygotes (–/–). In contrast, the p53 gene itself may be a direct target of carcinogens in squamous cells in the tongue and esophagus or in urothelial mucosa, thus sensitivity of heterozygotes (+/–) is increased by inactivation of wild type p53. Red square, mutant allele. White square, wild type allele. “X” indicates mutation of p53 gene induced by a carcinogen.

Acknowledgements: The authors thank Dr. Lawrence A. in p53-mutant mice. Curr Biol. 4: 1–7. 1994. Donehower for continuous collaboration and Drs. Masami 6. Purdie CA, Harrison DJ, Peter A, Dobbie L, White S, Howie Yamamoto, Hiroki Sakai, Norimitsu Shirai, and Takeshi SE, Salter DM, Bird CC, Wyllie AH, Hooper ML, and Iidaka for their expert technical assistance. We also are Clarke AR. Tumour incidence, spectrum and ploidy in mice grateful to Dr. Malcolm Moore for valuable discussion. This with a large deletion in the p53 gene. Oncogene. 9: 603–609. work was supported in part by a Grant-in-Aid for Cancer 1994. Research from the Ministry of Health, Labour and Welfare, 7. Donehower LA, Harvey M, Vogel H, McArthur MJ, Japan and a Grant-in-Aid from the Ministry of Education, Montgomery CA Jr., Park SH, Thompson T, Ford RJ, and Culture, Sports, Science and Technology of Japan. Bradley A. Effects of genetic background on tumorigenesis in p53-deficient mice. Mol Carcinog. 14: 16–22. 1995. 8. Harvey M, McArthur MJ, Montgomery CJ, Butel JS, References Bradley A, and Donehower LA. Spontaneous and carcinogen-induced tumorigenesis in p53-deficient mice. 1. Hollstein M, Sidransky D, Vogelstein B, and Harris CC. p53 Nat Genet. 5: 225–229. 1993. mutations in human cancers. Science. 253: 49–53. 1991. 9. Ozaki K, Sukata T, Yamamoto S, Uwagawa S, Seki T, 2. Srivastava S, Zou ZQ, Pirollo K, Blattner W, and Chang EH. Kawasaki H, Yoshitake A, Wanibuchi H, Koide A, Mori Y, Germ-line transmission of a mutated p53 gene in a cancer- and Fukushima S. High susceptibility of p53 (+/–) knockout prone family with Li-Fraumeni syndrome. Nature. 348: 747– mice in N-butyl-N-(4-hydroxybutyl)nitrosamine urinary 749. 1990. bladder carcinogenesis and lack of frequent mutation in 3. Donehower LA, Harvey M, Slagle BL, McArthur MJ, residual allele. Cancer Res. 58: 3806–3811. 1998. Montgomery CJ, Butel JS, and Bradley A. Mice deficient for 10. Mitsumori K, Onodera H, Shimo T, Yasuhara K, Takagi H, p53 are developmentally normal but susceptible to Koujitani T, Hirose M, Maruyama C, and Wakana S. Rapid spontaneous tumours. Nature. 356: 215–221. 1992. induction of uterine tumors with p53 point mutations in 4. Tsukada T, Tomooka Y, Takai S, Ueda Y, Nishikawa S, heterozygous p53-deficient CBA mice given a single Yagi T, Tokunaga T, Takeda N, Suda Y, Abe S, Matsuo I, intraperitoneal administration of N-ethyl-N-nitrosourea. Ikawa Y, and Aizawa S. Enhanced proliferative potential in Carcinogenesis. 21: 1039–1042. 2000. culture of cells from p53-deficient mice. Oncogene. 8: 11. Nishikawa T, Salim EI, Morimura K, Kaneko M, Ogawa M, 3313–3322. 1993. Kinoshita A, Osugi H, Kinoshita H, and Fukushima S. High 5. Jacks T, Remington L, Williams BO, Schmitt EM, Halachmi susceptibility of p53 knockout mice to esophageal and S, Bronson RT, and Weinberg RA. Tumor spectrum analysis urinary bladder carcinogenesis induced by N, N- 132 Susceptibility of the p53 Knockout Mice

dibutylnitrosamine. Cancer Lett. 194: 45–54. 2003. Nakamura Y, White R, and Vogelstein B. Chromosome 17 12. Shirai N, Tsukamoto T, Yamamoto M, Iidaka T, Sakai H, deletions and p53 gene mutations in colorectal carcinomas. Yanai T, Masegi T, Donehower LA, and Tatematsu M. Science. 244: 217–221. 1989. Elevated susceptibility of the p53 knockout mouse 28. Takahashi T, Nau MM, Chiba I, Birrer MJ, Rosenberg RK, esophagus to methyl-N-amylnitrosamine carcinogenesis. Vinocour M, Levitt M, Pass H, Gazdar AF, and Minna JD. Carcinogenesis. 23: 1541–1547. 2002. p53: a frequent target for genetic abnormalities in lung 13. Shirai N, Tsukamoto T, Yamamoto M, Iidaka T, Sakai H, cancer. Science. 246: 491–494. 1989. Yanai T, Masegi T, Donehower LA, and Tatematsu M. 29. Levine AJ, Momand J, and Finlay CA. The p53 tumour Tongue carcinogenic susceptibility of p53 deficient mice to suppressor gene. Nature. 351: 453–456. 1991. methyl-n-amylnitrosamine. J Toxicol Pathol. 15: 209–214. 30. Gudkov AV and Komarova EA. The role of p53 in 2002. determining sensitivity to radiotherapy. Nat Rev Cancer. 3: 14. Carmichael NG, Debruyne EL, and Bigot-Lasserre D. The 117–129. 2003. p53 heterozygous knockout mouse as a model for chemical 31. Dixon K and Kopras E. Genetic alterations and DNA repair carcinogenesis in vascular tissue. Environ Health Perspect. in human carcinogenesis. Semin Cancer Biol. 14: 441–448. 108: 61–65. 2000. 2004. 15. Kemp CJ. Hepatocarcinogenesis in p53-deficient mice. Mol 32. Hussain SP, Hofseth LJ, and Harris CC. Radical causes of Carcinog. 12: 132–136. 1995. cancer. Nat Rev Cancer. 3: 276–285. 2003. 16. Jerry DJ, Butel JS, Donehower LA, Paulson EJ, Cochran C, 33. Sancar A, Lindsey-Boltz LA, Unsal-Kacmaz K, and Linn S. Wiseman RW, and Medina D. Infrequent p53 mutations in Molecular mechanisms of mammalian DNA repair and the 7,12-dimethylbenz[a]anthracene-induced mammary tumors DNA damage checkpoints. Annu Rev Biochem. 73: 39–85. in BALB/c and p53 hemizygous mice. Mol Carcinog. 9: 2004. 175–183. 1994. 34. Kurz EU and Lees-Miller SP. DNA damage-induced 17. Yamamoto M, Tsukamoto T, Sakai H, Shirai N, Ohgaki H, activation of ATM and ATM-dependent signaling pathways. Furihata C, Donehower LA, Yoshida K, and Tatematsu M. DNA Repair (Amst). 3: 889–900. 2004. p53 knockout mice (–/–) are more susceptible than (+/–) or 35. Bartek J and Lukas J. Pathways governing G1/S transition (+/+) mice to N-methyl-N-nitrosourea stomach and their response to DNA damage. FEBS Lett. 490: 117– carcinogenesis. Carcinogenesis. 21: 1891–1897. 2000. 122. 2001. 18. Sakai H, Tsukamoto T, Yamamoto M, Shirai N, Iidaka T, 36. Taylor WR and Stark GR. Regulation of the G2/M transition Hirata A, Yanai T, Masegi T, Donehower LA, and by p53. Oncogene. 20: 1803–1815. 2001. Tatematsu M. High susceptibility of nullizygous p53 37. Dulic V, Kaufmann WK, Wilson SJ, Tlsty TD, Lees E, knockout mice to colorectal tumor induction by 1,2- Harper JW, Elledge SJ, and Reed SI. p53-dependent dimethylhydrazine. J Cancer Res Clin Oncol. 129: 335–340. inhibition of cyclin-dependent kinase activities in human 2003. fibroblasts during radiation-induced G1 arrest. Cell. 76: 19. Iidaka T, Tsukamoto T, Totsuka Y, Hirata A, Sakai H, Shirai 1013–1023. 1994. N, Yamamoto M, Wakabayashi K, Yanai T, Masegi T, 38. Bartek J and Lukas J. Mammalian G1- and S-phase Donehower LA, and Tatematsu M. Lack of elevated liver checkpoints in response to DNA damage. Curr Opin Cell carcinogenicity of aminophenylnorharman in p53-deficient Biol. 13: 738–747. 2001. mice. Cancer Lett. 217: 149–159. 2005. 39. Zhu Y, Alvarez C, Doll R, Kurata H, Schebye XM, Parry D, 20. Harris CC and Hollstein M. Clinical implications of the p53 and Lees E. Intra-S-phase checkpoint activation by direct tumor-suppressor gene. N Engl J Med. 329: 1318–1327. CDK2 inhibition. Mol Cell Biol. 24: 6268–6277. 2004. 1993. 40. Zhivotovsky B and Kroemer G. and genomic 21. McBride OW, Merry D, and Givol D. The gene for human instability. Nat Rev Mol Cell Biol. 5: 752–762. 2004. p53 cellular tumor antigen is located on chromosome 17 41. Mihara M, Erster S, Zaika A, Petrenko O, Chittenden T, short arm (17p13). Proc Natl Acad Sci USA. 83: 130–134. Pancoska P, and Moll UM. p53 has a direct apoptogenic role 1986. at the mitochondria. Mol Cell. 11: 577–590. 2003. 22. Sidransky D and Hollstein M. Clinical implications of the 42. Perfettini JL, Kroemer RT, and Kroemer G. Fatal liaisons of p53 gene. Annu Rev Med. 47: 285–301. 1996. p53 with Bax and Bak. Nat Cell Biol. 6: 386–388. 2004. 23. Lane DP and Crawford LV. T antigen is bound to a host 43. Vousden KH and Lu X. Live or let die: the cell's response to protein in SV40-transformed cells. Nature. 278: 261–263. p53. Nat Rev Cancer. 2: 594–604. 2002. 1979. 44. Ashkenazi A and Dixit VM. Death receptors: signaling and 24. Linzer DI and Levine AJ. Characterization of a 54K dalton modulation. Science. 281: 1305–1308. 1998. cellular SV40 tumor antigen present in SV40-transformed 45. MacLachlan TK and El-Deiry WS. Apoptotic threshold is cells and uninfected embryonal carcinoma cells. Cell. 17: lowered by p53 transactivation of caspase-6. Proc Natl Acad 43–52. 1979. Sci USA. 99: 9492–9497. 2002. 25. Rovinski B and Benchimol S. Immortalization of rat embryo 46. Galande S, Dickinson LA, Mian IS, Sikorska M, and Kohwi- fibroblasts by the cellular p53 oncogene. Oncogene. 2: 445– Shigematsu T. SATB1 cleavage by caspase 6 disrupts PDZ 452. 1988. domain-mediated dimerization, causing detachment from 26. Eliyahu D, Michalovitz D, Eliyahu S, Pinhasi-Kimhi O, and early in T-cell apoptosis. Mol Cell Biol. 21: Oren M. Wild-type p53 can inhibit oncogene-mediated focus 5591–5604. 2001. formation. Proc Natl Acad Sci USA. 86: 8763–8767. 1989. 47. Haupt S, Berger M, Goldberg Z, and Haupt Y. Apoptosis - 27. Baker SJ, Fearon ER, Nigro JM, Hamilton SR, Preisinger the p53 network. J Cell Sci. 116: 4077–4085. 2003. AC, Jessup JM, vanTuinen P, Ledbetter DH, Barker DF, 48. Storer RD, French JE, Haseman J, Hajian G, LeGrand EK, Tsukamoto, Hirata, Tatematsu 133

Long GG, Mixson LA, Ochoa R, Sagartz JE, and Soper KA. 1998. P53+/– hemizygous knockout mouse: overview of available 61. Furihata C, Tatematsu M, Saito M, Ishida S, Nakanishi H, data. Toxicol Pathol. 29 (Suppl): 30–50. 2001. Inada K, Tei H, Hattori M, Ito T, and Sakaki Y. Rare 49. Venkatachalam S, Tyner SD, Pickering CR, Boley S, Recio occurrence of ras and p53 gene mutations in mouse stomach L, French JE, and Donehower LA. Is p53 haploinsufficient tumors induced by N-methyl-N-nitrosourea. Jpn J Cancer for tumor suppression? Implications for the p53+/– mouse Res. 88: 363–368. 1997. model in carcinogenicity testing. Toxicol Pathol. 29 62. Kinzler KW and Vogelstein B. Cancer-susceptibility genes. (Suppl): 147–154. 2001. Gatekeepers and caretakers. Nature. 386: 761, 763. 1997. 50. Ohne M, Satoh T, Yamada S, and Takai H. Experimental 63. Levine AJ. p53, the cellular gatekeeper for growth and tongue carcinoma of rats induced by oral administration of division. Cell. 88: 323–331. 1997. 4-nitroquinoline 1-oxide (4NQO) in drinking water. Oral 64. Diwan BA and Blackman KE. Differential susceptibility of 3 Surg Oral Med Oral Pathol. 59: 600–607. 1985. sublines of C57BL/6 mice to the induction of colorectal 51. Dunnick JK, Hardisty JF, Herbert RA, Seely JC, Furedi- tumors by 1,2-dimethylhydrazine. Cancer Lett. 9: 111–115. Machacek EM, Foley JF, Lacks GD, Stasiewicz S, and 1980. French JE. Phenolphthalein induces thymic lymphomas 65. Diwan BA, Meier H, and Blackman KE. Genetic differences accompanied by loss of the p53 wild type allele in in the induction of colorectal tumors by 1,2- heterozygous p53-deficient (+/–) mice. Toxicol Pathol. 25: dimethylhydrazine in bred mice. J Natl Cancer Inst. 59: 455– 533–540. 1997. 458. 1977. 52. Marsella JM, Liu BL, Vaslet CA, and Kane AB. 66. Sevignani C, Wlodarski P, Kirillova J, Mercer WE, Susceptibility of p53-deficient mice to induction of Danielson KG, Iozzo RV, and Calabretta B. Tumorigenic mesothelioma by crocidolite asbestos fibers. Environ Health conversion of p53-deficient colon epithelial cells by an Perspect. 105 (Suppl 5): 1069–1072. 1997. activated Ki-ras gene. J Clin Invest. 101: 1572–1580. 1998. 53. Tennant RW, French JE, and Spalding JW. Identifying 67. Tsukamoto T, Tanaka H, Fukami H, Inoue M, Takahashi M, chemical carcinogens and assessing potential risk in short- Wakabayashi K, and Tatematsu M. More frequent beta- term bioassays using transgenic mouse models. Environ catenin gene mutations in adenomas than in aberrant crypt Health Perspect. 103: 942–950. 1995. foci or adenocarcinomas in the large intestines of 2-amino-1- 54. Finch GL, March TH, Hahn FF, Barr EB, Belinsky SA, methyl-6-phenylimidazo[4,5-b]pyridine (PhIP)-treated rats. Hoover MD, Lechner JF, Nikula KJ, and Hobbs CH. Jpn J Cancer Res. 91: 792–796. 2000. Carcinogenic responses of transgenic heterozygous p53 68. Yamada Y, Yoshimi N, Hirose Y, Kawabata K, Matsunaga knockout mice to inhaled 239PuO2 or metallic beryllium. K, Shimizu M, Hara A, and Mori H. Frequent beta-catenin Toxicol Pathol. 26: 484–491. 1998. gene mutations and accumulations of the protein in the 55. Venkatachalam S, Shi YP, Jones SN, Vogel H, Bradley A, putative preneoplastic lesions lacking macroscopic aberrant Pinkel D, and Donehower LA. Retention of wild-type p53 in crypt foci appearance, in rat colon carcinogenesis. Cancer tumors from p53 heterozygous mice: reduction of p53 Res. 60: 3323–3327. 2000. dosage can promote cancer formation. EMBO J. 17: 4657– 69. Dass SB, Bucci TJ, Heflich RH, and Casciano DA. 4667. 1998. Evaluation of the transgenic p53+/– mouse for detecting 56. Kemp CJ, Donehower LA, Bradley A, and Balmain A. genotoxic liver carcinogens in a short-term bioassay. Cancer Reduction of p53 gene dosage does not increase initiation or Lett. 143: 81–85. 1999. promotion but enhances malignant progression of 70. Park CB, Kim DJ, Uehara N, Takasuka N, Hiroyasu BT, and chemically induced skin tumors. Cell. 74: 813–822. 1993. Tsuda H. Heterozygous p53-deficient mice are not 57. Yamamoto M, Furihata C, Fujimitsu Y, Imai T, Inada K, susceptible to 2-amino-3,8-dimethylimidazo[4,5- Nakanishi H, and Tatematsu M. Dose-dependent induction f]quinoxaline (MeIQx) carcinogenicity. Cancer Lett. 139: of both pepsinogen-altered pyloric glands and 177–182. 1999. adenocarcinomas in the glandular stomach of C3H mice 71. Uehara T, Kashida Y, Watanabe T, Yasuhara K, Onodera H, treated with N- methyl-N-nitrosourea. Jpn J Cancer Res. 88: Hirose M, and Mitsumori K. Susceptibility of liver 238–244. 1997. proliferative lesions in heterozygous p53 deficient CBA 58. Tatematsu M, Furihata C, Katsuyama T, Mera Y, Inoue T, mice to various carcinogens. J Vet Med Sci. 64: 551–556. Matsushima T, and Ito N. Immunohistochemical 2002. demonstration of pyloric gland-type cells with low- 72. Kinzler KW and Vogelstein B. Colorectal Tumors. In: The pepsinogen isozyme 1 in preneoplastic and neoplastic tissues Genetic Basis of Human Cancer. Kinzler KW and of rat stomachs treated with N-methyl-N'-nitro-N- Vogelstein B (eds), McGraw-Hill, New York. 565–587. nitrosoguanidine. J Natl Cancer Inst. 78: 771–777. 1987. 1998. 59. Tatematsu M, Furihata C, Mera Y, Shirai T, Matsushima T, 73. Erdman SH, Wu HD, Hixson LJ, Ahnen DJ, and Gerner EW. and Ito N. Immunohistochemical demonstration of induction Assessment of mutations in Ki-ras and p53 in colon cancers of pyloric glands with low pepsinogen 1 (Pg 1) content in rat from azoxymethane- and dimethylhydrazine-treated rats. stomach by N-methyl-N'-nitro-N-nitrosoguanidine. Jpn J Mol Carcinog. 19: 137–144. 1997. Cancer Res. 77: 238–243. 1986. 74. Okamoto M, Ohtsu H, Kominami R, and Yonekawa H. 60. Totsuka Y, Hada N, Matsumoto K, Kawahara N, Murakami Mutational and LOH analyses of p53 alleles in colon tumors Y, Yokoyama Y, Sugimura T, and Wakabayashi K. induced by 1,2-dimethylhydrazine in F1 hybrid mice. Structural determination of a mutagenic Carcinogenesis. 16: 2659–2666. 1995. aminophenylnorharman produced by the co-mutagen 75. Okamoto M, Ohtsu H, Miyaki M, and Yonekawa H. No norharman with aniline. Carcinogenesis. 19: 1995–2000. allelic loss at the p53 locus in 1,2-dimethylhydrazine- 134 Susceptibility of the p53 Knockout Mice

induced mouse colon tumours: PCR-SSCP analysis with precancerous lesions and carcinoma among high-risk sequence-tagged microsatellite site primers. Carcinogenesis. populations in Henan, China. Cancer Res. 54: 4342–4346. 14: 1483–1486. 1993. 1994. 76. Kress S, Konig J, Schweizer J, Lohrke H, Bauer-Hofmann 82. Mandard AM, Hainaut P, and Hollstein M. Genetic steps in R, and Schwarz M. p53 mutations are absent from the development of squamous cell carcinoma of the carcinogen-induced mouse liver tumors but occur in cell esophagus. Mutat Res. 462: 335–342. 2000. lines established from these tumors. Mol Carcinog. 6: 148– 83. Greenblatt MS, Bennett WP, Hollstein M, and Harris CC. 158. 1992. Mutations in the p53 tumor suppressor gene: clues to cancer 77. Evans SC and Lozano G. The Li-Fraumeni syndrome: an etiology and molecular pathogenesis. Cancer Res. 54: 4855– inherited susceptibility to cancer. Mol Med Today. 3: 390– 4878. 1994. 395. 1997. 84. Mirvish SS. Role of N-nitroso compounds (NOC) and N- 78. Hussain SP and Harris CC. Molecular epidemiology and nitrosation in etiology of gastric, esophageal, carcinogenesis: endogenous and exogenous carcinogens. nasopharyngeal and bladder cancer and contribution to Mutat Res. 462: 311–322. 2000. cancer of known exposures to NOC. Cancer Lett. 93: 17–48. 79. Ruggeri B, DiRado M, Zhang SY, Bauer B, Goodrow T, and 1995. Klein-Szanto AJ. Benzo[a]pyrene-induced murine skin tumors exhibit frequent and characteristic G to T mutations Note added in proof in the p53 gene. Proc Natl Acad Sci USA. 90: 1013–1017. The manuscript by Hirata et al. has been accepted for 1993. publication: Hirata A, Tsukamoto T, Yamamoto M, Sakai 80. Poljak M, Cerar A, and Orlowska J. p53 protein expression H, Yanai T, Masegi T, Donehower LA, and Tatematsu M. in esophageal squamous cell papillomas: a study of 36 Organ-specific susceptibility of p53 knockout mice to N- lesions. Scand J Gastroenterol. 31: 10–13. 1996. bis(2-hydroxypropyl)nitrosamine carcinogenesis. Cancer 81. Gao H, Wang LD, Zhou Q, Hong JY, Huang TY, and Yang Lett. in press. CS. p53 tumor suppressor gene mutation in early esophageal