p53+/- Hemizygous Knockout Mouse: Overview of ...

Toxicologic Pathology http://tpx.sagepub.com/

p53 +/- Hemizygous Knockout Mouse: Overview of Available Data Richard D. Storer, John E. French, Joseph Haseman, Gerald Hajian, Edmund K. Legrand, Gerald G. Long, Lori A. Mixson, Ricardo Ochoa, John E. Sagartz and Keith A. Soper Toxicol Pathol 2001 29: 30 DOI: 10.1080/019262301753178465 The online version of this article can be found at: http://tpx.sagepub.com/content/29/1_suppl/30

Published by: http://www.sagepublications.com

On behalf of:

Society of Toxicologic Pathology

Additional services and information for Toxicologic Pathology can be found at: Email Alerts: http://tpx.sagepub.com/cgi/alerts Subscriptions: http://tpx.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav Citations: http://tpx.sagepub.com/content/29/1_suppl/30.refs.html

Downloaded from tpx.sagepub.com by guest on July 15, 2011

TOXICOLOGIC PATHOLOGY, vol 29(Suppl.), pp 30– 50, 2001 Copyright C 2001 by the Society of Toxicologic Pathology

p53+=¡ Hemizygous Knockout Mouse: Overview of Available Data RICHARD D. STORER,1 JOHN E. FRENCH,2 JOSEPH HASEMAN,2 GERALD HAJIAN,3 EDMUND K. LEGRAND,4 GERALD G. LONG,5 LORI A. MIXSON,1 RICARDO OCHOA,6 JOHN E. SAGARTZ,7 AND KEITH A. SOPER1 2

1 Merck & Co Inc, West Point, Pennsylvania 19486 National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina 27709 3 Schering Plough Research Institute, Kenilworth, New Jersey 07083 4 RW Johnson Pharmaceutical Research Institute, Raritan New Jersey 08869 5 Eli Lilly & Co, Green eld, Indiana 46140 6 P zer, Inc, Groton, Connecticut 06340 7 Monsanto, St. Louis, MO 63141

ABSTRACT The performance of the p53 / transgenic (knockout) mouse model was evaluated through review of the data from 31 short-term carcinogenicity studies with 21 compounds tested as part of the International Life Sciences Institute’s (ILSI) Alternatives to Carcinogenicity Testing (ACT) project, together with data from other studies which used comparable protocols. As expected based on the hypothesis for the model, a signi cant number (12/16 or 75%) of the genotoxic human and/or rodent carcinogens tested were positive and the positive control, p-cresidine, gave reproducible responses across laboratories (18/19 studies positive in bladder). An immunosuppressive human carcinogen, cyclosporin A, was positive for lymphomas but produced a similar response in wild type mice. Two hormones that are human tumorigens, diethylstilbestrol and 17b -estradiol, gave positive and equivocal results, respectively, in the pituitary with p53-de cient mice showing a greater incidence of proliferative lesions than wild type. None of the 22 nongenotoxic rodent carcinogens that have been tested produced a positive response but 2 compounds in this category, chloroform and diethylhexylphthalate, were judged equivocal based on effects in liver and kidney respectively. Four genotoxic noncarcinogens and 6 nongenotoxic, noncarcinogens were also negative. In total (excluding compounds with equivocal results), 42 of 48 compounds or 88% gave results that were concordant with expectations. The technical lessons learned from the ILSI ACT-sponsored testing in the p53 / model are discussed.

Keywords. p53 de cient; bioassay; short-term; 26-weeks; alternatives; carcinogenicity; testing.

INTRODUCTION The current ICH guidelines on testing for carcinogenicity, with their allowance for the potential use of alternative shortand medium-term in vivo carcinogenicity assays, have created an urgent need for further evaluation of the performance of these assays (7, 22). The International Life Sciences Institute (ISLI) recognized the need for further characterization of these models and formed a committee, the Alternatives to Carcinogenicity Testing Committee (ACT), to organize and direct an international collaborative effort to evaluate 21 well-characterized, non-proprietary compounds (mostly pharmaceuticals) in 4 transgenic models (RasH2, p53 / , Tg.AC, and XPA / ), the neonatal mouse model and the Syrian hamster embryo (SHE) cell in vitro assay (1, 19, 38, 42). In order to provide for a rigorous and consistent review of the resulting data from the sponsoring laboratories, assay working groups (AWG ) were formed comprising toxicologists, pathologists and statisticians from industry and government. The AWGs were charged with reviewing the draft reports from each study and independently evaluating the outcome of the study including the performance of the positive control compound, the rationale for and adequacy of the dose selection, the tumor incidence data, and the use of statistics. The overview of the available data presented here by the ILSI/HESI p53 AWG is intended to summarize the results of

that process for the compounds evaluated in the p53 / heterozygous knockout mouse model and to place these results in perspective together with all of the other available data for the model in studies that used comparable study protocols.

MATERIALS AND METHODS Animals and Husbandry C57BL/6TacfBR-[ KO] p53 N5 heterozygous mice and C57BL/6TacfBR-[KO] p53 N5 wild type mice were obtained from Taconic Farms, Inc., Germantown, NY, USA. Mice were generally 6 to 10 weeks old at the initiation of studies. Mice were supplied ad libitum with drinking water and the standard rodent diet used in the sponsoring laboratory throughout the study. The standard protocol for the 26-week assays recommended that all mice be genotyped prior to assignment to dose groups. Male mice were generally singly caged but female mice were group caged in some studies. Test Substances and General Protocol The test substances used in the ILSI ACT studies in the p53 model included all of the 21 listed compounds selected to be part of the ILSI/HESI ACT program. Several compounds (melphalan, diethylstilbestrol, cyclosporin A, and reserpine) included on the list had studies completed, and in one case published, prior to nalization of the ILSI ACT protocols (14, 43). With the exception of reserpine (43), the results of these studies were reviewed by the p53 Assay Working Group (AWG) in the same manner as for other compounds tested. p-Cresidine administered at 400 mg/kg/day by gavage in corn

Address correspondence to: Richard D. Storer, Merck Research Laboratories, Dept. of Safety Assessment, WP45-311, West Point, PA 19486; e-mail: richard [email protected].

30


0192-6233/01$3.00 $0.00

Vol. 29(Suppl. ), 2001

p53

/

: OVERVIEW OF AVAILABLE DATA

oil at a volume of 10 ml/kg 7 days per week, or benzene administered at 100 mg/kg/day by gavage in corn oil 7 days per week at a volume of 5 ml/kg, were the recommended positive controls. The positive controls used, if any, for each study are listed in Table 1. Standardized protocols were used for both the 4-week dose-range nding and 26-week short-term carcinogenicity studies (38). Protocols for the 4 compounds noted above for which studies were completed prior to the start of the ILSI ACT studies were similar (14, 43) although the duration of dosing was 24 – 25 weeks with an additional 4 – 6-week holding period prior to necropsy in the reserpine study, which was negative (43). The experimental design for ILSI ACT protocol studies called for 15 mice/sex/group at each of 3 dose levels with the recommendation of inclusion of control and high-dose groups of wild-type mice to allow for assessment of any differential responses based on the genotype. Generally, during the in-life phase of the studies, clinical signs, mortality, body weight, and in some cases food consumption and dietary intake of test substance (for diet studies ) were measured and/or recorded at regular intervals. In some studies, blood samples were taken and hematology and/or clinical chemistry parameters were measured. In most of the 26-week studies, exposure to the test article was con rmed by toxicokinetic analyses. At necropsy, tissues were examined macroscopically for evidence of gross lesions and the protocol indicated that sections of any masses greater than 2– 4 mm (and normal tissue from the same animal) should be frozen for possible molecular analysis. Organ weights were typically determined for 5 organs; brain, heart, kidney, liver, and gonads (testes or ovaries). Sections from a list of 40 different tissues were preserved, prepared for microscopic examination and evaluated for histopathological alterations (neoplastic and nonneoplastic ) as per the standard operating procedures of the sponsoring laboratories. Representative sections of all tissues from all mice in the control, positive control, and high dose groups, as well as all gross lesions were examined. Tissues from the low and mid-dose groups were also examined when treatment-related effects were noted in the high-dose group. Statistical Analyses and Interpretation of Tumor Incidence Data Study reports and tumor incidence data for all studies, with the exception of reserpine (43), were evaluated by the p53 AWG according to the recommendations of the Statistics Subcommittee and the ILSI ACT technical committee (36). A study was considered positive if a treatment-related increase in tumor incidence (combined benign and malignant) for a particular organ or tissue type was statistically-signi cant ( p 0.05) based on results of either a pair-wise comparison using Fisher’s Exact test (one-sided ) or a trend test (Cochran Armitage) where tumor incidence data was available for more than one dose level. Where an increase in tumor incidence for a particular organ or tissue did not reach statistical significance but the tumors were considered “rare” tumors (there were none in the control group ), then the presence of 2 tumors in that tissue at any one dose level was considered a positive response if the presence of other relevant biological evidence (non-neoplastic proliferative lesions ) in a substan-

31

tial number of mice at that dose level indicated a treatmentrelated effect of the test substance in the tissue. Tumors were classi ed as “rare” tumors if the incidence of spontaneous tumors by sex in the combined negative control groups for all reviewed ILSI p53 studies was less than 1%. If the tumor incidence for a rare tumor was a single tumor in one dose group with supporting biological evidence, or if the required additional biological evidence was lacking, the response was judged to be equivocal. RESULTS Overview of Status and Results for 26-Week ILSI ACT Studies in the p53 / Model In total, ILSI ACT study sponsors/participants completed 31 of the 32 studies initially planned for the 21 compounds chosen for evaluation in time for the data to be reviewed by the working group and reviewed herein. The 32 studies are summarized in Table 1, grouped by compound category. The dose levels evaluated and the route and frequency of administration are listed for the test compound as well as for the positive control compound, if any, tested in the sponsoring laboratory. Study results are listed as positive, negative or equivocal as judged by criteria for evaluation of responses in alternative models developed by the ILSI/HESI ACT steering committee (36). Where a quali cation on the overall validity of the study result for a particular chemical in the model is noted, the reason for the quali cation is described in the Results section. The principal reasons for a quali cation were 1) the absence of convincing data indicating a maximum tolerated dose (MTD) had been achieved in both sexes and 2) the absence of data indicating a sponsoring laboratory had either completed a study with a positive control compound that was judged to be positive or provided evidence to verify the genotype of the p53 / mice.

Incidence of Spontaneous Tumors The incidences of spontaneous tumors observed in p53 / mice in studies included in the ILSI ACT project are listed in Table 2. The overall spontaneous tumor incidence was low, 2.8% in males (n 283) and 6% in females (n 284) in studies without transponders and 8.0% in males and 11.3% in females in studies with transponders (n 150 ). The 3 most common spontaneous tumor types were subcutaneous sarcomas, lymphomas and osteosarcomas, as had been reported previously (29). The incidence of sarcomas is listed separately for studies in which microchip (transponder ) implants were used for animal identi cation and for studies in which other means of animal identi cation were used.

Performance of the Positive Controls p-Cresidine, an aromatic amine bladder carcinogen, was the most commonly used positive control compound in ILSI ACT studies in the p53 / model with 19 sponsors electing to use the recommended protocol of daily (7 days/wk ) gavage administration of 400 mg/kg of p-cresidine suspended in corn oil. The 400 mg/kg dose was established based on bladder tumor incidence data from a preliminary range- nding study conducted at Boehringer Ingelheim (41). The data for bladder tumor incidence in males and females in these 19 studies is summarized in Figure 1. A positive response was


32


Sponsor and contact

Schering M. Halleck

NIEHS S. Eastin Merck R. Storer

Novartis D. Lapadula

P zer R. Ochoa

Purdue Pharma & Springborn Laboratories R. Rush

Phenobarbital

Methapyrilene

1 500 1,000

1 50 250

SC 2x/week (mg/kg) Diet (ppm)

Gavage (mg/kg)

Diet (mg/kg/day)

Gavage (mg/kg)

Diet (mg/kg/day)

10 25 10 30 60 5 25 50

10 25 10 30 60 5 25 50

Gavage 5x/week (mg/kg) Diet (mg/kg/day)

Diet (%)

Ip 3x/week (mg/kg) Gavage (mg/kg)

8 16 32 0.3 1.5 100 200 350 0.14 0.7 1.4

8 16 32 0.3 1.5 100 200 350 0.14 0.7 1.4

Dose females

Gavage 2x/week (mg/kg)

Dose males

Positivey

Negative

20 100 200 31.3 62.5 93.8

20 100 200 31.3 62.5 93.l8

Negative

Negative

Positive (preliminary) positive

p-cresidine 400 mg/kg p-cresidine 3,750 ppm (diet) p-cresidine 400 mg/kg p-cresidine 400 mg/kg

Positive p-cresidine 400 mg/kg

Positive

Positive

Inadequate

Positive

Positive

na

Positive

Positive

na

Positive

Positive

na

Inadequate

Positive control

Benzene 100 mg/kg (7x/wk)

p-cresidine 3,750 ppm (diet) p-cresidine 400 mg/kg

None

p-cresidine 400 mg/kg


Positive Positive

None



Benzene 100 mg/kg (5x/wk) None

Negative

Negative

Negative

Positive

Positive

Positive controlb

heterozygous mouse model.

p53 / AWG review decision

/

0.1 0.1 Negative 0.5 0.5 2.0 2.0 17b -Estradiol Novartis Gavage 0.5 0.5 Equivocalyy (mg/kg) I. Nichols 2.0 2.0 5.0 5.0 Nongenotoxic rodent carcinogens: Compounds putatively noncarcinogenic in humans based on epidemiologic evidence Clo brate Merck Gavage 50 50 Negative (mg/kg) C. Boussiquet-LeRoux 125 200 250 400 Clo brate Glaxo Gavage 25 25 Negative (mg/kg) J. Allen 75 75 100 125 Phenobarbital Searle Diet 500 500 Negative (ppm) J. Sagartz 1,000 1,000

R.W. Johnson J. Oldham

17b -Estradiol

Human carcinogens: Hormones DES NIEHS S. Eastin DES Searle J. Sagartz

Cyclosporin A

Human carcinogens: Immunosuppressives Cyclosporin A NIEHS S. Eastin Cyclosporin A SKB L. Tierney

Phenacetin

Phenacetin

Melphalan

Human carcinogens: Genotoxic Cyclophosphamide Eli Lilly G. Long

Chemical

Route and frequency of administration (dose units)a

TABLE 1.—Results of 26-week ILSI ACT studies in the p53

(Continued on next page)

See “Results” for quali cations







p53 / AWG comments & quali cations


33

Aventis B. McCullough

Dieldrin

200 800 1,600 0.0005 0.001

200 800 1,600 0.0005 0.001

Dose females

p-cresidine 2,500 and 5,000 ppm (diet) p-cresidine 400 mg/kg

(Negative)

Positive Negative positive

Positive

Positive Positive

Positive Positive

p-cresidine 400 mg/kg p-cresidine 400 mg/kg p-cresidine 400 mg/kg

p-cresidine 400 mg/kg p-cresidine 400 mg/kg p-cresidine 400 mg/kg p-cresidine 400 mg/kg p-cresidine 400 mg/kg

—kidney papilloma.

Positive

Positive

Positive

Negative

Positive control


None


Positive controlb

Negative

p53 / AWG review decision

heterozygous mouse model. (Continued)

Dose frequency is 7 x/week (daily) unless otherwise noted. All p-cresidine studies by oral gavage in corn oil unless otherwise noted. —lymphoma; —lymphoma, sarcoma; y—pituitary adenoma, testicular tumors; yy—pituitary adenoma; —liver adenoma and carcinoma;

a

b

Diet (ppm)

Diet (%)

Diet (mg/kg/day)

Dose males

/

1 1 Negative 5 5 10 10 Nongenotoxic rodent carcinogens: Compounds putatively noncarcinogenic in humans based on mechanisic evidence Chloroform Dow Gavage 5x/week 28 24 Equivocal (mg/kg) B. Gollapudi 90 90 140 240 Chlorpromazine Dupont Gavage 2.5 2.5 Negative (mg/kg) D. Delker 5.0 5.0 10.0 10.0 Chlorpromazine Schering Gavage 2.5 2.5 Negative (mg/kg) M. Halleck 5.0 5.0 10.0 10.0 DEHP Dupont Diet 500 500 Equivocal (ppm) D. Delker 3,000 3,000 6,000 6,000 Haloperidol R.W. Johnson Gavage 0.5 0.5 Negative (mg/kg) J. Oldham 2.0 2.0 10.0 10.0 Haloperidol Novartis D. Lapadula Metaproterenol Eli Lilly Diet 0.06 0.05 Negative (%) G. Long 0.26 0.18 0.9 0.69 Mertaproterenol Huntingdon Diet 0.025 0.025 Negative (%) P. Ryle 0.5 0.5 1.0 1.0 1.5 1.5 Sulfamethoxazole P zer Diet 50 50 Negative (mg/kg/day) R. Ochoa 250 250 1,000 1,000 WY-14643 Sano Diet 25 25 Negative (ppm) T. Gray 50 50 100 100 Nongenotoxic, Noncarcinogens Ampicillin Bayer Gavage 350 350 Negative (mg/kg) M. Rinke 1,000 1,000 3,000 3,000 Mannitol Rorer Diet 2.5 2.5 Negative (%) B. McCullough 5.0 5.0 10.0 10.0 Sul soxazole Eli Lilly Diet 0.05 0.05 Negative (%) G. Long 0.5 0.5 2.0 2.0

Reserpine

Synthelabo Recherche R. Forster NIEHS J. French

Sponsor and contact

Methapyrilene

Chemical

Route and frequency of administration (dose units)a

TABLE 1.—Results of 26-week ILSI ACT studies in the p53



Study complete but not reported






p53 / AWG comments & quali cations

34

TOXICOLOGIC PATHOLOGY

STORER ET AL

TABLE 2.—ILSI ACT studies: Spontaneous tumor incidence in control p53 / mice.

TABLE 3.—ILSI ACT p53 in gavage studies (Merck).

/

studies: Positive results for p-cresidine control

% Incidence (# of mice)

Tumor type

Males

Females

Lymphoma Subcutaneus sarcomas —without transponders —with transponders Osteosarcoma Lung, adenoma Lung, rhabdomyosarcoma Liver, adenoma Brain, Meningeal sarcoma Olfactory neuroblastoma Granulocytic leukemia Pancreas, carcinoma Adrenal cortical adenoma Pheochromocytoma Prostate, sarcoma Prostate, leiomyosarcoma

1.7 (433) 1.0 (433) 0.4 (283) 2.0 (150) 0.5 (433) 0.0 (433) 0.0 (433) 0.2 (433) 0.0 (433) 0.0 (433) 0.2 (433) 0.2 (433) 0.2 (433) 0.2 (433) 0.2 (433) 0.2 (433)

2.9 (434) 3.3 (434) 1.4 (284) 6.7 (150) 0.7 (434) 0.5 (434) 0.2 (434) 0.0 (434) 0.2 (434) 0.2 (434) 0.0 (434) 0.0 (434) 0.0 (434) 0.0 (434) na na

obtained in 18 of the 19 studies based on either a statisticallysigni cant increase in bladder tumor incidence ( 4/15 ), or an increase ( 2/15 ) judged to be signi cant by the rare tumor criteria (36), in at least one sex. As expected, males were more sensitive than females with 18/19 studies in males showing a positive result as compared to 15/19 studies in females with positive results. Representative data for 2 studies that gave a clear, statistically-signi cant increase in bladder tumor response, and for one study which gave a positive tumor response in one sex only as judged by the rare tumor criteria, are shown in Tables 3 and 4 respectively. In the one negative study where no bladder neoplasms were found (see Table 5), there was a signi cant increase in nonneoplastic proliferative bladder lesions con rming that there was exposure to the test article. One additional study was initiated with the 400 mg/kg gavage dose of p-cresidine but the dose in this study was

p-Cresidine 400 mg/kg

Vehicle control

Bladder lesion incidencea

Merck (USA) study Hyperplasia Dysplasia Transitional cell carcinoma Merck (France) study Hyperplasia Dysplasia Transitional cell carcinoma Papilloma Squamous cell carcinoma Sarcoma

M

F

M

F

0 0 0

0 0 0

15 0 8b

15 1 4b

0 0 0 0 0 0

0 0 0 0 0 0

15 7 7b 4b 5b 1

15 1 4b 2 2 0

a Mice

with lesions; 15 mice/sex/group. statistically signi cant increase in tumor incidence as compared to the negative control, p 0.05. b

reduced to 50 mg/kg in study week 2 due to signi cant mortality and there were no bladder tumors found at necropsy. The unexpected mortality in this study was associated with gavage trauma attributed to failure to wipe the gavage needle and animals struggling due to the bitter taste of the test agent. Finally, one study (39) used dietary administration of p-cresidine at an average concentration of 3750 ppm and obtained a strong positive bladder tumor response in p53 / (see Table 6) relative to the minimal response in wild-type mice (1/15 mice with tumors). These data con rm previously reported ndings (43) that the p53 heterozygote is more sensitive to bladder tumor induction by p-cresidine than the wild type. Benzene, a known human and rodent carcinogen, was also used as the positive control agent in two studies at the dose level of 100 mg/kg (gavage in corn oil) recommended in the ILSI ACT protocol. At this dose level, with dosing either

FIGURE 1.—Bladder tumor incidence for p-cresidine positive controls (Gavage/400 mkd). Downloaded from tpx.sagepub.com by guest on July 15, 2011

Vol. 29(Suppl. ), 2001

p53

/


TABLE 4.—ILSI ACT p53 / studies: Positive results for p-cresidine positive control in gavage study (Sano ). p-Cresidine 400 mg/kg

Vehicle control


M

F

M

F

In ammation Urothelial hyperplasia Squamous metaplasia Transitional cell carcinoma

0 0 0 0

1 0 0 0

3 15 8 3

11 14 4 1

a

Mice with lesions; 15 mice/sex/group.

5 times per week (Table 7) or daily (Table 8), there were no statistically-signi cant increases in tumor incidence at any site. The combined increases in thymic hyperplasia and thymic lymphoma suggested a borderline effect in a suspected target tissue but the incidence of tumors was not statistically signi cant in either study. Since lymphoma is a common spontaneous tumor in this model, statistical signi cance is required in order to conclude that an increased incidence of tumors at the site constitutes a positive response to the test agent. Benzene at a higher dose of 200 mg/kg by gavage either 5 times per week (15) or daily (23) has been reported to give positive results in this model and it appears that the recommended dose of 100 mg/kg was too low based on the available data (15). In addition, benzene gave a robust thymic lymphoma response in inhalation studies sponsored by CIIT to examine the effect of dose rate for benzene (37). Data from this study showed a strong response at 9 months based on the endpoint of the cumulative mortality of mice with thymic masses. However, the outcome of this benzene inhalation study at study week 26 is unknown since there was no scheduled necropsy to determine tumor incidence at that time point. Therefore, it would not be appropriate to use these CIIT results to suggest that a 26-week study is inadequate for benzene, and possibly other genotoxic carcinogens, in the p53 / model. In summary, the overall performance of p-cresidine as a positive control agent for the p53 / model was judged to be adequate based on qualitatively consistent positive responses for bladder tumorigenesis across multiple laboratories in 95% (18/19 ) of the studies. The less robust responses in the two positive studies for which the increase in bladder tumor incidence was not statistically signi cant and the lack of a tumor response in one additional study were regarded as unsatisfactory for an agent being employed as a positive control. In the latter case of the negative result for the p-cresidine positive control, con rmation of the TABLE 5.—ILSI ACT p53 / studies: Negative results for p-cresidine positive control in gavage study (Dupont).

p53 p53

Vehicle control

/

De cient Wild-type

/

/

studies: Positive results for p-cresidine control

Vehicle control

p-Cresidine 3,750 ppm

M

F

M

F

0 0

10b

11b 1

0 0

1

with tumors (transitional cell carcinomas); 15 mice/sex/group. Statistically signi cant increase in tumor incidence as compared to the negative control, p 0.05. a Mice

b

genotype of the animals was not provided. This information is then critical to, but not necessarily suf cient for, an assessment of the validity of the study results (equivocal ) for the test agent (DEHP). Possible reasons why a negative result was obtained in this study other than chance and normal biologic variability were not readily apparent. It was noted that the extent of reduction in body weight gain for the p-cresidine groups in the negative study was less than that seen in other studies. There did not appear to be any correlation between bladder tumor incidence and age at initiation when using average body weight at study initiation as a surrogate parameter for age. There was speculation that the manner in which the p-cresidine oral dose form suspension in corn oil is prepared (mortar and pestle vs motorized homogenization ) might be a factor in determining the rate and extent of absorption of the test article after gavage administration but additional toxicokinetic studies would be required to investigate this possibility. Finally, the performance of benzene administered daily or 5x/week at 100 mg/kg by oral gavage in corn oil was clearly unsatisfactory as a positive control for a 26-week study and publication of data from additional studies would be needed to determine if increasing the dose to 200 mg/kg gives a reproducible tumor response at 26-weeks. Results for the 21 ILSI ACT Test Compounds The overall results (positive/equivocal/negative ) for each of the 31 studies (out of a total of 32 studies with the 21 ILSI ACT compounds ) that were completed and evaluated by the p53 AWG, are listed in Table 1. One study (reserpine/NTP) was completed and published (43) prior to the start of the ILSI ACT studies but detailed study reports were not available for review by the AWG. These results are discussed below by compound category with an emphasis on the overall conclusion for each compound where more than one study was performed. Two studies (haloperidol/Novartis) TABLE 7.—ILSI ACT p53 / studies: Negative results for benzene positive control in gavage study (Eli Lilly). Vehicle 0

M

F

M

F

In ammation Urothelial hyperplasia Atypia Apoptosis Neoplasms

0 0 0 0 0

0 0 0 0 0

12 15 6 13 0

13 11 5 10 0



p-Cresidine 400 mg/kg


a

TABLE 6.—ILSI ACT p53 in diet study (Searle) (39).

35

Benzene 100 mg/kgb

Lesion incidencea

M

F

M

F

Thymic lymphoid hyperplasia Lymphoma Subcutaneous sarcoma (no implants) Osteosarcoma

0 0 1

0 1 0

3 3 0

2 1 0

0

0

1

0

a

Mice with lesions; 15 mice/sex/group. b Dosing 5x/week.


36

TABLE 8.—ILSI ACT p53 / studies: Negative results for benzene positive control in gavage study (Novartis). Vehicle 0

Benzene 100 mg/kgb

Lesion incidencea

M

F

M

F

Atypical thymic hyperplasia Thymic lymphoma Subcutaneous sarcoma (implants) Osteosarcoma Oral cavity, squamous cell carcinoma

0 0 0

0 0 1

4 1 1

0 0 1

0 0

1 0

0 1

0 0

a b


STORER ET AL

Mice with lesions; 15 mice/sex/group. Dosing 7x/week.

was completed but has not as yet been reviewed by the p53 AWG. Human Carcinogens: Genotoxic Cyclophosphamide (positive): Cyclophosphamide was studied in p53 / mice in a single study using twice weekly gavage administration of doses of 8, 16, and 32 mg/kg in water. Control and high-dose groups for wild-type p53 / mice were included in the study. Benzene administered 5x/week at 100 mg/kg in corn oil served as the positive control in the sponsoring laboratory (Eli Lilly). The incidence of neoplasms in this study is shown in Table 9. Cyclophosphamide produced a statistically signi cant increase in the incidence of lymphomas with 4/15 male and 8/15 female, high-dose p53 / mice with tumors versus 0/15 and 1/15 for the controls respectively. In wild-type mice treated with cyclophosphamide 3/15 males and 0/15 females had lymphomas with none in controls. The positive control benzene was negative but, as previously discussed, the dose selected appears to have been too low. The study was accepted without quali cation by the p53 AWG as showing a valid positive result for the test compound. Melphalan (positive ): Melphalan was studied in p53 / mice in a single study using 3x/week ip administration of doses of 0.3 and 1.5 mg/kg in propylene glycol. Control and high-dose groups for wild-type p53 / mice were not included in the study. Melphalan produced a statistically signi cant increase in the incidence of lymphomas and sarcomas in male p53 / mice with 9/15 high-dose and 3/15 lowTABLE 9.—ILSI ACT p53

/

dose animals respectively with tumors versus 1/15 and 0/15 for the controls, respectively (14). Positive results had been reported by the sponsoring organization (NTP/NIEHS) for studies with the positive control compounds p-cresidine and benzene (15, 43) but these studies were not concurrent with this study (14). The study was accepted by the p53 AWG as showing a valid positive result in male mice with the quali cation that the negative (but suggestive ) results for female mice were inconclusive due to the high number of early death animals and the number of tissues that were as a result, missing or not readable due to autolysis. Phenacetin (negative ): Phenacetin was studied in p53 / mice in two separate studies using either daily gavage administration (Merck) with doses of 100, 200 and 350 mg/kg suspended in 0.5% methylcellulose or dietary administration at levels of 0.14, 0.7, and 1.4% in the diet (Schering Plough ). Control and high-dose groups for wild-type p53 / mice were included in both studies. p-Cresidine at 400 mg/kg in corn oil served as the positive control for both studies. There were no increases in tumor incidence for phenacetintreated mice (p53 / or wild-type) at any site in either study that were judged to be treatment related. The incidence of neoplasms in the Merck phenacetin study is shown in Table 10 as an example of results from a negative study. The positive control p-cresidine was positive in both studies. Both studies were accepted without quali cation by the p53 AWG as showing a valid negative result for the test compound at maximum tolerated dose levels.

Human Carcinogens: Immunosuppressives Cyclosporin A (positive): Cyclosporin A was studied in p53 / mice in three studies, two studies with dietary administration at dose levels of 10, 30, and 60 mg/kg body weight (Smith Kline Beecham) or 5, 25, and 50 mg/kg body weight (Novartis) and one older study using gavage administration 5x/week in olive oil:ethanol at doses of 5 and 25 mg/kg (NIEHS/NTP) (14). Control and high-dose groups for wild-type p53 / mice were included for comparison in the diet studies but not in the earlier gavage study. p-Cresidine at 400 mg/kg in corn oil served as the positive control in both diet studies whereas the NIEHS/NTP gavage study had p-cresidine and benzene (15, 43) as positive controls though they were not concurrent with the cyclosporin study (14). The incidences of neoplasms in the two diet studies are shown in

studies: Incidence of neoplasms in cyclophosphamide study (Eli Lilly). Cyclophosphamide (mg/kg)b

Vehiclea 0

8

16

32

Lesion incidencea

M

F

M

F

M

F

M

F

Lymphoma (wild-type) Lymphoma (p53 / ) Granulocytic leukemia Subcutaneus sarcoma (no implants) Thymus, histiocytic sarcoma Fibrosarcoma, meninges Endometrial stromal polyp

0 0 0 1

0 1 0 0

— 3 0 0

— 1 0 0

— 0 0 1

— 0 0 0

3 4c 0 2

0 8c 1 2

0 0 0

1 0 0

0 0 0

0 0 0

0 0 0

0 0 1

0 0 0

0 1 0

a Mice

with lesions; 15 mice/sex/group. Twice weekly gavage administration. c Statistically signi cant increase in tumor incidence as compared to the negative control, p b

0.05.


Vol. 29(Suppl. ), 2001

/

p53

: OVERVIEW OF AVAILABLE DATA /

TABLE 10.—ILSI ACT p53

37

studies: Incidence of neoplasms in phenacetin study (Merck). Phenacetin (mg/kg)

Vehiclea 0

100

200

350

Lesion incidencea

M

F

M

F

M

F

M

F

Benign meningioma Thymic lymphoma Subcutaneous sarcoma (implants) Adrenal cortex adenoma

0 0 1

0 2 2

0 0 1

0 0 1

0 0 2

0 0 1

1 0 1

0 0 4

0

0

0

0

0

0

1

0

a Mice

with lesions; 15 mice/sex/group.

Tables 11 and 12. Cyclosporin A produced a statistically signi cant increase in the incidence of lymphomas at the high dose in female mice in both studies and, in high-dose males, only in the Novartis study. Wild-type mice showed a similar pattern of results suggesting that the lymphoma response to cyclosporin A was independent of the p53 genotype. Negative results were obtained in the earlier NIEHS/NTP gavage study which was conducted at lower levels (14). The positive control p-cresidine gave a statistically signi cant bladder tumor response in the Novartis study and a non statistically signi cant but still positive response in the SKB study as judged by the rare tumor criteria (36). The results for the two gavage studies were accepted without quali cation by the p53 AWG as showing similar, valid positive results for the test compound in p53 / mice. The negative results in the earlier gavage study (non-ILSI protocol ) were considered to have been overridden by the positive results in the two, more recent, ILSI-protocol studies at higher dose levels.

Human Carcinogens: Hormones Diethylstilbestrol (positive): Diethylstilbestrol (DES) was studied in p53 / mice in two studies, one using dietary administration at dietary levels of 50 and 250 ppm for males and 500 and 1,000 ppm for females (Searle) and one older study using subcutaneous (sc) administration 2x/week in emulphor:ethanol:water at a dose of 1 mg/kg (NIEHS/NTP) (14). Control and high-dose groups for wild type p53 / mice were included for comparison in the diet study but not in the earlier sc study. p-Cresidine in the diet at 3750 ppm served as the positive control in the diet study while the NIEHS/NTP gavage study had p-cresidine and benzene (15, 43) as positive controls though they were not concurrent with the DES study (14). The incidences of pituitary lesions in females


/

and testicular lesions in males in the diet study are shown in Tables 13 and 14 respectively. Diethylstilbestrol produced increases in the incidence of nodular hyperplasias and adenomas in the pituitary in the low- and high-dose female mice and an increase in interstitial cell hyperplasias and tumors in the testes in males at the high dose. For both tumor types, the increases were not statistically signi cant but were nevertheless judged to be evidence of a positive response using the rare tumor criteria (36). Wild-type mice showed a similar pattern of results for the nonneoplastic (hyperplastic ) lesions but no tumors were found suggesting that the tumor response to DES in these tissues may be accelerated in the p53 / background relative to wild-type. Negative results were obtained in the earlier NIEHS/NTP sc study that was conducted at a low dose of 1 mg/kg twice weekly. No nonneoplastic proliferative lesions were seen in the pituitary or testes in this study (14). The positive control for the diet study, p-cresidine (diet; 3,750 ppm ), gave a statistically signi cant bladder tumor response in both males and female that was accelerated relative to the response in wild-type mice (see Table 6). The results for the diet study were accepted without quali cation by the p53 AWG as showing a valid positive result for the test compound in p53 / mice. The negative results in the earlier sc study (non-ILSI protocol ) were considered to have been overridden by the positive results in the more recent, ILSI-protocol studies at higher dose levels. 17-b -Estradiol (equivocal ): 17-b -Estradiol was studied in p53 / mice in two gavage studies with daily dosing at dose levels of 0.1, 0.5, and 2.0 mg/kg in 1% v/v 95% ethanol/0.8% hydroxypropylmethylcellulose (RW Johnson ) and at 0.5, 2.0, and 5.0 mg/kg in 1% v/v ethanol/0.5% hydroxypropylmethylcellulose (Novartis). Control and highdose groups for wild-type p53 / mice were included for

studies: Incidence of neoplasms in cyclosporin A study (SKB). Cyclosporin A (mg/kg/day)b

Controla 0

10

30

60

Lesion incidencea

M

F

M

F

M

F

M

F

Lymphoma (wild-type) Lymphoma (p53 / ) Osteoarcoma Fibrosarcoma subcutaneous) (implants) Pheochromocytoma Astrocytoma

1 1 0 0

0 0 0 0

— 0 0 0

— 0 0 0

— 0 0 0

— 1 0 0

0 1 1 0

3 4c 0 1

1 0

0 0

0 0

0 1

0 0

0 0

0 0

0 0

a Mice

with lesions; 15 mice/sex/group. Dietary administration. c Statistically signi cant increase in tumor incidence as compared to the negative control, p b

0.05.


38


STORER ET AL TABLE 12.—ILSI ACT p53

/

studies: Incidence of neoplasms in cyclosporin A study (Novartis). Cyclosporin A (mg/kg/day)b

Controla 0

5

15

50

Lesion incidencea

M

F

M

F

M

F

M

F

Lymphoma (wild-type) Lymphoma (p53 / ) Osteosarcoma Fibrosarcoma (Subcutaneous) (implants) Olfactory neuroblastoma Lung adenoma

1 2 0 0

0 1 1 2

— 0 0 0

— 0 0 2

— 0 0 0

— 0 1 0

5 8c 0 2

9c 8c 0 1

0 0

0 1

0 0

0 0

0 0

0 0

0 0

1 1

a Mice

b c

with lesions; 15 mice/sex/group. Dietary administration. Statistically signi cant increase in tumor incidence as compared to the negative control, p < 0.05.

comparison in both studies. p-Cresidine given by oral gavage at 400 mg/kg in corn oil served as the positive control in the RW Johnson study and benzene at 100 mg/kg by gavage in corn oil 7x/week served as the positive control in the Novartis study. There were no treatment-related neoplastic effects in any tissue in the RW Johnson study and there were no treatment related nonneoplastic (hyperplastic ) lesions in the pituitary. The top dose in the RW Johnson study was the maximum feasible dose based on the apparent limit of solubility of 17-b -estradiol in 95% ethanol and a limit of 1% ethanol in the nal dose form. A 2.5-fold higher dose (5.0 mg/kg) was achieved in the Novartis study, and in this study there were increased incidences of pituitary hyperplasias and a single adenoma in the female high-dose mice (see Table 15). The 1/15 incidence of adenoma though obviously not statistically signi cant was, when evaluated together with the evidence of a hyperplastic response, nevertheless judged to constitute evidence of an equivocal response using the rare tumor criteria (36). High-dose, wildtype mice in the Novartis study showed a similar pattern, but lower incidence, of nonneoplastic (hyperplastic ) pituitary lesions but no tumors were found suggesting that the equivocal tumor response to 17-b -estradiol in the pituitary in females may, as was true for DES, also be accelerated in the p53 / background relative to wild-type. The positive control for the RW Johnson study, p-cresidine (gavage; 400 mg/kg) gave a statistically signi cant bladder tumor response in both males and females while the benzene positive control in the Novartis study was negative (see Table 8). The results for the Novartis study were accepted by the p53 TABLE 13.—ILSI ACT p53

Pituitary lesion incidencea

Par Distalis Diffuse hyperplasia Nodular hyperplasia Adenoma Pars intermedia Diffuse hyperplasia Adenoma

/

AWG as showing an equivocal result for the test compound in p53 / mice with the quali cation that con rmation of genotype was not provided. The negative results in the RW Johnson study with the 2.0 mg/kg high-dose level were considered to have been overridden by the equivocal results in Novartis study that achieved a 2.5-fold higher top dose levels.

Nongenotoxic Rodent Carcinogens: Compounds Putatively Noncarcinogenic in Humans Based on Epidemiologic Evidence Clo brate (negative ): Clo brate was studied in p53 / mice in two separate gavage studies with doses of 50, 125, and 250 mg/kg (males) or 50, 200, and 400 mg/kg (females) of clo brate in 0.5% methylcellulose (Merck) or doses of 25, 75, 100 mg/kg (males) or 25, 75, 125 mg/kg (females) in 0.5% methylcellulose (Glaxo). The dose levels cited above for the Glaxo study were initiated on study day 14 and were reduced due to mortality from the initial dose levels used on study days 1 – 3 or 1 – 6 of 50, 250, and 400 mg/kg for males and 50, 200, and 500 mg/kg for females. Control and highdose groups for wild-type p53 / mice were included in both studies. p-Cresidine administered by gavage at 400 mg/kg in corn oil served as the positive control in both studies. There were no increases in tumor incidence in clo brate-treated mice (p53 / or wild-type) at any site in either study that were judged to be treatment related. The positive control, p-cresidine, produced statistically signi cant increases in bladder tumor incidence in both sexes in both studies. The

studies: Incidence of pituitary lesions in female mice in diethylstilbestrol study (Searle). DES

DES

Cont 0 p53 /

500 ppm p53 /

1,000 ppm p53 /

Cont 0 Wild-type

500 ppm Wild-type

1,000 ppm Wild-type

0 0 0

15 8 2

14 5 2

0 0 0

15 1 0

14b 2b 0b

0 0

14 1

15 1

0 0

13c 0c

12d 0

a

Mice with lesions; 15 female mice/group examined; unless otherwise noted. female mice/group examined. c 13 female mice/group examined. d 12 female mice/group examined. b 14


Vol. 29(Suppl. ), 2001

p53


/

/


39

studies: Incidence of testicular lesions in male mice in diethylstilbestrol study (Searle). DES

DES

Testes lesion incidencea

Cont 0 p53 /

500 ppm p53 /

1,000 ppm p53 /

Cont 0 Wild-type

500 ppm Wild-type

1,000 ppm Wild-type

Interstitial cell hyperplasia Interstitial cell tumor

0 0

4 0

4 2

0 0

2b 0

3 0

a b

Mice with lesions; 15 male mice/group examined unless otherwise noted. 13 female mice/group examined.

Merck gavage study was accepted without quali cation by the p53 AWG as showing a valid negative result for the test compound at maximum tolerated dose levels. The Glaxo study was accepted by the p53 AWG as showing a negative result for the test compound at the dose levels administered with the quali cation that a maximum tolerated dose may not have been achieved. Phenobarbital (negative ): Phenobarbital was studied in p53 / mice in two separate studies using dietary administration with daily dose levels of 20, 100, and 200 mg/kg body weight (P zer) or levels of 500 and 1,000 ppm continuously in the diet (Searle) (39). Control and high-dose groups for wildtype p53 / mice were included in the Searle study but not in the P zer study. p-Cresidine at 400 mg/kg in corn oil served as the positive control for the P zer study and p-cresidine in the diet at 3,750 ppm served as the positive control for the Searle study. There were no increases in tumor incidence for phenobarbital-treated mice (p53 / or wild-type ) at any site in either study that were judged to be treatment related. The positive control p-cresidine induced statistically signi cant increases in the incidence of bladder tumors in both sexes in both studies (see Table 6 for results from the Searle study). Both studies were accepted without quali cation by the p53 AWG as showing a valid negative result for the test compound at maximum feasible dose levels (based on sedation).

Methapyrilene (negative ): Methapyrilene was studied in p53 / mice in two separate studies using either daily gavage administration (Springborn Labs/Purdue Pharma) with doses of 31.3, 62.5, and 93.8 mg/kg of methapyrilene hydrochloride in water or dietary administration at target dose levels of 200, 800, and 1,600 mg/kg in the diet (Synthelabo Recherche). Control and high-dose groups for wild-type p53 / mice were included in both studies. p-cresidine administered by gavage at 50 mg/kg (CIT/Synthelabo study) and 400 mg/kg (Springborn Labs/Purdue Pharma) in corn oil served as the positive controls. There were no increases in tumor incidence in methapyrilene-treated mice (p53 / or wild-type) at any TABLE 15.—ILSI ACT p53

/

site in either study that were judged to be treatment related. The positive control p-cresidine was positive in both sexes in the Springborn Labs/Purdue Pharma study at 400 mg/kg but negative in the CIT/Synthelabo Recherche study (see discussion under “Performance of the Positive Controls” ). The Springborn Labs/Purdue Pharma gavage study was accepted without quali cation by the p53 AWG as showing a valid negative result for the test compound at maximum tolerated dose levels. The CIT/Synthelabo Recherche diet study was accepted by the p53 AWG as showing a valid negative result for the test compound at maximum tolerated dose levels with the quali cation that genotyping would be required to con rm the genotype of the p53 / mice given the failure of the positive control. Reserpine: Reserpine was studied in p53 / mice in a study sponsored by NIEHS using dietary administration at levels of 0.0005% and 0.001% continuously in the feed (43). The high dose level of 0.001% in the feed was reduced to 0.0005% after 4 weeks of treatment due to excessive weight loss and mortality. Control and high-dose groups for wildtype p53 / mice were included in the study. This study was completed prior to the initiation the ILSI ACT project and, as such, was not performed according to the ILSI protocol. The number of mice per dose group was 10 or less and the age at initiation (15 – 18 weeks of age) and the study duration (24 weeks plus 4 – 6 weeks post treatment holding interval) also differed substantially from those used in studies conducted under the ILSI protocol. p-Cresidine administered in the diet at 0.25 and 0.5% served as the positive controls in a study reported concurrently (43). There were no increases in gross lesions or tumor incidence in target organs in reserpinetreated mice (p53 / or wild type) that were judged to be treatment related. The positive control, p-cresidine, was positive in both sexes at the high dose level and in males only at the low dose level. The NIEHS reserpine study was not reviewed by the p53 AWG and the negative result reported for the test compound (43) would not appear, based on the protocol differences from the ILSI ACT protocol, to be acceptable

studies: Incidence of proliferative lesions in pituitary in female mice in 17b -estradiol study (Novartis). 17b -E

Pituitary lesion incidencea

Cont 0 p53 /

0.5 mg/kg p53 /

2 mg/kg p53 /

5 mg/kg p53 /

Cont 0 Wild-type

17b -E 5 mg/kg Wild-type

Pars Distalis Hyperplasia Focal hyperplasia Adenoma

0 0 0

0 0 0

0 0 0

12 1 1

0 0 0

5 0 0

a

Mice with lesions; 15 female mice/group examined unless otherwise noted.


40


STORER ET AL

as a valid negative result without the important quali cations noted above. Dieldrin (negative ): Dieldrin was studied in p53 / mice in a single study using countinuous dietary administration with dose levels in the feed of 1, 5, and 10 ppm. Control and high-dose groups for wild-type p53 / mice were included in the study design. p-Cresidine at 400 mg/kg in corn oil served as the positive control. There were dose-related increases in mortality in the high dose group for both sexes with only 6/15 males and 5/15 females surviving to study termination indicating that the top dose exceeded a maximum-tolerated dose. There were no increase in tumor incidence for dieldrin-treated p53 / mice at any site that were judged to be treatment related. Nonneoplastic, treatment-related effects were seen in liver in the mid- and high-dose levels that consisted of minimal to severe centrilobular hypertrophy. The positive control, p-cresidine, produced a statistically signi cant bladder tumor response in both sexes. The study was accepted as showing a valid negative result for the test compound at maximum tolerated dose levels with the quali cation that all tissues were not examined microscopically at the highest dose level with adequate survival to 26 weeks (5 ppm ). However, sections from the liver (the principal target organ for dieldrin toxicity, and carcinogenicity in long-term studies ), were examined for all surviving mice in the 1 and 5 ppm dose groups.

Nongenotoxic Rodent Carcinogens: Compounds Putatively Noncarcinogenic in Humans Based on Mechanistic Evidence Chloroform (equivocal ): Chloroform was studied in p53 / mice in a single study using 5x/week gavage administration in corn oil of doses of 28, 90, and 140 mg/kg for males and 24, 90, and 240 mg/kg for females. Control and high-dose groups for wild-type p53 / mice were included in the study. A positive control was not run in the sponsoring laboratory (Dow ). The incidence of neoplasms in this study is shown in Table 16. Chloroform did not produce a statistically signi cant increase in the incidence of tumors at any site. However, 2/15 high dose female p53 / mice had liver tumors as compared to 0/15 in control females (and 1/15 in control males). Since there was no other evidence of nonneoplastic, proliferative lesions in the liver, this result was judged to be equivocal evidence of a tumorigenic effect under the rare tumor criteria (36). There were no liver tumors found in the control or high-dose, wild-type mice. The p53 AWG concluded that the high dose for the females (240 mg/kg) TABLE 16.—ILSI ACT p53

/

was a maximum tolerated dose but that the effects produced by the high dose in males (140 mg/kg) (decreased kidney weight, kidney tubular degeneration, 8.6% decrease in terminal body weight, liver cell hypertrophy and vacuolization ) were marginal in terms of justi cation for the high dose level. As a result, the p53 AWG accepted the equivocal results of the study as valid with the quali cations that 1) the males may have tolerated a higher dose level and 2) that the genotype of the p53 / mice should be con rmed because no positive control was run in the sponsor’s laboratory. Chlorpromazine (negative ): Chlorpromazine was studied in p53 / mice in two separate studies (Dupont and Schering Plough ) using daily gavage administration of doses of 2.5, 5.0, and 10 mg/kg dissolved in water. Control and high-dose groups for wild-type p53 / mice were included in both studies. p-Cresidine at 400 mg/kg in corn oil served as the positive control for both studies. There were no increases in tumor incidence for chlorpromazine-treated mice (p53 / or wild-type) at any site in either study that were judged to be treatment related. The positive control, p-cresidine, produced a positive bladder tumor response in both sexes in each study. Both studies were accepted without quali cation by the p53 AWG as showing valid negative results for the test compound at maximum tolerated dose levels. Diethylhexylphthalate (equivocal ): Diethylhexylphthalate (DEHP) was studied in p53 / mice in a single study using continuous dietary administration in feed of 500, 3,000, and 6,000 ppm (Dupont ). Control and high-dose groups for wild-type p53 / mice were included in the study. p-Cresidine administered by gavage at 400 mg/kg in corn oil served as the positive control. DEHP did not produce a statistically-signi cant increase in the incidence of tumors at any site (see Table 17). However, 1/15 high-dose female p53 / mice had a renal transitional cell papilloma as compared to 0/15 in control mice. Because there was evidence of nephrotoxicity (hydronephrosis, atrophy, in ammation, and pelvic dilation) and nonneoplastic, proliferative lesions (transitional cell hyperplasia secondary to the hydronephrosis ) in the kidney (in both males and females), this result was judged to be equivocal evidence of a tumorigenic effect under the rare tumor criteria (36). There was also evidence of nephrotoxicity and nonneoplastic proliferative lesions in the kidneys in the high-dose, wild-type male mice but not in females. The positive control was negative with no tumors observed in the bladder in either sex (see Table 5). The sponsors commented that this result was in contrast to the results of another

studies: Incidence of proliferative lesions in chloroform study (Dow). Chloroform (mg/kg)

Control

Lesion incidencea

Liver, adenoma Liver, carcinoma Lung, hyperplasia, vsl Lung, adenoma Skin (subcutaneous), brosarcoma (implants)

0 M

0 F

28 M

24 F

90 M

90 F

140 M

240 F

1 0 0 0 0

0 0 0 1 0

0 0 0 0 0

0 0 0 0 0

0 0 0 0 0

0 0 1 1 0

0 0 0 1 0

1b 1 1 0 1

a b

Mice with lesions; 15 mice/sex/group. Two adenomas in one mouse. Downloaded from tpx.sagepub.com by guest on July 15, 2011

Vol. 29(Suppl. ), 2001

p53


/

/


41

studies: Incidence of proliferative lesions in kidney in female mice in DEHP study (Dupont). DEHPb

Kidney Lesion incidencea

Cont 0 p53 /

500 ppm p53 /

3,000 ppm p53 /

6,000 ppm p53 /

Cont 0 Wild-type

DEHPb 6,000 ppm Wild-type

Hyperplasia, ureter Hyperplasia, transitional cell Papilloma, transitional cell

0 0 0

— — —

0 0 0

1 3 1

0 0 0

0 0 0

a Mice b

with lesions;15 female mice/group. Dietary administration.

concurrent, p-cresidine positive control group (chlorpromazine study ) performed by the same laboratory using the same lot of p-cresidine where a strong bladder tumor response was observed. The p53 AWG accepted the equivocal results of the study as valid with the quali cation that the genotype of the p53 / mice should be con rmed because the positive control that was run with this study was negative.

Haloperidol (negative ): Haloperidol was studied in p53 / mice in two separate gavage studies with doses of 0.5, 2.0, and 10 mg/kg in water (RW Johnson ) and at undetermined doses in a separate study which is complete but not yet reported to the p53 AWG (Novartis). Control and high-dose groups for wild-type p53 / mice were included in both studies. p-cresidine administered by gavage at 400 mg/kg in corn oil served as the positive control in the RW Johnson study. There were no increases in tumor incidence in haloperidol-treated mice (p53 / or wild-type ) at any site in the RW Johnson study that were judged to be treatment related. The positive control, p-cresidine, produced statistically-signi cant increases in bladder tumor incidence in both sexes in the RW Johnson study. The RW Johnson gavage study was accepted without quali cation by the p53 AWG as showing a valid negative result for the test compound at maximum tolerated dose levels.

Metaproterenol (negative ): Metaproterenol was studied in p53 / mice in two separate studies (Eli Lilly and Huntingdon ) using continuous dietary administration at levels (for males/females) of 0.06/0.05, 0.26/0.18, and 0.92/0.69% in the diet (Eli Lilly) or 0.025, 0.5, 1.0, and 1.5% in the diet (Huntingdon ). Control and high-dose groups for wild-type p53 / mice were included in both studies. p-Cresidine at 400 mg/kg in corn oil served as the positive control in the Huntingdon study. No concurrent positive control was reported for the Eli Lilly study but the sponsoring laboratory had run benzene at 100 mg/kg by oral gavage as a positive control in another study in the same laboratory (see discussion for cyclophosphamide ). There were no increases in tumor incidence for metaproterenol-treated mice (p53 / or wild-type) at any site in either study that were judged to be treatment related. The tumor incidence data in the Eli Lilly study were unusual in that there was a high incidence of lymphoma (3/15 ) in the female vehicle control group (see Table 18). The positive control, p-cresidine, produced a positive bladder tumor response in both sexes in the Huntingdon study but benzene was negative in the Eli Lilly laboratory. The Huntingdon study was accepted without quali cation by the p53 AWG as showing valid negative results for the test compound at maximum tolerated dose levels. The Eli Lilly study was accepted as a valid negative with the quali cation

that genotyping would be required to con rm the p53 genotype of the mice.

/

Sulfamethoxazole (negative ): Sulfamethoxazole was studied in a single study (P zer) using continuous dietary administration in feed with intended doses of 50, 250, and 1,000 mg/kg/day. Control and high-dose groups for wildtype p53 / mice were not included in the study design. pCresidine at 400 mg/kg in corn oil served as the positive control that was conducted with the same lot of mice but in a separate laboratory from the sulfamethoxazole study. There were no increases in tumor incidence for sulfamethoxazole-treated p53 / mice at any site that were judged to be treatment related. The sponsor noted that the p53 / mice responded to sulfamethoxazole in a manner similar to that previously observed in other strains and species of rodents, with decreased T4 levels and a proliferative change in the thyroid (follicular cell hyperplasia ) presumably due to increased TSH levels (TSH levels were not measured). The positive control, pcresidine, produced a statistically signi cant bladder tumor response in both sexes. The study was accepted as showing a valid negative result for the test compound with the quali cation that a maximum tolerated dose level may not have been achieved in females where the mean body weights were only reduced to 93% of control at the high dose (84% for male high-dose animals).

Wy-14643 (negative ): Wy-14643 was studied in a single study (Sano ) using continuous dietary administration with concentrations in the feed of 25, 50, and 100 ppm. Control and high-dose groups for wild-type p53 / mice were not included in the study design. p-Cresidine at 400 mg/kg in corn oil served as the positive control. There were no increases in tumor incidence for Wy-14643-treated p53 / mice at any site that were judged to be treatment related. Nonneoplastic, treatment-related effects were seen in liver, mostly at the high-dose level in males and in the mid- and high-dose levels in females. These included periportal hypertrophy and hepatocellular vacuolation, hepatocellular necrosis (mostly single cell), Kupffer cell pigmentation, and increased mitoses of hepatocytes. The positive control, p-cresidine, produced a statistically signi cant bladder tumor response in both sexes. The study was accepted as showing a valid negative result for the test compound at maximum tolerated dose levels.

Nongenotoxic Non-Carcinogens Ampicillin (negative ): Ampicillin was studied in a single study (Bayer) using daily gavage administration at dose levels of 350, 1,000, and 3,000 mg/kg/day. Control and high-dose groups for wild-type p53 / mice were included in the study design. p-Cresidine at 400 mg/kg in corn oil served as the


42

STORER ET AL

positive control. There were no increases in tumor incidence for ampicillin-treated p53 / mice at any site that were judged to be treatment related. Nonneoplastic, treatmentrelated effects were seen in the gastrointestinal tract at the high-dose level. These included hyperplasias of the forestomach and glandular stomach in females and in ammatory in ltrates in the submucosa of the glandular stomach and in the cecal mucosa (frequently with cecal dilation ) in both sexes. The positive control, p-cresidine, produced a statistically signi cant bladder tumor response in males but was negative in females. The study was accepted as showing a valid negative result for the test compound at maximum tolerated dose levels. Sul sozazole (negative ): Sul soxazole was studied in a single study (Eli Lilly) using continuous dietary administration at concentrations in the feed of 0.05, 0.5, and 2.0%. Control and high-dose groups for wild-type p53 / mice were included in the study design. No concurrent positive control was reported, but the sponsoring laboratory had run benzene at 100 mg/kg by oral gavage as a positive control in another study in the same laboratory, but the results were negative (see discussion for cyclophosphamide ). There were no increases in tumor incidence for sul soxazole-treated p53 / mice at any site that were judged to be treatment related. There were also no treatment-related, non-neoplastic changes in the thyroid or any other tissues that were examined. The study was accepted as showing a valid negative result for the test compound at maximum tolerated dose levels with the quali cation that con rmation of the genotype is necessary given the failure of the positive control (benzene ) in the sponsor’s laboratory. Mannitol (negative ): Mannitol was studied in a single study (Aventis) using continuous dietary administration with concentrations in the feed of 2.5, 5.0, and 10.0 %. Control and high-dose groups for wild-type p53 / mice were included in the study design. p-Cresidine at 400 mg/kg in corn oil served as the positive control. There were no increases in tumor incidence for mannitol-treated p53 / mice at any site that were judged to be treatment related. There were also no treatment-related, nonneoplastic changes in any tissues that were examined. The positive control, p-cresidine, produced a statistically signi cant bladder tumor response in both sexes. The study was accepted as showing a valid negative result for the test compound at maximum tolerated dose levels.

DISCUSSION Diagnostic Pathology Issues Several issues in diagnostic pathology were encountered during the course of the evaluation and review of the histopathology across all of the studies. One of the more dif cult issues was the correct nomenclature, classi cation, and grouping of sarcomas for the purpose of evaluating tumor incidence by site. This is especially true since sarcomas are one of the most important spontaneously occurring lesions in p53 / mice. Although this issue did not appear to be pivotal for any of the ILSI ACT studies in terms of the determination of whether the test compound was positive, negative, or equivocal in the study, the situation may arise where the manner in which the sarcomas are grouped and analyzed changes the perceived outcome of the study. The


ILSI ACT Steering Committee discussed the broader issue of whether to analyze tumor incidence data both for individual tumor sites and also by total number of tumors per animal. The nontraditional approach of looking at the total number of tumors per animal was ultimately rejected after considerable discussion, the consensus opinion being that in most cases it would dilute the signi cance of ndings at a single tumor site. Most pathologists agreed however that grouping of some types of sarcomas was justi ed. Accordingly some general guidelines were adopted. In the p53 model, poorly differentiated sarcomas appear in the subcutis. These tumors can occur in association with implanted microchip transponders, but also occur spontaneously and in reaction to other treatments. The sarcomas not associated with transponders generally are not locally invasive and seldom metastasize. These tumors may show slight differentiation toward skeletal muscle (rhabdomyosarcoma), collagenous tissue ( brosarcoma ), endothelium (hemangiosarcoma), reticuloendothelial cell system ( brous histiocytoma ), or no recognizable differentiation (sarcoma, undifferentiated or not otherwise speci ed). The consensus of the p53 pathology subcommittee was that these tumors should be grouped, based on morphologic similarities of neoplasms in the ILSI studies and other studies with the p53 / model (28). Grouping these tumors would likely make them common rather than rare tumors. Sarcomas in deeper tissues should not be grouped with these (eg, brosarcoma of the meninges, osteosarcoma of bone, leiomyosarcoma and undifferentiated sarcoma in the prostate). It was recommended that malignant histiocytomas should also not be grouped with the dermal or subcuticular sarcomas, but should be grouped with malignant histiocytomas at other tissue sites. These guidelines are for grouping of sarcomas for the purpose of determining when the tumor incidence in a treatment group is increased relative to the control group. However, when compiling summary incidence tables, it would be preferable to leave all different diagnoses for a given site as separate and distinct entities, because they each re ect a separate histomorphology. This information would otherwise be lost if the different diagnoses were pooled under one diagnosis. A related issue is the problem of sarcomas associated with microchip transponder implants used for animal identi cation. As seen in Table 2, the spontaneous tumor data for vehicle control animals in the ILSI ACT studies shows an increased incidence of subcutaneous sarcomas in mice with transponders. Subcutaneous sarcomas are then classi ed as common tumors in both sexes in mice with transponders but are common tumors only in female mice, and not male mice, without transponders. Given the marked confounding effect of transponders in p53 / mice and the dif culty of dissecting out the relative contributions of the possible cooperating carcinogenic effects of the p53 genotype, the transponder, and the test chemical, it appears prudent to avoid the use of transponders in the p53 / model (note for example the confounding effect of the transponder in the data for phenacetin in Table 10). This issue and the possible mechanisms involved in transponder induced-tumorigenesis has also been addressed in the literature (4). The other critical issues in diagnostic pathology encountered during the review and evaluation of the data are not issues unique to the p53 model but center around dif cult


Vol. 29(Suppl. ), 2001

p53

/



/

43

studies: Incidence of lesions in metaproterenol study (Eli Lilly). Metaproterenol (mg/kg)

Control 0

100

400

1,500

Lesion incidencea

M

F

M

F

M

F

M

F

Lymphoma Rhabdomyosarcoma

0 0

3 0

0 0

0 0

1 0

1 1

0 0

0 0

a


diagnoses with lesions whose features place then at the juncture of two closely related diagnoses. For example, the differential diagnoses in lymphoid tissue between regenerative hyperplasia and preneoplasia or between atypical hyperplasia and lymphoma were especially problematic, especially because thymic lymphoma is an important spontaneous and induced tumor type in p53 / mice. Similarly, the differential diagnosis between nodular hyperplasia and adenoma was particularly dif cult for pituitary tumors.

Results for Other Compounds Tested in p53 / Mice Using Comparable Protocols Data on the carcinogenicity of the 21 ILSI ACT test chemicals in p53 / mice represents only a subset of the larger body of data from tests for the tumorigenic potential of chemicals in this model. In order to determine what other available data might be comparable to the ILSI ACT data for evaluation of the p53 / model, a review of the literature and other available data was conducted. The criteria for what constituted a “comparable” study were: that the p53 / heterozygous knock out mice used were the Donehower strain (8, 9, 10, 11) on a C57BL6 background (available commercially from Taconic, Germantown, NY); that a high dose that could reasonably be considered a near-maximally tolerated dose was tested; and that tumor incidence was analyzed, or could be evaluated from the data, in a cohort of 10 or more mice of at least one sex, within 24– 30 weeks from the beginning of the study. Using these criteria, data from studies on an additional 29 compounds were determined to be available and relevant to this evaluation. The results for these compounds are listed in Table 19. It should be noted that many, if not most, of these studies do not meet the rigorous criteria set forth for studies initiated under ILSI ACT protocols, particularly with respect to the extent of histopathologic evaluation, the rationale for high dose selection (eg, an MTD based on a 4-week range- nding study on wild-type mice), and the recommended number of mice per dose group (15). The negative results for the NIEHS-sponsored study with the rodent genotoxic carcinogen, glycidol, must be quali ed in this regard because they were from a study with only 7 (males) or 8 (females) mice per dose group (16). In addition, there was no increased mortality or reduction in body weight gain in the glycidol treatment groups in this study to suggest an MTD had been achieved.

Response to Genotoxic Carcinogens in p53 / Mice in Other Studies with Noncomparable Protocols Because the p53 / model is expected to respond to genotoxic carcinogens, any published results for a genotoxic carcinogen tested in the model under any protocol are of interest in terms of understanding and evaluating the perfor-

mance of the model. Table 20 lists results for an additional 13 compounds with published data from a study for which results at 26 weeks can be discerned. Note that for the compounds listed with positive results, most of the studies are in a p53-de cient mouse on a different genetic background. For the compounds listed with negative results, most of the studies used low doses, single doses and/or alternate dosing protocols (eg, single-dose, neonatal protocol ) and some showed positive responses after a 26-week time point. However, phenolphthalein has been shown to be negative after chronic high-dose exposure in 2 different p53 / knockout strains developed in Japan including one with a C57BL/6 background (32). The positive results for phenolphthalein in the US (Taconic ) p53 / model have been con rmed in both gavage and diet studies (18). Therefore, further research will be required to understand the mechanistic basis for this differential response to phenolphthalein between the U.S. and Japanese p53 / heterozygous knockout strains. Taken together, these differences in strain origin and protocol for the studies listed in Table 20 are considered signi cant with respect to the ILSI ACT protocol studies and, for this reason, the results are not considered together with the other available data when evaluating the overall performance of the p53 / model as a short-test for genotoxic carcinogens.

Results vs Expectations for the p53 / Model The expectation, as set forth in the literature for the performance of the p53 / model as a short-test for carcinogenicity (43 – 45), is that the model would show a selective response to genotoxic human and/or rodent carcinogens and that there would be a requirement for mutation and/or loss of the normal allele in the cells of a tumor induced by the action of the carcinogen. It was further postulated that nongenotoxic carcinogens and noncarcinogens would be negative when tested in the model (43). Two of the three (66% ) genotoxic human carcinogens that were included in the ILSI ACT project, melphalan and cyclophosphamide, gave positive results while phenacetin was clearly negative in 2 studies, one diet and one gavage. When these data are combined with the results of the other studies with genotoxic carcinogens with comparable protocols (Table 19), 12 of 16 compounds or 75% of all genotoxic carcinogens tested gave positive results in a 26-week study (Table 21). In addition to phenacetin the other negative compounds were 2,4-diaminotoluene, bromodichloromethane, and glycidol; however the negative results of glycidol must be quali ed, as previously discussed, by the small number of mice per treatment group and the lack of evidence that an MTD had been reached. Benzene (100 mg/kg) gave negative responses in 2 ILSI ACT studies where it was used as the positive control but data from other studies which used a


Chemical or agent


Diet

Gavage (0.5% methylcellulose)

Methylphenidate

Oxymetholone

Gavage 2x/wk (Corn oil/acetone 9:1) Gavage 5x/wk (Corn oil)

Diet Inhalation

2,4-Diaminotoluene Bromodichloromethane

Nongenotoxic rodent carcinogens 2,3,7,8-Tetrachlorodibenzo- pdioxin (TCDD) N-Methyloacrylamide

1x/wk Gavage 5x/wk (DiH2 0)

Topical/Skin 3x/wk for 30 weeks Drinking water

Gavage (0.5% Methylcellulose/0.2% Tween 80) Drnking water

Gavage 5x/wk (Corn oil) Gavage (Corn oil) Topical/Skin 2x/wk (Acetone) Gavage 1x (Citrate-buffered saline pH 4.5) Diet

50 ppm 250 ppm 500 ppm 125 mg/kg 625 mg/kg 1,250 mg/kg

25 mg/kg 50 mg/kg

0.25 ug

200 ppm 0.5 ppm 3.0 ppm 10.0 ppm 15.0 ppm

nd 25 mg/kg 50 mg/kg

0.0005%

0.002% 0.004% 0.0075% 0.25% nd

1 mg/kg 10 mg/kg 100 mg/kg

nd

100 mg/kg 200 mg/kg 200 mg/kg 400 mg/kg 12.5 mg 25.0 mg 90 mg/kg

Dose males

50 ppm 250 ppm 500 ppm 125 mg/kg 625 mg/kg 1,250 mg/kg

25 mg/kg 50 mg/kg

1.0 ug/kg

200 ppm 0.5 ppm 3.0 ppm 10.0 ppm 15.0 ppm

50 mg/kg 25 mg/kg 50 mg/kg

nd

13 kJ/m2

nd

200 ppm 375 ppm 750 ppm 3,000 ppm 12,000 ppm nd

100 mg/kg 200 mg/kg 200 mg/kg 400 mg/kg 12.5 mg 25.0 mg 90 mg/kg

Dose females

TABLE 19.—Results of other short-term studies in the p53 Route of administration and frequency (vehicle)a

7,12-Dimethylbez(a)anthracene Glycidol

Dimethylnitrosamine

UV-B

N-Butyl-N-(4-hydroxybutyl)Nitrosamine (BBN)

Urethane

Phenolphthalein

4-Vinyl-1-cyclohexene diepoxide (VCD) N-Methyl-N-nitrosourea (NMU)

p-Cresidine

Genotoxic carcinogens Benzene

44 /

p53 / resultsb

nd

nd

nd

nd nd

nd nd

nd

nd

nd

Positive for bladder neoplasms

nd

Negative

Negative. 24-week study

(41)

(16)

(43)

(14)

(14) (46)

(30) (16)

(20)

(25)

(34)

(6)

(13)

(2)

(43)

(41)

(15)

Reference

(Continued on next page)

Negative 24-week study plus 4 – 6-week hold Negative 24-week study plus 4 – 6-week hold

Positive for skin tumors at 26 weeks Positive for liver hemangiosarcomas No scheduled 26-week necropsy. F3 mice used. p53 / , CII double transgenic mice Protocol not comparable. Only 5 – 8 mice per dose group. Histopathology data not available Negative Negative

Positive for bladder neoplasms

Positive for hemangiomas, lung adenomas and hepatocellular carcinomas

Positive for lymphomas

Posivite for lymphomas (13 week study)

Positive for skin tumors

Positive for sarcomas

Comments & quali cations

nd

p53 / resultsb

heterozygous mouse model.


45

b

a

Gavage 5x/wk (water) Diet Topical/Skin 5x/wk (95% ethanol)

Diet Diet Drinking water

Dose frequency was 7x/week (daily) unless otherwise indicated. Results listed are for the overall evaluation of the results in the study.

Rotenone Oleic acid diethanolamine

Nongenotoxic, noncarcinogens Resorcinol

2,6-Diaminotoluene 8-Hydroxyquinoline 1-chloro-2-propanol

Diet

Inhalation 5x/wk

Chloroprene

Genotoxic noncarcinogens p-Anisidine

Topical/Skin 5x/wk (acetone)

Lauric acid diethylethanolamine

Cocunut oil diethanolamine

Gavage (0.5 % Methylcellulose/ 0.2% Tween 80) Topical/Skin 5x/wk (95% ethanol)

Diet

Pentachlorophenol

D -Limonene

Drinking water

Pyridine

Chemical or agent

Route of administration and frequency (vehicle)a

1,200 ppm 0.4 mg 0.8 mg 1.2 mg

225 mg/kg

0.225% 0.45% 200 ppm 3,000 ppm 250 ppm 500 ppm 1,000 ppm

2.5 mg 5.0 mg 7.5 mg 5.0 mg 10.0 mg 20.0 mg 2.0 ppm 12.8 ppm 80 ppm

250 ppm 500 ppm 1,000 ppm 100 ppm 200 ppm 400 ppm 250 mg/kg

Dose males

TABLE 19.—Results of other short-term studies in the p53

2.5 mg 5.0 mg 7.5 mg 5.0 mg 10.0 mg 20.0 mg 2.0 ppm 12.8 ppm 80 ppm

125 pm 250 ppm 500 ppm 100 ppm 200 ppm 400 ppm nd

Dose females

1,200 ppm 0.4 mg 0.8 mg 1.2 mg

225 mg/kg

p53 / resultsb

nd nd

nd

nd nd nd

nd

nd

nd

nd

nd

p53 / resultsb

heterozygous mouse model. (Continued)

0.225% 0.45% 200 ppm 3,000 ppm 250 ppm 500 ppm 1,000 ppm

/

Negative Negative

Negative

Negative Negative Negative

Negative

Negative

Negative (SA , MN ) Genotoxic? Negative

Negative

Negative

Negative

Comments & quali cations

(14) (40)

(14)

(14) (14) (40)

(43)

(16)

(40)

(40)

(6)

(40)

(40)

Reference

46

STORER ET AL

2-fold higher dose (200 mg/kg) indicate that benzene does give a positive response in the model (15, 23). Interestingly, the 2 hormones which are human carcinogens were either positive (DES) or equivocal (17-b -estradiol) in the model with evidence for accelerated tumorigenesis in the pituitary (DES and 17-b -estradiol) and/or testes (DES only ). Whereas DES has been shown to produce DNA adducts in some tissues (5) and genotoxic catechol metabolites of estrogens have been implicated in the mechanisms of estrogen tumorigenicity, these results in the p53 / model may indicate that both hormonal and genotoxic mechanisms play a role in estrogen tumorigenicity. Alternatively, at the low doses employed in these studies, genotoxicity may not be involved in the action of DES and 17-b -estradiol in these hormonally-responsive target tissues and the tumor data may then add some weight to the argument that the model is not speci c for genotoxic componds. The immunosuppressive human carcinogen, cyclosporin A, also gave a positive response for lymphomas in 2 diet studies with similar dose levels (50 – 60 mg/kg) whereas negative results were obtained in an earlier gavage study with dosing 5 days per week and a 2-fold lower top dose level (25 mg/kg). The wild type mice gave a quantitatively similar positive response in both studies suggesting that the p53 de cient genotype did not play an important role in the mechanism of tumorigenesis. This might be the expected result for an immunosuppressive agent if the cell types giving rise to the lymphomas were other than those giving rise to the T-cell thymic lymphomas for which the spontaneous rate is higher in p53 de cient versus wild-type mice. Interestingly, several of the lymphomas in the high-dose female mice in the SKB study were localized to small intestine or lung and did not present as thymic lymphomas. Analysis of the cyclosporin A-induced tumors for mutation of loss of heterozygosity of the p53 wild-type allele would be of interest in further clarifying the mechanism for this compound. For the class of nongenotoxic rodent carcinogens, none of the 12 ILSI ACT compounds tested nor any of the 10 other compounds in this class tested in non-ILSI studies with comparable protocols, gave positive results. Two of the ILSI ACT compounds, did however give equivocal results, chloroform and diethylhexylphthalate (for liver adenomas and kidney papilloma, respectively). For these tumor types in these tissues, there is no evidence, as yet, that p53 mice show accelerated tumorigenic responses to genotoxic carcinogens. Thus, while it remains possible that some non-genotoxic carcinogens may show slightly accelerated responses in p53 de cient mice in 26-week studies in target tissues for which the driving mechanism is toxicity and/or cell proliferation, there are no clear positive ndings in any study conducted to date to support this view. The negative results in the p53 / model with compounds such as phenacetin and WY-14643 that also produced kidney and liver toxicity, respectively, also do not support this idea. Finally, none of the 4 genotoxic noncarcinogens or 6 nongenotoxic, noncarcinogens tested to date in the p53 / mouse model in ILSI ACT sponsored studies or studies with comparable protocols have shown positive or equivocal results. The fraction of compounds grouped by category, that are positive in the assay, either in ILSI ACT sponsored


studies or studies with comparable protocols, is shown in Table 21. Results for one of the two hormones tested (DES) and for the immunosuppressive (cyclosporin A) are included under nongenotoxic carcinogens. The results for the compounds that were equivocal (17-b -estradiol, chloroform and DEHP) are not included for the purposes of this tabulation. In total, these data show that 42 out of 48 compounds or 88% gave results that were concordant with expectations for the model, namely that genotoxic human and/or rodent carcinogens would be detected as tumorigens in a 26-week study but that nongenotoxic carcinogens and noncarcinogens would not. The “false negatives” from this perspective of concordance with expectations were phenacetin, 2,4-DAT, bromodichloromethane, and glycidol (which may not have been tested adequately ). The “false positives” were cyclosporin A and diethylstilbestrol, both human carcinogens. However, as noted above, the positive results for DES in combination with reports in the literature indicating DNA adduct formation associated with exposure to DES (or a metabolite ) indicate that there might be a genotoxic, cocarcinogenic component to the action of DES and that the DES result might therefore not be a “false positive” with respect to concordance with the expectations for the model. Technical Lessons Learned from ILSI Protocol Studies with the p53 / Model Appropriateness of a 26-Week Study Duration: The appropriate duration for a “short-term” carcinogenicity assay is a key issue in the debate over how this type of study should be conducted. The ILSI p53 AWG acknowledged that the choice of 26 weeks for the ILSI ACT studies in the p53 / de cient mouse model was a somewhat arbitrary choice but it did re ect the satisfactory results from early pilot studies with the model which used 24- to 26-week exposure periods, with 4 – 6 additional weeks of observation in some studies. While this duration may not be optimal for speci c test substances, it does appear to be suf ciently long to detect reasonably potent agents. For example, the tumorigenicity of benzene was detected in 26 weeks in gavage studies using a 200 mg/kg dose, 5 – 7x/week (15, 23). However, the robust thymic lymphoma response between 6 and 9 months in inhalation studies with benzene sponsored by CIIT appeared to indicate that 6 months may not be optimal for this compound, a known human carcinogen (37). These data were based on the endpoint of the cumulative mortality of mice with thymic masses and it remains uncertain whether this study would have been positive had a full necropsy of 15 mice/sex/group been conducted at 26 weeks with microscopic examination of all tissues to detect early lesions. Therefore, it would not be appropriate to employ the CIIT data as a rationale for requiring a 9-month exposure period. The question of the optimal duration of studies in this model, and the additional, related question of the appropriate number of animals per dose group, were discussed. To the extent that the spontaneous tumor incidence remains low out to 8– 9 months of age in control p53 / mice, marginal gains in sensitivity to carcinogenic agents of lesser potency are likely to be achieved with longer duration studies and larger numbers of animals per dose group. Enhancements to the protocol of this type, resulting in the use of up to 20 – 25 mice per group or alternatively, 8 – 9 month studies, may ultimately be necessary in


Vol. 29(Suppl. ), 2001

p53

/


TABLE 20.—Response to genotoxic carcinogens in p53 Chemical or agent

Background strain

Positive Diethylnitrosamine Ethylnitrosourea 7,12-Dimethylbenz(a)anthracene

C3H CBA 129/C57BL6?

7,12-Dimethylbenz(a)anthracene /TPAb Negative IQb 6-Nitrochyrsene PhiPb Dimethylnitrosamine 239 Pu0 2 MeIQxb 60 Co radiation Phenolphthalein Phenolphthalein

/

mice: Results at 26 weeks in other experimental studies (with noncomparable protocols). p53 versus wild-type responsea

Tumor site

Liver Lung & uterus Mammary gland Skin

NIH

47

C57BL6 C57BL6 C57BL6 C57BL6 CBY/OYC 129/NIH/Ola CBA CIEA (C57BL/6)

p53 p53 p53

/

p53

/

/

/

p53 > p53 > p53

/

> p53

/

Comments and quali cations

Reference

Pituitary isograft used

(26) (31) (24)

/

/

p53

/

> p53

/

p53

/

> p53

/

(27)

Low-dose experiment Neonatal protocol Neonatal protocol Neonatal protocol Positive after 26 weeks Low-dose experiment Positive after 26 weeks

(33) (3) (3) (3) (17) (35) (28) (32) (32)

Accelerated or equivalent response in p53 / de cient versus wild-type mice. Abbreviations used: TPA, tetradecanoyl phorbol ester; IQ, 2-amino-3-methylimidazo[4,5-f]quinoline; PhiP, 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine; MeIQx, 2-amino-3,8dimethylimidazo[4,5-f]quinoxaline. a

b

order to achieve wider acceptance of, and con dence, in the sensitivity of the test system. Increasing the study duration in order to enhance sensitivity, however, would result in the loss of the existing control database and require further evaluation studies with genotoxic and non-genotoxic carcinogens and non-carcinogens using the “enhanced” protocol to examine both sensitivity and speci city. Appropriate Use of Positive Controls: The p53 / AWG concluded that it was certainly useful to include a positive control compound in the study design to con rm the tumor susceptible phenotype and genotype of the mice and it appears that regulatory agencies are requiring a positive control. In routine testing, protracted use and handling of known genotoxic carcinogens in testing facilities is not desirable and there will be pressures to not incorporate a positive control or use a single-dose protocol for the positive control. If a positive control is not used, con rmation of the genotype of the mice should be a required element of the protocol but even this may not be suf cient to guarantee a valid study for regulatory agency acceptance. For the positive control, the agent and dose should be chosen such that it is positive only in p53 / mice and not in wild-type in a 26-week study. p-Cresidine administered daily by gavage at 400 mg/kg in corn oil appears to be the best characterized positive control at present while benzene administered by gavage at 100 mg/kg TABLE 21.—Summary results for ILSI ACT and other compounds: Fraction of compounds positive in p53 / mice by category. Category

Genotoxic carcinogensa (Human and/or rodent) Nongenotoxic carcinogensa (Human and/or rodent) Genotoxic noncarcinogensa Nongenotoxic, noncarcinogensa a

Fraction positive

% positive

12/16

75%

b

2/22

9%b

0/4 0/6

0% 0%

Combined results for compounds in Tables 1 and 19. Excluding 3 equivocals (estradiol, chloroform, DEHP); includes “quali ed” negatives for dieldrin and reserpine. b

5-7 times per week appears clearly to not be an appropriate positive control. Publication of existing data for the higher dose of benzene (200 mg/kg) may provide suf cient evidence that this is also an appropriate positive control. Ultimately, the most practical positive control might be a single dose of a direct-acting genotoxin such as methylnitrosourea (MNU) at a dose level that produces a statistically signi cant increase in tumors in the p53 / mice but not in wild-type mice. However, the data to support this are not available at present; a dose-response study with MNU would rst need to be perfomed to nd an appropriate dose level with a clear differential response between p53 / and wild-type mice.

Criteria for a Valid Assay: Several criteria need to be applied when evaluating the validity of an individual assay. For studies with negative results, careful consideration must be given to the rationale for the high dose level and the extent to which the actual results of body weight, food consumption, clinical chemistry and non-neoplastic pathology indicate that a maximum-tolerated dose level was studied. Evidence for attainment of a maximum-tolerated dose level is an especially important criteria for a valid study since the data from the current evaluation studies indicate near maximum-tolerated dose levels are required to see effects with established genotoxic carcinogens such as benzene, cyclophosphamide, and p-cresidine. A second important criterion as discussed above is that the positive control should give a clear response. If it does not, or a positive control was not used, con rmation of genotype should be required. Where a positive control is used but fails to work, or when no positive control is used, and the correct genotype of the mice in all groups is con rmed, the results may nevertheless still carry less weight in the risk assessment process. Use of Nontumor Data in Study Evaluation: Nonneoplastic effects induced by the test agent as seen histologically, and to a lesser extent in clinical chemistry and hematology parameters, serve to identify target organs for pharmacologic and toxic effects. Most importantly, for the purposes of evaluating the tumorigenicity of the test agent, nonneoplastic and


48

STORER ET AL

preneoplastic proliferative lesions that appear to be on a histological continuum with any neoplasias seen in the same target organ can provide additional weight of evidence for a compound-induced tumorigenic effect. This is a judgment call by the pathologist, based on the weight of the histological evidence, which can confer biological signi cance to small numbers of tumors that do not achieve statistical significance. This is only true in practice for tumor sites where the spontaneous incidence is less than 1% (“rare” tumors under the ILSI ACT guidelines ) (36). Thus, an increase in pituitary adenomas in DES-treated female mice (see Table 13), though not statistically signi cant, was judged to be evidence for a tumorigenic effect of the test agent because of the supporting non-tumor histopathologic data showing increased incidences of diffuse hyperplasia and nodular hyperplasia. Use of Wild-Type Mice: 1. In Dose Range-Finding Studies. The ILSI ACT p53 / AWG saw no compelling evidence in its review of data in the study reports that were submitted to indicate that the use of wild-type mice in range- nding studies results in misleading information and/or inappropriate selection of doses for a 26-week study. In comparing body weight, mortality and other data for high-dose p53 / and wild-type mice there was also no consistent pattern of effects indicative of differential sensitivity of one genotype relative to the other. Whereas the p53 protein plays an important role in the response to DNA damage and other cellular stressors (15) and since there is evidence for a haploinsuf cient phenotype in p53 / cells (12, 47), it would not be unexpected or surprising to nd differential toxicity in p53-de cient versus wild-type mice. Differential toxicity has been observed for XPA / homozygous knockout mice (48) and there is limited evidence now from a non-ILSI sponsored study that p53 / heterozygous knockout mice may also show a similar increase in sensitivity to toxic stressors (46).

2. In 26-Week Studies. The ILSI ACT protocol for p53 studies included groups of control and high-dose treated wild-type mice in order to produce the necessary data to judge whether tumors induced by the test agent were accelerated in their development relative to any appearing in wild-type mice by 26 weeks. Where tumorigenesis is shown to be accelerated, it con rms a role for the p53 gene in the tumor response and adds to the weight of evidence that the test agent may act as a genotoxic carcinogen. This was true for the positive controls compounds, p-cresidine and benzene, and was also true for the one ILSI ACT genotoxic human carcinogen that was positive in the model (cyclophosphamide ) and for which wild-type high dose mice were included. Conversely, for the immunosuppressive human carcinogen Cyclosporin A, no signi cant acceleration of tumorigenesis (lymphomas ) was seen in the 2 out of 3 studies which were positive and included wild-type, high-dose mice. Certainly this is useful information. Finally, the nding for diethylstilbestrol that the combined incidence of pituitary lesions (adenomas and nodular hyperplasias ) appeared to be increased in p53 de cient as compared to wild-type, high-dose mice may also turn out to be useful information for risk assessment, possibly indicating a genotoxic as well as hormonal mechanism for tumor induction at this site.


Value of Additional Molecular Endpoints. Published data to date for several compounds indicates that molecular analyses of tumors for mutations in, or loss of heterozygosity for, the normal p53 allele in p53 / mice can provide additional information of value for understanding the mechanism of tumor induction by the test agent (12, 15). However, because there were few genotoxic agents studied in the ILSI ACT project that produced signi cant numbers of tumors of suf cient size for molecular analyses, little additional information has been generated as yet from the ILSI ACT studies. Efforts are underway however to apply newer technologies for molecular analysis of tumors (laser capture microdissection of tumors and quantitative TaqMan PCR assays) to some of the smaller lesions of interest observed in ILSI ACT studies. Based on the existing data, it appears that the nding of mutations or loss of heterozygosity (LOH) in a signi cant fraction of tumors does provide further weight of evidence for a direct or indirect genotoxic mechanism of action for a test compound. The converse does not appear to be true however. As reported for p-cresidine, the principal positive control employed in the ILSI ACT study, the absence of LOH or mutation in the wild-type allele in induced tumors does not appear to rule out a genotoxic mechanism. A genotoxin may, for example, target oncogenes or tumor suppressor genes other than the p53 gene, which when altered or deleted, act cooperatively with a p53 haploinsuf cient phenotype to accelerate tumor cell growth.

Use of Embedded Chip Transponders for Animal Identi cation. As discussed previously in the section on spontaneous tumor incidence, the increase in the background rate of subcutaneous sarcomas in mice with embedded chip transponders used for animal identi cation is problematic for the evaluation of tumor incidence in this model and is not therefore recommended. Anecdotal evidence indicates that the transponders can be displaced from the original site of implantation by handling and that transponder-induced tumors are not necessarily then found in close proximity to the chip. Thus, if a test agent has the potential to act independently to increase sarcoma incidence, the presence of transponders would confound interpretation of the study. ACKNOWLEDGMENTS Liaisons Dr. Joseph DeGeorge, Dr. James MacGregor, and Dr. Karl Lin (Food and Drug Administration, Rockville, MD) Study Sponsors Aventis (mannitol, dieldrin) Bayer AG (ampicillin) CIT/Synthelabo Recherche (methapyrilene) Dow (chloroform ) Dupont Pharmaecuticals (chlorpromazine, DEHP) Eli Lilly (cyclophosphamide, metaproterenol, sul soxazole ) GlaxoSmithKline (clo brate, MNU, cyclosporin A) Huntingdon Laboratories (metaproterenol) Merck and Co., Inc. (clo brate, phenacetin ) NIEHS/NTP (cyclosporin A, melphalan, DES, reserpine ) Novartis (estradiol, cyclosporin A, haloperidol ) P zer (sulfamethoxazole, phenobarbital )


Vol. 29(Suppl. ), 2001

p53

/


RW Johnson (estradiol, haloperidol ) Sano (WY-14643 ) Schering Plough (chlopromazine, phenacetin ) Searle/Monsanto (DES, phenobarbital ) Springborn Laboratories/Purdue Pharma L.P. (methapyrilene ) R EFERENCES

22. 23. 24.

25.

1. Aardema MJ, Isfort RJ, Thompson ED, LeBoeuf RA (1996). The low pH Syrian hamster embryo (SHE) cell transformation assay: A revitalized role in carcinogen prediction. Mutat Res 356(1): 5 – 9. 2. Allen J, personal communication, September, 2000. 3. Balachandra Dass S, Bucci TJ, He ich RH, Casciano DA (1999). Evaluation of the transgenic p53 / mouse for detecting genotoxic liver carcinogens in a short-term bioassay. Cancer Lett 143: 81– 85. 4. Blanchard KT, Barthel C, French JE, Holden HE, Moretz R, Pack FD, Tennant RW, Stoll RE (1999). Transponder-induced sarcoma in the heterozygous p53 / mouse. Toxicol Pathol 27(5): 519– 527. 5. Caldwell J, this volume. 6. Carmichael NG, Debruyne ELM, Bigot-Lasserre D (2000). The p53 heterozygous knockout mouse as a model for chemical carcinogenesis in vascular tissue. Environ Health Perspect 108(1): 61 – 65. 7. Contrera JF, DeGeorge JJ (1998). In vivo transgenic bioassays and assessment of the carcinogenic potential of pharmaceuticals. Environ Health Perspect 106(Suppl 1): 71 – 80. 8. Donehower LA (1996). The p53-de cient mouse: A model for basic and applied cancer studies. Semin Cancer Biol 7: 269– 278. 9. Donehower LA, Godley LA, Aldaz CM, Pyle R, Shi Y-P, Pinkel D, Gray J, Bradley A, Medina D, Varmus HE (1995). The role of p53 loss in genomic instability and tumor progression in a murine mammary cancer model. Prog Clin Biol Res 395: 1– 11. 10. Donehower LA, Harvey M, Slagle BL, McArthur MJ, Montgomery CA Jr, Butel JS, Bradley A (1992). Mice de cient for p53 are developmentally normal but susceptible to spontaneous tumors. Nature 356: 215– 221. 11. Donehower LA, Harvey M, Vogel H, McArthur MJ, Montgomery CA Jr, Park SH, Thompson T, Ford RJ, Bradley A (1995). Effects of genetic background on tumorigenesis in p53-de cient mice. Mol Carcinogen 14: 16 – 22. 12. Donehower LA, this volume 13. Dunnick JK, Hardisty JF, Herbert RA, Seely JC, Furedi-Machacek EM, Foley JF, Lacks GD, Stasiewicz S, French JE (1997). Phenolphthalein induces thymic lymphomas accompanied by loss of the p53 wild-type allele in heterozygous p53-de cient ( / ) mice. Toxicol Pathol 25: 533– 540. 14. Eastin WC, Haseman JK, Mahler JF, Bucher JR (1998). The National Toxicology Program evaluation of genetically altered mice as predictive models for identifying carcinogens. Toxicol Pathol 26(4): 461– 473. 15. French, JE, this volume 16. French JE, personal communication (glycidol, methylphenidate chloroprene), October, 2000. 17. Finch GL, March TH, Hahn FF, Barr EB, Belnisky SA, Hoover MD, Lechner JF, Nikula KJ, Hobbs CH (1998) Carcinogenic responses of transgenic heterozygous p53 knockoutmice to inhaled 239 Pu02 or metallic beryllium. Toxicol Pathol 26(4): 484– 491. 18. Furst SM, Blanchard KT, Lilly PD, Holden HE, Stoltz JH, Barthel C, Stoll RE (1999). Six month oral gavage and diet carcinogenic study with phenolphthalein in the heterozygous p53 / mouse. Toxicol Sci 48(1-S): A1202. 19. Gulezian D, Jacobson-Kram D, McCullough CB, Olson HM, Recio L, Robinson DE, Storer RD, Tennant RW, Ward J, Neumann D (2000). Use of transgenic animals for carcinogenicity testing: Considerations and implications for risk assessment. Toxicol Pathol 28(3): 482– 499. 20. Harvey M, McArthur MJ, Montgomery CA Jr, Butel JS, Bradley A, Donehower LA (1993). Spontaneous and carcinogen-induced tumorigenesis in p53-de cient mice. Nature Gen 5: 225– 229. 21. Harvey M, Vogel H, Morris D, Bradley A, Bernstein A, Donehower LA (1995). A mutant p53 transgene accelerates tumour development in het-

26. 27.

28. 29.

30.

31.

32.

33.

34.

35.

36. 37. 38.

39.

40.

41.

42.

49

erozygous but not nullizygous p53-de cient mice. Nat Genet 9: 305– 311. International Conference on Harmonisation (1998). Guidance on testing for carcinogenicity of pharmaceuticals. Fed Reg 63: 8983– 8986. Jacobsen-Kram D, personal communication, August, 1998. Joseph Jerry D, Butel JS, Donehower LA, Paulson EJ, Cochran C, Wiseman, RW, and Medina, D (1994). Infrequent p53 mutations in 7,12dimethyl[a]benzanthracene-induced mammary tumors in BALB/c and p53 hemizygous mice. Mol Carcinogen 9: 175– 183. Jiang W, Ananthaswamy HN, Muler HK, Kripke, ML (1999). p53 protects against skin cancer induction by UV-B radiation. Oncogene 18: 4247– 4253. Kemp CJ (1995). Hepatocarcinogenesis in p53-de cient mice. Mol Carcinogen 12:132– 136. Kemp CJ, Donehower LA, Bradley A, Balmain A (1993). Reduction of p53 gene dosage does not increase initiation or promotion but enhances malignant progression of chemically induced skin tumors. Cell 74:813– 822. Kemp CJ, Wheldon T, Balmain A (1994). p53-de cient mice are extremely susceptible to radiation-induced tumorigenesis. Nat Genet 8: 66– 69. Mahler JF, Flagler ND, Malarkey DE, Mann PC, Haseman JK, Eastin W (1998). Spontaneous and chemically induced proliferative lesions in Tg.AC transgenic and p53-heterozygous mice. Toxicol Pathol 26: 501– 511. Mirsalis JC, Hursey JA, Johnson A Fairchild DG, Kanazawa N Winegar, RA (2001). Evaluation of tumors and mutations in p53 / , CII double transgenic mice treated with various test agents. Toxicol 60(1): A969. Mitsumori K, Onodera H, Shimo T, Yasuhara K, Takagi H, Koujitani T, Hirose M, Maruyama C Wakana S (2000). Rapid induction of uterine tumors with p53 point mutations in heterozygous p53-de cient CBA mice given a single intraperitoneal administration of N-ethyl-N-nitrosourea. Carcinogenesis 21(5): 1039– 1042. Mitsumori K, Yasuhara K, Onondera H, Koujitani T, Takagi H, Hirose M, Usui T, Tamaoki N, Nomura T (2000). ILSI Health and Environmental Sciences Workshop on the Evaluation of Alternative Methods for Carcinogenicity Testing: Poster #19. Leesburg, VA. Morimura K, Salim EI, Yamamoto S, Wanibuchi H, Fukushima S (1999). Dose-dependent induction of aberrant crypt foci in the colons but no neoplastic lesions in the livers of heterozygous p53-de cient mice treated with low dose 2-amino-3-methylimidazo[4, 5- f ]quinoline.Cancer Lett 138: 81 – 85. Ozaki K, Sukata T, Yamamoto S, Uwagawa S, Seki T, Kawaski H, Yoshitake A, Wanibuchi H, Koide A, Mori Y, Fukushima S (1998). High susceptibility of p53( / ) knockout mice in n-butyl-N-(4-hydroxybutyl)nitrosamine urinary bladder carcinogenesis and lack of frequent mutation in residual allele. Cancer Res 58(17): 3806– 3811. Park, CB, Kim, DJ, Uehara, N, Takasuka, N, Hiroyasu, B-T, and Tsuda, H (1999). Heterozygous p53-de cient mice are not susceptible to 2-amino3,8-dimethylimidazo[4, 5- f ]quinoxaline (mIQx) carcinogenicity. Cancer Lett 139:177– 182. Popp J, this volume. Recio L, personal communication, October, 2000. Robinson D (1998). The International Life Science Institute’s role in the evaluation of alternative methodologies for the assessment of carcinogenic risk. Toxicol Pathol 26(4): 474– 475. Sagartz JE, Curtiss SW, Bunch RT, Davila JC, Morris DL, Alden CL (1998). Phenobarbital does not promote hepatic tumorigenesis in a twenty-six-week bioassay in p53 heterozygous mice. Toxicol Pathol 26: 492– 500. Spalding JW, French JE, Stasiewicz S, Furedi-Machacek M, Conner F, Tice RR, Tenant RW (2000). Responses of transgenic mouse lines p53 / and Tg.AC to agents tested in conventional bioassays. Toxicol Sci 53: 213– 223. Stoll RE, Holden HE, Barthel CH, Blanchard KT (1999). Oxymetholone: III. Evaluation in the p53 / transgenic mouse model. Toxicol Pathol 27(5): 513– 518. Storer RD (2000). Current status and use of short/medium term models for carcinogenicity testing of pharmaceuticals—Scienti c perspective. Toxicol Lett 112-113: 557– 566.


50

STORER ET AL

43. Tennant RW, French JE, Spalding JW (1995). Identifying chemical carcinogens and assessing potential risk in short-term bioassays using transgenic mouse models. Environ Hlth Perspect 103, 942– 950. 44. Tennant RW, Spalding J, French JE (1996). Evaluation of transgenic mouse bioassays for identifying carcinogens and noncarcinogens. Mut Res 365: 119– 127. 45. Tennant, RW (1998). Evaluation and validation issues in the development of transgenic mouse carcinogenicity bioassays. Environ Health Perspect 106 (Suppl 2): 473– 476.


46. Torti VR, Cobb AJ, Everitt JI, Marshall MW, Wong BA, Boorman GA Butterworth BE (2001). Bromodichloromethane inhalation in p53 / mice: Assessment of toxic and tumorigenic responses. Toxicologist 60 (1): A1326. 47. Venkatachalam S, Shi Y-P, Jones SN, Vogel H, Bradley A, Pinkel D, Donehower LA (1998). Retention of wild-type p53 in tumors from p53 heterozygous mice: Reduction of p53 gene dosage can promote cancer formation. EMBO J 17(16): 4657– 4667. 48. van Kreijl CF, this volume.
