prediction of rodent carcinogenesis: an evaluation of ......dence of a correlation between mouse...

Toxicologic Pathology, 32:393–401, 2004Copyright C© by the Society of Toxicologic PathologyISSN: 0192-6233 print / 1533-1601 onlineDOI: 10.1080/01926230490440934

Prediction of Rodent Carcinogenesis: An Evaluation of Prechronic LiverLesions as Forecasters of Liver Tumors in NTP Carcinogenicity Studies

D. G. ALLEN,1 G. PEARSE,2 J. K. HASEMAN,2 AND R. R. MARONPOT2

1Pathology Associates, A Charles River Company, Raleigh, North Carolina, USA, and2National Institute of Environmental Health Sciences and National Toxicology Program, Research Triangle Park, North Carolina, USA

ABSTRACT

The National Toxicology Program (NTP) developed the chronic 2-year bioassay as a mechanism for predicting the carcinogenic potential ofchemicals in humans. The cost and duration of these studies has limited their use to small numbers of selected chemicals. Many different short-termmethods aimed at increasing predictive accuracy and the number of chemicals evaluated have been developed in attempts to successfully correlate theirresults with evidence of carcinogenicity (or lack of carcinogenicity). Using NTP studies, the effectiveness of correlating prechronic liver lesions withliver cancer encompassing multiple studies using mice (83 compounds) and rats (87 compounds) was assessed. These lesions include hepatocellularnecrosis, hepatocellular hypertrophy, hepatocellular cytomegaly, bile duct hyperplasia, and hepatocellular degeneration, along with increased liverweight. Our results indicate that pooling 3 of these prechronic data points (hepatocellular necrosis, hepatocellular hypertrophy, and hepatocellularcytomegaly) can be very predictive of carcinogenicity in the 2-year study (p < 0.05). The inclusion of increased liver weight as an endpoint in thepool of data points increases the number of rodent liver carcinogens that are successfully predicted (p < 0.05), but also results in the prediction ofincreased numbers of noncarcinogenic chemicals as carcinogens. The use of multiple prechronic study endpoints provides supplementary informationthat enhances the predictivity of identifying chemicals with carcinogenic potential.

Keywords. Rat; mouse; prechronic toxicity; liver tumors; carcinogenicity; bioassay.

INTRODUCTION

Cancer persists as a worldwide disease and remains amongthe most common causes of death in the human population.Validated, dependable techniques that predict carcinogenicchemicals and provide information on human risk also re-main elusive. A wide range of research methods have beenexplored, each aimed at predicting the carcinogenic poten-tial of an untested chemical. For example, automated com-puter systems predict carcinogenicity based on mathematicalmodels created by statistical analysis to correlate relation-ships between chemical structure similarities to a defined bi-ological activity (Cunningham et al., 1998; Richard, 1998).In vitro methods such as the Salmonella mutagenicity assayand the micronucleus test are used to evaluate the ability ofa compound to interact with DNA. Other mammalian cell-based assays have been implemented for nongenotoxic car-cinogenicity predictions (Aardema et al., 1996; Foster, 1997).In vivo assays that reduce the term of the traditional 2-yearbioassay, so-called “medium-term” bioassays, have been re-searched (Ward and Ito, 1988). Transgenic mouse and ratmodels have been employed with the hopes of enhancingthe sensitivity of the 2-year bioassay (Tennant et al., 1995;Spalding et al., 2000). Many of these methods have provenbeneficial in the study of carcinogenic potential, but nonehave been sufficiently validated so as to supplant the chronic2-year bioassay as the benchmark carcinogenicity assay. Pit-falls of each assay have been recognized as limiting theirrole in carcinogenic predictions. Mathematical models and

Address correspondence to: Dr. Gail Pearse, NIEHS, P.O. Box 12233MD B3-06, 111 Alexander Drive, Research Triangle Park, NC 27709, USA;e-mail: [email protected]

structure activity relationships rely on generalizations of bi-ological activity based on chemical structure. Mutagenicityassays such as the Salmonella and Micronucleus assays areonly applicable to genotoxic carcinogens. Revelvant trans-genic models may be costly and difficult to obtain.

In contrast, there is an excellent concordance between theevidence of carcinogenicity in chronic bioassay studies usingcompounds that are known human carcinogens (Huff, 1999).In addition, direct comparisons among chemicals with di-verse structural and/or biological properties can be drawnsince the chronic bioassay is conducted by a defined protocoland on the same rodent strains. However, certain limitationsexist that contribute to many chemicals never reaching thestage of the chronic bioassay. These include the high cost ofa full 2-year bioassay in both sexes of rats and mice, and thetime interval between a chemical’s nomination for study andthe submission of a final report on findings. Therefore, theselimitations are not only financially burdensome, but also re-duce the throughput of potential chemicals because they areso laborious. In addition, animal welfare proponents cite theneed to minimize the number of animals subjected to study.

The dose and route of administration for the 2-year chronicbioassay is based largely on data obtained from prechronictoxicity studies using the same compound. Prechronic tradi-tionally encompasses both 14-day and 90-day exposures at awide range of doses with histopathological endpoints simi-lar to those in the chronic study. The results obtained in thisevaluation are used to determine the appropriate dose lev-els for the chronic study. The prechronic bioassay not onlyprovides important information for the study design of thechronic study, but also may provide insight into the expectedresults from the chronic study, including evidence of potentialcarcinogenicity (Ashby and Tennant, 1994).

393

at Society of Toxicologic Pathology on October 12, 2009 http://tpx.sagepub.comDownloaded from

http://tpx.sagepub.com

TAB

LE

1.—

Prec

hron

icm

ouse

liver

lesi

ons,

geno

toxi

city

,and

mou

seliv

ertu

mor

s.

394



CE

=cl

ear

evid

ence

ofca

rcin

ogen

icity

;SE

=so

me

evid

ence

ofca

rcin

ogen

icity

;E

E=

equi

voca

lev

iden

ceof

carc

inog

enic

ity;N

E=

noev

iden

ceof

carc

inog

enic

ity;F

=fe

ed;W

=dr

inki

ngw

ater

;G

=ga

vage

;I=

inha

latio

n;D

=de

rmal

;NT

=no

ttes

ted;

CY

M=

hepa

toce

llula

rcy

tom

egal

y;N

EC

=he

pato

cellu

lar

necr

osis

;BD

-HY

PR=

bile

duct

hype

rpla

sia;

HY

PT=

hepa

toce

llula

rhy

pert

roph

y;T

R=

Nat

iona

lTox

icol

ogy

Prog

ram

Tech

nica

lRep

ort;

CA

S=

Che

mic

alA

bstr

actS

ervi

ceR

egis

try

num

ber.

395



TAB

LE

2.—

Prec

hron

icra

tliv

erle

sion

s,ge

noto

xici

ty,a

ndra

tliv

ertu

mor

s.

396



CE

=cl

ear

evid

ence

ofca

rcin

ogen

icity

;SE

=so

me

evid

ence

ofca

rcin

ogen

icity

;E

E=

equi

voca

lev

iden

ceof

carc

inog

enic

ity;N

E=

noev

iden

ceof

carc

inog

enic

ity;F

=fe

ed;W

=dr

inki

ngw

ater

;G=

gava

ge;

I=

inha

latio

n;D

=de

rmal

;NT

=no

tte

sted

;CY

M=

hepa

toce

llula

rcy

tom

egal

y;N

EC

=he

pato

cellu

lar

necr

osis

;BD

-HY

PR=

bile

duct

hype

rpla

sia;

HY

PT=

hepa

toce

llula

rhy

pert

roph

y;D

EG

=he

pato

cellu

lar

dege

nera

tion;

TR

=N

atio

nalT

oxic

olog

yPr

ogra

mTe

chni

calR

epor

t;C

AS

=C

hem

ical

Abs

trac

tSer

vice

Reg

istr

ynu

mbe

r.

397



398 ALLEN ET AL. TOXICOLOGIC PATHOLOGY

This review of NTP studies was conducted as an attempt toidentify specific pathological endpoints in short-term toxic-ity that can be used individually and/or in concert with otherfindings to predict carcinogenicity in chronic studies. Thefocus of this review was to identify specific histopathologi-cal lesions that would provide investigators with criteria tostrengthen the justification for recommending a chemical forchronic study, or alternatively, to remove it from further con-sideration. This would serve to reduce the number of chemi-cals for chronic study, which in turn would reduce costs, andreduce the number of animals used. We have attempted to cor-relate specific hepatocellular pathology in prechronic studieswith carcinogenic endpoints in the chronic 2-year study. Wepresent data from chemicals tested by the U.S. National Tox-icology Program (NTP), encompassing both genotoxic andnongenotoxic compounds, as well as noncarcinogens. Wefocused specifically on liver carcinogenicity, as it representsthe most frequently diagnosed cancer in the reports that wereviewed, and thus maximizes our population size.

METHODS AND MATERIALS

Data SetRodent carcinogenicity bioassay data were obtained for

111 technical reports (TR) produced by the NTP over a10-year period, from March 1991 (TR-388) to July 2001 (TR-499). These represented the most recent final technical reportsto date when this evaluation was conducted. Only those re-ports from studies using B6C3F1 mice (Table 1) and/or Fisher344 rats (F344/N) (Table 2) were used for the analysis. In ad-dition, only those studies using both male and females froman individual species were evaluated (i.e., eliminating stud-ies on compounds delivered intravaginally, or those specifi-cally targeting single sex organs). Finally, only those studiesthat contained prechronic evaluations (studies ≤12 monthsduration) were considered. These criteria reduced the num-ber of compounds reviewed to 87 for rat, and 83 formice.

The NTP denotes evidence of carcinogenic activity using5 different categories. Two categories denote positive results:Clear Evidence (CE) and Some Evidence (SE). Studies thatare designated CE show a dose-dependent increase in neo-plasms, which may be malignant and/or benign. However, ifonly benign neoplasms are reported, sufficient evidence mustexist that these neoplasms progress to malignancy. Studiesthat are designated SE show a statistically significant, dose-dependent increase in neoplasms, albeit less than the levelof significance seen in CE studies. Another 2 categories de-lineate negative results: Equivocal Evidence (EE) and NoEvidence (NE). Studies that are designated EE show onlya marginal increase in neoplasms that may or may not bechemically related. Studies assigned NE show no chemicallyrelated increases in neoplasms. A fifth designation: Inade-quate Study (IS), is assigned to those studies that are flawedand/or limited in their capacity to define neoplastic changesand thus are not interpretable with respect to carcinogenic-ity. For the purposes of our evaluation, only those studies as-signed CE or SE were considered carcinogenic, while studiesassigned EE or NE were considered noncarcinogenic. Stud-ies designated IS were treated as untested in the species inquestion.

Specific hepatic histopathological lesions were evaluatedfor their correlation to evidence of carcinogenicity. These le-sions were selected as those that appeared most frequentlyamong the reports evaluated and were as follows: hep-atocellular cytomegaly, hepatocellular necrosis, bile ducthyperplasia, hepatocellular hypertrophy, and hepatocellu-lar degeneration (rats only). Increased liver weight (relativeand/or absolute) was also used as an endpoint for analy-sis because of its prevalence among the technical reportswe screened, as well as the positive correlation betweenliver weight and mouse liver cancer previously reported byElcombe et al. (2002). It should be noted that all lesionsreported in this study were as they appeared in the NTP tech-nical report for the respective compounds. No attempt wasmade to compare or contrast the nature and degree of similarlyrecorded changes. Furthermore, this review did not attempt tolink the dose that induced prechronic findings with the dosethat induced tumors.

Genotoxicity data were also evaluated using results fromthe Salmonella test and the Micronucleus assay, as these werethe most common genotoxicity assays employed among thetechnical reports evaluated. Compounds were scored as eitherpositive or negative in each assay. Although the Salmonellatest is often performed on multiple strains, a positive resultrecorded for any strain was interpreted as a positive result forthe assay.

Statistical AnalysisComparison tables (2 × 2) were constructed for each of the

parameters investigated (specified liver toxicity, genotoxicity,carcinogenicity) and a Fisher’s Exact Test was performed todetermine level of significance (defined as p < 0.05).

RESULTS

MiceThe best single predictor of liver cancer in mice was hepa-

tocellular hypertrophy (Table 3). Hepatocellular cytomegalyand hepatocyte necrosis also contributed, although the num-bers of positive findings were less than hypertrophy. Bileduct hyperplasia failed to identify any liver carcinogens thatwere not already identified by the other 3 nonneoplastic liverlesions. As a group, hypertrophy, cytomegaly, and necrosissuccessfully predicted 17 of the 27 liver carcinogens (10false negatives), with only 2 false positives (p < 0.0001).Adding increased liver weight as a predictor successfullyidentified 8 of the 10 false negatives as liver carcinogens.However, including increased liver weight resulted in theidentification of 16 additional false positives, for a totalof 18 false positives. As a single predictor, liver weightsuccessfully identified 18 of the 27 liver carcinogens, butalso identified 17 false positives.

With regard to the genotoxicity studies, there was no evi-dence of a correlation between mouse liver tumor chemicalsand Salmonella or micronucleus assay outcome. For exam-ple, of the Salmonella positive chemicals 27% (6/22) pro-duced mouse liver tumors compared with 33% (22/66) of theSalmonella negative chemicals that also produced liver tu-mors resulting in an insignificant difference. A similar resultwas observed for the micronucleus assay, although the samplesizes were much smaller, due to its exclusion from many of



Vol. 32, No. 4, 2004 PREDICTING RODENT LIVER TUMORS 399

TABLE 3.—Correlation between liver toxicity and carcinogenesis—mouse.

Liver neoplasia Cancer+ Cancer+ Cancer− Cancer−Liver toxicity Tox+ Tox− Tox+ Tox− Association

Hypertrophy 10 17 0 56 p < 0.001Necrosis 8 19 1 55 p < 0.001Cytomegaly 4 23 2 54 p = 0.084Bile duct-hyperplasia 4 23 1 55 p = 0.037Increased liver weight 19 8 17 39 p < 0.001HYPT+NEC+CYM 17 10 2 54 p < 0.001HYPT+NEC+CYM+LW 25 2 18 38 p < 0.001

CYM = hepatocellular cytomegaly; NEC = hepatocellular necrosis; BD-HYPR = bileduct hyperplasia; HYPT=hepatocellular hypertrophy; DEG=hepatocellular degeneration;LIVER WT, LW= increased liver weight.

the studies. In addition, none of the prechronic liver le-sions were correlated with either Salmonella or Micronucleusresults.

Therefore, our analysis indicated that a chemical showinga positive response in the 3 nonneoplastic liver lesions de-tailed previously in a prechronic study had very high likeli-hood (17/19 or 89.5%) of being a liver carcinogen in a chronicstudy. However, these lesions did not identify 10/27 or 37.0%of the liver carcinogens. If increased liver weight was alsogrouped with the liver lesions, a majority of the liver car-cinogens were identified (25/27, or 92.6%), but only 25/43,or 58.1%, of the positive prechronic findings correlated withliver cancer in the chronic studies (i.e., a large number offalse positives would be introduced).

RatsNo single prechronic liver lesion (when considered in-

dividually) was a strong predictor of liver cancer in rats.The most predictive lesion was hepatocellular hypertrophy(Table 4). As was seen in the mice, bile duct hyperplasia failedto contribute to predictions, and identified no carcinogens notdetected by other lesions. Hepatocellular degeneration wasalso a poor predictor of liver cancer. Grouping hepatocellularhypertrophy, hepatocellular necrosis, and hepatocellular cy-tomegaly (as in mice) resulted in 7 of the 11 (64%), rat livercarcinogens being correctly predicted (p < 0.01). However,this strategy also produced 16 false positives. Increased liverweight correctly predicted 8 of the 11 (73%) liver carcino-gens (p < 0.05), but also produced even more false positives,26, as well as producing 4 false negatives.

Therefore, as in mice, increased liver weight (when evalu-ated alone) was not as successful a predictive strategy as thegrouped strategy detailed before. Including increased liver

TABLE 4.—Correlation between liver toxicity and carcinogenesis—rat.

Liver neoplasia Cancer+ Cancer+ Cancer− Cancer−Liver toxicity Tox+ Tox− Tox+ Tox− Association

Hypertrophy 5 6 10 66 p = 0.02Necrosis 4 7 9 67 p = 0.056Cytomegaly 2 9 2 74 p = 0.076Bile duct-hyperplasia 2 9 4 72 p = 0.16Degeneration 2 9 5 71 p = 0.21Increased liver weight 8 3 26 50 p = 0.018HYPT+NEC+CYM 7 4 16 60 p = 0.006HYPT+NEC+CYM+LW 11 0 32 44 p < 0.001

CYM = hepatocellular cytomegaly; NEC = hepatocellular necrosis; BD-HYPR = bileduct hyperplasia; HYPT = hepatocellular hypertrophy; DEG = hepatocellular degenera-tion; LIVER WT, LW = increased liver weight.

weight in the grouped strategy corrected 3 of the 4 false neg-atives produced by the 3-lesion group, but also added falsepositives to bring the total to 32. Despite the high numberof false positives, the correlation between the grouped strat-egy (including increased liver weight) and rat liver cancer ishighly significant (p < 0.001).

Analysis of the rat data provided the same results as micewith respect to genotoxicity data—no significant correlationbetween liver tumors/toxicity and the 2 mutagenicity mea-sures were found. The only suggestion of correlation wasbetween liver tumors and Salmonella results (p > 0.15). Ofthe 24 positive Salmonella chemicals, 21% (5/24) producedliver tumors in rats compared with only 10% (7/71) of thechemicals with negative Salmonella results that producedliver tumors.

DISCUSSION

These results provide evidence that prechronic liver lesionsmay be used as a component in the search for predictors ofliver carcinogenicity in the chronic 2-year bioassay. In mice,a chemical showing a positive response in the 3 liver lesions(hepatocytomegaly, hepatocellular hypertrophy, and hepato-cellular necrosis) has a very high likelihood of being a car-cinogen in the chronic study. However, more than one-thirdof carcinogens would not be identified due to a propensity forfalse negatives using these criteria. If increased liver weightis also included as a lesion in the screening criteria, a majorityof these false negatives would be eliminated. This enhancedsensitivity would come at a cost because an increase in falsepositives would be created (Figure 1). In rats, the same 3liver lesions would again be very effective at predicting livercarcinogenesis while also producing fewer numbers of falsenegatives compared to mice. However, this apparent improve-ment in accuracy is offset by an increased occurrence of falsepositives. Inclusion of increased liver weight allows for all ofthe rat liver carcinogens evaluated in this study (11) to be suc-cessfully identified. As in mice, however, an improvement inthe numbers of carcinogens identified is accompanied by anincrease in the number of false positives (Figure 2). It shouldbe noted, however, that the predictivity seen in the rat datamight be slightly artifactual due to the low number of livercarcinogens (11) relative to mice (27).

It should be noted that clinical chemistry endpoints werealso explored as potential predictors of liver carcinogenic-ity. It is conceivable that significant changes in certain liverenzymes (e.g., alkaline phosphatase, lactate dehydrogenase,alanine aminotransferase, sorbitol dehydrogenase) could cor-respond to toxicity that may in turn correlate with specificliver lesions, and ultimately with liver cancer. However, uponexamination of all of the technical reports involved in thisstudy, it became apparent that several inconsistencies pre-cluded any feasible inclusion of these endpoints in the fi-nal analysis. For example, many of these studies have noprechronic liver chemistry data (samples were collected over12 months after initiation) or have no liver chemistry data atall. In the studies that do contain prechronic liver chemistrydata, there are a variety of endpoints that were collected, andthe same endpoints were not always collected. A completeset of consistently generated data might have strengthenedthe predictivity of the morphological parameters.



400 ALLEN ET AL. TOXICOLOGIC PATHOLOGY

FIGURE 1.—Correlation between mouse liver toxicity and mouse liver carcino-genesis: The impact of including increased liver weight group of mouse livertoxic lesions. (A) Tox = Pooled endpoints of hepatocellular necrosis, hypertro-phy, and cytomegaly. (B) Tox = Pooled endpoints of hepatocellular necrosis,hypertrophy, cytomegaly, and increased liver weight.

An intriguing finding is the fact that genotoxicity is notcorrelated with liver carcinogenesis in either rodent species.This conclusion implies that the liver carcinogens evaluatedare predominantly nonmutagenic in their mechanism of in-duction. However, the lack of micronucleus assay data forseveral chemicals mandates its exclusion from screening. Inaddition, there have been previous reports of compounds thatare genotoxic, based on positive Ames assay results, thatwere not found to be rodent liver carcinogens. Therefore, thevalidity of this conclusion could be questioned because itis solely dependent on Salmonella mutagenicity. Additionalgenotoxic endpoints could conceivably shift the associationbetween liver cancer and genotoxicity towards a more posi-tive correlation (Ashby, 1996).

As presented previously, 2 types of errors are inherentlyassociated with this type of evaluation. An error of false pos-itivity demonstrates an overprediction of cancer, while anerror of false negativity corresponds to an underprediction.Clearly, using this data analysis for liver cancer predictionmust be accompanied by careful scrutiny. In both rats andmice, inclusion of increased liver weight markedly enhances

FIGURE 2.—Correlation between rat liver toxicity and rat liver carcinogenesis:The impact of including increased liver weight group of rat liver toxic lesions.(A) Tox = Pooled endpoints of hepatocellular necrosis, hypertrophy, and cy-tomegaly. (B) Tox = Pooled endpoints of hepatocellular necrosis, hypertrophy,cytomegaly, and increased liver weight.

the probability of identifying carcinogens using this method.However, should this inclusion of liver weights be made ifit results in an overprediction of cancer? The primary ram-ification of this type of error would be the premature re-moval of a chemical from consideration when it actuallydoes not cause cancer. In contrast, while exclusion of in-creased liver weight reduces the number of false positives,an increase in false negatives results. The impact of such anerror could be the premature acceptance of a compound asa noncarcinogen when it is actually a carcinogen. A morerealistic scenario, however, would be the inclusion of thecompound in the 2-year chronic bioassay with the expecta-tion of noncarcinogenicity, only to learn that it is indeed acarcinogen.

With this information in hand, one must next determinethe best strategy with which to use this approach as a prog-nostic tool for carcinogenicity. The data suggest that theseprechronic liver lesions may provide another supplementaryinformation source for use in concert with other short-term as-says (e.g., in vitro assays) to create a pool of data. Compilingdata from multiple assays in this manner could conceivably



Vol. 32, No. 4, 2004 PREDICTING RODENT LIVER TUMORS 401

reduce the collective impact of the errors intrinsic to eachindividual assay. Perhaps an even more appropriate use ofthis data could be as an asset to study design of the chronicbioassay. Armed with the knowledge that a compound has asignificant chance of being a liver carcinogen, investigatorscould design experiments with more mechanistic approaches(in conjunction with the traditional protocol) that might shedsome light onto the actual mode of the cancer-causing agent.Finally, with the ever-mounting increases in budgetary con-straints, the reality exists that compounds under evaluationwill need to be prioritized. This method could potentially pro-vide investigators with evidence in support of or against con-tinued study by identifying the best candidates among a groupof compounds. Once again however, a decision would have tobe reached regarding inclusion of increased liver weights asa component in the screening and which type of error wouldbe most acceptable under these circumstances.

REFERENCES

Aardema, M. J., Isfort, R. J., Thompson, E. D., and LeBoeuf, R. A. (1996).The low pH Syrian hamster embryo (SHE) cell transformation assay: arevitalized role in carcinogen prediction. Mutat Res 356, 5–9.

Ashby, J. (1996). Alternatives to the 2-species bioassay for the identification ofpotential human carcinogens. Hum Exp Toxicol 15, 183–202.

Ashby, J., and Tennant, R. W. (1994). Prediction of rodent carcinogenicity for44 chemicals: results. Mutagenesis 9, 7–15.

Cunningham, A. R., Klopman, G., and Rosenkranz, H. S. (1998). Identificationof structural features and associated mechanisms of action for carcinogensin rats. Mutat Res 405, 9–27.

Elcombe, C. R., Odum, J., Foster, J. R., Stone, S., Hasmall, S., Soames, A. R.,Kimber, I., and Ashby, J. (2002). Prediction of rodent nongenotoxic car-cinogenesis: evaluation of biochemical and tissue changes in rodents fol-lowing exposure to nine nongenotoxic NTP carcinogens. Environ HealthPerspect 110, 363–75.

Foster, J. R. (1997). The role of cell proliferation in chemically induced carcino-genesis. J Comp Pathol 116, 113–44.

Huff, J. (1999). Long-term chemical carcinogenesis bioassays predict humancancer hazards. Issues, controversies, and uncertainties. Ann NY Acad Sci895, 56–79.

Richard, A. M. (1998). Structure-based methods for predicting mutagenicityand carcinogenicity: are we there yet? Mutat Res 400, 493–507.

Spalding, J. W., French, J. E., Stasiewicz, S., Furedi-Machacek, M., Conner, F.,Tice, R. R., and Tennant, R. W. (2000). Responses of transgenic mouselines p53(+/−) and Tg.AC to agents tested in conventional carcinogenicitybioassays. Toxicol Sci 53, 213–23.

Tennant, R. W., French, J. E., and Spalding, J. W. (1995). Identifying chemi-cal carcinogens and assessing potential risk in short-term bioassays usingtransgenic mouse models. Environ Health Perspect 103, 942–50.

Ward, J. M., and Ito, N. (1988). Development of new medium-term bioassaysfor carcinogens. Cancer Res 48, 5051–4.



prediction of rodent carcinogenesis: an evaluation of ......dence of a correlation between mouse...

Documents