The Utility of Genetically Modified Mouse Assays for Identifying Human Carcinogens: A Basic Understanding and Path Forward

James MacDonald*, John E. French{dagger}, Ronald J. Gerson{ddagger}, Jay Goodman§, Tohru Inoue, Abigail Jacobs||, Peter Kasper|||, Douglas Keller||||, Amy Lavin#,1, Gerald Long**, Bruce McCullough{dagger}{dagger}, Frank D. Sistare||, Richard Storer{ddagger}{ddagger} and Jan Willem van der Laan§§

* Schering-Plough Research Institute, Kenilworth, New Jersey 07033; {dagger} National Institute of Environmental Health Sciences, Research Triangle Park, North Carolina 27709; {ddagger} Endo Pharmaceuticals, Chadds Ford, Pennsylvania 19352; § Michigan State University, East Lansing, Michigan 48824; National Institute of Health Sciences, Tokyo 158-8501, Japan; || U.S. Food and Drug Administration, Center for Drug Evaluation and Research, Rockville, Maryland 20857; ||| Federal Institute for Drugs and Medical Devices (BfArM), D-53175 Berlin, Germany; |||| Sanofi-Synthelabo Research, Malvern, Pennsylvania 19355; # ILSI Health and Environmental Sciences Institute, Washington, DC 20005; ** Eli Lilly & Company, Greenfield, Indiana 46140; {dagger}{dagger} Aventis Pharmaceuticals, Inc., Bridgewater, New Jersey 08807; {ddagger}{ddagger} Merck Research Laboratories, West Point, Pennsylvania, 19486; and §§ National Institute of Public Health and the Environment (RIVM), 3720 BA Bilthoven, The Netherlands

Received August 21, 2003;
    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 REGULATORY PERSPECTIVES
 AREAS OF AGREEMENT AND...
 ASSAY SELECTION ISSUES
 PROTOCOL ISSUES
 CONCLUSIONS
 REFERENCES
 
The Alternatives to Carcinogenicity Testing Committee of the International Life Sciences Institute (ILSI) Health and Environmental Sciences Institute (HESI) conducted a large-scale, multinational collaborative research program to evaluate several genetically modified mouse assays for assessing the human carcinogenic potential of compounds. The data from this testing program have made an important contribution to the general understanding of how these models can be best applied in hazard identification; however, questions still exist regarding methodology and data interpretation. To address these issues, ILSI HESI hosted a February 2003 workshop on the Utility of Transgenic Assays for Risk Assessment. The purpose of this workshop was to reach an understanding of how data from genetically modified mouse models are viewed by different regulatory bodies in the pharmaceutical sector and, based on this understanding, to identify areas in which more experimental work may be needed to increase the utility of data derived from these assays. In the course of discussions, various data gaps related to model selection and protocol issues were identified. Based on the outcome of the workshop, various studies are proposed to provide data to improve the utility of currently available assays for cancer hazard identification and risk assessment purposes.

Key Words: carcinogenicity; p53+/-; knockout mouse; regulatory perspective; risk assessment; Tg.rasH2 transgenic mouse; genetically modified mice.


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 REGULATORY PERSPECTIVES
 AREAS OF AGREEMENT AND...
 ASSAY SELECTION ISSUES
 PROTOCOL ISSUES
 CONCLUSIONS
 REFERENCES
 
In 1996, the International Conference on Harmonization (ICH) Expert Working Group on Safety acknowledged the limited utility of conventional two-year rodent bioassays for assessing the human carcinogenic potential of chemicals, including pharmaceuticals, based on past positive findings that are now considered to have little or no relevance for human risk assessment. The ICH further acknowledged the potential of several new testing models to produce meaningful information for human cancer risk assessment. The group proposed a new scheme for the carcinogenicity testing of pharmaceuticals (ICH S1B). This scheme calls for one long-term rodent carcinogenicity study, plus an additional test for carcinogenic activity in vivo, consisting of either another long-term carcinogenicity study in a second rodent species, or a short- or medium-term rodent test, to be chosen from several available alternative models.

This guideline opened the way for scientists to use greater flexibility and judgment in choosing an approach for assessing carcinogenic potential. It also stimulated international interest in gaining experience and a greater understanding of the available methodologies for carcinogenicity testing, as these new methods had not been fully characterized, and scientific questions remained to be addressed regarding their most appropriate application. To address these issues, the International Life Sciences Institute (ILSI) Health and Environmental Sciences Institute (HESI) coordinated a large-scale, multinational collaborative research program to provide data needed to further understand the benefits and limitations of the proposed new models. Under the auspices of the ILSI HESI Alternatives to Carcinogenicity Testing (ACT) Committee, seven alternative models have been tested, including five genetically modified mouse models:

The results of this testing program, developed over 4 years, were previously discussed in Toxicological Sciences (Cohen et al., 2001Go), presented at a workshop, and as a special issue of Toxicologic Pathology (Vol. 29, supplement issue, 2001). This latter publication also contains comprehensive background information covering each of the alternative models. The ILSI HESI research program has added important new data to our understanding of how these models can be applied in assessing carcinogenic potential; however, questions still exist regarding how these models may best be used in a regulatory environment. To address these questions, the ILSI HESI ACT Committee hosted a February 2003 workshop on the Utility of Transgenic Assays for Risk Assessment in Washington, DC. This workshop involved approximately 75 participants from the United States, Europe, and Japan, including scientists from industry, government, and academia. The purpose of the workshop was to reach an understanding of how data from genetically modified mouse models are viewed by different regulatory bodies and, based on this understanding, identify areas in which more experimental work may be needed to increase the utility of data derived from these assays.


    REGULATORY PERSPECTIVES
 TOP
 ABSTRACT
 INTRODUCTION
 REGULATORY PERSPECTIVES
 AREAS OF AGREEMENT AND...
 ASSAY SELECTION ISSUES
 PROTOCOL ISSUES
 CONCLUSIONS
 REFERENCES
 
Perspectives of U.S., European, and Japanese regulatory agencies regarding the use of genetically modified assays for carcinogenic risk assessment, as expressed at the ILSI HESI ACT Workshop, are described below.

United States Food and Drug Administration (U.S. FDA)
Over the past 2 years, 25% of the proposed mouse carcinogenicity study protocols received by the FDA’s Center for Drug Evaluation and Research’s Executive Carcinogenicity Assessment Committee have been for an alternative model. Data from approximately 90 protocols and 24 completed genetically modified mouse or other alternative assays have been received and evaluated. Most of these protocols and completed studies have used the p53+/- assay. The p53+/- and neonatal mouse studies are generally considered appropriate for clearly or equivocally genotoxic drugs. The Tg.rasH2 assay is generally considered appropriate for either genotoxic or nongenotoxic drugs, based on the results of the ILSI HESI research program evaluation of this assay. The Tg.AC assay has generally been used for testing dermally applied drug products, although some drugs intended for systemic administration have been assayed using the dermal route.

Although most of the compounds tested in the p53+/- assay were genotoxic in the ICH battery (primarily in in vitro clastogenicity assays), results in 16 of 16 p53+/- assays received over the past seven years have been negative. The reasons for this are not currently understood. Results of traditional two-year carcinogenicity studies in rats have been received for a number of drugs that gave negative results in the p53+/- studies. Drug-related neoplastic findings in rats were seen for most of these compounds. Although a number of the results may be attributed to nongenotoxic mechanisms, some of the results have no apparent nongenotoxic mechanism. Consistent with the experience of some other investigators, p53+/- protocol issues that remain to be addressed include the variable performance of the positive control (mainly p-cresidine), the duration of the assay, the age of animals at study initiation, and the response to dermally applied drugs.

Results in three of the five Tg.AC studies have been positive, not including findings at distal sites. A fourth study was positive only at distal sites. Drug vehicle has been shown to affect local papilloma formation in response to 12-O-tetradecanoylphorbol-13-acetate (TPA) and various drug products. Also, in order to minimize local irritation and inflammation at the site of application, it is desirable not to exceed the dermal maximum tolerated dose (MTD) found in dose-ranging studies; however, there are examples showing dermal irritation and inflammation without papilloma formation, and examples of papillomas formed without preceding dermal irritation or inflammation. The apparent inability of the Tg.AC assay to distinguish nongenotoxic "promoters" from "complete" genotoxic carcinogens complicates the integration of this assay with results from traditional carcinogenicity assays and other data.

FDA experience with the Tg.rasH2 assay has been too limited to comment on here, although this assay is generally viewed as acceptable for carcinogenicity testing, based on the results of the ILSI HESI research program evaluation of this assay.

Results from these alternative studies have helped alleviate concern when the adequacy of a traditional 2-year study was questionable and repeating the two-year study was undesirable to both the drug sponsor and the FDA. Results have also helped alleviate concern when the traditional two-year rat study was adequate, but the findings were equivocal. Results from these assays have been used as a part of the weight of evidence in early assessment of genotoxicity study results before initiation of clinical trials. It should also be noted that results from genetically modified and other alternative assays should not be used in isolation, but rather, integrated with all available toxicology and pharmacology data and results from traditional carcinogenicity assays done in rats.

European Committee for Proprietary Medicinal Products Safety Working Party (CPMP SWP)
The following conclusions and recommendations are based on a comprehensive review by the CPMP SWP of the data produced from the ILSI HESI research program. Both the Tg.rasH2 and the p53+/- models are considered to be acceptable for regulatory use and likely to have an additive value to carcinogenicity assessment if the experiment is properly designed. The available data do not suggest that one model is more appropriate than another for a particular class of compounds, a particular mechanism of tumorigenic activity, or other specific conditions. It should be noted that these genetically modified models can also be used as an additional component in the assessment of potential genotoxic carcinogenicity. However, the outcome of an experiment with genetically modified animals should not be considered as THE decisive factor in the assessment of genotoxicity, but rather, as part of the weight of evidence in this assessment.

The Tg.AC model reacts inconsistently and incompletely to known human carcinogens. Although developed to be responsive at the site of application (i.e., the skin), the evaluation of studies using human carcinogens has included responses at other sites as well, in accordance with the profile of the compound. Nevertheless, the model is considered to be useful for screening the carcinogenic properties of dermally administered pharmaceuticals. The Tg.AC model cannot be recommended for oral studies with the forestomach as the reporter site. The XPA-/- and XPA-/-/p53+/- assays appear to be promising models, but more data from studies using acceptable protocols are needed.

As part of the above evaluation of alternative models, several points were raised regarding experimental design:

If a positive study outcome is defined on the basis of the "rare tumor criteria" only, a repetition of the study should be considered (especially when the historical control data on which the definition of a "rare tumor" is based are relatively limited).

Wild type animals should be included (except in the Tg.AC assay), preferably as control and high dose groups, in order to determine whether the outcome in genetically modified animals is a function of the modified genotype. This would provide additional and useful mechanistic information.

At present, positive controls should be included in the p53+/- and Tg.rasH2 models. As an alternative, approaches to verify the genotype of the test animals at a molecular level may be acceptable in the future.

The regulatory experience gathered in the European system consists mainly of requests for advice regarding study design and discussions with companies about the potential acceptability of such a test in the nonclinical package for a marketing authorization. As time progresses, it is anticipated that companies will include the results of these studies in their marketing authorization applications.

In Europe, companies can come to the European Agency for the Evaluation of Medicinal Products (EMEA) for advice with respect to their product safety-testing plan, including setup of carcinogenicity testing assays. While the ILSI HESI program was underway, three requests for advice were received, all regarding the applicability of newborn mice. The CPMP accepted all three proposals; however, it became clear later on that the FDA had expressed a different opinion in various cases, mainly because of genotoxicity testing data. Following completion of the ILSI HESI research project, the CPMP received four specific requests for advice regarding the p53+/- model. Issues related to how to handle estrogenic hormones, and the tissue specificity of the model.

Thus far, five product applications received for marketing authorization of new active substances include studies with p53+/- mice in their dossiers; all five studies showed negative results for carcinogenicity. These studies were accepted as contributing to the weight of evidence, in combination with results from the long-term rat study (available for four of five products) and data from the genotoxicity tests. A Tg.rasH2 study was also included in a dossier for which a p53+/- study was already present. The company carried out this additional study because of lack of experience with the p53+/- model with regard to compounds that may be specifically carcinogenic in the gastrointestinal tract. Finally, in a dossier of an orally administered compound, a dermal Tg.AC study was included. This study was not accepted as contributing to the weight of evidence, as the route of administration was considered to be inappropriate, and the study did not add to the body of information regarding the mechanism of tumor formation in the rat.

To conclude,

The experience obtained in the European system thus far is insufficient to allow for further conclusions.

Japanese Ministry of Health, Labor, and Welfare (MHLW)
Current regulatory requirements allow for the use of an alternative assay, in combination with a traditional 2-year assay, to assess a compound’s human carcinogenic potential. In choosing a genetically modified animal model for assessing carcinogenic potential, it is important to consider the model’s underlying mechanism of action and phenotypic characteristics. In general, the Tg.rasH2 model responds well to both genotoxic and nongenotoxic carcinogens, the p53+/- model is most responsive to complete carcinogens, and the Tg.AC model is of limited use due to its limited phenotype.

The general testing scheme for use of alternative carcinogenicity assays will depend on the compound’s mutagenicity. A mutagenic compound is evaluated using a multi-organ assay system, such as the p53+/- or Tg.rasH2 model, which can detect complete genotoxic carcinogens. The compound should also be tested in an assay, such as the MutaTM mouse or Big-BlueTM, which predicts the target tissues of effect, before being evaluated in a model aimed at assessing these specific target tissues (e.g., XPA-/- for skin or oral tissues, Tg.rasH2 for lung).

The p53-/- model is being used in studies conducted at the Japanese National Institute of Health Sciences (NIHS) aimed at understanding mechanisms of carcinogenesis, with emphasis on dose-response relationships and discerning possible thresholds. Additionally, researchers at NIHS are working toward developing a p53-/- model on a C3H/He background that might be capable of responding to genotoxic carcinogens within ten weeks of treatment.

Assays for evaluating the carcinogenic potential of nonmutagenic compounds are not readily available at this time. For now, a nonmutagenic compound should be evaluated in a multiorgan system, such as the Tg.rasH2, which can detect compounds acting through epigenetic mechanisms. In the future, the compound should also be evaluated using a cDNA microarray assay to predict possible target tissues, and then examined using a specific target tissue model.


    AREAS OF AGREEMENT AND DISAGREEMENT IN PERSPECTIVES
 TOP
 ABSTRACT
 INTRODUCTION
 REGULATORY PERSPECTIVES
 AREAS OF AGREEMENT AND...
 ASSAY SELECTION ISSUES
 PROTOCOL ISSUES
 CONCLUSIONS
 REFERENCES
 
To fulfill the objectives of the workshop, it was considered instructive to make a side-by-side comparison of the views expressed concerning the utility of each of the alternative models.

p53±/- Model
CPMP considers this model acceptable for regulatory purposes and does not limit its use for genotoxic compounds only. FDA and NIHS, on the other hand, consider this as an appropriate alternative model when dealing with compounds that are clearly or equivocally genotoxic only.

Tg.rasH2 Model
CPMP, FDA, and NIHS consider this as an appropriate alternative model for regulatory purposes for both genotoxic and nongenotoxic compounds.

Tg.AC Model
CPMP considers this model to be useful for screening the carcinogenic potential of dermally administered pharmaceuticals. In agreement with CPMP, FDA considers this model useful for dermally applied products; however, some data from products intended for systemic administration but assayed using the dermal route in the Tg.AC model, have been reviewed. NIHS expressed concerns regarding the stability of this model’s phenotype.

XPA-/- and XPA-/-/p53±/- Models
CPMP expressed the view that these models, while promising, require further development. For example, additional studies using known human carcinogens should be conducted. Neither FDA nor NIHS appear to have experience with these models.

Neonatal Mouse Model
CPMP has accepted proposals for the use of this model. FDA considers this to be an appropriate model in select limited circumstances for compounds that are clearly or equivocally genotoxic. NIHS has some experience with this model.

There is a need for further discussion to resolve differences between the CPMP and FDA concerning the acceptability of the p53+/- model for genotoxic versus nongenotoxic compounds. The FDA reserves the use of this model for compounds that are clearly or equivocally genotoxic, while CPMP considers the model acceptable for regulatory purposes without stipulating that the compound in question be shown to have genotoxic potential. In this context, it is important to note that the p53+/- model is considered inappropriate as a test for genotoxicity as such, but rather, is appropriate for establishing potential carcinogenicity.

Overall, the above perspectives show that there is now widespread agreement that alternative assays can and, indeed, do play an important role in carcinogen safety assessment. Importantly, general thinking has advanced beyond the notion that the traditional standard approach involving two species of rodent of both sexes exposed over their lifetimes is the only way to assess the carcinogenic potential of compounds in vivo. It is now evident that alternative models can be used as an integral component of an overall evaluation of carcinogenic potential. In addition, the above perspectives also indicate that there are potential areas in which more study of alternative models would be beneficial. Having reached an understanding of how these assays are currently used in safety risk assessment, it is important to consider some of the issues related to their use. Such an exercise will help to elucidate what the data gaps are for these assays and what additional studies may be done to increase their utility.


    ASSAY SELECTION ISSUES
 TOP
 ABSTRACT
 INTRODUCTION
 REGULATORY PERSPECTIVES
 AREAS OF AGREEMENT AND...
 ASSAY SELECTION ISSUES
 PROTOCOL ISSUES
 CONCLUSIONS
 REFERENCES
 
Regulatory perspectives regarding the use of alternative models are still evolving. Therefore, the decision as to which assay to choose becomes a function of the properties of the test chemical, the regulatory acceptance of the models, and the availability of the desired mouse strain. In special cases, data such as metabolism, pharmacokinetics, mechanisms of action, target tissues, and potential for molecular analysis of tumor tissues may play a role in assay selection. At the present time, the genotoxicity profile of a test chemical appears to play a large role in assay selection. Compounds that are equivocally or weakly genotoxic in one or more tests have been assayed largely in the p53+/- model, with a smaller number of Tg.rasH2 assays performed. Uncertainty exists regarding the best model to choose for weakly or equivocally genotoxic materials, especially if these results are only seen in vitro, in a single assay, or only at very high and/or cytotoxic doses. It may be useful to evaluate more genotoxic compounds that act through a wide variety of mechanisms of genotoxicity (i.e., oxidative damage, DNA intercalation, topoisomerase inhibitors, etc.). At this point, the Tg.rasH2 model is the preferred model for nongenotoxic compounds, and it responds to genotoxic compounds, too.


    PROTOCOL ISSUES
 TOP
 ABSTRACT
 INTRODUCTION
 REGULATORY PERSPECTIVES
 AREAS OF AGREEMENT AND...
 ASSAY SELECTION ISSUES
 PROTOCOL ISSUES
 CONCLUSIONS
 REFERENCES
 
The strengths and limitations of alternative mouse models for carcinogenicity assessment are a function of both the model’s invariant genetic background and the detailed protocol design variables guiding the conduct of a specific study. In this section, issues relating to current protocol design variables are evaluated, and modifications likely to optimize the value of these alternative models are considered. The specific objectives are to identify the most critical protocol design variables that, if altered, would likely improve the clarity, acceptability and confidence of a positive or negative tumorigenic response.

Study Duration
For the Tg.AC and Tg.rasH2 models, a 6-month study duration has been generally considered sufficient. However, it has also been suggested that 6 months might be suboptimal for the p53+/- assay, and that this protocol design variable should be further investigated. Several published studies indicate that background tumor findings remain low in p53+/- mice for 9 months (Donehower et al., 1992Go; Recio et al., 2000Go). These findings are further supported by the results of five 9-month studies recently completed by the National Toxicology Program (J.R. Bucher, personal communication). In addition, benzene inhalation studies have shown that the cumulative tumor incidence at 9 months (~70–80%) in moribund or early death benzene-exposed animals was higher than that previously seen at a scheduled 6-month necropsy (51%) after oral dosing (French, 2001Go). Unfortunately, tumor incidence was not assessed at 6 months in this inhalation study. Finally, a positive response was obtained in at least one sex in 18 of the 19 oral gavage studies with the p-cresidine positive control (400 mg/kg/day for 6 months), based on either a statistically significant increase in bladder tumor incidence, or an increase judged to be significant by criteria for rare tumors (Storer et al., 2001Go). Tumor incidence was statistically significant in 17/19 studies for males and in 15/19 studies for females (Storer et al., 2001Go); similar variability in the positive control response has been noted in regulatory studies (Sistare and Jacobs, 2003Go). Therefore, based on the results described above, careful investigation of the merits of a 9-month design as compared to the 6-month protocol for p53+/- studies may be warranted.

Age at Study Initiation
The age of the animals at the start of a study has not always been tightly and consistently controlled. Some studies, for example, have started dosing animals at 12 weeks of age, while the majority of studies began with animals less than eight weeks old. It may be useful to analyze existing data sets to evaluate the potential relationship between age and tumor response within or across the many p-cresidine studies.

Group Size
Morton et al.(2002)Go indicated that 25 Tg.rasH2 mice per group provide statistical power that is generally equivalent to that of a standard two-year mouse bioassay. It is suggested that 25 animals per group for p53+/- mice be adopted as the standard without the need for further specific study of this variable. It should be noted that 50 animals per group are currently used in the standard two-year rodent assay.

Positive Controls
There is a diversity of opinions as to the need to run a concurrent positive control with every study. Most agree that it would be advantageous for a positive control alternative to p-cresidine to be considered for the p53+/- assay. The positive controls for Tg.rasH2 (a single dose of N-methyl-N-nitrosourea [MNU]) and Tg.AC models (TPA 3x/week) are generally considered adequate, but it is proposed that the supplier/breeder might establish a robust phenotypic testing scheme for all alterative models to satisfy concerns regarding the integrity and stability of a genetically modified model’s responsiveness. This approach would not, however, address questions concerning the effects of vehicles on dermal responses to tumorigens in Tg.AC mice.

Tissue-Restricted Tumor Susceptibility
Results from a limited number of p53+/- studies done on different background strains confirm the generally accepted notion that the genetic background of the parental strain influences a model’s transgene-dependent tumor spectrum. It is important to test/identify/confirm the "blind spots" or tumor-resistant target organs and then decide on best next steps. Second-tier steps might include comparative assessments of the same genetic modification on different parental strain backgrounds, or of different transgenic constructs on the same strains. However, testing for likely "blind spots" in the existing models is considered a higher priority.

Maximum Tolerated Doses
It has been suggested that the dose of p-cresidine used as a positive control in the p53+/- assay (400 mg/kg/day) exceeds the MTD for p53+/- mice. The available data on the p53+/- and Tg.rasH2 models should be carefully assessed and tabulated for positive and negative studies, both at doses when MTDs are exceeded and at the next lower doses. It has also been suggested that the accuracy of MTD determinations derived from 30-day dose-range finding studies performed in wild-type parental strains be reviewed. It is further noted that the upper limits of a nonconfounding MTD in Tg.AC skin after topical exposures have not been well defined.

Rare Tumor Criteria
For the ILSI HESI research program, tumors at rare sites (background tumor incidence of 1% or below) were considered drug-related if 2 of 15 study animals developed tumors with concurrent evidence of drug-related hyperplasia at the same tissue sites in additional test mice. With an increase in group size to 25, confirmation that the background incidence remains below 1% is viewed as important. It should be noted that the lack of availability of a larger historical database, especially relating to potential drug-related marginal/equivocal tumor findings, is presently a deterrent to the expanded use of these models.


    CONCLUSIONS
 TOP
 ABSTRACT
 INTRODUCTION
 REGULATORY PERSPECTIVES
 AREAS OF AGREEMENT AND...
 ASSAY SELECTION ISSUES
 PROTOCOL ISSUES
 CONCLUSIONS
 REFERENCES
 
Based on the above discussion, it is apparent that there are a number of areas of general agreement with regard to the utility of genetically modified mouse models for risk assessment. These include agreement that:

Based on input provided by U.S., European, and Japanese pharmaceutical regulatory scientists, the most commonly used genetically modified mouse models for risk assessment today are the p53+/-, Tg.AC, and Tg.rasH2 assays. These are also the most well-characterized models available at this time. It is noted that a niche has been carved out for the Tg.AC model, which is being applied to a limited extent in testing the carcinogenic potential of dermally applied compounds. It is also noted that, at least in the United States, the greatest amount of regulatory experience has been associated with the p53+/- model. Regulatory bodies, however, express disparate views regarding the application of this model for risk assessment purposes. Specifically, the United States and Japan consider the p53+/- model acceptable for the testing of compounds shown to be clearly or equivocally genotoxic only; however, European regulatory bodies consider this model acceptable for all compounds.

In addition to general agreement regarding the utility of the p53+/-, Tg.rasH2, and Tg.AC models, a number of the general aspects of study design are also accepted. However, a number of specific details regarding group size, study duration, and data interpretation are under discussion. While it is generally accepted that group sizes should be increased to 20–25 animals per sex per dose group, the single most important outstanding issue is the duration of the p53+/- assay. It is not known from presently available data whether an increase in duration of this assay from 6 to 9 months would enhance the utility of the data derived.

In the course of discussions, the following data gaps related to assay selection and/or protocol issues have been identified for which more experimentation could be of value in increasing the utility of genetically modified mouse assays for risk assessment:

Further clarification around the issues listed above would help in the interpretation of results obtained using an alternative assay. For example, expanding the historical control animal tumor incidence databases would be helpful in interpreting the findings from alternative models. Additionally, it is noted that experience with alternative testing methods remains limited, which in turn, affects the decision to test for carcinogenic potential using these methods versus the traditional two-year rodent bioassay. Experimental work examining some of the above issues would assist in the evaluation of criteria by which specific assays are selected. For example, in order to address whether genotoxic potential should be considered in the selection of an assay (in particular, the p53+/- model), it would be useful to test more clastogenic compounds in this assay. The answer to the question of which alternative model to use will also become more clear as experience with the models grows.

Although further work is needed in the development of additional alternative models, the consensus from this workshop is that efforts should be focused on addressing some of the issues related to the more widely used alternative assays. In particular, questions related to the p53+/- and Tg.rasH2 models are considered the most important to address at this time.


    ACKNOWLEDGMENTS
 
The Utility of Transgenic Assays for Risk Assessment workshop was organized by the International Life Sciences Institute’s (ILSI) Health and Environmental Sciences Institute (HESI) Alternatives to Carcinogenicity Testing (ACT) Technical Committee. In addition to the authors of this paper, the ILSI HESI ACT Committee would like to acknowledge Dr. Christopher Portier of NIEHS for his contribution to the workshop, as well as the many workshop participants for their valuable input and critique. Acknowledgement and appreciation is also extended to the following organizations for their participation on the ILSI HESI ACT Committee: Aventis Pharmaceuticals, Bayer Corporation, Boehringer Ingelheim Pharmaceuticals, Inc., Bristol-Myers Squibb Co., Covance Laboratories, Dow Chemical Company, DuPont Haskell Laboratories, Eli Lilly & Co., Endo Pharmaceuticals, Genzyme Transgenics, GlaxoSmithKline, Hoffmann-La Roche, Inc., Merck and Co., Inc., Millennium Pharmaceuticals, Novartis Pharmaceuticals Corp., Pfizer Inc., The Procter & Gamble Company, Purdue Pharma L.P., R.W. Johnson Pharmaceutical Research Institute, Sanofi-Synthelabo, Inc., Schering-Plough Corporation, SmithKline Beecham Pharmaceutical, Solvay Pharmaceuticals, and Wyeth-Ayerst Research. ILSI HESI staff directly involved in this project include Dr. Amy Lavin and Ms. Cyndi Nobles.

ILSI HESI is a global branch of ILSI, a public, nonprofit scientific foundation that brings together scientists from academia, government, industry, and the public sector to solve problems with broad implications for the well being of the general public. ILSI HESI provides an international forum to advance the understanding and application of scientific issues related to human health, toxicology, risk assessment, and the environment. ILSI HESI is widely recognized among scientists from government, industry, and academia as an objective, science-based organization within which important issues of mutual concern can be discussed and resolved in the interest of improving public health. As part of its public benefit mandate, ILSI HESI’s activities are carried out in the public domain, generating data and other information for broad scientific use and application. ILSI HESI’s programs are supported primarily by its industry membership. ILSI HESI also receives support from a variety of government agencies from the United States and internationally. Further information about ILSI HESI can be found at http://hesi.ilsi.org/ or obtained by e-mailing hesi{at}ilsi.org.

The section in this article regarding Japanese Ministry of Health, Labor, and Welfare (MHLW) regulatory information is based on the Special MHLW Research Project for the development of a future strategy for the cancer bioassay.


    NOTES
 
1 To whom correspondence should be addressed at ILSI HESI, One Thomas Circle, 9th Floor, Washington, DC 20005. Fax: (202) 659-3617. E-mail: alavin{at}ilsi.org. Back


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 REGULATORY PERSPECTIVES
 AREAS OF AGREEMENT AND...
 ASSAY SELECTION ISSUES
 PROTOCOL ISSUES
 CONCLUSIONS
 REFERENCES
 
Cohen, S. M., Robinson, D., and MacDonald, J. (2001). Alternative models for carcinogenicity testing. Toxicol. Sci. 64, 14–19.[Abstract/Free Full Text]

Donehower, L. A., Harvey, M., Slagle, B. L., McArthur, M. J., Montgomery, C. A. J., Butel, J. S., and Bradley, A. (1992). Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumors. Nature 356, 215–221.[CrossRef][ISI][Medline]

French, J. E., Lacks, G. D., Trempus, C., Dunnick, J. K., Foley, J., Mahler, J., Tice, R. R., and Tennant, R. W. (2001). Loss of heterozygosity frequency at the Trp53 locus in p53-deficient (+/-) mouse tumors is carcinogen- and tissue-dependent. Carcinogenesis 21, 99–106.[CrossRef]

Morton, D., Alden, C. L., Roth, A. J., and Usui, T. (2002). The Tg.rasH2 mouse in cancer hazard identification. Toxicol. Pathol. 30, 75–79.[CrossRef][ISI][Medline]

Recio, L., Boley, S., Everitt, J., James, R. A., Janszen, D., Healy, L., Roberts, K., Walker, D., Pluta, L., and French, J. E. (2000). Cancer bioassay and genotoxicity of inhaled benzene in p53+/- and C57Bl6 mice. Toxicologist 54, 222.

Sistare, F. D., and Jacobs, A. C. (2003). Use of transgenic animals in regulatory carcinogenicity evaluations. In Alternative Toxicological Methods (H. Salem, and S.A. Katz, Eds.), pp. 391–412. CRC Press, Inc., Boca Raton, FL.

Storer, R. D., French, J. E., Haseman, J., Hajian, G., LeGrand, E. K., Long, G. G., Mixon, L. A., Ochoa, R., Sagartz, J. E., and Soper, K. A. (2001). p53+/- Hemizygous knockout mouse: Overview of the available data. Toxicol. Pathol. 29 (Suppl.), 30–50.[CrossRef][ISI][Medline]