Initial Development of Transgenic Mouse Models for Carcinogenicity Testing and a Review of the Regulatory Environment

Carcinogenicity studies are conducted to determine the tumorigenic potential of pharmaceuticals and chemicals, as required by regulatory authorities.1 Two-year bioassays in rats and mice have historically been used for carcinogenicity assessment. However, their relevance to human risk assessment is often questioned due to rodent-specific carcinogenic mechanisms and a very high incidence of background neoplastic and non-neoplastic lesions, which often confound the interpretation study results. The requirement of large quantities of each test article, the number of animals needed and average study costs exceeding $4 million have contributed to the dissatisfaction with two-year bioassays. Genetically engineered rodent models are attractive alternatives because they address all of these concerns.

Validation of genetically engineered mice for cancer hazard identification

Draft guidance for alternative short-term carcinogenicity tests involving transgenic mice was introduced in 1996 by the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH’s S1B).2

To fully characterize these models, the International Life Sciences Institute’s Health and Environmental Sciences Institute (ILSI HESI) formed a committee representing industry, regulatory agencies, government organizations and academic institutes. The committee formulated an international validation program for the Tg.rasH2 and Tg.AC transgenic mouse models, p53+/− and xpa–/– knockout mouse models, xpa–/–/p53+/− double knockout model and neonatal mouse assay.

Twenty-one compounds were tested in the ILSI HESI validation studies, including genotoxic and nongenotoxic human carcinogens, rodentonly carcinogens and noncarcinogens. Under contract with the National Toxicology Program, BioReliance (Rockville, Md.) was involved in testing many compounds in the p53+/− and Tg.AC models.

Conducted according to the ICH S1B guidelines, these studies ultimately resulted in the regulatory acceptance of Tg.rasH2, Tg.AC and p53+/− as primary models for carcinogenicity testing by the Committee for Proprietary Medicinal Products (CPMP), FDA and the Japanese Ministry of Health, Labour and Welfare.3,4 The Tg.rasH2, Tg.AC and p53+/− models subsequently became the most widely used models for carcinogenicity assessment.

The p53+/− model

The p53+/− model was created by Donehower et al.;5 these mice have one copy of the wild-type allele and one copy of a null allele. The hemizygous knockout animals have a very low incidence of spontaneous tumors up to nine months of age.6 At approximately 18 months of age, however, survival is reduced to about 50% due to an increase in the incidence of tumors.

Use of the p53+/− model is limited to mutagenic compounds because it has been shown to be ineffective for the identification of nongenotoxic carcinogens. Further, the model is not considered to be highly responsive for positive control material. For these reasons, the p53+/− model is no longer favored by the regulatory agencies.

The Tg.AC model

The Tg.AC mouse is a homozygous transgenic model carrying the v-Ha-ras oncogene fused to a mouse zeta-globin promoter. While not normally expressed in the skin, the transgene is significantly expressed in skin tumors. The skin of Tg.AC mice behaves as if genetically initiated, and develops papillomas and carcinomas in response to promoters, although the mechanism is poorly understood.7

Use of the Tg.AC model was limited to the assessment of pharmaceutical compounds intended for dermal administration. Due to a large number of false positive outcomes observed in studies submitted for regulatory approval, the Tg.AC model is no longer recommended as an alternative to the two-year bioassay.7 BioReliance is actively pursuing the possibility of using Tg.rasH2 mice as an alternative to the Tg.AC model for dermal carcinogens.

The Tg.rasH2 model

Tatsuji Nomura (1922–2013) and his team of scientists at the Japanese Central Institute for Experimental Animals in Tokyo originally developed the Tg.rasH2 transgenic mouse model.8 Tg.rasH2 hemizygous transgenic mice carry multiple copies of the human c-Ha-ras gene with their own promoter and enhancer.

The initial validation work for this model was mainly performed in Japan. However, since the acceptance of genetically engineered models for regulatory submission, use of Tg.rasH2 has increased significantly, and it now comprises more than half of the mouse carcinogenicity studies in the U.S.

Results from the ILSI HESI program and data generated from other Tg.rasH2 studies to date indicate that the model can be used for both genotoxic and nongenotoxic compounds. It is not susceptible to rodent carcinogens and has a very low false positive rate.9

BioReliance scientists have made significant progress in establishing the largest historical control database and improving the design of and methodologies for carcinogenicity studies in transgenic animals, specifically Tg.rasH2 mice.6,10-15

Advantages of the Tg.rasH2 model

There is considerable evidence that low false positive and false negative rates are produced with the Tg.rasH2 model.16,17 The combination of a two-year rat study with a transgenic six-month study produced zero false negatives, while a combination of two two-year chronic rodent bioassays in rats and mice produced a high number of false positives.18

Approximately 75% of positive neoplastic results obtained in two-year bioassays are considered to be irrelevant to humans.19 When 19 compounds were tested in the Tg.rasH2 model and the two-year rat bioassay, significant differences in outcome were obtained. Only two positive results were observed in the Tg.rasH2 model, both of which were expected based on mechanism, while 11 positive results were obtained in the two-year rat bioassay, nine of which were induced by rodent-specific mechanisms.19

The overall incidence of spontaneous tumors in six-month Tg.rasH2 studies was also considerably lower than that observed in conventional two-year mouse models. The percentage of tumor-bearing animals at study termination was close to 25% in Tg.rasH2, and approximately 70–80% for two-year mouse studies, respectively.6 BioReliance has also found that this model and study design yield a low mortality rate.

Consequently, as a humanized model for cancer risk assessment, the six-month bioassay using Tg.rasH2 mice provided a carcinogenicity assessment system with many advantages over the conventional two-year rodent assay. Furthermore, the six-month bioassay resulted in significant cost savings and flexibility during drug development by shortening the duration of carcinogenicity assessments, using fewer animals and reducing the amount of test articles and overall required resources.6,18

Regulatory requirements

The ICH S1B guideline, as adopted by the regulatory agencies in the U.S., Europe and Japan, calls for the use of two species for carcinogenicity assessment (rats and mice). Genetically engineered mouse models are now increasingly used as alternatives to the traditional two-year mouse bioassay.

Recently, there has been great interest in further refining the cancer risk assessment process based on a weight of evidence approach. Evaluation of available data by the ICH S1 expert working group (S1 EWG) has revealed that knowledge from pharmacologic mechanisms and toxicologic endpoints can provide sufficient information for predicting the outcome of two-year rat studies and their use in human cancer risk assessment.

Changes to ICH S1, which would use the weight of evidence approach, have been proposed and released for public comment. To evaluate the proposed changes to the ICH guidelines, data will be submitted from carcinogenicity assessment documents (CADs) for small-molecule compounds that require a two-year rat carcinogenicity study. Data collected from CADs will be used to formulate the changes to the ICH S1 guideline.

In addition to collecting information on two-year rat bioassays, pharmaceutical companies and BioReliance have suggested that more emphasis be placed on study data from transgenic mouse carcinogenicity studies. Based on these proposals, data from a six-month transgenic study may provide sufficient information to allow for the assessment of carcinogenicity in instances in which a two-year rat study would not be needed.

Based on current timelines, updates to guidelines should be available in 2017. If approved, it is expected that a two-year rat carcinogenicity study will not be needed for at least 40% of compounds, while the use of the Tg.rasH2 mouse model will significantly increase.

Advancement of carcinogenicity testing

BioReliance scientists have dedicated significant resources to increasing the understanding of Tg.rasH2 mouse pathology and refining study protocols and methodologies. The largest historical control database of neoplastic lesions and the only historical control database for nonneoplastic lesions seen in Tg.rasH2 mice were published as a result.6,10-15 The BioReliance historical control database now includes more than 1000 male and 1000 female Tg.rasH2 mice.

The historical control database is used in conjunction with statistical analyses to interpret the biological significance of tumor responses and reduce false positive and false negative results.6 In specific treatment groups, tumors may show statistically significant, albeit small, increases in incidence relative to controls. In these instances, defining the biological significance of these tumors may only be achieved by comparing their occurrence to the incidence seen historically, which can be attained through the use of a large database of study results. Regardless of the results of statistical analyses, a drug- or chemical-induced effect may be verified or nullified based on a comparison with the historical control range for that tumor type.

BioReliance scientists have also reported data on:

  • The importance of body weight parameters
  • Trend analyses
  • Reduction in the size of positive control groups
  • Evaluation of select tissues with respect to the Tg.rasH2 model
  • Recommendations for regulatory changes on dose selection for 26-week carcinogenicity studies in Tg.rasH2 mice.

Additional studies related to the performance of carcinogenicity studies are ongoing.

References

  1. Development & Approval Process. www.fda.gov/drugs/developmentapprovalprocess. Last accessed June 2, 2014.
  2. ICH Harmonised Tripartite Guideline. Testing for Carcinogenicity of Pharmaceuticals (S1B) 1997, July 16. Last accessed June 2, 2014.
  3. Yamamoto, S.; Urano, K et al. Validation of transgenic mice carrying the human prototype c-Ha-ras gene as a bioassay model for rapid carcinogenicity testing. Environ. Health Perspect. 1998 Feb, 106 suppl 1, 57–69.
  4. Morton, D.; Alden, C.L. et al. The Tg rasH2 mouse in cancer hazard identification. Toxicol. Pathol. 2002 Jan-Feb, 30(1), 139–46.
  5. Donehower, L.A.; Harvey, M. et al. Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours. Nature 1992 Mar 19, 356(6366), 215–21.
  6. Paranjpe, M.G.; Elbekai, R.H. et al. Historical control data of spontaneous tumors in transgenic CByB6F1-Tg(HRAS)2Jic (Tg.rasH2) mice. Int. J. Toxicol. 2013 Jan-Feb, 32(1), 48–57.
  7. Hansen, L.A.; Spalding, J.W. et al. A transgenic mouse model (TG.AC) for skin carcinogenesis: inducible transgene expression as a second critical event. Prog. Clin. Biol. Res. 1995, 391, 223–35.
  8. Saitoh, A.; Kimura, M. et al. Most tumors in transgenic mice with human c-Ha-ras gene contained somatically activated transgenes. Oncogene 1990 Aug, 5(8), 1195–2100.
  9. Cohen, S.M. Alternative models for carcinogenicity testing: weight of evidence evaluations across models. Toxicol. Pathol. 2001, 29 suppl, 183–90.
  10. Morton, D.; Alden, C.L. et al. The Tg rasH2 mouse in cancer hazard identification. Toxicol. Pathol. 2002 Jan-Feb, 30(1), 139–46.
  11. Paranjpe, M.G.; Denton, M.D. et al. Regulatory forum opinion piece: retrospective evaluation of doses in the 26-week Tg.rasH2 mice carcinogenicity studies: recommendation to eliminate high doses at maximum-tolerated dose (MTD) in future studies. Toxicol. Pathol. 2014 Nov 11. In press.
  12. Paranjpe, M.G.; Denton, M.D. et al. Relationship of body weight parameters with the incidence of common spontaneous tumors in Tg.rasH2 mice. Toxicol. Pathol. 2014 Oct, 42(7), 1143–52.
  13. Paranjpe, M.G.; Shah, S.A. et al. Incidence of spontaneous non-neoplastic lesions in transgenic CBYB6F1-Tg(HRAS)2Jic mice. Toxicol. Pathol. 2013, 41(8), 1137–45.
  14. Paranjpe, M.G.; Denton, M.D. et al. Trend analysis of body weight parameters, mortality, and incidence of spontaneous tumors in Tg.rasH2 mice. Int. J. Toxicol. 2014 Sep. 26. In press.
  15. Shah, S.A.; Paranjpe, M.G. et al. Reduction in the number of animals and the evaluation period for the positive control group in Tg.rasH2 short-term carcinogenicity studies. Int. J. Toxicol. 2012 Sep-Oct, 31(5), 423–9.
  16. Usui, T.; Mutai, M. et al. CB6F1-rasH2 mouse: overview of available data. Toxicol. Pathol. 2001, 29 suppl, 90–108.
  17. ICH S1: Rodent Carcinogenicity Studies for Human Pharmaceuticals— 2012 Nov 14.
  18. Jacobson-Kram, D. Cancer risk assessment approaches at the FDA/ CDER: is the era of the 2-year bioassay drawing to a close? Toxicol. Pathol. 2010 Jan, 38(1), 169–70.
  19. Morton, D.; Sistare, F.D. et al. Regulatory forum commentary: alternative mouse models for future cancer risk assessment. Toxicol. Pathol. 2013 Aug 21, 42(5), 799–806.

Scott Hickman is global market segment manager for toxicology, BioReliance (part of Sigma-Aldrich), 14920 Broschart Rd., Rockville, Md. 20850-3349, U.S.A.; tel.: 301-738-1000; fax: 301-610-2590; e-mail: [email protected]; www.bioreliance.com. BioReliance has more than 15 years of experience working with transgenic models for carcinogenicity assessment. Through the National Toxicology Program, the company was involved in the initial validation of the p53+/− and Tg.AC models and, later, validation of the Tg.rasH2 model. The validation work was part of the International Life Sciences Institute (ILSI) program, the results of which led to changes in the International Conference on Harmonisation (ICH) guidelines and acceptance by international regulatory bodies of these models as alternatives to the two-year mouse bioassay.