COC AD-HOC PAPER
ILSI'S COLLABORATIVE RESEARCH PROGRAM ON ALTERNATIVE MODELS FOR CARCINOGENICITY ASSESSMENT
|COMMITTEE ON CARCINOGENICITY OF CHEMICALS IN FOOD, CONSUMER PRODUCTS AND THE ENVIRONMENT
PRESENTATION TO COC ON 24 JUNE 1999 BY DR DENISE ROBINSON (HEALTH AND ENVIRONMENTAL SCIENCES INSTITUTE (HESI), INTERNATIONAL LIFE SCIENCES INSTITUTE (ILSI))
1. The Chairman (Professor Blain, University of Newcastle) welcomed Dr Robinson to the meeting. In 1997, the Committee had reached a number of conclusions on the utility of short-term carcinogenicity testing using transgenic animal models following an assessment of the available published literature (Mutation Research 403, 259-263, 1998). Dr Robinson had agreed to provide the Committee with an overview of the HESI program of work and the likely times for completion of the work. He explained that questions raised during and at the end of the meeting would be recorded in the report of the meeting.
2. Dr Robinson thanked the Committee for the invitation to make a presentation. Members were told that ILSI was a non-profit scientific research and education foundation with a number of work programs related to public health. Each program involved participation from industry, government and academia and the topics identified for research programs were those most likely to have an impact on public health. HESI was established in 1989 with the aim of investigating pathology, toxicology, human health and environmental risk issues. HESI identified the utility of short-term animal carcinogenicity tests as a key topic for discussion following the adoption by the International Conference on Harmonisation (ICH) of guidelines for the carcinogenicity testing of pharmaceuticals. These included the option for a short or medium term rodent test in addition to a long-term carcinogenicity test in a rodent species instead of requesting long-term carcinogenicity tests in two species. HESI established the Alternatives to Carcinogenicity Testing (ACT) committee in June 1996 as a steering group for this work program. It had been established that there was an inadequate database on performance of alternative assays, there was no accepted rationale for selection of a model for a second test and there were major differences between regulatory authorities in North America and those in Europe and other areas on the acceptability of the new models.
3. ACT set a number of objectives for the work programme, which included expanding the database on the alternative methods, establishment of common experimental approaches, information on the scientific basis of the new models and the need for public debate on the evolving state of the science. Advisors were appointed to the work programme who had been involved in the early stages of the test model development. Fifty-one laboratories from around the world, with approximately equal numbers from U.S.A, Europe (including U.K.) and Japan agreed to take part. The outline of the program was considered which included the following issues; i) identification of assay systems to be evaluated, ii) establishment of standardised protocols, iii) selection of test agents, iv) data collection systems, v) data review process, vi) joint data analyses. Some comments made by Dr Robinson on each of these six issues are given below :
Assay systems to be evaluated
4. The test systems selected for evaluation were those identified in the ICH guidelines.
Establishment of standardised protocols
5. Dr Robinson noted that transgenic mice cost on average approximately five times more than wild type mice. Given the scale of the HESI program, a number of pragmatic decisions were made regarding protocols. Thus three dose levels were studied per chemical and the number of animals used per dose level was 15 males and 15 females. A vehicle and high dose control in wild type animals were also included. Information from NTP bioassays and a 4-week range finding assay were used to assist in dose selection. In most instances this has resulted in adequate survival. One exception was chloroform where a number of animals had died at the high dose level within the 26 week-dosing period. Possible differences between wild type and transgenic mice with respect to the metabolism of chloroform were to be investigated. Expanded histopathology is being undertaken on control and high dose animals, with tissues from animals in the intermediate dose group retained. More recently it had been agreed that histological analysis of the jaw would be undertaken in Tg.AC mice due to several instances of odontoma (a rare tumour in mice) in previous studies. Genotyping of tumours will be undertaken.
Selection of test agents
6. HESI had selected well-studied compounds with extensive human data, predominantly pharmaceuticals.
Data collection systems
7. An Assay Working Group (AWG) had been established for each of the models under consideration in order to establish criteria for evaluating tests, to ensure a common format for data collection and consistency in histopathology. The AWGs would also act as a focal point for data review and analyses.
Data review process
8. The AWGs suggested additional areas for research. Proposals were reviewed from approximately 30 laboratories in respect of drug metabolism in transgenic mice, mutational analyses of tumours, tissue specificity of mechanisms, and compound specific mechanisms. Thus proposals funded in 1999 included, i) mechanisms of hormonal carcinogenesis in transgenic mouse models, ii) an evaluation of drug metabolising enzymes in transgenic and parental strains of mice, and iii) mechanism of carcinogenicity in p53 mice (tumour analysis).
Joint data analyses
9. Data review(s) were expected in late 1999/early 2000, Histopathology database development by mid 2000. Meetings of the Joint Steering Committee/AWGs in early 2000. Data analyses by around June 2000. An open public workshop was planned in Washington DC around November 2000 to tie in with the simultaneous ICH meeting.
10. The ICH guidelines had made an impact on the strategy for carcinogenicity tests being adopted by pharmaceutical companies. Information from the Centre for Medicines Research showed that 18 out of 38 companies asked were using alternative short-term tests. Out of the 49 companies using short-term tests, the predominant models being used were the p53 knock-out (26.6%) and rasH2 (20.4%), although tests using rat initiator/promoter (18.4%), neonatal rodent assay (16.3%), Tg.AC (12.2%), and the XPA deficient (6.1%), mouse models were being undertaken. Members heard that some data were available to show a drift in the response of Tg.AC hemizygous mice to the positive control (12-O-tetradecanoylphorbol 13 acetate) TPA. Southern Blot analysis of responder and non-responding Tg.AC mice showed evidence for a 2kb deletion in the transgene. It was not clear how prevalent the transgene instability was in earlier tests, although action was taken to re-derive the breeding colonies of mice. Dr Robinson noted that the transgene deletion in Tg.AC mice had been detected before any significant impact in the HESI program could have occurred.
11. The Chairman thanked Dr Robinson for her presentation and said that a summary would be published on the COC Internet site.
COC questions and discussion
12. Dr Venitt (Institute of Cancer Research, Sutton) asked what criteria had been used in the selection of the mouse strain used to generate transgenic animals. Dr Robinson noted that the p53+/- transgenic mice were developed at Baylor College of Medicine from the C57BL strain presumably as there was considerable background history in animal husbandry and response of this strain to carcinogens. She acknowledged that other strains of mice and transgenic models could be developed which would have specific uses, for instance models for specific carcinogenic mechanisms. Members agreed it was important not to confine developmental work to the models already under consideration. Professor Davies (Imperial College of Science Technology and Medicine, London) asked whether dosing transgenic animals up to the maximum tolerated dose (MTD) would inevitably repeat many of the problems that had been experienced with the NTP programme. Dr Robinson agreed that this might be the case, but there was sensitivity to the regulatory agencies expectations that studies yield comparable data to the available long-term carcinogenicity bioassays. The ICH guidelines on dose selection were the official guidance. Professor Renwick (University of Southampton) asked whether tumour yields were expected to be comparable to those obtained in long-term carcinogenicity bioassays. Dr Robinson said that this aspect would be considered in the data evaluation conducted during the HESI/ILSI programme.
13. Professor Boobis (Imperial College of Science Technology and Medicine, London ) noted that the skin of wild type (parent strain FVB/N) and Tg.AC mice were fundamentally different. He considered that little value could be ascribed to vehicle and high dose control studies in wild type animals as presumably no response would be expected in the parental strain. Dr Venitt (Institute of Cancer Research, Sutton) asked if the statistical power of the standardised protocol to detect carcinogens had been evaluated. Dr Robinson said that a fair amount of time had been spent considering this question. The calculations indicated that most genotoxic carcinogens should be detected using group sizes of 10 using the MTD. A pragmatic decision had been taken to use three groups of 15 male and 15 female animals, which should allow for the detection of most carcinogens rather than a higher number of animals but only two dose groups. Dr Sims (Medicines Control Agency, London) asked about the extent to which the selected test chemicals would be evaluated in all the test models under consideration. Dr Robinson considered that most (ca >90%) of chemicals would be tested in each of the test systems. The lowest number of chemicals being tested was in the Tg.AC mouse model. She noted that approximately 20% of chemicals were being tested in duplicate experiments at two or more laboratories, which would provide some data on interlaboratory performance of the test systems. Dr Stemplewski (Medicines Control Agency, London ) noted that intraperitoneal dosing was the traditional route of administration for the neonatal mouse model and asked about the extent of oral dosing that was planned. Dr Robinson indicated that the route of administration in the HESI studies was by oral gavage. Dr Robinson also noted that HESI was aware of the COC views on this model which had been placed on the Internet.
14. Dr Robinson noted that the Assay Working Groups would discuss generic issues regarding each test system when appropriate and before the overall data evaluation, for example the p53 group was discussing data on some chemicals, which had been recently tested. Professor Purchase (University of Manchester) asked what goals HESI expected to reach at the planned public workshop in November 2000. Dr Robinson considered that the HESI test programme would have made a significant contribution to the science of carcinogenicity testing if the results clearly identified the response profile of each of the models under evaluation and their use in a testing strategy for carcinogenicity. Dr Venitt noted this comment but emphasised that the Assay Working Groups would have to keep a close review on the test methods adopted in order that any deficiency in the generic standards prescribed before testing started did not end up in invalidating individual tests with a consequent loss of valuable data. Professor Boobis (Imperial College of Science Technology and Medicine, London) noted the problems with loss of transgene from Tg.AC hemizygous mice and asked whether the response to the positive control TPA and the genotype of animals in studies completed to date had been checked. Dr Robinson confirmed that this had been undertaken.
15. Members agreed that the HESI programme would yield a considerable amount of useful data regarding a better understanding of the scientific basis and value of short-term tests for carcinogenicity in transgenic animals. They looked forward to receipt of data from the HESI programme for their consideration. Members thanked Dr Robinson for the additional data on current strategies being adopted by the pharmaceutical industry and the recent data on positive control response of TPA in Tg.AC mice.
Secretariat October 1999