Professor P Farmer

Dr B Burlinson
Dr G Clare
Dr J Clements
Dr D Gatehouse
Dr N Gooderham
Ms M Langley
Dr I Mitchell
Dr E Parry
Professor D Phillips


Dr R J Fielder (Scientific DH)
Dr D Benford (Scientific FSA)
Mr J Battershill (Scientific DH)
Mr S Robjohns (Minutes)
Mr K N Mistry (Administrative)


Dr D Andrew (PSD)
Dr A Smith (HSE)
Dr H Stemplewski (MHRA)
Mr A Browning (VMD)
Mr R Alexander (NAW)

In attendance :

Dr K Fletcher (DH Tox Unit)
Ms C Mulholland (FSA item 3)




Announcements/Apologies for absence



Minutes of the meeting 2 October 2003 (MUT/MIN/03/3)



Draft COM statement on the mutagenicity of trivalent chromium and chromium picolinate (MUT/04/6)



PAVA: consideration of new mutagenicity data (MUT/04/7)



2,3-Dichloropropan-2-ol: consideration of new mutagenicity 8 data (MUT/04/8)





6.1 Update on joint COT/COC/COM meeting October 2001 (MUT/04/1)
6.2 Application to mechanism of benzene induced Leukaemia (MUT/04/2)



Genotoxic carcinogens and DNA repair at low doses. Advice requested from COM (MUT/04/3)



Draft OECD guidelines on in vitro micronucleus tests: Use of cells or culture as statistical unit? (MUT/04/5)



Significance of paper by Thilly on environmental mutagens (MUT/04/4)



Holding Committee meetings in open session. Experience of COT (MUT/04/9)



Any other business



Date of next meeting – Thursday, 27 May 2004



1. The Chairman welcomed Ms C Mulholland (Food Standards Agency), Mr R Alexander (National Assembly for Wales) and Dr K Fletcher (DH Toxicology Unit).

2. Members were reminded of the need to declare any interests before discussion of items.


3. The minutes were approved with minor amendments.


4. The COM at its last meeting considered a request from the Food Standards Agency to review the available mutagenicity data on chromium picolinate. The Committee was asked what additional mutagenicity studies might be appropriate for this food supplement. A draft statement had been produced. The Committee considered that the in vitro studies had given some indication of mutagenic activity and recommended that the critical tests should be repeated using commercial grade material (the previously tested material had been synthesised by the testing laboratory) before any definite conclusions could be drawn. Thus a further in vitro hprt assay and an in vitro chromosome aberration study were recommended with chromium picolinate (both tests using CHO cells and to international standards) to determine whether the published results could be confirmed with a commercial grade sample. It was also recommended that the hprt assay should include a 48-hour incubation experiment in the absence of S-9 in addition to the standard time points of analysis.

5. Any decision on further in vivo genotoxicity tests would depend on the results obtained in the requested in vitro studies. Members were informed that the requested in vitro studies were currently being conducted and that the results should be available at the next COM meeting in May 2004.

6. Members agreed the text of the draft statement on the mutagenicity of trivalent chromium and chromium picolinate subject to minor amendments but would wait until they had seen the results of the further in vitro tests before finalising the statement.


7. The minutes of this item now published - see paragraph 45 onwards.


8. Dr Clements declared a non-personal interest. 2,3-Dichloropropan-1-ol (2,3-DCP) is a member of a group of chemicals called chloropropanols. It was last considered by the COM and the COC in 2001, together with 1,3-dichloropropan-2-ol (1,3-DCP). COM concluded that it would be prudent to consider 1,3-DCP and 2,3-DCP as potentially genotoxic in vivo but agreed that they should be tested for genotoxicity in vivo using the approach set out in the COM guidelines. The COM had considered the results of in vivo studies from a rat bone-marrow micronucleus test and a liver UDS assay conducted with 1,3-DCP at its May 2003 meeting and had reached a conclusion that there was no evidence for in-vivo mutagenicity in the tissues assessed. The Committee was asked to consider the data from two in vivo mutagenicity assays (rat bone-marrow micronucleus test and a liver UDS assay) conducted with 2,3-DCP.

9. The Chair noted that 2,3-DCP had structural alerts for mutagenicity and was an in-vitro mutagen. He noted that in contrast to 1,3-DCP there were no carcinogenicity bioassays available with 2,3-DCP and this compound also exhibited differences in metabolism compared to 1,3-DCP. He asked members for comments on the metabolic activation of 2,3-DCP and then for comments on the newly submitted in vivo mutagenicity tests.

10. Members agreed that 2,3-DCP was metabolised to 2,3-dichloroacetaldehyde and from this to the corresponding acid. One research group had provided some in vitro data to suggest that induction of CYP2E1 resulted in 2,3-DCP mediated hepatotoxicity and glutathione depletion. (Hammond AH et al Chemico-Biological Interactions, 122, 107-115, 1999). Members noted the findings of Koga (Journal Univ Occup Environ Health, 14, 13-22, 1992) which suggested dechlorination/hydroxylation occurred but agreed these authors had not provided evidence for the formation of an epoxide. Members concluded that there were insufficient data to draw conclusions on the metabolic activation of 2,3-DCP but overall the evidence suggested metabolic activation of 2,3-DCP differed from 1,3-DCP.

11. The Committee considered the new in vivo rat bone-marrow micronucleus test. Members observed that the analyses of the dosing solutions suggested the concentration of 2,3-DCP was ca 76-88% of the nominal value. It was noted that the analyses had been undertaken approximately 5 months after dosing. The study authors considered that the lower than nominal concentrations in dosing solutions could have been due to analytical methodology used or the degradation of the sample during storage. Members agreed that although the analysis of dosing solutions was not optimal, this had not compromised the study as clear signs of toxicity had been recorded in dosed animals. Turning to the results of the study, members noted signs of toxicity and evidence for reduced PCE/NCE ratios particularly at higher doses. The Committee agreed there was no need to request analysis of stored plasma samples from this study as evidence of bone-marrow effects had been demonstrated. It was agreed that there was no evidence for a mutagenic effect in this study.

12. Members reviewed the in vivo rat liver UDS assay. It was agreed that this assay conformed to the relevant international guidelines, although it was noted that for one animal, at 110 mg/kg bw in the 12-14 hour trial, the number of cells scored was below the recommended number. However there were sufficient cells scored from all other animals in the study. There was clear evidence of toxicity at the top dose level used. There was no evidence of an increase in Net Nuclear Grain counts in treated animals. Positive controls gave appropriate responses. Overall the Committee concluded that there was no evidence for a mutagenic effect in this study.

13. The Secretariat was asked to draft a statement for the next meeting.



14. The COT/COC/COM held a joint symposium on the use of genomics and proteomics in toxicology on 8 October 2001. A statement outlining the conclusions reached has been published on the Committee websites and a full write up of the meeting was published in Mutagenesis. 2003 May;18(3):311. There were relatively little data on screening for mutagenicity available at the October 2001 joint meeting. Since then, a number of papers had been published and some preliminary data from the ILSI/HESI trial had become available. Members were asked to consider the new information in the context of updating the joint COT/COC/COM statement.

15. The Chair asked members to consider the information from the ILSI trial first. He noted a further paper detailing some more information from the ILSI/HESI trial had just been published online (Newton RK, Environmental Health Perspectives, February 2004).

16. Members were aware that the ILSI/HESI working group on genotoxicity had been established along with the hepatotoxicity and nephrotoxicity working groups with the objective of initiating research which could be used to inform debate on the potential use of cDNA microarray screening for a number of toxicological mechanisms. The ILSI/HESI genotoxicity trial had investigated 11 mutagens, which were direct acting mutagens via a number of mechanisms. It was evident that much had been learned about analysis of transcriptomic data and the reproducibility of data during the study. Most participants had migrated to using the cDNA Affymetrix U34 system during the study. Very few results from the ILSI/HESI trial had been published. Members noted that it had been reported that data from cDNA microarray analyses could be used to differentiate between classes of direct acting mutagens in L5178Y mouse lymphoma cells. However, members felt that the reported results needed to be published in peer reviewed journals along with information to validate the results before any definite conclusions could be drawn.

17. Members queried the use of p53 status of the cell lines used and commented that it would have been preferable to undertake basic research in primary cultures (eg peripheral blood lymphocytes) rather than genetically altered cell lines. Members heard that the choice of L5178Y tk+/- mouse lymphoma cells had been taken pragmatically as there was more background information on mutagenicity in this assay system for the chemicals chosen than any other test system. It was noted that for benzo(a)pyrene diol epoxide the available information suggested that in L5178Y cells, DNA adducts and mutations were more sensitive indicators of chemical induced genetic damage than data on the number of genes for which there were significant changes in mRNA levels that were reported with cDNA microarrays.

18. Members observed that mutagenic activity might be associated with changes in the expression of relatively few genes. Hence it was possible that the use of global cDNA arrays which contained several thousand genes and/or expressed sequence tags was inappropriate as it would be difficult to separate gene expression changes relevant to mutagenicity from background "noise" changes. Members agreed that the predominant problems with use of high density global cDNA microarray platforms for mutagen screening concerned reproducibility of identifying gene changes (both inter and intra laboratory variation in analyses of mRNAs), the confirmation of key gene expression changes by quantitative analysis of mRNA levels, and the statistical and bioinformatic approaches to identification of relevant gene expression changes. Members queried whether a two-fold change in gene expression, as commonly used, was appropriate for mutagenicity screening. Members also commented that it would be useful to explore time-dependent changes in gene expression changes for novel mutagens.

19. Members discussed the appropriate use of low density and high density cDNA array platforms. Low density arrays with targeted genes were most likely to be useful for screening, but the drawback would be that previously unidentified mechanisms of genotoxicity could be missed. High density arrays would be most useful for identifying gene expression changes associated with new mechanisms and hypotheses for further investigation with regard to chemical induced carcinogenicity/mutagenicity. Members agreed it was essential that experiments presented validation data from quantitative RNA measurements for key genes. It was recognised that many of the most recent publications would present data from quantitative polymerase chain reaction (RT-PCR) estimations of mRNA levels and would be undertaken in accordance with the Minimum Information About a Microarray Experiment (MIAME, Members commented that further information on gene expression changes in response to mutagens would be derived from estimating the levels of proteins and their activity; eg measuring cascades in metabolic pathways or the phosphorylation status of proteins.

20. Members also considered that developments were needed in statistical handling of data and bioinformatic presentation of data. Thus for example, should there be an agreed minimum expression change for reporting results and would this vary depending on the toxicological mechanism under consideration? Members noted that many research groups were beginning to use Affymetrix gene chips, which could provide rodent and human mRNA expression information using the same platform. Members also commented on the need for validation of sequences incorporated onto chips.

21. Turning to the published papers which had also been submitted. Members commented on the work published by Seidel and colleagues (Environ and Mol Mutagenesis, 42, 19-25, 2003.) The chemicals under study (hydrogen peroxide and bleomycin) induced non-specific and localised oxidative DNA damage, but no evidence for gene expression changes in DNA repair or cell cycle control genes had been reported using the 1.2K Clontech cDNA microarray. Lee M and colleagues (Environmental and Molecular Mutagenesis, 42, 91-97, 2003) reported on common gene expression changes in L5178Y cell treated with three chemicals (hydroxyurea, p-anisidine and paclitaxel) with differing in vitro and in vivo mutagenic potential and evidence for carcinogenicity. Although it was claimed that the data supported both common gene expression changes and also differences in gene expression changes between these chemicals it was not possible to discern a pattern consistent with DNA damage for any of these chemicals. Members considered that both of these papers were limited by the use of L5178Y cells which carried mutations in the tk gene and possibly also in the p53 gene. In a further paper Ellinger-Ziegelbauer H and colleagues (Toxicological Sciences advanced internet publication) had investigated the gene expression changes at various intervals during a 14 day exposure to each of four hepatocarcinogens. Gene expression changes included up-regulation of p53 associated functions and one DNA repair gene (06-methylguanine- DNA methyltransferase). Members considered the results of this study, although preliminary, were encouraging and that a genotoxic response pattern for hepatocarcinogens could be identified in in-vivo experiments.


22. The Committee considered this paper (Yoon BL et al Env Hlth Perspec, vol 111, 1411-1420, 2003) as part of the review of toxicogenomics. Heterozygous p53 male and female C57BL6 mice were mated to produce wild type (WT p53+/+) and p53 Knock-out mice. Groups of mice were exposed to 300 ppm benzene for 6h/day for 5 days/week for 2 weeks. It is presumed that heterozygous p53+/- not p53-/- mice were used in this study but this isn’t clearly reported. Mice were sacrificed on day 5, 12 and after a further 3 day recovery period (day 15). Samples of liver were snap frozen at –800C and retained for RNA extraction and isolation. Gene expression was evaluated using Affymetrix (gene array) at day 12 and the Incyte GEM system (using cDNA) on days 5,12 and 15. The study was the first in the current COT/COC/COM review to demonstrate gene expression changes during exposure, those which occurred during dosing and which persisted during recovery and those which occurred following termination of dosing. The evidence that p53 gene was critical to the effects of benzene was difficult to assess and it was not possible to draw any conclusions. Members felt that additional studies which had used transgenic mice, where key genes in the metabolism of benzene had been switched off, provided more relevant information. Of interest were the data in male mice, which suggested that epoxide hydrolase was essential for haematotoxicity. These latter data, if confirmed, suggested that metabolism downstream from benzene epoxide might be important with regard to the haematotoxicity and potentially leukemogenicity of benzene.


23. The COC considered a horizon scanning paper by the DH Toxicology Unit at its June 2003 meeting. One aspect considered was hormesis (the occurrence of "U" shaped dose-response curves). During its discussions the COC considered a paper by Calabrese EJ and Baldwin LA (Nature, vol 421, 691-692, 2003), which provided an argument for the occurrence of hormesis (i.e. the occurrence of "U" shaped dose-response curves at low doses).

24. The COC conclusions included the recommendation that the evidence for the induction of increased DNA repair (above background levels and with the potential to reduce the number of spontaneous mutagenic lesions) following exposure to very low levels of genotoxic carcinogens warranted further review, as a potential mechanism that could result in a "U" shaped dose response curve; they had asked the COM for advice. Paper MUT/04/3 presented a brief overview of an initial literature search in this area, covering in vitro and in vivo data. Members were asked to advise on the most suitable approach to evaluating the significance of DNA repair induction at low doses in the context of the hormesis hypothesis. Also members were asked whether any further work was justified in this area in view of the fact that any impact on risk assessment was likely to be small.

25. The Committee considered the position paper (MUT/04/3) on genotoxic carcinogens and the impact of DNA repair at low doses. This paper provided an overview of the literature (based on abstracts only) and revealed that there were both in vitro and in vivo studies, which reported modulation (either up or down) of DNA repair with a variety of chemicals. The results varied on a case-by-case basis and in some instances error prone DNA repair was induced and in some cases DNA repair was inhibited. Some in vitro studies had suggested a threshold below which DNA damage is repaired. Several in vivo studies had used repair proficient/deficient animals (in transgenic animals) and claimed the occurrence of thresholds for mutation. Other in vivo studies had provided evidence for error prone induced DNA repair.

26. Members noted the uncertainty over the dose-response curve for cancer arising from very low doses of genotoxic carcinogens down to zero exposure. Most of the evidence for very low dose effects came from radiation biology which supported the absence of a threshold. However it was unlikely this would apply to all in-vivo genotoxins. The Committee considered that is was still prudent to assume that there is no threshold for mutation for in-vivo mutagens and where a potential threshold mechanism could be identified, appropriate evidence should be provided on a case by case basis.

27. The COM considered that it was important to keep the issue of the absence of a threshold under review. They felt that it would be useful to carry out a literature search targeted on low dose effects of a few direct acting chemical mutagens on DNA adduct formation, mutation rates, and the significance of DNA repair mechanisms. The search should concentrate on low molecular weight compounds such as ethylene oxide and ethyl or methyl methanesulphonate, for which there was a rich database. Members agreed that bacteria would most likely demonstrate the most sensitivity to low doses of mutagens and had reservations as to whether mammalian cell systems would have sufficient sensitivity to detect evidence for an effect of DNA repair induction on the dose-response relationship for mutation. The Committee considered that natural variability would make the detection of any ‘U’ shaped curve very difficult.

28. The secretariat would carry out the requested literature search and inform the Committee of its findings, for further consideration at the next meeting.


29. At the 3rd International Workshop on Genotoxicity Testing (IWGT) held in 2002 it was reported that the working group on the in-vitro micronucleus test had reached consensus that the method had been adequately validated. This potentially removed the block on the UK proposals in 2000 for an OECD guideline in this area. The COM was provided with details of the agreement on the methodological issues in October 2003. A cause for concern was the use of the cell and not the culture as the statistical unit. The secretariat wrote to the Chair of the IWGT working group expressing the COM’s concern over this aspect, but the final report maintained the original view of the cell being the statistical unit. The report argued that this is a similar position to the in-vitro mammalian cytogenetics assay, where the cell is the statistical unit. A subgroup of the COM had been formed to look at this issue. It noted that the current OECD guideline for the in-vitro cytogenetic assay used the cell as the statistical unit whereas the current OECD guideline for the mammalian cell gene mutation assay does not give any clear preference.

30. Members agreed that for observational purposes the unit was the cell; it is the number of cells with micronuclei that is recorded. However, the major source of variability is likely to be the culture, and this should be considered in the analysis. The following wording was agreed. ‘The observational unit is the cell, but the unit for statistical analysis is the culture.’ This will be included in the draft proposals for an OECD guideline.


31. At the COM October 2003 meeting a paper in Nature by Professor Thilly (Biological Engineering Division, MIT) had been tabled. One member had drawn the attention of the secretariat to this paper. It was agreed that detailed discussion of the significance of the paper would take place at the February 2004 meeting. The paper argued that the evidence for environmental mutagens causing cancer in humans is very weak and was limited to sunlight and ionising radiation. Thilly suggested that environmental risk factors for cancer act through non-mutational processes, and the genetic changes that cause cancer in humans arise from endogenous processes. The paper questioned the value of mutagenicity assay in screening for potential carcinogens.

32. Members agreed that the alternative hypothesis proposed by Thilly, namely that environmental risk factors for carcinogenesis acted not by mutation but by selecting particular kinds of preneoplastic cells previously initiated by spontaneous mutation resulting in a positive net growth rate, was interesting and warranted consideration. Regarding the most important environmental risk factor for lung cancer, cigarette smoking, Thilly pointed out that there was little evidence that smoking increased point mutations in bronchial epithelial cells or in peripheral T cells in humans, where the hprt locus had been examined. Members felt that the hprt assay in human lymphocytes may not be sufficiently sensitive or appropriate for detecting smoking induced mutations. Additionally, Thilly had not carried out a comprehensive review of this evidence and at least one report showing a significant increase in mutations in the hprt locus in smokers had been omitted.

33. The Committee considered that Thilly has been selective in his use of the available data. Although he had accepted ionising radiation and sunlight as exceptions to his alternative hypothesis, there was a considerable body of evidence supporting the somatic mutation mechanism for chemical carcinogenesis that had not been taken into account. Further, Thilly had concentrated on only one type of mutation (gene mutation) and had not considered other examples, such as structural or numerical chromosome changes. It was widely accepted that clastogenic events, for example, are seen in smokers, and that it is crucial to control for smoking habits in any cytogenetics study in humans. No mention was made of the dominantly inherited conditions (and the changes in genetic material involved) that predispose to cancer in humans. Thilly had ignored the fact that essentially all clear in-vivo mutagens in somatic cells had also been shown to be carcinogens in animals when adequately tested. The fact that the majority of chemicals recognised as being carcinogenic in humans are also in-vivo mutagens had also been ignored.

34. Members concluded that arguments in the Thilly paper were not convincing and there was no justification to depart from the somatic mutation theory of chemical induced cancer or to change the COM recommended strategy for testing chemicals for mutagenicity.


35. Members were informed in 2002 that the COT would discuss proposals for future meetings to held in open session. The COT agreed in principle to hold open meetings during 2003. A more detailed protocol was devised and discussed by COT in February 2003 (Annex A COT/2003/06). The COT agreed subject to minor changes that the protocol was adequate and that they would hold meetings in open session from April 2003.

36. It was agreed that the protocol would be discussed at future COM and COC meetings with a view to those Committees following the same procedures for holding meetings in open session.

37. When discussion by all three Committees was complete it was intended that a Joint COT/COC/COM code on Openness would be agreed. A revised code of practice would then be published in the annual report.

38. The FSA representative informed the COM that the COT had held six open meetings since April 2003. Advertisements for open sessions had appeared in FSA News and on the FSA website. No applications from the general public had so far been received.

39. External observers had attended open COT meetings. A member of the press attended the April meeting and in July a representative of the Institute of Optimum Nutrition with an interest in the agenda item on tryptophan, had attended the COT meeting. COT members had not felt inhibited by the presence of observers. Industry representatives with an interest in a discussion item had attended on two occasions, and had been helpful in providing additional clarification on certain facts.

40. It was noted that the COM might discuss more commercial in-confidence items than the COT. The Committee agreed that if a data holder had concerns over confidential information that the relevant item could be discussed in closed session.

41. COM members had some concerns over open meetings and questioned whether there were any measures in place to prevent the attribution of comments to individual members. It was noted that the draft protocol specified that it was important to clarify with observers at the start of the meeting (or item) what was expected of them and that comments should be attributed to the Committee as whole and not individual members. The Committee agreed that the draft code of conduct for observers should also describe the role of assessors.

42. Overall members acknowledged that there was potential for observers to make adverse comments or to distort members’ views. However, members agreed that the protocol used by the COT for holding meetings in open session could also be used by the COM and that in principle, future COM meetings could be held in open session. Members were also content for an appendix summarising the protocol, produced in conjunction with the COC and COT to be added to the Code of Practice on Openness prior to its publication in the 2004 Annual Report.


43. The Committee was informed that the web address for DH advisory Committees would be changing. Members would be informed of the changes in due course. It was noted that ‘doh’ in the URL would be replaced with ‘dh’.


44. Date of next meeting Thursday, 27 May 2004.



45. In 2001 the Home Office requested advice from the COT on the health effects of a chemical incapacitant spray based on pelargonyl vanillylamide (PAVA), a synthetic equivalent of capsaicin, present in pepper. It is a sensory stimulant and is also used both as a food flavour (at up to 10 ppm) and as a topical rubifaciant in human medicine. The COT requested the advice of the COM on a package of mutagenicity data that had been commissioned by Sussex Police to support the health risk assessment of PAVA. The COM considered these mutagenicity data at its meeting in February 2002. The COM agreed that some further reassurance was necessary, in addition to the negative in vivo bone marrow micronucleus tests that had already been conducted. The Committee requested an in vivo liver UDS assay to provide this necessary reassurance.

46. Sussex Police commissioned the requested study and a full report was provided in Annex B. An in vivo liver UDS assay was carried out in rats using oral exposure and dose levels of 625 and 1250 mg/kg body weight. The top dose was the maximum tolerated dose, as determined from a range-finding toxicity study.

47. Members agreed that the study conformed to the relevant OECD test guideline (OECD Guideline No. 486: UDS test with mammalian liver cells in vivo) and was adequate. 1,2- Dimethylhydrazine dihydrochloride was used as the positive control. There was no evidence of the induction of DNA repair by PAVA, as measured by unscheduled DNA synthesis in this assay. The positive control clearly induced DNA damage at both time points.

48. The COM agreed that the information sought by the Committee to conclude that PAVA was not expected to be an in-vivo mutagen had now been provided. No further data were required on the mutagenicity of this compound.

49. These conclusions will be incorporated in the overall conclusions drawn by the COT when they consider new toxicity data on PAVA (relating to sensitisation and developmental toxicity) later this year.



Item Action Who
3. Mutagenicity of chromium picolinate
Reconsider draft statement when new in vitro data submitted
5. 2,3-Dichloropropan-2-ol Draft statement Secretariat
6. Toxicogenomics Inform COC/COT of COM views Secretariat
7. Low dose genotoxic carcinogens Prepare paper for May

DH Tox Unit/secretariat

8. OECD Guidelines on in-vitro micronucleus test Write to OECD group on COM view Secretariat
10. Holding committee meetings in open session
Revise protocol Secretariat