Identification and Evaluation of Potentially Genotoxic Agricultural and Food-related Chemicals

Abstract

The Food Safety Commission (FSC) was founded in 2003 to conduct the risk assessment of chemicals in food and food products and also residues of agricultural chemicals. Genotoxicity assessment is one component of the overall risk assessment process. Historically, genotoxicity assessment has been limited mainly to qualitative hazard identification. We are proposing a strategy for when the chemical is classified as a genotoxic carcinogen and the acceptable daily intake (ADI) cannot be set because a worldwide consensus has not been obtained on the existence of threshold for DNA direct-acting genotoxicity. To evaluate the mechanism(s) of carcinogenicity, it is important to make judgment whether genotoxicity, especially genotoxicity/mutagenicity resulting from direct reaction with DNA, is a key event or not in the carcinogenic process. Here, we focus on the residues of agricultural chemicals and discuss the strategy of how to evaluate and interpret genotoxicity, and provide guidance that we can use at the site of assessment. This paper presents the authors’ personal opinion and it does not necessarily represent the official opinion of the FSC. There are four independent expert working groups in the Expert Committee for evaluation of agricultural chemicals and the authors hope this paper will help to make evaluation fair and transparent across the working groups. Of course, other strategies to evaluate genotoxicity of food and food related chemicals, including residues of agricultural chemicals may also exist, and they should also be appreciated. The goal is scientifically sound, transparent, and fair evaluation and interpretation of genotoxicity, as an integral part of the risk assessment.

Introduction

The Food Safety Commission (FSC) was founded in 2003 to make risk assessment of food and related chemicals, independent from risk management. Before that time, mainly the Japanese Ministry of Health, Labour and Welfare (MHLW: former Ministry of Health and Welfare) conducted risk assessment in conjunction with risk management of chemicals and covered almost all areas including pharmaceutical drugs, agricultural chemicals, and food and food related chemicals. A principal aim of the FSC was to make risk assessment independent from risk management.

Evaluation and interpretation of chemical safety is generally based on the toxicological tests to identify hazards, including carcinogenicity, mutagenicity, reproductive toxicity, neurotoxicity, and other specified toxicological endpoints. Among them, carcinogenicity, mutagenicity, and reproductive toxicity are regarded as important hazards for assessing human health assessment and called “CMR effects”.¹⁾ Generally, CMR effects are well evaluated especially in the case of food and related chemicals, e.g., food additives, pesticides and veterinary drugs²⁾, which are consumed every day unlike pharmaceutical drugs which are used intermittently in most cases.

Mutagenicity is usually defined as induction of a heritable alteration in DNA, usually gene mutation or chromosomal aberration. Mutagens may damage DNA directly or indirectly, or may modify the mitotic apparatus with a resulting infidelity of maintenance of the genome balance in cells. Such damage is often a key initial event of carcinogenicity and termed initiation, and also has the potential to cause heritable adverse events in subsequent generations. When the chemical of interest has not been evaluated for carcinogenicity, then mutagenicity can be used as a predictor of carcinogenic potential. When a chemical is carcinogenic in experimental animals, then assessment of mutagenic activity is an important component of evidence used to determine the mechanism of the carcinogenicity. If the chemical does not show any mutagenic potential, then carcinogenicity must be induced by a non-mutagenic mechanism—and non-mutagenic mechanisms are generally considered to exhibit a threshold exposure level below which carcinogenicity does not occur. Because mutagenesis that results from direct modification of DNA is generally considered to lack a threshold (i.e., low exposures are considered to still carry some level of risk), determination of a mutagenic mode of action is a key element of cancer and reproductive risk assessment. The definition of a genotoxic carcinogen was proposed in the evaluation guideline of the food additive evaluation group of FSC as follows: “A genotoxic carcinogen is a chemical or its metabolite that reacts directly with DNA resulting in gene mutation or chromosomal aberration and the genotoxic effect(s) is considered to be the mechanism, at least a part of the mechanism, of carcinogenicity. It is necessary that genotoxicity is confirmed in vivo, preferably at the target organ of carcinogenicity.”

Gene mutations and chromosomal aberrations are fixed lesions in the DNA that can be transmitted to future generations of cells or organisms. The initial event of mutation is the interaction between DNA and the molecule of interest. The interaction is based on a probability that depends on the densities of both molecules. This may be the reason why mutation does not have real threshold concentration below which the event does not occur. DNA damage, adduct formation, DNA repair in relation to the DNA damage, and sister-chromatid exchange may be prior events to the mutation and many times such events are included in the definition of genotoxicity.

Genotoxicity is assessed by many test systems designed to detect specific endpoints. Genotoxic activity can be classified by the endpoint and the test systems, e.g., mutation, chromosomal aberration, DNA adducts, etc. in vitro or in vivo. In vitro assay systems include tests using microbial systems, for example Ames gene mutation (Salmonella/microsome) assay is the most well-known, and is widely used for initial assessment of chemical genotoxicity. The in vitro assay is more hazard identification oriented while the in vivo assay systems can provide information more relevant to assessment of human safety. Historically, the assessment of genotoxicity has been made qualitatively, just classification as positive or negative regardless of the potency of the effects, and used with the objective of hazard identification. Currently, the field is moving toward a more quantitative risk assessment that uses exposure-response relationships to determine the safety to humans based on the evaluation of exposure in relation to mutagenic potency.

Role of FSC

The FSC prepares dossiers of the safety of food related chemicals based on animal and some in vitro studies, and determines the acceptable daily intake (ADI) by voluntary specialists from each field. Risk assessment is conducted by the process of hazard identification, dose-relationship analysis, exposure analysis, and risk characterization. The relationship among FSC, MHLW, and the Ministry of Agriculture, Forestry and Fisheries (MAFF) and other related Ministries is described at the following website: http://www.fsc.go.jp/english/aboutus/roleofthefoodsaftycommission_e1.html. The FSC makes the exposure analysis based on, for example, the residue levels of pesticides from the test field and by national investigations of food consumption. After the report of safety information and of the ADI of chemicals to the MHLW, which decides how to manage to use and report back on the intake of the chemical relative to the ADI based on the estimated human consumption, e.g., for infants, adults, and aged people. Therefore, the FSC is able to understand the exposure levels of agricultural chemicals and make risk communications based on the risk characterization.

There are four independent Expert Committees for evaluation of agricultural chemicals under umbrella of the Executive Committee, one of which purpose is to make balance among four Expert Committees. The authors hope this paper will be a guidance to help makings evaluation fair and transparent across the working groups.

Basically, the FSC evaluates the safety of material as a single chemical and not as a complex mixture, to which people usually are exposed. There remains a great need to assess complex mixtures or combinations of chemicals to which humans are exposed. To consider such multiple chemicals exposure, the concept of threshold is important to evaluate overall risk if each chemical is evaluated separately, because, if each chemical has no threshold genotoxic effects the effects should be accumulating,and may reach the virtual human risk. Another important and difficult issue is the reliability of extrapolation from unrealistically high concentrations/dose experiments to the actual human exposure level. Generally, laboratory assays for genotoxicity are conducted at quite high concentrations/dose levels because of technical considerations.

Evaluation of the safety of agricultural chemicals is not simple. People usually consume vegetables, for example, after careful washing and cooking (heating). Therefore, the agricultural chemical and probably metabolite residues are diminished during preparation. Thus, actual exposures are often overestimated. This provides an additional margin of safety between our actual consumption and the estimated exposure, in relation to the ADI.

Genotoxicity Test Systems

To detect genotoxicity of chemicals many test methods have been developed and used in both research and regulatory science fields. There are two principal endpoints thought to be related directly to health risk—gene mutation and chromosomal aberration—among other supportive endpoints. The most important secondary endpoint is DNA damage that can lead to mutations or chromosomal aberrations. There is no assay system that detects all endpoints in one experiment. Therefore we use several assay systems to cover at least two major endpoints. This evaluation strategy is called a “battery” approach and has been accepted in the regulatory science field to avoid making false negative conclusions for the chemicals. Table 1 shows representative assay systems classified by endpoint and test materials.

Table 1.Representative Test Classified by Endpoints

	DNA damage	Gene mutation	Chromosomal aberration
In vitro	Rec assay Umu assay 32P-post labeling assay Alkaline elution assay Comet assay	Microbial reverse mutation assay (OECD TG 471)* HPRT gene mutation assay using Chinese hamster cells (OECD TG476) Mouse lymphoma TK assay (MLA, OECD TG is preparing)	Metaphase analysis using cellLine or primary culture (OECD TG473)In vitro micronucleus assay(OECD TG478)Mouse lymphoma TK assay(MLA, OECD TG is preparing)
In vivo	Unscheduled DNA sysnthesis Assay (OECD TG486) Alkaline elution assay comet assay (OECD TG is preparing)	Transgenic animal mutation model (OECD TG488)	Bone marrow metaphase analysis (OECD TG475) Rodent micronucleus assay (OECD TG474)

* Assay with underline is a component of standard test battery in fields of pesticides, agricultural chemicals, and pharmaceuticals

An important consideration is the choice of test systems. One major classification is in vitro vs. in vivo assays. Gene mutation was originally recognized and evaluated using plants by Mendel, and much basic mutation research has been conducted using microbial assays and plant cells. In the field of regulatory science, Bruce N. Ames developed the assay system to detect mutagenic chemicals using Salmonella typhimurium TA strain series in combination with an exogenous metabolic activation system to mimic mammalian metabolism (OECD TG471)³⁾. These tester strains were modified to increase their sensitivity to detect either base-change gene mutation or frame-shift mutation. The so-called Ames assay is performed in Petri plates, and so we call this an in vitro assay although it uses a whole organism (bacterium). To detect another endpoint, chromosomal aberrations, cultured Chinese hamster cells are frequently used because of the small number of large chromosomes, which can be easily scored for chromosomal aberrations. Recently, the selection of target cells has become an important issue for regulatory science because it has been recognized that the species or origin and also the status of the p53 gene can affect sensitivity/specificity⁴⁾. The currently employed in vitro assay systems have high performance and relatively low cost, and have proven valuable as reliable screening methods to detect mutation.

There is a current trend toward more extensive use of exposure-related risk assessment of genotoxic agents, and therefore more emphasis is being placed on results from the in vivo assay systems. The rodent micronucleus assay to detect chromosomal aberrations is well-established in the field of regulatory science^5,6) and has been incorporated into the standard battery for assessing chemical mutagenicity. Although the micronucleus assay has been widely used, the target organ has been limited exclusively to hematopoietic cells; e.g., bone marrow polychromatic erythrocytes and peripheral reticulocytes. The most important role of genotoxicity evaluation is to identify the mechanism of carcinogenicity and to determine whether it is dependent on genotoxicity or not. Accordingly, the use of only hematopoietic cells is insufficient to answer this question. The transgenic animal models (OECD TG488)⁷⁾ and the single cell gel electrophoresis assay (comet assay)(OECD TG in preparation) have become well recognized and are widely used because these assays can evaluate essentially any tissue of interest. Recently, the micronucleus assay has been extended to many organs (e.g., testes, skin, liver, digestive tract)⁸⁾. Thus three important endpoints of genotoxicity can be covered by in vivo assay systems: the comet assay detects DNA damage, the transgenic animal model detects gene mutation, and the micronucleus assay detects chromosomal aberrations. Because these assays can be applied to the organ that is the target for carcinogenicity, the genotoxicity assay can play an important role for chemical safety assessment for humans.

Although the importance of the in vivo assay systems is recognized, another important issue is animal welfare. Therefore, it is important to seek a balance between these two important issues, and attention has been given to minimizing animal use while at the same time obtaining the data necessary for risk assessment⁶⁾. Examples include: 1) The micronucleus assay using a tiny amount of peripheral blood without killing experimental animals can provide time-course information after treatment⁹⁾. This method can reduce animal usage. 2) Omission of the concurrently treated positive control animals in the assay is appropriate when a laboratory has demonstrated experience and pre-prepared scoring controls are included. 3) Several endpoints may often be combined into one assay; for example, the micronucleus assay and comet assay sharing animals for both endpoints¹⁰⁾. This also reduces animal use. 4) The possible ultimate method is the integration of genotoxic endpoints into the general toxicological study¹¹⁾. This not only reduces the number of animals to be used but improves the overall assessment and interpretation of chemical safety to humans by providing toxicity, exposure, and metabolic information in the context of the genotoxicity assay.

Assay systems used for regulatory safety assessments should be well validated and widely used for a long period. On the one hand, standardization of methods for research oriented studies is not critical. On the other hand, for regulatory science, the assay methods to be used must be well-characterized and standardized, including characterization of limitations and the possibility of false-positive/false-negative results. Validation studies should be designed based on the established strategy and should evaluate the technical aspects of the method, intra- and inter-laboratory reproducibility, and performance to detect genotoxic chemicals and also not to detect non-genotoxic chemicals^12,13).

Strategy of Assessment and Interpretation

1) Data Quality

At the FSC, risk assessment of food and food related chemicals, including agricultural chemicals, is routinely undertaken. One of the most important criteria is that the assessment should be done using high quality test data, especially when the data are obtained from the public literature. Adherence to test guideline (TG) and good laboratory practice (GLP) cannot fully guarantee the quality of data, but these are minimum requirements of qualified data. In the case of assessment of new agricultural chemicals, the majority of data are obtained using TG and done under conditions of GLP compliance, while in the case of food additives, it is often necessary to use published data obtained without consideration of TG nor GLP. Moreover, some chemicals may need to be assessed without any final reports of tests on the chemicals, using only the evaluation reports conducted by overseas authorities.

2) Expert Judgment and Low-dose Assessment

Some low level food additive chemicals, especially flavoring chemicals that have been evaluated and used in overseas countries, require the use of a strategy that uses only information about structural alerts, mode of action (MOA), and overall weight of evidence (WOE)¹⁴⁾ together with an established threshold of toxicological concern (TTC)^15,16,17). It is best to conduct risk assessments that consider mechanisms of all toxicological endpoints, referred to as mechanistic risk characterization. It is, however, usually difficult to explain all toxicological effects by credible mechanisms. It is also difficult to assess weak/marginal effects biologically, although statistical evaluation can be done especially at very high dose levels. For risk assessment, exposure analysis is important especially at low dose levels, to establish an acceptable Margin of Exposure (MOE)¹⁸⁾.

3) Threshold

One important consideration in genotoxicity risk assessment is the theoretical lack of threshold for DNA-reactive chemicals. Theoretically, reactivity is based on the probability of collision between DNA and the chemical substance. This cannot result in a “0” probability of introducing a lesion that could theoretically cause, for example, a cancer or birth defect. Organisms, however, have many defense mechanisms against genotoxic events. For example, DNA repair mechanisms rapidly remove lesions on the DNA to restore normal DNA. Such protective mechanisms suggest that even agents that directly introduce DNA lesions may have a threshold of exposure below which the probability of producing a health effect is insignificantly small.

Kirsch-Volders, et al. (2000)¹⁹⁾ proposed simple definitions for “real” threshold and “alleged” threshold as follows:

“Real’’ threshold: a concentration/dose below which the measured effect does not occur.

“Alleged’’ threshold: a concentration/dose below which the measured effect does occur, but cannot be detected, because the system is not sufficiently sensitive to discriminate it from spontaneous events.

Now it is widely accepted that the genotoxic chemicals that do not target DNA directly have thresholds, at least a practical threshold²⁰⁾. For example, spindle poisons, which induce numerical chromosomal aberration, target mitotic apparatus²¹⁾, topoisomerase family, which disturbs the fidelity of DNA unwinding and resulting in infidelity of DNA duplication^22,23), the imbalance of nucleotides in cells which also disturbs the fidelity of DNA duplication^24,25), etc., are thought to have such thresholds. Whenever the mechanism of mutagenesis is based on lesions in biomolecules other than DNA, it is generally accepted that a threshold can be identified.

For risk characterization, it is important how the threshold of genotoxicity is considered, especially in the case of direct-acting mutagens that interact with DNA. As mentioned above, mutagens that target macromolecules other than DNA can be assessed by considering exposure level relative to a defined threshold. Although international worldwide agreement has not yet been obtained to accept a threshold concept for direct DNA-reactive mutagens, appropriate approaches for estimating acceptable exposure levels to such mutagens are being discussed by international expert committees. As mentioned in the former section, there is already substantial evidence to demonstrate practical thresholds for at least some directly DNA-reactive mutagens^26,27). We think it is possible to assess such mutagens in the same way as other toxicological effects that exhibit thresholds^28,29), by determining a virtually safe dose (VSD), ensuring levels As Low As Reasonably Achievable (ALARA) in most cases, and/or establishing appropriate TTCs.

There are several cases in which evidence has shown a threshold concentration/dose, at least a practical threshold, below which the measured effect does not occur although the exposure is enough for some other effects occur^26,27). These cases include observations of DNA damage, bacterial gene mutation, and chromosomal aberrations. The rec assay provides a good example (Fig. 1). The test chemical is spotted on the filter paper (black circle on the right) and diffuses into the plate making a concentration gradient. Repair deficient Bacillus subtilis strain (M45) cannot survive in an area close to the test chemical spotted site where the concentration is high. In contrast, the repair proficient strain (H17) survives at the area even close to the spotted site where the chemical exists at higher concentration. This is evidence to show that there are protective mechanisms that would be expected to result in a threshold at low exposure levels.

Fig. 1.

rec-assay has been used widely in Japan to assess DNA damage using Bacillus subtilis wild strain H17 with repair capacity and strain M45 deficient DNA repair.

Kirsch-Volders, et al.¹⁹⁾ discussed the alleged threshold because any assay system may not be sufficiently sensitive to detect the endpoint at very low concentration/dose levels. However, Asano, et al. showed that increasing the number of cells analyzed in the peripheral blood micronucleus assay up to 1 million cells per animal by flow cytometer on three model chemicals still did not demonstrate an effect at low concentration (Fig. 2)³⁰⁾. Sensitivity of course depends on the unit evaluated, numbers of cells or animals. Although it may be possible to increase sensitivity using more animals or cells, the important issue is the biological relevance of minimal effects. Animal welfare considerations also place a limit on the extent to which numbers of animals are increased to obtain small, and possibly biologically insignificant, increases in sensitivity.

Fig. 2.

Mouse micronucleus assay using 3 model chemicals with different mode of action analyzed from 2000 cells per animal manually (red) and flow-cytometry up 1 million cells per animal with flow cytometer (black).

These apparent threshold phenomena are also observed in the bacterial reverse mutation (Fig. 3). Fig. 3 shows the outcomes of the Ames assay using TA1535 and YG7108, which is an O⁶-methylguanine DNA methltransferase-deficient strain. The YG7108 strain induced reverse mutation by N-methyl-N-nitro-nitrosoguanidine at concentration levels where the TA1535 strain did not respond²⁶⁾.

Fig. 3.

Dose response curves of reverse gene mutation using TA1535 and O⁶⁾-methylguanine DNA methltransferase deficient strain YG7108 on N-methyl-N-nitro-nitrosoguanidine

Finally, the authors would like to propose the following definitions of threshold, including the case of DNA direct-acting mutagens:

“Real” threshold: a concentration/dose below which the measured effect does not occur.

“Practical” threshold: a concentration/dose below which the measured effect is not observed (does not occur or cannot be detected by ordinary assay systems) and below which any effect is not considered biologically relevant.

According to the definition, for the purpose of risk characterization, we do not have to consider “no threshold” or “no risk” even for DNA direct acting mutagens. It may be necessary, however, to add larger safety-factors in the case of clearly mutagenic chemicals.

4) Standard Battery and Additional Studies

It is the basic strategy to apply a battery system for identification of hazard of agricultural chemicals. This battery includes a “microbial (S. typhimurium and E. coli) gene mutation assay (commonly called the Ames test) with and without an exogenous metabolic activation system (OECD TG471)”, an “in vitro chromosomal aberration assay (OECD TG473)”, and an “in vivo rodent micronucleus assay (OECD TG474)”. The mouse lymphoma TK assay (MLA)(OECD TG476/under preparation) and in vitro micronucleus assay (OECD TG478) are alternatives to the in vitro chromosomal aberration assay and evaluate equally the induction of structural and numerical chromosomal aberrations by test chemicals. Recently, other in vivo assay systems have been, or are being, developed and validated. The transgenic animal model to detect gene mutation (OECD TG488) and the comet assay to detect DNA damages (OECD TG under preparation) have been introduced. Both assays can be applied to any organs of animals, and thus it is possible to assess these endpoints in the organ that is the target of carcinogenicity. Moreover, the development of the micronucleus assay in tissues other than bone marrow has been achieved; moreover it can target tissues other than hematopoietic tissue, e.g., testes, skin, liver, and digestive tracts⁸⁾.

There are many other assay systems to detect genotoxic events. For example, gene mutation assays using insects, the sperm abnormality assay, cytogenetic assays using plants, the gene-conversion assay using yeast, the sister chromatid exchange assay, etc. These assay systems have neither been well validated nor used widely. Accordingly, the results of these assays are considered as supportive evidence of chemical genotoxicity and should be used on a case-by-case basis. The basic components of standard battery and assay systems with validated test guideline should be considered first to make risk characterization.

5) Identification of Genotoxic Hazards from Chemicals That May Impact Human Health

Standard hazard identification of chemicals for genotoxicity as the first step of risk assessment is shown in Table 2, which is based on the combination of the results from a standard test battery. As the current standard test battery consists of a bacterial gene mutation assay (i.e., Ames test), an in vitro chromosomal aberration (CA) assay, and an in vivo rodent micronucleus (MN) assay (see section 4) and Table 1), the following eight combinations of test outcomes are possible:

In all cases, quality of data (e.g., application of appropriate guidelines or GLP), chemical properties (e.g., purity, solubility, volatility)/class/structure, study design, mechanism or mode of action, dose-effect relationship, or biological relevance should be taken into account. For additional tests, weight and/or strength of evidence of the existing data should be considered on a case by case basis, and the additional test(s) should be selected from within well-validated standard tests.

• 1. Ames, negative; in vitro CA, negative; in vivo MN, negative
• The chemical is not genotoxic. For final conclusion, quality of data (i.e., reliability, relevance) should be taken into account.

• 2. Ames, negative; in vitro CA, positive; in vivo MN, negative
• The chemical is not considered to be genotoxic in vivo.
• Biological relevance of the positive findings should be taken into account. Structural or numerical chromosomal aberrations should also be checked. Actually, there should also be evidence that bone marrow is exposed to the relevant active agent or metabolite if the conclusion is to be that it is not considered genotoxic in vivo.

• 3. Ames, negative; in vitro CA, negative; in vivo MN, positive
• The chemical is genotoxic.
• Clastogenic or aneugenic effect in MN induction should be considered. The possibility of non-genotoxic mechanism in MN induction is also taken into account on a case by case basis. Expert judgment is needed.

• 4. Ames, negative; in vitro CA, positive; in vivo MN, positive
• The chemical is genotoxic.
• Germ cell mutagenicity test might be considered on a case by case basis. Or, evaluation of effect to the germ cells in data from reproductive toxicity studies might be considered. Expert judgment is needed.

• 5. Ames, positive; in vitro CA, negative; in vivo MN, negative
• The chemical is concluded to be neither genotoxic nor genotoxic in vivo.
• Consider specific bacterial metabolism of the chemical for positive results in the Ames test. The case needs further investigation in vitro and/or second in vivo test. If a negative result was obtained in the in vivo transgenic gene mutation model or in vivo comet assay in the suitable organs, the chemical is not considered genotoxic. If positive, the chemical is genotoxic and a germ cell mutagenicity test should be considered. Expert judgment is needed.

• 6. Ames, positive; in vitro CA, positive; in vivo MN, negative
• The chemical is concluded to be neither genotoxic nor genotoxic in vivo.
• Structural or numerical chromosomal aberrations should be checked. Need second in vivo test. If a negative result was obtained in the in vivo transgenic gene mutation model or in vivo comet assay in the suitable organs, the chemical is not considered genotoxic in vivo. If positive, the chemical is genotoxic and a germ cell mutagenicity test should be considered. Expert judgment is needed.

• 7. Ames, positive; in vitro CA, negative; in vivo MN, positive
• The chemical is genotoxic.
• A second in vivo test is required. If negative result was obtained in the in vivo transgenic gene mutation model or in vivo comet assay in the suitable organs, the chemical is not considered to cause gene mutation in vivo. If positive, a germ cell mutagenicity test should be considered. Expert judgment is needed.

• 8. Ames, positive; in vitro CA, positive; in vivo MN, positive
• The chemical is genotoxic, and is a possible germ cell mutagen.
• A germ cell mutagenicity test should be considered. If a positive result is obtained, the chemical is a germ cell mutagen and possible heritable mutagen. Expert judgment is needed.

Table 2.Genotoxic Hazard Categorization of Chemical Based on the Results of the Test Battery

No.	Ames	In vitro CA	In vivo MN	Judgement	Need for additional test(s)	Explanation
1	−	−	−	Not genotoxic	No	The chemical is not genotoxic. Quality of data (i.e., reliability, relevance) should be taken into account.
2	−	+	−	Not genotoxic	No	The chemical is not considered to be genotoxic in vivo . Structural or numerical chromosomal aberrations should be checked. Bone marrow exposure to the chemical or metabolite should also be checked.
3	−	−	+	Genotoxic	No	The chemical is genotoxic. Consider clastogenic or aneugenic effect in MN induction. The possibility of non-genotoxic mechanism in MN induction is taken into account on a case by case basis. Expert judgement is needed
4	−	+	+	Genotoxic	No	The chemical is genotoxic. Germ cell mutagenicity test might be considered on a case by case basis. Or, evaluation of effect to the germ cells in data from reproductive toxicity studies might be considered. Expert judgement is needed.
5	+	−	−	Not concluded yet	Yes	The chemical is not concluded to be either genotoxic or not genotoxic in vivo . Consider specific bacterial metabolism of the chemical for positive result in the Ames test. Need further investigation in vitro and/or second in vivo test. If negative result was obtained in the in vivo transgenic gene mutation model or in vivo comet assay in the suitable organs, the chemical is not considered genotoxic. If positive, the chemical is genotoxic and germ cell mutagenicity test should be considered. Expert judgement is needed.
6	+	+	−	Not concluded yet	Yes	The chemical is not concluded to be genotoxic in vivo or not. Structural or numerical chromosomal aberrations should bechekced. Need second in vivo test. If negative result was obtained in the in vivo transgenic gene mutation model or in vivo comet assay in the suitable organs, the chemical is not considered genotoxic in vivo . If positive, the chemical is genotoxic and germ cell mutagenicity test should be considered. Expert judgement is needed.
7	+	−	+	Genotoxic	Yes	The chemical is genotoxic. Need for second in vivo test. If negative result was obtained in the in vivo transgenic gene mutation model or in vivo comet assay in the suitable organs, the chemical is not considered to cause gene mutation in vivo . If positive, germ cell mutagenicity test should be considered. Expert judgement is needed.
8	+	+	+	Genotoxic	Yes	The chemical is genotoxic, and is a possible germ cell mutagen. Germ cell mutagenicity test should be considered. If positive result is obtained, the chemical is a germ cell mutagen and possible heritable mutagen. Expert judgement is needed.

In all cases, quality of data (e.g., application of guideline or GLP), chemical properties (e.g., purity, solubility, volatility)/class/structure, study design, mechanism or mode of action, dose-effect relationship, or biological relevance should be taken into account.The weight and/or strength of evidence of existing data should be considered for futher testing on a case by case basis.Additional test(s) should be selected from among well-validated standard tests.  If the chemical with possible genotoxic hazard (e.g., No. 3 to 8) is carcinogenic, it might be a genotoxic carcinogen. However, evaluation such as carcinogenic mode of action in the target organ or toxicokinetics data will be needed to determine that it is a “genotoxic” carcinogen.If the chemical is not carcinogenic, genotoxic hazrd of the chemical to human health will be low concern. Even if so, effects to germ cells should be considered on a case by case basis.

If a chemical that is a possible genotoxic hazard (e.g., No. 3 to 8) is carcinogenic, it might be a genotoxic carcinogen. However, evaluation such as carcinogenic mode of action in the target organ or toxicokinetics data will be needed to determine that it is a “genotoxic” carcinogen. If the chemical is not carcinogenic, the potential genotoxic hazard of the chemical to human health will be of low concern. Even if so, its effect to germ cells should be considered on a case by case basis.

For risk assessment of genotoxic hazard with respect to human health, exposure assessment and mode of action should be taken into account. Determination of a VSD, the TTC, and MOA or MOE will be considered as appropriate.

Discussion

The risk characterization should include proper hazard identification, dose-response relationship analysis, and consideration of exposure levels. Genotoxicity has been evaluated mainly by hazard identification in the course of chemical risk assessment. However, qualitative assessments should be considered in combination with evaluation of actual exposure levels in the context of the exposure-response relationship and with consideration of the practical threshold concept and agreed-upon exposure thresholds of concern. Generally, there is tendency to detect genotoxicity of chemicals at high concentrations/dose levels. The biological relevance of observed effects and the possibility to extrapolate to the actual exposure levels to humans should also be considered. To detect genotoxicity, there are many in vitro assay systems but we do not have good tools to convert from the results in in vitro assays to the in vivo situation—especially the extrapolations required to estimate the human ADI. In vitro assays may play an important role in characterizing hazard and defining mechanisms, thereby supporting the principals of animal welfare, but for accurate risk characterization in vivo assay systems play a toxicologically more important role. New in vivo assay systems and integration of multiple genetic endpoints into individual assays, including standard toxicological assays, can minimize animal use while providing the information necessary for accurate risk characterization^7,31). It is not necessary to discuss the importance of the concept of threshold. It is well accepted that there are thresholds for genotoxicants that target macromolecules other than DNA. While the existence of thresholds for genotoxicants that react directly with DNA is still under discussion, there is evidence showing the existence of thresholds, at least practical thresholds, for some direct acting genotoxic chemicals. In many cases in the field of toxicology, the TTC concept is accepted as a part of chemical safety assessment^20,32,33). The TTC concept might also be applicable to the assessment of genotoxicity in the same manner as with other toxicological endpoints. The concept of threshold is also relevant to the strategy of assessing the safety of complex mixtures. People are exposed to many kinds of chemicals concomitantly, but it is general practice to assess each single chemical independently. There is no clear evidence to show that complex mixtures give more potent genotoxicity than additive effects. If we accept the existence of threshold, even practical one, it makes easy to expand the idea to the complex mixture simply based on the exposure levels of the component.

For risk characterization, sound scientific data and reasoning are critical, but fairness and transparency are also important. These factors are especially important for risk communication, which is just as important as risk assessment and risk management. Guidance for risk characterization must include fairness and transparency among the teams that perform risk-based assessments of chemical safety.

References

• 1.  European Chemicals Agency (ECHA), CMR substances from Annex VI of the CLP Regulation registered under REACH and/or notified under CLP − a first screening, ECHA-12-R-01-EN, May 2012 Available at: http://echa.europa.eu/documents/10162/13562/cmr_report_en.pdf.
• 2.  European Food Safety Authority (EFSA), Opinion of the Scientific Committee on a request from EFSA related to A Harmonised Approach for Risk Assessment of Substances Which are both Genotoxic and Carcinogenic, The EFSA Journal (2005) 282, 1–31. Available at: http://www.efsa.europa.eu/en/efsajournal/pub/282.htm.
• 3.  OECD 1997. OECD Guidelines for the Testing of Chemicals. Bacterial reverse mutation test (TG 471). Available at: http://www.oecd-ilibrary.org/environment/test-no-471-bacterial-reverse-mutation-test_9789264071247-en.
• 4.   Kirkland  D,  Pfuhler  S,  Tweats  D, et al. How to reduce false positive results when undertaking in vitro genotoxicity testing and thus avoid unnecessary follow-up animal tests: report of an ECVAM workshop. Mutat Res. 2007; 628(1): 31–55.
• 5.  OECD 1997. OECD Guidelines for the Testing of Chemicals. Mammalian Erythrocyte Micronucleus Test (TG 474). Available at: http://www.oecd-ilibrary.org/environment/test-no-474-mammalian-erythrocyte-micronucleus-test_9789264071285-en.
• 6.  International Conference on Harmonization, S2(R1) Guideline, Guidance on genotoxicity testing and data interpretation for pharmaceuticals intended for human use, step 4 version, 9 November, 2011. Available at: http://www.ich.org/fileadmin/Public_Web_Site/ICH_Products/Guidelines/Safety/S2_R1/Step4/S2R1_Step4.pdf.
• 7.  OECD 2011. OECD Guidelines for the Testing of Chemicals: Transgenic rodent somatic and germ cell gene mutation assays (TG 488). Available at: http://www.oecd-ilibrary.org/environment/test-no-488-transgenic-rodent-somatic-and-germ-cell-gene-mutation-assays_9789264122819-en.
• 8.   Morita  T,  MacGregor  JT,  Hayashi  M. Micronucleus assays in rodent tissues other than bone marrow. Mutagenesis. 2011; 26(1): 223–230.
• 9.  The Collaborative Study Group for the Micronucleus Test Micronucleus test with mouse peripheral blood erythrocytes by acridine orange supravital staining: The summary report of the 5th collaborative study by CSGMT/JEMS·MMS. Mutat Res. 1992; 278(2–3): 83–98.
• 10.   Pfuhler  S,  Kirkland  D,  Kasper  P, et al. Reduction of use of animals in regulatory genotoxicity testing: Identification and implementation opportunities—Report from an ECVAM workshop. Mutat Res. 2009; 680(1–2): 31–42.
• 11.   Rothfuss  A,  O’Donovan  M,  De Boeck  M, et al. Collaborative study on 15 compounds in the rat-liver Comet assay integrated into 2- and 4-week repeat-dose studies. Mutat Res. 2010; 702(1): 40–69.
• 12.  Organisation for Economic Co-operation and Development (OECD) Series on testing and assessment Number 34, guidance document on the validation and international acceptance of new or updated test methods for hazard assessment, ENV/JM/MONO, 14 2005. Available at: http://search.oecd.org/officialdocuments/displaydocumentpdf/?doclanguage=en&cote=env/jm/mono(2005)14.
• 13.   Kirkland  D,  Kasper  P,  Müller  L,  Corvi  R,  Speit  G. Recommended lists of genotoxic and non-genotoxic chemicals for assessment of the performance of new or improved genotoxicity tests: A follow-up to an ECVAM workshop. Mutat Res. 2008; 653(1–2): 99–108.
• 14.   Kirkland  DJ,  Aardema  M,  Banduhn  N, et al. In vitro approaches to develop weight of evidence (WoE) and mode of action (MoA) discussions with positive in vitro genotoxicity results. Mutagenesis. 2007; 22(3): 161–175.
• 15.   Munro  IC,  Renwick  AG,  Danielewska-Nikiel  B. The Threshold of Toxicological Concern (TTC) in risk assessment. Toxicol Lett. 2008; 180(2): 151–156.
• 16.   Munro  IC,  Ford  RA,  Kennepohl  E,  Sprenger  JG. Correlation of structural class with no-observed-effect levels: a proposal for establishing a threshold of concern. Food Chem Toxicol. 1996; 34(9): 829–867.
• 17.   Munro  IC,  Kennepohl  E,  Kroes  R. A procedure for the safety evaluation of flavouring substances. Joint FAO/WHO Expert Committee on Food Additives. Food Chem Toxicol. 1999; 37(2–3): 207–232.
• 18.  EFSA Scientific Committee Statement on the applicability of the Margin of Exposure approach for the safety assessment of impurities1 which are both genotoxic and carcinogenic in substances added to food/feed, EFSA Journal 2012;10(3):2578. Available at: http://www.efsa.europa.eu/en/efsajournal/doc/2578.pdf.
• 19.   Kirsch-Volders  M,  Aardema  M,  Elhajouji  A. Concepts of thresholds in mutagenesis and carcinogenesis. Mutat Res. 2000; 464(1): 3–11.
• 20.  EFSA Scientific Committee Scientific opinion on genotoxicity testing strategies applicable to food and feed safety assessment, EFSA Journal 2011;9(9):2379. Available at: http://www.efsa.europa.eu/en/efsajournal/doc/2379.pdf.
• 21.   Parry  JM,  Fielder  RJ,  McDonald  A. Thresholds for aneuploidy-inducing chemicals. Advisory Committee on Mutagenicity of Chemicals in Food, Consumer Products and the Environment of the UK Department of Health. Mutagenesis. 1994; 9(6): 503–504.
• 22.  ECETOC 1997. Aneuploidy. ECETOC Monograph No.27, August 1997. European Centre for Ecotoxicology and Toxicology of Chemicals, Brussels.
• 23.   Aardema  MJ,  Albertini  S,  Arni  P, et al. Aneuploidy: a report of an ECETOC task force. Mutat Res. 1998; 410(1): 3–79.
• 24.   Yajima  N,  Kurata  Y,  Sawai  T,  Takeshita  Y. Comparative studies in induction of micronuclei by three genetically recombinant and urinary human erythropoietins. Mutagenesis. 1993; 8(3): 237–241.
• 25.   Yajima  N,  Kurata  Y,  Imai  E,  Sawai  T,  Takeshita  Y. Genotoxicity of genetic recombinant human erythropoietin in a novel test system. Mutagenesis. 1993; 8(3): 231–236.
• 26.   Sofuni  T,  Nohmi  T,  Ohta  T,  Hayashi  M. [Genotoxicity: Is a threshold concept applicable to evaluate the mutagenic activity of DNA-targeting substances!?] Environ Mutagen Res. 2005; 27: 61–73. Japanese.
• 27.   Gocke  E,  Müller  L. In vivo studies in the mouse to define a threshold for the genotoxicity of EMS and ENU. Mutat Res. 2009; 678(2): 101–107.
• 28.   Gillespie  Z,  Pulido  O,  Vavasour  E. Risk Assessment Approaches for Carcinogenic Food Contaminants. International Food Risk Analysis Journal. 2011; 1: 1–18.
• 29.   O’Brien  J,  Renwick  AG,  Constable  A, et al. Approaches to the risk assessment of genotoxic carcinogens in food: a critical appraisal. Food Chem Toxicol. 2006; 44(10): 1613–1635.
• 30.   Asano  N,  Torous  DK,  Tometsko  CR,  Dertinger  SD,  Morita  T,  Hayashi  M. Practical threshold for micronucleated reticulocyte induction observed for low doses of mitomycin C, Ara-C, and colchicine. Mutagenesis. 2006; 21(1): 15–20.
• 31.   Test Guideline  OECD, In vivo Mammalian Alkaline Comet Assay (in preparation).
• 32.  EFSA Scientific Committee Scientific Opinion on Exploring options for providing advice about possible human health risks based on the concept of Threshold of Toxicological Concern (TTC). EFSA Journal. 2012; 10(7): 2750. Available at: http://www.efsa.europa.eu/de/efsajournal/doc/2750.pdf.
• 33.   Dewhurst  I,  Renwick  AG. Evaluation of the Threshold of Toxicological Concern (TTC) − Challenges and approaches. Regul Toxicol Pharmacol. 2013; 65(1): 168–177.