Genotoxicity assessment: opportunities, challenges and perspectives for quantitative evaluations of dose-response data

Genotoxicity data are mainly interpreted in a qualitative way, which typically results in a binary classification of chemical entities. For more than a decade, there has been a discussion about the need for a paradigm shift in this regard. Here, we review current opportunities, challenges and perspectives for a more quantitative approach to genotoxicity assessment. Currently discussed opportunities mainly include the determination of a reference point (e.g., a benchmark dose) from genetic toxicity dose-response data, followed by calculation of a margin of exposure (MOE) or derivation of a health-based guidance value (HBGV). In addition to new opportunities, major challenges emerge with the quantitative interpretation of genotoxicity data. These are mainly rooted in the limited capability of standard in vivo genotoxicity testing methods to detect different types of genetic damage in multiple target tissues and the unknown quantitative relationships between measurable genotoxic effects and the probability of experiencing an adverse health outcome. In addition, with respect to DNA-reactive mutagens, the question arises whether the widely accepted assumption of a non-threshold dose-response relationship is at all compatible with the derivation of a HBGV. Therefore, at present, any quantitative genotoxicity assessment approach remains to be evaluated case-by-case. The quantitative interpretation of in vivo genotoxicity data for prioritization purposes, e.g., in connection with the MOE approach, could be seen as a promising opportunity for routine application. However, additional research is needed to assess whether it is possible to define a genotoxicity-derived MOE that can be considered indicative of a low level of concern. To further advance quantitative genotoxicity assessment, priority should be given to the development of new experimental methods to provide a deeper mechanistic understanding and a more comprehensive basis for the analysis of dose-response relationships.

[The One Health approach in the context of global commodity chains, crises, and food and feed safety]

The holistic view of food and feed safety, including animal health and environmental conditions, is an important pillar of the One Health approach. The terminology thus clearly goes beyond the prevention of spreading microbiological diseases, in which context it is often understood, and highlights that humans, animals, and the environment as well as their interaction should be considered in a transdisciplinary context.In terms of One Health, this discussion paper focuses less on microbiological risks, but rather on the connection to chemical risks in the food chain. This is illustrated by concrete examples of chemical contaminants (metals, persistent organic contaminants, natural toxins). The mechanisms of input and transfer along the food chain are presented.Minimizing the presence of contaminants and thus exposure requires international and interdisciplinary cooperation in the spirit of the One Health approach. Climate change, pandemics, shortages of raw materials, energy deficiencies, political crises, and environmental disasters can affect the entire food chain from primary production of plant and animal foods to further processing and provision of products to consumers. In addition to changing availability, this can also have an impact on the composition, quality, and safety of food and feed. Based on the effect on global commodity chains, vulnerable and resilient areas along the food chain become visible. In terms of the One Health approach, the aim is to increase safety and resilience along the food chain and to minimize its vulnerability.

Chronic dietary exposure to total arsenic, inorganic arsenic and water-soluble organic arsenic species based on results of the first German total diet study

For risk assessment purposes, the dietary exposure to total arsenic and inorganic arsenic was estimated within the first German total diet study (BfR MEAL Study) for the whole population in Germany. Therefore, occurrence data of 356 different foods from the BfR MEAL Study were combined with consumption data from German nutrition surveys. Due to the different toxicological potentials of other water-soluble organic arsenic species present in rice-based foods, fish and seafood, dietary exposure to dimethylarsinic acid, monomethylarsonic acid and arsenobetaine was assessed in consumers in Germany through such foods for the first time. Related to the bodyweight, dietary exposure to total arsenic and inorganic arsenic in infants and young children (0.5-<5 years) were higher than in adolescents/adults (≥14 years). The highest median exposure estimates to inorganic arsenic resulted for the age group of infants from 0.5 to <1 year under modified lower bound conditions and for young children from 1 to <2 years under upper bound conditions (0.17 μg kg^-1 bodyweight day^-1-0.24 μg kg^-1 bodyweight day^-1 and 0.26 μg kg^-1 bodyweight day^-1-0.34 μg kg^-1 bodyweight day^-1, respectively). 'Grains and grain-based products' (especially rice) were identified as the main contributors for dietary exposure to total arsenic and inorganic arsenic for all age classes. Especially, for infants and young children, high consumption of rice-based foods and fish fingers is driving the dietary exposure to dimethylarsinic acid. The dietary exposure calculations indicate that a further reduction of dietary exposure to inorganic arsenic and further investigations to water-soluble organic arsenic species are necessary.

Human sulphotransferases are involved in the activation of aristolochic acids and are expressed in renal target tissue

Use of herbal preparations containing Aristolochia species has led to progressive nephropathy and urothelial cancer in humans. Analysis of DNA adducts formed in human target tissues and studies in animal models have pointed out a major role of the secondary plant metabolites, aristolochic acids, in these effects. Only a minority of the users of Aristolochia-containing products developed nephropathy and cancer, suggesting differences in individual susceptibility. Differences in metabolic activation and inactivation frequently affect the susceptibility towards chemicals. Others have shown that the activation of aristolochic acids to DNA-reactive and mutagenic metabolites requires reduction of their aryl nitro group. The biological activity of numerous nitro- and aminoarenes, after appropriate phase I metabolism, is strongly enhanced in the presence of acetyltransferases or sulphotransferases (SULTs). In the present study, we demonstrate that expression of human SULTs in bacterial and mammalian target cells reinforces the mutagenic activity of aristolochic acids. Using Salmonella typhimurium TA1538 as the recipient organism, we identified the expression of all 12 human SULT forms. SULT1A1 led to the strongest increase in the mutagenicity of aristolochic acids. Some activation was also observed with SULT1B1, but not with the remaining forms. The role of SULT1A1 in the activation of aristolochic acids was corroborated using S. typhimurium TA100- and Chinese hamster V79-derived target cells engineered for expression of human SULT1A1 when compared with control cells. Furthermore, pentachlorophenol, an inhibitor of SULT1A1, strongly reduced the mutagenic effect of aristolochic acids in V79-hCYP2E1-hSULT1A1 cells. Moreover, we demonstrate that SULT1A1 and SULT1B1 are expressed in human kidney using immunoblot analysis, but their levels are substantially lower than in liver. Finally, we discuss the possibility that reactive sulphuric acid conjugates produced in other tissues are transferred to kidney and ureter.

Bioactivation of the heterocyclic aromatic amine 2-amino-3-methyl-9H-pyrido [2,3-b]indole (MeA C) in recombinant test systems expressing human xenobiotic-metabolizing enzymes

2-Amino-3-methyl-9H-pyrido[2,3-b]indole (MeAalphaC) and some metabolites were investigated for mutagenicity in mammalian cell lines and bacterial strains engineered for the expression of human enzymes. MeAalphaC induced gene mutations (studied at the hprt locus) in Chinese hamster V79-derived cells co-expressing cytochrome (CYP) 1A2 and sulphotransferase (SULT) 1A1 even at a concentration of 30 nM, but was inactive in cells co-expressing CYP1A2 and N-acetyltransferase (NAT) 1 or 2. MeAalphaC, tested in the presence of rat liver post-mitochondrial fraction, showed strongly enhanced mutagenicity in a Salmonella typhimurium strain expressing human SULT1A1 compared with the control (recipient) strain TA1538/1,8-DNP (deficient in endogenous acetyltransferase). Mutagenicity was also enhanced, although to a lesser extent, when NAT2 was expressed in the latter strain. The metabolite, 2-hydroxylamino-3-methyl-9H-pyrido[2,3-b]indole (N-OH-MeAalphaC) was a direct mutagen to strains TA1538 and TA1538/ 1,8-DNP. This mutagenicity was strongly enhanced in corresponding strains expressing SULT1A1. A moderate enhancement was observed when SULT1A2, SULT1B1, SULT1C2 or NAT2 were expressed in strain TA1538. The remaining enzymes studied (SULT1A3, 1C1, 1E1, 2A1, 2B1a, 2B1b, 4A1 and NAT1) did not indicate any activation of N-OH-MeAalphaC. Preliminary mutagenicity experiments in SULT-expressing S.typhimurium strains were conducted with other hydroxylated metabolites of MeAalphaC. The phenols, 6- and 7-hydroxy-MeAalphaC, were inactive under the conditions studied. The benzylic alcohol, 2-amino-3-hydroxymethyl-9H-pyrido[2,3-b]indole, was mutagenic in a strain expressing SULT1A1, but its activity was much weaker than that of N-OH-MeAalphaC. Thus, N-hydroxylation (e.g. mediated by CYP1A2) and sulpho conjugation (primarily mediated by SULT1A1) was the dominating activation pathway of MeAalphaC in model systems engineered for human enzymes. Some other SULT forms as well as NAT2 were also capable of activating N-OH-MeAalphaC, although with much lower efficiency than SULT1A1. Another minor activation pathway involved benzylic hydroxylation followed by sulpho conjugation by SULT1A1.

Activation of 3-nitrobenzanthrone and its metabolites by human acetyltransferases, sulfotransferases and cytochrome P450 expressed in Chinese hamster V79 cells

3-nitrobenzanthrone (3-NBA) is a potent mutagen and suspected human carcinogen identified in diesel exhaust and ambient air pollution. 3-aminobenzanthrone (3-ABA), 3-acetylaminobenzanthrone (3-Ac-ABA) and N-acetyl-N-hydroxy-3-aminobenzanthrone (N-Ac-N-OH-ABA) have been identified as 3-NBA metabolites. Recently we found that 3-NBA and its metabolites (3-ABA, 3-Ac-ABA and N-Ac-N-OH-ABA) form the same DNA adducts in vivo in rats. In order to investigate whether human cytochrome P450 (CYP) enzymes (i.e., CYP1A2), human N,O-acetyltransferases (NATs) and sulfotransferases (SULTs) contribute to the metabolic activation of 3-NBA and its metabolites, we developed a panel of Chinese hamster V79MZ-h1A2 derived cell lines expressing human CYP1A2 in conjunction with human NAT1, NAT2, SULT1A1 or SULT1A2, respectively. Cells were treated with 0.01, 0.1 or 1 microM 3-NBA, or its metabolites (3-ABA, 3-Ac-ABA and N-Ac-N-OH-ABA). Using both enrichment versions of the (32)P-postlabeling assay, nuclease P1 digestion and butanol extraction, essentially 4 major and 2 minor DNA adducts were detected in the appropriate cell lines with all 4 compounds. The major ones were identical to those detected in rat tissue; the adducts lack an N-acetyl group. Human CYP1A2 was required for the metabolic activation of 3-ABA and 3-Ac-ABA (probably via N-oxidation) and enhanced the activity of 3-NBA (probably via nitroreduction). The lack of acetylated adducts suggests N-deacetylation of 3-Ac-ABA and N-Ac-N-OH-ABA. Thus, N-hydroxy-3-aminobenzanthrone (N-OH-ABA) appears to be a common intermediate for the formation of the electrophilic arylnitrenium ions capable of reacting with DNA. Human NAT1 and NAT2 as well as human SULT1A1 and SULT1A2 strongly contributed to the high genotoxicity of 3-NBA and its metabolites. Moreover, N,O-acetyltransfer reactions catalyzed by human NATs leading to the corresponding N-acetoxyester may be important in the bioactivation of N-Ac-N-OH-ABA. As human exposure to 3-NBA is likely to occur primarily via the respiratory tract, expression of CYPs, NATs and SULTs in respiratory tissues may contribute significantly and specifically to the metabolic activation of 3-NBA and its metabolites. Consequently, polymorphisms in these genes could be important determinants of lung cancer risk from 3-NBA.

Metabolic activation of the environmental contaminant 3-nitrobenzanthrone by human acetyltransferases and sulfotransferase

3-Nitrobenzanthrone (3-NBA) an extremely potent mutagen and suspected human carcinogen identified in diesel exhaust and in airborne particulate matter was shown to form multiple DNA adducts in vitro and in vivo in rats. In order to investigate whether human N,O-acetyltransferases (NATs) and sulfotransferases (SULTs) contribute to the metabolic activation of 3-NBA we used a panel of newly constructed Chinese hamster lung fibroblast V79MZ derived cell lines expressing human NAT1, human NAT2 or human SULT1A1, as well as TA1538-derived Salmonella typhimurium strains expressing human NAT1 (DJ400) or human NAT2 (DJ460) and determined DNA binding and mutagenicity. The formation of 3-NBA-derived DNA adducts was analysed by (32)P-postlabelling after exposing V79 cells to 0.01 micro M 3-NBA or 0.1 micro M N-acetyl-N-hydroxy-3-aminobenzanthrone (N-Ac-N-OH-ABA), a potential metabolite of 3-NBA. Similarly up to four major and two minor adducts were detectable for both compounds, the major ones being identical to those detected previously in DNA from rats treated with 3-NBA. Comparison of DNA binding between different V79MZ derived cells revealed that human NAT2 and, to a lesser extent, human NAT1 and human SULT1A1, contribute to the genotoxic potential of 3-NBA and N-Ac-N-OH-ABA to form DNA adducts. However, the extent of DNA binding by 3-NBA was higher in almost all V79 cells at a 10-fold lower concentration than by N-Ac-N-OH-ABA, suggesting that N-Ac-N-OH-ABA is not a major intermediate in the formation of 3-NBA-derived adducts. 3-NBA showed a 3.8-fold and 16.8-fold higher mutagenic activity in Salmonella strains expressing human NAT1 and human NAT2, respectively, than in the acetyltransferase-deficient strain, whereas N-Ac-N-OH-ABA was only clearly (but weakly) mutagenic in Salmonella DJ460 expressing human NAT2. This finding suggests that N-Ac-N-OH-ABA is not a major reactive metabolite responsible for the high mutagenic potency of 3-NBA in Salmonella. Collectively our results indicate that O-acetylation and O-sulfonation by human NATs and SULTs may contribute significantly to the high mutagenic and genotoxic potential of 3-NBA. Moreover, the yet-unidentified four major 3-NBA-derived adducts may be DNA adducts without an N-acetyl group.

Conjugation of 4-nitrophenol and 4-hydroxylonazolac in V79-derived cells expressing individual forms of human sulphotransferases

Sulpho conjugation of xenobiotics is catalysed by enzymes of the SULT superfamily. We have studied the conjugation of two model compounds, 4-nitrophenol and 4-hydroxylonazolac, in cultures of V79-derived cell lines that individually express human SULT1A1 (alloenzymes *1 and *V), 1A2 (alloenzymes *1 and *2), 1A3, 1B1, 1E1, 2A1 and rat SULT1E1. 4-Nitrophenol was sulphonated in all recombinant cell lines used but not in the control cell line (V79p). The relative activity in the various cell lines strongly depended on the substrate concentration used (1-1000 μM). 4-Hydroxylonazolac was conjugated in the cell lines expressing the following enzymes, in this order, human SULT1E1>1A1 (*1>*V)>1A2 (*1>*2)>1A3. In all these cell lines, the rate of conjugation increased with the substrate concentration (1-100 μM) without reaching a saturation level. The mass spectrometric and fluorometric analyses used are very sensitive. Preliminary experiments demonstrate that activities can readily be measured in microtitre-plate cultures.

Human cytosolic sulphotransferases: genetics, characteristics, toxicological aspects

Cytosolic sulphotransferases transfer the sulpho moiety from the cofactor 5'-phosphoadenosine-3'-phosphosulphate (PAPS) to nucleophilic groups of xenobiotics and small endogenous compounds (such as hormones and neurotransmitters). This reaction often leads to products that can be excreted readily. However, other sulpho conjugates are strong electrophiles and may covalently bind with DNA and proteins. All known cytosolic sulphotransferases are members of an enzyme/gene superfamily termed SULT. In humans, 10 SULT genes are known. One of these genes encodes two different enzyme forms due to the use of alternative first exons. Different SULT forms substantially differ in their substrate specificity and tissue distribution. Genetic polymorphisms have been described for three human SULTs. Several allelic variants differ in functional properties, including the activation of promutagens. Only initial results are available from the analysis of SULT allele frequencies in different population groups, e.g. subjects suffering from specific diseases and corresponding controls.

Sulfotransferases: genetics and role in toxicology

The mammalian xenobiotic-metabolizing sulfotransferases are cytosolic enzymes, which form a gene superfamily (SULT). Ten distinct human SULT forms are known. Two SULT forms represent splice variants, the other forms are encoded by separate genes. Common functional polymorphisms of the transcribed region are known for two of the forms. We have expressed 16 separate rat and human SULTs as well as some of their allelic variants, in Salmonella typhimurium TA1538 and/or V79 cells, which are target cells of commonly used mutagenicity assays. The expressed SULTs activated numerous compounds to mutagens in both assay systems. However, some promutagens were activated by only one or several of the human SULTs. Pronounced differences in promutagen activation were also detected between orthologous rat and human SULTs, and between allelic variants of human SULTs.

Human phenol sulfotransferases hP-PST and hM-PST activate propane 2-nitronate to a genotoxicant

The industrial solvent 2-nitropropane (2-NP) is a genotoxic hepatocarcinogen in rats. The genotoxicity of the compound in rats has been attributed to sulfotransferase-mediated formation of DNA-reactive nitrenium ions from the anionic form of 2-NP, propane 2-nitronate (P2N). Whether human sulfotransferases are capable of activating P2N is unknown. In the present study we have addressed this question by investigating the genotoxicity of P2N in various V79-derived cell lines engineered for expression of individual forms of human sulfotransferases, the phenol-sulfating and the monoamine-sulfating phenol sulfotransferases (hP-PST and hM-PST) and the human hydroxysteroid sulfotransferase (hHST). Genotoxicity was assessed by measuring the induction of DNA repair synthesis and by analyzing the formation of DNA modifications. P2N induced repair synthesis in V79-hP-PST and V79-hM-PST cells, whereas induction of repair synthesis in V79-hHST cells was negligible. P2N also resulted in the formation of 8-aminodeoxyguanosine and increased the level of 8-oxodeoxyguanosine in V79-hP-PST cells, but not in the parental V79-MZ cells, which do not show any sulfotransferase activity. Acetone oxime, the tautomeric form of the first reduction product of 2-NP, 2-nitrosopropane, was inactive in all cell lines. The results show that the human phenol sulfotransferases P-PST and M-PST are capable of metabolically activating P2N (P-PST >> M-PST) and that the underlying mechanism is apparently identical to that resulting in the activation of P2N in rat liver, where 2-NP causes carcinomas. These results support the notion that 2-NP should be regarded as a potential human carcinogen.

Sulfation of thyroid hormone by estrogen sulfotransferase

Sulfation is one of the pathways by which thyroid hormone is inactivated. Iodothyronine sulfate concentrations are very high in human fetal blood and amniotic fluid, suggesting important production of these conjugates in utero. Human estrogen sulfotransferase (SULT1E1) is expressed among other tissues in the uterus. Here we demonstrate for the first time that SULT1E1 catalyzes the facile sulfation of the prohormone T4, the active hormone T3 and the metabolites rT3 and 3,3'-diiodothyronine (3,3'-T2) with preference for rT3 approximately 3,3'-T2 > T3 approximately T4. Thus, a single enzyme is capable of sulfating two such different hormones as the female sex hormone and thyroid hormone. The potential role of SULT1E1 in fetal thyroid hormone metabolism needs to be considered.

Activation of propane 2-nitronate to a genotoxicant in V79-derived cell lines engineered for the expression of rat hepatic sulfotransferases

2-Nitropropane (2-NP) is a genotoxic hepatocarcinogen in rats. The genotoxicity of the compound has been attributed to a sulfotransferase-mediated formation of DNA-reactive species from the anionic form of 2-NP, propane 2-nitronate (P2N). Several observations have suggested that sulfotransferases (SULTs) 1A1 and/or 1C1 may be important in the activation of P2N to a genotoxicant in rat liver, but a definite proof is lacking. In order to identify the sulfotransferase(s) of rat liver that are capable of activating P2N, we have investigated the genotoxicity of P2N in various V79-derived cell lines engineered for expression of individual forms of rat hepatic sulfotransferases. Genotoxicity was assessed by measuring the induction of DNA repair synthesis. 1-Hydroxymethylpyrene (HMP), which is metabolically activated by most sulfotransferases, served as a positive control. Neither P2N nor HMP induced DNA repair in the parental V79-MZ cells, which do not show any sulfotransferase activity. P2N was also inactive in V79-rHSTa and V79-rHST20 cells, which express specific hydroxysteroid sulfotransferases. By contrast, a clear and concentration-dependent induction of repair synthesis by P2N was observed in V79-rPST-IV and V79-rST1C1 cells, which express rat SULT1A1 and SULT1C1, respectively. HMP was genotoxic in all sulfotransferase-expressing cell lines. Acetone oxime (AO), the tautomeric form of the first reduction product of 2-NP, 2-nitrosopropane, was inactive in all cell lines. The results corroborate the essential role of sulfotransferases in the metabolic activation of P2N to genotoxic products and identify two rat sulfotransferases which are capable of catalyzing the activation step.

Chiral inversion of 1-hydroxyethylpyrene enantiomers mediated by enantioselective sulfotransferases

The benzylic alcohol 1-hydroxyethylpyrene (1-HEP) is activated to a mutagen by sulfotransferases. The sulfuric acid ester formed is difficult to detect, as it is rapidly hydrolysed back to the alcohol. Incubation of the individual enantiomers of 1-HEP with human hydroxysteroid sulfotransferase (hHST) or estrogen sulfotransferase (hEST), expressed in bacteria, led to the formation of the other enantiomer. The rates of sulfation were determined from the initial rates of chiral inversion of the alcohol, knowing that hydrolysis follows an SN1 mechanism and therefore produces racemic alcohol. hEST showed high enantioselectivity for S-1-HEP, whereas hHST strongly preferred the R-enantiomer. The rates of sulfation of the preferred enantiomers were high, similar to those for the prototype substrates of hEST (beta-estradiol) and hHST (dehydroepiandrosterone). Moreover, after a 30-min incubation of S-1-HEP with hEST, 95% of the recovered alcohol showed the R-configuration, indicating that several cycles of sulfation and hydrolysis had led to the depletion of one enantiomer and to the enrichment of the other enantiomer.

Sulfotransferase-mediated activation of mutagens studied using heterologous expression systems

Sulfation is a common final step in the biotransformation of xenobiotics and is traditionally associated with inactivation. However, the sulfate group is electron-withdrawing and may be cleaved off heterolytically in some molecules leading to electrophilic cations which may form adducts with DNA and other important cellular structures. Since endogenous sulfotransferases do not appear to be expressed in indicator cells of standard mutagenicity tests, rat and human sulfotransferases have been stably expressed in his- Salmonella typhimurium strain TA1538 and Chinese hamster V79 cells. Using these recombinant indicator cells, sulfotransferase-dependent genotoxic activities were detected with N-hydroxy-2-acetylaminofluorene, 2-acetylaminofluorene (in the presence of co-expressed rat cytochrome P450 1A2), hycanthone, 1'-hydroxysafrole, alpha-hydroxytamoxifen and various benzylic alcohols derived from polycyclic aromatic hydrocarbons. In several cases, it was critical that the reactive sulfuric acid conjugates were formed directly within the indicator cells, owing to the inefficient penetration of cell membranes. In other cases, spontaneous benzylic substitution reactions with medium components, such as halogenide ions or amino acids, led to secondary, membrane-penetrating reactive species. Different sulfotransferases, including related forms from rat and human, substantially differed in their substrate specificity towards the investigated promutagens. It is known that some sulfotransferases are expressed with high tissue and cell type specificities. This site-dependent expression together with the limitations in the distribution of reactive sulfuric acid conjugates may explain organotropic effects of compounds activated by this metabolic pathway.

Expression of human estrogen sulfotransferase in Salmonella typhimurium: differences between hHST and hEST in the enantioselective activation of 1-hydroxyethylpyrene to a mutagen

Various human sulfotransferases (hP-PST, hM-PST, hHST) and rat sulfotransferases (rPST-IV, rHSTa) have already been expressed in Ames' Salmonella strains (in particular in TA1538). Now a further strain, TA1538-hEST, which expresses the human estrogen sulfotransferase (hEST), has been constructed. This strain activated the primary benzylic alcohol 1-hydroxymethylpyrene (1-HMP) and the secondary benzylic alcohol 1-hydroxyethylpyrene (1-HEP) to mutagens. Human sulfotransferases hEST and hHST both activated 1-HEP, but they differed substantially in their enantioselectivity for this compound.