Report of the IWGT working group on strategy/interpretation for regulatory in vivo tests II. Identification of in vivo-only positive compounds in the bone marrow micronucleus test.

A survey conducted as part of an International Workshop on Genotoxicity Testing (IWGT) has identified a number of compounds that appear to be more readily detected in vivo than in vitro. The reasons for this property varies from compound to compound and includes metabolic differences; the influence of gut flora; higher exposures in vivo compared to in vitro; effects on pharmacology, in particular folate depletion or receptor kinase inhibition. It is possible that at least some of these compounds are detectable in vitro if a specific in vitro test is chosen as part of the test battery, but the 'correct' choice of test may not always be obvious when testing a compound of unknown genotoxicity. It is noted that many of the compounds identified in this study interfere with cell cycle kinetics and this can result in either aneugenicity or chromosome breakage. A decision tree is outlined as a guide for the evaluation of compounds that appear to be genotoxic agents in vivo but not in vitro. The regulatory implications of these findings are discussed.

Report of the IWGT working group on strategies and interpretation of regulatory in vivo tests I. Increases in micronucleated bone marrow cells in rodents that do not indicate genotoxic hazards.

In vivo genotoxicity tests play a pivotal role in genotoxicity testing batteries. They are used both to determine if potential genotoxicity observed in vitro is realised in vivo and to detect any genotoxic carcinogens that are poorly detected in vitro. It is recognised that individual in vivo genotoxicity tests have limited sensitivity but good specificity. Thus, a positive result from the established in vivo assays is taken as strong evidence for genotoxic carcinogenicity of the compound tested. However, there is a growing body of evidence that compound-related disturbances in the physiology of the rodents used in these assays can result in increases in micronucleated cells in the bone marrow that are not related to the intrinsic genotoxicity of the compound under test. For rodent bone marrow or peripheral blood micronucleus tests, these disturbances include changes in core body temperature (hypothermia and hyperthermia) and increases in erythropoiesis following prior toxicity to erythroblasts or by direct stimulation of cell division in these cells. This paper reviews relevant data from the literature and also previously unpublished data obtained from a questionnaire devised by the IWGT working group. Regulatory implications of these findings are discussed and flow diagrams have been provided to aid in interpretation and decision-making when such changes in physiology are suspected.

Region-specific chromatin decondensation and micronucleus formation induced by 5-azacytidine in human TIG-7 cells.

A human diploid lung fibroblast cell strain, TIG-7, has a heteromorphic chromosome 15 with an extra short arm carrying a homogeneously staining region (15p+hsr). We demonstrated previously that the 15p+hsr consists of an inactive and G+C-rich rDNA cluster characterized by fluorescence in situ hybridization (FISH) and various chromosome banding techniques. Thus, it was suggested that the region could contain highly methylated DNA. To observe methylation status on the target region directly under the microscope, we used a demethylating agent, 5-azacytidine (5-azaC), to induce decondensation of the chromatin containing methylated DNA. At 24 h after treatment with 0.5 microM 5-azaC, marked decondensation of the 15p+hsr was observed in almost all of the metaphases. Furthermore, we observed micronuclei, which were equivalent to the rDNA of the 15p+hsr demonstrated by FISH in the same preparation. In contrast, the DNA cross-linking agent mitomycin C (MMC) preferentially induced 15p+hsr-negative micronuclei. These findings indicated that chromatin decondensation and subsequent DNA strand breakage induced by the demethylating effect of 5-azaC led specifically to 15p+hsr-positive micronuclei.

Spindle poisons induce allelic loss in mouse lymphoma cells through mitotic non-disjunction.

Aneuploidy is an important contributor to reproductive failure and tumor development. It arises spontaneously or as a result of exposure to aneugenic agents through non-disjunction. Two spindle poisons, colchicine (COL) and vinblastine (VBL) are mutagenic in the mouse lymphoma assay (MLA), a gene mutation assay that targets the heterozygous thymidine kinase (tk) gene on chromosome 11 in mouse lymphoma L5178Y tk+/- 3.7.2c cells. To investigate the mechanisms of spindle poison mutagenesis, we analyzed the COL- and VBL-induced TK mutants at the molecular and cytogenetic level. Loss of heterozygosity (LOH) analysis employing a microsatellite region within the tk locus revealed that almost all mutants had lost the functional tk allele. To determine the extent of the LOH, we further examined LOH mutants for heterozygosity at nine microsatellite loci spanning the entire chromosome 11. Interestingly, every microsatellite marker showed LOH in all COL- and VBL-induced LOH mutants, suggesting that these mutants were generated by loss of the whole chromosome 11 through mitotic non-disjunction. Chromosome painting analysis supported this hypothesis; there were no mutants showing structural changes such as deletions or translocations involving chromosome 11. In contrast, spontaneous TK mutants followed from point mutations, deletions and recombinational events as well as whole chromosome loss. Our present study indicates that spindle poisons induce mutations through mitotic non-disjunction without structural DNA changes and supports a possible mechanism in which a recessive mutation mediated by aneuploidy may develop tumors.

Hepatocarcinogen quinoline induces G:C to C:G transversions in the cII gene in the liver of lambda/lacZ transgenic mice (MutaMouse).

Quinoline is carcinogenic to the liver in rodents, but it is not clear whether it acts by a genotoxic mechanism. We previously demonstrated that quinoline does induce gene mutation in the liver of lambda/lacZ transgenic mice. In the present report, we reveal the molecular nature of the mutations induced by quinoline in the lambda cII gene, which is also a phenotypically selectable marker in the lambda transgene. (The cII gene has 294bp, which enables much easier sequence analysis than the original lacZ gene (3kb)). The liver cII mutant frequency was nine times higher in quinoline-treated mice than in control mice. Sequence analysis revealed that quinoline induced primarily G:C to C:G transversions (25 of 34). Thus, we have confirmed that quinoline is genotoxic in its target organ, and the G:C to C:G transversion is the molecular signature of quinoline-induced mutations.

Requirement of wild-type p53 protein for maintenance of chromosomal integrity.

Chromosomal double-strand breaks (DSBs) occurring in mammalian cells can initiate genomic instability, and their misrepairs result in chromosomal deletion, amplification, and translocation, common findings in human tumors. The tumor-suppressor protein p53 is involved in maintaining genomic stability. In this study, we demonstrate that the deficiency of wild-type p53 protein may allow unrepaired DSBs to initiate chromosomal instability. The human lymphoblastoid cell line TK6-E6 was established by transfection with human papilloma virus 16 (HPV16) E6 cDNA into parental TK6 cells via a retroviral vector. Abrogation of p53 function by E6 resulted in an increase in the spontaneous mutation frequencies at the heterozygous thymidine kinase (TK) locus but not at the hemizygous hypoxanthine phosphoribosyl transferase (HPRT) locus. Almost all TK-deficient mutants from TK6-E6 cells exhibited loss of heterozygosity (LOH) with the hemizygous TK allele. LOH analysis with microsatellite loci spanning the long arm of chromosome 17, which harbors the TK locus, showed that LOH extended over half of 17q toward the terminal end. Cytogenetic analysis of LOH mutants by chromosome painting indicated a mosaic of chromosomal aberrations involving chromosome 17, in which partial chromosome deletions, amplifications, and multiple translocations appeared heterogeneously in a single mutant. We speculate that spontaneous DSBs trigger the breakage-fusion bridge cycle leading to such multiple chromosome aberrations. In contrast, no chromosomal alterations were observed in TK-deficient mutants from TK6-20C cells expressing wild-type p53. In wild-type p53 cells, spontaneous DSBs appear to be promptly repaired through recombination between homologous chromosomes. These results support a model in which p53 protein contributes to the maintenance of genomic integrity through recombinational repair.

Human monochromosome hybrid cell panel characterized by FISH in the JCRB/HSRRB.

The human monochromosome hybrid cell panel in the Japanese Collection of Research Bioresources (JCRB) consists of 23 mouse cell clones, each containing a different human chromosome (the Y chromosome is not yet included). The panel is currently distributed by the Human Science Research Resources Bank (HSRRB) in Osaka. In order to determine the state of the human chromosomes and to supply the information to investigators, we characterized the cells by fluorescence in-situ hybridization (FISH) with corresponding human chromosome-specific painting probes, and, in part, by reverse FISH with the hybrid total DNA hybridized onto human metaphase spreads. Here, we report the frequency of intact human chromosomes maintained in each hybrid and the retained subregions of corresponding human chromosomes with relative frequencies estimated by fluorescent intensity. We used specific painted patterns to classify each hybrid into tentative types with their frequencies showing the nature of each hybrid and the state of rearrangements. This characterization will provide valuable information to investigators using the panel.

In vivo genotoxicity of 2-amino-3,8-dimethylimidazo[4, 5-f]quinoxaline in lacI transgenic (Big Blue) mice.

2-Amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx), a heterocyclic amine found in cooked meat, is a strong mutagen in the Salmonella/microsome assay and was proven to be a hepatocarcinogen in rodents. We used the lacI transgenic (Big Blue(R)) mouse to investigate MeIQx genotoxicity in vivo. lacI mutant frequencies were examined in liver and colon after single intragastric administration of MeIQx (males) or 12 weeks of feeding in the diet (males and females). Micronucleus induction was monitored in the peripheral blood and cell proliferating activity was monitored by proliferating cell nuclear antigen (PCNA) immunostaining, but only after the intragastric administration. Intragastric treatment with MeIQx (100 mg/kg) did not increase mutant frequency (MF) in liver or colon but it did induce a slight but statistically significant increase in the incidence of micronucleated reticulocytes 48 h after the treatment. No apparent increase in PCNA-positive foci was observed in any of tissues analyzed 14 days after the treatment. Administration of MeIQx (300 ppm) in diet for 12 weeks, however, caused MF increases in liver and colon in male and female mice, with greater increases in the females. An increase was also obvious after 4 weeks, but only in females. The sex difference in MF is consistent with the fact that female mice are more susceptible to MeIQx carcinogenesis. These results demonstrated that in the transgenic mouse mutation assay, long-term feeding of MeIQx was more effective than single gastric exposures in revealing the compound's mutagenicity in the target organs of carcinogenicity and that sex differences in susceptibility can also be observed.

Report from the In Vitro Micronucleus Assay Working Group.

At the Washington International Workshop on Genotoxicity Test Procedures (March 25-26, 1999), the current methodologies and data for the in vitro micronucleus test were reviewed. From this, guidelines for the conduct of specific aspects of the protocol were developed. Because there are a number of important in vitro micronucleus validation studies in progress, it was not possible to design a definitive, internationally harmonized protocol at this time. Agreement was achieved on the following topics: Cells. The choice of cells is flexible, yet the choice of cell type should be justified and take into consideration doubling time, spontaneous frequency of micronuclei, and genetic background. Slide preparation. A fixation method that preserves the cytoplasm and cytoplasmic boundaries, and minimizes clumping should be used. Use of fluorescent DNA-specific dyes is encouraged for better detection of small micronuclei. Analysis. Micronuclei should have a diameter less than one-third of the main nucleus, and should be clearly distinguishable from the main nucleus. In the cytokinesis-block method, binucleated cells selected for analysis should have two clearly distinguishable main nuclei. Cells where the main nucleus(ei) is undergoing apoptosis should not be scored for micronuclei because the assumed micronuclei may have been the result of nuclear fragmentation during the apoptotic process. Toxicity. Cytotoxicity can be measured by various methods including cell growth, cell counts, nucleation (i.e., percent binucleated), division/proliferation index, confluence. A majority of the group recommended that the highest concentration should induce at least 50% cytotoxicity (by whatever measure is selected). Cytochalasin B. There is much debate regarding the use of cytochalasin B. For human lymphocytes, the use of cytochalasin B (6 microg/ml [lymphocytes cultured from whole blood cells] and 3-6 microg/ml [isolated lymphocyte cultures]) is recommended. For cell lines, because there were no definitive data showing a clear advantage or disadvantage of the use of cytochalasin B for a variety of chemicals, the majority opinion of the group was that at this time, the use of cytochalasin B for cell lines is considered optional. Further studies (many chemicals of a variety of potencies, tested both with and without cytochalasin B) are clearly needed to resolve this issue. Number of doses. At least three concentrations should be scored for micronuclei. Treatment/harvest times. At this time, there are not enough data to define the most appropriate treatment/harvest times. Following the principles of the in vitro metaphase assay (with or without metabolic activation), it was agreed that there was a need for a short treatment followed by a recovery time in the absence of test chemical, there was a need for a long treatment (maybe with and without recovery time), and ideally, treatment should cover cells in different cell cycle stages.

Appropriate levels of cytotoxicity for genotoxicity tests using mammalian cells in vitro.

Among standard battery genotoxicity assays, the in vitro chromosome aberration test and the mouse lymphoma tk assay (MLA) yield about fourfold higher incidences of positive test results than the bacterial reverse mutation test or in vivo bone marrow tests. This is a result of experience with submissions of 335 new pharmaceuticals to the German Federal Institute for Drugs and Medical Devices. While all of the standard systems have their value in detecting relevant genotoxins, there is no supportive evidence for DNA reactivity for a considerable number of in vitro clastogens and MLA positives. In particular the clastogenic response of such compounds is often associated with high cytotoxicity. This may invoke the need to change the approach to test for clastogenicity in vitro. A combination of measures such as (1) a change in the upper limits of cytotoxicity that are currently given in International Conference on Harmonisation (ICH) and Organization for Economic Co-Operation & Economic Development (OECD) guidelines, (2) the creation of a common ground of understanding for interpretation of in vitro (positive) test results, and (3) lowering the upper limits of test compound concentration irrespective of cytotoxicity may prove useful to ensure a sufficient reliability of genotoxicity testing with mammalian cells in vitro.

Semi-quantitative evaluation of genotoxic activity of chemical substances and evidence for a biological threshold of genotoxic activity.

In Japan, the Chemical Substances Control Law requires evaluation of the genotoxic potential of chemical substances semi-quantitatively by application of a ranking system. During the past 10 years under the law, 1049 new chemical substances were evaluated by a reverse mutation assay in bacteria (RMA) and a chromosome aberration test in cultured mammalian cells (CAT). Of them, 130 (12.4%) were positive in the RMA and 402 (38.3%) were positive in the CAT. Eighty (7.6%) were positive in both tests. Fifty (4.8%) were positive only in the RMA, 322 (30.7%) were positive only in the CAT, and 452 (43.1%) were positive in either the RMA or the CAT. Thus, the tests complement each other in detecting genotoxic substances in vitro. To explore the "threshold" concept, we compared the genotoxic responses of Salmonella typhimurium tester strains with and without DNA repair capacity. Recently constructed strains of TA1535 lacking O(6)-methylguanine DNA methyltransferase genes (ogt(ST) or ada(ST) and ogt(ST)) showed dose-related increases in the number of revertants induced by N-ethyl-N'-nitro-N-nitrosoguanidine, methyl methanesulfonate, dimethylnitrosamine, and ethylnitrosourea, while in the same dose ranges the parental strain TA1535 did not. This finding suggests that there is a threshold at which all DNA damage induced by low dose levels of genotoxic chemicals are repaired. That biological threshold seems to exist for both DNA and non-DNA targeting chemicals.

A micronucleus assay in sea urchin embryos.

We have developed a micronucleus assay for use in sea urchin embryos. The embryos at the early blastula stage (about 256 cells) were exposed to genotoxic chemicals overnight until control embryos have reached the gastrula stage. Then all embryos were suspended in 1 M urea, dissociated by pipetting, and fixed with methanol:acetic acid (9:1). The preparations were air-dried and stained with acridine orange. The test chemicals (mitomycin C [MMC], vinblastine and 1-beta-D-arabinofuranosylcytosine [Ara C]) induced clear micronuclei dose-dependently. The maximum frequency induced with MMC was 2-3% in Clypeaster japonicus and 1-2% in Hemicentrotus pulcherrimus.

Validation study of the in vitro micronucleus test in a Chinese hamster lung cell line (CHL/IU).

We conducted a collaborative validation study, under the auspices of the Japanese Ministry of Labour, on the in vitro micronucleus test to see if it could be used as an alternative to the in vitro chromosome aberration test for evaluation of chemical safety. We used the Chinese hamster lung cell line (CHL/IU), which is the most widely used system for the latter test in Japan, and evaluated 66 chemicals, including clastogens and polyploidy inducers. The cytochalasin B cytokinesis blocking method, which is commonly used in human lymphocyte culture, was applied to the established cell line, but did not improve the detection of chemically-induced micronuclei in continuously growing cells. The highest micronucleus frequencies were obtained at 48 or 72 h continuous treatments. In short treatments (6 h), a 42 h recovery time yielded the best responses. Concordance between the results of the micronucleus test and the chromosomal aberration test was satisfactorily high (88.7%), and we concluded that the in vitro micronucleus test could be used in place of the chromosomal aberration test as a simple and rapid method for detecting clastogens and aneugens in vitro. We also propose a protocol for the test.

UVB-induced gpt mutations in the skin of gpt delta transgenic mice.

Ultraviolet light B (UVB)-induced mutagenesis was studied in gpt delta transgenic mice, which contain the lambdaEG10 shuttle vector as a transgene. The mice were exposed to UVB at single doses of 0.3, 0.5, 1.0, 1.5, and 2.0 kJ/m(2). At 4 weeks after irradiation, the mutant frequencies (MF) of the gpt gene were determined in the epidermis and the dermis, and the gpt mutations in the epidermis were identified by DNA sequencing. The epidermis exhibited a higher sensitivity to UVB than the dermis at doses of 0.3 and 0.5 kJ/m(2) UVB: the MF of the epidermis were more than nine times higher than those of the nonirradiated mice, whereas the MF of the dermis were only two to three times higher than the nonirradiated level at the doses used. The UVB-induced mutation spectrum in the epidermis was dominated by G:C to A:T transitions at dipyrimidine sites, such as 5'-TC-3', 5'-CC-3', and 5'-T/C-CG-3'. Tandem transitions such as CC to TT were also observed. Interestingly, a remarkable bias towards the template strand of the gpt gene was observed in the single transitions at 5'-TC-3' and 5'-CC-3' sites, but not at 5'-T/C-CG-3' site. In contrast, G:C to A:T transitions at CpG sites and deletions were observed in nonirradiated mice. Hot spots of transitions were observed at different sites in UVB-irradiated and nonirradiated mice. These results indicate that gpt delta transgenic mouse is a suitable model for studying in vivo UVB-induced mutations at the molecular level.

Effects of O (6)-alkylguanine-DNA alkyltransferase deficiency in Escherichia coli as the host for the detection of mutations in lacI transgenic mice.

Transgenic mice are widely used to detect gene mutations in vivo induced by a variety of chemicals. It is known, however, that no mutagenicity of methyl methanesulfonate (MMS) is detected in epididymal sperm in various transgenic mice assays, although MMS induces the dominant lethal and specific locus mutations in male mice. To investigate the issue of whether unrepaired lesions in DNA of mature sperm can be transformed into mutations during replication of the lambda phage in Escherichia coli cells, we developed an E. coli strain YG5152, which is a derivative of strain SCS-8 but is deficient in the genes encoding O (6)-alkylguanine-DNA alkyltransferases. When lambda LIZalpha phages were treated with MMS or N-ethyl-N-nitrosourea (ENU) in vitro and infected to the E. coli strains, the mutant frequencies of lacI were markedly higher in strain YG5152 than in strain SCS-8. When Big Blue(trade mark) mice were treated with MMS (160 mg/kg) or ENU (125 or 250 mg/kg) and the phages rescued from mature sperm were infected to the strains, the mutation frequency (MF) of phages from ENU-treated mice at a dose of 250 mg/kg in strain YG5152 was about two times higher than that in strain SCS-8. However, no increase in the MF was observed in the MMS-treated mice even in strain YG5152. These results suggest that, although strain YG5152 efficiently detects ex vivo mutations caused by mutagenic alkyl adducts formed by MMS in lambda phage DNA, no detectable levels of mutagenic methyl adducts are present in mature sperm of MMS-treated mice. Possible reasons for this lack of mutagenicity of MMS in mature sperm using transgenic mice assays are discussed.

Procarbazine genotoxicity in the MutaMouse; strong clastogenicity and organ-specific induction of lacZ mutations.

Procarbazine, a drug used for cancer chemotherapy, is carcinogenic in rodent bioassays. We analyzed the mutagenicity of procarbazine in various organs and the clastogenicity of the drug in hematopoietic cells of the lacZ transgenic MutaMouse. This was part of the second collaborative study of the Mammalian Mutagenesis Study Group of the Japanese Environmental Mutagen Society on the transgenic mouse mutation assay. At 50 mg kg(-1), procarbazine induced micronuclei in hematopoietic cells, but it did not increase the lacZ mutant frequency (MF) in bone marrow. It was also negative in liver, testis, spleen, kidney, and lung. Five daily administrations of 150 mg kg(-1) yielded highly positive responses in the drug's target organs for carcinogenesis (lung, bone marrow, and spleen). Lower positive responses were detected in kidney, which is a minor target organ. Liver showed only a slight increase in lacZ MF and brain showed no increase. The testis MF more than doubled which suggest that procarbazine is mutagenic to germ cells. Thus, we demonstrated that procarbazine has a strong clastogenic effect in hematopoietic cells and is mutagenic in a variety organs after high dose treatment. The induced MF was especially high in procarbazine's target organs for carcinogenesis, which supports the relevance of the transgenic mouse mutation assay for the assessment of potential genotoxins in vivo.

Mutagenicity of 2-amino-1-methyl-6-phenylimidazo [4,5-b]pyridine (PhIP) in the new gpt delta transgenic mouse.

Gender differences and organ specificity of 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP)-induced mutagenesis were examined with the new gptdelta transgenic mouse (T. Nohmi, M. Katoh, H. Suzuki, M. Matsui, M. Yamada, M. Watanabe, M. Suzuki, N. Horiya, O. Ueda, T. Shibuya, H. Ikeda, T. Sofuni, A new transgenic mouse mutagenesis test system using Spi-and 6-thioguanine selections (Environ. Mol. Mutagen. 28 (1996) 465-470). In this mouse model, two distinct selections are employed to efficiently detect different types of mutations, i.e 6-thioguanine (6-TG) selection for point mutations and Spi-selection for deletions, respectively. In both selections, the highest mutant frequencies were observed in colon, followed by in spleen and liver. No increases in mutations were observed in testis, brain and bone marrow in PhIP-treated male mice. No significant differences in 6-TG and Spi- mutant frequencies were observed in colon and liver between male and female treated mice. The correlation between PhIP-induced mutagenesis and carcinogenesis in colon is discussed.

Evaluation of the mouse lymphoma tk assay (microwell method) as an alternative to the in vitro chromosomal aberration test.

In order to evaluate the utility of the mouse lymphoma assay (MLA) for detecting in vitro clastogens and spindle poisons and to compare it with the in vitro chromosomal aberration test (CA), we conducted an international collaborative study of the MLA that included 45 Japanese laboratories and seven overseas laboratories under the cooperation of the Ministry of Health and Welfare of Japan and the Japanese Pharmaceutical Manufacturer's Association. We examined 40 chemicals; 33 were reportedly positive in the CA but negative in the bacterial reverse mutation assay, six were negative in both assays and one was positive in both. We assayed mutations of the thymidine kinase (TK) locus (tk) of L5178Y tk +/- mouse lymphoma cells using the microwell method. According to our standard protocol, cells were exposed to the chemical for 3 h, cultured for 2 days and TK-deficient mutants were expressed in 96-well plates under trifluorothymidine. Each chemical was coded and tested by two or three laboratories. Among the 34 CA-positive chemicals, positive MLA results were obtained for 20 and negative results were obtained for nine. The remaining five chemicals were inconclusive or equivocal because of discrepant inter-laboratory results or reproduced discrepant results, respectively. Among the six CA-negative chemicals, one was negative in the MLA, two were positive and three were inconclusive. Thus, the MLA could detect only 59% (20/34) of CA-positive chemicals. We concluded that the MLA was not as sensitive as the CA. Some MLA-negative chemicals evoked positive responses in the CA only after long continuous treatment. These might also be genotoxic in the MLA with long continuous treatment. Improvement of the MLA protocol, including alteration of the duration of the treatment, might render the MLA as sensitive as the CA.