Nitric oxide complexes in the interaction between primary and secondary tumor of L5178Y lymphoma.

Heme and non-heme Fe-NO complexes were observed in regard to the growth of primary and secondary solid tumors and ascites of murine L5178Y lymphoma. The complexes were detected by electron paramagnetic resonance spectroscopy at liquid nitrogen temperature. Primary solid tumors and secondary solid tumors or ascites were inoculated on the same day, or with a delay. The primary tumor inhibited growth of the secondary solid tumor only if the latter was inoculated with a delay, which did not correlate with the change of the types, nor with the increase in the level of Fe-NO complexes detected in the tissue, suggesting a "non-immunological" character of this inhibition. In some animals with solid tumors, spontaneous ascites developed. This process resulted in a marked decrease in the level of Fe-NO complexes in the solid tumor tissue. The primary solid tumor, however, did not influence the growth of secondary ascites, but intensified NO generation in the ascites of animals with partial removal of ascitic fluid. This experimental group survived 2.2 days longer than the control group without primary solid tumor. Our research revealed that the presence of Fe-NO complexes in the interaction between primary and secondary tumor strongly depends on the form of the tumor: solid or ascitic, and that murine L5178Y lymphoma may serve as a convenient model for the research on "concomitant immunity" against in vivo growing tumors. This is the first EPR study on "concomitant immunity" in regard to tumor-tumor and tumor-ascites interactions in vivo.

Activity of 3'-(3-cyano-4-morpholinyl)-3'-deaminoadriamycin in sensitive and resistant L5178Y lymphoblasts in vitro.

An L5178Y murine lymphoblast cell line resistant to 3'-(3-cyano-4-morpholinyl)-3'-deaminoadriamycin (MRA-CN), L5178Y/MRA-CN, was isolated and characterized. L5178Y/MRA-CN cells were 9.6-fold resistant to MRA-CN compared with parental cells. The resistant cell line also displayed 2-fold resistance to 3'-(4-morpholinyl)-3'-deaminoadriamycin but was not cross-resistant to Adriamycin or chlorambucil. Uptake of MRA-CN was slightly reduced in the resistant cells compared to sensitive cells, but the distribution of the drug within the cells was unchanged. DNA interstrand cross-linking by MRA-CN was not significantly different in the sensitive and resistant cell lines, but MRA-CN was slightly less effective in inhibiting both DNA and RNA synthesis in L5178Y/MRA-CN cells compared with parental cells. NADPH cytochrome P-450 reductase activity was increased in L5178Y/MRA-CN cells compared to parental cells, while the activity of DT-diaphorase was decreased in the resistant cells. The levels of glutathione and glutathione S-transferase activity were increased in the resistant cells compared to sensitive cells; however, pretreatment of L5178Y/MRA-CN cells with buthionine sulfoximine to reduce the glutathione level did not reverse the resistance of these cells to MRA-CN. MRA-CN induced DNA fragmentation that was characteristic of apoptosis in both L5178Y and L5178Y/MRA-CN cells at equitoxic drug concentrations. However, apoptosis occurred more rapidly in L5178Y/MRA-CN cells compared with parental cells. Thus, MRA-CN induces apoptosis in L5178Y cells, and this effect may be important for the anti-tumor activity of this agent. In contrast, DNA interstrand cross-linking does not appear to be the primary mechanism responsible for the cytotoxicity of MRA-CN in these cells.

Inhibition of repair of bleomycin-induced DNA strand breaks by 2'-deoxycoformycin and its effect on antitumor activity in L5178Y lymphoblasts.

We have observed previously that treatment of plateau-phase L5178Y murine lymphoblasts in vitro with 2'-deoxycoformycin plus deoxyadenosine (dCF/dAdo) can inhibit the repair of X-irradiation-induced DNA single-strand breaks (SSB) in these cells and that this effect is associated with synergistic cell kill. In this study we examined the effect of a combination treatment of plateau-phase L5178Y cells with bleomycin (BLM) plus dCF/dAdo. Incubation of BLM-treated cells with dCF/dAdo resulted in significant inhibition of the repair of BLM-induced DNA SSB. However, an additive, but not a synergistic, increase in cell kill was observed when cells were treated with a combination of BLM plus dCF/dAdo.

Cellular recovery in two sub-lines of the L5178Y murine leukaemic lymphoblast cell line differing in their sensitivity to ionizing radiation.

Cellular recovery was assessed in two sublines of L5178Y murine lymphoma cells of differing radiosensitivity (LY-S and LY-A4) using low dose-rate irradiation and split-dose experiments. No increase in cell survival was observed in the LY-S cell line until the dose-rate was reduced to 2 cGy/min, whereas in the LY-A4 cell line 20 cGy/min was low enough to detect changes in survival. The extent of this change, as assessed by dose reduction factors at 2 logs of cell kill, was greater in the LY-A4 cell line. Fitting these data with the incomplete repair model of Thames led to anomalous values for the half-time of repair. In split-dose experiments the maximum observed recovery ratio increased as a function of dose in a manner that is consistent with the linear-quadratic equation. As was found previously with radiosensitive human tumour cells, the LY-S cell line showed more split-dose recovery at any given dose than the LY-A4 cell line.

Specific interaction between gangliotriaosylceramide (Gg3) and sialosyllactosylceramide (GM3) as a basis for specific cellular recognition between lymphoma and melanoma cells.

In view of the possible role of glycosphingolipids in defining the specificity of cell-cell interactions, the key molecules for recognition of cell surface glycosphingolipids have been studied. In addition to previously suggested recognition mechanisms involving endogenous lectins and glycosyltransferases, an alternative possibility, based on carbohydrate-carbohydrate (Lex-Lex) interaction, has been raised (Eggens, I., Fenderson, B., Toyokuni, T., Dean, B., Stroud, M., and Hakomori, S. (1989) J. Biol. Chem. 264, 9476-9484). We now report a highly specific interaction between gangliotriaosylceramide (Gg3, GalNAc beta 1----4Gal beta 1----4Glc beta 1----Cer) and sialosyllactosylceramide (GM3, NeuAc alpha 2----3Gal beta 1----4Glc beta 1----Cer). The interaction requires a bivalent cation (Ca2+ or Mg2+) and can be inhibited by sialosyl 2----3 lactose, anti-GM3 antibody (DH2), anti-Gg3 antibody (2D4), or EDTA. The strength of interaction between GM3 liposome and the Gg3-coated plastic surface was highly density-dependent. The mouse lymphoma L5178 AA12 cell line (high expressor of Gg3) interacted specifically with the mouse B16 melanoma cell line (high expressor of GM3). The interaction was inhibited by 5 mM sialosyllactose, anti-GM3 antibody, anti-Gg3 antibody, and EDTA in analogy to GM3-Gg3 interaction. L5178 AV27, a genetically related variant clone which does not express Gg3, showed no interaction with B16 cells. Untreated AA12 cells, but not 2D4-treated AA12 cells or AV27 cells, interacted with GM3 coated on the plastic surface. These findings suggest a specific interaction between AA12 cells and B16 cells based on Gg3-GM3 interaction.

Characterization of L5178Y murine lymphoblasts resistant to quinone antitumor agents.

The exact contribution of the quinone group to the activity of quinone antitumor agents remains uncertain. Two L5178Y murine lymphoblastic cell lines resistant to the model quinone antitumor agent, hydrolyzed benzoquinone mustard, and one partial-revertant cell line were isolated and characterized. The antitumor activity of hydrolyzed benzoquinone mustard has been shown previously to be due to its ability to induce free radical mediated DNA strand breaks. Resistant cells were obtained by growing a cloned L5178Y parental cell line in media containing increasing concentrations of hydrolyzed benzoquinone mustard. L5178Y/HBM2 cells were selected from L5178Y cells growing in media containing 0.2 mM drug, while L5178Y/HBM10 cells were selected from cells growing in media containing 1.0 mM drug. The L5178Y/HBMR cells were obtained by growing L5178Y/HBM10 cells in media without hydrolyzed benzoquinone mustard. The resistant cell lines, L5178Y/HBM2 and L5178Y/HBM10, were 2.5- and 6-fold less sensitive, respectively, to hydrolyzed benzoquinone mustard compared to parental cells, and this was accompanied by a decrease in the formation of DNA single and double strand breaks by this drug. The partial-revertant cell line, L5178Y/HBMR was 2.9-fold less sensitive to hydrolyzed benzoquinone mustard compared to parental cells. Drug uptake appeared to be lower in the resistant cells compared to parental cells. The resistant cells had a slightly elevated level of superoxide dismutase activity compared to parental cells, but there was no increase in the mRNA for superoxide dismutase nor any amplification of the gene for this enzyme. Intracellular catalase activities of the L5178Y/HBM2 and L5178Y/HBM10 cells were elevated by 1.25- and 2.6-fold, respectively, and the increased enzyme activity in the L5178Y/HBM10 cells appeared to result from a 3.6-fold increase in mRNA for this enzyme. Glutathione peroxidase activity was slightly elevated in L5178Y/HBM2 cells, but was unchanged in the other resistant cells. The L5178Y/HBM2 and L5178Y/HBM10 cells showed increased concentrations of glutathione and elevated levels of glutathione transferase activity. The resistant cell lines also had DT-diaphorase activity that was 3- and 24-fold higher in L5178Y/HBM2 and L5178Y/HBM10 cells, respectively, compared to sensitive cells. However, cytochrome P-450 reductase activity and the ratio of reduced to oxidized pyridine nucleotides was unchanged in the resistant cell lines. The partial-revertant cell line, L5178Y/HBMR, showed approximately the same level of resistance to hydrolyzed benzoquinone mustard as the L5178Y/HBM2 cells.(ABSTRACT TRUNCATED AT 400 WORDS)

Murine tumor-induced platelet aggregation and coagulation: mechanisms, inhibitors, and species differences.

This study examined the platelet-aggregating and procoagulant activities of two hematogenously disseminating tumors, a mouse lymphoblastic leukemia (L5178Y) and a mouse renal adenocarcinoma (RAG). Tumor-induced human platelet aggregation was inhibited by addition of the following agents to platelet-rich plasma (PRP): a calcium channel blocker (verapamil), a chelator of divalent cations (EDTA), stimulators of adenylate cyclase (2-fluoroadenosine and forskolin), and inhibitors of cAMP phosphodiesterase (oxagrelate and papaverine). The platelet-aggregating activities of both cell lines were completely blocked by treatment of the cells with heat, sonication, phospholipase A2, and Triton X-100. These data suggest that L5178Y and RAG cell-induced human platelet aggregation are dependent on a heat-labile phospholipid component of the tumor cell membrane. L5178Y cells had greater platelet-aggregating activity in human plasma than in rat or mouse plasma, whereas RAG cells had greater procoagulant activity in rat or mouse plasma than in human plasma. The procoagulant activity of RAG cells in rat and mouse plasma was demonstrated by three lines of evidence: RAG cells induced heparinized PRP to clot; the thrombin inhibitor DAPA lengthened of the clotting time and the lag time before aggregation; and RAG cells shortened of the recalcification time of the plasma. The above data indicate that RAG cell-induced murine platelet aggregation and coagulation is dependent on thrombin generation.

Predefined gene transfer for expression of a glycosphingolipid antigen by transfection with a cosmid genomic library prepared from a cell line in which the specific glycosphingolipid is highly expressed.

The deliberate transfer of globotriaosylceramide (Gb3) expression in mouse lymphoma L5178 cells was achieved by transfection with a cosmid DNA library prepared from human Burkitt lymphoma Ramos cells in which Gb3 was highly expressed. The recipient mouse lymphoma cells did not contain Gb3 but did contain its direct precursor, lactosylceramide. The transfected cells expressed Gb3, detected both chemically and immunologically, and contained human DNA detected by an Alu sequence probe. This model demonstrates a general method for studying glycosyltransferase genes and other factors necessary for the expression of glycosphingolipid antigens.

Reduced pinosome formation in a cell-fusion-impaired mutant of mouse lymphocytic cell line L5178Y.

Accumulation and release of a fluid-phase marker ([14C]-sucrose) were studied in a subline of the mouse lymphocytic cell line L5178Y and in a polyethylene glycol-resistant, intercellular-fusion-impaired mutant of this line. The mutant was found to accumulate [14C]-sucrose at a significantly slower rate than the parent. Analysis of release of preloaded label shows that the reduced rate of accumulation is due to a correspondingly low level of internalization, i.e., pinosome formation, rather than to a reduction in delivery of label to lysosomes. Cell-to-cell fusion and pinosome formation both involve a fusion event initiated at the extracellular surface of the plasma membrane, and we propose that the coordinate reduction in both processes suggests that they are mechanistically related.

Electric pulse-induced fusion of mouse lymphoma cells: roles of divalent cations and membrane lipid domains.

Mouse leukemic lymphoblasts (L5178Y) brought into close contact by dielectrophoresis underwent cell fusion following the application of electrical pulses in the presence of electrolytes. The electrically fused cells became spherical after switching off the dielectrophoretic field. Fusion between a cell vitally stained with Janus Green and that with Neutral Red resulted in the homokaryon with a mixed color. Intracellular potentials simultaneously recorded from the two cells located on both sides of the homokaryon were identical. The fusion efficiency was remarkably dependent upon temperature, displaying a discontinuity at about 11 degrees C in the Arrhenius plot. The extracellular application of phospholipase-A2 or -C suppressed the fusion yield. Thus, it appears that the phospholipid domains play a crucial role in the electric pulse-induced cell fusion. Treatment of the cells with proteolytic enzymes markedly enhanced the fusion yield, presumably due to removing the glycocalix and/or giving rise to fusion-potent, protein-free lipid domains. The presence of millimolar concentrations of divalent cations (irrespective of Mg2+ or Ca2+) as well as of micromolar concentrations of Ca2+ (but not Mg2+) was prerequisite to the resealing of membranes suffered from electrical breakdown upon exposure to electric pulses. In addition, extracellular Ca2+ (but not Mg2+) ions at more than micromolar concentrations were indispensable for the cell fusion.

Microelectrophoretic study of radiation-induced DNA damages in individual mammalian cells.

Mammalian cells were after irradiation suspended in melted agarose, and casted on microscope slides. The slides were after gelling at 0 degree C immersed in a neutral detergent solution which lysed the cells. A weak electric field (5 V/cm) was then applied over the gel for 5 minutes. The DNA in the gel was stained with the fluorescent dye acridine orange and gives a green emission in a microscope photometer. DNA had migrated towards the anode and this migration was more pronounced in irradiated than in control cells. The differences in migration pattern were quantitatively measured. The lower detection limit was below 0.5 Gy and a plateau in the dose-effect curve was reached at about 3 Gy. In repair experiments residual DNA damage could be observed after postirradiation incubation for 60 minutes. The advantages of the method is: no radioactive labelling and only a few number of cells is required.

Characterization of L5178Y cell phenotypes isolated during progression of the tumor-dormant state in DBA2 mice.

During the course of the L5178Y tumor-dormant state in DBA/2 mice, there is continual selection of a tumor cell subpopulation ("emergent" phenotype) from the uncloned original L5178Y population used to initiate the tumor-dormant state. In vivo and in vitro experiments show that the emergent-phenotype tumor cells are less capable than "original"-phenotype cells, which constitute the majority of the L5178Y cell inoculum, of restimulating cytolytic T-lymphocyte (CTL) activity in tumor-dormant mice and are less susceptible to lysis by those CTL. Both original- and emergent-phenotype tumor cells are capable of eliciting an immune CTL response in naive mice, but again emergent-phenotype cells are poorly lysed by this response. As a consequence of these characteristics, emergent-phenotype cells rapidly form ascitic tumors when used as a challenge in L5178Y cell-immunized mice and cannot establish a tumor-dormant state. Results presented here and in previous publications indicate that CTLs are the major host cells involved in the selection of emergent-phenotype L5178Y cells during the course of the tumor-dormant state. Heterogeneity of the tumor cell challenge inoculum is important in establishing the L5178Y tumor-dormant state. The state, once established, is maintained by an immune CTL response which is continuously being restimulated by the strongly immunogenic original-phenotype L5178Y cells. The tumor-dormant state terminates when the less immunogenic and more immunoresistant emergent-phenotype tumor cells predominate and escape the waning immune response.

Somatic hybridization between human and mouse lymphoblast cells produced by an electric pulse-induced fusion technique.

The technique of electric pulse-induced cell fusion (electro-fusion) was used to obtain heterokaryons between normal human lymphoblasts (HSC93) and mouse leukemic lymphoblasts (MCN151). The two types of cells were brought into contact in the cell suspension by dielectrophoresis with an alternating electric field (0.8 kV/cm, 100 kHz) in the presence of calcium ions and pronase E. Cell fusion was induced by giving two successive electric pulses (3.3 and 5 kV/cm, 10 microsec). Prior treatment of human (but not mouse) lymphoblasts with neuraminidase improved fusion efficiency. Differential staining of the two types of cells with Janus Green and Neutral Red showed that about 40% of the viable fused cells underwent heterokaryonic fusion. We concluded that electrofusion is an efficient method for obtaining heterokaryons from human and mouse lymphoblasts.

Effect of asparaginase on cell membranes of sensitive and resistant mouse lymphoma cells.

High concentrations of Escherichia coli asparaginase (80 U/ml) altered the binding of concanavalin A (Con A) to L 5178Y murine lymphoma cells that are sensitive to the cytotoxic action of this enzyme. Incubation of the asparaginase sensitive line in asparagine-free media or media containing Acinetobacter glutaminase-asparaginase did not alter the Con A binding of these cells. Escherichia coli asparaginase had no effect on Con A binding of two asparaginase resistant L5178Y cell lines that were isolated and maintained in asparagine depleted or asparaginase containing medium. The E. coli asparaginase preparation inhibited protein and glycoprotein biosynthesis to comparable degrees. It did not have proteolytic or glycolytic activity. Escherichia coli asparaginase did not alter the binding of wheat germ, soybean or ricin agglutinins to any of these cell lines. These data suggest that high concentrations of E. coli asparaginase have a specific effect on the Con A receptor in the sensitive line.

Phenotypic shifts in the L5178Y lymphoma population during progression of the tumor-dormant state in DBA/2 mice.

We evaluated the phenotypic changes that take place in the L5178Y tumor cell population during the establishment, maintenance, and termination of the L5178Y tumor-dormant state in DBA/2 mice. Since there is considerable variation among DBA/2 mice in the course of the L5178Y tumor-dormant state, analysis of these phenotypic changes required prospective studies on individual mice. Therefore, we developed a partial peritoneal lavage technique which could be performed repeatedly on an individual mouse without killing the mouse and without terminating the tumor-dormant state. The partial peritoneal lavage technique provided cell samples that were quantitatively representative of the entire peritoneal cell contents without altering the relative proportions or total number of host peritoneal cell subpopulations. Tumor cell clones were isolated from the uncloned L5178Y cell inoculum that was used to initiate the tumor-dormant state, and from the partial peritoneal lavage performed on individual mice at regular intervals throughout the tumor-dormant state until each mouse developed an ascitic tumor. We found that the uncloned L5178Y cell population is heterogeneous, with the majority of the cells [approximately 80% of the clonal populations ("original" phenotype)] able to compete as well as the uncloned L5178Y cell population for immune cytolytic T-lymphocyte activity, and a smaller percentage [approximately 20% of the clones ("emergent" phenotype)] competing significantly less but similar to the tumor cell population which grows out as an ascitic tumor at the termination of the tumor-dormant state. However, both "original" and "emergent" phenotypes encompass a range of subpopulations which have the respective characteristics. There was a progressive enrichment in the tumor cell population for "emergent"-phenotype cells as mice progressed through the tumor-dormant state. In the 30-day period prior to tumor emergence, only "emergent"-phenotype cell clones were isolated. These data indicate that, during the course of the tumor-dormant state, there is continual selection by immune cytolytic T-lymphocytes for "emergent"-phenotype cells, which preexist in the L5178Y cell population used to initiate the tumor-dormant state, and that the progressive enrichment of "emergent"-phenotype cells in the peritoneal cavity is associated with and may be responsible for termination of the tumor-dormant state.

Locus specificity of mutagenicity of 2,4-diaminotoluene in both L5178Y mouse lymphoma and AT3-2 Chinese hamster ovary cells.

2,4-Diaminotoluene, a hepatocarcinogenic aromatic amine, was tested for mutagenic potential at both the autosomal tk locus and the sex-linked hgprt locus of both L5178Y 3.7.2C mouse lymphoma cells and Chinese hamster ovary (CHO) AT3-2 cells. This compound was mutagenic in both cell types at the tk locus but not at the hgprt locus. Mutagenic activity was observed in L5178Y cells only in the absence of exogenous metabolic activation, but was observed in CHO-AT3-2 cells both with and without activation.

Enhanced suppressor macrophage activity associated with termination of the L5178Y cell tumor-dormant state in DBA/2 mice.

Both cytolytic T-lymphocytes and cytolytic macrophages have been implicated in the long-term maintenance of L5178Y cells in a tumor-dormant state in DBA/2 mice. Eventually, however, the tumor-dormant state terminates, and all mice develop ascitic tumors. In an evaluation of the mechanisms involved in termination of the tumor-dormant state, we detected in the peritoneal cavity of many tumor-dormant mice macrophages with increased capacity to suppress the in vitro generation of a secondary anti-L5178Y cell cytolytic T-lymphocyte response. The incidence of macrophage-mediated immunosuppressive activity in individual tumor-dormant mice was related directly to the number of tumor cells in the peritoneal cavity of those mice. Furthermore, in tumor-dormant mice harboring fewer than 5 X 10(4) L5178Y cells, the detection of macrophage-mediated immunosuppressive activity was a prognostic indicator of termination of the tumor-dormant state and development of an ascitic tumor. These data suggest that peritoneal macrophage-mediated immunosuppressive activity, through inhibition of cytolytic T-lymphocyte generation in vivo, contributes to the termination of the tumor-dormant state and development of ascitic tumors.

A comparison of the agar cloning and microtitration techniques for assaying cell survival and mutation frequency in L5178Y mouse lymphoma cells.

Microtitration methods for assaying cell survival and mutation frequency to ouabain resistance, 6-thioguanine resistance and 1-beta-D-arabinofuranosyl cytosine resistance in L5178Y mouse lymphoma cells were compared to the standard agar cloning technique. The two methods gave essentially similar results for untreated cells, and after treatment with ethyl methanesulphonate and 4-nitroquinoline 1-oxide. Potential advantages of the microtitration method as a routine assay system are discussed.

Mutations resistant to bromodeoxyuridine in mouse lymphoma cells selected by repeated exposure to EMS. Characteristics of phenotypic instability and reversion to HAT resistance by 5-azacytidine.

Mutations induced by repeated EMS treatments were investigated by using mouse L5178Y cells. The frequency of TGr mutations increased linearly with the number of EMS treatments whereas the yield of BrdUrdr mutations showed a curvilinear dose-response curve. The BrdUrdr frequency was roughly proportional to the square of the TGr frequency and the results were compatible with the hypothesis that BrdUrdr colonies isolated after 6 EMS treatments, however, were unstable. When BrdUrdr colonies that had arisen in BrdUrd medium after 2 weeks' incubation were isolated in normal medium, the descendant cells showed a nearly normal level of thymidine incorporation and low plating efficiencies of about 1% in BrdUrd medium. In contrast, after isolation of the same colonies in BrdUrd medium, a low level of thymidine incorporation and high plating efficiencies in BrdUrd medium were observed in the descendant cells. Reverse selection from BrdUrdr to HATr was accomplished with frequencies of 10(-6)-10(-3) for the descendants grown in BrdUrd medium, and AzaCyd treatment drastically increased the reversion frequency to nearly 10(-1). Further re-revertants from HATr to BrdUrdr were also found with frequencies of 10(-3)-10(-2) without treatment. These results indicate that the initial BrdUrdr cells did not result from inactivation of the thymidine-kinase gene but that the mode of gene expression was altered in some way.

The genetic toxicology of Kathon biocide, a mixture of 5-chloro-2-methyl-4-isothiazolin-3-one and 2-methyl-4-isothiazolin-3-one.

Kathon biocide, an aqueous solution containing a mixture of 5-chloro-2-methyl-4-isothiazolin-3-one and 2-methyl-4-isothiazolin-3-one in an approximate ratio of 3:1, was tested for mutagenic activity in Salmonella typhimurium, L5178Y mouse lymphoma cells in culture and Drosophila melanogaster. Tests also were conducted for chromosome aberrations in vivo on mouse bone marrow cells, for DNA damage/repair in primary rat hepatocytes in culture, and for morphological transformation in C3H 10T1/2 cells in culture. Kathon biocide produced point mutations in the absence of a rat-liver metabolizing system in bacteria (strain TA 100) and mammalian cells in culture. In the presence of rat-liver metabolizing system a 10-fold higher concentration was required to induce point mutations in mammalian cells in culture. No mutagenic activity was observed with the metabolizing system and S. typhimurium. Negative results were obtained in the sex-linked recessive lethal assay in Drosophila, the in vivo cytogenetic assay in mice, the unscheduled DNA synthesis assay in cultured rat hepatocytes, and the in vitro cell transformation assay.