New approaches to assessing the effects of mutagenic agents on the integrity of the human genome.

Heritable genetic alterations, although individually rare, have a substantial collective health impact. Approximately 20% of these are new mutations of unknown cause. Assessment of the effect of exposures to DNA damaging agents, i.e. mutagenic chemicals and radiations, on the integrity of the human genome and on the occurrence of genetic disease remains a daunting challenge. Recent insights may explain why previous examination of human exposures to ionizing radiation, as in Hiroshima and Nagasaki, failed to reveal heritable genetic effects. New opportunities to assess the heritable genetic damaging effects of environmental mutagens are afforded by: (1) integration of knowledge on the molecular nature of genetic disorders and the molecular effects of mutagens; (2) the development of more practical assays for germline mutagenesis; (3) the likely use of population-based genetic screening in personalized medicine.

Future approaches to genetic toxicology risk assessment.

Short-term genetic toxicology tests were developed for the purpose of identifying chemical carcinogens in the environment. After two decades of development and validation, the tests are well-established in routine testing schemes, but our views of their utility for safety evaluation have undergone re-assessment. The correlation between identified mutagens and identified carcinogens has turned out to be significantly less than one. Processes or mechanisms that are not directly genotoxic appear to play a role in carcinogenesis. While short term test data are still components of the assessment of carcinogenic risk, genetic damage also has been recognized as important in its own right, in relation to heritable genetic risk and other health-related effects, such as aging, reproductive failure and developmental toxicity. The revolution in molecular biology and genetic analysis occurring over the past 20 years has contributed to the wealth of new information on the complexities of cell regulation, differentiation, and the carcinogenic process. These technologies have provided new experimental approaches to genetic toxicology assessments, including transgenic cell and animal models, human monitoring, and analysis of macromolecular interactions at environmentally relevant exposures. The potential exists for the development of more efficient and more relevant genetic toxicology testing schemes for use assessing human safety. A delineation of contemporary needs, a modern view of the elements of cancer induction, and an examination of new assays and technologies may provide a framework for integrating new approaches into current schemes for evaluating the potential genetic and carcinogenic risk of environmental chemicals.

Photoactivated gilvocarcin V induces DNA-protein crosslinking in genes for human ribosomal RNA and dihydrofolate reductase.

The nature of DNA interactions with photoactivated gilvocarcin V has been analyzed at the gene level in both rRNA and dihydrofolate reductase genes of human fibroblasts, utilizing a modified Southern hybridization technique. Neither interstrand DNA crosslinking nor RNA linkage to DNA was detected. However, we consistently observed in both genes retarded DNA bands appearing in a dose-dependent fashion following exposure to photoactivated gilvocarcin V. These retarded bands were enhanced when genomic DNA was prepared without proteinase K treatment, suggesting involvement of protein in this DNA interaction. Because these bands disappear following proteinase K treatment, it is probable that photoactivated gilvocarcin V induces DNA-protein crosslinking.

Examination of alpha-carbonyl derivatives of nitrosodimethylamine and ethylnitrosomethylamine as putative proximate carcinogens.

Metabolites produced by enzymic oxidation are believed to be responsible for the mutagenicity and carcinogenicity of N-nitrosamines. Although alpha-hydroxy compounds are often considered, a related and more stable oxidation product, the alpha-carbonyl compound, was studied here. The alpha-carbonyl derivatives of nitrosodimethylamine (NDMA) and ethylnitrosomethylamine (oxidized at either the methyl or the ethyl group) were synthesized. The derivatives were methylnitrosoformamide (MNFA), ethylnitrosoformamide (ENFA) and methylnitrosoacetamide (MNAA). These compounds were then studied as potential toxic, mutagenic and carcinogenic intermediates. All three compounds were very potent directly acting mutagens to Salmonella typhimurium TA1535. Mutational Fingerprints in Escherichia coli of MNFA and ENFA (but not MNAA) matched those produced by SN1-type methylating and ethylating compounds respectively. The latter results indicate that the two alkylnitrosoformamides could be intermediates in the mutagenicity of the parent nitrosamines. In animal studies the putative metabolite MNFA was more acutely toxic than NDMA in F344 rats. In chronic experiments with MNFA in F344 rats and Syrian golden hamsters, tumors of the forestomach were induced by oral administration in most animals (except female hamsters) within 8 months. The properties of these oxidized derivatives of N-nitrosamines are consistent with expectations for proximate carcinogenic intermediates.

Mutagenicity of glyceryl trinitrate (nitroglycerin) in Salmonella typhimurium.

The recent finding that the clinical nitrovasodilator, glyceryl trinitrate (GTN), is mutagenic in Salmonella typhimurium strain TA1535 has been examined in closer detail, with emphasis on its mechanism of action. GTN increased the number of His+ revertants to a maximum of 4 times over background at a GTN dose of 5 mumol/plate. Hamster liver S9 depressed the toxicity of high GTN doses and increased the maximum number of revertants to 5 times over background at 10 mumol/plate. GTN did not cause significant reversion in any of the six other S. typhimurium strains tested (TA1975, TA102, TA1538, TA100, TA100NR, YG1026), although signs of toxicity were observed. Therefore, the mutagenicity of GTN was manifest only in the repair-deficient (uvrB and lacking in pKM101) strain which is responsive to single base changes. Oligonucleotide probe hybridization of TA1535 revertants showed that virtually all of the GTN-induced mutants contained C-->T transitions in either the first or second base of the hisG46 (CCC) target codon, with a preference for the latter. A similar mutational spectrum was seen previously with a complex of spermine and nitric oxide (NO) which releases nitric oxide. This suggests that NO, which can be derived from GTN via metabolic reduction, may be responsible for GTN's mutagenic action. The known NO scavenger oxymyoglobin did not substantially alter the dose response of GTN, indicating that extracellular NO was not mediating reversion. The data are consistent with the hypothesis that intracellular nitric oxide is responsible for the observed mutations.

DNA deaminating ability and genotoxicity of nitric oxide and its progenitors.

Nitric oxide (NO), a multifaceted bioregulatory agent and an environmental pollutant, can also cause genomic alterations. In vitro, NO deaminated deoxynucleosides, deoxynucleotides, and intact DNA at physiological pH. That similar DNA damage can also occur in vivo was tested by treating Salmonella typhimurium strain TA1535 with three NO-releasing compounds, including nitroglycerin. All proved mutagenic. Observed DNA sequence changes were greater than 99% C----T transitions in the hisG46 (CCC) target codon, consistent with a cytosine-deamination mechanism. Because exposure to endogenously and exogenously produced NO is extensive, this mechanism may contribute to the incidence of deamination-related genetic disease and cancer.

Discrimination of mutagenic intermediates derived from alkylating agents by mutational patterns generated in Escherichia coli.

Reactive intermediates (ultimate mutagens/carcinogens) generated by alkylating agents are unstable and difficult to characterize by chemical means. We have used a genetic system to distinguish the in vivo interactions of eight carcinogenic methylating agents and five ethylating agents by the patterns of induced mutations at different target sites in Escherichia coli WU3610. For this multiple locus assay, target sites were an amber (TAG) and an ochre (TAA) triplet, DNA encoding five suppressor tRNA anticodons, and one unidentified locus. Most of the mutations could be classified as specific sequence changes at the target loci by suppressor analysis using T4 bacteriophage. Ratios of the slopes of dose-response curves for induced mutations were used to generate a profile of preferred sites for mutagenesis independent of mutagen potency. 'Mutational fingerprints' derived from different methylating and ethylating agents were compared, as evidence for the existence of common intermediates responsible for their biological effects. Six methylating agents thought to act via SN1 mechanisms were found to generate similar mutational patterns, indicative of a common mechanism, while two methylating agents reacting via SN2 mechanisms gave different patterns. The mutational fingerprints of SN1- and SN-type ethylating agents were also distinct. Mutational fingerprints may be useful in distinguishing the interactions of different ultimate mutagens.

Inhibition of herpes virus plaquing capacity in human diploid fibroblasts treated with gilvocarcin V plus near UV radiation.

The capacity of human fibroblasts to support plaque formation by Herpes simplex virus following treatment of the cells with gilvocarcin V, a polyaromatic C-glycoside, plus near ultraviolet radiation (UVA, 320-400 nm) was examined. Gilvocarcin V, plus UVA radiation, effectively inhibited host cell capacity at concentrations five orders of magnitude lower than that of 8-methyoxypsoralen required for capacity inhibition at similar levels of UVA radiation. This result extends the observation of unusual biological potency of UVA-activated gilvocarcins from bacterial cells to human cells.

An ESR study of the visible light photochemistry of gilvocarcin V.

Photolysis of gilvocarcin (GV) at 405 nm in argon saturated dimethylsulfoxide (DMSO) or 50% DMSO-water solutions in the presence of the sodium salt of 3,5-dibromo-2,6-dideutero-4-nitrosobenzene sulfonic acid (DBNBS-d2) generates the CH3-DBNBS-d2.spin adduct. It is postulated that this spin adduct is produced by photoreduction of DMSO by GV and the consequent formation and trapping of the generated methyl radicals. Gilvocarcin V also photoreduces oxygen and methyl viologen with quantum yields of 0.019 and 0.0012 respectively. The quantum yield for singlet oxygen formation by GV in DMSO, determined by measuring the rate of production of the nitroxyl radical produced by the reaction of 2,2,6,6-tetramethylpiperidinol with singlet oxygen, was found to be 0.15. Thus, GV photochemistry proceeds by both Type I and Type II pathways which could contribute to the reported GV phototoxicity in biological systems.

Induction of lambda prophage near the site of focused UV laser radiation.

DNA damage from photon scatter or beam spread during UV excimer laser irradiation was investigated using the induction of bacteriophage lambda in E. coli BR339. Prophage induction in these cells leads to the production of beta-galactosidase which can be detected colorimetrically by the application of appropriate substrates. An agar surface overlayed with BR339 cells was placed at various distances from the focal point of a converging lens and exposed to either 193 or 248 nm laser radiation. Energy densities ranging from approximately 5 mJ/cm2 to 30 J/cm2 were used. Ablation with 193 nm laser radiation produced an 800 microns wide clear 'trench' surrounded by a 500 microns zone of cells in which lambda had been induced. Following ablation with 248 nm laser radiation, the zone of induction was several millimeters wide. Exposures to 193 nm radiation at 170 mJ/cm2/pulse produced visible ablation of the agar surface at 1.7 J/cm2. Lambda induction was observed surrounding cleared ablation areas. The presence of induction in this system suggests that both 248 and 193 nm excimer laser radiation delivered at high energy densities has sufficient spread or scatter to damage DNA in cells surrounding areas of ablation.

Photosensitized DNA breaks and DNA-to-protein crosslinks induced in human cells by antitumor agent gilvocarcin V.

The antitumor agent gilvocarcin V (GV) is photoactivated to a genotoxic form by low fluences of near-ultraviolet radiation. Activation of GV by monochromatic 450-nm radiation causes two specific DNA changes in human P3 cells in culture as shown by alkaline elution techniques: single-strand breaks (i.e., alkali-labile sites plus frank strand scissions) and DNA-to-protein covalent bond crosslinks. When GV is present with the cells during irradiation, the yields of these damages are increased. Fluence and concentration studies show that the induction of both DNA lesions occurs at unusually low concentrations of drug and fluences of radiation. Both breaks and crosslinks are readily detectable after exposure to less than 100 kJ m-2 of 405 nm-radiation at a GV concentration of 7.5 X 10(-9) M. These results indicate a possible potential for use of GV in human tumor photochemotherapy.

Relative mutagenic and prophage-inducing effects of mono- and di-alkyl nitrosoureas.

Mutagenic capacity, prophage-inducing ability, and decomposition rates of mono- and di-alkylnitrosoureas with the same alkyl groups were studied in aqueous systems using S. typhimurium strain TA1535 and E. coli strain BR339 (lambda). Slower decomposition rates of nitrosodialkylureas compared with monoalkylnitrosoureas were not reflected in reduced mutagenicity, with the exception of the methylating agents. With the monoalkylnitrosoureas, mutagenicity varied with length of the alkyl chain, and was greatly increased by oxygen substituents. Among the dialkylnitrosoureas, mutagenesis and decomposition rates were dependent on the substitutent in the N-1 position, adjacent to the nitroso group. Prophage induction, on the other hand, was dependent on a substituent in the N-3 position. The results suggested the existence of two distinct mechanisms for DNA damage, one SOS-dependent and the other SOS-independent, by which different dialkylnitrosoureas acted. Neither in vitro assay system was a reasonable model for the carcinogenic activity of these compounds.

Micro-BIA, a colorimetric microtiter assay of lambda prophage induction.

Escherichia coli that are lysogenic for a lambda-lacZ fusion phage produce beta-galactosidase, product of the lacZ gene, upon induction of the prophage by DNA-damaging agents. The miniaturization of a quantitative, colorimetric beta-galactosidase (prophage) induction assay (BIA) is presented. Induction assays are performed in microtiter wells with the aid of multichannel pipetting devices. Results are shown with screening strain BR513 (uvrB delta envA) and a strain, BR339 (uvrB delta lexA3ind-) which exhibits enhanced induction. A method developed for strain BR339 utilizes bacteria stored frozen in log phase, permeabilized in vitro, and used immediately; with this method, 2 consecutive assays may be completed in 1 working day. Mutagens utilized for the model studies included 4NQO, ENNG, daunorubicin, bleomycin, acetoxy-AAF, B[a]P, DMBA, and DEN (the last three in the presence of liver S9). Induced levels of beta-galactosidase were monitored using a vertical light path photometer that measured color absorbance in each microtiter well. Alternatively, color intensity could be determined by using a color chart prepared for this assay. These values were then plotted to generate dose-response curves. Considerable savings in labor and materials are achieved with the method described, one which may be used as a screen for DNA-damaging chemicals. Automated equipment and computers may be used to advantage with this assay, but they are not required.

Lack of genetic and in vitro metabolic activity of potently carcinogenic azoxyalkanes.

4 carcinogenic azoxyalkanes (azoxymethane, azoxymethane and the 2 mixed methyl-ethyl compounds) were examined for activity in the Salmonella histidine reversion assay and in a lambda-lacZ prophage induction assay. Because azoxyalkanes are isomeric with nitrosodialkylamines, and might be expected to generate the same active intermediates, their biological activity was investigated under conditions which would allow direct comparison with these well-studied carcinogens. However, none of the azoxyalkanes, which are liver carcinogens, showed significant activity in either microbial assay in the presence of liver S9. In addition, metabolism studies with liver microsomes or hepatocytes indicated that the compounds were metabolized only to a small extent, if at all, under the conditions examined. This inactivity of the azoxyalkanes contrasts with the considerable activity in these assays - and the substantial metabolism - of the isomeric nitrosodialkylamines, also liver carcinogens. These results suggest that the carcinogenic action of azoxyalkanes proceeds through alternative metabolic pathways that are not adequately modeled by the assays and in vitro conditions used here.

Activation of antitumor agent gilvocarcins by visible light.

Gilvocarcins that are antitumor agents are activated by low doses of visible light to induce bacteriophage lambda in Escherichia coli. This result is dependent on interaction with DNA. Gilvocarcin M, an analog without antitumor activity, failed to induce the prophage after light exposure, thus demonstrating a correlation between photosensitizing and antitumor activities. These results raise several possibilities regarding the mode of action of gilvocarcins as antitumor agents in vivo, involving light or enzymatic activating systems, which could be exploited in human cancer therapy.

Genetic and biochemical factors affecting the induction of bacteriophage lambda by N-nitroso compounds.

As part of our effort to validate a biochemical (prophage) induction assay (BIA) as a screening test for carcinogens, we have tested more than 100 N-nitroso compounds. An enzyme, beta-galactosidase, is induced as an indirect consequence of DNA damage to the host, as part of the 'SOS' response. Besides the obvious practical importance of detecting this class of carcinogen, there is the question of the mechanism by which these compounds work. Mutagenesis by one compound, N-methyl-N'-nitro-N-nitrosoguanidine, is known to proceed by both SOS (recA)-dependent and SOS-independent pathways. Mispairing due to O6 alkylation of guanine is thought to be responsible for the SOS-independent pathway; however, there has been little consideration of recA-dependent functions, of which phage induction is one. Although nitrosamides could be detected as phage inducers in our assay, N-nitrosamines in the presence of rat liver 9 000 X g supernatant usually gave no response. We found that the use of several mutant strains, particularly a lexA mutant, in combination with hamster liver 9 000 X g supernatant, allowed us to detect most N-nitrosamines reasonably well, in either a spot test or a quantitative tube assay. Induction in a lexA strain was most unexpected, since this mutation usually diminishes the expression of SOS functions induced by ultra-violet light. Because the genetic and biochemical conditions that favour phage induction are different from those that favour mutagenesis, it seems likely that the lesions in DNA leading to the two biological end-points are different.

Platinum complexes with anticancer potential and their evaluation by a colorimetric lambda prophage induction assay.

A simple biochemical phage induction assay (BIA) showed significant activity with 90% of the antitumor platinum compounds tested and lack of activity for all Pd(II) compounds and Pt(II) cationic complexes, compounds that are expected to be inactive. Structure-activity relationships for a large number of chemicals can be studied simultaneously by this simple, rapid, inexpensive and quantitative biochemical assay. Fifty-three platinum complexes were tested, including a number of ethylenediamines synthesized for this work. The magnitude of inducing activity varied over a 25-fold range; differences among analogs reflected structural differences in a chemically consistent manner. Seven platinum complexes showed greater activity than that of cis-diamminedichloroplatinum(II) (cisplatin, cis-DDP), while other compounds appeared to be substantially less toxic. The assay was predictive for most compounds with very high or very low activity in vivo against L1210. For compounds with intermediate levels of activity, no correlation between inducing and antitumor activity was observed.