DNA adducts: effects of low exposure to ethylene oxide, vinyl chloride and butadiene

Abstract

Dose–response relationships of genotoxic agents differ greatly depending on the agent and the endpoint being evaluated. Simple conclusions that genotoxic effects are linear cannot be applied universally. The shape of the molecular dose of DNA adducts varies from linear, to supralinear, to sublinear depending on metabolic activation and detoxication, and repair of individual types of DNA adducts. For mutagenesis and other genotoxicity endpoints, the dose–response reflects the molecular dose of each type of DNA adduct, cell proliferation, as well as endogenous factors that lead to mutagenesis such as the formation and repair of endogenous DNA adducts. These same factors are important when interpreting the shape of dose–response data for carcinogenesis of genotoxic agents, however, tumor background variability adds additional complexity. Endogenously formed DNA adducts may be identical to those formed by chemicals, as in the case of vinyl chloride and ethylene oxide, or they may be those associated with oxidative stress. Data presented in this paper demonstrate that the exogenous number of adducts induced by 5 days of exposure to 10 ppm vinyl chloride is only 2.2-fold greater than that present as a steady-state amount in unexposed control rats. Similar data are shown for ethylene oxide. Extremely sensitive methods have been developed for measuring the molecular dose of genotoxins. These methods can detect DNA adducts as low as 1 per 10⁹ to 10¹⁰. However, in view of the high number of endogenous DNA adducts that are present in all cells, it is unlikely that causal relationships can be attributed to very low numbers of such DNA adducts. Effects of both exogenous and endogenous DNA adducts need to be factored into the interpretation of chemical exposures.

Introduction

There are often large differences between the amount or dose of a chemical that humans are exposed to and the dose employed in toxicity and carcinogenicity studies used for risk assessment. This greatly exaggerated exposure of test animals is necessary due to the insensitivity of bioassays involving 50 animals per dose group to predict risks for much larger populations of humans. However, the magnitude of this difference often results in large uncertainties associated with such risk assessments. If a better understanding of critical mechanisms involved in the induction of mutations and cancer can be gained, it can be used to improve the accuracy of human risk assessment [1]. This paper will demonstrate the utility of such knowledge regarding the dose–response of DNA adducts of three well established genotoxic carcinogens, ethylene oxide (EO), vinyl chloride (VC), and butadiene (BD).

Genotoxic chemical carcinogens can either be direct acting (ultimate carcinogens) or require metabolic activation (procarcinogens) [2]. As shown in Fig. 1, the chemical must be absorbed and distributed within the body, where it undergoes metabolic activation if required. This metabolic activation is often tissue and cell-type specific. The ultimate carcinogen is also subjected to detoxication. It is the balance between the exposure to or production of the ultimate carcinogen and the extent of detoxication that determines the amount of ultimate carcinogen available to bind to DNA or other biomarkers of exposure, such as proteins. Since both metabolic activation and detoxication are enzymatic processes, it is possible to have saturation, induction, or depletion, such that a nonlinear dose–response is created between high and low exposure. There is additional potential for nonlinearities due to similar changes in DNA repair. Finally, nonlinear responses in tumor induction can be brought about by changes in cell proliferation, such as are often associated with cytotoxic or mitogenic events.

In this case study, EO is a direct alkylating agent that does not require metabolic activation, VC requires metabolic activation to form chloroethylene oxide, and BD can undergo multiple steps of metabolic activation and detoxication.

There are a limited number of chemicals that have had the molecular dosimetry of DNA adduct formation studied in detail. The alkylating agents were recently reviewed by La and Swenberg [3]. Much of the data are based on single-day dosing, so that alterations in gene expression of proteins involved in biotransformation or DNA repair may not be identified. The development of sensitive analytical methods that do not require administration of radioactive carcinogens has greatly enhanced the ability to investigate molecular dosimetry in repeated dose and even chronic exposure studies. The data sets discussed below involve up to 4 weeks of dosing, so that changes in gene expression and substrate depletion have time to occur. The exposures cover a minimum of one and one-half orders of magnitude and at low exposures that are within approximately one order of magnitude of current occupational limits. Thus, the data presented become highly relevant for risk assessment in an occupational setting, but are still several orders of magnitude away from most environmental exposures.

Recent improvements in methodology for DNA and protein adducts have allowed the clear demonstration that endogenous formation of DNA adducts is a common phenomenon in unexposed animals and humans 1, 4, 5. This is of interest when considering the low dose effects of genotoxic chemicals for several reasons. Mathematical approaches to low dose estimations of risk have long suggested that if identical or very similar processes exist as background risk, that any additional exposure will be additive and will be expected to be linear at low dose [6]. On the other hand, if a chemical induces extremely low numbers of exogenous DNA adducts under the conditions that cancer is induced, the chance that these adducts play a key role in carcinogenesis seems very unlikely, since much higher numbers of endogenous DNA adducts will dominate the induction of mutations. In other cases, a chemical may cause very small numbers of direct DNA adducts at high doses, but induce orders of magnitude more indirect DNA adducts, such as those induced by oxidative stress. Under such conditions, it is much more likely that the indirect DNA adducts are causally related to tumor induction. Since oxidative stress is usually greatly exacerbated by the depletion of host cellular defenses frequently associated with toxic exposures, this mechanism may have little or no relevance to low exposures where no such depletion exists [7]. Finally, some of the endogenous DNA adducts formed are identical to those formed by genotoxic chemicals. In fact, two, if not all three of the chemicals discussed below fall into this category. N-7-(2-Hydroxyethyl)guanine (HEG) is the major adduct formed by EO [8]. It is also formed endogenously from ethylene generated by gut microflora, lipid peroxidation and metabolism of methionine and hemin. The etheno adducts formed by VC are also caused by lipid peroxidation 9, 10. While clear evidence of endogenous DNA adducts has not been shown for BD, 1,2,3-trihydroxybutane N-terminal valine adducts have been demonstrated in several species, including humans with no exposure to BD [11]. In the cases of EO and VC, the number of endogenous DNA adducts has been measured and data will be presented comparing the number of adducts resulting from the lowest exogenous exposure with the number of identical endogenous DNA adducts present at steady state in unexposed animals. These data change the paradigm for risk assessment from one of extrapolation to one of interpolation.