DNA adducts are biomarkers of genotoxicity, and there is an ongoing discussion whether the dose response for DNA-adduct levels also predicts the dose response for mutation or even cancer. A linear low dose-effect relationship is plausible for the rate of DNA-adduct formation, but numerous factors modulate the dose response for successive endpoints. Furthermore, background levels of DNA damage must be taken into account for the assessment of exposure-related increments.
For the methylating agents, we found that the dose resulting in a doubling of the background levels of “biological” endpoints was much higher than the dose required to double DNA methylations. For instance, a 150-fold higher dose of methylnitrosourea (MNU) was required to double the background of micronuclei when compared to the formation of O6-mdGuo. For methyl methanesulfonate (MMS) we observed a threshold-like dose response at low dose for the induction of tk-/- mutants, even with an initial decrease below background. Fitting by a hockey stick threshold model was not significantly better than linearity, but the observed non-monotonic dose response indicated some “protective” cellular reaction at lowest dose. For oxidative stress the situation was different. Here the “biological” endpoints were more sensitive than measurements of the DNA adducts 8-oxodGuo and ÎµdAdo, and did not show deviation from linearity. In contrast, 8-oxodGuo showed a threshold-type increase after cumene hydroperoxide (CHP) treatment.
In summary, we observed non-monotonic and threshold-like shapes of dose-response relationships in addition to linearity, which indicates the complexity of the response of a cell to exposure to a chemically reactive agent. DNA adduct data may therefore indicate the potential of a chemical for genotoxicity and mutagenicity but are not sufficient to predict dose-response or potency for DNA damage in general or for successive endpoints. When expressed as increment over background, data on DNA adducts can over- or underestimate the genotoxic or mutagenic potency.
Particular caution must be exercised when continuous response variables such as biomarker data for genotoxicity are considered for the dose response of cancer incidence. Incidence measures are binary variables, where each individual contributes the count “one” (tumor) or ” zero” (no tumor). The dose-response curve separates two subgroups, the susceptible and the resistant (under the given conditions). The shape of the “curve” therefore is a mirror image of the distribution of individual susceptibilities for the rate of the process of carcinogenesis. Measures of early endpoints such as DNA damage or mutant frequencies disregard a number of factors that modulate individual susceptibility and cannot therefore represent the dose response for tumor incidence.
A Brink, B Schulz, H Stopper, WK Lutz, Biological significance of DNA adducts investigated by simultaneous analysis of different endpoints of genotoxicity in L5178Y mouse lymphoma cells treated with methyl methanesulfonate, Mutation Research – Fundamental and Molecular Mechanisms of Mutagenesis 2007, 625, 94-101.
A Brink, B Schulz, K Kobras, WK Lutz, H Stopper, Time-dependent effects of sodium arsenite on DNA breakage and apoptosis observed in the comet assay, Mutation Research/Genetic Toxicology and Environmental Mutagenesis 2006, 603, 121-128.
A Brink, U Lutz, W Voelkel, WK Lutz, Simultaneous determination of O6-methyl-2′-deoxyguanosine,
8-oxo-7,8-dihydro-2′-deoxyguanosine, and 1,N6-etheno-2′-deoxyadenosine in DNA using on-line sample preparation by HPLC column switching coupled to ESI-MS/MS, Journal of Chromatography B 2006, 830, 255-261.
WK Lutz, RW Lutz, ME Andersen, Dose-Incidence Relationships Derived from Superposition of Distributions of Individual Susceptibility on Mechanism-Based Dose Responses for Biological Effects, Toxicological Sciences 2006, 90, 33-38.