This is the second of two reports from the International Workshops on Genotoxicity Testing (IWGT) Working Group on Quantitative Approaches to Genetic Toxicology Risk Assessment (the QWG). The first report summarized the discussions and recommendations of the QWG related to the need for quantitative dose-response analysis of genetic toxicology data, the existence and appropriate evaluation of threshold responses, and methods to analyze exposure-response relationships and derive points of departure (PoDs) from which acceptable exposure levels could be determined. This report summarizes the QWG discussions and recommendations regarding appropriate approaches to evaluate exposure-related risks of genotoxic damage, including extrapolation below identified PoDs and across test systems and species. Recommendations include the selection of appropriate genetic endpoints and target tissues, uncertainty factors and extrapolation methods to be considered, the importance and use of information on mode of action, toxicokinetics, metabolism, and exposure biomarkers when using quantitative exposure-response data to determine acceptable exposure levels in human populations or to assess the risk associated with known or anticipated exposures. The empirical relationship between genetic damage (mutation and chromosomal aberration) and cancer in animal models was also examined. It was concluded that there is a general correlation between cancer induction and mutagenic and/or clastogenic damage for agents thought to act via a genotoxic mechanism, but that the correlation is limited due to an inadequate number of cases in which mutation and cancer can be compared at a sufficient number of doses in the same target tissues of the same species and strain exposed under directly comparable routes and experimental protocols.

This report summarizes the discussion, conclusions, and points of consensus of the IWGT Working Group on Quantitative Approaches to Genetic Toxicology Risk Assessment (QWG) based on a meeting in Foz do Iguaçu, Brazil October 31-November 2, 2013. Topics addressed included (1) the need for quantitative dose-response analysis, (2) methods to analyze exposure-response relationships & derive point of departure (PoD) metrics, (3) points of departure (PoD) and mechanistic threshold considerations, (4) approaches to define exposure-related risks, (5) empirical relationships between genetic damage (mutation) and cancer, and (6) extrapolations across test systems and species. This report discusses the first three of these topics and a companion report discusses the latter three. The working group critically examined methods for determining point of departure metrics (PoDs) that could be used to estimate low-dose risk of genetic damage and from which extrapolation to acceptable exposure levels could be made using appropriate mode of action information and uncertainty factors. These included benchmark doses (BMDs) derived from fitting families of exponential models, the No Observed Genotoxic Effect Level (NOGEL), and "threshold" or breakpoint dose (BPD) levels derived from bilinear models when mechanistic data supported this approach. The QWG recognizes that scientific evidence suggests that thresholds below which genotoxic effects do not occur likely exist for both DNA-reactive and DNA-nonreactive substances, but notes that small increments of the spontaneous level cannot be unequivocally excluded either by experimental measurement or by mathematical modeling. Therefore, rather than debating the theoretical possibility of such low-dose effects, emphasis should be placed on determination of PoDs from which acceptable exposure levels can be determined by extrapolation using available mechanistic information and appropriate uncertainty factors. This approach places the focus on minimization of the genotoxic risk, which protects against the risk of the development of diseases resulting from the genetic damage. Based on analysis of the strengths and weaknesses of each method, the QWG concluded that the order of preference of PoD metrics is the statistical lower bound on the BMD > the NOGEL > a statistical lower bound on the BPD. A companion report discusses the use of these metrics in genotoxicity risk assessment, including scaling and uncertainty factors to be considered when extrapolating below the PoD and/or across test systems and to the human.

Aryl radicals and arenediazonium ions are suspected to react with cellular DNA, resulting in C8-arylguanine adducts. 8-Phenyl-2'-deoxyguanosine (8-PhdG) was synthesized as a model adduct by reacting dG with benzenediazonium chloride and incorporated into oligodeoxynucleotides using phosphoramidite techniques. A site-specifically modified oligodeoxynucleotide containing a single 8-PhdG was then used as a template for primer extension reactions catalyzed by the intact (exo+) or 3'-->5' exonuclease-free (exo-) Klenow fragment of Escherichia coli DNA polymerase I and mammalian DNA polymerase alpha (pol alpha). Although primer extensions catalyzed by the Klenow fragments were retarded at the position of 8-PhdG, most of the primer extension passed the lesion to form the fully extended products. In contrast, primer extensions catalyzed by pol alpha were strongly blocked opposite the lesion. The fully extended products formed during DNA synthesis were analyzed to quantify the miscoding specificities of 8-PhdG. The exo- Klenow fragment incorporated primarily dCMP, the correct base, opposite 8-PhdG, along with small amounts of incorporation of dAMP. Two-base deletions were also observed. In contrast, the exo+ Klenow fragment incorporated dCMP opposite the lesion. When pol alpha was used, 8-PhdG promoted small amounts of misincorporation of dAMP and dGMP as well as one- and two-base deletions. The duplex containing 8-PhdG.dG was thermally and thermodynamically more stable than dG.dG. The duplex containing 8-PhdG.dA was thermodynamically more stable than dG.dA. We conclude that 8-PhdG is a weak miscoding lesion, capable of generating G-->T and G-->C transversions and deletions in cells.