Adverse Outcome Pathway-Informed Integrated Testing to Identify Chemicals Causing Genotoxicity Through Oxidative DNA Damage: Case Study on 4-Nitroquinoline 1-Oxide

Abstract

Adverse outcome pathways (AOPs) provide a framework to organize and weigh evidence linking molecular interactions of toxicants in cells to adverse outcomes relevant to risk assessment or regulatory decision-making. Applying this framework facilitates the interpretation of data produced using new test methods. We used an existing AOP (AOP #296) that describes how oxidative DNA damage leads to mutations and chromosomal aberrations to develop an integrated testing strategy to evaluate whether a chemical operates through this pathway. We exposed human TK6 cells to increasing concentrations of 4-nitroquinoline 1-oxide (4NQO), a tobacco mimetic that causes oxidative DNA damage, in a time-series design. We measured oxidative DNA damage and strand breaks using the high-throughput CometChip assay with and without formamidopyrimidine DNA glycosylase (Fpg), alongside analyses of micronucleus (MN) frequency by flow cytometry, and mutations by error-corrected sequencing (duplex sequencing-DS). Our analysis shows how these methods can be combined to quantify 4NQO-induced, concentration- and time-dependent increases in: (a) oxidative DNA damage (occurred early and at low concentrations); (b) strand breaks (remained elevated to 6 h post-exposure); (c) MN frequency (at 24 h); (d) mutation frequency (at 48 h); and (e) C > A transversions consistent with expected substitutions induced by oxidative DNA lesions. The time series shows the repair of oxidative DNA damage with persistent strand breaks remaining at 6 h. Overall, we provide an example of an AOP-informed testing strategy and contribute to the quantitative understanding of AOP #296. We also demonstrate the value of DS as an effective approach for mutagenicity assessment.

Keywords: adverse outcome pathway; duplex sequencing; error‐corrected sequencing; genetic toxicology; in vitro toxicology; new approach methodologies.

FIGURE 1

Flow diagram of adverse outcome pathway 296 (AOP 296) “Oxidative DNA damage leading to mutations and chromosomal aberrations” (modified from Cho et al. 2022). The molecular initiating event is shown in green, the key events in blue, and the adverse outcomes in red. The accompanying table shows the methods that were used to measure each of the key events in this study.

FIGURE 2

Percent DNA observed in comet tails after 2, 4, or 6 h exposures to 4NQO with Fpg enzyme treatment (in red) and without Fpg enzyme treatment (in blue). A: Each bar represents the average tail DNA (%) of all technical replicates from three experiments (n = 3). The error bars represent the standard error of the mean of all replicates. Statistically significant (p < 0.05) increases from the solvent control at each time point is indicated by asterisks (*** for p < 0.001, ** for p < 0.01, and * for p < 0.05). B: Each point represents the average tail DNA (%) of all technical replicates from three experiments (n = 3). Hashmarks (#) indicate statistically significant differences in tail DNA (%) between the + and −Fpg conditions at the same concentration (### for p < 0.001, ## for p < 0.01, and # for p < 0.05).

FIGURE 3

A: Average percentage of micronuclei (orange bars) and relative survival (blue line) observed after a 24‐h exposure to 4NQO (n = 3). Error bars represent the standard errors of the mean. Statistically significant increases from the solvent control are indicated by asterisks (*** for p < 0.001 and * for p < 0.05). B: Mutation frequency per base pair in TK6 cells after exposure to 4NQO or vehicle control (0, DMSO) for 24 h. Cells were sampled 24 h after the end of the 24‐h exposure (i.e., 48 h after the start of the experiment). Each bar represents a replicate and asterisks indicate statistically significant increases from the control (p = 0.02 to 1.6e‐7).

FIGURE 4

Mutation frequency of each subtype of base substitution (main figure), insertions, deletions, and multinucleotide variants (MNV) (inset figure). Cells were sampled 24 h after the conclusion of the 24‐h exposure (48 h total). Each bar represents a replicate.

FIGURE 5

Trinucleotide spectra in TK6 cells after exposure to 4NQO or vehicle control for 24 h. Cells were sampled 24 h after a 24 h exposure (48 h total). Mutation frequencies were averaged for each exposure concentration (n = 3 for controls and n = 2 for exposed). The substitution subtype is listed at the top, with the two flanking nucleotides shown along the bottom. The Y axis indicates the proportion of each substitution type within the entire population of mutations recovered. Gray bars on the right indicate concentration in μg/mL.

FIGURE 6

Contribution of COSMIC signatures to the reconstruction of experimental trinucleotide mutation spectra in TK6 cells exposed to DMSO or 4NQO for 24 h, determined using the SigProfiler online application.