Benchmark Dose Analysis of DNA Damage Biomarker Responses Provides Compound Potency and Adverse Outcome Pathway Information for the Topoisomerase II Inhibitor Class of Compounds

Visualization strategies to aid interpretation of high-dimensional genotoxicity data

This article describes a range of high-dimensional data visualization strategies that we have explored for their ability to complement machine learning algorithm predictions derived from MultiFlow® assay results. For this exercise, we focused on seven biomarker responses resulting from the exposure of TK6 cells to each of 126 diverse chemicals over a range of concentrations. Obviously, challenges associated with visualizing seven biomarker responses were further complicated whenever there was a desire to represent the entire 126 chemical data set as opposed to results from a single chemical. Scatter plots, spider plots, parallel coordinate plots, hierarchical clustering, principal component analysis, toxicological prioritization index, multidimensional scaling, t-distributed stochastic neighbor embedding, and uniform manifold approximation and projection are each considered in turn. Our report provides a comparative analysis of these techniques. In an era where multiplexed assays and machine learning algorithms are becoming the norm, stakeholders should find some of these visualization strategies useful for efficiently and effectively interpreting their high-dimensional data.

Effect of cell treatment procedures on in vitro genotoxicity assessment

So far, the majority of in vitro toxicological experiments are conducted after an acute 24 h treatment that does not represent a realistic human chemical exposure. Recently, new in vitro approaches have been proposed to study the chemical toxicological effect over several days in order to be more predictive of a representative exposure scenario. In this study, we investigated the genotoxic potential of chemicals (direct or bioactived clastogen, aneugen and apoptotic inducer) with the γH2AX and pH3 biomarkers, in the human liver-derived HepaRP cell line. We used different treatment durations, with or without a three-day recovery stage (release period), before genotoxicity measurement. Data were analysed with the Benchmark Dose approach. We observed that the detection of clastogenic compounds (notably for DNA damaging agents) was more sensitive after three days of repeated treatment compared to one or three treatments over 24 h. In contrast, aneugenic chemicals were detected as genotoxic in a similar manner whether after a 24 h exposure or a three-day repeated treatment. Globally, the release period decreases the genotoxicity measurement substantially. For DNA damaging agents, after high concentration treatments, γH2AX induction was always observed after a three-day release period. In contrast, for DNA topoisomerase inhibitors, no effect could be observed after the release period. In conclusion, in the HepaRP cell line, there are some important differences between a one-day acute and a three-day repeated treatment protocol, indicating that different cell treatment procedures may differentiate chemical genotoxic mechanisms of action more efficiently.

Comparative analysis of micronucleus induction and DNA damage biomarkers in TK6 and A375 cells using flow cytometry

Previously, we introduced an alternative adherent A375 cell line for clastogenicity and aneugenicity testing using a high content imaging platform. To further characterize the performance of A375 cells, we investigated the sensitivity and specificity of A375 and TK6 cells by directly comparing micronucleus (MN) induction, cytotoxicity (relative cell counts, viability, and apoptosis), clastogenicity (γH2AX), and aneuploidy markers (pH 3, MPM-2, and polyploidy) using flow cytometric methods. We evaluated 14 compounds across different mechanisms (non-genotoxic apoptosis inducers, clastogens, and aneugens with either tubulin binding or aurora kinase inhibiting phenotypes) at 4-h and 24-h post treatment. Both aneugens and clastogens tested positive for micronucleus induction in both cell lines. Apoptosis continued to be a confounding factor for flow cytometry-based micronuclei assessment in TK6 cells as evidenced by positive responses by the three cytotoxicants. Conversely, A375 cells were not affected by apoptosis-related false positive signals and did not produce a positive response in the in vitro micronucleus assay. Benchmark dose response (BMD) analysis showed that the induction of micronuclei and biomarkers occurred at similar concentrations in both cell lines for clastogens and aneugens. By showing that A375 cells have similar sensitivity to TK6 cells but a greater specificity, these results provide additional support for A375 cells to be used as an alternative adherent cell line for in vitro genetic toxicology assessment.

Impact of experimental design factors on the potency of genotoxicants in in vitro tests

Previous studies have shown that differences in experimental design factors may alter the potency of genotoxic compounds in in vitro genotoxicity tests. Most of these studies used traditional statistical methods based on the lowest observed genotoxic effect levels, whereas more appropriate methods, such as the benchmark dose (BMD) approach, are now available to compare genotoxic potencies under different test conditions. We therefore investigated the influence of two parameters, i.e. cell type and exposure duration, on the potencies of two known genotoxicants [aflatoxin B1 and ethyl methanesulfonate (EMS)] in the in vitro micronucleus (MN) assay and comet assay (CA). Both compounds were tested in the two assays using two cell types (i.e. CHO-K1 and TK6 cells). To evaluate the effect of exposure duration, the genotoxicity of EMS was assessed after 3 and 24 h of exposure. Results were analyzed using the BMD covariate approach, also referred to as BMD potency ranking, and the outcome was compared with that of more traditional statistical methods based on lowest observed genotoxic effect levels. When comparing the in vitro MN results obtained in both cell lines with the BMD covariate approach, a difference in potency was detected only when EMS exposures were conducted for 24 h, with TK6 cells being more sensitive. No difference was observed in the potency of both EMS and aflatoxin B1 in the in vitro CA using both cell lines. In contrast, EMS was more potent after 24 h exposure compared with a 3 h exposure under all tested conditions, i.e. in the in vitro MN assay and CA in both cell lines. Importantly, for several of the investigated factors, the BMD covariate method could not be used to confirm the differences in potencies detected with the traditional statistical methods, thus highlighting the need to evaluate the impact of experimental design factors with adequate approaches.

Estimating provisional margins of exposure for data-poor chemicals using high-throughput computational methods

Current computational technologies hold promise for prioritizing the testing of the thousands of chemicals in commerce. Here, a case study is presented demonstrating comparative risk-prioritization approaches based on the ratio of surrogate hazard and exposure data, called margins of exposure (MoEs). Exposures were estimated using a U.S. EPA’s ExpoCast predictive model (SEEM3) results and estimates of bioactivity were predicted using: 1) Oral equivalent doses (OEDs) derived from U.S. EPA’s ToxCast high-throughput screening program, together with in vitro to in vivo extrapolation and 2) thresholds of toxicological concern (TTCs) determined using a structure-based decision-tree using the Toxtree open source software. To ground-truth these computational approaches, we compared the MoEs based on predicted noncancer TTC and OED values to those derived using the traditional method of deriving points of departure from no-observed adverse effect levels (NOAELs) from in vivo oral exposures in rodents. TTC-based MoEs were lower than NOAEL-based MoEs for 520 out of 522 (99.6%) compounds in this smaller overlapping dataset, but were relatively well correlated with the same (r² = 0.59). TTC-based MoEs were also lower than OED-based MoEs for 590 (83.2%) of the 709 evaluated chemicals, indicating that TTCs may serve as a conservative surrogate in the absence of chemical-specific experimental data. The TTC-based MoE prioritization process was then applied to over 45,000 curated environmental chemical structures as a proof-of-concept for high-throughput prioritization using TTC-based MoEs. This study demonstrates the utility of exploiting existing computational methods at the pre-assessment phase of a tiered risk-based approach to quickly, and conservatively, prioritize thousands of untested chemicals for further study.

A new imaging platform (iScreen) allows for the concurrent assessment of micronucleus induction and genotoxic mode of action in human A375 cells

Genotoxicity testing guidelines require the assessment of the clastogenic and aneugenic potential of compounds. While in vitro micronucleus assays detect both types of endpoints, it requires labor-intensive microscopic scoring and does not discriminate between the two modes of actions. Here, we present a novel high-content imaging platform in A375 human cells that addresses the need for rapid scoring while providing additional mechanistic information. We evaluated the new platform with 12 compounds, three compounds from each mechanistic class (clastogen, aneugen tubulin binder, aneugen aurora inhibitor, and nongenotoxicant) following 4- and 24-h compound treatments. The approach we developed is first discriminating between genotoxicant and nongenotoxicant using an image analysis algorithm to quantify micronucleus induction below a 60% cytotoxicity cutoff. Then it uses centromere protein A (CENPA) staining for the genotoxic compounds to discriminate between aneugens and clastogens. Lastly, we use phosphorylated histone H2AX Ser139 (γH2AX) staining to confirm clastogenicity and changes in phosphorylated histone 3 Ser10 (pH 3) and increases in polyploidy in mitotic cells to discriminate between aneugens that bind tubulin from those that affect aurora kinases. All compounds were correctly classified, and we showed by using benchmark dose-response analysis that the imaging platform in A375 cells is at least as sensitive as the MicroFlow® assay in TK6 cells for genotoxicant but appears to be more specific for the nongenotoxicants. A detailed comparison of the cell lines and a more comprehensive validation with a much larger compound set, predictive and dose-response modeling will be presented in the future.

The use of benchmark dose uncertainty measurements for robust comparative potency analyses

The Benchmark Dose (BMD) method is the favored approach for quantitative dose-response analysis where uncertainty measurements are delineated between the upper (BMDU) and lower (BMDL) confidence bounds, or confidence intervals (CIs). Little has been published on the accurate interpretation of uncertainty measurements for potency comparative analyses between different test conditions. We highlight this by revisiting a previously published comparative in vitro genotoxicity dataset for human lymphoblastoid TK6 cells that were exposed to each of 10 clastogens in the presence and absence (+/-) of low concentration (0.25%) S9, and scored for p53, γH2AX and Relative Nuclei Count (RNC) responses at two timepoints (Tian et al., 2020). The researchers utilized BMD point estimates in potency comparative analysis between S9 treatment conditions. Here we highlight a shortcoming that the use of BMD point estimates can mischaracterize potency differences between systems. We reanalyzed the dose responses by BMD modeling using PROAST v69.1. We used the resulting BMDL and BMDU metrics to calculate "S9 potency ratio confidence intervals" that compare the relative potency of compounds +/- S9 as more statistically robust metrics for comparative potency measurements compared to BMD point estimate ratios. We performed unsupervised hierarchical clustering that identified four S9-dependent groupings: high and low-level potentiation, no effect, and diminution. This work demonstrates the importance of using BMD uncertainty measurements in potency comparative analyses between test conditions. Irrespective of the source of the data, we propose a stepwise approach when performing BMD modeling in comparative potency analyses between test conditions.

Experiments in the EpiDerm 3D Skin In Vitro Model and Minipigs In Vivo Indicate Comparatively Lower In Vivo Skin Sensitivity of Topically Applied Aneugenic Compounds

Risk management of in vitro aneugens for topically applied compounds is not clearly defined because there is no validated methodology to accurately measure compound concentration in proliferating stratum basale keratinocytes of the skin. Here, we experimentally tested several known aneugens in the EpiDerm reconstructed human skin in vitro micronucleus assay and compared the results to flow cytometric mechanistic biomarkers (phospho-H3; MPM2, DNA content). We then evaluated similar biomarkers (Ki-67, nuclear area) using immunohistochemistry in skin sections of minipigs following topical exposure the potent aneugens, colchicine, and hesperadin. Data from the EpiDerm model showed positive micronucleus responses for all aneugens tested following topical or direct media dosing with similar sensitivity when adjusted for applied dose. Quantitative benchmark dose-response analysis exhibited increases in the mitotic index biomarkers phospho-H3 and MPM2 for tubulin binders and polyploidy for aurora kinase inhibitors are at least as sensitive as the micronucleus endpoint. By comparison, the aneugens tested did not induce histopathological changes, increases in Ki-67 immunolabeling or nuclear area in skin sections from the in vivo minipig study at doses in significant excess of those eliciting a response in vitro. Results indicate the EpiDerm in vitro micronucleus assay is suitable for the hazard identification of aneugens. The lack of response in the minipig studies indicates that the barrier function of the minipig skin, which is comparable to human skin, protects from the effects of aneugens in vivo. These results provide a basis for conducting additional studies in the future to further refine this understanding.