February’s Editor's Choice article is “A comparison of classical and 21st century genotoxicity tools: A proof of concept study of 18 chemicals comparing in vitro micronucleus, ToxTracker and genomics-based methods (TGx-DDI, whole genome clustering and connectivity mapping),” (https://onlinelibrary.wiley.com/doi/10.1002/em.22418) by Ashley Allemang, K. Nadira De Abrew, Yuqing K. Shan, Jesse M. Krailler, Stefan Pfuhler.
Hazard identification using data from 2 to 3 in vitro genotoxicity tests is one of the early steps in chemical risk assessment. Although these in vitro tests exhibit high sensitivity, they also tend to display low specificity. Generally, an in vivo rodent follow-up test is the preferred method to confirm in vitro positive results for a given test article. However, the use of in vivo follow-up tests are often limited due to animal welfare concerns as well as cost and time constraints; meaning further evaluation of useful chemicals might be abandoned. Recently developed genomic analyses might be used to complement in vitro genotoxicity data to provide better predictivity of in vivo genotoxic outcomes. The Allemang study evaluated how well three novel genomics/biomarker-based methods can distinguish different classes of genotoxic chemicals to predict in vivo genotoxic outcomes when used in conjunction with the in vitro micronucleus test. These genomic analyses, when applied to genotoxicity testing, have the potential to supplement or possibly replace conventional in vitro tests and, more importantly, reduce the need for in vivo genotoxicity testing.
In their February 2021 article, Allemang and colleagues evaluated the effects of 18 chemicals with different genotoxic modes-of-action (MOA) in the TK6 cell line using a flow-cytometry based micronucleus (MN) assay. The MN assay detects the effects of treatments that produce micronuclei containing chromosome fragments from DNA breakage (clastogenicity) or the loss of whole chromosomes (aneugenicity). Microarray gene expression data from these chemically treated TK6 cells was then evaluated using three different analysis methods. The MN and genomic based methods were also compared to the ToxTracker assay, which provides mechanistic insight into genotoxicity via the expression of GFP-tagged genes involved in DNA damage response (Bscl2, Rtkn), oxidative stress (Srxn1, Blvrb), protein damage (Ddit3) and p53 activation (Btg2) in mouse embryonic cell lines via flow cytometry
The genomic analysis component of the study consisted of data from TGx-DDI, connectivity mapping (cMap), and hierarchal clustering of all differentially expressed genes (DEGs). The TGx-DDI approach uses the expression profiles of 64 biomarker genes as a signature for distinguishing DNA damage-inducing (DDI) from non-DDI compounds. The cMap approach uses a gene expression signature for a compound and concentration combination that is obtained by searching against the CMap database containing gene expression profiles from multiple cell types treated with small-molecule compounds. Finally, the microarray data were used to perform a hierarchical clustering of all differentially expressed genes (DEGs) and dendrograms were created for all DEGs to group genes by maximum or complete linkage clustering.
The objective of the study by Allemang and colleagues was to compare findings from currently available genotoxic assessment methods to novel genomics methods following treatment with 18 chemicals with different genotoxic MOAs. The ideal method, or combination of methods, would accurately predict the expected in vivo genotoxicity outcome. Individually or in combination, the methods could correctly distinguish most, but not all, classes of genotoxicants. For instance, the TK6 MN assay correctly identified 14/18 chemicals while the ToxTracker assay correctly identified 13/18 compounds. However, the ToxTracker assay, TGx-DDI biomarker and DEG whole genome approaches demonstrated a higher predictive potential when used in combination with the MN assay, increasing the correct calls to 17/18, 16/18 and 17/18 respectively. While this work is still at the proof-of-concept stage and conducted with a limited battery of test chemicals, the results so far are encouraging. These findings suggest that integrating some combinations of these tests into early genotoxicity hazard identification steps could serve to protect public health while simultaneously reducing reliance on animal studies and preserving commercial interests.