Genotoxins induce binucleation in L5178Y and TK6 cells

The in vitro micronucleus assay using imaging flow cytometry and deep learning

The in vitro micronucleus (MN) assay is a well-established assay for quantification of DNA damage, and is required by regulatory bodies worldwide to screen chemicals for genetic toxicity. The MN assay is performed in two variations: scoring MN in cytokinesis-blocked binucleated cells or directly in unblocked mononucleated cells. Several methods have been developed to score the MN assay, including manual and automated microscopy, and conventional flow cytometry, each with advantages and limitations. Previously, we applied imaging flow cytometry (IFC) using the ImageStream® to develop a rapid and automated MN assay based on high throughput image capture and feature-based image analysis in the IDEAS® software. However, the analysis strategy required rigorous optimization across chemicals and cell lines. To overcome the complexity and rigidity of feature-based image analysis, in this study we used the Amnis® AI software to develop a deep-learning method based on convolutional neural networks to score IFC data in both the cytokinesis-blocked and unblocked versions of the MN assay. We show that the use of the Amnis AI software to score imagery acquired using the ImageStream® compares well to manual microscopy and outperforms IDEAS® feature-based analysis, facilitating full automation of the MN assay.

Evaluation of the automated MicroFlow® and Metafer™ platforms for high-throughput micronucleus scoring and dose response analysis in human lymphoblastoid TK6 cells

The use of manual microscopy for the scoring of chromosome damage in the in vitro micronucleus assay is often associated with user subjectivity. This level of subjectivity can be reduced by using automated platforms, which have added value of faster with high-throughput and multi-endpoint capabilities. However, there is a need to assess the reproducibility and sensitivity of these automated platforms compared with the gold standard of the manual scoring. The automated flow cytometry-based MicroFlow^® and image analysis-based Metafer™ were used for dose response analyses in human lymphoblastoid TK6 cells exposed to the model clastogen, methyl methanesulfonate (MMS), aneugen, carbendazim, and the weak genotoxic carcinogen, ochratoxin A (OTA). Cells were treated for 4 or 30 h, with a 26- or 0-h recovery. Flow cytometry scoring parameters and the Metafer™ image classifier were investigated, to assess any potential differences in the micronucleus (MN) dose responses. Dose response data were assessed using the benchmark dose approach with chemical and scoring system set as covariate to assess reproducibility between endpoints. A clear increase in MN frequency was observed using the MicroFlow^® approach on TK6 cells treated for 30 h with MMS, carbendazim and OTA. The MicroFlow^®-based MN frequencies were comparable to those derived by using the Metafer™ and manual scoring platforms. However, there was a potential overscoring of MN with the MicroFlow^® due to the cell lysis step and an underscoring with the Metafer™ system based on current image classifier settings. The findings clearly demonstrate that the MicroFlow^® and Metafer™ MN scoring platforms are powerful tools for automated high-throughput MN scoring and dose response analysis.