RE: Recommendations, evaluation and validation of a semi-automated, fluorescent-based scoring protocol for micronucleus testing in human cells (Mutagenesis, 29, 155–164, 2014)

We are writing in regard to the article by Seager et al. (1) that assesses and validates the Metafer slide scanning system, developed by MetaSystems (Altlussheim, Germany), for the automated scoring of micronuclei (MN) in the in vitro cytokinesis-block micronucleus assay. Specifically, we would like to highlight a further use of the Metafer system that is not widely known and may facilitate the efficient and objective measurement of cytotoxicity in cells exposed to genotoxicants. As part of Test Guideline 487 ‘The In Vitro Mammalian Cell Micronucleus Test’, the Organisation for the Economical Cooperation and Development (OECD) specifies that concurrent measures of cytotoxicity and/or cytostasis should be quantified when the cytokinesis-block method is used (2). Annex 2 of the guideline outlines calculations for cytokinesis-block proliferation index (CBPI) and the replicative index that rely on the enumeration of mononucleated, binucleated and multinucleated cells in the treated and control cultures. In their study, Seager et al. seeded satellite cultures of their cells in order to calculate the relative population doubling and relative increase in cell count as indicators of cytotoxicity/cytostasis. We would like to suggest that it would have also been possible to score the CBPI concurrently and automatically using the Metafer slide scanning system on the same slides used for scoring MN. In 2011, MetaSystems released a white paper outlining the option of using their software to count the nuclei in cells for the purposes of calculating the CBPI (3). However, the white paper contains only cursory information, and the classifier provided by MetaSystems requires modifications based on the cells of interest. With guidance from MetaSystems, we have optimised a classifier for use with a murine epithelial lung cell line, the details of which are shared below so that other investigators can also employ the platform to score both MN frequency and CBPI. Prepared slides were stained with 500nM of propidium iodide (PI) for 5min, then stained with 200ng/ml of 4′-6-diamidino-2-phenylindole (DAPI) solution for 5min and cover slips affixed with VectaShield® mounting medium (either with or without DAPI). It should be noted that this dual staining protocol addresses one of the troublesome issues noted by Seager et al. That is, when manually verifying scored slides, it can be difficult to identify whether a MN belongs to one cell or another. PI effectively defines cell boundaries by staining the cytoplasm and thus facilitates the assignment of MNs to particular cells (Figure 1). A grey scale image of binucleate cells with cytoplasm stained by PI. The staining not only allows for the number of nuclei in a cell to be enumerated but also facilitates the assignment of MN to the correct cell. Slides were scanned on the Metafer4 platform (v3.8.6) with a Carl Zeiss Axio Imager M1 microscope, equipped with a CoolCube 1 charge-coupled device camera and Märzhäuser motorized stage that scans eight slides unattended (MetaSystems). Slides were first scored with the MNScore function to obtain the number of MN in binucleated cells and subsequently with the MetaCyte function to detect the number of nuclei in each cell for calculation of CBPI. When optimising the MetaCyte classifier, we found that first modifying the Integration Time, and specifically the Minimum Integration Time, resulted in the most improvements in cell image capture. Pressing ‘c’ on the keyboard while in the gallery allowed the captured cell contours to be verified. In addition to Integration Time, the Saturation Area (which is typically higher for the red channel), the Object Threshold (estimated using the Classify Fields Function), the Camera Gain Factor (inversely proportional to integration time and dependent on camera type) and Cell Size (estimated using training data and the Classify Fields Function) were also important parameters that influenced cell imaging. The values used for these parameters are listed below. System settings: Grid Focus: FL10, Fine Focus: FL10 Capture tab: Use Automatic Objective Change, CCD Camera Gain Factor: 8% Exposure tab: Minimum Integration Time CS: 0.0400 S1: 0.0044 Cell Selection tab (Single Cells sub-tab): Minimum Object Area: 550 Maximum Object Area: 5000 Maximum Concavity Depth: 0.150 Maximum Aspect Ratio: 1.430 CS Object Threshold: 5% Although these parameters have been optimised for use with a murine epithelial lung cell line, it is expected that similar classifiers could be established for other cell lines through modification of these same parameters. The exact nature of the adjustments, however, would depend on the optical and geometric properties of the stained cells and sub-cellular structures. Although there are an extensive number of parameters that can be modified, including beyond what is listed here, MetaSystems technical staff are knowledgeable about their function and efficacy with respect to scoring metrics and are a valuable resource. When scoring cells, the MetaCyte software first acquires images based on the PI channel to determine the cell boundaries and then acquires images based on the DAPI channel to determine the number of nuclei within each cell. Because the system requires the acquisition of images under two colour channels, and it conducts several relatively complex calculations, the scanning process can take a substantial amount of time. For example, whereas the system can score MN in 1000 cells in 2–5min, counting 500 cells for the CBPI can take upwards of 20–40min per slide, depending on slide quality. Having an automated slide feeder can mitigate this inconvenience as slides can easily be set up to run during off hours (e.g. overnight). MetaSystems suggests that the MNScore and MetaCyte scans be linked in the classifier so that the same slide region is scanned. However, OECD protocols recommend that 1000 binucleated cells per culture be scored for detecting MN, whereas only 500 cells need to be scored for the CBPI. With the linked classifier, 1000 cells must be scanned for both MN and CBPI detection, unnecessarily increasing scan time. Following the MetaCyte scan, images need to be manually verified in the gallery for accuracy, a task that is quickly accomplished. However, unlike the MNScore classifier for MN scoring, cells that are incorrectly categorised by the MetaCyte classifier cannot be reclassified (i.e. they can only be deleted). Therefore, the user must ensure that a sufficient number of cells are scored (e.g. 550) to achieve successful scoring of at least 500 cells after deletion of cells that were incorrectly categorised. Once set up, the Metasystems’ slide scanning and image analysis system permits the enumeration of mononucleated, binucleated and multinucleated cells and subsequent calculation of indices of cytotoxicity/cytostasis. The semi-automated slide scanning has the benefit of scoring numerous cells with limited subjectivity and little user involvement. The CBPI classifier is a natural complement to Seager et al.’s scoring protocol for the assessment of MN, and together, the two provide a useful indicator of clastogenicity and cytotoxicity following exposure to test agents. Conflict of interest statement: None declared.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.