Analysing the cellular stress response, and in particular DNA repair mechanisms, now requires approaches capable of quantifying multiple parameters at the single-cell level. Quantitative Image-Based Cytometry (QIBC) addresses this need: applied by fluorescence microscopy to large populations of asynchronous adherent cells, it makes it possible to determine the cell-cycle stage of each cell while simultaneously measuring the signals of proteins of interest. This technique nevertheless remains largely dependent on high-content microscopy platforms and proprietary analysis software. The main obstacles lie in automating acquisition and analysis, which requires reliable segmentation of each nucleus coupled with unbiased extraction of multiple parameters.
The protocol described here offers a complete, open-access solution, spanning cell preparation, treatment and fixation through to staining, imaging and image analysis. It relies on a wide-field fluorescence microscope configuration for automated multicolour acquisition and on scripts developed for the Fiji software. The key analysis step is based on StarDist, a deep-learning tool trained to recognise nuclei. Unlike the conventional thresholding method, which uses only fluorescence intensity, StarDist exploits multiple morphological features, ensuring accurate and fully automated segmentation even in the presence of densely clustered nuclei, with minimal user intervention. The total DAPI intensity per nucleus serves as an estimate of DNA content, and cross-referencing with the mean EdU signal allows cells to be assigned to the G1, S and G2 phases. Additional parameters, such as the mean intensity or the number of foci per nucleus for the proteins studied, are then extracted and visualised in a multiparametric manner.
The validity of the approach is illustrated using the γH2AX and 53BP1 signals in the context of the DNA damage response. The authors find approximately twice as many γH2AX foci in irradiated cells in G2 as in G1, consistent with the doubling of DNA content; they observe specific γH2AX activation in S phase after replication stress induced by camptothecin; and they show that the number of 53BP1 foci normalised to DNA content is higher in G1 and then declines over the course of the cycle. The authors note certain limitations: the default StarDist model is suited to cell lines with ovoid-to-rounded nuclei, such as U-2 OS or RPE1-hTERT, and has not been tested on other cell types, while the identification of discrete foci remains challenging in samples with high density or heterogeneous intensity. Together, this constitutes a rapid, automated QIBC analysis procedure, accompanied by publicly available Fiji scripts.