μclear imaging plates

Cells were seeded in tissue culture-treated 384-well μClear imaging plates (Greiner BioOne, Frickenhausen, Germany) and incubated under standard tissue culture conditions during 24 h at 37 °C. Then, cells were treated with the indicated compounds and after 6 or 24 h of incubation, cells were fixed with 4% formaldehyde solution containing 1 μM Hoechst 33342 overnight at 4 °C. The fixative was changed to PBS, and the plates were subjected to automated image analysis. For automated fluorescence microscopy, a robot-assisted Molecular Devices IXM XL BioImager (Molecular Devices, Sunnyvale, CA, USA) equipped with Sola light sources (Lumencor, Beaverton, OR, USA), adequate excitation and emission filters (Semrock, Rochester, NY, USA), and a 16-bit monochromes sCMOS PCO.edge 5.5 camera (PCO, Kelheim, Germany) and a ×20 PlanAPO objective (Nikon, Tokyo, Japan) was used to acquire nine view fields/well, followed by image processing with the custom module editor of the MetaXpress software (Molecular Devices). For the latter, the images were segmented and analyzed for GFP, RFP granularity or global fluorescence intensity (depending on the dyes) by comparing the standard deviation of the mean fluorescence intensity of groups of adjacent pixels within the cytoplasm of each cell to the mean fluorescence intensity in the same ROI using the MetaXpress software (Molecular Devices).

One day before treatment, 2,500 U2OS cells stably co‐expressing either CALR‐GFP or HMGB1‐GFP with H2B‐RFP per well were seeded in 384‐well μClear imaging plates (Greiner Bio‐One) and let adhere. The next day, cells were treated and CALR‐GFP and HMGB1‐GFP cells were observed by live‐cell microscopy as described before with a frequency of image acquisition at one image per hour for 12 and 24 h, respectively. The images were segmented and analyzed with R using the EBImage and flowcatchR packages from the Bioconductor repository (https://www.bioconductor.org). H2B‐RFP was used to segment nuclei, and the obtained mask was either used to measure GFP intensity in the nuclear compartment (HMGB1‐GFP) or as a seed to segment the cytoplasmic compartment (CALR‐GFP). Then, a top‐hat filter was applied and the area of CALR‐GFP^high regions was measured. HMGB1‐GFP nuclear fluorescence intensity of single cells tracked over time was normalized to its value at first time point.

U2OS wild‐type cells were seeded at 8,000 cells per well in 96‐well μClear imaging plates (Greiner Bio‐One) and let adhere for 24 h before treatment. Cells were treated with a large range of concentrations for 24 h and then stained by the addition of propidium iodide (P4864, Sigma‐Aldrich) at a final concentration of 1 μg/ml and Hoechst 33342 at 2 μg/ml for 30 min. Plates were centrifuged in order to bring detached cells to the focal plane and then, images were acquired by automated microscopy using adequate filter sets as described above. The images were segmented with R by means of the EBImage package. Nuclei were segmented based on Hoechst 33342 signal; then, nuclear area and fluorescence intensities (in DAPI and Cy3) were measured. The assessed parameters were used to cluster cells as healthy (normal‐sized, Hoechst^low, PI⁻), pyknotic (condensed, Hoechst^high, PI⁻), or dead (PI⁺). The number of healthy cells was then used to establish dose–response models, by fitting the data points with a 4‐parameter log‐logistic model. The model was then used to calculate the IC₆₀ (concentration for which 40% of cell population is healthy) for each drug.