Abstract 2829: Multifaceted characterization of ferroptosis using live-cell imaging and AI-driven analysis

Ferroptosis is a form of regulated cell death which is distinct from apoptosis and therefore represents a new mechanism by which cancer cells can be targeted and killed. Live-cell imaging was used to capture the dynamic and non-discrete morphology changes associated with ferroptosis in a range of cell types, and to characterize this mechanism of cell death and its contrast with apoptosis. Incucyte® AI Cell Health Analysis uses a trained neural network (NN) to segment cells in phase contrast images. The NNs were trained on a wide range of cell types, including healthy and non-viable, and therefore adapts robustly to unseen cells, yielding highly accurate delineation across a variety of morphologies as well as live/dead classification of ferroptotic cells. Using Incucyte® Advanced Label-free Classification software to perform machine learning-based classification of cells based on morphology, the characteristic ferroptotic transition state was selectively identified using images of cells treated with FIN56. This method was used for label-free quantification of compound efficacy in HeLa cells, comparing apoptosis inducer CMP to ferroptosis inducers Erastin (ERA), FIN56 and RSL3. Pharmacological validation of the mechanism of action showed ferroptosis inhibitor Ferrostatin-1 (FERR) inhibited ERA-induced cell death but not CMP-induced cell death; apoptosis inhibitor Z-VAD-FMK displayed converse selectivity. In addition, we demonstrated that ferroptosis inducer ERA, but not apoptosis inducer staurosporine (STP), was sensitive to the density of cells in the assay. ERA EC50 in SK-OV-3 cells decreased from 2.2 x 10-6 to 8.2 x 10-6 as the cell density increased from 500 to 3,000 cells per well while STP EC50 remained constant at 1.3 x10-7 at all cell densities. Fluorescence labeling of live and dead cells was explored to evaluate ferroptosis in 2D monolayer as well as 3D spheroid culture. Fibroblasts were treated with FIN56 in the presence of Incucyte® Cytotox Dye which enters non-viable cells via permeable cell membranes. Rapid ferroptosis was observed with the area of green fluorescence experiencing a rapid increase over 12h post treatment, and cell death was concentration-dependent. Ramos B cell line and adherent A549 cells were labeled with nuclear restricted orange and Near IR fluorophores, respectively. Upon treatment with ferroptosis inducers, both Ramos Nuclight Orange cells and A549 Nuclight NIR spheroids displayed a decrease in fluorescence intensity, consistent with cell death. Ferroptotic cell death in the spheroids was also characterized by an increase in brightfield area as the spheroids lost compactness, in contrast to apoptotic cell death which reduces the total spheroid area. Live-cell analysis provided a valuable platform for characterizing the dynamic process of ferroptotic cell death, yielding insight into the cellular mechanisms, and enabling quantification across multiple cell types. Citation Format: Jasmine E. Trigg, Gillian Lovell, Nicola Bevan. Multifaceted characterization of ferroptosis using live-cell imaging and AI-driven analysis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 2829.