Regulated cell death modalities such as apoptosis and necroptosis play an important role in regulating different cellular processes. Currently, regulated cell death is identified using the golden standard techniques such as fluorescence microscopy and flow cytometry. However, they require fluorescent labels, which are potentially phototoxic. Therefore, there is a need for the development of new label-free methods. In this work, we apply Digital Holographic Microscopy (DHM) coupled with a deep learning algorithm to distinguish between alive, apoptotic and necroptotic cells in murine cancer cells. This method is solely based on label-free quantitative phase images, where the phase delay of light by cells is quantified and is used to calculate their topography. We show that a combination of label-free DHM in a high-throughput set-up (~10,000 cells per condition) can discriminate between apoptosis, necroptosis and alive cells in the L929sAhFas cell line with a precision of over 85%. To the best of our knowledge, this is the first time deep learning in the form of convolutional neural networks is applied to distinguish—with a high accuracy—apoptosis and necroptosis and alive cancer cells from each other in a label-free manner. It is expected that the approach described here will have a profound impact on research in regulated cell death, biomedicine and the field of (cancer) cell biology in general.

Regulated cell death (RCD) has a fundamental role in development, pathology, and tissue homeostasis. In order to understand the RCD mechanisms, it is essential to follow these processes in real time. Here, atomic force microscopy (AFM) is applied to morphologically and mechanically characterize four RCD modalities (intrinsic and extrinsic apoptosis, necroptosis, and ferroptosis) in murine tumor cell lines. The nano-topographical analysis revealed a distinct surface morphology in case of necroptosis, ∼ 200 nm membrane disruptions are observed. Using mechanical measurements, it is possible to detect the early onset of RCD. Combined elasticity and microrheology analysis allowed for a clear distinction between apoptotic and regulated necrotic cell death. Finally, immunofluorescence analysis of the cytoskeleton structure during the RCD processes confirm the measured mechanical changes. The results of this study not only demonstrate the possibility of early real-time cell death detection but also reveal important differences in the cytoskeletal dynamics between multiple RCD modalities.

Human and animal health rely on balancing cell division and cell death to maintain normal homeostasis. This process is accomplished by regulated cell death (RCD), whose imbalance can lead to disease. Currently, the most frequently used method for analyzing RCD is fluorescence microscopy. This method has limitations and potential side effects due to the presence of fluorescent labels. Furthermore, fluorescence often lacks specificity and may have side effects. In the quest to overcome such difficulties, label-free approaches have come into focus.Here, Raman microscopy in combination with machine learning is used to investigate RCDs, where biochemical molecular “fingerprints” are investigated with a focus on the vibrations of atoms in molecules. Three different and unique RCD types with different genetic and biochemical machinery, namely, ferroptosis is studied in comparison with apoptosis, and necroptosis in the murine fibroblast line L929sAhFas. Interestingly, during ferroptosis, a decrease in the wavenumber at 939 cm^-1 was observed, which is associated with a potential reduction in the expression of collagen – a compound essential in multiple diseases. Data analysis was performed by machine learning (ML), here SVMs, where the model utilizing the spectra directly into a support vector machine (SVM) outperforms other SVM strategies correctly predicting 73% of all spectra. Other methods: PCA-SVM (principal component analysis-SVM), peak fitting-AUC-SVM (area under the curve) and peak fitting-spectral reconstruction-SVM rendered prediction accuracies of ~52%, ~43%, and 61%, respectively. Peak fitting has the additional benefit of enabling the biological interpretation of Raman scattering peaks by using the area under the curve, although at a loss of general accuracy. The potential of Raman microscopy in biology, in combination with machine learning pipelines, can be applied to a broader field of cell biology, not limited to regulated cell death.

In recent years, the importance of the investigation of regulated cell death (RCD) has significantly increased and different methods are proposed for the detection of RCD including biochemical as well as fluorescence assays. Researchers have shown that early stages of cell death could be detected by using AFM. Although AFM offers a high single‐cell resolution and sensitivity, the throughput (<100 cells/h) limits a broad range of biomedical applications of this technique. Here, a microfluidics‐based mechanobiology technique, named shear flow deformability cytometry (sDC), is used to investigate and distinguish dying cells from viable cells purely based on their mechanical properties. Three different RCD modalities (i.e., apoptosis, necroptosis, and ferroptosis) are induced in L929sAhFas cells and analysed using sDC. Using machine learning on the extracted parameters, it was possible to predict the dead or viable state with 92% validation accuracy. A significant decrease in elasticity can be noticed for each of these RCD modalities by analysing the deformation of the dying cells. Analysis of morphological characteristics such as cell size and membrane irregularities also indicated significant differences in the RCD induced cells versus control cells. These results highlight the importance of mechanical properties during RCD and the significance of label‐free techniques, such as sDC, which can be used to detect regulated cell death and can be further linked with sorting of live and dead cells.