High throughput screening (HTS) methodologies allow for simplified analysis of numerous conditions. The development of the arrayed cellular microenvironment technology has allowed for multi-factorial investigations into cellular processes through controlling the cellular environment. Here, we describe the development of a HTS cellular microarray technology and its application to (1) the identification and analysis of an optimal, defined matrix composition for human embryonic stem cells (hESCs), (2) the testing of synthetic polymers for their ability to support hESC growth, and (3) the development of a functional genomics screening platform. Through examination of gene expression of hESCs cultured on the defined matrix identified, we characterized maintenance of pluripotency and demonstrated differentiation potential. Additionally, polymers were systematically screened for ability to support hESC culture. Long-term culture methods on discovered polymers were examined through culture on hydrogels. Furthermore, we investigated a novel method of siRNA screening using this technology. The versatile arrayed cellular microenvironment technology can be utilized to understand the complex factors that modulate cellular activity and influence cell adherence, proliferation, and fate.

Cellular events underlying neurodegenerative disease may be captured by longitudinal live microscopy of neurons. While the advent of robot-assisted microscopy has helped scale such efforts to high-throughput regimes with the statistical power to detect transient events, time-intensive human annotation is required. We addressed this fundamental limitation with biomarker-optimized convolutional neural networks (BO-CNNs): interpretable computer vision models trained directly on biosensor activity. We demonstrate the ability of BO-CNNs to detect cell death, which is typically measured by trained annotators. BO-CNNs detected cell death with superhuman accuracy and speed by learning to identify subcellular morphology associated with cell vitality, despite receiving no explicit supervision to rely on these features. These models also revealed an intranuclear morphology signal that is difficult to spot by eye and had not previously been linked to cell death, but that reliably indicates death. BO-CNNs are broadly useful for analyzing live microscopy and essential for interpreting high-throughput experiments.

Cell death is a critical process that occurs normally in health and disease. However, its study is limited due to available technologies that only detect very late stages in the process or specific death mechanisms. Here, we report the development of a family of fluorescent biosensors called genetically encoded death indicators (GEDIs). GEDIs specifically detect an intracellular Ca²⁺ level that cells achieve early in the cell death process and that marks a stage at which cells are irreversibly committed to die. The time-resolved nature of a GEDI delineates a binary demarcation of cell life and death in real time, reformulating the definition of cell death. We demonstrate that GEDIs acutely and accurately report death of rodent and human neurons in vitro, and show that GEDIs enable an automated imaging platform for single cell detection of neuronal death in vivo in zebrafish larvae. With a quantitative pseudo-ratiometric signal, GEDIs facilitate high-throughput analysis of cell death in time-lapse imaging analysis, providing the necessary resolution and scale to identify early factors leading to cell death in studies of neurodegeneration.

Microscopy is a central method in life sciences. Many popular methods, such as antibody labeling, are used to add physical fluorescent labels to specific cellular constituents. However, these approaches have significant drawbacks, including inconsistency; limitations in the number of simultaneous labels because of spectral overlap; and necessary perturbations of the experiment, such as fixing the cells, to generate the measurement. Here, we show that a computational machine-learning approach, which we call "in silico labeling" (ISL), reliably predicts some fluorescent labels from transmitted-light images of unlabeled fixed or live biological samples. ISL predicts a range of labels, such as those for nuclei, cell type (e.g., neural), and cell state (e.g., cell death). Because prediction happens in silico, the method is consistent, is not limited by spectral overlap, and does not disturb the experiment. ISL generates biological measurements that would otherwise be problematic or impossible to acquire.