Abstract
Metabolic homeostasis requires engagement of catabolic and anabolic pathways consuming nutrients that generate and consume energy and biomass. Our current understanding of cell homeostasis and metabolism, including how cells utilize nutrients, comes largely from tissue and cell models analyzed after fractionation, and that fail to reveal the spatial characteristics of cell metabolism, and how these aspects relate to the location of cells and organelles within tissue microenvironments. Here we show the application of multi-scale microscopy, machine learning-based image segmentation, and spatial analysis tools to quantitatively map the fate of nutrient-derived 13C atoms across spatiotemporal scales. This approach reveals the cellular and organellar features underlying the spatial pattern of glucose 13C flux in hepatocytes in situ, including the timeline of mitochondria-ER contact dynamics in response to changes in blood glucose levels, and the discovery of the ultrastructural relationship between glycogenesis and lipid droplets.