Machine learning-based morphological quantification of replicative senescence in human fibroblasts.

Although aging has been investigated extensively at the organismal and cellular level, the morphological changes that individual cells undergo along their replicative lifespan have not been precisely quantified. Here, we present the results of a readily accessible machine learning-based pipeline that uses standard fluorescence microscope and open access software to quantify the minute morphological changes that human fibroblasts undergo during their replicative lifespan in culture. Applying this pipeline in a widely used fibroblast cell line (IMR-90), we find that advanced replicative age robustly increases (+28-79%) cell surface area, perimeter, number and total length of pseudopodia, and nuclear surface area, while decreasing cell circularity, with phenotypic changes largely occurring as replicative senescence is reached. These senescence-related morphological changes are recapitulated, albeit to a variable extent, in primary dermal fibroblasts derived from human donors of different ancestry, age, and sex groups. By performing integrative analysis of single-cell morphology, our pipeline further classifies senescent-like cells and quantifies how their numbers increase with replicative senescence in IMR-90 cells and in dermal fibroblasts across all tested donors. These findings provide quantitative insights into replicative senescence, while demonstrating applicability of a readily accessible computational pipeline for high-throughput cell phenotyping in aging research.

An open access, machine learning pipeline for high-throughput quantification of cell morphology.

Cell morphology is influenced by many factors and can be used as a phenotypic marker. Here we describe a machine-learning-based protocol for high-throughput morphological measurement of human fibroblasts using a standard fluorescence microscope and the pre-existing, open access software ilastik for cell body identification, ImageJ/Fiji for image overlay, and CellProfiler for morphological quantification. Because this protocol overlays nuclei with their cell bodies and relies on coloration differences, it can be broadly applied to other cell types beyond fibroblasts. For details on the use and execution of this protocol, please also refer to Leung et al. (2022).1.

Chronic stress-driven glucocorticoid receptor activation programs key cell phenotypes and functional epigenomic patterns in human fibroblasts.

Chronic environmental stress can profoundly impact cell and body function. Although the underlying mechanisms are poorly understood, epigenetics has emerged as a key link between environment and health. The genomic effects of stress are thought to be mediated by the action of glucocorticoid stress hormones, primarily cortisol in humans, which act via the glucocorticoid receptor (GR). To dissect how chronic stress-driven GR activation influences epigenetic and cell states, human fibroblasts underwent prolonged exposure to physiological stress levels of cortisol and/or a selective GR antagonist. Cortisol was found to drive robust changes in cell proliferation, migration, and morphology, which were abrogated by concomitant GR blockade. The GR-driven cell phenotypes were accompanied by widespread, yet genomic context-dependent, changes in DNA methylation and mRNA expression, including gene loci with known roles in cell proliferation and migration. These findings provide insights into how chronic stress-driven functional epigenomic patterns become established to shape key cell phenotypes.