Exploring Real-time Cell Mass Dynamics as a Biomarker: Apoptosis, Necroptosis and Cell Type Recognition

Abstract

Cells are the basic units of life, and their growth, function, and regulated cell death are central to health and disease. Most assays rely on labels or endpoints, obscuring the exact transition from life to death and limiting real time analysis for drug discovery, tissue engineering, and personalised medicine. Cell mass regulation is linked to gene expression, heterogeneity, lineage, and cell death, but it remains underused as a biomarker because tools for accurate, continuous measurements in physiological conditions, including liquids, are scarce.

This thesis addresses these limits using an inertial picobalance for high resolution, real time, label free mass measurements. First, we identified and corrected a key artifact in liquid and non liquid operation: far out of band laser emission misplaces the beam waist, reducing photothermal excitation efficiency and biasing mass. Our correction increased photothermal efficiency up to eightfold, improving sensitivity to subtle mass changes (Chapter 3). We then validated accuracy using calibration blocks across liquids of different density (Chapter 4), achieving more than 96% accuracy even when the fluid was denser than the sample.

With the optimized system, cell mass dynamics provide a universal, real time biomarker for regulated cell death (Chapter 5). Distinct fluctuations separate apoptosis from necroptosis, and a criterion pinpoints when cells cease metabolic and structural activity. Extending this to identity, we trained a deep learning classifier on nine mammalian cell lines, reaching 85% accuracy in cell type recognition (Chapter 6).

Overall, this work establishes cell mass dynamics as a powerful, label free readout for dissecting cell death pathways and identifying cell types in real time, surpassing conventional assays and enabling advances in cellular biophysics and translational research.