UC San Diego

In instances where tissue availability is limited or surgical risks to the patient are significant, the application of multiplexed biomedical imaging can be a challenge due to limitations in spectral bandwidth of fluorophores and histological stain chroma. This dissertation presents a label-free multimodal optical imaging platform that combines stimulated Raman scattering (SRS), second harmonic generation (SHG), and multiphoton fluorescence (MPF) in a single microscopy setup. Using this home built multimodal microscopy, the dissertation investigates subcellular organelle morphology, their molecular composition, and metabolic dynamics in triple negative breast cancer cells under tandem nutrient modulation, as well as assesses various biomarkers such as 3D mesangial expansion, collagen fiber thickening, oxidative stress, and lipid dysregulation in diabetic kidney disease. From subcellular to tissue levels, the platform provides spatial, biomolecular, and metabolic insight across disease models without relying on exogenous labels or serial sectioning. Developing and employing analytical techniques such as chemometric spectro-microscopy such as relative entropy, penalized reference matching, and stimulated Raman histology, this work advances quantitative bioimaging while preserving the flexibility of traditional methods through its non-destructive properties. Ultimately, this custom platform optimizes informatics while minimizing sample utilization and disruption, unlocking novel biological insights that were previously reliant on more intricate sample preparation, expensive reagents, and abundant tissue availability.