Multispectral Fingerprinting Resolves Dynamics of Nanomaterial Trafficking in Primary Endothelial Cells.

Intracellular vesicle trafficking involves a complex series of biological pathways used to sort, recycle, and degrade extracellular components, including engineered nanomaterials (ENMs) which gain cellular entry active endocytic processes. A recent emphasis on routes of ENM uptake has established key physicochemical properties which direct certain mechanisms, yet relatively few studies have identified their effect on intracellular trafficking processes past entry and initial subcellular localization. Here, we developed and applied an approach where single-walled carbon nanotubes (SWCNTs) play a dual role-that of an ENM undergoing intracellular processing, in addition to functioning as the signal transduction element reporting these events in individual cells with single organelle resolution. We used the exceptional optical properties exhibited by noncovalent hybrids of single-stranded DNA and SWCNTs (DNA-SWCNTs) to report the progression of intracellular processing events two orthogonal hyperspectral imaging approaches of near-infrared (NIR) fluorescence and resonance Raman scattering. A positive correlation between fluorescence and G-band intensities was uncovered within single cells, while exciton energy transfer and eventual aggregation of DNA-SWCNTs were observed to scale with increasing time after internalization. An analysis pipeline was developed to colocalize and deconvolute the fluorescence and Raman spectra of subcellular regions of interest (ROIs), allowing for single-chirality component spectra to be obtained with submicron spatial resolution. This approach uncovered correlations between DNA-SWCNT concentration, dielectric modulation, and irreversible aggregation within single intracellular vesicles. An immunofluorescence assay was designed to directly observe the DNA-SWCNTs in labeled endosomal vesicles, revealing a distinct relationship between the physical state of organelle-bound DNA-SWCNTs and the dynamic luminal conditions during endosomal maturation processes. Finally, we trained a machine learning algorithm to predict endosome type using the Raman spectra of the vesicle-bound DNA-SWCNTs, enabling major components in the endocytic pathway to be simultaneously visualized using a single intracellular reporter.