Single-Particle Characterization of Engineered Extracellular Vesicles Using Halotag-Linked Fluorophores to Investigate Cargo Loading

Background: Extracellular vesicles (EVs)—secreted nanoscale particles that natively transport cellular contents between cells—are promising cell-free delivery vehicles which can be engineered to incorporate and deliver specific protein and nucleic acid cargos (Fig 1a). In order to realize this potential, methods for efficiently loading and analyzing EV cargo are required. To date, single vesicle characterizations remain rare and challenging, and thus we lack understanding of population heterogeneity and quantitation of display/loading on a single particle basis. To help address this need, we developed a robust, modular method for engineering EVs to surface display a broad variety of natural and synthetic molecular cargos using the HaloTag display system. HaloTag is a modified bacterial haloalkane dehalogenase which forms a covalent bond via click chemistry with a synthetic HaloTag ligand (Fig 1b). Thus, display of a HaloTag on EVs enables surface functionalization with any HaloTag ligand.

Methods: The HaloTag protein was genetically fused to an EV-enriched membrane protein and stably expressed in HEK293FT cells via lentiviral transduction. EVs were harvested from producer cells by differential centrifugation, and HaloTag presence in cells and EVs was confirmed by Western blot. Functionality of the HaloTag fusion protein on EVs was tested via labeling with a fluorescent HaloTag ligand (Alexa Fluor 488). A cell-impermeable fluorescent dye was used to ensure that EV fluorescence was attributable to properly displayed HaloTag fusion proteins on the vesicle surface. Fluorescently labeled HaloTag EVs and controls were measured using single vesicle flow cytometry (Apogee Micro-PLUS).

Results: We observed that ~80% of HaloTag EVs were fluorescent in each EV subpopulation analyzed (exosomes and microvesicles) conferring a 20-fold increase in overall fluorescence compared to non-HaloTag vesicle controls (Fig 1c). Alexa Fluor 488 quantification beads were used to determine the absolute number of fluorophores, and therefore the (minimum) number of HaloTag proteins, per vesicle. On average, ~1000 HaloTag proteins were measured per EV for each EV population. We also confirmed this quantitation by super-resolution microscopy. One would expect a theoretical maximum of 3,000-6,000 HaloTag proteins to fit, based upon geometric estimates, on an 100-150 nm EV. By comparison, antibody labeling of the same HaloTag construct yielded a significantly lower level of fluorophore labeling per EV and the number of EVs labeled, with ~20% of vesicles labeled and with an interquartile mean of ~300 molecules per vesicle. This comparison suggests that antibody-mediated quantitation of EV surface display substantially underestimates true display rates, potentially due to steric hindrance or the limitations of reversible (antibody-mediated) vs. irreversible (HaloTag-mediated) labeling.

Implications: Our HaloTag-based platform is capable of simple, rapid, and specific surface functionalization of EVs. This system enables direct quantification of surface protein expression, which will be useful in the field of EV engineering for developing sophisticated drug and biological therapeutic delivery vehicles. Furthermore, the versatility of HaloTag functionalization facilitates the development of a wide variety of modified EVs using a single engineering platform.