Cell surface vesicles outperform standard carriers for delivering proteins and gene editors

Cells talk to each other partly by releasing tiny membrane-wrapped packets called extracellular vesicles. These particles carry proteins, RNA, and other molecules to neighboring cells, and for years researchers have been trying to co-opt this delivery system for medicine - loading vesicles with therapeutic cargo and letting cells do the rest. The challenge has been that not all extracellular vesicles are equal. Cells generate them through at least two distinct routes, and until recently nobody had rigorously compared which route produces better delivery vehicles.

A team led by Professor Shiro Suetsugu at the Nara Institute of Science and Technology (NAIST) resolved that question with a direct head-to-head comparison. Their results, published in Nature Communications on December 8, 2025, show that vesicles budding from cell-surface protrusions - generated through a pathway dependent on the I-BAR domain protein MIM - are dramatically more efficient at delivering functional protein cargo to recipient cells than the more conventional endosome-derived vesicles associated with the CD63 protein.

Two pathways, one comparison

The two classes of vesicles form in different cellular compartments. Endosome-derived vesicles originate inside the cell, within structures called multivesicular bodies, before being expelled. Protrusion-derived vesicles bud directly from the cell surface at the tips of membrane protrusions. Both types end up outside the cell, but their biogenesis - and, as this study shows, their delivery capabilities - differ substantially.

The NAIST team cultured human cells, isolated both vesicle types, loaded them with defined protein cargo, and added them to recipient cells. Using advanced live-cell and super-resolution imaging, they tracked how the vesicles entered target cells through endocytosis, moved through endosomal compartments, and released their cargo into the cytoplasm.

Protrusion-derived vesicles consistently outperformed their endosomal counterparts. Cargo loaded into protrusion-derived vesicles escaped from late endosomes into the cytoplasm far more effectively, where it remained active and functional.

Active proteins and gene editors delivered without viral components

One test cargo was Rac1, a protein that regulates cell migration. Protrusion-derived vesicles delivered active Rac1 to recipient cells, where it functioned normally and stimulated movement. The endosome-derived vesicles delivered comparatively little active protein.

The more striking test involved Cas12f, a compact genome-editing enzyme from the CRISPR family. Delivering CRISPR enzymes into cells is a persistent challenge in gene therapy - viral vectors carry safety risks, and most non-viral alternatives are inefficient. Protrusion-derived vesicles transported Cas12f with dramatically higher functional efficiency on a per-protein basis than the endosomal vesicles, achieving genome-editing activity without the use of viral vectors or viral fusogenic proteins.

"In this study, we were able to clearly demonstrate that vesicles generated from cell-surface protrusions function as a highly efficient and natural protein delivery system," Suetsugu said. "By directly comparing them with endosome-derived vesicles, we gained new insight into how cells achieve efficient cytosolic protein transfer without relying on viral mechanisms."

Implications for protein and gene therapy

The significance lies partly in what these vesicles are not. They are not engineered viral particles. They are not synthetic nanoparticles. They are structures the cell makes naturally, using its own membrane machinery. That origin may reduce immunogenicity and other safety concerns that complicate viral delivery systems.

"Our findings show that cells might already possess a remarkably effective, virus-free delivery mechanism," Suetsugu concluded. "By understanding and harnessing this natural system, we may be able to develop safer and more precise strategies for genome editing, regenerative medicine, and protein-based therapeutics."

Several important qualifications apply. All the experiments described in this paper were conducted in cultured cells - the jump from cell culture to animal models to eventual clinical use involves many additional steps, each with its own challenges. The study does not report data from living organisms, and the efficiency advantages seen in vitro may not fully replicate in more complex biological environments. Scaling up vesicle production, ensuring consistent cargo loading, and verifying that cargo remains active across target tissue types are all engineering problems that remain to be addressed.

The paper also does not claim that MIM-dependent protrusion vesicles are ready for therapeutic use - it establishes their mechanism and comparative efficiency as a scientific foundation that others can build on.