Page 162 - Read Online

Ribovski et al. Extracell Vesicles Circ Nucleic Acids 2023;4:283-305 https://dx.doi.org/10.20517/evcna.2023.26 Page 7

was detected and suggested to promote fusion [58,80,81] .

By showing the absence of endosomal permeabilization, using a galectin-3 assay (see Techniques to study
endosomal permeabilization), and detecting cargo exposure to the cytosol, using GFP-loaded EVs and GFP
fluobody-expressing recipient cells (see Techniques to study EV-membrane fusion and cargo delivery), Joshi
et al. showed that EVs are taken up via endocytosis and a fraction of them (~25%) release their cargo in a
non-disruptive manner. Moreover, the site of cargo release was identified through correlative light and
electron microscopy (CLEM), revealing EV back fusion with endosomes .
[22]

Bonsergent et al. corroborated these findings, showing a very similar efficiency of the cargo release process
(30%), but using a different analytical method (NanoLuc; see Read out systems for functional delivery of RNA
cargo) [59,82] . Cells incubated with EVs loaded with NLuc-HSP70 showed only ~1% uptake in recipient cells
and the uptake never reached saturation even at very high concentrations (100 ug/ml), leading the authors
to suggest that EVs are endocytosed nonspecifically, i.e., without associating to a specific receptor, in line
[70]
with a previous study . Using cell fractionation, they showed that roughly 20%-30% of the internalized EV
cargo was present free in the cytosol of the cells. This release was not because of endosomal
[22]
permeabilization as EVs did not induce galectin-3 punctae, again corroborating results from Joshi et al .
Interferon-inducible transmembrane (IFITM) proteins IFITM 1 and 3 showed 80%-90% colocalization with
internalized EVs when overexpressed in HEK293T cells [83,84] , and their presence inhibited the cargo release
from EVs. IFITM proteins are known to inhibit fusion reactions during viral infection [83,85,86] , and may
inhibit EV membrane fusion through similar mechanisms. Overall, the rate-limiting step in EV cargo
delivery appears to be the EV uptake (~1%), while the content delivery is much more efficient.

When studying EV cargo release upon interaction with plasma membrane sheets, efficient cargo release was
[82]
observed, which was pH-dependent and proteinase K-sensitive . Using another cell-free assay, Morandi et
al. provided interesting insights into the process of endosomal fusion in EV cargo release [22,68,87] . Using a
fluorescence resonance energy transfer (FRET) assay (see Techniques to study EV-membrane fusion and
cargo delivery), EVs were shown to fuse with large unilamellar vesicles (LUVs) mimicking late endosomal
composition, i.e., without cholesterol and with LBPA, a lipid with pH-dependent fusogenic properties.
Fusion only happened at acidic pH and was significantly reduced when the EVs were treated with
proteinase K, suggesting that EV membrane proteins are required in this fusion process, in line with
previous reports [59,82] . It is important to note that the recipient LUVs did not contain proteins, meaning that
the proteins in the endosomal membranes may be dispensable for fusion and the proteins in EVs may play a
more structural than functional role, which was suggested before . EVs did not fuse with LUVs with early
[58]
endosomal membrane composition, underscoring the importance of specific lipids in the fusion
process [22,87] .

Nevertheless, there is speculation of a putative EV fusogen or set of fusogens akin to viral fusion proteins
[88]
that drive the process of membrane fusion . Most viral fusogens require activation by low pH, and
similarly, EV fusogens may get activated by acidic pH. Further, the EV membrane fusion process follows the
canonical intermediates of viral fusion, namely (A) close contact between lipid bilayers; (B) formation of a
hemifusion diaphragm by fusion with the outer leaflet; (C) initial fusion pore formation by fusion of the
inner leaflet; (D) expansion of the fusion pore resulting in content mixing [89-92] . All in all, EV fusion features
that play a role in cargo release are strikingly similar to that of some viruses. And both EVs and viruses
hijack the host cell machinery for the expression of their genetic material. That putative EV fusogens may
exhibit structural and molecular similarities to viral fusogens has already been suggested in many
studies [22,26,59,69] .