Table 2.

Characteristics of molecular tools and mechanisms for loading Cas proteins into EVs.

Name	Advantages	Drawbacks	Prospects
EXPLORs [201]	-Highly efficient -Utilize reversible protein-protein interaction modules -Transient protein docking into EVs	-Have not been used for CRISPR/Cas9 -sgRNA delivery is not addressed -Blue light is toxic to the cells	-Can be coupled with other light-induced dimerization (LID) or chemically-induced dimerization (CID) systems for sgRNA packaging -Cycles of blue light may be less toxic to producer cells
NanoMEDIC [202]	-Very first demonstration of successful exosome engineering for packaging and delivering CRISPR/Cas9 -Very efficient packaging of both Cas9 and sgRNA-Very efficient in vitro and in vivo genome editing -Proven activity in vivo-Cleared in vivo within 3 days -Scalable system in chemically-defined media with suspension cell culture	-Use HEK293T, a transformed cell line -Transformed cell lines produce exosomes with pro-oncogenic properties -Use rapamycin, an immunosuppressive drug with a number of potential adverse effects, to induce dimerization of domains. Rapamycin may potentially be packaged into EVs or alter exosome composition -Tissue-specific targeting upon systemic delivery has not been investigated -Reliance on HIV-1 Tat/Gag to drive sgRNA expression imposes the risks of toxicity both to producer cell lines and target cells -HIV-1 Tat/Gag may alter exosome composition -Co-produces Cas9 and sgRNA in the same cell	-Can be potentially expanded to clinically relevant EV-producing cell lines -Packaging of rapamycin into EVs and its effects on exosomes still needs to be defined -Any type of CRISPR/Cas system can be packaged
Tags for post-translational modification [203]	-Simple and feasible even for large proteins	-Have not been used for CRISPR/Cas9 -Most likely cell type-specific -Efficiency is unclear-Functionality in target cells unclear	-Simple and feasible -Applicability for Cas proteins needs to be defined
GEDEX or stochastic packaging [204]	-Very first demonstration of CRISPR/Cas9 RNP stochastic packaging into exosomes -Packaging of both Cas and sgRNAs -Efficient in vitro and in vivo genome editing -Scalable	-Utilize transformed cell lines -Transformed cell lines produce exosomes with pro-oncogenic properties -Tissue-specific targeting upon systemic delivery has not been investigated -Co-produce Cas9 and sgRNA in the same cell	-Very simple (overexpression of CRISPR/Cas components) -Any type of CRISPR/Cas system can be packaged
WW-Ndfip1 interaction [87]	-Efficiently delivers Cre-recombinase to target cells -Tested in vivo -Very simple (very short fusion peptides)	-Has not been used for CRISPR/Cas9 -Utilize mouse cells; not studied in human cells -Not studied with CRISPR/Cas packaging -Do not contribute to sgRNA packaging -Overexpressed Ndfip1 is required -Ndfip1 is toxic to producer cells -Ndfip1 interacts with numerous pro-oncogenic and pro-apoptotic factors	-Ndfip1 is toxic to producer cells -Ndfip1-WW interaction needs to be rationally engineered
ARMMs [205]	-Simple loading of protein and RNA cargo into vesicles -Efficient packaging of CRISPR/Cas RNPs -Efficient genome editing -Scalable -ARMMs may enter cells by direct fusion -Cargo bypasses endolysosomal pathway	-Use transformed cell lines -Transformed cell lines may produce vesicles with pro-oncogenic properties -Tissue-specific targeting upon systemic delivery has not been investigated -Co-produces Cas9 and sgRNA in the same cell	-Very simple packaging -Effects of ARRDC1 expression on producer cells and vesicle composition need to be addressed -Benefits of ARMMs over exosomes in terms of scalability and production need to be addressed
Nanoblades [206]	-Very limited carry-over of cellular proteins or overexpressed RNAs -Can potentially be produced from non-transformed cell lines -Have been combined with BaEV and VSV-G for improved delivery -Tested in vivo -Complex homologous DNA templates to generate knock-ins	-Carry-over of cellular RNAs (including those with pro-oncogenic potential) has not been investigated -Virus-like particles (viral origin) with membrane-associated proteins -Competition between HahMLV and Gag-PolMLV potentially reduces Cas packaging per particle -MLV protease may non-specifically cleave SpCas9 and reduce activity -Cas9 and sgRNA co-produced in the same cell	-Any type of CRISPR/Cas system can be packaged -Demonstrated for SpCas9 and dCas9-VPR
VEsiCas [207]	-Efficient Cas9 and sgRNA packaging -Very simple and easy-to-use fusion of Cas9-VSV-G and sgRNA-expressing constructs -Efficient, on-target genome editing -Tested in vivo	-Use HEK293T, a transformed cell line -Generated EVs are not exosomes; their properties and interaction with target cells need to be determined -Tissue-specific targeting upon systemic delivery has not been investigated -Quantity needed and quality of VEsiCas remain to be investigated -Composition of VEsiCas and co-packaging of potentially toxic proteins is not clear -Cas9 and sgRNA co-produced in the same cell	-Can be potentially expanded to clinically relevant EV-producing cell lines -Any type of CRISPR/Cas system can be packaged
Gesicles [208]	-Transfer Cas9:sgRNA RNPs -Efficient genome editing in target cells -Simple packaging system	-Use HEK293FT, a transformed cell line -Evidently less effective than NanoMEDIC -Cas9 protein half-life is reduced -<1% of produced gesicles contain RNPs -Carry-over of producer proteins and RNAs is possible -Use potentially toxic A/C heterodimerizer -Cytotoxicity and immunogenicity have not been studied -Not tested in vivo -Cas9 and sgRNA co-produced in the same cell -No tissue-specific targeting reported	-Potentially consist of a vesicle population mixed with cell waste as evidenced by increased gesicle formation following transfection