Exosomes as a biovector : a new potential for drug delivery
Exosome loading and modification
One of the many interesting uses of EVs is their use as drug delivery vectors. Our team pioneered this field in 2010’s, showing the proof of concept of parental cell loading with nanoparticles and model drugs to load EVs [88, 89]. In general, EVs could be used for chemotherapy or other drugs, protein, RNA and miRNA or gene delivery. The potential interest is to protect against drug degradation, target a specific tissue  or evade an immune response against the drug of interest.
However, up to now, loading methods are relatively inefficient or artefactual, whatever the method used (incubation , parental cell loading , transfection , permeabilization , electroporation , sonication , etc), leading to the loss of a large amount of the drug to encapsulate. In this particular topic, a lot can be learned from the comparison between EVs and liposomes. As an example, whereas lipophilic drugs are quite easy to encapsulate in liposomes, amphiphilic drugs are rather difficult, able to cross back the membrane. It should be the same in EVs, but results published rarely discuss the encapsulation efficiency. The case of hydrophilic compounds is really interesting and more complex. Liposome loading methods for these compounds are divided in two methods, the active encapsulation that uses a pH gradient, or the ability of the drug to precipitate in the liposome interior media to obtain a high concentration of the drug in the liposome. On the contrary, the passive methods consist in producing the liposome in the desired environment, enabling the encapsulation of a portion of the surrounding media in the liposome during the process. The passive methods are limited by the maximal solubility of the drug, and the loss of a portion of the drug to be encapsulated. Transposing this in EVs, the use of active encapsulation is difficult (it needs a change in the interior Exosomes medium, and would therefore need a loading method as a first step), and the use of passive encapsulation leads to similar conclusions, i.e. loss of cargo and low concentration inside the liposome. In particular, the maximal concentration of EVs is rather limited compared to liposomes, meaning that most of the drug is lost during the process. A simple calculation based on classical values from the field give an idea of the efficiency that can be achieved: for a concentration of 1013 EVs/mL (which is already huge), a mean size of 150 nm, we see that the volume occupied by EVs is roughly 1.8 x 10-2 mL, i.e. 1.8% of the volume, therefore the maximal encapsulation that can be expected is 1.8% of the drug. Moreover, EV loading with hydrophilic compounds involves the creation of a hole in the membrane (electroporation, saponin, sonication, etc), with a potential loss of the EV content.
Once again the idea of targeting EVs to a specific tissue benefit from a comparison with the liposome/nanoparticle field. The main rule in the field of targeting is that to achieve a potent specific targeting one needs a long half-life in the blood, avoiding rapid nonspecific clearance by the reticulo-endothelial system to ensure the potential specific interaction with the receptor of interest. This long half-life allows the nanoparticle to circulate at least once in the blood vessels that distribute the place to target, and maybe to specifically bind on it. On the contrary, a short half-life impeaches the nanoparticle to circulate throughout the body and to have a chance to reach the desired site of action, precluding any specific targeting. As an example, the enhanced permeability rate (EPR) effect, i.e. the enhanced delivery of nanoparticles in tumors due to a leaky vasculature and defective lymphatic drainage is effective only in the case of prolonged circulation time. Getting back to the EVs, we clearly see that EVs do not fulfill this first criteria of long half-life. Their half-life of intravenously-injected EVs from various cell source is rather in the 3-30 min range, for example EVs from neuro2a cell derived EVs have a <5 min half-life , and as expected, they mainly distribute in the liver and the spleen, major actors of the reticulo-endothelial system, in a nonspecific way. Therefore, there is a crucial need to enhance the circulation time of EVs to investigate their still potential targeting properties. This would allow us to investigate the potential of MSC derived EVs in the targeting of inflamed tissue, as described by their parental cells , or their blood brain barrier crossing effect, that is for the moment limited by the very low amount of EVs that was shown to be distributed in the brain (maximum 0,5 to 1% ). The half-life increase would also allow a much better targeting effect of the previously published EV modifications like expressions of RVG peptide expression on its surface to target specific acetylcholine receptor that showed interesting but limited results .
Once again, we propose the use of surface modification as a way to enhance half-life of vesicles, for example using PEGylation, as proposed by articles from Koojimans et al [96, 97] describing the use of PEG grafting at the surface of EVs. Following this direction, our proposition of PEG induced EV/PEGylated liposome fusion to produce PEGylated hybrids seems promising in vitro.
 N. Luciani, C. Wilhelm, F. Gazeau, The role of cell-released microvesicles in the intercellular transfer of magnetic nanoparticles in the monocyte/macrophage system, Biomaterials, 31 (2010) 7061-7069.
 A.A. Faraj, F. Gazeau, C. Wilhelm, C. Devue, C.L. Guérin, C. Péchoux, V. Paradis, O. Clément, C.M. Boulanger, P.-E. Rautou, Endothelial Cell–derived Microparticles Loaded with Iron Oxide Nanoparticles: Feasibility of MR Imaging Monitoring in Mice, Radiology, 263 (2012) 169-178.
 O.P.B. Wiklander, J.Z. Nordin, A. O’Loughlin, Y. Gustafsson, G. Corso, I. Mäger, P. Vader, Y. Lee, H. Sork, Y. Seow, N. Heldring, L. Alvarez-Erviti, C.I.E. Smith, K. Le Blanc, P. Macchiarini, P. Jungebluth, M.J.A. Wood, S.E.L. Andaloussi, Extracellular vesicle in vivo biodistribution is determined by cell source, route of administration and targeting, Journal of Extracellular Vesicles, 4 (2015) 10.3402/jev.v3404.26316.
 X. Zhuang, X. Xiang, W. Grizzle, D. Sun, S. Zhang, R.C. Axtell, S. Ju, J. Mu, L. Zhang, L. Steinman, D. Miller, H.-G. Zhang, Treatment of Brain Inflammatory Diseases by Delivering Exosome Encapsulated Anti-inflammatory Drugs From the Nasal Region to the Brain, Molecular Therapy, 19 (2011) 1769-1779.
 T.A. Shtam, R.A. Kovalev, E.Y. Varfolomeeva, E.M. Makarov, Y.V. Kil, M.V. Filatov, Exosomes are natural carriers of exogenous siRNA to human cells in vitro, Cell Communication and Signaling : CCS, 11 (2013) 88-88.
 M.J. Haney, N.L. Klyachko, Y. Zhao, R. Gupta, E.G. Plotnikova, Z. He, T. Patel, A. Piroyan, M. Sokolsky, A.V. Kabanov, E.V. Batrakova, Exosomes as drug delivery vehicles for Parkinson’s disease therapy, Journal of Controlled Release, 207 (2015) 18-30.
 O.P.B. Wiklander, J.Z. Nordin, A. O’Loughlin, Y. Gustafsson, G. Corso, I. Mäger, P. Vader, Y. Lee, H. Sork, Y. Seow, N. Heldring, L. Alvarez-Erviti, C.I.E. Smith, K. Le Blanc, P. Macchiarini, P. Jungebluth, M.J.A. Wood, S.E.L. Andaloussi, Extracellular vesicle in vivo biodistribution is determined by cell source, route of administration and targeting, Journal of Extracellular Vesicles, 4 (2015) 26316.
 J.-S. Schoefinius, B. Brunswig-Spickenheier, T. Speiseder, S. Krebs, U. Just, C. Lange, Mesenchymal Stromal Cell-Derived Extracellular Vesicles Provide Long-Term Survival After Total Body Irradiation Without Additional Hematopoietic Stem Cell Support, Stem Cells, 35 (2017) 2379-2389.