sEVs are secreted by all cells and are originated as intraluminal vesicles during the process of multivesicular body formation. The biogenesis of sEVs has been addressed by many publications (Baumann, 2021; Brites, 2020; Russell et al., 2019; van Niel et al., 2018). Briefly, the biogenesis of sEVs (Figure 2A) consists of three different stages: (1) the formation of endocytic vesicles from the plasma membrane, (2) the inward budding of the endosomal vesicle membrane that maturates in multivesicular bodies (MVBs), which consist of intraluminal vesicles (ILVs), and (3) the fusion of MVBs with the plasma membrane (instead of being sent to degradation into the lysosomes), allowing the release of these ILVs, including sEVs, into the extracellular space (Zhang Y. et al., 2019). The pattern of nanospherical membrane-type is derived from the parent cells from which the sEVs are formed. Their respective cargo will likewise vary depending on the cell type of origin and status (Kang et al., 2021). The sEV membrane is formed by two layers of proteins and lipids, including cholesterol, phospholipids, glycerophospholipids, and sphingolipids that maintain its stability and structure (Figure 2B). Their lipid composition accounts for their unique rigidity (Skotland et al., 2019; Lin Y. et al., 2020).

After sEV release by exocytosis, their cargo is protected from enzymes like proteases and ribonucleases by their lipid bilayer membrane (Lorenc et al., 2020) (Figure 3). Once at the extracellular space, sEVs are transferred into the recipient cells through interaction with proteins that facilitate subsequent endocytosis by specific processes, such as receptor interaction, membrane fusion, and internalization. The internalization step of sEVs by the recipient cells can occur via receptor-ligand interactions, direct fusion of membranes, or internalization via endocytosis, and is normally dependent on the cell type, as recently reviewed by Malloci and others (Malloci et al., 2019). For example, the selective transfer of sEVs from oligodendrocytes and their subsequent uptake by microglia through a macropinocytosis mechanism does not require binding to the specific receptor (Fitzner et al., 2011). Another example is the injection of oligodendroglia sEVs in mouse brain, which results in a functional retrieval of sEV cargo in neurons through a clathrin-dependent endocytosis (Fruhbeis et al., 2013). Another study refers to the release of EVs from primary cortical astrocytes and microglial cells being triggered by ATP-mediated activation of P2X7 receptors after contact with phosphatidylserine at the cell surface (Bianco et al., 2009). sEVs carrying a multitude of proteins, such as myelin proteins, as well as RNA, are released from oligodendrocytes and endocytosed by neurons (Frohlich et al., 2014). In addition, the release of serotonin from neurons has been implicated in the release of microglial sEVs, upon binding of serotonin to specific receptors found in the microglia, suggesting a neurotransmitter dependent release (Glebov et al., 2015). To sum up, sEV uptake is determined by multiple mechanisms, together with cell-dependent different combinations of strategies. These specificities should be considered when designing sEV-based therapies.

sEVs exert unique biological activities by stimulating regeneration and modulating pathological conditions, properties that can be explored for medicinal purposes. Namely, they are able to transfer RNA and proteins from donor cells to other cells in the surrounding milieu (Simons and Raposo, 2009). Furthermore, EVs derived from mesenchymal stem cells (MSCs) appear particularly useful in enhancing recovery from various injuries. MSCs are commonly used as a source of sEVs because they can recapitulate the biological activity of MSCs and have been used as a cell-free therapeutic (Harrell et al., 2021). For example, stem cells may modulate their biological effects through the delivery of genetic information that will alter the gene expression of the target cells (Zhao et al., 2019). In addition, EVs from both immune and non-immune cells are shown to suppress or stimulate both adaptive and innate immunity, which effects likely depend on the environmental context, as well as on the type of EV from a particular immune cell (Kim et al., 2005; Bu et al., 2015). sEVs are potential fingerprints of their originating cells and their composition largely depends on the donar cell, although it can also be determined by the cell metabolism and its needs, as well as be influenced by cellular and environmental factors (Bell and Taylor, 2017). As an example, release of miRNAs from cells into sEVs can occur either passively or actively. Therefore, it is very important to choose the cell lines that will act as the source of sEVs, once the donor cell will determine their ability to selectively target cells (Sancho-Albero et al., 2019). It has been shown that the interaction between the drug delivery systems is affected by the EV surface proteins (Yin et al., 2013). The ability of different types of EVs to modulate immune responses allows their therapeutic potential as a tool with theragnostic applications in nanomedicine (Thomas et al., 2021). Therefore, we may enrich sEVs with promising curative molecules or engineering cell lines to produce sEVs with a specific and desired target specificity (Gao et al., 2018).

The use of sEV technology as drug delivery vehicles must meet two basic requirements: first, a reproducible and scalable isolation with established production protocols to achieve high purity and elevated yield of a defined population of sEVs for clinical use; and second, the high drug-loading content in sEVs must be enough to obtain a therapeutic response (Armstrong and Stevens, 2018; Lin Y. et al., 2020; Wu Z. et al., 2021). Cells may produce higher or lower number of sEVs depending on several factors, but what will be important is that the production of enormous quantities of vesicles does not modify cell characteristics and behavior. Alterations in the composition and function of the sEVs may be performed using environmental parameters, such as pH gradients, mass transfer, or hydrodynamic force during scale-up processes. The number of sEVs released depends on the cell type, physiological state, and microenvironmental conditions. The production increases in cancer cells under hypoxic conditions, in melanoma cells under acidic microenvironment, and under cell stress induction with calcium ionophores (Zaborowski et al., 2015). It was indicated that immature dendritic cells produce a limited number of sEVs in the µg range (Alvarez-Erviti et al., 2011). However, MSCs secrete more elevated amounts, usually in the milligram range (Chen et al., 2011), making them a more relevant source for the production of therapeutic EVs targeting neurodegenerative diseases. Nevertheless, it should be taken into account that MSCs may undergo senescence after a few passages, a condition that will lead to the production of sEVs with impaired regenerative capacity as compared to those from younger cells (Severino et al., 2013). The confluence of cell cultures also plays a critical role in the biogenesis and secretion of sEVs. Cell culture in a pre-confluent state showed a decreased sEV secretion compared to the confluent one, potentially due to changes in cholesterol metabolism (Llorente et al., 2013). In addition, sEV secretion rate from the 3D cultured cells is better when cells are close to confluence (Thippabhotla et al., 2019).

Contact inhibition between cells can also reduce sEV release, since these cells become quiescent and do not actively divide (Whitford and Guterstam, 2019). Apart from the individual characteristics of the cell culture and of the cell type that originates the sEVs, some methods are designed to increase their total yield, such as intracellular calcium production (Savina et al., 2003), external stress (de Jong et al., 2012), cytoskeletal blocking (Parolini et al., 2009), drug stimulation (Datta et al., 2018; Wang et al., 2020), and induction of gene expression factors (Boker et al., 2018). In spite of increasing the sEV yield, such methods should be considered with caution, since they may also alter sEV composition, therapeutic efficacy, and safety (Emam et al., 2018; Whitford and Guterstam, 2019). Another aspect to take into consideration is the increased risk of contamination with apoptotic bodies (Emam et al., 2018; Whitford and Guterstam, 2019), although their presence was referred to also have important regulatory roles, and can be a mechanism to immunomodulate dying cells (Caruso and Poon, 2018).

Some recent studies report the relevance of neural cell-derived sEVs to be used in neurodegenerative conditions (Huo et al., 2021). The direct conversion of somatic cells into induced neural precursor cells (iNPCs) provides abundant production of sEVs (Ma et al., 2019; Yuan P. et al., 2021), mainly if using bioreactors (Baghbaderani et al., 2011; Miranda et al., 2016). These sEVs after transplantation revealed to have therapeutic properties in ischemic stroke (Zhang et al., 2020). Lately, studies considered that milk is a viable and safe source of sEVs for therapeutic delivery. For instance, breast milk was shown to contain EVs in large quantities and with an enriched cargo in miRNAs (Jiang et al., 2021). Moreover, it has been demonstrated that milk sEVs improve oral drug bioavailability by protecting their cargo against degradation by low pH, RNases, or conditions that mimic digestion in the gastrointestinal tract (Izumi et al., 2012). Other biofluids, such as saliva, blood, and urine, were also demonstrated to be an alternative source of sEVs. Nevertheless, each body fluid has a clearly distinct vesicle profile and it is not clear what is the delivery potential and safety of these systems for continuous or long-term continuous exposure (Yuana et al., 2013).

Applying the sEV-based therapies to clinics has been slowed down due to sEV low availability and difficulties for large-scale production. Beyond the use of small molecule modulators capable of enhancing sEV secretion, e.g., norepinephrine, N-methyldopamine, and forskolin recently described for MSCs (Wang et al., 2020), other upscaling approaches and technologies for their production have been reported, assuring not only high quality but also high quantity (Andriolo et al., 2018; Mendt et al., 2018). These strategies involve standardized production methods that increase the volume of cell culture from flasks to containers, bioreactors, or hollow fibers, aiming production toward a good manufacturing practice–grade (GMP-grade) of sEVs. Different methods have allowed researchers to isolate and characterize sEV populations in a more accurate way. First works describing methods for the production, purification, and characterization of clinical grade sEVs (cGMP) emerged in the beginning of 2000 (Lamparski et al., 2002; Navabi et al., 2005).

Bioreactors and hollow-fiber perfusion bioreactors have been used in cell culture systems for a long time and they improve sEV yield. These reactors support large numbers of cells at high densities in a continuous mode without the need of splitting or subculturing these cells (Yamashita et al., 2018; Yoo et al., 2018). Using a hollow fiber bioreactor, Watson and others showed that the yield of sEVs obtained from the conditioned media of human embryonic kidney 239 cells (HEK293 cells) was 100 × 10⁷ particles per µg EV protein, instead of 20 × 10⁷ by the conventional flask culture (Watson et al., 2016). In another study with MSCs, the authors showed that once compared to the conventional methods, the bioreactor led to a 5.7-fold increase of EV concentration in the conditioned medium, with a threefold increase in the number generated per cell (MSC from bone marrow originated an average of 3 × 10¹¹ EV particles per 60 ml final volume) (de Almeida Fuzeta et al., 2020). Additional advantages include less risk of contamination and less labor, besides the higher yield in a short incubation period. However, ideal criteria for production of large-scale GMP-grade EVs, such as scalability, reproducibility, size distribution, safety, surface charge, and purity of the resulting product still remain to be determined (Yamashita et al., 2018). Nevertheless, biomanufacturing of cells engineered to produce therapeutic proteins is well established, and practices or lessons from this area of industry could be transferred into the production of sEVs to be used in biomedicine.

Stability of products is mandatory and required for biomedical applications, such as the case of drug delivery systems. There are few research groups trying to understand the relationship between storage conditions and the integrity of sEVs (Jeyaram and Jay, 2017). The ISEV recommends that sEVs must be suspended in phosphate buffered saline (PBS) and stored at −80°C (Witwer et al., 2013). This recommendation was based on previous works showing that the structural and biological stability of the sEVs stored at −80°C for periods of up to 5 months have not been affected. Mendt and collaborators showed that the number and size distribution of sEVs after freezing (for 45 d or 6 months) and thawing are not affected as evaluated by Nanoparticle Tracking Analysis (NanoSight) and electron microscopy compared with freshly prepared sEVs and for up to 48 h when stored at room temperature or 4°C (Mendt et al., 2018). Recently, it has been shown that sEVs should be stored at 4°C or -20°C for short-term preservation; for longer periods of storage for therapeutic application, the most suitable temperature of preservation of sEVs is −80°C (Wu JY. et al., 2021). The effects of storage do not seem to depend on the sEV cellular or tissue source (Maroto et al., 2017).

A correct cryopreservation of biological material intended for therapeutic use is also crucial. When frozen and kept properly, specimens maintain their viability and prevent osmotic damage through the use of cryoprotectants, such as sucrose or trehalose (Kusuma et al., 2018). These cryoprotectants do not penetrate in the cell membrane. Instead, they stabilize and preserve the cell membranes, allowing regulated extracellular ice growth during controlled cooling, thus protecting sEVs from cryodamage (Bosch et al., 2016). There is still a critical need to develop protocols for collection and storage of sEV samples due to the still undefined standard operation procedures for their preservation conditions after isolation. Emerging techniques for EV storage include the lyophilization and freeze-dried sEVs can be stored at −20°C, a technique considered as a cost-saving strategy and holding great promise for translation into therapeutic protocols (Yuan F. et al., 2021). However, such studies are preliminary and require further investigation and standard protocols. More work is needed in this area and the impact on long-term storage and cold chain processes shall be considered.

Standard flow cytometry is one of the most common and prevalent tools used to analyze the origin, size, and morphology of lEVs, while high resolution flow cytometry and imaging flow cytometers have been used for sEVs, as reviewed in MISEV 2018 guidelines (Théry et al., 2018) and recently published (Botha et al., 2021) Such a high-throughput, multi-parametric technique quantitates thousands of single cells or particles and quickly analyzes both their relative size and granulation (Ko et al., 2016; Tertel et al., 2020; Botha et al., 2021).

The working principle of a flow cytometer is that a laser beam with a specific wavelength is directed through a stream of fluid containing suspended particles and the scattered light is converted to an intensity-associated voltage pulse that can be quantified later. The degree of light scattering depends on the presence of particles in the samples, and sEVs can be quantified and/or classified accordingly to antigen expression levels using specific fluorescently labeled antibodies (Konoshenko et al., 2018). This same approach is also used by some commercially available kits and enables parallel multiple surface biomarker detection with different fluorescent antibodies. However, sEVs are too small to allow that the standard flow cytometry captures the florescence signal, which can be increased through the usage of immunoconjugated beads (Ko et al., 2016). Flow cytometry is a methodology that can be combined with molecular methods increasing statistical robustness and allowing simultaneous analysis of various antigens. Nevertheless, there is still a lack of robust protocols reported, and many studies regarding flow cytometry of sEVs have incomplete methodological descriptions and insufficient calibration and standardization processes to generate the data, as indicated in MISEV guidelines for flow cytometry experiment with EVs (MIFlowCyt-EV) (Welsh et al., 2020). Some studies have used engineered CD63eGFP-labeled sEVs and imaging flow cytometry as a robust multi-parametric detection and quantification of single sEVs and sEV subsets in heterogeneous samples (Görgens et al., 2019). In recent years, ExoView® platform was proposed. It allows the addition of fluorescent antibodies on single sEVs and other EVs. No purification is required, allowing us to quantify the expression of relative proteins with a single fluorescent antibody sensitivity, while simultaneously measuring their size and number (Harkonen et al., 2019). Some of the biophysical approaches routinely used for the characterization of sEVs are focused on measuring the size distribution and morphology. Dynamic light scattering (DLS) allows the assessment of sEV size distribution based on the intensity of the scattered light (Frisken, 2001). Similar to DLS, the nanoparticle tracking analyzer (NTA) uses a laser beam to illuminate sEVs in the sample (Filipe et al., 2010). The suspended sEVs are illuminated by a laser, and the intensity of the light fluctuates over time since sEVs go through Brownian motion, that is, they are then identified and related by movement and particle size based on the Stokes-Einstein equation, where the diffusion velocities are inversely proportional to the size of particles. Both methods provide accurate and sensitive size distributions, but while DLS gives the diameter range of the analyzed vesicles, NTA is capable of tracking a single sEV overcoming polydispersity problems (Frisken, 2001; Filipe et al., 2010). Lately, sEV protein detection has been possible thanks to the quantum dot method coupled with immunomagnetic capture and enrichment. sEVs are captured by magnetic beads based on CD81 protein expression and subsequently detected by fluorescent spectroscopy (Vinduska et al., 2021).

Transmission electron microscopy (TEM) has a higher resolution compared with the other electronic microscopy techniques and has similar procedures for fixation and contrast enhancement. It is widely used to characterize the structure, morphology, and size of various biological components (Figure 1) (Théry et al., 2006). In the scanning electron microscopy (SEM) technique, a focused electron beam causes electron emission from the sEV surface. Samples are chemically or cryogenically fixed followed by dehydration and sputter-coated with a thin layer of gold or carbon for imaging (Chuo et al., 2018). These electrons are collected and magnified using a special lens. Usage of immune-gold labeling in TEM and SEM, as well as cryo-TEM allows the observation of EVs in their hydrated-like native state (Linares et al., 2017).

ExoCarta was previously described in 2009 and is routinely used by researchers to validate and/or characterize their findings in sEVs (Mathivanan and Simpson, 2009). ExoCarta is an online database that allows various groups to establish sEV markers with detailed information about lipids, proteins, and RNA sequences that were identified in specific sEV preparations (Mathivanan et al., 2012). Up to now, approximately 10,000 different proteins and more than 3,000 mRNAs have been characterized (http://www.exocarta.org/). ExoCarta contains annotations on the study that identified the molecule, the sEVs isolated from a particular sample, the employed methods of isolation and identification, and the date of publication. Free web-based resources like Exocarta, and others such as EVpedia and Vesiclepedia, help the scientific community in elucidating the molecular mechanisms and pathophysiological effects of EVs, specifically of sEVs, on target cells, as well as on their therapeutic potential (Mathivanan et al., 2012; Simpson et al., 2012). Recently, the topics on EVs have been updated, providing suggestions on molecular markers for their characterization based on protein composition, as well as on isolation and standardization techniques (Théry et al., 2018).

Western blot is used to separate and identify proteins and can be applied to detect and confirm the presence and expression levels of the sEV-specific proteins in purified sEV samples of tissue homogenate or extracts (He et al., 2014). After sEV isolation, they can be lysed and used for proteomics or be transferred onto a membrane for Western blotting. The specific markers frequently used as sEV markers are the characteristic proteins that belong to their biogenesis process and include, besides those pointed out in Section 3.1.5, Rab GTPase, annexins, flotillin, and the endosomal sorting complexes required for transport (ESCRT) pathway-dependent protein Alix (Simpson et al., 2009). It is a useful technique to identify proteins that are associated with sEVs once it can process multiple proteins at the same time. It is also important to state that although Western blot is often used to detect and confirm the presence of proteins in sEVs, it cannot completely exclude contaminants from different vesicles (He et al., 2014; Zhang P. et al., 2019). An alternative to Western blot, ELISA may be performed, which is a less time-consuming technique that also requires a smaller amount of sample. This technique uses sEV target specific antibodies immobilized to a solid surface to capture sEVs, and is followed by labeling with a detection antibody (Logozzi et al., 2020). Nevertheless, both Western blot and ELISA have some disadvantages, such as low specificity and quality, as well as higher costs.

One of the important characteristics of sEVs is that they may contain nucleic acids (DNAs, RNAs, miRNAs) that will play a crucial role in cell-to-cell communication and gene regulation in the target cell. These nucleic acids can be easily quantified by real-time quantitative reverse transcription polymerase chain reaction (RT-qPCR). This technique has several advantages since it does not need high sample volumes, it has higher sensitivity and resolution capacity, while allows a high-throughput analysis and absolute quantification (Chevillet et al., 2014). RT-qPCR is also important for the characterization of sEVs after transcriptomics. miRNAs have been described either as biomarkers and targets for modulation in several pathologies, including the neurodegenerative diseases (Brites, 2020), which turn these EVs as important tools for disease diagnosis and better understand the disease spread. Recent studies reported other forms of DNA (e.g., genomic and mitochondrial DNA) inside sEVs (Thakur et al., 2014; Sansone et al., 2017), although the exact role of them and the underlying mechanisms are still to be elucidated. Preliminary studies relate mitochondrial DNA as a biomarker of cancer aggressiveness (Arance et al., 2021) and its depletion in sEVs from astrocytes was associated with cell malfunction and spread of pathology (Ha et al., 2021).

The therapeutic agents (drugs, proteins, and nucleic acids) can be incorporated into cells that encapsulate the material during the production of sEVs. This strategy is more limited, since it depends on the cellular processes that are responsible for the production of sEVs as well as the cargo loading of the molecules. It also requires that the cellular source (cell line, primary culture) used is compatible and retains viability during the production of sEVs. In this case, after incubation with the therapeutic agents, the donor cells secrete sEVs that will be loaded with both biologically produced components and/or the drug. Such an approach requires that the donor cell can tolerate high concentrations of the drugs in order to be effective (Pessina et al., 2011). As so, it is also difficult to control the loading efficiency of the therapeutic molecule, which may suffer degradation in the host cells. Nevertheless, a major advantage is to specifically target sEVs against disease mechanisms (Garcia-Manrique et al., 2018). Another approach is to transfect the donor cells by genetic engineering with the drug-encoding DNA, which can then be expressed and sorted into sEVs. sEVs full of proteins encoded by the inserted genes can be obtained by isolating and purifying the sEVs through the natural packaging processes. Some disadvantages in this case are that the drug entrance is limited for its encoding DNA yield, depends on the transfection efficiency, and the transfection agents may lead to a reduction of cell viability (Johnsen et al., 2014; Batrakova and Kim, 2015).

The simplest strategy for loading therapeutic agents into sEVs is mixing them with the free drugs, as first demonstrated by Pascucci and others in 2014. Paclitaxel was selectively taken up by MSCs and then incorporated into the released sEVs, at a concentration sufficient to inhibit the growth of tumor cells in vitro (Pascucci et al., 2014). The process can also be used to load sEVs with small molecules. For cargo loading after isolation, sEVs are isolated and purified from biological fluids, such as cell culture media, plasma, serum, or milk for the downstream processing. Several methods have been proposed to encapsulate therapeutic agents in sEVs with the common goal of permeating its membrane. These techniques include freeze-thaw cycles, saponin permeabilization, sonication, extrusion, and electroporation procedure. The process of loading should allow the maintenance of the structure and activity of the drugs.

In the method of freeze-thaw cycles, drugs are incubated with sEVs at room temperature for a fixed amount of time, and the mixture is subsequently frozen at −80°C or in liquid nitrogen, and re-thawed at room temperature. This process is repeated for at least three cycles to ensure drug encapsulation. However, the method can induce aggregation of the sEVs, while the drug loading efficiency is generally lower than that of sonication or extrusion methods. sEVs are shown to resist multiple freeze-thaw cycles and are stable when stored from −20°C to −80°C and when subjected to multiple freeze-thaw cycles (Haney et al., 2015; Sato et al., 2016).

Another way to encapsulate drugs in sEVs is to perform a permeabilization with saponin. Saponin, as a natural surfactant, interacts with membrane-bound cholesterol, creating pores that increase the permeability of sEV membrane, thus favoring the encapsulation (Chen et al., 2021). The method appears to not be affected by sEV morphology. However, such an approach is described to have poor stability in vivo, as well as low encapsulation yield when saponin is added (Fuhrmann et al., 2015; Haney et al., 2015).

Sonication consists of a method where sEVs derived from donor cells are mixed with drugs or proteins using a sonicator probe. The mechanical shear force from the sonicator probe induces deformation, compromises the membrane integrity of the sEVs, and allows the drug to flow into the sEVs. Nevertheless, in some cases, drugs can also be found in the outer layer of the sEV membrane. There are no significant changes in the structure and content of their membranes after sonication, and the drug sEV formulation has shown to be kept stable under various conditions for over a month (Kim et al., 2016; Liu et al., 2019).

Another method is extrusion. In this method, sEVs from donor cells are mixed with a drug, and the mixture is loaded into a syringe-based lipid extruder and extruded through membranes with 100–400 nm porous size, under a controlled temperature. During the extrusion, the sEV membrane is disrupted and vigorously mixed with the drug, resulting in drug loading into the sEVs. However, the use of this method can lead to changes in EV size, composition, and delivery capacity (Fuhrmann et al., 2015; Haney et al., 2015).

Finally, in electroporation, an electrical field (short, high-voltage pulses) is applied to a suspension of sEVs and the drug cargo of choice, creating small pores in their membrane, thereby facilitating the passage of cargo into the lumen of the sEVs. The integrity of the vesicle membrane is then recovered, resulting in the formation of drug-loaded vesicles. Electroporating a mixture of drug and sEVs at 1000 kV for 5 min was demonstrated to successfully load the drug into the sEVs (Kim et al., 2016; O'Loughlin et al., 2017).

Lately, efficient delivery of siRNAs and miRNAs into sEVs was shown to be achieved with commercial transfection kits, as the Exo-Fect^TM from System Biosciences (SBI; System Biosciences, Palo Alto, CA) (Chivero et al., 2020). It was demonstrated to be more efficient than electroporation, heat shock, saponin, or cholesterol-mediated, at least for miRNAs, where more than 1000-fold upregulation was achieved, as compared to native sEVs (de Abreu et al., 2021).

sEVs are stable and long-circulating endogenous nanocarriers that provide protection of the drug cargo from degradation, while increasing drug delivery to the targeted tissues. One of their major advantages is that they can cross the BBB, thus penetrating into the CNS. On the other hand, some disadvantages may arise from the fact that some components carried by natural sEVs are incompatible with therapeutic purposes, with difficulties in the purification methods and also that cells release relatively low quantities of sEVs (Zhang Y. et al., 2019). In this context, liposomes may be an alternative, as they are small artificial vesicles of spherical shape that can be created from cholesterol and natural non-toxic phospholipids. Due to their size and unique structure, liposomes can compartmentalize and solubilize both hydrophilic and hydrophobic materials. However, liposomes also have some limitations due to their poor stability under shelf and in vivo conditions (Akbarzadeh et al., 2013). Modified sEVs with other nanostructures, such as liposomes, may be considered a new outlook of biological drug delivery systems. They are inspired in EVs but represent novel biological nanocarriers. sEV-mimetic nanovesicles contain only the crucial components of natural EVs and are highly biocompatible with efficient targeting ability (Garcia-Manrique et al., 2018). These EVs are obtained by: i) subjecting cells to physical processes producing EVs of nano-dimensions; ii) fusing the membranes of sEVs and liposomes; and iii) coating nanoparticles with a lipid bilayer of cell plasma membranes (Luan et al., 2017; Colao et al., 2018), as detailed in this section.

Significant efforts have been made to deliver small therapeutic molecules/drugs and diagnostic agents into the brain. The discovery of different types of EV cargo and their ability to interact and be taken up by specific cells has led investigators to concentrate on the potential of sEVs as delivery vehicles for therapeutic applications (Luan et al., 2017; Yamashita et al., 2018). sEVs have an efficient capability to cross the BBB, to deliver their intact content into a specific target cell, while also remain stable in the peripheral circulation, thus contributing to be considered potential attractive nanocarriers in the treatment of several CNS diseases (Alvarez-Erviti et al., 2011; Kojima et al., 2018; Busatto et al., 2021). A study carried out in 2016 reported the transfer of a tight junction protein from the endothelial cells to the leukocytes through EVs, thus supporting their cell-to-cell migration (Paul et al., 2016). In another work using rats with a fluorescently tagged protein expressed selectively in the brain tissue, the authors recovered the labeled sEVs in the blood of those animals and demonstrated that sEVs crossed the BBB in a bi-directional manner (Gomez-Molina et al., 2019). Chen et al. (2016) also showed that sEVs are internalized by brain microvascular endothelial cells through endocytosis, by using confocal microscopy and co-localizing sEVs with endosomes (Chen et al., 2016). In addition, sEVs are internalized by target cells through a variety of endocytic pathways. In particular, they use endogenous receptors that are highly expressed at the surface of the BBB, such as transferrin and insulin receptors, confirming sEVs as having a key role in intercellular communications and passage through the BBB in both directions (Yamashita et al., 2018). However, little is still known about the mechanistic details of sEV migration across the BBB and further investigation is required.

The content of sEVs can reflect the cell of origin and depends on the physiological or pathological conditions during their formation, suggesting horizontal transfer of genetic information (Zhang Y. et al., 2019). However, sEVs may also be enriched with a particular set of proteins and microRNAs (miRNAs), through selective mechanisms of protein cargo sorting controlled by specific post-translational modifications (PTMs) (van Niel et al., 2006). miRNAs are short nucleotide sequences of non-coding RNAs and can regulate gene expression in various cell types, constituting a critical factor in intracellular communication between origin and target cells (Simpson et al., 2009; Hessvik and Llorente, 2018; Zhang Y. et al., 2019). In the nervous system, a range of cell types have been shown to release miRNA-containing sEVs, including Schwann cells, microglia, oligodendrocytes, astrocytes, and neurons. For instance, miR-21, usually implicated in the microglial anti-inflammatory response, is transferred from neurons to microglia in a process that is mediated by sEVs (Fernandes et al., 2018). On the other hand, the most abundant miRNA of the CNS, the miR-124, is released from neurons as an sEV cargo and is taken up by astrocytes (Morel et al., 2013). Another study proposed that the delivery of miR-155 from microglia to adjacent cells may be mediated by sEVs (Cunha et al., 2016). In this study, the authors found that the expression profile of inflammation-associated miRNAs in sEVs recapitulated those in the cells on exposure to lipopolysaccharide (LPS), a pro-inflammatory stimulus. Increased expression of miR-21 released by Schwann cells in sEVs constitutes an important feature of the repair program of these cells, contributing to axonal regeneration and functional recovery after nerve injury (Lopez-Leal et al., 2020). miRNA profiles identified in sEVs of the CSF/blood in neurodegenerative disorders reinforce the important role of specific sEVs-miRNAs in the regulation of the neuroinflammatory processes (Ciregia et al., 2017; Yoo et al., 2018). In particular, we recently showed that increased levels of miR-124 in neurons derived from induced pluripotent cells generated from AD patients reduces amyloid precursor protein (APP) gene expression, tau hyperphosphorylation, and prevents dendritic spine deterioration (Garcia et al., 2021). In contrast, upregulation of miR-124 in SOD1G93A mutant motor neurons as a model of amyotrophic lateral sclerosis (ALS) evidenced to associate with neurodegeneration and homeostatic cell imbalance (Vaz et al., 2021). Therefore, miRNA sorting in sEVs may either work as a protective mechanism or be a sign of injury to neighboring cells and is dictated by the cell’s own metabolism. Therefore, these sEVs-containing miRNAs are cell specific and may be involved in many biological and pathological processes, as further detailed in the next section. All these factors should be considered when choosing the cell source of sEVs to be used as therapeutic carriers.

To date, cell-derived EVs provide multiple advantages over traditional synthetic delivery vehicles and potential new drug delivery methods, such as limited immunogenicity, natural composition, small size (nanoscale), enhanced stability in circulation, slightly negative zeta potential for long circulation, and deformable cytoskeleton. As drug delivery carriers, sEVs have the ability to cross many biological barriers, such as the BBB (Elliott and He, 2021; Yamashita et al., 2018). The main advantages of sEVs to be used as drug delivery systems in neurological diseases were addressed in previous sections and are indicated in Figure 3. To note, however, that the low residence time of sEVs in the circulation because of their rapid clearance by phagocytosis may compromise their efficiency and the generation of invisible sEVs has been attempted (Parada et al., 2021). We next will summarize the application that sEVs may have to prevent, halt, or even regenerate neural cell malfunction in age-associated neurodegenerative diseases and brain tumors.