EVs have garnered considerable attention as a mechanism of intercellular communication, and as candidates for therapeutic development as cell-free therapies [26]. In this study, we investigated the potential of hUCMSCs as a source of therapeutic EVs, and provided an in-depth evaluation of the influence of culture and harvest conditions on the final product characteristics, including physical parameters and proteomic profiles.
Compared with other tissue sources of MSCs, hUCMSCs have many advantages, such as the high availability of cord tissue, the high proliferative profile of cells, and their history of successful clinical translation, indicating safety and potential for beneficial therapeutic effects for different diseases [27–29]. As the proof of concept, the safety and potential efficacy of hUCMSCs have been extensively studied in patients with severe Coronavirus disease-19 (COVID-19) in previous studies, including our own work [18, 29, 30]. The same hUCMSCs utilized as the source of EVs in here have been previously utilized to treat a patient with advanced critical COVID-19, showing significant immunomodulatory effects [18]. Thus, we suggest that switching from the therapeutic use of MSCs to their EVs could lead to a superior safety profile, and provide several advantages in terms of logistics, as EV products can be safely stored without significant loss of function; therefore, they can be made available as an off the shelf medicinal product.
A major requirement in the field of therapeutic EVs is the realization of optimal conditions for large-scale manufacturing with high productivity and product lot-to-lot consistency [31]. The use of repeated harvesting protocols is highly desirable to enhance the yield of EVs manufactured in each lot. Considering the possible variability related to the culture harvesting time in the characteristics of the purified EVs, we compared the characteristics of the EVs obtained 24, 48, and 72 h after the introduction of the EV collection medium. The harvest points studied here are within the range reported in the literature, usually comprising different intervals ranging from 12 h to 7 days [32]. The harvesting time is also influenced by the cell seeding density and prolonged culturing under starvation conditions, which could lead to the loss of viability and an increase in apoptotic bodies during EV preparation. Using the protocol, we depicted consistent production of hUCMSC-EVs for 72 h, with a progressive increase in productivity and yield, while also maintaining cell viability. Similar results regarding the EV productivity have recently been reported by others using different cell sources [33].
Regardless of the harvesting timepoints, the data indicate that hUCMSC-EVs were successfully isolated from hUCMSCs and met the minimum criteria to be classified as EVs [7]. hUCMSC-EVs showed a rounded cup-like morphology with an average diameter of < 200 nm at all evaluated timepoints; however, reduced diameters of hUCMSC-EVs were observed at 48 and 72 h compared with those at 24 h. Immunophenotyping and proteomics showed positive expression of tetraspanins (CD63, CD9, CD81, and CD82) and markers of the origin of MSCs (CD44 and CD90), independent of the harvesting time [33, 34]. Proteomic data of hUCMSC-EVs revealed a small set of differentially expressed proteins at 24 h compared with that at 48 h and 72 h, which were remarkably similar. Two clusters of proteins enriched at 24 h were related to phagocytosis, positive regulation of B cell activation, innate immune response, and complement activation. The possible influence of these small differences on the safety, and therapeutic profile were not evaluated as we decided to pool the different harvests as recommended by the EVOLVE guidelines for the preclinical biodistribution and safety analyses [35].
The analysis of proteomic data in hUCMSC-EVs indicated the presence of proteins involved in multiple cellular pathways relevant to health and disease. The GO analysis revealed that most proteins found in hUCMSC-EVs were components of the EVs, plasma membrane, cytoplasm, lysosomes, ribosomes, and cytoskeleton, and were involved in signal transduction, energy metabolism, innate immune response, and several other BPs. The analysis of bioinformatic data identified the abundance of proteins that are involved in several processes, such as angiogenesis, immune response, response to mechanical stimulus, wound healing, cell differentiation, and response to cytokine cell-matrix adhesion. Fibronectin 1 (FN1) facilitates cell migration through tissues [36], and exosomal FN1 mediates the mitogenic activity of MSC-derived exosomes [37]. The main components of the glycolytic pathways, GAPDH and PKM, which are known to be exported by EVs, were also present, and are listed among the 10 most frequently identified proteins in EVs (http://microvesicles.org/) [24]. ANXA2 is a protein present in the cell composition of the membrane, extracellular exosome, and EVs. It is involved in many cellular processes, including membrane trafficking events [38]. The involvement of ANXA2 in the loading of RNA into exosomes has been experimentally demonstrated [39]. The extracellular matrix protein, LGALS3BP, is known to interact with a set of membrane molecules, such as integrins, fibronectin, laminins, collagen, nidogen, and galectins [39]. Traditional complement proteins, such as C3, are an important component of the innate immune system, as complement activation results in the generation of activated protein fragments that play a role in inflammatory reactions, immune complex clearance, and antibody production [40]. Structural and cytoskeletal proteins ACTN1, filamins, such as filamin A, which act as scaffolds for various signaling molecules implicated in cell motility, transcription[41], and mechanical sensing, as well as the most abundant translational factor EEF1A1, which regulates the synthesis of proteins, translation machines, cell proliferation, and apoptosis, were also present in the sample [42, 43].
By analyzing protein associations, we identified 8 hubs comprising these proteins (e.g., FN1 and EEF1A1) and ACTB, COL1A1, HSP90AA1, PSMA3, PSMA7, and ITGB. β-Actin is a major cytoskeletal filament protein encoded by the actin beta (ACTB) gene, and is an important player in cell motility, migration, and gene expression [44]. Molecular chaperones, especially Hsp90, are an evolutionarily conserved class of proteins that assist normal folding, intracellular protein disposition, and proteolytic turnover of the key regulators of cell growth [45]. Lauwers et al. [46] demonstrated that Hsp90 facilitates the transport of multivesicular bodies toward the plasma membrane and enhances exosome secretion. PSMA3 and PSMA7 are proteasome subunits that belong to the 26S proteasome complex. This protease complex is part of the ubiquitin proteasome system, that is the principal proteolytic system responsible for the functional modification and degradation of cellular proteins and processes such as proliferation, growth, differentiation, gene transcription, signaling, and apoptosis [47, 48]. Proteasomes are also present and active in the extracellular compartments, including EVs [49]. Integrin subunit beta 1 (ITGB1) adheres to collagens, laminins, fibronectin, and other glycoproteins. Notably, ITGB1 plays a role in cell adhesion, cell-matrix adhesion and is abundant in exosomes with different origins [50]. These hub proteins are associated with effector immune mechanisms, transport, localization, stress response, cellular activation, and response to stimulus.
The proteomic analysis identified proteins that have previously been enriched in both small EVs (e.g., CD9, CD81, CD63, annexins, ALIX, and aldolase A) and the non-vesicular fractions (GAPDH, PKM, HSP90, EEF2, PGK1, and clathrin) [24]. Although a significant overlap of protein content has been reported in small EVs and non-vesicle fractions, the results suggest that our samples included a mixture of small EVs and non-vesicle components, which is expected for the protocols of isolation and purification used here. Furthermore, a mixture of exosomes and small microvesicles (CD81+/CD63+/CD9+, and annexin A1+) can be expected based on the analysis of protein content [24]. The nanoflow analysis presented here supports the frequency of classic exosomes estimated in approximately 20–30% of the EVs in the preparations. Quantitatively monitoring the frequency of exosome marker expression in each batch, rather than qualitative measurements, may be important to ensure the lot-to-lot consistency [24, 35].
Despite the increasing interest and the developments in the field of therapeutic EVs, only few studies have evaluated the biodistribution of EVs in-vivo, which is a critical step in preclinical development [51]. Our results demonstrated that when intravenously administered to naive mice, hUCMSC-EVs accumulated largely in the liver, spleen, and lungs, which is consistent with previous observations [52]. The identification of EVs in other organs and tissues has also been reported in protocols using higher doses, which could be related to the sensitivity of the methods [53]. EVs injected intravenously have been reported to be cleared by the reticuloendothelial system, and may influence the local or systemic processes of injury and inflammation [54]. Whether this mechanism of clearance by innate immune cells could be a part of the mechanisms of immune regulation promoted by MSC-EVs still requires further investigation.
In this study, the preclinical toxicology of hUCMSC-EVs was analyzed to identify possible adverse effects. Our results indicate that hUCMSC-EVs are safe in terms of the immune response and toxicity, after either a single systemic IV dose or even after 3- and 6 weeks-long repeated administrations of hUCMSC-EVs. This result is consistent with previous reports, including a study wherein the repeated administration of hUCMSC-EVs showed no signs of immunogenicity [25]. Previous studies evaluating the toxicity of EVs from different sources corroborate our data, demonstrating their safety after IV administration in a single dose or in repeated administrations, even at extremely high doses, such as 2 × 1012 EVs/200 µL/mouse, with no reports of acute or subacute toxicities in immunocompetent mice. Thus, through feasible doses, it is not possible to determine the level where adverse effects were observed [55–57]. This emphasized the increased safety profile of EVs compared with MSCs, for which the maximum tolerated dose has been demonstrated, which is related to their pro-thromboembolic activity at high doses [58].
Immune toxicity may occur with biological medicines (e.g., monoclonal antibodies and advanced therapy products), and must therefore be evaluated as part of the preclinical toxicity evaluation of EV products, especially considering that EVs target immune cells as part of their mechanism of action [59, 60]. Our data show that no significant differences were observed between the groups (Plasma-Lyte and hUCMSC-EVs) for any of the tested immune cell populations, indicating that treatment with hUCMSC-EVs did not alter the composition of the immune cells in the spleen. Our results are in accordance with a previous study that demonstrated neither toxicity nor induction of an immune response in immunocompetent mice after repeated administrations of HEK293-derived EVs [61]. Our biodistribution data show that circulating hUCMSC-EVs were present in the spleen 24 h after IV administration. These interactions with distinct spleen cell populations were expected to trigger differential physiological responses and alter local signaling at the autocrine and paracrine levels.