Could Mesenchymal Stem Cell-Derived Exosomes Be a Therapeutic Option for Critically Ill COVID-19 Patients?

Coronavirus disease 2019 (COVID-19) is a pandemic viral disease originated in Wuhan, China, in December 2019, caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The severe form of the disease is often associated with acute respiratory distress syndrome (ARDS), and most critically ill patients require mechanical ventilation and support in intensive care units. A significant portion of COVID-19 patients also develop complications of the cardiovascular system, primarily acute myocardial injury, arrhythmia, or heart failure. To date, no specific antiviral therapy is available for patients with SARS-CoV-2 infection. Exosomes derived from mesenchymal stem cells (MSCs) are being explored for the management of a number of diseases that currently have limited or no therapeutic options, thanks to their anti-inflammatory, immunomodulatory, and pro-angiogenic properties. Here, we briefly introduce the pathogenesis of SARS-CoV-2 and its implications in the heart and lungs. Next, we describe some of the most significant clinical evidence of the successful use of MSC-derived exosomes in animal models of lung and heart injuries, which might strengthen our hypothesis in terms of their utility for also treating critically ill COVID-19 patients.


Introduction
The novel coronavirus 2019 (2019-nCoV) has reached pandemic proportions across the world after originating in Wuhan, the capital of China's Hubei province, in December 2019 [1,2]. Initially called 2019-nCoV, the World Health Organization (WHO) subsequently adopted the official name severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) for indicating the virus, and the term coronavirus disease 2019 (COVID-19) for identifying the virus-associated disease [3]. As of 7 May 2020, the WHO has reported almost 3,634,172 confirmed cases of COVID-19 with 251,446 confirmed deaths in 215 countries/areas/territories worldwide [4].
The clinical spectrum of COVID-19 is highly variable-in addition to mild, severe, and critical forms, asymptomatic or paucisymptomatic infections have been described as well [5,6]. Milder clinical conditions are commonly characterized by fever, dry cough, myalgia or fatigue, headache, and mild pneumonia, whereas the severe form of the disease is associated with dyspnea, acute respiratory distress syndrome (ARDS), and hypoxemia (low level of oxygen in arterial blood) [1,7]. The most critical cases experience respiratory failure requiring mechanical ventilation and support in the intensive care mice was correlated with myocardial disfunction [46]. Considering the similarity with SARS-CoV, it has been speculated that SARS-CoV-2 infection might also down-regulate ACE2 expression in lung and heart, thus leading to the pathological processes of lung and cardiac injuries [27]. To date, no experimental or clinical data have evidenced that using ACEI/ARB therapy makes patients more susceptible to the virus. Therefore, several leading cardiovascular societies have strongly urged to not discontinue intake of RAS inhibitors in the event the patient develops COVID-19 [47,48]. J. Clin. Med. 2020, 9, x FOR PEER REVIEW 4 of 28 regulated after SARS-CoV infection in mice, and this was accompanied by increased pulmonary vascular permeability and pulmonary edema [45]. Similarly, cardiac ACE2 down-regulation following SARS-CoV infection in mice was correlated with myocardial disfunction [46]. Considering the similarity with SARS-CoV, it has been speculated that SARS-CoV-2 infection might also downregulate ACE2 expression in lung and heart, thus leading to the pathological processes of lung and cardiac injuries [27]. To date, no experimental or clinical data have evidenced that using ACEI/ARB therapy makes patients more susceptible to the virus. Therefore, several leading cardiovascular societies have strongly urged to not discontinue intake of RAS inhibitors in the event the patient develops COVID-19 [47,48]. , angiotensin (1-7); AT1R, ang II type 1 receptor; ACEIs, angiotensin-converting enzyme inhibitors; ARBs, angiotensin receptor blockers; ARDS, acute respiratory distress syndrome. Green arrows indicate that ACEIs/ARBs increase ACE2 levels in the heart, therefore increasing the susceptibility of cardiac cells to SARS-CoV-2 infection [42,43]. Blue dotted hammerhead indicates the hypothetical effect of SARS-CoV-2 on ACE2 expression in lung and heart, which is based on the reported effect of SARS-CoV in the same body districts [45,46].

Anti-Inflammatory and Immunomodulatory Properties of MSCs and MSC Secretome
MSCs are thought to prevent or reduce the cytokine storm in COVID-19 patients, owing to their powerful anti-inflammatory and immunomodulatory functions [8]. MSCs exert these effects by directly interacting with different cells of innate and adaptive immunity, including T cells, B cells, dendritic cells (DCs), macrophages, and natural killer cells, and by indirectly releasing many types of inflammatory mediators by paracrine secretion [49][50][51][52]. Many studies have described a differential regulation by MSCs on the different T cell subsets [53][54][55][56][57]. MSCs inhibit effector T (Teff) cell proliferation induced by mitogens or alloantigens by causing cell cycle arrest at the G1 phase [53,58]. Another explanation for this immunosuppressive capacity is the loss of CD25, the alpha-chain of the IL-2 receptor, which is cleaved from the activated T cell surface by MSC-secreted matrix metalloproteinases [59]. This leads to blockage of the IL-2 cytokine signaling pathway required for T cells activation, expansion, and differentiation. Interestingly, such T cell-suppressing properties of MSCs seem to require the presence of inflammatory cytokines in the microenvironment, which provoke the production of several T cell-attracting chemokines and inducible nitric oxide synthase Figure 1. Schematic diagram showing the renin-angiotensin system (RAS) cascade and the effects on the cardiovascular system and lung. ACE1, angiotensin-converting enzyme 1; ACE2, angiotensin-converting enzyme 2; ang I, angiotensin I; ang II, angiotensin II; ang (1-7), angiotensin (1-7); AT1R, ang II type 1 receptor; ACEIs, angiotensin-converting enzyme inhibitors; ARBs, angiotensin receptor blockers; ARDS, acute respiratory distress syndrome. Green arrows indicate that ACEIs/ARBs increase ACE2 levels in the heart, therefore increasing the susceptibility of cardiac cells to SARS-CoV-2 infection [42,43]. Blue dotted hammerhead indicates the hypothetical effect of SARS-CoV-2 on ACE2 expression in lung and heart, which is based on the reported effect of SARS-CoV in the same body districts [45,46].

Anti-Inflammatory and Immunomodulatory Properties of MSCs and MSC Secretome
MSCs are thought to prevent or reduce the cytokine storm in COVID-19 patients, owing to their powerful anti-inflammatory and immunomodulatory functions [8]. MSCs exert these effects by directly interacting with different cells of innate and adaptive immunity, including T cells, B cells, dendritic cells (DCs), macrophages, and natural killer cells, and by indirectly releasing many types of inflammatory mediators by paracrine secretion [49][50][51][52]. Many studies have described a differential regulation by MSCs on the different T cell subsets [53][54][55][56][57]. MSCs inhibit effector T (Teff) cell proliferation induced by mitogens or alloantigens by causing cell cycle arrest at the G1 phase [53,58]. Another explanation for this immunosuppressive capacity is the loss of CD25, the alpha-chain of the IL-2 receptor, which is cleaved from the activated T cell surface by MSC-secreted matrix metalloproteinases [59]. This leads to blockage of the IL-2 cytokine signaling pathway required for T cells activation, expansion, and differentiation. Interestingly, such T cell-suppressing properties of MSCs seem to require the presence of inflammatory cytokines in the microenvironment, which provoke the production of several T cell-attracting chemokines and inducible nitric oxide synthase (iNOS) from MSCs, so that T cells migrate into proximity of these cells [60]. At the same time, MSCs have been shown to induce the survival and expansion of regulatory T (Treg) cells, a subset of T cells involved in the suppression of proliferation and cytokine production by Teff cells [56]. Therefore, Treg cells foster the MSC-mediated immunosuppressive effect. In addition to directly interacting with T cells, MSCs also modulate the adaptive immune response by acting on antigen-presenting cells (APCs), such as DCs, monocytes, and macrophages, by shifting them to regulatory phenotypes characterized by T cell-suppressive properties [61,62].
The spectrum of regulatory factors secreted by MSCs is collectively defined as the MSC secretome, and include a complex array of soluble molecules, such as anti-inflammatory cytokines, angiogenic growth factors, antimicrobial peptides, and lipid mediators. A growing body of evidence nowadays suggests that some of these molecules are packaged into cell-secreted vesicles, known as extracellular vesicles (EVs) [63][64][65]. Besides apoptotic bodies, the two main types of EVs released by MSCs include exosomes and microvesicles (MVs). Exosomes (30-100 nm) are derived by fusion of multi-vesicular bodies with the plasma membrane, whereas MVs (100-1000 nm) are formed by cellular membrane budding, and contain cellular cytoplasm. All these EVs are released into the extracellular microenvironment, where they exert biological effects in a paracrine and endocrine manner, similarly to the soluble components. For this reason, a broader definition of MSCs secretome encompasses the entire spectrum of bioactive factors secreted by MSCs, which consists of both the soluble and the extravesicular elements. It has now been demonstrated that MSCs are also able to transfer functional mitochondria or mitochondrial DNA (mtDNA) to target cells, thus rescuing aerobic respiration in cells with non-healthy mitochondria or regulating T cell functions [66,67]. Following systemic injection, some MSCs accumulate in the lung, where they release these soluble mediators, potentially recovering the pulmonary microenvironment, protecting alveolar epithelial cells, and counteracting pulmonary fibrosis, thus resulting in a final improvement of lung function [8]. Moreover, distant injured organs, such as the cardiovascular system, can also benefit from them, by virtue of the secretory abilities of these cells.
To date, two studies have investigated the employment of MSCs in severely affected SARS-CoV-2 patients, with both reporting remarkable reversal of symptoms within a few days [10,68]. In one of these works, the levels of biochemical indicators of liver and myocardium damage (aspartic aminotransferase, creatine kinase activity, and myoglobin) returned to reference levels 4 days after MSC treatment [10]. The authors demonstrated that the cells expressed high levels of anti-inflammatory and angiogenic factors, such as transforming growth factor-beta (TGF-β), hepatocyte growth factor (HGF), leukemia inhibitory factor (LIF), fibroblast growth factor (FGF), vascular endothelial growth factor (VEGF), epidermal growth factor (EGF), brain-derived neurotrophic factor (BDNF), and nerve growth factor (NGF), further demonstrating their potent immunomodulatory abilities.
It has been reported that MSCs are generally resistant to viral infections compared to their differentiated progeny, probably due to intrinsic expression of IFN-stimulated genes (ISGs) [69]. Among these genes, those coding for proteins of the interferon-induced transmembrane (IFITM) family prevent viruses from traversing the lipid bilayer of the cell and accessing the cytoplasm, thus impairing viral infection [70]. These antiviral proteins limit infection in cultured cells by many viruses, including SARS-CoV, dengue virus, Ebola virus, influenza A virus, and West Nile virus. However, some studies have reported that human MSCs are permissive to other viruses, for example avian influenza viruses H1N1 and H9N5 and respiratory syncytial virus (RSV), losing vitality and compromising their immunomodulatory activities [71,72]. In the case of SARS-CoV-2, the advantage in using MSCs seems to be additionally related to the absence of ACE2 receptors on the cell surface, which precludes their recognition by the virus. Notably, in the study of Leng and coworkers, the cells remained negative for ACE2 also after transplantation in infected patients [10].

MSC-Derived EVs as a Therapeutic Pption for Critically Ill COVID-19 Patients
Although MSCs seem to be refractory to SARS-CoV-2 infection, in order to bypass the impact of viruses on MSCs, an interesting therapeutic strategy could consider the use of the MSC secretome. Among the bioactive factors released by MSCs, EVs, exosomes in particular, have gained remarkable interest in recent years because they enable more efficient communication and targeting than soluble molecules [73]. EVs, by virtue of their lipid bilayer membrane, better protect their molecular cargo of proteins and genetic material from environmental degradation (i.e., from trypsin or nuclease digestion) when compared to soluble molecules. Encapsulation within EVs may also facilitate delivery and targeting of these bioactive factors to distant recipient cells, mediated by binding of the EV surface proteins to cells that express appropriate receptors [74].
MSC-derived exosomes offer several advantages over traditional cell-based therapies. First, exosomes are considered safer than cells, because they are biocompatible, non-immunogenic, and lack the potential for endogenous tumors and emboli formation [75]. In addition, exosomes are physiologically more stable than cells, because their multiple membrane adhesion proteins allow for efficient binding in the target tissues during transplantation. Thanks to their resistant membrane, exosomes maintain their integrity during freezing and thawing procedures, making long-term storage without biological degradation possible [76]. In this context, a process has recently been proposed that combines ultrafiltration and lyophilization and is able to convert MSC secretome into a freeze-dried, ready-to-use powder [77,78]. The same research group also suggested the possibility to administer EVs by inhalation in the treatment of respiratory diseases [79]. This route of administration would benefit from lower invasiveness and pain, faster onset of action, and use of lower doses to achieve the same therapeutic effect when compared to oral or injection therapies. In this regard, a pilot clinical trial (NCT04276987) will be conducted in China for exploring the safety and efficiency of aerosol inhalation of MSC-derived exosomes in comparison to conventional treatment in 30 severe patients with COVID-19.
Another advantage of MSC-derived exosomes over whole-cell therapy is that, to improve their therapeutic potential, exosomes could potentially be modified with various types of cargos, including mRNA, microRNA (miRNA), and proteins, tailored to the disease process of interest [73]. In one pioneering work, exosomes incorporating the S protein have been explored as a novel vaccine approach against SARS-CoV infections [80]. The immunogenicity and efficacy of the S-containing exosomes were tested in mice, where they induced neutralizing antibody titers. Finally, from an economical point of view, MSC-derived exosome therapy might enable development of cheaper treatments other than the expansion and maintenance of individualized clonal cell populations [73]. This aspect is particularly important when a global pandemic has to be managed, as in the case of COVID-19. In Section 4, we provide an overview of the currently available evidence on the effects of MSC-derived exosomes in pre-clinical models of lung and heart injuries, which are the body districts most affected by SARS-CoV-2.

MSC-Derived EVs from Patients with Metabolic Disorders
Diabetes mellitus (DM) represents the most common inflammatory and chronic metabolic disorder worldwide, and continues to increase in number and significance-it is estimated that there will be 693 million persons with DM by 2045 [81]. Type 2 diabetes mellitus (T2DM) accounts for 90-95% of all cases of diabetes and results from a progressive defect in insulin production and insensitive response of the body to insulin [82]. Accumulating evidence shows that such a state of insulin resistance (IR) is closely related with obesity [83]. Obesity, mainly visceral adiposity, is, indeed, one of the most important comorbidities in diabetic patients.
People with diabetes have a higher overall risk of infections that result from compromised innate cell-mediated immunity; impaired phagocytosis by neutrophils, macrophages, and monocytes; and impaired neutrophil chemotaxis and bactericidal activity [84]. Regarding COVID-19, it is currently unknown whether patients with diabetes have a higher susceptibility to the virus; nonetheless, there is evidence of higher risk for both infection and disease severity [85].
As stated above, there is growing interest in the use of MSC-derived EVs as a therapeutic tool for the management of several diseases. However, because EV cargo usually reflects parent cell characteristics, and these are influenced by the metabolic state of source cells, it is reasonable to consider the risks associated with the employment of MSC-derived EVs from patients with coexisting metabolic disorders such as T2DM. In effect, clinical studies have found differences in the number and composition of EVs isolated from the adipose tissue of obese patients and from animal models of obesity [86][87][88][89]. For example, MSC EVs isolated from a swine model of metabolic disorder were found to be enriched with mRNAs associated with inflammation, such as those coding for the integrin family proteins, or proteins of the FGF signaling [87]. These MSC-derived EVs also showed a distinctive miRNAs cargo, being enriched in miRNA-targeting genes involved in the development of metabolic disease and its complications, including diabetes, obesity, and insulin signaling [88]. Apart from influencing the mRNA and miRNA content, metabolic disorder also alters packaging of proteins into porcine MSC-derived EVs, promoting the inclusion of pro-inflammatory proteins, such as those involved in acute inflammatory response, cytokine production, and leukocyte transendothelial migration [86]. The limitations of these works reside in the small sample size and short duration of metabolic disease compared to the human condition; therefore, further studies would be needed to draw clear conclusions. However, in humans also, analysis of adipose tissue-derived EVs demonstrated that obesity alters their cargo of mRNAs, miRNAs, and proteins [90]. In particular, the differentially expressed miRNAs contained in the isolated EVs stimulated up-regulation of Wnt/β-catenin and TGF-β signaling pathways, which are related to inflammation, into A549 lung epithelial cells.
Overall, these observations suggest that diabetes and metabolic disorders might alter the MSC-derived EV cargo, which in turn might compromise their anti-inflammatory and immunodulatory potential both in the endogenous microenvironment and after autologous transplantation.

MSC-Derived Exosomes and Respiratory Lung Injuries
Acute lung injury (ALI) and ARDS are major causes of respiratory failure in critically ill ventilated patients, with an estimated 60-day mortality rate of 32% [91]. ARDS is also one of the most common complications in severely affected COVID-19 patients. The term ARDS is often used interchangeably with ALI; nevertheless, ARDS should be reserved for the most severe form of the disease [92]. Bacterial or viral infections are the most common causes of ALI and ARDS; however, they can also be initiated by aspiration of gastric contents, toxic inhalation, lung contusion, or trauma [92]. The acute phase (the first 1-6 days) of the diseases is characterized by injury to both the pulmonary endothelium and the alveolar epithelium, the two barriers forming the alveolar-capillary barrier. In healthy lung microvessels, the pulmonary endothelium is maintained by vascular endothelial cadherin (VE-cadherin), an endothelial-specific adherens junction protein, whereas the alveolar epithelial barrier has E-cadherin junctions and is substantially less permeable than the endothelial counterpart [93,94]. During lung injury, VE-cadherin bonds are destabilized by increased expression of thrombin, TNF-α, VEGF, and signals from leukocytes. At the same time, E-cadherin epithelial junctions are disrupted by neutrophil migration, which causes injury, apoptosis, and membrane denudation. This ultimately results in increased epithelial permeability, leading to accumulation of protein-rich edema fluid in the alveoli, and in turn to an impairment in gas exchange and to hypoxemia [95]. Dysregulated immune activation has also been implicated in the pathogenesis of ALI/ARDS. In the air space, macrophages release pro-inflammatory cytokines and chemokines, which act locally to stimulate chemotaxis and activate neutrophils. Activation of neutrophils leads to the release of numerous cytotoxic products, such as reactive oxygen species, cationic peptides, eicosanoids, and proteolytic enzymes, which may further damage the alveolar epithelium [96].
Resolution of ALI/ARDS aims at removing alveolar edema fluid, repairing the epithelial and endothelial barriers, and removing inflammatory cells and exudate from the air spaces [95]. To date, management of ALI/ARDS includes lung protective ventilation, prone positioning, neuromuscular blockade, and extracorporeal membrane oxygenation. Mechanical ventilation represents the mainstay treatment in ALI/ARDS, and consists in the application of positive-end expiratory pressure for optimizing arterial oxygenation. It has been evidenced that ventilation with a low tidal volume (6 mL/kg) gives better results when compared to traditional tidal volume (12 mL/kg) [97]. Indeed, the use of lower tidal volumes during ventilation may reduce injurious lung stretch and the release of inflammatory mediators. Prone positioning enhances arterial oxygenation by improving alveolar ventilation/perfusion matching. Nevertheless, this treatment should be used with caution and should be reserved for patients with critical hypoxemia, since it does not improve survival or decrease the duration of lung ventilation. All these therapeutic options remain primarily supportive; on the other hand, alternative treatments with glucocorticoids, surfactants, inhaled nitric oxide, antioxidants, protease inhibitors, or other anti-inflammatory agents had proven unsuccessful in reducing mortality or improving ALI/ARDS outcomes [96].
In terms of promising novel strategies, MSC-based approaches have been explored for the management of ALI/ARDS. The benefit of MSC therapy appears to be related to a decrease in pro-inflammatory cytokines and to an increase in anti-inflammatory cytokines, particularly IL-10 [98]. MSCs release prostaglandin E 2 , which in turn stimulates secretion of IL-10 by monocytes and alveolar macrophages [99]. Moreover, administration of MSCs seems to be effective in normalizing lung endothelial and epithelial permeability to protein, as well as in reducing pulmonary edema and increasing the rate of alveolar fluid clearance [100].
Recently, MSC-derived exosomes have been demonstrated to have comparable and even greater effects than cells themselves in improving inflammation and injury in a variety of pre-clinical lung disease models, including ALI/ARDS (Table 1). For the completeness of information, we have to specify that some of these works take into account the entire spectrum of EVs that, in addition to exosomes, also includes MVs. This is because as of yet there are no standardized methods for isolation, quantification, and characterization of EVs, or for discriminating MVs and exosomes. Consequently, in the majority of these pre-clinical studies, EVs, exosomes, and MVs are collectively referred to as EVs.
MSC-derived EVs have been proven to be beneficial in both bacteria-and virus-induced ALI/ARDS. A large number of studies have employed an endotoxin-mediated in vivo model to investigate the effects of MSC-derived EVs for ALI/ARDS. In one of the first works, ALI was induced in C57BL/mice using the intratracheal (IT) instillation of endotoxin (4 mg/kg) from Escherichia coli (E. coli) [101]. MVs were isolated from the conditioned medium of human bone marrow-derived MSCs with two sequential ultracentrifugations at 100,000× g for 1 h. Then, 30 uL of MVs, corresponding to the vesicles released by 3 × 10 6 MSCs, were administrated intratracheally or intravenously in mice. After 48 h, MSC-derived MVs reduced lung inflammation and reduced edema to the same levels as MSCs themselves, which were used as a positive control. Furthermore, MVs also decreased the influx of neutrophils and MIP-2 levels in the alveolar fluid, indicating a reduction in inflammation. Surprisingly, the therapeutic effects of the MVs were comparable, regardless of route of administration. The authors suggested that the mechanism underlying the therapeutic effect of MVs might be in part mediated by the transfer of keratinocyte growth factor (KGF) mRNA into the injured alveolar epithelium, with subsequent expression of the protein. KGF is an epithelial-specific growth factor released from MSCs, which has been shown to reduce lung edema and inflammation in various ALI models [102][103][104]. In the same study, the effect of MVs was additionally evaluated in RAW 264.7 cells, a mouse macrophage cell line. Treatment with 30 µL of MSC-derived MVs to endotoxin-stimulated RAW 264.7 cells reduced the levels of TNF-α and MIP-2, and concomitantly increased the production of the anti-inflammatory cytokine IL-10 at 6, 12, and 24 h compared with endotoxin-stimulated mouse macrophages [101]. In the work of Tang and colleagues, ALI was induced in C57BL/mice by the instillation of lipopolysaccharide (LPS) from Pseudomonas aeruginosa at 4 mg/kg intratracheally [105]. MVs released from human bone marrow MSCs were isolated by two sequential ultracentrifugations at 100,000× g for 1 h, then intratracheally administrated in endotoxin-injured mice. IT administration of MSC-derived MVs improved the lung inflammation induced by LPS in mice, including the influx of white blood cells and neutrophils, and MIP-2 secretion. In that study, the authors found that the transfer of angiopoietin-1 (Ang-1) mRNA by MVs was essential for the reduction of inflammation and the restoration of alveolar-capillary barrier. Ang-1 plays a key role in vascular stabilization, since it reduces endothelial permeability and suppresses leukocyte-endothelium interactions [106]. Furthermore, MSC-derived MVs showed immunomodulatory effects on RAW 264.7 cells in vitro by inhibiting TNF-α mRNA production and promoting the mRNA levels of IL-10 after 3 h [105]. Collectively, these two studies suggest that the beneficial immunomodulatory effect of MSC-derived MVs in ALI is strongly dependent on KGF and Ang-1 mRNA transfer into injured endothelial cells.
It has been demonstrated that, apart from mRNAs, the therapeutic effect of EVs is also mediated by the transfer of functional mitochondria to target cells. MSCs have been reported to naturally transfer mitochondria to recipient cells through different mechanisms-incorporated within EVs, via cell-to-cell contact through tunneling nanotubes, or through direct release of naked mitochondria into the extracellular microenvironment [107,108]. In the work of Phinney and colleagues, transmission electron microscopy images evidenced structures consistent with the morphology of mitochondria inside MVs over 100 nm in size, previously isolated from the conditioned medium of human MSCs after centrifugation at 10,000× g for 1 h [109]. The authors found that these mitochondria were loaded in the cytoplasm into LC3-containing MVs, which migrated towards the cell periphery and were incorporated into outward budding blebs in the plasma membrane. The MSC-derived MVs contained functionally active mitochondria that were taken up by macrophages and resulted in improved bioenergetics after oxidative stress increment. In particular, the transfer of human MSC-derived mitochondria involved fusion with mitochondria inside macrophages, suggesting that the mitochondrial membrane was not collapsed. The same study also confirmed that mitochondria were not packaged within exosomes; rather, exosomes were able to deliver mtDNA, which in mammals has an average size under 100 nm [110]. In the context of lung injury, the group of Morrison and colleagues revealed that MSC-derived EVs protected mice against LPS-induced ALI by altering alveolar macrophage (AM) polarization from the pro-inflammatory M1 phenotype towards the M2 anti-inflammatory phenotype [111]. In detail, EVs were obtained from human bone marrow MSCs after ultracentrifugation at 100,000× g for 2 h. These were used for pre-treating AM, which were then intranasally administrated to LPS-injured mice. The MSC-derived EVs increased phagocytic activity by macrophages and reduced their secretion of TNF-α and IL-8, two major pro-inflammatory cytokines related to ARDS severity [112,113]. The transfer of functional mitochondria contained in EVs, associated with the promotion of oxidative phosphorylation, was supposed to be the mechanism responsible for the observed effects in macrophages [111].
The work of Monsel and colleagues was the first to evaluate the effect of MSC-derived MVs in an infectious ALI model [114]. Indeed, the authors instilled live E. coli bacteria into the trachea; then, they intravenously administrated 90 uL of MVs, corresponding to the vesicles released by 9 × 10 6 human bone marrow MSCs. MV injection improved survival and reduced the bacterial load, as well as the influx of white blood cells, neutrophils, and MIP-2 levels, in the injured alveolus of C57BL/mice. The authors reported that the effect was in part mediated by KGF mRNA shuttled by the vesicles into target cells, as described in their previous study [101]. In addition to testing in mice, the effect of MSC-derived MVs was also investigated in human monocytes and AT2 cells. MV treatment increased the percentage of phagocytosis of human monocytes against E. coli bacteria, thus reducing the bacterial count, and decreased TNF-α secretion. Furthermore, MVs showed a beneficial effect on injured human AT2 cell metabolism through the restoration of intracellular ATP levels to control levels. In these primary in vitro cultures, the uptake of MVs was mediated by CD44, which was essential for the observed therapeutic effects. CD44 is the hyaluronic acid receptor expressed in almost every cell type including MSCs [51,115]. The results of this work suggested that MVs, similarly to their parent cells, act through different mechanisms on the basis of anti-inflammatory, anti-microbial, and metabolomic effects.
MSC-derived EVs have also shown reparative properties on microvascular endothelial and epithelial cells, which are often severely injured in the lung during ALI, and are associated with increased mortality in ARDS patients. Hu and coworkers investigated the effects of MVs isolated from human bone marrow MSCs on human lung microvascular endothelial cells (HLMVECs) in vitro [116]. The cells were injured by cytomix, a mixture of the most biologically active cytokines found in ALI pulmonary edema fluid (IL-1β, TNF-α, and IFN-γ at 50 ng/mL), and simultaneously exposed to increasing doses (30 or 60 uL) of MVs using a transwell co-culture system [117]. Administration of MSC-derived MVs restored protein permeability of HLMVECs by preventing the reorganization of cytoskeleton protein F-actin into "actin stress fibers" and the loss of tight and adherens junction proteins (zonula occludens-1 and VE-cadherin, respectively) following inflammatory injury. The internalization of MVs via cluster of differentiation (CD)44 receptor, as well as the subsequent transfer of Ang-1 mRNA into injured HLMVECs, were required for the observed therapeutic effects.
The study of Khatri and colleagues is interesting because ARDS was induced in pigs after infection with a mixed swine (H3N2, H1N1) and avian (H9N5, H7N2) influenza viruses (SwIV) [118]. Pigs are often used as large animal pre-clinical models for several human diseases, including respiratory diseases, due to their close similarity in anatomy, physiology, and immunology to humans [119]. In addition, influenza virus pathogenesis and clinical signs are similar to those observed in humans. In that work, EVs (80 µg/kg) isolated from swine bone marrow MSCs with two ultracentrifugation steps at 25,000 rpm for 70 min were intratracheally administrated in pigs 12 h after SwIV inoculation. MSC-derived EVs were found to inhibit influenza virus replication and shedding in pigs 3 days post-infection. As in other studies, EVs also modulated inflammatory cytokine and chemokine production in the lungs, as demonstrated by reduction in TNF-α and CXCL10 protein levels, and increase in IL-10 protein levels. Unfortunately, there are not yet pre-clinical data on the effects of MSC-derived EV administration in models of coronavirus respiratory infection, mostly due to the lack of an established animal model [120].
From the studies discussed above, it emerged that the rationale for using MSC-derived exosomes, MVs, or EVs in ALI/ARDS is based on several processes, many of which are shared with those identified in the parent MSCs. These include immunomodulation and anti-inflammatory properties on host tissue, reduction of the permeability of alveolar epithelium and endothelium, improvement of alveolar fluid clearance, enhancement of macrophage phagocytosis, and tissue repair through direct mitochondrial transfer with host cells (Figure 2). reduction in TNF-α and CXCL10 protein levels, and increase in IL-10 protein levels 3 days post-infection. [118] potential therapy in the management of other lung diseases, such as bronchopulmonary dysplasia, pulmonary arterial hypertension, idiopathic pulmonary fibrosis, and asthma, which have been recently revised in Worthington [121] and Behnke [122]. Moreover, in these lung diseases, the most common effects of MSC-derived EVs were decreased inflammation and restoration of the lung architecture, achieved through the reduction of fibrosis and increase of vascularization and alveolarization.  [101,105,114,118], reduction of the permeability of alveolar epithelium and endothelium [116], improvement of alveolar fluid clearance [114], enhancement of macrophage phagocytosis [111], and tissue repair through direct mitochondrial transfer with host cells [111].

MSC-Derived Exosomes and Heart Injuries
Acute myocardial injury has been described as the most common cardiovascular complication in COVID-19 patients [13]. Myocardial injury is defined as an elevation in serum levels of highsensitive cardiac troponin (cTn) above the 99th percentile upper reference limit, although over the years it has also been identified through an increase in different cardiac enzymes and/or electrocardiographic abnormalities [123]. The injury is considered acute if there is a dynamic rise and/or fall of cTn values. When acute myocardial injury is caused by myocardial ischemia, it is designated as acute myocardial infarction (AMI). On the contrary, myocardial injury not related to ischemic events may arise secondary to many cardiac conditions, such as myocarditis [123].  [101,105,114,118], reduction of the permeability of alveolar epithelium and endothelium [116], improvement of alveolar fluid clearance [114], enhancement of macrophage phagocytosis [111], and tissue repair through direct mitochondrial transfer with host cells [111].
Apart from ALI/ARDS, there have been several investigations using MSC-derived EVs as a potential therapy in the management of other lung diseases, such as bronchopulmonary dysplasia, pulmonary arterial hypertension, idiopathic pulmonary fibrosis, and asthma, which have been recently revised in Worthington [121] and Behnke [122]. Moreover, in these lung diseases, the most common effects of MSC-derived EVs were decreased inflammation and restoration of the lung architecture, achieved through the reduction of fibrosis and increase of vascularization and alveolarization.

MSC-Derived Exosomes and Heart Injuries
Acute myocardial injury has been described as the most common cardiovascular complication in COVID-19 patients [13]. Myocardial injury is defined as an elevation in serum levels of high-sensitive cardiac troponin (cTn) above the 99th percentile upper reference limit, although over the years it has also been identified through an increase in different cardiac enzymes and/or electrocardiographic abnormalities [123]. The injury is considered acute if there is a dynamic rise and/or fall of cTn values. When acute myocardial injury is caused by myocardial ischemia, it is designated as acute myocardial infarction (AMI). On the contrary, myocardial injury not related to ischemic events may arise secondary to many cardiac conditions, such as myocarditis [123].
Analyzing several reports from China, a considerable proportion of patients (12-27.8%) presented elevated cTn levels, and most of them required ICUs and showed higher in-hospital mortality [7,[124][125][126]. The mechanisms of myocardial injury are not well established but likely involve direct or indirect processes and/or their combination (Figure 3). Myocardial infection by SARS-CoV-2 resulting in cardiomyocyte death and inflammation has been proposed as a possible direct mechanism, although, to date, there are no data demonstrating the presence of SARS-CoV-2 within myocardial tissue [127]. Nevertheless, a previous autopsy study in patients who died from SARS identified the viral RNA in 35% of the post-mortem human heart samples, providing evidence for direct myocardial injury by the virus [46]. In addition, patients carrying SARS-CoV in their hearts died considerably earlier, suggesting that viral infiltration in the myocardium was associated with a more aggressive course of illness. Systemic inflammatory response or respiratory failure and hypoxemia can represent indirect mechanisms leading to increased cardiac stress and myocardial inflammation [127,128]. In a couple of studies, biopsies taken from heart tissue of COVID-19 patients evidenced mononuclear inflammatory infiltrates, mainly associated with regions of cardiomyocyte necrosis, which identifies myocarditis according to Dallas criteria [129][130][131]. Nevertheless, acute lymphocyte infiltrates were not observed in the myocardium of SARS-CoV-2-infected patient autopsy. died considerably earlier, suggesting that viral infiltration in the myocardium was associated with a more aggressive course of illness. Systemic inflammatory response or respiratory failure and hypoxemia can represent indirect mechanisms leading to increased cardiac stress and myocardial inflammation [127,128]. In a couple of studies, biopsies taken from heart tissue of COVID-19 patients evidenced mononuclear inflammatory infiltrates, mainly associated with regions of cardiomyocyte necrosis, which identifies myocarditis according to Dallas criteria [129][130][131]. Nevertheless, acute lymphocyte infiltrates were not observed in the myocardium of SARS-CoV-2-infected patient autopsy.
Other aspects of COVID-19 in cardiac involvement include blood pressure abnormalities and arrhythmias, ranging from tachycardia and bradycardia to asystole [127]. Very recently, it has also been suggested that there is a link between SARS-CoV-2 infection and Kawasaki disease (KD), especially in pediatric patients [132]. Although KD is a disease of unknown etiology, infections are considered to be one of the predisposing factors [133]. The disease predominantly affects children under 5 years of age and causes inflammation in the walls of medium-sized arteries, primarily the coronary arteries, those that supply blood to the heart muscle. Consequences of AMI are loss of cardiomyocytes and adverse remodeling of the extracellular matrix, which contribute to the reduction of pumping of the heart and further heart failure. Nowadays, the best therapeutic strategy for reducing AMI is timely and effective myocardial Other aspects of COVID-19 in cardiac involvement include blood pressure abnormalities and arrhythmias, ranging from tachycardia and bradycardia to asystole [127]. Very recently, it has also been suggested that there is a link between SARS-CoV-2 infection and Kawasaki disease (KD), especially in pediatric patients [132]. Although KD is a disease of unknown etiology, infections are considered to be one of the predisposing factors [133]. The disease predominantly affects children under 5 years of age and causes inflammation in the walls of medium-sized arteries, primarily the coronary arteries, those that supply blood to the heart muscle.
Consequences of AMI are loss of cardiomyocytes and adverse remodeling of the extracellular matrix, which contribute to the reduction of pumping of the heart and further heart failure. Nowadays, the best therapeutic strategy for reducing AMI is timely and effective myocardial reperfusion. However, this treatment induces oxidative stress and inflammation, thus leading to further cardiomyocyte death, myocardial remodeling, and decreased cardiac function, a phenomenon known as myocardial reperfusion injury [134].
Over the last years, management of AMI using stem cell therapy was found to prevent myocardial cell apoptosis, promote local neoangiogenesis, and reduce the local inflammatory response [135][136][137]. Similarly to what was described above for lung injuries, the beneficial effect of stem cells seems to be largely attributable to the secreted EVs. Since the first description of the therapeutic potential of MSC-derived exosomes in a mouse model of myocardial ischemia/reperfusion (I/R) injury in 2010, several studies have subsequently reported cardio-protective effects of MSC-derived EVs in AMI animal models (Table 2) [138]. In one of these works, a single intravenous (IV) injection of MSC-derived exosomes in a mouse AMI model led to decreased infarct size, enhanced nicotinamide adenine dinucleotide (reduced form) (NADH) and ATP levels, and reduced oxidative stress, which are hallmarks of reperfusion injury [139]. All these events seemed to be associated with the exosome-mediated activation of the pro-survival phosphoinositide 3-kinase/protein kinase B (PI3K/Akt) signaling pathway, which resulted in an enhancement of myocardial viability and prevented adverse remodeling after myocardial I/R injury. Importantly, intact but not lysed exosomes were responsible for the improved cardiac function after AMI induction.
Another important mechanism by which MSC-derived EVs contribute to ischemic myocardial repair is through stimulation of neovascularization, as shown in the work of Bian and colleagues [140]. Neovascularization refers to processes, such as vasculogenesis, angiogenesis, and arteriogenesis, that are associated with migration and proliferation of endothelial cells. In line with these findings, Ma and coworkers also demonstrated that exosomes isolated from Akt-transfected MSCs accelerated angiogenesis in a rat myocardial infarction model [141]. The authors suggested that platelet-derived growth factor D (PDGF-D), which was enriched in MSC-derived vesicles, was mainly responsible for the Akt exosome-mediated improvement of myocardial repair. A more recent study by Xuan and colleagues identified Notch1 as a potent modulator of angiogenesis and cardiomyocyte proliferation into ischemic mice hearts following coronary heart ligation [142]. The role of Notch1 signaling in inducing cardiac angiogenesis during ischemia and enhancing survival of cardiac cells is well established [143]. The injection of MSC-derived EVs over-expressing Notch1 intracellular domain (NICD) in ischemic myocardium led to decreased infarct size, improved cardiac function, and increased arteriole density in the peri-infarct area, 1 month after AMI [142]. Moreover, Teng and coworkers indicated that the beneficial effect of MSC-derived exosomes on infarcted rat hearts is mainly dependent on their angiogenesis-promoting activity [144]. In their study, the authors proved that exosomes also act by restraining the inflammatory response. In agreement with the results of Arslan and colleagues, they also demonstrated that fresh exosomes achieved a better therapeutic effect with respect to frozen exosome preparations.
Several studies agree that reduced fibrosis and apoptosis of myocardial cells are other important effects of the EV-mediated ischemic cardiac repair [145][146][147][148]. In particular, Zhao and coworkers showed that human umbilical cord MSC-derived exosomes improved cardiac function and reduced cardiac fibrosis by preventing cardiomyocyte apoptosis and promoting cell proliferation in the border zone of infarcted rats [146]. The effect mediated by exosomes was attributed to the up-regulation of the anti-apoptotic protein B cell lymphoma 2 (Bcl-2) in the myocardial cells. Other works have proposed that specific functional miRNAs contained into EVs and shuttled to target injured cells are primarily responsible for the beneficial effects. For example, Feng and colleagues found that miR-22 was up-regulated in MSC-derived EVs, and it possibly reduced cardiac apoptosis and fibrosis in an AMI mouse model via inhibition of methyl cytosine-phosphate-guanine (CpG)-binding protein 2 (Mecp2) expression [145]. In that study, the authors isolated EVs from MSCs subjected to ischemic pre-conditioning, which is an effective approach to potentiate survival and regeneration of these cells in an ischemic environment. Yu and coworkers identified miR-19a as the molecular mediator able to restore cardiac function and reduce infarct size in a rat model of AMI [147]. The cardio-protective role of miR-19a was mediated by down-regulation of target genes, phosphatase and tensin homolog (PTEN), and Bcl-2 interacting mediator of cell death (BIM) in cardiomyocytes and subsequent activation of the Akt and extracellular signal-regulated kinase (ERK) signaling pathways. In their study, exosomes were isolated from MSCs over-expressing GATA binding protein 4 (GATA-4), a transcription factor able to regulate miRNA expression in MSCs and increase their survival in an ischemic environment [149]. Apart from observing reduced cardiac fibrosis and reduced inflammation in infarcted rat hearts 7 days after exosome injection, Shao and colleagues identified a panel of miRNAs, which were similarly up-or down-regulated in MSCs and the derived exosomes [148]. On the other hand, other miRNAs, such as miR-21 and miR-15, resulted as being differentially expressed between exosomes and MSCs, with this potentially explaining why MSC-derived exosomes demonstrated superior beneficial effects when compared with treatment with their parent cells.
Several studies have demonstrated that autophagy also has an important role in mediating the therapeutic effects of MSC-derived exosomes. Autophagy is known to be an important mechanism in cardio-protection, and dysregulated autophagy is associated with a variety of CVD [150]. In particular, it has been demonstrated that exosomes reduce apoptosis and the myocardial infarct size, as well as improve cardiac function by inducing cardiomyocyte autophagy both in vitro and in vivo [151,152].
Collectively, the described studies documented reduction in infarct size with improved recovery of cardiac function, reduction of fibrosis and apoptosis, stimulation of angiogenesis, and decreased infiltration of macrophages and other immune cells into the injured heart regions following treatment with MSC-derived EVs (Figure 4). When comparing the properties of exosomes recovered from different MSCs sources, those isolated from adipose tissue samples exhibited the strongest cardio-protective effects [153]. [151] 1 ischemia/reperfusion; 2 left coronary artery; 3 intravenous; 4 glucogen synthase kinase-3; 5 c-Jun Nterminal kinase; 6 acute myocardial infarction; 7 left anterior descending; 8 platelet-derived growth factor D; 9 cardiac MSCs over-expressing Notch1 intracellular domain; 10 methyl cytosine-phosphateguanine binding protein 2; 11 B cell lymphoma 2; 12 phosphatase and tensin homolog; 13 phosphatebuffered saline; 14 microtubule-associated protein 1 light chain 3 beta.

Conclusions and Future Perspectives
The recent coronavirus COVID-19 global pandemic has driven the need for novel urgent therapies. MSCs and their derivatives are being evaluated for the treatment of a number of diseases that currently have limited or no therapeutic options. MSC-derived EVs (exosomes and MVs) have recently attracted great attention because, similarly to their parent cells, they possess strong anti-inflammatory, immunomodulatory, and pro-angiogenic abilities, just to name a few. However, compared to MSCs themselves, EVs hold many biological and technological advantages. EV administration is considered safer than MSC transplantation, lacking some of their negative side-effects, and they are more stable than MSCs themselves, allowing for easier handling and storage. Over the last years, a plenty of pre-clinical studies in animal models have demonstrated that the administration of MSC-derived EVs significantly reduced lung inflammation and pathological impairment subsequent to different types of lung injury, as well as resulted in improved cardiac function after acute myocardial injury. However, several challenges still need to be overcome to make the transition from animal models to humans possible. For example, standardized techniques for isolation, characterization, and quantification, as well as criteria for establishing dose, quality control, and storage conditions of MSC-derived EVs, are required before these can be advanced to the clinic. To date, it is difficult to compare and analyze studies employing MSC-derived EVs since there is a large degree of heterogeneity in EV preparations, and because MSC-derived EVs differ depending on tissues and donors from which the cells are isolated.
Regarding COVID-19, the lack of an established animal model of coronavirus-induced lung injury requires a more prudent and careful use of MSC-derived EVs. In this context, a significant issue is to establish under what circumstances and with what criteria to administer MSC-derived EVs. For example, which population among COVID-19 patients to target and when to start EV administration. Moreover, there remains the challenge to clarify the optimal route of EV administration that, in the case of lung diseases, mostly occurs through IT instillation or IV injection, although the possibility of EV inhalation has recently been explored. To date, no studies have investigated the biodistribution and the in vivo metabolic fate of EVs following IT instillation. On the other hand, systemic IV injection has been shown to deliver EVs primarily to the spleen and liver, then to the gastrointestinal tract and lungs, followed by renal and hepatic clearance in mice [138,154]. Apart from the administration route, another major issue concerns the optimal EV therapeutic dose. Considering that the average therapeutic dose of MSCs for treating lung injuries is 10 × 10 6 cells/kg per body weight, the amount of cells required to generate enough EVs to achieve the equivalent effect of MSCs is generally 5-10 times higher [155]. This necessitates large scale production of MSC-derived EVs. Although this could be implemented with the use of bioreactors for MSC expansion, different bioreactor culture conditions would result in alterations of EV content, which in turn may impact on the therapeutic efficacy. Another challenge to consider for the administration of MSC-derived EVs to COVID-19 patients is the need to manufacture a safe and reproducible therapeutic product. Since the production of EVs requires the use of living cells, these have to be cultured under good manufacturing practice (GMP)-compliant procedures to preserve the quality and safety standards criteria. Therefore, EV production must follow the same rigorous scientific and ethical guidelines that apply to MSCs, and any therapy based on MSC-derived EVs needs to be approved by the national regulatory agencies to demonstrate its safety and efficacy. In light of these observations, in our opinion, the use of MSC EVs could be contemplated to treat critically ill patients with ARDS requiring mechanical ventilation or ICU support, or patients with recognized risk factors, such as pre-existing CVD, or cardiovascular complications, for whom standard therapeutic approaches have not proven resolutive.

Conflicts of Interest:
The authors declare no conflict of interest.