Acute respiratory distress syndrome (ARDS) is the major cause of acute respiratory failure in the intensive care units (ICUs) and carries a high mortality rate (Thompson et al., 2017). Risk factors for this condition include infection, trauma or other systemic conditions. Management is exclusively supportive and the lack of approved pharmacological therapies reflects major deficiencies in our understanding of the pathogenesis of ARDS. Although studies defining appropriate ventilator management have improved patient outcomes, the incidence and mortality associated with ARDS remains unacceptably high. In late December 2019, an outbreak of a novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was first identified. Its manifestations can range from asymptomatic or mild respiratory infection to severe ARDS and death. With a high rate of transmission and mortality, the SARS-CoV-2 infection (COVID-19) has become a global threat which demands effective treatments beyond supportive therapies (Zheng, 2020).

The pathological hallmark of ARDS is the diffuse alveolar damage with an early alveolar epithelial-capillary barrier disruption. ARDS is characterized by pulmonary edema and alveolar collapse accompanied by ventilation-perfusion mismatch and severe arterial hypoxemia (Thompson et al., 2017). In these patients, the injured alveoli are characterized by the presence of an intense inflammatory response with leukocyte infiltration, activation of pro-coagulant processes in the alveolar air spaces and in the microvasculature, and damage of epithelial and endothelial cells. All these events contribute to the breakdown of the alveolar-epithelial barrier and, consequently, to the formation of alveolar protein-rich edema. Such pulmonary edema is a major factor for hypoxemia and one of the earliest events in ARDS. Once ARDS develops, patients may present pulmonary vascular dysfunction and pulmonary hypertension which is independently associated with poor outcomes in patients with ARDS (Bull et al., 2010). Uncontrolled inflammation can also lead to the massive release of inflammatory mediators (such as interleukin [IL]-1β, IL-6, IL-8 or tumor necrosis factor α [TNF-α]), causing the so-called “cytokine storm” which results in vascular inflammation, thrombosis and vasodilation and may lead to multiorgan dysfunction (Thompson et al., 2017; Zheng, 2020). Besides the formation of thrombi in the pulmonary microvasculature, disseminated intravascular coagulation (DIC) is also a frequent complication in patients with severe ARDS that can be clinically expressed by excessive hemorrhage and ischemic necrosis of extremities. In addition, the onset of DIC has been associated with deterioration of lung functions with worsening hypoxemia and a progressive fall in pulmonary compliance (Gando et al., 2004). Platelet activation and pulmonary-capillary endothelial-platelet interaction may be one of the major and earliest mechanisms involved in both local lung and systemic coagulation in these patients. Furthermore, increasing evidence shows the relevant role of activated platelets and their vasoactive and procoagulant products in the increase of pulmonary capillary permeability that contributes to lung edema formation in ARDS (Zarbock et al., 2006).

In the lung, the alveolar-capillary membrane separates the alveolar airspace from the capillary lumen and comprises a complex architecture optimized to exert its multiple functions that include gas exchange (oxygen is diffused into the capillaries and carbon dioxide released from the capillaries into the air space). The alveolar-capillary membrane has several layers: a lining fluid layer containing the surfactant, the epithelial barrier and its basement membrane, a thin interstitial space with a biologically active extracellular matrix (ECM), a capillary basement membrane and the capillary endothelium (Weibel, 1973; Maina and West, 2005; Knudsen and Ochs, 2018). Between these epithelial and endothelial layers there are also resident and migratory leukocytes, as well as a population of mesenchymal stromal cells, such as pericytes and resident fibroblasts (Barron et al., 2016).

The alveolar epithelium is a tight barrier that restricts the passage of water, electrolytes, and small hydrophilic solutes from the insterstitium to the air spaces, allowing at the same time the diffusion of carbon dioxide and oxygen (Taylor and Gaar, 1970; Weibel, 2015). The alveolar epithelium is composed by flat alveolar type I (ATI) cells and cuboidal shaped alveolar type II (ATII) cells. The ATII cells secrete surfactant, the critical factor that reduces surface tension, enabling the alveoli to remain open and facilitating gas exchange. ATII also can differentiate to replace damaged type I cells (Matthay et al., 1993; Mason and Dobbs, 2016). Both ATI and ATII cells have the capacity to absorb excess fluid from the airspaces to the interstitium by vectorial ion transport, primarily promoted by epithelial sodium channels (ENaC) and basolateral Na⁺/K⁺-ATPase pumps (Matthay, 2014; Matthay et al., 2019). Thus, proper function of these channels is essential for the reabsorption of edema and are key elements for a clear improvement of the disease in patients with sepsis and ARDS (Ware and Matthay, 2001; Zeyed et al., 2012). However, this capability of alveolar fluid clearance not only depends on intact ion transport channels in the epithelial cells, but also on the integrity of the junctions between ATI and ATII cells. Once into the insterstitium, the edema fluid is removed to systemic circulation by the lymphatics. Although more permeable than the epithelium, the capillary endothelium in the alveoli forms a semipermeable barrier that limits the extravasation of plasma and its macromolecules from the vascular lumen to the instertitium. Both epithelial and endothelial barrier functions and permeability are governed by intercellular junctions (Herrero et al., 2018). These intercellular junctions between neighboring cells in the epithelium and endothelium are mainly formed by apical tight junctions (TJs) and the underlying adherens junctions (AJ), and linked to the cellular cytoskeleton via numerous adaptor proteins. AJs are composed of cadherins, mainly vascular endothelial cadherin (VE-cadherin) that regulate the paracellular transport between blood and interstitium, consequently determine leukocyte migration and edema formation during ARDS (Millar et al., 2016).

In general, TJs control paracellular transport, maintain cellular polarity, establish separate intercellular compartments, regulate a variety of intracellular signals, and control the transcellular transport. Occludins, claudins, and zonula occludens (ZO) are essential components of tight junctions in the alveolar epithelium that constitute the main structure to regulate the passage of water and solute from the interstitial to the alveolar space, and to prevent the passage of pathogens and toxins from the air space into the systemic circulation (Zemans and Matthay, 2004; Yanagi et al., 2015; Herrero et al., 2018). Thus, alteration of the epithelial TJs results in protein-rich edema formation, and passage of infectious agents, exogenous toxins and endogenous products into the systemic circulation, exposing other organs and contributing to multiorgan failure (Denker and Nigam, 1998; Schneeberger and Lynch, 2004; Herrero et al., 2018). Endothelial junctional proteins also play important roles in tissue integrity as well as in vascular permeability, leukocyte extravasation, and angiogenesis (Wallez and Huber, 2008). Specifically, intercellular cell adhesion molecule- 1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1) control the adhesion of leukocytes to endothelial cells and facilitate their subsequent transendothelial migration via the platelet-endothelial cell adhesion molecule-1 (PECAM-1), contributing to the inflammatory process (Millar et al., 2016; Villar et al., 2019).

The ECM, composed of a highly dynamic complex of fibrous proteins, glycoproteins, and proteoglycans, is crucial to maintain the epithelial and endothelial barrier function by regulating cell-cell interactions. ECM also modulates cell survival, proliferation, migration and differentiation, and play an important role in tissue repair. Resident fibroblasts in the interstitium are mainly responsible for ECM production (White, 2015; Barron et al., 2016). Changes in the composition and mechanic properties of the ECM have been shown to modify the expression of TJs and the barrier function in alveolar epithelial and endothelial cells, contributing to lung edema formation (Rocco et al., 2001; Pelosi et al., 2007; Koval et al., 2010; Mammoto et al., 2013). Post-translational modifications (e.g., enzymatic and chemical crosslinking, transglutamination, glycation and glycosylation, oxidation, and citrullination), are known to affect the structural and/or functional diversity of ECM proteins. The structural organization and degradation of the ECM are controlled by the proteolytic action of many proteases, including the superfamily of metalloproteinases that in turn comprises adamalysins (e.g., ADAMS, ADAMTS), and matrix metalloproteinases (MMPs) and their inhibitors (the tissue inhibitors of metalloproteinases-TIMPs). These enzymes are expressed by inflammatory cells (mainly macrophages) and stromal cells (including fibroblasts, endothelial and epithelial cells). In this line, a growing body of evidence shows a role of MMP-1, -2, -3, -7, -8, and -9, expressed by inflammatory, mesenchymal and epithelial cells, in the development and repair of the alveolar-capillary damage in ARDS (Davey et al., 2011). The oxidation of ECM proteins by reactive oxygen species (ROS) further modulates characteristics of the ECM. It is well known that increased levels of ROS are present in the alveoli of patients with ARDS that may alter ECM properties, having an impact on cell-ECM interaction and alveolar permeability (Watson et al., 2016).

In this interstitial space the pericytes and resident fibroblasts, not only play a relevant role in the maintenance of the normal vascular and epithelial functions in the lung, but also contribute to lung injury in pathological conditions. Pericytes underlie and envelop capillaries, forming intimate contacts with adjacent endothelial cells that control vessel integrity, angiogenesis and capillary permeability. Resident fibroblasts are mainly located beneath epithelial cells or are scattered through the interstitium. These mesenchymal stromal cells generate and remodel extracellular matrix, regulate the vasculature, help maintain and restore the epithelial barrier structure and function, and control immune cell activity and migration. These cells are the main source of myofibroblasts in the lung during development and after injury, contributing in the latter to lung fibrosis (Rock et al., 2011; Hung et al., 2013; Barron et al., 2016).

The lung has two main types of macrophages that reside in different anatomical compartments, namely interstitial and alveolar macrophages. Whereas alveolar macrophages (AM) are predominantly of embryonic origin, interstitial macrophages (IM) are mainly derived by blood monocytes. Both types of macrophages have an important role in host-defense as part of the local innate immune response in the lung (Evren et al., 2020). AM are in close communication with the alveolar epithelial cells (AECs) by different mechanisms including mediators, membrane glycoproteins and their receptors, gap junction channels, and extracellular vesicles (EVs). Thanks to this communication, AM regulate the alveolar epithelial barrier function and permeability, while exerting other functions such as barrier immunity, surfactant clearance, and removal of foreign particles. In pathological conditions, activated AM cause epithelial cells to produce anti-microbial peptides and a variety of mediators that not only modulate immune responses, but also contribute to epithelial barrier dysfunction via alteration of the TJ proteins (Bissonnette et al., 2020). Interstitial macrophages, located in the space between the lung epithelium and capillaries in the airway, perform antigen presentation and contribute to tissue remodeling in the lung (Koch et al., 2017). In contrast to AM, the location of IM in the alveoli and their contribution to ARDS have not been totally elucidated.

Multiple clinical and experimental studies have provided relevant information about the different mechanisms involved in acute lung injury (ALI) and recovery. One of the main conclusions is that recovery from lung injury requires the repair of the endothelial barrier as well as the reestablishment of the normal functions of the alveolar epithelial barrier capable of secreting surfactant, preventing the passage of infectious/toxics agents into systemic circulation, and removing alveolar edema (Herrero et al., 2018; Matthay et al., 2019). Increasing studies are focusing on EVs released by pulmonary structural cells, immune cells and mesenchymal stem cells and their potential contribution to the development or resolution of lung injury.

Extracellular vesicles represent a heterogeneous range of membrane enclosed spheres of varying size that are secreted by a variety of cell types, including T cells, B cells, dendritic cells, platelets, mast cells, epithelial cells, endothelial cells, neuronal cells, cancerous cells, oligodendrocytes, Schwann cells, embryonic cells, and mesenchymal stem cells (MSCs) (Raposo and Stoorvogel, 2013; Borgovan et al., 2019). They are shed or secreted from these cell types under various physiologic and pathologic conditions into the circulation or surrounding body fluids, including blood or bronchoalveolar lavage fluid (BALF) (Witwer et al., 2013; Cocucci and Meldolesi, 2015; Konala et al., 2016; Borgovan et al., 2019).

The International Society of Extracellular Vesicles (ISEV) establishes the minimum requirements for the collection and pre-processing of samples and for the separation, concentration, and characterization of EVs, as well as the steps to demonstrate that a function is associated specifically with EVs (Thery et al., 2018; Nieuwland et al., 2020). According to the ISEV, three main sub-groups of EVs have been classified based on their size, membrane composition and biogenesis, as shown in Table 1 (Raposo and Stoorvogel, 2013; Witwer et al., 2013). Apoptotic bodies (50–5000 nm) are the largest EVs and are formed during cellular apoptosis by cell membrane-blebbing. Apoptotic bodies contain histones and genomic DNA. Microvesicles (MVs; 100–1000 nm) are shed via the outward blebbing of the plasma membrane, allowing retention of the membrane proteins of the parent cell. MVs are rich in the surface marker CD40, integrins and selectins as well as cholesterol, sphingomyelin, and ceramide. Exosomes (40–120 nm) are the smallest subgroup and are released after multiple vesicular bodies fuse with the plasma membrane. Exosomes may express distinct biomarkers, including tetraspanins (CD61, CD63, or CD81), ESCRT proteins (TSG101 and Alix), flotillin, and heat shock proteins, as well as high acetylcholinesterase activity (Crescitelli et al., 2013; Raposo and Stoorvogel, 2013; Cocucci and Meldolesi, 2015; Yanez-Mo et al., 2015; Borgovan et al., 2019).

Extracellular vesicles are carriers of biologically active molecules (nucleic acids, proteins and lipids), whose composition vary based on the parent cell phenotype and biological state (Keerthikumar et al., 2016; Pathan et al., 2019; Xie et al., 2020). Proteins carried by EVs include chemokines, and inflammatory cytokines, integrins, growth factors, enzymes or even cytoskeletal components (Gutierrez-Vazquez et al., 2013; Keerthikumar et al., 2016; Maas et al., 2017; Mardpour et al., 2019). Nucleic acid cargo in EVs comprise mitochondrial and genomic DNA, small non-coding RNA species (such as microRNA or tRNA, small nucleolar RNA, and small nuclear RNA) and long non-coding RNA species (Kalra et al., 2012; Keerthikumar et al., 2016; Xie et al., 2020). EVs are also an important source of lipids, including sphingomyelin, ceramides, phosphatidylserine (PS), cholesterol or saturated fatty acids (Gutierrez-Vazquez et al., 2013; Chatterjee et al., 2020b).

Increasing evidences suggest that EVs mediate intercellular communication by transferring their cargo to recipient cells and are able to modulate physiological and pathophysiological process in both parent and recipient cells, including the induction and the resolution of lung injury and inflammation (Lee et al., 2018a; Lanyu and Feilong, 2019). In this review, we will discuss the potential pathophysiological and therapeutic role of EVs in ARDS. Since exosomes and MVs can overlap in the size range and current methods are unable to separate these populations efficiently, the ISEV has encouraged using the generic term of “extracellular vesicles” (Thery et al., 2018). Following these recommendations, we will use the term EVs, without making distinctions between exosomes or MVs, since the vast majority of studies have not been able to determine the specific biogenesis pathway for a given vesicle.

Increased levels of EVs have been detected in BALF following either pulmonary viral and bacterial infections (Mills et al., 2019) or injury (hyperoxia, ventilator-induced lung injury-VILI, acid installation) (Lanyu and Feilong, 2019; McVey et al., 2019). Interestingly, a study carried out in mice revealed that BALF-EVs from sterile stimuli (hyperoxia-induced ALI) were mainly derived from ATI epithelial cells (enriched in cytokeratin and podoplanin), whereas BALF-EVs from infection-induced injury were mainly released by AM (enriched in CD68). This study showed that these two types of EVs activated AM through Toll-like receptor (TLR) pathways, although differential TLR pathways and signaling cascades were involved. In this regard, epithelial cell-derived EVs (non-infectious stimuli-induced EVs) activated TLR2, Myd88, TNF-α, and IL-6 expression in recipient macrophages, whereas macrophage-derived EVs (infectious stimuli-induced EVs) upregulated TLR6, TLR9, CD80, IL-1β, and IL-10 in recipient macrophages (Lee et al., 2018b). Both TLR2 and TLR6 are key receptors in the nuclear factor kappa B (NFκB)-mediated inflammation in many lung diseases, including ARDS (Lafferty et al., 2010).

A growing body of evidence shows that BALF-EVs significantly contribute to the development of lung inflammation and the progression of pulmonary damage during ARDS (Lee et al., 2018b). In patients with ARDS, secretory phospholipase A2 (sPLA2) is increased in BALF-EVs at an early state compared to patients without ARDS (Papadopoulos et al., 2020). sPLA2 is an oxidative stress-induced inflammation factor that activates inflammation and hydrolyzes lung surfactant phospholipids, contributing to lung collapse (Kim et al., 1995). In mouse models of ventilator-induced lung injury (VILI), BALF-EVs are enriched in IL-1β, IL-6, and TNF-α compared to non-ventilated control animals (Dai et al., 2019). Certain miRNAs were also detected in BALF-EVs of lipopolysaccharide (LPS)-induced ARDS in animals, specifically miR-466g, miR-466m-5p, miR-155, and miR-146a (Shikano et al., 2019). BALF-EVs enriched with miR-466g and miR-466m-5p activated the nod-like receptor family pyrin domain containing 3 (NLRP3) inflammasome in macrophages culture, inducing the IL-1 production by macrophages (Shikano et al., 2019). Cultured AECs treated with BALF-EVs enriched in miR-155 and miR-146a overexpress pro-inflammatory mediators, such as TNF-α and IL-6, and downregulate the expression of zonula occludens 1 (ZO-1) (Yuan et al., 2018), suggesting their role in epithelial barrier disruption.

Circulating EVs are also increased in ARDS models, including infection-related, endotoxin-induced and sterile lung injury (Letsiou et al., 2015; Moon et al., 2015; Pan et al., 2017; Jiang et al., 2019; Chatterjee et al., 2020a). Pneumonia induced by Escherichia coli in an ex vivo model of perfused human lung causes the release of EVs, enriched in TNF-α and IL-6, into the perfusate. Interestingly, these pneumonia-induced EVs, mainly derived from endothelial cells and platelets, are capable to cause lung damage when administered to healthy lung (Liu et al., 2019). Indeed, increased levels of endothelial-derived EVs have been associated with endothelial dysfunction (Letsiou and Bauer, 2018), gas-exchange deterioration (Cabrera-Benitez et al., 2015; Pan et al., 2017) and higher mortality (Sun et al., 2012) in ARDS patients. Also, it is crucial to determine the cellular origin of circulating EVs, since a protective role for some EVs has also been identified. For example, Shaver et al. (2017) found that elevated concentrations of EVs in plasma of patients admitted to ICU were associated with a reduced risk of developing ARDS, whereas increased levels of leukocyte-derived EVs in BALF, during the early stages of the disease, were associated with increased survival in ARDS patients (Guervilly et al., 2011).

Extracellular vesicles can also serve as vehicles of mature infectious virus particles between cells, which can increase the infectivity to host cells (Santiana et al., 2018; Altan-Bonnet et al., 2019). In fact, EVs derived from endothelial cells have been reported to transfer angiotensin-converting enzyme 2 (ACE-2) and contribute to the SARS-CoV-2 virus spreading (Wang et al., 2020). EVs can also modulate the inflammatory response triggered by virus infection. For example, rhinovirus infection of human bronchial epithelial cells results in the release of EVs containing tenascin-C, an immunomodulatory ECM protein which has the ability to induce inflammatory cytokine production (Mills et al., 2019).

Altogether, EVs seem to be a relevant mechanism in the pathogenesis of ARDS since they can activate inflammation, alter the alveolar capillary barrier and lead to alveolar edema in the lung (the summary of the main effects induced by EVs in ARDS is provided in Table 2). After lung injury, the release of lung-derived EVs into the systemic circulation also might contribute to damage in distal organs. Finally, we should consider that EVs derived from BALF, although less accessible than those EVs from serum/plasma, might reflect better the complete status of the lung microenvironment and be more useful to elucidate the pathophysiology of ARDS (Thompson et al., 2017).

As discussed in the previous section, there is growing evidence that EVs play an important role in the pathogenesis of ARDS. However, there is also some evidence suggesting that some endogenous EVs may also have a protective effect in ARDS patients (Mahida et al., 2020a). For example, some clinical studies have reported that total leukocyte-derived EVs are associated with a better prognosis in patients with sepsis (Lashin et al., 2018) or ARDS (Guervilly et al., 2011; Mahida et al., 2020a, b). Notably, circulating leukocyte-derived EVs expressing α2-macroglobulin (A2MG) isolated from septic patients were shown to reduce endothelial cell permeability and increased bacterial phagocytosis by neutrophils in vitro (Lashin et al., 2018). Similarly, some studies have reported anti-inflammatory effects by EVs derived from neutrophils on AM and AECs (Gasser and Schifferli, 2004; Eken et al., 2010; Mahida et al., 2020a, b). Indeed, transfer of miR-223 by neutrophil-derived EVs to airway epithelial cells has been shown to reduce protein permeability and inflammatory cytokine release in vitro and in vivo (Neudecker et al., 2017). Finally, several studies have demonstrated that EVs from endothelial progenitor cells (EPCs) also reduce inflammation and lung injury in several models of ALI (Wu et al., 2018; Zhou et al., 2018, 2019).

Mesenchymal stem cells (MSCs) are multipotent progenitor cells that can be isolated from multiple tissues including bone marrow, adipose tissue, and umbilical cord tissue, blood and perivascular tissue. The source may impact the immunomodulatory effects, proliferation properties and therapeutic benefits of MSCs. Bone marrow-derived MSCs were the first type of MSCs isolated and are the most widely used for lung injury (Monsel et al., 2016; Mohammadipoor et al., 2018). Administration of MSCs induces potent anti-inflammatory and immunomodulatory effects and has been proven to decrease lung injury and increase survival in several preclinical models of ALI (Pati et al., 2011; Curley et al., 2012; Jackson et al., 2016; Xiao et al., 2020). Thus, MSCs were able to maintain the integrity of the lung microvascular barrier (Li C. et al., 2019) and reduce the formation of pulmonary edema (Gupta et al., 2007) in several models of endotoxin-induced ALI due to a down-regulation of pro-inflammatory cytokines, such as TNF-α or MIP-2, and an increase in the production of the anti-inflammatory cytokine IL-10 (Gupta et al., 2007). Moreover, MSCs also attenuated neutrophil-predominant inflammation and lung injury in an in vivo rat model of VILI (Lai et al., 2015). In Influenza mice models, MSCs also limited alveolar inflammation (Li et al., 2016) and prevented the downregulation of sodium and chloride transported in infected cells, reducing the impairment of alveolar fluid clearance and attenuating lung injury (Chan et al., 2016).

Despite these long-lasting therapeutic effects in a wide variety of in vivo experimental models, the engraftment rate of MSC has been shown to be extremely low (Monsel et al., 2016; Mastrolia et al., 2019). This unexpected low engraftment rate represented a major challenge in explaining the beneficial effects of MSCs (since in most cases, the cells were only temporarily resident in the host). Indeed, administration of MSC-derived conditioned medium (MSC-CM) recapitulates the therapeutic effects of MSCs in ARDS and other inflammatory lung diseases through activation of anti-inflammatory, pro-survival and anti-apoptotic pathways (Monsel et al., 2016; Byrnes et al., 2021). MSC-CM can mitigate the inflammatory response of the injured endothelium by preserving the integrity of the vascular barrier, restoring the normal status of membrane adhesion molecules β-catenin and VE-cadherin, and preventing inflammatory cell binding to endothelial cells (Pati et al., 2011). Moreover, it has been demonstrated the capability of MSC-CM to reduce the secretion of TNF-α by macrophages mediated by IL-1RA (Ortiz et al., 2007). The administration of MSCs and MSC-CM improved lung endothelial barrier integrity and the rate of alveolar fluid clearance in an ex vivo model of perfused human lungs injured by endotoxin. This effect was, at least partially, mediated by the release of keratinocyte growth factor (KFG) which contributes to restore sodium dependent alveolar fluid transport (Lee et al., 2009). Administration of MSC-CM also exerted protective effects following VILI, decreasing lung inflammation (as evidenced by the reduction in alveolar inflammatory cell counts, the decrease in IL-6 and TNF-α production or the increase in IL-10), improving systemic oxygenation and enhancing alveolar fluid clearance through a mechanism which may also be mediated through KGF secretion (Curley et al., 2012). Intratracheal administration of MSCs or MSC-CM also reduced the total number of infiltrated neutrophils in BALF and attenuated the formation of lung edema in mice with endotoxin-induced ALI (Ionescu et al., 2012). In contrast, in a rat model of E. coli-induced pneumonia administration of MSCs reduced the infiltration of neutrophils and the total amount of protein in BALF, whereas MSC-CM was less effective, improving animal survival but without significant mitigation of the severity of lung injury or inflammation (Devaney et al., 2015).

Overall, we now have a large body of evidence suggesting that MSCs act through paracrine effects, rather than (trans)differentiating and incorporating into the host tissue. In this regard, the discovery that EVs released by MSCs act as carriers of bioactive molecules (Maas et al., 2017), opens the possibility to develop new therapies based on the use of stem cells but cell-free and therefore potentially safer and amenable to standardization (Mohammadipoor et al., 2018; Lanyu and Feilong, 2019). In line with this, administration of MSCs-derived EVs (MSC-EVs) are also showing promising results in experimental models of ALI from different etiologies (Lee et al., 2019; Table 3). Accordingly, administration of human MSC-EVs to mice model of ALI induced by IT administration of endotoxin attenuated the influx of inflammatory cells, decreased the total protein content in BALF and reduced the extravascular lung water content (Zhu et al., 2014). In this study, Zhu et al. (2014) demonstrated that IT administration of MSC-EVs was able to reduce inflammation and restore the barrier function in an endotoxin-induced ALI model through a mechanism which involved KGF. In a later study, these authors confirmed that IV administration of MSC-EVs also improved survival and reduced lung protein permeability in a model of E. coli-induced pneumonia through a mechanism also mediated by KGF (Monsel et al., 2015). In addition, MSC-EVs were able to restore protein permeability in primary cultures of human alveolar epithelial type II cells, induce anti-inflammatory M2 polarization in a macrophage cell line (Zhu et al., 2014) and enhance bacterial phagocytosis by human monocytes (Monsel et al., 2015).

The direct effects of MSC-EVs on activated macrophages have also been confirmed in human monocyte–derived macrophages in the presence of BALF from ARDS patients and in murine AM (Morrison et al., 2017). In this study, the authors found that EVs derived from MSCs can modulate macrophage polarization through a mechanism involving the transfer of functional mitochondria and upregulation of oxidative phosphorylation. Furthermore, administration of AM treated with MSC-EVs, but not control macrophages, protected against endotoxin-induced lung injury, significantly reducing the protein content and the total number of neutrophils in BALF (Morrison et al., 2017). Phinney et al. (2015) also demonstrated that mitochondrial transfer by MSC is facilitated by the simultaneous release of EVs containing microRNA which suppressed TLR signaling and desensitized macrophages to the ingested mitochondria. Using miRNA microarray analysis, the authors explored the content of MSC-derived EVs obtained from five human donors, identifying 156 microRNAs (45 upregulated; 111 downregulated) that differed in abundance between EVs compared with their parent MSCs. The EVs were enriched in miR-451, miR-1202, miR-630, and miR-638, whereas miR-125b and miR-21 showed the largest reduction in MSC-EVs (Phinney et al., 2015).

MSC-EVs also have protective effects on pulmonary structural cells. In a murine model of haemorrhagic shock, both MSCs and their EVs were shown to attenuate vascular permeability in injured lungs through inhibition of RhoA GTPase activity in lungs, but through differential activation of proteins and pathways (Potter et al., 2018). In a swine model of influenza virus-induced ALI, MSC-EVs were able to inhibit the replication of influenza virus in epithelial cells and to significantly reduce the infiltration of inflammatory cells to the lungs, the thickening of alveolar walls and the number of collapsed alveoli in infected swine (Khatri et al., 2018). MSC-EVs were also effective on reducing the H₂O₂-induced epithelial cell death in vitro and hyperoxia-induced lung injuries in vivo, such as impaired alveolarization and angiogenesis, increased cell death and inflammatory responses. These protective effects against hyperoxic lung injury were apparently mediated by the transfer of vascular endothelial growth factor (VEGF) contained within these vesicles into pericytes, AM and ATII cells (Ahn et al., 2018). In endotoxin-induced lung injury, the protective effects induced by MSC-EVs on pulmonary edema were mediated partly by transfer of KGF mRNA to airway epithelial cells and angiopoietin-1 (Ang-1) mRNA to endothelial cells (Zhu et al., 2014; Tang et al., 2017; Mahida et al., 2020b). In endothelial cells stimulated with endotoxin in vitro, MSC-EVs augmented the expression of endothelial intercellular junction proteins, reduced endothelial cell apoptosis and inhibited the production of inflammatory cytokines through a mechanism involving the hepatocyte growth factor (HGF) gene carried inside the vesicles (Wang et al., 2017). Increasing evidence suggest that MSC-EVs improve alveolar-capillary barrier properties through restoration of mitochondrial functions via mitochondrial transfer. In a model of severe pneumonia induced by E. coli, MSC-EVs also augmented intracellular ATP levels in injured AECs which is critical to restore fluid clearance and surfactant production (Monsel et al., 2015). Although the mechanism was not elucidated in this study, the presence of mitochondria is critical for EV ability to reduce lung injury and restore mitochondrial respiration in the lung tissue (Dutra Silva et al., 2021). Indeed, mitochondrial transfer by MSC-EVs has been shown to play a key role in the amelioration of lung injury by modulating the phenotype of macrophages (Jackson et al., 2016) or restoring bioenergetic functions in AECs, ECs, and ex vivo cultured precision cut lung slices (Islam et al., 2012; McVey et al., 2019; Dutra Silva et al., 2021). Another explanation for the barrier-enhancing effect of MSC-EVs can be attributed to an increased VE-cadherin expression and enhanced VE-cadherin and β-catenin interaction, supporting barrier integrity (Potter et al., 2018; Simmons et al., 2019). MSC-EVs restored protein permeability across human lung microvascular endothelial cells (HLMVECs) exposed to an inflammatory insult (a mixture of IL-1β, TNF-α, and interferon γ [IFN-γ]) in part by maintaining inter-cellular junctions and preventing actin stress fiber formation. Incorporation of MSC EVs into HLMVECs through the surface receptor CD44 was required for restoration of protein permeability. The therapeutic effect of MSC EVs was associated with the transfer of Ang-1 from the EVs to the injured HLMVECs with subsequent secretion of the anti-permeability factor (Hu et al., 2018).

MSC-EVs have been shown to restore alveolar fluid clearance and reduce edema in ex vivo perfused human lungs rejected for transplantation and those injured by bacterial pneumonia (Gennai et al., 2015; Park et al., 2019). Similarly, E. coli-induced pneumonia in mice was ameliorated by MSC EVs via the well-documented barrier-stabilizing and anti-inflammatory effects of KGF (Monsel et al.,. 2015; McVey et al., 2019). Prophylactic treatment with MSC-EVs increased survival in rats undergoing traumatic lung injury; inflammatory cytokines, infiltrating leukocytes and the degree of pulmonary edema were all reduced (Li Q.C. et al., 2019).

Although MSCs have an innate potential to induce and/or contribute to regeneration, this potential is now known to be greatly influenced by diverse extrinsic factors such as the tissue source of the MSCs, the health status and age of the MSCs donor, the batch/lot of serum used for the in vitro culture of the MSCs, passage number, oxygen concentration, and the presence/absence of a pro-inflammatory environment when the MSCs are infused. Indeed, after transplantation, MSCs must confront a harsh environment which limit their survival and compromise their ability to migrate toward damaged tissues leading to unsatisfactory therapeutic results. In the pursuit of strategies to enhance the therapeutic potential of MSCs, preconditioning strategies are gathering increasing attention. In this context, one of the major challenges in MSC-based therapies is to develop methods that mimic the injury environment, without compromising cell quality and function. Thus, currently explored in vitro preconditioning strategies include environmental stimuli (such as exposure to hypoxia), treatment with cytokines or pharmacological agents, physical factor preconditioning or genetic engineering (summarized in Table 4; Ferreira et al., 2018; Han et al., 2019; Gorgun et al., 2021; Rolandsson Enes et al., 2021).

Extracellular vesicles play a crucial role in the onset and progression of inflammation and pulmonary damage during ARDS. Extensive research has identified EVs released from structural (alveolar epithelial and endothelial cells) and immune cells with a potential pathophysiological role in the disease. Thus, EVs are increased following lung injury and have been shown to induce pulmonary inflammation by promoting the recruitment and activation of immune cells, to impair alveolar-capillary barrier function by altering intercellular junction and to contribute to the structural and functional damage by inducing apoptosis in AECs and endothelial cells (Figure 1 and Tables 2, 3). EVs derived from AECs, endothelial cells and platelets can also promote coagulation, leading to the formation of microthrombi in the lungs. Furthermore, the release of lung-derived EVs into the systemic circulation might contribute to damage in distal organs and DIC. However, although the field is rapidly progressing, standardized methods to isolate the different types of EVs and determine the cellular origin of EVs are still missing. These pitfalls lead to a lack of reproducibility between different studies and slow down our understanding of how changes in the release of specific EVs or the composition of their cargo translate into differences in the functional effects of these vesicles and their specific impact on the progression and resolution of ARDS.

Extracellular vesicles also offer a unique opportunity to develop new therapeutics for the treatment of ARDS. There is increasing attention in exploiting the therapeutic potential of EVs derived from MSCs. A large number of preclinical studies have demonstrated that these EVs are able to reduce pulmonary inflammation, stimulate alveolar fluid clearance and to attenuate the lung injury through mechanisms involving the transportation of regulatory molecules (VEGF, HGF, KGF, Ang-1 or specific miRNas) and the transfer of functional mitochondria (Figure 2 and Table 3). A summary of the main findings supporting the use of MSC-derived EVs for the treatment of ARDS is provided in Table 3. Preconditioning and genetic engineering strategies are currently being evaluated to improve the therapeutic potential of these EVs, as shown in Table 4. However, the heterogeneity among the MSCs and the EVs preparations (including differences in the source, intrinsic donors’ conditions or EVs isolation procedures), make comparisons among different studies difficult. Based on these promising preclinical studies, several clinical trials are currently evaluating the safety and efficacy of MSCs and/or MSC-EVs in patients with ARDS, including 5 clinical trials valuating the effects of MSC-EVs in patients with COVID-19¹. First reports suggest that administration of MSCs is safe in patients with ARDS (Zheng et al., 2014) and may limit the cytokine storm (by reducing the levels of C-Reactive Protein, IL-6 or TNF-α) and improve oxygenation in severe COVID-19 patients (Sengupta et al., 2020). However, further studies are still needed to determine their optimal dosage, timing and route of administration.

In summary, EVs play an essential role in the pathophysiology of ARDS by modulating the onset and the progression of lung inflammation and injury. EVs also offer an extraordinary opportunity for developing novel therapeutic agents. EVs-based therapies, in contrast to administration of living cells, may be safer and easier to manage. Furthermore, MSC-EVs can be modified to deliver specific proteins or miRNA (Table 4) which could be directed to particular cell types by specific EV surface markers. Although this is an exciting field with a high translational potential, the implementation of standardized methods for EVs isolation and manipulation is essential in order to achieve reproducible results which may be validated in large multicenter clinical trials.

SE-R, LM, and RH conceived and designed the revision. SE-R, PG-R, LM, and RH contributed to writing of the manuscript. LM, RH, JAL, and FP-V revised the manuscript and approved the final version of the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.