Systemic lupus erythematosus (SLE) is a chronic autoimmune disease, associated with multiorgan damage and variable clinical manifestations (1). SLE is a multifactorial disorder, in which the phenotype is modulated by a combination of genetic, epigenetic, environmental, hormonal and immunoregulatory factors. SLE is characterized by a dysfunction of the immune system, notably a presence of autoreactive T cells and hyperactive B cells, leading to a loss of tolerance, production of autoantibodies against self-antigens formation and deposition of immune complexes, as well as a sustained systemic inflammation. SLE is classified as interferonopathy, since type I interferons (IFN-I) play a crucial role in the development of the disease (2). The chronic production of IFN-I, especially IFN-α, is a key characteristic of SLE and contributes to the autoimmune process (3). IFN-I stimulates the activation of interferon producing cells, such as plasmacytoid dendritic cells (pDCs), which are responsible for elevated levels of IFN-α in blood plasma and organs (4). The excessive signaling of IFN-I results in the increased expression of various pro-inflammatory cytokines, chemokines, and markers of immune cell activation. This, in turn, contributes to the dysregulation of the immune system and the generation of autoantibodies. Additionally, prolonged exposure to IFN-I can boost the activation and survival of autoreactive B cells and encourage the differentiation of T cells into pro-inflammatory subsets, sustaining the autoimmune response (5). SLE manifests through periods of flares and remissions, with symptoms showing considerable variation among individuals. Treatment objectives revolve around managing symptoms, mitigating inflammation, and safeguarding against organ damage. Due to its clinical heterogeneity and complex pathogenesis, SLE remains hard to diagnose and the available treatments show limited efficiency.

Long considered as platelet debris, EVs have gained critical interest from the scientific community in the recent years (6). EVs are released by all cellular organisms and can be defined by their heterogeneity. Indeed, apart from their structural definition, a cell secreted particle enclosed by a lipidic bilayer, EVs differ from their sources, biogenesis, biophysical, biochemical characteristics and functional activity (7). A standard classification can be established based on the biogenesis of EVs: exosomes are generally small sized extracellular vesicles, from endosomal origin, released by multivesicular bodies (MVBs). Ectosomes can be defined as extracellular vesicles stemming from the plasma membrane. EV biogenesis pathways and mechanisms of interaction with target cells have been extensively reviewed (8, 9). EVs are heterogenous in size and have a diameter varying in the range of the nanometer. However, EVs are not the only nanometer-sized particles secreted by cells and there is still no specific markers allowing an efficient separation of EVs. Thus, guidelines have been set by the international community regarding ways of purifying and characterizing EVs. Transmembranal proteins such as tetraspanins, cytosolic proteins such as TSG101, and other cell-dependent proteins have been used to demonstrate the EV nature and to some extent the degree of purity of an EV preparation (10). EVs have been linked with a variety of cellular functions, and can affect other cells with their surface proteins or the cargo encapsulated by EVs. EVs carry diverse proteins, nucleic acids and lipids from their parent cell, which can in turn be delivered to the recipient cell. EVs have a fundamental role in the immune system and in immune-related diseases, highlighting their potential either as a new biomarker or a therapeutic tool (7).

In the field of mesenchymal stromal cells (MSCs), EVs have emerged as a potential avenue for cell-free therapy. MSCs have been used for their immunosuppressive capabilities, with utilization in clinical trials and availability on the market. Nonetheless, the use of MSCs can carry some drawbacks and EVs have been looked into as a cell-free alternative. More specifically, as EVs carry the same biological molecules as their parent cells, the effect of MSCs is partially mediated by them with paracrine actions. EVs have shown to have the same immunomodulatory and regenerative potential than their parent cells, and have thus become a promising alternative to MSCs themselves for therapies (11, 12). More broadly, the secretome embraces not only the previously mentioned EVs but also a variety of proteins, lipids, and nucleic acids. These components may be either inside or adsorbed to EVs, or freely present in the secretome (13). This added complexity of which fractions holds the therapeutical activity and the issues of standardization regarding means of production, concentration, and characterization in the field of EVs make any conclusion towards the effect of MSC derived EVs or MSC derived secretome a complex equation.

In order to understand which fractions hold the immunomodulatory potential, the next part focuses on having a critical look over the many parameters influencing MSC-EV immune activity. Some scientific papers use the term exosomes to describe their fractions, while they could be more accurately defined as EVs or even secretome. Various downstream processes such as size exclusion chromatography (SEC), polyethylene glycol (PEG) and ultracentrifugation broadly used by the scientific community separate selectively the content of the secretome (14), and alter the composition of the protein corona of EVs (15). The protein corona is a set of proteins and other molecules adsorbed to the EVs. These proteins can be bioactive and mediate part of the immunomodulatory potential (16, 17). All of this heterogeneity of methods brings at the same time strength to the immunomodulatory response of EVs as a global effect, but brings hardship as to determine what is driving this specific effect. A number of studies showed the difference of potency between the secretome, the EV free fraction, and the fractions containing EVs. Papait et al. showed that amniotic MSCs derived conditioned media (e.g secretome) and the EV free fraction (meaning the supernatant after ultracentrifugation) maintained their immunomodulatory potential by inhibition of T cell, promotion of Tregs, shifted monocytes towards M2 instead of M1, but also reduction of the maturation of dendritic cells. The EV fraction collected after ultracentrifugation had no effect even though these EVs were uptaken (18). Another study has found no effect of the three fractions on the inhibition of T cell proliferation (19). These differences in results might be from methodological differences, specifically in the dose parameter (e.g. protein amount in ug or particles concentration measured by Nanoparticles Tracking Analysis (NTA)), cell sources of EVs, or downstream processes. A recent study has shown that the immunomodulatory potential of MSCs is independent of EVs, which runs counter to most results in the literature using EV-enriched fractions (20). González-Cubero et al. showed that EVs and soluble fractions from conditioned media promote an in vitro anti-inflammatory modulation in intervertebral disc degeneration in a “highly synergistic way”, thus highlighting that the use of the whole secretome rather than isolated EVs might be more beneficial for therapy (21). While it is still hard to understand what exactly mediates the immunomodulatory potential, the effect might come from a synergistic combination of both the soluble factors from the secretome and the EVs.

Other factors have also been shown to have an impact on the immunomodulatory potential of EVs or secretome. Indeed, the donor of primary cells, the source of MSCs, the passage doubling number, and the production method (which will be addressed later on) can influence the secretome content and thus the potency of EVs (22–25). Immortalized MSCs might be a solution for reproducible batches of secretome/EVs for therapeutic use (26).

The characterization of the secretome of MSCs shows the presence of both pro and anti-inflammatory molecules. While the secretome of MSC-EVs has shown to have high levels of IL-6 and IL-8 (27), a recent study showed that the conditioned media derived from umbilical cord (UC)-MSC promoted anti-inflammatory macrophages polarization. This despite a mostly pro-inflammatory profile of cytokines, though the authors have only investigated surface markers of M2 macrophages (28). MSC-EVs contain nucleic acids, which may bind to TLR7 and 9 as foreign nucleic acids and trigger a pro-inflammatory cascade. However, this review hasn’t found any reports of inflammation induced by MSC-EV-associated DNA (29).

The methods to assess EV immune potency are also critical. The dose of EV is differently calculated, either with particles concentration, protein concentration, or even EVs per receptor cells, which may result in some disparities in results (30). Some studies tend to isolate a specific cell type for their in vitro potency, though immune cells do not uptake homogenously EVs. Monocytes seem to uptake the highest proportion of MSC-EVs (18, 31, 32). Nonetheless, even though EVs are not uptaken by lymphocytes, studies on MSC-EVs and on EV-free fractions added to PBMCs have shown an immunosuppressive effect on T cells, thus independent of the uptake (18, 33, 34). Others have shown that MSC-EVs failed to suppress lymphocyte proliferation (35). The immunomodulatory effect of MSCs on B cells is independent of secreted EVs (36). This reinforces that the global immunomodulatory effect might not be entirely mediated by EVs, but also through soluble factors of the secretome. The effect of EVs on T cells could also be mediated in an indirect way, through the actions on EV-uptaking immune cells (37, 38).

Regarding the immunoregulatory effect of MSC-EVs on immune cells, some reviews have already discussed it thoroughly, either as a comprehensive overview (39), or more precisely in the case of SLE (40). Thus, in the next part, this review will focus on the immunomodulatory factors in the secretome of MSCs.

The immunomodulatory potential of MSC-EVs could be driven by a broad range of bioactive molecules, including proteins, nucleic acids, and lipids. Many evidences suggest the importance of specific bioactive proteins mediating the immunomodulatory effect of MSC-EVs. This review focuses on the supposed localization of the bioactive molecules, though it is important to remember that soluble proteins secreted by MSCs are also probably carried by EVs. The lack of studies comparing EV-free conditioned media and purified EV fractions makes the localization of all bioactive molecules a hard task. ( Figure 1 ) is a proposition of the immunomodulatory bioactive molecules of MSC-derived secretome and EVs.

MSCs have an important role in tissue repair and can be mobilized to the sites of tissue damage and inflammation. Thus, MSCs in vivo are often subject to inflammatory stimuli such as but not limited to, damage-associated molecular patterns (DAMPs), pathogen-associated molecular patterns (PAMPs), pro-inflammatory cytokines from activated immune cells, or hypoxia (136, 137). Attempts to recreate this environment in vitro have been used to increase the immunomodulatory potential of MSC-EVs. As a matter of fact, MSCs have high plasticity regarding their immunomodulatory potential: an anti-inflammatory environment might inhibit the immunosuppressive activity of MSCs, while a pro-inflammatory one will enhance it (138, 139). This can explain why some immunomodulatory molecules such as IDO are only produced under a pro-inflammatory environment. ( Table 1 ) summarizes all the priming methods for MSC-EV enhanced immunomodulatory potential.

Priming or conditioning MSCs with different methods have shown to induce different EV release, membrane markers, differential uptake and activation of T cell subsets (167). Priming also modifies the miRNA and protein EV cargo (168–170). Importantly, the source of MSCs and the inter-donor variability within the same source of MSCs has an impact on the response to priming and thus the improvement of immunosuppressive effect of MSC-EVs. Peltzer et al. showed that PCA failed to discriminate groups between MSC- EVs from 5 donors without priming, with IFN-γ priming and hypoxia, regarding their differential miRNA expression, showing that inter-individual variability was stronger, especially regarding their response to priming (171). Gorgun et al. showed that while priming had a significant effect on the secretome of MSCs, it did not majorly affect the miRNA in their EVs (172). A recent study was able to discriminate hypoxia and normoxia group looking at their miRNA, possibly underlying a donor disparity between studies (173). Jin et al. showed that priming overcame the MSC inter-donor variability by looking at gene expression (174). Priming with a cocktail of cytokines resulted in two different responses in different donors of MSCs (140). These studies show that priming MSC has significative changes on the biophysical, biochemical and bioactive properties of MSC-EVs, but the changes might be hidden by donor- and source-dependent differences.

One way to recreate the inflammatory conditions of tissue damage is to target Toll-Like Receptors (TLRs), which recognize various types of molecules such as DAMPs and PAMPs. LPS-preconditioned MSC-EVs were more efficient in converting THP-1 to M2 phenotype in vitro, and in relieving inflammation in vivo than untreated MSC-EVs, notably through miR-let-7b (141). LPS-primed MSC-EVs were significantly more efficient at increasing M2 and reducing M1 polarization in vitro, and were more effective in relieving post-infarction inflammation in mice (142). Hwang et al. showed that TLR-3 and 4 primed MSCs secretomes were more successful in reducing pro-inflammatory cytokines from LPS-induced macrophages. EVs were key in increasing the percentage of M2 (143). Notably, Kink et al. showed that the concentration of LPS for priming MSC resulted in EVs having different effects on macrophage receptors expression, which might explain the effect of the next studies (144). Indeed, some studies show that TLR-priming does not always result in enhanced immunosuppressive effect. Recent results from Cui et al. showed that LPS primed MSC-EVs induced a M1 and not a M2 phenotype (145). Previous results from Waterman et al. showed a pro-inflammatory phenotype MSCs after TLR-4 priming, and anti-inflammatory phenotype after TLR-3 priming (146). All in all, these results show that priming through TLRs might be an interesting method to prime MSC-EVs, though some optimization of concentration of priming might be important to achieve the highest enhanced immunosuppressive effect.

A second way to prime MSCs is to use pro-inflammatory cytokines. A study by Cheng et al. showed that priming with IFN-γ and TNF-α resulted with higher suppression of T cell proliferation and induced a different protein profile with higher levels of anti-inflammatory proteins such as TSG-6 and A20 (147). Ragni et al. showed that IFN-γ priming changes the proteins secreted by MSCs and the miRNA content of their EVs, resulting in a diminution of M1 polarization (148). Priming with IL-1β resulted in higher anti-inflammatory activity of EVs, with a significant increase of mir-147b, which partially mediated the immunomodulatory effect of primed MSC-EVs in osteoarthritis cells (149). IL-1β priming on mouse MSC resulted in an enhanced macrophage polarization to M2 in vivo and in vitro, and a better therapeutic effect on septic mice by MSC-EVs. Equivalent priming also induced higher expression of miR-21, which mediated the effect of MSC-EVs on sepsis (150). Nonetheless, conditioning does not always result in better activity: priming with IFN-γ did not induce a better therapeutic effect in T-cell modulation activity, and induced a loss of protection against ischemic acute kidney injury (151). A few teams have also tried cocktails of different pro-inflammatory molecules, compared to one-molecule priming. A combination of IFN-γ and TGF-β priming resulted in a higher proportion of Treg cells after treatment with EVs, but also elevated levels of IFN-γ, IL-10 and IDO within these EVs. The combination of priming resulted in better immunomodulatory effect than EVs derived from untreated MSC, or MSCs treated with only one of the molecules (101). Hackel et al. showed a variation of response between MSCs when treated with multicytokine combination of IFN-γ, TNF-α and IL-1β. One group of 3 donors secreted more PD-L1 with the full priming compared to two-cytokines priming, and the other group of 3 other donors responded equally with 3 and 2-cytokine priming. Nonetheless, priming still induced higher levels of PD-L1 and PD-L2 compared to non-primed EVs. Higher therapeutic effects of primed MSC-EVs were mediated by PD-1 ligands (140). Priming MSCs with pro-inflammatory components has been shown to modify the whole secretome, elevating the level of anti-inflammatory molecules in the secretome and in EVs, thus mediating higher potency and therapeutical effects of MSC-EVs. Priming with higher number of cytokines seems to induce a higher immunomodulatory effect of EVs.

Other molecules have also been used to prime MSCs. Conditioning with Atorvastatin resulted in a better cardioprotective effect on infarcted rat heart, with a better inhibition of TNF-α and Il-6 in the tissue of the infarct zone (152).

Genetic modifications of MSC producing EVs can improve the immunomodulatory potential of MSC-EVs. A few studies already described above have targeted specific immunomodulatory proteins such as IDO (100), IL-10 (44), TSG-6 (52), TGF-β, PTX3, miR-let7-5p and miR-21-5p (25), CD73 (66), PD-L1 (77–79), in order to secrete more of these specific proteins and induce a better immunoregulatory activity of MSC-EVs. Genetic modifications can also target the expression of bioactive miRNAs. Overexpression of miRNA-181a induced better inhibition of the inflammatory response, increased the percentage of Tregs among PBMCs, and delayed ischemic damage in vivo (153). Overexpression of miRNA-126 promoted functional recovery after stroke by suppressing microglia activation (154). Genetic modifications can also target molecules which are not constitutively expressed by MSCs. A study has shown that modified MSCs for the expression of TRAIL secrete EVs with TRAIL, and are able to induce apoptosis in various cell lines (155).

Conditioning of MSC in hypoxia has also been shown to improve the immunomodulatory potential of MSC-EVs. Hypoxic environment leads to increased levels of hypoxia-inducible factors, notably HIF-1a, which regulates many physiological pathways, including angiogenesis (175). The effect of hypoxia on the immune-modulatory properties of BM-MSCs has already been reviewed (176). Extracellular vesicles derived from hypoxia conditioned MSCs were more effective in decreasing levels of pro-inflammatory cytokines and in shifting microglia from M1 to M2 polarization, compared to non-treated MSC-derived EVs (156). Same results were observed by another group (157). Hypoxia was also shown to increase the expression of HGF (172), which we have already discussed the immunomodulatory effect. Gómez-Ferrer et al. showed that double-primed hypoxia + inflammation MSC-derived EVs were more efficient in repolarizing M1 to M2-like phenotype than single-primed inflammation MSC-EVs. These EVs promoted healing in a TNBS induced mouse colitis, partially through reduction of pro-inflammatory cytokines (80). Hypoxic UC-MSC-derived EVs inhibited more efficiently maturation of DCs (177). Hypoxia treated AD-MSC EVs were more efficient in reducing macrophage infiltration, reducing levels of Il-6, though MCP-1 levels were higher, compared to non-treated MSC-EVs, in renal tissue after ischemia reperfusion injury (158). Thus, hypoxia conditioning of MSCs could be a relevant technique to enhance the MSC-EV immunosuppressive functions for chronic diseases.

Finally, a few studies have reported higher potency with EVs derived from MSC cultured in 3D. 3D culture of MSCs englobes a great variety of techniques, including but not limited to microcarriers, diverse scaffolds, microgels, and spheroids, which can be then cultured in bioreactors or other types of vessels. 3D culture better mimics the natural cell conditions, and increases the levels of secreted angiogenic and immunomodulatory factors (178).

Bulati et al. were able to differentially cluster between IFN-γ primed and 3D spheroid cultured MSCs by looking at their miRNA EVs (179). Many studies have shown that MSCs have increased immunomodulatory potential after 3D spheroid culture (159–161). EVs derived from pro-inflammatory primed MSC spheroids had an enhanced anti-inflammatory effect by decreasing the expression of NF-κB, IL-8 and IL-6 in TNF-induced inflammation in HK2 cells, compared to 3D cultured only EVs and 2D cultured EVs (162). Nonetheless, Kusuma et al. have shown contrasting results where 3D spheroid cultured MSC-EVs produced significantly less IDO, and that overall, these EVs had a lower immunosuppressive and therapeutic potency than 2D MSC-EVs (163). Further studies are needed to understand the impact of MSC-EVs derived from spheroids.

Regarding MSC culture in bioreactors, 3D cultured in WAVE bioreactor MSC-EVs induced same decrease in M1 markers expression in macrophages than 2D MSC-EVS, but had enhanced suppression of CD8+ T cell proliferation (164). Hollow Fiber bioreactor systems have gained popularity for the 3D culture of MSC for EV research. 3D cultured MSC-EVs using this system were more efficient than 2D cultured MSC-EVs in alleviating acute kidney injury, notably by reducing inflammatory factors, repressing T cell and macrophage infiltration (165). Sun et al. also used the hollow fiber system, and found that 3D-MSC EVs exhibited a stronger anti-inflammatory effect on stimulated monocytes, but also in acute myocardial infarction rats (166). Thus, 3D culture in bioreactor seems to enhance the immune-modulatory potential of MSC-EVs.

Other than MSCs, other EVs have been showing immunosuppressive potential for a therapeutic approach. Namely, tumor cells like sarcoma cells evade the immune system, which is partially mediated by the release of EVs (180, 181). EVs derived from Ewing Sarcoma induced a pro-inflammatory response on myeloid cells, but impaired the maturation and function of dendritic cells (182). Droste et al. wrote an in-depth review about tumor derived EVs, their effect on immune cells and how in vivo animal models help understand the potential of these EVs (183).

Based on their broad immunosuppressive effects and the possibility to enhance them through a variety of techniques, MSC-EVs could be a promising therapeutic solution for the treatment of SLE.

The emerging field of EVs presents a promising avenue for therapeutic treatment. EVs carry a variety of membrane and soluble proteins, and play a key role in immune processes. Thus, EVs could be used for diagnostic as a biomarker, or a therapeutic tool. More specifically, MSC-EVs mediate a broad immunosuppressive effect, showcasing their potential as a cell-free therapy for SLE. Further techniques such as pre-conditioning of MSCs, genetic modification or EV engineering could enhance their immunomodulatory activity and could be applied to further therapeutic applications of EVs.

However, a more specific understanding of whether the immunoregulatory activity is mediated by EV-associated bioactive molecules, soluble factors, or both is needed. The impact of the many sources of heterogeneity in EV studies on these immune mediators should be investigated. The EV field still suffers from barriers such as standardization of isolation and characterization methods, Good Manufacturing Practice (GMP)-compliant large scale production, or specific guidelines for validation of EVs as a therapeutic tool, which in turn hinders the use of EVs for the treatment of SLE. Furthermore, regarding SLE, the role of EVs on pDC activation should be further investigated to understand their potential role on chronic production of IFN-I in SLE. Similarly, a fine characterization of pDC derived EVs content should be carried out prior considering the use of EVs as therapeutic strategy. The effect of MSC-EVs regarding the regulation of interferon production of pDCs should be investigated. Finally, methodological studies on the dosage and administration interval of MSC-EVs in SLE are still essential to advance their therapeutic development.

CW: Writing – original draft, Writing – review & editing. IS: Writing – original draft, Writing – review & editing. FG: Writing – original draft, Writing – review & editing. J-PH: Writing – original draft, Writing – review & editing. TF: Writing – original draft, Writing – review & editing.