Extracellular Vesicles isolated from Mesenchymal Stromal Cells Modulate CD4+ T Lymphocytes Toward a Regulatory Profile.

Mesenchymal stromal cells (MSCs) can generate immunological tolerance due to their regulatory activity in many immune cells. Extracellular vesicles (EVs) release is a pivotal mechanism by which MSCs exert their actions. In this study, we evaluate whether mesenchymal stromal cell extracellular vesicles (MSC-EVs) can modulate T cell response. MSCs were expanded and EVs were obtained by differential ultracentrifugation of the supernatant. The incorporation of MSC-EVs by T cells was detected by confocal microscopy. Expression of surface markers was detected by flow cytometry or CytoFLEX and cytokines were detected by RT-PCR, FACS and confocal microscopy and a miRNA PCR array was performed. We demonstrated that MSC-EVs were incorporated by lymphocytes in vitro and decreased T cell proliferation and Th1 differentiation. Interestingly, in Th1 polarization, MSC-EVs increased Foxp3 expression and generated a subpopulation of IFN-γ+/Foxp3+T cells with suppressive capacity. A differential expression profile of miRNAs in MSC-EVs-treated Th1 cells was seen, and also a modulation of one of their target genes, TGFbR2. MSC-EVs altered the metabolism of Th1-differentiated T cells, suggesting the involvement of the TGF-β pathway in this metabolic modulation. The addition of MSC-EVs in vivo, in an OVA immunization model, generated cells Foxp3+. Thus, our findings suggest that MSC-EVs are able to specifically modulate activated T cells at an alternative regulatory profile by miRNAs and metabolism shifting.


Transmission Electronic Microscopy (TEM)
After ultracentrifugation, MSC-EVs were resuspended in a 2% paraformaldehyde solution. The particles suspension was dripped onto carbon-coated electron microscopy screens and adsorbed for 20 min. The screens were fixed with glutaraldehyde 1% and washed with deionized water. Subsequently, the screens were contrasted with uranyl acetate for 10 min and rinsed again with distilled water and air dried. The images were acquired and observed in a JEOL 1200 EX II transmission electron microscope at 80 kV.

Detection and Incorporation of EVs
MSC-EVs were labeled with the fluorescent red dye PKH26 (Sigma) following the manufacturer's instructions and subjected to ultracentrifugation for the washes required to remove excess of dye. Labeled EVs were added to the culture of naive CD4 + T lymphocytes purified by FACS sorting (FacsAria-BD) and activated with anti-CD3 (BD-clone145-2C11) and anti-CD28 (BD-clone 37.51) for evaluation of the internalization through the imaging by confocal microscopy (Zeiss LSM 780-NLO). Lymphocytes were monitored overnight for approximately 15 h.

miRNA PCR Array and in Silico Analysis
Total RNA of Th1-differentiated cells was isolated using miRNeasy Mini Kit (Qiagen). cDNA was synthesized from 300 ng of the mRNA using the miScript II RT kit (Qiagen). PCR array of miRNAs were run in 96-well plates for each sample (3 samples per group) following the instructions of the miScript MIMM-111Z-T Cell & B Cell Activation miRNA PCR Array (Qiagen) assay. Analysis were performed using the Qiagen website. miRNAs differentially expressed between the groups were selected and only the two miRNAs more significantly upregulated and two more significantly downregulated were considered. Three different online databases (TargetScan7 [32], miRDB [33] and Starbase [34]) were used to obtain the targets of these miRNAs. The InteractiVenn [35] website was used to select the intersections between the target genes empirically obtained from the different databases. Enrichr [36] platform was used to obtain the correlations between these target genes and possible signaling pathways in which these genes are involved. Thus, 4 signaling pathways (KEGGS pathway) that could be modulated by the miRNA expression changes were found. Analyzing the signaling pathways involved, we selected 10 of the target genes to evaluate their expression by RT-PCR.

Glycolytic Stress Test-Seahorse
Extracellular acidification rates (ECAR) were measured using Extracellular Flux Analyzers (Seahorse Bioscience). After Th1 differentiation, cells were plated to XF assay media without glucose. Glucose 10 mM (Sigma-Aldrich), 1-µg/mL oligomycin (Sigma-Aldrich), 22 mM 2-deoxiglicose (2-DG) (Agilent) and media were added in this order using the ports on the XF96 cartridges. The data were collected using the XF Reader software (Seahorse Bioscience). Glycolysis was calculated by the difference between ECAR rates after glucose injection and the basal rate (before glucose injection). Glycolytic capacity was calculated by the difference between ECAR rates reached after oligomycin injection and the ECAR rates reached before glucose injection. The glycolytic reserve was calculated by the difference between glycolytic capacity and glycolysis rate. Finally, ECAR rate prior to glucose injection was determined as non-glycolytic acidification.

Mitochondrial Membrane Potential Detection Assays
Mitochondrial membrane potential ( ψ m ) was detected in cells after differentiation. Cells were incubated with TMRE (Abcam) (0.3 µM) or MitoTracker Deep Red FM (Life Technologies) (0.3 µM) for 30 min at 37 • C. For evaluation by FACS the cells were also labeled with live/dead and anti-CD4 and the fluorescence intensity of TMRE and MitoTracker was calculated by MFI (Median of Fluorescence). To get the images the cells were labeled concomitantly with Hoescht, washed and immediately subjected to imaging of 5 fields per well using a 40x magnification in the InCell Analyzer 2200 (GE) equipment. For quantitative analysis of fluorescence intensity and percentage of positive cells, InCell Investigator software was used. As a control, cells were incubated with CCCP (Carbonyl cyanide m-chlorophenyl hydrazone) (10 µM) for 30 min.

Statistical Analysis
Data were analyzed by ANOVA or Student's t-test. All results are presented as mean and standard deviation. Values of p < 0.05 were considered statistically significant.

Isolation and Characterization of MSC-EVs
The adipose tissue-derived MSCs used in this study was provided by a cell bank which cells were previously characterized by our group [38]. A scanning electron microscope was used to demonstrate that MSCs release EVs of different sizes and origins ( Figure 1A,B). TEM analysis showed that MSC-EVs presented a spheroid shape, with a bi-lipid membrane structure and varied sizes, representing a mixed population of smaller vesicles, that could be exosomes (approximately 100 nm) and larger vesicles, for example, microvesicles (between 100 and 1000 nm) ( Figure 1C). MSC-EVs presented a mean size of 150-200 ηm with concentration, analyzed by NTA, of approximately 10 11 particles/mL ( Figure 1D). Expression of MSCs markers by EVs was detected using FACS. Calibration beads of 1 µm were used to adjust the parameters of size (FSC) and granularity (SSC) ( Figure 1E). Over half of EVs expressed CD9 ( Figure 1F) and the CD9 + particles were positive for CD73, CD90 and CD105 and negative for CD45 ( Figure 1G-J), as observed at MSCs [2]. In addition, The CytoFLEX also was used to detect and better characterize the MSC-EVs. Using Gigamix beads the MSC-EVs were gated according to their corresponding size ( Figure 1K,L) and they expressed classical EVs markers as annexin and CD9 ( Figure 1M,N).

MSC-EVs are Incorporated by CD4 − T Cell and Alter Lymphocyte Proliferation and Differentiation
To investigate the biodistribution of MSC-EVs, confocal microscope assay was performed and the MSC-EVs were visually incorporated by CD4 + T lymphocytes (Figure 2A,B). In order to evaluate the functional effect of the MSC-EVs on lymphocytes proliferation, total splenocytes were labeled with Cell trace violet and stimulated with anti-CD3. The proliferation was evaluated by FACS 3 days after the stimulus. MSC-EVs were able to reduce the proliferation of activated CD4 T lymphocytes by approximately 50% (Figure 2C by approximately 50% (Figure 2C-E).
Since we observed that MSC-EVs inhibited CD4 + T cell proliferation, we next sought to assess whether the MSC-EVs would influence on T cell differentiation. Then, we sorted naïve CD4 + T cells and polarized them to differentiate to Th1, Th17 and Treg cells in the presence of MSC-EVs. The presence of MSC-EVs significantly affected Th1 differentiation as observed by the decreasing of IFNγ production ( Figure 3A,B), although nothing was observed regarding the IL-17A production in the presence of MSC-EVs ( Figure 3C,D). In the Treg differentiation, it was detected higher numbers of Foxp3 + cells within MSC-EVs-treated group when compared to untreated ones, however without statistical differences ( Figure 3E,F). Further, to access the precise effect of MSC-EVs in differentiated Tregs, we purified mature Tregs (Foxp3 + GFP + ) by cell sorting from spleen and lymph nodes of FOXP3-GFP knock in mice and stimulated them with plated-bound anti-CD3 and soluble anti-CD28 in the presence of MSC-EVs. Again, we did not observe any statistical improvement of the foxp3 expression, even with higher number of Foxp3 + cells detected in the MSC-EVs-treated group ( Figure  3G,H). Thus, these results altogether indicate that MSC-EVs can modulate the Th1 differentiation.  Since we observed that MSC-EVs inhibited CD4 + T cell proliferation, we next sought to assess whether the MSC-EVs would influence on T cell differentiation. Then, we sorted naïve CD4 + T cells and polarized them to differentiate to Th1, Th17 and Treg cells in the presence of MSC-EVs. The presence of MSC-EVs significantly affected Th1 differentiation as observed by the decreasing of IFN-γ production ( Figure 3A,B), although nothing was observed regarding the IL-17A production in the presence of MSC-EVs ( Figure 3C,D). In the Treg differentiation, it was detected higher numbers of Foxp3 + cells within MSC-EVs-treated group when compared to untreated ones, however without statistical differences ( Figure 3E,F). Further, to access the precise effect of MSC-EVs in differentiated Tregs, we purified mature Tregs (Foxp3 + GFP + ) by cell sorting from spleen and lymph nodes of FOXP3-GFP knock in mice and stimulated them with plated-bound anti-CD3 and soluble anti-CD28 in the presence of MSC-EVs. Again, we did not observe any statistical improvement of the foxp3 expression, even with higher number of Foxp3 + cells detected in the MSC-EVs-treated group ( Figure 3G,H). Thus, these results altogether indicate that MSC-EVs can modulate the Th1 differentiation. Cells 2020, 9, x 9 of 29

MSC-EVs Induce Foxp3 Expression in Th1 Differentiated Cells
In order to understand how MSC-EVs inhibited Th1 differentiation we decided to investigate the generation of Foxp3 + cells during Th1 differentiation in the presence of MSC-EVs, since this is one

MSC-EVs Induce Foxp3 Expression in Th1 Differentiated Cells
In order to understand how MSC-EVs inhibited Th1 differentiation we decided to investigate the generation of Foxp3 + cells during Th1 differentiation in the presence of MSC-EVs, since this is one defined mechanisms of MSCs immune regulation [6,8,9]. As expected, the addition of MSC-EVs decreased IFN-γ production visualized by reduced number of red dye-labeled cells ( Figure 4A,C,D). Notably, MSC-EVs addition in CD4 + naïve T cells polarized to Th1 increased the frequency Foxp3 + cells and expanded Foxp3 expression (green dye-labeled cells) ( Figure 4A,C,E). This finding of increase in Foxp3 + cells in Th1 cells in the presence of MSC-EVs was confirmed by confocal microscopy and by flow cytometry (Figure 4A,B). The addition of MSC-EVs increased the frequency of Foxp3 + cells ( Figure 4C,E). Surprisingly, IFN-γ and Foxp3 double-positive cells were also identified after treatment with MSC-EVs ( Figure 4A,B,F).
Expression of IFN-γ and Foxp3 was also evaluated by RT-PCR, as well as Tbet (Th1-specific transcription factor) expression. Although a reduction of IFN-γ by FACS and confocal microscopy was demonstrated, we did not see a difference in the expression of IFN-γ mRNA ( Figure 4G). The same was observed for Tbet expression ( Figure 4I). However, the Foxp3 increase was confirmed by real-time PCR analysis ( Figure 4H).
In an attempt to confirm functional regulatory profile of these MSC-EVs-modulated T cells generated, we performed a suppression assay using differentiated Th1 cells in the presence or absence of MSC-EVs in co-culture with total splenocytes at different ratios ( Figure 5). As control, Tregs cells were purified by sorting and co-cultured with total splenocytes, also in different ratios ( Figure 5C-F). In the 2:1 ratio, we observed that MSC-EVs-modulated T cells (Foxp3 + IFN-γ + ) in the presence of MSC-EVs were able to suppress CD4 + T cell ( Figure 5A) and CD8 + T ( Figure 5B) proliferation. The suppressive effect of these cells was similar to Tregs at 1:4 ratio ( Figure 5E).
These results suggest that cells differentiated to Th1, in the presence of MSC-EVs, are reprogrammed to a more regulatory profile, decreasing the frequency of IFN-γ producing cells and increasing Foxp3 expressing cells, which are able to suppress the proliferation of total splenocytes.

The Global Analysis of miRNA Array Reveals Possible Targets of miRNA Regulation in Lymphocytes Treated with MCS-EVs
The transfer of miRNA has been considered the main mechanism by which EVs exert their effects [39]. Therefore, in order to detect differentially expressed miRNAs in Th1 cells and possible molecular signature involved in the Th1 regulation by MSC-EVs, a specific platform for detection miRNAs associated with signaling pathways in B and T cell activation (QIAGEN) was used. This kit comprises a panel with 84 miRNAs involved in the differentiation of lymphocytes ( Figure 6A). We detected 5 upregulated and 53 downregulated miRNAs in T cells differentiated into Th1 in the presence or not of MSC-EVs (fold change> 1.5) ( Figure 6B). Further, we verified the most regulated transcript and only 3 miRNAs had significantly its expression decreased and 13 showed to be statistically increased (p < 0.05) ( Figure 6C,D). In an attempt to detect signaling pathways and genes regulated by these differentially expressed miRNAs, an in silico analysis was carried out, in which the most modulated miRNAs were addressed (upregulated: miR-19a-3p and miR106a-5p and downregulated: miR23a-3p and miR-21a-5p). The target genes of each miRNA were found in 3 different and independent databases ( Figure 7A,C). Putative target genes commonly found in these 3 databases were analyzed in the ENRICHR website to detect possible signaling pathways involved ( Figure 7B,D). Among the pathways regulated by the positively regulated miRNAs (pathways that would be less active), the AMPK and MAPK pathways were selected ( Figure 7B and Table S1), whereas the pathways regulated by reduced miRNAs (pathways that would be more active), the TGF-β pathway and FoxO pathway were selected ( Figure 7D and Table S2). The target genes that participate in each pathway analyzed were identified and some genes were validated by RT-PCR (Tables S1 and S2). TGFBR2 expression was the only gene evaluated which had its expression regulated ( Figure 7E-G).

The Global Analysis of miRNA Array Reveals Possible Targets of miRNA Regulation in Lymphocytes Treated with MCS-EVs
The transfer of miRNA has been considered the main mechanism by which EVs exert their effects [39]. Therefore, in order to detect differentially expressed miRNAs in Th1 cells and possible molecular signature involved in the Th1 regulation by MSC-EVs, a specific platform for detection miRNAs associated with signaling pathways in B and T cell activation (QIAGEN) was used. This kit comprises a panel with 84 miRNAs involved in the differentiation of lymphocytes ( Figure 6A). We Expression of AMPK pathway genes PPP2R5E ( Figure 7F) and PI3KR3 ( Figure 7G), which are target genes of the upregulated miR-19a-3p, was not altered. Since TGFBR2 is regulated by miRNA 23a-3p, which was downregulated in the presence of MSC-EVs, its expression was increased in the MSC-EVs treated group ( Figure 7E). These results confirm, in part, the participation of miRNAs in Th1 cell modulation by MSC-EVs and suggests that TGF-β signaling pathway (miRNA-23a-3p/TGFBR2) may modulate Th1 differentiation.

Treatment with MSC-EVs Alters the Metabolism of Differentiated T Cells to Th1, Decreasing Mitochondrial Membrane Potential and Glycolysis
It is well known that TGF-β pathway can regulate metabolic process through the regulation of mTOR pathway, which can regulate T cell metabolism. Therefore, in order to further investigate

Treatment with MSC-EVs Alters the Metabolism of Differentiated T Cells to Th1, Decreasing Mitochondrial Membrane Potential and Glycolysis
It is well known that TGF-β pathway can regulate metabolic process through the regulation of mTOR pathway, which can regulate T cell metabolism. Therefore, in order to further investigate possible determinant mechanisms by which MSC-EVs exert effects on T lymphocytes, as a consequence TGF-β pathway modulation observed in the miRNA profile, we searched for specific metabolic and mitochondrial changes in T cells. Genes involved in metabolic regulation (ACLY, HK2, PKM2 and PDk1) were evaluated ( Figure 8A-D) and PKM2 showed reduced expression after treatment with EVs ( Figure 8C), while no statistical difference was observed in the other molecules evaluated.  Th1 cells depend on glycolysis to produce IFN-γ [40,41]. On the other hand, inhibition of glycolysis favors Treg differentiation [40]. Therefore, we performed a Seahorse analysis in order to obtain the extracellular acidification rate (ECAR), for direct measurement of glycolysis index. A prominent decreasing in ECAR rate in the cells differentiated in the presence of MSC-EVs was observed ( Figure 8E). In addition, we detected several metabolic parameters decreased in T cells treated with MSC-EVs, as glycolysis rate ( Figure 8F), glycolytic capacity ( Figure 8G) and glycolytic reserve ( Figure 8H), considering that no statistical difference was seen at non-glycolytic acidification ( Figure 8I) rate. All these data confirm that T cells are less glycolytic after MSC-EVs addition and this may be affecting their IFN-γ production.
To verify whether MSC-EVs can further impact mitochondrial metabolism, lymphocytes were labeled with TMRE and MitoTracker and analyzed by FACS and microscopy (InCell Analyzer). T cells treated with MSC-EVs demonstrated a decrease in mitochondrial membrane potential ( ψ m ), represented by the reduced TMRE and MitoTracker fluorescence intensity when evaluated by FACS ( Figure 9A,B) and by low frequency of TMRE-positive cells when evaluated by microscope ( Figure 9E,G). Besides no difference in intensity of TMRE fluorescence ( Figure 9D,G), this was confirmed by the decrease of MitoTracker fluorescence ( Figure 9C,F). As a control, CCCP, a decoupling agent inhibiting oxidative phosphorylation, was added and a ψ m reduction was observed.
These results suggest that MSC-EVs can regulate metabolic pathways in Th1 differentiated cells, which may also be related to the reduction of IFN-γ and generation of Foxp3 + cells and to the modulation of the TGF-β pathway.

MSC-EVs Treatment Expands Tregs In Vivo
To verify the regulatory effects of MSC-EVs on Th1 differentiation in vivo, we choose a murine immunization model with OVA. After 7 days, the animals were euthanized, and the draining lymph nodes were collected for analysis. It was detected that immunization with OVA significantly increased the number of infiltrating cells in the lymph nodes ( Figure 10A). After 7 days, the number of CD4 + T cells expressing IFN-γ after immunization showed a tendency to increase and no effect was observed after MSC-EVs injection ( Figure 10B). On the other hand, the number of Foxp3-expressing CD4 + T cells was dramatically elevated after treatment with MSC-EVs ( Figure 10C). Finally, these results confirm the ability of MSC-EVs to expand Foxp3 + T cells during an inflammatory response in vivo, as it was seen in the in vitro assays.

Discussion
In this study, we demonstrate the ability of MSC-EVs to regulate Th1 cells potentially via modulation of miRNA profile associated TGF-β pathway and metabolism shifting. MSC-EVs reduced differentiation to Th1, generating cells that express Foxp3 with reduced IFN-γ production and increased suppressive capacity. The addition of MSC-EVs changed miRNAs expression in the Th1 cells, increasing TGFBR2 expression, as a consequence of the reduction of mir-23a-3p. The regulation of TGF-β pathway can be related with the regulation of metabolic pathways. Accordingly, we showed a reduction in glycolytic and mitochondrial metabolism in the cells differentiated in the presence of MSC-EVs. Our data confirm the capacity of MSC-EVs in regulating T cells and open new perspectives to the use of EVs as a therapeutic alternative to MSCs.
During EVs isolation, we excluded the larger vesicles and possible cellular debris of our MSC-EVs population by centrifuging the cell supernatant at 2.000× g. After centrifugation at 100.000× g, we obtained a mixture of EVs with the presence of exosomes, exomeres and MVs. Opting for not do the 10,000× g ultracentrifugation, we don't differentiate medium size particles from smaller size ones [42]. Accordingly, the presence of different vesicle sizes was detected by scanning and transmission electron microscopy and NTA. Analysis of size distribution by NTA suggests, however, a predominant presence of smaller particles. Although several studies indicate that different fractions of EVs may have different [43] or even opposites effects, [44,45], the choice of using this mixture of particles is mainly because EVs may be released simultaneously by the cells. Additionally, the yield of EVs from MSC culture supernatant is not high and large amounts of particles from purified EVs are needed to obtain modulatory effect. The mixed population of EVs obtained showed the molecules from parental cells, expressing MSCs [27,46] markers, and were mostly positive for Annexin and CD9, classical EVs markers.

MSC-EVs Treatment Expands Tregs in Vivo
To verify the regulatory effects of MSC-EVs on Th1 differentiation in vivo, we choose a murine of CD4 + T cells expressing IFN-γ after immunization showed a tendency to increase and no effect was observed after MSC-EVs injection ( Figure 10B). On the other hand, the number of Foxp3-expressing CD4 + T cells was dramatically elevated after treatment with MSC-EVs ( Figure 10C). Finally, these results confirm the ability of MSC-EVs to expand Foxp3 + T cells during an inflammatory response in vivo, as it was seen in the in vitro assays.

Discussion
In this study, we demonstrate the ability of MSC-EVs to regulate Th1 cells potentially via modulation of miRNA profile associated TGF-β pathway and metabolism shifting. MSC-EVs reduced differentiation to Th1, generating cells that express Foxp3 with reduced IFN-γ production and increased suppressive capacity. The addition of MSC-EVs changed miRNAs expression in the Th1 cells, increasing TGFBR2 expression, as a consequence of the reduction of mir-23a-3p. The regulation of TGF-β pathway can be related with the regulation of metabolic pathways. Accordingly, we showed a reduction in glycolytic and mitochondrial metabolism in the cells differentiated in the presence of MSC-EVs. Our data confirm the capacity of MSC-EVs in regulating T cells and open new perspectives to the use of EVs as a therapeutic alternative to MSCs.
During EVs isolation, we excluded the larger vesicles and possible cellular debris of our MSC-EVs population by centrifuging the cell supernatant at 2.000 g. After centrifugation at 100.000 g, we obtained a mixture of EVs with the presence of exosomes, exomeres and MVs. Opting for not do the 10,000 g ultracentrifugation, we don't differentiate medium size particles from smaller size ones [42]. Accordingly, the presence of different vesicle sizes was detected by scanning and transmission electron microscopy and NTA. Analysis of size distribution by NTA suggests, however, a predominant presence of smaller particles. Although several studies indicate that different fractions of EVs may have different [43] or even opposites effects, [44,45], the choice of using this mixture of particles is mainly because EVs may be released simultaneously by the cells. Additionally, the yield of EVs from MSC culture supernatant is not high and large amounts of particles from purified EVs are needed to obtain modulatory effect. The mixed population of EVs obtained showed the molecules from parental cells, expressing MSCs [27,46] markers, and were mostly positive for Annexin and CD9, classical EVs markers. PKH dyes have been widely used for labeling extracellular vesicles [47,48], however, their use has been controversial. Recent studies have shown that nanoparticles of dye can also be internalized [49,50], but in a lower extension as EVs-labeled particles [49]. When we evaluated the incorporation of MSC-EVs by lymphocytes using PKH26, we found an accumulation of EVs on the cell surface. Although it is not possible to differentiate whether there was accumulation in the cytosol or fusion with the plasma membrane, studies have suggested that EVs naturally associate with the lymphocyte membrane [43,51] exerting their modulatory effects. In this sense, the addition of MSC-EVs significantly reduced the proliferation of T cells. The literature is controversial regarding the effects of MSC-EVs on lymphocyte proliferation. Some studies have shown inefficiency of MSC-EVs in inhibiting T cell proliferation [52][53][54], others have demonstrated lower efficiency of EVs when compared to MSCs [55][56][57][58] while others have reported ability to inhibit lymphocyte proliferation [59][60][61]. The difference between these studies may be due to variations in the dose and types of EVs and T cell sources used.
Functionally, authors have reported that MSC suppression on T-cell proliferation is more related to cell cycle inhibition than apoptosis induction [62]. To confirm that our EVs were not killing the cells, we performed live/dead labeling by cytometry and did not see changes in cell viability following the addition of MSC-EVs in our culture conditions ( Figure S1). In order to evaluate whether the MSC-EVs could alter specifically T cell response, we differentiated naïve CD4 + T cells into different T cell subtypes: Th1, Th17 and Treg. It was previously shown that MSCs or MSC-EVs can alter the balance between different T cell subtypes [63,64]. In a study using PBMC cells from type I diabetes patients, MSC-EV administration decreased Th17 response while increased Treg response [65]. Based on that, we differentiated lymphocytes to a pro-inflammatory condition (Th1) and as it was shown previously with MSCs treatment [66,67], a significant reduction of IFN-γ-secreting cells was observed after treatment with MSC-EVs. We then differentiated to another pro-inflammatory condition, Th17, but we did not see effect on T cell differentiation. When we evaluated the effect over Treg differentiation, we detected a trend to increase the percentage of cells expressing Foxp3. The same was observed when we evaluated the effects directly on purified Tregs from the spleen. The different behavior of EVs in each condition may indicate a dependency of a favorable environment for them to perform their adequate functions, which we believe being preferentially associated with inflammatory microenvironment. For instance, in a model of skin transplantation, MSC-EVs expanded Tregs and increased animals survival only in animals that received the graft [68]. Accordingly, the presence of cytokines may influence the effects of MSC-EVs. One study observed that the presence of TNF along with EVs may promote greater effect on target cells than using EVs or TNF alone, suggesting that they may act synergistically [69]. Therefore, cytokines present in a Th1 condition, such as IL-12 or IFN-γ, may have intensified the effects of MSC-EVs.
In addition, knowing the ability of MSCs to induce Tregs [9,66,70], we also sought to evaluate the expression of Foxp3 transcription factor in T cells under inflammatory conditions. Interestingly, the addition of MSC-EVs induced a significant increase in Foxp3 expression in Th1-differentiated cells, suggesting that MSC-EVs can induce a regulatory transcriptional profile in Th1-type cells. In our study, the reduction of Tbet expression was not demonstrated after the addition of MSC-EVs. However, the co-expression of Tbet and Foxp3 transcription factors, by itself, may suggests the induction of more regulatory cells. Recent studies have demonstrated the existence of a type of hybrid cells that co-express transcription factors of different T subtypes, exerting preferentially an anti-inflammatory effect [71][72][73]. It was demonstrated that the expression of Tbet in Tregs (Foxp3 + ) is important for these cells to regulate Th1 response and, consequently, regulate IFN-γ production [71,74]. Tregs can also produce IFN-γ as a way to regulate the inflammatory response [58] or generate operational tolerance [75]. Accordingly, the co-expression of Foxp3 and IFN-γ in MSC-EVs treated cells, suggests the generation of a different hybrid cell subset that may have regulatory function. In this context, the suppression assay of CD4 + and CD8 + T cell proliferation confirmed that these generated Foxp3 + IFN-γ + T cells are functional and have regulatory profile. Therefore, the treatment with MSC-EVs generate a rare subtype of Th1 cells that at same time that produce IFN-γ and express the regulatory transcription factor Foxp3, which make them acquire regulatory properties, regulating lymphocyte proliferation. However, the molecular mechanisms of how these cells behave should be further investigated.
In order to investigate possible mechanisms involved in the MSC-EVs effects over Th1 cells, the PCR miRNA array platform associated to T and B cell activation pathways was used as a tool to identify miRNAs and signaling pathways involved in the regulation of Th1 response. We observed that MSC-EVs addition reprogrammed T cell miRNAs profile as observed during modulation of the immune system upon activation [76,77]. In silico enrichment analysis, focusing on the target genes of the most up or downregulated miRNAs, showed possible signaling pathways involved in this process. Among the pathways linked to downregulated miRNAs was observed the TGF-β pathway, which is involved in the generation of Tregs through the binding of transcription factors in the regulatory elements at the Foxp3 locus [78,79]. Therefore, we looked for genes that activate these pathways and we detected difference in TGFBR2 (TGF-β receptor 2) expression, a gene targeted by miR-23a-3p. This receptor has important role in the control of Th1 cells. Its absence in T cells led to a more inflammatory condition, with greater production of IFN-γ by CD4 + T cells [80,81]. Additionally, expression of a dominant negative form of this receptor also decreased the suppressive capacity of Tregs [82]. These data suggest that the regulation of miRNAs expression and their target genes can be one mechanism by which MSC-EVs are modulating Th1 cells. Specifically, the miRNA regulation of TGFBR2 expression may be regulating IFN-γ production from Th1 cells after MSC-EVs addition.
TGF-β pathway is also involved in the inhibition of mTOR pathway, which regulates metabolic process in immune cells, favoring anabolic processes and promoting glycolysis [83]. Recently, the correlation between metabolism and immune cells has been more explored. After activation, Naïve T cells need substrates for the biosynthesis of proteins, lipids and nucleic acids essentials for cell proliferation and, hence, to accelerate this process, they switch their metabolic profile to anabolic state, mainly dependent on aerobic glycolysis [84]. This process, known as the Warburg effect [85], occurs when the cell performs glycolysis even in the presence of oxygen. Specifically, for Th1 cells, the glycolytic pathway appears to be the most important. IFN-γ production is regulated by GAPDH, a glycolytic pathway enzyme [86] and Glut-1 [41]. In contrast, Tregs is dependent mainly of the metabolism of lipids (beta oxidation), which occurs in the mitochondria [84,87] and the inhibition of glycolysis favors Treg differentiation [40]. Therefore, in order to verify if MSC-EVs treatment and its consequent signal trough TGF-β pathway could alter T cell metabolism, we evaluated the expression of genes associated with different metabolic pathways. Hexokinase 2 (HK2) is an enzyme that catalyzes the first step of glycolysis, ATP citrate lyase (ACLY) is an important enzyme in lipid biosynthesis [88], Pyruvate dehydrogenase kinase 1 (PDK1) is an enzyme that phosphorylates and inactivates pyruvate dehydrogenase, the enzyme responsible for the first step of citric acid cycle-TCA [89]-and pyruvate kinase isozyme 2 (PKM2) catalyzes the last glycolysis reaction [90]. Although we have not identified any significant differences in the expression of ACLY, PDK1 and HK2, we cannot affirm that there is no metabolic changes in T cells treated with MSC-EVs, since the regulation of these enzymes may act principally at the protein level [91][92][93]. However, importantly, we detected a difference in PKM2 expression. This enzyme has been related to the Warburg effect, especially in cancer cells [94]. PKM2 has also functions not related to metabolism, being associated with activation of inflammatory response [95], mainly acting as co-activator of HIF-1α [96,97]. Accordingly, it is demonstrated that under normoxia conditions, the absence of HIF-1α increase Tregs differentiation [40]. Finally, PKM2 can act as a regulator of mTOR pathway. A recent study has shown that the reduction of PKM2 inhibits PI3K/Akt [98] signaling, which may also justify the reduction of IFN-γ in cells treated with MSC-EVs. Additionally, when we evaluated specifically the glycolytic metabolism, a reduction of ECAR, glycolysis and glycolytic capacity and glycolytic reserve was shown, corroborating with PKM2 expression levels reduced. These data also suggest that the metabolic switch required to cytokine production was not achieved and may explain the IFN-γ reduction observed.
To further investigate the effects of MSC-EVs over T cell metabolism, we looked at mitochondrial metabolism through the detection of mitochondrial membrane potential ( ψ m ). The evaluation of membrane potential of Th1-differentiated cells demonstrated a reduction in cells treated with MSC-EVs, which suggests a lower participation of mitochondrial metabolism in this process. Mitochondrial metabolism results in the generation of ROS (reactive oxygen species), which is important for the optimal activity of NFAT and NF-kB [99,100], involved in the signaling for T cell activation and cytokine production [101], such as IFN-γ [102]. In this context, the addition of MSC-EVs may have altered the metabolism of the T cells, reducing their activation and IFN-γ production. Additionally, in a recent study, Sukumar and colleagues isolated cells differentiated for Th1 in vitro, sorting them according to the mitochondrial potential. Cells with high ψ m produced four times more IFN-γ compared to low ψ m cells and cells with low ψ m showed a reduction in mTORC1 activity [103]. In another study, the addition of TGF-β reduced glycolytic metabolism in thymus Tregs, as a consequence of the reduction of mTOR pathway activation [104]. These data corroborate our findings that cells producing less IFN-γ, after treatment with MSC-EVs, have lower ψ m and may also suggest the participation of TGF-β pathway in this regulation.
Finally, to evaluate the MSC-EVs effect in vivo, we used an OVA immunization experimental model. Since we believe that the major effect of EVs depends on an inflammatory environment, we looked for a model that would give us a Th1 response. Therefore, we used Montanide ISA adjuvant [37], which induces predominantly a Th1 response [105]. According to what was discussed above, we believe that MSC-EVs effects are dependent on the presence of specific cytokines related to the Th1 response. When MSC-EVs were injected concomitantly with OVA, an increase in the total number of cells expressing Foxp3 was seen, confirming the ability of MSC-EVs to generate regulatory cells expressing Foxp3-as observed in vitro in a Th1 differentiation. However, we did not observe any effect on the IFN-γ production in this context. Increasing the dose may lead to a more pronounced effect. Nevertheless, these data corroborate our in vitro studies and demonstrate the ability of MSC-EVs to induce Foxp3 + cells under inflammatory conditions in vivo.
In summary, in this report we demonstrated that MSC-EVs have the capacity to modulate lymphocytes. We verified that Th1 cells reduced the production of IFN-γ-and surprisingly-they started to express Foxp3. As part of their regulatory mechanism, we have shown that MSC-EVs induced modifications in the miRNA profile, decreasing miR-23a-3p expression and increasing the expression of its target gene, TGFBR2, suggesting active participation of TGF-β pathway in this regulation. This pathway, in turn, may inactivate the mTOR pathway and, as a consequence, alter the metabolic profile of T cells treated with MSC-EVs. These cells presented a reduction in the glycolytic metabolism, as well in the mitochondrial metabolism, what can be related to PKM2 modulation. This molecular crosstalk can explain the presence of less activated Th1 cells with lower IFN-γ production. Concluding, our findings suggest that MSC-EVs are able to specifically modulate activated T cells at an alternative regulatory profile by miRNAs and metabolism shifting. Thus, we can speculate that MSC-EVs can induce immunological tolerance, in vivo, contributing for their future use as an alternative therapy.
Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4409/9/4/1059/s1, Table S1. Upregulated miRNAs' target genes selected for analysis and their corresponding signaling pathways. Table S2. Downregulated miRNAs' target genes selected for analysis and their corresponding signaling pathways. Figure S1. Cells viability after MSC-EVs treatment: Live and dead staining of cells differentiated for Th1 after MSC-EVs treatment. Data representative of 4 independent experiments. Video S1. Animation of incorporation of PKH26-labeled MSC-EVs by purified naive CD4+ cells: Images were captured at 15 min interval over the course of 15h and collated in sequential order. See Figure 2A