Toll-like receptors are the best-studied pattern recognition receptors that specifically recognize pathogen-associated molecular patterns, such as double-stranded RNA (TLR3). Activation of TLR3 by the poly(I:C) ligand in human macrophages, dendritic cells, and epithelial cells is known to trigger MyD88-independent signaling cascades [35, 36], unlike other TLRs. Activation of TLR3 results in the nuclear translocation of NF-κB and secretion of proinflammatory cytokines, which ultimately leads to the maturation of dendritic cells [37]. Thus, TLR3 modulates the innate and adaptive immune responses.

TLR activation in MSCs can modulate their immune properties. Binding of the TLR3 ligand through the adapter protein TICAM-1/TRIF induces the expression of IFN type I, including in MSCs [19, 36, 38]. At the same time, IFN-α/β expression can be induced by activation of additional cytosolic receptors in the presence of dsRNA: retinoic acid-induced gene-I (RIG-I/DDX58) and melanoma differentiation-associated antigen 5 (MDA5/IFIH1) [37, 39]. It is known that secreted MSCs IFN-β can have an anti-inflammatory effect by increasing the number of T-reg cells [40] and reducing the proliferation of activated T cells when cocultivated with MSCs [19].

Activation of TLRs on effector immune cells (i.e., T and B cells) usually results in the secretion of proinflammatory cytokines [41]. However, MSCs are immunoplastic and, similar to macrophages, can exhibit immunosuppressive or proinflammatory properties, depending on the level of inflammation. This is probably why the priming of TLRs on MSCs can enhance both proinflammatory potency and immunosuppressive properties, depending on the chosen modes of experimental exposure to TLR ligands [17, 19, 42].

T-cell activation is associated with increased tryptophan metabolism. WARS1 and IDO1 are the main enzymes involved in tryptophan metabolism and can be activated in MSCs by the inflammatory environment [28, 43, 44]. Increased production of WARS1 is known to interact with IDO1 to regulate immune responses in vivo; namely, IDO1 catalyzes the degradation of tryptophan into kynurenine, and WARS1 is associated with its transport into cells [45]. It has been shown that TLR3 activation in MSCs leads to STAT1 phosphorylation and subsequent activation of IDO1. The change in the balance between the activity of WARS1 and IDO1 in primed MSCs determines the decrease in the availability of tryptophan for protein synthesis and thus mediates the immunosuppressive effects of MSCs on T cells during their cocultivation [46, 47]. Therefore, in this study, the expression levels of IDO1 and WARS1 were assessed during the first optimization stage of the MSC priming protocol.

According to our results, TLR3 priming in all studied modes led to activation of WARS1 (Figs. 2b, 6) and IDO1 (Figs. 2a, 6). We noted that the greatest response to IDO1 and WARS1 was observed after 3 h of incubation of cells with 1 and 10 µg/mL poly(I:C). It was first shown that WARS1 can be activated by poly(I:C), similar to preconditioning with IFN-γ, in MSCs [28]. In addition, our results are consistent with a number of papers in which poly(I:C) exposure resulted in increased IDO1 expression [17, 19, 48, 49]. However, priming protocols differed in all cases. Thus, Lombardo and DelaRosa noted an increase in IDO1 activity upon stimulation with 10 μg/mL poly(I:C) but not at 1 μg/mL [16]. At the same time, Michalis Mastri et al. showed that exposure to poly(I:C) can trigger differential trophic responses [50]. The contradictions between our results and those of Liotta et al., who did not observe an effect of TLR3 priming on IDO1 activation [51], may be due to differences in the concentration and duration of exposure to the stimulatory ligand. Since TLRs perceive exogenous and endogenous stress signals, the intensity and duration of these signals can differently affect the properties of MSCs [17], which are the so-called "inflammatory switches". Thus, when priming TLR, it is important to normalize the concentration and time of exposure to the ligand, since these parameters can determine the immune status of MSCs.

We also investigated the expression of TSG-6, which is involved in extracellular matrix remodeling and is a potent inhibitor of neutrophils [29]. TSG-6 is known to have anti-inflammatory activity through modulation of chemokines/cytokines produced by damaged tissue [52]. As a result, secreted MSCs TSG-6 can suppress the development of Th1 by directly inhibiting the activation of T cells or by suppressing the activation of antigen-presenting cells, disrupting the translocation of NF-κB to the nucleus [53]. To our knowledge, this is the first study to show that TLR3 priming in MSCs leads to an increase in TSG-6 expression and that the expression level is modulated by the poly(I:C) preconditioning protocol (Fig. 2c). PD-L1 expression also depended on the activation protocol used (Fig. 2d). PD-1 is expressed on various activated immune cells (including T cells, B cells, macrophages, DCs, etc.), interacting with ligands (PD-L1 and PD-L2), and these activated immune cells are depleted. Thus, the intensity of inflammation decreases in chronic infection and autoimmunity [54]. Increased PD-L1 expression can enhance the immunosuppressive function of MSCs by inducing T regulatory cells and modulating cytokine expression.

Prostaglandin E2 is another key MSC cytokine involved in suppressing the proliferation of activated T cells [18, 55, 56]. MSC treatment with dsRNA activates PTGES2 and PGES, which are involved in PGE2 production [57]. At the same time, the effect of this cytokine on T cells is associated with the inhibition of intracellular calcium release [58], a decrease in the activity of p59 protein tyrosine kinase, and the level of IL-2 secretion [59]. The source of stem cells determines the level of constitutive PGE2 secretion. Intact MSCs from Wharton's jelly express less PGE2 than MSCs isolated from the bone marrow (MSCs-BM) [60] and, as a result, do not suppress T-cell proliferation [61]. This may explain our conflicting results with those of Liotta et al., who did not observe the effect of TLR3 priming on PGE2 levels owing to their high constitutive expression in MSCs-BM [51]. In contrast, in our study, stimulation with 10 μg/ml poly(I:C) for 1 or 3 h led to an increase in PTGES2 expression and PGE-2 levels, which was consistent with previously obtained data for adipose tissue-derived MSCs (Fig. 2e, f) [55, 56].

The effect of MSCs on cells of the immune system is achieved through direct intercellular interactions, as well as through paracrine signaling. In addition, MSCs isolated from adipose tissue studied in this work have the most significant secretory function compared to MSCs from the bone marrow or umbilical cord [62], and their availability and fewer ethical problems make them more attractive for cell therapy [63]. In this study, a protocol of short-term (3-h) treatment with poly(I:C) (10 µg/mL) resulted in maximum production of anti-inflammatory cytokines and enhancement of the immunosuppressive properties of MSCs, which is consistent with previously published data [17, 18, 49, 59]. However, conflicting results have been reported in a number of works. Thus, Raphaëlle Romieu-Mourez et al. reported the formation of a proinflammatory MSC phenotype upon TLR3 priming [20], Liotta et al. noted a decrease in immunosuppressive effects [51], and DelaRosa and Lombardo stated that there is no significant effect of poly(I:C) exposure on MSCs modulating immune responses in vitro [25]. As noted earlier, the intensity and duration of TLR3 activating signals can affect MSC properties in different ways. Thus, Raphaëlle Romieu-Mourez et al. studied the effect of the culture medium of primed MSCs on peripheral blood mononuclear cells (PBMCs) and primary macrophages. Liotta et al. used purified CD4+ T cells, while DelaRosa and Lombardo cocultured MSCs with PBMCs and purified CD4+ and CD8+ T-cell fractions. Simultaneously, in the first case, stimulation was carried out for 6 h at a poly(I:C) concentration of 20 µg/mL, which can lead to increased inflammation [50]. Liotta et al. exposed MSCs for 5 days with a TLR agonist (concentration not specified), and DelaRosa and Lombardo evaluated the effect of priming after 72 h of exposure. We noted that long-term exposure to an agonist (24 h) reduced the expression levels of characteristic immunosuppressive markers of MSCs (Fig. 2), which may be associated with the effect of reverse regulation and the decrease/absence of immunosuppressive properties of MSCs. We used linear Jurkat cells to modulate T cell responses, which prevented non-T cells from influencing PBMCs. In this study, we evaluated the effectiveness of various TLR3 stimulation protocols, assuming that MSC immunosuppression is primarily mediated by secreted factors. The optimal exposure was within 3 h of treatment with 10 µg/ml poly(I:C). The enhanced immunosuppressive activity of prMSCs activated according to this protocol was confirmed when MSCs were cocultured with Jurkat cells. It was shown that prMSCs, more actively than MSCs, reduced the proportion of proliferating (by 18%) and increased apoptosis (by 29%) activated T cells but not intact T cells (Fig. 3). T-cell activation leads to increased IL-2 production. PrMSCs inhibited T-cell activation by reducing the expression and secretion of this cytokine (Fig. 3d, f). These data confirm the functionality of MSCs as “inflammatory switches”: TLR3 priming enhances the immunosuppressive effects of MSCs under conditions of inflammation.

Type I IFN signaling is associated with the induction of the immunosuppressive cytokine IL-10 by macrophages [64]. It has been noted that stimulation of TLR3 in MSCs can also increase the secretion of this cytokine [18]. However, the mechanisms underlying the direct immunosuppressive effects of IL-10 are not fully understood [65, 66]. IL-10 is known to inhibit the activation of CD4+ T cells by suppressing the production of IL-2 and signaling CD28 [67], and myeloid cells treated with IL-10 lose their ability to respond to lipopolysaccharide (LPS) [68]. We showed that when cocultivated with J+ prMSCs and MSCs, the level of IL-10 in the medium did not differ but was higher than that after cocultivation with J− and in intact T cells, similar to the data of Yang et al. [69] (Fig. 3e). It is likely that the secretion of this cytokine can be blocked by newly synthesized IL-10 [33], thus limiting its amount in a conditioned environment [70]. It is also known that IL-10 acts at low doses and mainly locally to induce immune responses. We also did not detect the secretion of IL-10 by intact MSCs or prMSCs, similar to the results of Lombardo et al. [16]. At the same time, it was previously shown that IL-10, as a component of the MSC secretome, plays the role of an immunosuppressant, reducing the secretion of IL-2 by activated Jurkat cells [33], and may also be associated with their differentiation into T-reg cells [34]. In addition, increased secretion of IL-10 by transduction of MSCs with the MIG retroviral vector (MSCV-IRES-GFP) or transfection of IL-10 mRNA increases their anti-inflammatory effect in a model of acute graft-versus-host disease or acute respiratory distress syndrome [70, 71].

In the next stage, we performed a proteomic study of MSCs after TLR3 priming according to the chosen protocol and compared the results with those for intact cells. Poly(I:C) mimics dsRNA released during viral replication; thus, TLR3 stimulation activates interferon regulatory factor 3 (IRF3). Activation of IRF3 leads to an increase in the expression of primary response genes (interferon-β (IFN-β), interferon-stimulated gene 15 (ISG15), etc.), which in turn initiates autocrine/paracrine activation of secondary genes (MX1, MX2) [72]. This pathway plays a key role in the antiviral response of cells. In this study, the impact of poly(I:C), according to the results of a proteomic study, led to the activation of ISG15, ISG56 (IFIT1), ISG54 (IFIT2), ISG60 (IFIT3), MX1 and MX2, which are necessary to maintain the viability of MSCs. However, IFIT1 can also negatively regulate the expression of TNF-α and other inflammatory cytokine genes in LPS-treated macrophages while simultaneously stimulating IFN-β expression and the subsequent interferon gene program. It is likely that IFIT1 regulates an important balance between the inflammatory and IFN gene programs to promote an optimal innate immune transcriptional response to microbial infection [73]. The chronic type I IFN signature is also known to increase the expression of PD-L1 and IL-10 in dendritic cells and macrophages, modulating the anti-inflammatory environment [74]. Thus, the level and duration of viral load may influence the immunomodulatory properties of MSCs. We noted a decrease in the number of proliferating T cells after coculture with prMSCs, probably due to the functioning of IFIT1. The TLR3 activation protocol obtained in this study allowed maintenance of the immunosuppressive status of cells.

In addition, exposure to poly(I:C) can lead to the activation of IRF3 through the cytosolic receptors RIG-I/DDX58 and MDA-5/IFIH1 in MSCs [75–77]. Activation of RIG-I strongly induces the OASL protein, which stimulates the production of type I IFN and its antiviral response [78, 79]. STAU1 expression, which we obtained from the results of proteomic analysis, is known to positively regulate the expression of immune response genes (IFIT2, IFIT3, and OASL) [80], and its knockdown leads to inhibition of the activation of the porcine IFN-β promoter [81]. TRIM25, in turn, mediates both RIG-I/DDX58 and MDA-5/IFIH1 ubiquitination and is required for RIG-I-mediated interferon production and antiviral activity [82]. It is important to note that activation of RIG-I/DDX58 and MDA-5/IFIH1 modulates the expression of proinflammatory cytokines, such as IL-6 [77], and immunosuppressive molecules, such as IDO1 [83]. We noted increased expression of IDO1 earlier in the qRT‒PCR results, and IDO1 was also present in the proteome results as a protein that appeared in prMSCs after exposure to poly(I:C). Thus, we showed that TLR3 functions normally in MSCs, as its activation under the conditions of the obtained protocol triggers type I IFN signaling cascades.

However, type I IFN, along with stimulating antiviral functions, limits damaging immune responses that can lead to tissue pathology and excessive tissue damage. MSCs play a key role in tissue repair and must remain viable to perform their functions. We noted an increase in the amount of DNJA1 protein, which negatively regulates BAX translocation from the cytosol to the mitochondria in response to cellular stress and thus probably additionally protects cells from apoptosis [84]. The inflammatory environment appears to lead to oxidative stress (OS) in MSCs, as we observed an increase in GFTP2 protein, a marker of OS [85]. At the same time, activation of TLR signaling also greatly increases the expression of the antioxidant defense and DNA repair protein SOD2 [86]. We also observed an increase in PNPT1 protein, the import function of which is known to be necessary for mitochondrial respiration and cell viability [87], and TMX, which plays an important role in host defense against OS-relate