Recently years, stem cell treatments have become an important therapeutic strategy for the treatment of various proinflammatory and autoimmune diseases because of their powerful immunomodulatory properties via the suppression of T cells, B cells, natural killer (NK) cells, and antigen presenting cells [26, 27]. Such immunological effects have been shown primarily in vitro but also in vivo, in a number of experimental disease models such as EAE [1012], CIA [8], and experimental colitis [9, 28]. Despite the in-vitro and in-vivo evidence for a therapeutic effect of MSCs, their precise mechanism of action and the profile of their adverse effects as immunomodulatory agents are still poorly understood.

Recently, it has been demonstrated that stimulation of human MSCs with poly(I:C) and LPS induces activation of NF-kB, mitogen-activated protein kinases (MAPK), and protein kinase B (AKT) signaling pathways. Activation of these pathways was associated with the induction and secretion of different patterns of cytokines and chemokines, suggesting that LPS could promote the activation of immune responses while poly(I:C) could suppress it [29]. Similarly, Waterman et al. [20] demonstrated that human MSCs polarize into a proinflammatory or anti-inflammatory phenotypes, according to the specific TLR3 or TLR4 activation in vitro. This functional phenotype was also shown in vivo, in experimental models of diabetes [30] and ovarian cancer [31]. These findings suggest that pretreatment of MSCs with TLRs could be a powerful and innovative therapeutic tool for the treatment of autoimmune and proinflammatory pathologies. In the present study, we evaluated the immunomodulatory effect of murine MSCs after treatment with TLR3 and TLR4 agonists in vitro and in a mouse model of multiple sclerosis. Our results demonstrated that pretreatment of MSCs with poly(I:C) enhances their immunosuppressive capacity in vitro and that intraperitoneal injection of these MSCs significantly reduces the severity of EAE. In contrast, LPS pretreatment of MSCs induces a significant decrease in their immunomodulatory function in vitro and completely reverses the therapeutic immunosuppressive effect of MSCs in vivo.

Diverse studies have shown that murine MSCs express different functional TLRs, such as TLR1–TLR8 [32]. Our data showed that murine MSCs cultured to 80–90 % confluence in complete culture medium express significant levels of mRNA for TLR3 and TLR4, and that the expression level of TLR4 was higher than that of TLR3, similar to the pattern described by Pevsner-Fischer et al. [32]. In addition, we demonstrated that pretreatment of these TLRs for 1 hour with their respective agonists differentially affects the in-vitro immunosuppressive capacity of murine MSCs. First, we observed that untreated MSCs were functionally capable of inhibiting the proliferation of activated T cells, confirming what has been published previously [7]. Once the inhibitory capacity of the MSCs on T-cell proliferation was confirmed, we evaluated the effect of MSCs pretreated for 1 hour with poly(I:C) or LPS. MSCs pretreated with poly(I:C) were able to significantly increase the inhibitory capacity of MSCs on T-cell proliferation by approximately 33 % with respect to untreated MSCs. Conversely, MSCs pretreated with LPS completely reversed the immunosuppressive effect of untreated MSCs and induced a significant, and dose-dependent, increase in T-cell proliferation.

These results demonstrate that brief, in-vitro LPS stimulation of murine MSCs induces a proinflammatory phenotype, similar to the effects previously shown by Waterman et al. [20], using human MSCs.

To better understand the effect of activation of TLR ligands on the immunomodulatory activity of MSCs, we measured NO production in the absence or presence of splenocytes stimulated with ConA as well as the expression and levels of proinflammatory cytokine IL-6. In the absence of splenocytes, no differences were observed in NO secreted by untreated or pretreated MSCs. However, in the presence of splenocytes, we detected a significant increase in NO production induced by MSCs pretreated with poly(I:C) but not by those pretreated with LPS, which had lower NO production compared with untreated MSCs. On the other hand, our results indicated that the expression of IL-6 increased after stimulation of MSCs with LPS and was inhibited after stimulation of MSCs with poly(I:C). Taken together, these data provide evidence of a based anti-inflammatory phenotype for MSCs pretreated with poly(I:C) and an opposite, proinflammatory phenotype for MSCs stimulated with LPS, which show a loss of capacity to inhibit T-cell proliferation, a higher expression of IL-6, and nonsignificant NO secretion. MSCs have been identified as immunomodulating cells because they inhibit the generation and function of Th1 and Th17 cells and increase Treg cell formation [3336]. Previous studies from our laboratory showed that MSCs cocultured with CD4+ T cells grown in conditions polarizing them towards Th1 or Th17 lineages exert strong Th1 immunosuppression but have little effect on Th17 cells [22, 25]. Here, we evaluated the immunomodulatory effect of TLR3 and TLR4-pretreated MSCs on Th1 and Th17 differentiation and proliferation in vitro. We observed a strong capacity of MSCs pretreated with poly(I:C) to inhibit Th1 and Th17 differentiation and proliferation, which was even more pronounced than the effect of untreated MSCs. Conversely, MSCs pretreated with LPS showed a diminished capacity to inhibit Th1 and Th17 differentiation and proliferation.

Recently, we studied the therapeutic effect of MSC administration on EAE, showing that the injection of MSCs at the time of disease onset induces a significant improvement in the clinical signs of the disease [12]. In the present study, using the same mouse model, we studied whether the administration of MSCs pretreated with poly(I:C) or LPS generated distinct therapeutic effects in vivo. Our results demonstrated that MSCs pretreated with poly(I:C) significantly reduce the clinical signs of EAE and that pretreatment of MSCs with LPS completely reverses the therapeutic immunosuppressive effect of MSCs. Furthermore, when we evaluated the cumulative score and the weight loss of the animals in each group, we found the same pattern that again highlighted the ability of MSCs stimulated with poly(I:C) to increase the immunosuppressive capacity of the MSCs. Poly(I:C) stimulation generated a decrease in the score and weight loss in the treated animals, while LPS caused an increase in clinical signs and a high percentage of weight loss in animals. In addition, we investigated the relationship between the treatments of the animals with respect to Th1 and Th17 proinflammatory cell subsets in the lymph node of EAE mice as a way to account for the observed results. We found a significant decrease of the Th1 and Th17 subsets induced by the administration of untreated MSCs, although these differences were significant only in the case of Th17 cells. No significant differences were observed in the expression of Th1 and Th17 cells when EAE mice were injected with poly(I:C)-pretreated MSCs in comparison with the expression in untreated MSCs. In contrast, the treatment of EAE mice with LPS-pretreated MSCs completely reversed the effect on the Th1 and Th17 subset cells induced by untreated MSCs.