Characterization of signaling proteins secreted during ASC and PBMC co-

Signaling proteins such as cytokines and chemokines are strongly involved in processes mediating the immunogenicity and immunosuppression of ASCs. Thus, a set of signaling proteins was analyzed from the co-cultures of ASCs and PBMCs and mono-cultures of ASCs and PBMCs alone. It has been demonstrated in previous studies, that the local inflammatory environment is highly important in the regulation of MSC plasticity, and especially the expression of cytokines IFN-γ, TNF-α, and IL- 6 are important in regulating the immunosuppressive phenotype of ASCs (Crop et al., 2010a). IFN-γ has been shown to induce MHC-I and MHC-II expressions that further increase the antigen-presenting capacity and immunogenicity of MSCs (Chan et al., 2008; Le Blanc et al., 2003b).

In our study III, the secretion of IFN-γ, TNF-α, and IL-6 was increased after co- culture of ASCs and PBMCs compared with mono-cultures. Furthermore, IFN-γ and TNF-α levels were higher in indirect co-culture compared with direct co- cultures, which suggest the inhibitory effect of direct contact between ASCs and lymphocytes with respect to the secretion of IFN-γ and TNF-α. A similar phenomenon of reduced IFN-γ and TNF-α levels after direct co-culture of CD4+T- cells and human umbilical cord MSCs was reported earlier (Wu et al., 2014). In this previous study, T-cells were first activated for high IFN-γ and TNF-α secretion via PHA/IL-2 exposure, followed by co-culture with MSCs that reduced these secretion levels. It may be speculated that, similar to PHA/IL-2, the indirect contact between ASCs and lymphocyte changes the local inflammatory environment and activates PBMCs for higher IFN-γ and TNF-α secretion, as demonstrated by Crop et al. (Crop

et al., 2010a); however, direct contact between ASCs and lymphocytes will reduce the activation.

In contrast to IFN-γ and TNF-α secretion, direct contact between ASCs and PBMCs induced cells toward higher IL-6 production in our studies. Additionally, IL- 6 has several functions because it is involved in inflammation and infection responses and also in the regulation of metabolic, regenerative, and neural processes (Scheller et al., 2011). Thus, IL-6 can either support or suppress inflammation depending on the context, and the anti-inflammatory effect is mediated in co- operation with IFN-γ by inhibiting the secretion of TNF-α (Ulich et al., 1991).

Through production of IL-6, MSCs prevent monocyte differentiation toward antigen-presenting immunogenic cells and drive differentiation toward an anti- inflammatory IL-10-producing cell type (Melief et al., 2013b). In our study, IL-6 secretion was higher in FBS conditions compared with XF/SF conditions, which is in agreement with the strong suppression potential observed in FBS media. In addition, high IL-6 levels are correlated with low TNF-α levels and vice versa. Based on our studies, IL-6 is more heavily involved in ASC-mediated suppression functions.

The secretion levels of IFN-γ and TNF-α may be decreased by the cytokine TGF- β1, which is an important factor in maintaining immune tolerance, in addition to several other functions in control of cell growth, proliferation, differentiation, and apoptosis (Wahl et al., 2006). In addition, proliferation and activation of T-cells can be diminished by the effect of TGF-β1 (Gilbert et al., 1997; Tiemessen et al., 2003).

In our study, significantly lower levels of TGF-β1 were secreted in co-cultures with XF/SF-expanded ASCs, which was in line with the immunosuppression results in which significantly less suppressive potential was observed with ASCs expanded in XF/SF conditions. These results were in accordance with the increased IFN-γ and TNF-α and decreased IL-6 secretion levels that were measured in co-cultures containing XF/SF-expanded ASCs.

Additionally, our study demonstrates the secretion profile of chemokines CXCL8, -9, -10 and CCL2 and -5 in co- and mono-cultures of ASCs and lymphocytes. It was demonstrated that the family of chemotactic cytokines known as chemokines regulate cell adhesion and chemotaxis of all types of leucocytes, including hematopoietic precursors, mature leucocytes of the innate and naive immune system and memory and effector lymphocytes (Moser and Willimann, 2004). Thus, breakdown in the control of leukocyte mobilization may lead to chronic inflammatory diseases. However, by controlling the secretion of chemokines, ASCs may have the capacity of immunomodulation. The secretion of several chemokines

has been shown to increase when ASCs are co-cultured with lymphocytes (Crop et al., 2010a). Our results support these previous studies and further demonstrate that the direct contact between ASCs and lymphocytes is particularly required for stronger chemokine secretion and T-cell recruitment. Overall, chemokine secretion was strongly increased in direct co-cultures compared with indirect co-cultures, and interestingly, greater secretion of CXCL-8 and -10 was significantly more secreted in cultures with XF/SF expanded ASCs. It is likely that high IFN-γ levels that were detected in XF/SF media have induced XF/SF cells for higher CXCL-8 and -10 secretion, which would be typical example of IFN-γ induced chemokine secretion.

As described earlier, our study demonstrates that XF/SF cells were more immunogenic compared with FBS-expanded cells and less immunosuppressive compared with those in both FBS and HS conditions. Thus, it may be speculated that higher IFN-γ and CXCL-8 and -10 levels detected in XF/SF conditions are associated with weaker suppressive potential and higher immunogenicity.

The secretion of CCL2 and CCL5 was lower in HS conditions compared with FBS and XF/SF conditions. These chemokines are involved in MSC chemotaxis and cell migration to the sites of injury or inflammation (Anton et al., 2012). The chemokines CXCL8, CCL2 and CCL5 may be produced by immune cells, which induce extracellular matrix degradation and further facilitate the increased migration of MSCs through the extracellular matrix. However, the effect of CCL2 and CCL5 on cell homing to inflammation sites could probably be better demonstrated in vivo.

Based on our results, it is difficult to conclude how the lower levels of CCL2 and CCL5 that we observed in HS conditions affect the immunogenicity of ASCs.

The role of IDO (Crop et al., 2010a) has been demonstrated in MSC-mediated immunosuppression and is often secreted in response to IFN-γ. However, in our studies, IDO was secreted at significantly higher levels in XF/SF and HS media compared with FBS conditions, which contradicted the result of the strong immunosuppressive potential observed in FBS media. It may be speculated that the high initial concentration of IDO that was detected in basal culture media could have interfered with the result or that other factors in the co-culture reaction assay had a stronger effect on IDO secretion. Furthermore, a greater amount of IDO was clearly measured in co-cultures of ASC and lymphocytes compared with mono-cultures of ASCs only, suggesting that ASCs do not readily express IDO without activation, which is in line with previous results demonstrated with MSCs (Meisel et al., 2004).

According to previous observations, IFN-γ levels and IDO secretion were correlated, as also observed in our study. Furthermore, Meisel et al. concluded that IDO-mediated T-cell inhibition depends on MSC activation (Meisel et al., 2004), and

consequently, modulation of IDO activity may alter the immunosuppressive properties of MSCs.

The secretion of galectin-1 and -3 was included in the characterization panel because their expression was clearly linked to the immunosuppressive potential of ASCs (Sioud et al., 2011a; Sioud, 2011b). However, it has been shown that in addition to suppressive functions, galectin-3 also has T-cell proliferative functions (Hernandez and Baum, 2002). It was therefore rational that the secretion patterns of galectin-1 and -3 were somewhat different in our studies, and galectin-3 was less secreted overall. However, the secretion of both galectin-1 and -3 was higher in direct and indirect co-cultures compared with mono-cultures of ASCs and PBMCs alone. Galectin-1 suppresses immune responses and induces T-cell apoptosis, and thus, increased galectin-1 concentrations and the detected suppression potential in the co-cultures were expected. According to our results, galectin-3 expression is not as good an indicator of immune suppression as galectin-1, but galectin-1 concentrations were better correlated with the suppression assay results.

Taking these cytokine secretion results together, the final determination of inflammatory responses may be elicited through a combined action of direct cell-cell contacts and secretion of soluble factors following modulation of the local inflammatory environment. Moreover, the suppressive capacity of ASCs is likely a result of several factors that can act in a synergistic manner. Our study clearly demonstrated that direct versus indirect contact between ASCs and lymphocytes has an effect on the secretion of suppressive soluble factors. Although this result was expected, the difference was substantially strong, especially for the chemokine secretion. Puissant et al. have previously shown that ASCs do not spontaneously release suppressive factors, but instead, interaction between ASCs and lymphocytes is required for the secretion (Puissant et al., 2005). Moreover, Puissant et al. showed that the MLR-conditioned media was not able to suppress the lymphocyte proliferation; however, the suppressive capacity of ASCs was retained in indirect MLR. In our study, increased secretion levels of IFN-γ and TNF-α were observed in indirect culture compared with direct culture, which was an interesting result. It may be speculated that the local inflammatory environment in indirect MLR is more beneficial for inducing the immunosuppressive phenotype of ASCs compared with direct MLR. In line with our result, Crop et al. reported that secretion of suppressive factors such as IDO was increased in indirect co-culture compared with direct MLR (Crop et al., 2010a).

Nevertheless, additional studies on the cell signaling mechanisms associated with ASC-mediated immunosuppression should be performed to draw a conclusion on

the importance of direct versus indirect contact. To underline a few key factors, the low secretion of TGF-β1 as well as high secretion of INF-γ and chemokines CXCL- 8 and -10 that were measured in XF/SF conditions may be associated with increased immunogenicity under XF/SF conditions. Furthermore, it may be speculated that the low IFN-γ and high IL-6 secretion that was observed in FBS conditions may be linked to stronger suppressive potential of ASCs. Still, relatively small differences were observed between different serum conditions, and the effect of direct versus indirect contact between ASCs and lymphocytes on signaling protein secretion was more evident. The results of ASC immunogenic properties and cytokine secretion under different culture condition are illustrated in Figure 31.

Figure 31. Schematic illustration of the immunogenic properties of ASCs and signaling protein secretion under different culture conditions. (A) Despite the low immunogenicity in all studied conditions, increased proliferation of PBMCs was observed in HS and XF/SF conditions. The high secretion of IFN- γ and chemokines CXCL-8 and -10, as well as low secretion of TGF-β1 in XF/SF conditions may be associated with increased immunogenicity. (B) A higher ASC number was required for suppressive effects in HS and XF/SF conditions compared with FBS medium. The low IFN-γ and high IL-6 secretion that was measured in FBS conditions may be associated with stronger suppressive potential.