The effect of TLR3 priming conditions on MSC immunosuppressive properties

Abstract

Background: Mesenchymal stromal cells (MSCs) have regenerative and immunomodulatory properties, making them suitable for cell therapy. Toll-like receptors (TLRs) in MSCs respond to viral load by secreting immunosuppressive or proinflammatory molecules. The expression of anti-inflammatory molecules in MSCs can be altered by the concentration and duration of exposure to the TLR3 ligand polyinosinic-polycytidylic acid (poly(I:C)). This study aimed to optimize the preconditioning of MSCs with poly(I:C) to increase immunosuppressive effects and to identify MSCs with activated TLR3 (prMSCs).

Methods: Flow cytometry and histochemical staining were used to analyze MSCs for immunophenotype and differentiation potential. MSCs were exposed to poly(I:C) at 1 and 10 μg/mL for 1, 3, and 24 h, followed by determination of the expression of IDO1, WARS1, PD-L1, TSG-6, and PTGES2 and PGE2 secretion. MSCs and prMSCs were cocultured with intact (J^-) and activated (J⁺) Jurkat T cells. The proportion of proliferating and apoptotic J⁺ and J^- cells, IL-10 secretion, and IL-2 production after cocultivation with MSCs and prMSCs were measured. Liquid chromatography-mass spectrometry and bioinformatics analysis identified proteins linked to TLR3 activation in MSCs.

Results: Poly(I:C) at 10 μg/mL during a 3-h incubation caused the highest expression of immunosuppression markers in MSCs. Activation of prMSCs caused a 18% decrease in proliferation and a one-third increase in apoptotic J⁺ cells compared to intact MSCs. Cocultures of prMSCs and Jurkat cells had increased IL-10 and decreased IL-2 in the conditioned medium. A proteomic study of MSCs and prMSCs identified 53 proteins with altered expression. Filtering the dataset with Gene Ontology and Reactome Pathway revealed that poly(I:C)-induced proteins activate the antiviral response. Protein‒protein interactions by String in prMSCs revealed that the antiviral response and IFN I signaling circuits were more active than in native MSCs. prMSCs expressed more cell adhesion proteins (ICAM-I and Galectin-3), PARP14, PSMB8, USP18, and GBP4, which may explain their anti-inflammatory effects on Jurkat cells.

Conclusions: TLR3 activation in MSCs is dependent on exposure time and poly(I:C) concentration. The maximum expression of immunosuppressive molecules was observed with 10 µg/mL poly(I:C) for 3-h preconditioning. This priming protocol for MSCs enhances the immunosuppressive effects of prMSCs on T cells.

Keywords: Immunosuppression; Jurkat; MSCs; Poly(I:C); TLR3; Toll-like receptors.

Fig. 1

Phenotypic characterization of human adipose-derived MSCs. a The fibroblast-like morphology of adhesion cells was studied by phase-contrast microscopy. Scale 100 µm. b Multilineage differentiation of MSCs. Histochemical staining of calcium deposits (alizarin red) and neutral lipid accumulation (oil red O) after 14 days of osteogenesis and adipogenesis induction. Scale 100 µm. c Flow cytometry analysis of TLR3 expression: after 10 µg/mL poly(I:C) stimulation for 3 h – prMSCs (red line) and without stimulation – MSCs (constitutive expression, blue line), isotypic control—IC (green line). d, e Immunophenotypic characteristics of MSCs and prMSCs: negative expression of CD34, CD45, and HLA-DR markers (top panel); positive expression of CD73, CD90, and CD105 markers (bottom panel)

Fig. 2

Evaluation of the influence of the TLR3 priming protocol on the production of immunosuppressive factors. Expression of the immunosuppressive effectors IDO1 (a), WARS1 (b), TSG-6 (c), PD-L1 (d) and PTGES2 (e) genes, secretion of PGE-2 (f), activation of PD-L1 (g) in prMSCs cells. Control—unprimed MSCs. Data for various stimulation protocols: 1 μg/mL and 10 μg/mL poly(I:C) for 1,3 and 24 h (a-e). The analysis was performed 2 h (c–e) or 24 h (a, b, f, g) after preconditioning with poly(I:C). Error bars indicate SEM. *p < 0.05, **p < 0.01 and ***p < 0.005 compared to the control

Fig. 3

Influence of MSCs and prMSCs on activation (IL-2 production), the number of proliferating and apoptotic Jurkat cells, and the level of anti-inflammatory IL-10 cytokine. The number of cells in the G1-phase (left) and G2/M-phase (right) as a percentage of nonactivated Jurkat cells (a, c) and activated PMA/PHA Jurkat cells (b, c) (flow cytometry). Level of relative mRNA expression (d) and secretion (f) of IL-2 by Jurkat cells (qRT‒PCR and ELISA). Amount of cytokine IL-10 (e) in the conditioned medium (ELISA). The number of living cells (bottom right), cells in the early apoptosis stage (bottom left), and cells in the late apoptosis stage (top left) as a percentage of nonactivated Jurkat (g, i) and activated PMA/PHA Jurkat (h, j) (flow cytometry). Int—noncocultured cells, MSCs/prMSCs—after cocultivation with MSCs/prMSCs. Error bars indicate SEM. n.s.: not significant (p > 0.05); *p < 0.05, **p < 0.01 and ***p < 0.005 compared with the control. ^#p < 0.05, ^##p < 0.01 and ^###p < 0.005 compared with the MSCs

Fig. 4

Volcano plot of the 1686 proteins identified by LC–MS/MS. Proteins with unchanged expression are marked in gray, those upregulated in prMSCs are marked in red, and those downregulated in prMSCs are marked in blue

Fig. 5

Functional classification of the changed proteins in prMSCs. Biological process in Gene Ontology term (a), according to Reactome pathways (b) and cellular component in Gene Ontology term (c). The top 10 most significant biological processes and localization are presented. Enrichment p values were adjusted by Benjamini–Hochberg false discovery rate correction

Fig. 6

Analysis of STRING interactions for 22 proteins that were more abundant in prMSCs compared to MSCs: the network was built based on high density (0.7), and the edges of the network represent the significance of the interaction. The disabled nodes in the network have been hidden. All networks were enriched using the intersection of 8612 genes present on all platforms as a background, as well as data from experimental protein‒protein interactions, text mining, and curated databases. Proteins from the network of interactions that form the type I interferon signaling cluster according to the biological process (GO) are marked in blue. Proteins associated with interferon alpha/beta signaling and negative regulators of DDX58/IFIH1 signaling are highlighted in red, according to the local network cluster (STRING)