Role of gamma-secretase in human umbilical-cord derived mesenchymal stem cell mediated suppression of NK cell cytotoxicity

Abstract

Background: Mesenchymal stem cells (MSCs) are increasingly considered to be used as biological immunosuppressants in hematopoietic stem cell transplantation (HSCT). In the early reconstitution phase following HSCT, natural killer (NK) cells represent the major lymphocyte population in peripheral blood and display graft-vs-leukemia (GvL) effects. The functional interactions between NK cells and MSCs have the potential to influence the leukemia relapse rate after HSCT. Until date, MSC-NK cell interaction studies are largely focussed on bone marrow derived (BM)-MSCs. Umbilical cord derived (UC)-MSCs might be an alternative source of therapeutic MSCs. Thus, we studied the interaction of UC-MSCs with unstimulated allogeneic NK cells.

Results: UC-MSCs could potently suppress NK cell cytotoxicity in overnight cultures via soluble factors. The main soluble immunosuppressant was identified as prostaglandin (PG)-E2. Maximal PGE2 release involved IL-1β priming of MSCs after close contact between the NK cells and UC-MSCs. Interestingly, blocking gamma-secretase activation alleviated the immunosuppression by controlling PGE2 production. IL-1 receptor activation and subsequent downstream signalling events were found to require gamma-secretase activity.

Conclusion: Although the role of PGE2 in NK cell-MSC has been reported, the requirement of cell-cell contact for PGE2 induced immunosuppression remained unexplained. Our findings shed light on this puzzling observation and identify new players in the NK cell-MSC crosstalk.

Figure 1

Phenotype of UC-MSCs. UC-MSCs were detached using accutase and stained with antibodies against CD14, CD19, CD34, CD45, CD73, CD90, CD105, and HLA-DR (shaded histograms) or isotype control antibodies (filled histograms).

Figure 2

Suppression of NK cell cytotoxicity by UC-MSCs. A: NK cells were cultured alone or with MSCs at NK cell: MSC ratio of 5:1, 10:1, and 20:1. Following overnight co-culture, CD107a degranulation assay was performed with K562 target cells. The bar graphs represent the percentage of CD107a⁺ NK cells (n = 5) B: NK cells were cultured without MSCs or with MSCs at NK cell: MSC ratio of 10:1. CD107a degranulation assay was performed with K562 target cells (n = 5). Following overnight co-culture, CD107a expression on CD56 ^bright NK cells was analysed. The bar graphs represent the percentage of CD107a⁺ cells (n = 9). C, D: NK cells were cultured overnight with or without MSCs. NKG2D expression on the NK cells was analysed by flow cytometry. The bar graphs depict the mean fluorescence intensity (MFI) of NKG2D staining on CD56 ^dim (C; n = 9) and CD56 ^bright (D; n = 8) NK cells. E, F, G, H: NK cells were cultured overnight with or without MSCs. Perforin (E; n = 7), granzyme A (F; n = 7), granzyme B (G; n = 5), and granzyme K (H; n = 7) content of the NK cells was analysed by flow cytometry.

Figure 3

Mechanism of MSC-mediated suppression of NK cell degranulation. A: NK cells were cultured in normal medium or in MSC cm or in NK-MSC cm. CD107a degranulation assay was performed with K562 as target cells. Representative dot plots of the alterations in NK cell degranulation are shown (n = 3). B: NK cell-mediated cytotoxicity was determined by chromium release assay in different conditioned media (n = 3). C: UC-MSCs were cultured overnight with or without NK cells. Intracellular staining for COX-2 was performed and analysed by flow cytometry. One representative figure and bar graphs (n = 5) depicting COX-2 upregulation in UC-MSCs cultured with NK cells in comparison with COX-2 expression in “unprimed” UC-MSCs is shown. D: NK cells were cultured with or without MSCs or co-cultured in presence of NS-398 or DMSO. CD107a degranulation assay was performed with K562 target cells. The bars represent the effect of COX-2 inhibition on degranulative capacity of NK cells (n = 5). E: NK cells from four different donors were cultured for 16 hours and culture supernatants were collected. The bars represent the IL-1β levels measured in the supernatants (n = 4). F: MSCs were co-cultured with NK cells in presence of IL-1β neutralising antibody or matched isotype antibody. The bars illustrate the effect of IL-1β blocking on COX-2 expression (n = 4).

Figure 4

Effect of gamma-secretase on IL-1 signalling and PGE2 content. A: UC-MSCs were cultured in presence of different doses of IL-1β stimulation with or without DAPT (gamma-secretase inhibitor). UC-MSCs were detached using accutase and stained to analyse intracellular COX-2 expression (n = 5). B: Gamma-secretase activity in UC-MSCs was silenced using siRNA targeting a catalytic subunit of gamma-secretase (PSEN-1). Scrambled siRNA (SCR) was also introduced into UC-MSCs as control. After 24 hours, PSEN-1 or SCR siRNA treated MSCs were stimulated with IL-1β overnight. UC-MSCs were detached using accutase and stained to compare intracellular COX-2 expression with or without IL-1β stimulation (n = 3). C: UC-MSCs were stimulated with IL-1β for 40 minutes, with or without 2.5 hours pre-treatment with DAPT. MSCs were detached using accutase, immediately fixed, permeabilised and stained to analyse intracellular pJNK levels (n = 4). D: NK cells were cultured with or without MSCs or co-cultured in presence of 6 μM DAPT, 10 μM DAPT or DMSO. CD107a degranulation assay was performed with K562 target cells. The bars represent the effect of gamma-secretase inhibition on degranulative capacity of NK cells (n = 5). E: NK cells were cultured in NK cell conditioned media (NK cm) or in NK-MSC conditioned media (NK-MSC cm) or NK-MSC conditioned media where they were co-cultured in presence of 10 μM DAPT or DMSO (NK-MSC-10 μM DAPT cm or NK-MSC-DMSO cm respectively). Chromium release assay was performed with K562 target cells (n = 3). F: PGE2 concentration in MSC cm, NK-MSC cm or NK-MSC-10 μM DAPT cm as determined by competitive ELISA (n = 3).