Differential and transferable modulatory effects of mesenchymal stromal cell-derived extracellular vesicles on T, B and NK cell functions

Abstract

Mesenchymal stromal cells (MSCs) are multipotent cells, immunomodulatory stem cells that are currently used for regenerative medicine and treatment of a number of inflammatory diseases, thanks to their ability to significantly influence tissue microenvironments through the secretion of large variety of soluble factors. Recently, several groups have reported the presence of extracellular vesicles (EVs) within MSC secretoma, showing their beneficial effect in different animal models of disease. Here, we used a standardized methodological approach to dissect the immunomodulatory effects exerted by MSC-derived EVs on unfractionated peripheral blood mononuclear cells and purified T, B and NK cells. We describe here for the first time: i. direct correlation between the degree of EV-mediated immunosuppression and EV uptake by immune effector cells, a phenomenon further amplified following MSC priming with inflammatory cytokines; ii. induction in resting MSCs of immunosuppressive properties towards T cell proliferation through EVs obtained from primed MSCs, without any direct inhibitory effect towards T cell division. Our conclusion is that the use of reproducible and validated assays is not only useful to characterize the mechanisms of action of MSC-derived EVs, but is also capable of justifying EV potential use as alternative cell-free therapy for the treatment of human inflammatory diseases.

Figure 1. MSC immunomodulation is mediated by paracrine molecules.

(a) Schematic representation of Transwell^® system with MSCs in the bottom well and IECs in the top well. A 0.4 μm-porous membrane was used to prevent cell-cell interaction and permit soluble molecule exchange. Sorted-IECs (T, B and NK cells) were stimulated with specific stimuli and cultured alone or in the presence of resting or primed allogeneic MSCs. At the end of co-culture, IEC proliferation was assessed using carboxyfluorescein succinimidyl ester (CFSE) dilution method, as described in Materials and Methods section. CFSE fluorescence was analyzed after 6 days for T (at 10:1 T/MSC ratio) and NK (at 1:1 NK/MSC ratio) cells (b,d, respectively), while for B cells (c) the fluorescence was detected after 4 days of co-culture (at 1:1 B/MSC ratio). The same IEC:MSC ratios were maintained to assess the effect of MSC paracrine molecules on sorted-T, -B and -NK cells (b–d, respectively) proliferation by use of Transwell^® 24 system. The results are expressed as relative proliferation percentage of IECs, normalized to IEC cultured alone (100%). Error bars represented mean ± SEM of twelve independent experiments for standard immunological assays and four independent experiments for Transwell^® assays. ^***P < 0.001.

Figure 2. Internalization of MSC-derived EVs by IECs.

Resting and primed PKH26-MSCs were cultured in presence of unstimulated PBMCs or sorted-T, -B or -NK cells in order to assess the transfer of MSC-derived EVs to IECs. After 4 days, the cells were harvested and labeled with anti-CD45, anti-CD3, anti-CD14, anti-CD56, anti-CD19 to identify the different IEC lineage inside unfractionated PBMCs (a); anti-CD45 were used for sorted-IECs (b–e). TOPRO-3 was added to detect viable cells. The EV-uptake by IECs was detected as percentage of lineage specific^pos/PKH26^pos IECs by FACS. (c) Representative immunofluorescence staining of CD45^pos/PKH26^pos IECs. At the end of co-cultures, cells were detached and labeled with anti-CD45 (green) and TOPRO-3 (blue nuclei) to assess the incorporation of PKH26-EVs (red). Scale bars: 5 μm. Images were obtained by LSM 710 confocal microscopy (Zeiss) at 63x magnification. (d) EV-internalization by stimulated CFSE labeled IECs was evaluated after 6 days for T and NK cells and after 4 days for B cells. (e) CFSE plot representative of three independent experiments, showing the localization of EVs inside IEC generation as percentage of CFSE^pos/PKH26^pos IECs. Error bars represented mean ± SEM.

Figure 3. Inhibition of EV secretion impairs MSC immunosuppression.

(a) Schematic representation of EV isolation protocol. (b) EV gate was carried out by using size-calibrated fluorescent beads ranging from 0.1 μm to 0.9 μm. The number EVs were calculated using 4.3 μm TruCount beads, which are shown in the upper right corner. The absolute count of EVs was subtracted to background noise events from 0.22 μm-filtered PBS. (c) Count of EVs obtained from MSCs treated for 48 hours with 10 μM GW4869, 10 μM imipramine, 60 μM DEVD and relative control vehicles, including: DMSO for GW4869 and DEVD; H2O for Imipramine. The results are expressed as percentage of relative EV-release inhibition, normalized to number of EVs obtained from untreated MSCs (100%). Resting and primed MSCs treated with GW4869, imipramine and DEVD were cultured in presence of activated CFSE labeled T, B and NK cells (d–f, respectively) in order to assess the effects of the inhibition of EV release on the immunomodulatory properties of MSCs. The results are expressed as relative proliferation percentage of IECs, normalized to IEC cultured alone (100%). Error bars represented mean ± SEM of four independent experiments. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001.

Figure 4. Characterization of MSC-derived EVs.

(a) Size distribution of resting and primed EVs obtained by NTA. (b) Immunoblot analysis of Giantin, LAMP1, Alix, GRP78, HSP70 and CD9 expression in resting and primed MSCs and purified EVs. This blot is representative of three independent experiments showing the same trends. (c) Representative plots of the immunophenotypic analysis of MSC-derived EVs showing the expression profile of a specific exosome marker (CD63), mesenchymal stromal cell markers (CD73, CD90 and CD105), adhesion molecules (ICAM-1 and VCAM-1) and MHC class I and II (HLA-ABC and HLA-DR, respectively). The histograms display the isotopic controls (dotted curve) and specific markers of resting (filled curve) and primed (black curve) EVs. (d) Graph showing quantitative differences of EV release between resting and primed MSCs. Data represented as mean ± SEM of ratio between number of EVs and number of cells of origin obtained from 33 independent experiments. ^***P < 0.001.

Figure 5. Immunosuppressive properties of MSC-derived EVs.

EVs were purified from resting and primed MSCs and added to unfractionated PBMCs or sorted-T, B and NK cells that were activated by specific stimuli (1 × 10⁴:3 × 10⁶ IEC:EV ratio). At the end of co-cultures, the cells were harvested and IEC proliferation was assessed by FACS analysis. CFSE fluorescence was analyzed after 4 days for PBMCs, T and B cells (a–c, respectively), while for NK cells (d) the fluorescence was analyzed after 6 days of co-culture. The results are expressed as relative proliferation percentage of IECs, normalized to IEC cultured alone (100%). Error bars represented mean ± SEM of six for PBMCs, and five for sorted-T, NK cells and B cells independent experiments. ^*P < 0.05, ^**P < 0.01, ^***P < 0.001.

Figure 6. Effect of EVs on immunomodulatory properties of MSCs.

(a) Methodological approach to induce MSC priming with EVs (as described in Materials and Methods section). (b) Immunophenotypic analysis of MSCs exposed to resting or primed EVs. IFN-γ/TNF-α-treated MSCs were used as positive control. The histograms show the isotopic controls (dotted curve) and specific markers of control (filled curve) or treated (black curve) MSCs derived from three independent experiments. (c) Resting and primed EVs were incubated with resting MSCs to induce the immunosuppressive phenotypic switch. After 48 hours, sorted-T, -B and –NK cells were added to MSC cultures to assess their immunomodulatory properties. IFN-γ/TNF-α-treated MSCs were used as positive control. CFSE fluorescence was analyzed after 6 days for T and NK, while for B cells the fluorescence was analyzed after 4 days of co-culture. The results are expressed as percentage of relative increase in MSC immunosuppression, normalized to the effect of untreated MSC on IECs (100%). Error bars represented mean ± SEM of five independent experiments. ^*P < 0.05, ^***P < 0.001.