Plasma-derived DNA containing-extracellular vesicles induce STING-mediated proinflammatory responses in dermatomyositis

Objectives: Extracellular vesicles (EVs) are lipid bilayer membrane vesicles that are present in various bodily fluids and have been implicated in autoimmune disease pathogenesis. Type I interferons (IFN), specifically IFN-β, are uniquely elevated in dermatomyositis (DM). The stimulator of interferon genes (STING) works as a critical nucleic acid sensor and adaptor in type I IFN signaling with possible implications in autoimmune diseases such as DM. In the current study, we investigated whether circulating EVs contribute to proinflammatory effects in DM, whether these proinflammatory responses are mediated by the STING signaling pathway, and if so, by what mechanism STING is activated. Methods: We collected and characterized EVs from plasma of healthy controls (HC) and DM patients; analyzed their abilities to trigger proinflammatory cytokines release by ELISA, and explored STING signaling pathway activation using immunoblot and immunofluorescent staining. STING signaling pathway inhibitors and RNAi were used to further investigate whether STING was involved in EVs-triggered proinflammatory response. DNase/lipid destabilizing agent was utilized to digest EVs and their captured DNA contents to evaluate how EVs triggered STING-mediated proinflammatory response in DM. Results: EVs isolated from DM plasma triggered proinflammatory cytokines including type I IFN release with STING signaling pathway activation. The activated STING pathway was preferentially mediated by dsDNA captured by EVs. Suppression of STING or its downstream signaling proteins attenuated the EVs-mediated proinflammatory response. Conclusions: Plasma-derived, DNA containing-EVs induced STING-mediated proinflammatory effects in DM. Targeting the STING pathway may be a potential therapeutic approach for DM.

. The concentration and average size of large extracellular vesicles-derived from DM plasma and HC plasma. (A) Total number of lEVs derived from healthy plasma (5.358x10 9 ± 6.745x10 9 particles/mL, n = 6) and DM plasma (2.786x10 9 ± 3.140x10 9 particles/mL, n = 7). (B) Average size of lEVs derived from healthy plasma (177.0 ± 5.539nm, n = 6) and DM plasma (173.5 ± 5.092 nm, n = 7). Data were represented as mean ± SD. Comparison between two groups was analyzed by the Student t test. (A) Immunoblot image showed different surface makers CD81, CD9, CD63 expression level from same number (3.125x10 8 particles) of sEVs derived from HC plasma (n = 3) and DM plasma (n = 3). (B) Relative intensity of CD81 expression level from same number of sEVs derived from HC plasma (n = 5) and DM plasma (n = 7). (C) Relative intensity of CD63 expression level from same number of sEVs derived from HC plasma (n = 5) and DM plasma (n = 7). (D) Relative intensity of CD9 expression level from same number of sEVs derived from HC plasma (n = 5) and DM plasma (n = 7). Data were represented as mean ± SD. Comparison between two groups was analyzed by the Student t test.

Figure S3. Proinflammatory cytokines captured by HC or DM plasma-derived EVs.
Both HC and DM plasma-derived EVs captured IFNβ and IL-6 were undetectable, TNFα captured by HC and DM plasma-derived EVs is very low and has no significant difference. HC plasma-derived EVs (n = 6) and DM plasma-derived EVs (n = 6). Data were represented as mean ± SD. Comparison between two groups was analyzed by the Student t test.

Figure S4. HC or DM plasma-derived EVs had no effects on cell viability of PBMCs.
PBMCs were placed in 96-well plates at a density of 1.5x10 6 cells/mL, 100 μL/well. After stimulation of cells with 5 μL of large EVs or small EVs derived from 50 μL of HC or DM plasma for 15 h, the plates were incubated with 10 μL/well of WST-8 in the incubator with 95% air and 5% CO2 at 37℃ for 1 h, and then measured at 460 nm absorbance wavelength. HC plasma-derived EVs (n = 3) and DM plasma-derived EVs (n = 3) were studied. Data are represented as mean ± SD. Comparison between two groups was analyzed by the Student t test.

Figure S5. EVs internalized by PBMCs.
Cells were stimulated with 12.5 μL of precipitated CM-DiI dye following ultracentrifugation or CM-DiI dye-labeled EVs for 15 h. The cells were fixed, incubated with 1 μg/mL of DAPI in PBS for 1 min, washed 3 times, and then studied by using confocal microscopy. PBMCs were placed in EZ slide at a density of 1.5x10 6 cells/mL, 500 μL/well. The upper panel showed no interaction of the precipitated CM-DiI dye with PBMCs; the lower panel showed that CM-DiI dye-labeled EVs could be properly internalized by PBMCs (Scale bar = 10 μm).

Figure S6. H-151 suppressed 2'3'-cGAMP-induced STING signaling pathway phosphorylation and proinflammatory cytokine release in PBMCs.
PBMCs were placed in 6-well plates at a density of 1.5x10 6 cells/mL, 2 mL/well. The cells were pretreated with/without 1 μM of H-151 for 1 h, and then stimulated in the presence/absence of 10 μg/mL or 20 μg/mL of 2'3'-cGAMP for 15 h. After stimulation, the cells were collected for immuno-blot to check the changes in the STING signaling pathway. ELISAs were performed to examine proinflammatory cytokines release in the supernatants. Data in (B,C,D) represent mean ± SD. *P < 0.05 **P < 0.01 ***P < 0.001 between groups as indicated. Comparison among three or more groups was performed using ANOVA, followed by Student-Newman-Keuls test.

Figure S7. DM plasma-derived large extracellular vesicles induced STING-mediated IFNβ production in PBMCs.
(A) lEVs derived from DM plasma induced IFNβ release in PBMCs. H-151 attenuated IEVstriggered IFNβ release in PBMCs. (B) STING antagonist H-151 and lEVs derived from DM plasma had no effect on TNFα or (C) IL6 release in PBMCs. (D) Immunoblot image showed that lEVs derived from DM plasma induced Serine 366 phosphorylation of STING in PBMCs. STING antagonist H-151 attenuated lEVs-triggered STING phosphorylation. Data in (A,B,C) represent mean ± SD. **P < 0.01 between groups as indicated. Comparison among three or more groups was performed using ANOVA, followed by Student-Newman-Keuls test.

Figure S8. FAM-siSTING could be transfected and decreased STING expression in PBMCs.
PBMCs were placed in EZ slide at a density of 1.5x10 6 cells/mL, 500 μL/well. After cells were transfected by FAM-labelled siSTING for 36 h, the cells were fixed, permeabilized, and stained for STING. Cells were then incubated with 1 μg/mL of DAPI in PBS for 30 s, washed 3 times, and then studied by using fluorescent microscopy. (Scale bar = 31 μm).

Figure S9. DM plasma-derived sEVs triggered STING phosphorylation with IFNβ production in WT (B6) mice eluted peritoneal macrophages other than STING KO (Sting1 gt/gt ) mice peritoneal macrophages.
WT mice and STING KO mice eluted peritoneal macrophages were stimulated with/without 50 μL of DM plasma-derived sEVs for 15 h. (A) Immunoblot image showed that DM plasmaderived sEVs triggered STING phosphorylation in B6 mice eluted peritoneal macrophages other than Sting1 gt/gt mice eluted macrophages (B). B6 mice expressed total STING while Sting1 gt/gt mice dose not expressed total STING (C). sEVs triggered more IFNβ production in B6 macrophages but not in Sting1 gt/gt macrophages (D). Data in (B,C,D) represent mean ± SD. **P < 0.01 ***P < 0.001 between groups as indicated. Comparison among three or more groups was performed using ANOVA, followed by Student-Newman-Keuls test.