Increased extracellular vesicles release in the plasma of SLE patients

The images of transmission electron microscope (Fig. 1A) and the results of tunable resistive pulse sensing assay (Fig. 1B) showed that the average particle size of EVs in the plasma of SLE patients was 120.0 nm, which was consistent with the size of typical small EVs (< 200 nm) [18]. The immunoblotting results showed that SLE patient-derived EVs expressed EVs-specific markers (CD63, CD81, CD9, and Alix) (Fig. 1C). We further compared the levels of EVs in the plasma of SLE patients with different disease severity by using an ELISA-based assay (Fig. 1D). Significantly increased levels of circulating EVs were observed in the plasma of SLE patients compared with those in HC subjects and were associated with disease severity (active SLE: n = 30, 1.62 ± 0.36 × 109 EV particles/ml; inactive SLE: n = 20, 0.96 ± 0.26 × 109 EV particles/ml versus HC: n = 20, 0.59 ± 0.23 × 109 EV particles/ml, P < 0.005).

Fig. 1

SLE patient-derived immune complexes induced platelets activation and extracellular vesicles production through FcγIIA. (A)Transmission electron micrographs of purified extracellular vesicles (EVs) isolated from plasma of SLE patient. The scale bar in the image represents 50 nm. (B) The size distribution of EVs isolated from the plasma of SLE patient was analyzed using a tunable resistive pulse sensing assay. (C) Expression of EV-specific surface markers (CD63, CD81, CD9, and Alix [ALG-2-interacting protein X]) and a platelet-specific marker (CD41) in EVs from plasma was analyzed using immunoblotting. (D) Comparison of EVs levels in the plasma of SLE patients with different activity and HC subjects by using enzyme-linked immunosorbent assay (ELISA)-based assay. (E and F) The levels of (E) platelet factor 4 (PF4), and (F) CD62p in the plasma of SLE patients. (G) A positive correlation between the levels of CD62p and EVs release was shown in SLE patients. (H) Increased percentages of pEVs (CD41+CD63+) were released in the plasma of patients with SLE and are associated with disease severity. (I to K) SLE patient-derived immune complexes (ICs) induced normal platelets activation, which is accompanied by increased pEVs release through FcγIIA. All experiments were performed in triplicate, and the data are presented as the mean ± SD. The densitometric analysis of immunoblotting were presented in Additional file 2. *P < 0.05, **P < 0.01, ***P < 0.005. aSLE, active SLE; iSLE, inactive SLE; HC, healthy control

Increased extracellular vesicles in SLE are associated with platelet activation

Increased levels of platelet activation markers (e.g., platelet factor 4 [PF4], and CD62p) were detected in the plasma of SLE patients and were associated with disease severity (Fig. 1E and F). Moreover, a positive correlation between the levels of CD62p and EVs release was shown in SLE patients (r = 0.67, P < 0.0001, Fig. 1G). Elevated percentages of pEVs (CD63+CD41+) were shown in patients with active SLE (66.94 ± 10.01%, Fig. 1H), compared to those in inactive SLE patients (54.38 ± 7.62%, P < 0.005) or HC subjects (17.89 ± 6.45%, P < 0.005).

We observed that increased CD62p was induced in human platelets after treatment with SLE patient-derived ICs (Fig. 1I), suggesting that SLE ICs could activate platelets. To examine whether SLE ICs-induced platelet activation may contribute to pEVs release through binding to FcγRIIA, normal human platelets were treated with SLE ICs for 4 h in the presence or absence of anti-FcγRIIA antibody. The pEVs were isolated from supernatant using human CD63 Dynabeads magnetic separation technology (Thermo Fisher Scientific, USA). The levels of FcγRIIA, CD62p, and CD41 were significantly induced in platelets after treating with SLE ICs (left panel, Fig. 1J), which were accompanied by increased pEVs release (right panel, Fig. 1J). These effects were suppressed in the presence of anti-FcγRIIA antibody. We further confirmed our observations by using ICs isolated from different SLE patients (n = 6) and HC subjected (n = 6), and measured the levels of pEVs (CD63+CD41+) release using ELISA-based analysis (Fig. 1K, P < 0.005).

SLE patient-derived pEVs induced NETs formation

Next, we compared the levels of NETs DNA release in SLE patients with different disease activity. Elevated levels of NETs DNA were detected in SLE patients with different severity (active SLE: n = 30, 322.5 ± 92.5 ng/ml, P < 0.005; inactive SLE: n = 20, 190.0 ± 39.8 ng/ml, P < 0.005; Fig. 2A), compared to those in HC subjects (n = 20, 114.7 ± 46.4 ng/ml). To examine the effect of SLE patient-derived pEVs on NETs formation, human neutrophils were stimulated with active SLE-, inactive SLE-, and HC subject-derived pEVs (approximately 5 × 108 particles) for 4 h, respectively. Significantly increased NETs formation was shown in neutrophils after stimulation with SLE pEVs (active SLE: 390.32 ± 27.75 ng/ml; inactive SLE: 326.30 ± 28.71 ng/ml, Fig. 2B), compared with those in HC subject-derived pEVs (235.31 ± 15.01 ng/ml, P < 0.05). This effect decreased in the presence of the endocytosis inhibitor cytochalasin D (372.30 ± 11.02 versus 213.80 ± 9.97 ng/ml, P < 0.005, Fig. 2C) and the vacuolar type H+-ATPase inhibitor bafilomycin A1 (219.42 ± 10.68 ng/ml, P < 0.005), suggesting that SLE pEVs induced NETs formation by uptake and activation of endolysosomal Toll-like receptors (e.g., TLR3, 7, 8 and 9).

Fig. 2

SLE patient-derived pEVs induced NETs formation. (A) Increased NETs DNA release in plasma of SLE patients with different disease activity. (B) Increased levels of NETs DNA were detected in normal human neutrophils after treatment with SLE patient-derived pEVs (approximately 5 × 108 particles), and (C) this effect was decreased in the presence of the endocytosis inhibitor cytochalasin D (Cyt D, 10µM) or the vacuolar type H+-ATPase inhibitor bafilomycin A1 (BafA1, 100nM). (D) A dramatically elevated level of TLR8 was shown in normal neutrophils after treatment with SLE pEVs, compared to those with healthy control (HC) subject-derived pEVs treatment or mock control cells. (E and F) SLE pEVs induced NETs formation through TLR8 activation (E, left panel). This effect was diminished in TLR8 knockdown cells (E, right panel), and (F) in the presence of Cyt D or the TLR8-specific inhibitor Cu-CPT9a (10µM). (G) pEVICs enhanced ICs-induced NETs formation, and this effect was suppressed in the presence of Cu-CPT9a, or in (H) TLR8 knock-down cells. The scale bar in the IFA image represents 5 μm. All experiments were performed in triplicate, and the data are presented as mean ± SD. The densitometric analysis of immunoblotting were presented in Additional file 2. *P < 0.05, **P < 0.01, ***P < 0.005

Given that mature neutrophils express all TLRs except TLR3 [19], we analyzed the intracellular expressions of TLR7, TLR8, and TLR9 in human neutrophils after treatment with pEVs from SLE patients (n = 6) and HC subjects (n = 6). Compared to TLR7 and TLR9, a dramatically elevated level of TLR8 was shown in neutrophils after treatment with SLE pEVs (MFI, 111.31 ± 13.87 versus 24.47 ± 5.02, P < 0.005, Fig. 2D). We further confirmed the effect of TLR8 on SLE pEVs-induced NETs formation by using TLR8 knockdown cells. As shown in Fig. 2E, SLE pEVs-induced NETs-associated proteins (e.g., MPO, peptidylarginine deiminases 4 [PAD4], and citrullinated histone H3 [citH3]) were almost completely suppressed in TLR8-knockdown cells.

Next, we explored whether SLE ICs-induced pEVs (pEVICs) may contribute to NETs formation. Significantly increased NETs formation was induced in normal neutrophils after treatment with pEVICs (Fig. 2F), and this effect was suppressed in the presence of cytochalasin D or the TLR8-specific inhibitor Cu-CPT9a, but no effect was shown in cells after treatment with the TLR7-specific inhibitor IRS661. Our results suggested that pEVICs may carry specific ssRNA to induce NETs formation directly by uptake and TLR8 activation. Additionally, it is worth noting that dramatically increased ICs-primed NETs formation was shown in the presence of pEVICs (565.2 ± 24.74 ng/ml versus 427.7 ± 21.59 ng/ml, P < 0.01, Fig. 2G), which suggested that pEVICs play a crucial role in hyperactive NETs formation. This effect was reduced slightly under IRS661 treatment (487.8 ± 20.63, P < 0.05), but was almost completely suppressed in the presence of Cu-CPT9a (315.4 ± 5.01, P < 0.01, Fig. 2G). We confirmed this observation by using TLR8 knockdown neutrophils (Fig. 2H and Supplementary Fig. S1), suggesting that ssRNA carried by pEVICs contribute NETs enhancement through TLR8 activation.

SLE patient-derived pEVs carried tsRNA to induce NETs formation

Yang et al. identified that several EV-carried tsRNAs were upregulated in SLE patients, but their pathogenic role was unclear [9]. Based on their results, we choose three candidates (tRF-His-GTG-1, tRF-chrM.Pro-TGG and tRF-Val-AAC-1-M7) for further study. We assessed the association between these tsRNA candidates, pEVs, and SLE disease. Human platelets were stimulated with different SLE ICs (n = 3) for 1 h, and the levels of tsRNAs loaded into pEVs were quantified using QRT-PCR (Fig. 3A). There was no significant difference in the expression of tRF-Val-AAC-1-M7. But significantly increased levels of tRF-His-GTG-1 (7.42 ± 0.75-fold, P < 0.005) and tRF-chrM.Pro-TGG (1.67 ± 0.04-fold, P < 0.01) were both shown in pEVs released from platelets after stimulation with SLE ICs, and this effect was dose-dependent (Fig. 3B). We validated the levels of tRF-His-GTG-1 and tRF-chrM.Pro-TGG in pEVs by using QRT-PCR (Fig. 3C). Elevated tRF-His-GTG-1 (active SLE: 4.65 ± 2.49-fold, P < 0.005; inactive SLE: 1.84 ± 1.32-fold, P < 0.05; HC: 1.00 ± 0.76-fold), and tRF-chrM.Pro-TGG (active SLE: 2.60 ± 1.57-fold, P < 0.005; inactive SLE: 1.22 ± 0.60-fold; HC: 1.00 ± 0.39-fold) expression was revealed in pEVs of SLE patients and associated with disease severity (SLEDAI score) (tRF-His-GTG-1: AUC = 0.86, 95% CI = 0.77–0.94, P < 0.0001; tRF-chrM.Pro-TGG: AUC = 0.92, 95% CI = 0.85–0.99, P < 0.0001, Fig. 3D). We further analyzed the association between pEVs-carried tsRNAs and NETs DNA release. The expression of tRF-His-GTG-1 was positively correlated with NETs DNA release (r = 0.63, P < 0.01, Fig. 3E), but tRF-chrM.Pro-TGG was not.

Fig. 3

Elevated levels of tRF-His-GTG-1 in SLE patients-derived pEVs induce NETs formation through TLR8 activation and ROS production. (A) Increased levels of tRF-His-GTG-1, and tRF-chrM.Pro-TGG were shown in EVs released from normal platelets after stimulation with SLE patients-derived ICs (I) compared to those in mock control cells (M), and (B) this effect was dose-dependent. (C) Elevated tRF-His-GTG-1 and tRF-chrM.Pro-TGG expression were revealed in pEVs of patients with SLE, and associated with disease severity. (D) Receiver operating characteristic (ROC) curve analysis of tRF-His-GTG-1 and tRF-chrM.Pro-TGG expression in SLE patients with different severity (SLEDAI score), respectively. (E) The expression of tRF-His-GTG-1 was positively correlated with NETs DNA release levels in patients with SLE. (F to H) The tRF-His-GTG-1 mimic or mimic control (10µM) were loaded into human platelet-derived EVs (pEVs) using electroporation. Human neutrophils were treated with pEVs that carried-tRF-His-GTG-1 mimic or mimic control in the presence of tRF-His-GTG-1 inhibitor, tsRNA inhibitor control, or TLR8 inhibitor for 4 h. (F) NETs formation was observed using confocal microscopy (upper panel) and quantified by the MPO-DNA PicoGreen assay (right below panel). The intracellular tRF-His-GTG-1 in the neutrophils were measured by QRT-PCR (Left below panel). (G) The levels of cytosolic ROS were detected using flow cytometry with dihydrorhodamine (DHR) 123 staining. (H) The expression of intracellular TLR8 and NETs-associated proteins was analyzed by using immunoblotting. Immunoblotting bands from β-actin were densitometrically measured by ImageJ to determine the lane normalization factor for samples. The scale bar in the IFA image represents 5 μm. The image shown is from a single experiment that is representative of at least three separate experiments. Data are presented as the mean ± SD. The densitometric analysis of immunoblotting were presented in Additional file 2. *P < 0.05, **P < 0.01, ***P < 0.005. ns, non-significant

Platelet-derived tRF-His-GTG-1-induced NETs formation is TLR8 dependent

Therefore, we focused on the effect of pEVs-carried tRF-His-GTG-1 on NETs formation. We loaded tRF-His-GTG-1 mimics or tsRNA control mimics into human pEVs by using electroporation [20]. The tRF-His-GTG-1-loaded pEVs were then added to human neutrophils. After 4 h, significantly increased intracellular tRF-His-GTG-1 levels (left lower panel, Figs. 3F and 4A), accompanied by elevated NETs formation (upper panel, Fig. 3F) and NETs DNA release (right lower panel, Fig. 3F), were observed in neutrophils with the addition of tRF-His-GTG-1 mimics-loaded pEVs. This effect was almost completely suppressed in the presence of the TLR8-specific inhibitor or tRF-His-GTG-1 inhibitor. Our results showed that ICs primed platelet-derived tRF-His-GTG-1 could induce NETs formation directly through EVs transmission and TLR8 activation.

Fig. 4

pEVs-carried tRF-His-GTG-1 enhanced NETs formation and induced IL-1β/IL-8/interferon-α upregulation by TLR8 activation. (A) Human platelet-derived EVs (pEVs) were loaded with Cy5-labeled tRF-His-GTG-1 mimic (red) or mimic control (red) and then co-cultured with human neutrophils for 4 h. Endogenous TLR8 was stained with TLR8 antibodies (green). The tRF-His-GTG-1 transmission and endogenous TLR8 in neutrophils was detected by using the immunofluorescence assay. The scale bar in the image represents 10 μm. (B) Human neutrophils were treated with tRF-His-GTG-1 mimic, mimic control, TLR8- (ssRNA40), TLR7- (R837), or TLR7/8- (R848) agonist, respectively. The RNA-binding protein was collected using the RNA-protein pull-down assay kit and analyzed by immunoblotting with specific TLR7-, and TLR8 antibodies. (C and D) The HEK-Blue hTLR8 cells were stimulated with pEVs from the indicated individuals or (D) indicated tsRNA-loaded pEVs for 24 h. NF-κB activation was evaluated in terms of luciferase activity compared to control cells. (E to G) Human neutrophils were treated with pEVs from the indicated individuals or indicated tsRNA-loaded pEVs in TLR8 knock-down cells, or in the presence of TLR8 agonist for 24 h. The (E and G) intracellular and (F) extracellular proinflammatory cytokines, and interferon α were analyzed using immunoblotting and ELISA, respectively. (H to J) tRF-His-GTG-1 inhibitor suppressed SLE-ICs/SLE-ICs-primed pEVs induced (H, I) NETs formation/ hyperactivation, and (J) IL-1β, IL-8 and interferon α production. Immunoblotting bands from β-actin were densitometrically measured by ImageJ to determine the lane normalization factor for samples. The scale bar in the IFA image represents 5 μm. The image shown is from a single experiment that is representative of at least three separate experiments. Data are presented as the mean ± SD. The densitometric analysis of immunoblotting were presented in Additional file 2. **P < 0.01, ***P < 0.005

Reactive oxygen species (ROS) are essential in the regulation of NETs formation in SLE [12]. We showed that pEV-carried tRF-His-GTG-1 induced cytosolic ROS production in neutrophils through TLR8 activation (Fig. 3G). The phosphorylation of p47phox is a key factor in ROS production, which is regulated by the TLR8-mediated ERK (extracellular-signal-regulated kinase) and p38 MAPK (mitogen-activated protein kinase) signaling pathways [21]. We showed that elevated levels of ERK and p38 MAPK activation were induced in cells with tRF-His-GTG-1-loaded pEVs treatment, accompanied by increased p47phox phosphorylation and the enhancement of NETs-associated proteins. These effects were inhibited efficiently in the presence of tRF-His-GTG-1 inhibitor (Fig. 3H), or Cu-CPT9a (Supplementary Fig. S2).

tRF-His-GTG-1 interacts with TLR8

Next, we assessed the association between SLE ICs and tRF-His-GTG-1 expression in neutrophils. Increased levels of intracellular tRF-His-GTG-1 were revealed in normal neutrophils after treatment of ICs from different SLE patients (n = 5), and the effect was dose-dependent (P < 0.005, Supplementary Fig. S3A). A dramatically elevated intracellular tRF-His-GTG-1 was detected in ICs-primed neutrophils in the presence of pEVICs (P < 0.01, Supplementary Fig. S3B), and this effect was diminished in the presence of cytochalasin D (P < 0.005, Supplementary Fig. S3B). We further examined whether tRF-His-GTG-1 could bind to TLR8 on neutrophils by using immunofluorescence assay. Platelet-derived EVs were loaded with Cy5-labeled tRF-His-GTG-1 mimic or mimic control and then co-cultured with human neutrophils for 4 h. The tRF-His-GTG-1 transmission and endogenous TLR8 in human neutrophils were detected by using immunofluorescence assay. As shown in Fig. 4A, colocalization of tRF-His-GTG-1 and TLR8 was detected in neutrophils, and this effect was inhibited in the presence of tRF-His-GTG-1 inhibitor or Cu-CPT9a. We further performed RNA-protein pull-down assay to confirm that tRF-His-GTG-1 could induce endogenous TLR8 expression and interact with TLR8 directly (Fig. 4B).

pEVs carried tRF-His-GTG-1 induced IL-1β/IL-8/interferon-α upregulation in neutrophils through TLR8 activation

IL-1β, IL-8 and IFN-α were significantly increased in neutrophils of SLE patients and correlated with disease activity [22, 23], which were regulated by NF-κB and IRF7 signaling pathways through TLR8 and adaptor protein MyD88 activation, respectively [24]. Therefore, we examined whether SLE pEVs could regulate NF-κB activity by TLR8 activation. As shown in Fig. 4C, elevated NF-κB activity was revealed in cells with SLE pEVs modulation upon TLR8 engagement by using the HEK-hTLR8 cell model (InvivoGen, USA). In addition to NF-κB, increased levels of IRF7 activation were detected in neutrophils with SLE pEVs (Fig. 4E), accompanied by upregulated IL-1β, IL-8, and IFNα (Fig. 4E and F). These effects were diminished in TLR8-knockdown cells (Fig. 4E) or in the presence of Cu-CPT9a and tRF-His-GTG-1inhibitor (Fig. 4F). Consistent with this, tRF-His-GTG-1-loaded pEVs also had the same effects on NF-κB activation (Fig. 4D), the phosphorylation of p65 subunits and IRF7 activation (Fig. 4G), resulting in elevated IL-1β, IL-8, and IFNα. The above effects were completely inhibited in the presence of the Cu-CPT9a, or tRF-His-GTG-1 inhibitor.

tRF-His-GTG-1 inhibitor suppressed SLE-derived ICs primed NETs formation and IL-1β/IL-8/interferon-α production

Finally, we assessed the effects of tRF-His-GTG-1 inhibitor on SLE ICs (n = 5) and/or ICs-primed pEVs on NETs formation/hyperactivation. As shown in Fig. 4H, SLE ICs and/or ICs-primed pEVs induced NETs formation/ hyperactivation was significantly suppressed in the presence of tRF-His-GTG-1 inhibitor (P < 0.005), accompanied by downregulated IL-1β, IL-8, and IFNα (Fig. 4I and J).