CLEC5A and TLR2 are critical in SARS-CoV-2-induced NET formation and lung inflammation

Coronavirus-induced disease 19 (COVID-19) infects more than three hundred and sixty million patients worldwide, and people with severe symptoms frequently die of acute respiratory distress syndrome (ARDS). Recent studies indicated that excessive neutrophil extracellular traps (NETs) contributed to immunothrombosis, thereby leading to extensive intravascular coagulopathy and multiple organ dysfunction. Thus, understanding the mechanism of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)-induced NET formation would be helpful to reduce thrombosis and prevent ARDS in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. We incubated SARS-CoV-2 with neutrophils in the presence or absence of platelets to observe NET formation. We further isolated extracellular vesicles from COVID-19 patients' sera (COVID-19-EVs) to examine their ability to induce NET formation. We demonstrated that antagonistic mAbs against anti-CLEC5A mAb and anti-TLR2 mAb can inhibit COVID-19-EVs-induced NET formation, and generated clec5a−/−/tlr2−/− mice to confirm the critical roles of CLEC5A and TLR2 in SARS-CoV-2-induced lung inflammation in vivo. We found that virus-free extracellular COVID-19 EVs induced robust NET formation via Syk-coupled C-type lectin member 5A (CLEC5A) and TLR2. Blockade of CLEC5A inhibited COVID-19 EVs-induced NETosis, and simultaneous blockade of CLEC5A and TLR2 further suppressed SARS-CoV-2-induced NETosis in vitro. Moreover, thromboinflammation was attenuated dramatically in clec5a−/−/tlr2−/− mice. This study demonstrates that SARS-CoV-2-activated platelets produce EVs to enhance thromboinflammation via CLEC5A and TLR2, and highlight the importance of CLEC5A and TLR2 as therapeutic targets to reduce the risk of ARDS in COVID-19 patients.

Thrombotic complication is a major cause of morbidity and mortality in patients with COVID-19 [9]. Even though the pathogenesis of thrombo-emboli formation Open Access *Correspondence: slhsieh@gate.sinica.edu.tw Sung et al. Journal of Biomedical Science (2022) 29:52 in COVID-19 patients is still unclear, recent studies indicate that excessive NET production is associated with thrombo-emboli formation in human diseases. It has been reported that elevated levels of cell-free DNA, myeloperoxidase, and citrullinated histone H3 are noted in the sera of COVID-19 patients, and higher levels of NET formation correlated with disease severity [10,11]. Thus, targeting excessive NET formation is speculated to reduce pulmonary inflammation and thrombosis in COVID-19 patients [12]. Because platelets are hyperactivated in critically ill COVID-19 patients [13,14], and platelets-neutrophil interactions play critical roles in endothelial damage and immunothrombosis of COVID-19 patients [15][16][17], we are interested to understand the molecular mechanism of platelets-mediated enhancement of NET formation and pulmonary inflammation in SARS-CoV-2 infection. Recently, SARS-CoV-2 was shown to activate platelets to enhance thrombosis in vivo [18], and serum levels of platelet-derived EVs (PLT-EVs) strongly associated with severity of SARS-CoV-2 infection [19]. Moreover, sera from COVID-19 patients trigger NET formation in neutrophils isolated from healthy donors [20]. These observations suggest that SARS-CoV-2 may activate platelets to release EVs, thereby induce NET formation and immunothrombosis in COVID-19 patients.
Here we report that COVID-19 EVs express abundant markers of activated platelets, and induce NET formation via CLEC5A and TLR2. Simultaneous blockade of CLEC5A and TLR2 inhibited SARS-CoV-2-induced NET formation in vitro, and thromboinflammation and fibrosis were dramatically attenuated in clec5a −/− /tlr2 −/− mice. These observations suggest that EVs from virus-activated platelets play as endogenous danger signals to trigger NETs and inflammatory reactions via CLEC5A and TLR2, and blockade of CLEC5A and TLR2 may become a promising strategy to attenuate SARS-CoV-2-induced thrombus and coagulopathy in COVID-19 patients in the future.

Isolation of human primary neutrophils and platelets
Blood was drawn from healthy donors (unvaccinated with SARS-CoV-2 vaccine) into an anticoagulant ACDcontaining syringe (ACD: blood ratio = 1:6, v/v), platelet-rich plasma was collected by centrifuge at 230 × g for 15 min. Pellet of platelets was harvested by centrifugation at 1000 × g for 10 min, then suspended in Tyrode's buffer. For human neutrophils, whole blood was laid on the Ficoll-Paque (GE Healthcare, 45-001-748) and centrifuged at 500 × g for 15 min to get red blood cells (RBCs)-granulocytes-rich layer. After RBC lysis, neutrophils were washed and suspended in RPMI containing 10% serum from unvaccinated blood type A/B healthy donors. The protocol was approved by the Human Subject Research Ethics, Academia Sinica (AS-IRB-BM-20025).

Isolation of mouse primary neutrophils and platelets
Platelets-rich plasma from murine peripheral blood was collected into ACD-containing Eppendorf and centrifugated at 230 × g for 4 min. Platelets were washed once and harvested by centrifugation at 1000 × g for 4 min, pellet was suspended in Tyrode's buffer. For neutrophil isolation, bone marrow was collected and incubated with RBC lysis buffer. Bone marrow cells were suspended in a 45% Percoll solution, then laid on 52%, 63%, and 81% Percoll, and were further centrifugated at 1000 × g for 30 min. Neutrophils were harvested from layer 3 of the Percoll gradient, followed by washing with HBSS twice before being suspended in RPMI containing 10% FBS.

SARS-CoV-2 propagation
SARS-CoV-2 Taiwan/4/2020 was propagated in Vero E6 cells. Viral titer was determined by observation of the cytopathic effect (CPE) in Vero cells. This strain was used in all of the experiments.

Collection of tissues for RNA isolation
Mice were sacrificed at 3 days and 5 days post-infection. For RNA isolation, lung was dug into TRizol-containing MagNA Lyser Green Beads (Roche) for tissue homogenization and further isolated RNA using TriRNA Pure Kit (Geneaid) according to the manufacturer's instruction. cDNA was synthesized using the RevertAid First Strand cDNA Synthesis Kit and the real-time PCR was performed as followed conditions: 95 °C for 5 min, followed by 30 cycles of 15 s at 95 °C, 30 s at 58 °C, and 30 s at 72 °C. The primer sequences were listed in Additional file 2: Table S1. Data were shown as fold change compared to mock after normalized to GAPDH.

Immunohistochemistry (IHC)
Lung tissue was fixed in 10% paraformaldehyde for 48 h and embedded in paraffin subsequently. Tissue sections were deparaffined and rehydrated before H&E stain and multiple-color fluorescent staining using Opal 7-Color IHC Kits (Akoya bioscience). Samples were incubated with primary antibody (1:50) at 4 °C overnight, followed by incubation with secondary antibody (1:100) at room temperature for 1 h. The Opal fluorescent dye was applied according to the vendor's instructions. Images were captured with a Leica confocal microscope with white light laser system (TCS SP8X-FALCON) and exported using the Leica Application Suite X software. The NET/thrombosis quantitation and cell population analysis by MetaMorph ™ image software.

Collagen deposition
Lung sections were de-paraffined and re-hydrated before being stained with Picro Sirius Red Stain Kit (#ab150681, Abcam), and images were captured by a light microscope with polarized light (Nikon). Quantification of collagen was performed by MetaMorph ™ , and the level of collagen deposition was presented as area (μm 2 ) of collagen under 20× and 40× magnification, repectively. For Masson's trichrome stain, tissue sections were stained with Trichrome Stain (Masson) Kit (Sigma, # HT15-1KT), the level of pulmonary fibrosis was evaluated by the modified Ascroft scale as previously described [33].

Isolation of extracellular vesicles (EVs)
Plasma from healthy donors and COVID-19 patients with severe pneumonia were centrifugated at 3500 × g for 15 min to remove cells and debris. Supernatants were further centrifuged at 100,000 × g for 1.5 h at 4 °C. Pellets were washed with saline and centrifuged at 100,000 × g for 1.5 h at 4 °C, followed by resuspension in 1 ml of saline. The protein concentration of EVs was determined by DC protein assay (Bio-Rad) according to the manufacturer's instruction.
FACSVerse ™ and presented as mean fluoresce intensity (MFI). Statistical analysis was calculated with an unpaired and nonparametric Student's t-test with Mann-Whitney test. ns: no significant difference, *p < 0.05, **p < 0.01.

Visualization and quantification of NET structure
Cells were immersed in fixation buffer (containing 4% paraformaldehyde) overnight, followed by permeabilization using 0.5% Triton X100 in PBS, then incubated with anti-MPO antibody (1:100), anti-citrullinated histone antibody (1:100), and Hoechst 33342 (1:100,000). The level of NETs was calculated using the histone image captured by a Leica confocal microscope with white light laser system (TCS SP8 X-FALCON), and analyzed by MetaMorph ™ software.

Mass spectrometry analysis and ingenuity pathways analysis (IPA)
Mass spectrometry analysis was performed in the Mass Spectrometry Core Facility located in the Genomic Research Center, Academia Sinica. In brief, EVs samples were lysed by RIPA solution containing phosphatase and protease inhibitors. Before the mass spectrometry analysis, samples were washed in PBS, followed by trypsin digestion before subjected to LTQ Orbitrap XL mass spectrometer (Thermo Fisher Scientific Inc.). Data were further analyzed by the Ingenuity Pathways Analysis (IPA) software. The significance of p-value was calculated by the right-tailed Fisher's Exact Test and shown as -log (p-values).

Statistical analysis
All the numbers of samples or mice were described in the figure legend, and the statistical significance was calculated using GraphPad Prism (version 9.0) software (GraphPad Software Inc., San Diego, CA, USA). Data were presented as mean ± SEM and the statistical significance was measured using an unpaired and nonparametric Student's t-test with Mann-Whitney test. In all of the experiments, *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

SARS-CoV-2 induces robust NET formation in the presence of platelets
To understand the role of platelet in SARS-CoV-2-induced NET formation, SARS-CoV-2 were incubated with human neutrophils in the presence or absence of autologous platelets. At 5 h post-incubation with SARS-CoV-2 (MOI = 1) (2nd panel from left, Fig. 1a), SARS-CoV-2 alone induced colocalization of citrullinated histone (Cit-H3), chromosomal DNA, and myeloperoxidase (MPO) within neutrophils. In contrast, robust aggregated NETs were observed in the presence of platelets (4th panel from the left, Fig. 1a), which is distinct from the thread-like NET structure induced by DV/platelets [26]. The detail of NET structure was presented as separated panels in Additional file 1: Fig. S1 & S2, and the level of NETs was measured by Cit-H3 area (μm 2 ) (Fig. 1b). These observations demonstrated the critical role of platelets in SARS-CoV-2-induced NET formation.
We further asked whether SARS-CoV-2 spike protein contributed to platelet activation by incubating SARS-CoV-2 spike pseudotyped virus with neutrophils for NETosis assay. We found that platelets still enhanced SARS-CoV-2 spike pseudotyped virus-induced NET formation, though the enhancing effect was less obvious than SARS-CoV-2. Blockade of TLR2 (green bar) and CLEC5A (blue bar) partially inhibited NET formation, simultaneous blockade of CLEC5A and TLR2 abolished SARS-CoV-2 spike pseudotyped virus-induced NET formation dramatically (Fig. 1e). This observation suggests that SARS-CoV-2 spike protein co-activates CLEC5A and TLR2 to induce NET formation.

COVID-19 EVs induced NET formation via CLEC5A
As COVID-19 sera was reported to induce NET formation [20], and serum levels of platelet-derived EVs

Table 1 Proteins expressed in COVID19-EVs
The mass spectrum analysis was performed using the LTQ Orbitrap XL mass spectrometer (Thermo Fisher Scientific Inc.) and the proteomics data were further analyzed by Ingenuity Pathways Analysis (PLT-EVs) correlated with disease severity [19], we asked whether COVID-19 EVs induced NET formation was via CLEC5A and TLR2. Firstly, we compared the protein components of EVs isolated from serum samples of COVID-19 patients (COVID-19 EVs) and normal individuals (HC-EVs) by mass spectrometry, and data were analyzed by 'Ingenuity Pathway Analysis' (IPA, QIAGEN) software. We found that molecules involved in platelet degranulation (CD9, platelet factor 4 (PF4)), aggregation, and activation [CD9, CLEC1B (also known as CLEC2)] were upregulated dramatically in COVID-19 EVs, while none of these proteins were detectable in EVs from healthy donors (Fig. 2a). In addition to these upregulated molecules, proteins specifically expressed in COVID-19 EVs were listed in Table 1. We further confirmed the IPA results by flow cytometry analysis. We found that platelet activation markers (CD41a/b, CD62p), integrins (CD9, CD29, CD49e), adhesion molecules (CD31), and other activation marker (CD45, CD69) were upregulated (Fig. 2b, c). The change of surface marker expression with statistical significance is shown in Table 2. We then incubated neutrophils with HC-EVs and COVID-19 EVs, respectively, to compare their abilities to induce NET formation. While EVs from healthy control were unable to induce NET formation, COVID-19 EVs induced robust NET formation, which was blocked efficiently by anti-CLEC5A mAb, but not anti-TLR2 mAb (Fig. 2d). This observation suggests that COVID-19 EVs have potent activity to induce NET formation via CLEC5A.

Attenuation of lung collagen deposition in mice deficient of CLEC5A and TLR2
In addition to thromboinflammation, SARS-CoV-2 infection resulted in lung injury and pulmonary fibrosis, including thickening of basement membranes and deposition of collagen [34]. Thus, we were interested to understand whether CLEC5A and TLR2 contributed to SARS-CoV-2-induced collagen deposition. At day 5 post-infection, severe thickening of alveolar cell wall and cell infiltration were noted in WT mice (upper middle, Fig. 4a), while these phenomena were attenuated in clec5a −/− tlr2 −/− mice after SARS-CoV-2 infection (upper right, Fig. 4a). We further examined the extent of collagen deposition by Masson's trichrome (lower panels, Fig. 4a) and Picro Sirius Red Staining (Fig. 4b). We observed yellow-orange birefringence (type I collagen thick fiber) and green birefringence (type III collagen, thin fiber) in lung tissues of WT mice after SARS-CoV-2 infection (middle panels, Fig. 4b).
In contrast, collagen deposition was attenuated in clec5a −/− tlr2 −/− mice (right panels, Fig. 4b). The quantification of collagen deposition and pulmonary fibrosis (modified Ascroft scale) [33] was shown in Fig. 4c and d, respectively. These observations suggested that CLEC5A and TLR2 play critical roles in SARS-CoV-2-induced lung fibrosis. The quantitation of collagen area in lung was shown in Fig. 4b. Thus, we concluded that SARS-CoV-2 activated CLEC5A and TLR2 to induce lung inflammatory reactions, thrombosis, and collagen deposition, and PLT-EVs further enhanced NET formation via CLEC5A. These observations further suggested that EVs from virus-activated platelets are potent endogenous danger signals to enhance inflammatory reactions in vivo.

Discussion
It has been shown that lung is responsible for 50% of platelet biogenesis or 10 million platelets per hour [35]. Moreover, platelets were hyperactivated in clinically ill COVID-19 patients [13,14], and contributed to coagulopathy in COVID-19 patients [18]. Thus, we are interested to understand the mechanism of platelet-mediated immunothrombosis in SARS-CoV-2 infection. In this study, we demonstrated that platelets enhanced SARS-CoV-2-induced NET formation (Additional file 1: This phenomenon was not observed in DV/plateletsinduced NET formation, suggesting the aggregated floating NETs may contribute to microemboli in COVID-19 patients. Furthermore, SARS-CoV-2 and COVID-19 EVs induced NET formation via CLEC5A and TLR2, and clec5a −/− tlr2 −/− mice were resistant to SARS-CoV-2-induced lung inflammation and collagen deposition. These observations suggest that platelets contribute to SARS-CoV-2-induced inflammation significantly, and blockade of CLEC5A and TLR2 may be beneficial to alleviate thromboinflammation and reduce intravascular coagulopathy in COVID-19 patients. It has been shown that serum EV level correlates with disease severity in COVID-19 patients [36,37]. Moreover, Syk inhibitor R406 was shown to prevent NETosis of healthy donor neutrophils stimulated with COVID-19 patient plasma [38]. In this study, we further demonstrated that COVID-19 EVs express abundant markers of Fig. 3 Attenuation of SARS-CoV-2-induced thromboinflammation in CLEC5A and TLR2 deficient mice. C57BL/6 mice (WT) and clec5a −/− tlr2 −/− mice were inoculated with AAV-hACE2 for 14 days, followed by intranasal inoculation of SARS-CoV-2 (8 × 10 4 PFU/per mice). Tissues were collected at 3 days and 5 days post-infection. N = 3 of each group a The level of proinflammatory cytokines and chemokines were measured by real-time PCR and presented as fold change (compared to AAV-hACE2 uninfected mice/mock). b-d NET structure and thrombus were detected by Hoechst. 33342 (blue), anti-MPO antibody (green), anti-citrullinated histone H3 (red), anti-CD42b antibody (yellow) (b), and images were captured by a confocal microscope and subjected to determine the area of MPO (c) and CD42b (d) using MetaMorph ™ software. e Cell infiltrated to lung at 3 d.p.i.. Interstitial macrophage (interstitial MФ) was defined as CD11b + CD64 + F4/80 + cells; monocyte-derived dendritic cell (DC)/macrophage (MФ) was defined as CD11b + CD64 + Ly6C + ; Ly6C + monocyte was defined as Ly6C + . The cell number of each cell population was calculated using the multiple fluorescent staining image and analyzed by software MetaMorph ™ , and the data was presented as cell number/ per 664225 (815 × 815). Scale bar is 200 μm. Data are represented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, **** p < 0.0001 (Student's t-test) activated platelets, enhance NET formation via CLEC5A and TLR2 (Fig. 3d). These observations are in accord with our previous report that DV activates platelets to release EVs, which are critical endogenous danger signals to trigger NET formation and inflammatory reactions via Sykcoupled CLEC5A and TLR2 [26,27].
Recently, SARS-CoV-2 was shown to bind ACE2 in platelets to enhance NETosis [18]. However, mass spectrometry [39] and RNA-seq analyses [40,41] did not detect ACE2 and TMPRSS2 in platelets and megakaryocytes. Increasing evidence indicates that lectins play critical roles in virus-induced systemic inflammation and NET formation [42]. C-type lectins (DC-SIGN and L-SIGN) and sialic acid-binding immunoglobulin-like lectin 1 (SIGLEC1) were shown to function as attachments receptors by enhancing ACE-2 mediated infection [43]. Furthermore, SARS-CoV-2 was reported to exacerbate inflammatory responses in myeloid cells through C-type lectin receptors and Tweety family members [44]. We have shown that DV was captured by DC-SIGN to activate the CLEC2 to release EVs from platelets [26]. Moreover, DV has been shown to be captured by C-type lectins heterocomplex (DC-SIGN and mannose receptor) to trigger inflammation and NETosis via CLEC5A and TLR2 [26,27,45]. Thus, it would be very interesting to test whether CLEC2/DC-SIGN complex is responsible for SARS-CoV-2-induced platelet activation in the future.
It has been reported that clinical symptoms, laboratory features and autopsy findings between dengue fever and COVID-19 are similar and are difficult to distinguish, and some patients who were initially diagnosed with dengue, but were later confirmed to have COVID-19 [46]. Moreover, autopsy of COVID-19 patients demonstrated that cells involved in the pathogenesis of COVID-19 and dengue are similar [47]. In this study, we found that platelets and platelet-derived EVs play critical roles in the pathogenesis of SARS-CoV-2. Unlike the thread shape of NET formation caused by DV, SARS-CoV-2-induced robust aggregated NET formation in the presence of platelets. Because the aggregated NETs may detach from SARS-CoV-2-induced thrombi and form micro-emboli in vivo, this observation can explain why intracoagulopathy was only found in COVID-19 patients, but not in DVinfected patients. Thus, inhibition of platelet activation may become a novel strategy to attenuate virus-induced lung inflammatory reactions in the future.

Conclusions
This work clearly demonstrates that PLT-EVs from COVID-19 patients can enhance thromboinflammation via CLEC5A and TLR2, and further shows that PLT-EVs from virus-activated act as endogenous danger signals to cause systemic inflammation. As serum levels of PLT-EVs strongly associated with severity of SARS-CoV-2 infection [19] and thromboemboli contribute to the morbidity of post-acute sequelae of COVID (PASC) [9], CLEC5A and TLR2 are promising therapeutic targets to attenuate thromboinflammation and reduce the risk of post-acute COVID-19 syndrome in the future.