Extracellular vesicles from CLEC2-activated platelets enhance dengue virus-induced lethality via CLEC5A/TLR2

Platelet-leukocyte interactions amplify inflammatory reactions, but the underlying mechanism is still unclear. CLEC5A and CLEC2 are spleen tyrosine kinase (Syk)-coupled C-type lectin receptors, abundantly expressed by leukocytes and platelets, respectively. Whereas CLEC5A is a pattern recognition receptor (PRR) to flaviviruses and bacteria, CLEC2 is the receptor for platelet-activating snake venom aggretin. Here we show that dengue virus (DV) activates platelets via CLEC2 to release extracellular vesicles (EVs), including exosomes (EXOs) and microvesicles (MVs). DV-induced EXOs (DV-EXOs) and MVs (DV-MVs) further activate CLEC5A and TLR2 on neutrophils and macrophages, thereby induce neutrophil extracellular trap (NET) formation and proinflammatory cytokine release. Compared to stat1−/− mice, simultaneous blockade of CLEC5A and TLR2 effectively attenuates DV-induced inflammatory response and increases survival rate from 30 to 90%. The identification of critical roles of CLEC2 and CLEC5A/TLR2 in platelet-leukocyte interactions will support the development of novel strategies to treat acute viral infection in the future.

The manuscript entitled entitled "CLEC2-CLEC5A/TLR2 axis is critical in dengue virus-induced inflammation and lethality" by Sung et al. proposed the novel axis of CLEC2-CLEC5A/TLR2. The authors proposed a novel CLEC2-CLEC5A/TLR2 axis as a critical signaling pathway for dengue virus-induced inflammation. The authors also mentioned that the simultaneous blockade of CLEC5A and TLR2 may reduce dengue virus-induced inflammation and lethality. Although the study is potentially important from the clinical point of view, this reviewer feels that the manuscript would be much benefitted if the authors provide additional clarifications to following concerns.
Comments 1) There is no abstract. 2) No direct evidence is provided for the description "CLEC2-CLEC5A/TLR2 axis" throughout the manuscript.
3) It would be better to improve the quality of Figure 4A. 4) Figure 4D is missing in the current manuscript. 5) The authors may want to use Stat1-/-TLR2-/-mice to clarify the requirement of TLR2 in the infection experiment. 6) The each value in Figure 7B should be presented as actual concentration rather than fold change.
Reviewer #2 (Remarks to the Author): In this paper the authors are interested in understanding how Dengue virus (DV) activates platelets and leukocytes and whether DV-induced leukocyte-platelet interaction contributes to disease severity. In this paper the authors show that DV activates a molecule called CLEC5A in neutrophils to induce NETosis which is further enhanced by the presence of platelets. It was already shown that CLEC5A can interact directly with the dengue virion and induce proinflammatory cytokine expression (Chen et al., Nature 2008). Moreover anti-CLEC5A monoclonal antibodies inhibit DV-induced plasma leakage, and vital-organ haemorrhaging reducing the mortality of DV infection by about 50% in STAT1-deficient mice. The authors suggest that DV activates CLEC2 in platelets to release extracellular vesicles (EVs), which further enhance NET formation and pro-inflammatory cytokine production via activating CLEC5A and TLR2 in both macrophages and neutrophils. Blocking CLEC2-CLEC5A/TLR2 axis may be a novel strategy to reduce tissue damage of patients suffering from severe DV infection. This is a beautiful story with clearly presented figures and the content is potentially highly relevant for our understanding of DV pathogenesis. While the data is compelling and demonstrates a roles for a CLEC-2 platelet and CLEC5-neutrophil axis in NET formation, I have several questions about the mechanistic evidence that may need addressing.
Major concerns -The authors claim that DV-activated platelets require CLEC2 to generate EVs that in turn induce NETosis via CLEC5. I am not convinced yet that alternative activation of platelets will not lead to similar NET formation. Could the authors show that CLEC2-independent activation of platelets (PMA/von Willebrand factor/PAR-1 agonists?) have simialr effects? -The authors show that DV itself also activates NET formation (Fig 2A) which is increased by adding CLEC2-activated PLTs. It is unclear though how the authors separate DVs form the DVactivated PLTs and MVs. One may assume that the MVs and EXOs preparations will be contaminated with intact DVs. Can the authors rule out this possibility? How are the amount of material, PLTS, MVs and EXOs compared? -While CLEC2 mAb and CLEC2-/-platelets lose their ability to cause NET formation, it remains unclear what the critical molecule(s) on the platelets and or EVs is/are that support (CLEC5dependent) NET formation. Can the authors identify critical molecules on the PLTs/MVs or EXOs that activate the NET formation? Im asking because one may hypotesize that CLEC2 activation of PLTs may indiuce selected molecules for secretion via EVs.

Minor
What is the concentration of isotype control in Fig 2A?' The western blots of the MVs and EXs are not extremely convincing, in particular CD63 bands are suboptimal what is the predicted size maybe use non-reducing conditions. In addition, other EV protein markers such as HSP70 or CD81 and TSG101 or ALIX should be shown.
Finally it would behoove the authors to show Electron Microscopy images of the PLT, MV and EXO preparations to get an idea of their composition/purity.

Reviewer #3 (Remarks to the Author):
Sung and colleagues analyzed the involvement of CLEC-2 and CLEC-5 in DENV-induced platelet activation and platelet-mediated NET extrusion. Similar to platelets, platelet-derived extracellular vesicles were also able to increase NET extrusion. The participation of platelet CLEC-2 and neutrophil CLEC-5 and TLR-2 in NET extrusion in vitro is very convincing. The participation of CLEC-5 and TLR-2 in NET extrusion and increased vascular permeability is also very convincing in vivo. However, there is no evidence for the participation of platelets or platelet-CLEC-2 in the release of NET, increased vascular permeability or mortality in vivo. Even though the data presented here are of interest, some concerns must be addressed.
Major issues: 1-The authors mention platelet-specific CLEC-2 KO mice (CLEC2flox/PF4cre) in material and methods (line 337). Were platelet-specific CLEC-2 KO mice protected from DV infection? Were they susceptible to platelet activation, NET extrusion in vivo and increased vascular permeability? In addition, platelet depletion experiments would be useful to clarify the importance of platelets to NET extrusion and vascular damage in vivo.

2-Identification of extracellular vesicles (EVs) as microparticles (MPs) or exosomes is a crucial
aspect of the author's conclusions. Importantly, the authors do not present any data on characterization of platelet-derived MPs and exosomes (size, number, surface markers, ultrastructure…). Figure 3, the authors show that CD62P and CD63 expression is increased in MPs from DVinfected platelets. Did DV infection change only the expression of surface molecules on MPs or also change the number of EVs released from infected platelets? 4-Platelets were recently shown to productively replicate the DV (Simon et al, Blood 2015). Since exosomes are in the same size range than DV, how can the authors differentiate exosomes from DV particles released from infected platelets? As shown in Figure 2 and 4, exosomes from infected platelets activated neutrophils depending mainly on CLEC-5 expression, which was very similar to neutrophil activation by DV alone. The authors should consider the possibility that DV contamination in isolated exosomes are driving NET extrusion in this experiment. 5-Is the ability to induce NET a feature exclusive of CLEC-2-activated platelets? Are MPs from platelets activated by other agonists also able to induce NET? 6-The authors investigated the ability of EVs from infected platelets to increase endothelial cell permeability ( Figure 5) or cytokine release (Figure 7) in vitro. However, they never describe whether they are using MPs or exosomes in this assay. Please clarify this issue.

3-In
7-Was vascular permeability in vivo ( Figure 5D) and animal survival ( Figure 7D) significantly reduced in stat1/clec5 double KO compared to stat1 KO mice? Was this difference still present after anti-TLR2 treatment?
Minor issues: 1-The discussion can be improved with information on other receptors for DV on platelets. In addition to CLEC-2, another C type lectin receptor, namely DC-SIGN, have been implicated in DV and HIV binding and internalization by platelets; and in DV-induced platelet activation (Simon et al, Blood 2015, Hottz et al, JTH 2013Chaipan et al, J Vir 2006;Boukour et al, JTH 2006;reviewed in Hottz et al, Front Med 2018).
2-Increased levels of cell-free histones have been shown to correlate with disease severity in dengue. In addition, circulating cell-free histones contribute platelet activation in dengue patients (Trugilho et al, Plos Path 2017). The authors should discuss how platelet-induced NET extrusion makes a perfect explanation for increased levels of cell-free histones in dengue, and how histones can reciprocally increase platelet activation. This is an interesting study showing that TLR2 and CLEC5a are PRRs for Dengue virus. The work shows that DV induces platelets to make exosomes and vesicles which induce neutrophils to make NETs. Both human and mouse systems are used. This is a well accomplished study. I have really only one issue with this entire study. The role of DNAse-1.
1) It is not clear how adding DNAse-1 which chops up DNA would reduce the amount of citrullinated-H3. This is surprising as DNAse should only get rid of DNA.
2) Secondly the DNAse reduced permeability and various other biology. How does this happen. Presumabley it is not naked DNA that is the biological effector. Presumably proteases and various other molecules are released and so why is DNAse so effective.
3) I would like to see how effective DNAse-1 is in the mouse model for mortality and various other parameters.