key: cord-0788396-st8mlz21
authors: Barrett, Tessa J.; Cornwell, MacIntosh; Myndzar, Khrystyna; Rolling, Christina C.; Xia, Yuhe; Drenkova, Kamelia; Biebuyck, Antoine; Fields, Alexander T.; Tawil, Michael; Luttrell-Williams, Elliot; Yuriditsky, Eugene; Smith, Grace; Cotzia, Paolo; Neal, Matthew D.; Kornblith, Lucy Z.; Pittaluga, Stefania; Rapkiewicz, Amy V.; Burgess, Hannah M.; Mohr, Ian; Stapleford, Kenneth A.; Voora, Deepak; Ruggles, Kelly; Hochman, Judith; Berger, Jeffrey S.
title: Platelets amplify endotheliopathy in COVID-19
date: 2021-09-08
journal: Sci Adv
DOI: 10.1126/sciadv.abh2434
sha: cb295a951e7d5311021d5aec7cd8c9742b1323a9
doc_id: 788396
cord_uid: st8mlz21

Given the evidence for a hyperactive platelet phenotype in COVID-19, we investigated effector cell properties of COVID-19 platelets on endothelial cells (ECs). Integration of EC and platelet RNA sequencing revealed that platelet-released factors in COVID-19 promote an inflammatory hypercoagulable endotheliopathy. We identified S100A8 and S100A9 as transcripts enriched in COVID-19 platelets and were induced by megakaryocyte infection with SARS-CoV-2. Consistent with increased gene expression, the heterodimer protein product of S100A8/A9, myeloid-related protein (MRP) 8/14, was released to a greater extent by platelets from COVID-19 patients relative to controls. We demonstrate that platelet-derived MRP8/14 activates ECs, promotes an inflammatory hypercoagulable phenotype, and is a significant contributor to poor clinical outcomes in COVID-19 patients. Last, we present evidence that targeting platelet P2Y(12) represents a promising candidate to reduce proinflammatory platelet-endothelial interactions. Together, these findings demonstrate a previously unappreciated role for platelets and their activation-induced endotheliopathy in COVID-19.

Coronavirus disease 2019 (COVID- 19) is an acute viral illness caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Intense systemic inflammation and hypercoagulability are hallmarks of severe COVID-19 (1, 2) . High rates of arterial and venous thrombosis are frequently seen in SARS-CoV-2 infection (3) (4) (5) . The etiology of COVID-19-associated coagulopathy remains controversial and is likely to be complex, with contributions from many cellular and noncellular mediators.

The vascular endothelium provides a crucial interface between the blood compartment and tissues. The endothelial cell (EC) monolayer lining blood vessels prevents thrombosis and maintains hemostasis; however, following activation, ECs support the initiation of coagulation and thrombosis, hallmarks of severe COVID-19. Hospitalized COVID-19 patients have increased circulating markers of EC activation, including von Willebrand factor (vWF), plasminogen activator inhibitor 1 (PAI1), intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule-1 (VCAM-1), and P-selectin (6) (7) (8) (9) (10) . Autopsies of COVID-19 patients reveal significant EC apoptosis and loss of EC tight junction integrity in the pulmonary microvasculature (11) . Despite the suggestion that COVID-19associated endotheliopathy is mediated by SARS-CoV-2 EC infection (12, 13) , recent evidence suggests that indirect (nonviral) mechanisms drive EC activation in COVID-19 (14) .

Platelets are anucleate blood cells classically known by their roles in hemostasis and thrombosis. Platelets also maintain vascular function and integrity by bidirectional interactions with ECs and leukocytes (15) . Under inflammatory conditions, the cross-talk between platelets, coagulation, and the endothelium exacerbates local and systemic inflammation (16) (17) (18) (19) (20) . Previously, we found that SARS-CoV-2-platelet interactions drive a hyperactive platelet phenotype, and in the context of viral infections, platelets initiate EC proinflammatory activation (21) . Given that activated platelets release inflammatory factors in the local microenvironment, we hypothesized that platelet-EC interactions in COVID-19 promote endotheliopathy.

Herein we demonstrate that platelets from COVID-19 patients are hyperactive and that platelet-released factors activate microvascular ECs, promoting an inflammatory hypercoagulable phenotype. Investigation of the platelet transcriptome revealed significant differential expression of thousands of genes in COVID-19 and enrichment of pathways related to organelle/granule release, metabolism, and immune effector function. An investigation into platelet candidate genes that code for abundantly released proteins found up-regulation of S100A8 and S100A9 in platelets from COVID-19 patients. In vitro, we confirmed that SARS-CoV-2 could alter megakaryocyte gene expression and up-regulate S100A8 and S100A9. Consistent with an increase in transcription, the heterodimer protein product of S100A8/A9, myeloid-related protein 8 (MRP8)/MRP14, is released to a greater extent by platelets from COVID-19 patients relative to controls. MRP8/14 activates microvascular ECs and weakens EC cell-cell contacts resulting in increased permeability of the endothelium (22) , thus representing an important pathogenic mediator of inflammatory platelet-EC interactions in COVID- 19. Exposure of ECs to platelet releasate from COVID-19 patients induced significant transcriptomic and phenotypic changes relative to exposure with platelet releasate from controls. Consistently, we found a direct association between platelet-released MRP8/14 and EC activation. EC enrichment in pathways characteristic of cell-cell tight junctions, coagulation, and proinflammatory processes is correlated with the dysregulated COVID-19 platelet transcriptome and the expression of the RNA and protein product of S100A8/9. These transcriptomic changes were reflected in the increased production of classical inflammatory cytokines interleukin 6 (IL6) and IL8 and the correlation of the platelet and endothelial transcriptome with platelet activation. Furthermore, in a cohort of 291 hospitalized COVID-19 patients, we found that circulating MRP8/14 levels were associated with subsequent thrombosis and death following multivariable adjustment. Therapeutic targeting of platelet P2Y 12 decreases S100A8/A9 platelet mRNA and reduced proinflammatory plateletendothelial interactions.

Our data collectively indicate that platelets amplify EC dysfunction and mediate thrombo inflammation in patients with COVID-19. Strategies to mitigate platelet activation may reduce COVID-19associated endotheliopathy and improve clinical outcomes linked to thromboinflammation.

Platelet hyperactivity is increasingly recognized as a hallmark of COVID-19 (10, (23) (24) (25) . Consistently, we found increased platelet reactivity in COVID-19 patients as characterized by increased platelet surface expression of P-selectin and CD40 (Fig. 1, A and B, and table S1; n = 9 COVID-19 patients and 7 controls) and increased circulating platelet-heteroaggregates (platelet-leukocyte, plateletneutrophil, and platelet-monocyte aggregates; Fig. 1C ). In addition, COVID-19 patients had increased spontaneous platelet aggregation and aggregation in response to submaximal ADP and epinephrine (Fig. 1D ).

Patients infected with SARS-CoV-2 are noted to have an endotheliopathy (11, 12) . Under inflammatory conditions, plateletendothelium cross-talk exacerbates inflammation and aberrant immunity and impairs the ability of ECs to maintain vascular homeostasis. The interconnection between platelets and the vasculature in COVID-19 is supported by autopsy data showing plateletrich thrombi in macro-and microvasculature (26) , characterized by platelet activation and adhesion to the endothelium (Fig. 1 , E and F), increased vascular dysfunction, and thrombosis (6, 7, 10, 23, 26, 27) . Recent data demonstrate that direct endothelial infection of SARS-CoV-2 is unlikely to occur (14) , suggesting an indirect mechanism. Thus, we investigated platelet-EC interactions in the setting of COVID-19.

Human microvascular ECs were incubated with platelet releasate from COVID-19 patients or controls, and EC RNA sequencing (RNA-seq) was performed ( Fig. 1G and table S1; n = 7 COVID-19 patients and 7 controls). Treatment of microvascular ECs with COVID-19 releasate significantly altered the EC transcriptome, with 1288 transcripts differentially expressed (adjP < 0.01, 485 transcripts down-regulated, 803 transcripts up-regulated; Fig. 1H ). Top differentially expressed transcripts separate ECs treated with COVID-19 platelet releasate from controls (Fig. 1I ). Gene set enrichment analysis (GSEA) of differentially expressed genes revealed altered pathways associated with cell-cell tight junctions, coagulation, and proinflammatory processes (Fig. 1J) , hallmarks of thromboinflammation. To further evaluate gene modules dysregulated in COVID-19, we used weighted correlation network analysis to obtain eigengene representing the relative coexpression of gene sets on a per sample basis (28) . When comparing the eigengene values of these gene sets between COVID-19 and controls, the apical junction, coagulation, and tumor necrosis factor  (TNF) signaling pathways show significantly higher expression in ECs treated with COVID-19 platelets relative to controls (Fig. 1K) . EC genes involved in inflammation, immunity, and hypercoagulability (e.g., VWF, MMP11, MMP2, TIMP1, ICAM1, and ICAM2) were differentially expressed in platelet COVID-19 releasate-treated ECs (Fig. 1L ).

Consistent with others, our group identified a prothrombotic and proinflammatory platelet transcriptome in patients with COVID-19 [GSE176480 (23) ]. We, therefore, integrated our platelet and EC RNA-seq analyses to gain mechanistic insight underlying plateletinduced EC activation. Focusing on the coagulation pathway, we clustered differentially expressed EC-induced coagulation genes (adjP < 0.05) between COVID-19 patients and controls. Unsupervised clustering of these genes resulted in a clear separation of groups ( Fig. 2A) . To identify COVID-19-regulated platelet transcripts associated with EC coagulation, we investigated platelet transcripts identified as being differentially expressed in COVID-19 patients relative to controls (fig. S1 and table S2; n = 8 COVID-19 patients and 10 controls). Correlating our differentially expressed EC coagulation genes ( Fig. 2A) with differentially expressed platelet genes in COVID-19 patients (Fig. 2B) , we identified three distinct platelet gene clusters (C1 to C3; table S3). Hypergeometric GSEA tests were then performed on each of our three gene clusters to explore their functional makeup. C1, composed of 920 differentially expressed platelet genes, was enriched in pathways associated with organelle/granule release and oxidative phosphorylation, pathways characteristic of an activated platelet phenotype (Fig. 2C ) (29) . The C2 gene cluster (413 differentially expressed genes) was enriched in platelet pathways linked to granule secretion and immune effector functions, while the C3 gene cluster (715 differentially expressed genes) was primarily composed of pathways associated with cytokine and interferonsignaling responses and immune effector processes (Fig. 2 , D and E). Collectively, these data highlight how the COVID-19-activated platelet phenotype is directly associated with EC coagulation. To further explore the relationship between the transcriptomic signature of COVID-19 platelets and ECs, we repeated the above analysis across up-regulated EC pathways. Consistent in each analysis, three distinct blocks of platelet genes were identified on the basis of unsupervised clustering. Cluster one (C1) shared a 523-platelet gene signature across all EC-up-regulated enriched pathways ( Fig. 2F and table S4 ). GSEA revealed the shared platelet ECactivating signature to be enriched in gene sets linked to membrane protein targeting, oxidative phosphorylation, and viral gene expression (Fig. 2G) . These data demonstrate that platelet-activating pathways linked to granule/vesicle release, metabolism, immune effector functions, and response to viral infections induce an EC activation phenotype consistent with the described COVID-19 endotheliopathy. 

MMP11 GNB2 VWF USP11 MMP2 SIRT2 TIMP1 MAFF CRIP2 BMP1 GSN PLAU CLU CSRP1 PEF1 WDR1 CTSB DUSP14 MMP1 LAMP2 LTA4H ITGA2 CFH ARF4 Expression z score −4 −2 0 2 4 −1 −0.5 0 0.5 1 MMP11 GNB2 VWF USP11 MMP2 SIRT2 TIMP1 MAFF CRIP2 BMP1 GSN PLAU CLU CSRP1 PEF1 WDR1 CTSB DUSP14 MMP1 LAMP2 LTA4H ITGA2 CFH ARF4 Cohort Control COVID-19 COVID

Platelet S100A8/A9 is induced by SARS-CoV-2 and increased in COVID-19 patients We next sought to identify platelet-released factors that induce EC dysfunction. Gene ontology pathway analysis of differentially expressed transcripts reveals up-regulation of the platelet degranulation pathway in COVID-19 ( fig. S3A ). We therefore explored platelet candidate genes that code for abundantly secreted proteins that are differentially expressed in COVID-19 patients. Using platelet RNA-seq metrics of differential expression between COVID-19 and controls (matched for age, sex, race, and comorbidities; table S2, log 2 FC and adjP value) and overall abundance (base mean) of the gene, we evaluated the top 50 candidates (baseMean > 323, log 2 FC > 1.6, adjP < 1.03 × 10 −4 ). Cross-referencing this list of candidate genes with a secretome database (30) reduced this number to six genes-S100A8, CST7, TIMP1, S100A9, SMPD1, and CETP (Fig. 3A) . Of these six genes, S100A8 and S100A9 were the only two genes that met all of the above criteria and were also members of the 523-C1 platelet gene signature (Fig. 2F) . S100A8 and S100A9 genes code for MRP8 and MRP14 and were among the top 4 identified transcripts enriched in platelets from COVID-19 patients to have a released product (Fig. 3, A and B ). MRP8 and MRP14 heterodimerize to produce the inflammatory product MRP8/14 (calprotectin), an inflammatory protein present within platelet granules and released by activated platelets (31, 32) .

To explore whether viral-megakaryocyte interactions could account for changes to the platelet transcriptome, human CD34derived megakaryocytes were incubated with SARS-CoV-2 and with a coronavirus responsible for the common cold (CoV-OC43). Treatment led to the detection of SARS-CoV-2 mRNA within the megakaryocytes; however, no uptake of CoV-OC43 was found ( fig. S4A ). In both primary human megakaryocytes and a megakaryocyte cell line (Meg-01), treatment with SARS-CoV-2 increased S100A8 and S100A9 expression (P < 0.01; Fig. 3C and fig. S4B ). Conversely, treatment with CoV-OC43 had no effect ( Fig. 3C and  fig. S4B ), suggesting a specific SARS-CoV-2-mediated effect.

We next measured MRP8/14 release from platelets and found that COVID-19 platelets released significantly increased levels of MRP8/14 ( Fig. 3D ; P < 0.001). There were also increased circulating plasma levels of MRP8/14 in COVID-19 patients versus controls ( Fig. 3E ; P < 0.001, n = 9 COVID-19 patients and 7 control). A positive correlation between platelet-released MRP8/14 and plasma levels was found ( Fig. 3F ; R = 0.72, P = 0.0037), suggesting that plateletreleased MRP8/14 contributes to the pool of circulating MRP8/14 levels. These data indicate that direct interaction with SARS-CoV-2 increases S100A8/A9 transcription, MRP8/14 production, and release from megakaryocytes and platelets in patients with COVID-19.

We next explored the association between platelet-released MRP8/14 and EC activation. S100A8/A9 platelet mRNA expression significantly correlated with the expression of the dysregulated pathways seen in COVID-19-exposed ECs (Fig. 4A) , highlighting the central role of MRP8/14 in driving EC dysfunction. This is further supported by the positive correlation between platelet-released MRP8/14 and COVID-19 EC-enriched pathways (Fig. 4B) .

Next, we correlated various metrics of platelet activation and clinical measures to enriched EC pathways to explore further the association between platelet activation and EC dysfunction in COVID-19. In addition to platelet-derived MRP8/14, we measured factors known to be released from platelets that are increased in COVID-19 patients (IL6, P-selectin, and sCD40L) (10, 23, 24) . Expression of dysregulated EC pathways following exposure to COVID-19 platelet releasate was most significantly correlated with platelet-released MRP8/14 (Fig. 4C) . Assessment of well-described platelet activity measures-platelet-leukocyte heteroaggregates, surface P-selectin, CD40, PAC-1, CD42b, and platelet aggregation-revealed that although all of these measures correlated with the dysregulated transcriptomic signature, circulating monocyte platelet aggregates (MPAs) were most associated with the demonstrated transcriptomic alterations (Fig. 4C) . These data are consistent with previous findings that platelet-derived MRP-14 promotes MPA formation and a proinflammatory monocyte phenotype (31) . Circulating neutrophil platelet aggregates (NPAs) were weakly correlated with enriched EC dysfunction pathways, suggesting that neutrophils are not a significant contributor to platelet-mediated endotheliopathy in COVID-19.

MRP8/14 activates microvascular ECs and weakens EC cell-cell contacts, resulting in increased permeability of the endothelium (22) . In further support of an activated endothelium, ECs treated with platelet releasate from COVID-19 patients secrete more significant quantities of the inflammatory cytokines IL8 and IL6 (P < 0.05; Fig. 4D ). A positive correlation existed between platelet-released MRP8/14 and microvascular EC IL8 and IL6 production (P = 0.001 and P = 0.007, respectively; Fig. 4E ), suggesting that platelet-derived MRP8/14 is a driver of EC activation. MRP8/14 is a ligand for innate receptors, including Toll-like receptor 4 (TLR4), CD36, and receptor for advanced glycation end products (RAGE) (31, 33) . Blocking of EC TLR4 and RAGE signaling did not suppress plateletmediated EC IL6 or IL8 cytokine production ( fig. S5 ). In contrast, silencing CD36 significantly suppressed COVID-19 platelet-mediated EC IL6 and IL8 cytokine production (Fig. 4F) . These data suggest that platelet-released MRP8/14 mediate the proinflammatory effect of COVID-19 EC inflammation via CD36. The clinical relevance of these findings is further supported by COVID-19 autopsy data showing significant lung endothelial IL6 production ( fig. S6 ).

To examine the clinical relevance of our findings, we investigated the association between MRP8/14 levels in a cohort of hospitalized COVID-19 patients. Plasma samples were obtained on the day of admission from 291 patients hospitalized with COVID-19 within the New York University (NYU) Langone Health network (Fig. 5A) Fig. 5B and table S7 ). Among COVID-19 patients, MRP8/14 was significantly higher in patients who went on to develop an adverse clinical event (thrombosis, critical illness; Fig. 5, C and D, and fig. S7 ). Following adjustment for age, sex, race/ethnicity, body mass index (BMI), diabetes, chronic obstructive pulmonary disease (COPD)/asthma, history of coronary artery disease or cancer, antiplatelet and anticoagulant therapy, plasma MRP8/14 levels were independently associated with thrombosis [adjusted odds ratio (adjOR), 1.58; 95% confidence interval (CI), 1.27 to 1.99; P < 0.001] and other adverse events (Fig. 5E ). When stratified into quartiles, COVID-19 patients with the highest HBB  IFITM3  IFITM2  MYL9  ATOX1  S100A8  S100A10  CST7  IFI27  SAMD14  CA1  HCST  M AOB  TIMP1  TMSB10  HPSE  HBD  FCGR2A  TSPO  SELENON  RPLP1  KLHDC8B  MT1X  ARL2  NDUFS5  NDRG3 AL355512

Platelet S100A8 & S100A9 SARS-CoV-2 + megakaryocytes S100A9 S100A8 platelets versus control platelets based on baseMean expression, log 2 (FoldChange), and adjusted P value between COVID-19 and control platelets. Subset zooms in on the top 50 candidates, with S100A8 and S100A9 highlighted. (B) Platelet mRNA expression of S100A8 and S100A9. N = 7 samples per group; ***P < 0.001 FDR-adjusted P values output from DESeq2. (C) CD34-derived megakaryocyte mRNA expression of S100A8 and S100A9 following treatment with PBS, SARS-CoV-2, or CoV-OC43 for 24 hours. N = 3 samples per group. **P < 0.005, ***P < 0.001 by one-way analysis of variance (ANOVA . S8 ). Of the 247 patients who were discharged, the mean hospital stay was 13.1 days. Circulating MRP8/14 was higher in patients with a longer length of stay (Fig. 5, F and G) . After multivariable adjustment, patients in the highest quartile of MRP8/14 at baseline were in the hospital for an excess of 12.9 (6.8 to 19.0) days compared to quartiles 1 to 3. In a subset of 107 hospitalized COVID-19 patients, circulating MRP8/14 was measured longitudinally. In patients without a clinical event, there was no difference between admission and day 14 (Fig. 5H) , while in those who subsequently died, MRP8/14 increased ≈8-fold between baseline and day 14 (P = 0.002; Fig. 5I ). Similar trends were also found for those who went on to experience a thrombotic event and other adverse events ( fig. S9, A to C) . Together, these data indicate that circulating MRP8/14 levels are higher in COVID-19 patients and associate with thrombosis, critical illness, and length of stay among those hospitalized. P2Y 12 inhibition reduces platelet S100A8 and S100A9 and EC activation As COVID-19 patients remain at increased risk for thrombosis, critical illness, and death with and without anticoagulant therapy (3), we investigated the effect of antiplatelet therapy on platelet S100A8 and S100A9 mRNA and platelet-induced EC activation. A cohort of healthy controls received aspirin (n = 63) daily or ticagrelor (n = 49) twice daily for 4 weeks (table S7). Blood was collected and platelet RNA was extracted at baseline and following 4 weeks of therapy. While aspirin had no significant effect (Fig. 6A) , ticagrelor led to a significant reduction in both expression of platelet S100A8 and S100A9 mRNA (P < 0.0001; Fig. 6 , B and C). To assess whether antiplatelet therapy could suppress the platelet-mediated proinflammatory effects on ECs, isolated platelets were incubated with aspirin, a P2Y 12 antagonist (AZD1283), or a glycoprotein IIb/ IIIa inhibitor (eptifibatide). Incubation of platelet releasate generated in the presence of different antiplatelet drugs found that the P2Y 12 antagonist had the greatest reduction in proinflammatory EC gene expression (29% reduction in IL6, 51% reduction in IL8, and 32% reduction in CCL20; Fig. 6D ).

Severe COVID-19 is associated with a hypercoagulable state and robust inflammatory response that predisposes patients to macroand microthrombotic events (34) (35) (36) (37) (38) . The coordinated activation of the inflammatory and thrombotic response has been termed thromboinflammation. Levels of prothrombotic acute phase mediators, including fibrinogen, vWF, and D-dimer, are increased in COVID-19 patients, implicating the endothelium, platelets, and the coagulation system as likely mediators of thromboinflammation in COVID-19. The resulting endotheliopathy is thought to link inflammation, immune dysregulation, and thrombosis, and be a driver of poor clinical outcomes in COVID-19 patients (39, 40) . Despite suggestions that SARS-CoV-2 triggers endotheliopathy (12, 13) , recent experimental evidence supports indirect activating mechanisms driving EC injury in COVID-19 (14) .

Platelets are first responders to vascular injury and act to bridge the immune system and thrombosis via activation and release of hemostatic and inflammatory mediators. Activated platelets secrete a host of proinflammatory mediators in the local microenvironment, altering the endothelium's inflammatory and adhesive properties (18, 19) . Activated platelets isolated from patients with inflammatory disease, such as systemic lupus erythematosus (SLE), HIV, and psoriasis, induce a proinflammatory EC phenotype (16, 20, 21) . We and others previously reported that COVID-19 is associated with a hyperactive platelet phenotype (23, 24) , characterized by increased platelet activation markers and a proinflammatory, prothrombotic transcriptome. Given the interconnection between activated platelets and the microvasculature, we explored whether platelet-released factors contribute to COVID-19 endotheliopathy.

Our study details how COVID-19 hyperactive platelets release factors that induce a proinflammatory and hypercoagulable endothelium (Fig. 7) . The change in EC transcription directly correlated with alterations in the platelet transcriptome, connecting differences seen in these complementary cell types. Sequencing of the platelets used to incubate with ECs found a set of 523 platelet genes that are consistently associated with the proinflammatory EC transcriptomic pathway changes. This block of genes is potentially a unique COVID-19 platelet signature that merits future exploration and study.

Unbiased screening of candidate platelet genes differentially expressed in COVID-19 led to the identification of S100A8 and S100A9, and their protein product MRP8/14, as a likely candidate of platelet-mediated endotheliopathy. Platelet S100A8/A9 mRNA and platelet-derived MRP8/14 were increased in patients with COVID, and their expression and translation positively correlate with changes seen in EC coagulation and inflammatory pathways and the production of IL6 and IL8, proinflammatory cytokines that regulate EC adhesion molecule expression (41) . While circulating plasma MRP8/14 was once considered only to be leukocyte-derived (42) , platelets and megakaryocytes serve as an additional source of MRP8/14 (31, 32, 43, 44) . While circulating plasma levels of MRP8/14 alone correlated with EC coagulation and inflammatory pathways, platelet S100A8/A9 mRNA and releasate MRP8/14 levels were more tightly associated with EC dysfunction directly, suggesting the importance of platelet-derived MRP8/14 on EC dysfunction. Our present study does not rule out other cellular and noncellular contributors of COVID-19 endotheliopathy, which may act synergistically with MRP8/14. For example, serum antiphospholipid (aPL) antibodies in COVID-19 sera were recently reported to induce adhesion Aspirin S100A8 S100A9

Ticagrelor S100A8 S100A9 

Log fold change S100A8 S100A9 S100A8 S100A9 Fig. 6 . P2Y 12 inhibition reduces platelet S100A8 and S100A9 expression and platelet-mediated EC activation. A cohort of patients received (A) aspirin daily (n = 63) or (B) ticagrelor (90 mg) twice daily for 4 weeks (n = 49). Blood was collected and platelet RNA was extracted before the commencement of the study and following 4 weeks of antiplatelet therapy. Plots depict normalized expression counts of S100A8 and S100A9 at baseline and following 4 weeks of antiplatelet therapy. ****P < 0.0001 as determined by paired t test. (C) Forest plot of log fold change (LogFC) between baseline and follow-up platelet S100A8 and S100A9 expression. Data are means and 95% CI. (D) Expression of IL6, IL8, and CCL20 following exposure of ECs to platelet releasate generated in the presence of aspirin (1 mM), AZD1283 (1 M, P2Y 12 inhibitor), or eptifibatide (18 M, glycoprotein IIb/IIIa inhibitor) for 6 hours. Data expressed relative to platelet releasate-treated cells in the absence of platelet inhibitors. n = 6 unique releasate donors. Data are means ± SEM, expressed relative to the platelet releasate-treated ECs; *P < 0.05, **P < 0.01, ****P < 0.0001 by paired t test.

molecule expression on ECs (45) . Previous studies have found that aPL antibodies promote platelet activation and their release of MRP8/14 into the microenvironment, providing evidence of additional mechanisms by which interaction of platelets with the local inflammatory environment in COVID-19 contributes to EC dysfunction (32, 46) . We find that the direct interaction of SARS-CoV-2 with primary megakaryocytes induces S100A8/A9 expression, a finding that has clinical implications given our observation that megakaryocytes from COVID-19 patients contain SARS-CoV-2 virions (26) , and single-cell sequencing data reporting up-regulation of S100A9 in circulating megakaryocytes in COVID-19 (27) . Intriguingly, up-regulation of S100A8/A9 does not occur following exposure of megakaryocytes to a coronavirus responsible for the common cold (CoV-OC43), highlighting that, independent of indirect mechanisms (e.g., systemic inflammation), SARS-CoV-2 can mediate proinflammatory changes to megakaryocytes and platelets (47, 48) . Our data indicate that increased platelet S100A8/A9 mRNA translates to increased platelet MRP8/14 release, thus indicating that direct interactions of megakaryocytes with SARS-CoV-2 enhance proinflammatory platelet-EC interactions. Consistently, MRP8/14 is increased in activated platelets and platelets isolated from patients with peripheral artery disease and ST-elevation myocardial infarction. In vitro, MRP8/14 induces platelet surface P-selectin (31), a major ligand in the platelet-EC interaction. In addition, MRP-14 −/− knockout mice had reduced expression of P-selectin in response to collagen and arachidonic acid, which was independent of platelet aggregation (44) .

The prothrombotic and proinflammatory nature of platelet MRP8/14 is supported by studies finding platelet transcript expression of S100A8/A9, positively correlates with plasma MRP8/14 concentrations, and is predictive of future atherothrombotic events (31, 43) . MRP8/14 levels are increased in patients with cardiovascular diseases, metabolic disease, and several inflammatory conditions, including rheumatoid arthritis, SLE, and psoriasis (31, 43, (49) (50) (51) . Prior studies have found increased MRP8/14 in patients infected with SARS-CoV-2 (52, 53) . In the current study, we find platelet S100A8 and S100A9 mRNA to be increased in COVID-19 patients and a corresponding increase in platelet-released MRP8/14, all of which correlate with circulating MRP8/14. Among 291 patients hospitalized with COVID-19, circulating MRP8/14 increased during hospitalization. When measured at baseline, MRP8/14 correlates with incident thrombotic events, severe disease, and length of stay. In contrast to a small increase in MRP8/14 among patients with nonsevere illness, levels increased up to eightfold in COVID-19 patients who subsequently died.

Recently, the therapeutic potential of suppressing MRP8/14 signaling has been suggested as an approach to reduce COVID-19associated thromboinflammation (54) . To understand the clinical potential of our findings, we investigated the (i) effect of antiplatelet therapy on S100A8/A9 platelet mRNA and (ii) in vitro effect of antiplatelet treatment on platelet-induced EC activation. In contrast to aspirin, targeting of P2Y 12 with ticagrelor reduced S100A8 and S100A9 platelet mRNA. In vitro experiments found that P2Y 12 inhibition significantly attenuated the COVID-19 platelet-mediated EC activation. The ability of aspirin to attenuate the platelet-EC interaction was less pronounced. The benefit of reducing platelet MRP8/14 release via targeting of P2Y 12 is supported by murine studies, which found that the antithrombotic effect of the S100A9 vaccination is equivalent to that with clopidogrel (55) . Furthermore, given evidence that P2Y 12 inhibitors reduce platelet release of proinflammatory -granule contents (56) and the formation of platelet-leukocyte aggregates (57) , the beneficial effect of this approach is hypothesized to be multifaceted. These data collectively suggest that P2Y 12 inhibition will mitigate platelet activation, reduce platelet-released MRP8/14, and ultimately suppress platelet activity and EC activation and improve clinical outcomes linked to COVID-19 thromboinflammation. While observational data suggested a potential benefit of aspirin on outcomes in hospitalized COVID-19 patients (58, 59) , a large randomized trial of nearly 15,000 participants found that aspirin did not improve its primary end point of 28-day mortality (60) . The clinical effect of a P2Y 12 inhibitor strategy is being investigated in ACTIV4a (NCT04505774).

Our study describes how platelets isolated from hospitalized COVID-19 patients release significant quantities of MRP8/14 that can act in a paracrine fashion on nearby ECs promoting endotheliopathy. Several limitations exist with our current study: (i) the relatively small sample size used for sequencing; (ii) the cellular source of plasma MRP8/14 cannot definitively be attributed to platelets and may originate from other cells, including neutrophils; (iii) it is unknown whether the observed MRP8/14 induced endotheliopathy occurs in other viral settings; (iv) a clinical diagnosis of thrombosis during hospitalization may be underestimated because imaging studies were limited due to concerns of transmitting infection or competing risk of death; and (v) the effect of antiplatelet therapies on the COVID-19 platelet-induced EC activation was not explored. Future studies will aim to address these limitations to further our understanding of the interaction of platelets with SARS-CoV-2 and other viral illnesses and the endothelium. An ongoing multinational trial is underway exploring the effect of P2Y 12 inhibition on clinical outcomes of hospitalized COVID-19 patients (NCT04505774).

In conclusion, we define a novel role for platelet-induced endotheliopathy in COVID-19. The activated platelet phenotype observed in COVID-19 consistently induces a proinflammatory and dysfunctional endothelium. Platelet-derived MRP8/14 is increased in COVID-19 and induces endothelial injury. Circulating MRP8/14 has the potential to serve as a useful biomarker of thrombosis and severity of disease in patients infected with SARS-CoV-2. Plateletdirected therapy, specifically P2Y 12 inhibitors, may represent a particularly attractive therapy because of its effect on platelet S100A8/A9 and platelet-induced endotheliopathy (Fig. 7) .

Clinical cohort no. 1: A cohort of hospitalized COVID-19 patients was recruited from NYU Langone Health between 12 March and 2 May 2020. SARS-CoV-2 infection was confirmed by reverse transcription polymerase chain reaction (RT-PCR), in accordance with current standards. All COVID-19 patients and non-COVID-19 controls were recruited under study protocols approved by the NYU Langone Health Institutional Review Board. Each study participant or their legal authorized representative gave written informed consent for study enrollment in accordance with the Declaration of Helsinki. For COVID-19 patients, enrollment criteria included age greater than 18, hospital admission, positive SARS-CoV-2 testing, and informed consent. COVID-19 patients were monitored until discharge or death.

Clinical cohort no. 2: Healthy volunteers were recruited at the Duke Clinical Research Unit (DCRU; Durham, NC) (61). Written informed consent was obtained from each patient, and the studies were approved by the Duke University Health System institutional review board. Study participants (30 to 75 years) did not have a history of a bleeding disorder, gastrointestinal bleeding, intracranial bleeding, or known prior gastric ulcer without documented resolution; did not have known severe hepatic impairment; had not undergone surgery within the past 6 months; had not undergone prior gastric bypass surgery that could interfere with study drug absorption; and did not have an aspirin allergy or known intolerance to aspirin. Participants could not be taking any prescription medications other than oral contraceptives; could not exhibit regular use of tobacco or nicotine products or any known active use of illicit substances; and could not have a known pregnancy or be currently breastfeeding. Participants who were already taking aspirin underwent a 4-week washout period, with the approval of their primary care physician before enrollment in the study.

Circulating MPA, platelet P-selectin, and platelet CD40 were identified in citrate anticoagulated blood as previously described (62) . Detailed methods are provided in the Supplementary Materials. Platelet aggregation Aggregation was measured using light transmission aggregometry (Helena AggRAM, Beaumont, TX) at 37°C under stirred conditions as previously described (31) . Platelet aggregation was assessed in response to 0.1 M adenosine diphosphate (ADP), 0.1 M epinephrine, and phosphate-buffered saline and reported as aggregation at 180 s.

Citrated anticoagulated blood samples was allowed to rest at room temperature for 15 min after phlebotomy. Platelet-rich plasma (PRP) was obtained by centrifugation of blood at 200g for 10 min. PRP was added to 1:10 (v/v) ACD-A [tris-sodium citrate (25 g/liter), glucose (20 g/liter), and citric acid (14 g/liter)] and spun at 1000g for 10 min. The platelet pellet was washed in Tyrode's buffer (137 mM NaCl, 2.8 mM KCl, 1 mM MgCl 2 , 12 mM NaHCO 3 , 0.4 mM Na 2 HPO 4 , 5.5 mM glucose, and 10 mM Hepes, pH 7.4) and 1 M prostaglandin E 1 (Sigma-Aldrich), before centrifugation at 1000g for 10 min and resuspension at 500,000/l in Tyrode's buffer containing 2 mM CaCl 2 , and incubation at room temperature for 30 min. Platelet-secreted factors were then collected by isolation of the supernatant following centrifugation at 1000g for 3 min. Platelet supernatants were stored at −80°C before use. Protocol for generation of platelet releasate in the presence of platelet inhibitors is provided in the Supplementary Materials.

Platelet and endothelial RNA-seq details for the NYU and Duke cohorts and corresponding bioinformatic analyses are provided in the Supplementary Materials.

Blood samples were collected in PST tubes (BD Diagnostics), and plasma was collected by centrifugation, stored at 5°C for 5 days, aliquoted, and stored long term at −80°C. Plasma levels of MRP8/14 were measured by a LEGENDplex bead-based immunoassay (BioLegend, product no. 740891), according to the manufacturer's instructions.

EC cytokine production IL6 and IL8 concentrations in the microvascular EC supernatants were measured by a LEGENDplex bead-based immunoassay (BioLegend, product no. 740907), according to the manufacturer's instructions.

Purified peripheral blood CD34 + cells were obtained from the New York Blood Center (New York, NY). Cells were cultured in StemSpan SFEM II (STEMCELL Technologies) supplemented with stem cell factor (100 ng/ml; R&D Systems) and thrombopoietin (TPO; 50 ng/ml; R&D Systems) for 5 days. Cells were cultured with TPO (50 ng/ml) only from days 5 to 11. Megakaryocytes were incubated with SARS-CoV-2 or CoV-OC43 on day 11. SARS-CoV-2 or CoV-OC43 details are provided in the Supplementary Materials.

We used Student's t test or the Mann-Whitney U test for the analyses of continuous variables, as appropriate. Differences between categorical variables were calculated with the use of chi-square or Fisher's exact test, as appropriate. Correlation analyses were performed by the Spearman rank correlation method. Logarithmic transformation of the data was performed when appropriate. Logistic regression was used to assess the association between MRP8/14 and outcomes. Results are presented as ORs along with their 95% CIs. Two-sided P values of <0.05 were considered statistically significant. Statistical analyses were performed with R version 3.5.2.

Supplementary material for this article is available at https://science.org/doi/10.1126/ sciadv.abh2434

View/request a protocol for this paper from Bio-protocol.

Thromboinflammation and the hypercoagulability of COVID-19

Thrombosis in hospitalized patients with COVID-19 in a New York City health system

Thrombosis in hospitalized patients with viral respiratory infections versus COVID-19

Thromboelastography profiles of critically Ill patients with coronavirus disease 2019

Endotheliopathy in COVID-19-associated coagulopathy: Evidence from a single-centre, cross-sectional study

Circulating markers of angiogenesis and endotheliopathy in COVID-19

Elevated expression of serum endothelial cell adhesion molecules in COVID-19 patients

Severe COVID-19 infection associated with endothelial activation

Biomarkers of platelet activity and vascular health associate with thrombosis and mortality in patients with COVID-19

Autopsy findings and venous thromboembolism in patients with COVID-19: A prospective cohort study

Endothelial cell infection and endotheliitis in COVID-19

Inhibition of SARS-CoV-2 infections in engineered human tissues using clinical-grade soluble human ACE2

Lack of evidence of ACE2 expression and replicative infection by SARSCoV-2 in human endothelial cells

Circulating platelets as mediators of immunity, inflammation, and thrombosis

Platelets and the immune continuum

Platelets: Signaling cells in the immune continuum

Platelets as cellular effectors of inflammation in vascular diseases

Activated platelets induce endothelial cell activation via an interleukin-1 pathway in systemic lupus erythematosus

Platelet transcriptome profiling in HIV and ATP-binding cassette subfamily C member 4 (ABCC4) as a mediator of platelet activity

Myeloid-related proteins 8 and 14 induce a specific inflammatory response in human microvascular endothelial cells

Platelet gene expression and function in patients with COVID-19

Platelets can associate with SARS-Cov-2 RNA and are hyperactivated in COVID-19

Platelets and COVID-19: Inflammation, hyperactivation and additional questions

Megakaryocytes and platelet-fibrin thrombi characterize multi-organ thrombosis at autopsy in COVID-19: A case series

Rosenstiel; HCA Lung Biological Network; Deutsche COVID-19 Omics Initiative (DeCOI), Longitudinal multi-omics analyses identify responses of megakaryocytes

WGCNA: An R package for weighted correlation network analysis

A review of the mitochondrial and glycolytic metabolism in human platelets and leukocytes: Implications for their use as bioenergetic biomarkers

Platelet-derived MRP-14 induces monocyte activation in patients with symptomatic peripheral artery disease

Platelet-derived S100A8/A9 and cardiovascular disease in systemic lupus erythematosus

Mrp8 and Mrp14 are endogenous activators of Toll-like receptor 4, promoting lethal, endotoxin-induced shock

Supported by the ESC Working Group on Pulmonary Circulation and Right Ventricular Function, COVID-19 and thrombotic or thromboembolic disease: Implications for prevention, antithrombotic therapy, and follow-up

Incidence of thrombotic complications in critically ill ICU patients with COVID-19

Lille ICU Haemostasis COVID-19 Group, Pulmonary embolism in patients with COVID-19

Pulmonary vascular endothelialitis, thrombosis, and angiogenesis in Covid-19

Thrombocytopathy and endotheliopathy: Crucial contributors to COVID-19 thromboinflammation

COVID-19 is, in the end, an endothelial disease

Severe COVID-19 Is a microvascular disease

Cellular and oxidative mechanisms associated with interleukin-6 signaling in the vasculature

S100A9/S100A8: Myeloid representatives of the S100 protein family as prominent players in innate immunity

Platelet expression profiling and clinical validation of myeloid-related protein-14 as a novel determinant of cardiovascular events

Platelet-derived S100 family member myeloid-related protein-14 regulates thrombosis

Endothelial cell-activating antibodies in COVID-19. medRxiv

Intracellular events in platelet activation induced by antiphospholipid antibodies in the presence of low doses of thrombin

Platelet mRNA: The meaning behind the message

Megakaryocyte and platelet transcriptomics for discoveries in human health and disease

Neutrophil-derived S100 calcium-binding proteins A8/A9 promote reticulated thrombocytosis and atherogenesis in diabetes

Myeloid-related protein 8/14 complex is released by monocytes and granulocytes at the site of coronary occlusion: A novel, early, and sensitive marker of acute coronary syndromes

Increased serum levels of S100A8/A9 and S100A12 are associated with cardiovascular disease in patients with inactive systemic lupus erythematosus

Elevated calprotectin and abnormal myeloid cell subsets discriminate severe from mild COVID-19

Neutrophil calprotectin identifies severe pulmonary disease in COVID-19

DAMPening mortality in COVID19: Therapeutic insights from basic cardiometabolic studies on S100A8/A9

Therapeutic vaccine against S100A9 (S100 calcium-binding protein A9) inhibits thrombosis without increasing the risk of bleeding in ischemic stroke in mice

The central role of the P(2T) receptor in amplification of human platelet activation, aggregation, secretion and procoagulant activity

Inhibition of ADP-induced P-selectin expression and platelet-leukocyte conjugate formation by clopidogrel and the P2Y12 receptor antagonist AR-C69931MX but not aspirin

Aspirin use is associated with decreased mechanical ventilation, intensive care unit admission, and in-hospital mortality in hospitalized patients with coronavirus disease

Decreased in-hospital mortality associated with aspirin administration in hospitalized patients due to severe COVID-19

Influence of sex on platelet reactivity in response to aspirin

Circulating monocyte-platelet aggregates are a robust marker of platelet activity in cardiovascular disease

German Research Foundation) (RO 6121/1-1 to C.C.R.). This work was also supported, in part, by NYU CTSA grant UL1TR001445, from the National Center for Advancing Translational Sciences (NCATS)