Mechanisms of immunothrombosis in COVID-19

Abstract

Purpose of review:

Coronavirus disease 2019 (COVID-19) is an infectious disease caused by severe acute respiratory syndrome coronavirus-2. Over the past year, COVID-19 has posed a significant threat to global health. Although the infection is associated with mild symptoms in many patients, a significant proportion of patients develop a prothrombotic state due to a combination of alterations in coagulation and immune cell function. The purpose of this review is to discuss the pathophysiological characteristics of COVID-19 that contribute to the immunothrombosis.

Recent findings:

Endotheliopathy during COVID-19 results in increased multimeric von Willebrand factor release and the potential for increased platelet adhesion to the endothelium. In addition, decreased anticoagulant proteins on the surface of endothelial cells further alters the hemostatic balance. Soluble coagulation markers are also markedly dysregulated, including plasminogen activator inhibitor-1 and tissue factor, leading to COVID-19 induced coagulopathy. Platelet hyperreactivity results in increased platelet-neutrophil and -monocyte aggregates further exacerbating the coagulopathy observed during COVID-19. Finally, the COVID-19-induced cytokine storm primes neutrophils to release neutrophil extracellular traps, which trap platelets and prothrombotic proteins contributing to pulmonary thrombotic complications.

Summary:

Immunothrombosis significantly contributes to the pathophysiology of COVID-19. Understanding the mechanisms behind COVID-19-induced coagulopathy will lead to future therapies for patients.

Introduction

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is a novel coronavirus from zoonotic origin, first reported in December 2019 in Wuhan, China [1]. One year after the World Health Organization (WHO) declared the spread of SARS-CoV-2 a global health emergency, over 113 million cases have been reported worldwide, causing more than 2.5 million deaths as a result of coronavirus disease 2019 (COVID-19) [2]. While most COVID-19 patients experience only mild symptoms such as fever and cough, more severe cases present with substantial pulmonary complications, such as pneumonia and acute respiratory distress syndrome (ARDS), as well as an abnormal high rate of thrombosis [3–6*]. A meta-analysis of 36 studies calculated the incidence of venous thromboembolism (VTE) as high as 28% for patients admitted to the intensive care unit (ICU). In the non-ICU setting, the estimated incidence of VTE was 10% [7]. Although less common, arterial thrombotic events such as ischemic stroke, myocardial infarction and systemic arterial embolism have also been reported [8, 9]. Numerous autopsy reports of COVID-19 patients described increased thrombi in pulmonary arteries [10–15] and small fibrin thrombi in the alveolar capillaries [16**–18]. While comparable observations were made for lungs from influenza A/H1N1-mediated ARDS patients [19, 20], alveolar capillary microthrombi were 9 times more prevalent in the COVID-19 patients [16**]. In agreement with this, a retrospective study from the Netherlands found a twofold higher 30-day VTE rate in hospitalized COVID-19 patients versus influenza patients [21*]. Similarly, an increased incidence of ischemic stroke in COVID-19 patients compared to influenza patients was reported [22]. The observation that many COVID-19 patients with thrombotic complications were on standard prophylactic anticoagulation has raised the debate on whether therapeutic anticoagulation or alternative anti-thrombotic strategies should be considered [23–25]. However, properly controlled randomized clinical trials have yet to be finalized.

Clearly, thrombotic complications, and more specifically pulmonary thrombosis, are a hallmark of severe COVID-19 pathology and contribute to mortality. Pulmonary thrombi have been linked to the development of hypoxemia in the early stages of ARDS in COVID-19 and are estimated to contribute up to 10% of all COVID-19 related deaths [26*]. In this review, we will present an overview of emerging concepts in COVID-19 pathophysiology that explain the increased susceptibility to thrombotic complications. To this end we will discuss putative mechanisms and therapeutic approaches of pulmonary endotheliopathy, COVID-19 coagulopathy, platelet hyperreactivity and immunothrombosis (Figure 1).

Figure 1. Overview of COVID-19 pathophysiology.

Severe COVID-19 is characterized by pulmonary complications, including acute respiratory distress syndrome (ARDS), and thrombosis. Increased susceptibility to thrombotic complications is caused by synergistic interplay of endotheliopathy, platelet hyperreactivity, immunothrombosis and coagulopathy. VWF, von Willebrand factor. TF, Tissue factor. PAI-1, Plasminogen activator inhibitor-1.

SARS-CoV-2 induced endotheliopathy

In normal blood vessels, the endothelial cell monolayer is designed to prevent pathological thrombus formation. Endothelial dysfunction is a principal determinant of microvascular dysfunction by shifting the vascular equilibrium towards vasoconstriction and a pro-coagulant state. Data from autopsy studies in COVID-19 have identified marked endothelial cell apoptosis [27**], together with loss of endothelial cell tight junctions in the pulmonary microvasculature [16**, 28]. SARS-CoV-2 infects the host cells using one of its spike glycoproteins to bind to the angiotensin converting enzyme 2 (ACE2) [29], which is expressed on several cell types, including endothelial cells [30*]. SARS-CoV-2 first binds ACE2 and is then cleaved by the serine protease TMPRSS2 to coordinate entry into the host cell [31**]. Varga and colleagues were the first to provide evidence of viral elements within endothelial cells [27**]. The presence of SARS-CoV-2 virus within endothelial cells suggests direct viral effects may contribute to the endothelial injury [32, 33]. Subsequent perivascular inflammation may provide a detrimental negative feedback loop, as plasma from COVID-19 patients has been shown to induce rapid endothelial cell cytotoxicity [34]. Notably, unique features of the lung endotheliopathy of COVID-19 have become apparent [16**, 35]. Compared to influenza patients, COVID-19 patients had increased endothelial disruption and alveolar capillary micro- and macrothrombi were more prevalent [16**, 18].

So far, no marker of endothelial dysfunction has been found to truly discriminate non-COVID-19 ARDS from COVID-19 ARDS [36, 37]. Nevertheless, several markers of endothelial dysfunction correlate with admission to the intensive care unit (ICU) and poor outcomes in COVID-19 patients [38–41]. Von Willebrand factor (VWF) has been studied the most, and appears to strongly predict outcomes. VWF is a multimeric glycoprotein produced exclusively by endothelial cells and megakaryocytes and plays an important role in hemostasis [42]. To maintain hemostasis, VWF acts as a bridge between the activated endothelium and platelets [43]. Besides its hemostatic role, VWF has recently been recognized as an effective mediator of inflammation. VWF actively participates in inflammatory processes by recruiting leukocytes at sites of vascular inflammation and plays a critical role in thromboinflammation [44]. Importantly, the activity of VWF is restricted by the metalloprotease ADAMTS13, which cleaves newly released highly active VWF multimers into smaller, less thrombogenic and inflammatory multimers [45]. Inefficient digestion of large VWF multimers is known to play a critical role in the pathogenesis underlying microvascular occlusion in thrombotic thrombocytopenic purpura (TTP) [46], sickle cell disease [47] and sepsis [48].

In COVID-19 patients, VWF levels are markedly increased [37, 39*, 49–57*], while ADAMTS13 levels are significantly decreased [40, 49, 52, 58]. This imbalance results in elevated levels of highly reactive VWF multimers in the circulation, which spontaneously bind platelets, leading to thrombotic complications. Importantly, VWF hyperreactivity is strongly correlated to disease severity [52, 57*, 59, 60] and has been demonstrated in several independent studies to predict thrombotic complications [6*, 57*, 61], oxygen requirement [62] and mortality [41, 55, 60, 63, 64]. Similar to COVID-19, a mounting VWF/ADAMTS13 imbalance, culminating in the accumulation of platelet-rich microthrombi has been shown to induce secondary thrombotic microangiopathy in sepsis [48]. These observations have fueled the hypothesis that COVID-19 is a thrombotic microangiopathy similar to thrombotic thrombocytopenic purpura (TTP) [65, 66]. However, unlike TTP, levels of ADAMTS13 are not fully reduced, and platelet counts are not decreased in most COVID-19 patients [67]. Nevertheless, the observed mild systemic VWF/ADAMTS13 imbalance might be exacerbated locally [68], e.g., in the pulmonary vascular bed where both virus-mediated endothelial damage [69, 70] and subsequent hypoxia [71] increase VWF release. Therapeutically, targeting excessive VWF with Caplacizumab, which inhibits platelet binding to VWF, might be an attractive avenue to treat SARS-CoV-2 associated thrombotic microangiopathy [72]. Alternatively, plasma exchange has been proposed to restore hemostatic balance and reduce levels of circulating inflammatory markers, as is done for several thrombotic microangiopathies [73].

Hemostatic abnormalities result in a procoagulant phenotype in COVID-19

Despite abundant hematological abnormalities, SARS-CoV-2 RNA was rarely detected in blood samples [74], implying an important pathological cross-talk between the disrupted endothelium, the COVID-19 associated cytokine storm and other components of blood [75]. Several receptors expressed on endothelial cells function to promote anticoagulant pathways, including thrombomodulin (TM). TM is a constitutively expressed transmembrane receptor with important anticoagulant and antifibrinolytic activities [76]. Numerous studies have reported increased levels of soluble thrombomodulin (sTM) in COVID-19 patients, indicating active shedding from the endothelial surface [36, 39*, 77]. Interestingly, plasma concentrations of soluble thrombomodulin correlate with in-hospital mortality in COVID-19 patients [39*]. Besides sTM, SARS-CoV-2 infection was also reported to affect other modulators of coagulation such as plasminogen activator inhibitor 1 (PAI-1) [78–85], tissue-type plasminogen activator [81, 83, 85, 86], angiopoietin-2 [84, 87] and tissue factor (TF) [79]. Mechanistically, IL-6 is thought to be the causative cytokine for these observations, and clinical trials targeting IL-6 in COVID-19 patients have shown promising results [88–91]. IL-6 is known to directly impact the endothelium inducing increased levels of PAI-1 [82**], fibrinogen [92], and TF [93]. Changes in these hemostatic parameters result in an increased potential to form fibrin clots. This is evident in vivo by elevated levels of circulating D-dimers [94], a biomarker of both fibrin formation and degradation. Remarkably, even though the presence of D-dimers suggests fibrin is actively being lysed, autopsy studies revealed numerous fibrin depositions in the pulmonary microvasculature and alveolar space of lungs from patients with COVID-19 [10–15]. This suggests that dysregulation of the coagulation and fibrinolytic pathways is a major aspect of COVID-19 pathogenesis and is not just a bystander to SARS-CoV-2 infection. Interestingly, a recent study observed sustained hypercoagulable changes in COVID-19 patients 4 months after hospital discharge, suggesting that patients might be susceptible for thrombotic complications well beyond recovery [95*]. Importantly, these findings suggest plasma from convalescent COVID-19 patients used in transfusion medicine could be hypercoagulable and future studies examining the effect of “long” COVID-19-induced coagulopathy on plasma transfusion are necessary.

Bleeding complications due to disseminated intravascular coagulation (DIC) are a common cause for death in severely ill septic patients [96] and has been a concern in COVID-19 patients [97]. Similar to COVID-19, critically-ill sepsis patients exhibit exaggerated inflammation and are at an increased risk of thrombotic complications. However, while patients with sepsis frequently present with thrombosis and bleeding that stem from consumption of platelets and coagulation factors [96, 98], this is only rarely seen in patients with COVID-19 [99]. Indeed, severe COVID-19 patients usually present with normal or only mildly reduced platelet counts, high fibrinogen levels and increased or unaltered prothrombin times [100–104]. Moreover, when we retrospectively compared plasma from septic and COVID-19 patients we observed significant differences between the two diseases [36]. Compared to healthy donors, patients with COVID-19 showed increased thrombin generation and increased endogenous plasmin generation. In contrast, septic patients had similar thrombin generation and endogenous plasmin generation compared to healthy donors. Both groups had increased fibrinogen levels resulting in enhanced fibrin formation in both COVID-19 and sepsis. However, sepsis patients had significant resistance to fibrinolysis compared to COVID-19. These data suggest fundamental differences in the pathophysiology of these diseases and highlight the need for tailored treatment for COVID-19-associated coagulopathy [105**, 106]. Accordingly, a number of clinical trials were initiated to determine the best antithrombotic strategy [107–111]. While these trials are still ongoing, interim results have shown that early initiation of prophylactic anticoagulation was associated with a decreased risk of 30-day mortality and no increased risk of serious bleeding events [112]. Furthermore, an important distinction in the therapeutic benefits for moderate and severely ill hospitalized patients has emerged. In a large clinical trial conducted in more than 300 hospitals worldwide (incorporating the REMAP-CAP, ATTACC and ACTIV-4a trials), full dose heparin given to moderately ill patients hospitalized for COVID-19 was safe and reduced the requirement of vital organ support. In contrast, full dose heparin did not reduce the need for organ support in critically ill COVID-19 patients requiring ICU support and concerns were raised regarding an increased bleeding risk [113].

Platelets become hyperreactive during SARS-CoV-2 infection

Thrombocytopenia is common in viral infections [114] and considered a prognostic marker for mortality in ICU patients [115]. Surprisingly, most COVID-19 patients present with normal platelet counts or only mild thrombocytopenia [116]. Nevertheless, extensive meta-analyses indicate that low platelet counts are associated with more severe COVID-19 and an increased risk for adverse outcomes [116, 117*]. A possible explanation for these findings could be consumption of platelets during micro- and macro-vascular thrombosis, as several studies reported the presence of platelets in thrombi found in multiple organs of COVID-19 autopsy cases [53*, 118, 119*].

Morphologically, platelet ultrastructure appeared normal during SARS-CoV-2 infection. However, RNA-sequencing revealed significant changes in the platelet transcriptome of COVID-19 patients [120*]. Many differentially expressed transcripts appear to be unique for COVID-19 compared to other infectious diseases. However, the exact cause for altered gene expression remains unknown [121]. One possible explanation would be direct interactions of SARS-CoV-2 with platelets or megakaryocytes. Several platelet surface receptors have reported to mediate binding and entry of various viruses in platelets resulting in platelet activation [122]. Interestingly, only a subset (8–24%) of COVID-19 patients had detectable SARS-CoV-2 RNA in their platelets [120*, 123*]. Nevertheless, hyperreactive platelets are common in SARS-CoV-2 infected patients [56, 118, 120*, 123*–125]. We and others reported P-selectin expression, a marker for platelet activation, was significantly increased on resting platelets from COVID-19 patients [56, 120*, 124*]. Additionally, COVID-19 platelets were more prone to activate after in vitro exposure to weak agonists as evidenced by significantly increased P-selectin exposure, platelet aggregation, IL-1β and soluble CD40 ligand production [56, 120*, 123*]. Intriguingly, the increased reactivity to low concentrations of thrombin seems to be specific to COVID-19 ARDS patients [126]. These observations lead to several clinical trials targeting platelets, with early results showing reduced mortality in COVID-19 patients treated with aspirin [127, 128].

The exact mechanism for hyperreactive platelets in COVID-19 is still unknown. Zhang and colleagues provided evidence for a direct interaction of SARS-CoV-2 with ACE-2 on platelets, resulting in platelet activation through the MAPK pathway [125]. However, it is unlikely this suggested pathway is the sole contributor to platelet activation in COVID-19, considering several other groups were unable to detect ACE-2 RNA or protein on platelets or CD34⁺-derived megakaryocytes [120*, 123*, 129]. Intriguingly, not all markers of platelet activation were consistently up during COVID-19 pathology. PAC-1 binding, a marker for integrin α_IIbβ₃ activation, was reduced in platelets from COVID-19 patients compared to healthy donors after in vitro stimulation with different agonists [56, 120*, 130]. In addition, COVID-19 platelets also had a reduced ability to become procoagulant compared to healthy donor platelets [131]. The question remains if the reduced ability to form procoagulant platelets directly causes the hyper-aggregating platelets observed during COVID-19 [132] or whether exhausted platelets recirculate in the blood after activation and sequester in the lung [118]. Of interest to the latter hypothesis is the observation of increased circulating platelet extracellular vesicles [123*] and apoptotic platelets [133] in COVID-19 patients. The latter were induced by antibodies in COVID-19 sera and correlated with thrombotic events [133].

Finally, platelet activation during viral infections also triggers inflammation. Release of α-granule stored molecules, such as cytokines, chemokines and several surface molecules contribute to the recruitment and activation of immune cells [122]. Consistent with increased platelet P-selectin in COVID-19 patients, several studies also observed increased platelet-leukocyte aggregates [56, 120*, 124*, 130, 134]. Particularly, platelet-monocyte interactions are of interest as platelets from severe COVID-19 patients induce TF expression in monocytes, which correlated with disease severity and mortality [124*]. Moreover, platelet degranulation might contribute to the cytokine storm described in COVID-19 as many of the cytokines involved in viral infections were reduced in platelets from COVID-19 patients, indicating active release [123*].

Inflammation induced thrombosis contributes to outcomes in COVID-19

An important and common observation from COVID-19 autopsy studies is the involvement of leukocytes within platelet-rich macro and micro-vascular occlusions in the lung, kidney and heart [135, 136]. Circulating platelet-leukocyte complexes are common in COVID-19 and are thought to contribute to disease progression [56, 118, 120*, 124*, 134]. In particular, platelet-neutrophil interactions are known to induce the formation of neutrophil extracellular traps (NETs) [137–139]. NET formation is a form of neutrophil cell death, triggered to trap and kill invading pathogens [140]. NETs are extracellular webs of decondensed chromatin decorated with histones and antimicrobial proteins. However, NETs also capture blood cells and form a procoagulant and prothrombotic scaffold [139, 141] associated with detrimental collateral damage in sepsis and other thromboinflammatory diseases [142]. Several studies have identified NETs as important constituents of micro- and macro-vascular thrombi [14, 143–146] and bronchoalveolar lavage fluid [53*] in COVID-19 patients, even after virus clearance from the lungs [147]. Additionally, biomarkers of NETs are increased in serum [148] and plasma [49, 59, 147, 149–151] from COVID-19 patients and correlate with disease severity[144], were associated with thrombosis [152, 153] and predicted outcomes [150*–152]. Neutrophils from COVID-19 patients readily made NETs, unlike healthy neutrophils, and plasma from COVID-19 patients induced in vitro NET formation by healthy neutrophils [53*, 151]. Importantly, NET levels were reduced to healthy donor levels in convalescent patients suggesting NETs resolve after COVID-19 [53*]. Lastly, several studies have found reduced levels of DNases in critically ill COVID-19 patients [149, 154]. Together these data indicate SARS-CoV-2 infection creates a hostile environment priming NET formation. Besides platelet-neutrophil interactions, other proposed mechanisms of NET formation in COVID-19 include direct interactions of SARS-CoV-2 with neutrophils [146, 155], complement-mediated NET formation [156, 157], circulating anti-phospholipid antibodies [158] or monocyte-driven hypercytokinemia [118, 159].

Importantly, early comparative studies have found four times more immunothrombotic occlusions containing NETs in COVID-19 cases compared to influenza pneumonia underscoring neutrophil-driven immunothrombosis as a key element of severe COVID-19 [160*]. Besides prothrombotic effects, the release of NETs in the lower respiratory tract of SARS-CoV-2-infected patients can trigger direct cytotoxic effects on the alveolar epithelium and endothelium resulting in loss of alveolar integrity [135, 137, 161–163]. This further contributes to the injurious vicious cycle of endotheliopathy induced coagulopathy and highlights the important therapeutic potential of targeting NETs in COVID-19 [164]. Interestingly, the corticosteroid dexamethasone [165] and the gout drug colchicine [166], which both were shown to reduce mortality in COVID-19 patients, are both NET inhibitors [167–169]. Furthermore, degrading NETs may represent an additional therapeutic option for NET-mediated lung injury in COVID-19 in the form of the FDA-approved Dornase alfa (recombinant DNase1) for which currently two COVID-19 clinical trials are underway (NCT04359654; NCT04409925). DNase1 was previously shown to reduce thrombosis and improve mortality in preclinical models of immunothrombosis by degrading NETs [137, 170–172]. Of-note, also low-molecular weight heparin degrades NETs and was shown to block COVID-19 plasma-induced NET formation in vitro [152]. Lastly, plasmapheresis might also devoid NET inducers circulating in COVID-19 patients; however, so far this has not been explored.

Conclusions

The cross-roads between infectious diseases and coagulation have become increasing intertwined during the COVID-19 pandemic. The combination of endothelial dysfunction, increased circulating procoagulant proteins, platelet hyperreactivity and altered inflammatory cell function appear to contribute to thrombotic complications in COVID-19 patients. Understanding the mechanisms behind COVID-19-induced coagulopathy will help develop novel therapies to treat immunothrombotic complications. In addition, future studies examining the acute versus chronic nature of the alterations in coagulation will have significant implications for transfusion medicine.

Keypoints.

• COVID-19 patients are at increased risk for thrombotic complications.

• COVID-19-specific coagulopathy and lung endotheliopathy contribute to prothrombotic changes via different mechanisms than sepsis-related disseminated intravascular coagulation.

• Hyperreactive platelets and neutrophil extracellular traps contribute to detrimental immunothrombosis.