Platelets and SARS-CoV-2 During COVID-19: Immunity, Thrombosis, and Beyond

Abstract

COVID-19 is characterized by dysregulated thrombosis and coagulation that can increase mortality in patients. Platelets are fast responders to pathogen presence, alerting the surrounding immune cells and contributing to thrombosis and intravascular coagulation. The SARS-CoV-2 genome has been found in platelets from patients with COVID-19, and its coverage varies according to the method of detection, suggesting direct interaction of the virus with these cells. Antibodies against Spike and Nucleocapsid have confirmed this platelet-viral interaction. This review discusses the immune, prothrombotic, and procoagulant characteristics of platelets observed in patients with COVID-19. We outline the direct and indirect interaction of platelets with SARS-CoV-2, the contribution of the virus to programmed cell death pathway activation in platelets and the consequent extracellular vesicle release. We discuss platelet activation and immunothrombosis in patients with COVID-19, the effect of Spike on platelets, and possible activation of platelets by classical platelet activation triggers as well as contribution of platelets to complement activation. As COVID-19-mediated thrombosis and coagulation are still not well understood in vivo, we discuss available murine models and mouse adaptable strains.

It is now long established that the symptoms of COVID-19 are unpredictable and heterogeneous, with infections ranging from asymptomatic to significant hypoxia with acute respiratory distress syndrome (ARDS)¹ in more severe cases. ARDS is characterized by difficulty breathing, low blood oxygen levels, increased permeability of tight junctions and protective barrier loss, cytokine release (cytokine storm), inflammation, and thrombosis as a result of vascular injury.² ARDS also significantly increases the risk of secondary bacterial and fungal infections. Respiratory failure and thrombocytopenia in ARDS accounts for 70% of COVID-19-related mortalities.² The remaining COVID-19-related deaths result from a sepsis-like disease mediated through a cytokine storm in response to the viral infection and secondary infections. Multiple studies have identified increases in inflammatory cytokines such as IL (interleukin)-6, TNFα (tumor necrosis factor alpha), IL-6, IL-1β, and IL-18,^3–5 and steroid administration⁶ or anti-IL-6 monoclonal antibodies⁷ are associated with increased survival. Initial findings of severe lung necrosis and hyperinflammation, increased vascular damage, thrombosis, and coagulation led to trials therapeutically targeting classical platelet activation or the coagulation cascade. The incidence of thrombotic outcomes in the cumulative 30 days post symptom presentation in patients with COVID-19 is 25% to 30%,^8,9 while the incidence in influenza patients over the same period of time is approximately 11%.⁸ A detailed breakdown of the incidence of thrombotic events, including both arterial thromboses and venous thromboembolism, shows an increased risk of 17 to 33 (95% CI) in the first 7 days post-diagnosis, which then falls to 3 to 9 (95% CI), reaching 1.4 to 3 (95% CI) after week 9.⁹ Patients with COVID-19 who present to the intensive care unit exhibit an increased risk of ischemic stroke and disseminated intravascular coagulation with elevated C-reactive protein, D-dimer, and P-selectin but without any notable change in platelet levels or thromboplastin/prothrombin time that are typical for disseminated intravascular coagulation.¹⁰ Surprisingly, however, neither antiplatelet,¹¹ including P-selectin level reducing,¹² nor anticoagulant trials¹³ had a uniform beneficial effect, and failure of beneficial effect often coincided with time of infection and levels of inflammation. Antiplatelet treatments for instance had no benefit on mortality in patients with COVID-19.¹¹ Anticoagulant therapy, on the other hand, has been reported to improve mortality in moderate COVID-19 hospitalized, non–intensive care unit patients, but not in critically ill patients.¹³ These clinical trial outcomes, in addition to minimal changes in prothrombin or thromboplastin time and platelet levels, highlight the need for uncovering nonclassical pathways of platelet activation. In this review, we will summarize the prothrombotic, proinflammatory, and immune impact of SARS-CoV-2 and COVID-19 on platelets.

Overview of Platelets and Their Major Functions

Platelets are cell fragments, lacking nuclear content, which are generated by their megakaryocyte (MK) precursor and are primarily recognized as the blood component regulating hemostasis and thrombosis. Since platelets do not have a nucleus, most of their protein and RNA content is prepackaged by the MKs.¹⁴ The vascular importance of platelets has been attributed to their essential role in hemostasis and its pathogenic outcome of thrombosis and intravascular coagulation. These roles of platelets are manifested in pathological vascular occlusion in the setting of myocardial infarction, stroke and deep vein thrombosis.¹⁴ In addition to their well-described hemostatic function, platelets are active participants in the immune response to microbial organisms and foreign substances. Indeed their pathogen-mediated prothrombotic/procoagulant outcomes are so crucial that certain microbes have evolved to manipulate their function leading to disaggregation.¹⁵ Although incompletely understood, the immune role of platelets is a delicate balance between their response to pathogens and their functional contribution to thrombosis and coagulation.¹⁴

The major function of platelets is to provide the necessary support to maintain proper vascular hemostasis. Hemostasis is the process mediated by platelets to prevent leakage into the interstitial tissue by keeping blood circulating within the vessel. Platelets promote hemostasis through 3 sequential processes: adhesion, activation, and aggregation.¹⁶ Adhesion to the subendothelial matrix occurs upon endothelial disruption in the vessels.¹⁶ Two important proteins that aid in platelet adhesion are vWF (von Willebrand factor) and collagen. Circulating vWF immobilizes and sticks to the injury site upon endothelial damage, allowing platelet adhesion through the GP (glycoprotein)-IV-V complex. Platelets can also adhere to subendothelial collagen through α2β1 and GPVI.¹⁷ Platelet activation is induced when binding to both vWF and collagen. As a result of platelet activation, release of platelet granules mediated by GPVI leads to an increase of platelet activators like ADP (adenosine diphosphate; dense δ-granules), vWF (α-granules, in addition to vWF from endothelial cells), and thromboxane A2 that can activate other nearby platelets.¹⁸ Thromboxane A2 also aids in vasoconstriction to slow blood flow during endothelial damage. The binding of ADP and other potent platelet activators to their receptors also leads to platelet aggregation. Through an internal signaling mechanism, ADP can bind to the P2Y1 or P2Y12 receptors, ultimately leading to conformational change of the α2bβ3 receptor on the platelet surface.¹⁹ The active conformation of this receptor binds fibrinogen, allowing platelets to aggregate together and form a thrombus to prevent bleeding.

The pathological outcome of hemostasis is manifested in thrombosis and thrombotic outcomes. When vessel damage extends into the adventitial layer, TF (tissue factor), expressed on smooth muscle and adventitial cells, is released and generates thrombin through factor FVII converting it to VIIa consequently activating the intrinsic coagulation cascade.^13,14 During respiratory viral infections (influenza or COVID-19), TF can also be released from bronchial epithelial cells^15,16 and can circulate in extracellular vesicles with increased activity.^17–20 Thrombin, through the thrombin receptors PAR1 (protease-activated receptor 1; in humans), PAR3 (in mice), and PAR4, can lead to platelet degranulation, shape change and 3D thrombus formation as well as processing of fibrinogen to fibrin that contributes to clot retraction and tight plug formation on the injured vessel.²⁰ Arterial thrombosis occurs under high shear flow and is characterized by platelet-rich thrombi formed around ruptured atherosclerotic plaques and damaged endothelium.¹⁸ Venous thrombosis occurs under low shear flow, mostly around an intact endothelial wall. Venous thrombi are fibrin rich, encapsulating a large amount of red blood cells in addition to activated platelets.¹⁸

In addition to their well-recognized role in hemostasis and thrombosis, platelets also have innate and adaptive immune functions.²¹ Platelets contain multiple types of molecular pattern recognition receptors that recognize and initiate responses to pathogens. These include TLRs (toll-like receptors), NLRs (nucleotide-binding oligomerization domain-like receptors), and CLRs (C-type lectin receptors).²¹ Platelets express the transcripts for all TLRs although not all human platelets express all TLRs at the same time.²² Platelet surface TLR2 and TLR4 recognize surface components of invading pathogens while endosomal TLR3, TLR7, TLR8, and TLR9 recognize nucleic acids from the pathogens. Activation of platelet-TLR2,²³ -TLR7,²⁴ and -TLR9²⁵ lead to formation of heterotypic aggregates (HAGs) with leukocytes. Platelet-monocyte HAGs can contribute to TF expression on monocytes, leading to an increased prothrombotic response.²⁶ Platelet-TLR7 also contributes to activation of the complement system by releasing C3 stored in platelet α-granules.²⁷ Platelets contain other complement proteins in their α-granules including C4a, C1-I, and complement H,^28,29 further contributing to activation and regulation of the complement cascade.

In addition to thrombosis, platelets play a major role in immunothrombosis, or the process by which the immune system mediates thrombosis through platelets and neutrophils. At the core of immunothrombosis is the process of neutrophil DNA release, termed NETosis.²¹ The released DNA coated with highly positive histones serves as a mesh on which pathogens are trapped, neutralized, and removed from the circulation. However, in addition to pathogens, platelets also respond and attach to the mesh, forming aggregates during infection. These aggregates of platelets and netting neutrophils can be attached to the vessel³⁰ or circulate freely²⁷ contributing to prothrombotic events in occluded or inflamed vessels during infection or sterile inflammation.²¹ Platelets play a central role in this process not only by reacting to the netting neutrophils but also by initiating or contributing to NETosis. We now understand that the presence of platelets, through platelet-TLR4, shortens the time to lipopolysaccharide-mediated NETosis from hours to minutes.³¹ Platelet-TLR2 contributes to NETosis while platelet-TLR7 solely mediates the initial stages of NETosis.²⁷

Our understanding of platelet contribution to adaptive immunity is much more limited. Platelets can release serotonin and PF4 (platelet factor 4), both known to contribute to T cell activation. Activated platelets have been shown to be important in T cell rolling through direct binding via P-selectin.³² Platelets also play a role in CD4+ T cell differentiation and have been shown to increase the production of proinflammatory cytokines from them.³³ Platelets can also indirectly influence T cell activation by binding to professional antigen-presenting cells like dendritic cells and promoting their maturation via CD40L binding.³⁴ These platelets may also be able to activate dendritic cells to become antigen presenting.³⁵

Platelets also play a central role in influencing inflammation. They achieve this through the release of proinflammatory cytokines and chemokines in their α-granules and through EV release. Activation of platelets through many receptors (thrombotic or pattern recognition receptor) leads to release of platelet granule content or to full degranulation. ATP released from δ-granules can act as a danger associated molecular pattern to recruit macrophages and contribute to inflammasome activation in lymphocytes causing IL-1β release.³⁶ Platelets themselves can also release IL-1β as a function of TLR4³⁷ activation contributing to endothelial cell permeability.³⁸ Activation of the complement cascade by releasing C3 leads to generation of C3a and C5a components which are potent activators of inflammation by regulating pro- and anti-inflammatory cytokine secretion from cells in a local manner.³⁹ Activation of the C3a receptor on platelets can also contribute to thrombosis by mediating platelet adhesion, spreading, and Ca²⁺ influx.⁴⁰ Activation of the platelet-C5a receptor leads to inhibition of angiogenesis through the release of PF4.⁴¹ Thus, platelets can control neovascularization during inflammation at certain stages of infection.

Platelets From Patients With COVID-19

Platelets from patients with COVID-19 have been described as having an increased prothrombotic potential. Higher prothrombotic potential is evident in the increased propensity for aggregation in the presence of ADP,^42,43 epinephrine,⁴² TRAP,⁴³ and low concentrations of IIa,⁴⁴ and increased aggregation⁴² and adhesion on collagen (Table 1).⁴⁴ Although the propensity for aggregation is increased, the levels of classical α- (PF4) and δ-granule (serotonin) contents in platelets is reduced,⁴⁴ suggesting that these platelets may be exposed to certain activation signals. Increased plasma levels of proteins that could also be platelet-derived, such as sCD40LG (CD154),⁴⁶ TxB2,^46,47 P-selectin,^46,48 and vWF^48,49 have also been reported. Consistently, platelet activation is evident by increased expression of P-selectin,^43,45,47 CD40,⁴⁵ and CD63⁴⁷ on the surface of platelets from patients with COVID-19 (Table 1). In the presence of thrombin, platelets from patients with COVID-19 show an increased potential to release sCD40L⁴⁴ and IL-1β⁴⁴ with severity.⁴⁴ On a molecular level, thrombin also led to increased phosphorylation of Erk1/2, p38-MAP3K, and Cpla2 (cytosolic phospholipase A2) when platelets from patients with COVID-19 were compared with controls.⁴³ Activation of platelets is also evident by the presence of increased levels of platelet-derived extracellular vesicles in the plasma of patients with COVID-19.⁴⁴ It is important to note that the prothrombotic potential in most of the listed studies was assessed in comparison to healthy controls. Since patients with COVID-19 are severely ill, and many of the patients are aged and have comorbidities, it is challenging to delineate whether the origin of the increased prothrombotic potential is due to COVID-19 itself, the severity of disease, or comorbidities.

Table 1.

Changes in Platelets From Patients With COVID-19 When Compared With Healthy Controls

Increased expression of P-selectin and CD40LG on platelets results in the direct interaction with immune cells through PSGL1 and CD40, respectively.²¹ As mentioned in the previous section, these interactions are referred to as HAGs and the proportion of platelets to a leukocyte is often more than 2 to 1.^24,47 Blood from patients with COVID-19 shows increased HAGs between platelets and neutrophils,^43,45,50 platelets and monocytes,^43,45,47 and platelets and lymphocytes.⁴³ Platelet-monocyte HAGs in the patients with COVID-19 are also associated with higher expression of TF,⁴⁷ a primary trigger for coagulation, clot formation, and stabilization. One of the surprising findings that came from this infection is the previously unreported direct HAGs of platelets with CD4 and CD8 T cells.⁴³ How platelets influence or change these T-cells by directly interacting with them is still unclear and warrants further investigation.

With respect to their immune and inflammatory function, platelets from patients with COVID-19 were characterized by increased levels of interleukins such as IL-7, and chemokines such as CXCL1 (Groα), CCL22 (MDC), CCL5 (RANTES), and a decrease of proinflammatory cytokines such as IFN (interferon)-β, IFN-γ, IL-12β, IL-1β, and CX3CR1 in the plasma of the patients.⁴⁴ Some of these cytokines/chemokines are responsible for mediating the immune response during viral infection, chemotaxis of leukocytes such as neutrophils (CXCL1⁵¹) or lymphocytes (CX3CR1⁵² and CCL22⁵³), activation of T-cells (CD40LG, IL-7, IL-12β) and thus adaptive immunity.

Blood from patients with COVID-19 also shows evidence of NETosis^54,55 and platelets forming aggregates around the netting neutrophils,⁵⁰ highlighting their contribution to immunothrombosis. Immunothrombosis characterized by netting neutrophils, associated with platelets and fibrin, is also evident in lungs, kidney, and heart.⁵⁶ Immunofluorescent images of the lung tissue show platelets surrounding these netting structures and forming microthrombi.⁵⁰ Netting neutrophils from patients with COVID-19 also have increased TF deposition and serum from patients with COVID-19 can increase the expression of TF in neutrophils from healthy patients in a complement C3-dependent manner.⁴⁸ The exact contribution of platelets to COVID-19-induced NETosis has not been established. From previous studies with other respiratory viruses, such as influenza,²⁷ we can speculate that complement C3, can be released from platelets (thus present in the serum⁵⁴), and contributes at least partially to the increased NETosis. Complement 3 (C3) components⁵⁷ and PF4⁴⁴ are elevated in the plasma of patients with COVID-19 and C3-inhibition in COVID-19 patient plasma results in the reduction of NETosis.⁴⁹ As mentioned earlier, platelets can release C3 as a result of the recognition of viral ribonucleic acid (vRNA) by TLR7. C3, in turn, leads to activation of NETosis within 15 to 30 minutes depending on the donor,²⁷ platelet PF4 contributes to the physical compaction of NETosis.⁵⁸ Finally, TF deposited on COVID-19-initiated netting neutrophils⁴⁸ can serve as an initiator of platelet activation and thrombosis contributing to thrombotic outcomes.⁵⁹ Whether platelets release and contribute to the overall C3 pool and activation as a function of SARS-CoV-2 vRNA recognition remains to be established. It is clear, however, that there is a potential for the direct contribution of platelets to netting microthrombi that underlines the etiology of COVID-19 organ microthrombosis.

Platelets and SARS-CoV-2 vRNA

Viral RNA from SARS-CoV-2 has been found in the circulation and inside platelets from patients with COVID-19. Depending on the molecular method used, vRNA can vary in detection. Screening of vRNA in platelets by RT-qPCR methods using primers for different genes of the viral genome shows the presence of intact vRNA regions in 10% to 25% of the patients.^43,44 Screening of platelets using an RNAseq and tiled amplicon approach enriching for SARS-CoV-2 RNA shows that the vRNA is present in platelets from all patients tested (17 out of 17).⁶⁰ Using the same method shows the presence of vRNA in plasma in only 24% of the same patients and half of those patients did not survive the infection.⁶⁰ Consistently, a bigger cohort of 71 patients shows the presence of the vRNA in plasma of 27% of the patients and detection of SARS-CoV-2 vRNA was associated with disease severity.⁶¹

Presence of virus inside platelets from patients with COVID-19 was detected by transmission electron microscopy.^42,60,62 The fact that the sequenced vRNA in platelets is fragmented but covers almost the entire genome, and that qPCR primers often do not detect a uniform presence of vRNA in platelets, suggests that platelets break down the virus upon internalization. This observation proposes that platelets may function to remove SARS-CoV-2 vRNA that may be present in the circulation, halting its ability to infect or to get into other cells. It further proposes that the virus is unable to infect platelets, that is, produce functional progeny. The lack of viral reproduction or presence of infectious virus in platelets is supported by the inability of SARS-CoV-2-positive platelets from severe patients with COVID-19 to infect epithelial cells (VeroE6 or Calu-3)⁶³ 24 to 48 hours post-incubation. Vero E6 and Calu-3 cell lines traditionally are used to assess viral reproduction. Overall, transmission electron microscopy fails to detect any of the final stages of general viral replication defined as virion assembly and egress from platelets.⁶⁰ It can be speculated that platelets may serve as a dead-end for the virus due to their poor translational capability, lack of nucleus, and fast response. It is plausible to propose that platelets may also be necessary to prevent viremia by engulfing, removing, and digesting the virus. When engulfment and internalization of the virus is impaired in platelets, then the virus could reach other nucleated cells and have the potential to infect or induce a proinflammatory antiviral response in them. About 6% of monocytes in blood from patients with COVID-19 are infected with the virus.⁶⁴ It has been shown in vitro that frozen platelet-rich EDTA-plasma from patients with COVID-19 could contribute to the presence of infectious virus in macrophages,⁶² although it is unclear if the process of freezing halts platelet ability to neutralize the virus within.

Overall, digested, noninfectious SARS-CoV-2 vRNA is found in platelets regardless of severity while vRNA in plasma is associated with severity. We can speculate that platelets may be necessary to remove the virus⁴⁵ but may become dysfunctional with constant exposure to environmental signals⁶⁵ stemming from the infection.

Possible Receptors for SARS-CoV-2 Internalization and Platelets

The putative receptor for SARS-CoV-2 internalization and infection has been described to be ACE2 (angiotensin-converting enzyme 2).⁶⁶ ACE2 is necessary for the binding of the Spike protein on the virus, and the initiation of proteolytic cleavage by TMPRSS2 leads to initiation of endocytosis in cells.⁶⁶ This molecular system is particularly important for the internalization of the alpha- to delta variants. Mutations in Omicron on the other hand lead to viral RNA entrance in epithelial cells as a result of interaction of the virus with ACE2 without requiring the proteolytic activity of TMPRSS2.⁵⁶ Viral internalization assessed by vRNA presence, but not necessarily infection, has been reported in many cells throughout the body. However, many of these cells have minimal or completely lack ACE2 expression. Studies have identified ASGR1 (Kd=94.8 nmol/L) and KREMEN1 (Kd=19.3 nmol/L)⁶⁷ among 5054 human membrane proteins as additional binding partners for Spike protein. Kd for ACE2 was determined to be 12.4 nmol/L.⁶⁷ These 2 surface proteins serve as less efficient receptors, but efficiently contribute to infection when ACE2 is absent.⁶⁷ AXL⁶⁸ and CD209/DC-SIGN⁶⁹ have also been suggested as additional candidates for SARS-CoV-2 receptors. However, infection with the intact native SARS-CoV-2 virus downstream of C209 or AXL have yet to be shown. Platelets, mostly smaller than 2 µm,⁷⁰ have detectable levels of mRNA for ASGR1,^43,70 KREMEN1,^43,70 and AXL,⁷⁰ as KREMEN1 is increased in the platelets of patients with COVID-19⁴³ when compared with healthy subjects. Transcript levels of DC-SIGN^43,70 and ACE2^43,70 are below detection limits using sequencing or qPCR. However, DC-SIGN at protein levels has been detected on platelets^71–73 and is important for platelet activation in the presence of Dengue.⁷² ACE2 protein expression is described below.

Basigin, or CD147, is ubiquitously expressed and gained a significant amount of recognition as a novel receptor for viral entrance and infection in lungs using murine models expressing human CD147.⁷⁴ More recent studies, however, have provided proof that CD147 on human cells does not interact with SARS-CoV-2 Spike protein^75,76 and removing this receptor from lung epithelial cells does not change the propensity for infection.⁷⁶ Platelets express CD147⁷⁷ and studies have shown that anti-CD147 antibodies reduce platelet activation in the presence of Spike protein.⁷⁸ Since this receptor does not interact with Spike directly, it proposes that CD147 may be downstream of the actual receptor for SARS-Cov-2 and internalization of vRNA in platelets cannot be attributed to it, although activation⁷⁹ (direct or indirect) cannot be excluded.

The expression of ACE2, the putative receptor for SARS-CoV-2, on platelets is minimal and is mostly detected at the protein level. RNAseq or qPCR of platelets from healthy individuals or patients with COVID-19 have failed to detect uniform expression of ACE2 in platelets. Our own RNAseq studies using patients with COVID-19 have shown similar results. Using 3 different TaqMan gene expression assays for ACE2, we detected the presence of ACE2 mRNA in platelets; however, the 3 primer pairs never overlapped and detected ACE2 in the same individual, suggesting that either the RNA half-life is short and it is degraded quickly, or the efficiency of the primers is not the same. The latter could also be explained by a low copy number of the mRNA. RNA studies by sequencing and qPCR are often a function of RNA stability, abundance, and primer recognition for a particular mRNA. Utilizing confocal-microscopy and immune fluorescence on blood from patients with COVID-19 and healthy subjects, it has been shown that platelets from healthy subjects have scarce protein expression of ACE2.⁶⁰ Surface ACE2 protein levels in myocardial infarction or patients with COVID-19 were more easily detected. The antibodies used in this study were optimized using an epithelial cell line that was transfected to stably express both ACE2 and TMPRSS2.⁶⁰ Western blot analysis of platelets from healthy subjects were also able to detect ACE2 and TMPRSS2 in some individuals⁶⁰ while the levels in others was below detection.^43,60 Although ACE2 and TMPRSS2 are detected in some individuals, it really does not explain the uniform presence of vRNA in platelets.

Other mechanisms that may contribute to viral internalization do not require expression of ACE2 on platelets. It has been shown that platelets can internalize SARS-CoV-2 virions attached to platelet-derived microparticles.⁶⁰ In vitro studies show that virion-microparticle internalization happens at a later time after direct virion internalization.⁶⁰ This observation proposes that platelets may be primed by the virus to release microparticles necessary for additional removal of virions from the circulation. As infection progresses, viral particles may also be internalized as a result of clearance of IgG-virion immune complexes. In this case, the virions become internalized as a result of FcγRIIa on platelets by the Fc fragment of IgG.^80,81 Crosslinking of FcγRIIa leads to clathrin-mediated internalization of the FcγRIIa:IgG-virion immunocomplexes and the subsequent transport of the virions to the lysosomes.^80,81 This mechanism has been described in platelets with influenza virions⁸² and is a plausible explanation for the presence of vRNA in people up to 19 days post-positive qPCR test.

Overall, it can be hypothesized that as the second most abundant blood component, platelets have evolved multiple mechanisms to ensure removal of virions from the circulation.

Platelet Response to SARS-CoV-2

As mentioned earlier, viruses are sensed by the host cell through pathogen recognition receptors on or inside the host cell. The role of platelet-TLRs in sensing SARS-CoV-2 is yet to be analyzed. Studies from nucleated cells have shown that TLR2 on bone marrow–derived macrophage cells is able to recognize the envelope of SARS-CoV-2 and blocking TLR2 signaling increases survival and reduces inflammatory cytokine secretion.⁸³ Platelet-TLR2 activation by specific agonists is known to increase platelet aggregation potential and HAG formation with neutrophils and monocytes.²³ Whether SARS-CoV-2 is able to induce a similar activation of platelet-TLR2 remains to be established. In vitro studies using synthetic Spike protein also suggest possible recognition and activation of TLR4 in THP1 cells,⁸⁴ but whether this process occurs with intact and infectious virions is unclear. Lipopolysaccharide-activated TLR4 in platelets may contribute to increased activation and IL-1β release²¹ while calprotectin-activated TLR4 can lead to platelet pyroptosis.⁸⁵ It is unclear if SARS-CoV-2 can activate platelet-TLR4 and what that means for platelet activation.

In view of the presence of digested vRNA in human platelets, the most relevant receptors at the beginning stages of infection are TLR7 or TLR8. It has been proposed that single-stranded vRNA fragments generated from the viral genome of SARS-CoV-2 can induce inflammation via TLR7/8 in the lungs of mice and promote human dendritic cell activation as well as IFN-α and inflammatory-cytokine production from them.⁸⁶ It remains to be established if the intact virus is sensed in a similar fashion, particularly when the primary interferon response in patients seems to be impaired. Platelets have functional TLR7^24,25,27 that is known to sense viruses such as influenza²⁷ or EMCV,²⁴ leading to HAG formation with neutrophils,^24,25 and platelet-mediated NETosis through the release of complement C3²⁷ from platelet granules. SARS-CoV-2 replication-induced vRNA is sensed by MDA5 in epithelial cells,⁸⁷ but, since SARS-CoV-2 does not replicate in platelets, it is unclear if MDA5 plays any direct role in platelet activation.

The host cell has evolved to utilize various sensing mechanisms and there is never just 1 receptor responsible for the manner in which a cell responds. In addition, the receptor responsible for viral internalization is not necessarily responsible for activation. As mentioned previously, the direct sensors of SARS-CoV-2 in platelets are still unknown. What is now understood is that isolated platelets from healthy human subjects, incubated with infectious SARS-CoV-2 (WA-alpha-strain), do not form 3D aggregates.⁶⁰ The presence of infectious virus, however, induces activation of programmed cell death pathways in platelets. Apoptosis and necroptosis markers, Casp (Caspase)3 and phosphorylated-MLKL (pMLKL) respectively, colocalized with SARS-CoV-2 virions in the same platelets. Similarly, active-Casp3 and MLKL were evident in platelets from patients with COVID-19.⁶⁰ Transmission electron microscopy of platelets from either patients with COVID-19 or healthy subjects, incubated with SARS-COV-2 showed morphological features consistent with apoptosis such as membrane blebbing or with lytic cell death (necroptosis and pyroptosis, Figure 1), as evidenced by lack of intact membrane, membrane rupture and visible release of platelet cytoplasmic content (Figure 1).⁶⁰ transmission electron microscopy also captured release of extracellular vesicles such as microparticles and exosomes.⁶⁰ The contribution of programmed cell death pathways to platelet EV-generation is not clear, nor if Casp3 may directly contribute to MP release. When it comes to pMLKL activation, studies from other cells have suggested that, in addition to formation of oligomerization and formation of membrane pores that lead to necroptosis, pMLKL can self-restrict⁸⁸ the levels of necroptosis by generation of exosomes that excrete pMLKL,⁸⁸ limiting the damaging activity of this protein. Platelets incubated with SARS-CoV-2 indeed show exosome release from the platelet cytoplasm raising the necessity of further exploration of this EV generation mechanism. RNAseq from our lab shows increased levels of the pyroptotic initiator caspase, Casp1⁶⁰ in patients with COVID-19 and protein levels of Casp1 and noncanonical inflammasomal Casp4 and Casp5⁶³ may also be increased in patient’s platelets. Future studies are necessary to establish the immune versus thrombotic outcomes of all of these pathways.

Figure 1.

Transmission electron micrographs (TEMs) of platelets incubated with SARS-CoV-2 or from a patient with COVID-19. Platelets were incubated with infectious SARS-CoV-2 (WA strain; alpha variant); control represents the media in which the virus was eluted post purification. TEMs show release of cytoplasmic content post membrane rupture that, in general, is the last step of lytic programmed cell death in nucleated cells. These TEMs suggest signaling beyond classical platelet activation. Figure is borrowed from Koupenova et al⁶⁰ (Figure 5 and Figures S2 and S4) and is copyright of AHA Journals (CircRes).

It is important to mention that in nucleated cells, necroptosis and pyroptosis lead to release of danger-associated molecular patterns. These danger associated molecular patterns are proteins such as calprotectin (S100A8/A9).⁸⁹ S100A8/A9 belong to the family of 24 calcium-binding-S100 proteins that have various functions related to inflammation, cell death, endothelial activation, etc⁹⁰ Platelets from patients with COVID-19 have been shown to passively release S100A8/A9 at much higher levels than platelets from healthy donors.⁴⁵ Mechanistically, S100A8/A9 can induce platelet pyroptosis through a TLR4-GSDMD axis (Figure 2) leading to excessive release of inflammatory cytokines.⁸⁵ S100A8/A9 can also mediate procoagulant platelet function by induced exposure of phosphatidyl serine (PS), and release of PS-positive microvesicles through GP1b and CD36 on platelets.⁹¹ The presence of S100A8/A9 in plasma of patients with COVID-19 has also been associated with disease severity⁹² and increased platelet levels may contribute to amplified endotheliopathy,⁴⁵ MI⁹³ and other thrombotic⁹⁴ complications. COVID-19 plasma also contains platelet-derived HMGB1, and increased levels were assessed to be an independent predictor of worse clinical outcome.⁷⁰ Proper proteomics are necessary to assess if SARS-CoV-2 can lead to direct release of S100A8/A9 or HMGB1 from platelets through necroptosis⁶⁰ or pyroptosis rather than by classical platelet activation.

Figure 2.

Platelet activation during COVID-19 and contribution to thrombosis. Various classical platelet agonists become available as a result of the infected environment. Among these are ATP and ADP (adenosine diphosphate) released from dying, infected or damaged cells. ATP (not shown in the figure) is also quickly converted to ADP. TF (tissue factor) generated by damaged vessels and epithelial cells can lead to thrombin (IIa) generation by the extrinsic coagulation cascade. ADP and IIa can act on their respective receptors on platelets: ADP on P2Y12 and P2Y1 and IIa on PAR1 and PAR4. The result is increased granule release and consequent P-selectin and CD154 surface expression leading to interaction with leukocytes (heterotypic aggregation [HAG]) in addition to increased activation of the fibrinogen receptor, α2bβ3 (GPIIb/IIIa or CD41/CD61) leading to increased adhesion and aggregation. Additionally, platelet-monocyte HAGs lead to increased TF expression in monocytes, further contributing to the coagulation potential of the environment. These processes can be controlled by antiplatelet drugs. In addition to the response to environmental cues, platelets can also internalize the virus partially through ACE2 (angiotensin-converting enzyme 2) and partially by picking up microparticles with attached virions. Viral RNA in platelets can then initiate programmed cell death pathways in platelets leading to Casp3 activation, MLKL (Mixed Lineage Kinase Domain Like Pseudokinase) phosphorylation and consequently to extracellular vesicle release. Phosphorylated MLKL can oligomerize and form pores on the surface leading to intracellular content release. Molecules such as S100A8/A9, generally located in the cytoplasm of cells, can then be released in the environment. Extracellular vesicles and S100A8/A9 can lead to endothelial damage which further amplifies thrombosis and coagulation. Casp3 activation can lead to PS (phosphatidyl serine) surface translocation on platelets. Platelets with translocated PS, and perhaps EVs with PS on their surface can provide the surface for increased coagulation. It is unclear if classical platelet agonists have an effect on the response of platelets in their direct interaction with the virus. Illustration credit: Ben Smith.

Overall, we can conclude that programmed cell death pathways are activated in platelets post-exposure to SARS-CoV-2 (Figure 2) possibly in an attempt to control viral replication. One can hypothesize that, as a result of environmental factors due to damage and high inflammation, overactivated platelets may be less capable of removing infectious virus from the circulation. As a result, infectious virions can reach distal organs and, although not causing productive infection, can still lead to a highly inflammatory response as a function of viral sensing mechanisms, some related to cell death.

Platelet Response to SARS-CoV-2-Spike Protein

It is now understood that long COVID patients may have a persistent presence of Spike S1 protein (but not other SARS-CoV-2 proteins) in the circulation up to 15 months.^95,96 It has been hypothesized that the Spike protein may be released from cells that continue to have a low grade lingering infection; however, it is unclear what predisposes the immune system to fail in subduing the infection. The biochemical properties, protein modifications and physiological effect of the Spike protein present in the circulation is still not well understood. Additionally, it remains to be established whether this circulating Spike protein has the potential to be thrombogenic. Since thrombotic risk subsides significantly after 4 weeks⁹ and the levels in the circulation are low⁹⁶ it is unlikely that S1 reaches critical concentrations to become a thrombotic risk.

Nevertheless, studies using Spike-pseudovirus or free Spike protein as a model of interaction between platelets and SARS-CoV-2 have gained certain momentum. Platelets in the presence of Spike-pseudovirus release HMGB1⁷⁸ and vWF⁷⁸ and promote platelet-monocyte HAGs as well as inflammatory cytokine release from monocytes.⁹⁷ It was proposed that the S1 on the pseudovirus interacts with platelets through CD42b⁹⁷ or CD147.⁷⁸ Recombinant full⁹⁷ and S1 protein⁷⁸ incubated with platelets induces platelet aggregation,⁷⁸ as well as increased release of thromboxane A2, sNOS2, and H₂O₂ in a TLR4-dependent manner.⁹⁸ These studies propose that if the Spike protein reaches critical concentrations in the circulation it may have a prothrombotic effect by using surface receptors that the actual virus does not use. As mentioned, CD147 on human cells does not interact with SARS-CoV-2 Spike protein.^67,68 It is important to point out that platelets incubated with the infectious purified SARS-CoV-2 alpha strain did not undergo aggregation at a concentration of 1 infectious viral particle per 10 platelets.⁶⁰ Thus, although Spike-pseudoviruses and recombinant protein treatment of platelets are necessary to address principal platelet-Spike interactions, unless reproduced by treatment with SARS-CoV-2 virus itself, they should only be used to determine how the Spike protein alone has the potential to interact with platelets.

Programmed Cell Death Pathways and Platelet Number During COVID-19

In traditional immune cells, programmed cell death would lead to either nonlytic (low inflammatory) or lytic (highly inflammatory) death. The process would ultimately result in reduced cell number. If platelets were undergoing cell death in a traditional immune cell manner, we would expect some level of thrombocytopenia to be present in patients with COVID-19. Meta-analysis of 9 COVID-19 studies has associated reduced platelet count with severe COVID-19 and mortality.⁹⁹ However, the thrombocytopenia in patients is mild and in certain cases completely absent. As these studies are generally based on 1-time nonsynchronized blood draw, it is impossible to assess if there is a transient drop in platelet number, that may be compensated by increased production. Indirect support for this hypothesis comes from the wide range of TPO (thrombopoietin) levels measured in patients with COVID-19.⁴³ Certain patients are also reported to have an elevated platelet number. Studies in mice with another positive strand RNA virus (EMCV) have shown that there is a transient small drop in platelet count that fluctuates between 2 and 24 hours post-infection²⁴ and may be compensated thereafter. Reduced platelet count could be attributed to the formation of platelet-leukocyte aggregates,^24,43 NETing-microthrombi,⁵⁰ or to activation of platelet programmed cell death. Evidence of Casp3 and MLKL activation in platelets may be assessed by extracellular vesicle generation and thus reduced platelet size (1.5 µm) rather than their death. Small platelets indeed coexist with unusually large platelets in blood from patients with COVID-19⁶⁰ (Figure 3). Future studies are necessary to assess the contribution of programmed cell death pathways to transient thrombocytopenia, and platelet size, and to determine how programmed cell death occurs in a platelet.

Figure 3.

Platelets from COVID-19 patients vary in size. Consistent with the reported increased mean platelet volume (MPV), we find platelets bigger than 5 µm. We also find platelets smaller than 2 µm proposing a loss of platelet content that may be attributed to activation of programmed cell death pathways. Figure is borrowed from Koupenova et al⁶⁰ and is copyright of AHA Journals (CircRes).

Megakaryocytes and SARS-CoV-2: Implications for Platelet Number and Function

Platelet number is a function of production versus clearance, and aggregation versus de-aggregation, all of which can be happening simultaneously. Animal models (discussed below) are necessary to assess these processes in vivo. As mentioned earlier, TPO is increased in various patients with COVID-19 and the increased incidence of atypically large platelets⁶⁰ or platelets with high mean platelet volume⁴² are a hallmark of this disease. Large, reticulated platelets have been associated with thrombotic outcomes, but the evidence that these platelets are causative of thrombosis is indirect and their occurrence remains poorly understood. It is unclear what processes regulate the increased size of these large platelets and what processes regulate their production during infection, or which MKs produce them. What is evident is that MKs are able to sense the presence of infection either indirectly through increased signals from an overactivated immune system or through direct interaction⁴² with the virus. Platelet-producing MKs are found predominantly in the marrow of flat bones¹⁰⁰ and also have been described in lungs^21,101 and spleen. MKs residing in the interstitial lung tissue reach ploidy of 4 N and have a dendritic cell-like phenotype;²¹ MKs residing in the bone marrow can reach ploidy of 128 N and are known to be the major producer of platelets.¹⁰⁰ COVID-19 studies have claimed that MKs reside in the vasculature of lungs, brain, and heart. Although not impossible, a lack of morphological features indicative of MK polyploid presence, and reports of syncytia (fusion of leukocytes) in SARS-CoV-2 infected lungs,^102–104 make these reports difficult to interpret.

Platelet Activation by Antibodies as a Result of SARS-CoV-2 Infection

In addition to internalization, antibodies against SARS-CoV-2 can also contribute to platelet activation and the prothrombotic response. As infection progresses, antibodies against the virus increase in the circulation and the quantity and quality of these antibodies vary from person to person. Studies have shown that high antibody titers in patients with COVID-19 are associated with disease severity and increased mortality.^105,106 Relevant to the increased antibody titer in human platelets is activation of FcγRIIa (the only Fc receptor expressed by platelets) by immunocomplexes of IgG antibodies and influenza which can lead to increased platelet activation as assessed by increased surface P-selectin expression, 12-HETE and microparticle release.⁸² Whether this occurs with SARS-COV-2 is unclear. However, incubation of platelets with immunocomplexes of Spike-IgG antibodies in combination with recombinant Spike protein show an increase in the formation of thrombotic aggregates on vWF.¹⁰⁷ Interaction with FcγRIIa and level of vWF-mediated prothrombotic potential was dependent on the type of altered anti-Spike IgG glycosylation.¹⁰⁷ Particularly relevant to platelet activation are immunocomplexes of IgG with hypofucosylation and hypergalactosylation.¹⁰⁷ These observations raise the possibility of variations in platelet activation as a result of differences in the adaptive immune response, becoming relevant at later stages of infection.

Platelet Activation by Classical Platelet Activators That May Be Generated by SARS-CoV-2 Infection

Severe COVID-19 has been characterized by extensive organ damage, increased inflammation, and increased levels of cytokines. Cell damage leads to an increased level of molecules such as ADP and ATP that serve as classical thrombotic agonists. Activation of platelet P2RX1 by ATP can result in platelet shape change favoring adhesion while activation of P2Y1 and P2Y12 receptors by ADP can lead to increased platelet aggregation (Figure 2). Cytokines such as IL-6, IL-1β, and TNFα, which are increased in the plasma of patients with COVID-19,^3,108 can also contribute to platelet hyperactivity and hypercoagulability.^65,109,110

TF can increase in the circulation as a result of infection either by monocytes⁴⁷ or by infected lung epithelial cells.¹¹¹ TF, also known as F3 of the (extrinsic) coagulation cascade, ultimately leads to thrombin generation and direct activation of platelets. In addition to platelet activation, thrombin converts fibrinogen to fibrin, ultimately enabling FXIII to crosslink fibrin and stabilize the platelet-formed clot. Studies utilizing chemically-inactivated SARS-CoV-2 collected from infected A549-ACE2 epithelial cells, show that the viral particle can associate with TF in plasma, leading to platelet activation in a dose-dependent manner.¹¹² Although not impossible, it would be interesting to establish if infectious SARS-CoV-2 with antigenic potential and thus with proteins in their proper tertiary and quaternary structure^113,114 is able to engage in the same association with TF (Figure 3). However, it is clear that noninfectious altered viral particles, often generated during infection, may contribute to overall platelet activation¹¹² regardless of engagement with Spike receptors on the surface of the platelet.

Platelet Contribution to Coagulation During SARS-CoV-2 Infection

In addition to their direct interaction with the virus and response to the viral particles carrying TF, platelets also provide the necessary surface for coagulation to occur. Platelets play a major role in controlling the thrombin burst and fibrin-clot generation, stabilization, and dissolution, all achieved in a localized manner. Exposure of PS on the platelet surface is necessary to support platelet-dependent thrombin generation.^115,116 However, PS may not account for all variations in coagulation outcomes and other factors are necessary to support coagulation through FXa and thrombin generation.^115,116 Patients with COVID-19 show an increased presence of PS-rich platelet microparticles that may be depleted in nonsevere patients.⁴⁴ It can be hypothesized that platelets from patients with COVID-19 with increased Casp3 activity⁶⁰ would also have elevated PS exposure on their surface, proposing an increased potential for coagulation (Figure 2). These patients’ coagulation activity may deplete the free circulating platelet-derived PS-microparticles.

Interestingly, although the prothrombotic potential of patients with COVID-19 platelets is increased,^42–44 the overall procoagulant potential, measured by dual activation with IIa and convulxin, is significantly reduced.¹¹⁷ However, patients with COVID-19 who present to the intensive care unit exhibit an increased risk of ischemic stroke and deep vein thrombosis. Similar to coagulation outcomes in deep vein thrombosis, COVID-19-associated coagulopathy is characterized by C-reactive protein, D-dimer, and P-selectin. However, contrary to deep vein thrombosis, COVID-19-associated coagulopathy presents without any notable change in platelet levels or thromboplastin/prothrombin time.¹⁰ These observations propose that there may be an altered thrombolysis response in patients with COVID-19.¹⁰ The contribution of platelets to thrombolysis in COVID-19 remains to be elucidated, but it is important to mention that platelets contain various proteins that can have an effect on fibrinolysis. Some of these thrombolysis-related proteins are PAI-1, α₂-antiplasmin, thrombin activatable fibrinolysis inhibitor,¹¹⁸ etc. Future studies are necessary to elucidate the contribution of platelets to dysregulated coagulation during COVID-19.

Overall, dysregulated platelet activation by the local environment may play a critical role in dysregulated thrombotic outcomes not only locally but can propagate distal outcomes such as myocardial infarction and stroke. COVID-19 increased oxidative stress assessed by NOX2 activation in the circulation is associated with increased thrombotic outcomes.³ Understanding how the virus may directly activate platelets, how platelets may feed into each cascade activation during different stages of infection, and how each cascade step may activate them should be carefully evaluated in order for proper targeting of platelet-mediated coagulation.

Gut Permeability and Thrombosis in Patients With COVID-19

SARS-CoV-2 is known to productively infect not only the bronchial epithelium but also enterocytes in the gut lining leading to a large amount of virions in the intestine.¹¹⁹ As a result of infection, the intestinal barrier becomes damaged, permeable to bacterial toxins and treatment of patients with COVID-19 with antibiotics is associated with bacteremia.¹²⁰ In a cohort of 81 patients with COVID-19, endotoxemia measured by zonulin (a marker of gut permeability) and lipopolysaccharide presence in plasma correlated with thrombotic events.¹²¹ Bacteremia, in turn, is known to have thrombotic manifestations.¹²² Overall, endotoxemia, and bacteremia in certain patients under antibiotic treatment, may be additional contributors to thrombosis in patients with COVID-19.

Murine Models and Viral Strains for Studying COVID-19

What we have learned about platelets and their contribution to thrombosis and coagulation throughout the pandemic has greatly contributed to our overall understanding of how platelets may respond to the infection. In the next stage of understanding, it is crucial to utilize the appropriate murine models and the appropriate viral strains that can be used in mice. The range in severity of COVID-19, in addition to long COVID manifestations has presented challenges in modelling COVID-19 in animal models that recapitulate human pathologies.

Small animal laboratory species were initially tested for susceptibility to SARS-CoV-2 infection; however, many early studies in these models failed to mimic the severe disease seen in hospitalized patients. Indeed, golden syrian hamsters, ferrets, and nonhuman primates develop mild to moderate viral disease when challenged with the original Wuhan strain of SARS-CoV-2.¹²³ Early experiments revealed that common laboratory mice were not permissive to SARS-CoV-2 infection via murine ACE2 binding. Thus, multiple approaches were developed to express human ACE2 (hACE2) in the respiratory tract transiently or through genetic introduction of hACE2 expression. The emergence of the Omicron family of variants has also impacted analysis of Omicron-related disease because of its intrinsic properties that prevent it from establishing advanced infections in the lower respiratory tract in hACE2 experimental models.¹²⁴ Here, we discuss the current murine models and viral strains relevant to studying platelets and COVID-19.

Adenoassociated Virus Expression of Human ACE2

As discussed, SARS-CoV-2 is unable to infect conventional laboratory mice via the murine ACE2 receptor. When SARS-CoV emerged, studies demonstrated that laboratory mice were permissive to SARS-CoV infection; however, C57BL/6 mice do not develop severe disease. Early studies in 2020 utilized an adenoassociated virus (AAV)-hACE2 vector to express hACE2 in the respiratory tract. Single-dose infections with AAV-hACE2 for 14 days results in robust expression of hACE2 in the respiratory tract of C57BL/6 mice. Although transient, the use of this approach facilitates the expression of hACE2 in knockout mice to characterize the type I IFN response raised by SARS-CoV-2.¹²⁵ Following AAV expression of hACE2, SARS-CoV-2 infection results in disease that recapitulates severe COVID-19. AAV-ACE2-infected mice exhibit rapid weight loss and death at 8 days post-infection. Further, analysis of lung tissue reveals disease pathologies that recapitulate human COVID-19 such as diffuse peribronchial infiltrates, expansion of pulmonary myeloid-derived inflammatory cells and activated lymphoid cells in lung tissue such as CD4+, CD8+, and natural killer cells. Despite these findings, it is unclear what role thrombosis plays in this transient expression model.¹²⁵

Transgenic Models Expressing hACE2

During the SARS-CoV outbreak in 2003, transgenic mice expressing hACE2 under the control of the epithelial cell K18 (cytokeratin-18) promoter were developed to experimentally assess SARS-CoV pathogenesis (Table 2).¹³⁶ Given that SARS-CoV-2 also infects host cells via hACE2, K18-hACE2 mice were rapidly evaluated to model the pathogenies of SARS-CoV-2. K18-hACE2 mice infected with SARS-CoV-2 exhibit rapid weight loss and succumb to severe disease by 8 days post-infection.¹³⁷ At low intranasal doses, K18-hACE2 mice also display severe weight loss and lethality, albeit with slower kinetics. Infection of K18-hACE2 mice with SARS-CoV-2 mirror a significant number of symptoms observed in the AAV-ACE2 infection model.¹²⁶ Although the majority of viral replication occurs in the respiratory tract, vRNA is also detected in other organs including the heart, spleen and kidney. Immune cell infiltration is observed in the lung tissue in addition to lesions in the alveolar spaces and alveolar wall thickening resulting in interstitial lung disease.^{124,126–128} By days 6 to 8, infected mice display severe viral pneumonia with pan-alveolar inflammation, immune cell infiltration, and edema. As discussed, vRNA can also be detected in the heart tissue of infected K18-hACE2 mice; however, histological abnormalities in the heart tissue are inconsistent and it is unclear if age or gender of the mice are limiting factors in the infection reaching the heart.¹²⁸ In some studies, hypereosinophilia is also detected. In addition, lengthening of prothrombin time is preceded by increased D-dimer concentrations throughout the course of infection.^126,138 Indeed, K18-hACE2 mice display a progressive increase in the circulating plasma levels of D-dimer, indicative of hypercoagulation, vascular permeability and hemorrhagic infarction. Thus, the K18-hACE2 transgenic model could be utilized for assessing the role of the environment in SARS-CoV-2-induced thrombosis.¹²⁹

Table 2.

Models to Study COVID-19 in Mice

In addition to the conditional expression of hACE2 in lung epithelial cells, hACE2 has been constitutively expressed in murine models using the CAG promoter (Table 2).¹³⁰ CAG promoter–driven hACE2–transgenic (CAG-hACE2) mice are highly susceptible to SARS-CoV-2 and develop severe COVID-19 in response to intratracheal tract infection.¹³⁰ CAG-hACE2 develop severe symptoms such as weight loss and mortality. CAG-hACE2 also display acute lung injury and high levels of cytokine production following infection. Indeed, CAG-hACE2 mice display significant neutrophil and lymphocyte infiltration to the lung tissue. In addition, severe pneumonia and alveolar wall thickening with progressive infiltration of inflammatory cells are observed in lungs of mice infected with higher doses of SARS-CoV-2.¹³⁰

The third murine knock-in model expresses hACE2 in place of the mouse Ace2 gene generated via targeted mutagenesis (Table 2). In this case, hACE2 expression is regulated by the endogenous mouse Ace2 promoter and enhancer elements.¹²⁶ Interestingly, infection with SARS-CoV-2 in this model does not result in significant loss of body weight. However, viral replication analysis reveals that SARS-CoV-2 can replicate efficiently in the upper respiratory tract of hACE2 knock-in mice when compared with wild-type mice expressing murine Ace2. Higher viral loads are also detected in the lung tissue. Viral RNA is detected in the brain tissue albeit at lower levels. However, severe lung pathology is detected at the latter stages of the infection. Thus, hACE2 knock-in mice only recapitulate mild forms of infection that is self-limiting in mice¹²⁶ and further supports the notion that this particular virus utilizes many forms of entry in human cells that may not necessarily result in productive infection.

Given the wide-ranging outcomes of SARS-CoV-2 infection in transgenic ACE2 mouse models an emphasis should be placed on the relative copy number of ACE2 in the resulting model. Indeed, both K18-hACE2 and CAG-hACE2 mice express much higher copy numbers of human ACE2 when compared with human ACE2 knock-in mice controlled by the endogenous Ace2 promoter. Based on these variations, the appropriate dosing and mouse model should be considered when studying thrombosis in experimental models of COVID-19.

Murine adaptable MA10 SARS-CoV-2

Although advantageous for rapidly assessing therapeutics, the K18-hACE2 transgenic model is difficult to assess the contribution of genetic factors without intercrossing with knockout mice. In light of this, a mouse adapted SARS-CoV-2 strain was created that can establish infection in laboratory mice. MA10 was derived from the Wuhan isolate of SARS-CoV-2 with the introduction of 2 amino acid mutations within the RNA-binding protein of the Spike protein to facilitate binding to murine ACE2 (Table 2). Through 10-generation serial infections of BALB/c mice, the resulting strain, MA10, acquired mutations in NSP4 (T285I), NSP7 (K2R, E23R), Spike (Q493K), and ORF6 (F7S).¹³² Infection of BALB/c mice with MA10 results in weight loss and interstitial lung disease with age-related severity. Indeed, infection of 10-week-old mice results in minimal weight loss followed by complete recovery. However, 20-week-old mice display significant weight loss irrespective of dose. Furthermore, approximately 50% of aged mice fail to control MA10 and succumb to the infection. Thus, the use of MA10 in aged mice recapitulates the age and sex-dependent sensitivity to SARS-CoV-2 infection and COVID-19 in humans.¹³¹ Approximately 40% of symptomatic and asymptomatic COVID-19 survivors develop postacute sequelae, termed postacute sequelae of Covid-19 or long COVID, with persistent symptoms such as chest pain, cognitive decline or brain fog, and chronic interstitial lung disease.¹³⁹ Interestingly, infection of 1-year old BALB/c mice results in a postacute sequelae of Covid-19-like condition that persists for up to 4 months. One-year-old BALB/c mice infected with low-dose MA10 display significant weight loss with a survival rate of 75%. Histological assessment of the mice reveal that they retain consistent heterogeneous lung disease up to 120 days.¹³² Key persistent histological features include hypercellularity with immune cell infiltration (tertiary lymphoid structures), abundant smooth muscle actin positive fibroblasts and collagen deposition, which is characteristic of pulmonary fibrosis. Thus, the MA10 infection model may be useful for modeling SARS-CoV-2 in genetic knockout models and for assessing the key pathologies associated with long COVID.

MISTRG6 Humanized Mouse Model

MISTRG6 mice were originally engineered via a human/mouse homolog gene replacement strategy. MISTRG6 mice allow for expression of M-CSF (for the development of monocytes and tissue macrophage development), GM-CSF and IL-3 (in lung alveolar macrophages). MISTRG6 also allows for SIRPα, which permits tolerance to human cells, in addition to TPO and IL-6 for improved engraftment and humoral responses on a Rag2/γ common chain deleted background.^133,134 Expressing these components allows for the successful engraftment of human hematopoietic stem and progenitor cells, to reconstitute an immune system that recapitulates human immunity (Table 2). Recently, this model has been adopted to express hACE2 via AAV as previously described.¹²⁵

Hematopoietic stem and progenitor cell-engrafted MISTRG6 mice infected intranasally with SARS-CoV-2 exhibit immune responses and pathologies associated with COVID-19. Indeed, MISTRG6/AAV-ACE2 mice display severe weight loss, accumulation of vRNA, fibrotic lung disease, and T cell lymphopenia.¹³⁵ In contrast to other AAV dependent or transgenic expression of hACE2, severe chronic weight loss persists for 35 days post infection. Furthermore, MISTRG6-hACE2 humanized mice displayed a more severe form of interstitial lung disease when compared with the previously discussed infection models. Indeed, at days 14 to 28 post infection, MISTRG6-hACE2 humanized mice display increased collagen deposition and a deterioration of the alveolar structures. Thus, these mice recapitulate the severe lung pathology and ARDS seen in patients with severe COVID-19, with possibilities for evaluation of thrombosis and coagulation in this organ. However, to date, markers of thrombosis have not been evaluated in the MISTRG6-hACE2 humanized model.

Conclusions

With regard to platelet activation during infection with various pathogens, there is always the question of the contribution of the infected highly inflammatory environment versus direct interaction of platelets with the virus. Direct interactions propose exploration of avenues of platelet activation that differ from or contribute to classical platelet activation. This is particularly important when it comes to understanding why antiplatelet or P-selectin targeted therapies are limited in their effect on COVID-19 thrombosis or survival. Lack of effect of these platelet-targeting therapies propose that during infection other functions beyond thrombosis, namely immunity and antiviral response, should be of focus. As a community, we have well characterized the altered platelet prothrombotic and procoagulant potential in COVID-19. The effect on platelets with time as a function of direct interaction with the virus versus the environment during all stages of infection continues to be under-evaluated. Future studies utilizing the appropriate murine models and or viral strains are necessary to properly address platelet-mediated immunity and how this platelet function may contribute to the dysregulated impact of infection on thrombosis and coagulation. Only then, can we begin to address the proper targeting of pathways that go beyond classical antiplatelet therapies.

Article Information

Sources of Funding

M. Koupenova is supported by NIHLBI R01 HL153235 and F. Humphries is supported by the Charles Hood Foundation Child Health grant.

Disclosures

None.