SARS-CoV-2 Ion Channel ORF3a Enables TMEM16F-Dependent Phosphatidylserine Externalization to Augment Procoagulant Activity of the Tenase and Prothrombinase Complexes

Severe SARS-CoV-2 infection is complicated by dysregulation of the blood coagulation system and high rates of thrombosis, but virus-intrinsic mechanisms underlying this phenomenon are poorly understood. Increased intracellular calcium concentrations promote externalization of phosphatidylserine (PS), the membrane anionic phospholipid required for assembly and activation of the tenase and prothrombinase complexes to drive blood coagulation. TMEM16F is a ubiquitous phospholipid scramblase that mediates externalization of PS in a calcium-dependent manner. As SARS-CoV-2 ORF3a encodes a presumed cation channel with the ability to transport calcium, we hypothesized that ORF3a expression by infected host cells perturbs the cellular calcium rheostat, driving TMEM16F-dependent externalization of PS and enhancing procoagulant activity. Using a doxycycline-inducible system, synchronized expression of ORF3a in A549 pulmonary epithelial cells resulted in a time-dependent augmentation of tissue factor (TF) procoagulant activity exceeding 9-fold by 48 hours (p < 0.0001), with no change in TF cell-surface expression. This enhancement was dependent upon PS as determined by inhibition with the PS-binding protein lactadherin. Over 2-fold enhancement of prothrombinase activity (p < 0.0001) was also observed by 48 hours. ORF3a increased intracellular calcium levels by 18-fold at 48 hours (p < 0.0001), as determined by the intracellular calcium indicator fluo-4. After 16 hours of ORF3a expression, more than 60% of cells had externalized PS (p < 0.001) without increased cell death, as quantified by flow cytometry following annexin V binding. Immunofluorescence microscopy staining for ORF3a, annexin V, and nuclei confirmed ORF3a expression within internal and cell surface membranes and increased PS externalization. PS externalization was insensitive to the pan-caspase inhibitor z-VAD-FMK, and there was no evidence of apoptotic activation as determined by caspase-3 cleavage. By contrast, ORF3a expression did not augment coagulation in cells deficient in the calcium-dependent phospholipid scramblase TMEM16F. Similarly, ORF3a-enhanced TF procoagulant activity (p < 0.01) and prothrombinase activity (p<0.05) was completely abrogated using TMEM16 inhibitors, including the uricosuric agent benzbromarone that has been registered for human use in over 20 countries. Live SARS-CoV-2 infection of A549-ACE2 cells increased cell surface factor Xa generation at MOI 0.1 (p < 0.01) but not MOI 0.01 or following heat inactivation of the virus, and RNA sequencing confirmed ORF3a induction without increased F3 expression. RNA sequencing of human SARS-CoV-2 infected lung autopsy and control tissue (n= 53) confirmed these findings in vivo. Immunofluorescence staining for ORF3a and KRT8/18 and CD31 in SARS-CoV-2 infected human lung autopsy specimens demonstrated ORF3a expression in pulmonary epithelium and endothelial cells, highlighting the potential pathologic relevance of this mechanism. Here we demonstrate that expression of the SARS-CoV-2 accessory protein ORF3a increases the intracellular calcium concentration and TMEM16F-dependent PS scrambling to augment procoagulant activity of the tenase and prothrombinase complexes. Our studies of human cells and tissues infected with SARS-CoV-2 support the pathologic relevance of this mechanism. We highlight the therapeutic potential to target the ORF3a-TMEM16F axis as with benzbromarone to mitigate dysregulation of coagulation and thrombosis during severe SARS-CoV-2 infection. Disclosures Schwartz: Miromatrix Inc: Membership on an entity's Board of Directors or advisory committees; Alnylam Inc.: Consultancy, Speakers Bureau. Schulman: CSL Behring: Consultancy, Research Funding.

Procoagulant activities of skeletal and cardiac muscle myosin depend on contaminating phospholipid

Recent reports indicate that suspended skeletal and cardiac myosin, such as might be released during injury, can act as procoagulants by providing membrane-like support for factors Xa and Va in the prothrombinase complex. Further, skeletal myosin provides membrane-like support for activated protein C. This raises the question of whether purified muscle myosins retain procoagulant phospholipid through purification. We found that lactadherin, a phosphatidyl-l-serine–binding protein, blocked >99% of prothrombinase activity supported by rabbit skeletal and by bovine cardiac myosin. Similarly, annexin A5 and phospholipase A2 blocked >95% of myosin-supported activity, confirming that contaminating phospholipid is required to support myosin-related prothrombinase activity. We asked whether contaminating phospholipid in myosin preparations may also contain tissue factor (TF). Skeletal myosin supported factor VIIa cleavage of factor X equivalent to contamination by ∼1:100 000 TF/myosin, whereas cardiac myosin had TF-like activity >10-fold higher. TF pathway inhibitor inhibited the TF-like activity similar to control TF. These results indicate that purified skeletal muscle and cardiac myosins support the prothrombinase complex indirectly through contaminating phospholipid and also support factor X activation through TF-like activity. Our findings suggest a previously unstudied affinity of skeletal and cardiac myosin for phospholipid membranes.

Procoagulant Activity of Skeletal Muscle Myosin Is Not Caused By Contaminating Phosphatidylserine-Vesicles

Skeletal muscle myosin (SkM) can bind factor (F)Xa and FVa, thereby providing a surface that promotes thrombin generation by the prothrombinase complex (FXa:FVa:Ca++) that cleaves prothrombin. A recent BLOOD paper (Novakovic & Gilbert, 2020) asserted that this activity of SkM preparations is entirely due to contaminating phosphatidylserine (PS)-containing phospholipid vesicles in SkM preparations because annexin V and lactadherin neutralized the ability of SkM to enhance prothrombin activation. However, annexin V and lactadherin are certainly not monospecific for binding PS as they are globular proteins that can bind other lipids and many proteins. Without any PS measurements or use of any reagents specific for PS (e.g., monoclonal (mAb) anti-PS antibodies), that report, in a gross overinterpretation of its incomplete data, dismissed any direct role for myosin for SkM's procoagulant activity. When we previously observed that annexin V is inhibitory of SkM's support for prothrombinase activity, we sent a sample of SkM (Cytoskeletal Inc) to Avanti Polar Lipids for quantitation of the PS content based on liquid chromatography-mass spectrometry. That analysis showed that only a small amount of PS was present in the SkM, approximately 0.90 µmol PS per 40. µmol SkM. This amount of PS present in SkM preparations is not enough to explain SkM's procoagulant activity. For example, standardized purified prothrombinase reaction mixture assays show that 10 nM SkM enables formation of 3 nmol thrombin/min while 0.22 nM PS (in 1.1 nM phosphatidylcholine (PC)(80%)/PS(20%) vesicles) enables formation of only 0.4 nmol thrombin/min (Figure 1A). We directly assessed the role of contaminating PS for SkM's prothrombinase support using the well characterized anti-PS mAb 11.31 (aka mAb PGN632). When the ability of mAb 11.31 to inhibit prothrombinase enhancement by SkM or, in controls, by PC/PS vesicles was determined, the data showed that, in controls, mAb 11.31 at 1.0 nM severely inhibited PC/PS vesicle's enhancement of prothrombinase by > 90% (Figure 1B). The dose-response gave an inhibitory IC50 value of 0.2 nM which is near this mAb's reported Kd of 0.17-0.35 nM for PS, establishing the potent ability of this anti-PS mAb to neutralize PS procoagulant activity. However, there was no substantial inhibition of SkM's enhancement of prothrombinase by the anti-PS mAb 11.31 at up to 1 nM mAb 11.31 (Figure 1B). This indicates that contaminating, PS-containing vesicles are not a significant factor for SkM-dependent enhancement of prothrombin activation by the SkM preparation -- in direct contradiction of the assertion of Novakovic & Gilbert (BLOOD 2020). That recent report was correct that annexin V inhibits the ability of SkM to enhance prothrombinase as well as the ability of PC/PS vesicles to do so (Figure 1B). So it was the overinterpretation of the annexin V and lactadherin data as well as the failure to provide any direct measurement of PS that were problematic in that report. The observation that annexin V and lactadherin inhibit SkM's enhancement of prothrombinase merits further studies to understand what may be their mechanistic influences. Another informative test for an essential role for PS in SkM's enhancement of prothrombinase involves the use of FXa that lacks its gamma-carboxyglutamic acid (Gla) domain because this N-terminal domain of FXa is required for FXa binding to PS-containing phospholipid vesicles. Data in Figure 1A show that SkM, but not PS-containing PL vesicles, supports the prothrombinase activity of des-Gla Domain (DG)-FXa which lacks its Gla domain. Dose-response data show that, in prothrombinase assays, DG-FXa has only 1% activity in the presence of PC/PS vesicles but has 25-35% activity in presence of SkM (Figure 1A). These data indicate that SkM's procoagulant activity does not absolutely require the Gla domain of FXa whereas PS-containing vesicles do require the Gla domain of FXa for significant activity. These data prove that the recent assertion that PS vesicle contamination, not myosin, explains SkM's procoagulant activity represents an overinterpretation of annexin V and lactadherin data and is simply wrong. In conclusion, both new data here, i.e., PS content of SkM preparations and data for effects of anti-PS mAb 11.31 (Figure 1B) and previous data (Deguchi et al, Blood 2016 and J Biol Chem, 2019), affirm that the myosin protein is a key factor for SkM's ability to enhance prothrombin activation. Figure 1 Disclosures No relevant conflicts of interest to declare.

Cardiac Myosin Acts Is a Potent Procoagulant in Vitro and In Vivo

Thrombin generation and fibrin formation can cause occlusive thrombosis and myocardial infarction is caused by occlusive thrombi. Exposure and release of cardiac myosin (CM) are linked to myocardial infarction, but CM has not been accorded any thrombotic functional significance. Skeletal muscle myosin (SkM), which is structurally similar to CM, was previously shown to exert procoagulant activities (Deguchi H et al, Blood. 2016;128:1870), leading us to undertake new studies of the in vitro and in vivo procoagulant activities of CM. First, the setting of hemophilia A with its remarkable bleeding risk was used to evaluate the procoagulant properties of CM. In studies of human hemophilia plasma and of murine acquired hemophilia A plasma, CM was added to these plasmas and tissue factor (TF)-induced thrombin generation assays were performed. Plasmas included human hemophilia A plasma and C57BL/6J mouse plasma with anti-FVIII antibody (GMA-8015; 5 microgram/mL final). CM showed strong procoagulant effects in human hemophilia A plasma, which is naturally deficient in factor VIII (<1% FVIII). The addition of only CM (12.5-200 nM) greatly increased thrombin generation in a manner comparable to addition of only recombinant FVIII. In the wild-type C57BL/6J mouse plasma, anti-FVIII antibody greatly reduced TF-induced thrombin generation, as reported. When CM (12.5-200 nM) was added to mouse plasma containing anti-FVIII antibodies, TF-induced thrombin generation was concentration-dependently restored. To study the in vivo hemostatic ability of SkM, an acquired hemophilia A mouse model was employed. Intravenous injection of anti-FVIII antibody (GMA-8015; 0.25 mg/kg) or control vehicle was given retro-orbitally to wild type C57BL/6J mice at 2 hours prior to tail cutting. The distal portion of the tail was surgically removed at 1.5 mm tail diameter to induce moderate bleeding. Tails were immersed in 50 mL of saline at 37 degrees. Total blood loss was measured as the blood volume collected during 20 min normalized for mouse weight (microL/g). Mice given only anti-FVIII antibody had more blood loss (median = 6.7 microL/g) compared to control mice (median < 2 microL/g) (Figure). In this mouse model receiving anti-FVIII antibody, CM (5.4 mg/kg) injected at 15 min prior to tail cutting significantly reduced the median blood loss from 6.7 to 2.0 and 3.2 microL/g, respectively (p < 0.001 for each myosin) (Figure). Thus, these studies provide in vivo proof of concept that both CM and SkM can reduce bleeding and are procoagulant in vivo. Second, studies of the effects of CM on thrombogenesis ex vivo using fresh human flowing blood showed that perfusion of blood over CM-coated surfaces at 300 s-1 shear rate caused extensive fibrin deposition. Addition of CM to blood also promoted the thrombotic responses of human blood flowing over collagen-coated surfaces, evidence of CM's thrombogenicity. Further studies showed that CM enhanced thrombin generation in platelet rich plasma and platelet poor plasma, indicating that CM promotes thrombin generation in plasma primarily independently of platelets. To address the mechanistic insights for CM's procoagulant activity, purified coagulation factors were employed. In a purified system composed of factor Xa, factor Va, prothrombin and calcium ions, CM greatly enhanced prothrombinase activity. Experiments using Gla-domainless factor Xa showed that the Gla domain of factor Xa was not required for CM's prothrombinase enhancement in contrast to phospholipid-enhanced prothrombinase activity which requires that Gla domain. Binding studies showed that CM directly binds factor Xa. In summary, here we show that CM is procoagulant due to its ability to bind factor Xa and strongly promote thrombin generation. In summary, CM acts as procoagulant by its ability to bind factor Xa and strongly promote thrombin generation both in vivo an in vitro. These provocative findings raise many questions about whether and how the protective pro-hemostatic properties or the pathogenic prothrombotic properties of CM contribute to pathophysiology in the coronary circulation. This discovery raises many questions about CM and coronary pathophysiology, and future CM research may enable novel translations of new knowledge regarding CM's procoagulant activities for coronary health and disease. Figure Disclosures Mosnier: The Scripps Research Institute: Membership on an entity's Board of Directors or advisory committees, Patents & Royalties. Ruggeri:MERU-VasImmune Inc.: Equity Ownership, Other: CEO and Founder.

Development of a Plasma-Based Assay to Measure the Susceptibility of Factor V to Inhibition by the C-Terminus of TFPIα

Background Factor V (FV) is proteolytically activated to FVa, which assembles with FXa in the prothrombinase complex. The C-terminus of tissue factor pathway inhibitor-α (TFPIα) inhibits both the activation and the prothrombinase activity of FV(a), but the pathophysiological relevance of this anticoagulant mechanism is unknown. FV Leiden (FVL) is less susceptible to inhibition by TFPIα, while overexpression of FV splicing variants with increased affinity for TFPIα (FV-short) causes bleeding. Objective This study aims to develop a plasma-based assay that quantifies the susceptibility of FV(a) to inhibition by the TFPIα C-terminus. Materials and Methods FV in highly diluted plasma was preactivated with FXa in the absence or presence of the TFPIα C-terminal peptide. After adding prothrombin, thrombin formation was monitored continuously with a chromogenic substrate and prothrombinase rates were obtained from parabolic fits of the absorbance tracings. TFPI resistance was expressed as the ratio of the prothrombinase rates with and without peptide (TFPIr). Results The TFPIr (0.25–0.34 in 45 healthy volunteers) was independent of FV levels. The TFPIr increased from normal individuals (0.29, 95% confidence interval [CI] 0.28–0.31) to FVL heterozygotes (0.35, 95% CI 0.34–0.37) and homozygotes (0.39, 95% CI 0.37–0.40), confirming TFPI resistance of FVL. Two individuals overexpressing FV-shortAmsterdam had markedly lower TFPIr (0.16, 0.18) than a normal relative (0.29), in line with the high affinity of FV-short for TFPIα. Conclusion We have developed and validated an assay that measures the susceptibility of plasma FV to the TFPIα C-terminus. Once automated, this assay may be used to test whether the TFPIr correlates with thrombosis or bleeding risk in population studies.

Plasma protein S residues 37–50 mediate its binding to factor Va and inhibition of blood coagulation

SummaryProtein S (PS) is an anticoagulant plasma protein whose deficiency is associated with increased risk of venous thrombosis. PS directly inhibits thrombin generation by the blood coagulation pathways by several mechanisms, including by binding coagulation factors (F) Va and Xa. To identify PS sequences that mediate inhibition of FVa activity, antibodies and synthetic peptides based on PS sequence were prepared and employed in plasma coagulation assays, purified component prothrombinase assays, binding assays, and immunoblots. In the absence of activated protein C, monoclonal antibody (Mab) S4 shortened FXa-induced clotting in normal plasma but not in PS-depleted plasma. Mab S4 also blocked PS inhibition of FVa-dependent prothrombinase activity in purified component assays in the absence or presence of phospholipids and inhibited binding of PS to immobilised FVa. Epitope mapping identified N-terminal region residues 37–67 of PS as this antibody’s epitope. A peptide representing PS residues 37–50 inhibited FVa-dependent prothrombinase activity in a noncompetitive manner, with 50% inhibition observed at 11 µM peptide, whereas a peptide with a D-amino acid sequence of 37–50 was ineffective. FVa, but not FXa, bound specifically to the immobilised peptide representing residues 37–50, and the peptide inhibited binding of FVa to immobilised PS. These data implicate PS residues 37–50 as a binding site for FVa that mediates, at least in part, the direct inhibition of FVa-dependent procoagulant activity by PS.