Endothelial cell activation, Weibel-Palade body secretion, and enhanced angiogenesis in severe COVID-19

Background Severe COVID-19 is associated with marked endothelial cell (EC) activation that plays a key role in immunothrombosis and pulmonary microvascular occlusion. However, the biological mechanisms through which SARS-CoV-2 causes EC activation and damage remain poorly defined. Objectives We investigated EC activation in patients with acute COVID-19, and specifically focused on how proteins stored within Weibel-Palade bodies may impact key aspects of disease pathogenesis. Methods Thirty-nine patients with confirmed COVID-19 were recruited. Weibel-Palade body biomarkers (von Willebrand factor [VWF], angiopoietin-2 [Angpt-2], and osteoprotegerin) and soluble thrombomodulin (sTM) levels were determined. In addition, EC activation and angiogenesis were assessed in the presence or absence of COVID-19 plasma incubation. Results Markedly elevated plasma VWF antigen, Angpt-2, osteoprotegerin, and sTM levels were observed in patients with acute COVID-19. The increased levels of both sTM and Weibel-Palade body components (VWF, osteoprotegerin, and Angpt-2) correlated with COVID-19 severity. Incubation of COVID-19 plasma with ECs triggered enhanced VWF secretion and increased Angpt-2 expression, as well as significantly enhanced in vitro EC tube formation and angiogenesis. Conclusion We propose that acute SARS-CoV-2 infection leads to a complex and multifactorial EC activation, progressive loss of thrombomodulin, and increased Angpt-2 expression, which collectively serve to promote a local proangiogenic state.


| I N T R O D U C T I O N
SARS-CoV-2 is responsible for the COVID-19 pandemic that has already resulted in >200 million cases and >5 million deaths worldwide. Mortality is primarily due to severe bilateral pneumonitis, resulting in acute respiratory distress syndrome. Postmortem studies on COVID-19 have demonstrated the presence of fibrin-rich microthrombi disseminated throughout the pulmonary microvasculature. [1][2][3] Under normal conditions, endothelial cells (EC) function to prevent thrombosis. In patients with fatal COVID-19, evidence of significant endotheliopathy has also been observed at autopsy. [4] Its features include EC swelling, disruption of normal tight EC junctions with enhanced-barrier permeability, and enhanced EC apoptosis. [5] Significantly increased intussusceptive angiogenesis within the lungs is another hallmark of severe COVID- 19. In keeping with the postmortem findings, biomarkers of EC activation are significantly elevated in acute COVID-19. [6] For example, plasma von Willebrand factor antigen (VWF:Ag) and VWF propeptide levels are increased by 8to 10-fold in patients with COVID-19 requiring intensive care unit (ICU) admission. [7,8] VWF:Ag and VWF propeptide levels also correlate with COVID-19 severity.
[10] Acute EC activation triggers WPB exocytosis and secretion of these glycoproteins into the vessel lumen, where they impact local hemostasis, inflammation, and angiogenesis. [11] Although endotheliopathy has been implicated as playing a key role in COVID-19 pathogenesis, the biological mechanisms through which SARS-CoV-2 causes EC activation and damage are not well defined. Early electron microscopy studies reported the presence of SARS-CoV-2 within EC. [12] However, more recent studies suggest that ECs are actually resistant to direct infection. [13] In this study, we investigated EC activation in patients with acute COVID-19, and specifically focused on how proteins stored within WPB may impact key aspects of disease pathogenesis.

| Patients
Consecutive adult patients (aged >18 years) with acute COVID-19 (confirmed by a positive SARS-CoV-2 polymerase chain reaction test) admitted to St James's and Beaumont Hospitals, Dublin, between March 21 and May 6, 2020, were eligible for enrolment. Children and pregnant or breastfeeding women were excluded from the study. A control group of asymptomatic healthy controls (n = 15) was also recruited. The study was

Essentials
• Severe COVID-19 is associated with endotheliopathy that plays a key role in immunothrombosis.
• Endothelial cell activation correlates with COVID-19 severity and results in Weibel-Palade body exocytosis.
• Plasma alterations result in endothelial cell activation and Weibel-Palade body secretion and may promote local angiogenesis. approved by the local hospital research ethics committees (references REC 2020-03 and REC 17/06). Written informed consent was obtained from all the patients. For unconscious patients who were intubated and ventilated in the ICU, assent to participate in the study was obtained initially from the next of kin if patients lacked capacity and informed consent was retrospectively obtained wherever possible from participants once capacity was regained. All participants received standard dose lowmolecular-weight heparin thromboprophylaxis according to weight and renal function as previously described. [7] Patients with COVID-19 treated at the ward level were classified as "moderate," whereas patients requiring ICU admission were classified as "critical." Regarding respiratory support, moderate cases required supplemental oxygen (median FiO 2 requirement: 54%), whereas critical patients required intubation and ventilation. Patients who died from COVID-19-related complications were classified as "fatal" (Table). Blood samples were collected in 3.2% sodium citrate tubes.
Plasma VWF:Ag levels were determined by standard ELISA as before. [7] Plasma angiopoietin-2, osteoprotegerin, and soluble thrombomodulin (sTM) levels were measured using commercial ELISAs performed according to the manufacturers' instructions.
After washing (×3 with sterile phosphate-buffered saline) to remove the plasma, HUVECs were incubated in RM for 30 minutes. Supernatants were collected and secreted VWF:Ag levels were determined.
Values from control and patient plasma-treated cells were normalized and all results were presented as a percentage of secretion.

| Immunofluorescence
HUVECs were treated with control or COVID-19 plasma for 24 hours prior to fixing with 3% PFA. Cells were permeabilized, blocked, and stained with mouse anti-human VWF (kindly provided by Professor Jan Voorberg, Amsterdam, Netherlands) and rabbit anti-human angiopoietin-2. After incubation with secondary antibodies (donkey anti-mouse 488 and donkey anti-rabbit 568) and DAPI for nucleus staining, confocal microscope images (Zeiss LSM 710 NLO) were analyzed with ImageJ.

| Tube formation assay
HUVECs were trypsinized, resuspended in RM, control, or COVID-19 plasma, and then seeded onto ibidi μ-Slides Angiogenesis precoated with Matrigel. After 6 hours, images were collected and analyzed with an Angiogenesis Analyzer plugin for ImageJ.
Statistical analyses were performed using the Kruskal-Wallis test for multiple comparisons and the Spearman rank correlation in GraphPad Prism 9.0 (GraphPad Software) with a P value of <.05 considered statistically significant.
Angpt-2 is another EC glycoprotein stored in WPB reported in recent studies that binds directly to the A1 domain of VWF. [14] With respect to severe COVID-19, Angpt-2 is a ligand of the Tie-2 receptor and has been shown to destabilize ECs, with effects on EC survival, vascular stability, endothelial permeability, and angiogenesis. [15] Similar to VWF and previous reports, [16,17] plasma Angpt-2 levels were significantly increased in patients with critical and fatal COVID-19 (median levels: 5.60 ng/mL in fatal and 3.87 ng/mL in critical COVID-19 vs 1.42 ng/mL in healthy controls) ( Figure 1B).
Interestingly, in contrast to VWF, plasma Angpt-2 levels were not significantly elevated in patients with moderate COVID-19 compared to controls ( Figure 1B).
Osteoprotegerin is an EC protein that binds to VWF and thus is also recruited into WPBs. Osteoprotegerin has an established role in regulating bone remodeling by serving as a decoy receptor for RANKL. [18] Recent studies have also identified novel vascular functions for osteoprotegerin. For example, osteoprotegerin causes enhanced adhesion molecule expression on ECs and also influences EC apoptosis. [19] We observed that plasma osteoprotegerin levels were markedly elevated in patients with acute COVID-19 ( Figure 1C).
Consistent with the VWF data, osteoprotegerin levels were significantly elevated in all 3 subgroups of COVID-19 patients (median levels: 3.82 ng/mL in fatal COVID-19, 4.22 ng/mL in critical COVID-19, and 2.30 ng/mL in moderate COVID-19 respectively) ( Figure 1C). To our knowledge, this represents the first report of significantly elevated plasma osteoprotegerin levels in severe COVID-19. However, markedly increased osteoprotegerin levels have also recently been described in children with cerebral malaria, in which they correlated with disease severity. [20] Importantly, cerebral malaria is also characterized by fulminant endotheliopathy that ultimately leads to microvascular occlusion. Consistent with the hypothesis that acute COVID-19 is associated with marked EC activation and secondary WPB exocytosis, significant correlations were KARAMPINI ET AL. observed between plasma VWF:Ag, Angpt-2, and osteoprotegerin levels, respectively ( Figure 1D-F). Of note, although all 3 WPB biomarkers were markedly elevated in patients with critical or fatal COVID-19, only VWF:Ag and osteoprotegerin were significantly increased in cases with moderate COVID-19. These findings are interesting given that emerging data suggest that different subpopulations of WPB that differ in terms of their stored cargo proteins may exist within ECs. [21] In addition, heterogeneity in WPB distribution has also been described for lung microvascular EC. [22,23] Moreover, WPB secretion may differ depending upon whether it is T A B L E Demographic, clinical, and laboratory characteristics of patients with COVID-19.
To investigate whether alterations in plasma in severe COVID-19 play a role in EC activation, human ECs were treated with COVID-19or control-plasmas ex vivo. We observed a significant increase in VWF secretion when HUVECs were incubated with plasma from patients with critical or fatal COVID-19 ( Figure 2A). Furthermore, ex vivoinduced VWF secretion correlated significantly with the endogenous plasma VWF:Ag levels observed in patients with acute COVID-19 ( Figure 2B). In contrast, we found that incubation of plasma from moderate COVID-19 cases did not stimulate enhanced VWF secretion from HUVECs (Figure 2A). This is interesting because endogenous plasma VWF:Ag levels were significantly elevated in moderate COVID-19 ( Figure 1A). These data suggest that mediators in plasma play an important role in promoting EC activation in the later stages of COVID-19, but that other mechanisms may be more important in early-stage endotheliopathy. As acute COVID-19 progresses, a positive feedback loop is established wherein EC activation leads to coagulation activation and proinflammatory changes in plasma, which in turn then exacerbate endotheliopathy. Incubation with fatal COVID-19 plasma resulted in a significant reduction in secreted levels of Angpt-2 from HUVECs ( Figure 2C). To further investigate the effects of acute COVID-19 plasma on EC, immunofluorescent staining for VWF and Angpt-2 and intracellular Angpt-2 levels were assessed.

A significant increase in intracellular Angpt-2 levels was observed in
HUVECs incubated with fatal COVID-19 plasma compared to control, moderate, or critical COVID-19 plasma ( Figure 2D). This finding is consistent with previous studies that reported a significant increase in Angpt-2 transcription in HUVECs incubated with plasma from patients with severe COVID-19. [26] The majority of this Angpt-2 remained colocalized with VWF within WPBs ( Figure 2E). Together, these Angpt-2 data are interesting as angiopoietins have been reported to bind TM and inhibit its anticoagulant function. [27] Thus, in severe COVID-19, multiple mechanisms are likely to contribute to a significant reduction in generation of anticoagulant and anti-inflammatory activated protein C generation on the EC surface.
Autopsy studies have shown that acute COVID-19 is associated with a significant increase in angiogenesis. [1] In particular, intussusceptive angiogenesis and sprouting within the pulmonary vasculature were markedly elevated in COVID-19 compared to patients who had died from other types of viral infections. [1] Because VWF, Angpt-2, and osteoprotegerin have all been reported to influence angiogenesis, [15,18] we next studied whether COVID-19 plasma affected EC angiogenesis in a Matrigel-based tube formation assay. We observed a significant increase in all of the angiogenic parameters (nodes/meshes/branches per field) studied for HUVECs treated with plasma from patients with fatal COVID-19 compared to controls. In contrast, no significant changes in any of the angiogenesis parameters were observed in HUVECs treated with plasma from patients with critical COVID-19 ( Figure 3). Altogether, these findings suggest that COVID-19 plasma not only has the potential to exacerbate endotheliopathy but also to significantly promote angiogenesis.
In conclusion, we provided evidence that multifactorial EC acti-  incubation with control and fatal COVID-19 plasma. (C) Significantly increased nodes, meshes, and branches per field after HUVECs incubated with either fatal or critical COVID-19 plasma. All data shown have been normalized compared to control plasma-treated HUVECs. Comparisons between groups were assessed by one-way ANOVA test (ns, not significant: **P < .01; ***P < .001). HUVEC, human umbilical vein endothelial cell; WST, Water-soluble Tetrazolium 1.