key: cord-1045304-spg5r0xp
authors: Nader, Danielle; Kerrigan, Steve
title: Vascular dysregulation following SARS-CoV-2 infection involves integrin signalling through a VE-Cadherin mediated pathway
date: 2022-03-15
journal: bioRxiv
DOI: 10.1101/2022.03.15.484274
sha: 20f3cda75e880e5827732649df5c9d86a583347a
doc_id: 1045304
cord_uid: spg5r0xp

The vascular barrier is heavily injured following SARS-CoV-2 infection and contributes enormously to life-threatening complications in COVID-19. This endothelial dysfunction is associated with the phlogistic phenomenon of cytokine storms, thrombotic complications, abnormal coagulation, hypoxemia, and multiple organ failure. The mechanisms surrounding COVID-19 associated endotheliitis have been widely attributed to ACE2-mediated pathways. However, integrins have emerged as possible receptor candidates for SARS-CoV-2, and their complex intracellular signalling events are essential for maintaining endothelial homeostasis. Here, we showed that the spike protein of SARS-CoV-2 depends on its RGD motif to drive barrier dysregulation through hijacking integrin αVβ3. This triggers the redistribution and internalization of major junction protein VE-Cadherin which leads to the barrier disruption phenotype. Both extracellular and intracellular inhibitors of integrin αVβ3 prevented these effects, similarly to the RGD-cyclic peptide compound Cilengitide, which suggests that the spike protein – through its RGD motif – binds to αVβ3 and elicits vascular leakage events. These findings support integrins as an additional receptor for SARS-CoV-2, particularly as integrin engagement can elucidate many of the adverse endothelial dysfunction events that stem from COVID-19.

The interaction between the spike protein of SARS-coronavirus-2 (SARS-CoV-2) and endothelial cells 2 has been widely demonstrated to be a critical driver in vascular dysregulation observed in We were the first to describe a pattern of impaired vascular functionality following SARS-CoV-2 4 infection, and theorized that the major endothelial adherens junction protein, VE-Cadherin, was 5 involved 1 . A plethora of data now confirms this finding, where the disruption of junction proteins leads 6 to reduced endothelial barrier integrity and subsequent monolayer permeability, elucidating the vast 7 cardiovascular complications and septic shock experienced in severe COVID-19 2 . Although the 8 canonical ACE2 receptor has been implicated in driving this reaction, another potential mechanism of 9 action involves an integrin-mediated pathway. These heterodimeric transmembrane proteins are key 10 regulators of haemostasis, angiogenesis, proliferation, and inflammation. Activated through binding an 11 RGD-containing ligand, integrins can control downstream signalling transduction cascades that tether Interactions between recombinant spike proteins and integrins were performed as described previously 1 . 28

Briefly, a 96-well microplate was coated with 25 ng of integrin αVβ3 (3050-AV, R&D Systems) 29 overnight at 4°C and blocked in 5 % dry milk in 0.1 % Tween 20-PBS. Anti-αVβ3 mAB (MAB1876-30 Z, 1:100), anti-β3 mAB (sc-46655, 1:100), Cilengitide (0.0005-0.05 µM), GLPG-0187 (0.05-10 µM) 31

and Spike proteins (40591-V08H41, 40591-V08H23, 50 nM) were added and washed. After incubation 32 for 1 hour with AlexaFluor 405-labeled spike protein antibodies (FAB105403V, 1:100), absorbance 33 was measured at 405 nm. 34 4

Following spike protein engagement of integrins, reduced expression of some intercellular junction 69 proteins (JAM-A and Connexin-43) has been detected in cerebral microvascular cells, alongside 70 downregulation of the major adherens junction protein VE-Cadherin, which functions to mediate cell-71 cell adhesion 2 . Furthermore, evidence suggests SARS-CoV-2 spike protein directly induces this 72 hyperpermeability through its RGD site, as treatment using the RGD peptide compound Cilengitide 73 reduced inter-endothelial gaps and restored barrier function 1 . However, RGD-recognizing integrins are 74 known to spatio-temporally coordinate the intracellular cycling of specific RhoGTPases to control VE-75

Cadherin without stimulating its downregulation 11 . To investigate this matter, we evaluated whether the 76 endothelium could experience hyperpermeability through VE-Cadherin localisation during SARS-77

CoV-2 infection. Cell-surface or external VE-Cadherin levels were drastically reduced following viral 78 infection for 24 hours, and blocking the RGD-binding site of integrins on host endothelial cells 79 prevented this phenomenon ( Fig. 2A,B) . To measure non-surface bound VE-Cadherin, we utilised an 80 acid-wash immunofluorescence staining protocol that enables the detection of internalised proteins. The 81 amount of VE-Cadherin trafficked to intracellular compartments was significant in viral infected cells, 82

revealing that intricate VE-Cadherin dynamics are likely involved in modulating endothelial 83 permeability during COVID-19. 84

Treating the infected cells with the αVβ3 integrin antagonist Cilengitide reduced the amount of internal 86 VE-Cadherin to nearly uninfected basal levels ( Fig. 2A,B) . Our findings reveal that the spike protein 87 binding to integrin αVβ3 directly triggers the integrin-mediated VE-Cadherin pathway in endothelial 88 cells responsible for controlling vascular permeability. Moreover, the RGD site of the spike protein 89 drives this pathway. Our data additionally revealed that overall VE-Cadherin levels were consistently 90 stable in both healthy and infected vascular endothelial cells (Fig. 2C) . 91

Focal adhesion kinase (FAK) and Proto-oncogene tyrosine-protein kinase Src (Src) are non-receptor 93 tyrosine kinases that localise to the integrin β tail, and are widely implicated in coordinating integrin 94 signalling transduction in response to an external stimuli such as adhering to an RGD ligand. Since 95 these proteins regulate the RhoGTPases that control VE-Cadherin internalisation, we investigated 96 whether inhibiting these proteins associated with the integrin could also encourage similar events of 97 vascular barrier protection as the integrin antagonist. Inhibition of FAK and Src prevented endothelial 98 hyperpermeability in response to SARS-CoV-2 infection over 24 hours. Furthermore, this reduction 99 was comparable to Cilengitide (Fig. 2D) . 100 101

Clinical observations of viral endotheliitis, pulmonary thrombosis, hypoxia, oedema, and acute cardiac 103 injury in patients with severe COVID-19 is indicative of a dysfunctional endothelial barrier, which 104 establishes it as a vascular disease 12 . The relationship between SARS-CoV-2 spike protein and its host 105 receptor ACE2 has been well defined, and a dual-receptor mechanism has been proposed with another 106 cell surface receptor, integrins. In particular, integrins αVβ3 and α5β1 recognize the RGD motif 107 uniquely expressed by SARS-CoV-2, which mediates infection of epithelial and endothelial cells in 108 vitro and in vivo 1,5-8 . Subsequently, inter-endothelial junction weakening and hyperpermeability has 109 been observed, which likely elucidates the pulmonary and cardiovascular complications in COVID-110 19 1,2,7 . Inhibiting spike protein attachment hinders this response 1,7 . Therefore we sought to describe the 111 pathway correlating integrins to COVID-19 vascular dysregulation via VE-Cadherin. 112

Our work has identified the downstream signalling transduction cascade that links integrins directly to 113 the observations of vasculopathy in COVID-19. Firstly, the spike proteins of Delta and Omicron SARS-114

CoV-2 variants of concern are still highly recognized by integrin αVβ3 as they both retain the RGD 115 site. Both Cilengitide and integrin neutralising antibodies similarly blocked spike binding to integrins, 116 revealing that this interaction is likely RGD-dependent. Other integrin antagonists such as GLPG-0187 117 have successfully displayed efficacy in reducing spike protein infection when used at high 118 concentrations, confirming its involvement as a spike protein receptor 8 . However, when tested at similar 119 concentrations to Cilengitide (0.05µM), it failed to prevent attachment. This may be due to the broad-120 spectrum activity of GLPG-0187 compared to the highly specific affinity Cilengitide has towards αVβ3 121 (IC50=0.58 nm). Similarly to the α5β1 antagonist ATN-161, Cilengitide has undergone clinical trials 122

for the treatment of glioblastoma where it was greatly tolerated by patients due to its notable safety 123 profile 13 . This has critical implications for COVID-19 treatment. Neutralizing antibodies recognize the 124 ACE2 binding interface located on the spike protein surface (residues 437 -507), and therefore 125 mutations affecting receptor recognition often result in antibody evasion. Some in vivo and in vitro 126 success has been observed using antibody cocktails to reduce viral load in SARS-CoV-2, particularly 127

Etesevimab and Bamlanivimab combined therapy 14 . However, both antibodies were sensitive to 128 mutations found in circulating variants of concern B.1 .351 and B.1.617.2, and B.1.1.529 was partially  129 or completely resistant to 100% of neutralizing monoclonal antibodies 15 . The RGD (403-405) motif is 130 located within the spike receptor binding domain and is conserved across >99% of variants. As vaccine 131 and immune-induced immunity is a key contributor to viral evolution, developing a compound that 132 targets less immunodominant epitopes such as the RGD motif could be a more effective strategy against 133

Our group was the first to identify that intercellular proteins were involved in SARS-CoV-2 135 pathogenesis and this likely corresponded to the hyperpermeability observed across the endothelium in 6 COVID-19 1 . We previously identified the major adherens junction protein VE-Cadherin to be directly 137

impacted during viral infection, where it was notably missing from its expected occupancy at the cell-138 cell contacts. When endothelial cells typically undergo angiogenesis and cell migration is required, the 139 Rho GTPases Rac1 and RhoA tightly regulate stress fiber formation, whose spatially coordinated 140 activation are triggered by integrins 11 . Subsequently, VE-Cadherin can undergo translocation into 141 intracellular compartments via clathrin-mediated endocytosis, upon integrin activation and downstream 142 transduction signalling involving the major focal adhesion proteins FAK and Src 3 . We propose that 143 SARS-CoV-2 hijacks integrins via its RGD motif and controls its signalling cascade to command 144 endothelial permeability. This ensures severe hypoxia and serum leakage, circulatory collapse and 145 organ failure, which are key indicators of sepsis development. Critically, sepsis-related morbidity has 146 been significantly attributed to COVID-19 deaths in both ICU and non-ICU patients. To explicate this 147 matter, we evaluated the involvement of VE-Cadherin following spike protein infection. Although 148 previous data suggests that VE-Cadherin was downregulated to some extent alongside other gap and 149 tight junction proteins, here we showed that SARS-CoV-2 spike protein binding to integrin αVβ3 did 150

Cadherin organization by triggering its internalisation, which led to the dysfunctional barrier phenotype. 152

RhoA in infected venous endothelial cells by downregulating Rac1, which promoted permeability and 154 leakage 7 . Following integrin ligation, VE-Cadherin is known to coordinate with Rac1 to inhibit RhoA 155 to regulate cell spreading 16 (Fig. 3) . This is consistent with our own findings where we have found an 156 association between αVβ3 and cell permeability, a process tightly controlled by VE-Cadherin. 157

Additionally, several tyrosine sites across the cytoplasmic tail of VE-Cadherin undergo phosphorylation 158 via FAK and Src proteins, and elevated phosphorylation of Y658 and Y731 accounts for the majority 159 of barrier breakdowns 17 . Pharmacological inhibition of Src and its substrate FAK was effective in 160 stabilizing VE-Cadherin at cell surface due to the significantly reduced endothelial permeability during 161 SARS-CoV-2 infection. However, it has been reported that halting Src-mediated phosphorylation of 162 VE-Cadherin is not especially sufficient to fully repair the endothelial barrier, suggesting a myriad of 163 intracellular protein inhibitors would be required to suspend this occurrence 18 . Therefore, we suggest 164 that eliminating the initial contact between viral and host proteinsuch as using integrin antagonists -165 would be more effective. 

SARS-CoV-2 uses major endothelial 269 integrin alphavbeta3 to cause vascular dysregulation in-vitro during COVID-19

Degradation of Junctional Proteins That Maintain Endothelial Barrier Integrity

Vitronectin increases vascular permeability by promoting VE-cadherin 275

Mechanisms of VE-cadherin processing and degradation in microvascular 278 endothelial cells

In Vivo protection from SARS-CoV-2 infection by ATN-161 in k18-hACE2 280 transgenic mice

The Integrin Binding Peptide, ATN-161, as a Novel Therapy for 282

SARS-CoV-2 Infection

The spike protein of SARS-CoV-2 induces endothelial inflammation through 285 integrin alpha5beta1 and NF-kappaB signalling

Integrin activation is an essential component of SARS-CoV-2 infection

multicentre, randomised, open-label, phase 3 trial

Effect of Bamlanivimab as Monotherapy or in Combination With 306

Etesevimab on Viral Load in Patients With Mild to Moderate COVID-19: A Randomized 307

Considerable escape of SARS-CoV-2 Omicron to antibody neutralization

Vascular endothelial-cadherin regulates 311 cytoskeletal tension, cell spreading, and focal adhesions by stimulating RhoA

Tyrosine phosphorylation of VE-cadherin prevents 314 binding of p120-and beta-catenin and maintains the cellular mesenchymal state

Src-induced tyrosine 317 phosphorylation of VE-cadherin is not sufficient to decrease barrier function of endothelial 318 monolayers