Fig 1.
SARS-CoV-2 binds to human epithelial cells and causes permeability.
(A) SARS-CoV-2 (1.08 x 105 TCID50/mL) was added to either a control surface (BSA) or human epithelial cells (Caco 2). Cells were allowed to adhere to immobilised SARS-CoV-2 and lysed with pNPP, a fluorescent substrate against alkaline phosphatase expressed within cells. The fluorescent signal emitted by pNPP correlated to the number of cells adhered and was read at 405 nm. Epithelial cells significantly interacted with SARS-CoV-2 (P<0.001; paired t-test, N = 3). (B) Caco-2 was seeded onto transwell inserts and infected with SARS-CoV-2 at MOI = 0.4. Permeability was measured using Fluorescein isothiocyanate-dextran (FITC-Dextran, 40kDa) across 0 hours, 24 hours, and 48 hours. FITC-Dextran passes through the epithelial cells into the lower chamber and is proportionate to their permeability. The extent of permeability was measured by quantifying the fluorescent levels of FITC-Dextran at 490/520 nm. Cell permeability significantly increased throughout the time course (P<0.01; P<0.0001; ANOVA f-value = 46.84, N = 3).
Fig 2.
Sequence comparison between SARS-CoV and SARS-CoV-2 highlights the novel K403R mutagenesis.
(A) Overall schematic drawing of the SARS-CoV-2 spike protein shows the N-terminal domain (NTD), RGD motif (Arg-Gly-Asp), receptor binding domain (RBD), receptor binding motif (RBM), sub-domain 1 (SD1), sub-domain 2 (SD2). The RGD motif resides within the receptor binding domain but adjacent to the ACE2-binding region (RBM). (B) Pairwise sequence alignment using EMBL-EMBOSS Needle contrasts the spike protein of SARS-CoV and SARS-CoV-2. Identical amino acids are signified by (*), dissimilar amino acids are signified by (-), RGD motif (yellow), RBM region (blue).
Fig 3.
In-silico molecular interactions of SARS-CoV-2 spike protein with integrin αVβ3.
(A) Ribbon representation of SARS-CoV-2 spike protein (orange) and receptor binding domain (cyan) (PDB ID 6M0J) with RGD motif rendered as stick (red). (B) Ribbon representation of SARS-CoV spike protein (blue) and receptor binding domain (green) (PDB ID 5XLR) with KGD residues rendered as stick (purple). (C) Ribbon representation of integrin αVβ3 (yellow) with ligand-binding pocket rendered in space-fill mode. (D) Docking the integrin αVβ3 (yellow) to receptor binding domain of SARS-CoV-2 spike protein (orange). A 90° orientation is also shown to visualise top view of the complex which highlights where the RGD motif sits (red). Integrin αVβ3 –SARS-CoV-2 interaction surface displaying contact points with RGD motif. Protein structures were constructed using PyMol.
Fig 4.
The RGD motif of SARS-CoV-2 spike protein mediates interaction to human endothelial cells.
(A) The interaction between purified viral spike protein and endothelial integrin protein was assessed through an immunofluorescence binding assay. Recombinant SARS-CoV-2 spike protein was immobilised and its binding to recombinant αVβ3 protein was measured using an anti-αVβ3 fluorescent antibody (LM609-AF488, 1:100). Binding was analysed by measuring absorbance at 450 nm. The spike protein bound to αVβ3 showed significant interaction (P<0.0001; ANOVA f-value = 99.94, N = 3). (B) An in-vitro infection model investigated the adhesion potential of SARS-CoV-2 (1.08 x 105 TCID50/mL) to endothelial cells. Sheared human aortic endothelial cells were activated with the cytokine TNFα to induce a pro-inflammatory state similar to that experienced in COVID-19 sepsis. Cilengitide was administered in 10-fold increments (0.05 μM– 0.0005 μM). A pNPP binding assay indicated % binding according to fluorescent signal measured at 405 nm. Statistical significance was found between the no drug control and each Cilengitide treated group. Since 0.0005 μM eliminated the host-viral interaction, it was chosen as optimal concentration for subsequent experiments (P = 0.001, ANOVA f-value = 16.58, N = 3).
Fig 5.
Vascular dysregulation occurs during SARS-CoV-2 infection and can be prevented using an αVβ3 antagonist, Cilengitide.
(A) Both treated and untreated endothelial cells were seeded, sheared, and activated in transwell inserts to simulate the physiological conditions of blood flow and stress experienced by the vasculature in-vivo. Following Cilengitide treatment (0.0005 μM) for one hour, cells were inoculated with SARS-CoV-2 (1.08 x 105 TCID50/mL) for 24 hours. Fluorescein isothiocyanate-dextran (FITC-Dextran, 40kDa) was added, which passed through the endothelial cell monolayer into the lower chamber, proportionate to the barrier’s permeability. The extent of permeability was measured by quantifying the fluorescent levels of FITC-Dextran at 490/520 nm. The significant rise in permeability levels indicated severe loss of barrier integrity following SARS-CoV-2 infection, but was restored after treating cells with Cilengitide (P = 0.005; ANOVA f-value = 16.27, N = 3). (B) VE-cadherin levels in uninfected, infected, and treated cells were assessed as a marker of barrier integrity and quantified based on mean fluorescence. The vasculature experienced severe vessel leakage following viral infection, due to significantly reduced expression of VE-Cadherin. When treated with Cilengitide for one hour, barrier permeability was restored back to uninfected levels (P = 0.02; ANOVA f-value = 32.63, N = 3). (C) Vascular-endothelial cadherin (VE-Cadherin) expression on sheared, activated endothelial cells was assessed using immunofluorescence microscopy. Treated cells received 0.0005 μM Cilengitide. Infected cells were inoculated with SARS-CoV-2 for 24 hours. Cell junctions were stained with anti-VE Cadherin antibody (green) and 4,6-diamidino-2-phenylindole (blue) for nuclei staining. Scale bar at 50 μm. Uninfected cells were first visualised using fluorescent microscopy at magnification 43X (top panel). Infected endothelial cells experienced loss of monolayer connections, by reduced expressed of VE-Cadherin (middle panel). Blocking the viral RGD to host αVβ3 pathway with Cilengitide re-established the barrier integrity after 24 hours (bottom panel).