TIP peptide (a.k.a. Solnatide, AP-301), is a 17 amino acid cyclic synthetic peptide and a direct ENaC activator with the sequence CGQRETPEGAEAKPWYC and is provided by BCN, Barcelona, Spain. NADPH oxidase 2 inhibitory peptide gp91dstat is purchased from Anaspec (Freemont, CA, USA). Recombinant SARS-CoV2 spike protein RBD is purchased from Raybiotech (Peachtree Corners, GA, USA). Recombinant full length Spike protein is purchased from Creative Diagnostics (Shirley, NY, USA). Rabbit-anti-human ACE-2 polyclonal antibody and the human ACE-2 (CT) antibody blocking peptide are from PromoKine (PromoCell, Cat#: PK-AB718-3227, Heidelberg, Germany).

Human lung microvascular endothelial cells (HL-MVEC), cell medium and endothelial cell growth supplement were purchased from Lonza (East Rutherford, NJ). Cells were cultured in humidified 5% CO₂ at 37°C using the appropriate culture medium.

Mice are housed in accordance with the National Institutes of Health (NIH) guidelines in the AAALAC-accredited experimental animal facility at Augusta University in a controlled environment. All mouse studies described in this study are approved by the Institutional Animal Care and Use Committee at Augusta University. Global humanized ACE-2 mice, in which mouse ACE2 has been globally replaced with human ACE2 (global hACE-2 knock in mice) have been generated using CRISPR-Cas9 in the Transgenic and Genome Editing Core facility at Augusta University. These mice express human ACE-2 in several cell types, including pulmonary endothelial cells. All mice are genotyped by PCR amplification of tail DNA. Every effort is made to minimize animal suffering and reduce the number of animals used.

24h post i.p. RBD injection (500 μg/kg), i.t. instillation of 10⁶ CFU of D39 S. pneumoniae (26) (Institute of Medical Microbiology, Justus-Liebig University, Giessen, Germany) is performed. After sedation of the mice with i.p. ketamine (100 mg/kg) and xylazine (20 mg/kg) solution in 0.9% normal saline, animals are placed in an intubation platform, the tongue is carefully grasped with a curved blunt-ended forceps in an upward and leftward position to gain visualization of the larynx, and a fiber-optic illuminator is positioned over the trachea to trans-illuminate the tracheal opening. Mice are intubated using a 20-gauge catheter under a Zoom stereo microscope. A volume of 20-30 μl, containing either saline (vehicle), 10⁶ CFU of D39 S. pneumoniae or 10⁶ CFU of D39 S. pneumoniae with TIP peptide (2.5 mg/kg) is instilled. The catheter is removed, and the mice are maintained in the same position on the intubating platform for at least 30 s. Then, mice are removed from the platform and placed on a heating pad for a 3-4h recovery period. Animals are returned to isolated cages in the respective animal facilities for 24 h.

Tissue collection is performed in mice deeply anesthetized by i.p. injection of a ketamine (70-80 mg/kg) and xylazine (2-4 mg/kg) cocktail in sterile PBS at 24h post S. pneumoniae instillation. Following thoracotomy, lungs are removed and used for Evans Blue Dye extraction and immunohistology. Final euthanasia is guaranteed by exsanguination and organ removal.

HL-MVEC are plated on coverslips inserted into wells of a 24-well plate. The next day, cells are treated for 2h with either vehicle (ctrl) or RBD (5 μg/ml, 2h), in the presence or absence of TIP peptide (50 μg/ml), the latter of which is applied either 30 min before or 30 min after RBD. Five min before the end of the above-mentioned treatments, CM-H2DCFDA solution (5 µM, Thermo Fisher, Cat#C6827) in serum-free media is added and cells are incubated for 15-30 min. Vehicle-treated cells without incubation with CM-H2DCFDA solution are taken as negative control (No H2DCFDA control). Coverslips with cells are washed twice with PBS, mounted on slides with DAPI-containing mounting media, and immediately imaged using a Zeiss 780 confocal microscope. Images of five randomly selected microscopic fields are captured (excitation/emission: 490 nm/525 nm) and fluorescence intensity is quantified using the Image-Pro Plus software (Media Cybernetics, Bethesda, MD).

We use cell surface biotinylation to determine if RBD (2 μg/ml) reduces surface hACE-2 expression in commercially available HL-MVEC and we investigate whether TIP peptide can prevent a RBD-induced reduction. Confluent cells are grown on permeable supports and washed three times with cold PBS buffer before adding 0.5 mg/ml sulfo-SS-biotin (Pierce) in borate buffer (85 mm NaCl, 4 mm KCl, 15 mmNa₂B₄O₇, pH 9.0) to the apical surface while the basolateral compartment was exposed to media containing 5% (v/v) fetal calf serum. The experiment is performed at 4°C with gentle agitation for 15 min, and the procedure is repeated. Labeling is stopped at time zero by adding 5% fetal calf serum in endothelial cell medium. One filter sample from each group is scraped and lysed at times 0, 30 min, 1h, 2h and 4 h and 24h. Cells are extensively washed in PBS buffer, harvested, and lysed in buffer B (PBS with 0.1% SDS, 1% Nonidet P-40, and 0.5% sodium deoxycholate) containing protease inhibitors. Cellular debris is removed by centrifugation (1200 × g, 5 min). Biotin-labeled proteins are precipitated by incubating with prewashed streptavidin coupled to magnetic beads for 18 h at 4 °C with gentle agitation. The beads are collected with a magnetic stand and removed and the supernatant saved for analysis. 300μl of Buffer B is added to the tube and gently mixed. Beads are then collected and supernatant is discarded. This washing step is repeated twice. The biotin-streptavidin complex is then lysed by boiling in buffer containing 100 mm dithiothreitol and 5% SDS. Beads are magnetically separated and supernatant containing cell surface proteins is saved. The proteins are then separated in 7.5% SDS-PAGE, transferred to nitrocellulose, and probed with specific antibodies. Lysates are kept at 4°C until the last samples (24h) are collected. All samples from 0 to 4h from one experiment are run on one gel. The 24 h samples are run separately. Western blot densities are determined with a Licor imager and ACE-2 band densities are quantified with the FIJI variant of Image-J (27). Biotinylation densities of untreated, RBD-treated, and TIP-treated cells decrease exponentially with time. Initial densities at time 0 are set to unity and densities are normalized to the amount of biotinylated ACE-2 and values are plotted on a semi-log plot to determine rate of biotin loss.

Whole cell recording. Patch electrodes are pulled on Narishige vertical puller producing electrodes which when filled with saline are approximately 10 MOhms. A giga-Ohm seal is formed on a HL-MVEC and the patch membrane is disrupted with a short pulse of current to form a whole cell patch. PClamp 10.7 is used to apply a step voltage protocol and current voltage relationships are determined. Solutions for whole-cell patch-clamping recording from HL-MVEC are in mM NaCl: 140; KCl: 5; CaCl₂:1; MgCl₂:1; HEPES: 10. pH=7.4. The pipette solution is NaCl: 5; KCl: 140; CaCl₂: 3; MgCl₂:1; EGTA: 5; HEPES: 10. pH=7.4; final Ca²⁺concentration is 100 nM.

Varying concentrations of peptide are perfused on the apical surface of the cells to a final concentration of 0, 2, 5, 10, 25, or rarely 50 μg/ml; i.e., 0, 5.2, 10.4, 26, 52, 130, or 260 nM. The holding potential is -40 mV with voltage steps from -100 to +60 in steps of 10 mV. The current at -100 mV is recorded to measure ENaC response. Amiloride (10 μM) is added at the end of the protocol to determine the component of the current due to ENaC.

Single channel cell-attached recording. Single channel events are recorded with an Axopatch 1D and then digitized at 4000 Hz with an Axon 1440 digitizer directly to disk storage. Patch pipettes with a resistance of 6–10 MΩ are fabricated from filamented borosilicate glass capillaries (TW-150F; World Precision Instruments) with a two-stage vertical puller (PP-2; Narishige, Tokyo, Japan). Cells are visualized with Hoffman modulation optics (Modulation Optics, New Haven, CT, USA) on a Nikon Diaphot. Negative pressure is applied to achieve a cell-attached patch with a seal resistance of greater than 10 GΩ after making contact between the pipette tip and the cell surface. The extracellular bath solution consists of a saline solution (150 mM NaCl, 5 mM KCl, 1 mM CaCl₂, 2 mM MgCl₂, 5 mM glucose, and 10 mM HEPES, adjusted to pH 7.4). The patch pipette solution consists of a saline solution (140 mM NaCl, 2 mM MgCl₂, and 10 mM HEPES, adjusted to pH 7.4). The cell-attached patch configuration is used for single-channel experiments, and voltages are given as the negative of the patch pipette potential. The initial patch pipette potential is +40 mV, as such driving cations inward across the membrane. ENaC activity and open and closed times within a patch are calculated using pCLAMP 10 software (version 10.7, Molecular Devices, San Jose, CA, USA). We use the product of the number of channels (N) times the single channel open probability (P_o ) as a measure of channel activity within a patch. This product is calculated without making any assumptions about the total numbers of channels in a patch or the open probability time (P_o ) of a single channel. The total number of functional channels (N) in a patch is estimated by observing the number of peaks in the current-amplitude histogram over the entire duration of the recording, after which P_o can be calculated from NP_o and N.

Lungs are flushed through the right ventricle of the heart with warm PBS to be clear of blood, fully inflated with 10% formalin via intratracheal instillation, and excised en bloc after trachea ligation. Lungs are subsequently fixed in 10% formalin overnight and included in paraffin. Lung sections are cut at 5 µm and mounted on glass slides. Sections are incubated in an oven at 80°C for 30 min, deparaffinized with xylene and cleaned in ethanol and water. Deparaffinized sections are placed overnight in antigen retrieval solution (10 mM sodium citrate buffer, pH= 6) at 65–80°C. Sections are then removed from the oven to room temperature and allowed to cool for about 20 min, before proceeding with immunostaining. We use a primary anti-fibrin mouse monoclonal antibody (clone 59D8, Cat. No. MABS2155, Sigma-Aldrich) and a commercial kit (M.O.M.® ImmPRESS® HRP (Peroxidase) Polymer Kit, MP-2400, Vector), that contains a mouse Ig blocking reagent paired with a specialized, ready-to-use, one-step M.O.M. ImmPRESS Peroxidase Polymer reagent. Detection of fibrin deposition is achieved using diaminobenzidine (DAB) substrate (ImmPACT® DAB Substrate Kit, Peroxidase (HRP), SK-4105, Vector), which produces a brown reaction product. Sections are counterstained with hematoxylin, dehydrated, cleared in xylene and mounted (Vecta Mount, H-5000, Vector) for microscopic analysis.

Human Tissue factor ELISA kit is from Abcam (Cat # Ab220653).

All data are presented as means ± SD. Data are analyzed by GraphPad InStat software (GraphPad Software Inc.). Comparisons between groups are analyzed using one or two-way analysis of variance (ANOVA) with the Tukey’s post hoc test. P < 0.05 is defined as statistically significant.

We have previously shown that lung capillary endothelial cells express all three subunits of the epithelial sodium channel (ENaC): α, β and γ²⁵). In order to investigate whether TIP peptide can actually activate amiloride-sensitive Na+ uptake, we applied voltage steps in 10 mV increments from -100 mV to +60 mV to whole cell patches on HLMVEC cells and measured the current before and after applying increasing concentrations of TIP peptide. As shown in Figure 1B , TIP peptide increases the cellular current at all potentials. We then added 10μM amiloride to block ENaC channels. Amiloride reduces the whole cell current to a level below that of untreated cells showing that there is significant amiloride-sensitive current before addition of TIP peptide in HLMVEC and that amiloride can also block all of the TIP-induced current. The reversal potential for the amiloride-sensitive currents were all very positive consistent with the currents being due to ENaC.

We have recently shown that activation of endothelial ENaC can preserve barrier function in lung capillaries in the presence of bacterial toxins (24, 25). Previously, others have shown that the spike proteins of both SARS-CoV and SARS-CoV2 can inhibit ENaC activity in Xenopus oocytes expressing the α, β and γ subunits of human ENaC (28, 29). However, whether the RBD fragment of the spike protein S1 subunit can also reduce ENaC activity in primary HL- MVEC and whether ENaC can still be activated by TIP peptide in the presence of SARS-CoV2 RBD in these cells has not been investigated. As shown in single channel patch clamp measurements in Figure 2A , basal ENaC activity in HL-MVEC is strongly reduced after a 5 min treatment with recombinant SARS-CoV2 RBD (RayBiotech, 2 μg/ml, middle panel) ( Figure 2B ) and this is significantly relieved after the subsequent addition of TIP peptide (50 μg/ml) for 6 min ( Figure 2C ). Figure 2D shows that, as anticipated from whole cell currents, TIP peptide significantly increases ENaC open probability but without a significant change in the number of channels per patch. Figure 2E show the relationship between voltage and the amplitude of single channel events. Despite changing the open probability, TIP does not change the amplitude of single channel events which implies that TIP does not alter ENaC’s conducting pore. The slight curvature is characteristic of ENaC current and due to more Na⁺ charge carriers on the outside of the cell than on the inside. A summary of the effects of RBD and TIP from 8 experiments like the one shown in Figures 2A–C is in Figure 2F . RBD significantly inhibits ENaC open probability (Po) and that this is restored by subsequent TIP peptide addition (*: p<0.05 vs. vehicle control or TIP treated group, n=9 per group). We could detect a similar inhibitory activity with full length recombinant SARS-CoV2 spike protein (Creative Diagnostics, data not shown), confirming previous findings by others (29). Data from whole cell ( Figure 1 ) and single cell patch clamp measurements ( Figure 2 ) convince us that HL-MVEC express amiloride-sensitive low conductance ENaC channels and that these channels can be inhibited by RBD and that TIP peptide can restore their activity.

Human ACE-2 is the main receptor and point of entry for SARS-CoV2 in the alveolar compartment of the lungs, as such motivating strategies to reduce its expression in alveolar epithelial cells in COVID-19 (30). However, reduced hACE-2 expression in endothelial cells, as can occur through hACE-2 ectodomain shedding (31–33), will shift the ACE2/ACE balance towards generation of barrier-disruptive angiotensin 2 (Ang 2). Increased Ang 2 binding to the angiotensin type 1 receptor (AT1R) in capillary endothelial cells can trigger a signaling cascade, culminating in the shedding of additional ACE-2 and further loss of barrier function. Circulating plasma Ang 2 levels are markedly elevated in COVID-19 patients, as compared to healthy controls and correlate with the severity of lung injury (34). Consistent with data reported by other groups (33), RBD (2 μg/ml) time-dependently reduces surface hACE-2 expression in HL-MVEC. We used surface biotinylation ( Figures 3A, C ) to estimate the rate of loss of biotinylated hACE2 from the surface membrane. Loss of hACE2 surface expression is actually 5 times faster in the presence of RBD (2 μg/ml) compared to vehicle-treated cells. As demonstrated in Figures 3B, C , co-treatment with TIP peptide (50 μg/ml) significantly preserves hACE-2 surface expression in cells treated with RBD. Figure 3D documents the specificity of the antibody signal, which is abrogated upon preincubation of the antibody with the peptide immunogen against which it is was raised (1:1 molar ratio).

RBD binding to hACE2 shifts the hACE2/hACE1 balance towards Ang 2 generation, the circulating levels of which are increased in severe COVID-19 patients (30–34). Ang 2 activates PKC, which in turn phosphorylates cytosolic subunits of NADPH oxidase 2 (NOX2) and fosters their recruitment to gp91^phox required for NOX2 activation (35). NOX2 activation, as measured by soluble NOX2-derived peptide (sNOX2-dp) was shown to be higher in COVID-19 patients versus controls and in severe versus non-severe COVID-19 (36). As shown in Figures 4A, B , HL-MVEC treated for 2h with SARS-CoV2 RBD (5 μg/ml) significantly increase their reactive oxygen species (ROS) generation, as compared to vehicle control, TIP peptide-pretreated (-30 min) or post-treated (+1h) cells and as measured by H₂DFCDA staining.

The predominant mechanisms inducing endothelial barrier impairment include, first, phosphorylation of regulatory myosin light chain (MLC) (catalyzed by either Rho kinase or Ca²⁺-dependent MLC kinase, which causes actin cytoskeleton rearrangement and formation of actin stress fibers) and, second, microtubule depolymerization, which causes disassembly of adherens junction proteins, such as vascular endothelial (VE)-cadherin (24, 37). Figure 5 demonstrates that in accordance with findings from others, RBD (5 μg/ml) treatment induces a small, but significant increase in HL-MVEC monolayer permeability within 6h, as measured in electrical cell-substrate impedance sensing (ECIS). TIP peptide (50 μg/ml), as well as the NOX2 inhibitor gp91dstat (10 μM), significantly strengthen barrier function when applied to the cells 1h after RBD. These results indicate that oxidative stress induced by RBD is crucially involved in its barrier disruptive activity and that NOX2 activation plays a prominent role in this.

COVID-19 ARDS has a significantly higher prevalence of microthrombi in pulmonary capillaries as compared to non-COVID ARDS (5–12). The procoagulant glycoprotein tissue factor (TF) is a major mediator of thrombosis, since it interacts with circulating coagulation factor VII to trigger extrinsic coagulation. Under conditions of severe infection and inflammation, as in severe COVID-19, TF can be released in the lungs not only by monocytes, macrophages and epithelial cells, but also by endothelial cells. TF pathway inhibitor (TFPI) -predominantly expressed by endothelial cells- may be consumed and degraded in these conditions, as e.g. in patients with sepsis (38, 39). TF generation is significantly increased in plasma and lungs from severe COVID-19 patients as compared to mild COVID-19 or control patients (8, 39, 40). Endothelial cells were recently proposed to be a critical player contributing to severe thrombosis in the lungs of COVID-19 patients (5). NF-kB-dependent induction of TF gene transcription in endothelial cells requires increased NOX activity (40), as such providing a link between endothelial oxidative stress and coagulopathy. Whether SARS-CoV2 RBD, which increases ROS generation in HL-MVEC, augments TF generation in human lung MVEC has not been investigated. Our data from a representative Western blotting study in Figure 6A and from densitometry from three independent experiments in Figure 6B demonstrate that RBD (5 μg/ml) increases TF protein generation within 6h in HL-MVEC and that this can be blunted by TIP peptide-mediated ENaC activation. Using a TF ELISA kit with a special extraction buffer, the difference between the RBD and the vehicle or RBD+TIP groups is found to be even more pronounced ( Figure 6C ).

Although mice are not hosts for SARS-CoV-2, expression of human ACE-2 is sufficient, in itself, to confer susceptibility to the virus in this species (41). In order to investigate whether RBD can sensitize lung endothelium for pneumococcal pneumonia-induced ARDS and coagulopathy, we have generated transgenic mice, in which mouse ACE2 has been globally replaced with human ACE2 (global hACE-2 knock in mice), using CRISPR-Cas9 ( Figures 7A, B ). These mice also express human ACE-2 in pulmonary endothelial cells. 8-10 wk old female global hACE2 knock in mice (hACE2 KI) are injected on day 0 intraperitoneally (i, p.) with RBD of SARS-CoV2 Spike protein (500 µg/kg), followed 1h later by either TIP peptide (2.5 mg/kg) or Saline i.p. On day 1, mice receive intratracheal instillation of either a low dose of Streptococcus pneumoniae (Sp., 10⁶ CFU) or saline and on day 2 animals are euthanized. In a first pool of mice, lungs are flushed intravenously with warm PBS, inflated intratracheally with 1-1.5 ml of 10% formalin and excised en bloc after trachea ligation. Lungs are fixed in formalin overnight and included in paraffin for histology evaluation and immunostaining for fibrin deposits. The second pool of mice is selected for capillary leak measurements following euthanasia, which is performed by intravenous infusion of Evans Blue dye, as described (24). As shown in Figure 8A , at 48h after the start of the study, both RBD (given i.p), and Sp. (i.t. instillation) induce a minor increase in fibrin deposits (red arrows) in lung tissue. However, the combination of both induces a significantly higher fibrin signal. TIP peptide treatment, 1h after RBD, significantly reduces lung fibrin deposition in the combined RBD/Sp. group ( Figure 8A ). Intraperitoneal injection of 500 μg/kg RBD induces a significant, but modest increase in capillary leak in 8-10 wk old female global hACE2 KI mice 48h post injection ( Figure 8B ). However, when RBD is combined with an i.t. instillation of a low dose of S. pneumoniae (10⁶ CFU, does not induce significant barrier dysfunction by itself) 24h post RBD, a profound capillary leak occurs. TIP peptide, given 1h post RBD i.p., significantly mitigates capillary leak induced by the RBD/S. pneumoniae combination. Taken together, these data indicate that RBD can sensitize hACE2 mice for indirect and direct barrier dysfunction in pneumococcal pneumonia.