Fig 1.
Experimental design of this study.
Human pulmonary microvascular endothelial cells (HPMVECs) were exposed to sera obtained from patients with COVID-19. Samples were from those patients who were admitted to the ICU and eventually died. A. Sera from patients with COVID-19 (COVID-19 serum) or healthy individuals (normal serum) were used to supplement Dulbecco’s modified Eagle’s Medium (DMEM). B. The in vitro HPMVEC COVID-19 model: HPMVECs were exposed to media supplemented with COVID-19 serum or normal serum. In separate experiments, HPMVECs were pre-treated with either NADPH oxidase 2 (NOX2) inhibitor PIP-2 encapsulated in liposomes (PIP-2 carrying liposomes) or blank liposomes. This was followed by exposure to COVID-19 serum or normal serum. All cells were monitored for endothelial activation (i.e., ROS production), endothelial inflammation (i.e., appearance of pro-inflammatory phenotype), and cell death.
Fig 2.
ICAM-1 expression in postmortem lung tissue of COVID-19 affected and non-COVID-19 affected individuals.
Representative images of the immunofluorescence in lung sections are displayed in four vertical panels. Sections were stained with ICAM-1 (green, first vertical panel), PECAM-1 (red, endothelial marker, second vertical panel). The third vertical panel is a merged panel (yellow) which shows the extent of ICAM-1 expression in the endothelium. In the fourth vertical panel, insets are magnified to show ICAM-1 expression (green) around individual cells (DAPI nuclear stain, blue). A. Lung section from a COVID-19 affected individual. Upper horizontal panel: Representative field showing a blood vessel at low magnification. Marged panel shows ICAM-1 colocalized along the endothelial layer (yellow). Magnified inset shows ICAM-1 expression (green) around the cells of the vessel wall (white arrow). Middle horizontal panel: Higher magnification images along a small section of a vessel wall. Yellow of the merged panel indicates ICAM-1 on the endothelial layer (white arrowhead). Lower horizontal panel: isotype control (non-immune antibody of the same type/dilution as the anti-ICAM-1 or anti-PECAM-1 used in the upper and middle panels). B. Lung section from an individual without respiratory disease. Upper and Middle horizontal panels: Lower and higher magnification images along a small section of a vessel wall show low ICAM-1 expression (green). Lower horizontal panel: isotype control (non-immune antibody of the same type/dilution as the anti-ICAM-1 or anti-PECAM-1 used in the upper-middle panels). C. Quantification of ICAM-1 expression using MetaMorph Imaging Software. Data are from postmortem lung tissue of five COVID-19-affected individuals and five non-COVID-19-affected (cardiac transplant, breast cancers, renal cancer, sarcoid diagnosed at autopsy) individuals. Four sections were stained and imaged for each sample. Within each lung section, five fields were imaged and analyzed, and the fluorescence intensity of the green signal (representing ICAM-1) within the red region was quantified. For this, the red fluorescent areas were first outlined. Next, the green fluorescence intensity within that area was measured. This integrated intensity was normalized to the area (that had red fluorescence). This was considered ICAM-1 around the endothelium. Data are shown as a box & whiskers plot. The group average and median are indicated by a plus sign and horizontal bar, respectively. Data are obtained from n = 5 each of COVID-19-affected and non-COVID-19-affected individuals. *p<0.01 as compared to non-COVID affected lungs. Group differences were evaluated using ANOVA followed by a post hoc t-test.
Fig 3.
Expression of inflammatory moieties in an in vitro HPMVEC COVID-19 model.
For all immunostaining experiments, HPMVEC were exposed to COVID-19 or normal serum. This implies media supplemented with COVID-19 serum or normal serum as described in Materials and Methods Section. HPMVEC were washed after serum exposure, fixed, and immunostained. Images were acquired using confocal microscopy. Scale bar is 25 μm. A. ICAM-1 expression in representative images of HPMVECs after a 2 h exposure to COVID-19 serum or normal serum. Isotype controls are included. Due to low fluorescent signals in the isotype staining, DAPI labeling is provided to visualize cells in the field. B. ICAM-1 expression quantification by Image J. The fluorescence intensity was normalized to the field area. C. NLRP3 subunit in representative images of HPMVECs after a 2 h exposure to COVID-19 serum or normal serum. Isotype controls are included. Due to low fluorescent signals in the isotype staining, DAPI labeling is provided to visualize cells in the field. D. NLRP3 expression quantification from intensity of fluorescent signal by Image J. The fluorescence intensity was normalized to the field area. Inset is enlarged for both A. and C. to show expression of ICAM-1 and NLRP3 in the cytoplasm/perinuclear and nuclear membranes respectively. E. P-selection expression in representative images of HPMVECs after a 2 h exposure to COVID-19 serum or normal serum. Isotype controls are included. Due to low fluorescent signals in the isotype staining, DAPI labeling is provided to visualize cells in the field. F. P-selection expression quantification by Image J. The fluorescence intensity was normalized to the field area. Data in B, D, and F. show the integrated fluorescence intensity obtained as integrated intensity in a field normalized to the field area (i.e., sum of the area occupied by cells within the field). Data are shown as a box & whiskers plot. The group average and median are indicated by a plus sign and horizontal bar, respectively. Data for these experiments were obtained from N = 4–5 individuals in each category (COVID-affected or healthy individuals). Three separate experiments were conducted for each individual’s serum sample and the average of the three trials was considered to represent the values for that sample. *p≤ 0.01 when compared to healthy (normal) serum.
Fig 4.
Endothelial cell death in the in vitro HPMVEC COVID-19 model.
A. HPMVEC were labeled with SYTOX™ Green (100 nM) and exposed to COVID-19 serum or normal serum. This implies media supplemented with COVID-19 serum or normal serum as described in Materials and Methods Section. Representative images of the same field were acquired at 1 and 30 min by confocal microscopy. Scale bar = 25 μm. B. Quantification of the fluorescent signal from SYTOX™ Green was carried out by Image J. The integrated intensity of all cells in the field was normalized to the field area (area of the cells). Data are shown as a box & whiskers plot. Data were obtained from N = 4 individuals in each category (COVID-19 affected and healthy individuals). *p<0.01 as compared to healthy serum.
Fig 5.
ROS production in the in vitro HPMVEC COVID-19 model.
A. HPMVEC were exposed for 1 h COVID-19 serum and normal serum. This implies media supplemented with COVID-19 serum or normal serum as described in Materials and Methods Section. HPMVEC were labeled with CellROX™ Green (10 μM) for 30 min and imaged for fluorescence at λex 488 nm. Scale bar is 25 μm. B. Negative and Positive Controls to observe (CellROX)ROS fluorescence patterns in ECs in response to high and low ROS: Untreated and H2O2 treated (final concentration of H2O2 for low = 50 μM; high = 500 μM). After H2O2 addition, HPMVEC were labeled with CellROX™ Green and imaged. C. Quantification of ROS from COVID-19 and normal serum by integrating fluorescent signals using Image J, with fluorescence intensity normalized to field area of the cells. Data are shown as a box & whiskers plot. Data were obtained from N = 4 of each category (COVID-19 affected and healthy individuals). Three independent experiments were carried out for each individual’s serum sample and the average of the three trials was considered to represent the values for that sample. *p≤ 0.05 when compared to healthy serum.
Fig 6.
Effect of PIP-2 (NOX2 inhibitor) on Reactive Oxygen Species (ROS) production.
ROS was measured in separate experiments using two different dyes CellROX™ Green and dihydroethidium (DHE). HPMVEC were pre-treated with either blank liposomes or PIP-2 carrying liposomes for 3 h followed by (media supplemented with) COVID-19 serum for 1 h. A. Cells were then labeled with CellROX™ Green and imaged for fluorescence. Scale bar 25 μm. “COVID-19” denotes cells pre-treated with blank liposomes followed by COVID-19 serum; “COVID-19 + PIP-2” denotes cells pre-treated with PIP-2 carrying liposomes followed by COVID-19 serum. B. Quantification of the Cell-ROX™ Green fluorescent signal by Image J. The integrated intensity in a field was normalized to the field area (i.e., area occupied by cells within the field). Data are shown as a box & whiskers plot. Data were obtained from n = 3 samples in each category (COVID-19 and COVID-19 + PIP-2). *denotes p≤ 0.05 when compared COVID-19. C. Superoxide by DHE: Cells were exposed for 1 h to (media supplemented with) COVID-19 serum or normal serum (from healthy individuals). In separate experiments, cells were pre-treated with either blank liposomes or PIP-2 carrying liposomes for 3 h followed by media supplemented with COVID-19 serum. This was followed by labeling with DHE (10 μM) which is oxidized by superoxide to form 2-hydroxyethidium (2-OH-E+) (λex 520 nm/λem 590–620 nm). Scale bar = 50 μm. “- PIP2” denotes cells pre-treated with blank liposomes followed by COVID-19 serum; “+ PIP-2” denotes cells pre-treated with PIP-2 carrying liposomes followed by COVID-19 serum. D. Quantification of the DHE fluorescent signal (red). A total of n = 5 serum samples of each cohort (COVID-19/healthy or blank liposomes/PIP-2 carrying liposomes) were assessed. For each serum sample, three independent experiments were carried out. For each experiment 3–4 fields were summarized. Data were obtained as arbitrary fluorescence units and were normalized to fold increase over -PIP2 (blank liposomes followed by COVID-19 serum). Data are presented as mean ± SD of n = 5 samples each with and without PIP-2. *denotes p≤ 0.05 when compared to COVID-19.
Fig 7.
Rac1 translocation to endothelial membrane.
A. Immunofluorescent labeling of Rac 1 (green) and membrane marker flotillin (red) in HPMVEC. HPMVEC were pre-treated with blank liposomes or PIP-2 carrying liposomes for 3 h followed by 1 h incubation with (media supplemented with) COVID-19 serum. Cells were fixed and immunostained with anti-Rac-GFP (1:100, green) and anti-flotillin (1:250). The secondary antibody was conjugated to Alexa 594 (1:200; red). DAPI (blue) was used to stain the nuclei. Rac1 is visualized as a green signal. Membranes were visualized using flotillin (red). The yellow colabel (white arrow) indicates the colocalization of Rac1 with the EC membrane. “COVID-19” denotes cells pre-treated with blank liposomes followed by COVID-19 serum; “COVID-19 + PIP-2” denotes cells pre-treated with PIP-2 carrying liposomes followed by COVID-19 serum. B. Rac translocation was analyzed by the ImageJ program. For each field, the perimeters of 3–4 cells were outlined (as shown by the dotted line). The green fluorescence intensity (Rac signal) along the cell border was measured and normalized to the cell perimeter. Three separate experiments were conducted for each sample. The average number from the three independent experiments was considered to represent the values for that sample. Data were obtained from N = 5 samples for each category (COVID-19 +PIP2 and COVID-19). Data are shown as a box & whiskers plot. The group average and median are indicated by a plus sign and horizontal bar, respectively. *p<0.05 as compared to COVID-19.
Fig 8.
PIP-2 blocks ICAM-1 expression in an in vitro model of HPMVEC COVID-19.
A. HPMVEC were treated with either blank liposomes or PIP-2 carrying liposomes for 3 h and then treated with COVID-19 serum or normal serum (from healthy individuals) for 2 hrs. Cells were then fixed, and immunostained to measure ICAM-1 expression. Scale bar = 25 μm. B. Quantification of the fluorescent signals using ImageJ. The green fluorescence intensity (representing ICAM-1 expression) was integrated across each field (area occupied by cells within the field). “-PIP-2” denotes cells pre-treated with blank liposomes followed by COVID-19 serum; “+ PIP-2” denotes cells pre-treated with PIP-2 carrying liposomes. Data are shown as a box & whiskers plot. Data were obtained from n = 3 samples in each category (COVID-19, COVID-19 + PIP-2, healthy, healthy + PIP-2). * p≤ 0.05 and #p≤ 0.001 as compared to COVID-19.
Fig 9.
The role of NOX2-ROS in regulating the NLRP3 inflammasome in COVID-19.
NLRP3 subunit and caspase-1 in HPMVEC after exposure to COVID-19 serum or normal serum (from healthy individuals), with or without PIP-2 pre-treatment. HPMVEC were treated with either blank liposomes or PIP-2 carrying liposomes for 3 h and then treated with COVID-19 serum or normal serum for 2 hrs. Cells were then fixed, and immunostained for A. NLRP3 using anti-NLRP3 B. Caspase-1 using antibody against activated caspase. Scale bar = 25 μm. C. Quantification of NLRP3 and caspase-1 was done by quantifying the intensity of the fluorescence signal (normalized to area occupied by cells) using Image J. “COVID-19” denotes cells pre-treated with blank liposomes followed by COVID-19 serum; “COVID-19 + PIP-2” denotes cells pre-treated with PIP-2 carrying liposomes followed by COVID-19 serum; “Healthy” denotes cells pre-treated with blank liposomes followed by serum from healthy individuals; “Healthy+ PIP-2” denotes cells pre-treated with PIP-2 carrying liposomes followed by serum from healthy individuals. Data are shown as a box & whiskers plot. Data was obtained from n = 3 subjects in each category (COVID-19, COVID-19 + PIP-2, healthy, healthy + PIP-2). *p≤0.05 as compared to COVID-19.
Fig 10.
NOX2-ROS in modulation of endothelial cell death in an in vitro model of COVID-19.
HPMVEC were treated with either blank liposomes or PIP-2 carrying liposomes for 3 h. Cells were labeled with SYTOX™ Green (100 nM) and exposed to media supplemented with COVID-19 serum for 30 min. A. Images of the same field were acquired after 1 and 30 min by confocal microscopy. Scale bar for all images is 25 μm. B. Quantification of the fluorescent signals was carried out by Image J and normalized to area of cells. Data are shown as a box & whiskers plot. The group average and median are indicated by a plus sign and horizontal bar, respectively. Data were obtained from N = 4 subjects for each category (COVID-19, COVID-19 + PIP-2). *p<0.05 as compared to COVID-19.
Fig 11.
Overview of endothelial inflammation in COVID-19 and the potential role of novel agent PIP-2 in endothelial protection.
A. NADPH oxidase 2 (NOX2) comprises of membrane (NOX2) and cytosolic components. Peroxiredoxin 6 (Prdx6) is known to be phosphorylated by various inflammatory stimuli (including TNF-α, which is appreciably high in COVID-19 serum). Phospho-Prdx6 (pPrdx6) translocates to the membrane and its aiPLA2 activity converts membrane phosphatidylcholine (PC) to lysophosphatidylcholine (lyso-PC). Lyso-PC is catalyzed to lysophosphatidic acid (LPA). LPA binds to its receptor on the cell membrane, and the resulting signaling cascade leads to Rac phosphorylation, which enables assembly of the cytosolic components of NOX2. The assembled enzyme reduces molecular oxygen to superoxide which then dismutates to hydrogen peroxide (H2O2) which can participate in extracellular and intracellular signaling cascades. PIP-2 inhibits the aiPLA2 activity of Prdx6 and thus blocks NOX2 activation and ROS production as reported by us earlier [26]. B. A cytokine storm is well established in the systemic circulation in COVID-19. Cytokines possibly activate the pulmonary endothelium via a NOX2 activation pathway that leads to increased expression of ICAM-1 and the NLRP3 inflammasome. ICAM-1 is known to regulate immune cell adherence while the NLRP3 pathway drives endothelial cell death (pyroptosis). Blockade of the NOX2-ROS axis by preventing NOX2 activation (that occurs via PLA2-Rac) can abrogate this pathway potentially protecting against endothelial damage and injury.