Search

Eosinophilic inflammation is a characteristic feature of asthma. Integrins are highly versatile cellular receptors that regulate extravasation of eosinophils from the postcapillary segment of the bronchial circulation to the airway wall and airspace. Such movement into the asthmatic lung is described as a sequential, multistep paradigm, whereby integrins on circulating eosinophils become activated, eosinophils tether in flow and roll on bronchial endothelial cells, integrins on rolling eosinophils become further activated as a result of exposure to cytokines, eosinophils arrest firmly to adhesive ligands on activated endothelium, and eosinophils transmigrate to the airway in response to chemoattractants. Eosinophils express seven integrin heterodimeric adhesion molecules: α4β1 (CD49d/29), α6β1 (CD49f/29), αMβ2 (CD11b/18), αLβ2 (CD11a/18), αXβ2 (CD11c/18), αDβ2 (CD11d/18), and α4β7 (CD49d/β7). The role of these integrins in eosinophil recruitment has been elucidated by major advances in the understanding of integrin structure, integrin function, and modulators of integrins. Such findings have been facilitated by cellular experiments of eosinophils in vitro, studies of allergic asthma in humans and animal models in vivo, and crystal structures of integrins. Here, we elaborate on how integrins cooperate to mediate eosinophil movement to the asthmatic airway. Antagonists that target integrins represent potentially promising therapies in the treatment of asthma.

Plasma fibronectin enhances platelet thrombus formation on surfaces coated with collagen. We investigated the role of fibronectin assembly in this process. Platelets adherent to fibrillar type I collagen, but not platelets adherent to von Willebrand factor (VWF), supported assembly of plasma fibronectin under static conditions. At a shear rate of 1250 s–1, platelets adherent to collagen assembled coperfused plasma fibronectin and formed larger thrombi in a fibronectin-concentration–dependent manner, with a maximum effect at 250 μg/mL. Enhanced thrombus formation on collagen was blocked by a peptide that binds to the N-terminal region of fibronectin and inhibits fibronectin assembly. Cross-linking of fibronectin to collagen prior to exposure to platelets had no effect on thrombus formation. Collagen-induced platelet thrombus formation at a shear rate of 5000 s–1required coperfusion with VWF and did not result in assembly of coperfused fibronectin. VWF-mediated increase in platelet thrombi on collagen was not enhanced and indeed was somewhat attenuated by coperfused fibronectin at a shear rate of 5000 s–1. These results indicate that, at moderately high but not very high shear rates, fibronectin assembly in platelet aggregates that form in response to collagen enhances thrombus formation and serves as an alternative to VWF-mediated enhancement.

To learn how plasma fibronectin stabilizes platelet-rich thrombi in injured mesenteric arterioles of mice, we studied the impact of plasma fibronectin on platelet thrombus formation ex vivo in a parallel flow chamber. Thrombi were greater on surfaces coated with fibrin cross-linked to fibronectin by activated factor XIII than on surfaces coated with fibrin lacking cross-linked fibronectin or with fibronectin alone. Platelet thrombi were even greater when plasma fibronectin was perfused with platelets, resulting in deposition of the perfused fibronectin in platelet thrombi. The effect of perfused fibronectin on thrombogenesis was lost if fibronectin deposition was blocked by coperfusion with the N-terminal 70-kDa fragment of fibronectin or a peptide based on the functional upstream domain of protein F1 of Streptococcus pyogenes. Increases in thrombus formation were dependent on a platelet activator such as lysophosphatidic acid, amount of fibronectin cross-linked to fibrin, and concentration of fibronectin in the perfusate. The dependency of fibronectin concentration extended into the range of fibronectin concentrations associated with increased risk of coronary artery disease. At such concentrations, the 2 mechanisms for insolubilization of plasma fibronectin—cross-linking to fibrin and assembly by adherent and aggregating platelets—synergize to result in many-fold enhancement of platelet thrombus formation.

Protein F1 is a surface protein ofStreptococcus pyogenesthat mediates high affinity binding to fibronectin (Fn) and facilitates S. pyogenesadherence and penetration into cells. The smallest portion of F1 known to retain the full binding potential of the intact protein is a stretch of 49 amino acids known as the functional upstream domain (FUD). Synthetic and recombinant versions of FUD were labeled with fluorescein isothiocyanate and used in fluorescence anisotropy experiments. These probes bound to Fn or the 70-kDa fragment of Fn with dissociation constants of 8–30 nm. Removal of the N-terminal seven residues of FUD did not cause a change in binding affinity. Further N- or C-terminal truncations resulted in complete loss of binding activity. Analysis of recombinant versions of the 70-kDa fragment that lacked one or several type I modules indicates that residues 1–7 of the 49-mer bind to type I modules I1 and I2 of the 27-kDa subfragment and the C-terminal residues bind to modules I4 and I5. Fluorescein isothiocyanate-labeled 49-mer also bound with lower affinity to large Fn fragments that lack the five type I modules of the 27-kDa fragment but contain the other seven type 1 modules of Fn. These results indicate that, although FUD has a general affinity for type I modules, high affinity binding of FUD to Fn is mediated by specific interactions with N-terminal type I modules.

F1 is an adhesin of Streptococcus pyogeneswhich binds the N-terminal 70-kDa region of fibronectin with high affinity. The fibronectin binding region of F1 is comprised of a 43-residue upstream domain and a repeat domain comprised of five tandem 37-residue sequences. We investigated the effects of these domains on the assembly of fibronectin matrix by human dermal fibroblasts, MG63 osteosarcoma cells, or fibroblasts derived from fibronectin-null stem cells. Subequimolar or equimolar concentrations of recombinant proteins containing both the upstream and repeat domains or just the repeat domain enhanced binding of fibronectin or its N-terminal 70-kDa fragment to cell layers; higher concentrations of these recombinant proteins inhibited binding. The enhanced binding did not result in greater matrix assembly and was caused by increased ligand binding to substratum. In contrast, recombinant or synthetic protein containing the 43 residues of the upstream domain and the first 6 residues from the repeat domain exhibited monophasic inhibition with an IC50of ∼10 nm. Truncation of the 49-residue sequence at its N or C terminus caused loss of inhibitory activity. The 49-residue upstream sequence blocked incorporation of both endogenous cellular fibronectin and exogenous plasma fibronectin into extracellular matrix and inhibited binding of 70-kDa fragment to fibronectin-null cells in a fibronectin-free system. Inhibition of matrix assembly by the 49-mer had no effect on cell adhesion to substratum, cell growth, formation of focal contacts, or formation of stress fibers. These results indicate that the 49-residue upstream sequence of F1 binds in an inhibitory mode to N-terminal parts of exogenous and endogenous fibronectin which are critical for fibronectin fibrillogenesis.

ABSTRACTWe have previously shown that mutation of the two tyrosines in the cytoplasmic domain of integrin subunit ß1 (Y783 and Y795) to phenylalanines markedly reduces the capability of ß1A integrins to mediate directed cell migration. In this study, ß1-dependent cell spreading was found to be delayed in GD25 cells expressing ß1AY783/795Fcompared to that in wild-type GD25-ß1A. Focal adhesion kinase (FAK) tyrosine phosphorylation and activation were severely impaired in response to ß1-dependent adhesion in GD25-ß1AY783/795Fcells compared to that in wild-type GD25-ß1A or mutants in which only a single tyrosine was altered (ß1AY783For ß1AY795F). Phosphorylation site-specific antibodies selective for FAK phosphotyrosine 397 indicated that the defect in FAK phosphorylation via ß1AY783/795Flies at the level of the initial autophosphorylation step. Indeed, ß1A-dependent tyrosine phosphorylation of tensin and paxillin was lost in the ß1AY783/795Fcells, consistent with the impairment in FAK activation. In contrast, p130CASoverall tyrosine phosphorylation was unaffected by the ß1 mutations. Despite the defect in ß1-mediated FAK activation, FAK was still localized to focal adhesions. Taken together, the phenotype of the GD25-ß1AY783/795Fcells resembles, but is distinct from, the phenotype observed in FAK-null cells. These observations argue that tyrosines 783 and 795 within the cytoplasmic tail of integrin subunit ß1A are critical mediators of FAK activation and cell spreading in GD25 cells.

The β1-null fibroblastic cell line GD25 and its derivatives were studied to gain an understanding of the roles of β1 and β3 integrins in the initial (1-hour) contraction of collagen gels. Stable transfectants of GD25 cells expressing the β1A splice variant of β1 (β1A-GD25) did not express α2β1A and did not adhere to collagen. After transfection of α2 into β1A-GD25 cells, the α2β1A-GD25 transfectants contracted collagen gels in the presence of serum, whereas β1A-GD25 cells did not. The GD25 parental cells, however, also contracted collagen gels. Collagen gel contraction by GD25 cells was blocked by antibodies to αvβ3 or a RGD-containing peptide, indicating that αvβ3 is the integrin responsible for mediation of contraction by GD25 cells. Collagen gel contraction by α2β1A-GD25 cells was not inhibited by antibodies to αvβ3 or RGD-containing peptide, but was inhibited by anti-α2 antibody. Flow cytometry demonstrated negligible expression of αvβ3 by β1A-GD25 and α2β1A-GD25 cells when compared to GD25 cells. Platelet derived growth factor (PDGF) and sphingosine-1-phosphate (S1P) enabled gel contraction by α2β1A-GD25 and GD25 cells, respectively, in the absence of serum. PDGF-stimulated contraction by α2β1A-GD25 cells was attenuated in the presence of inhibitors of phosphatidylinositol-3-kinase whereas such inhibitors had no effect on S1P-stimulated contraction by GD25 cells. These experiments using the β1-null GD25 cells and β1A and α2β1A transfectants demonstrate that α2β1A and αvβ3 independently mediate collagen gel contraction and are regulated by different serum factors and signaling pathways.

Introduction:Fibronectin (FN) has key roles in cell adhesion, differentiation, migration, alignment and immune modulation in the developing heart, but is of low abundance in the postnatal heart. Replenishing injured myocardium with insoluble FN could be a novel therapeutic approach; however, methods to produce ECM enriched with insoluble FN have remained elusive. Our team recently developed a method to produce decellularized fibronectin-rich (dFN) sheets assembled by dense human cardiac myofibroblast (CMF) cultures. We discovered that the transmembrane protein, Sushi Containing Domain 2 (SUSD2), is broadly expressed on cultured CMFs.Hypothesis:We hypothesized that SUSD2 abundance predicts whether CMFs produce intact dFN sheets.Methods:Primary CMFs were isolated from healthy left ventricles of human cadaveric donors. CMFs were sorted by median fluorescence intensity (MFI) into SUSD2-Bright and SUSD2-Dim populations using flow cytometry. Our previously reported methods to produce dFN sheets were applied to both populations, and the resulting ECM was graded: 0=no ECM, 1=fragmented, 2=friable, and 3=intact sheet (Figure).Results:SUSD2 was detected on 98±1% of CMFs among 7 strains with each evaluated in triplicate. The average MFI across all samples was 167,000 ± 51,000. CMFs that produced Grade 0-2 (n=4) and Grade 3 (n=3) dFN sheets had an average MFI of 62,000 and 270,000, respectively (P=<0.0001). SUSD2-Bright cells sorted from natural SUSD2-Dimmer strains produced intact sheets. Conversely, SUSD2-Dim cells sorted from natural SUSD2-Brighter strains failed to produce intact sheets.Conclusions:SUSD2 abundance on CMFs predicts intact dFN sheet production. Ongoing protein analyses of CMFs and their matrices across varying ages, gender and health status will clarify SUSD2’s previously unidentified role in CMFs and ECM assembly, which has important implications for tissue-engineering and cardio-therapeutic applications.

• Six coagulation proteins were measured in 79 consecutive patients referred to the coagulation service for suspected disseminated intravascular coagulation. Antithrombin III, plasminogen, and α2-plasmin inhibitor were measured with fluorescent substrate assays. Fibronectin, prothrombin, and protein C were measured with electroimmunoassays. Using history and physical findings and the results of a coagulation screen (prothrombin time, partial thromboplastin time, fibrinogen, fibrin[ogen] degradation products, platelet count, and peripheral smear), the 79 patients were classified into five categories: no disseminated intravascular coagulation (n = 21), elevated fibrin(ogen) degradation products without other evidence of coagulopathy (n = 44), defibrination syndrome (n = 9), microangiopathic thrombocytopenic purpura (n = 4), and primary fibrinolysis (n = 1). Because the sensitivity and specificity of each of the proteins could not easily be compared, receiver operating characteristic (ROC) curves and areas under the ROC curves were calculated for each of the six proteins as well as for the tests of the coagulation screen. The ROC curves indicated that, apart from plasminogen, the other coagulation proteins provided little additional information about the classification of the coagulopathy.(Arch Intern Med 1986;146:1997-2002)

The effect of removal of Ca2+bound to thrombospondin (TSP) on the protein structure in solution has been investigated using ESR spin‐label techniques. A maleimide spin label was selectively attached to the free thiol group presumably near the carboxyl‐terminal domain in which Ca2+‐binding sites are situated. The ESR spectra of spin‐labeled TSP showed that the bound label undergoes a relatively fast rotational motion with an effective rotational correlation time in the nanosecond time regimes. Removal of bound Ca2+in TSP by dialyzing spin‐labeled TSP from a Ca2+‐containing buffer into an EDTA‐containing buffer resulted in an increase in the mobility of the bound label by a factor of 2.3. The data suggest that EDTA chelation of bound Ca2+in TSP induces a conformational change of TSP at least near the site of spin labeling.

Activated protein C is a potent anticoagulant and profibrinolytic enzyme that can be derived from the vitamin-K-dependent serine protease zymogen, protein C, by the action of thrombin. Protein C antigen concentration was determined in plasmas from normals (n =40) and from 38 patients with intravascular coagulation as evidenced by positive FDP (> 10 μg/ml). Plasma protein C was 4 μg/ml in normals and was significantly depressed (<2 SD below the mean of normals) in 19 of the 38 patients. Of 15 patients with suspected intravascular coagulation but normal FDP, protein C was decreased in 5 individuals; 3 of these 5 patients had liver disease. Based on these results, we suggest that extensive activation of the coagulation system in vivo causes a significant consumption of protein C, presumably due to its activation by thrombin and subsequent clearance.

The effect of removal of Ca 2+bound to thrombospondin (TSP) on the protein structure in solution has been investigated using ESR spin-label techniques. A maleimide spin label was selectively attached to the free thiol group presumably near the carboxyl-terminal domain in which Ca 2+-binding sites are situated. The ESR spectra of spin-labeled TSP showed that the bound label undergoes a relatively fast rotational motion with an effective rotational correlation time in the nanosecond time regimes. Removal of bound Ca 2+in TSP by dialyzing spin-labeled TSP from a Ca 2+-containing buffer into an EDTA-containing buffer resulted in an increase in the mobility of the bound label by a factor of 2.3. The data suggest that EDTA chelation of bound Ca 2+in TSP induces a conformational change of TSP at least near the site of spin labeling.

We labeled surface glycoproteins of human platelets by the neuraminidase-galactose oxidase/borotritiide and the periodate/borotritiide methods. When labeled platelets were treated with 1-nM thrombin, a minor glycoprotein weighing 68,000–85,000-d was lost from the surface, and a soluble glycoprotein weighing 57,000- 68,000-d was found in the supernatant. Treatment of platelets with ADP, collagen, or the calcium ionophore A23187 did not cause loss of the 68,000–85,000-d glycoprotein from platelet surfaces or appearance of the 57,000–68,000-d glycoprotein in the supernatant. However, trace amounts of the intact 68,000–85,000-d glycoprotein were found in the supernatants of platelets that were not treated with thrombin. The numerous effects of thrombin on platelets could be initiated by cleavage and release the thrombin-sensitive glycoprotein.

A difference in recognition of the adhesive glycoprotein vitronectin (also called S-protein, serum spreading factor, and epibolin) by monoclonal antibody 8E6 (Hayman EG, et al, Proc Natl Acad Sci USA 80:4003, 1983) was investigated using a competitive enzyme-immunosorbent assay and immunoaffinity chromatography. Recognition of vitronec-tin in serum was approximately 50-fold greater than recognition of vitronectin in plasma. Recognition of vitronectin incubated with heparin, thrombin-antithrombin III complex, or heparin and thrombin-antithrombin III complex together was 2.5-, 7-, or 32-fold greater, respectively, than recognition of vitronectin alone. Thrombin or antithrombin III by itself did not induce the antigenic change. Factor Xa-antithrombin III was less effective than thrombin-anti-thrombin III in induction of the change. Dextran sulfate and fucoidan were more potent than heparin in induction of the antigenic change, whereas dermatan sulfate, hyaluronic acid, heparan sulfate, chondroitin sulfate, or keratan sulfate were less effective. Immunoblotting analysis of serum and of vitronectin incubated with thrombin and antithrom- bin III demonstrated the presence of complexes composed of vitronectin and thrombin-antithrombin III that could only be dissociated with reducing agent. N-ethylmaleimide completely blocked the formation of the presumably disulfide-bonded complexes and partially blocked the antigenic change. Both non-disulfide-bonded and disulfide-bonded vitronectin bound to antibody-Sepharose from a mixture of vitronectin and thrombin-antithrombin III. Treatment of vitronectin with 8 mol/L urea resulted in enhanced recognition by the monoclonal antibody. Thus, the 8E6 antibody reacts with an epitope that is preferentially expressed by noncovalently and covalently linked vitronectin/thrombin-antithrombin III complexes and by urea-treated vitronectin. The change in vitronectin induced by thrombin-antithrom-bin III, therefore, is a physiological correlate of urea treatment and of adsorption of vitronectin onto tissue culture plastic (as is done in cell adhesion assays). The change may be important for expression of vitronectin activity.1988 by Grune & Stratton, Inc. 0006-4971/88/7203-0003$3.00/0

Disseminated intravascular coagulation (DIC) was induced in both normal and asplenic rhesus monkeys by intravenous challenge with Streptococcus pneumoniae.Our observations in the infected monkeys have led us to conclude that (1) pneumococcal capsular polysaccharide (PCP), immune complexes and complement may not have primary roles in the initiation of DIC; (2) intact pneumococci may be catalysts for the development of DIC; (3) the initial event in DIC may be activation of Hageman factor; and (4) evidence of activation of Hageman factor‐dependent systems is present regardless of severity of infection.

Plasma fibronectin binds to gelatin and is thought to mediate the removal of blood-borne particles by the reticuloendothelial system. This removal has primarily been studied using gelatin-coated colloidal particles. Fibronectin also binds to gram-positive bacteria. We tested the possibility that fibronectin might act as an opsonin for phagocytosis of bacteria by neutrophils. When fibronectin in physiologic concentrations was present in reaction mixtures containing Staphylococcus aureusand human neutrophils, staphylococci adhered to, but were not ingested by, neutrophils. Chemiluminescence of mixtures of neutrophils and S. aureuswas enhanced by fibronectin. Although S. aureuswere not killed by attachment, the number of colony-forming units fell as the bacteria were concentrated on the neutrophil surfaces. Washing of the neutrophils did not remove attached staphylococci. Sonication, however, released bound S. aureusand returned the number of colony-forming units to control values. Demonstration of the external location of the staphylococci was accomplished using lysostaphin. S. aureus,which became associated with neutrophils in the presence of fibronectin, were completely susceptible to lysis by lysostaphin; whereas, S. aureusthat became associated with neutrophils in the presence of serum were resistant to lysis by lysostaphin. Cytochalasin-B did not interfere with S. aureusattachment to neutrophils. When Escherichia coliwas the test organism, neither attachment nor chemiluminescence was increased by the presence of fibronectin. Thus, fibronectin mediated the attachment of S. aureusto human neutrophils, but did not promote neutrophil phagocytosis of bacteria.

Vitronectin, also known as serum-spreading factor or S-protein, mediates cell adhesion and inhibits formation of the membrane-lytic complex of complement and the rapid inactivation of thrombin by antithrombin III in the presence of heparin. Vitronectin is normally present in plasma at a concentration of approximately 300 μg/mL.The investigators quantified plasma vitronectin with an enzyme-linked immunosorbent assay and visualized reduced and nonreduced vitronectin by immunoblotting after separation of plasma or serum by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The concentration of plasma vitronectin was markedly reduced in some patients with disseminated intravascular coagulation, especially in those with liver failure; it was near normal in patients with metastatic cancer and acute leukemia. Patients with vitronectin levels < 40% normal invariably had low fibrinogen and antithrombin III and a prolonged prothrombin time. In both normal and patient plasmas there was heterogeneity in the ratio of the 75,000- and 65,000-mol wt polypeptides of reduced vitronectin: 18% had mostly the 75,000-mol wt polypeptide, 59% had roughly equal amounts of the two polypeptides, and 22% had mostly the 65,000-mol wt polypeptide. This polymorphism is inherited and appears to be due to two alleles that are present with approximately equal frequency. The blotting patterns of vitronectin in reduced and nonreduced plasmas were largely unaltered in plasma of patients with defibrination syndrome, fibrinolysis, liver failure, sepsis, metastatic cancer, and acute leukemia. There was no evidence of fragmentation of vitronectin or formation of the disulfide-bonded complex of vitronectin and thrombin-antithrombin III that is found when blood is clotted. Thus these results corroborate in vitro observations that the liver is the major source of plasma vitronectin, suggest that vitronectin may become depleted during disseminated intravascular coagulation, and define a genetic polymorphism of vitronectin.

The arrays of proteins adsorbed from plasma onto a series of polystyrene copolymeric latexes were analyzed by enzyme-linked immunosorbent assay (ELISA) of washed beads and immunoblotting of proteins desorbed from the beads and separated by polyacrylamide gel electrophoresis (PAGE). Beads were prepared by continuous emulsion polymerization in the absence of surfactant. Coomassie brilliant blue staining of gel electropherograms of desorbed proteins indicated that the presence of small amounts of comonomers (1 to 10 mole%) significantly influenced the composition of the adsorbed protein layer. Immunoblotting revealed that fibrinogen, fibronectin, and vitronectin were adsorbed by all surfaces investigated. C3 and Clq adsorption varied significantly with copolymer composition. The ELISAs revealed that although the concentrations of vitronectin and fibronectin in plasma are similar, the extent of vitronectin adsorption from 70% to 85% plasma was greater by two orders of magnitude than fibronectin adsorption. Vitronectin adsorbed on carboxylic acid-containing copolymers reacted more strongly with a conformationally sensitive antivitronectin monoclonal antibody (MoAb) than vitronectin adsorbed to polystyrene and was more susceptible to cleavage by plasma protease(s). The results show that vitronectin is a major protein adsorbed from concentrated plasma and that small changes in the chemical composition of a copolymer profoundly affects the extent and nature of protein adsorption from complex mixtures such as plasma.

Vitronectin (serum spreading factor), a major serum cell adhesion molecule, was compared with S-protein, the inhibitor of the C5-9 membrane attack complex. Data from the literature indicate that S-protein and vitronectin are alpha globulins with the same aminoterminal residues, amino acid compositions, and concentrations in normal plasma (150 to 250 µg/mL).Both proteins have been reported to interact with the thrombin-antithrombin complex. The cDNA sequences of vitronectin and S-protein were recently determined and found to be almost identical. In the present studies, rabbit-anti-S-protein and a mono-clonal antibody to vitronectin both recognized 65,000- and 75,000-molecular weight (mol wt) polypeptides when plasma or serum proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to nitrocellulose paper. The 65,000 and 75,000-mol wt polypeptides bound more avidly from serum than plasma to monoclonal anti-vitronectin or heparin coupled to agarose. The presence or absence of the polypeptides constituted a major difference between the heparin-binding proteins of serum and plasma. When complement-activated serum and unactivated serum were separated by gel filtration, vitronectin coeluted with C9 in high-mol-wt fractions of activated serum but not unactivated serum. Purified S-protein was recognized by the monoclonal antibody to vitronectin and promoted spreading of human skin fibroblasts. Both vitronectin and S-protein were degraded by thrombin. On the basis of immunological and functional, as well as biochemical, properties, therefore, S-protein and vitronectin are the same.

We performed RNA-seq and high-resolution mass spectrometry on 128 blood samples from COVID-19-positive and COVID-19-negative patients with diverse disease severities and outcomes. Quantified transcripts, proteins, metabolites, and lipids were associated with clinical outcomes in a curated relational database, uniquely enabling systems analysis and cross-ome correlations to molecules and patient prognoses. We mapped 219 molecular features with high significance to COVID-19 status and severity, many of which were involved in complement activation, dysregulated lipid transport, and neutrophil activation. We identified sets of covarying molecules, e.g., protein gelsolin and metabolite citrate or plasmalogens and apolipoproteins, offering pathophysiological insights and therapeutic suggestions. The observed dysregulation of platelet function, blood coagulation, acute phase response, and endotheliopathy further illuminated the unique COVID-19 phenotype. We present a web-based tool (covid-omics.app) enabling interactive exploration of our compendium and illustrate its utility through a machine learning approach for prediction of COVID-19 severity.